r nrY fe i. . . ., / x ^ VH & tXW ^j r ~ v i. "^f- yyb*& [^ ^Ki/^- ^T^ ^ REESE LIBRARY V * j ,,K THE K^ffk^Tttfk^l TTNTVFRSITY OF CALIFC ,- *>& -stS^ " -i? - ^f - ** " L&r * r * : - (, ^ -* - a> .-- Ii>. * f :* " ^JM3S^^"S /^ "Vv-, .VL-S* " ff^^ff-r- \.-f " V^- ^SJC v V.-f* v> V^SiC-^iSX V_ -^"V* ff5i& / r<:/ ^-i *T/*_ ^^ y^i - r/ /* *^o dSKL ^ r r / yv **3rT L ^ ^^ C. ^iife& v* i?^C ^ssf^ S ^ *^ x < <p^5 ->V/ 7 i ^^fe -V>^. <r* ^i: - A-. &* ^%^.^ ^% ^ . W^k<- ""^ ^ S5*" * \j> , -,- ^ ^SS* < *" 1 \> V v < ^3jfe^r "4^ 4fe ^ ^.>. " 1 :-n . .t-J* -. V ,:n t-ff ^fy-^4^ *"T^ W^. -& -- <* ^ VJ ^ -A- J2f / (*: ^tr*^ j ! i*. 4^ ^^^-AJ^ t*. *^ & a ^ -^%^a^ ^JS?e v .*^.r ^^ &- f -^/r<^.< ^^o v ^! ( , *$&& V ^j Vix- ^?> *^4^ ap>^i*i -^ "^^* <fjk , (^ ^* . ^Q ^^ ^H-*^ ; >^o^\f?^S?J^ fc^iSi^j^ ^>^ A r jgv>, ^ i? .-i> j*^ r ^^f*5. V r^t^K^^r^ Vi** ^ Xr TwS >-xf v.^ :; .<-. > *^<P y / <v. - ., P9^^i Vo- v i?^Vo .JlK"^ T^^ fe .(!x%^^J(x%^?^Ju^ ^^^wg^Lfr w REPORT ON THE INVESTIGATIONS INTO THE PURIFICATION OF THE OHIO RIVER WATER AT LOUISVILLE KENTUCKY MADE TO THE PRESIDENT AND DIRECTORS OF THE LOUISVILLE WATER COMPANY BY GEORGE W. FULLER PUBLISHED UNDER AGREEMENT WITH THE DIRECTORS NEW YORK D. VAN NOSTRANL) COMPANY 1898 TABLE OF CONTENTS. INTRODUCTION T Location, scope, and principal dates of the investigations. Nature of systems of purification tested. Character of the War ren, Jewell, and Western systems. The Harris device. The Mark-Brownell de vices. The MacDougall polarite system. Methods and devices of the Water Company. The development of water purification ; its state at the beginning of these tests and an historical resume. Most important types of filters abroad. Principles of sand filtration without coagulants. English filters. Modi fication of English filters. English filters in America. American filters. Efficiency of American filters and their cost. Conditions under which the tests were conducted. List of chapters. CHAPTER I. COMPOSITION OF THE OHIO RIVER WATER.. 15 Character of watershed. Freshets in the Ohio River, with annual comparisons, 1861-96. Plan of analytical work. Physi cal character of Ohio River water. Chemical character of Ohio River water, with tables of daily analyses. Special chemical anal yses. Biological character of Ohio River water. Results of microscopical analyses. Species of bacteria found. Number of bac teria in Ohio River water by days. CHAPTER II. DESCRIPTION OF THE APPLICATION OF CHEM ICALS TO THE OHIO RIVER WATER BY THE SEVERAL SYSTEMS OF PURIFICATION 40 Kinds of chemicals used. Composition of chemicals used. Devices used by the re spective systems for the application of chemicals. Uniformity in rate of applica tion of the chemicals. Strengths of chemi cal solutions, their uniformity, and daily averages for the respective systems. Aver age daily amounts of sulphate of alumina used by the respective systems. Average daily amounts of lime used by the Jewell system. CHAPTER III. DECOMPOSITION AND SUBSEQUENT DISPOSAL OF THE ALUM OR SULPHATE OF ALUMINA APPLIED TO THE OHIO RIVER WATER.... 53 The general action of the chemical upon its addition to the water. Reduction of alkalinity by the applied chemical. Ab sorption of applied chemical by suspended matters. Methods of daily tests for excess of chemical. Presence of traces of unde- composed chemical in filtered water at rare intervals. Undecomposed chemical in efflu ent in practice inadmissible and inexcu sable. CHAPTER IV. COAGULATION AND SEDIMENTATION OF OHIO RIVF.R WATER BY ALUMINUM HYDRATE FORMED BY THE DECOMPOSITION OF THE APPLIED ALUM OR SULPHATE OF ALU MINA 57 General action of a coagulant. Coagula tion. Sedimentation. Devices for coagu lation and sedimentation in the respective systems. Purification by sedimentation in the several systems. Bacteria in the efflu ent of the Warren and Jewell settling-cham bers. Special sedimentation experiments. Bacteria in the Louisville water supply as drawn from city taps, with the average puri fication effected by the distributing system. TABLE OF CONTENTS. CHAPTER V. DESCRIPTION OF THE FILTERS THROUGH WHICH THE RIVER WATER PASSED AFTER COAGULATION BY ALUMINUM HYDRATE AND PARTIAL PURIFICATION BY SEDIMEN TATION 70 General account of the leading features of all the filters. The Warren filter. The Jewell filter. The Western gravity filter. The Western pressure filter. CHAPTER VI. SUMMARY OF THE VARIOUS PARTS OF THE RESPECTIVE SYSTEMS, AND A RECORD OF REPAIRS, CHANGES, AND DELAYS 89 List of the principal parts of the purifica tion station. Schedules of the devices and appurtenances employed for application of chemicals, ^sedimentation and coagulation, and filtration, respectively, in each system. Repairs and changes of the respective sys tems. Delays in operation during the tests. Summary of the time occupied in various ways during the tests. CHAPTER VII. THE MANNER OF OPERATION OF THE RE SPECTIVE SYSTEMS OF PURIFICATION, AND THE AMOUNT OF ATTENTION GIVEN THERETO . 96 General manner of operation. Detailed method of operation of each of the systems. Mechanical devices used to aid in the operation of the respective systems. Atten tion given to the respective systems. CHAPTER VIII. COMPOSITION OF THE OHIO RIVEK WATER AFTER TREATMENT BY THE RESPECTIVE S\ STEMS OF PURIFICATION, AS SHOWN BY CHEMICAL, MICROSCOPICAL, AND BACTERIAL ANALYSES ; TOGETHER WITH A TABULATION OF THE MOST IMPORTANT DATA ON THE OPERATION OF THE RESPECTIVE SYSTEMS.. . 108 Description of tables. Results of chem ical analyses of effluents of the respective systems. Results of mineral analyses of effluents. Results of bacterial analyses of effluents. Records of operation of the re spective systems by runs, with summaries of the leading analytical results for each run. CHAPTER IX. SUMMARY OF THE PRINCIPAL DATA UPON THE EFFICIENCY AND ELEMENTS OF COST OE PURIFICATION BY THE RESPECTIVE SYS TEMS, OF THE OHIO RIVER WATER, DIVIDED INTO TWENTY PERIODS, ACCORDING TO THE CHARACTER OF THE UNPURIFIED WATER ; TOGETHER WITH A DISCUSSION OF SOME OF THE MORE IMPORTANT FEATURES 215 Description of summaries. Periods into which the investigations are divided. Daily appearance of the effluents. Daily amounts ot organic matter in the river water and the percentage removal by the respective sys tems. Daily number of bacteria in the river water and effluent of each of the respective systems, together with their bacterial effici ency. Summaries of the leading results for each of the twenty periods. Total quantities for the entire investigation of 1895-96 and leading averages. Final summaries of re sults. Outline of methods followed in the discus sion. Quality of Ohio River water after purification, with reference to the efficiency of the respective systems, and the general effect of this method of purification on the character of the effluent. Prominent factors which influenced the quality of the effluent and cost of purifica tion, among which were the following : Com position of the river water ; application of alum or sulphate of alumina ; quantity of ap plied alum or sulphate of alumina; provisions for the removal of suspended matter from the river water by sedimentation ; degree of coagulation of the partially subsided water as it entered the sand layer ; sand layers of the several filters ; rate of filtration ; regula tion and control of operation of the filters ; loss of head ; washing the sand layer ; and the effect of proper attention. Comparison of the elements of cost of purification of 25 million gallons of the Ohio River water daily by the respective systems, based on the results of ten months tests of 25o,ooo-gallon systems. General conclusions in regard to the tests of 1895-96. Applicability of the American method to the clarification and purification of Ohio River water. General defect of all systems in the lack of proper provisions for subsidence and its effect on the application TABLE OF CONTENTS. of this method of purification to the Ohio River water. Comparison of the principal devices of the respective systems. Quality of the filtered Ohio River water. Final con clusions on the 1895-96 data. CHAPTER X. DESCRIPTION OF THK HARRIS MAGNETO- ELECTRIC SYSTEM OF PURIFICATION, AND A RECORD OF THE RESULTS ACCOMPLISHED THEREWITH 276 General description. Spark drum. Elec trolytic tanks. Magnets. Electrodes. Re sults accomplished by the system. CHAPTER XI. DESCRIPTION OF THE DEVICES OPERATED BY THE HARRIS COMPANY IN JULY, 1896, AND A RECORD OF THE RESULTS ACCOMPLISHED THEREWITH 280 First Experiments. Situation on July i, 1896. Description of devices Nos. i, 2, 3, 4, and 5, and records of the results accom plished therewith, respectively. Summary of results accomplished with device No. 5. Status of the situation on Aug. i, 1896. CHAPTER XII. INVESTIGATIONS BY THE WATER COMPANY IN AUGUST INTO THE PRACTICABILITY AND ECONOMY OF THE DEVICES OPERATED BY THE HARRIS COMPANY 292 The direct and indirect effect of the application of electricity upon the bacteria and organic matter in the river water, and in the purification of the water through the formation of aluminum hydrate from alu minum electrodes. Comparison of the coagulating power of aluminum hydrate formed electrolytically and from sulphate of alumina, respectively. Action of electro magnets. Rate and regularity of electrolytic formation of aluminum hydrate. Amount of metallic aluminum wasted in this electro lytic process. Relative cost of aluminum hydrate formed electrolytically and from chemicals. On the amount of power re quired on a large scale to produce aluminum hydrate by means of electricity. CHAPTER XIII. DESCRIPTION OF THE MARK AND BROWNELI. ELECTROLYTIC DEVICES, AND A RECORD OF THE RESULTS ACCOMPLISHED THEREWITH 301 An account of preliminary experiments made in the laboratory of the Louisville Manual Training School. Description of elec trolytic appliances used in connection with the Jewell filter and of the conditions under which the tests were conducted. Electric generating plant. Electrolytic cells. Iron electrodes. Outline of the operation of these devices. Summary and discussion of results obtained. Deposition of iron hydrate at the bottom of the cells and its subsequent loss. Irregularities of the flow of water through the cells. Variations in the conductivity of the river water. Electric power used by these devices. General status of this proc ess at the close of these tests. Brownell electrodes. Mark electrodes. Brief com parison of the relative efficiency of iron and aluminum electrodes. Records of oper ation and results of analyses. CHAPTER XIV. DESCRIPTION OF THE MACDOUGALL POLARITE SYSTEM, AND A RECORD OF THE RESULTS ACCOMPLISHED THEREWITH 318 Preliminary plans. General description. Detailed description of iron tank with baffle- plates, and of the clay extractor. Descrip tion of the polarite filter, and modifications of the same. Character of the filtering mate rials. Composition of polarite used. Opera tion of the system, general description and detailed records. Special method of clean ing the Jewell filter. Quality of the Ohio River water after treatment by this system. Results of analyses in detail. Applicability of this method for the purification of the Ohio River water. CHAPTER XV. DESCRIPTION OF THE METHODS AND DEVICES OF THE WATER COMPANY, TESTED DURING 1897, AND A RECORD AND DISCUSSION OF THE RESULTS ACCOMPLISHED THEREWITH . . 333 Status of the problem on March 10, 1897. Objects of the investigations of 1897. Plan of presentation of the results. General de- TABLE OF CONTENTS. scription of the devices employed. Detailed description of the settling basins; devices for the application of chemical solutions; solutions used ; and devices for the applica tion of electrolytic treatment. The arrange ments for use of the Jewell filter. General adaptability of devices employed, with their limitations. Description of the methods and conditions of operation of these devices. General notes on the records of operation. Results accomplished by these devices. Tables of analyses. Summary of the results of analyses showing the amount of suspended matter and number of bacteria in the river water as it passed through the several set tling basins. Final summary of the leading results of the operation of the devices. Discussion of the results, arranged for con venience tinder fifteen sections. Section No. i. Purification of the Ohio River water by plain sedimentation. Limited evidence available. Probable effect of 24 and 48 hours plain subsidence. Efficiency of basins used. Effect of character of sus pended matter on removal by plain subsid ence. Effect of conditions of plain subsid ence on percentage removal. General con clusions. Section No. 2. Account of the commer cial chemicals available as coagulants for the Ohio River water, and of their behavior when applied to the water. Classification of metals in their applicability to the purifica tion of Ohio River water. Most suitable compounds capable of producing coagulat ing precipitates, and a description of their behavior. Section No. 3. General description of electrolysis. Special methods and devices for electrolysis arranged by the Water Com pany. Fundamental laws and principles of electrolysis as applied to the electrolytic formation of hydrates of iron and aluminum in the Ohio River water. Section No. 4. Detailed account of the electrolytic formation of iron hydrate in Ohio River water. Passivity of iron elec trodes to the ions of Ohio River water. Cause of passivity initial and acquired. Form in which the iron leaves the electrodes. Influence on the process of oxygen, free car bonic acid, hydrogen, and solubility of the initial iron compounds. Form in which the iron leaves the electrolytic, cell. Natural limitations of the electrolytic treatment with iron electrodes. Rate of decomposi tion of iron at the positive electrodes. Rate of deposition of iron on the nega tive electrodes. Rate and uniformity of formation of available hydrate. Influ ence on the formation of hydrate of the potential difference between the plates, of the current density, of the composition of the iron, of the composition of the river water, of reversing the direction of the elec tric current, and of allowing the electrodes to remain out of service. Metal wasted in the process. Electric resistance of films of iron oxide. Power wasted in the process. Effect of this process on subsidence, filtra tion, and composition of the filtered water. Conclusions. Section No. 5. Detailed account of the electrolytic formation of aluminum hydrate in Ohio River water. Passivity of aluminum electrodes. Rate and form in which alu minum leaves the positive pole. Form and rate of deposition of aluminum on the nega tive pole. Influence of the composition of the river water on the formation of hydrate and of scales. Influence on the formation of hydrate of the presence of scale. Influ ence on the process of reversing the direc tion of the electric current. Metal wasted in the process. Influence of scale and de posit on the amount of power required. Electric power wasted in the process. Con clusions. Section No. 6. Relative efficiency of available coagulants. General review of available coagulants. Relative efficiency in connection with 24 hours subsidence, of sulphate of alumina and sulphate of iron ; persulphate of iron and electric current with iron electrodes, and sulphate of alumina and electric current with alu minum electrodes. Relative efficiency in connection with filtration, of sulphate of alumina and persulphate of iron; sulphate of alumina and electric current with iron elec trodes; and sulphate of alumina and electric current with aluminum electrodes. Conclu sions. Section No. 7. Economical application of coagulants to aid in the removal of sus pended matter by sedimentation. Relative TABLE OF CONTENTS. efficiencies in sedimentation of different amounts of coagulants. Section No. 8. Effect of the period of co agulation of the river water before filtration on the results of filtration. Section No. 9. Degree of coagulation of the water before filtration, and the minimum amount of coagulant required for that pur pose. Section No. 10. The conditions of suc cessful filtration by the American system. Amount of suspended matter in the water reaching the sand layer and the coagulation of the same. Rate of filtration. Available head negative and positive. Cleaning the sand layer. Application of caustic soda. Character of the sand layer. Section No. u. Quality of the efflu ent after proper sedimentation, coagula tion, and filtration, independent of the na ture of the coagulant. Appearance. Taste and odor. Organic matter. Mineral mat ter. Gases. Algae and other micro-or ganisms. Bacteria. Undecomposed coag ulants. Storage of effluent. Corrosion of rnetal receptacles by the effluent. Loss of the partial protective influence of the suspended matter in the river water against corrosion. Adaptability of the effluent for boiler use. Uniformity in quality of the effluent. Section No. 12. Manner in which the nature of the coagulant affected the quality of the effluent. Section No. 13. Amounts of the differ ent available coagulants which would be re quired, with optimum conditions of sub sidence and filtration, to purify satisfactorily the Ohio River water. Section No. 14. Degree to which the sev eral coagulants in their respective amounts would affect the quality of the effluent, with its practical significance, and a consideration of the advisability and cost of the removal of the added constituents. Section No. 15. Comparative costs of equivalent amounts of the available coagu lants, and an estimate of the yearly cost of treatment of the Ohio River water by each of them. CHAPTER XVI. FINAL SUMMARY AND CONCLUSIONS 438 Character of the unpurified Ohio River water. Applicability to the purification of the Ohio River water of the three methods investigated during these tests. Imperative ness of the use of coagulants. Relative adaptability of American and English types of filters. Removal of coarse matters by plain subsidence. Most suitable coagulant for use with the Ohio River water. Prepa ration and application of solutions of sul phate of alumina. Coagulation and subsid ence. Coagulation and filtration. The op timum period of coagulation. Total annual average amounts of sulphate of alumina re quired for coagulation. Filtration. Essen tial features of an American filter for the successful filtration of 25 million gallons of Ohio River water daily. Quality of the purified Ohio River water. Final conclu sions. APPENDIX. Technical description of methods used for the collection of samples and of the princi pal features in the methods of analyses. Tables for the conversion of the various unit quantities employed 445 ILLUSTRATIONS. Plate No. I. Plan of Grounds of Exper imental Station. " " II. Plan of Warren Gravity System. " " III. Section of Warren Gravity System. IV. Plan of Jewell Gravity System. " " V. Section of Jewell Gravity System. " " VI. Plan of Western Pressure and Gravity Systems. " " VII. Section of Western Pressure and Gravity Systems. " " VIII. Typical Areas of Strainer Floors. WATER PURIFICATION AT LOUISVILLE. TO THE PRESIDENT AND DIREC TORS OF THE LOUISVILLE WATER COMPANY. GENTLEMEN: Herewith is presented the full report of your representative upon the results accom plished during the recent tests by the several filters or systems in the purification of the Ohio River water, together with such de scriptions, comments, and conclusions as are deemed pertinent to the subject. The following niters or systems of water purification were investigated, named in the order in which they were installed at the pumping station of this Company, where the tests and investigations were conducted: 1. The Jewell Eilter, of the O. H. Jewell Filter Company, 73 Jackson St., Chicago, 111. 2. The Warren Filter, of the Cumberland Manufacturing Company, 220 Devonshire St., Boston, Mass. 3. The Western Gravity Filter, of the Western Filter Company, St. Louis, Mo. 4. The Western Pressure Filter, of the Western Filter Company, St. Louis, Mo. 5. The Harris Magneto-Electric System, of the John T. Harris Company, of New York City. 6. The Palmer and Brownell Water Puri fier, of Palmer and Brownell, Louisville, Ky. 7. The MacDougall Polarite System, of John MacDougall, Montreal, Canada. On Oct. I, 1895, the writer took charge of the tests and investigations, which had for their purpose the determination of the quality of the river water after purification on a prac tical scale by each of the filters or systems, and the collection and compilation of such data as would indicate the cost of construc tion and operation of these filters or systems of purification. The first three weeks were devoted chiefly to the construction and equip ment of a suitable laboratory, in which chemi cal, bacteriological, and microscopical analy ses of the water could be made after the most approved methods. From Oct. 21, 1895, to Aug. i, 1896, daily tests were made, practically without interrup tion, of the filters or systems which were then ready. On Oct. 21, 1895, tlle Jewell and Warren filters were the only ones in readi ness for operation. The two Western filters were tested beginning December 23, 1895 the date when their construction was com pleted. No tests were made of the Harris system until June 24, 1896, when it was first offered for official inspection. The greater part of the month of August, 1896, was devoted to an investigation by the Water Company into the practicability of the principles employed in certain devices op erated by the Harris Company during the preceding month. This was made necessary by the incompleteness of the evidence which had been accumulated upon this point by Aug. i, the close of the tests as originally provided for. September, October, and November, 1896, were occupied in the preparation of this re port, so far as it relates to work done up to that time. In December, 1896, special tests and inves tigations were made relating to the action of WATER PURIFICATION AT LOUISVILLE. purified water in the corrosion of boilers and pipes, and in the incrustation of steam-boilers. An examination was also made of an experi mental electrolytical device for water purifi cation, submitted for inspection to the Water Company by Profs. Palmer and Brownell in their laboratory at the Louisville Manual Training High School. From January i to March 10, 1897, at tention was given to the construction and ex amination of electrolytical devices for water purification, designed by Profs. Mark and Brownell of Louisville, and to the investiga tion of points of practical significance con nected therewith. The tests of these elec trolytical devices were the outcome of the inspection of the above-mentioned laboratory experiments made by Profs. Palmer and Brownell in December, 1896. When the tests of these electrolytical de vices as designed by Profs. Mark and Brown ell were brought to a close on March 10, 1897, it had been decided to investigate the Mac- Dougall Polarite System as soon as a test plant could be constructed. It was also ar ranged on that date that the intervening time, before the polarite system was completed, should be occupied in constructing and test ing devices designed by the officers of the Water Company. This work, which was carried on solely by the Water Company, was intended, as far as possible, to be a practical demonstration of some of the leading con clusions drawn from the foregoing tests, and to extend our knowledge along several im portant but not thoroughly understood lines, so far as time permitted. Owing to several unavoidable delays, the construction of the devices of the Water Company was not com pleted until April 10. They were then tested until May 10, when the MacDougall Polarite System was offered for official examination. This system was tested from May 10 to 19, and from May 28 to June 12, inclusive. The remainder of the time up to August I, 1897, the date of the final close of these in vestigations, was devoted to work upon the devices of the Water Company referred to above. Since August i, the time has been devoted to the preparation of this report so far as it relates to work done after January i, 1897. NATURE OF THE SYSTEMS OF PURIFICATION WHICH WERE TESTED. Before recording the results accomplished by these several methods of purification it is necessary to show in general terms how they differed from each other and from those which have been employed elsewhere. At present the custom prevails to a large extent of calling all devices for water purification by the name of filters. In a majority of cases fil tration of some kind is employed in the process of purification, but none of the de vices tested by this company at this time consisted of plain filtration, as the term is properly used. Filtration alone means simply the passage of the water taken from its source through a layer of sand or similar material. This process, which is briefly outlined in the following pages, has been successfully em ployed for many years in Europe, where the yield of filtered water per acre of filtering sur face is about 2,000,000 gallons per 24 hours. When river water, which contains much mud, clay, and other suspended matters, reaches the sand layer, the pores of the sand become clogged so that it is soon necessary to scrape off a layer of the surface sand and accumula tions which are deposited on it. This treat ment would apparently be required at frequent intervals in the filtration by this method of the Ohio River water when in its muddiest con dition, even after the water had been sub sided for several days, and a large reserve area would probably be necessary to maintain the city supply. The cost of this reserve area of filters, and of the scraping of the sand surface, would probably be great if this method were adopted here, as I understand was indicated to be the case from experiments made by your Chief Engineer, Mr. Charles Hermany, at the Crescent Hill Reservoir dur ing the summer and fall of 1884 and spring of 1885. In the systems of purification which were recently tested by this company there were tried a number of different methods that were claimed to make the cost of purification less than by sand filtration, such as has been adopted in many European cities in purifying river waters less muddy than that of the Ohio River. With one exception, in all IN TROD UCTION. of these systems of purification filtration through sand or quartz was made a very prominent portion of their respective meth ods, although the various filters were con structed and operated differently from those used in Europe. But the principal difference between the European method of filtration and all but one of those tested at Louisville lies in the fact that in these test systems the water was coagulated by chemical or electro- lytical treatment, so that a portion of the suspended matter could settle out more rap idly in basins before the water reached the sand layer; and, further, so that the water could be filtered through sand about fifty times as fast as x when no coagulation was afforded the water. In the method of filtra tion in which there was no coagulation of the water by chemical (or electrolytical) treat ment, use was made of two filters for the water to pass through in turn, and this second filter contained a layer of material called po- larite, in addition to the sand. The rate of filtration through the polarite filter was about one-half as fast as through those filters re ceiving coagulated water. THE WARREN, JEWELL, WESTERN GRAVITY, AND WESTERN PRESSURE FILTERS. In the Warren, Jewell, Western Gravity, and Western Pressure Filters the general method of procedure was identical, and sub stantially as follows: Sulphate of alumina (or alum) was added to the river water, as it entered the devices in quan tities varying with the character of the water. By combining with lime naturally dissolved in the river water the sulphate of alumina formed a white, gelatinous, solid compound, called hydrate of alumina. This latter com pound gradually coagulated the suspended matter in the river water, in a manner similar to the well-known action of white of egg when added to turbid coffee. In the settling basins, where the river water first entered, this co agulation progressed so that, as the water left the settling basins and entered the sand layer, the river water had lost some of the mud sus pended in it, and the mud and clay which it did contain were formed into flakes of suffi cient size to allow a very rapid flow of water through the sand layer, with satisfactory re sults. The claim that this method of water purification was more economical for the Ohio River water than those practised in Europe was based on the assertion that com paratively small amounts of sulphate of alu mina permitted a very great reduction in the necessary area of filtering surface. While in the general method of procedure these four systems were the same, yet they were different in the manner in which the practical details were carried out. That is to say, there were different devices for the ap plication of sulphate of alumina; the settling basins differed in size and arrangement; and in the filters themselves the sand layers were different in depth and size of grain, and were cleaned in somewhat different ways. Detailed accounts of these filters are given beyond, but these statements show the general status of the matter. It may also be added here that when these tests were begun there were no means of telling which system had the best practical devices; or, indeed, whether any of them was adapted to a satisfactory and reasonably economical purification of the Ohio River water at this point. THE HARRIS DEVICE. In the Harris Magneto-Electric System of water purification, which was tested for a short period in June, 1896, no use was made of filtration. It consisted in treating the river water directly with an electric (spark) discharge, and the subsequent passage of the water through iron tanks, in which were car bon electrodes, and on which were placed powerful magnets. The electric current was supposed to destroy the germs and the or ganic matter, while the mud, clay, and silt were to be separated out from the water by the repellent action of the magnets. In July, 1896, the Harris Company made some experiments with the application of electricity to the purification of the Ohio River water by a method in which the hy drate of alumina (formed in the case of the other filters by the decomposition of sulphate of alumina by lime, as stated above) was pre pared by the electrolytic decomposition of metallic aluminum. It was apparently the WATER PURIFICATION AT LOUISVILLE. intention, so far as appliances permitted, to coagulate the water by the same chemical compound as in the sulphate of alumina treat ment, and then proceed with subsidence and filtration in a manner similar to that em ployed in the case of the other filters, the only difference in the methods being in the manner of application of chemicals: in one case a commercial chemical product was em ployed, while in the other case the coagulat ing compound was made by the electrolytic action on the pure metal. Aside from the question of cost the electric treatment has certain advantages, which will be explained subsequently. THE MARK-BROWNKLL DEVICES. During the months of January, February, and March, 1897, electrolytical devices, de signed by Profs. Mark and Brownell, were constructed and tested. These devices were an improvement in several ways over those of the Harris Company. Their only differ ence in general method was the substitution of iron electrodes for aluminum electrodes. The electric current produced hydrate of iron, a compound similar to hydrate of alumina in its coagulating properties, and it was claimed that this would materially reduce the cost of purification. THE MACDOUGALL POLARITE SYSTEM. The results of the tests at this point indi cated the desirability of reducing the amount of the coagulating chemicals whether pro duced electrolytically or from commercial products, and, if possible, doing away with them altogether. It was claimed by Mr. MacDougall that, judging from experience in purifying the water of the river Nile and of some English streams, the Ohio River water could be economically purified, without the use of coagulating chemicals, by his polarite filter. By this method the river water was passed through a settling tank (replaced later by a coke strainer) to remove the coarsest matter, thence the water was passed at a rapid rate through a sand filter in order to remove further the particles of mud, silt, and clay. The partially clarified water was finally passed at a slower rate through a filter con taining a polarite layer with sand layers above and below it. This polarite is an iron ore which has been treated by a patent pro cess. By doing away with the use of coagu lating chemicals, and their attending cost, and at the same time securing a rate of filtra tion many times greater than in the case of plain sand filtration, the advantage of polar ite as a filtering material was claimed to be great. The polarite filter was tested from May 10 to 19, and May 28 to June 12, 1897. METHODS AND DEVICES OF THE WATER COMPANY. During the time which was required for the construction of the polarite filter, advan tage was taken of the opportunity for the Water Company to test some of their own plans which had arisen as an outcome of the foregoing tests. These methods and plans are described fully beyond, in Chapter XV, but their objects may be briefly outlined as follows: 1. A reduction in the cost of purification by a removal of the bulk of the mud, silt, and clay from the river water before it reaches the filters, thereby doing away with the ne cessity of a large reserve portion of a filter plant, to be used only at times of muddy water, with its cost of installation and opera tion. Experiments upon a small scale with the removal of the bulk of the mud by sub sidence alone were made during the early summer of 1896, and gave very promising re sults. 2. The most economical and efficient method of application of coagulating chemi cals, in connection with subsidence, to pre pare the water for filtration at a rapid rate, with reference to the best period for the co agulation of the matter suspended in the water. 3. The relative economy, advantages, and disadvantages of different coagulating chemi cals prepared in various ways. The investigations along these lines were carried to a logical end so far as was con sidered possible under the existing condi tions. IN TROD UCTION. THE STATE OF DEVELOPMENT OF WATER PURIFICATION AT THE TIME OF THESE TESTS. While much careful attention has been given to the art of water purification for more than sixty years, yet the general solution of the problem on a practical basis for large cities is far from satisfactory or complete at its present stage of development. This is due partly to varying effects of the adopted processes with different natural waters, partly to the lack of a widely practical and scientific understanding of the influence of a number of factors of the processes themselves, and partly to the great cost involved in the construc tion of adequate filtration works. With a river water of such exceedingly great varia tions in its composition as that of the Ohio River, and with proprietary systems of puri fication about which so little accurate in formation was available, these tests at Louis ville were bound to be pioneer work in a large measure. Nevertheless, valuable data were obtained. But to understand the significance of these data, and to give them their true value in the line of studies necessary to place this line of work on a satisfactory basis and capable of general application, it is essen tial to trace the development of this subject up to this time. BRIEF HISTORICAL RESUME. The filtration of public water supplies was first adopted at London, England. The date of adoption of filtration at London has been generally regarded in this country as 1839. But it is now known that a sand filter, one acre in area, was put in service in 1829, the year following the appointment of the first Royal Commission on the Quality of the Metropolitan Water Supply. This Commis sion recommended the filtration of the Thames water, and the filter referred to above was constructed by the Chelsea Water Com pany in compliance therewith. Progress in the adoption of filtration was slow until after 1849. During this year there was a severe cholera epidemic, and in August, 1849, Dr. Snow first formally announced the theory that drinking water, polluted" from those ill or dead of cholera, was the chief means of propagation of this disease. Fol lowing this the nitration of river-water sup plies advanced less slowly. After December 31, 1855, filtration of all river water supplied to the Metropolitan District of London was made compulsory by an Act of Parliament passed in July, 1852. Since this date rapid advance in the adop tion of filtration for public water supplies has been made in Europe. The population of the European cities now supplied with fil tered water aggregates from fifteen to twenty millions, or more. After the severe epidemic of cholera at Hamburg in 1892, caused largely by the polluted Elbe water, the Im perial Board of Health of Germany ordered that all public water supplies in that country drawn from rivers or lakes should be filtered. During the last thirty years there has been a marked increase in the efficiency of these systems of purification of European water supplies, owing to improvements in both the construction and the operation of the filters. The first important step in this direction was taken at London in 1871, when Parliament made provision for systematic examinations at frequent intervals of the filters and the fil tered water. The greatest progress, how ever, has been made during the past dozen years. This has been due to the establish ment of the germ theory of disease, and the general recognition by sanitarians that such diseases as typhoid fever and cholera are transmitted largely by drinking water. And, further, rapid developments in the new science of bacteriology have made it possible to apply this science in the solution of problems in water purification, so as to yield results of substantial and practical value. The recognition of the need of reliable in formation from an engineering, chemical, and bacteriological standpoint to facilitate the adoption, construction, and operation of puri fication systems has led to several important investigations. These have been made in England, Germany, and America. One of the objects of filtration, in many instances in Europe, has been to remove mud, silt, and clay from river water. In many cases, however, filtration has also been di rected to protect the water consumers from WATER PURIFICATION AT LOUISVILLE. those diseases which are carried by the water. This is a very important matter in Europe, where the population has become very dense around the great cities. In America, with its comparatively sparse population, it is not as a rule so pressing at present. But disastrous experience in Europe with some filters built in the early days of water purification show clearly that all niters should be capable at all times of protecting the health of the con sumers from water-borne diseases. GENERAL DESCRIPTION OF THE MOST IM PORTANT TYPE OF FILTERS ABROAD. Filters such as were introduced into Eng land, and which have since been employed regularly there and in many places "on the Continent, consist substantially of a large, open basin ranging as a rule from about 0.5 to 1.5 acres in area and jo feet or more in depth. In cold climates, such as in Northern Germany, they are covered, to afford protec tion from ice and frost. The bottom of the basins are made practically water-tight. On the bottom of these basins, drains and pipes are suitably arranged so as to- conduct the water from the filter to a collecting well or reservoir, located at some convenient place near the filter. In some cases, instead of using lateral pipes with perforations or open joints, the water is taken to the main tinder- drain through an arrangement of dry-laid bricks. , Over the underdrains are placed succes sively layers of broken stone and gravel, the depth of each of which varies usually accord ing to the construction of the underdrains. The size of the stone and gravel in turn be comes gradually finer toward the top, in or der that they may better serve their purpose of supporting the layer of sand which rests upon the gravel. The thickness of this layer of sand placed upon the gravel varies in dif ferent filters from about 2 to 5 feet. There is also some variation in the size ot the sand grains in the filters of the different cities. In the operation of the filters water flows or is pumped from the river or sedimentation basin onto the filter, and stands several feet in depth above the surface of the sand. The water passes downward through the sand, gravel, and broken stone, in turn, and thence through the underdrains, collecting well, or reservoir, and pumps (if such are necessary) to the consumer. The rate at which the water Hows by gravity through the filter is generally controlled and made fairly uniform by regulating devices on the outlet pipe from the filter. After a time, when a geater or less quan tity of water has passed through the filter, there appears at and near the surface of the sand an accumulation of silt and other mat ters which were suspended in the water when it reached the filter. Eventually this accumu lation becomes so great that the interstices of the sand are clogged so that an adequate quantity of water cannot pass through the filter. When this condition of affairs obtains the inlet water is shut off. The water stand ing on the filter is allowed to drain to some distance below the surface of the sand, and workmen remove with shovels and wheel barrows the upper layer of the clogged sand ordinarily to a depth of about 0.5 to 0.75 inch. The main body of the sand is cleaned only by the re moval of organic matter through the action of bacteria. The filter is slowly filled with water after the surface has been scraped, either by applying the unfiltered water at the top or by letting filtered water flow in from below. This latter procedure, where the con struction of the filter will permit it, is much the better, because it tends to prevent the formation of channels in the sand, due to the escaping air which enters the pores of the sand upon draining. Such channels are very objectionable, because they allow the water to pass through them without satisfactory pu rification. Once or twice a year the layer of sand is restored to its original thickness by either replacing the removed sand after thorough washing, or adding new clean sand. In some of the important filters of this type use is made of coagulating chemicals, but the rate of filtration is comparatively slow about 2,000,000 gallons per acre daily. In Holland chemicals for coagulation have been used to some extent. INTRODUCTION. PRINCIPLES OF SAND FILTRATION WITHOUT THE USE OF COAGULATING CHEMICALS. There has never yet been given an accurate and concise definition of the principles by which water is purified by the type of filters just described. The reason of this appears to be that there are several factors which have to be taken into consideration; and the relative practical value of these factors seems to vary under different local conditions. As we now understand the subject, the principles of puri fication by this type of filtration involve three significant phases, namely: A. Mechanical or Physical. B. Biological. C. Chemical. A. Mechanical or Physical. There are at least two important actions of a mechanical or physical nature which aid in the purifica tion of water by this type of filtration, namely : 1. A straining action, by which there are removed from the water those small sus pended particles which may be called large when compared with the size of the inter stices in the sand layer. 2. An adhesive action, by which there are removed those suspended particles, including the bacteria, which are far smaller than the interstices of the sand layer through which the water passes. This very important adhesive action is in fluenced by several varying factors and is not thoroughly understood. Its efficiency in filtration, furthermore, is associated to a con siderable degree with chemical and biological conditions, as noted below. B. Biological. This aspect is of practical significance by virtue of its action in remov ing organic matter which, in places beneath the upper surface, accumulates as films around the sand grains. The removal of or ganic matter by oxidation and nitrification appears to be a factor in causing indirectly the death of bacteria, which are mechanically arrested by the adhesive action of the sand grains. By some it has been claimed that the bacteria pass into a gelatinous form, the zoogloea stage; and, being attached to the sand grains, they facilitate thereby the re moval of bacteria in the active vegetable stage, and of minute suspended particles by means of adhesion. C. Chemical. The chemical side of filtra tion deals with the removal of dissolved or ganic matters and, together with the bacteria, with the removal of organic matters, accumu lated on the sand grains. In many cases it appears that an action, more or less chemical in its nature, between certain ingredients in the water and certain ingredients of the sand causes the formation of films, containing or ganic matter, around the sand grains. This facilitates the mechanical removal from water of bacteria by the adhesive action mentioned above; and it is also probable that this puts the organic matter in a position where the bacteria may do their work of destroying it to better advantage. In the early days of filtration the mechani cal and chemical aspects of the subject were the only ones which received attention. Since the dawn of bacteriology much attention has been given to the question as to how far the biological side aided in the accomplishment of purification by filtration. Biological theo ries advanced rapidly. By some it was claimed that the whole process was a bio logical one. These theories, however, reached a point which was untenable, and for several years the mechanical and chemical phases have been regaining more nearly their true significance. ENGLISH FILTERS. The type of filters which we have been con sidering, and which was introduced at Lon don, England, by James Simpson in 1829, is called by various names. The number has become so great that they are very confusing. The principal names given to this type of fil ters are as follows: 1. Filter beds. 2. Sand beds. 3. Sand filters. 4. Artificial sand filters. 5. Natural sand filters. 6. Slow sand filters. 7. Biological filters. 8. English filters. 9. European filters. Of these various names all have more 01 WATER PURIFICATION AT LOUISVILLE. less significance, although some of them con vey an impression which is not altogether correct. Thus, " biological filter " in the light of our present knowledge is an unfor tunate name, because it gives undue promi nence to one of several phases of the process. Quite recently " natural sand filters " has been used by many to designate this type of filters. This expression has considerable significance in that there is imitated in these filters the process in nature by which spring water and other ground waters are purified by filtration through the upper layers of the earth. The use of this name is not strictly correct in this connection, because these fil ters are actually of artificial construction, and the processes go on under conditions widely different from those in nature. Natural fil tration for public water supplies is correctly applied only to those cases where galleries or wells are located in the earth near a river or lake, where the water is naturally filtered, either from the adjoining body of water or, more frequently, from the ground on the land side, where it is naturally filtered through the earth before it reaches the place of collection. Natural filters as thus described are success fully used in France and in some places in this country where the geological conditions are favorable. The water obtained from driven wells is also similarly purified. It seems practically impossible to find a name which will specifically characterize the construction and operation of this type of ni ter, now that so many modifications in filters have been introduced. In view of this fact, it is believed that " English filters " is the best name to apply to them, and we shall use this name throughout this report. This type of filter is distinctly of English origin, and Eng lish engineers and English capital introduced it on the Continent of Europe at an early date, at Berlin, St. Petersburg, Altona, and other places. For this reason and the fact that there are several modifications in some of the Continental filters we prefer to call them English rather than European filters. MODIFICATIONS IN EUROPE OF THE ORIGI NAL ENGLISH FILTERS. In England there have been no marked changes in the construction of filters, al though some attempts have been made to re place sand with other materials, such as carbide of iron, and polarite. Improvements in the efficiency of filtration for the most part have come, however, from more careful op eration, and from extensions in the sedimen tation basins. In the latter instance there is a notable reduction in the cost of filtration of turbid and muddy river waters. On the Continent of Europe, however, a number of modifications in the original filters have been introduced. The more important ones are as follows: Tours, France. In 1856 two filters were, put in service at this place. They were de signed with the view to having the accumu lation of mud, etc., on the surface of the sand removed by forcing filtered water up through the sand from the bottom, instead of having it scraped off with shovels as in English fil ters. This idea never worked well in practice at this place, owing to insufficient pressure to force the water up through the filter. These filters were abandoned after a time, owing ap parently to a failure to provide sedimentation basins in which the sediment in the river water could subside by gravity. Holland. In several places in Holland, notably at Leeuwarden, Groningen, and Schiedam, and also at Antwerp, in Belgium, alum has been used at times to aid in the pu rification of colored and polluted water by English filters. The practical effect of the application of alum is entered into in detail in a subsequent portion of this report, and it is the purpose here only to record its use in Holland. Anderson Process. This process has been in use more or less regularly for some years at Antwerp, Belgium, and in some small towns in Europe. Quite recently a large plant has been installed at Paris, France, for the purification of water from the river Seine. Essentially this process consists of passing the water first through a revolving cylinder containing iron filings. The carbonic acid in the water dissolves some of the iron, form ing ferrous carbonate. By the air con tained in the water this salt of iron is oxidized more or less rapidly to ferric hy drate. The iron when changed into this INTRODUCTION. solid, gelatinous form combines with much of the organic matter, and, like aluminum hy drate formed from alum or sulphate of alu mina, coagulates the suspended matter, and makes it easier to filter the water subse quently through English filters. ENGLISH FILTERS IN AMERICA. Although there are ten or twelve com paratively small filters in America, more or less resembling English filters, it may be safely stated that this system of water purifi cation has never become well established in this country. Among the principal reasons of this are the following: 1. The question of cost. 2. The general absence of State or Federal Boards constituted with adequate authority to enforce the protection of citizens from pol luted water supplies, as is the case in the more thickly populated countries of Europe. 3. The absence of severe cholera epidem ics, such as have led a number of European cities to adopt filtration with haste. For a number of years sufficient informa tion has been available to show that prac tically any water may be satisfactorily puri fied by English filters, provided sufficient sedimentation is first employed in the case of very turbid or muddy waters, and that the rate of filtration is sufficiently low. With re gard to the question of expense, however, it has been, and is still, difficult to estimate even approximately the cost of construction and operation of filters which will purify a turbid or muddy water satisfactorily. The reason of this is that the various elements of cost differ widely with the local conditions, and especially with the character of the water to be purified. There are two noteworthy points to be mentioned in connection with English filters in America. In the first place this type of water purification was well described in a report by an American engineer. This gentleman, now deceased, was Mr. James P. Kirkwood, Chief Engineer of the Water Com mission of St. Louis, Mo. In December, 1865, he was instruct edby.the commissioners to pro ceed to Europe and examine into this ques tion of water filtration, with a view to apply ing this information in connection with the purification of the water supply of St. Louis. The publication of the report made a very valuable work of reference, which has been used by both American and European en gineers. The work was of such a high grade that it was translated into German in 1876. In it there are several important suggestions which have led to improvements in the con struction and operation of this type of filters. The most noteworthy of these points are that the removal of mud and silt from the water by subsidence in basins before the water reaches the filters reduces the cost, and in creases the efficiency of filtration; and, fur ther, that the efficiency of the operation is enhanced by maintaining by suitable devices a uniform flow of water through the filter. During the past six years, furthermore, the most extensive experimental investigations upon the purification of water by slow filtra tion through sand, unaided by treatment with a coagulant, have been made in America at the Lawrence Experiment Station of the State Board of Health of Massachusetts. These investigations have yielded a large fund of information on the purification of such clear but polluted waters as that of the Mer- rimac River. Another factor which has recently served to explain in part the slowness with which American cities have adopted puri fication systems for their water supplies is the fact that there has appeared in America within the last dozen years another type of water filter. This type of filter is described below. It is spoken of as the " mechanical," " alum," and " rapid sand " filter. None of these names is particularly appropriate, and in distinction from the Eng lish filters we shall refer to it as the American filter. Both types of filters unquestionably pos sess merit. But as to their relative merits for the purification of waters in general, or of any particular water, we have little or no in formation to guide us. In the absence of facts there have arisen in connection with the subject numerous statements and opinions, many of which are partisan and erroneous. This unfortunate state of affairs has recently done much to retard the adoption of muni- WATER PURIFICATION AT LOUISVILLE. cipal systems of water purification, and will probably continue to do so until reliable com parable data are available. AMERICAN FILTERS. This type of filters is the outgrowth of schemes to purify water for industrial and manufacturing purposes. Its development up to this time has been tentative to a marked degree, and has been in the hands of several competing business corporations. In 1883 it first attracted the attention of those con nected with public water supplies. At that time it consisted essentially of a large circular tank in which there was a layer of sand sup ported by a perforated bottom. Its chief characteristic, other than small size, in dis tinction from English filters, was the fact that the sand layer was cleansed of the accumu lated materials removed from the river water by forcing water under pressure up through the layer of sand. In this respect it resem bled the filters constructed in 1856 at Tours, in France. Patents were taken out in 1884 to cover a modification which consisted of the applica tion of alum, a salt of iron, or other similar coagulating chemical, to the water, just be fore it passed through the layer of sand. The custom of applying alum to coagulate water, in order to facilitate the removal of foreign matter, has been practised in various ways for many centuries in different parts of the world, and the description of it in scientific literature began about seventy years ago. The apparent object of the application of chemicals under the stated conditions are un derstood to be a reduction in the cost of treatment, by doing away with subsidence basins, and by diminution of the area of filter ing surface. This type of filters was first employed in the treatment of a public water supply at Somerville, N. J., in 1885. Since that time many towns and small cities have adopted systems of this general type. At present it is said that over 100 town and municipal plants are in operation, but among this number there are none for large cities. In the last ten years many modifications have been introduced by the several compet ing companies. These modifications, more or less protected by patents, relate for the most part to devices for supporting the sand layer at the bottom; the introduction of filtered water under pressure below the sand layer, to enable the filter to be cleaned by a reverse flow of water; and of agitating devices to stir the sand during washing, and thus aid the cleansing process. In the present filters of the several companies the coagulating chemi cals are applied at points differently located with reference to the sand layer, and with varying provisions to secure not only more complete coagulation, but also to effect a re moval of some suspended matter before the water is filtered. To this general account of the American filters it may be added that a majority of them are gravity filters where the water tlows by gravity through a sand layer placed in an open tank. In some cases, however, pressure filters are used. The pres sure filters, in addition to customary devices, consist of a sand layer placed in a closed compartment, so that the water can be forced through the filter under pressure, thereby avoiding, it is claimed, additional pumping under some conditions. Compared with the English filters, the American filters at present show the follow ing principal differences: T. The American filters are aided by the application to the water of a coagulating chemical, which makes it possible to filter through sand at a much more rapid rate, and thereby the necessary area of filter is much reduced. 2. The American filters are cleaned In- passing a stream of water upward through the sand, with or without accompanying agi tation, rather than by scraping off the surface layers, as in the case of the English filters. There are of course many other features of difference, such, for example, as the strain ers at the bottom, to hold back the sand, and at the same time furnish an exit for the fil tered water; but the two points stated above are the principal differences. EFFICIENCY OF AMERICAN FILTERS, AND COST OF THEIR OPERATION. At the beginning of the Louisville tests INTRODUCTION. there were no available data which would show whether or not the American type of filter was capable of purifying the Ohio River water; or which of the several companies had the best filter for sale; or whether any of the American filters were capable of purifying; the Ohio River water at a reasonable cost. It is true that some scattering data indicated a satisfactory purification of certain waters by this type of filters, but there was other information pointing to work of an inferior grade. With regard to the question of cost, practically nothing was available which would be of any service in considering the purification of such an exceedingly variable water as that of the Ohio River. On the one hand, it was claimed that somewhat similar muddy waters were purified at a comparatively low cost by this type of filter; while, on the other hand, it was known that a system of purification installed at New Orleans by one of the prominent American filter-makers had for some reason been a failure. What the exact facts and con ditions of purification were at the several places where this type of filter had been tried could not be learned. In fact there is reason to believe that they were not accurately known. Very early in these tests the results of some tests of an American filter made at Provi dence, R. I., were available. The Providence work was of much value in indicating that it was possible with some waters and some con ditions to accomplish a satisfactory purifica tion by this type of filter. But the Pawtuxet River water, so far as can be learned from the limited analytical evidence as to its character, is very much easier to purify than the Ohio River water. And it may be safely stated that a thoroughly satisfactory solution of the prob lem of purifying the Pawtuxet water could not by any means serve as an adequate guide for the purification of the Ohio water. An attempt was also made to learn the relative advantages of the English and American types of filters in purifying the local water, but the conditions were such that in this re spect the work at Providence led to no de cisive conclusions of value. With regard to the Harm Magneto-Elec tric System it was said that an experimental device at Brooklyn had been successful in pu rifying the local water, but no accurate idea could be obtained as to the cost of treatment. The Mark and Brownellelectrolytical device, in which the current of electricity was applied to the water through iron electrodes, was on the same principle as the Webster Process for sewage purification. Eight or nine years ago the Webster Process was claimed in England to be very promising, but for the past few years little or nothing had been heard about it. As already stated, this electrolytical de vice replaces the application of chemicals, but it was used in connection with American fil ters. This portion of the test, therefore, re fers to the coagulation preceding filtration. The MacDougall Polarite System had never been tried in America, but fragmentary ac counts of its trial in England and Egypt in dicated that it probably had some advantages. Such were the conditions found by the Louisville Water Company when they made these tests, with a view to finding a practi cable method of purifying the Ohio River water as delivered to the citizens of Louis ville. CONDITIONS UNDER WHICH THE TESTS WERE CONDUCTED. The investigations and tests described in the following portion of this report were all conducted at the pumping station of the Water Company, about three miles above the city of Louisville on the Kentucky shore of the Ohio River. A plan of the ground at the pumping station is presented on Plate I. The Water Company constructed six tem porary buildings, four of which were occupied by the companies which offered purification systems for examination. One of them was equipped as the laboratory of the Water Company, and under the direction of the writer was furnished with all apparatus and supplies necessary for analytical work in tin s line after the best modern methods. The remaining building contained a pump with which filtered water under pressure was sup plied for washing the filters. Steam, and Ohio River water taken from the force main at a point about 390 feet from the intake, were supplied by the Water Company to these buildings. All the piping leading to and from WATER PURIFICATION AT LOUISVILLE. the buildings, the meters on the water-pipes, and sewer connections were also furnished by the Water Company. During the period preceding Aug. i, 1896, the four buildings above mentioned were oc cupied by the Warren Filter, the Jewell Fil ter, the two Western Filters, and the Harris Magneto-Electric System of purification, re spectively, named in the order of their loca tion, beginning at the laboratory. This order was used in the current note-books, and as a matter of convenience the several systems of purification will be referred to in this report in the above order. In view of the fact that filtration alone was not the only method em ployed in the purification processes, the term system of purification will be used in this re port in speaking of the entire experimental devices of each company, rather than the term filter alone. By the terms of the contracts be tween the Water Company and the other companies the latter installed their respective systems, each having a capacity of 250,000 gallons per twenty-four hours, in separate temporary houses at their own expense. The several companies, further, managed and operated their systems without any expense, risk, or responsibility to the Water Company. Authority, however, was reserved by the Water Company to make such rules and regulations as it deemed advisable in con ducting these tests on a fair competitive basis, and to allow its representatives unrestricted access to the several systems at all times, in order that such information could be obtained as was deemed necessary in the premises. Mr. George A. Soper was engineer in charge of the Warren Filter. Messrs. Will iam M. Jewell and Ira H. Jewell, officers of the Jewell Filter Company, were in charge of the Jewell Filter. Mr. Charles T. Whittier was chemist in charge of the Western Sys tems. The writer wishes to express his ob ligations to these gentlemen for their cour teous cooperation with him in conducting these tests. The Warren Filter and the two Western Filters were removed promptly at the close of the competitive tests. Aug. i, 1896. The Jewell Filter and the Harris Systems were not removed at once. During August, 1896, ar rangements were made with the Harris Com pany to utilize some of their appliances in supplementary tests made by the Water Com pany, and, when the investigations were re sumed at the beginning of 1897, arrange ments were also made by the Water Company whereby these two experimental plants could be used, in order to guard against delays. In the tests of electrolytical devices from Jan. i to March 10, 1897, the professional services of Profs. Mark and Brownell were retained by the Water Company to devise necessary electrical appliances, and to consult with officers of the Water Company with re gard to their operation. These new devices were installed at the expense of the Water Company in the temporary house formerly occupied by the Warren Filter. The settling tank, clay extractor, and po- larite filter used in the MacDougall Polarite System were constructed by Mr. MacDougall at his own expense, and in connection with the Jewell Filter were operated under the su pervision of Mr. John MacDougall. The tests of methods and appliances car ried on solely by the Water Company were made under the direction of the Chief En gineer and Superintendent, Mr. Charles Her- many, and of the Chief Chemist and Bacteri ologist. A large share of the engineering portions of the work was done under the gen eral supervision of Mr. Hermany, and the writer desires to express his obligation to him for much valuable advice and assistance upon the entire work obtained in frequent confer ences throughout the progress of the investi gations. A large amount of construction and repair work in the course of the tests was ably done by Mr. John Wiest, the engineer in charge of the pumping station. There was a considerable difference in the amount of work to be done at various times throughout the tests, and, accordingly, the number of assistants employed by the Water Company varied from time to time during the two years of investigation. The number ranged from two to seven, and averaged a little more than three, exclusive of the stenographer and porter. In addition to the construction and repair work on pipes and meters, the Water Company funished an engineer during the competitive tests to operate the wash- water pump. IN TROD UCTJON. .5 The following gentlemen were engaged as assistants during these investigations, and to their faithfulness and industry a large share of the success of the work is due: Mr. Chas. L. Parmelee, Assistant Engineer. Mr. Robert S. Weston, Assistant Chemist. Dr. Hibbert Hill, Assistant Bacteriologist. Mr. Joseph W. Ellms, Assistant Chemist. Mr. George A. Johnson, Clerk and Assist ant Bacteriologist. Mr. Reuben E. Bakenhus, Assistant. Mr. Harold C. Stevens, Assistant. All analytical work connected with these investigations was done in the laboratory of the Louisville Water Company, with the ex ception of the necessary mechanical analyses of filtering material, which were made at the Lawrence Experiment Station by Mr. Harry W. Clark. At the outset of these investigations it was arranged that bi-weekly reports of progress should be made by the Chief Chemist and Bacteriologist to the Directors of the Water Company, and to them alone. Early in the competitive tests the Jewell Filter Company and the Cumberland Manufacturing Com pany requested the Water Company to keep them informed as to the daily results accom plished by their respective niters. The Water Company, in response to this request, offered to furnish the operators of the filters tran scripts of analytical results obtained from their own filters (without any comments, sum maries, or conclusions), provided the filter companies would reimburse the Water Com pany for the additional expense incurred, and that the transcript of the results would not be used within a stated period for any purpose other than as an aid to the intelligent operation of their respective filters. The Jewell Filter Company and the Cumberland Manufacturing Company accepted, but the Western Filter Company declined, this propo sition. In compliance therewith the amount of analytical work was increased, beginning Feb. I, 1896. In order that the present report may be more readily understood it is divided into six teen chapters, as shown below, giving the full results of the investigations in their logical order. With the exception of Chapter I, on the composition of the Ohio River water, and upon which additional data were obtained in 1897, the first twelve chapters are presented in substantially the same form as prepared in 1896. It will be seen that the remaining chapters deal with the work of the current year. In the appendix are recorded the methods of analyses employed, and several other matters of purely technical interest. The chapters into which the report is di vided are as follows: I. Composition of the Ohio River water. II. Description of the application of chemicals to the Ohio River water by the respective systems of purifi cation. III. Decomposition and subsequent dispo sal of the alum or sulphate of alu mina solutions applied to the Ohio River water. IV. Coagulation and sedimentation of the Ohio River water by aluminum hy drate, formed by the decomposition of the applied alum or sulphate of alumina. V. Description of the filters through which the river water passed after coagulation by aluminum hydrate, and partial purification by sedimen tation. VI. Summary of the various parts of the respective systems, and a record of the repairs, changes, and delays. VII. The manner of operation of the re spective systems of purification, and the amount of attention given there to. V r ITT. Composition of the Ohio River water after treatment by the respective systems of purification as shown by chemical, microscopical, and bac terial analyses; together with a tabulation of the most important data upon the operation of the re spective systems. IX. Summary of the principal data upon the efficiency and elements of cost of purification, by the respective systems, of the Ohio River water, divided into twenty periods, accord ing to the character of the unpuri- fied water; together with a discus- 4 WATER PURIFICATION AT LOUISVILLE. sion of some of the more impor tant features. X. Description of the Harris Magneto- Electric System of purification, and a record of the results accomplished therewith. XI. Description of the devices operated by the Harris Company in July, and a record of the results accomplished therewith. XII. Investigation by the Water Company in August, 1896, into the practi cability and economy of the devices operated by the Harris Company. XIII. Description of the Mark and Brownell electrolytical devices, and a record of the results accomplished there with. XIV. Description of the MacDougall Po- larite System, and a record of the results accomplished therewith. XV. Description of the methods and de vices of the Water Company, tested during 1897, and a record and discussion of the results ac complished therewith. XVI. Final summary and conclusions. Appendix, containing technical records of methods of analyses, etc. COMPOSITION OF OHIO Rll ER WATER. CHAPTER I. COMPOSITION OF THE OHIO RIVER WATER. THE water of the Ohio River at Louisville varies widely from time to time in its compo sition. This variation is caused by a number of factors, among which are the following: 1. The size and varying geological forma tion of the watershed. 2. The number of comparatively large tributaries which drain areas of distinctly un like geological character. 3. The amount of precipitation (rain and snow). 4. The distribution of the precipitation over the watershed. 5. The condition of the soil at the begin ning of heavy rain-storms. 6. The amount and rate of precipitation during single storms. 7. The stage of the river. 8. The velocity of flow of the river. 9. Agitation of the water in the river, due to wind-storms, etc. The watershed of the Ohio River above Louisville is about 85,000 square miles in area. This area includes portions of the States of New York, Pennsylvania, Ohio, In diana, North Carolina, Virginia, West Vir ginia and Kentucky. Wide extremes in geological formation exist in the watershed. At Pittsburgh the Alleghany and Monon- gahela rivers unite to form the Ohio River. West of the city of Pittsburgh the drainage of this portion of the watershed finds its way into the Ohio River through thirty-three principal tributaries and a great number of smaller affluents. East of Pittsburgh there are numerous affluents to the two main streams, but they are correspondingly small in size. The total population resident on this watershed above Louisville is estimated at 4,500,000, of which 1,575,000 is contained in 220 towns and cities, according to the census of 1890, increased 15 per cent, for the six years of the present decade. The nearest city discharging sewage into the water which passes this pumping station is Madison, In diana, situated about 50 miles above Louis ville, with a population of about 12,000. The next city is Frankfort, Kentucky, situated on the Kentucky River 67 miles from its mouth. This city, has a population of about 10,000. The Kentucky River joins the Ohio about 57 miles above Louisville. The nearest large centre of population discharg ing sewage into this water supply is at Cin cinnati, Ohio. Opposite this city are the cities of Newport and Covington, Kentucky. Their aggregate population (three cities) is about 420,000, and they are distant above Louisville about 150 miles by river. At the pumping station of this Company where the tests and investigations were con ducted the Ohio River is about 1700 feet wide and 20 feet in average depth at low water. At the Ohio Falls, which are about three miles below the pumping station and opposite the city of Louisville, the river is about 4400 feet wide at low water. When very heavy freshets or floods occur in the Ohio River in this locality they cause the river to overflow its banks at the pumping station, and reach to the bluffs which run parallel to the river on the Kentucky side. The width of the river is then about 5500 feet. The rises and floods in the Ohio. River, with their associated factors, produce wide and rapidly changing variations in the com position of the river water. Owing to the fact that the composition of the river water is a prominent factor in the cost of purification, analyses were made practically every day dur ing these tests of the water before its appli cation to the systems of purification. Before giving attention to the results of analyses, however, the question of frequency and depth of freshets or floods is to be considered. 1 6 WATER PURIFICATION AT LOUISVILLE. FRESHETS OR FLOODS IN THE OHIO RIVER. PLAN OF ANALYTICAL WORK. Freshets or floods may be considered as stages or depths of water in the river which are above the normal. Their frequency and magnitude depend upon a series of factors connected with the rainfall on the watershed, and are very irregular. By virtue of the in fluence which they exert indirectly upon the cost of purification, and the method leading to the most efficient and economical results, they are worthy of very careful consideration. This is especially true in connection with this report, because inspection of the data pre sented in the following table shows that during the period covered by the first portion of these tests the magnitude of freshets or floods was below the normal for the past thirty-six years. Practically speaking, this means that the average amount of mud, silt, and clay sus pended in a given volume of the Ohio River water during these investigations and tests was abnormally small. In the following table is given a summary of the number and magnitude of the freshets or floods which occurred in the Ohio River at Louisville during the past thirty-six years, 1861 to 1896, inclusive. This was obtained from curves prepared annually by the Water Company from data obtained daily at the pumping station during the entire period. These annual curves were made from plot- tings of a convenient scale, in which the ab scissae correspond to the number of days in a year, and the ordinates to depths of water above the low-water level. By connecting the points plotted in this manner curves have been obtained which show the depth of the river water for each day of each year of this period. The number and extent of the fresh ets orjloods were obtained by noting those portions of the curve corresponding to rising, fairly stationary, and falling stages or depths of water in the river. The end of a given freshet or flood is shown by a return to the normal depth of water, or by the beginning of another freshet or flood quickly following the one in question. To obtain the depth of a given freshet or flood from these curves the difference in elevation is noted between the initial and highest point of the given portion of the curve. In the determination of the composition of the Ohio River water, as shown by analyses, attention was directed to the physical, chemi cal, and biological characters of the water. Physical Character. Upon this point the examinations included observations on the ap pearance and character of the matters in sus pension, and on the odor, color, taste, and temperature of the water. Chemical Character. The chemical analy ses included the determinations of the total amount by weight of the mineral and organic matters dissolved and suspended in the water; the amount of organic matter in solution and in suspension; the form in which the nitro gen was present in its passage through the cycle from crude organic matter (albuminoid ammonia) to completely mineralized matter (nitrates); the alkalinity, due chiefly to the carbonates and bicarbonates of calcium and magnesium, which indicated the amount of alum that could be effectually decomposed by the water; and the amounts present of chlo rine, dissolved alumina, iron, and fixed residue on evaporation after ignition to burn up the organic matter and effect incidental changes. These determinations compose the regular sanitary and technical chemical analysis of water for work of this class. In addition to the regular chemical analy ses, as stated above, there were made from time to time as occasion presented special sanitary and technical analyses. Among these were included the determination of the amounts of free (atmospheric) oxygen and car bonic acid gases, dissolved in the water; and the amounts of incrusting constituents of the water in connection with its adaptability for use in boilers. Mineral analyses were also made of several samples of water which were representative of different grades in the wide range of com position of the water met with in these inves tigations. These analyses consisted of the de termination of the principal metallic elements and the acids present in the mineral com pounds contained in the river water. Biological Character. The biological analy ses consisted chiefly of the determination of the numbers of bacteria present in the water, COMPOSITION OF OHIO RIVER WATER. 1 8 IVATKR PURIFICATION AT LOUISVILLE. and of the examination of the species of bac teria with special reference to their connec tion in the causation of disease. Microscopical examinations were also made from time to time to learn the numbers and kinds of alga:, diatoms, infusorise, etc., pres ent in the river water. These microscopical analyses differ distinctly from the bacterial analyses in that the former relate solely to those relatively large micro-organisms which may be counted and classified with the aid of comparatively low powers of the microscope; while the bacteria are so small (about o.oooi inch in length) that they require for their enumeration and classification special meth ods of laboratory procedure. Preceding the several tables showing the results of analyses there will be found ex planatory notes, calling attention to the na ture of the principal points of practical significance. At the close of the report is an appendix in which is presented a record of some of the more important features of the analytical methods from a technical stand point. The plan of analytical work, which has been briefly outlined in the foregoing para graphs, may be made plainer by the following synopsis: Synopsis of Analytical Work. 1. Physical: Appearance, odor, color, taste, and temperature. 2. Chemical: Regular sanitary and technical analyses. Special sanitary and technical analyses. Mineral analyses. 3. Biological: Microscopical examinations. Quantitative bacterial analyses. Identification of species of bac teria, with special reference to the causation of disease. Place of Collection of Samples of River Water for Analysis. Samples of river water for analysis were collected from a tap on a 6-inch pipe. This tap was kept open during working hours. The 6-inch pipe was about 230 feet in length, and connected with the force main leading to the distributing reservoir at Crescent Hill. From the intake to the point where the 6-inch pipe branched from the force main the dis tance was about 390 feet. In this distance the water passed through the pump well and the pump which was operated to supply the city. The intake of the water supply is located 3.5 feet below the low-water stage and about 100 feet from the Kentucky shore at low water. Manner of Collection of Samples of River Water for Analysis. After the investigations were well under way it was the general custom to collect on each working day, from the above-described place, one sample of water for regular chemi cal analysis, and two or more samples for the determination of the numbers of bacteria. When the systems of purification were in op eration night and day samples of water for both chemical and bacterial analyses were col lected once in six hours. In the case of the chemical samples four portions were mixed together to give a representative sample for the day. Samples of water for other analytical pur poses were collected from time to time as occasion demanded, and as the pressure of regular work allowed of their analysis. All samples were placed in a large ice-box during the period which intervened between their collection and their analysis. PHYSICAL CHARACTER OF THE OHIO RIVER WATER. The most noticeable of the physical charac ters of the Ohio River water is its appearance with regard to the matters suspended in it. At no time was the river water clear and free from suspended matters. During October and the greater part of November, 1895, the water was comparatively clear; but even at that time it had a distinct turbidity due to the presence of minutely divided particles. The first heavy rains caused the water to be come muddy. From that time until the close of the investigations the appearance of the COMPOSITION OF OHIO RIVER WATER. river water possessed a wide range of rapidly changing variability. As a means of expression of the relative ap pearance of the Ohio River water the use of adjectives fails utterly. The best idea of the varying appearance of the water is obtained from the results of the daily determination of the weight of the matter suspended in it. These results form a portion of the regular chemical analyses; and reference is made to the following tables in which they are pre sented, and to an explanation of them in the note which precedes the tables. Here it will suffice to state that the weight of organic and mineral matter suspended in the water ranged from i to 5,311 parts per million. The ratio between the weights of the maximum and minimum suspended matter, therefore, was 5,311 to i. The appearance of the suspended matter it self was quite different from time to time, ranging from a light gray to a dark red color. A series of factors influenced the appearance in this regard. Prominent among them was the character of the soil on which the rain fell. The extreme conditions of muddy water in connection with the appearance of the sus pended matter were noted in March and in May, 1896. During March heavy rains fell throughout the valley. All the tributaries were in flood, and during the last days of the month, when the velocity of the Ohio River was great, the water had a decided red appearance. These particles were comparatively large and came, apparently, from the upper portion of the water shed. In April and May, 1896, there was a period of extended drought and the surface of the earth was very dry. The rains which came during the last week in May produced muddy water, which contained an immense number of minutely divided particles of a light gray color. This gave the water a yellowish ap pearance. Some of the particles were smaller than bacteria and measured under the micro scope less than o.ooooi inch in diameter. Naturally enough this water was very difficult to clarify. Between these extreme conditions of ap pearance there was a wide range of intermedi ate conditions, depending upon the relative influence of the series of factors outlined on page 15. During 1897 there was a still greater range than during 1895-96 in the amounts and character of the suspended matter in the river water, although at no time were the clay par ticles finer than in May, 1896. Further, it appears that the heavy mud is most prevalent during the winter and early spring, while the fine clay prevails in the late spring and summer. Odor of the River Water. The Ohio River water, when it was not heated, possessed as a rule a faint odor, the intensity of which was somewhat variable. Occasionally the odor was quite pronounced, but often no odor could be detected. Dur ing the fall, winter, and early spring the odor was usually musty, sometimes aromatic and resinous. After the rains in the spring the odor had a vegetable character at times. Upon heating the river water the odor be came stronger, especially in the case of the vegetable odor noticed during warmer weather. In practically no case, however, was the odor disagreeable, or stronger than would be expected in a surface water of this kind. Color of the River Water. It is the suspended particles in the river water which give to it a color. This has al ready been referred to under the appearance of the water. When the water is freed from its suspended particles it is practically colorless. In the following tables containing the re sults of the regular chemical analyses will be found a record of the amount of dissolved color, expressed in units of the platinum- cobalt standard. These color results were obtained after the suspended particles had been removed by the passage of the water through a fine paper filter or a Pasteur filter. Taste of the Rh cr Water. Disregarding the suspended matter, the taste of the river water is satisfactory, al- WATER PURIFICATION AT LOUISVILLE. though the salts dissolved in it, especially the lime, give a slight taste which is noticed by those accustomed to drinking a softer water. There is at times a slight earthy taste to the water. The suspended matter cannot be regarded as other than objectionable. But after per sons become familiar with this kind of water there appear to be comparatively few com plaints, except when the water is very muddy. Temperature of the River Water. The results of observations on the tem perature of the river water, expressed in de grees centigrade, are presented in the tables beyond, containing the results of the regular chemical analyses. CIIFMICAL CHARACTER OF THE OHIO RIVER WATER. In the next set of tables there are pre sented the results of the regular chemical analyses of the river water from a sanitary and technical standpoint. The times of collection, the temperature, and the color results will be readily under stood from the foregoing pages. They are recorded here as a matter of convenience. The remaining columns contain the resul s of the several chemical determinations. An outline of the analytical methods used in these chemical determinations will be found in the appendix. In order that the practical signifi cance of these results may be understood more clearly a brief explanation of them will be given. Explanation of the Results of Chemical Analyses. The several points will be taken up in the order in which they appear in the tables. Form of Expression. All of these results are expressed in parts per million. The exact meaning of this is that one million parts of water by volume contained the several sub stances in parts by weight to the extent in dicated by the figures. These results may be converted into grains per United States gallon (231 cubic inches) by dividing by 17.1. Oxygen Consumed. The results of this de termination indicate the amount of organic matter present in the water. By analytical methods there is measured the amount of oxy gen which is actually consumed by the or ganic matter in the water, as it is converted into a comparatively stable form not readily capable of farther decomposition by ordinary means. As nitrogen cannot be oxidized by this method these results are generally considered to be indicative of the amount of organic matter of a carbonaceous nature. Nitrogen as Albuminoid Ammonia. When water containing organic matter of a nitrog enous nature is distilled with a strong alka line solution of potassium permanganate, the organic nitrogen is changed to ammonia. This ammonia is spoken of as " albuminoid ammonia," and the results of determinations by this method indicate the amount of or ganic matter of a nitrogenous nature. A comparison of the results of analyses by the last two methods indicates that the nature of the organic matter in the river water varied considerably, according to the relative results by these methods for its determination. It will also be noted in the tables that the results by the second method show the amounts of organic matter in suspension and in solution, respectively. Comparatively speaking, the amount of nitrogenous organic matter in solution is fairly constant, although it varied somewhat at different seasons of the year. Nitrogen as Free Ammonia. Upon the dis tillation of the river water without chemicals there is obtained in the distillate a small quan tity of ammonia. This is known as the " nitrogen in the form of free ammonia." It measures the amount of nitrogenous organic matter which has undergone the initial step in the decomposition of organic matter by na ture. This decomposition in nature is accom plished in the presence of oxygen by bacteria which eventually convert crude organic mat ter into harmless mineral matter. Nitrogen as Nitrites. These results show the amount of organic matter that is in the COMPOSITION OF OHIO RIVER OOOOOOOOOOOOOOOOOOOOOOOOOOO-ii-iN" ooooooooooooooooooooooooooooooooooooooooo i O r- O w O O co O ** 00 ON 1-1 O O co - O ddo doooooooooooooooooooooooooo pspuadsns .0 m \r> m in r- -t >n"r- in uS ir>oOO"OOOOOOOOOOOOO^O -O O 808806000000000000000000000000 -o o CO N -fOOOO N Tj- -t C4 N O O O tot 2 o" ^ "o oO" V o v o3"o *o ?" 5 1 o 1 oo o o o o IH o o o So 0*0 o o o > sO O O O -t pauinsuo^ ua3XK(> oooooo oo o^ o O"">O O O O i-- -t 0*0 f-* *> 0^1- w m-tinoco o t-i w ci r^c^c-i -to ^xno I^ON co-to oo - bo^^bc^ ^"^ be WATER PURIFICATION AT LOUISVILLE. COMPOSITION OF OHIO RIVER WATER. a: O S O ^ w ^ 1 U. w c g. o- o>o o o o f> o EU|Uin|V P3A10SS|Q OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO O -r O o o^ -< *- o ~O tn^o I O c> papusdsns ooooooooo oooooob oo o ooob oobooooooooooccO -ooooooooo i- i- o o o o c? ? b ? ? o 3" o" o f o o o o oooooooooo WATER PURIFICATION AT LOUISVILLE. OOOOOOOOOOOOOOOOOOOOO-OOOOOOOOOOOOOOO i. w . O <n o- - i- O t---<OOOaoa^OOi -O -J--1-O M T O CO CNI-I m*O pspUDdsns o o o r^ -< - < x s "5. "a. E e rt rt SOOOOgOOOOOOOOOOOOOOOOOOOOOOOO"--""-- ooooSoooooooooooooooooqqqooooooooooo 5 1 ???" o > o > o ?o o > o ??o 1 o o o o cfo o o > o o 1 o > o > o > ?o oo "0 ^o ? O OOO3 O COOOOOOOO^OOOOOOOOOOOOOOOOOOOO - OOOO O f) in O O is^ COMPOSITION OF OHIO RIVER WATER. > s o 5 O T oooooooooooooooooooocoooooooooooooooooo papuadsns M p i Q O H - c O O ! papuadsns o o o i- e^w Ot*~)M O w Oco O 1 * O^co 3C 1 C 1 O S O^- < O^OOO ( - 1 ~ | C MO | - | O 1 - | OOOW O -fWOaO * ci MOO "T -f T M -1- O Cl O fi O W O fl -too O < ***** o-**** ooooooooocoooooooooooooooooo WATER PURIFICATION AT LOUISVILLE. o o o ? o o* o o o o o* o 1 o o o* o o 00*0 o o o o papuadsns o o o o o"6 o o o"o"b o v en -* tn woo C4 -1- * T O CH O O -t W O O O O O -TO O -*** - -- ~ O O O O cnu->r-O O - O O O i- - COMPOSITION OF OHIO RIVER WATER. - 7 * s 1 papusdsns |>nJ. paAiossiQ popuadsns O O NCOOOOOOO 1- OOOOOOOOOOOOOOOOOOOO O O O O OOOOOOO O O O O "- O O HiOOOOOOOOOOOOOOoOOOO O O O O OOOOOOO O O CCCOCOCOOOCQ OD OO OO CO e^CQCjOOQ IVATF.R PURIFICATION AT LOUISVIl.Lb. o o o ooo o ooooo o ooo ooo o o o o H ~ coo O O O 000 O 0000 000 000 O O o ooo o COMPOSITION OF OHIO RIVER WATER. 9 w O O O O 1-1 O O o _ O O O O O -1- 0*0 o S & "8 S-o -papuadsns t >o o o o o o 660 O O v> 8 S 5 o ~ 5 5 c^ o o i^ r^ i^ 3 WATER PURIFICATION AT LOUISVILLE. o oooooo o o o o oo o o ooo o o s- s r^ O t O t O o o a* ooo~ do" odd So o 000 o ooo ? o o oooooo o ooo o o s papuadsnij H O O oo o. o o o o o O O O O u~. O o o o o O 000 ooo r^ so o w. 888888 ?, ooooooooo oooooo COMPOSITION OF OHIO RIVER IVATER. pspuadsng pspusdsns o o o o o o o o o o o o S 2 o o o O O WATER PURIFICATION AT LOUISVILLE. second intermediate stage of the process in nature by which organic matter is converted to mineral matter. Nitrogen as Nitrates. From these results there is learned the amount of organic matter which has been completely oxidized and changed into the form of mineral matter. Chlorine. Chlorine is usually supposed to be present in water as common salt for the most part. Some of it very likely comes from mineral deposits on this watershed. Salt is also present in sewage, and this is one of the reasons why it is accurately determined. The determination of chlorine is also of value in studying the composition of a water by virtue of the fact that it is not affected by any ordinary conditions which waters meet in nature. It is always soluble, and cannot be oxidized or reduced. For this reason it does not pass through a cycle of changes as does nitrogen. A comparison of the nitrogen and chlorine is therefore instructive. Residue on Evaporation. The residue on evaporation shows the total weight of the solid matter which the water contained. In the following tables the total residue is sub divided into suspended and dissolved residues Attention is especially called to the sus pended residue on evaporation. This shows the weight of the matters suspended in the water, and gives a good general idea of what the relative appearance of the water was on the different days. Fixed Residue after Ignition. After weigh ing the total residue of the water upon evapo ration it is the custom to heat the platinum dish containing the residue to a dull red point and again weigh it. By this means the fixed residue is obtained. Formerly it was supposed that this ignition burnt off the organic matter, and the differ ence in weight of the contents of the dish would give the amount of organic matter present in the water. This is not true, how ever, because the ignition volatilizes certain mineral constituents of the water, which would be erroneously figured as organic matter. Nevertheless, the fixed residue on evapora tion appeared to be of some value in studying the comparative composition of the mineral constituents of the water. Alkalinity. This determination is one of great importance in connection with the puri fication of water by the method under investi gation. These results show the amount of carbonates and bicarbonates of calcium and magnesium which were present in the water. It is these compounds which decompose alum or sulphate of aluminum, as is explained in Chapter III. The results are expressed in terms of cal cium carbonate (lime). They are somewhat similar to the " temporary hardness " deter mination by the " soap method." Dissolved Alumina. No appreciable amount of dissolved alumina could be found in the river water. The results of the tests are recorded, however, as they are of value in the study of the question as to the passage of alum through the systems into the filtered water. Iron. The results of the determination of iron are of value in showing variations in the composition of the suspended matter in the water. Practically all of the iron was con tained in the suspended matters. Special Chemical Analyses. There were several sanitary and technical problems under consideration, which re quired special chemical analyses from time ( o time, as follows: 1. Atmospheric oxygen dissolved in the water. 2. Carbonic acid gas (carbon dioxide) dis solved in the water. 3. Those dissolved chemicals in the water which give to the water its " permanent hard ness " and its power to produce incrustations in steam-boilers. The first set of these data was fairly com plete, from a practical point of view, during 1895-96, and the analyses were made less fre quently in 1897. With regard to the latter sets of data, however, the evidence early in 1897 showed that these constituents were so variable and of such importance that they were included in the regular analyses. As a matter of convenience, however, the results are recorded here; but reference to the fore going tables will show their relation to other constituents of the water. The significance COMPOSITION OF OHIO RIVER WATER. 33 of these results is explained and discussed in subsequent chapters. Dissolved Oxygen in the River Water. These results are of value in connection with the preservation of the quality of the water after purification, and for a comparison of the water before and after treatment, especially in the electrolytic process with iron elec trodes. The amount of atmospheric oxygen which may be contained in the river water is limited by the saturation of oxygen gas in the water; and the saturation depends chiefly on the temperature (and pressure), the amount of oxygen necessary for saturation decreasing as the temperature increases. In the adjoin ing table the amounts of oxygen gas dis solved in the river water are expressed in parts by weight per million parts of water by volume, and in percentages of the amounts necessary for the saturation of the water at the actual temperature at the time of col lection. Carbonic Acid Gas Dissolved in the River Water. As an aid in an investigation into the influence of carbonic acid gas (carbon di oxide) upon the corrosive action of the river water before and after purification by differ ent methods, the amount of this gas which was naturally dissolved in (he river water with the formation of carbonic acid was deter mined with results given on page 34. It de veloped in the course of the tests that the determinations of carbonic acid gas are of more importance than was generally sup posed to be the case formerly, and. as stated above, the analyses were made with more fre quency in 1897 than during the first part of the work. In passing it may be noted that at times the weight of carbonic acid gas dis solved in the river water equalled and even exceeded the weight of all solid matters dis solved in the water. Hardness of the River Water. The hard ness of a water depends upon the presence of dissolved salts of calcium (lime) and magne sium. These salts consist of the carbonates, bicarbonates, sulphates, chlorides, and ni trates. The bicarbonates are carbonates which are held in solution by carbonic acid. For many years it has been the custom to subdivide hardness into " temporary hard ness " and " permanent hardness." Tempo- PARTS PER MILLION OF ATMOSPHERIC OXY GEN DISSOLVED IN THE OHIO RIVER WATER, WITH PERCENTAGES SHOWING THE RELATION BETWEEN THE AMOUNTS FOUND AND THOSE NECESSARY FOR SAT URATION AT ACTUAL TEMPERATURES. Date. 181,5. Tempera- Dissolved Atmospheric Oxygen. Parts per Million. Percentages which the of the Amounts Required for Saturation. Satura- Found. Dec. 3 3-4 2.32 10.3 78 " 4 4-5 2-57 9-5 76 " 6 4-2 2.69 0.2 80 " 9 4.1 2.88 1.2 . 87 1896 Jan. II 2.1 3.52 1.6 86 Feb. 10 5-9 2.19 1-3 93 " 15 7-0 1.90 1.6 97 " 26 3-4 3.01 3- IOO Mar. 4 4.4 2.61 1.6 92 " II 6.5 2.01 0.8 90 " 19 5-2 2.38 2-4 IOO Apr. 9 9.9 1.16 O. 2 91 May 6 23.1 8.47 7-2 85 " 4 24.0 8.34 6.6 79 " 23 24.5 8.28 6.2 75 " 29 24.7 8.25 5-9 71 June 5 24.2 8.32 6-3 76 " 10 24-7 8.25 66 So " 18 25-3 8.20 6.4 78 " 24 26.8 8.02 6.4 80 July 9 25-5 8.17 5-9 72 " 18 25.6 8.15 5-8 7i 1897 April 10 -3 10.81 10. 1 93 " 20 it. 6 10.75 10. I 94 " 29 16.3 9.61 8.5 88 June 4 20.9 8.81 8.7 99 " 5 21.8 8.65 8-7 IOO " 17 26.7 8.04 7-4 92 " 18 29.1 7-73 6.6 85 " 25 26.0 8. II 6-7 82 " 27 25 3 8.20 6.9 84 11 28 25-3 8.20 6.6 81 " 3 26.2 8.09 8.0 99 July 1 26.6 8.05 6.7 83 1 3 27.6 7-95 6.2 78 1 9 30.0 7.76 5-4 69 3 26.5 8.05 6.8 84 4 27.2 7.98 7.0 88 10 26. i 8.10 8.0 99 20 26.8 8.02 8.0 IOO " 21 27-5 7.96 8.0 IOO " 23 27.7 7-94 6.5 82 " 27 26.2 8.09 4-6 57 rary hardness is caused by bicarbonates of lime and magnesia which are precipitated upon boiling, due to the expulsion of car bonic acid gas. The remaining salts of lime and magnesia, as stated above, have been re garded as permanent hardness. The practical significance of the above- stated salts of lime and magnesia is twofold in connection with these investigations, namely: 34 WATER PURIFICATION AT LOUISVILLE. AMOUNT OF CARBONIC ACID GAS (CARBON DIOXIDE) DISSOLVED IN THE OHIO RIVER WATER. (Parts per Million.) Date. 1896. Car- Acid Gas. Date. 1897. Car- Acid Gas. Date. ,897. bonic Acid June 18 30.8* April 3 53-5 May 29 IOI.6 " 22 26.4* " 7 79.6 June I 66.6 " 24 27.7* " 8 46.0 2 90.5 " 27 29.7* " 9 gi.o 4 82.7 July 3 30.6* " 10 So.o 7 89.0 " 8 21. I* " 12 65.0 " 10 133.0 Nov. 28 83.0 " 13 44.0 " it 107.6 Dec. 10 98.0 " 14 75-7 " 15 98.8 1897 " 15 88.3 " 16 103.3 Feb. 16 80.4 " 16 50.2 " 17 107.6 Mar. 2 63.4 " 21 41.2 " 18 82.7 " 3 59.0 22 42.7 " 19 106.3 4 67.8 23 43-0 " 20 100.3 " 5 49-3 25 55 o 21 100.3 11 6 47-6 27 94-9 22 100.3 7 51-4 2 9 85-9 23 107.4 II 99-5 May 4 S6.S 24 100.3 12 S8.o 7 57-4 25 105.6 13 122.4 " 8 1 10.6 26 "3-7 15 45-8 " 9 66. 7 " 27 I2O.O 16 33-4 " 10 72.1 " 28 92.1 19 38.8 " 13 65.2 " 30 105.9 20 42.6 " ! 4 76.6 July i 93-9 22 46.4 " 15 50.8 : 2 75-3 23 40.4 " IS 67.3 3 73.1 24 44-9 " 19 71.8 6 100.4 25 41.9 " 21 88.7 7 106. j 26 36.5 " 22 95-9 8 99.9 27 47.0 " 23 94-3 12 71. S " 29 56.6 " 26 So. 2 15 47.0 " 30 So.o " 27 So.o 16 28. S Apiil t 53-6 " 28 107.3 17 49-4 * The results of June and July. 1896, were obtained by the Pettenkoffer method, without the Trillich modi fication, and are probably much too low. 1. It is the carbonates and bicarbonates of lime and magnesia in the river water which possess the power of decomposing such ap plied chemical products as alum and sulphate of alumina, and thereby forming the gelati nous hydrate of aluminum that acts as a co agulant. 2. It is the remaining salts (sulphates, chlorides, and nitrates) of lime and magnesia which are connected with the formation of in crustations when the water is used in steam- boilers. By the old Clark method of getting the bi- . carbonates, called temporary hardness, the full power of the water to decompose the commercial chemicals stated above is not re corded, because it does not include the car bonates. In practice it is found that the car bonates will decompose 3 grains or more of sulphate of alumina per gallon. To use a method which shows only the bicarbonates is, therefore, inadmissible; and Hehner s method was employed. This method furnished what is required, that is, both the carbonates and bicarbonates. For the sake of explicitness these results are recorded as the alkalinity of the water in the foregoing tables of analyses. As the Ohio River possesses no carbonate or bicarbonate of soda or potash, the full alkalinity of the water is due to the carbon ates and bicarbonates of lime and magnesia. By the old Clark method the carbonates of lime and magnesia are recorded with the " in- crusting constituents " or " permanent hard ness." The facts show that these two com pounds are permanent, but they form a sludge, and not an incrustation, in steam-boil ers. By the Hehner method, which was em ployed in these investigations, the carbonates are not included in the following table of re sults, which, in the absence of a better name, are termed the " incrusting constituents " of the water. These results, which are dis cussed in Chapter XV, are expressed accord ing to the conventional method in equivalent parts of calcium carbonate. A further con sideration of the methods of analyses will be found in the appendix. The dissolved salts of lime and magnesia are a so of importance in connection with the consumption of soap when the water is used for washing purposes. This point is practi cally uninfluenced by the purification pro cesses under consideration, but the range of variation in this soap-consuming ingredient may be noted by taking the sum of the alka linity and incrusting ingredients. See first table on page 35. These results approximate the total hardness results obtained by the Clark method. Mineral Analyses of the Ohio River Water. A record of the results of the determination of the mineral constituents of the river water is presented in the next table. Eight samples were analyzed with as much completeness as circumstances allowed; and the results show very clearly the marked variations which the COMPOSITION OF OHIO RIVER WATER. INCRUSTING CONSTITUENTS OF THE OHIO RIVER WATER. (Parts per Million.) Incrust- Incrust- Incrust- Date. ing D;ite. ing Date mg 1895. Constit 1897. ,897. Constit uents. UelHS. uents. Dec. 9-1 1 43-9 Mar. 25 10.0 May 19-20 IO.S 1896 " 26 13-9 21 II.9 May 6 43-0 " 29 12.8 " 21-23 15.9 M 33-8 " 30 33-3 " 23-24 IO. 2 " 22 40.1 Apr. I 29.1 25J 20.2 " 2g 41.1 " 2-3 18.2 " 25-26, 14.5 June ii 44.0 3-4 II. 27 16.7 18 30.0 5 29.9 " 27-28] 19.3 July 30 35-0 6 23.4 " 28-29 16.8 1897 " 6-7 30.0 31 I I7 n Feb. 17 18.7 " 8 19.8 June i f 1 ? " . 22 24.7 " 9 14.2 2-3 17-6 " 23 17.2 i "9-0 13 3 4-5! 23.5 " 24 21.2 j I 9.0 7-8 22.5 " 25 It, I 2 12.7 9-10. 23.8 " 26 10. " 13- 4 If).0 11-12 28.8 " 27 S.o M- 5 2O.O 21 28.8 Mar. I 15.8 15- 6 iS.i 22 3I.O 2 10.0 20-21 15-5 23 31-8 3 8.0 21-22 17.5 24 27.8 4 12. 22 12. 25 21.9 5 25-3 22-23 14.6 27 19.0 (, 34-6 23-24 U.3 28 17-5 7 1 6.0 27 12.7 29-30 25-5 9 23.4 28 17.0 July 2 20. " 10 17.9 29 32.0 6 20.9 " n 30.0 " 29-30 21-7 7 47.0 " 12 5 33-u 14.0 30 May I f- 20.0 9-10 12-13 II. O 12.8 16 20.5 4 23.0 14-15 14.8 17 24.6 5 17.0 l6j 14.0 18 36.0 " 6-7 15.5 17-18 12.3 19 16.7 8-y 15.8 " 19-20 24.2 20 22.0 13 I6.I " 21-22 24.2 22 12.4 14 23.2 " 23-24 43.8 23 9.0 15 "3 24 10.8 " 17-19 9.0 composition of the river water possessed dur ing these investigations. I The sample which was collected on May 29 and 30, 1896, was analyzed both before and after nitration through fine filter-paper. At this time the water contained a large amount of very finely divided particles; and it was probably the most difficult water to purify without subsidence that was encountered dur ing the whole work. The sample was col lected just after a heavy rain, following an extended period of drought. From March 23 to 29, inclusive, the sam ple for analysis was prepared by mixing equal small portions of the river water collected every six hours. During this time the sys tems of purification were in operation night and day. By automatic devices samples of filtered water were collected, representing the entire period. The analyses of the filtered water are presented in Chapter VII. According to the general custom the re sults of the determination of the various ele ments in the water, both metallic and acid, are expressed in the following table of analyses in the form of oxides (except the chlorine). As would naturally be expected in the water of a river, the watershed of which offers such a wide range in the possibilities for dif ferent kinds of rock disintegration and sur face erosion, the relative amounts of the mineral constituents are seen to vary widely. This is shown very forcibly in the following table, in the case of suspended matters, by a comparison of the ratio existing between the alumina and the oxide of iron. RESULTS OF MINERAL ANALYSES OF THE OHIO RIVER WATER. (Parts per Million.) Periods of Collection. ,8 95 ,896 Oct. 28 No v M. NOV. 23 Nov. 29. Dec. 9 Dec. 20. Jan",, Feb. 7 Feb. 27. Feb. 28 Mar. 18. Mar. 23 Mar. 29. May 29 and 30. Unfil- tered. Filtered. Silica (SiO,) II .2 0.2 3-7 25.9 0.4 3-9 [o, 58.2 28.4 3-7 7-3 1.4 49.0 13.8 227.2 21 .6 15-7 0,) 39.6 206.7 26.0 70.0 42 f) 6.4 5-9 299.5 39-4 76.6 2.2 I . I 31 7 14.0 8.5 18.1 325-3 32.8 131.4 3-4 2-9 69.2 28.1 93 0.6 o Trace. 47-7 3-9 Oxide of iron (Fe a Oj) Alumina (Al a O 3 ) Oxide of nickel (NiO) Lime (CaO) . . 35-5 13.0 32.7 11.4 Magnesia (MgO) . Potash (K?O) Chlorine(Cl) 65.4 3-1 44 o 39-9 4.8 43.0 18.0 39-9 5-o 29.8 24.3 Trace. 10.7 22.3 20.8 19.4 Trace. 6.4 16.9 19.7 28.2 1.2 8.2 ii. 6 25-9 37-4 Trace. 5.6 14-7 21.3 23-3 1.6 10.4 3-3 38.7 18.1 3-7 10.4 33 38.7 18.3 Trace. Nitric acid (NjO) . Carbonic acid (combined) (CO*). ... P WATER PURIFICATION AT LOUISVILLE. BIOLOGICAL CHARACTER OF THE OHIO RIVER WATER. Determinations of the number of bacteria in the river water, and a study of their relation to disease, occupied the gerater part of the at tention with regard to this portion of the work. Microscopical examinations of the water, however, were made from time to time to learn the numbers and kinds of the larger micro-organisms which were present. Microscopical Examinations of the River Water. In the next table (see p. 37) there are pre sented the results of the microscopical exam ination of the river water for the presence of algae, diatoms, etc. As already stated, these micro-organisms are much larger than the bacteria, and may be classified by the aid of low powers of the microscope. It will be noted that the algae (cyanophyeae and chlorophyceae) and diatoms, which are visually abundant in surface water during hot weather, were present in only very limited numbers after the last of May, 1896. The reason of this was, undoubtedly, that the large amount of suspended matter in the water prevented the sunlight, which is necessary for their development, from reaching them. In 1897 no microscopical examinations of the river water were made during its muddy condition. The single analysis on June n, however, when the water was very clear, com paratively speaking, shows the range in num bers and kinds of organisms which may be expected under such conditions. Identification of Species of Bacteria in the River Water, -with special reference to their Causation of Disease. With regard to this portion of the biologi cal analyses attention was especially directed to the detection of bacteria which are dan gerous or suspicious from a hygienic point of view. It is to be stated that during the low water in the river in November, 1895, and again during the last part of April and early part of May, 1896, when there had been a drought for a month or more, there was found in the river water B. coli communis. This germ is the most prominent one in sewage, and it is also the most abundant species in the fecal discharges of man and certain domestic animals. On May i, five days after the be ginning of the period when this germ was re peatedly found in the river water, an exam ination of the tap water in the city also showed its presence there. At high stages of the river and when the water was very muddy the results of numer ous examinations for sewage bacteria were negative in a great majority of cases. Nevertheless, B. coli communis was found in the river water on June 30, 1896, and closely allied kinds of bacteria were noted from time to time during the investigations. The evi dence indicates that with muddy water and high stages of the river there are conditions which aid in causing the disappearance to a marked degree of those germs which appear to come from the entrance of sewage into the river above the pumping station. The results of tests for the presence of the germs of typhoid fever and other diseases were negative in all cases. It is not to be understood, however, that these negative re sults are adequate proof that disease germs were entirely absent from the river water. The reason of this lies in the natural limita tion of the most approved laboratory methods, which at best allow an examination of only a very small portion of the quantity of water flowing in the stream. These comments are especially true in view of the fact, as stated above, that at times of low water there were present intestinal bacteria. In 1897 twelve tests for B. coli communis were made between January 2 1 and February 4, with negative results in each case. From April i to 9, nine more tests were made in which this germ was found in three instances. The question of the classification of the nu merous but harmless species of bacteria in the writer received as much attention as time al lowed. Owing to the fact that there were several other lines of work which yielded re sults of greater importance from a practical standpoint, this question was not made one of constant study. Nevertheless, the investiga tions on the detection of dangerous or suspi cious species, and on the comparison of the species of bacteria in the water before and COMPOSITION OF OHIO RIVER WATER. RESULTS OF MICROSCOPICAL ANALYSES OF THE OHIO RIVER WATER. (Number of organisms per cubic centimeter.) Date of Collection. 1 1896. February. 1896. March. 896. April. 1896. May. 1846. June. 1896. July. 8 1897. lunc. jj ,6 4 ,, 19 Number of Sample. 277 280 1 304 329 354 382 404 459 471 516 544 564 585 629 651 684 711 jooi D 340 ft 60 18 38 112 44 6 24 2 2 659 "5 69 165 37 16 96 60, 8 8 9 16 5 40 40 8 M 16 8 6 16 2 2 Tabellaria So I 173 I 8 3 g Cyclotella 20 20 8 2 g -.2 8 g g ft I I 17 8 8 289 2O o 8 2 I 8 o o 2 8 26 352 20 4 16 2 O g 22 88 8 g 26 20 M3 2 3 19 I 2 13 g O 60 60 20 8 8 at 2 2 2 2 2 3 12 12 6 10 IO 8 8 16 2 2 O 2 O 2 2 2 o 6 | 2 8 20 4 24 2 4 IO 8 8 2 2 8 8 16 Cilliata g 2 8 3 o o 4 2 O o 2 o o 8 o 16 g O O I 13 I i 3 2 I Rotifera g 40 8 2 2 3 o 12 4 24 8 16 . . . . 6 6 O Ova 8 2 8 4 IO 44 7 2 72 IO 400 6 Ver 200 5 y abi 168 8 nda 28 10 nt in II all c 70 3 ases. 21 5 168 8 66 IO 496 96 8 9 3 5 80 9 6 3 6 3 993 26 Amorphous matter after purification, necessarily involved a con siderable effort in this direction. Comparison of the results of diagnostic tests used for the identification of species of bacteria with the available published descriptions of bacteria indicated, so far as the similarity of labo ratory methods would allow, the presence of several new species as well as a consider able number which have been found else where. The following list of bacteria and yeasts, together with the results of the microscopical examinations already presented in this chap ter, indicate the nature of the microscopic llora of the Ohio River water: Bacillus arborescens (Frankland). aurantiacus (Frankland). coli communis (Kscherich). flavescens (Pohl). fluorescens liquefaciens (Fliigge). fluorescens non-liquefaciens (Ei- senberg). fulvus (Zimmerman). " janthinus (Zopf). lactis serogenes (Escherich). " mesentericus ruber (Globig). WATER PURIFICATION AT LOUISVILLE. Bacillus nebulosus (Wright). prodigiosus (Ehrenberg). proteus vulgaris (Hauser). " radiatus aquatilis (Zimmerman). " ramosus (Frankland). " rubidus (Eisenberg). " subtilis (Ehrenbcrg). " venenosus (Vaughan). " violaceus (Frankland). Cladothrix castrana (Cohn). dichotoma (Cohn). Micrococcus aquatilis (Bolton). cremoides (Zimmerman). Proteus fluorescens (Jager). Sarcina lutea (Schroeter). Saccharomyces cerevisoa (Mayen). Rosa Hefe. COMPOSITION OF OHIO RIVER WATER. 39 Quantitative Bacterial Analyses of the River Water. The average results for each day of the determination of the numbers of bacteria in the river water are recorded in the following table: AVERAGE RESULTS OF DAILY DETERMINATIONS OF THE NUMBER OF BACTERIA PER CUBIC CENTIMETER IN THE OHIO RIVER WATER. Date. October. November. December. January. February. March. April. May. Jun. 1 8 800 6 800 8 600 126 6 228 8 l87 637 6600 888 3 600 800 1 80 16 I 6 600 18 1 66 184 4 Soo d8 D 58 8 ooo 18 8 i(x> 28 ^8 06 12 (XX) 81 * ^uo _ Date. July. August. December. February. March. April. May. jun. July. -4 5 g 8 700 " 8 300 16 t Soo 6fio 18 1 1 800 46 8OO 6 loo 26 28 8 600 4 WATER PURIFICATION AT LOUISVILLE.. CHAPTER II. DESCRIPTION OF THE APPLICATION OF CHEMICALS TO THE WATER BY THE SEVERAL SYSTEMS OF PURIFICATION. OHIO RIVER WITH the systems of purification examined during the first portion of these tests, the ap plication of chemicals is a matter of funda mental importance for two reasons: 1. Chemicals are absolutely necessary un der normal conditions for successful filtration of water through sand at the rapid rate em ployed in American filters. 2. The application of chemicals to facilitate the subsidence of suspended matters, in such muddy water as that of the Ohio River, and to insure efficient filtration, makes their use the principal item in the cost of purification of this water by these systems. The topics which are considered in this chapter are as follows: Kinds of chemicals used. Composition of chemicals. Effect of the application of alum, or sul phate of alumina, to the Ohio River j water. Devices used by the respective systems for the application of chemicals to the Ohio River water. Uniformity in the rate of application of chemicals by the respective devices. Strength of solutions of chemicals applied to the river water by the respective sys tems. Average daily amounts, in grains per gal lon, of sulphate of alumina applied to the river water by the respective systems. This chapter deals with the problem as it stood on Aug. i, 1896. KINDS OF CHEMICALS USED. During these tests (1895-96) three kinds of chemicals were used: i. Sulphate of alumina (trade name, " basic sulphate of alumina "). 2. Potash alum. 3. Lime. Electrolytically formed chlorine and scrap- iron were also used in an experimental way with the Jewell System for a few hours on each of several days. Sulphate of alumina, of different lots and brands, was used regularly in the Warren and Jewell systems, except for a few hours on Feb. TO in the case of the Warren System, when potash alum was employed. During the time when the Western Company made use of their first device for the application of chemicals, potash alum was used instead of sulphate of alumina, because the former was less soluble in water, and therefore more adaptable under the circumstances, as will be shown beyond. Potash alum was used by the Western Sys tem up to May 20, and from June 4 to 6, inclusive. Sulphate of alumina was used dur ing the remainder of the test. During a greater part of the time from Feb. 8 to April i , inclusive, lime was added to the river water with the sulphate of alu mina in the case of the Jewell System. The object of this, apparently, was to improve the effect of the application of the sulphate of alumina, and to guard against the passage of the latter through the system into the filtered water. Sulphate of alumina, known commercially as basic sulphate of alumina, and potash alum are approximately of equal cost. The former contains no potash, less sulphuric acid and water of crystallization, but more alumina, as is shown in the table of analyses in the next section. It is the available (soluble in water) alumina in these two chemicals which give to them their efficiency in connection with the purification of such water as that of the Ohio APPLICATION OF CHEMICALS TO THE OHIO RIVER WATER. River. For this reason sulphate of alumina is better and more economical to employ for this purpose. COMPOSITION OF CHEMICALS. The average results of duplicate analyses of potash alum crystals used in the Western System, as stated above, are presented in the following table. For the purpose of com parison the theoretical percentage composi tion of pure potash alum is also given. PERCENTAGE COMPOSITION OF POTASH . ALUM USED IN THE WESTERN SYSTEM. Source of Sample. Alum Used in Western System. Pure Alum (theoretical). Matter insoluble in water Available alumina (AljOj). . . . Sulphuric acid (SO 3 ) O.O2 10.72 34.06 45.69 IO.OO o o.oo 10.77 33-76 45-54 9-93 o Water(HjO) Potash (KjO) Lime (CaO) Oxide of iron (FeiOj) These results show that the potash alum used in the Western System was absolutely pure, practically speaking. In the next table are presented results of analyses of the sulphate of alumina used in the several systems. In the Warren System use was made of one brand obtained in three principal lots. For the most part in the Jewell System use was made of one brand, but a different one from that of the Warren Sys tem, and also obtained in three principal lots. The second brand (lot No. 4) was used in the Jewell System alternately with the regular brand from June 20 to 30, inclusive, and from July 6 to ii, inclusive. With the Western System use was made of several lots, for the most part of the same brand as that employed in the Warren System. In order to compare the composition of these commercial products with that of the theoretical sulphate of alumina, the percent age composition of the latter is given in the following table: PERCENTAGE COMPOSITION OF THE SEVERAL LOTS OF SULPHATE. OF ALUMINA. System. Number of Lot. Matter Insoluble in Water. Available Alumina (Al,0,). Sulphuric Acid (SO 3 ). Water (H a (J). Timrirnm Oxide of Iron (KejO,). I 2 3 1 2 3 4 2 o.oO 0. IO O.O2 1.9S 0.63 0.4" 2. 17 0.30 0.00 17.88 17.90 17.86, 16.39 16.19 if). 12- lS.62 17.20 "5-32 39-87 38.61 37-72 37- 9 6 37-87 37--4S 42.20 37-64 36.04 41 .22 42-75 44.62 43.46 45-28 46.08 36 . 90 44-92 48.64 o.oo Trace 0.08 Trace Trace Trace 0.02 O.OO 0.43 0.32 o.oo 0.20 0.1X3 O.OO "34 0.24 o.oo ,. lewell ,. Western Sulphate of alumina (theoretical).. In connection with the above tables it is to be noted that each lot of commercial sulphate of alumina contained considerably more available alumina than the theoretical sulphate of alumina. It is this portion (avail able alumina) of these compounds that gives to them their efficiency for this particular pur pose; and, on an average, these commercial products contained about 60 per cent, more available alumina than the pure potash-alum crystals, analyses of which are presented in a foregoing table. Some of the sulphate of alumina used in the Jewell System contained more alumina than is indicated by the above results. But as it was insoluble in water it was of no value, and is recorded with other matters as " matter insoluble in water." These lots of sulphate of alumina differed in part from the theoretical sulphate of alumina, in that they contained less water of crystalliza tion owing to the process of their prepara tion. This fact alone caused the former to be richer in available alumina than the latter. The increase in the available alumina in the commercial products above that in the theo retical sulphate of alumina was also greater than the corresponding increase in sulphuric acid. This point doubtless explains the origin of the trade name, basic s.ulphate of alumina. The ratio of the alumina (A1 2 O 3 ) to the acid (SO 3 ) in each lot is shown in the following table, with the corresponding ratio in the theoretical sulphate of alumina taken as one: 4- WATER PURIFICATION AT LOUISVILLE. System. Lot. Warren . . i . . 2 Ratio. 1.05 1.09 Western . 2 1.07 System. Lot. Ratio. Jewell. . . i i. 02 " ... 2 I.OI ... 3 i.oi " ... 4 1.04 In all comparisons and tabulations throughout this report the chemicals applied to the river water for the purpose of coagula tion and sedimentation are expressed in terms of sulphate of alumina. Wherever potash alum was used the results are converted into their respective equivalents of sulphate of alumina, on the basis that 10 parts of the lat ter are equal, from a practical point of view, to 1 6 parts of potash alum. EFFECT OF THE APPLICATION OF ALUM OR SULPHATE OF ALUMINA TO THE OHIO RIVER WATER. When alum or sulphate of alumina is ap plied to the Ohio River water, it is decom posed for the most part by the lime (calcium carbonate and bicarbonate) dissolved in the water, and there is formed a white, gelatinous precipitate of aluminum hydrate. The mag nesium carbonate and bicarbonate in the water also decompose the alum in the same manner as does the lime. Carbonic acid gas is liberated by the decomposition of the alum, but remains dissolved in the water as- free acid. The lime and magnesia which combine with and decompose the alum pass into the form of soluble neutral sulphates. The tiny, sticky particles of aluminum hydrate group themselves together; and around the infinite number of centers of coagulation are gathered together more or less completely the matters suspended in the water, including the bac teria, thereby forming flakes of greater or less size and weight. Neither before or after its decomposition has the alum or sulphate of alumina, in amounts which would be permis sible in the purification of municipal water supplies, a germicidal effect on the bacteria, but simply aids in their removal, through sub sidence and filtration, by their envelopment in the gelatinous flakes. Several factors ex erted a marked influence upon the coagula tion, upon the subsequent sedimentation, and finally upon the effect of the remaining co- agula in the water as it was filtered at a rapid rate through the sand. The application of chemicals to the Ohio River water where this system of purification is employed is of fundamental importance, in fluencing both the efficiency and cost of the system, and the whole subject in its different phases will be discussed in detail beyond. At this time it is the purpose simply to point to the matter in a very general way, in order to make plainer and to bring out the signifi cance of the following account of the devices used in the initial step in the process of puri fication. DEVICES USED IN THE RESPECTIVE SYSTEMS FOR THE APPLICATION OF CHEMICALS TO THE OHIO RIVER WATER. In this section is given a brief description of the principal features of the several devices for the application of solutions of alum and other chemicals. An account of the effi ciency with which this was accomplished will be found in subsequent sections of this chap ter. In the following chapters are described the decomposition and subsequent disposi tion of the applied alum or sulphate of alu mina, and also the effect which this application produced in connection with the purification of water. Warren Device. Sulphate of alumina was applied to the water in the Warren System just after the river water entered the settling basin. The current of water in the inlet water- pipe revolved a small propeller wheel lo cated in the mouth of the pipe. This wheel turned, by means of two sets of beveled gears, a specially designed pump, working in a pump box on the floor above. This pump box received the sulphate of alumina solution from the mixing tanks, the flow from which was regulated by a float valve. From the pump the solution of sulphate of alumina passed through a lead pipe discharging by gravity into the water in the settling basin, opposite the center and approximately 5 inches from the end of the inlet water-pipe. Two white-pine mixing tanks were located on the floor over the settling basin and ad- APPLICATION OF CHEMICALS TO THE OHIO RIVER WATER. 43 jacent to the pump box. These tanks were used alternately, solutions being made up in one tank while the other was in service. Fil tered water pumped from the filtered-water reservoir was used for dissolving the sulphate of alumina; and stirring was done in all cases by hand. The depth of the tanks was 4.5 feet and the diameter about 4 feet. Owing to unsatisfactory working of the meter on the pipe through which the chemicals entered the water, and its final abandonment, glass gauges were employed for measuring the quantity of solution used. Calibrations of the tanks showed an average capacity for each o.i foot in depth of 1.24 and 1.18 cubic feet in tanks A and B, respectively. Owing to the distance which the outlet pipe was above the floor of the tanks, the lower portion of the so lution in the tanks could not be used. The quantity of solution left in the tanks each time a new solution was prepared varied con siderably, but generally amounted to some thing less than 3 cubic feet. Pump Box. The solution of sulphate of alumina flowed into the pump box from the mixing tanks, the flow being regulated by a 2-inch float valve of vulcanized rubber. The pump box was about 2.9 feet long by 1.2 feet wide. The depth of solution in the pump box was capable of regulation by varying the distance of the float from the center of the valve, the float arm being adjustable in length. This was intended as a means of varying the rate of application of solution by increasing or decreasing the depth of immer sion of the pump arms. In practice it was not found successful, as the float valve was too irregular to make such an adjustment feasible. The maximum and minimum depths of solution, at the overflow of the pump box and the lowest level at which the device op erated, were 1 1 and 5 inches, respectively. Propeller. The propeller wheel was a small screw wheel of about 0.5 foot outside diameter, set on a horizontal shaft directly in the mouth of the inlet water-pipe. When the Warren System was first installed a 5-blade wheel was tried, but this was taken out and a 7-blade wheel substituted Nov. 25, 1895. This wheel was made of cast brass. 0.5 foot in diameter by 0.2 foot deep. There were seven blades, each pitched so that the circumferential distance between their edges was 2 inches. Connections from this pro peller wheel to the shaft of the chemical pump above were made by two sets of small beveled gears and a vertical shaft. Pump. The pump was a patented device constructed of vulcanized rubber. It was made up of six hollow curved arms, each of which lay in a plane perpendicular to the horizontal shaft on which the pump revolved, and which were connected respectively to six tubes placed parallel to the same shaft. The shape of the curved arms was approximately that of two straight pipes 5 and 3 inches long, respectively, making an angle of 45 and con nected by a circular curve of 3 inches radius. The inside diameter of these arms and of the horizontal tubes was 0.5 inch. In operation the pump was revolved by the propeller wheel just described. The shaft of the pump was located upon the top of the pump box, the solution filling the pump box ordinarily to from I to 3 inches below the pump shaft. As the pump revolved, each arm was filled as it entered the solution, and as the end was the first part to leave the solution it trapped some of it, the amount varying with the height of the solution in the box. As the pump turned the liquid was dropped back into the arm and emptied out of the horizontal tube into a funnel at the side of the box. To this funnel was connected a lead pipe through which the solution flowed into the settling basin, discharging opposite the mouth of the inlet water-pipe. Manner of Control of the Application of Chemicals. When the flow of river water. into the settling chamber was not too low (above 18 cubic feet per minute) the discharge of the solution from the pump was supposed to be proportional to the admission of river water. The amount of solution pumped was changed by the removal or insertion of rub ber stoppers into the ends of the hollow arms of the pump. In some instances half-stop pers (halved lengthwise) were inserted. Marked changes in the application of alum called for new solutions of different strength. Elevations. The relative elevations in feet of the more important points, referred to the bottom of the sand layer of the filter as the datum plane, were as follows: 44 IVATER PURIFICATION AT LOUISVILLE. Maximum flow line in mixing tanks.. 14.80 Minimum flow line in mixing tanks. . . . 10.66 Center of chemical pump 10.60 Maximum flow line in pump box 10.58 Center of discharge in the settling basin . 1.03 Jeivell Device. The solution of sulphate of alumina was pumped into the inlet water-pipe against a pressure of about 60 pounds at a point about lofeet inside the settling chamber. Before the entrance to the settling chamber of the river water containing the solution it passed through a meter and two valves on the main inlet pipe. Mixing Tanks. Sulphate of alumina solu tions were prepared alternately in two cypress tanks 5.5 feet deep and 3.5 feet in average diameter. Filtered water, taken from the outlet pipe just as it left the filter, was used for dissolving the commercial product after it had been broken into small pieces. This was facilitated at times by heating the water with steam which was allowed to enter the water-pipe just before it reached the tanks. Pump. From *he tanks the solution was pumped into the inlet pipe by a 3.5 by 4.5 by 6-inch single pump, the suction pipes of which reached to within about i inch of the bottom of the tanks. The ends of the suction pipes were capped with screens. The steam supplied to the pump was kept at practically a constan . pressure by means of a regulating valve. Feed Pipe. The feed pipe from the pump to the inlet pipe was a heavy lead pipe 0.75 inch in diameter. At first all fittings were of wrought iron, but owing to corrosion by the sulphate of alumina they repeatedly broke, and at the close of the test practically all fit tings were of brass. Manner of Control. At the outset it was the custom to start the pump at a speed to deliver the desired quantity of solution and keep it under general control by means of a float on the water above the sand layer. This float was connected by a chain and pulley with a valve regulating the flow of steam to the pump. This was not a success, and the float was abandoned during the latter part of March, and the application controlled by fre quent regulations of the steam-valve, inspec tions of quantities from the meters and of the speed of the pump being used as guides. Ordinarily, changes in the speed of the pump would allow the desired arrangement in the application of sulphate of alumina, but in ex treme cases the strength of the solution was altered. Elevations. The relative elevations in feet of the more important points, with the bot tom of the sand layer of the filter as the datum plane, were as follows: Maximum flow line in mixing tanks. - 6.90 Minimum flow line in mixing tanks. - 12.06 Center of discharge into inlet pipe. . - 11.13 Western Systems. As only one settling chamber was used for both the Western Pressure and Western Gravity systems a single device for the appli cation of alum was sufficient. Two separate and distinct devices, however, were used. The first was used from the beginning of the operation of these systems up to April 7, and the second, following an extended period of modifications and repairs, was in service from May 7 till the close of the tests. First Western Device. On the main inlet water-pipe to the set tling chamber there was a 6-inch gate valve which caused a difference in pres sure above and below it. From above this valve a o.5-inch brass pipe led to the alum tank, which was a cast-iron vertical cylinder 1 foot in inside diameter and approximately 2 feet deep. The alum tank had a top open ing with a cover constructed like a hand-hole fitting. The diameter of the opening was 4.5 inches. The brass pipe above mentioned connected with the alum tank at the top and extended into it about I foot. From the top of the alum tank a second o. 5-inch brass pipe led to the inlet pipe below the valve above men tioned. Suitable valves cut off the flow in both brass pipes; allowed access to the alum tank; and aided in controlling the flow of alum solution. A mercury column in a celluloid tube was used to indicate the difference in pressure in the two brass pipes. By this ar rangement the alum solution was applied to the river water in the inlet pipe about 10 feet from the settling chamber. Operation. In use the alum tank was kept filled to a greater or less depth with crystals APPLICATION OF CHEMICALS TO THE OHIO RIVER WATER. 45 of potash alum put in through the hand-hole at the top. Differences in pressure in the inlet water-pipe before and after this by-pass caused the flow of a small quantity of water through the tank whereby the alum was dis solved and carried over into the inlet water- pipe. The only means of regulating the quantity of alum solution applied was by dif ferences in the pressure of the water flowing through this by-pass. Second Western Dei-ice. The entire device for the application of chemicals was changed during April, the final arrangement being as follows: The separate pipe for the admission of river water to the system, which was introduced Feb. 29, was broken and a duplex pumping engine inserted on the pipe line. To this du plex pump were attached auxiliary pumps by which the solution of chemicals was forced into the main inlet water-pipe beyond the pumps and about 30 feet from the settling chamber. The duplex water pump forced the river water, containing the alum or sul phate of alumina solution, through the set tling chamber, and also through the pressure filter, or into the top of the gravity filter. Mixing Tanks. Two pine tanks each 4 feet deep by 3 feet in diameter were used alter nately for the purpose of preparing 1 the solu tions. The solutions were made with filtered water taken from the outlet pipe near its exit from the pressure filter, or with river water when the pressure filter was not in operation. Afain Water Pumps. The main pumping engine was a Worthington single-expansion duplex engine. The principal dimensions were as follows: Diameter of steam cylinder 9 inches Diameter of water cylinder 8.5 Length of stroke 10 Steam was supplied by a i. 5-inch asbestos- covered pipe. The exhaust, a 2-inch pipe, was open to the atmosphere. Auxiliary Chemical Pumps. The device used for pumping the solution of alum or sul phate of alumina consisted of small plunger extensions of the main piston-rods on the water pumps above described. These worked in pockets in which they caused al ternate suction and pressure. The valve sys tem was a pair of cup valves in the same casting, one opening to allow flow from, and the other to allow flow into, the plunger pocket. These valves were located just out side of the plunger chamber. Piping System. From the mixing tanks a system of o.5-inch brass pipes led to the aux iliary pumps. The arrangement was such that either tank could be used, and either one or both of the auxiliary pumps op erated. From the pump the solution of chemicals was forced through a system of o.75-inch brass pipes into a glass tube. A 3- inch brass air-chamber in the system equal ized the flow. This glass tube was connected at the bot tom with a brass pipe which led into the in let water-pipe, discharging in the center of the inlet pipe through a tee set with its long arm with the current. The glass tube was also connected at the top with the top of a 6-inch air-chamber in the inlet water-pipe. A body of air was always maintained in this chamber, and there was a corresponding ver tical air column in the glass tube, as the level of the chamber and of the tube were the same. By this arrangement the alum or sulphate of alumina solution was discharged through an air column, thus making the flow of the solu tion plainly visible. Manner of Control. As the pumps dis charged a constant quantity, regulation of the application of chemicals was obtained by re lief pipes and valves through which the excess of solution was returned to the mixing tanks and pumped over again. Elevations. The relative elevations in feet, with the bottom of the sand layer of the pres sure filter as the datum plane, were as follows: Maximum flow level in mixing tanks. . .0.00 Minimum flow level in mixing tanks. -3.34 Center of pumps - 2.70 Center of discharge into inlet + 1.80 Jcivcll Device for Application of Lime. The device used for adding lime to the river water was modified a number of times during its use. At first it consisted simply of an ordinary barrel and suitable piping as described below. The barrel was located on the upper floor. Unslaked lime was put in the barrel and a stream of water let in at the 4 6 WATER PURIFICATION AT LOUISVILLE. bottom. The flow of water into the barrel was regulated by a float valve. Near the top of the barrel a pipe led to a connection with the suction pipe from the sulphate of alumina tanks. To the lower end of this pipe was at tached a glass cylinder in order to make visible the rate of flow. The mixed milk of lime and sulphate of alumina solutions were forced into the main inlet water-pipe by the same pump. At first the solution was stirred by hand, but later an aspirator was intro duced. When the use of lime was resumed on March 21 the milk of lime solution was pumped through a separate pipe into the inlet water-pipe just outside the settling chamber; and the entire lime system was independent of the sulphate of alumina sys tem. Elevations. The relative elevations in feet, with the bottom of the sand layer as the da tum plane, were as follows: Discharge level in the lime barrel + 13.00 (approx.) Center of modified dis charge into inlet pipe. . - 5.95 The device used for the application of iron consisted of a cast-iron tank, approximately j foot in diameter and 3 feet long, filled with scrap-iron. The piping was so arranged that the suction from the sulphate of alumina tanks, the suction from the lime barrel, and the force main from the sulphate of alumina pump could be connected with this tank. The iron could be thus introduced at any desired point in the flow of chemicals. It can hardly be said to have been in actual service, but was tried on Feb. 10 and 12, and again for 2.5 hours on Feb. 22. Its use was discontinued on the latter date on account of the very evident presence in the effluent of dissolved iron. Jewell Device for the Application of Chlorine. This device consisted of a set of small U- shaped tubes, in which a common salt solution was decomposed by an electric current. A constant flow of water was maintained through the tubes. The water dissolved the hypochlorites and carried them with it to the water in the top of the filter. The apparatus was never used regularly, but was tried on Jan. 21 and 22, and for very short periods at later dates. On Jan. 22 available chlorine was applied in this way during the morning at the rate of o. i part per million by weight of applied water. UNIFORMITY IN THE RATE OF APPLICATION OF CHEMICALS IN THE RESPECTIVE SYSTEMS. Owing to the marked and comparatively sudden variations in the quality of the river water, the rate of application of chemical so lution was varied by necessity from time to time. But with regard to uniformity when the quality of the river water was practically the same the observations revealed several points of an unsatisfactory nature. The amount of attention which was given to the devices for application of chemicals, further more, was found to be a very important fac tor in most instances. Speaking in general terms, the application of chemicals by the Warren device was fairly satisfactory. Its chief merit lay in the fact that it was automatic. It had a number of shortcomings, however. The rate of appli cation of chemicals during short periods was variable, due to varying heights of solution in the pump box. When the river water en tered the settling chamber at a rate of ess than about 18 cubic feet per minute the pro peller wheel could not be depended upon to operate the pump regularly. So far as was learned the propeller was reasonably uniform in its action when the flow was greater than that stated above. The manner of regulating the operating area of the open arms of the pump by means of rubber stoppers was crude, and under some circumstances would limit the serviceability of the device in purifying such a water as that of the Ohio River. The operations of the Jewell device showed c 1 early that its efficiency was closely depend ent upon the attention which it received. During the latter part of the test it received APPLICATION OF CHEMICALS TO THE OHIO RIVER WATER. 17 sufficient attention to make its operation satisfactory. At. times during the first part of the tests, however, the application of chem icals was very erratic. In some instances the rate of application of sulphate of alumina varied five or six hundred per cent, on the same day when the quality of the water was about the same. Complications arose when lime and sul phate of alumina were both applied by the same device in the Jewell System. At times it appeared that the two chemicals entered the water alternately. With the first Western device control of the application of alum was repeatedly lost, even with a laborer spending the greater part of his time watching it. The Western Com pany abandoned this device in April. The second Western device gave fairly satisfactory results, although it was necessary to give close attention both to the stuffing boxes between the water pumps and chemical pumps, and to the relief valves on the pipes through which the excess of solution was re turned to the mixing tanks. This feature of the application of chemicals is a very important one, both with regard to the cost and the efficiency of purification, and will be discussed in Chapter IX. STREN GTH OF SOLUTIONS OF CHEMICALS APPLIED TO THE RIVER WATER IN THE RESPECTIVE SYSTEMS. Samples of the alum and sulphate of alu mina solutions used in the several systems were collected at frequent but irregular inter vals for examination. The specific gravity of the samples was determined by the aid of a Sartorius balance, and these readings were converted into percentages of applied chemicals from tables of factors which were checked from time to time. In all 1632 de terminations were made of the strength of applied chemical solutions. In the next set of tables there are given the daily averages of the percentage strength of the solutions used in each system. Before considering these results there are several comments to be made. In the Jewell System the uniformity of strength of the sulphate of alumina solutions was very satisfactory, as a rule, during a greater part of the test. During the first portion, however, the variations in the strength of the solution were quite confus ing. For the most part the uniformity of strength of the sulphate of alumina solutions used in the Warren System was fairly satis factory. Small variations in the strength of even consecutive solutions, however, were repeatedly noted. This was due, apparently, to the complicated system of preparation of the solutions, which involved the considera tion of certain quantities of solution in each tank when its use was stopped to prepare new solutions, and to too much dependence upon hydrometer readings. With the first device used in the Western System the variations in the strength of the alum solutions were so great that this factor placed the whole system at a great disadvan tage at times, in spite of the comparatively close attention which it received. This device, which consisted of allowing a small stream of river water in the inlet pipe to flow through a by-pass in which was placed an iron cyclinder containing potash alum, showed marked weaknesses, among which were the following: 1. The solubility of potash alum crystals varied with the temperature of the river water in the alum tank. 2. The strength of the solution applied to the river water varied with the period of time that the river water remained in the alum tank; that is, it varied inversely with the rate of application of the alum solution. 3. The strength of the solution varied with the amount of potash alum crystals in the tank. 4. There were no ready means of knowing how much alum was being applied to the water; and in several instances the alum crys tals in the cylinder became exhausted, or very nearly so. On many days the strength of the solution was quite uniform, especially during the latter half of the period when this device was used, and when, at times, a small gas flame was placed beneath the inlet pipe leading to the alum tank to increase the temperature of the water. Similar results to the following were 4 8 W AT ER PURIFICATION AT LOUISVILLE. frequently obtained, however, during the first two months that the system was in operation. Date. Hour Percent age of Alum in Date. Hour. *K?"f Alum in 1896. Applied Solution. 1896. Applied Jan. 24. 10.28 A.M. 7.8 Feb. 20. 11.25 A.M. 3-9 "35 " 5.0 1.20 J .M. 5-7 11 12.32 I .M. 6.2 3- 30 3-3 " 1.30 " 3-2 4.00 4-6 14 3-"7 5-1 4-3 5-2 5.00 5-2 5-30 5-1 The following results, obtained on Feb. 26, are more representative of the last portion of the period when more care was taken to pre vent the exhaustion of alum from the tank. Percentage of Percentage of Hour. Alum in Applied Hour. Alum in Applied Solution. Solution. 9.3O A.M. 5-9 2.30 P.M. 6. 5 1 1 . OO " 4-9 3.00 6.4 12.30 P.M. 5-9 3-3 (>. 5 I . OO 6.1 4.00 5-7 1.30 " 6.2 4.30 6.6 2.00 " 5-9 5.OO 6-5 With the second device for the application of chemicals in the Western Systems, there came an improvement in the uniformity of the strength of solutions, but it was not thor- DAILY oughly satisfactory. This was due in part to mistakes in weighing out Jhe alum or sul phate of alumina, and in part to accidental dilution of the solutions after their prepara tion. An important factor in these varying strengths of solution, the effect of which is difficult to estimate accurately, was the flow of river water from the water pumps through the stuffing boxes into the pumps containing the solution of alum or sulphate of alumina. This was repeatedly observed and guarded against in part by frequent packing of the stuffing boxes. The maximum leak noted was about i gallon per hour for one of the two pumps. Under the conditions on that day, May 20, the dilution from this single pump amounted to about 10 per cent, of the full quantity of solution in one tank when full. With more nearly normal rates of flow of water and of alum solution this percentage would be somewhat less, but on the other hand the effect of the dilution from this source became cumulative, owing to the pas sage through the pumps of an excess of the solution, and its return through relief pipes leading back to the mixing tanks. The last portion of the solution from the tank was thereby more diluted than the first. AVERAGES OF THE PERCENTAGE STRENGTH OF THE SOLUTIONS OF SULPHATE OF ALUMINA APPLIED TO THE RIVER WATER IN THE WARREN SYSTEM. l>y. June. July. 2. SO JO 5 1 .25 I. 80 I. 80 18 2. 2O .60 2. IO 60 * 80 I .80 2 5 I- 06 I. 10 1-90 2.00 2.50 70 .00 2.OO 2.80 6O , .60 APPLICATION OF CHEMICALS TO THR OHIO RIVER WATER. f- DAILY AVERAGES OF THE PERCENTAGE STRENGTH OF THE SOLUTIONS OF SULPHATE OF ALUMINA APPLIED TO THE RIVER WATER IN THE JEWELL SYSTEM. Day. October. November. December. January. February. March. April. May. June. July. o 80 " 0.33 o 60 6 o 85 8 0.29 i 12 0.34 0.44 0.31 o 85 0.28 0.64 16 0.30 0.40 O.6o 18 o 60 0.30 0.88 o 88 21 0.24 0.30 0.88 O.6o O.2O 0,65 I 4O 0.30 0.26 0.80 0.58 fin 26 0.38 0.39 0.80 O.So 27 28 0-39 0.80 o 80 0.65 0-45 .20 O.6o 0.60 0.90 i. So ?n 0.95 0.37 i .60 0.70 I .20 DAILY AVERAGES OF THE PERCENTAGE STRENGTH OF THE SOLUTION OF ALUM AND OF SULPHATE OF ALUMINA APPLIED TO THE RIVER WATER IN THE WESTERN SYSTEMS. Day. January.* February. March. April. May. June. July. I 6.0 6.7 2 6 8 o 3 6.0 3 8 5 8 5 6 2 6 6 I 7 -j s 8 6 6 10 5.2 6 7 it 6 2 3 2 6 o 8 2 6 13 6 2 6 9 14 4-3 6.0 6 8 15 6.8 4.2 2.2 2.9 16 7.2 17 5.8 4-6 6.6 2.2 2.O 18 5.8 6 5 1.8 19 6 8 20 5 ( > 2.8 21 5.8 (, 3 22 5 - 2 o 6 23 5.6 o 6 24 5 -5 5.6 25 4.0 5-4 6.0 0.6 2. I 26 6 i 6 Q o 8 27 6.0 6 o 28 5.2 6.9 2g 6 8 30 2 6 2 6 31 5.8 7- 1 2.0 * The meter on the alum pipe was not attached till Jan. 10; and from Dec. 23 to that time ihe amount of alum applied to the water was computed from the weight of alum put into the alum tank. WATER PURIFICATION AT LOUISVILLE. AVERAGE DAILY AMOUNTS, IN GRAINS PER GALLON, OF SULPHATE OF ALUMINA APPLIED TO THE OHIO RlVER WATER IN THE RESPECTIVE SYSTEMS. The daily average results of the determina tion of the amounts of sulphate of alumina used in the respective systems (from October, 1895, to July, 1896, inclusive) are presented in the following- tables. In the case of the Western Systems, when both niters were in operation, the average results refer to the filters in common. As already explained alum was used during a large portion of the time in the Western Systems, and in all cases it is converted into equivalent amounts of sulphate of alumina. As a matter of convenience these results are expressed in grains per gallon. They may be changed to parts per million by mul tiplying by 17.1, and to pounds per million gallons by multiplying by 143. AVERAGE DAILY AMOUNTS, IN GRAINS PER GALLON, OF SULPHATE OF ALUMINA APPLIED TO THE OHIO RIVER WATER IN THE WARREN SYSTEM. Day. October. November. December. January. February. March. April. May. June. July. 0.88 6. 53 I.l6 4.05 3.82 o 78 5 85 5- 9 5.03 3 eg 5 89 i 66 4 18 I 85 3.27 6 o 68 3 86 1.83 3- 12 3.36 3 5 6 i 86 I 28 8 4 46 2.88 i. 08 3.87 1.89 1 . 59 4.78 2.73 2 85 i 88 4.00 3- 2 9 2.32 2.40 3.04 I 67 o 18 4 83 i 81 1.78 3.18 c 76 i 66 3 28 2.87 16 4 48 i 89 2.68 2.83 18 o 88 3 61 2.32 2.76 20 2.88 1.58 i .15 2.65 2.80 j 82 o 83 o 85 1.68 4.33 i 58 6 37 3.64 2. 15 6.45 i 36 o 84 7.41 0.66 2.38 7.40 26 T 6.1 A 6c i 28 28 3 60 1 58 o 87 5.16 4.84 3-95 4.42 4 08 5 7 8 4. 17 APPLICATION OF CHEMICALS TO THE OHIO RIVER WATER. 51 AVERAGE DAILY AMOUNTS, IN GRAINS PER GALLON. OF SULPHATE OF ALUMINA APPLIED TO THE OHIO RIVER WATER IN THE JEWELL SYSTEM. Day. October. November. December. January. February. March. April May. Tune. July. l 4 68 o 80 5.87 6 71 2 82 i . 13 4-55 5-94 6.33 4.16 1.62 4.31 " 64 I 69 5 06 6 3.56 0.98 4.05 1.65 4.35 4.67 o 88 i 18 I 64 8 2 68 i 6r 6.36 o 81 i .30 1.61 5. go 6.17 2 08 2.18 O qq i 61 i 62 o 58 .38 1.64 3. Si 6.36 i 38 o 85 i 61 15 2.8l .31 5.78 1 6 i 28 i 62 5 64 2 48 4 82 18 i 60 5 Si 3 08 2 85 o 86 i 68 i 83 o 89 .25 I .93 8.32 25 I 66 0.98 1.65 10.83 26 o 84 i 08 i 8-} 27 i 65 1.48 8 75 28 .36 3 38 6 95 29 2.06 4 ^3 6.24 7.58 3 0.41 4 3S 4.85 8.72 AVERAGE DAILY AMOUNTS, IN GRAINS PER GALLON, OF SULPHATE OF ALUMINA APPLIED TO THE OHIO RIVER WATER IN THE WESTERN SYSTEMS. Day. January. February. March. April. May. June. July. 4 38 4 87 o 63 2 83 2 64 6 0.82 3 16 4.28 S <>2 8 4.61 O.8O 6.32 o 66 3- 16 II o 62 5-77 1.48 5-S6 I 06 o 83 4-84 15 5- 10 16 i 66 4.18 S -36 18 I .24 0.84 O.5I i .52 3.66 20 i .24 2.78 0.83 5.39 3.70 o 87 22 1.58 3-29 4.46 23 3.83 7.08 24 0.81 1.36 4.56 3.64 8.31 25 I 63 5.20 26 3 7 8 27 i 58 r 16 2 87 5.38 28 2.60 5.36 29 6 18 6 98 6 71 11 3.83 2 WATER PURIFICATION AT LOUISVILLE. AMOUNTS OF LIME USED IN THE JEWELL SYSTEM. DAILY AVERAGES OF PERCENTAGE STRENGTHS OF LIME SOLUTION USED IN THE JEWELL SYSTEM. Date. 1896. Strength. Date. 1896. Strength. Date. 1896. Strength. Date. 1896. Street!,. February 8 0.52 February 18 0.55 February 27 0.49 March 7 0.44 " 10 o. 52 19 0-55 28 0.49 " 21 o. 19 " II 0.52 " 20 0.49 29 0-53 " 23 0.30 " 12 0.52 21 0.49 March 2 0.44 " 25 0.40 13 o-55 " 22 0.49 3 0.44 26 0.25 14 0.55 24 0.49 4 0.44 " 27 0.21 15 o-55 25 o. 52 " *, 0.44 31 0.25 i? 0-55 26 0.52 6 0.44 April i 0.30 The daily average results of the amounts of lime used in the Jewell System are pre sented in the following table: AVERAGE DAILY AMOUNTS, IN GRAINS PER GALLON, OF LIME APPLIED TO THE OHIO RIVER WATER IN THE JEWELL SYSTEM. Date. 1896. Amount. Date. 1896. Amount. Date. 1896. Amount. Date. 1896. Amount. February 8 0.85 February 18 1.05 February 27 0.32 March 7 0.23 " 10 I. 3 8 19 1.17 28 0.03 21 0.28 II 1.67 20 0.88 29 o.i 8 23 0.72 " 12 1.56 " 21 0.47 March 2 0.19 25 0.49 13 0.52 " 22 0.32 3 0.17 26 1.05 H 0.72 24 0.46 4 0.19 27 1.41 15 0.92 25 0.38 5 0.18 31 2.17 17 0.82 " 26 o.53 " " 0,3 April i o.S I DECOMPOSITION AND SUBSEQUENT DISPOSAL OF THE ALUM. 53 CHAPTER ill. DECOMPOSITION AND SUBSEQUENT DISPOSAL OF THE ALUM OR SULPHATE OF ALUMINA APPLIED TO THE OlIIO RlVER WATER. IN some localities objections have been raised to the use of alum and of sulphate of alumina in the purification of public water supplies. The ground for this has been that some of the dissolved chemicals passed through the filter, appeared in the filtered water, and were liable to injure the health of the water c6nsumers. While this might be true with some waters, it can be positively stated that it should not, and need not, be the case with the Ohio River water at Louisville. The reasons for this are that there is dissolved in the Ohio River water an ample supply of lime and magnesia to combine with, and to decompose, more sulphate of alumina than would be necessary to apply under conditions giving a satisfactory and economical purifi cation of the water by the general method under consideration. The lime and magnesia which are found in the river water, in a form capable of decom posing sulphate of alumina, are present as carbonates and bicarbonates. In the tables of chemical analyses of the river water in Chapter I the amounts of these constituents are recorded as alkalinity. These compounds which give to the water its alkalinity possess the power of decomposing sulphate of alumina and alum, by virtue of the fact that the sul phuric acid of the applied chemicals is much stronger than the carbonic acid of the alkaline compounds. The result is that the sulphuric acid (the strong acid) combines with the lime or magnesia (the strong bases); the alumina is thus disengaged and, in the presence of the water, forms aluminum hydrate, which soon appears in the form of a white gelatinous solid compound; and the carbonic acid (the weak acid) remains in the water as free acid. Tak ing sulphate of alumina, the more efficient of the two chemicals, as an example, and view ing the results of its application to the Ohio River water in the light of the expressions used in water analysis, we find that the follow ing principal changes occur: 1. The alkalinity is reduced by the displace ment of carbonic acid by sulphuric acid, which, combining with the lime and mag nesia, forms neutral sulphates. 2. The alumina is freed from sulphuric acid when the latter unites with the alkaline constituents, and appears as the gelatinous, solid aluminum hydrate, which possesses the power of coagulating the suspended matters in the water. In its solid form the aluminum hydrate is removed subsequently by sedimen tation and filtration. 3. The incrusting constituents arc in creased, due to the sulphuric acid uniting with the lime and magnesia. 4. The free carbonic acid is increased on account of the liberation of this acid when the alkalinity is reduced, with a resulting increase in the incrusting constituents. At the in stant of liberation it exists as a gas, but it im mediately takes up water and forms free car bonic acid. It is noted above that these are the principal changes. If the river contained no suspended matter or dissolved organic matter, there would be no other action; and these changes, furthermore, would be proportional to the amount and composition of the applied sul phate of alumina. In practice it is found that the particles of suspended matter, by an ab sorptive or mordanting action, dispose of some of the sulphate of alumina, without the above decomposition taking place. This secondary action will be mentioned be yond. Considering the primary decomposition (in a water free from suspended matters and 54 WATER PURIFICATION AT LOUISVILLE. dissolved organic matter), the reduction in alkalinity in parts per million, for i grain per gallon of each lot of the chemicals used in these tests, would be proportional to the sul phuric acid, as follows: REDUCTION OF ALKALINITY (LIME AND MAG NESIA) BY ONE GR<UN PER GALLON OF EACH LOT OF COMMERCIAL SULPHATE OF ALUMINA. System. Number of Percentage of Sulphuric Acid. Reduction in Alka inity. Parts per Million. Warren I 39-87 8-53 2 38.61 8.27 3 37-72 8.07 Jewell I 37.96 8.12 2 37-87 8. II 3 37-46 8.02 4 42. 2O 9.04 Western i 37-72 8.07 2 37.64 8.06 Under the above-mentioned conditions, the increase in incrusting constituents, expressed in the usual way in parts per million, would be exactly proportional to the reduction in alkalinity; and the increase in carbonic acid, expressed in parts per million by weight of carbon dioxide, would be 44 per cent, of the reduction in alkalinity. Independent of the absorptive and mor danting action with suspended matters and certain dissolved organic matters, the changes in the river water, upon the addition of I grain per gallon of the first lot of sulphate of alumina used with the Warren System, may be illustrated as follows: COMPARISON IN PARTS PER MILLION OF IMPORTANT CONSTITUENTS OF THE OHIO RIVER WATER BEFORE AND AFTER TREAT MENT WITH ONE GRAIN PER GALLON OF SULPHATE OF ALUMINA. Befor< Alkalinity Incrusting constituents... Carbonic acid (carbon dioxide) After Treatment. 51.47 (decrease) 33.53 (increase) 63.75 (increase) When suspended matter, especially clay, is present in the water, there is a certain amount of the sulphate of alumina absorbed without its decomposition by the alkaline constitu ents. This causes the alkalinity to be reduced in amounts less than that indicated above. The degree of reduction, furthermore, varies with the amount and character of the matter in suspension. The Ohio River contains so little dissolved organic matter that this factor probably does not cause the actual reduction in alkalinity to depart from the theoretical more than about 5 per cent. But the varying composition of the Ohio River water, with re gard to suspended matter, causes a variable relation to exist between the theoretical re duction in alkalinity, due to complete decom position of the applied sulphate, and the actual reduction after a portion of it has been absorbed by the suspended matter. The significance of this is shown in the following table, where the average results of several de terminations are presented, in which the amount of sulphate of alumina was sufficient to produce complete coagulation. PERCENTAGES WHICH THE ACTUAL REDUC TION OF ALKALINITY BY SULPHATE OF ALUMINA WERE OF THE THEORETICAL, WITH OHIO RIVER WATER CONTAINING AMOUNTS OF SUSPENDED D IFF ERF. NT MATTER. 200 400 vhich the Actual Further information upon this point was obtained in 1897, and the results are recorded in Chapter XV. Combining the above information con cerning the decomposition of sulphate of alu mina with the varying amounts of alkalinity in the river water, as recorded in Chapter I, it will be seen that the greatest amount of sulphate of alumina which can be safely ap plied to the ( )hio River water is about 4 grains per gallon for a minimum; the normal ranges from (> to 10 grains; and the maxi mum about 15 grains per gallon. From the outset of these investigations, the importance of determining accurately the presence or absence of the applied alum or sulphate of alumina in the filtered water was clearly recognized. There are two methods which can be utilized in the solution of this problem. These methods, both of which were carefully applied to samples of the filtered water day by day, are as follows: i. The determination of the alkalinity of the effluents. This proved, if the effluents DECOMPOSITION AND SUBSEQUENT DISPOSAL OF THE ALUM. 55 were alkaline, that no undecomposed sulphate of alumina was present. If the effluent was acid, however, the opposite of this was true. 2. The test for alumina in the effluent, ac cording to Richards logwood and acetic acid test. In the following table are presented all re sults in which any trace of alumina (A1 2 O 3 ) was found in the filtered water, together with the corresponding alkalinity or acidity of each sample. The amount of sulphate of alumina equivalent to each amount of alumina is also recorded. It will be noted, in those cases where the effluents were acid, that the amounts of alumina and sulphate of alumina were abnormally low. This may have been due to changes in the sulphate of alu mina in passing through the filter, or to in accuracies of the logwood test in measuring such small quantities of alumina, or both. SUMMARY OF ALL RESULTS SHOWING ALUMINA BY THE LOGWOOD AND ACETIC ACID TEST, WITH THE CORRESPONDING ALKALINITY IN THE EFFLUENT OF EACH SYSTEM. Date. .896. Alkalinity. Parts per Million. Parti per Mil lion. Sulphate of Grains per Gallon. Date. 1896. Alkalinity. Parts per Million. Pam per Million. Sulphate of Grains per rn. .11. JEWELL EFFLUENT. February 15 " 20 6.5 17.0 O.2 0. I 0.07 0.04 March 27 59-2 0. I 0.04 March 16 12.0 O. I 0.04 " 28 31-9 O. I 0.04 " 20 I3.I O.I 0.04 April 2 5-5 0.3 0.17 " 21 IO. I O. I 0.04 3 1-3 0.2 0.07 23 4-9 O. I 0.04 4 Acidity 4.7 0.7 0.25 24 4.0 O. I 0.04 6 4.0 0.2 0.07 April 2 0.2 0.5 0.18 June 2 28.9 O. I 0.04 3 Acidity o.i 0.3 O. I I 3 39-7 0. I 0.04 6 Acidity 3.1 0.8 0.27 10 22.5 0.3 O.II " 7 3-5 0.5 o.iS 16 24.0 O.I 0.04 July 3 10.2 o. r 0.04 July I Acidity 1.8 o-3 O. 1 I 2 Acidity 3.9 0.4 0.13 WKSTKK.N r.KAVlTY EM- LU K.VI . 3 4.6 o.S 0.27 July 2 ; 6.1 O.2 0.07 8 8.8 O. I 0.04 " 3 6.0 O. I 0.04 13 4-1 0.3 O. II M Acidity i.o 0.3 O.II WESTERN PRKSSURK EFFLUENT. 15 2-5 O.2 0.07 April 3 3.0 O.2 0.07 17 IO.O 0.4 0.13 " 4 Acidity i.o 0.4 0.13 20 15.2 O.2 0.07 " 6 5.0 0.2 0.07 21 17.1 O.I 0.04 22 14.0 O. I 0.04 JEWELL EFFLUENT. 23 8.9 o.i 0.04 February 10 43 -o O.2 0.07 24 Acidity 2.1 0.3 o.i I " 20 14.9 0.2 0.07 25 Acidity 6.0 0.3 o.ii March 24 15-4 O. I 0.04 27 Acioity 3.8 0.3 o.ii 26 51.1 0. 1 0.04 28 Ai-i iny l.i) 0.2 0.07 The above results show that on 2, 8, o, and i days, respectively, the Warren, Jewell, Western Gravity, and Western Pressure effluents were acid, and contained, therefore, undecomposed sulphate of alumina. The acidity on these several occasions was due, of course, to the application of sulphate of alumina in amounts exceeding that capable of decomposition by the alkaline constituents of the river water. On a practical basis of operation, such applications would be inex cusable. It will also be noted from the results in the last table that on 10, 22, 2, and 2 days, respec tively, the Warren, Jewell, Western Gravity, and Western Pressure effluents contained slight traces of alumina, although the efflu ents were alkaline. In a majority of these cases the effluents were sufficiently alkaline to decompose several grains of sulphate of alumina, and the passage of the latter through the filter in an undecomposed form, under these conditions, was an impossibility. So far as we could learn, these slight traces of alu mina in alkaline effluents were due to the passage through the filter of minute flakes of the insoluble aluminum hydrate, and to their subsequent solution by the reagents used in the logwood and acetic acid test. So far as their direct and inherent influence is con- WATER PURIFICATION AT LOUISVILLE. cerned, these slight traces of alumina in an alkaline effluent cannot be regarded as objec tionable. In conclusion it may be stated that the experience obtained during these tests shows clearly that the Ohio River water contains a sufficient amount of alkaline compounds to decompose adequate quantities of sulphate of alumina; that the alumina appears as a solid gelatinous body, which coagulates the mud silt and clay, and subsequently is completely removed, practically speaking, by sedimenta tion and nitration; and that the sulphuric acid combines with lime and magnesia to form neutral sulphates of those bases, while an equivalent amount of carbonic acid is formed, and remains dissolved in the water. In a very few instances very slight amounts of unde- composed sulphate of alumina were found in the effluent of these systems. This was due to faults of construction of the systems, and of their operation, which must be improved as explained in Chapters IX and XV. Under the conditions of efficient and eco nomical purification of this water, the pres ence of undecomposed sulphate of alumina in the filtered water would be not only inadmis sible, but inexcusable. COAGULATION AND SEDIMKNTA TION BY ALUM1N UM HYDRATE. 57 (II XI TKK [V. COAGULATION AND SEDIMENTATION OF OHIO RIVER WATER HY ALUMINUM HYDRATE FORMED BY THE DECOMPOSITION OF THE APPLIED ALUM OR SULPHATE OF ALUMINA. IT has been already shown in the preced ing chapters how the geiaunous precipitate of aluminum hydrate is lormed, and reference has also been made to the disposal of the alumina in this solid form by subsidence and filtration through sand. In this chapter it is the purpose to explain the nature of coagula tion and sedimentation, to describe the de vices in the several systems where the coagu lation and sedimentation were accomplished, and to point out the practical results which may be obtained by the aid of aluminum hydrate in the purification of such water as that of the Ohio River. Coagulation is the term generally used to express the action which is produced by the application of alum to water. In general terms this action has al ready been described, but in detail it is as fol lows: When the applied alum solution comes in contact with the dissolved lime and magnesia in the river water, the former is immediately decomposed by the latter, which are present in excess. At the instant of the decomposi tion of the alum it forms aluminum hydrate. The latter is also dissolved in the water at the time when the decomposition or reaction takes place. The great bulk of aluminum hy drate passes very quickly into the form of solid matter. To chemists a solid compound, separating out by the action on each other of two soluble chemical compounds, is known as a precipitate. At first this precipitate of aluminum hydrate is present as innumerable, minute, white particles of a gelatinous and sticky nature; and it is not until the alum has been decomposed and converted into this form that its application in the purification, by this* system, of such water as that of the Ohio River begins to be of practical value, by its accomplishment of the initial step in the purification, viz.: COAGULATION. The process of coagulation consists of a gradual grouping together of the tiny parti- cies of aluminum hydrate which surround, and at the same time envelope, the mineral matter, organic matter, bacteria, and other organisms suspended in the river water. Aluminum hydrate (or, perhaps, sulphate of alumina) also combines with some of the dis solved organic matters, and adds them to the mass of coagulated material. Coagulation of muddy water by aluminum hydrate may be more easily understood by comparing it to the clarification of turbid coffee to which the white of eggs has been added. The completeness with which coagulation takes place depends upon the relation of sev eral factors. But it may be stated that, disre garding the question of cost, it is possible with sufficient coagulation followed by suf ficient sedimentation to render very muddy water perfectly clear. In actual practice the economical aspects must be considered, and coagulation should be carried to such a de gree, and under such conditions, that sedi mentation will remove the most mud and other suspended matter for the least money. For efficient and economical filtration through sand at a rapid rate," attention must be given also to the coagulated masses in the water as it reaches the filter. The degree of coagulation is influenced and controlled by several factors. Primarily it is controlled by the amount of alum which is applied and decomposed into aluminum hy drate. It is also influenced, among other . WATER PURIFICATION AT LOUISVILLE. things, to a marked degree by the amount of suspended matter in the river water, the relative character of the suspended particles, and the period during which coagulation and subsequent sedimentation may take place. The last factor is of great economical im portance. Coagulation by itself effects no purification in the sense that it removes from the water any objectionable matters. it is simply an initial and very important step in the purifi cation of waters heavily charged with sus pended matters, by which the way is paved for economical and efficient purification by means of sedimentation and filtration through sand. SEDIMENTATION. Sedimentation consists solely of the sub sidence due to gravity of the suspended mat ters after coagulation. It is a process of purification in that it removes from the water objectionable matters. It occurs in part simultaneously with coagulation in these sys tems, but the greater part of the sedimenta tion follows coagulation. In the following pages are described the portions of each system where the coagula tion and sedimentation took place. lieyond this are given some results showing the puri fication effected by the coagulation and sedi mentation in the respective systems, and an account of the relative value of the above- named factors, together with the results of some special experiments made with the view to demonstrating more clearly the economi cal significance of these portions of this method of purifying the Ohio River water. These descriptions will be more clearly un derstood by reference to the accompanying plates. WARREN SETTLING BASIN. The settling basin was rectangular in plan, 1 2. i feet by 12.0 feet, and 10.25 feet deep. It was constructed entirely of yellow pine. The bottom was made of planks 2.5 inches thick, and the sides for a distance of 5.1 feet above the bottom were 2.5 inches thick; above this height they were 5 inches thick. The basin was strongly braced inside and out by white-pine posts (6 by 8 inches), and the foundation was made of timbers of the same size. Iron rods 0.375 inch in diameter ex tended across the basin to stay it. Two par titions or baffle-walls divided the basin into three sections as shown on the plan. These walls did not extend entirely across the basin, but left an opening at -one end of 2.67 feet. These partitions were made of i-inch sheathing fastened to 1.75 by 6-inch posts. The general arrangement of the basin is shown on the plans, as are also the locations of the inlet and outlet water-pipes. Inlet Water-pipe. The river water entered the basin through a 6-inch pipe, connecting- just inside the basin wall with a 6-inch bal anced valve controlled by a float. There was from 45 to 65 pounds pressure on the inlet water-pipe, which branched from the force main leading to the Crescent Hill Reservoir. A small propeller-wheel located in a 6-inch nipple, which extended 5 inches from the valve, drove the chemical pump as previously described. Outlet. The outlet was a box channel, 3.4 by i.i feet in section. Its crest was 8.7 feet above the floor of the basin. From the basin outlet an 8-inch cast-iron pipe led to the filter, where it connected with an 8-inch pipe, which in turn connected with the central well. The passage of water from the basin to the filter was controlled by an 8-inch valve in this pipe. Elevations. The relative elevations in feet, with the bottom of the sand layer as the da tum plane, were as follows: Center of inlet water-pipe at basin. . . + 1.02 Floor of basin - 1.98 Top of basin +8.52 Average maximum water level + 8.02 Crest of outlet (mudsill) + 6.72 Center of pipe leading to the filter. . . . + 1.60 Depth. The depths of the chamber in feet were as follows: At level of mudsill 8.7 Average maximum water level 10.0 Total depth of chamber 10.5 Area. The areas of the chamber in square feet were as follows: COAGULATION AND SEDIMENTATION BY ALUMINUM HYDRATE. 59 At floor level 139-7 At level of mudsill H3-6 At average maximum water level 47-9 Capacity. The capacities of the chamber in cubic feet were as follows: Below level of mudsill 1229 Below average maximum level of water. 1422 Total capacity ! 459 These do not include the outlet channel, the contents of which were 33.9 cubic feet. Storage Period. Assuming complete dis placement, the length of time required for water to pass through the basin at the con tract rate (250,000 gallons per 24 hours) was 61 minutes. The distance from the inlet to the outlet along center lines was 36.6 feet. Concentrated solutions of common salt and of various aniline colors were added to the water on several occasions as it entered the basin, and careful observations made to learn the time taken for passage through the basin. It was found that the first water so charged passed through to the outlet at the contract rate of flow in about 15 minutes. The water containing the greatest amount of these solu tions passed through in 58 minutes, but the dilution was so great that it was just 2 hours before the last traces of the substances dis appeared from the water as it left the settling basin. In explanation of the short period which elapsed before the first appearance of the substances at the outlet it is to be stated that in the baffle-wall opposite the mouth of the main inlet water-pipe there was an open ing, i to 2 square inches in area, through which passed an iron rod 0.5 inch in diameter. Drainage. A 4-inch flap valve located in one corner of the settling basin was used as a sludge outlet. No provision was made to drain to this valve, the floor of the chamber being level. Cleaning. No special arrangements were made for cleaning. JEWELL SETTLING CHAMBER. The settling chamber together with the filter was included in one large circular wooden tank, 14.0 feet high and 13.5 feet out side diameter. The sides were made of 3 by g-inch cypress staves, and the bottom of two layers of 3-inch pme planks. The hoops were eleven in num ber; the first and sixth were 4.5 byo.iSinches, the second, third, fourth, and fifth were 4 by o. 12 inches, and the upper five were 3 by 0.12 inches. All of them were made of wrought iron. At a distance of 6.79 feet above the floor of the tank was a second floor 3 inches thick, which formed the lower floor of the filter tank. It was supported on eight 8 by 8-inch white-pine posts, four 8 by lo-inch timbers forming the floor-beams. The lower part of the tank was used as the settling chamber; the floor and sides of the tank forming the bottom and sides of the set tling chamber, respectively, while the bottom floor of the filter formed the top of the set tling chamber. The general dimensions as shown on the plan were: diameter, 13.0 feet, and height, 6.89 feet. Inlet Water-f>ife. This pipe was of wrought iron and was 5 inches in diameter. It was con nected to the side of the chamber by a flange joint. The river water contained in it was under from 45 to 65 pounds pressure, and was taken from the force main leading to the dis tributing reservoir. Inside the chamber there was a single- seated valve operated by a float in the filter above, and designed to control the flow into the settling chamber. Chamber Outlet. The outlet from the chamber was in the center, through an 8-inch central well passing no through the filter. Elevations. The elevations in feet, with the bottom of the sand laver as the datum plane, were as follows: Center of inlet water-pipe. . . -6.05 Floor of chamber - 7.61 Roof of chamber - 0.82 Height. The total height of the chamber was 6.79 feet, and the height under the beams supporting the floor of the filter was 5.96 feet. Area. The gross area of the settling cham ber was 132.7 square feet. The area deduct ing the supports for the filter floor was 129.2 square feet. Capacity.- The total capacity of the cham ber was about 879 cubic feet. WATER PURIFICATION AT LOUISVILLE. Storage Period. Assuming complete dis placement of the water, the length of time re quired for the water to pass from the inlet to the outlet pipe at the contract rate (250,- ooo gallons per 24 hours) was 36 minutes. The flow of water through the chamber was traced by the application of salt and various aniline colors in experiments similar to those described in the case of the Warren System. Under the above-stated contract rate of flow the water charged with these substances ap peared at the outlet in about 8 minutes after their application at the inlet; the period of passage of the water containing the greatest amount of these substances when it reached the outlet was 22 minutes; and the last ap preciable trace of these substances in the water as it left the outlet disappeared in 48 minutes after application at the inlet. Inspection. A manhole was provided at a convenient location to allow of inspection of the settling chamber. Drainage. An 8-inch valve was connected to the side of the chamber at the bottom by means of a flange joint. This valve discharged into a barrel connected with the sewer. The settling chamber could be drained completely through this valve provided its contents were quite liquid. The floor of the chamber was level, however, so that mud and slime had to be swept or shoveled out. Cleaning. It was intended to flush out the settling chamber by allowing waste wash- water to flow over into the central well from the filter above, and discharge into the set tling chamber. A curved half-pipe 4 inches in diameter, which was fastened to and turned with the agitator shaft, was used as a trough to direct the flow to different parts of the chamber. This did not prove effective, how ever, and the method of cleaning resorted to was by hand, aided by occasional flushings from the inlet water-pipe. WESTERN SETTLING CHAMBER. The Western Pressure Filter and the set tling chamber used by both filters were con tained in a large steel cylinder made of 0.62- inch plates. It was 22.5 feet long and 8.0 feet in inside diameter. The ends were dome- shaped, curving outwards 1.25 feet. This cylinder was divided in the center by two curved partitions. The partition plates touched at the center and were bolted to gether. The vertical joints were all lapped, and the horizontal ones were all butt-joints with two cover plates. Two lines of staggered rivets 0.75 inch in diameter were used throughout. The total weight of the empty cylinder was said to be 27,000 pounds. The west half of the cylinder was used for the settling chamber, the east half for the pressure filter. This chamber was 11.15 ^ eet long in the center, 8.71 feet long on the sides, and 8.0 feet in inside diameter. Supply of River Water. The supply for the Western Systems was at first furnished by the same pipe which supplied the Warren and Jewell systems with river water under from 45 to 65 pounds pressure from the force main to the Crescent Hill Reservoir. The varia tions in pressure were due to the variations in draft on the supply-pipe. Objections were made to this by the West ern Filter Company on account of the varia tions in pressure caused by the operation of the other systems. Therefore, on Feb. 29, 1896, a new 4-inch river-water pipe was laid from the force main, to be used solely by the Western Systems. After the change the pressure was held very closely between 60 and 65 pounds. Up to April 7 this pipe was connected di rectly with the settling chamber. Among the changes made during the period from April 7 to May 8 was the introduction of a Worthington pump on this pipe. This was done with the view to obtaining better equali zation of the pressure in the water-pipe, and also to operate a pair of auxiliary plunger pumps which were used for applying the chemical solution as already described in Chapter II. Pumping Engine. The pumping engine was of the the H. R. Worthington pattern. The main dimensions were g-inch steam- cylinder, 8.5-inch water cylinder and lo-inch stroke. It was a single-expansion duplex en gine. The steam was supplied by a i. 5-inch covered pipe. The exhaust was a 2-inch pipe, open to the atmosphere. Inlet Water-pipe. As first used the inlet to the settling chamber was a simple pipe with a COAGULATION AND SEDIMENTATION BY ALUMINUM HYDRATE. 61 flange joint screwed on to the bottom of the cylinder. With the other changes in April this was modified, and a distributing pipe was inserted in the chamber. This distributer was formed by a 6-inch nipple 12 inches long screwed into the upper side of the Mange joint above mentioned; a 6-inch tee with its long arm horizontal, and two lengths of 6-inch pipe each 2.5 feet long capped at the outer end. On each side of the nipple and pipes was a line of holes 1.5 inches in diameter. There were two holes in the tee, four in each nipple and one in the center of each cap. The center of the holes was 1.30 feet above the floor of the chamber. Outlet. The outlet from the settling-cham ber was a simple 6-inch pipe connected to the top of the cylinder by a flange joint. This pipe led down to the front of the cyl inder and joined a tee, to the ends of which were attached the pipes leading to the press ure and gravity filters, respectively. Elevations. The principal elevations in feet, with the bottom of the sand layer of the pressure filter as the datum plane, were as follows: Center of inlet pipe -0.80 Bottom of cylinder (inside) . . -2.15 Top of cylinder (inside) + 5.85 Capacity.- The capacity of the chamber was 503 cubic feet. Storage Period. Assuming complete dis placement in the chamber the storage interval at the contract rate (250,000 gallons per 24 hours) was 22 minutes. Experiments with salt and various aniline colors, similar to those described in connec tion with the other systems, were made to trace the flow of water through the settling chamber. At the contract rate of flow, as stated above, the water containing these ap plied substances first appeared at the outlet in about 2 minutes after their application; the water containing the largest proportion of these substances passed through the chamber in 9 minutes; and the last appreciable traces of the salt and dyes disappeared from the water at the outlet in 27 minutes after their application to the water at the inlet. With both filters in operation at the con tract rate, the storage period in this chamber would be only one-half as long as stated above. Inspection. The settling chamber could be inspected by removing a manhole placed about in the center of the upper front quad rant. A hand-hole was also provided, occupy ing a similar position in the lower quadrant. Drainage. The construction was such that there was no convenient method of draining the chamber. The inlet pipe at its lowest point was, however, provided with a tee, one arm of which was plugged. By removing this plug the water could be drained out to the level of the distributing pipe. As no arrange ments were made to carry off the water from this plug it was used but little. The usual method was by siphonage through the man hole in the upper part of the chamber. Cleaning. Handwork was mainly relied upon for cleaning. By a system of valves and piping, connection was formed between the wash-water and chamber outlet pipes so that wash-water could be turned in from above to aid in flushing the chamber. PURIFICATION OF THE OHIO RIVER WATER BY SEDIMENTATION IN THE SEVERAL SYSTEMS. As sedimentation is an intermediate step in the complete purification by this system of the Ohio River water, and as it varies widely ac cording to the existing conditions, this phase of the tests was not made the subject of de tailed daily study. Attention was given to the matter in a general way, however, with the view to learning its practical significance. Inspection showed very quickly that the de gree of purification of the Ohio River water by sedimentation was a variable factor so far as the removal of mud was concerned. With the same river water sedimentation increased with the amount of aluminum hydrate formed from the decomposed alum or sul phate of alumina. This would be naturally expected, of course, because the greater the number of minute gelatinous particles, forming centers of coagulation, the greater would be the size and weight of the coagulated masses or flakes; and, in turn, the greater and heavier 62 WATER PURIFICATION AT LOUISVILLE. these flakes the more quickly would they sub side by gravity to the bottom of the settling chambers. At times the Ohio River water had sus pended in it large quantities of very fine silt and clay, of which the individual particles sometimes ranged as small as o.ooooi inch in diameter. It was after heavy rains following a long period of drought that water of such a character was found. With the same amounts of aluminum hydrate in two samples of river water, one of the character just de scribed, and the other a more nearly normal water containing the same amount by weight of larger suspended matter, the latter water is far more purified by coagulation and sedi mentation in the same period of time than is the former water. With plain sedimentation, without coagulation, similar results would be obtained; and the explanation of the results just described is that the coagulation was quite incomplete. With a water containing an innumerable quantity of very finely divided particles, the period necessary for coagulation is unusually long; and it appears that, in some cases at least, the bulk of the aluminum hy drate together with the larger suspended par ticles subside before a large portion of the fine particles is coagulated. Another factor which produces a marked effect upon the degree of sedimentation is the period of time during which the coagulation and subsidence take place. The actual stor age periods under normal conditions for the respective systems have already been pre sented in an earlier portion of this chapter. These storage periods were complicated in a good many cases by washing, repairing, and modifying the filters and by delays occasioned by the filters being out of service during the night (except March 24-30 and April 27 to June 6) and on Sundays. During the six weeks continuous run (Sun days excepted), from April 27 to June 6, twenty-nine sets of bacterial analyses were made of the river water before treatment, and of the corresponding water after it has passed through, under normal conditions, the War ren settling basin and the Jewell settling chamber, respectively. At the outset the fa cilities for taking samples of the water after passage through the Western settling cham ber were not wholly satisfactory. During the latter part of the period (May 28 to June i) eight samples were taken from this system. The average results of these analyses are com pared with the corresponding ones from the other two systems just after the next table. In the next table are recorded the results of the individual analyses with the percentages of removal, in the full set of tests of the War ren and Jewell systems upon this point. The average quantities of sulphate of alumina ap plied by each system on the different days are given in Chapter II. It will be noted that these results, which are tabulated below, are quite variable with re gard to the percentages of removal. This was due in part to the amount of applied sulphate of alumina in relation to the quality of the river water; and also to the fact that there were in the water small flakes of coagulated NUMBERS OF BACTERI A PER CUBIC CENTI METRE IN THE OHIO RIVER WATER BEFORE AND AFTER PASSAGE THROUGH THE WARREN SETTLING BASIN AND THE JEWELL SETTLING CHAMBER. RESPECT IVELY, WITH THE PERCENTAGES OF REMOVAL. Dale. _ in Water fi m 1896. Warren Jrwdl River. Settling Settling Warren. Jewell. Basin. Chamber. April 28 5 70" 5 Soo 4 3< o 25 29 7 loo 2 800 2 8OO 60 60 3 3 700 2 300 1 700 38 54 May 2 5 600 I TOO 4 too 80 27 2 9 OOO 2 8OO 7400 69 18 4 7 5oo 2 9OO 4 700 6 1 37 <; 9 ooo 6 OOO 4900 33 46 6 4 900 I 800 4400 63 10 7 5 ooo 4 2(X> I 700 16 66 1 1 6 900 2 40O 3 500 65 49 12 7 100 I 7<w> 4 coo 76 44 13 4 200 i ooo 3 400 7f> 19 M ? Son I 500 ; 3 too 74 47 15 7 ?oo I 300 I 500 83 So if) 10900 I 800 2 400 83 78 18 9 5o I Soo 4 600 8 1 52 19 7 800 5 40" 3 400 26 53 20 4 700 4 400 2 3OO 06 ;i 21 5 900 2 700 4300 54 27 22 4 600 i 800 2 400 61 48 28 14 900. 6 800 II 900 54 20 29 33900 6800 21 2OO 80 37 29 23 600 I 400 1 5 900 94 33 30 28 7ooj 5 too 20 300 82 29 30 21 8ooi 5 100 15 200 77 30 Ji ne 3 18 900, 2 200 4 100 88 78 59 900! 2 800 4 100 72 59 5 6 200 2 400 3 500 61 44 5 5 ooo i 600 3 100 68 38 Averages 10500 3100 5 9o 61 43 COAGULATION AND SEDIMENTATION BY ALUMINUM HYDRATE. material, containing bacteria enveloped within and around them, and which were of necessity broken up in an incomplete and irregular manner as they were mixed with the culture medium for bacterial analysis. That is to say, it was practically impossible to get all the bacteria in these Makes separated into single cells so that each colony on the culture medium should represent only one bacterium, as the method of analysis called for. During the period from May 28 to June I, inclusive, when the river water contained the greatest amount of very finely divided par ticles, and when it was most difficult to coagu late, the average results of bacterial purifica tion by coagulation and sedimentation in the three systems were as follows: System. Number of Bacteria per Cubic Centimeter. Percentage Remoral. Warren 5 ooo So Jewell 1 6 900 31 Western 16 600 32 (River) 24600 On June 5 and 6 four samples of river water before and after passage through the Warren settling basin and the Jewell settling chamber, respectively, were collected under normal conditions and mixed together for chemical analysis. The results of the analyses showed that 59 and 18 per cent., respectively, of the suspended matter in the river water were removed in these two systems by coagu lation and sedimentation. These results show very forcibly the great economical importance of long storage peri ods in order to allow the coagulated material to subside, especially as the removal in the Warren System of more than three times that in the Jewell System was effected with only 65 per cent, of the sulphate of alumina em ployed in the latter system. The examinations of the removal of sus pended matter of the river water in the West ern settling chamber indicated that it was more variable than in the case of the other systems, but on an average about equal to that by the Jewell settling chamber. During June and July several sets of analy ses, both chemical and bacterial, were made of the river water before and after it had re mained over night or over Sundav in the sev eral respective settling chambers. The re sults of the analyses bore out the, observations as to the appearance of the water in the set tling chambers, showing in practically every case a removal of more than 90 per cent, of both bacteria and suspended mineral and organic matter, while in several instances the removal was more than 99 per cent. These last data show very conclusively the great economical importance of coagulation and sedimentation in the purification of such muddy water as that of the Ohio River. They also show the superiority in this re spect of the Warren over the other systems, owing to a longer storage period in the set tling chamber, during which sedimentation takes place. Furthermore, this evidence is abundant proof that in all these systems the storage period in the settling basin and cham bers is too short by far to allow full benefit and economy to be derived from sedimenta tion. Owing to the great practical importance of sedimentation, some special experiments were made for the purpose of obtaining more in formation on this subject, as will be found in the next section of this chapter. During 1897 additional experiments were made, and the re sults are recorded in Chapter XV. SPECIAL INVESTIGATIONS UPON THE DEGREE OF PURIFICATION OF THE OHIO RIVER WATER BY SEDIMENTATION UNDER VARYING CONDITIONS, BOTH WITH AND WITHOUT COAGULATION BY ALUMINUM HYDRATE, AND WITH SPECIAL REFER ENCE TO THE INFLUENCE OF THE PERIOD OF SUBSIDENCE. This set of experiments was made with the aid of a settling pipe, 20 inches in diameter and 24 feet deep, placed in the boiler house. Suitable piping arrangements were made to allow flushing, filling, and draining the pipe, and at the sides of the pipe was placed a series of pet cocks through which samples of water could be drawn at stated distances from the bottom. The results of these experiments are given in the following table. Except in those cases where the regular samples (numbers in paren theses) for daily analyses of the river water 6 4 WATER PURIFICATION AT LOUISVILLE. were used, special serial numbers were given to samples collected for this purpose. The distances from the bottom of the pipe to the tap from which the sample was taken are given under the heading, source of sample. Analyses were made for the determination of the total suspended solids (insoluble residue on evaporation), and also of the number of bacteria in the water. As explained above, the latter determination was complicated by the presence of masses of suspended matter in the water which made it difficult to sepa rate the individual bacteria. Another factor affecting the determination of the percentage removal of the bacteria was the high and un equal temperature of the water in the pipe at different heights and at different times during the same experiment. The temperature probably exerted a retarding influence upon subsidence, but, on the other hand, the gen eral conditions of sedimentation in a small tank are more favorable than in a large basin or reservoir. In those cases where no coagulants were applied it is probable that, under the conditions of practice with longer periods of subsidence, the variation in the amount of suspended matter in the water at different depths would be reduced materially. COAGULATION AND SEDIMENTATION BY ALUMINUM HYDRATE. 65 RESULTS OF SEDIMENTATION EXPERIMENTS. Experiment. Applied Sulphate of Grains per Gallon. Sample. Period of Subsidence. Hours. Tempera- Deg rees C. Suspended Solids. Bacteria. Number. Date. 1896. Number. Source. Parts per Million. Per Cent Removed. Per Cubic Per Cent Centimeter. Removed. I 2 3 4 5 6 7 8 9 10 ii May 29 June I June 2 June 4 June 6 June 10 June 10 June ii June II June 12 June 12 O.O 0.0 0.0 O.O 0.0 4.0 o.o 3-0 2.0 2.0 3.0 I 2 3 4 5 6 7 8 9 10 1 1 13 14 5 16 18 19 20 22 23 24 25 27 28 29 31 32 33 34 36 37 38 40 41 (626) 50 5i 52 53 54 57 58 59 60 61 62 63 64 66 67 69 70 71 72 73 74 75 77 78 79 80 (632) 81 83 84 86 87 88 River. 2 feet 4 11.75 " 20 2 " 4 11.75 " 20 River 2 feet 11.75 " 20 River 2 feet 11.75 " 20 2 " 11.75 " 20 River 2 feet 11.75 " 20 2 " 11.75 " 20 " River 2 feet 11.75 " 20 2 " 11.75 " 20 River 2 feet 6 11.75 " 20 2 " 2O " "75 " River 2 feet 11.75 " 20 " River 2 feet 20 " 2 " 2O " "-75 " River 2 feet 11.75 " 20 2 " 20 " ir.75 " 11.75 " 11.75 " River 2 feet 20 " 2 " 20 " River 2 feet O 24 24 24 24 48 48 48 48 O 24 24 24 O 24 24 24 48 48 48 24 24 24 48 48 48 O 24 24 24 48 48 48 O I I I I 3 J O 18 18 18 O I I 3 3 4-5 o i i i 3 3 6 12 16 o i i 3 3 i 590 32O 45-8 286 262 170 51-5 55-6 71.2 1 60 90 390 254 226 94 936 "24 72. S 84.7 34-9 42.1 50.2 22.6 580 5 4 504 338 22S 38.0 45.1 4 6.1 63-9 75.6 30.O 33-5 31.0 29.2 33-o 35-8 32.8 34-8 36.0 220 218 150 186 122 92 22O 31-3 31-9 53-1 41.9 61.9 71.2 I 4 6 130 118 114 88 80 413 17 19 20 9 9 10 "5 298 33-6 40.9 46.4 48.2 60.0 63.6 29.8 32.0 33-7 2 gOO 2 700 fOO 52.4 55-7 91 .8 95-9 95-4 95.1 95-4 97.8 97-7 95.1 I 6OO I OOO I 300 400 87.4 92. i 89.8 96.9 900 92.9 256 198 152 296 14.1 33-6 45.6 8 500 700 400 19 18 ii ii 9 252 93.6 93-9 96.3 96.3 97-0 91.8 95-3 18 14 ii 4 3 3 2 2 92.8 94-4 95.6 98.4 98.8 98.8 99.2 99.2 4 ioo 500 300 9 19 6 6 245 91.9 91.9 97-4 97-4 87.8 92-7 10 95-9 66 WATER PURIFICATION AT LOUISVILLE. RESULTS OF SEDIMENTATION EXPERIMENTS. Continued. Experi ment. Applied Sulphate of Sam pie. Period of Tempera- Suspende d Solids. Bact -ria. Date Grains per Hours. Degrees C. 1896. Gallon. Source. Mtllio P n. Removed. Removed. s 2 feet 78 o 06 80.3 8l 2 OS 88 84 6 2O " 89.6 6 89 6 800 I.O (642) River. 46 2 feet 96 6 6 98 6 June 16 0.5 (646) River. 8 800 2 feet 63 6 108 76 8 80 3 64 7 80 3 15 1 .0 River. 248 8 300 64 3 800 85 g 86 7 28 88 7 16 June 17 0.5 116 266 266 118 20 " 119 2 " 3 98 96 121 1 1 -75 " 18 81.6 86.5 17 June 18 1 .5 (651) River. 278 91 8 800 89 6 95.7 98.9 18 June 18 0.25 126 127 128 2 feet 20 " I I 299 o 12 2OO 20 8 Ig 0.25 (655) River. 132 2 feet I 84 80.0 3 800 84 80 o 63 8 134 2 " 3 20 June ig 0.75 136 River. 2 feet 138 20 " I 20 " 3 182 28 6 141 11.75 " 20 80 81.3 21 June 20 2.O (658) River 142 2 feet i 37 87.5 89.8 I 91 8 2 " 3 18 92.8 145 22 I .O 146 River. 25 8 M7 148 2 feet 20 2 " I i 28.O 35-5 4 6 48 82.9 82.2 3 500 2 500 63.9 74.2 98 8 23 June 20 O.O 151 261 152 2 feet 22 23. 7 153 H-75 " COAGULATION AND SEDIMENTATION^ BY ALUMINUM HYDRATE. RESULTS OF SEDIMENTATION EXPERIMENTS. Continued. 6 7 Experiment. Applied Sulphate of Grains per Gallon. Sample. Period of Subsidence. Hours. Tempera- Degrees C. Suspended Solids. Bacteria. Number. Date. ,896. Number. Source. Parts per Million. Per Cent Removed. Per Cubic Per Cent Removed. 23 24 25 26 27 28 29 30 31 32 33 34 35 June 20 June 23 June 23 June 24 June 24 June 25 June 26 June 27 June 29 July I July 2 July 6 July 6 o.o 3.0 1 .0 2.O 0-75 0.75 0-75 0.75 O.O 0.0 1.0 2.0 0.75 154 55 156 157 (660) 158 159 1 60 161 162 163 164 165 1 66 (668) 167 1 68 169 170 171 172 173 174 175 176 (678) 177 178 179 1 80 181 (681) 182 183 184 185 1 86 (684) 187 188 189 i go 191 192 (686) 193 194 195 196 (692) 202 203 204 205 206 207 208 209 (704) 210 211 212 213 2I 4 215 216 217 218 20 feet 2 " 11.75 " 20 " River 2 feet 20 " 2 " 20 " River 2 feet 20 " 2 " 20 " River 2 feet 20 " 2 " 20 " River 2 feet 20 " 2 " 20 " 11.75 " River 2 feet 20 " 2 " 20 11.75 " River 2 feet 20 " 2 " 20 " 11.75 " River 2 feet 20 " 2 " 20 " 11.75 " 11.75 " River 2 feet 20 " 2 " 20 " River 2 feel 20 " River 2 feel 20 " 2 " 2O " 11-75 " River 2 feet 20 " 2 " 20 " River 2 feel 20 " 2 " 20 " 22 48 48 48 I I 3 3 o i i 3 3 i i 3 3 i i 3 3 20 O I I 3 3 24 o i i 3 3 24 i i 3 3 30 48 3 3 24 24 o 24 24 o I I 3 3 17 o i i 3 3 o i i 3 3 37-2 129 125 67 48 235 31 25 9 5 218 42 39 32 32 446 54 41 22 22 267 52.6 52.1 74-3 81.6 3900 900 900 65.2 92.0 92.0 26.2 26.2 30.0 86.8 89-3 96.2 97-8 400 500 95-7 94-7 27.0 27.0 32.0 8OOO 4600 39 2 2OO 80.7 82 i 85.3 85.3 42.5 51-3 72.5 27.4 27.0 36.5 87.7 90.6 94.8 94.8 3 too 8 ooo 4400 4 200 13 Soo 39-2 13-7 17.6 I 4 8 M7 58 58 T 9 321 59.6 59-8 84.1 84.1 94.8 8 loo II 200 7 800 IO *00 2800 5 300 41.4 18.8 43-5 26.0 69.7 27-3 62.2 109 66.1 46 85.5 35 8 9-i 28 91.2 2 400 2 COO 3 Sou 500 6 ooo 2 400 2 300 700 I 400 400 7900 6 loo 7 700 4300 54-7 56.6 28.2 90.6 193 114 52 48 2 4 358 358 232 100 88 34-8 58.2 82 4 83-7 92.0 0. 35-2 72.1 75-3 60.0 61.6 88.3 76.6 93-3 26.5 27.0 36.0 22.8 2.6 45-6 18 55 493 465 276 193 856 435 33 6 636 1552 1385 i 248 796 43 244 40 30 12 8 395 209 199 "5 103 95-1 400 94-9 25-5 26.0 34-8 2.4 7-9 45-3 61.8 10 900 9 200 3 ooo 4 200 12.8 26. 5 76.0 66.4 26.5 29-3 38.2 23-5 25.8 31-3 49-2 60.7 8 400 4 200 5-2 15-3 23.7 51-3 73-8 26.3 26.0 34-5 6300 2 2OO I 3OO I IOO 600 Q500 7 100 7400 3600 4 loo 83.6 87.7 95.1 96.7 65.0 79-4 82.5 90.5 47.1 49-5 70.9 73-9 25-2 22.0 62.1 56.8 68 WATER PURIFICATION AT LOUISVILLE. RESULTS OF SEDIMENTATION EXPERIMENTS. Concluded. Exper men, Applied Sulphate of San pie. Period of Tempera- Suspend d Solids. Bact Number. Date. 1896. Grains per Number. Source. Hours. Degrees C. Parts per Million. Pei Cent Per Cubic Per Cent Removed. Si 8 16 Inlv i (707) 222 2 feet 3 26.5 438 188 16.5 26 i 3700 31-5 224 225 226 "75 " "75 " n. 75 " River 24 48 7 2 34-0 34-3 35-7 22g 77 153 087 56-3 66.2 70.8 3500 1 IOO 2 400 35-2 7Q.6 55-5 ^S 227 (T2t) II .75 feet River 24 34-6 21 45-7 I 300 6 Soo 84. 1 23O 231 2 feet 20 " 3 28 27-3 37-5 1 66 154 20.2 26.0 64 4 7 500 5 500 8.5 32.8 a6 34 6 6 1 ic 8 79 8 600 (73O) July 18 235 (744) 11.75 feet 24 35- S 28 637 93-o I COO gi .6 248 249 250 (767) 11.75 feet "75 " 11.75 " River 72 120 144 33-4 35-5 36.7 202 142 131 68.2 77-7 79-4 2 IOO 600 7OO 72.4 92.1 90.8 251 252 11.75 feet "75 " 24 144 32-9 3S-9 872 195 74-0 94-2 5 100 10 85-0 99-9 COAGULATION AND SEDIMENTATION BY ALUMINUM HYDRATE. 69 NUMBER OF BACTERIA PER CUBIC CENTI METER IN THE OHIO RIVER WATER AFTER PASSAGE THROUGH THE DIS TRIBUTING RESERVOIR AND A PORTION OF THE DISTRIBUTING PIPE OF THE LOUISVILLE WATER COMPANY. In the next table are recorded the results of bacterial analyses of tap water collected in the city of Louisville. In all cases the water was allowed to run from the faucet for some minutes before the sample was col lected. The place of collection was 419 West Chestnut Street up to Feb. i, 1896, and at 820 South Second Street for the remainder of the time. On its way to the city the river water is pumped to the Crescent Hill Reservoir, which has a capacity of 100 million gallons, equiva lent to about six times the average daily con sumption of the city. The chief value of these results is that they show a removal by subsidence and passage through the distributing pipes of about 80 per cent, of the bacteria originally present in the water as it was pumped from the river. In this connection reference may be made to the results of bacterial analyses of the river water already presented in Chapter I. NUMBER OF BACTERIA PER CUBIC CENTIMETER IN THE TAP-WATER OK THE CITY OF LOUISVILLE, BY DAYS, IN 1895-96. pn . ay. 128 I OOO i 600 116 I 2OO I IOO 800 I QOO 8OO 1 08 i 848 6 119 2 5OO 2900 6 700 2 OOO 6 900 I 5OO goo I 800 8 178 I SOO I ";oo 228 i 700 2 000 i 200 1 18 -no 118 4 3 13 104 I OOO IOO 400 3 5oo 6 200 2 OOO I 3OO 2 200 2 3OO 136 i 800 2 2OO I 2(X) 2 20O 2 IOO 1 6 i6j I OOO 800 3 800 6OO 2 6OO I OOO 18 368 800 QOO I 6OO 700 I Soo 20 321 nf> I 300 I 400 4 800 3300 800 4 ooo 9700 2 7OO 2 IOO I IOO 588 I OOO * ooo 288 I IOO 400 r 700 2 OOO 26 27 28 456 700 I OOO i 300 7 ioo I 300 1 800 2 OOO 3500 29 4 i8 800 10900 8 200 600 2 OOO 6o<> 1 600 i QOO 900 3900 2 2OO I 2OO WATER PURIFICATION AT LOUISVILLE. CHAPTER V. DESCRIPTION OF THE FILTERS THROUGH WHICH THE RIVER WATER PASSED AFTER COAGULATION BY ALUMINUM HYDRATE AND PARTIAL PURI FICATION BY SEDIMENTATION. IT has already been explained in the intro duction that this method of purification con sisted of several parts, each of which, to quite a degree, was distinct in its nature and in its application. The last part of the process is the passage of the water downward through a layer of sand either by gravity or pressure, in order to remove from the water the bac teria, aluminum hydrate, mud, clay, and other suspended matters. This final operation is properly called nitra tion. It is erroneous, however, to speak of the entire process as filtration, or mechanical filtration, because, so far as waters like the Ohio River are concerned, filtration is only one of several steps in a process of economical purification. This (American) type of filtration differs in several respects from the older (English) type of filtration which has been adopted and studied in Europe and several places in this country. There are two chief differences, namely: 1. In American filters the aluminum hy drate remaining in the water as it flows from the settling chamber to the filter, by virtue of its gelatinous nature, enveloping the bacteria and other suspended matters, makes it prac ticable to allow the water to pass through the sand at a much more rapid rate than in the case of English filters. 2. In American filters the accumulation of matters which are removed from the water by the sand (bacteria, aluminum hydrate, mud, clay, and other suspended matters) is in turn removed from the sand by the passage of a re verse current of water through the sand from the bottom to the top, either with or without accompanying agitation of the sand. The corresponding accumulations in English fil ters of foreign matter from the water, located for the most part at and near the surface of the sand, are removed, practically speaking, by scraping the surface of the sand layer with a shovel or similar implement to a depth ordi narily of 0.5 inch or thereabouts. By corresponding accumulations on the sand in the English filters is meant, ordina rily, the various kinds of material noted, with the exception of the aluminum hydrate; al though it is not to be forgotten that aluminum hydrate, formed from the decomposition of alum added to river water, was used for some years in I lolland in connection with the puri fication of public water supplies by English filters. The systems which were investigated dur ing these tests are included in the American type of filtration and are all divided into three main divisions. Each division includes the devices used for carrying out one step of the process, and it is to be noted that it was only in the design and construction of these de vices that these systems differed. These di visions may be outlined as follows: 1. An arrangement for the preparation and delivery of the chemicals. This included preparation tanks; pumps or other devices for delivering the solutions to the river water; pipes and fittings; valves and other regulat ing devices; scales, gauges, hydrometers, etc. 2. A chamber or basin for the reception of the treated water and in which coagulation and sedimentation took place to a greater or less degree. This included all the necessary inlet, outlet, and drain pipes, and the devices used for controlling the flow of water through the basin. DESCRIPTION Of FILTERS. 3. A filter and appurtenances. This division included a tank which contained the sand layer and water to be hitered; a system of strainers for removing the water trom the sand; a system for distributing the wash- water beneath the sand layer; and, in the case of the Warren and Jewell systems, a set of rakes with operating mechanism for stirring the sand. All piping, valves, and regulating- devices which pertained to the filter are in cluded in this division. in Chapter 11 the devices included in the first division have been described in consid erable detail. The second division has been presented in Chapter IV, and it now remains to present the third division, which is the sub ject of this chapter. In the next chapter will be found a sum mary of the principal parts of winch each division of each system was composed, to gether with a record of the repairs, changes, and delays noted during these tests. The manner of operation of these systems is given in Chapter VT1, where a more com plete description of the special regulating de vices is also presented. The niters of the respective systems are described in the order which has been fol lowed heretofore. All elevations used are in feet and refer to the bottom of the respective sand layers as the datum plane. The accompanying draw ings will facilitate an understanding of the several niters and their respective appurte nances. THE WARREN FILTER AND APPURTENANCES. The filter was placed in a circular wooden tank. About 1.5 feet from the bottom was the strainer system, which was made of per forated copper plates with suitable wooden supports. The layer of sand which was used as a filtering medium was placed upon the strainer system. Above the sand there was an open compartment which contained the water to be filtered. During filtration the water passed by gravity from the upper compart ment through the sand layer and strainer system into a closed compartment situated between the strainer system and the bottom of the tank, from which a pipe connected with the weir box, where the rate of filtration was regulated. For washing, the water was re moved from the upper compartment and wash-water admitted under pressure into the lower chamber, from which it forced its way up through the sand. After passage through the sand the wash-water was removed from the upper compartment by drain pipes. Dur ing washing the sand was stirred by rakes which were supported at the top of the tank. Plans and sections of the Warren System will be found on Plates II and 111, respectively. Filter Tank. The filter tank was made of alternate cypress and pine staves, the bottom being entirely of cypress. It was 10.6 feet in diameter on the inside and 9.75 feet deep inside. The staves were 2.62 inches thick and 6 inches wide. They were strongly bound with iron hoops, six in number. The hoops were 0.6 inch thick and 2 inches wide. About 1.5 feet from the bottom of the tank were wooden pieces which served as a support for a copper strainer floor. In the open com partment above the perforated copper floor was the layer of sand. The closed compart ment beneath the copper floor was the fil- tered-water chamber, through which the fil tered water passed as it made its exit from the filter. The filtered water used for washing the filter also passed through this closed com partment as it was pumped upward through the sand. From the bottom of the tank a central well 4.33 feet in height extended through the fil- tered-water chamber, strainer floor, and the sand layer. For a distance of 1.17 feet from the bottom the diameter of this well was 2.42 feet, and above this point 1.71 feet. Across the top of the tank" lay two timbers, one 12 by 12 inches and the other 6 by 12 inches, on which rested the bulk of the ap pliances for the operation of the agitator. The main vertical shaft, to which the rakes were fastened, was supported at the top by these timbers and guided at the bottom by a casting on the upper end of the central well. The height of water above the sand during filtration was normally about 5.75 feet. As described beyond, the total available acting head was 4.17 feet. The above description in general terms shows the relation to each other of the vari- WATER PURIFICATION AT LOUISVILLE. ous devices located in the filter tank. The de tails of these devices and their piping con nections are as follows: Inlet Water-pipe. The main inlet water- pipe was 8 inches in diameter and conducted the \vater by gravity from the outlet of the settling basin. This pipe led into and across the bottom of the riltered-water chamber, at the bottom of the filter tank, and connected with the central well by a flange joint. Arrangement for the Exit of the Filtered Water. After passage through the sand the water passed through the strainer system, composed of perforated copper plates and wooden supports; next through the filtered- water chamber; and thence through an 8- inch pipe to the weir box. From the weir box the water passed through about 65 feet of 5-inch pipe to the filtered-water reser voir. Strainer System. The original strainer sys tem consisted of punched copper plates sup ported by a network of radial and circum ferential wooden braces. Details of the ar rangement of these braces can best be under stood by an examination of the accompany ing drawings (Plates II, III, VIII). The radial supports were 2.25 by 2.75 inches, with the long side set vertically. They were supported at the center on a shoulder made for that purpose in the central well, and at the periphery on a ring made of short wooden sections nailed to the inside of the tank. They fitted tightly together at the cen tral well and were 7 inches apart at the pe riphery. On top of these supports was laid a second set of ribs, each 1.37 by 0.75 inches, with the long side set vertically. Between this upper set of ribs were laid pieces 1.25 by 0.75 inches, set perpendicular to the radius at their center. These circumferential spacers were sup ported at each end by the main radial beams, and were level on top with the upper radial strips. The spacers were not accurately cir cumferential, but were really a series of short chords. The perforated copper plates were placed on top of the ribs and spacers, and were fas tened to them. They were made in sections of the size of the space subtended by the radial ribs. The joints of the plates were over these ribs and were protected by copper strips 1.12 inches wide. Exit Area. The orifice area of the copper plate system was made up of about 681,900 punched holes. These holes averaged 0.043 inch (i.i millimeter) in diameter. They averaged 10.5 per linear inch radially, and 7.5 per linear inch at right angles thereto. The size and spacing varied considerably, but the above figures are averages of numer ous determinations at different parts of the strainer area. Using these figures as a basis of computation, the total orifice area of the copper-plate system, including all holes ex posed on the upper side, was 1032 square inches. No possible method of determining how much water passed through the holes directly over the supports was found. It seems probable, however, that the weight of the sand would press the plates sufficiently close to the supports to obstruct the passage of water to a considerable extent. Deduct ing all such holes, the net area was 923 square inches. On April 12, 1896, a finer sand was put in service, and it was found to be too fine to use with the original perforated copper plates de scribed above. Accordingly an auxiliary strainer device, consisting of a fine brass gauze, was added. This gauze was laid di rectly over the copper plates, the same copper strips being used to keep it in place. Owing to inability to secure readily a sufficient quan tity of gauze of the desired size two sizes were used. The first portion, which was used to cover about 80 per cent, of the area, had 65 meshes to the linear inch. For the remaining 20 per cent, of the area a gauze which had 80 meshes per linear inch was used. By the introduction of the brass gauze the determination of the available exit area of the strainer system was complicated. There were two extreme areas which may be considered, the true area utilized being somewhere be tween the two, apparently. 1 . The brass gauze may be assumed to have reduced the exit area of the perforated copper plates. In this case the gauze is assumed to have allowed water to pass through only those portions of it which were directly above the holes in the copper plate. 2. It may be assumed that the entire exit DESCRIPTION OF FILTERS. 7.; area of the holes in the copper plates was available, and that the water could in all cases pass more or less freely between the gauze and the copper plates. The second supposition appears to be more nearly correct, because, no matter how closely the gauze was pressed upon the copper plate, unless the wires were flattened, innumerable channels must have existed through which the water could flow more or less freely. Filtered-water Chamber. Directly beneath the strainer floor and forming the bottom of the filter tank was a closed compartment which was used as a collecting chamber for the filtered water, and also as a distributing chamber for the wash-water. It was simply the space left in the construction of the tank, no finishing being used. The total depth of the chamber was 1.5 feet, but the upper part was largely obstructed by the braces of the strainer floor, below which the depth was i.i feet. The area of the base of this chamber was somewhat less than that of the main tank, on account of restrictions by the wooden rim which supported the outer side of the strainer floor and by the central well. The area was 70.6 square feet. The total capacity of the chamber was 94.7 cubic feet, including the spaces between the supports of the strainer floor. The chamber could be drained through the waste-water pipe to within 0.6 foot of the bottom. No arrangements were made for complete draining. A small hand-hole was provided in one side of the tank for the purpose of inspection. The only method for cleaning was by forc ing filtered water into the chamber and allow ing it to flow out through the waste-water pipe. Weir Box. The weir box was an open, rectangular compartment constructed at the northwest corner of the settling basin, and built in connection therewith of the same ma terial. It was connected with the filtered- water chamber by an 8-inch pipe, 8 feet in length. It was 5.71 feet long by 2.75 feet wide, in side dimensions. The weir partition ran across the short dimension, dividing the box into an inlet and an outlet side. The inlet side was 2.67 by 2.75 feet, and the outlet 3.04 by 2.75 feet. As first constructed the weir was a fixed one with its crest approximately at an elevation of 6.00 feet. On Nov. 25 it was lowered to approximately 5.5 feet. With other changes previous to Nov. 25 a movable weir was inserted. This weir was made of an iron plate moving in guides at the sides, its position being controlled by a worm shaft operated by a wheel on the floor over the settling basin. It had an available ver tical movement from an elevation of 3.85 to the maximum water level (elev. 8.02), a dis tance of 4.17 feet. The nominal crest was 2.1 feet wide, but on account of leakage in the guides its actual width was probably about 2.5 feet. A 3-inch valve connecting the two sides was put in the bottom of the weir box Feb. 12 to allow more complete draining of the filter before washing. The center of this valve was at elevation 1.13. From the weir box the water flowed through about 65 feet of 5-inch pipe to the filtered-water reservoir. Outlet for Filtered Waste Water. At such times as in the opinion of the operator the filtered water was not of a satisfactory charac ter, a 3-inch pipe leading from the filtered- water chamber to the sewer was used in place of the main outlet through the weir box. Sand Layer. During the test several changes were made in the sand layer, the kind of sand, the thick ness of the layer, and the area of the surface, all having been changed. Kinds of Sand Used. At the opening of the test the sand layer was composed of sand No. I. This was removed Jan. 22, and sand No. 2 put in place and used up to April 13. On April 17 sand No. 3 was put in place. This was used up to July 25, when 2 inches of sand containing 23 parts of No. 3 and one part of a very fine sand were added. No. i was natural sand; the other two were crushed quartz. Mechanical analyses of these sands gave results which are presented on the next page. Thickness of Sand Layer. The thickness of the sand layer varied from three causes: i. Addition of sand by the operators of the filter. 74 WATER PURIFICATION AT LOUISVILLE. MECHANICAL ANALYSES OF THE SANDS USED IN THE WARREN FILTER. Finer than 2.04 millimeters... " 0.93 " 0.462 " " 0.316 " " 0.182 .-._ . (Ten per cent finer than ) Effective J dia ^ eter in mi ili m e- 100. 100. 4-2 0.3 O. I ent. by > 100. IJ.O 0.2 eight. IOO. 95.5 6.5 0.9 0.56 2. Losses of some of. the finer portions of the sand during the process of washing. 3. Increased compactness of the sand layer. At different places on the surface of the sand layer the thickness varied owing to the action of the rake-teeth of the agitator. As the agitator revolved during washing a small portion of the sand was moved from the cen tral part of the layer towards the periphery. The effect of this action was cumulative. Ob servations made on Jan. 20, Feb. 14, April 13, May 22, and July 17 showed differences in the elevations of the surface of the layer at the central well and at the periphery, ranging from i to 4 inches. On Nov. 25, 1895, the average thickness of the sand layer was about 2.36 feet. This thickness was increased on Jan. 3, 1896, by the addition of 0.6 inch of new sand (No. i). The average thickness of the layer of sand No. 2 on Jan. 25 was 1.86 feet. On Feb. 12 this thickness was increased to 2.25 feet. With the third lot of sand the thickness of the layer when new, April 17, was 2.17 feet; on May 22, 2.12 feet; and on July 17, 2.00 feet. On July 25, 0.25 foot of mixed sand was added. The average thickness at the close of the test was 2.25 feet. Area. As first arranged the sand layer extended to the wall of the filter tank and the surface area of the sand was equal to the area of the filter tank, excepting the central well. 1.8 feet in diameter. This area was approxi mately 85.70 square feet. Practically all of the tests were made after the completion on Nov. 25 of a new collecting gutter to carry the wash-water to the sewer-pipe. This made the diameter of the sand layer 10.1 feet includ ing the central well above noted. Allowing 2 square inches for each of ifi teeth which extended into the sand bed dur ing filtration, the net area of the surface of the sand was 77.36 square feet. From April 17 to July 25 the rake-teeth barely pierced the sand layer, thus increasing the area to 77.50 square feet. Device for Cleaning the Sand Layer. The device for cleaning the sand layer by washing comprised the following principal parts which are described in turn below: 1. Pipes through which filtered water was pumped from the filtered-water reservoir into the filtered-water chamber, and thence through the strainer system into the bottom of the sand layer. 2. Auxiliary slotted pipes, located at the bottom of the sand layer just above the strainer system, through which for a short time part of the wash-water was pumped, with the view to getting more uniform dis tribution. 3. A collecting gutter and pipes to carry to the sewer the last portion of the water on the sand layer just after draining the filter prior to washing, and the wash-water after it had passed through the sand layer. 4. An agitator with the necessary mechan ism for stirring the sand during the process of washing. 5. An engine, with pulleys, belting, and shafts, to operate the agitator. Wash-water Supply Pipe. The wash-water taken from the filtered-water reservoir was pumped through an 8-inch pipe. This ar rangement was used by all the filters in com mon. From the pump to a point on this pipe where a separate pipe branched to the Warren System was about 100 feet; a 4-inch pipe TO feet in length led from this point to the fil tered-water chamber beneath the sand. At first a 3-inch valve was located on this pipe just outside the filter tank. With this 3-inch valve on the 4-inch pipe the distribution of wash-water was not satis factory. The restriction in the pipe caused by the small valve gave something of a nozzle effect, so that the stream of water entered the filtered-water chamber with sufficient velocity to strike the outer wall of the central well, and be deflected up through a compara tively small area of the strainer svstem and DESCRIPTION OF F I HERS. 75 of the sand layer. To remedy this difficulty, and to increase the loss of pressure in the piping, the 3-inch valve was replaced on Feb. 12 by a 4-inch valve. Auxiliary Slotted Pipes for the Distribution of Wash-water. In addition to putting a larger valve on the wash-water pipe on Feb. 12, there were also introduced at the same time supplementary pipes to convey a portion of the wash-water to different points at the bottom of the sand layer. A 2-inch brass pipe branched from the main wash-water supply just outside the filter tank, the flow being regulated by a 2-inch valve. This pipe entered the tank above the perfo rated copper floor on which the sand rested. It connected directly with a ring of 2-inch iron pipe made of tees and eighth bends. The water was distributed by six i-inch slotted tubes of brass and the inlet pipe noted above, which was also slotted. The seven pipes or tubes were laid radially, spaced equally around the central well, and fitted into the respective tees in the iron ring encircling the central well. Two rows of longitudinal slots 90 apart extended the entire length of each tube. They averaged 2 inches in length, 0.031 inch in width, and were approximately 3 inches apart. All of the tubes were capped at the ends, and in the center of each cap was a hole 0.031 inch in diameter. The tubes were first set with the slots on the under side, with the center of the tube 2.25 inches above the per forated copper bottom at the inner end and 3 inches above at the outer end. On Feb. 19 the tubes were reversed, bring ing the slots on the upper side. The entire device was removed on Feb. 21. Collecting Gutter and Central Well. The cir cular collecting gutter was constructed of wood and galvanized sheet iron. A lining of pine staves 0.25 foot thick, extending from the strainer floor to 2.35 feet above it, was placed inside the filter tank. On the side of the lining towards the inner well a strip of galvanized iron was tacked, its upper edge extending 0.5 foot above the staves. The space thus formed between the metal strip and the main wall of the tank was used as a gutter. The upper edge of the metal strip, or, in other words, the discharge level, was at elevation 2.85 feet. At three equidistant points collections were made with this gut ter to a 3-inch pipe which partially encircled the filter tank on the outside. This 3-inch pipe connected by means of a special casting with a branch from the inlet pipe from the settling basin to the filter, which in turn connected with the sewer. By means of a tee and suitable valves on the 6-inch inlet pipe to the filter, this pipe was connected with the sewer, thus allowing the use of the central well (1.8 feet in diameter) for the removal of unfiltered waste and wash water. During the tests the crest of the central well was changed three times. When the depth of the sand was increased on Feb. 12 the height of the well was also increased about 4 inches. It was lowered again on Feb. 14 to the original height in order to try the effect on the sand of discharging all the water dur ing washing through the well. On Feb. 21 it was raised to the same level as the crest of the collecting gutter. Agitator. The agitator consisted essen tially of two horizontal rake-arms with eight teeth each, and the necessary mechanism to raise and lower the rakes, and to revolve them as desired. Power was furnished by a small engine, and transmitted by a 6-inch belt to a counter-shaft, from which another belt 6 inches wide led to the driving pulley of the mechanism. For simplicity in presentation a general description of the operating mechan ism is given, referring to each of the parts by serial numbers. Following this, the several parts are tabulated, and their leading dimen sions given. The rake-arms were hung on the main ver tical shaft, which was supported at the upper end by the frame of the machine, and guided at the lower end by a collar on the top of the inlet well. This shaft was turned around by means of a large bevel gear (i), a lug on which fitted into a vertical slot in the shaft. By this arrangement the shaft could be raised or lowered without interfering with the rotary motion. To drive the gear, a pinion (2) on the shaft which carried the main driving pul ley drove a gear (3) on a lower parallel hori zontal shaft. At the end of this shaft was a bevel pinion (4) which drove the rotating WA1ER PURIFICATION AT LOUISVILLE, gear. It will be noticed that this arrange ment necessitated rotation of the main verti cal shaft with its rakes whenever any part of the mechanism was in operation. For raising or lowering the main vertical shaft, power was transferred by gearing from the horizontal driving shaft to an upper parallel shaft, on the end of which a bevel pinion (5) drove the raising and lowering gears. For transferring the power two duplicate sets of gears con nected the main driving shaft with the upper shaft. Either of these sets could be used as desired, or they could both be out of opera tion, hand levers controlling their position. Each set consisted of the driving gear (6); two idle gears (7) and (8), and the driven gear (9); (6) and (9) were the same for both sets, and (7) and (8) duplicates in each set. in the original machine gear (8) was omitted, the in creased length of vertical motion of the modi fied machine, with its necessitated increase in height of the frame, requiring the introduc tion of the second gear. The raising and lowering gear proper con sisted of a large bevel gear (10), and a sleeve on the main shaft. This sleeve was made of babbitt metal cast on the main vertical shaft in the following manner: In the upper end of the main shaft (the lower end of which held the rake-arms) were cut nine circular slots, each I inch wide and 0.35 inch deep. The first one was 1.5 inches from the top of the shaft. Below this the slots were spaced 2.5 inches apart. On the shaft thus prepared was cast a sleeve of babbitt metal 0.75 inch thick and 25 inches long. In casting, a ver tical slot 1.5 inches wide and 0.5 inch deep was left in this sleeve. This slot engaged a lug on the framework of the machine, and prevented rotation of the sleeve, the steel core (main shaft) rotating within the sleeve. On the face of the sleeve a helical thread was cut, with three threads to the inch; this formed the worm, which engaged and was driven by a similar thread on the inside of the lifting gear, that worked freely on a loose bearing plate. As above described, a bevel pinion on an upper horizontal shaft drove this gear. For the purpose of stopping the vertical motion of the main shaft automatically, lugs were provided on the vertical shaft, which at the limits of motion (top or bottom) oper- aied sets of levers, which disengaged the set of idle gears which were in operation, trans ferring power from the driving shaft to the upper parallel shaft. The main dimensions of the gears and pin ions numbered in the above description are as follows: 1. Bevel gear, 35 inches in diameter with 72 teeth. 2. Pinion, 5.75 inches in diameter with 14 teeth. 3. Gear, 26 inches in diameter with 50 teeth. 4. Bevel pinion, with 13 teeth. 5. Bevel pinion, with 15 teeth. 6. Gear, 4.25 inches in diameter with 24 teeth. 7. Gear, 6.25 inches in diameter with 36 teeth. 8. Gear, 8.25 inches in diameter with 48 teeth. 9. Gear, 8.25 inches in diameter with 48 teeth. 10. Bevel gear, 16.5 inches in diameter with 66 teeth. Rakes. Attached by means of a collar and socket bolts to the main vertical shaft were the rake-arms. These were two in number and were set 180 apart. Two shorter arms on the other diameter carried tie-rods to strengthen the rake-arms. The rake-teeth were of cast iron. The original teeth were 27 inches long, but later a change was made, and 35-inch teeth were inserted. There were eight teeth on each arm. They were wedge- shaped in section, the back being rounded. On the original teeth a wedge-shaped shoulder was cast 13.25 inches from the upper part of the teeth. This was not used in the longer teeth. As first used, the rakes in the upper position were clear of the sand, and in the lower posi tion they averaged 5 inches from the strainer floor. On Feb. 21 they were lowered so that they came within 2 inches of the floor, longer rakes being introduced at the same time to allow for this greater penetration. The lift of the original machine was found to be too small with this new arrangement to raise the rake-teeth clear of the sand; and on April 13 a new machine was put in service as noted above, giving approximately 8 inches greater DESCRIPTION OF FILTERS. 77 lift of the rakes. At the close of the test with the sand layer 27 inches thick, the rake-teeth remained about i inch in the sand at the upper position. Engine and Belting. The engine was a Carlisle single-cylinder, fly-wheel engine. The size was 5.75 by 6 inches, with 77 per cent, cut-off. The engine drove a 6-inch belt over a 12-inch pulley 8.25 inches wide. From the engine a 6-inch rubber belt drove a 2O-inch pulley on a 2.5-inch counter-shaft. Another 6-inch belt from a 1 6-inch pulley on the counter-shaft drove an 1 8-inch pulley on the agitator machinery. Elevations. The different elevations in feet, referred to the bottom of the sand layer as the datum plane, were as follows: Bottom of sand layer (top of strainer floor) o.oo Floor of filtered-water chamber - 1.48 Sand surface (average, Aug. i, 1896). +2.27 Crest of central well and circular gut ter + 2.85 Lower end of rake-teeth (agitator up) . + 1.94 Top of filter tank + 8.27 Average maximum water level + 8.02 Lower floor (main-house floor) - 1.77 Center of inlet pipe at filter - 0.94 Center of outlet pipe at filter -0.85 Center of wash- and waste-pipes at filter - 0.98 Lowest position of weir + 3.85 Highest position of weir (available as outlet + 8.02 Crest of outlet channel from settling basin (mudsill) + 6.72 THE JEWELL FILTER AND APPURTENANCES. The layer of sand forming the filtering medium was held in a wooden tank set in the upper compartment of the main tank. The roof of the settling chamber served as a sup port for a layer of bricks and cement which covered the strainer manifold, and formed a support for the sand layer. Between the inner and outer tanks was a space which was used as a collecting gutter for draining. A central well connected the compartment above the sand with the settling chamber. During fil tration the water passed downward through the sand and the strainer system by gravity. The total available acting head was about 14 feet, of which 5.5 feet were positive (above the bottom of the sand layer), and 8.5 negative. Plans and sections of this system are shown on Plates IV and V, respectively. The rate of filtration was regulated by valves on the pipe from the strainer system. When the filter required washing, the water in the com partment above the sand, about 2.5 feet deep, was removed and wash-water admitted to the strainer system under pressure. Wash-water was then forced up through the sand and dis charged into the space between the two tanks, from which it was removed to the sewer. During washing the sand was stirred by a set of rakes supported by beams at the top of the main tank. Filter Tank. The filter tank was of cypress, 12.15 ^ ee t m inside diameter, 5.0 feet high on the outside, and 3.41 feet deep above the strainer floor. It was made of 3-inch staves and was strongly bound by three hoops, each 3 inches wide and 0.125 mcri thick. At its bot tom, the space between the filter tank and the main tank (about 0.3 foot wide) was filled by a wooden ring 0.33 foot thick. This ring served to brace the bottom of the staves, and also prevented any lateral movement. There was no floor in this tank, the staves resting upon the roof of the settling chamber. The spaces between the pipes of the strainer sys tem were filled with a layer of brick and cement, supported by the roof of the settling chamber. This brick and cement layer in turn supported the sand. The strainer system, consisting of a set of pipes to collect the water from the cups, and cups through which the water passed from the sand layer, was laid on the roof of the settling chamber, and covered by the layer of bricks and cement, the face of which was flush with the top of the strainer cups. The set of parallel pipes were all connected to a special cast-iron pipe which ran across the filter tank. On one side this casting was connected by a suitable joint with the outlet pipe. This pipe conveyed the water to the outside of the main tank, where it connected WATER PURIFICATION AT LOUISVILLE. with a cross, which was also connected to the outlet pipe leading to the nltered-water res ervoir; to the waste-water pipe leading to the sewer; and the wash-water supply pipe. A central well extended from the settling cham ber through the strainer floor and the sand layer to about 1.4 feet above the sand. At the top of the main tank were two tim bers, on which rested the bulk of the appli ances for the operation of the agitator. These timbers were supported at either end by suit able iron brackets fastened on the inside of the wall of the main tank, the upper face of the timbers being flush with the top of the main tank. Ordinarily the water above the sand layer partly submerged these timbers. The main vertical shaft, to which were fast ened the rake-arms of the agitator, was sup ported at the top by these timbers, and guided at the bottom by a ring on the inlet well. The above description in general terms shows the relation to each other of the various devices located in the filter tank. The details of these devices and the piping connections were as follows: Inlet Water-pipe. The inlet water-pipe was a central well 0.67 foot in diameter and 4.5 feet high. It was made of cast iron. Arrangements for the Exit of the Filtered Water. After passage through the sand the water passed through the strainer system, consisting of 444 strainer cups and suitable collecting pipes, to a connection with a 5-inch pipe. This pipe connected with a cross out side the filter. From the cross there were about 8 feet of 4-inch pipe leading to the au tomatic controller, from which about 65 feet of 5-inch pipe led to the filtered-water reser voir. Strainer System. The strainer system con sisted of brass strainer cups screwed into a set of collecting pipes. The shape and size of these cups, of which there were 444, is shown on the drawings. The face of the cup was covered with a punched aluminum bronze plate, the plate being secured to the cup by a ring which was riveted to the cup flange. The strainer cups were screwed directly into the collecting pipes, the arrangement of which is shown on the drawings. A central casting was fastened to the filter floor by six o.75-inch studs. In general form this casting was a hollow annular ring with flange joints on two ends of one diameter. To each of these flanges was attached a length of 5-inch pipe, each length being made in three sections 2 feet, i foot, and 2 feet long, respectively. The central section in one side was a nipple, and in the other a tee with the short arm passing down through the filter floor. The outlet pipe connected to this arm. Running from the 5-inch pipes above de scribed, and also from the central casting, was a system of i. 5-inch pipes, 23 on each side of the large pipes. These pipes were of different lengths to fit the inner circumference of the filter tank. The shortest was 1.33 feet and the longest 5.0 feet. They were spaced 0.5 foot from center to center. All of the pipes were capped at the ends. The strainer cups were screwed into the tops of the whole sys tem as above described; six in the central casting, twenty in each of the large pipes, and the remainder in the smaller pipes. These cups were all spaced approximately 6 inches from center to center except in the central casting. The distribution was very uniform; the great est distance from any cup to the nearest other cup, or from any point on the floor to the nearest cup, was 6 inches. The shortest dis tance between any two cups (4 inches) was at the central casting. Exit Area. The diameter of the opening of the strainer cups was 1.69 inches, and the area 2. 24 square inches. The aluminum bronze plate was punched with twenty holes to the linear inch, the holes averaging 0.028 inch (0.70 millimeter) in diameter and 0.0006 square inch in area. The orifice area per single strainer cup was therefore 0.54 square inch, giving a total area for the entire system of 240 square inches. The passage through the neck of the strainer cup was 0.188 inch in diameter. The total area of the whole system was 12.26 square inches, equivalent to a small fraction less than that of a 4-inch pipe. The major portion of the strainer system was covered with cement, the space between the pipes being filled with bricks. This formed the filter floor. It was level in the main, and flush with the top of the strainer cups. Where the cups were set into the large pipes and central casting they were 1.8 inches DESCRIPTION OF FILTERS. 79 higher than where set into the small pipes. There was no cement over the large pipes or central casting. Outlet Pipe. The outlet pipe was a 5-inch cast-iron pipe connected to the short arm of the tee in the main pipe of the strainer system. It was made up of 6.2 feet of straight pipe set vertically, a U trap and a length of 5-inch horizontal pipe connecting with a 5-inch cross outside the main tank. The whole length was about 11.5 feet to the center of the cross. From the cross about 8 feet of 4-inch pipe connected with the automatic controller. The filtered-water meter was located on this pipe. Automatic Controller. The automatic con troller consisted of a galvanized-iron tank, set vertically, open at the top, and arranged with a sharp-edge orifice at the bottom; an ar rangement of the outlet piping to discharge into the top of this tank; a funnel under the tank, on top of the pipe to the filtered-water reservoir, to collect the discharge from the orifice; a butterfly valve on the outlet pipe above the tank; a balance arm, operating the butterfly valve, one end of the arm supporting a weight, the other a copper can; and a con nection from the base of the tank with an ad justable overflow which discharged into the can on the balance arm. The device was de pendent on the rate of overflow into the can on the balance arm, an increase in height of water in the main tank increasing the over flow, thus increasing the amount of water in the small can, which caused a movement of the balance arm and a consequent closing of the valve. A 4-inch pipe connected with the outlet pipe just before the latter reached the con troller. It was used when the necessary act ing head fell below that available with the controller. This pipe emptied into the sewer. From the controller the water flowed by gravity through about 65 feet of 5-inch pipe, emptying into the filtered-water reservoir in side the house for the wash-water pump. Outlet for Filtered Waste Water. A 4-inch pipe connected with the cross above men tioned and conveyed such water as. in the opinion of the operator, was not of a satisfac tory character directly to the sewer. Sand Layer. During the test the sand was changed twice. The area and thickness were modified somewhat during the test by changes in the tank itself, due to warping. Kinds of Sand Used. At the beginning of the test the sand layer was composed of sand No. 4. On Feb. i this was removed and sand No. 5 put in its place. This was used till July 3. Sand No. 13 was put in service July 6 and used for the remainder of the test. Sand No. 13 was a natural sand; the other two were crushed quartz. Mechanical analyses of these sands gave the following results: MECHANICAL ANALYSES OF THE SANDS USED IN THE JEWELL FILTER. No. 4. No. 5. No. 13. Per cent, by weight. Finer than 2.04 millimeters 100.0 100.0 100.0 " 0.93 " 74.2 91.0 95.5 " " o . 462 " 19.5 1 1 . o 1 6 . 6 0.316 " 1.4 1.4 1.4 0.42 0.45 0.43 Thickness of Sand Layer. The thickness of the sand layer varied slightly, due to increased compactness during use and slight wastes during washing. The nominal thickness was 34 inches. On Feb. 28 it averaged 34 inches. On July 6 the new sand layer was reported as 34 inches thick, but a measurement on July 8 gave only 32 inches. The thickness at the close of the test was 30.5 inches. The sand layer was quite uniformly level, only 0.25 inch difference having been re corded between the center and the periphery. During reverse motion the rake-arms cut fur rows in the surface varying in depth from 0.25 to 0.75 inch. The impact of the water over the crest of the inlet well also caused a slight depression at about i foot from the well. The rake-arms, during filtration, pene trated the surface of the sand layer from 3 to 5 inches. Area. An average of several determina tions gave 115.8 square feet as the area of the sand surface. Device for Cleaning the Sand Layer. The device for cleaning the sand by wash ing comprised the following principal parts, which are described in turn below; 8o WATER PURIFICATION AT LOUISVILLE. 1. Pipes through which filtered water was pumped from the filtered-water reservoir into the outlet piping system. 2. The strainer system already described, which was used as a system for the distribu tion of the wash-water beneath the sand layer. 3. A collecting channel to convey to the sewer the wash-water after it had passed through the sand. 4. An agitator with the necessary mechan ism for stirring the sand during washing. 5. An engine to drive the main shaft. Wash-water Supply Pipe. The wash-water taken from the filtered-water reservoir was pumped through an 8-inch pipe. From the pump to the point where a separate pipe branched to the Jewell and Western systems was about 60 feet. From this point 10 feet of 5-inch pipe led to a point where a separate pipe, made up of about 4 feet of 5-inch pipe, a meter, and about 3 feet of 4-inch pipe, led to a connection outside the main tank. Device for Distributing the Wash-water under the Sand Layer. The device used for dis tributing the water beneath the sand layer comprised the outlet pipe, main casting, set of parallel pipes, and strainer cups, employed during filtration as the collecting strainer sys tem. As this device has already been described, it will not be repeated here. For the purpose of breaking the nozzle effect of the neck of the strainer cups, a small casting consisting of a ring and four arms connecting at the center was put in the cup when it was made. (See Plate VTIT.) The total area of the necks of the strainer cups was equal to a 4.1 2-inch pipe, or 68 per cent, of the wash-water supply pipe. Collecting Clianncl. The space between the filter tank and the main tank was used as a collecting channel, the water overflowing the edge of the inner tank. This channel was nominally 0.33 foot wide, but owing to warp ing and other displacements of the inner tank it varied from 0.2 to 0.4 foot. A suitable valve controlled the flow from this channel to the sewer through an 8-inch pipe about 9 feet long. Agitator. The agitating device consisted of a set of four rake-arms hung from a vertical shaft on the upper end of which was a hori zontal gear engaging a worm on a horizontal shaft. This shaft was driven by a small en gine. These portions of the agitator are next taken up and described in detail. During the first part of the test (up to June 2) a double-thread worm was used. On this date a single-thread worm was installed. The dimensions of this worm were:- Outside length, 4 inches; pitch, I inch; smallest diameter, 2.75 inches; and largest diameter, 4 inches. Both worms were of steel. The dimensions of the gear were: Outside diameter, 16.5 inches; inside, 16.188 inches; and pitch, i inch. The ratio of revolutions of the agitator shaft to revolutions of the main driving shaft was i : 49. The central portion of this gear was of iron, and the teeth were of bronze metal. The vertical shaft which carried the rake- arms was i .8 1 inches in diameter. The weight of the shaft and rake-arms was sup ported by the bearing of the gear above men tioned, the whole system being hung from this support. At the lower end a collar working on the inlet pipe leading from the settling chamber served as a guide. Attached to the vertical shaft above men tioned was a casting, in which there were sockets holding the rake-arms, four in num ber. The casting was fastened to the shaft by two set-screws, and it was also sup ported by a collar 1.75 inches wide fastened to the shaft by two set-screws. The arms were steel rods 1.75 inches in diameter. They were fastened into the sock ets by key bolts. There were two long and two short arms, set alternately about 90 apart. One of the long arms was 4.67 feet long, the other 4.33 feet long. The short arms were 2.17 feet and 1.58 feet long, re spectively. The longest arm had seven teeth, the next six teeth, and each of the short arms three short teeth and chains. On the long arms the teeth averaged 3.69 feet in length be low the center of the arms. Short teeth (2 feet long) were used on the short arms, each hav ing 22 inches of o.44-inch chain attached. The teeth were made of iron bars, 0.87 inch square in section, set so that one diagonal was tangent to the arc of movement. They were attached to the arms by wrist-joints allowing DESCRIPTION OF FILTERS. 8t them to turn freely in a left-hand direction, but holding them vertically when the move ment of the agitator was left-handed, the teeth turning in a right-handed direction on the arms. By this device the teeth were made to penetrate the sand to the full depth at once. When in their lowest position the distance between the teeth and lowest portion of the sand was about 0.25 foot. Engine.- -The engine was a small, double- cylinder, reversible, marine engine, with both pistons connected directly to a single hori zontal shaft by crank arms set at 90. The main dimensions of the engine were: Cylin ders, 3 inches in diameter; stroke, 4.125 inches; and cut-off at 85 per cent. On the outer end of the shaft there was a fly-wheel 2 feet in diameter, having an approximate weight of 90 pounds. The driving shaft was 1.23 inches in diam eter. From the center of the engine to the center of the worm the distance was 5.625 feet. Elevations. The different elevations in feet, referred to the bottom of the sand layer as the datum plane, were as follows: Bottom of sand layer (top of strainer floor) o.oo Top of filter tank (wash-water over flow) + 3.41 Sand surface (average Aug. i, 1896). . +2.54 Crest of central well + 3.68 Center of rake-arms + 3.93 Lower end of rake-teeth (during wash ing) + 0.24 Average maximum water level + 5.27 Lower floor (main-house floor) - 9.13 Center of supply pipe at settling basin. - 6.05 Center of outlet, wash and waste pipes at cross - 6.38 THE WESTERN GRAVITY FILTER AND APPURTENANCES. Water from the common settling chamber used for both Western Systems passed through this filter by gravity. The sand layer was contained in a vertical wooden tank, and the. open compartment in the tank above the sand contained the water to be filtered. During filtration the water passed down ward through the sand and was collected by a manifold of slotted brass tubes into a single outlet pipe, through which it flowed to the sewer. There were two outlets on this pipe, for use as a filtered-water outlet and a waste- water outlet, respectively, as the operator deemed advisable. When it seemed advisable to wash the sand the supply of water from the settling chamber was shut off, and the filter allowed to drain more or less completely. The water remaining above the sand was drawn off by means of a circumferential gutter at the periphery, an outlet from which connected with the sewer. Wash-water was then intro duced into the wash-water distributing system and forced up through the sand. The wash- water after passing up through the sand over flowed into the collecting gutter, from which it passed into the sewer. The total available acting head was about 14 feet. Before entering into a more detailed de scription of this filter, it is necessary to state that under the name of the Western gravity filter two essentially different filters were examined. The original filter (operated up to March 22) differed from the final filter (put in ser vice July 2) in the following points: 1 . Location of the sand layer, the final filter having its sand layer 7.0 feet higher than the original one. 2. Washing device, the final filter having a special arrangement for distributing the wash- water, while in the original filter the strainer manifold for the collection of filtered water alone was used. On account of the many modifications inci dental to the changes above noted, it seems best to consider two filters, Western gravity filter (A) and Western gravity filter (B). What has already been said applies to both niters. All elevations used are in feet and re fer to the level of the bottom of the sand layer of the Western pressure filter as the datum plane. The drawings (Plates VI and VII) give a plan and section of Western gravity filter (B), with reference lines to the 82 WATER PURIFICATION AT LOUISVILLE. location of the sand layer and the strainer floor of Western gravity filter (A). Western Gravity Filter (A). This filter was placed in a circular wooden tank which was open at the top. About one foot of the lower portion of the tank was filled with a layer of broken stone, concrete, and cement, by which the sand layer was supported. A manifold of slotted brass tubes which formed the strainer sys tem was half buried in the cement. The inlet pipe entered the tank at the top and dis charged into a circumferential trough, the crest of which was 1.69 feet above the sand and 8.24 feet below the top of the tank. The upper portion of the tank held the water to be filtered, a column normally about 8 feet deep. Filter Tank. The filter tank was made of pine staves 2.75 inches thick and 4 inches wide. It was smaller at the top than at the bottom, being m.o feet in inside diameter at the base and 9.5 feet in inside diameter at the top. The depth was 14.37 f eet - It was bound strongly by ten iron bands, each 0.25 inch thick, and ranging from 3.5 inches in width at the bottom to 2.5 inches in width at the top. /;//</ Water-pipe, The supply pipe to the filter connected with the outlet pipe from the settling chamber, and passed up over the edge of the tank and down on the inside, discharg ing into the circumferential trough. This pipe, from its junction with the outlet from the settling chamber, was 4 inches in diam eter. From the settling chamber to the point where this pipe began there were about 10 feet of 6-inch pipe. The total length of pipe from the settling chamber to the discharge in the filter tank was about 41 feet. Flow through this pipe was regulated by a hand- valve and by a plug operated by a float on the water in the filter tank. Arrangements for the Exit of the Filtered Jl atcr. After passing through the sand, the water passed through the strainer system, consisting of a manifold of slotted brass tubes; a rectangle of iron pipes into which these tubes were screwed, and which served as col lecting pipes; and an outlet pipe connecting the rectangle with a branch where two dis charge pipes, the filtered-water and waste- water pipe, respectively, connected with the sewer. Strainer System. The strainer system was a manifold of slotted brass tubes screwed into a rectangle 5 by 7 feet of 6-inch wrought-iron pipe. The drawings show the arrangement of these tubes in Western gravity filter (13). They were arranged in almost the same man ner in Western gravity filter (A), except that short lengths of tube were screwed into the outside of the rectangle also. The tubes were 1.5 inches in inside diame ter and laid in a bed of concrete, the surface of the concrete being just above the center of the tubes. The slots were circumferential, five rows of slots in each section, two above the cement floor and three below, the lower ones of course being covered up with con crete. They were cut from the inside by a circular saw making them wider and longer on the inside of the tube than on the outside. There was considerable variation in the length of the slots, and the widths differed by nearly 50 per cent. An average of several deter minations gave a length of 0.719 inch and a width of 0.024 inch. The slotted sections were spaced 0.125 inch from center to center. (See Plate VIII.) Exit Area. The total length of cut tubing was 727 inches. The exit area per linear inch was 0.272 square inch, making the total orifice area 198 square inches. Outlet Pi fie. The outlet was a 4-inch pipe connecting with the strainer manifold in the middle of one of the short sides of the rectangle. From this point it led out through the side of the tank and to the front of the filter, a distance of about 8 feet, where it branched into a filtered-water and a filtered waste-water pipe, the two latter pipes con necting with the sewer 6 feet beyond. Sand Layer. The sand layer was the same throughout the use of this filter. Sand No. 6, a natural sand, was used. Mechanical analysis of this sand gave results which are presented on the next page. Thickness of Sand Layer. The nominal thickness of the sand layer was about 3 feet. The thickness as determined January 16 was DESCRIPTION OF FILTERS. MECHANICAL ANALYSIS OF THE SAND USED IN THE WESTERN GRAVITY FILTKR(A). Number 6. I crci-nl. by weight. 100.00 90. ( 19.00 ... 3.60 o oo Finer than 2.04 millimeters " -93 ^ " 0.46 " 0.316 " "0.182 " Effective j Ten per cent, finer than / n< size ( diameter in millimeters )" 36 inches above the strainers. On March 20 about 2 inches were scraped off the surface after the close of the day s operations. The sand surface was level. Area. The area of the sand surface was that of an unbroken circle g feet 10 inches in diameter, which is equal to 75.94 square feet. Device for Cleaning the Sand Layer. The device for cleaning the sand comprised the following principal parts, which are de scribed in turn below: 1. Pipes through which the filtered water was pumped from the filtered-water reservoir to the wash-water distributing pipes. 2. A system of piping to distribute the water under the sand and thus cause its dis tribution through the sand layer during its upward passage. 3. A collecting gutter and pipes to carry to the sewer the water remaining above the sand after draining, and the wash-water after passage through the sand during washing. Wash-water Supply I ipc. The wasli-wat er, taken from the filtered-water reservoir, was pumped through 60 feet of 8-inch pipe and 55 feet of 5-inch pipe to a point of connection with the outlet pipe. Wash - water Distrihnting Pipes. The strainer system of slotted brass tubes was used as a wash-water distributing system. Collecting Gutter. A circular wooden gut ter, 12 inches deep and made of o.375-inch pine boards, was fastened to the inner wall of the tank 0.6 foot above the sand. This was used to carry off the wash-water after passage through the sand, and a pipe at the front with a suitable valve connected it with the sewer. Elevations. The different elevations in feet, referred to the bottom of the sand layer of the Western pressure filter as the datum plane, were as follows: Bottom of sand layer (top of strainer floor) - 0.78 Sand level (average March 22, 1896). + 2.22 Crest of collecting gutter + 3.91 Top of tank +12.15 Average maximum water level + 1 1.90 Lower floor (main-house floor) - 2.22 Center of outlet pipe at filter - 0.95 Western Gravity Filter (B\ The second filter operated under the name of the Western gravity filter differed from the first one in the location of the sand layer and the device for distributing the wash- water. The manner of operation was practi cally the same, except that a special wash- water distributing device was used. Connec tion was also made from the wash-water pipe to the collecting strainer system, whereby the latter could be used to distribute wash-water if desired, and also in order to loosen the sand around the strainers by forcing water through them. The normal depth of water above the sand during filtration was 3 feet. Filter Tank. The tank used was the same as that used by the Western gravity filter (A). Strainer Floor. The strainer floor was lo cated 8.37 feet above the house floor, or 7 feet higher than in the first filter. This was accomplished by building a second flooring of 3-inch pine planks supported by eight 4 by 6-inch pine posts. The lower part of the tank was not used with this filter, but was kept filled with water throughout the remainder of the test. On the wooden floor was laid a layer of broken stone and concrete faced with cement. The wash-water system, consisting of distributing pipes and ball nozzles, was buried in this cement layer. The top of the cement was flush with the face of the nozzles. The strainer system, consisting of a manifold of slotted brass tubes, was laid on top of the cement floor. The relation of the inlet, outlet, and waste- water pipes to the sand layer was the same as in the Western gravity filter (A). Arrangements for the Exit of the Filtered WATER PURIFICATION AT LOUISVILLE. Water. After passage downward through the sand the water flowed into the strainer tubes, a manifold of which covered the bot tom of the filter tanks. From this manifold a single pipe led to the filtered-water and waste- water outlets as in Western gravity filter (A), the main change being the insertion of 7 more feet of pipe necessitated by the increased elevation of the sand layer. Strainer System. The strainer system was composed of slotted brass tubes set in a rectangle of 6-inch wrought-iron pipes. It was laid on top of the cement floor, however, and not imbedded in it, the entire slotted area of the tubes being utilized. Instead of con forming to the sides of the tank as in Western Gravity Filter (A), the strainer tubes formed a rectangle. The strainer tubes were nominally spaced 12 inches from center to center, but, as will be seen from the drawing (Plate VI), there were several places on the floor of the filter where the nearest tubes were more than i foot apart. On the side of the rectangle at the center, the distance to the nearest strainer slot was approximately 15 inches, while at the corners the distance was about 18 inches. In general the arrangement covered the center of the bed uniformly, but was not well ar ranged to drain the sand at the periphery. Exit Aim. The total length of strainer tubes used was 39.3 feet. The orifice area per linear inch was 0.680 square inch, making the total area about 320 square inches. (For de tails of the strainer system see Plate VIII.) Sand Layer. The sand layer was made up of a mixture of sands in the following manner: Approxi mately 12 inches of a natural sand (No. 9) were put into the filter and washed for six minutes. One inch of fine material was then scraped off the top and discarded. Sample No. 10 was taken after this sand had been washed and scraped. The sand which was used in Western gravity filter (A), (No. 7), was then screened through a No. 24 sieve, and all of that which would pass through it (about one-half) was discarded. About fwo feet in depth of the screened sand were then put in the filter and the coarse and fine washed Finer than 3.90 mm IOO.O IOO.O IOO.O IOO.O I 2 . 04 100.0 95.5 100.0 98.0 i 0-93 95.5 66. o 93.0 89.0 0.46 22.0 25/0 35.0 lg.0 0.316 1.7 12.0 4.0 3.9 0.182 o.o i.o o.i o.g 0.105 o.o o.o o.o o.o ("Ten per cent"! Effective! finerthandi-l g size ] ameter in [ [ millimeters J together for ten minutes. Finally, on April 30, enough more of the screened sand was added to make the layer three feet thick. It was then thoroughly washed and ready for service. Sample No. 1 1 w from the final sand when ready for use. Sample No. 14 was collected at the close of the test, Aug. i, 1896. Mechanical analyses of these sands gave the following results: MECHANICAL ANALYSES OF THE SANDS USED IN THE WESTERN GRAVITY FILTER (B). No. 7. No. 9. No. 10. No. ji. No. 14. Per cent, by weight. Thickness of Sand Layer. The nominal thickness was 36 inches. On July 25, how ever, the thickness was found to be only 31 inches. The same thickness was found at the close of the test, August i. Area. On acount of the sloping sides of the filter tank the area was less than in West ern gravity filter (A), being 72.78 square feet. Device for Cleaning the Sand Layer. The device for cleaning the sand consisted of the following principal parts, which are de scribed in turn below. 1. Pipes through which the wash-water was supplied to the wash-water distributing system. 2. A system of piping and ball nozzles used to distribute the wash-water under the sand layer. 3. A secondary system for the distribution of the wash-water under the sand layer (com prising the strainer manifold). 4. A collecting gutter and pipes to carry to the sewer the last portion of the water re maining on the sand after draining prepara tory to washing the filter, and also to carry off the wash-water after its passage through the sand. Wash-water Supplv Pipes. The same pip ing was used as in the Western gravity filter DESCRIPTION OF FILTERS. (A) to convey the filtered water to the con nection with the wash-water pipe at the filter. During the test of this filter, however, un- filtered wash-water was used, except on the last day, July 30. For the use of unfiltered water a connec tion was made from the main inlet pipe to the settling chamber with the wash-water pipe at the meter, the filtered wash-water pipe being disconnected. Beyond the meter there were two branches taken from the main wash-water supply pipe, one of which was connected with a 6-inch pipe which led to the washing device in the filter tank, a distance of about 27 feet. On this pipe was located a swing check valve with a sand pocket. The other branch from the main wash-water supply pipe was con nected to the outlet pipe from the strainer system. When filtered water was used the main supply pipe was disconnected, and in its place connection was made with the filtered-water pipe used in the Western gravity filter (A). Alain Wash-water Distributing Device. The main wash-water distributing device con sisted of a manifold of pipes, feeding eighty- two ball nozzles distributed over the strainer floor as shown in the drawing. The entire system up to the face of the ball nozzles was covered by the cement floor. Sections of the nozzles are shown on the drawings. The total orifice area at the neck of the nozzles was made up of eighty-two 0.5- inch pipes equaling a 5.i88-inch pipe. All of the balls were of solid rubber, and had a diameter of 1.625 inches. The construction allowed them a rise and fall of about 0.5 inch. With the ball at full height the orifice area was approximately 1.4 square inches for each nozzle. The inner face of the nozzle was ground to allow the ball to make a close fit and so shut off the sand and water during filtration. Secondary Wash-ivater Distributing Svstcm. A connection was made so that the strainer tubes could be used as wash-water distribu ters, but the main washing was given through the ball nozzles, the water being turned through the strainers for the last minute only. Collecting Gutter. This was the same as was used in the Western gravity filter (A), but it was located 6.33 feet above its former position. Elevations. The different elevations in feet, referred to the bottom of the sand layer of the West ern pressure filter as the datum plane, were as follows: Bottom of sand layer (top of strainer floor) + 6. 1 9 Sand surface (average Aug. i, 1896). + 8.78 Crest of collecting gutter + 10.24 Top of tank +12.15 Lower floor (main-house floor) - 2.22 Center of inlet pipe (highest point). . -f 12.35 Center of outlet pipe at filter + 6.32 Center of outlet (lowest point) - 1.62 Center of outlet (discharge) - 1.62 THE WESTERN PRESSURE FILTER. This was a portion of a continuous series of pipes and compartments through which the water passed in the process of purification. There was no restriction of the pressure from the beginning to the end (outlet) of the en tire system, except such as was caused by the resistance of the piping, sand layer, and strainer system. The filtering medium was placed in one- half of a closed steel cylinder, the other half of which was used as a settling chamber. A supply pipe for this filter connected with the cylinder at the top by a flange joint. In the lower part of the filter chamber was a layer of broken stones, concrete and cement. The strainer system, consisting of a set of s otted brass tubes, was half buried in this concrete layer, the surface of which formed the floor for the sand layer. During filtration the water was admitted under a pressure of from 45 to 65 pounds into the portion of the chamber above the sand layer. After Feb. 29, 1896, the pressure was kept quite uniformly between 60 and 65 pounds. The outlet from the strainer system was then opened and the difference in pres sure caused the water to pass downward through the sand layer, through the slots in the strainer tubes and thence through the collecting pipes and outlet to the sewer. The WATER PURIFICATION AT LOUISVILLE. average total available acting head was about 140 feet, as the full pressure in the supply pipe was available, and the filtered water was dis charged into the sewer. At such times as it seemed necessary or advisable to wash the filter, a valve on the supply pipe from the settling chamber was closed. At the same time a valve on a^branch from this pipe which led to the sewer was opened. Wash-water was then let into the outlet system through connections between the two pipes, forced up through the sand and out through the inlet pipe and branch to the sewer. Filter Chamber. The filter chamber was cylindrical in section with dome-shaped ends. The principal inside dimensions were: Length in the center, 11.15 f eet : length on the sides, 8.71 feet; diameter, 8.00 feet. The inlet pipe entered the top of the chamber at the center of the compartment. In the lower portion of the compartment was placed a layer of broken stones, concrete and cement, about 2.1 feet thick in the center. The strainer system, con sisting of a frame of iron pipe and a set of slotted brass tubes, was half buried in this layer. ( )n top of this floor was the sand layer, and the upper portion of the compartment for a space about 1.7 feet high contained the water to be filtered. The inlet pipe and a branch therefrom was used as an outlet for waste water during washing. Inlet Water-pipe. The inlet pipe was 6 inches in diameter and conducted the water from the outlet of the settling chamber over to and into the filter chamber at the top. It was about 29 feet from the point where it con nected with the settling chamber to the con nection with the filter chamber. Connection with the steel shell was made by a flange, riveted to the shell. The pipe screwed into this flange. At first there was no provision for breaking the flow, but it was soon found that the impact of the water caused consid erable disturbance in the sand surface, and a 6-inch nipple 4 inches long was screwed into the flange from the inside. A 6-inch tee was screwed on the nipple, the long arm of the tee running parallel with the sides of the cham ber. Arrangements for the Exit of the Filtered Water. After passage through the sand, the water was collected by a manifold of slotted brass tubes set in a frame of iron pipe made in the form of a letter H, 9.0 feet long and 3.5 feet wide. From the center of the cross- piece of the H a single outlet pipe led down through the shell of the cylinder and out in front, where it rose above the floor, dividing into two outlets, for the effluent and filtered waste water, respectively, both of which con nected with the sewer. Strainer System. The strainer system was made of a manifold of slotted brass tubes screwed into two lines of 6-inch pipe. The arrangement is shown on the drawings. The tubes were 1.5 inches in diameter. They were partially imbedded in a concrete floor, the floor line being just above the center of the tubes. The slots were circumferential, five slots in each section, two of them above the floor and three below. The lower ones were of course covered up. They were cut from the inside by a circular saw, making the slot wider on the inside than on the outside, or, in other words, wedge-shaped. As the depth of cutting varied considerably, the width and length of the slots varied by quite a percentage. An average of many determinations gave a width of 0.024 and a length of 0.719 inch. The slotted sections were spaced 0.125 mc h from center to center. Exit Area. The area was made up of 731 linear inches of strainer tube, containing, per linear inch, 16 slots of an area of .017 square inch each, giving a total orifice area of 199 square inches. Outlet Pipe. The strainer manifold as above described connected by a tee in the center to a 6-inch downcomer, which went through the bottom of the filter and con nected with a pipe which passed out from under the filter, and branched up above the floor. The upward bend was made by a tee, the long arm of which was horizontal. To the outer end of the long arm the wash-water pipe was joined. Just above the floor the outlet pipe entered a cross. The opposite arm of this cross connected to the inlet pipe. The two horizontal arms connected to the outlet and waste pipes, respectively. These two pipes passed directly to the sewer. From the strainer to the cross the distance was about 9.5 feet. DESCRIPTION OF FILTERS. Outlet and Waste Discharges. From the opposite sides of the cross branched the out let and waste-water pipes, 4 inches in diam eter. They both led directly to the sewer, a distance of about 4.5 feet. Sand Layer. Kinds of Sand Used. The character of the sand was changed twice, a slight amount of coarser material being added the first time and some line sand the second time. Up to April 8, the layer was composed of sand No. 6, a natural quartz sand. Sand No. 8 was the same sand after use, the sample having been collected from the sand which was removed April 8. The new sand layer put in service May 8 was made up in the following manner: Approximately 12 inches in depth of sand No. 9, a natural sand, were put into the filter and well washed. All of the old sand was then put back, and on top of this 12 inches of the original sand (No. 6) were added. The sand layer was then washed for 10 minutes under unusually high pressure, enough sand being washed out, it was estimated, to lower the level from 3 to 4 inches. On June 3 about 6 inches of the original sand (No. 6) were added. Sample No. 15 was taken of the sand in use at the close of the test. Mechanical analyses of these sands gave the following results: MECHANICAL ANALYSES OF THE SANDS USED IN THE WESTERN PRESSURE FILTER. No. 6. No. No. 9. No. 15. Perc : by wei(, h t. nerthan 3.90 i nillimeters IOO O IOO O IOO.O ii ) 2.04 1 IOO O IOO O 95-5 IOO " 0-93 96 o 93 o 66.0 95 .5 0.46 "9 o 14 5 25.0 15 I 0.316 3 .6 i o 12. O 4 0.182 .0 o I I .O o .0 0.105 o o o o o.o o .0 fective I size j Ten percent finer ] than diameter in [ o millimeters. ) 39 o 43 0.30 o 44 Thickness of Sand Layer. The thickness of the sand layer was changed twice intention ally. On Jan. 13 it was 4.00 feet deep. Prac tically no change took place from that date until April 8, when the sand was removed. The new layer put in service May 8 was esti mated to be 4.85 feet thick before it was washed. On June 3 the thickness was found to be 4.27 feet or 7 inches less, of which it was estimated that 4 inches was caused by settling and 3 inches by removal in washing. On this date ten sacks of sand (No. 6) were added. The thickness as determined July 15 was 4.71 feet. After the close of the test, Aug. i, 1896, it was found to be 4.12 feet. Sand Surface. The major portion of the sand surface was level. In the center the im pact of the water from the inlet formed a de pression, about 3 feet in diameter and 3 inches to 4 inches deep. The introduction of a tee on the inlet pipe remedied this trouble. Area. The determination of the available filtration area of the sand layer was compli cated by the following facts: The sides and ends of the layer were curves, and every change in thickness changed the area of the layer. The sand surface did not form a sharp junction with the side wall, but for a depth, apparently, of from i to 2 inches the sand curved away from the wall. As the sides and ends of the sand layer were curved, the upper surface was less in area than a section at the center of the chamber. Inasmuch as it is the surface of the sand which removes the major portion of the sedi ment, and as it is customary in this connec tion to use the surface of the sand as a basis of computation, it is deemed advisable to use this as the filtration area. Further, as the thickness varied considerably during the test it was thought advisable to take the area as first measured and use it in current work, cor recting at the close if necessary. The area used was determined from measurements taken Jan. 13, which were as follows: Length at the side, 9.07 feet; length at the center. 10.35 f eet - an( l width, 6.67 feet. The laps at the ends somewhat reduced the area, the surface as determined being 66.22 square feet. Redetermination after the close of the test gave an area of 65.30 square feet, the change being due to the increased depth caused by adding more sand. This area was determined from the following measurements: Approximate radius of ends of sur face 8.92 feet. Middle ordinate of curve 0.60 Length of rectangular portion of surface 9.03 Total length of surface at center. . 10.23 Width of surface at center 6.67 " WATER PURIFICATION AT LOUISVILLE. For the purpose of comparison the follow ing areas are presented: Area of sand surface (Jan. 13, 1896) 66.2 square feet. Area of sand surface (Aug. i, 1896) 65.3 Area of surface of strainer floor 72.8 Area of maximum hori zontal section of filter chamber 83.3 The difference between the areas as deter mined Jan. 13 and Aug. i, 1896, is only 1.4 per cent., and inasmuch as the related obser vations were liable to a greater percentage error, it has been thought best to use the orig inal area of 66.22 square feet in all computa tions. On May 8 the area was probably 10 per cent, less than this, but it increased rapidly for three or four days, owing to loss of sand during washing, and was probably not over 2 or 3 per cent, less than this during the balance of the test. Device for Cleaning the Sand Layer. The device used for cleaning the sand layer by washing comprised the following parts: 1 . Pipes through which the wash-water was conveyed to a connection with the distribut ing pipes. 2. Pipes for the distribution of the wash- water under the sand layer during washing. 3. An exit pipe for the wash-water after it had passed through the sand. Wash-ivater Supply Pipe. During the major portion of the test filtered water was used as wash-water. From June 4 to July 27, inclusive, unfiltered water was used. The supply of filtered water was pumped from the filtered-water reservoir, through the same pipes as supplied the gravity filter, to the wash-water meter, a distance of 115 feet. From this point a separate 6-inch pipe led to a connection with the outlet pipe of the pressure filter, a distance of about 18 feet. For the use of unfiltered wash-water the connection with the. filtered-water supply pipe at the meter was disconnected and replaced by a connection with the main river-water pipe leading to the settling chamber. Pipes for the Distribution of IV ash-water. The outlet pipe conveyed the wash-water to the strainer tubes, which were used to dis tribute the water under the sand layer during washing. Exit Pipe for Wash-water. The inlet or supply pipe was used as an exit pipe for the wash-water after its upward passage through the sand. About 8 feet from the connection of this pipe with the shell of the cylinder a 6-inch pipe branched over to the sewer, a dis tance of about 10 feet. Suitable valves were provided on these pipes to allow their use as desired. Elevations. The different elevations in feet, referred to the bottom of the sand layer as the datum plane, were as follows: Bottom of sand layer (surface of strainer floor) o.oo Surface of sand layer (Aug. i, 1896). . +4.12 Inner side of cylinder at top + 5.84 Center of discharge tee on inlet pipe . . + 5.25 Lower floor (main-house floor) - 2.22 Center of outlet discharge at sewer. . . - 1.22 SUMMARY OF THE VARIOUS PARTS OF THE SYSTEMS. 89 CHAPTER VI. SUMMARY OF THE VARIOUS PARTS OF THE RESPECTIVE SYSTEMS, AND A RECORD OF REPAIRS, CHANGES AND DELAYS. IT is stated in the introduction to this re port that each of the systems described in the foregoing chapters represents the same method of purification, and that they differed only to a certain degree in the various devices employed to put into practical use the same fundamental principles. In this chapter it is the purpose to present a list of all the parts comprised in each of the divisions of the re spective systems employed in carrying out this method of purification, which consists of three steps, viz.: 1. Application of chemicals to the river water. 2. Coagulation and sedimentation. 3. Filtration. The schedules on the following pages (Tables Nos. i, 2, and 3) show the various parts comprising a system which in each case was installed to purify 250,000 gallons of river water per 24 hours, according to the contracts with the Water Company. They are of value not only as a matter of record, but also as an indication of the attention necessary for their satisfactory operation and maintenance. It will be understood that these schedules refer only to the systems and their immediate con nections. In accordance with the contracts between the several Filter Companies and the Water Company the latter provided the following portions of the experimental plant, in addi tion to the laboratory: 1. The houses in which the systems were installed, and the necessary foundations on which they rested. 2. All water and steam used by the systems in their operation. 3. All steam and water pipes leading to and from each system. 4. All meters for the measurement of the water. 5. A reservoir of 142,000 gallons capacity for storage of the filtered water for washing the filters. 6. A pumping engine of 3,000,000 gallons capacity per 24 hours to deliver filtered water under pressure to the large filters for wash ing^ The general location of the chief portions of the experimental plant, including the stor age reservoir and wash-water pump, is shown on Plate I. In Tables Nos. i, 2, and 3 are given lists of the principal devices employed in the respec tive systems for carrying on the three steps of (his method of purification. Table No. I includes all the principal de vices used in connection with the application of the chemicals to the river water. Table No. 2 contains a list of the prin cipal appurtenances of the settling cham bers. Table No. 3 is a tabulation of the principal devices and appurtenances of the filters. In the tabulations under this head, only the leading dimensions of the various devices are given, and reference is made in all cases to Chapters IT, IV, and V, where descriptions will be found, and to the drawings on which these devices are shown. All devices which were in design or construction peculiar to the respective system in which they were used are marked with a star (*). All other devices are understood to be such as are in common 9 o WATER PURIFICATION AT LOUISVILLE. TABLE No. 1. DEVICES FOR THE APPLICATION OF THE CHEMICALS TO THE RIVER WATER BY THE RESPECTIVE SYSTEMS. Warren. Jewell. Western-Device No. i. Western-Device No. 2. Mixing Tanks. Pump Boxes. Propeller Wheels. Pumps. Air Chambers. Air Compressors. Gauges. Glass Sight Tubes. Platform Scales. Steel Rods. Two circular white-pine tanks, 4 feet in diameter, 4.5 feet deep. One rectangular wooden box, 2.9 by 1.2 feet, by i.o feet deep. One seven-bladed screw- wheel made of cast brass, diameter 0.5 foot, depth 0.2 foot.* One vulcanized-rubber pump.* Two circular cypress tanks 3.5 feet in diameter, 5.5 feet deep. One ordinary " half-bar rel" of oak, iron-bound. One vertical iron cylinder, Two circular white-pine i foot in diameter, 2 feet tanks, 3 feet in diameter, deep.* 1 4 feet deep. Two single-acting steam- pumps. Size, 3.5 by 4.5 by 6.0 inches. One Worthington pumping engine. Size 8.5 by 9 by 10 inches. Two auxiliary pumps; plunger extensions of the piston-rods of the main pump. One cast-brass cylinder, 3 inches in diameter, 6 inches long. One cast-iron cylinder, 6 inches in diameter, 2 feet long- One wrought-iron cylinder, 5 inches in diameter, 4 feet long, with fittings. Two wooden depth gauges. One o. 5-inch glass sight tube, 1.5 feet long with brass fittings. One 24O-pound platform scale. Two glass sight gauges, i.o inch in diameter, 4 feet long, with brass fit tings. Two wooden depth gauges. One o. 75-inch glass sight tube, i.o foot long, with brass fittings. One 24O-pound platform scale. One celluloid mercury sight gauge with fittings; diameter approximately 0.25 inch. One 24o-pound platform scale. 6 feet o. 5-inch steel rod. One 24O-pound platform scale. Brass Pipes. Brass Fittings. Brass Valves. Iron Pipes. Iron Fittings. Iron Valves. Lead Pipes. Rubber Valves. gears.* Two 4 inch steel bevel gears.* 12 feet i. 5-inch brass pipe. Twelve i.5-inch brass fittings. Two i. 5-inch brass plugs. 24 feet i. 5-inch iron pipe. 10 feet o. 5-inch brass pipe. Five o. 5-inch brass fittings. Two o. 75-inch brass valves. Three o. 5-inch brass valves. One o.5-inch brass check valve. 95 feet o. 75-inch iron pipe. 30 feet o. 5-inch iron pipe. 10 feet i.o-inch iron pipe. 10 feet 0.5 inch brass pipe. Eight o. 5-inch brass fit tings. Twooo-inch brass valves. 10 feet o.75-inch brass pipe. 6 feet 0.5 -inch brass pipe. Nine o.75-inch brass fit tings. Twenty-two o. 5-inch brass fittings. Twoo. 75-inch brass valves. Two cast-brass valve chambers with valves. Six o. 5-inch brass valves. One o.5-inch brass check valve. 20 feet o.75-inch iron pipe. 12 feet o. 5-inch iron pipe. Twelve o. 75-inch iron fit tings. Eight o. 5-inch iron fittings. Four o.75-inch iron valves. Two o. 5-inch iron valves. One o.5-inch iron check valve. Two i. 5-inch iron valves. 7 feet i. 5-inch lead pipe. Dne i. 5-inch rubber float valve. Fifteen o. 75-inch iron fit tings. Twenty o. 5-inch iron fit tings. Five o. 75-inch iron valves. Two o. 5-inch iron valves. Three i.o-inch iron valves. One i.o inch iron steam regulating valve. 25 feet o. 75-inch lead pipe. SUMMARY OF THE VARIOUS PARTS OF THE SYSTEMS. 9 1 TABLE No. 2. DEVICES FOR THE COAGULATION AND SEDIMENTATION OF THE RIVER WATER BY THE RESPECTIVE SYSTEMS. Warren. Jewell. Western Device No. i. Western Device No. j. Settling One open rectangular The lower portion of a The settling chamber Same as device No. I. Basins or wooden basin, 12.1 feet circular wooden tank, formed one half of a Chambers. by 12. o feet, by 10.25 13.5 feet in diameter, steel cylinder, with feet deep. 14.0 feet high. dome-shaped ends, S.o; Main dimensions of set feet in diameter, 22.5 tling chamber : Diam long. eter, 13.0 feet ; depth, Inside dimensions of the 6.89 feet. settling chamber : Length in center, 11.15 feet; length on the side, 8.71 feet ; diameter, 8.0 feet. Iron Rod. 48 feet o.375-inch iron rod. Iron 2 feet 6-inch iron pipe. One quarter bend of 30 feet S-inch iron pipe. 24 feet 6-inch iron pipe. Pipes. halved 4-inch iron pipe. 23 feet 4-inch iron pipe. Two feet 8-inch iron pipe. 3 feet 4-inch iron pipe. 0.5 foot 5-inch iron pipe. Iron Four 6-inch iron fittings. Two 8-inch iron fittings. Six 6-inch iron fittings. Seven 6-inch iron fittings. Fittings. Three 4-inch iron fittings. Two 5-inch iron fittings. Six 4-inch iron fittings. Six 4-inch iron fittings. Iron One 5-inch balanced iron One 8-inch iron valve. One 6-inch iron valve. One 6-inch iron valve. Body valve operated by float. One 5-inch single seated One 4-inch iron valve. Valves One 4 inch Map valve valve operated by float. One 5 inch iron valve. TABLE No. 3. DEVICES FOR THE FILTRATION OF THE COAGULATED AND PARTIALLY SUBSIDED WATER BY THE RESPECTIVE SYSTEMS. Warren. Jewell. Western Gravity (A). Western Gravity (B). Western Pressure. Filter Tanks. One open circular wooden One open circular wooden One open circu- Same as listed One half of the tank ; 10.6 feet in diam tank, 12.15 feet in diam lar wooden under Western steel cylinder eter, 9.75 feet deep. eter, 5.0 feet deep. tank, 9.5 feet gravity (A). listed under set- in diameter at tlingchambers. the top, 10.0 Size of fi 1 1 er feet in diam compartment : eter at the bot Length in cen tom, 14.37 feet ter, 11.15 feet; deep. length on sides, 8.71 feet; di ameter, 8.00 feet. Sand Layers. Area, 77.36 square feet ; thickness, 2.25 feet ; vol Area, 115.8 square feet; thickness, 2.54 feet; vol Area, 75.94 square feet ; Area, 72.78 square feet ; Area, 85.30 square feet; ume, 6.5 cubic yards. ume, 10.9 cubic yards. thickness, 3.0 thickness, 2.58 thickness, 4.12 r 88.25 square feet copper feet; volume, i feet; volume, feet ; volume, plate, 0.031 inch thick, 8.4 cubic 7.ocubicyards. 12. o cubic punched with 0.043 inch yards. yards. holes, 78.6 to the square inch.* Strainer Systems. 6 1. 9 square feet brass gauze, 65 meshes to the linear inch. Seven iron castings for strainer manifold.* 444 strainer cups.* 39.5 feet 1.5- inch slotted brass pipe.* 38.0 feet 1.5- inch slotted brass pipe.* 61 feet i. 5-inch slotted brass pipe.* 15.5 square feetbrassgauze, 80 meshes to the linear inch. 172.5 feet copper strips 1.12 inches wide. Belts. 60 feet 6-inch rubber belt. WATER PURIFICATION AT LOUISVILLE. TABLE No. 3. Continued. Warren. Jewell. Western Gravity (A). Western Gravity (B). Western Pressure. Engines. )ne vertical single-cylinder 3ne double-cylinder re engine. versible marine engine. Diameter of cylinder, 5.75 Diameter of cylinder, 3.0 inches ; stroke, 6.0 inches ; stroke, 4.125 inches. inches. Gears. Dne 35-inch iron gear.* 3ne steel worm, single One 26-inch iron gear.* thread : length, 4 inches; One 16. 5-inch iron bevel pitch, I inch ; smallest gear. diameter, 2.75 inches ; Two 8.25-inch iron gears.* largest diameter, 4 One 6.25-inch iron gear.* inches.* One 6.o-inch iron bevel One gear made of iron and gear.* bronze metal : outside One 5.75-inch iron gear.* diameter, 16.5 inches ; One 4.25-inch iron gear.* pitch, i inch.* One 3.o-inch iron bevel gear.* Pipes. One length 8-inch iron pipe, 8 feet 4-inch iron pipe. 27 feet 4-inch 48 feet i. 5-inch 3 feet 4-inch iron 4 feet long.* 2 feet 5-inch iron pipe. iron pipe. iron pipe. pipe. One length 8-inch iron pipe, .8 feet 8-inch iron pipe. 28 feet 6-inch 25 feet 4-inch 62 feet 6-inch 3.75 feet long.* iron pipe. iron pipe. pipe. One iron central well : 2 feet 8 inch iron 60 feet 6-inch height, 4.33 feet ; diam pipe. iron pipe. eter, 2.42 and 1.71 feet.* g feet 8-inch iron One length 8-inch iron pipe. Pipe- 6.75 feet long.* 4 feet 2-inch brass pipe. 2 feet 3-inch iron pipe. 12 feet 3-inch iron pipe.* CUnftc 4 feet 4-inch iron pipe. shaft onaiis. i.o feet 2. 5-inch steel shaft. 17.2 feet i. 75-inch stee 1.81 feet 1. 25-inch steel shaft. shaft. 8 feet i. 25-inch steel shaft. 2.21 feet 1. 75-inch steel shaft. 3.87 feet 2.25-inch steel shaft. I pvf>r<5 Two sets shifting levers.* Rakes. Two rake arms.* One iron casting for sup Two stiffener arms.* port of rake-shafts.* Two tie-rods 0.75 inch in Thirteen wrought-iron diameter, 5.5 feet long. teeth, 3.69 feet long, 0.87 Sixteen rake-teeth, 35 inches inch square. long.* Six wrought-iron teeth, 2.00 feet long, 0.87 inch square. Pulleys One 20 inch ulle it feet o. 44-inch iron chain. One i8-inch pulley. One friction clutch for 18- inch pulley. One i6-inch pulley. One 12-inch pulley. ma n Special Castings for vertical shaft, 25 inches gear and shaft. Agicator. long, 0.75 inch thick Two cast-iron shaft sup cast with a helical thread ports. three threads to the inch.* Framework for agitator machinery, with bearing plates. Fittings. Five 8 inch iron fittings. Three 8-inch iron fittings. Six 6-inch iron Two 8-inch iron Twenty-five 6- Seven 4-inch iron fittings. Four 5-inch iron fittings. fittings. fittings. inch iron fit Twelve 3-inch iron fittings Eighteen 4-inch iron fit Nine 4-inch iron Six 6-inch iron tings. Four 2-inch brass fittings. tings. fittings. fittings. Four 4-inch iron Sixteen 2-inch iron fittings Five 4-inch iron fittings. One 8 by 6 by 3-inch tee.* fittings. Twelve i. 5-inch iron fittings. SUMMARY OF THE VARIOUS PARTS OF THE SYSTEMS. 93 TAHLE No. 3. Concluded. Warren. Jewell. Western Gravity (A). Western Gravity (B). Western Pressure. Brass and One 2-inch brass valve. One 8-inch iron valve. One 8-inch iron One 8-inch iron Five 6-inch iron Iron Body Two 8-inch iron valves. Three 5-inch iron valves. valve. valve. valves. Valves. One 6-inch iron valve. Two 4-inch iron valves. Four 4-inch iron Two 6-inch iron Two 4-inch iron One 4-inch iron valve. valves. valves. valves. One 3-inch iron valve. Four 4-inch iron valves. One 4-inch brass plug with float and float arm. Special One open rectangular One controller. Outlet wooden box, 5.71 feet by Regulating 2.75 feet, by 10.25 feet Devices. deep. One iron plate, 2.25 feet wide, 4.5 feet long. One worm shaft and wheel with stand. Broken Stone, 2 . 2 cubic yards brick and 2 n rnhir varHc , . . . Brick and cement. broken stone broken stone 3.0 cubic yards broken stone Cement. and concrete. and concrete. and concrete. Special 30 feet i-inch slotted brass Devices for pipe.* nozzles ; brass Distributing 4.5 feet 2-inch slotted brass castings with Wash-water. I pipe.* rubber balls.* RECORDS OF THE REPAIRS AND CHANGES OF THE VARIOUS DEVICES OF THE RESPEC TIVE SYSTEMS. The next two topics of this chapter, dealing respectively with the repairs and changes which were made during these investigations, are closely allied to each other. The majority of repairs were coincident with changes of more or less importance. As a matter of con venience for reference the repairs and changes are listed separately, so far as it is practicable to do so, on the basis that repairs related to work done on devices which had temporarily failed to serve their purpose, and that changes refer to the installment of new devices or por tions thereof where the old devices did not give results satisfactory to the operators. The repairs and changes, with the total periods occupied, are listed in the next two tables. The periods refer to 10 working hours per day from the time that regular operations ceased until they began again. Whenever possible repairs and changes were made outside of the regular hours of operation or during delays due to other causes. When repairs were made at such times the period occupied is es timated, and marked with a star (*). In some instances the periods for repairs and changes were caused in part by failure to provide necessary materials promptly. RECORDS OF THE PERIODS OCCUPIED IN RE PAIRS OF THE VARIOUS DEVICES OF THE RESPECTIVE SYSTEMS. Device Repaired. Total eriod Occ Jpied. Warren. Jewell. Western. *8.0 *10 8 * - Controller *2.O Inlet Ifj.0 *IO.O o*.S *IO.O RECORDS OF THE PERIODS OCCUPIED IN CHANGES OF THE VARIOUS DEVICES OF THE RESPECTIVE SYSTEMS. Warren. 1895, Nov. 12 to 25. 93 hours 30 min utes. Mainly to change sand layer and modify filter tank. 1896, Jan. 23 to 25. 20 hours 25 min utes. Mainly to change sand layer and modify central well. 94 WATER PURIFICATION AT LOUISVILLE. 1896, Feb. ii to 13. 19 hours o minutes. Mainly to introduce auxiliary wash- water distributing system, add new sand and raise central well. Feb. 14. 48 minutes. Mainly to change agitator teeth. Feb. 15. 50 minutes. Mainly to modify auxiliary wash-water distrib uting system. Feb. 21. 3 hours 34 minutes. Mainly to remove auxiliary wash-water dis tributing system, change rakes and modify central well. March 17. 10 minutes. Mainly to change position of rake-arms. April 13 to 20. 59 hours 30 minutes. Mainly to change sand layer and strainer system. April 23. i hour 35 minutes. Mainly to change agitating devices. April 25. 3 hours o minutes. Mainly to change agitating devices. 1896, Jan. 31 to Feb. 4. 26 hours 25 min utes. Mainly to change sand layer. Feb. 14. 47 minutes. Mainly to change outlet valves. June 2. i hour. Mainly to change worm gear of agitator. July 3 to 5. 30 hours o minutes. Mainly to change sand layer. Other changes were made outside of the regular hours of operation at various times. These were mainly changes in chemical pump, lime apparatus, other chemical devices, float in main tank, chemical feed-pipe fittings, fas tenings for rake-arms; the total time so oc cupied was about 35 hours. Western Gravity. March 22 to July 2. 765 hours. During this period the main changes made were in the strainer floor, sand layer, wash-water distributing systems, and piping systems. The changes in this filter were complete on May 8, but the filter was not put in official operation till July 2. Western Pressure. April 7 to May 8. 221 hours 45 minutes. During this period the main changes were made in the devices for the ap plication of the chemicals, the supply, and the distributing piping systems. Records of the Delays of Operation during the Tests. In this section is presented a record of the delays which were met with, and a summary of the time occupied in various ways during these investigations. After the work was well begun it was arranged that the systems should be operated as continuously as prac ticable from 9.00 A.M. to 5.30 P.M. on each week day, unless the Water Company re quested otherwise. From March 24, 9.00 A.M., to March 30, 5.30 P.M., the operations were requested to be continuous, as was the case from 9.00 A.M on Monday to 4.00 P.M. on Saturday for each of six weeks begin ning April 27. A number of repairs and changes by the operators of the several sys tems, and operations and observations by the Water Company, reduced somewhat the available period of operation as outlined above. The chief factors which caused tie- lay were: 1. Repairs and Changes. These have al ready been referred to above. Some of the principal ones necessarily extended into the regular periods of operation. The minor ones frequently were made outside the hours of regular operations, as will be noted from the differences in total time consumed as shown in the first two tables and in the final summary. 2. Removal of Sediment which had Subsided to the Bottom of the Settling Chambers. The average time required by the respective sys tems for this operation was as follows: War ren, 3 hours; Jewell, 2 hours; Western, 6 hours. In a majority of cases the settling chambers were cleaned at times of washing or when other causes delayed the regular op eration of these devices. The delays at such times were therefore less than the actual time required to clean the chambers. Exclusive of operations under prescribed conditions, the SUMMARY OF THE VARIOUS PARTS OF THE SYSTEMS. 95 chambers were cleaned on the following dates: Warren System: December 19, 1895; January 23, 1896; April 13, April 28, June 9, July 2, July 23, and July 28. Jewell System: December n, 1895; Feb ruary 28, 1896; April 25, July 3, and July 17. Western System: December 31, 1895; January 13, 1896; April 7, June 3, June 8, June 24, and July 23. 3. Sterilization of the Sand La\cr. This occurred three times in the case of the Jewell System, on October 30, 1895; January 8, 1896; and February 28, 1896. About four hours were required for the operation each time, and the sand was allowed to cool over night. Sterilization was not attempted in any of the other systems. 4. Change of Water in Settling Chambers. This was occasioned in some instances by the conditions prescribed by the Water Com pany during the period from May 18 to June 6. The water was usually changed dur ing the time of washing the niters in order to make the delay as small as possible. The periods required for the operation depended on the rate which was being maintained and the size of the respective settling chambers. The dates when these operations took place were: Warren System: May 18, 19, 20, 21, 25, 27, 28, and 29; June i, 2, and 4. Jewell System: May 18, 19, 22, 26, 28, 29, and 30; June 2, 4, and 5. Western Pressure System: May 28, 29, and June 4. 5. Observations and Operations b\ the Water Company. These included inspection of systems, collection of sand samples, special tests, repairs of meters and pipes, and ex amination of various details. The total periods of delay were: Warren, 33.2 hours; Jewell, 56.2 hours; Western Gravity, 15.6 hours; Western Pressure, 22.6 hours. The following table gives a summary of the total periods available for operation of each system, the periods during which the respec tive systems were in actual operation, and the periods during which the above-mentioned causes of delay interfered with the regular operation of the systems. It is to be noted that in this table the actual total time used in operation is presented, while in the final summary in Chapter IX only the period occupied by operations in cluded in averages is given. SUMMARY OF THE TIME OCCUPIED IN VARIOUS WAYS DURING THE TESTS IN DAYS OF 24 HOURS. Warren. Jewell. Western Gravity. Pressure Oct. 21, 95 Aug. i, i/> 103.56 91 .63 i .09 8.40 O.2I Oct. 21, 95 Aug. i, 96 102.54 94.00 0.79 3-89 0.16 Dec. 23/95 Aug. i, 96 65-45 26.50 0. 31.88 0.05 Dec. 23, 95 Aug i, 96 79.46 66.05 o. 10 9.25 0.05 0.03 0.94 o 2.12 O.92 Pe iod available for operation (by arrangement) Pe iod used for repairs Period used to change water in settling chambers L38 0. o. 0. 10 2.34 0.50 o. 0.34 0.65 0. 5.67 0.70 Period used for sterilizing sand layer Periods when by request of Filter Company systems were out of service .... WATER PURIFICATION AT LOUISVILLE. CHAPTER VII. THE MANNER OF OPERATION OF THE RESPECTIVE SYSTEMS OF PURIFICATION AND THE AMOUNT OF ATTENTION GIVEN THERETO. THE method of water purification investi gated in these tests, generally called up to this time " mechanical filtration," has been held by some to be so simple that practically no atten tion is required for its satisfactory operation. To many, however, the name conveys a dif ferent impression, that of a mechanism or combination of mechanical devices, for the perfect working of which, like that of any other appliance, careful and systematic su pervision must be maintained. It is the purpose of this chapter to show that the latter supposition is correct so far as it relates to the purification of the unsettled Ohio River water, because for the efficient maintenance of the systems examined during these tests constant care and regulation were necessary; and, further, that without this, ir regularities, often highly detrimental both to the character of the effluent and the cost of treatment, were bound to occur. The following topics will be presented in this connection: 1. The general manner of operation of the different systems. 2. The mechanical devices installed and used to aid in the operation of the systems. 3. The attention given to the systems throughout the tests. Under section No. i will be presented a general outline of the manner of operation of the different parts of the respective systems. Section No. 2 includes a detailed descrip tion of the special valves and other devices used to regulate or control the different op erations. These have already been referred to briefly under the different portions of Chapters II, IV. and V. Section No. 3 will include statements of the number of men employed by each system throughout the test. THE GENERAL MANNER OF OPERATION OF THE RESPECTIVE SYSTEMS. The general manner of operation of all the systems represented at these tests may be de scribed as follows: 1. The treatment of the river water with alum or sulphate of alumina for the purpose of obtaining coagulation and subsequent sedi mentation. 2. The filtration of the coagulated water, partially purified by sedimentation, through a layer of sand. 3. The washing of the filter (sand layer). Nos. i and 2 were carried on simulta neously, but were quite separate in their methods of control. In the following pages these different op erations for the respective systems will be described in order. Operation of the Warren System. Application of Sulphate of Alumina. The river water was supplied under pressure to the Warren System through a 5-inch pipe which was enlarged to 6 inches at the settling basin. The passage of the water through this pipe was controlled by a 5-inch gate valve, and a 6-inch balanced valve on the mouth of the in let pipe, the balanced valve being operated by a float in the settling basin. As has already been described in Chapter II, the arrange ment used for the application of the sulphate of alumina solution comprised a propeller wheel in the mouth of the inlet pipe; a pump on the upper floor operated by the propeller wheel : and a pair of tanks in which the solu tion was made, and from which it flowed by gravity to the pump box. The operation of the whole device was MANNER OF OPERATION OF THE PURIFICATION SYSTEMS. 91 automatic, as the current of water upon enter ing the basin operated the propeller; and it ii turn drove the pump, to which the solutior flowed from the chemical tanks by gravity. For the successful operation of this portion of the system, regulation of the rate of inflow of the river water was required. Control ol the strength solution was also required. The amount of sulphate of alumina applied to the water was regulated in two ways: 1. By varying the strength of solution. 2. By varying the number of arms on the pump into which stoppers were inserted, to prevent the entrance of the solution into the arms. This is more fully described in Chap ter II. It will be seen that the design of this por tion of the system called only for the initial application and the regulation of the river water into the settling basin, by hand; atten tion to the preparation of the sulphate of alumina solution; and the adjustment of the [jump to deliver a suitable quantity of chemi cals. The balance of the work was automatic. The special construction of the automatic de vices will be described in the next section. From the settling basin the water passed through a pipe to the central well of the filter, and thence to the top of the sand. Filtration. Starting with a clean sand layer just after washing, the settling basin full of chemically treated water, and all valves closed, the first operation was to open a valve on the inlet pipe from the settling basin to the filter, allow the water to fill the central well, over flow on top of the sand and slowly rise in the open compartment above the sand. It was necessary to let the filter fill slowly to avoid disturbance of the sand surface. From 8 to 15 minutes were occupied in filling the filter, the average time being about 10 minutes. As soon as the water reached within about 0.5 foot of the maximum level, a valve on the waste pipe was opened slowly and filtration begun. At the same time the valve on the in let pipe to the filter was opened wide. Dur ing the latter and greater portion of the tests no water was wasted in this system following a wash of the filter, and the filtered water was turned immediately into the main outlet pipe leading to the weir box. When wasting was practiced the rate of flow of water was regu- lated by hand by means of a 4-inch gate valve on this pipe. When the water became satis factory in appearance, the valve in the waste pipe was closed, and a valve on the main out let pipe, leading to the weir box, opened. This valve was opened slowly, allowing the filtered water to enter the weir box and rise on the inlet side thereof. As soon as the water began to flow over the crest of the weir, the valve on the outlet pipe was opened wide, and the rate of filtration regulated by means of the weir. The entire system was then in operation. The water was treated with suU phale of alumina as it entered the settling basin through which it flowed on its way to the sand layer. After it was filtered, it was discharged over the weir. The rate of flow of water through the en tire system was regulated solely by the mova ble weir, which was used only for this pur pose and not as a measuring device. The height of this weir was adjusted at varying in tervals, depending largely on the amount of suspended matter in the water flowing into the filter. As a general rule, at intervals of half or three quarters of an hour it was low ered an amount necessary to maintain the de sired rate of filtration, the meter on the pipe from the weir chamber to the filtered-water reservoir being used for the determination of the actual rate. During filtration the water passed freely from the settling basin to the compartment at the top of the filter, and stood at the same level in each. When it was considered necessary to waste the filtered water during filtration, the valve on the pipe connecting the filtered-water chamber beneath the filter and weir box was closed, and the valve on the waste pipe opened, the rate of flow of water being regu lated by hand. When the effluent became clear, the change was made to the main outlet pipe as described above in starting filtration. If the effluent did not become clear in a rea sonable length of time the filter was prepared or washing in the manner described below. Decision to IVasli the Filter. This decision was one which required considerable judg- nent. During the whole test no case was re corded where the Warren filter was washed on account of the entire available head having WATER PURIFICATION AT LOUISVILLE. been used, and the rate falling below the de sired quantity. In fact, less than 60 per cent, of the available head obtained with the weir (4.17 feet) was ordinarily utilized. In general it may be said that the only immediate guide to the decision to wash the filter at any par ticular time was the appearance of the efflu ent. In passing it may be stated, further, that the decision as to washing was influenced in a measure by several other features, the rela tive importance of which varied from time to time. These features related largely to the quality of the river water as it flowed from the settling basin to the filter, and especially in connection with the relative amount of aluminum hydrate present in the water at that point. The significance of these features will be mentioned beyond. Preparations for Washing the Filter. When it was decided to wash the filter the valve on the inlet pipe to the filter from the settling basin was closed. The water above the sand was then allowed to filter off through the sand, the rate being carefully regulated to the normal in order to maintain as good a charac ter of effluent as possible. This was continued until the water was drained down as far as de sired. With the use of the weir alone, there was left at least 2 feet of water above the sand. By the introduction of the valve (Feb. 12) in the weir chamber further drainage was made possible, only about 0.5 foot of water being left above the sand when this valve was used. During draining the settling basin was allowed to fill till the float closed the valve on the inlet pipe. Washing of the Filter. During the drain ing of the filter, preparatory to washing, the engine used for operating the agitator machinery was " warmed up." As soon as the filter was drained, the engine was started at full speed and the friction-clutch of the agi tator engaged. This started the agitator, which was allowed to turn a partial revolu tion before the lowering gear was engaged. The agitator then slowly descended, revolving at the same time. After about one revolution the wash-water \vas admitted into the filtered- water chamber at the bottom of the filter. The pressure of the water forced it up through the perforated bottom into and through the sand layer, thus loosening the sand. The agi tator continued to descend until it reached the full depth into the sand, when a system of levers automatically disengaged the lowering gears. At times these gears were thrown so far that the raising gears were engaged, neces sitating adjustment by hand. Washing was continued till the sand was, in the opinion of the operator, cleansed suf ficiently. During this time the power given to the agitator machinery was left constant, and the amount of wash-water admitted was regulated so as to maintain a regular rate of revolution of the agitator, usually from six to eight revolutions per minute. The rate of ad mission of wash-water was regulated by a valve operated by hand. The maximum, minimum, and average vertical velocities of the wash- water used (estimating 45 per cent, of the sand layer occupied by water) were 4.05, 0.86, and 1.79 linear feet per minute, respectively. As soon as the bed was cleansed to the de sired degree, the lifting gears were engaged and the rakes raised out of the sand. It was customary at this time to supply a little more steam to the engine and to admit a little extra wash-water, as the greatest load came on the agitating machinery when lifting the rakes. The construction of the machinery was such that the rakes could not be lifted vertically out of the sand, but must continue to revolve while rising. As the rakes approached their highest position, steam was gradually shut off from the engine. For some time it was cus tomary to shut off the wash-water when the rakes were about three-fourths out of the sand. In the early part of February, however, it was found that by continuing the supply of wash-water till the agitator was fully raised and stopped, the ridges in the sand formed by the rake-teeth were lessened. After this date it was customary to shut off the wash- water gradually as the rakes ascended, to stop the agitator by disengaging the friction- clutch as soon as the rakes were fully raised, and then to shut off the wash-water entirely. The engine was then stopped. The filter was now ready for filling with water from the settling basin preparatory to filtration, as has been described. MANNER OF OPERATION OF THE PURIFICATION SYSTEMS. Operation of the Jewell System. The same method of presentation of the operation of this system will be followed as was used with the Warren System. Application of Sulphate of Alumina. The river water was supplied to the Jewell System, under a pressure of about 60 pounds, through a 5-inch pipe. The pipe through which the solution of sulphate of alumina was pumped joined the inlet pipe at a point about 10 feet from the entrance to the settling chamber. The river water and chemical solution had to pass through the inlet meter and two valves before they reached the settling chamber. Two valves (a hand valve and an automatic valve) were used to control the flow through the inlet pipe. The first was a simple globe valve used to regulate the flow when starting the system or to shut off the river water upon stopping operations. The other valve was situated in the mouth of the inlet pipe within the settling chamber. It was controlled by a float in the compartment above the sand in the filter, and was relied upon to regulate the rate of admission of the water into the system as soon as the water rose high enough to set the float in operation. Regulation of the rate of admission of the water to the settling chamber also controlled its passage through the chamber and entrance to the fil ter. The rate of application of the chemicals to the river water was regulated solely by the speed of the pump used for that purpose. For large changes in the amount of chemicals ap plied to the water, the strength of the solu tion was varied. The speed of the pump was adjusted by regulation of a steam throttle- valve, the pressure of the steam being held nearly constant by a regulating valve on the main steam-pipe. During the early part of the test, the throttle-valve on the pump for the delivery of sulphate of alumina was con trolled by a float at the top of the filter. This was found to be unsatisfactory and hand regu lation was relied upon throughout the balance of the test. The rate of feeding the sulphate of alumina solution for short intervals was de termined by counting the strokes of the pump, the delivery of which was approxi mately 2.1 cubic inches per stroke. For longer periods control was obtained by com parison of the readings of the meter used for measuring the amount of solution and the meter on the inlet water-pipe. To start or stop the application of the sulphate of alumina solution the throttle-valve was opened or closed. The chemical pump was started just before the valve on the inlet water-pipe was opened, and stopped immediately after the latter was closed. A check-valve on the chemical feed pipe prevented the flow of water through this pipe from the inlet water- pipe when the pump was stopped. It will be noted that the regulation of the entrance of river water was automatic, but that the regulation of the application of the sulphate of alumina required adjustment by hand of the throttle-valve of the pump. The application of the mixed lime and sul phate of alumina was regulated at first in the same manner as the application of sulphate of alumina, one pump being used for the de livery of both solutions. No adequate means were provided to regu late the relative quantities of the two solu tions. This was remedied in the latter part of March by the use of an entirely separate arrangement for delivering the lime, includ ing a separate pump and piping system. Filtration. The water after passing through the settling chamber rose up through the cen tral well and overflowed in the compartment of the filter above the sand. Its flow was regulated by the entrance of the river water into the settling chamber, and this in turn was controlled for the most part by the float valve described above. Starting with a clean sand layer just after washing, the settling chamber filled with chemically treated water, and all valves closed, filtration was proceeded with as fol lows: The valve on the inlet water-pipe leading to the settling chamber was first opened and the sulphate of alumina pump started. This caused the water to rise in the central well and overflow on top of the sand in the filter. As soon as the water had reached its normal height above the sand, the outlet was opened, and filtration begun. This process of filling the filter usually occupied about 6 minutes, the time being dependent upon the rate used. WATER PURIFICATION AT LOUISVILLE. In filling the filter the rate of flow was regu lated by hand to the required amount, as the float valve did not operate unless the water was almost at its normal height in the filter. As the main outlet pipe and the waste-water pipe were simply different branches of the same pipe leading from the manifold in which the filtered water was collected beneath the sand, no special difference in the operation occurred whether the main outlet pipe or waste pipe was used. It was customary in this system to turn the filtered water directly into the main outlet pipe. This operation will therefore be described next. The flow through the main outlet pipe was controlled by a valve operated by hand, which was supplemented during the latter portion of the test by an automatic controller. In start ing filtration this valve was opened and the rate of flow regulated to the desired quantity. This was the only means of regulating the rate of filtration up to April 10, when the auto matic controller was introduced. This device, which will be described in the next section, was so arranged that a variation in the flow through it closed a valve automatically, if the flow increased, or opened it if the flow de creased. After the introduction of this de vice, it was customary to open wide the valve on the main outlet pipe as soon as the con troller was in operation. Owing to the pres sure required to operate this controller, the head available for filtration was reduced about 4 feet. To obviate this difficulty there was placed on the main outlet pipe a by-pass which cut out the controller. This by-pass was used when the available head fell below that necessary when the controller was in op eration. Under such circumstances the valve on the main outlet pipe was used to regulate the rate of filtration by hand. Whenever it was considered necessary to waste the filtered water, the valve on the main outlet pipe was closed and the valve on the waste pipe opened, the rate of flow being ad justed by hand. Filtration was continued until one of the two following conditions appeared: either the resistance of the filter, due to accumula tions of matters removed from the water, be came so great that the desired rate could not be maintained with the available head, or the appearance of the effluent became unsatisfac tory in the opinion of the operator. The determination of the course to be pur sued under these circumstances rested with the judgment of the operator of the system. At times it was found that the appearance of the effluent might fail for a short period and then improve. Wasting the effluent for a short time was often tried under these condi tions. When the available head fell below that necessary to maintain the desired rate of flow, one of the two following operations was adopted: either the filter was washed or the surface of the sand agitated. Surface agitation consisted in trailing the agitator (generally by hand) in a reverse di rection for about one revolution. By this means the surface of the sand was disturbed by the rake-teeth which rested upon it, and the layer of sediment on the top of the sand was broken up more or less. During this op eration the passage of water through the sys tem was stopped, but the water above the sand was never removed. Application and filtration of the water were immediately re sumed, the whole operation occupying from i to 3 minutes. Several times during the early part of the test continuous agitation of the surface of the sand during filtration was tried. It cannot be said to have been a nor mal procedure, however. In deciding whether to agitate the surface of the sand or to wash the filter, many con siderations had to be borne in mind. Decision to Agitate the Surface of tlic Sand. With a decreasing rate of flow, owing to increasing resistance of the filter, and a satis factory appearance of the filtered water, the question as to whether it was better to agitate the surface, or wash the filter, involved a con sideration by the operator of the following factors: 1. Length of the last run and amount of water filtered during the same. 2. Cause for last wash. 3. Success of surface agitation on last run, if tried. 4. Length of present run and amount of water filtered. 5. Appearance of the water flowing from the settling chamber to the top of the filter. MANNER OF OPERATION OF THE PURIFICATION SYSTEMS. It was found that under some conditions two and sometimes three surface agitations between washings were successful; while at other times the disturbance of the surface caused a deterioration in the character of the effluent, which did not improve. The degree of coagulation of the water as it entered the sand layer seemed to be a controlling factor. Decision to Wash the Filter. Several factors influenced this decision, which was in general only reached after a careful study of the vary ing conditions under which the system was being operated. Unsatisfactory appearance of the effluent and a utilization of the total available head were the immediate guides to washing. The quality of the river water be fore and after filtration, as shown by inspec tion and analytical results, was an important factor. Preparations for Washing the Filter. When it was decided to wash the filter the valve on the inlet water-pipe was closed, the chemical pump stopped, and the water above the sand allowed to filter off. When the water, in the opinion of the operator, was seriously defi cient in quality it was drained out through the waste pipe for filtered water or drawn off from above the sand by means of the collect ing gutter which connected with a pipe lead ing to the sewer. As a rule, however, the water while draining the filter was allowed to pass into the main outlet pipe to the fil- tered-water reservoir. During the draining of the filter the engine used for driving the agitator was " warmed up." Washing the Filter. As soon as the water above the sand was drained off, the engine was started at full speed in reverse motion. The wash-water was then turned into the out let pipe and allowed to force its way up through the sand. As soon as it appeared on the surface of the sand (generally about i minute) the engine was reversed, and the rake-teeth turning on the arms penetrated the sand to their full available length. The agitator was continued in operation, stirring the sand throughout the wash. Up to May i the rate of delivery of the wash-water was regulated to maintain a certain pressure on the sand. After this date the valve on the wash-water pipe was left wide open during washing. The agitator was operated nor mally at a speed of eight to nine revolutions per minute. Washing was continued until the sand layer was cleansed sufficiently, in the opinion of the operator, when the valve on the wash-water pipe was closed. The agitator engine was immediately reversed, and the rake-teeth were thrown to the surface by the resistance of the sand, which very quickly settled into place. The engine was then stopped. In washing this filter the wash-water was passed upward through the sand layer (esti mating 45 per cent, of the layer as occupied by water) at the following vertical velocities in linear feet per minute: Maximum, 2.58; minimum, 0.42; average, 1.37. Operation of the Western Systems. The arrangement of the Western Systems, as has already been stated, was such that the supply of river water, the apparatus for the application of alum or sulphate of alumina, and the settling chamber were used by both filters in common. The first portion of the description will therefore apply to both the gravity and pressure systems. Application of Alum Original Device. In the original device the river water was sup plied under pressure through a 6-inch pipe leading directly to the settling chamber. The flow of water through this pipe was controlled by a valve operated by hand. From this pipe, on each side of the valve, a small brass pipe led to the alum tank. The inlet pipe to the tank, from the upper side of the valve, passed through the cover of the tank, and projected into it about I foot. The outlet pipe simply passed through the cover of the alum tank and entered the inlet water-pipe below the valve. This tank was water-tight, and in it were placed crystals of potash alum. After the addition of the alum the cover was re turned, and the water from the main inlet water-pipe was admitted to the alum tank through the small brass inlet pipe. A valve was placed on each of the brass pipes to regu late the flow through them. To start the supply of river water and alum solution to the settling chamber, the valve on the inlet water-pipe was opened nearly wide, but not completely so, thus leaving a differ- WATER PURIFICATION AT LOUISVILLE. ence in pressure on the two sides of the valve. The valve on the brass pipe leading to the alum tank was normally left open. When the valve on the inlet water-pipe was opened, the valve on the brass pipe leading from the alum tank to the inlet water-pipe was also opened. The difference in pressure caused a current of water to pass from the inlet water-pipe through the alum tank and back to the inlet water-pipe. It was considered that this water in passing through the alum tank formed a saturated alum solution. The rate of flow of this solution was regulated by hand, the valve on the brass outlet pipe from the alum tank being used for this purpose. The actual rate of flow was observed by the aid of a small meter. No regulation of the water entering the settling chamber was attempted, as the rate of entrance of water was controlled solely by the rate of removal of water from the set tling chamber, which was kept constantly rilled and under nearly the full pressure. The water passed through the settling chamber and out at the top into the chamber outlet pipe, w.hich branched in front of the chamber into a supply pipe for the gravity filter and a supply pipe for the pressure filter. Unless the whole system was taken out of operation for a time no change was made in the valves on the main inlet pipe at all, except to regulate the rate of application of alum as described above. Application of Alum or Sulphate of Alumina Second Device. As has been described in Chapter II, the second device used in this system for the treatment of the river water with alum or sulphate of alumina consisted mainly of two mixing tanks and a pair of small pumps, together with suitable piping. The arrangement of the inlet pipes was changed but little so far as general operation is concerned. Unless the system was to be taken out of operation, the river water had free passage into the settling chamber at all times. The only change in this portion of the system was the operation of the main water pump. As soon as the valve on the pipe leading from the settling chamber to either filter was opened, and the water drawn from the settling chamber, the pump was started in operation by opening the steam throttle-valve. This valve was opened to a regular position and allowed to remain there till the draft on the settling chamber was stopped, when the pump slowed down owing to the increased pressure to pump against. The throttle-valve was then closed. The device for the application of alum or sulphate of alumina was put in operation simultaneously with the main water pump, as the chemical pumps were simple extensions of the piston-rods of the water pumps. When starting, the valves cutting off the supply of chemical solution from the tanks to the pumps were opened, the air in the pumps blown off by means of petcocks, and the device was then in operation. The method of regulation of the chemical application, as in the Warren and Jewell sys tems, was either by changes in the strength of solutions, or by varying the rate of applica tion. Setting aside leakages, the discharge from the chemical pumps was practically con stant. The method used to control the rate of application was to regulate by hand the rate of return from the pumps to the mixing tanks of a portion of the solution. Suitable piping with simple cocks was provided for this purpose. Observations of the rate of flow of the solution were made by the aid of a meter. A glass tube forming part of the pip ing system made visible the flow of the chemi cal solution. The operation of filtering the partially puri fied (settled) water by the gravity and pres sure filters, and the manner of washing these filters, is next presented. The modification in the former filter in April makes it advisable to consider it as two separate filters, gravity filters (A) and (B). Operation of the Western Gravity Filter (A). The water from the settling chamber flowed through a 4-inch pipe into the open circum ferential gutter in the compartment above the sand, from which it overflowed on top of the sand layer. The flow through this pipe was controlled by a 4-inch gate valve operated by hand, and a 4-inch plug operated by a float in the open compartment -referred to above. Filtration. Starting with a clean sand layer just after washing, and all valves closed, the operation of filtration was proceeded with as follows: MANNER OF OPERATION OF THE PURIFICATION SYSTEMS. 103 The gate valve on the inlet from the set tling chamber to the filter was opened and the filter allowed to fill slowly to the desired height. As a usual rule, filtration was begun when the open compartment above the sand was half to three-qarters full, and it was then allowed to fill till the flow was stopped by the plug operated by the float. This plug was intended to regulate the rate of flow of the water applied to the filter. The valve on the filtered waste-w ater pipe was next opened and the water allowed to pass downward by gravity through the sand. Filtration was usually begun at about half of the normal rate or less. This rate was held fairly constant, the regulation being by hand adjustment of the valve, until the water began to appear clear. The rate of filtration was then slowly increased to the normal. As soon as the water became clear at the normal rate of flow, the valve on the waste-water pipe was closed and the valve on the main outlet pipe opened. Filtration was continued, the valve on the main outlet pipe being regulated by hand from time to time to maintain the desired rate, till it became necessary or desirable to wash. Decision to Wash the Filter. It may be stated that the cause of washing this filter was in practically all cases the exhaustion of the full available head for filtration. When the accumulations on and in the sand caused a resistance so great that, with the outlet valve wide open, the rate of filtration fell below that which, in the opinion of the operator, it was economical or desirable to maintain, the filter was washed. Preparation for Wasliing the Filter. When it was decided to wash the filter, the valve on the pipe from the settling chamber to the filter was closed, and the water above the sand was allowed to filter out. This was, of course, done at a constantly decreasing rate, owing to the increasing resistance and the decreas ing head. The time occupied in this operation was often comparatively long, as high as i hour and 40 minutes being recorded, while intervals of from I to 1.5 hours were quite common. The variations in the quantity of unfiltered waste water (water remaining upon the sand after draining prior to washing) are interesting in this connection and will be pre sented in tables in Chapter VIII. After having drained through the sand layer as long as seemed desirable, the re mainder of the water above the sand was drawn through the circumferential gutter and pipe which led from it to the sewer. Washing the Filter. The operation of washing this filter consisted in opening the valve on the wash-water pipe to its full extent and letting the water under pressure into the strainer system beneath the sand. From the strainer system it forced its way up through the sand, stirring it more or less by the cur rent of the water. This was continued till the sand was sufficiently cleansed in the judgment of the operator, when the wash-water was shut off. The maximum, minimum, and aver age vertical velocities of the wash-water used with this filter (estimating 45 per cent, of the sand layer as occupied by water) were 3.42, 1.22, and 2.22 linear feet per minute, respec tively. The filter was then ready for filling with water from the settling chamber prepara tory to filtration, as has been described. Operation of the Western Gravity Filter (B). Filtration. This operation was similar to that followed in Western gravity filter (A), as the methods of regulating the flow into the filter, starting filtration and the regulation of the rate of filtration were all the same as in the original filter. In this filter, however, washing was often found advisable before the available head was used up. Decision to Wash the Filter. This question was, as with the other systems, a matter of judgment with the operator. Either the un satisfactory appearance of the effluent or "the decrease in rate of filtration on account of re sistance of the filter were used as immediate guides for the determination of the time of washing. The general features of operation, character of river water, amount and charac ter of water filtered, and several other allied factors, however, were all considered, as a rule. Preparation for Washing the Filter. The manner of preparing to wash the filter was practically the same as in the original I0 4 WATER PURIFICATION AT LOUISVILLE. filter. Owing to the small column of water (only about 3 feet) maintained above the sand, there was but little to drain out, and the op eration was performed quite rapidly, eight minutes being the longest time recorded. As a rule, however, no attempt was made to drain the filter, the water above the sand being drawn off through the gutter and drain-pipe to the sewer, as far down as the top of the col lecting gutter. The filter was then ready for washing. Washing the Filter. As soon as the filter was drained, the valve controlling the supply of wash-water was opened, and the water ad mitted freely into the ball-nozzle washing system. The water was allowed to pass up through the sand under full pressure, until the sand was considered to be cleansed to the desired degree. The wash-water was then shut off from the ball-nozzle system, and turned into the strainer system for a period of about one minute. The valve on the wash- water pipe was then closed, and the filter was ready for filling with water from the settling chamber preparatory to filtration, as has been described. In washing this filter an average vertical velocity of the wash-water (estimating 45 per cent, of the sand layer as occupied by water) of 2.25 linear feet per minute was maintained, and the range was from 1.64 to 2.81 linear feet per minute. Operation of the Western Pressure Filter. The supply for this filter was taken from the settling chamber through a 6-inch pipe, and admitted to the filter under full pressure (about 45 pounds to 65 pounds). The flow through this pipe was cut off when desired by a valve operated by hand. Filtration. Starting with a clean sand layer just after washing, and with all valves closed, the operation was as follows: Water from the settling chamber was ad mitted to the filter through the inlet pipe. The valve on this pipe was opened wide by hand. As the pressure filter in reality was only one of several closed compartments through which the water passed, there was no filling or draining the tank such as took place in the gravity filters. As soon as the valve on the inlet water- pipe was open, the valve on the waste-water pipe was opened, and filtration begun. The most distinctive point in the operation of this filter was the pressure. As noted above, full pressure was carried in the water above the sand, and the valve on the outlet pipe was opened only enough to cause a difference in pressure sufficient to allow the desired amount of water to pass through the filter. The available head in this filter was approximately 1 1 5 feet for the ordinary minimum pres sure. In starting filtration it was always cus tomary to waste the effluent for a short period. As in the operation of the Western gravity filter, the rate of wasting was usu ally about half of the normal rate until the water began to be clear. The rate was then increased up to the desired quantity; the valve on the waste-water pipe was closed; and the valve on the outlet pipe was opened. Filtration was continued until it was found necessary or desirable to wash the filter. The rate of filtration was regulated from time to time by hand adjustment of the valve on the outlet pipe. Decision to Wash the Filter. In this filter two factors were principally used as guides to determine when to wash it. The one most often relied upon was the appearance of the filtered water. The other, which was used mainly during the early part of the test, was the loss of head due to the resistance of the sand layer. As has been explained, the mini mum available head was about 1 15 feet. This was never used entirely, however, as it was deemed advisable to wash the filter when the loss reached about 50 feet. In some cases when the appearance of the filtered water was unsatisfactory it was found that by reducing the rate of flow for a short time the water would again become clear. Under these con ditions it was customary to waste at a low rate until the appearance of the water was again satisfactory, when filtration was resumed. Preparation for Washing the Filter. The preparation for washing was to close the valve on the main outlet pipe or waste-water pipe (whichever was in use), close the valve on the pipe leading from the settling chamber to the filter, and then open the valve on the MANNER OF OPERATION OF THE PURIFICATION SYSTEMS. branch which led from the inlet pipe to the sewer. Washing the Filter. The operation of washing the filter consisted of turning the wash-water under full pressure into the strainer system beneath the sand. The pres sure of the water forced it up through the sand layer, from which it passed out through the inlet pipe and its branch to the sewer. Washing was continued until the operator, judging from the appearance of the water dis charging into the sewer, thought the sand was cleansed sufficiently. The valve on the wash-water pipe was then closed, shutting off the -supply of wash-water. The valve on the sewer pipe was next closed and the filter was at once ready for use. On account of the curved sides of the filter chamber the velocity of the wash-water varied at different points in the sand layer. The maximum, minimum, and average vertical velocities in linear feet per minute were as follows: At strainer floor, maximum 4.48, minimum 1.40, average 2.68; at maximum horizontal section of sand layer, maximum 4.52, minimum 1.23, average 2.33; at sand surface, maximum 5.68, minimum 1.54, aver age 2.93. It will be seen that no automatic devices were employed in connection with the West ern Pressure System. Filtered water was used for washing the filter during the major portion of the tests, but during the period from June 24 to July 27 unfiltered river water, admitted at the full pressure in the main, was used for this pur pose. THE MECHANICAL DEVICES INSTALLED AND USED TO AID IN THE OPERATION OF THE RESPECTIVE SYSTEMS. Under this section the special devices used by the several systems will be described in the following order: Devices to regulate the admission of river water to the systems. Devices to regulate the flow of water from the settling basins or chambers to the filters. Devices to regulate the admission of chemi cal solution. Devices to regulate the rate of filtration. Devices Used to Regulate the Admission of River Water. Warren System. The flow of river water under pressure into the settling basin was regulated by a 5-inch valve operated by hand, and a common 6-inch balanced valve operated by a float in the settling basin. The float was a cylinder fixed on the end of an arm fastened to the side of the basin. The arm was ap proximately 3 feet long. About 0.75 foot from the fixed end of the arm a chain con nected to the valve stem. The relative motion of the float and valve was therefore about 4 to i. Jewell System. The flow of river water under pressure into the settling chamber was regulated by a 5-inch valve operated by hand and a 5-inch single-seated valve operated by a float in the open compartment of the filter above the sand. The float was a metal cylin der attached to an arm 3.67 feet long, one end of which was fastened to the beam supporting the agitator machinery. From the other end of the cylinder extended an arm to which was attached a chain leading down through a small vertical pipe to the valve mechanism on the inlet water-pipe. The total length from the fixed end of the float to the point where the chain was fastened was 5.17 feet. The chain was fastened on the long arm of a bell- crank lever, the arms of which were i .84 feet and 0.21 foot long, respectively. The short end was fastened by a wrist-pin to an arm o. 18 foot long, which in turn connected with the valve plate by means of a wrist-pin. The valve plate was connected with a fixed arm hinged on one side of the frame. The dis tance from the wrist-pin on the valve plate to the hinge about which the valve moved was 0.33 feet. This combination of levers caused a motion of the valve, the relation of the movement to that of the float depending upon the position of the float. When the latter was down, the valve was given its greatest pro portional movement. When the float ap proached its highest position, the short arm of the bell-crank and the arm attached to the valve approached a straight line, and the movement of the valve became very small. Western Systems. The supply for these systems was regulated only by controlling io6 WATER PURIFICATION AT LOUISVILLE. the rate of flow to the filters as described be yond. The main water pump aided somewhat in maintaining a uniform pressure of the water, but cannot be said to have been used as a device for regulating the flow of water. Devices Used to Regulate the Flow of Water from the Settling Chambers to the Respective Filters. Only one special device can be said to have been used directly for this purpose, viz., the plug operated by a float in the Western gravity filter. In operation, however, the device used by the Jewell System for regulat ing the admission of water to the settling chamber acted as a controlling device in this connection, as did also the similar mechanism in the Warren System. Devices Used to Control the Application of Alum or Sulphate of Alumina. Warren System. The sulphate of alumina solution admitted to the pump box was regu lated by a valve operated by a float in the pump box. The float arm was vertical, the relation between the position of the valve and the level of the solution being adjustable by changing the length of the vertical arm. The rate of discharge of solution by the pump for each revolution was capable of variation by means of the insertion of stoppers in the ends of the pump arms. This operation was, of course, performed by hand. The pump itself was the main regu lating device, because it was operated by the flow of the water in the main inlet water-pipe, and its rate of revolution was designed to be proportional to the flow of water through the pipe. Jewell System. During the first five months of the test the movement of a float at the top of the filter, as described above, was relied upon to regulate the rate of application of the solution of sulphate of alumina, by ad justing the throttle-valve on the chemical pump. After this was abandoned, hand regu lation alone was used. Western Systems. In the original device used by these systems the method used for regulating the application of alum solution was by hand adjustment of the valve on the alum pipe. The arrangement of the device allowed adjustment by regulating the differ ence in pressure in the two alum pipes, which was done by changing the position of the valve on the main inlet water-pipe. This was only used in starting operation, how ever. In the second Western device the method of adjusting the application of chemicals was by hand regulation of the return or relief out let valves. Devices Used for Regulating the Rate of Filtration. Warren System. The regulating device used in this system was a movable weir. This has already been fully described in Chapter V. As a rule it was adjusted about every half-hour. A valve operated by hand was used for regulating the rate of wasting. Jeivell System. As described in section I of this chapter, the rate of filtration was regu lated by means of hand valves up to April 10, 1896. On this date a device called an " auto matic controller " was installed. The ar rangement of this device was as follows: SBA\ ,i3;iy aqi UIQJJ adid }3[}no UIEUI 3t[j, raised up and the end turned down, so that it discharged into a galvanized iron tank through a 4-inch butterfly valve. Through this pipe the flow was regulated by the valve, which in turn was controlled by the position of a bal ance arm on the valve stem. The iron tank, i foot in diameter, was hung from the outlet pipe by four arms. Its upper end was open. The outlet from this tank was a sharp-edged orifice at the bottom, through which the water was discharged into a galvanized iron funnel, which led the water into the pipe con nected with the filtered-water reservoir. The regulation of the flow was obtained by the butterfly valve above mentioned, the posi tion of the valve been controlled by the level in the tank in the following manner: The balance arm of the valve held an iron weight at one end and a copper cylinder at the other. This copper cylinder had a discharge port and funnel at the bottom. From the bottom of the iron tank a flexible pipe, with an overflow, fed into the small copper cylin- MANNER OF OPERATION OF THE PURIFICATION SYSTEMS. 107 der above mentioned, and the water in the cylinder flowed back into the funnel on the pipe to the reservoir. The small overflow pipe was adjustable to any desired height so that it would cause the desired rate of flow through the orifice in the bottom of the iron tank. The parts were so proportioned that if the water remained con stantly at the desired level, the overflow into the small cylinder was just sufficient to keep the water in this cylinder at a height neces sary to balance the weight on the other end of the lever arm, thus keeping the valve open the required amount. When the flow in creased or decreased, the overflow became greater or smaller, and the level of the water in the small cylinder therefore increased or decreased. The balance arm moved corre spondingly, thus closing or opening the valve. Western System. Hand valves alone were used for regulating the rate of filtration in the Western Systems. THE ATTENTION GIVEN TO THE RESPECTIVE SYSTEMS. The general manner of operation of the respective systems, and the special devices in stalled to aid the operators, have already been presented. In this section, the number of men which the several companies considered necessary to employ to operate the respective systems will be given. It is not the pur pose of this section to present any compara tive statements for the purpose of showing the number of men necessary to operate any modification or enlargement of these systems, but to show clearly the amount and character of supervision deemed necessary by the dif ferent filter companies for the operation of their respective systems. From March 24 to 29, inclusive, the sys tems were operated day and night. For six weeks beginning April 27 the systems were operated continuously from 9 A.M. on Mon day to 4 P.M. on Saturdays of each week. The Number of Men Engaged in the Operation of the Respective Systems. Warren System. Throughout the test this system was in charge of a trained engineer who was also a chemist. He was assisted by one regular helper. During the continuous run in March, the system was operated at night by a superintendent of the company. During the six weeks run the regular assist ant took charge at night, and a second helper was employed during the day. Jewell System. From the beginning of the test till Nov. n, 1895, the system was oper ated by one man who was a chemist. After Nov. 1 1 an officer of the company was in charge. He was assisted till Nov. 18 by the original man in charge. On that date an other chemist was employed in place of the original chemist. This chemist was replaced during the second week in December by a mechanic. A boy was employed after the first week in December. This force of two men and a boy remained up to March 23, when a chemist was employed. No change was made in this force of three men and a boy through out the balance of the test except during the continuous run in March. During this continuous run the system was operated during the day by an officer of the company assisted by the chemist, and during the night by another officer assisted by the mechanic and the boy. During the six weeks continuous run the force of three men was divided into three watches, the boy assisting about the place during the day. Western Systems. These systems were in charge of a trained chemist throughout the test. He was assisted in the routine operation by one and sometimes two men. Up to March 24 only one helper was employed. A new man was employed to assist during the day at the beginning of the first continuous run, the original helper being on duty at night. Two men were employed from this time till the first week in July. A boy was also employed to assist at night throughout the six weeks continuous run. The influence on the results accomplished by the systems of the attention received will be considered further in Chapter IX. io8 WATER PURIFICATION AT LOUISVILLE, CHAPTER VIII. COMPOSITION OF THE OHIO RIVER WATER AFTER TREATMENT BY THE RESPECTIVE SYSTEMS OF PURIFICATION, AS SHOWN BY CHEMICAL, MICROSCOPICAL, AND BAC TERIAL ANALYSES; TOGETHER WITH A TABULATION OF THE MOST IMPORTANT DATA ON THE OPERATION OF THE RESPECTIVE SYSTEMS. IN this chapter is recorded the main bulk of the detailed results of the observations and examinations made during the investigations. These results are presented in a series of tables as follows: Table No. I, results of regular chemical analyses, indicating the sanitary and technical characters of the water after purification. Table No. 2, results of mineral analyses of the water after purification. Table No. 3, results of microscopical ex aminations of the water after purification. Table No. 4, results of bacterial analyses of the water after purification, and a record of conditions under which each sample was col lected. Table No. 5, records of the operation of the respective systems, including a brief summary of the analytical results and also the amount of sulphate of alumina used during each run. The next chapter, in which these results are summarized and discussed, also contains some analytical and other results which were ob tained in connection with special points which are outlined in the discussion. This chapter deals solely with the principal detailed rec ords. In the case of each table there are a number of explanations and points to which attention is called, as stated in the following paragraphs. It may again be stated that effluent refers to the water, after its passage through one of the systems, which was passed into the outlet provided for the finished prod uct. It is further to be noted that whenever filtration or any related term is used, it refers to effective filtration, i.e., the filtration of water which was passed into the outlet pro vided for the finished effluent. Table No. i. Samples. The samples of the several effluents, for regular chemical analysis, are listed in serial numbers. The same series of numbers was used for both the untreated river water and the effluents. During the process of analysis samples were designated by these numbers only, and the source of the samples was not known to the analysts. The condi tions under which the samples were collected are presented as a matter of convenience by reference to Table No. 5, which includes the results of bacterial analyses of corresponding samples. During a portion of the time samples for chemical analysis were collected continuously by an automatic device which is briefly described in the appendix. In such cases the period covered by the sample is re corded. In cases where several portions of effluent were mixed together for a single average sample, the water was kept in an ice-box dur ing the period which intervened between the times of collection and of analysis. Methods of Analysis. Substantially the same methods were employed for the analysis of effluents as were used in the case of the untreated river water. They are presented briefly in Chapter I. The only point to be mentioned is that in the present tabulations there are included results indicating the ap pearance of the effluents. These results are given under the heading. " Degree of Clear ness." They were obtained by careful inspec tion, aided by a diaphanometer such as is briefly described in the appendix, where an outline of the methods followed will be found. COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 109 The significance of the five degrees of clear ness is substantially as follows: Degree No. i, brilliant. Degree No. 2, clear. Degree No. 3, slightly turbid. Degree No. 4, turbid. Degree No. 5, very turbid. These expressions relate to perfectly clear water as a basis of comparison, and have nothing to do with such expressions as might be applied to the muddy river water. No ob jection could be raised by consumers with re gard to the appearance of the effluents when it was represented by any of the first three degrees of clearness; with the fourth, the turbidity would probably be noticed by con sumers at times, but not uniformly. Color. The color of the effluents, so far as related to dissolved matters in the water, was very slight, as is also true of the Ohio River water before treatment. Some of the color re sults were unavoidably increased by minute particles suspended in the water. Carbonaceous Organic Matter Oxygen Con sumed. The carbonaceous organic matter in the effluents, as indicated by the oxygen con sumed, was satisfactory, practically without exception, and was less than that dissolved in the river water. Nitrogenous Organic flatter Albuminoid Ammonia. The nitrogenous organic matter in the effluents, as indicated by the nitrogen in the form of albuminoid ammonia, was also satisfactory as a rule, and less than that dis solved in the river water. Very little or no organic matter was suspended in the effluents ordinarily. In effluents which had either of the first two degrees of clearness it is recorded as zero. In the other three degrees of clear ness it was appreciable; but it was too small for measurement in a satisfactory manner, and accordingly blanks are inserted in the tables under these conditions. Nitrogen as Free Ammonia and Nitrites. As a rule there was a slight reduction in the effluents, as compared with the river water, in these compounds, which represent inter mediate steps in the conversion of organic matter in its crude form into completely oxidized mineral matter. Nitrogen as Nitrates. There was sub stantially no change in the water before and after treatment with regard to the amount of nitrogen in the form of nitrates. This deter mination indicates the amount of organic matter which is completely oxidized ; and it is not to be expected that the amounts would change after treatment of the water by a process in which the organic matter is re moved mechanically not by oxidation and nitrification. Chlorine. The chlorine in the water was not affected by the treatment. Residue on Evaporation. The suspended matter in the river water was completely re moved in a majority of cases by the treat ment, as well as some of the dissolved mat ters. Whenever the effluents had a degree of clearness of No. 4 or No. 5, there was an appreciable amount of mineral matter sus pended in it, but it could not be satisfactorily measured. I i.rcd Residue on Evaporation. These re sults are given as a matter of record for com parison with corresponding results of the water before treatment. Alkalinity. The alkalinity of the effluents, caused chiefly by lime (carbonate and bicar bonate of calcium), was less than that of the river water by a quantity almost directly pro portional to the amount of sulphate of alu mina used in the treatment of the water. In some instances the alkalinity was exhausted, due to an excess of sulphate of alumina, and the effluents were acid. Dissolved Alumina. As a rule the analyses indicated the effluents to be completely free from dissolved alumina, although at times mere traces were noted in the course of anal ysis. The question of alkalinity and dissolved alumina have already received careful consid eration in Chapter III. Iron. With the possible exception of very slight traces of dissolved iron, all of the iron in the river water was removed except when the effluents were quite turbid. In many cases it appeared that iron was contained in the par ticles which made the effluents turbid. Table No. 2. In this table are recorded the results of mineral analyses of the several effluents dur- WATER PURIFICATION AT LOUISVILLE. ing the period of continuous operation, from March 23 to 29, inclusive. During this time samples of the effluents were collected con tinuously by automatic devices. As a matter of convenience for comparison, the mineral analyses of the corresponding river samples are also presented. These results are of value in showing the constituents which composed the mineral matter in the water before and after treat ment. Table No. 3. This table contains the results of micro scopical analyses of the effluents for algae, diatoms and such micro-organisms as may be enumerated and classified with the aid of the microscope, and without the aid of special methods such as are necessary in the case of the bacteria. Practically speaking, the effluents were free from this class of living organisms. Table No. 4. The results of the determinations of the numbers of bacteria in the effluents are re corded in this table. The samples were given a number in the series which included also the samples of river water. In addition to the hour of collection of the sample a rec ord is also given of the " run " during which the collection was made. This facilitates a detailed study of the entire records, including those of the following table. A run was re garded during these investigations as com prising all the normal operations of the respective systems, from the first opening of the valve on the filtered-water pipe, following a wash, to the next succeeding similar opera tion. The rate of filtration, expressed in cubic feet per minute and million gallons per acre per 24 hours, and the loss of head at the time of collection of the samples are also presented. It will be noted that the loss-of-head observa tions were not made at the outset of the inves tigations. This was caused by unavoidable delays in providing proper facilities for ob taining full sets of observations. The period of time occupied in filtration, and the quan tity of water filtered, between the resumption of filtration following the preceding washing of the sand layer and the collection of the samples, are each recorded. Under the re marks will be found comments upon special features in the operation of the respective systems in association with the sample ana lyzed. A series of letters will also be noted under the column of remarks. They serve as a guide in showing the basis upon which the average bacterial efficiencies of the respective systems were computed, as follows: A. Those abnormal results which were ob tained at the extreme end of a run just prior to washing, and which are not included in averages. B. The results of special samples collected in special places, and of those taken after the system had been out of operation for periods of greater or less duration, both of which were therefore not included in averages. C. When two sets of bacterial samples were collected, one set taken " all at once " and the other collected by an automatic sampler and covering a long period, only one set of results were used for official averages. Those results designated by the letter C were used only as checks. D. Those results were excluded which were obtained at times when the operations were under conditions known to be abnormal, and which were in the majority of cases caused by the Water Company. E. Long series of results on certain runs, when the automatic samplers were in use, were excluded from the daily averages, but were used exclusively in obtaining the aver ages for those particular runs. Table No. 5. This table contains the records of the operation of the respective systems tabulated in the form of runs. As stated above, all normal operations of the respective systems from the first opening of the valve on the fil tered-water pipe following a wash to the next succeeding similar operation composed a run, according to the system of records employed in these investigations. The several head ings, under which the data upon each indi vidual run are grouped, are defined as fol lows: COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. Period of Operation. This includes all the time devoted to normal operation of the sys tem. Period of Service. The time during which water passed through the pipe provided for the finished product, i.e., the period of effec tive filtration. Period of Wash.- The time occupied in preparing the filter for filtration, comprising the time occupied in washing the sand layer, filling the filter, and wasting the filtered water when considered necessary. Period of Delay. The time which was not used in normal operation of the system from the beginning to the end of the run. Quantities of Water. These are all ex pressed in cubic feet as actually recorded by the meters, except the unfiltered waste water, which was determined from gauge observa tions. Applied Water. The total quantity of river water treated by the system. Filtered Water. The total quantity of fil tered water turned into the outlet provided for the finished product. Filtered Waste-water. The total quantity of filtered water which was wasted. Unfiltered Waste-water. The total quantity of unfiltered water which was removed from above the sand layer prior to washing. The remaining headings are self-explana tory, but attention may be called to the sum maries for each run of the following data, dealing with the efficiency and economy of purification: 1. The amount of sulphate of alumina ap plied to the river water in grains per gallon. 2. The estimated amount of suspended matter, in parts per million, in the river water. 3. The average number of bacteria per cubic centimeter in the river water and in the effluents. 4. The maximum and minimum number of bacteria per cubic centimeter found in the effluents. 5. The average bacterial efficiency, which was computed by taking the percentage which the difference in the average numbers of bacteria in the river water and in the effluent was of the average number of bacteria in the river water. Special features are noted under " Remarks." Those runs marked with a star (*) were made under abnormal conditions and are excluded from subsequent averages and summaries. WATER PURIFICATION AT LOUISVILLE. UOJI 8888*o 88888888?8S8Sg858-^^c?^5 < S^n-3 < SSS ooooooooooooooooooooooooooooooooooooo Buimn V P3A|OSSI(J ooooooooooooooooooooooooooooooooooooc X]IUI|EJHV C-O CT-C^CT^OOOO^C^O^OO^ OOO co crt oo O O Ooo t-^ r~- t^- tn \rt u~t >rj tn in 7n in m m in m s = P3AIOSS1Q ??H^1 ^Hsffsl : ^ : : : S : : : ff : : : : : : :::::: iSji papuadsns OOOOOOOOOOOOOO -O---O---O t/5 S E^ |Kinx E?5if?!s3ff3^5*5Ss$a5 j j :& j :5 : : : ? : i : :5 H t/3 > 6. > w . paAIOSSTQ Hl~~<?; isiJUi M M i ; M ML- i ! w > D l- papusdsns oooooooooooooo--o-*-o-*-o u w & |oj. HIHSSI&R=5ffaaalHI ; ! i&j H j H ! I N DOooir>u-iOO-n-iOO-*OT^-.t^-TTt- . -o -t- * -to tJ ol auuono SS^S^^K^S^S^S LUiiji^IiLili^ u X H 31HJ11 N 000000060006006606600 -o -o -6 -- O (/) . c 7 ~^ limy 4j O O O O O O O*C O O*??" o o o ^^r^ ^ So ^Hr-?^ w w 1 ^ w w 2 "2 ww W 0) s I l -l *-* ~ O i- 1 0) ~ *! ^ I PJAIOSSIQ w w c S. <D H fc t ~ <; O nj a s < papusdsns OOOOOOOOOOOQQO Q -OOQOO OO H u^i w e IEJO.L O NOOco -tO -1- W tOOco CO -tO OO-Ocoo O D O O M -1-n M O >-vCoOOvO f) >. < < 55 < -r . .0 . - .o-^^ CT U < - - SS3UJE3|; ) JO 33J33(I a saaaS-iQ :::::: o : : : : : moo o-oo -ooow-tNooto.nm-T.r>Hi^M-t : : : : u njujsduiaj. . . ; ; . . ; .cj^ooo^-rco^coc.r.^r.c.c.^^^cnu,. . . . b O to H D i/) Pf l|^ ** s ^ s ^ tn ^ U OS u Sol | li* uEg K" ^ ui . t-- Q oo O^H N cO tjnoco OtH WI^-O i~>O^M c^ jOvO CT>O M d ^ enO t^co O O O O O O i f , C4 C 1 Cl Cl C) <TI Cl C-1 C^CIM HMMMMMMMMMHCInMn O 2 Q 43q un N 1 E M3S 10 "2 ^S, g 3"8 S 1 ?,S^?S5o o^ftKS<g > S ra c?S> w So- 5= ? "8 o 2 2 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 113 $ E 7 .5 2 .2 <0 in = r c " w * u "X rt rs > i pspuadsng tn * * -O -O -O -"O"-" MI - l OOOONOOOOO t - OC 6 -o -oo O -o odoodooodddddoo 666000 c oooooooooooooooooooooooooooooowoooooooo O O O papujdsng O O O OOOOOOOO -OOOOOOOO OOOOOOOO -OOOOOOOO T- - o-co oo O - oo o N TO o o" "o <?? o o & o S 1 ? o" o" o" ? "8 o > 5 o r o > oooooo? pspuadsns 3UJE.1Q ,0 MJ33Q O Q O OOO OOOOOOOO---OOOOOOOO oooooooo O -O -OOOOOOOOOOO -OOOOOOOOOOO OOOOOOOOO Q A WATER PURIFICATION AT LOUISVILLE. oooooooooooooooooooooooooooooooooooooo njy psAlossjQ OOOOOOOOOOOOOOOOinOOOMMiHMOOOOOOOOO OOOOOO OOOOOOO OOOOOO OOOOOO OOOOOOO O OOOOOO . 4 oo?o ooooooooooooo > ? o> C3 "oo^oo ooooooo oo ooooo OON^t -1-O O ooo oboo o o q o o o -O-t-t <? OOOT ooo* oooooo ^OO M * to OOcxJooO tOO OtO Oooo tttO t oooooooooooooooooooooooooo OO oo oo ?o"o o 0*0 *o o"o"o t oo 1 o"oo > o o o 1 "c"o"o o"o q Ht o o o o 0*0 o o oo 00* 3U4E310 JO 33433Q 3 S33j33a o o o o o o o o o o o o o o COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 115 Se pspuadsng oooo ooo dodoooooooooooooooooooooooooooo oooooooooooooooooooooooooooooooooooooo 00000 pspusdsns o o o O O oooooooooooooooooooooooooooo ooooooooooooooooooooooooooo iO *-< o c* r^ 1- O < S31141IM SB o o q o o o o q o o 5 q o q o q o 5 o o O "t" ~f O * O o o<?o"ooS S <?oooooooooooooo > S o 1 ooooooooooooooooo^oooooooooooo oooo coooo paHinsuo^ uaaxxo dd6d6dd66dd6dd-d66" -" ->i-> -. - <3l o oooo M o o o oo oo ^o o o o ?" o"o o " o o o -" : : !?1. < :." SS-2 1 : : """2"^ - : : ^ r!i O O t< to rl- . ^L " T 1 ?T c ?T_7777.7--7T? t l J < "1 t i < ?JL > n6 WATER PURIFICATION AT LOUISVILLE. papuadsng OOOOOOOOOOOOOOOOOOOOOOOi-iOOOOOOOOOOOOOOO oooooooooooooooooooooo ooooooooooooooooocoooo oooo oooooooooo 0000000000 ) O " T * O "-"OOWc^-t-rr- ooooooooooooooooocooooooocooooooooooooo .., o o oo & o"8 o o" o"o o" o"? ? o boo b -^- -to co O O -i- o o <? o o o ooo o oo o^o c o v b o o o c o o S"o " "8 o~ o~ o"o o" & & jsqumN IB|JSS o o o &o J o cTcf o o o ^ ? S 1 o"o o"?"?- o > o r o>? " 1 o > o o T o oo"o" o"2 o" 1- Jj t -t- i o" 5 a- o> ^ M I I I I I O O O M M COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 117 s/ S; o 2 :? 8888808888888888888888^ 8 8 s : 8 oooooo ooooooooooooooooooo ooodo o o" "6 O O O O f O O co O oooooooooooooooooooooooooooo pjpuidsns pspuidsns papuadsns oS?8 : "3 t > o 5. t" "SSSR . m t r^ . 00000 jOOOOOOOOOOOOOO o o o o o o ^oSo?s ; ?""" 2, t" - S S S R Si ~S PPS ct t : i?;?^:? ijJ^3aM8*&S3 S: aff!j O oooooooooooooo o o o o o o O O CMC i- ~ "-- ; ci : - I O O O O O -iO *O ro M M W PH o O O O O ooooooooooooooooooooooooooo 85t-OOQ QQQQOQQOQO | - ll ^ - -t-HiHi ooooo ooooooooooooooooo o o o o o o o SS : O O O O O O OOOOOOOOOOOOOOOOOOOO > S^-i t -i-iM^? O * -t <D 30 O O C d e^i N 3UJMQ JO 33J3SQ JO B U*M 4 rt CC/1 O ^nuiOl^- 1 - 1 >-vO O C^ O C <*"> "- 1 >- 10 < : 1 1 1^ N- *O CO -t M C* M M M C4 C4 M ->O o O t^ t~* r^ r^oo u8 WATER PURIFICATION AT LOUISVILLE. 1- in OM to r^ Tf -M-co g>co o O CO CO IS ? M O r* R m m.n mg. O;N 0000000000000000 o o o o O o O O O O O O O oooooooooooooooo o o o o O O O O O O N ^.n-*oOMOoeoqwcoo mm * in o K co. 11 en o HI o R ," K?,?, SSSSS,aS,S,K ss "3 -T 5 SSS S>S5 O i-i O O - O O o * I-*- o * c*i c*i CN o^ r~-o o C 1 >-< . >,< rH (I) &. On" ID Q, HI O O ?H,Ii o C OH. < o < o o HI HI > in -t r^ -t HI o O O O O O O M O O OO O -OOO HI w o -i-i *co in -) w O HI o HI COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 119 O Q "-" C<" en C 1 O O ** w "* r>- Tf o *O CO ^O s CO W Tfr^cOM in * w \r> \r> N M r* in * co C OOO OO- OOOf> -tOnNO>-OOO - - OOOOOOOOOOOOOOOO^ -tOOO ONOOOOOOOONOOOOOOOOOOO 00000000000066666 s g ~ .BUIUiniV P3AIOSS1Q OOOOOOOOWOOOOOOOOOOOOOOOOOOOOOOOOOOO^Ow P3A10SSIQ pspuadsng O r^ O O O f> O> O papuadsns O ^ O O M o o o o o oooooooooooooo "OOC^OOOOOC oooooooooooooo o ** o c- o pnox auuojiQ oo to o o OOOOC050 D V JJ ^.%^ * doddddo o paAIOSSjQ "TO -1- vi 0000 5"? 3" O N ?? oootoo-rotOTi--t-rrfoo oo f o oo????oo?o o O papuadsns ! : 1 ; : i i 1 i |!0 X 1-cn -1- >xi v O O O O C - ( ha i p : CONONWOOtC ^o o 1 ??? 1 ??;: O N : ?? T O O -TCO -TO -TO T T f -KO O^ t ^ooo?oo???o oooo 1 oo oc oo O "* O 0* r~ r^O CO Tt^W - N-T-T-T (no O O*OO "^ O OOOOOOOOOOOOOO- O - OOO o N TO o* ( TNOOOOOO O *-" _ . WATER PURIFICATION AT LOUISVILLE. "8 o" ?"o o "o ? q SJq q qSqo Soqqqo qqqqSSoqoo oqcSqqq ddddddddddddddddddoooooooooooooooooooo - OOOOOf^Nr-NOOOOOOOOOOOOOOOOOOOOOOOOOOO oooooooooooooooooooooooooooooo O oo I-H to c O O - O -< " -< O "O NO pspuadsns OOOOOOOOOOOOOOOOOOOOOOOOOOOOOO > o a* o o H o 888888888888888888118888888888888 coooooooobooooooooooooooooooooocooooo o OOOOCOOOxO^OCOOOOOOOOOOOOOOOCOOO pauinsuoo U^AXQ " M c 8 o o M c o o --- COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 121 poooooo oqqqoqqowoqqqqoqqqqqoqqqqqo~oq 6666600 6 06006666066666 6 66666666666666 OOOOOOOOOOOOOOOOOOMMiOOOOc^OOOO^OOOOOOOO papusdsns ooooooooooooooooooo OOOOOOOOOOOOOOOO O OOOOOOOOOOOOOOOOOOO o Ouir^o O"t"O O CT r^ ooooooooooooooooooooooooooooooooooooo oooooooooooooooooSoooooooooooooo ooo ooo o > o r o"o o" I 6 o o o o o o? 1 ? o 6"o >1 o"o o > "o o o S o 0*3 o"o"S o ? o" G^C* o * 1-1 ci r "O "1-TNOCO ^>co HH *roo TO>- *O O-1-OO "T"t-O u^-T-ftn r fi \n O OOOOOO- OOOi- OOOOi-iOOOO -OOOOOOOOOOOOO-O" WATER PURIFICATION AT LOUISVILLE. gj^ s?> L7 - -t - - - -- j - S- 3- 3s 55 8 of. >FS S H" o o o o : o o o O O O o oooooooo o o o o S^M^ S 8i - ~ "; S ; " "S-S-Ssao S o * ^s O> O O t -,. -T ~ f . n ^ o -t ^.S-S R??^ RS.3S o o o o o O O o D oooooooo 0000 O f^O /-,-fO -r oo-oo-oooodd6d6 a> o * S ^ oooooo l oooooooooooo?o r o oo"o o 1 OOOOOOOOOOOOCOi "^O O O >> CO o o o o ** o O Cl rf TO O v ooooooooooooooo i O oo -T r- O^ tug M^ Jo oooooooo-oooooooooooooooo r u^ O> T COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 123 1" l>s s 6 o * c " w ft * < I is 8-5 ~ O - - Ir^MN fweo -r^o - 0^00-0000 ooooo -oooo -oooooooooooooo ooooo o o o o -oooooooooooooo o o o 1 ^ o o?o > ?ooooooooooooSoo?5 v o 1 oo o o* ? ?~o ooooo linsj 1FU3Q I *nO O i^ r^-tr) c/j c> N M c^ i^j -t- -t in mO l^ l^oo O C?" O - CH to f"i -t inO l^^^S.^* ] I2 4 WATER PURIFICATION AT LOUISVILLE. H S q q w w d 6 d d papuadsns o a : C oooooooooo o o o o OOOOO -OOOOOi-ii-iOO COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 125 H " U) ~ "o 8?q8qqq T 8q2q > q q > 8" q "q q <? o"q^-C ZT 5 q"q q q q q S 00001-006606006666666666666606660666006 00000000 00NN Xu Sh oooooooo oooooooo O O O O O O popuadsns oooooooo OOOOcoOOO -Tc*-]c*">CH-t--rc*">r-- oooooooo c*^ ^f -TCO o o o o o o f o > o r o o ?o o"o o 1 ?" o o o o o cTo r S 1 o ?5 > o"o o~o l "o o > o 1 o o o 1 8 o o"o o o" o o " o o o o o o ESQ jo aajijaa 111 r>. a--" w - 126 WATER PURIFICATION AT LOUISVILLE E H a * oooooo -oooo oooooo o o o o 6 o d d d o d 6 d d d d 6 6 d 6 d 6 d 6 d 6 o o o o o oooooooooooooooooooooooooooo oooooo OOOOOO-O -OOOOOOOO-O-OOOM- COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 127 TABLE No. 2. RESULTS OF MINERAL ANALYSES OF THE OHIO RIVER WATER BEFORE AND AFTER PURIFICATION BY THE RESPECTIVE SYSTEMS. (Parts per Million.) ce of Sample. Silica (SiO.,) Oxide of iron .(Fc 3 Oj) Alumina (AljO,) Oxide of manganese (MnO) Oxide of nickel (NiO) Lime (CaO) Magnesia (MgO) Soda (Na,O) Potash (K,O) Chlorine . . .(Cl) Nitric acid (N,O 5 ) Carbonic acid, combined (CO 2 ) Sulphuric acid (SO,) Phosphoric acid (PjO s ) Ri 299 . 50 39-45 76.55 2. 2O I .09 3 -/o 13.98 8.48 18.15 5-57 14.67 2I.2S 23.28 0.79 \v , Kfflu 4-25 0.15 2.05 none less than I .o 30 . 60 7-03 5.02 8.19 5-54 14.67 7-75 35-71 0.98 Jewell Kffluent. Western I res. Eff. 4 00 o. 1 1 0.24 trace less than I 33-22 6.64 3.56 7.85 5-78 13.89 13.12 33-37 o.Si tra less thi 34 6, 8 41. 0.47 TABLE No. 3. RESULTS OF MICROSCOPICAL ANALYSES OF THE EFFLUENTS OF THE RESPECTIVE SYSTEMS (Number of Organisms per Cubic Cemimeter.)* 1896 Feb. 18 9 26 March 4 March 4 April May June April May June July Feb. March May June uly Number of Sample. Total Number, Effluent of the Warren System. 278 Diatomaceae: Synedra, 8; Chlorophyceae: Protococcus, 2; Infusoria: Parts of cases, 6 281 Oiatomaceae: Navicula and Cymbclla pr 305 No organisms present 330 Diatomaceae. Synedra, pr; Fungi: Crenothrix, pr 355 Diatomaceae: Synedra, 3 383 Diatomaceae: Meridion, pr; Miscellaneous: Anguillula, I 405 Miscellaneous: Anguillula, pr 460 No organisms present 518 Chlorophyceae: Protococcus, I ; Miscellaneous: Vegetable Hairs, 23 545 Diatomaceae: Synedra, i; Cymbella, i; Chlorophycene: Proto coccus, i 567 Chlorophycete: Protococcus, pr; Infusoria: Trachelomonas, pr. . . 588 Diatomaceas: Synedra and Cyclotella, pr 630 Chlorophyceae: Conferva, pr 652 Verities: Ploima, 26; Miscellaneous: Zoospores, 16 Effluent of Jewell System. 282 Diatomaceae: Syr.edra, Navicula, pr 306 .Diatomaceae: Pleurosigma, Cymbella, pr; Chlorophyceae: Proto coccus, Scenedesmus, pr 328 :No organisms present 353 iNo organisms present 381 iNo organisms present 406 [Fungi: Mould hyphae, pr 461 i No organisms present 472 Chlorophyceae: Protococcus, pr 517 |Chlorophyceae: Protococcus, 16; Miscellaneous: Anguillula, I .... 546 | Miscellaneous: Zoospores, pr 566 jDiatomacese: Synedra, pr; Chlorophyceae: Prolococcus, i; Scene desmus, 2; Miscellaneous: Zoospores, pr 587 Diatomaceae: Synedra, pr; Cyclotella, i; Chlorophyceae: Proto coccus, 2; Pandorina, pr; Endorina, pr; Vermes: Plorima, pr.. 63 1 Chlorophyceae: Protococcus, pr 653 No organisms present 685 Infusoria: Monas, 12; Miscellaneous: Zoospores, pr 709 No organisms present Effluent of Western Gravity System. 307 No organisms present. . . 331 No organisms present 350 Chlorophyceae: Spyrogyra, I 380 Chlorophyceae: Protococcus, pr Effluent of Western Pressure System. 308 No organisms present 332 No organisms present 350 No organisms present 374 Chlorophyceae: Piotococcus, I 407 No organisms present 565 Chlorophyceae: Protococcus, 5 586 Diatomaceae: Cyclotella, pr 654 No organisms present 718 Chlorophyceae: Protococcus. pr . .._._ *pr present. 16 pr. o pr. 3 i pr. o 24 3 pr. pr. pr. 42 128 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. RESULTS OF BACTERIAL ANALYSES OF THE EFFLUENTS OF THE RESPECTIVE SYSTEMS. Warren System. Ra eof j 8 y Collected. Filtr; lion. OJ 15 U, Period of ^ bi" U aj s Number I Js. a ||| a of j ^ " E 3 Washing. ) "- " Remarks. z Date. Hour. Run. u! O U 3 o Hours and Minutes. Is3 P |s = k? $ Jj 03 u i * 1895 Oct. 21 12-35 P-M. I 2h 35m. 47 D 6 " 21 3.40 " I 5h. 4001. 34 " 22 11.51 A.M. 2 4 . o I5m. 88 I I " 22 12.07 l -M. 2 4.0 132 3im. 32 12 " 22 I. 10 " 2 3-5 129 ih. 34m. 54 16 " 22 o rn 2 4h. 1401. 64 21 " 23 J J^ IO.OO A.M. 2 I.O us 4h. 24m. 36 23 " 23 11.13 " 2 2.O 121 5h. 37m. 48 25 " 23 12.05 r - M - 2 I.O "5 6h. 2gm. 44 27 " 23 1.25 " 2 2.0 121 7h. 4gm. 34 29 " 23 2-33 [ 2 2.0 121 8h. 57m. 72 31 " 23 2 2.O 121 loh. 54m. 72 35 " 24 9.56 A.M. 3 I.O "5 i6m. 239 34 36 " 24 10.11 " 3 3.0 126 3im. 518 99 37 " 24 I0. 5 I " 3 3.0 126 ih. um. 1 504 37 39 " 24 12.22 I .M. 3 2.O 121 2h. 42m. 3472 49 41 " 24 I. SO " 3 23-5 I2g 4h. lom. 5 515 196 43 " 24 3-<>5 " 3 24.0 132 5)1. 25m. 7326 53 44 " 24 4.01 3 25.0 137 6h. 2im. 8535 53 " 24 5.o(> " 3 7h. 26m. 10 197 2 3 5 " 25 9.58 A.M. 3 25.0 137 7h. 53m. 19787 36 52 " 25 I 1. 12 " 3 25.0 137 gh. 07m. 12 600 48 54 " 25 12.12 " 3 25.0 137 loh. O7m. 14 117 38 55 " 25 1.28 r.M. 3 25.0 137 nh. 23m. I60I6 34 57 " 25 2.38 " 4 27.0 148 28m. 427 93 60 " 25 3-35 " 4 24.0 132 ih. 25m. I 825 46 64 " 25 4-34 4 25.0 137 2h. 24m. 32/4 33 66 " 26 10.56 A.M. 4 25.0 137 4h. 26m. 7 205 65 69 " 26 1. 12 P.M. 4 24.0 132 6h. 42m. 9493 31 71 " 26 4.40 " 4 20.0 no loh. lom. 13 527 32 73 " 28 12.10 " 5 26.0 M3 45m. i 127 61 74 " 28 1. 10 " 5 26.0 M3 ih. 45m. 2 671 28 75 " 28 2.30 " 5 23.0 126 3h. 05m. 4554 28 76 " 28 3.23 " 5 25.0 138 3h. s8m. 5 802 3 79 " 29 12.33 " 6 27.0 148 ih. 33m. 2357 42 81 " 2g 2.00 " 6 28.0 154 3h. oom. 4676 20 82 " 29 3-25 " 6 JI..O M3 4h. 25m. 7 O2 3 18 85 " 30 9.55 A.M. 7 24.0 132 55m. i 290 46 86 " 3 IO.I8 " 7 27.0 148 ih. iSm. i 866 58 87 " 30 10-37 " 7 26.0 43 ih. 37m. 2330 18 88 " 30 11.17 " 7 28.0 154 2h. I7in. 3385 15 Sg " 3 12.25 I -M. 7 27.0 148 3h. 25m. 5 118 10 30 1.40 " 7 24.0 132 4h. 4om. 7052 II 93 " 30 3-22 " 7 28.0 154 6h. 22m. 8629 16 97 " 30 4.30 " 7 7h. 30:11. 10 689 14 I OS " 31 2.31 " 8 23.0 126 5h. 26m. 6262 9 113 Nov. i 12.07 8 25.0 137 7h. I2m. 8 OK 13 I I [ " i 3-15 " 8 loh. 2om. 12 23 14 Shut outlet 3.15 r.M. IlS " i 4.03 " 9 22.0 121 38m. 1062 16 I2C " i 4-33 " 9 22.0 121 ih. o8m. I 732 II 124 " 2 11.07 A.M. 9 24. c 132 2h. 52m. 4099 15 I2t " 2 12.31 r.M. 9 26.0 143 4h. i6m. 6 232 14 I2S " 2 1.26 " 9 24.C 132 5h. nm. 7 599 43 130 " 2 3.48 " 9 24. c 132 7h. 33m. 9344 19 13. " 4 2.32 " o 26. C 143 ih. 25m. i 802 42 135 " 4 3-37 " 27. c 148 2h. 3Om. 3474 38 13: " 5 9.23 A.M. o 24. c 132 3h. oim. 4 i6c 20 " 5 9.46 " o 26. c 43 3h. 24m. 4 73^ II 142 " 5 IO.O7 " o 3h. 45m. C 25 10 14: " 5 10.32 " 25. c 137 4h. lom. 5 82C 12 " 5 11.30 " o 24. c 132 5h. oSm. 7 2IC 14 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 129 TABLE No. 4. Continued. Warren System. Rat eof *j u Col ected. Filtr uion. i/i . J ^ M <u Period of - ^ ^ 1 Number S. _o a T3 Last 3 Z i. > e Run. u . 5 " ^ I Washing. 3-?\L O.U Remarks. 5 Date. Hour. tj it! Hours and Minutes. Is ! P X o i a " 1 03 151 Nov. 5 12.50 P.M. IO 24.0 132 6h. 28m. 9079 13 153 5 3-50 " IO 16.0 88 9h. 28m. II 744 10 rcS " 6 18 I 50 159 6 11.13 " II 24.0 132 ih. 58m. 2552 15 100 6 I.I9 P.M. II 24.0 132 4h. 04111. 5661 17 161 6 2.55 " II 24.0 132 5h. 4om. 7952 24 165 " 7 9. 10 A.M. II 24.0 132 6h. 0301. 8489 138 1 66 7 9.27 " II 25.0 137 6h. 2om. 8940 82 167 7 9-49 " II 24.0 132 6h. 42m. 9452 54 168 7 10.27 II 24.0 132 7h. 2om. 10385 21 173 7 11.25 A.M. II 23.0 126 8h. iSm. ii 725 30 175 7 12.24 P.M. 12 30.0 16=; O2m. 26 196 176 7 12.30 " 12 27.0 148 o8m. 191 54 177 7 12.50 " 12 28.0 154 28m. 751 48 178 7 I.og " 12 24.0 132 47m. i 245 61 1 80 7 1-35 " 12 26.0 143 ih. 13111. 1938 46 189 7 2.27 " 12 25.0 137 2h. 0501. 3342 31 193 7 3.00 " 12 24.0 132 2h. 38m. 4 180 35 197 8 9.50 A.M. 12 24.0 132 3h. 4301. 5 "3 49 199 8 11.05 " 12 23.0 126 4h. 58m. 6813 42 202 " 8 12 "* 7 P M I 2 6h. 30111. 8 822 K 206 " 8 1 " j 1 2.13 13 24.0 132 oym. 152 1 80 208 8 2.27 " 13 23.0 126 2im. 542 70 210 " 8 2.50 " 13 6.0 33 44m. i 050 2f 214 " 9 11.33 A.M. 13 24.0 132 ih. 22tn. 2015 4f 218 9 1.27 P.M. 13 25.0 137 3h. i6m. 4802 50 221 9 2.30 " 13 24.0 132 4h. igm. 6328 58 227 ii IO.47 A.M. 13 25.0 137 7h. oim. 10215 224 230 " ii II. IO " 3 23.0 126 7h. 24m. 10745 136 231 " ii 2.32 " 14 24. o 132 i 637 41 30-1 25 1 5 35m I 58.1 ^J 306 " 25 II.OO " 15 24.0 146 ih. I5m. i 096 468 3" 11 25 12.20 P.M. 15 20.0 121 2h. 35m. 2 044 390 313. " 25 1.40 " 15 21. O 127 3h. 55m. 3 7: 414 315 " 25 3-30 " 15 25.0 152 5h. 45m. 5732 294 318 " 26 9.28 A.M. 15 23.0 I4O 6h. i8m. 6745 496 322 " 26 10.20 " 15 21. O 127 7h. igm. 7901 372 324 " 26 H-34 " 15 :< LI 121 Sh. 24m. 9437 328 326 " 26 2.IX) P.M. "5 20.0 121 loh. 5om. 12445 344 332 " 27 9.27 A.M. 15 II. O 67 nh. 52m. 13667 6Sc 335 " 27 10.24 15 18.0 109 I2h. 49m. 14546 452 338 27 n-53 " 5 25.0 152 I3h. 42m. 15890 446 340 " 27 1.44 P.M. 15 23-5 143 I5h. 33m. 18488 564 343 27 3-17 " 15 20. o 121 I7h. o6m. 20 310 512 347 29 9.52 A.M. 1 6 II. 67 O2m. 31 I 302 348 29 10.03 " 1 6 IO.O 6 1 I3m. 132 1 169 349 29 10.14 " 1 6 IO.O 61 2401. 248 90! 350 29 10.23 " 1 6 IO.O 6 1 33iii. 355 1092 351 29 10.32 " 16 IO.O 6 1 42m. 507 876 352 29 10.43 " 16 21. O 127 53m. 762 760 354 29 10.54 1 6 2<).i 121 ih. 04111. I 006 344 356 29 1 2. 06 P.M. 1 6 23-5 143 2h. i6m. 2458 624 358 29 1.58 " 1 6 23.0 140 4h. o8m. 4968 448 368 30 IO.4I A.M. 16 - ( I 4 6 7h. 17111. 9416 632 370 " 30 11.51 " 16 Jli.i 158 8h. 2710. u 043 840 372 " 30 1.39 P.M. 16 25.0 152 loh. ism. 13 691 984 376 Dec. 2 9.47 A.M. 16 20. o 121 I2h. ism. 16 533 77 378 " 2 10.48 16 24.0 I 4 6 I3h. i6m. 17949 945 381 " 2 12-35 P.M. 16 23.5 143 I 5 h. O 3 m. 20 462 875 383 " 2 2.38 " 16 23.0 140 I7h. o6m. 23 122 i 078 385 3 9 35 A.M. 16 25.0 152 i8h. o6m. 24401 826 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Warren System. Rat . O f S 8 Collected. Film tion. <S1 . Z ^ u c iJ Period of * e fl Number 8. S. a a Last Slcti ^ si e 3 7, Date. Hour. Run. fc 3 O 3 its X Washing. Hours and Minutes. || II Remarks. C fi = O.B I ~>4u u;j </) u S J U, CO 1895 387 Dec. 3 10.35 A.M. 16 25.0 152 igh. o6m. 25866 532 389 3 11.41 " 1 6 24-5 149 2oh. I2m. 27334 665 391 3 1.02 P.M. 1 6 22. O 133 2ih. 33m. 29349 I 036 392 3 2.O9 " 1 6 24.0 146 22h. 4om. 30 929 I l6g Shut inlet 2. 01 P.M., out 395 3 3-00 " 17 21.0 127 I2m. 213 462 let 2.21 P.M. 396 3 3.10 " 17 22. O 133 22m. 401 490 397 3 3.2O " 17 22. O 133 32m. 598 392 398 3 3-31 " I/ 25.0 152 43m. 836 406 399 3 3-4 " 17 24.0 146 52m. i 036 399 400 3 3-50 " 17 21. O 127 ih. O2m. i 259 399 401 3 4-49 " 17 23.0 140 2h. oim. 2498 334 4<>3 4 10.40 A.M. 17 21. O 127 3h. 49111. 4898 3i? 405 4 II.06 " 17 2O. O 121 4h. ism. 548J 37S 406 4 11.26 " 17 21. O 127 4h. 35m. 5908 357 407 4 11.45 " 17 22.0 133 4h. 54m. 6315 322 408 4 1. 12 I .M. 17 25.O 152 6h. 2im. 8 281 5iS 409 4 2.52 " 17 21. O 127 Sh. oim. 10591 594 412 4 4.19 " 18 24.0 I 4 6 nm. 321 548 413 4 4.29 " 18 22. O 133 2im. 533 47" 4i5 4 4.42 " 18 24.0 I 4 6 34m. 886 320 416 4 4.50 " 18 23.0 140 42m. 95< 3* 4 7 4 5.OO " 18 2O. O 121 52m. 1136 254 418 4 5.II " 18 23.O 140 ih. 03m. i 239 280 421 " ^ 9.58 A.M. 18 2h. oom. 2 33 260 423 5 10.42 " 18 24.0 I 4 6 2h. 54m. 3468 1 80 426 5 11.52 " 18 24.O I 4 6 3h. 54m. 5 124 236 428 5 2.44 P.M. 18 24.O I 4 6 6h. 46m. g 161 336 435 5 3-40 " 18 24.0 I 4 6 yh. 42m. 10464 472 439 6 11.07 A.M. 18 22. O 133 loh. 38m. 13 12; 386 441 6 11.17 " 18 24.0 I 4 6 loh. 48m. 14727 690 444 6 12.50 P.M. 19 18.0 log 1 2m. 131 524 445 6 I.OO " 19 18.0 109 22m. 295 476 446 6 I.IO " 19 20. o 121 32m. 522 476 447 1.20 " 19 22.5 126 42m. 738 476 448 6 1.30 " 19 23.0 140 52m. 97< 440 45" 6 1.40 " 19 24.0 I 4 6 ih. O2m. i 175 412 451 6 3.3S " 19 22. O 133 3h. oom. 3934 660 454 " 7 10.17 A.M. 19 23.0 140 5)1. 5im. 7932 450 456 7 12.27 P.M- 19 25.0 152 8h. oim. ii 981 412 459 7 3-5<> " 20 23.0 140 nm. 182 254 462 9 10.10 A.M. 20 20. o 121 2h. 1 901. 3082 324 464 9 11.14 " 20 21.0 127 3h 23m. 4445 250 466 9 12.17 I .M- 2O 24.0 I 4 6 4h. 26m. 5 7<>5 270 469 9 i-57 " 20 24.0 I 4 6 6h. o6m. 8o6c 272 47i 9 3-3 " 20 24.0 I 4 6 7h. 39m. 1037 324 475 o 9.17 A.M. 2O 23- I4O gh. I3m. 12 8IC, 312 477 " o 0. 2O " 20 23.O 140 loh. i6m. 14 oil 406 484 " 2.23 P.M. 2 22.0 133 2om. 325 288 485 " o 2-33 " 2 21. O 127 3Om. 53< 344 486 " o 2.43 " 2 2O. O 121 4om. 779 330 487 2-53 " 2 22.0 133 5om. i 032 324 488 " o 1.03 " 2 24.0 I 4 6 ih. oom. i 247 246 489 " o 1.13 " 2 24.0 I 4 6 ih. lorn. 1494 1 88 49 1 " 2.10 " 2 24.0 I 4 6 2h. O7m. 2845 266 493 " o 3-24 " 2 24.0 I 4 6 3h. 2im. 4564 244 496 " I 9.24 A.M. 2 24.0 I 4 6 5h. 3im. 7489 196 498 " I II. II " 2 24.0 I 4 6 7h. l8m. 9994 210 500 502 " I " I 12. l8 P.M. I.lS " 2 2 23.O 2O. O 140 121 8h. 25m. gh. 25m. ii 56 12 92; I 9 6 2O2 Shut inlet 1.18 P.M., out 505 " I 2. 5 S " 2 21.0 127 54m. i 137 I 4 6 let i. 21 P.M. 509 " 2 9.40 A.M. 2 24.0 I 4 6 4h. 07m. 5452 128 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. TABLE No. 4. Continued. Warren System. Rate of S S Collected. Filtration. K | . U 1/5 ii Period of u 5f u > J& Number Q. o ex d ServiceSinc " !H ** g w "S w " u Last _r ** tj 4> ~ Remarks. 3 Run V -J O S = X Washing. *gfc a 6 "Z. Date. Hour. tl 13 Hours and Minutes. *- V- & SS q II (J ~.3u rt U X u s a " ,J E m 1895 5" Dec. 12 12.04 P M - 22 24.0 146 6h. 3im 8 85 228 513 " 12 3-05 " 23 22. C 133 ih. oim I 273 igc 518 " 13 IO.I6 A.M. 23 22. C 133 4h. 4Om 6i8c 1 08 521 13 4.38 P.M. 24 22. C 133 52m 1073 136 525 14 lO.Og A.M. 24 23. c 140 2h. 52m 3811 108 527 14 12.59 P - M - 24 23. 140 5h. 42m 7672 135 53i 14 3-33 " 25 20. 121 I4m 281 274 534 " If) 9.30 A.M. 25 21. 127 2h. 25m. 3 log 160 536 " 16 11-37 " 25 23- 140 4h. 32m. 6 09. 126 538 " 16 2.31 P.M. 25 21. 127 7h. 26m. 9961 172 54 " 16 5.17 " 26 23. 140 ih. o8m. 1492 1 20 543 " 17 9.30 A.M. 26 23. 140 ih. 5om. 2430 ill 546 17 12.58 P.M. 26 23- 140 5h. i8m. 7156 170 548 . " 17 3.21 " 27 23. 140 ih. o6m. I 260 no 551 17 4-37 " 27 22. 133 2h. 22m. 3001 148 554 " 18 g.2O A.M. 27 22. 133 3h. 4om. 4731 197 556 " 18 10.41 " 27 23- 140 5h. oim. 6 509 196 558 " 18 1.16 " 27 33 7h. 36m. 9382 185 559 " 18 2.32 " 28 22. 133 I3m. 234 236 560 " 18 2.42 " 28 22. 133 23m. 444 294 56i " 18 2.52 " 28 22.0 33 33m. 658 274 562 " 18 3-02 " 28 22.0 133 43m. 874 2 2O 563 " 18 3.12 " 28 2.O 133 53m- i 098 I 7 8 565 " 18 3.22 " 28 2.0 133 ih. 03111. i 306 2 7 8 568 " 18 4-39 " 28 3- 140 2h. 2om. 2 g64 I 5 8 57 " 9 9.25 " 2S 3-0 140 3h. 3im. 4474 l6g 580 9 3.23 P.M. 28 4.0 146 7h. 39111. IO 212 142 582 19 4.40 " 28 4.0 146 8h. 56m. 11987 165 587 1 20 10.14 A.M. 29 I.O 127 I2m. 234 426 589 20 11.59 " 29 4.0 146 ih. 57m. 2 590 103 591 20 2. 02 P.M. 29 3-o i |n 4h. oom. 5 397 188 594 " 2O 3-53 " 29 4-0 146 5h. sim. 7902 720 597 " 21 g.ig A.M. 29 I.O 127 7h. 47m. 10460 344 600 " 21 4-01 " 30 2.0 133 3h. 3im. 4 627 164 603 " 23 9.18 " 30 2.O 73 5h. I4m. 6551 260 606 ; 23 10.39 " 30 g.O H5 6h. 35m. 7 754 150 612 23 12.34 P.M. 30 o.o 121 8h. 3om. 9932 132 616 23 3-33 " 30 8.0 49 nh. 2gm. 13316 150 Shut inlet 3.20 P.M., out 619 24 9.29A.M. 31 3.0 140 ih. 4gm. 2466 63 let 3.40 P.M. 627 24 12.48 P.M. 31 4.0 146 5h. o8m. 6433 78 628 24 3.07 " 31 I.O 127 7h. 27m. 9772 86 637 " 26 lO.Og A.M. 31 22. 133 toh. 48m. 13578 95 638 " 26 11.49 " 32 22. O 133 32m. 688 59 639 " 26 12.04 ! -M. 32 23.0 140 47m. 994 60 644 " 26 3-49 " 32 22.0 133 4h. 32tn. 6i34 244 651 " 27 IO.O7 A.M. 32 21. O 127 7h. ism. 9392 i 701 655 27 12.34 P.M. 32 21. O 127 gh. 4201. 12 757 i 530 660 " 2 7 2-57 " 32 I2h. 05111. 15 815 924 Shut outlet 2.57 P.M. 665 " 27 3-42 " 33 24.0 146 3001. 641 230 666 27 3-57 " 33 22.0 133 45m. 978 128 671 27 4-55 " 33 22. O 133 ih. 43111. 2 223 1 60 674 " 28 9.56 A.M. 33 22.0 133 3h. iim. 4131 428 682 " 28 11.56 " 33 21.0 127 5h. iim. 6776 i 728 683 28 3.06 P.M. -IT 8h. 2im. II Oil I 350 Shut inlet 3.04 P.M., out 692 " 30 10.56 A.M. JJ 34 22. 133 2h. 3im. 2 884 402 let 3.24 P.M. 696 30 12.15 r -M. 35 21. 127 14111. 179 474 697 1 30 12.44 " 35 22.0 133 43m. 804 2IO 7 " 30 1-59 " 35 24.0 146 ih. 58m. 2 561 502 705 1 30 4-53 " 36 23.0 140 l8m. 378 210 707 3 5., 6 " 3f> 22.0 133 4im. 959 170 711 31 10.45 A.M. 36 23.0 140 2h. 34tn. 4708 406 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Warren System. Rate of ti Collected. Filtration. fc 1 J V Number g_ 1 s. T3 Period of Service Since 1 1 . u u . E of *j "^ V rr 5 Last rt f- } v ~ Remarks. p Run. OJ 0; 6 s 3 X Washing. "* - k* a Date. Hour. fc 5 Hours and 1; ~ ~ T S ~n .y c .2 v-" VI Minutes. v "a ^ X \7. u 5 a? --(-) n 1895 717 Dec. 31 1.39 P.M. 37 17.0 103 I5m. 215 440 718 " 31 2.09 " 37 23.0 140 45m. 835 278 1896 730 Jan. 2 10.56 A.M. 38 22.5 136 lorn. 130 784 731 2 11.26 " 38 25.0 152 4om. 690 600 740 " 2 2.40 P.M. 38 18.0 109 3)1. 54111. 5 260 i 400 743 " 2 4.06 " 39 20. 121 I5m. 222 405 749 3 IO.2I A.M. 39 21.0 127 3h. 0701. 3722 i So 753 3 1.22 P.M. 40 2O. O 121 iSm. I 86 220 754 3 I. 5 8 " 40 2O.5 124 5im. 956 97 765 4 11.55 A.M. 41 18.0 109 I4m. 118 I/O 772 4 2.26 P.M. 21.5 130 2h. 45m. 2558 133 776 6 11.55 A.M. 42 17.0 103 3om. 415 1 02 780 6 3.27 P.M. 42 17.0 103 4h. O2m. 3891 112 785 7 12.24 " 43 14.0 85 4h. igm. 4 159 43 789 7 3-52 " 44 12.0 73 2h. I7m. 2 218 66 794 8 11.58 A.M. 45 16.0 97 55m. 858 25 799 8 2.28 P.M. 45 16.0 97 3h. 2501. 3235 68 802 8 2.49 " 45 16.0 97 3h. 46m. 3529 84 809 9 IO.2I A.M. 46 18.0 1 09 2h. 2sm. 2 316 10 814 9 1.26 P.M. 47 16.5 IOO 1501. 21O 53 816 9 1.47 " 47 15.0 gl 36m. 500 21 826 10 1.05 " 48 14.0 85 09 m. 119 152 827 " 10 1.42 " 48 15.0 9 46m. 699 36 836 " ii II. 21 A.M. 49 17-5 1 06 2h. 53111. 2 630 36 843 " 13 12.30 P.M. 51 16.0 97 15111. 198 35 844 13 1. 00 " 51 15.0 91 45m. f>33 31 846 ; 13 2.OI " 51 16.0 97 ih. 46m. i 608 10 850 3 4-57 " 52 16.0 97 i SRI. 207 78 856 14 11.56 A.M. 52 17.0 103 3h. 44in. 3516 M 859 14 1.57 r.M. 53 16.0 97 I5m. 186 31 863 4 2.27 " 53 17.0 103 45m. 7 4 32 867 4 3-25 " 53 17.0 103 ih. 4301. i 606 25 871 15 10.40 A.M. 54 16.0 97 ISm. 181 62 875 T5 II. 10 " 54 16.0 97 4 5 m. 657 12 879 15 12.59 r.M. 54 16.0 97 2h. 34111. 2327 4g 88 1 " 15 2.54 " 54 14.0 85 4)1. 2gm. 4 124 77 886 " 16 10.43 A.M. 55 16.5 IOO 2h. 34m. 2356 78 891 " 16 12.58 P.M. 55 14.0 85 4h. 4gm. 4486 94 896 " 16 2.17 " 56 16.0 97 15111. 203 73 897 " 16 2.47 " 56 16.5 IOO 45m. 663 60 898 " 16 2.59 " 56 16.0 97 57m. 853 58 915 " 17 11.37 A.M. 57 17.0 103 15111. 176 126 920 17 I2.O7 P.M. 57 17.0 103 45m. 675 924 17 1.02 " 57 16.0 97 ih. 4om. 5 5 "65 931 17 2. II " 57 15.0 91 2h. 4gm. 2 615 i/i 935 " 17 3-53 " 58 "7-5 1 06 1 7m. 240 82 939 ; 17 4.23 " 58 15.0 91 47m. 660 76 941 17 4.58 " 58 16.0 97 ih. 22m. I 205 68 943 " 18 IO.O8 A.M. 58 16.0 97 2h. 54m. 2 620 38 952 " 18 I. CO P.M. 59 16.0 97 15111. 217 67 956 " 18 1.30 " 59 14.0 85 45m. 677 70 961 " 18 2.50 " 59 15-5 94 2h. 05111. 1887 64 968 " 20 10.38 A.M. 60 15.0 9 2h. iSm. 204O 102 971 " 20 2.05 P.M. 61 16.0 97 I5m. 176 1 08 973 " 2O 4.09 " 61 15.0 91 2h. igm. 2086 146 982 " 21 12.10 " 62 16.0 97 15111. 181 86 983 " 21 12.40 " 62 16.0 97 45m. 691 32 985 " 21 4.12 " 63 16.0 97 34m. 453 36 99 22 g.2I A.M. 63 16.0 97 2h. 17111. 2073 33 997 " 22 2.12 P.M. 64 17.0 103 ih. 35111. 1464 36 1028 " 25 2.27 " 65 ig.o 2h. O3m. 2418 60 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 133 TABLE No 4. Continued. Warren System. Rate of _ u Collected. Filtration. S c u fc. .5 & Number S. ! Period of Service Since S|- 3 o u . E of a y u Last > rt 4J - Remarks. a Run. ,** 6 S I K Washing. Hours - nd fek K E J Date. Hour. * | 0*^3 ^ Minutes. " 5 S s e 2jg is.? i = JU "u (/> <> s U. n 1895 032 Jan. 27 10.03 A.M. f>5 16.0 97 5h. 52m 6 108 II- 37 27 11.22 " 66 16.0 97 15111 216 39 038 27 11.52 " 66 16.0 97 45m 666 177 039 1 27 1.05 I .M. 66 ii, .1 97 ih. sSm. 1836 41 044 27 4.06 " 66 15.0 91 4h. sgm. 4636 412 050 " 28 9.45 A.M. 67 16.5 100 ih. 2gm. i 226 225 "57 " 28 3-33 I -M. 68 15-0 gl ih. lorn. 954 536 058 " 28 4.29 " 68 16 o 97 2h. o6m. I 924 650 065 1 29 10.09 A.M. 68 16.0 97 3 h. sSm. 3504 327 068 29 1.55 I -M. 69 15-0 91 ih. 55m. i 828 570 071 29 5-15 " 70 14- 85 I7m. 219 679 074 3 10.56 A.M. 7" 14.0 85 2h. 28m. 2 279 79 076 " 30 12.56 P.M. 70 14.0 85 4h. 2Sm. 4259 76 079 30 2.54 " 71 15-5 94 5&m. 648 58 083 31 10.42 A.M. 71 10. 61 5 h. I 4 m. 4788 169 A. Shut inlet 10.42 A.M.. 090 31 2.42 P.M. 72 7O.O 121 2h. O5m. 1998 59 outlet 10.56 A.M. 092 31 3-43 " 72 16.0 97 3h. oom. 2948 55 095 Feb. i 9-57 A.M. 72 16.0 97 5h. 46m. 5298 85 096 " i 12. II I .M. 73 16.0 97 47m. 699 39 099 " i 2.40 " 73 16.0 97 3 h. i6m. 2959 69 103 i 4-55 " 74 14.0 85 3om. 415 30 107 3 10.12 A.M. 74 16.0 97 2h. 3om. 22Q5 174 no 3 1.10 I .M. 74 16.0 97 5h. ism. 4875 196 116 3 4-55 " 75 17.0 103 ih. i8m. I 140 225 121 4 10.12 A.M. 75 15-5 94 3 h. O 5 m. 2 850 240 124 4 II-45 " 75 16.0 97 4h. 3801. 4290 606 127 4 2.25 P.M. 75 16.0 97 7h. i8m. 6730 555 131 4 5-13 " 76 16.0 97 ih. 24m. I 273 4740 I 3 6 5 lO.Og A.M. 76 18.0 109 2h. 5im. 2873 256 140 5 "45 " 76 18.0 109 4h. 27111. 4513 752 144 5 3.O2 P.M. 77 16.0 97 2h. 3om. 2570 510 149 5 5-04 " 78 14.0 85 05111. 29 816 155 6 1O.05 A.M. 78 17.0 103 ih. 36m. I 339 232 161 6 12.19 I .M. 78 17.0 103 3h. som. 3519 173 163 6 3-10 " 79 17.0 103 34m. 43 207 1 68 6 4.12 " 79 16.5 IOO ih. 36111. i 421 270 173 7 10.10 A.M. 79 16.0 97 4h. ogm. 3 701 237 177 7 1.25 P.M. 80 16.0 97 ih. 33m. i 646 338 183 7 5.22 " Si 16.0 97 ih. oom. 886 512 187 8 10.27 A.M. 81 16.0 97 2h. I7m. 2049 55f> 191 8 2.10 P.M. 82 16.0 97 43m. 677 194 195 8 3.00 " 82 18.0 109 ih. 27111. 1417 1 80 198 8 4 . 4 6 " 82 14.0 85 3h. 1301. 3047 53<> 203 " IO 10.12 A.M. 82 16.0! 97 4h. 57m. 4797 275 207 " 10 12.56 P.M. 83 ifi.o 97 54m. 800 no 211 " IO 3.10 " 83 14.0 85 3h. oSm. 2980 384 Shut inlet 3.03 P.M., out 215 " IO 4-57 " 84 17.0 3 57m. 862 348 let 3.23 P.M. 258 " 13 2-34 " 85 21 23m. 468 182 263 13 5.19 " 85 19.0 15 3h. oSm. 3 588 678 I). Application of chemi 265 14 10.19 A.M. 85 19. o 15 4(1. 23m. 4928 I 670 cals unsatisfactory: 269 14 1. 12 P.M. 86 21. 27 ih. 44m. 2047 57 chemical meter out 273 14 3-14 " 86 19-5 18 2h. 58111. 3457 40 of order. 283 15 IO.O7 A M. 86 21. O 27 4(1. 23m. 5<x>7 84 287 15 1.25 P.M. 8? 22-5 36 ih. torn. 1487 go 291 15 2-57 " 87 20. o 21 2h. 42m. 3397 243 295 15 5.24 " 88 19.0 15 ih. 32m. 1673 302 17 IO.06 A.M. 88 20.0 21 2h. 54m. 4 173 log 306 17 1.35 P.M. 89 I." IS 33m. 521 125 310 17 3.07 " 89 20. O 21 2h. osm. 2 311 87 320 iS 1O.2O A.M. 90 V.I 09 3om. 550 20 324 18 11-55 " 90 I9.O 15 2h. osm. 2 260 56 34 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Warren System. Ka e of - u Collected. Filtr s c u b. Period of *% (_j \ Number of - gs. _ d " Tast "" W;islnn R . js.fc Remarks. 3 Kun. *> ~ ojji g PH Until sand 0^0 A 8 55 D:ite. Hour. 1 i.= o Minutes. t ~ Us.? i 1,33 U % U *?. j h CO 1895 1328 Feb. IS 2. 2O I .M. , 9 7.0 103 4!). 3om. 4790 198 1333 " 18 4-55 " 9 8.0 109 ih. 39m. I 804 67 i en " 1 8 5.05 " 9 18.0 109 ih. 49111. 954 Oi i ;i ; " i<) 10. 12 A.M. 9 9.0 "5 3h. ism 3434 39 1347 9 11.31 " 9 S.i 109 4h. 34111. 4894 104 1351 9 3.04 r.M. 92 18.5 112 ih. 42111. i 829 55 1358 9 5.10 " Q2 iS.o 109 3)1. 4801. 4069 382 1362 " 20 II .OO A.M. 92 iS.o 109 5h. 43m. 6029 156 1 (06 " 20 12.03 I .M. 93 iS.o 109 15111. 209 116 1368 " 20 12.18 " 93 18.0 109 30 m. 459 104 1369 " 20 I2 -33 " 93 18.0 log 45111. 739 287 1370 " 2O 12.40 " 93 18.0 109 52111. i 009 560 1371 " 2O 1.04 " 93 18.0 109 ih. i6m. 1309 79 1375 " 20 2.06 " 93 18.0 109 2h. i8m. 2389 135 1377 " 20 3.08 " 93 17.0 103 3h. 2om. 3469 MI 1382 " 20 4.08 " 93 17. ( 103 4h, 2om. 4489 325 1387 " 20 5- 5 " 94 17-5 1 0() oSm. 90 200 i ;.> " 21 9.5*. A.M. 94 16.5 100 Ih. 2om. i 330 48 i ;<>! " 21 12.42 r.M. 14 17.0 103 4)1. o6m. 4 160 34 i I M " 21 5-9 " 95 18.5 I 12 14111. 227 "47 i |c>S " 22 10.20 A.M. 95 18.0 109 ih. 55in. 2077 88 M" " 22 1 . 1 8 r.M. 95 18.0 R>) 4)1. 53m. 5337 "5 1412 " 22 3.01 " 95 19. 115 6h. 3(>m. 7 77 528 M 4 " 22 4-53 " 96 7-5 1 06 ih. 14m. 3 5 94 1422 " 24 10.27 A.M. 96 17-5 U)6 3h. 1701. 3480 79 4 2 3 " 24 i . 10 r.M. 96 18.5 112 Oh. oom. 0440 77 M3 " 24 5.16 " 97 18.5 I 12 ih. 35111. i On 02 437 ! 25 10.26A.M. 97 20. o 121 3)1. lOin. 3631 4" 44 -5 1.14 I .M. 98 19.0 U5 48111. 734 52 445 3.07 " 98 7-5 1 06 ah. 41111. 2 824 0.( 1452 " 25 5.08 " 98 18.5 112 4h. 4 2m. 4994 27 1456 " 2f> 10.2(> A.M. s "7-5 1O6 (>h. 3001. 0904 122 1460 " 2(> n.43 " 98 15.0 9 7h. 47111. 8 214 2 94 1464 " 26 3.04 P.M. 99 18.0 109 ih. 47111. 1823 OS 147 " 2t> 5.23 " 99 iS.o 109 4h. 06111. 4363 326 1478 " 27 10.3(1 A.M. 99 18.0 109 5h. 49m. 193 34 M79 27 1.43 r.M. 00 =4-5 49 44111. 923 30 i r i 2.58 " oo 24.0 146 ih. 59111. 2 773 53 1488 " 27 5.07 " 01 23.0 140 17111. 34 4- 1490 " 28 10.3!) A.M. 01 25. 152 2)1. lOm. 33" 21 1502 " 28 3.23 I .M. 02 25. ( 152 ih. 56m. i 241 33 1507 " 28 5.00 " 02 23.5 43 3h. 33111. 4961 4 3 1512 " 2C) 10.30 A.M. 03 -..> I5Z ih. 29111. 2 320 138 1516 29 1.34 I .M. "4 - I > 146 24111. 494 1 7 1520 29 3-14 " "4 25.0 152 ah. 04111. 3024 187 1528 " 29 5.15 " 05 24.0 146 17111. 396 211 113 Mar. 2 9.33A.M. i>5 -- .- 135 ill. 03111. i 6oO I 3 1536 2 10.21 " 05 25.0 152 ih. 51111. a 766 33 1540 " 2 1.33 P.M. 01) 24.5 M9 ih. o7m. i 549 1544 " 2 3.12 " Of) - 140 ah. 40m. 3859 676 1549 2 5.00 " 7 25.0 152 22m. 457 333 557 3 10.37 A.M. 07 24.0 146 2h. 23111. 3 547 55 1561 3 12.15 r.M. 07 24.0 146 4)1. oim. 5897 405 5<>5 3 3-io " 08 25.0 152 ih. 27111. 2025 95 1570 " 3 5.10 " 08 24.0 146 3h. 2701. 4845 553 I57& 4 10.44 A.M. oS 21-5 130 4h. 54in. 0999 005 1581 4 12.58 I .M. 09 23.0 140 ih. 14111. i 747 So 1584 4 3.19 " 09 24.1 146 3h. 35m. 5093 i 39. 1589 4 5 3 " 10 - 1 146 ih. 05111. 1452 116 1595 5 10.30 A.M. 10 25. t 152 3h. 02m. 4442 187 I 599 " 5 12.49 r.M. II 23.0 140 43111. 930 93 lOob 5 3.29 II 23.0 140 3h. 23111. 4050 590 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. TABLE No. 4. Continued. Warren System. 35 Rate of j s Collected. Filtration. [i. Jj Number. I IS. Period of service Since fc c 1 3 O ^ a 25 of Sc Ill I Last Washing. Hours and tl! s.5 2 a Remarks. Date. Hour. o c ; 1.3: *o Minutes. sa s 5 c 3 IS is JJ<3 %" u Ji j b m 1895 1612 Mar. 5 5.18 r.M. 112 !6.0 158 iom. 169 615 1616 6 IO.32 A.M. 112 !3-o 140 ih. 54111. 2 819 48 1621 6 12.38 P.M. 112 23.0 140 4)1. oom. 5 699 615 1625 6 3 .l6 " H3 26.0 158 ih. 51111. 2737 106 1633 " 6 5-25 " H4 24.5 149 37" - 808 59 1637 7 10.40 A.M. 114 24.0 146 2li. 2om. 3468 39 1641 7 12.53 l -M. 114 23.0 140 4h. 33tn. 6 508 485 If 44 7 3.10 " 5 24.0 146 ih. i6m. I 750 40 1649 7 5-15 115 24.0 146 3h. 2im. 4 710 154 1656 9 10.58 A.M. H5 23.5 M3 5h. 33111. 7890 620 1 66 1 9 12.5O I .M. 116 24.5 149 ih. i6m. I 738 40 1670 9 3-40 " i 7 24.0 146 3(1111. 837 35 1671 9 5.04 117 24.0 146 2h. oom. 2757 64 1678 10 IO.I9 A.M. "7 24.0 146 3h. 45m. 5 357 139 1682 " 10 1.33 I .M. US 25.0 152 ih. 43m. 2344 68 1686 " 10 3-07 " 18 24.5 149 3h. 17111. 4605 935 1693 " 10 5.15 9 25.0 152 ih. 29111. 2 085 61 1699 " ii IO.I8 A.M. 9 25.0 152 3]]. <)2m. 4375 300 1706 " ii 3.17 P.M. 21 20.0 121 ih. 34m. I 872 310 1713 " ii 5.05 22 21.5 130 29111. 562 60 1719 " 12 10.15 A.M. 22 19-5 118 2h. 09111. 2 622 89 1723 " 12 12.54 I -M 23 9.0 H5 ih. igm. I 439 26 1727 " 12 3.22 " 23 7.5 1 06 3h. 47m. 4069 485 1732 " 12 5.oS " 24 20.0 121 ih. 07m. I 311 24 1739 " 3 IO.23 A.M. 24 8.5 112 2h. 55m. 3 551 137 1743 13 1.09 P.M. 25 9-5 us ih. 17m. 1389 26 1747 3 3-10 " 25 18.5 112 3h. i8m. 3639 I OOO 1752 13 5.OI P.M. 26 19.5 n8 ih. 1301. i 377 1 1 1758 M IO.27 A - M 26 18.5 112 3h. 09111. 3827 137 1764 4 1. 06 P.M. 127 19.0 US ih. 2im. I 480 37 1770 M 3.10 " 127 19-5 118 ... 3h. 25m. 3870 45i 1778 1 M 4.50 " 128 20.0 121 ih. o6m. i 182 19 1784 " 16 IO.22 A.M. 128 ig.O 115 3h. oSm. 3642 40 1790 " 16 1.09 P.M. 129 -.c log ih. 49111. 2019 34 1796 " 16 3-12 " 129 17.0 103 3h. 52m. 4 169 172 1803 " If) 5.02 " 130 20.0 121 26m. 443 60 1809 " 17 9.31 A.M. 130 20.0 121 ih. 19111. I 533 8 1810 17 10.24 " 130 20.0 121 2h. I2m. 2583 32 1816 7 I. I I P.M. 131 18.5 112 iSm. 264 5 1822 " I? 3-15 " 131 . 0.1 121 2h. 22m. 2734 34 1834 " 18 9.31 A.M. 132 ,.> I , .. , 3im. 551 23 1835 " 18 IO.26 " 132 18.0 109 ih. 26m. I 611 31 1840 " 18 1. 06 " 132 20.0 121 4(1. o6m. 4 801 51 1846 " 18 2.06 " 133 19.0 US iom. 203 9< 1847 " 18 3.20 " 133 20. 124 ih. 2401. i 623 s( 1852 " 18 4.21 " 133 19. 118 2h. 25m. 2793 43 1353 " 18 4-59 " 133 20.0 121 3h. O3m. 3 523 50 I85S " 19 9.31 A.M. 133 20.0 121 3h. 55m. 4<>i3 65 I86c 19 9-45 " 133 19. 118 4h. ogrri. 4863 78 1861 19 9-55 133 20. 121 4h. 19111. 5063 43 1 86s 19 10.44 34 2O. 121 iom. 185 150 l86f 19 10.56 " 34 2O. 121 22m 415 10; 186- 19 11.05 " 134 I 9 . 118 3im 585 91 187; 19 2.08 P.M. 135 2O. 124 1 3m 217 26 1878 9 3-M " 135 19. US ih. nun i 547 It i88( " 20 9.30 A.M. 136 20. 121 3om 610 4: 188- " 2O IO.2I " 136 20. 121 ih. 2im I 580 14-1 189- " 20 1.03 P.M. 137 18. 109 ;8m 978 13- i8g< " 20 2.18 " 137 18. log 2h. I3m 2 318 ii 190* " 2O 3.27 " 138 16. IOO 3om 497 29! ; 190. " 2O 4.10 " 138 18. 112 ih. 13m I 277 ii< i 3 6 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Warren System. I a z Collected. Number of Run. Rate of Filtration. 73 a J Period of Washing. Hours and Minutes. c </5 si s.s . s| u * -JU ii. 3 u u . o. w a 8 rt ^ m Remarks. | 11 u l*v ogis * I** Date. Hour. 1906 1912 1916 1918 1925 1926 1927 1929 1934 1935 1939 1940 1945 1946 1951 1952 1958 1961 1965 1968 1972 1975 1976 1977 1978 1979 1980 1981 1984 1985 198(1 1987 1988 1992 1997 2OOO 2004 2007 2OII 2020 2030 2034 2039 2042 2046 2049 2053 2056 2064 2075 2O82 2098 2102 2105 2IOg 2IJ2 2113 2114 2115 2116 2117 1896 Mar. 20 " 21 " 21 " 21 " 21 " 21 " 21 " 21 " 23 23 23 1 23 23 23 23 2 3 24 24 24 24 24 " 25 ; 25 25 25 1 25 : 25 25 24-25 25 25 " 25 25 25 " 25 " 25 25 25 " 25 " 25-26 " 26 " 26 " 26 " 26 " 26 " 26 26 " 26-27 " 27 27 27 27 27 27 " 27 " 28 " 28 " 28 " 28 " 28 " 28 4.46 P.M. 10.40 A.M. 11.58 " 12.51 P.M. 3 37 " 4.03 " 4-33 5-03 " 9-37 A.M. 10.20 " II. IO " 11.58 " 1. 01 P.M. 2.53 " 4.27 5.00 " 9 A.M. to 11.30 A.M. II.3O " " 2.3O P.M. 2.30 P.M. " 5.30 " 5.30 " " 8.30 " 8.30 " " 11.30 " 12.21 A.M. 1.04 " I. 14 1.24 I .39 1.54 2.24 " II.3O P.M. to 2.30 A.M. 2.54 A.M. 3-54 " 4.12 " 2.30 A.M. to 5.30 A.M. 5.30 " " 8.30 " 8.30 " " 11.30 " II.3O " " 2.30 P.M. 2.30 P.M. " 5.30 " 5 30 " " 8.30 " 8.30 " " 71.30 " 11.30 " " 2.3OA.M. 2.30 A.M. " 5.30 " 5.30 " " 8.30 " 8.30 " " 1I.3O " 11.30 " " 2.30 P.M. 2.30 P.M. " 5.30 " 5-30 " " 8.30 " 8.30 " " II.3O " 11.30 " " 2.30A.M. 2.30 A.M. " 5.30 " 5.30 " " 8.30 " 8.30 " " 11.30 " 11.30 " " 2.30 P.M. 2.30 P.M. " 5.30 " 5.30 " " 8.30 " 8.30 " " 11.30 " I2.O3 A.M. 12.37 " 12.47 12.57 " 1. 12 " 1.27 " I 3 8 39 139 140 141 141 141 141 142 142 142 142 143 43 144 144 M4 144-145 145-146 146 146-147 147 148 I 4 8 148 148 148 148 147-148 148 148 148 148-149 149 150 150-151 I5I-I52 152-153 153-154 154-155 155-156 156-157 157-158 I5S 159 1 60 160-161 161-162 162-163 163 163-164 164-165 165-166 166-167 167 167 168 168 1 68 1 68 168 18.5 18.5 18.0 19.0 18.0 19.0 18.0 18.0 15-5 18.0 18.0 18.0 iS.o 18.0 19.5 18.5 17-7 !6. 4 16.3 19.3 18.4 17.0 17-5 18.0 20. o 18.5 18.0 18.0 16.4 18.0 18.0 15-0 15.8 19.9 18.2 17.6 15-5 15.6 17-3 17-5 18. 1 18.3 17.8 17-7 18.4 18.4 15.1 17.6 18 3 112 112 log "5 109 115 109 109 94 109 109 109 109 109 118 112 107 99 99 U7 in 103 1 06 109 121 112 log log 99 log 109 91 96 120 no 103 94 114 95 105 1 06 log in 108 107 in in 9i 107 ih. 4901. ih. lorn. 2h. 28m. iSm. 06 m. 32m. ih. 02m. ih. 32m. O7m. 5om. ih. 4Om. 2h. 28m. I4m. 2h. o6m. 13111. 46m. 1877 i 227 2647 266 97 507 1087 I 597 138 828 i 708 2558 183 2 153 215 845 in 138 148 220 820 229 116 150 147 5i 60 60 440 5 435 73 25 21 127 66 61 286 153 170 97 116 75 64 52 90 196 252 103 62 71 85 201 81 865 74 87 121 205 142 228 89 69 59 209 91 62 189 3og no 169 207 535 no 97 59 U5 C. E. E. E. E. E. E. E. E. E. D. Application of chemi cals unsatisfictorv 3h. osm. lorn. 2om. 30111. 45m. ih. oom. ih. 30m. 3373 131 3ii 481 75i i 071 i 581 2h. oom. 3h. oom. 3h. i8m. 2 IOI 3241 3491 on Run No. 154; chemical meter out of order. C. Shut inlet 12. 03 A.M., E. [outlet 12. u A.M. E. E. E. E. 17.8 17.8 18.7 19.4 17.6 21.0 iS.o 18.0 iS.o 18.0 18.0 1 08 108 "3 "7 107 127 109 109 109 log 109 3h. 4901. lorn. 2om. 30111. 45m. ih. oom. 4050 124 324 504 784 I 004 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 37 TABLE No. 4. Continued. Warren System. Serial Number. Collected. Number Run. Ra Filti leof i h j Period of ServiceSince Last Washing. Hours and Minutes. c t c K K 2 :=~U b !5 3 u w * "s S" Remarks. oi i u a O u != c< o o U I is.? Date. Hour. 2118 2Iig 2120 2123 2124 2125 2126 2130 2135 2138 2142 2145 2154 2157 2160 2164 2168 2172 2181 2184 2188 2191 2195 2198 2202 2207 2209 2215 2219 2223 2228 2233 2236 2241 2246 2249 2254 2258 2261 2266 2270 2275 2280 2285 2288 2293 2298 2301 2309 2312 2313 2314 2315 2316 2317 2318 2319 2321 2330 2333 2336 1896 Mar. 28 " 28 " 27-28 " 23 " 28 " 28 28 " 28 " 28 " 28 " 28 " 28 " 28 " 28-29 " 29 29 " 29 29 1 29 29 " 29 1 29-30 30 1 30 1 30 1 30 30 31 1 31 31 April i I " I " 2 " 2 2 3 3 3 4 4 4 6 6 6 7 7 7 8 8 8 8 8 8 8 8 8 8 8 9 9 1.57 A.M. 2.27 " II.3O P.M. to 2.30 A.M. 3.27 A.M. 4.27 ; 4-33 " 2.30 A.M. to 5.30 A.M. 5.30 " " 8.30 " 8.30 " " 11.30 " 11.30 " " 2.30 P.M. 2.30 P.M. " 5.30 " 5.30 " " 8.30 " 8.30 " " 11.30 " 168 168 167-168 168 1 68 168 168-169 169 169-170 170-171 171 171-172 172-173 18.0 18.0 16.8 18.0 17-5 16.0 17.1 18.1 18.4 19.1 18.5 18.3 18.0 log log 102 109 1 06 97 104 no in 116 112 III ic.g ih. 3om. 2h. oom. I 494 2 IO4 107 134 126 1 60 221 2O6 83 77 114 155 169 187 135 21 40 67 4 6 125 3OO 69 I So 52 92 E. E. E. E. E. E. Shut inlet 12. 08 P.M., toulet 12.24 P- M - C. C. C. 3h. oom. 4(1. oom. 4h. o6m. 3184 4244 4354 2.3O A.M. " 5.3O " 5.30 " " 8.30 " 8.30 " " 11.30 " 11.30 " " 2.30 P.M. 2.3O P.M. " 5.3O " 5.30 " " 8.30 " 8.30 " " 11.30 " 11.30 " " 2.30A.M. 2.30 A.M. " 5.30 " 5.30 " " 8.30 " 8.30 " " II.3O " II.3O " " 2.30 P.M. 2.30 P.M. " 5.30 " 9.35 A.M. " 11.30 A.M. 11.30 " " 2.3O P.M. 2.30 P.M. " 5.30 " 9.15 A M. " 11.30 A.M. 11.30 " " 2.30 P.M. 2.30 P.M. " 5.30 9.15 A.M. " 11.30 A.M. II.3O " " 2.30 P.M. 2.30 P.M. " 5.30 " 9.15 A.M. " 11.30 A.M. I 1.30 " " 2.30 P.M. 2.30 P.M. " 5.30 " 9.15 A.M. " 11.30 A.M. I 1.30 " " 2.30 P.M. 2.30 P.M. " 5.30 " Q.I5 A.M. " 11.30 A.M. 11.30 " " 2.30 P.M. 2.30 P.M. " 5.30 " 9.15 A.M. " II.3O A.M. 11.30 " " 2.30 P.M. 2 30 P.M. " 5.30 " 9. 3O A.M. " II. 30 A.M. 12.10 P.M. 12.4!) " 12.49 " 12.52 " 12.55 " 12.58 I.I3 " 1.28 " II.3O A.M. to 2.30 P.M. 2.30 P.M. " 5.30 " g.IS A.M. " 11.30 A.M. 11.32 A.M. 173-174 174-175 175 175-176 176 177 177-178 178 178-179 179 179-180 180-181 181 182 182-183 184-185 185 186 187-188 188 188-189 189-190 190-191 191 192 193 193 194 94 194-195 195 195 195-196 196 196 196 197 197 197 97 197 97 197 196-197 197 197 97 17-3 17.9 17-8 16.2 18.9 17-5 17.2 17.9 19.0 17-3 17.1 17-4 17.9 17-8 17.2 17-5 17-8 17.4 16.6 17-8 17.6 17.4 18.4 17.2 18.0 17.0 18.0 18.2 lS.2 16.6 iS.o 17.0 17.0 18.5 17.2 14.0 18.0 19.0 18.5 18.0 18.0 18.0 18.0 17-4 7-5 7-3 18.0 105 1 09 jo8 98 H5 106 104 108 H5 104 103 105 1 08 1 08 104 1 06 108 105 101 1 08 1 06 105 in 104 109 103 109 no 1 10 101 log 103 103 112 104 85 log U5 112 109 109 109 log 105 1 06 104 109 M5 JI3 "7 147 325 635 300 220 270 102 157 205 93 77 76 1 80 87 81 25 37 39 37 25 59 51 124 91 127 117 93 81 33 53 37 70 I 1 8h. 04m. 03111. 06 m. 09 m. I2m. I5m. 3om. 45m. 8673 38 88 148 198 248 5i8 778 7h. igra. 7738 138 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Warren System. Rate of t Collected. Filtration. h i/i . | 1 Number a I s - i Period of ServiceSince Last |I J U u . a of Run. s 6 " X Washing. ^1^ i> ~- Remarks. ? Date. Hour. t a c< o Hours and Minutes. SsS if I i u ^ a cT J i 25 1" iSg6 2337 April g 11.47 A.M. 197 17.0 103 7h. 34m. 8008 126 2339 9 12.02 P.M. 197 15.0 91 7h. 4gm. 8 198, 158 Shut inlet 11.52 A.M., 2342 g II.3O A.M. to 2.30 P.M. 197198 18.4 T T T 7 J C. [outlet 12.08 P.M. 2345 198 l8 .0 Tnn 126 C. 2350 9 " IO 10. ig A.M. 198 18.0 log 6h. 2401. 6798 44 2352 " IO 10.49 " 198 18.0 109 6h. 54m. 7298 51 2354 " 10 ii. ig " 198 18.0 log 7h. 24m. 7838 87 2355 " IO 9.2O A.M. to II.3O A.M. 198 17 I C. 2357 " IO 11.49 A.M. 198 A / . J 18.0 log 7h. 54m. 8338 86 2359 " 10 13.31 P.M. 199 18.0 log O3m. 19 32 2360 " 10 12.34 " 199 !8. 5 112 o6m. 79 55 2361 " IO 12-37 199 18.5 112 ogm. 139 47 2362 " 10 12.40 " igg 18.5 112 I2m. igg 29 2363 " 10 12.43 " igg 18.5 112 15111. 259 39 2364 " IO 12.58 " 199 18.0 log 3om. 519 16 2365 " 10 I.I3 " igg 18.0 log 45m. 789 15 2366 " 10 " IO 2. 2O " 199 18.0 1 7 O log ih. 52m. 2 OOgi 23 C. 2370 " IO 3.20 P.M. 199 1 1 y 18.0 log 2h. 52m. 3 099 29 2371 " 10 4.2O " igg 17-5 1 06 3h. 5201. 4 149 30 2372 " 10 5.2O " 199 17.0 103 4h. 52m. 5 229 44 " IO 2.3O P.M. to 5.3O P.M. 17 7 C. 2377 2379 " II " II 11.30 A.M. 9.15 A.M. to 11.30 A.M. igg IOQ 1 1 1 18.0 17.8 107 109 1 08 7h. 32m. 8099 17 if, C. 2381 II 12. OO M. igg 18.0 log 8h. 02m. 8629 16 2383 " II 12.30 P.M. 199 18.0 log 8h. 32m. 9119 13 2385 " II I.OO " 199 18.0 log gh. O2m. 9 709 25 2387 " II I.3O " igg 18.0 log gh. 32m. 10 159 12 2396 " II 11.30 A.M. to 2.30 P.M. 17.2 104 32 C. 2400 " II 3-45 P.M. 200 16.0 97 03m. IO 65 2401 " II 3.48 " 200 18.0 log o6m. 60 41 2402 II 3-51 2OO 18.0 109 ogm. 120 62 2403 " II 3-54 " 2OO 18.0 109 I2m. 170 53 2404 " II 3-57 " 200 18.0 log I5m. 230 37 2406 II 4.12 " 20O 18.0 tog 3Om. 500 19 2407 " II 4.27 2OO 18.0 log 45m. 77O 20 2409 " II 4-57 20O 18.0 log ih. ism. 1310 ig 2411 " II 2.30 P.M. to 5.30 P.M. 2OO 17.1 104 28 C. 2452 " 20 10.25 A.M. 2OI 26.0 158 ih. 23m. 2 O20 30 2455 " 20 11.58 " 2OI 26.0 158 . . . . 2h. 56m. 4400 24 2458 20 2.54 P.M. 201 27.0 164 - - - 5h. 52m. 8 goo 48 2463 21 9.30 A.M. 202 22-5 136 . . . . 3om. 622 46 2465 " 21 10.22 " 2O2 23.0 140 . . . . ih. 22m. I 782 59 2468 " 21 12.35 P.M. 2O2 24.0 146 3h. 35m. 4 832 64 2472 21 1.48 202 25.0 152 .... 4h. 48m. 6 472 96 2475 " 21 2-55 " 2O2 24.0 146 5h. 55m. 8 002! 147 2479 22 g.SI A.M. 203 23.0 140 5im. I 136 104 2481 " 22 10.45 203 23.0 140 ih. 45m. 2356 igi 2484 " 22 12.32 P.M. 2O3 23.0 140 3h. 32m. 4 7g6 174 2486 " 22 1.22 " 203 23-5 143 4h. 22m. 5 gi6 162 2489 " 22 2.56 " 203 22.0 133 5h. 5&m. 8 076 ig8 2495 " 23 II. 18 A.M. 2O4 23.0 140 33m. 680 53 2497 23 12.47 ! -M. 2O4 24.0 146 2h. O2m. 2 760 73 2499 23 2.00 " 204 24.0 146 3h. ism. 4410 107 2502 ! 2 3 3 02 " 204 24.0 146 4h. 1701. 5 830 167 2504 " 23 4-47 " 2O4 24.O 146 6h. O2m. 8 240 151 2508 24 9.31 A.M. 204 24.0 146 yh. i6m. g 920 356 2510 24 "43 " 204 23.0 140 gh. 28m. 12960 375 2513 " 24 I. Og P.M. 2O4 21-5 130 loh. 54m. 14 830 460 2516 24 2.44 205 23.0 140 4im. 882 220 2518 2 4 4.41 205 23-5 143 2h. 38m. 3572 330 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. TABLE No. 4. Continued. Warren System. 39 1 X 2522 2524 2526 2531 2534 2537 2539 2543 2547 2550 2556 2560 2563 2588 2598 2602 2607 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2629 263: 2632 2633 2634 2635 2637 2639 2641 2642 2643 2644 2645 2646 2650 2656 2659 2661 2663 2670 2693 2702 2712 2718 2721 Collected. Number of Run. 205 205 205 206 206-207 207 207-208 209 209 210 Rate of Filtration. o a _! Period of Service Sincf Last Washing. Hours and Minutes. c fc*. ."is && O* u " " = ~u 3 o^- !fi (8^ B Remarks. si. 5 II u jk CS = *= = V - s Date. Hour. 1896 April 25 " 25 25 27 " 27 " 27-28 " 28 " 28 " 28 " 28-29 " 29 " 29 " 29 " 29-30 " 3 " 30 " 30 May i I " I " I " i " I " i " i i i i " i " i " i " i " i " i " i " i Apr. 3O-May I May i " i " i i " i " i I i " i " i " i i i " i " 2 2 " 2 2 4 4 " 4-5 5 5 5 IO. 15 A.M. I2.4O P.M. 1.44 " 9.30 A.M. to 3.30 P.M. 3.30 P.M. " g.oo " 9.30 " " 3.00A.M. 3. 00 A.M. " g.OO " g.OO " " 3.OO P.M. 3-OO P.M. " g.OO " 9.00 " " 3.00A.M. 20. 20. o 20. o 21.3 21.2 20. 6 22.3 21. g 21.8 20.5 121 121 121 12g 128 125 34 33 32 24 4h. 42m. 7h. O7m. 8h. ilm. 6 292 9172 10 462 231 371 324 M3 5OO 559 Application of chemicals unsatisfactory on run No. 208 ; alum meter out of order. C. This series of results on run No. 215 used in obtaining the average bacteria for this run, but not for the day. Shut inlet 2.25 P.M., out let 2.48 r M. From May 2-9, inclusive, the results of both single samples and those collected by the sampler were used to obtain the average bacteria for days and for runs. 547 410 390 l57 9.00 " " 3.00 P.M. 3.00 P.M. " g.oo " g.oo " " 3.00A.M. 3. oo A.M. " g.oo " g.oo " " 3.00 P.M. 3.00 P.M. " g.oo " 12.46 A.M. 1.58 " 2.OO " 2.02 " 2.04 " 2.O6 " 2.08 " 2.10 " 2.12 " 2.14 " 2.16 " 2.18 " 2.20 " 2.22 " 2.24 " 2.26 " 2.31 " 2.41 " 2.56 " 2 1 1-2 1 2 212 212-213 213 21.3 23.0 20.3 20.6 20.8 20. g 21. 19. o 24.0 20. o 2O.O 2O. O 2O. O 20.O, 2O. O 21. 2I.O 21. 21. 21. 2 .0 2 .O 2 .O 2 .0 2 .0 29 40 23 25 26 26 27 15 46 25 25 21 21 21 21 27 27 27 27 27 27 27 27 27 27 172 146 95 1 20 76 go 193 187 243 170 199 123 144 144 71 64 57 86 145 85 94 IOI 53 107 96 98 i") 45 53 87 IcH 83 III 132 140 1 66 158 191 77 95 140 149 183 72 56 43 5i 66 214 214 215 215 215 215 215 215 215 215 215 215 215 215 215 215 215 215 215 215 loh. 34m. 02m. O4m. oom. oSm. lorn. I2m. 14111. i6m. iSm. 2om. 22m. 24m. 26m. 28m. 3om. 35m. 45m. ih. oom. 13 136 12 52 92 132 182 222 262 302 352 392 432 472 512 552 602 802 I OI2 I 322 3.56 A.M. 4.56 " 5.56 " 6.56 " -.56 " 8.56 " 3.OO A.M. to g.OO A.M. g. 56 A.M. 10.56 " 11.56 " 12.56 P.M. 1.56 " 2.43 " g.OO " 3.45A.M. g.oo " 12.40 P.M. 3.00 " 9.15 A.M. to 3-OO P.M. 3.00 P.M. " g.oo " 9.00 " " 3.00A.M. 3.00 A.M. " g.oo " 12.31 P.M. g.oo A.M. to 3.00 P.M. 215 215 215 215 215 215 215 215 215 215 215 215 215 216 217 217 217 217 217-218 218-219 219 2lg-220 220 220 2 .0 2 .0 2 .O 2 .OJ 2 .O| 2 .O 20. 6 20.5 20.5 2O.O 21. 21.0 14-0 21.5 21. 27 27 27 27 27 27 25 24 24 21 27 27 85 30 27 2h. oom. 3h oom. 4h. oom. 5h. oom. 6h. oom. 7h. oom. 2432 3692 4952 6 182 7442 8662 8h. oom. gh. oom. loh. oom. nh. oom. I2h. oom. I2h. 47m. 5h. 44m. i8m. 5h. 33m. gh. I3m. Iih. 33m. 9902 II 162 12352 13652 14862 I57I2 7 188 364 6744 12374 14244 21.0 21 21.7 20.5 20.3 20. 6 21. 20.5 127 127 131 124 122 124 127 124 7h. som. 9769 140 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Warren System. Ra teof 4J S Collected. Filtr ation. fc ~ .y ^ ~ ^>"^~ Period of u * 3 Number o. C 0. T3 ServiceSince ii-- ^ ^ ,_ e 3 of Run. s ^s^ Washing. 11 jLs Remarks. 2 Date. Hour. k 3 -) C sill Hours and Minutes. 13 4.,-- ii <$ U s a " J " u E 1896 2726 May 5 2728 2730 6 6 T2.I5 A.M. 221 21 .O 127 6h. 55m. 8 876 69 2732 " 6 6.OO A.M. 222 21 .O 127 52m. I 076 35 2738 6 12. 2O P.M. 222 21 .O 127 jh. I2m. 8 846 IOO 2740 " 6 3.00 " 222 20 J 124 gh. 52m. 12 186 52 2748 6 ooo 227 5O 7 2751 6-7 1 0,6 2754 " 7 2757 1 7 3.25 A.M. 223 224 22. 27 33 26 urn. 216 V" IOO 2758 " 7 9 OO A.M. 224 21. 27 5h. 46m. 7 226 2762 7 3.OO P.M. 224 21. O 27 nh. 4601. 14790 30 2763 " 7 g.oo A.M. to 3.00 P.M. oo 27 2770 " 7 3.OO P.M. " g.OO 224-225 * / 128 2771 " 7 2777 " 7-8 g.oo P.M. 9.OO P.M. to 3-OO A.M. 225 21. 127 4h. 3om. 5836 51 2778 8 3.OO A.M. 225 21.0 27 loh. 3001. 13 326 167 2783 " 8 T T-2 2784 " 8 9-00 A.M. 226 21.0 27 4h. 3901. 5674 1 -* J 17 2789 8 g.OO A.M. to 3 OO P.M. 226 oo o 2T 2790 " 8 2797 " 8 3.OO P.M. 227 24.0 227228 46 .... O 5 m. IOO Application of chemicals 2801 " 8-0 g.OO " 3.OO A.M. 228 20 ( 27 unsa is ac ory on run 2802 2807 " 9 n 3.OO A.M. 228 2O.5 124 0-7 . . . . 7h. 4im. 9545 58 out of order. 2812 9 9.00 A.M. 229 20.5 * / 24 . . . . 5h. ogm. 6482 43 2817 9 3.00 P.M. 229 20.0 21 uh. ogm. 14322 49 Shut inlet 2.53 P.M., out 2823 4 3.OO P.M. 230 J2I.O 27 .... 6h. I2m. 7 958 44 let 3.13 P.M. 2829 g.oo " 230 21. O 27 I2h. I2m. 15 578 41 2832 9.51 " O2m. 13 54 2833 9-53 " 231 21.5 3 04m. 2834 9-55 231 22. O 33 06 m. 93 88 2835 1 9 57 231 22.0 33 o8m. I33i 71 2836 9-59 " 231 20.5 24 lorn. 173 49 2837 IO.OI 231 21. O 127 I2m. 213 49 2838 10.03 " 231 21. 127 I4m. 253 48 2S 3 g[ 10.04 " 231 21.5 130 1 I5tn. 273 30 2840 10.05 " 231 22. O 133 .... 7.6m. 303 33 2841 10.07 " 231 22.0 133 iSm. 343 55 2842 10.09 " 231 21.5 130 2om. 383 23 2843 1 10.11 " 231 21-5 130 22m. 423 32 2844 " 10.13 " 231 21. 127 .... 24111. 473 47 2845 10.15 " 231 21.0 , 127 26m. 513 25 284*: 10.17 " 231 2T.O 127 28m 553 33 2847 10.19 " 231 21.0 27 3Otn. 603 27 2848 10.24 " 231 21.0 27 35m. 713 33 2849 10.34 " 231 21. C 27 45m 963 36 2850 10.49 " 231 21. C 27 ih. oom. I 243 27 2851 11.49 " 231 21. C 27 2h. oom. 2483 76 2853 12.49 A.M. 231 Jl .< 127 3h. oom. 3733 48 2854 1.49 " 231 21.0 127 4h. oom. 4 973 29 2855 " 2.49 " 231 21. 127 5h. oom. 6 283 32 2856 3.OO " 231 21. oj 127 .... 5h. urn 6443 31 2859 3-49 " 231 21.0 127 .... 6h. oom. 7 503 32 2860 1 4-49 " 231 21 .0 127 7h. oom 8713 36 2861 5-49 " 231 20. 5 24 . . . .j 8h. oom. 9913 41 2863 6.49 " 231 20.5 24 . . . . j gh. oom. II 183 43 2864 1 7-49 " 231 21.0 27 . . . J loh. oom 12433 42 2865 8.45 " 231 21. 27 .... loh. 56m. 13 683 73 2870 12. OO M. 232 21. O 27 2h. nm 2790 123 2878 8.44 P.M. 233 23.0 40 lom 164 47 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 141 TABLE No. 4. Continued. Warren System. Rate of i Collected. Filtration. ^ 2 t Number S. O O. c Period of 2-S.j U ^ 1 3 f. Date. Hour. 3 Pi I Last Washing. Hours and Minutes. SI! t ~ Remarks. S 2jg ^ a? 8 JU ^CJ & U S Z m 1896 2880 May 13 2.0O A.M. 233 21.0 127 5h. 26m. 6534 44 2884 " 13 8.00 " 234 21.0 127 3h. 59^- 4961 31 288g 13 1. 00 P.M. 234 21.0 127 8h. 5gm. ii 231 13 2895) 13 7.00 " 235 21.5 130 4h. 03m. 5136 II 28gg| 14 3.OO A.M. 236 21.0 127 3h. 13111. 3899 10 2904 1 14 g.oo " 236 21. O 127 gh. I3m. n 38g 15 2go8 14 2.00 P.M. 237 22.0 133 ih. 47m. 2 330 18 2gi2 " 4 S.oo " 237 21.5 130 7h. 47m. 9940 24 2918 " 15 1. 00 A.M. 2 3 8 21. O 127 5&m. i 139 52 2922 15 8.00 " 238 21-5 130 7h. 56m. g 759 59 2g2& 15 II. OO " 2 3 8 21.5 130 loh. 56m. 13 609 47 2930 15 5-og P.M. 239 21. 127 4h. 56m. 6 250 14 2960 " 15 11.00 " 239 20. 121 loh. 47m. 13 580 1 6 2g6g " 16 5.OO A.M. 240 21.0 127 3h. I3m. 3 162 25 2g8o " 16 10 oo " 240 22.5 I 3 6 8h. I3m. 9362 17 2ggl " 16 3.00 P.M. 240 21.0 127 I3h. I3m. 15 652 4 2997 " 18 12. OO M. 241 I6. 5 IOO 0.6 i6m. 212 103 3004 " 18 3.05 P.M. 241 18.0 109 I.O 3h. 2im.< 3 372 53 3007 " 18 6.00 " 241 16.5 IOO 1.2 6h. i6m. 6 212 192 D. 3012 " 18 g.u " 242 16.0 97 0.6 I5m. 2IO 181 D. Application of chem- 3I4 18 12. OO " 242 16.5 IOO o.g 3h. 04m. 3040 152 D. icals unsatisfactory 3017 19 3-00 A.M. 242 16.0 97 6h. 0401. 6 no 79 D. on runs Nos. 242 3020 19 4-57 " 243 16.5 IOO 0.6 osm. 33 "7 D. and 243; chemical 3021 19 5-07 " 243 16.5 IOO 0.6 I5m. 223 127 D. meter out of order. 3023 19 6.00 " 243 16.5 IOO 0.7 ih. o8m. I 183 118 D. 3026 19 8.30 " 243 17-5 1 06 I.O 3h. 38m. 3593 74 D. 3031 19 12. OO M. 243 16.5 IOO i. 5 7h. o8m. 7033 81 u. 3034 19 3.00 P.M. 243 16.5 IOO 2.0 loh. o8m. 9913 98 D. 3040 19 6.00 " 243 17-5 1 06 2.2 iih. 04m. 10763 58 D. 345 19 9-3 " 244 16.0 97 0-3 osm. 88 log 3046 19 9- 3 " 244 16.0 97 0.6 I5m. 238 185 3048 19 12.00 " 244 16.5 IOO 0.8 3h. O2m. 3018 41 3052 " 20 3.00A.M. 244 17-5 106 i. 5 6h. O2m. 6038 40 3056 " 20 6.00 " 244 18.0 109 i.g 8h. 42m. 8688 45 3059; 20 8.30 " 244 17-5 1 06 2.2 nh. I2m. ii 148 75 3068 1 " 20 12.00 M. 244 15-5 94 2.6 I4h. 42m. 14508 116 3071! " 20 3.00 P.M. 244 16.5 IOO 3-6 I7h. 42m. 17348 13 3079: 20 7-35 " 245 15-0 91 0.7 osm. 93 107 3080 " 2O 7-45 245 17.0 103 0.6 ism. 243 78 3081 " 20 g.oo " 245 16.5 700 0.8 ih. 3om. i 533 47 3085 " 20 12.00 " 245 17-5 106 1.2 4h. 3om. 4563 66 3088 " 21 3.00 A.M. 245 18.0 log 1.4 7h. 3om. 7393 57 3092 " 21 6.00 " 245 17.0 103 loh. 3om. 10 263 49 3og7 " 21 8.30 " 245 17-5 106 2-7 I3h. oom. 12513 62 3103 " 21 12.32 P.M. 246 16.0 97 0.5 osm. n in 3104 " 21 12.41 246 16.0 97 o-5 iSm. 161 20 1 " 21 3.00 " 246 16.5 IOO 2h. 3301. 2491 52 3111 " 21 6.00 " 246 16.5 100 1. 1 5h. 3301. 5431 42 3"4 " 21 g.oo " 246 17.0 103 i .5 8h. 33m. 10541 102 3120 " 22 1.48 A.M. 247 15-7 91 0.5 osm. 97 52 3121 " 22 1.58 " 247 1 -.<> 109 0-5 ISm. 227 67 3122 " 22 3.00 " 247 17-5 106 0.7 i h. 1 7m. i 227 17 3126 " 22 6.00 " 247 17-5 106 I.O 4h. I7m. 4 127 27 3i2g " 22 8.30 " 247 17. n 103 1-3 6h. 47m. 6577 33 3136 " 22 12.00 M. 247 I6. 5 too 1.6 loh. 1701. 10018 41 3139 " 22 2.50 P.M. 247 22.0 133 2.O I3h. 07m. 12838 22 3144 " 22 3-47 " 248 16.0 97 1 1 . " o8m. 97 135 3145 " 22 3-52 " 248 16.0 97 o-5 I3m. 177 f>7 3M7 " 22 6.00 " 248 16.0; 97 0.6 2h. 2im. 2237 36 3150 " 22 g.oo " 248 17-0, 103 0.8 5h. 2im. 5 267 48 3154 " 22 12.00 " 248 16.5 too 1.2 8h. 2im. 8 127 52 I 4 2 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Warren System. Ra teof j 8 Collected. Flit ation. S jj ^ W L. Period of v. bi y ji Number JL ! & . "S Service Since Last sss S oi 6 Date. Hour. Run. s ul II Washing. Hours and Minutes. fc* .2.1 Remarks. D ^ u = a? 2 E" 5 n 1895 3157 May 23 3-OO A.M. 248 16.0 97 1.3 Ilh. 2im. II 017 44 3160 " 23 6.00 " 248 16.5 IOO 1-7 I4h. 2im. 3937 30 3162 " 23 8.30 " 248 16.0 97 1.9 l6h. Sim. 16257 24 3174 " 25 I2.OO M. 249 ig.5 18 ih. O2m. I 443 49 3177 25 2.OO P.M. 249 19.5 18 1. 1 3h. O2m. 3613 21 3181 25 6.OO 249 20.0 21 2.0 7h. 02m. 8463 39 3184 " 25 8.00 " 249 2O. O 21 2.5 gh. O2m. 9743 35 3188 " 25 I2.OO " 249 2O. O 21 3-5 I3h. O2m. 15 513 46 3191 " 26 2.00 A.M. 249 19-5 18 3-9 15(1. O2m. 17833 35 3194 " 26 4.36 " 250 20. o 21 0.6 05m. 107 46 3195 " 26 4.46 " 250 20. o 21 0.7 I5m. 357 50 3197 " 26 6.00 " 250 20. O 21 o.g ih. 2gm. I 777 30 3202 " 26 8.30 " 250 20. O 21 3h. 5gm. 4707 56 3208 " 26 IO.OO " 250 2O. O 21 2. 2 5h. 2gm. 6 557 31 3212 " 26 2 OO P.M. 250 19-5 18 3-0 gh. 2gm. 11347 29 3215 " 26 4.OO " 250 20.0 21 3-7 Ilh. 2gm. 13727 33 3219 " 26 6.O5 " 251 I8. 5 12 0.7 O5m. 68 94 3220 " 26 6.15 " 251 ig.o 15 0.7 15111. 228 37 3221 " 26 8 oo " 251 19.0 15 1. 1 2h. oom. 2288 41 3224 " 26 10.00 " 25 19.5 18 1-5 4h. oom. 4638 39 3228 " 27 2.OO A.M. 251 19. o 15 2.4 8h. oom. g 188 26 3231 " 27 4.OO " 251 19. o 15 3-2 loh. oom. n 488 28 3236 27 7.30 " 251 19.5 18 2.O I3h. 3om. 15718 36 3240 " 27 11-51 252 19.5 18 0.7 05 rn. 37 gi2 3243 " 27 12. OI P.M. 252 20. o 21 0.7 I5m. 347 23 3245 " 27 3.00 " 252 19.5 18 1-5 3h. I4m. 3 727 19 3255 27 6.00 " 252 19-5 18 2. I 6h. I4m. 7 2g7 36 3258 27 g.oo " 252 20. o 21 3- gh. I4m. 10847 24 3264 " 27 12. OO P.M. 252 20. o 21 3-8 I2h. I4m. 14337 60 326g " 28 3.og A.M. 253 19. o 15 0.6 osm. 54 63 3270 " 28 3-19 " 253 19.5 18 0.7 I5m. 234 3 1 3272 " 28 6.00 " 253 19-5 18 1. 1 2h. 56m. 3374 45 3275 " 28 7-30 " 253 19-5 18 1-3 4h. 26m. 5094 go 3279 " 28 10.00 " 253 19-5 18 1.6 6h. 5&m. 8 174 62 3294 28 2.32 P.M. 254 20. o 21 0.7 I5m. 67 247 D.* 32g5 " 28 2.32 " 254 20. o 21 0.7 I5m. 77 207 B.* Collected from weir 82g6 " 28 2.42 " 254 20.0 21 0.7 25m. 267 328 D.* box. T-")7 " 28 4.00 " 254 20.0 21 0.8 ih. 43m. i 827 340 D.* 3305 " 28 8.00 " 254 2O. O 21 1 . 3 5h. 43m. 6577 185 D.* 3314 28 IO.2O " 255 20. O 21 0.7 2Sm. 384 312 D.* 3315 3324 " 28 29 12.33 A.M. 10.50 P.M. 256 255 ig.o 20.0 15 21 0.6 0.8 05111. 58m. 45 974 570 680 D.* D.* Shut inlet 10. 48P.M., 3325 " 29 12.43 A.M. 256 19-5 18 0.7 I5m. 235 325 [).* outlet 10.53 P - M - 3332 2g 2.OO " 256 19.0 15 0.8 ih. 32m. i 735 220 D.* 3343 2g 4-44 " 257 19-5 18 0.7 15m. 200 157 3355 29 7.30 " 257 ig-5 18 I.O 3h. oim. 3 520 88 3360 29 12. OO M. 258 20.0 21 0.9 2h. 3im. 2950 89 3303 29 2.00 P.M. 258 2O. O 21 1.2 4h. 3im. 5 3 139 3367 2g 6.OO " 259 lg-5 18 I.O 2h. 32in. 3070 181 3373 1 2g 8.00 " 259 19-5 18 1.2 4h. 32m. 5490 80 3378 " 2g 12.OO " 260 19.5 18 0.8 ih. igm. i 505 293 5386 30 2.50 A.M. 261 19-5 18 0.7 nm. 294 227 3389 " 30 6-55 " 262 20.0 21 0.8 1 6m. 32g 246 3399 : 30 IO.IO " 262 19.5 18 1.2 3h. 3im. 4219 92 3402 " 30 12.10 P.M. 263 2O. O 21 o.g O7m. 112 93 3406 June 12. OO M. 263 21.0 27 i.g 6h. 57tn. 9 322 35 3408 3.OO P.M. 263 2O. O 21 2-5. gh. 57m. II g52 64 3413 6.00 " 264 23.O 40 o.g 3Om. 666 31 3415 " g.oo " 264 23.0 40 3h. 3om. 4806 50 3420 12. OO " 265 23.0 40 I.O 5gm. i 349 73 3422 " 2 4.00 A.M. 266 22.5 36 I.O 35m. 79 in Prescribed amount of chemicals insufficient. COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 143 TABLE No. 4. Continued. Warren System. Ra ,-,,, . u Collected. Filtr ation. S c O It. Period of ti .5 V Number 6. Is. . a ServiceSince " ifi " j e D z Date. Hour. Run. ol O S 3 ? = I Last Washing. Hours and Minutes. > 3 " fe u <- a Remarks. 2 I i = 8.? S Us rt^ X u 2 _] u. CQ 1895 3425 June 2 6.45 A.M. 26 7 23.0 140 o.g ogm. 260 e>4 3428 " 2 IO.2O " 268 23.0 140 4.0 ih. I2m. 1656 67 3431 " 2 11.51 " 269 24.0 146 9-7 lom. 1 80 181 3435 " 2 4.30 P.M. 271 23-5 143 I.O 22m. 568 121 3440 " 2 6.50 " 272 18.0 109 0-7 22m. 325 177 3444 " 2 10.47 " 273 18.0 log 0.7 27m. 414 49 3446 " 3 3.30 A.M. 274 17.0 103 0.7 3/m. 529 81 3450 " 3 6.00 " 274 16.5 100 0.9 3h. o7m. 284g 57 3453 " 3 9.00 " 275 16.5 IOO ih. 42m. i 797 77 3457 " 3 12.00 M. 275 17-5 106 I.O 4h. 42m. 4837 69 3459 " 3 2.00 P.M. 275 16.0 97 1.2 6h. 42111. 6797 71 3461 " 3 2. 5 8 " 275 17.0 103 7h. 40111. 7897 61 Shut inlet 2.42 P.M., out 3462 " 3 4-3" " 276 15-5 94 0.7 ih. oom. 944 36 let 3.00 P.M. 34<>7 " 3 6.00 " 276 17.0 103 I.O 2h. 3Om. 2424 3471 " 3 9.00 " 276 17.0 103 1.2 5h. 3om. 5444 "38" 3475 " 3 10.50 " 276 18.5 112 1.2 7h. 2om. 7304 66 / 3477 " 3 12.00 " 277 16.5 IOO 0.5 osm. 41 28 3480 " 4 12.10 A.M. 277 17.0 103 0.7 I5m. 201 33 343i " 4 3.OO " 277 16.5 IOO I.O 3h. osm. 3 101 25 3486 " 4 6.00 " 277 16.5 IOO 1.2 6h. osm. 5921 26 3489 " 4 7.00 " 277 7h. O5m. 6 941 57 3491 " 4 9.00 " 278 I h. ogm. i I ig 24 3494 " 4 10.35 " 278 18.0 109 I.O 2h. 44m. 2679 34 3503 " 4 3.52 P.M. 278 17.0 103 0.7 8h. oim. 8 28g 36 3504 " 4 5.52 " 278 15-5 94 2-9 loh. oim. 10 189 39 3508 " 4 8-35 " 279 20. o 121 o.g 4&m. 884 28 3534 " 4 12. OO " 279 20. o 121 1-5 4h. nm. 4924 16 3538 " 5 3.20 A.M. 280 19.5 118 0.6 nm. 201 43 3542 " 5 6.00 " 280 19-5 118 I.I 2h. 5im. 3251 33 3546 " 5 g.OO " 280 20.0 121 1-5 5h. 5im. 6 801 63 3553 " 5 4.0O P.M. 281 23.0 140 1.8 5h. 33m. 7521 87 3558 " 5 IO.OO " 282 23.0 I4O 1.8 5h. 55m- 6608 40 3585 " 6 2.27 A.M. 283 22.5 136 1-5 3h. 27m. 4 522 21 3591 " 6 7.03 " 283 23.0 I4O 2.1 8h. 03m. 10642 4 6 3598 " 6 7-57 " 284 7-5 45 I.I O2m. 15 31 3599 " 6 i 7.57 " 284 O2m. 15 84 3600 " 6 7-59 " 284 20. o 121 o.S 04111. 55 50 3601 " 6 8.01 284 20. o 121 0.8 o6m. 95 34 3602 " 6 8.01 " 284 20. o 121 0.8 o6m. 95 3603 " 6 8.03 " 284 20.0 121 0.8 o8m. 135 25 3604 " 6 8.05 " 284 2O. O 121 o.g lom. 175 27 3606 " 6 8.07 " 284 25.0 152 0.9 I2m. 225 19 3607 " 6 8.09 " 284 25.0 152 0.9 I4m. 275 14 3608 " 6 8. II " 284 23.0 140 0.9 l6m. 320 16 3609 " 6 8.13 " 284 23.0 140 o.g i8m. 365 12 3610 " 6 8.15 " 284 25.O 152 o.g 2om. 415 9 3611 " 6 8.17 " 284 23.0 140 o.g 22m. 460 23 3612 " 6 8.19 " 284 23.0 I4O o.g 24m. 505 ii 3613 " 6 8.21 " 284 22.5 136 o.g 26m. 550 8 3614 " 6 8.23 " 284 22-5 136 o.g 28m. 595 ii 36i5 " 6 8.25 " 284 23.0 140 o.g 3om. 640 21 3616 " 6 8.27 " 284 23.0 140 o.g 32m. 685 14 3617 " 6 8.32 " 284 22.5 136 o.g 37m. 795 16 3618 " 6 8.42 " 284 23.O 140 o.g 47m. i 025 16 3619 " 6 8.57 " 284 22.0 133 I.O ih. O2m. i 345 73 3622 " 6 9-55 284 23.0 I 4 1-3 2h. oom. 2645 14 3623 " 6 10.55 " 284 23.0 140 i-5 3h. oom. 3995 27 3626 " 6 "55 284 23.0 I4O 1.8 4h. oom. 5415 25 3627 " 6 12.55 P.M. 284 23.0 140 2.O 5h. oom. 6795 21 3628 " 6 1-55 " 284 23-5 143 2. 2 6h. oom. 8 205 3 6 3631 " 6 2-55 284 23-5 143 2.4 7h. oom. 9595 12 144 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Warren System. Rate of J u Collected. Filtration. Si t !/> . 15 jj Number a s. Period of Service Since t.s" - a 3 of Run. " a ls= K Last Washing. Hours and !l! O. 4; Remarks. __ Date. Hour. o c 8 SB o Minutes. v y/ S 2 IS = u i ^JCJ "U $ U i 3 S 1895 3634 June 6 3.17 P.M. 284 20. O 121 7h. 22m. 10085 42 Shut inlet 3.10 P.M., out- 3656 9 12.45 285 23.0 140 1-3 3h. nm. 3871 171 outlet 3.27 P.M. 3659 9 5.00 " 286 22. 5 136 0.8 lorn. 186 159 3668 IO 11.07 A.M. 286 22. O 133 1.2 2h. 43m. 3656 50 3671 " 10 I.OO P.M. 286 21-5 130 1.6 4h. 3&m. 6316 57 3675 " IO 3-3 " 287 22.5 136 o.g 44m. i 015 40 3f>8l " II 10.28 A.M. 287 21-5 130 1.3 4h. I2m. 5 705 25 3684 " II I.OO P.M. 287 22. ? 136 1.8 6h. 44m. 9245 112 368 7 | " II 2.15 " 288 22. O 133 0.8 26m. 499 16 3692 II 3.40 " 288 22. 5 136 I.O ih. sim. 2439J 16 3697 " 12 10.18 A.M. 288 22.5 136 i.. 4h. 5gm. 6 729; 27 3704 " 12 2.40 P.M. 289 22.5 136 o.S 33m. 689 53 3711 " 3 10.11 A.M. 289 23.O 140 1. 1 4h. 34m. 6 219 241 3718 13 12.58 I .M. 290 22-5 136 I.O ih. 4801. 2417 log 3724 13 2.53 " 290 23.0 140 l.i 3h. 43m. 5087 48 373 13 5.28 " 2g i 23.O 140 0.8 28m. 503 355 3734 15 g.oo A.M. 67 B. From usual nlace * 3735 g.oo 216 iK. From weir box.* 3736 " 15 g.oo " 73 B. From filtered - water 3740 " 15 10.12 " 2gi 22.5 136 o.g ih. 42m. 2 223 91 chamber.* 3743 15 12.20 PrM. 291 23.0 140 i .^ 3h. som. 5 203 356 3747 15 2.58 " 2g2 23.0 140 o.c Som. i 130 41 3753 15 4-3 " 292 22.0 133 i.; 2h. 22m. 3 240 42 3759 " 16 IO.22A.M. 292 24.0 146 i ... 4h. 44m. 6 560 35 3766 - 16 12.55 P.M. 293 22.5 136 0.8 24m. 439 31 3767 - 16 3-25 " 293 23.O 140 1.3 2h. 54m. 3 969 28 3772 " 16 4-30 " 293 23.0 140 1.8 3h. sgm. 5 479 49 3776 " 17 10.05 A.M. 2g3 23.5 143 1.8 6h. 04111. 8449 57 378o " 17 12.55 P.M. 2g4 22.5 136 0.8 32m. 674 Si 3783 " 17 2.51 " 294 22.5 136 1.2 2h. 28m. 3 454 89 3791 " 17 4.20 " 2 g4 23.0 140 1.6 3h. 57m. 54U 71 3796 " 18 IO.08 A.M. 295 23-5 143 0.9 ih. o8m. i 439 177 3801 " 18 12.32 I .M. 2g5 23-5 143 1.2 3h. 32m. 4399 43 3809 " 18 2-44 " 295 23.0 140 i. 7 5h. 44m. 7 789 22 3818 " 19 9.58 296 22-5 136 o.g ih. 23m. i 737 61 3824 " 19 12.40 296 4h. O5m. 4662 in Shut inlet 12.24 P.M. .out j - 3829 " 19 2.57 " 297 23.0 140 I . I ih. 5001. 2 315 61 let 12.41 P.M. 3845 " 9 4.26 " 297 23.0 140 I. ", 3h. igm. 1 4495 61 3853 20 9.26 A.M. ?n-7 4h. 49m. 6999 171 3856 " 20 II. IO " 298 23.0 140 o.g 36m. 756 77 3861 " 2O 12.38 P.M. 298 24.0 146 1. 1 2h. 04m. 2 746 39 3862 " 2O 12.38 " 268 From filtered - water 3870 " 20 3.22 " 298 23.5 143 i. 5 4h. 48m. 6 536 77 chamber. 3874 " 20 4.36 " 298 24.0 146 1.6 6h. 02m. 8 256 65 [chamber. f 3881 " 22 9-OO A. M. 690 B. From filtered - water 3882 " 22 g.oo " 78 jB. From weir box.f 3883 " 22 g.oo " 145 IB. From usual olace.-r 3887 3888 22 " 22 10.08 A.M. 10.08 " 299 o nn 2O.5 124 .. ih. oSm. i 508 go 70 3893 " 22 1. 12 P.M. 2gg 23.0 140 1.4 4h. I2m. 5 738 61 3894 " 22 1. 12 " 2Q9 go Fi. From filtered - water 3898 " 22 3.OO " 299 23.5 143 1.8 6h. oom. 8268 IO2 chamber. 3899 " 22 3 . oo * 525 [3. From filtered - water 3902 " 22 4-5^ " 299 23.5 143 2.O 7h. 56m. 10888 75 chamber, [chamber. 3903 " 22 4.56 " Q- B. From filtered - water 3912 " 23 10.03 A.M. 299 299 gh. 33m. T"! IlS 49 Shut inlet 10.03 A.M. 3913 " 23 10.05 " 299 15-0 91 gh. 35m. 13 168 56 39M 23 IO.O7 " 299 25.0 152 gh. 37m. 13 218 61 39 r 5 23 IO.O9 " 299 20.0 121 gh. 3gm. 13258 49 * Collected before the filter was in operation, and after period of rest of 39 hours 30 minutes. t " " " " " " " " " 39 " 51 COMPOSITION OF OHIO RIVRR WATER AFTKR PURIFICATION. 45 TABLE No. 4. Continued. Warren System. Rate MI o3 S Collected. Filtration. il . 2 1 Number , S. 1 o Period of ServiceSinc i)L- 5 . s 7. Run. 8,j O u = X Last Hoif and ^>t i Remarks. Date. Hour. u C u i o Minutes. S3 S c E is = ^ ; 1 ^ J(3 "u u S J m 1896 3gif June 23 IO. II A.M. 299 20.1 ) 121 gh. 4im. 13 298 83 39n " 23 10.13 " 299 2O. C ) 121 gh. 43m. 13338 106 391? " 23 10.15 " 299 2O. ( ) 121 gh. 45m. 13 373 92 391? 23 10.17 299 20. C 121 gh. 47m. 1340* 150 3920 " 23 10. 19 " 299 2O. C 121 gh. 4gm. 13438 72 3921 " 23 10.21 299 2O. C 121 .... gh. 5im. 13473 g2 Shut outlet 10.22 A.M. 3924 " 23 II.O9 300 22.= 136 o.S 2om. 495 560 3926 " 23 1.25 P.M. 300 23-5 43 I. 2 2h. 3 6m. 3655 410 393 " 23 3-15 " 300 23.0 140 1. 7! 4h. 26m. 6 225 3935 23 5.OO " 300 23.0 140 1.9 6h. nm. 8738 58" 3938 24 IO.I4 A.M. 3OO 23-5 143 1. 9 7h- 55m. 10895 132 3942 " 24 11.15 " 300 8h. 56m. 12 382 33O Shut outlet 11.15 A.M. 3943 24 I 1 . 2O 345 3947 " 24 12.36 " 300 301 23.0 140 0.8 38m. 713 275 chamber. 3954 " 24 3-20 " 301 23.O I4O 1-3 3h. 22m. 4593 240 3904 24 4-45 301 23.0 I4O 1.5 4h. 47m. 6493 ig5 ! [outlet 9.54 A.M. 3977 " 25 9.40 " 301 .... 6h. I2m. 8 393 Shut inlet 9.37 A.M., 3978 " 25 52 ( B. From filtcred-watcr 4O7g in O TO 301 4081 " 3 V" J^ O.3O " j I 4086 " 30 11.27 " 302 25-5 155 0.8 05 rn. 57 37 4087 3 1 1 2O * " 30 11.32 " 302 26.0 158 0.8 lorn. 217 51 chamber. 4089 " 30 "-37 302 26.O I 5 8 0.8 1 5m. 347 33 4090 " 3 11.42 " 302 23.0 140 o.g 2om. 497 21 4091 3 11.47 " 302 23.0 140 o.g 25m. 617 28 4<*;2 " 3 11.52 302 23.0 140 o.g 3om. 727 28 4"93 " 3" 11.57 302 23.0 140 o.g 35m. 837 42 4094 11 30 2. 02 P.M. 302 24.0 146 o.g 40111. 897 26 4095 " 3 2.07 302 22.0 I 33 O. U 45111 967 25 4096 " 30 2.12 " 302 23.1 140 0.9 5om. 1047 5 4097 " 30 2.17 " 302 23-0 140 o.g 55m. I 167 42 4098 " 30 2.22 " 302 23.0 140 O.g ih. oom. I 307 29 4099 " 30 2.38 " 302 23.0 140 O.g ih. i6m. I 667 51 4104 " 3 2.47 " 302 23.0 140 1. 1 3h. 2501. 4657 62 4log " 30 4.25 33 23.O 140 37m. 788 ; 40 4113 July I IO 22 A.M. 33 23.0 140 1. 1 3h. o6m. 4228 4122 I I.I5 P.M. 3<>4 23.0 I4O o.6i 43m. 934 4131 I 3.17 " 3"4 23-5 M3 i.i 2h. 45111. 3774 4147 2 11.28 A.M. 306 23-5 143 1 1 m. 159 49 4151 2 12.33 P-M. 306 24.0 I 4 6 o.g ih. i6m. I 6gg 130 4156 2 3.03 " 307 23.0 140 i.o ih. 28m. I gso 105 4164 3 IO.IO A.M. 308 23.0 140 1. 1 2h. 32m. 3493 81 4167 3 11.08 " 308 23.0 140 i.i 3h. 30111 4823 57 4186 " 3 1.45 P.M. 309 23-5 43 1 .0 ih. 42m. 3443: 76 4196 3 3-45 310 22. 5 I 3 6 0.7 07m. 117 18 4197 3 4-52 310 23.0 140 0.8 ih. 14111. I 607 167 [chamber.* 421 rj 6 9-OO A.M. 178 B. From filtered water 4203 6 g.oo 4204 6 g.oo 122 B. From weir box.* 4216 " 6 2.26 P.M. 311 21.0 127 0.8 osm. 55 1 86 4217 " 6 2.31 " 3" 22.0 133 o.S 1 lorn. 135 145 421.S 6 2.36 " 3" 22. 5 136 0.8 15111. 255 112 4219 6 2 41 3" 22-5 I 3 6 0.8 2om. 355J 77 4220 6 2.46 " 3" 22.5 I 3 6 o.S 25m. 465 42 4221 6 2.51 " 311 22.5 I 3 6 0.8 3om. 565 42 4222 " 6 2.56 " 311 J22.5 I 3 6 0.8 35m. 705 40 4223 6 3.01 " 1 22.5 136 o . g 4om . 815 38 4224 6 3-o6 " 3" 23.O 140 o.g 45m. j 945 52 * Collected before the filter was in operation, and after the period of rest of 63 hours and 30 minutes. 146 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued, Warren System. Ra teof V X Collected. Filtr S ~ . t- in u Penod of t. y 3 S Number 6. o a "S Last s^s >- V e 3 Run. ~ III K Sly 6 Remarks. .2 Date. Hour. 3 .1 L.I Minute^ sj! S S i 3 = a? I " u ffl 1896 4225 July 6 3.II P.M. 3" 23.0 140 0.9 5om. io6 5 63 4226 6 3.16 " 3" 23.0 140 I.O 55m. 1 185 46 4227 6 3.21 " 22.5 136 I.O ih. oom. 1 295 36 4228 6 3.26 " 3" 22.5 136 I.O ih. osm. 1 405 31 4230 6 3-51 " 311 23.0 140 1. 1 ih. 30m. 2005 39 4233 6 4.21 3" 23.0 140 I.I 2h. oom. 2 705 26 4240 6 5.25 " 311 22.0 133 1.2 3h. 04111. 4175 54 4245 7 10.00 A.M. 311 23.0 140 1-4 4h. ogm. 5685 55 4252 7 I.OO P.M. 312 22.5 136 0.8 iSm. 303 112 4255 7 3.00 " 312 23-0 140 i.i 2h. i8m. 3253 46 4259 7 5.13 " 313 23.0 140 0.8 lom. 1 60 114 4266 8 10.55 A.M. 313 23.0 140 1. 1 2h. 22m. 3350 51 4267 8 12-35 P.M. 314 22. O 133 0.8 o6m. 88 118 4271 " 8 3-50 " 314 22.0 33 1. 1 3h. 2im. 4438 47 4274 " 8 5.00 " 315 22.0 133 0.8 I4m. 232 62 4279 9 10.20 A.M. 315 22.5 136 i .0 2h. 04111. 2 622 39 4282 9 12. II P.M. 315 22.0 133 1-3 3h. 55m. 5 232 52 4285 9 I.Og " 316 lg-5 118 0.7 osm. 53 9 4286 9 I.I4 " 316 21.5 130 0.7 lom. 143 20 4287 9 I.I9 " 316 22.0 133 0.8 I5m. 263 55 4288 9 1.24 316 22.5 136 0.8 2om. 373 51 4289 9 1.29 " 316 23-5 M3 0.8 25m. 483 74 4290 9 1-34 " 316 22.5 136 0.8 3om. 603 46 4291 9 1-39 " 316 22.5 136 0.8 35m. 713 82 4292 9 1.44 316 23.0 140 0.9 4001. 823 58 4293 9 1.49 " 316 22.5 136 0.9 45m. 943 39 4294 9 1.54 " 316 23.0 140 0.9 5om. i 053 58 4295 9 i-59 " 316 23.0 140 0.9 55m. i 173 52 4296 9 2.04 " 316 23.0 140 I.O ih. oom. I 283 24 4297 9 2.09 " 316 23.O 140 I.O i h. osm. i 513 64 4298 9 2.24 " 316 23-5 143 1. 1 ih. 2om. i 773 42 4299 9 2-39 " 316 23.0 140 I.I ih. 35m. 2 113 33 4300 9 2.54 316 23.0 140 1. 1 ih. som. 2 553 37 4301 9 3.09 " 316 23.0 140 I.I 2h. osm. 2 913 79 430ia 9 3-24 316 23.0 140 1.2 2h. 2om. 3 163 45 4305 9 3-39 " 316 22. O 133 1.2 2h. 35m. 3 513 40 4306 4308 9 9 3.54 4.09 " 316 316 23.0 19. o 140 1.2 2h. som. 3 h. O 5 m. 3943 4 183 75 M3 Shut inlet 4.O7P.M., out 43M 10 11.07 A.M. 317 22.5 136 1.2 2h. 49m. 3693 54 let 4.24 P.M. 4317 10 1. 01 P.M. 317 22. O 133 1.6 4h. 43m. 6283 137 4320 " 10 3.10 " 318 23.O 140 I. I ih. igm. I 708 25 4323 " 10 5-05 " 318 24-5 149 1-5 3h. I4m. 4378 69 4328 " ii 10.31 A.M. 3 I8 22.5 136 1.8 5h. lom. 6988 41 4333 " ii 12-59 P - M - 319 22-5 136 1. 1 ih. oim. i 469 36 4346 " ii 3.12 " 319 23.0 140 1-5 3 h. I 4 m. 4639 43 43^7 13 IO.IO A.M. 320 22.5 136 0. I ih. lom. i 508 53 4370 13 11.44 " 320 23.0 140 1.2 2h. 44m. 3658 20 4375 13 3.33 P.M. 321 23.0 140 0.8 3om. 639 42 4376 ! 3 5.08 " 321 22-5 136 i.i 2h. osm. 2819 140 4396 IQ.lS A.M. 321 23.0 140 1.2 3h. 45m. 6h. 32m. 5 139 8 919 31 51 Shut outlet 1.05 P.M. 4409 4422 " M 3.21 " 322 22.0 133 I.I ih. 54m. 2538 17 4424 " J 4 4-55 322 23.0 140 1-4 3h. 28m. 4728 4442 " 15 I. II 323 23.0 140 O.g 35m. 732 16 4443 15 2.04 " 323 23-5 143 I.O ih. 28m. i 972 37 4448 IS 3.12 " 323 23.0 140 1.2 2h. 3&m. 3 562 37 4456 " 16 9-35 A.M. 324 22.5 136 O.g 35m. 686 34 4459 " 16 11.04 " 324 23.0 140 I. I 2h. 04111. 2 756 15 4469 " 16 I. II P.M. 324 23.0 140 I. 5 4h. nm. 5 726 4477 " 16 2-54 " 325 22.5 136 0.8 lom. 184 17 4478 16 2.59 " 325 22.0 133 0.8 I5m. 273 32 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 47 TABLE No. 4. Continued. Warren System. Rate of S u Collected. Filtration. u c u t/5 . 1: Number I !. i Period of Service Sine Last . ho c o u . Run. X 5jjl= X *lfa - R Remarks. 3 Date. Hour. 3 5 jtla 1 Minutes. jjjl P X u i -J 03 1896 . 4479 July 16 3.04 P.M. 325 22. C 133 0.8 2om 374 6? 4480 " 16 3.09 " 325 22.0 133 0.9 25m 504 4 448 1 " 16 3-14 325 23.0 140 o.g 3om 614 33 4482 " 16 3.19 " 325 23.0 140 o.g 35m 724 38 4483 " 16 3-24 " 325 22.5 136 o.g 4om . 844 30 4484 16 3.29 " 325 22.5 136 o.g 45m. 954 59 4485 16 3-34 325 23-5 143 I.O 5om. 1074 40 4486 " 16 3-39 " 325 23.0 140 I.O 55m. i 184 18 4487 " 16 3-44 325 23.0 14" I.O ih. oom. I 304 27 4488 " 16 3-59 " 325 23.0 140 i.o ih. ism 1654 23 4489 " 16 4-14 325 23.0 140 i.i ih. 3om i 994 ii 4490 " 16 4.29 " 325 23.0 140 1. 1 ih. 45m 2344 33 4493 " 16 4-44 " 325 23.0 140 1.2 2h. oom 2694 17 4494 " 16 4-59 " 325 23-0 140 1.2 2h. 1501 354 15 4497 " 16 5.14 325 23.0 140 i. 2h. 3om. 3394 29 4498 " 16 5.29 " 325 23.0 140 i-3 2h. 45m. 3 744 35 453 " 7 2-37 " 325 23-5 M3 1-3 3h. 46m. 5 124 ii 4525 " 18 11-37 A.M. 326 22.5 I 3 6 O.g 3im. 708 2 3 4546 " 18 1.52 P.M. 326 23.O 140 1.2 2h. 46m. 3728 2 4563 " 18 5-12 " 327 23-5 M3 I.I ih. I7m. 1683 4570 " 20 11.07 A.M. 327 22.5 136 1.4 3h. 42m. 4983 22 4575 " 2O 1.48 P.M. 328 23.0 143 0.8 i6m. 305 5 4576 " 20 3-24 " 328 23.0 143 i.i ih. 52m. 2 515 21 458i " 20 5-12 " 328 21.5 130 3-6 3h. 4om. 4955 48 4603 " 21 11.07 A.M. 328 23-5 M3 1.8 6h. osm. 8195 93 4608 " 21 I.I4 P.M. 329 23-5 M3 I.O ih. nm. i 562 57 4613 " 21 3-19 " 329 23.0 140 1.4 3h. i6m. 4512 86 4616 " 21 5-10 " 330 22.5 136 o.g 2gm. 616 217 4619 " 22 11.02 A.M. 330 23.0 140 i.i 2h. 38m. 3 626 383 4627 " 22 3-47 I -M. 332 23.0 140 o.g 55m. i 236 818 4633 " 22 4.52 " 332 23.0 140 1. 1 2h. oom. 2 756 1055 4637 " 23 II. l6 A.M. 333 21.0 127 o.g 5om. 939 367 4643 23 12.54 P.M. 333 23.0 140 i.i 2h. 28m. 3 I0 9 420 4645 " 23 3-9 " 334 23.0 140 I.O 44m. 875 916 4649 " 23 4-59 " 335 17.0 103 0.8 I5m. 143 685 4682 " 24 1.32 " 336 15-5 94 0.6 2om. 225 288 4689 " 24 4.20 " 337 15-5 94 0.7 [Om. 101 900 4690 24 4-30 " 337 16.0 97 0.7 2om. 361 1300 4691 " 24 4.40 " 337 [6.0 97 0.8 3om. 421 2000 46913 " 24 4.50 " 337 Id M 97 0.8 4Om. 2500 4692 24 5.00 " 337 I6. 5 IOO o.g 5om. 731 22OO 4695 24 5.10 " 337 16.0 97 o.g ih. oom. 891 392 4696 24 5-25 " 337 15-5 94 I.O ih. ism. i 131 1034 Shutinlet 5. 20 .P.M., out 4/05 " 25 11.05 A.M. 338 17.0 103 ih. i8m. 1086 let 5.30 P.M. 4710 " 25 I.I7 P.M. 338 15-0 91 2.2 3h. 30m. 2906 4l6 4712 " 25 3.15 " 339 13-0 79 2.O 43m. 508 378 4715 " 25 4-43 " 339 14.0 85 4.0 2h. Mm. i 698 504 [chamber.* 4722 " 27 g.oo A.M. I B. From filtered-water 4723 " 27 g.oo " 8 B. From usual place.* 4725 " 27 9.25 " 22 3. From filtered-water 4727 " 27 11.49 " 341 14.5 88 0.8 39m. ^33 26 chamber. 4732 " 27 2. II P.M. 341 16.0 97 0.7 3h. orm. 2^63 268 4733 " 27 3-02 " 341 19. o "5 o.g 3h. 52m. 3623 624 4753 " 28 9-45 A.M. 343 22.0 133 o.g I5m. 226 61 4754 " 28 9.50 " 343 23.5 M3 I.O 2om. 33 394 4755 " 28 9-55 343 23-5 I.O 2$m. 446 113 4756 " 28 IO.OO " 343 23-5 143 I.O 3om. 566 36 4758 " 28 10.05 " 343 23-5 143 I.O 35m. 696 44 4759 28 IO. IO " 343 23.5 143 I.O 4Om. 816 53 4760 " 28 10.15 " 343 23.O 140 I.O 45m. 916 60 * Collected before the filter was in operation, and after a period of rest of 40 hours and 7 minutes. i 4 8 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Warren System. Rate of j a Collected. Filtration. % .S 1 ai Number o. C l- d Period of 4J ^ u 3 Date. Hour. Run" iS 3 5 5 I Last Washing Hours and Minutes. Sl| c Remarks. 1 u i J " u n 1896 761 July 28 10.20 A.M. 343 23.0 140 I.O 5001. 1 046 66 762 " 28 0.25 " 343 23.0 140 I.O 55m. 1 1561 7 763 28 0.30 " 343 23.0 140 I.O ih. oom.j i 266 77 765 " 28 0-45 " 343 23.0 140 I.O ih. 15111. i 606 31 7" " 28 I.OO " 343 22.5 136 1. 1 ih. 30111. 1 956 40 7 ") " 28 1.15 " 343 22.5 136 I.O ih. 45m. 2396 21 77 " 28 1.30 " 343 22.5 136 I.I 2h. oom. 2 636 59 773 " 28 1-45 " 343 23.5 143 I . 2 2h. I5tn. 2986 29 774 " 28 2.OO M. 343 23.5 143 1 .3 2h. 3om. 3 336 27 77 " 28 2.15 P.M. 343 23.0 140 1.3 2h. 45m. 36861 53 \m " 28 2.30 " 343 23.0, 140 1-3 3h. oom. 4 046! 52 77<) " 28 2-45 " 343 23.0 140 1-4 3 h. I 5 m. 4 376 61 1780 " 28 I.OO " 343 23.0 140 1.4 3h. 3001. 4 746 25 784 " 28 J..I5 " 343 23.0 140 1-4 3h. 45m. 5096 31 17^- " 28 1.30 " 343 22.5 136 i . 5 4h. oom. 5 416 46 17-7 " 28 2.OO " 343 23.O 140 1.6 4h. 3om. 6146 14 17- " 28 2.30 " 343 21.0 127 i-7 5h. oom. 6826 18 17T- " 28 3-OO " 343 21 .O 127 i-7 5h. 3om. 7476 33 1795 " 28 5.O2 " 344 21-5 130 0.9 1 7m. 254 18 Pi " 29 II. 10 P.M. 344 22.5 136 i .3 2h. 55m. 4024 51 I.-4I 29 1.25 " 345 22.0 133 1. 1 ih. I5m. i 601 90 r = - 29 2.56 " 345 22. O| 133 1-3 2h. 46111. 3661 10 1862 29 5.05 " 345 23.O 140 1.6 4)1. 55m. 6 581 58 ,< 1 30 I. ig " 346 23-5 143 I.O 37m. 774 12 1 S 7 " 30 3-43 " 346 23-5 143 i .3 3h. oim. 4 134 15 4882 P-7 . . 3I 1 31 II.O9 A.M. 1.58 P.M. 347 347 22. 5 22.5 136 136 I .2 1-7 2h. ogm. 4h. 58m. 2874 25 6 764 7 4893 " 31 3-44 347 23.5 143 2.0 6h. 44m. 9 274! 66 Jewell System. 1895 2 Oct. 21 10.47 A.M. , 25 . o IOI 76 21 12^30 P.M. 2 S 2 4 " 21 3-46 " 2 3h. 43m. 38 6 " 22 9.45 A.M. 2 26.0 105 4h. 28m. 6405 14 9 " 22 11.25 2 28.0 114 6h. o8m. 8465 42 3 " 22 1.34 P.M. 2 22. O 89 8h. 17111. ii 425 66 " 22 1-47 " 2 23.5 95 Sh. 30111. 11697! 62 Agitated surface of sand 15 " 22 3.05 " 3 J- , 114 1 8m. 504 84 layer at 1.39 P.M. i? " 22 4.00 " 3 28.0 114 ih. 1301. I IIO 49 20 " 23 9.28 A.M. 3 3O.O 122 2h. O2m. 2 782 no 22 23 10.57 " 3 30.0 122 . . . . 3 h. 3im. 5 319 55 24 " 23 n.53 " 3 29.0 118 4h. 27m. 7088 53 26 " 23 1.20 P.M. 3 29 o 118 5h. 54m. 9496 56 28 " 23 2.30 " 3 28.5 116 7h. 0401. ii 475 42 30 23 4.17 3 29.0 118 8h. 5im. 15 104 38 33 23 5-2O " T 9h. 54tn. 16 853 39 38 " 24 12.12 " 3 29.0 118 loh. 52m. 18423 77 40 " 24 1.30 " 3 29.0 118 I2h. lorn. 20 710 67 45 " 24 4.06 " 3 29.0 118 I4h. 46m. 25097 52 47 24 5.12 i 1 5h. 52m. 27 022 40 49 " 25 9.53 A.M. 3 30.0 122 i6h. 32m 28 105 34 51 25 11.07 " 3 29.0 118 I7h. 46111 30324 24 53 25 12.05 P.M. 3 26.O 105 i8h. 44m.! 31 908 28 56 25 1.32 " 3 23.0 93 2oh. iim 33 948 27 58 " 25 2.52 " 4 30.0 122 25m 786 28 59 25 3-30 " 4 3O.O 122 ih. 03111 1971 36 63 " 25 4.27 4 28.0 114 2h. oom 3638 29 65 " 26 10.52 A.M. 4 20.0 8l 4h. I4m 7 "5 10 68 " 26 1.03 P.M. 5 2<> . < 105 i h. I3m i 951 32 7 " 26 4-35 " 5 25.0 IOI 4h. 4jm 5785 12 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. TABLE No. 4. Continued. Jewell System. 149 1 E z 1 72 78 80 92 9-1 95 98 107 109 no 114 "7 IIQ 123 125 127 129 147 148 152 156 170 171 "79 184 186 187 190 192 194 95 198 201 20 3 207 209 213 217 2 2O 225 226 229 232 235 236 237 238 239 240 243 244 245 246 247 248 2513 2516 252 253 256 257 Collected. Number Run. Rate of Filtration. I -j Period of Service Since Last Washing. Hours and Minutes. "" ti fell % u> u %-:(j X 3 ^ P Remarks. 8. U. 3 I* u Js. ] III CL 7 2 Date. Hour. 1895 Oct. 28 " 2g " 29 " 30 " 30 " 3 " 30 " 31 " 31 " 31 Nov. i : : 2 :: 2 :: i \ s I ! 7 7 ! . 7 8 8 8 8 " 8 9 9 9 II " II " II " 12 " 12 " 12 " 12 " 12 " 12 " 13 13 "3 " 13 13 >3 " 14 "4 " "4 14 15 5 10.54 A.M. 12.05 P.M. 1-55 " 3- 7 " 3.50 " 4.18 " 5-25 " 1.43 " 2.38 " 4.16 " 1.52 " 4.00 " 4.3 " II .02 A.M. 12.27 I .M 1.23 " 3-27 " 11.04 A.M. 11.24 " 12.54 P.M. 3-57 " IO.5O A.M. 11.20 " I . 13 I .M. 1.59 " 2.O9 " 2.18 " 2.36 " .2.50 " 3-09 " 3-32 " II .OO A.M. 12.35 r.M. I . 12 " 2.23 " 2.46 " II .30 A.M. i . 18 r.M. 2.27 " 9.17 A.M. 9.50 " 11.06 " 2.50 P.M. 10.44 A.M 1 1 .05 " H-35 " 12.00 M. I . 15 I .M. 3.10 " g.2O A.M. 9.50 " 10.27 " 11.27 " I. I 8 P.M. 2.52 " g.02 A.M. 9.50 " 12.00 M. 3.00 I .M. IO.50 A.M. 12.55 P.M. 5 5 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 7 7 7 8 8 8 8 8 8 8 8 S 8 9 9 9 9 9 9 9 9 9 o o o o o o o o I I I 6h. 19111. Sh. 39m loh. 2901. 8 916 10507 12 776 772 1585 I 998 2 926 6527 7 600 9061 ii 790 600 i 35i 4906 7 39 8846 10547 14362 14874 17 241 20 531 22 002 22 761 25432 59 334 549 I 012 I 355 1852 2439 4878 777 7845 303 842 2 534 5 244 6863 8469 9267 10772 15681 43 640 i 411 2040 4 212 6527 10 308 10857 II 764 13290 15783 18 105 19505 20482 23763 2390 8 167 11417 40 54 20 a 4 2 3 3 16 16 ii 27 M 17 28 23 38 520 138 68 102 540 268 186 182 124 34 156 132 33 178 192 222 193 107 226 128 128 57 Hoo 864 396 19? 1356 178 308 283 260 244 106 136 106 104 74 140 172 116 82 68 92 Sterilized filter on this day. Agitated surface of sand layer at 2.38 P.M. Agitated surface of sand layer at 11.22 A.M. Agitated surface of sand layer from 10.42 A.M. to i .56 P.M. Agitated surface of sand layer all day, Nov. n. 26.0 21 .O 24.O 24.0 24.O 27.0 22. 16.0 14.0 21.0 28.0 28.0 32.O 3O.O 26.0 27.0 27.0 27.O 27.0 28.0 26.O 24.O 22. O 26.O 25.0 105 85 97 97 97 109 89 65 57 85 114 "4 130 122 105 109 109 109 109 114 105 97 89 i5 IOI 02m. I2m. 2im. 39m. 53m- ih. I2m. ih. 35m. 2h. 2im. 3h. 56m. 4h. 33m. 1 1 in 34"i. ih. 41111. 3(1. 2gm. 4h. 3801. 5 h. 5 6m. 6h. 2gm. 7h. 45<n. lib. 29111. O2m. 23m. 53"i- Ih. i8m. 2h. 33m. 4h. 28m. 6h. 54m. 7h. 24m. 8h. oim. gh. oim. loh. 52m. izh. 26m. i3h. ism. I4h. 03m. i6h. 1301. ih. 43111. 5h. 3gm. 7h. 44111. 24.0 24.0 24.0 25.0 27.0 2O. O 21 .O 25.0 25.0 26.0 29.0 24.0 25.0 24.0 24.0 21 .0 25.0 25.0 25.. 25.0 25.0 25.0 -). 24.0 25.0 23.0 20.0 24-5 25.0 24.0 25.0 24.0 23." 97 97 97 IOI 109 81 85 IOI IOI 105 IlS 97 IOI 97 97 85 IOI IOI IOI IOI IOI IOI 97 97 IOI 93 81 99 IOI 97 IOI 97 93 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Jewell System. Rate of 8 u Collected. Filtration. .5 .H ^ -^ " * k. Period of i. bi u Number a o a -o Service Since S-s *j ^ u a Run. S,j ti 5 X Last Washing. Hours and g a Remarks. _ Date. Hour. u c o K o Minutes. Ss ja i c ~ a cT 1 - JU rt U $ U S J [i. n 1895 258 Nov. 15 3.29 I .M. II 22.0 89 lOh. l8m. 14 961 86 " 16 I [ 25.0 101 18 267 116 , " 1 6 > M 1 1 20 584 112 203 " 16 32 -3 p M 1 1 146 204 267 " 18 . ^ J i . . 9.41 A.M. 1 1 25 .O Tnr 28 137 276 268 " 18 1 1 . OO 1 1 24.0 Q7 30 124 168 260 " 18 12-35 P.M. II 24. o y / 97 32 329 116 270 " 18 3.O5 " II 25.O IOI 36 138 nS 27T 10 0.30 A.M. 1 1 24. 5 37 271 318 */ J 274 A V IO. 2O " 1 1 24. o 99 07 38 43 168 277 " 20 12.31 P.M. 12 25.0 y / IOI 2im. 412 34 278 " 20 12.51 " 12 25.0 IOI 4im. 984 294 279 " 20 1.05 12 25.0 IOI 55m. i 276 54 280 " 2O 1 . 2O " 12 25.0 IOI ih. lom. i 714 68 281 " 20 2.30 " 12 25-0 IOI 2h. 2om. 1909 49 284 " 21 g.22 A.M. 12 24.5 99 4h. oim. 6 117 142 285 " 21 IO.IO " 12 24.0 97 4h. 49m. 7237 72 286 " 21 12.03 l -M. 12 25.0 IOI 6h. 42m. 10061 76 287 " 21 2.OO " 12 25.0 IOI 8h. 3gm. 13072 50 290 " 22 2.22 " 12 24.0 97 nh. o6m. 16687 9 291 " 22 3.32 " 12 25.0 IOI I2h. i6m. 18487 36 294 " 2 3 9-21 A.M. 12 26.0 105 I2h. 32m. I933I 394 295 " 23 10.24 " 12 26.0 105 I3h. 35m. 21 046 44 296 23 I.I5 P.M. 12 24.0 97 i6h. i8m. 24837 58 299 " 23 3.40 " 12 20.0 81 iSh. 26m. 27 4O2 77 301 " 25 9-45 A.M. 12 19-5 79 i8h. 43m. 28 CO2 378 305 25 10.40 " 12 20.0 81 igh. 38m. 28 927 420 307 25 11.45 A.M. 13 25.0 IOI 07111. 175 440 308 " 25 n-55 13 25-0 IOI 17111. 413 368 309 " 25 12.05 P.M. 13 24.5 99 27m. 622 364 310 " 25 12.15 " 3 24.0 97 37111. 848 390 312 " 25 1-35 13 24.0 97 ih. 57m. 2605 366 314 " 25 3-20 " 13 26.0 105 3h. 42m. 5356 484 320 " 26 g.22 A.M. 13 27.0 log 4h. oSm. 6 460 748 321 " 26 IO.I5 " 13 25.0 101 5h. oim. 7474 512 323 " 26 11.27 " 13 24-0 97 6h. 1311. 9204 664 325 " 26 1.48 P.M. 13 24.0 97 Sh 34m. 12 664 394 328 " 26 3-15 " 3 23.0 93 lOh. oim. 14633 386 331 " 27 9.2O A.M. 13 25-0 IOI loh. 3601. 15 617 754 333 " 27 IO.I6 " 13 26.0 105 lib. 32m. 17 078 875 337 " 27 11.45 " 13 25-5 103 I3h. oim 9 333 T358 339 34 2 27 " 27 1.33 P.M. 3. 12 " 13 13 23-5 95 I4h. 49m. i6h. 28m. 22 034 24 259 972 704 345 " 29 9.14 A.M. 13 25.0 IOI i6h. 48m. 24 652 2280 346 29 9-45 " 13 24.0 97 I7h. igm. 25407 665 Agitated surface of sand 353 " 29 10.51 13 21. O 85 i8h. 25m. 26 945 343 layer from g.2S A.M. to 355 " 29 12.01 P.M. 13 22.0 89 igh. 35m. 28587 444 9.43A.M. 357 " 29 1.47 " 13 24-0 97 2ih. 2im. 30880 328 362 " 30 9.48 A.M. 14 24.0 97 I2m. 217 700 363 " 30 9.58 " 14 24.0 97 22m. 511 558 364 " 3 10.08 14 24.0 97 32111. 712 540 365 " 30 IO.I8 " M 24.0 97 42m. 990 528 366 " 30 10.28 " 14 24.0 97 52m. i 233 560 367 " 30 10.38 " 14 24-5 99 ih. O2m. i 368 546 369 " 30 11.43 " 14 26.0 105 2h. 07m. 3112 658 37i " 30 1.32 P.M. 14 24.0 97 3h. 56m. 5 808 834 375 Dec. 2 9.42 A.M. 14 25.0 IOI 8h. 35m. 12 2I3 ; 448 377 " 2 10.43 " 14 25.0 101 gh. 26m. 13684: 322 3 So " 2 12.29 P - M - 14 24.0 97 nh. I2m. 16273; 376 382 " 2 2.32 " 14 24.0 97 I3h. ism. 19 148; 294 386 3 10.31 A.M. 14 25.0 IOI I5h. 05m. 21 901 392 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 151 TABLE No. 4. Contimted. Jewell System. Rate of -; > Collected. Filtration. fc j .2 - ~"C " S Period of !. M " i Number a S & 1 Service Since " J " ^ *" s Run So "rt 4J O U 3 X Washmfj. ^Jfc S.S Remarks. z Date. Hour. fc 3 Hours and Minutes. E. inXl I l c Is -~ G. fT % 233 "u u 7. J m 1895 388 Dec. 3 11.37 A.M. 14 24.0 97 i6h. nm. 23 537 350 390 3 12.55 P.M. 14 22. O 89 I7h. 2gm. 25437 322 393 3 2.15 " 14 22. C 89 iSh. 4gm. 27 206 385 Agitated surface of sand 411 4 3-07 14 25.0 IOI 2oh. O2m. 29122 372 layer at 1.32 P.M. 414 4 4-37 14 23.0 93 2ih. 32m. 3i 350 290 420 " 5 9-47 A.M. 14 23.O 93 23h. ogm. 33716 264 422 5 10.38 " 14 23.0 93 24h. oom. 34 2g 150 425 5 11.47 " 14 21.01 85 25h. ogm. 36313 232 Agitated surface of sand 427 " 5 2.40 P.M. 15 24.0 97 lom. 271 270 layer at 12.23 l - M - 429 5 2.50 " 15 24.0 97 2om. 502 244 430 5 3.OO " 15 25.0 IOI 3001. 761 192 431 " 5 3.10 " 15 26.0 105 4Om. I 023 298 432 " 5 3.20 " 15 24.0 97 5om. i 248 280 433 5 3 30 " 15 24.0 97 ih. oom. I 482 368 436 5 3.46 " 15 25.0 IOI ih. i6m. i 856 156 438 " 6 10 04 A.M. 15 Jfj. 105 ih. 43m. 2 6g2 240 442 " 6 11.27 " 15 24.0 IOI 3h o6m. 4827 194 449 6 1.36 P.M. 15 24.0 97 5h. ism. 8 029 296 452 | 6 3-45 " 15 22.0 89 7h. 24m. ii 206 236 453 7 9.25 A.M. 15 25.0 IOI gh. 29111. 14 124 274 455 7 12.24 P.M. 15 23.0 93 I2h. 28m. 1 8 442 124 458 7 12-55 " 15 24.0 97 I4h. 4gm. 21438 864 Agitated surface of sand 461 9 10.05 A.M. 15 22.0 89 i8h. nm. 26 096 1 60 layer at 2.47 P.M. 465 9 II.I8 " 15 22.0 89 igh. 22111. 27655 144 467 " 9 12. 2O P.M. 15 23.0 93 2Oh. ism. 28825 192 Agitated surface of sand 468 9 I. 4 8 " 15 21 .O 85 2ih. 43m. 30 683 164 layer at 11.47 A.M. 472 9 3.38 " 15 21. O 85 23h. 26m. 32836 172 Agitated surface of sand 478 10 10.40 A.M. 16 24.0 97 1401. 592 224 layer at 3.14 P.M. 479 " 10 10.50 " 16 24.0 97 24m. 796 176 480 " 10 11.00 " 16 24.0 97 34m. i 092 214 481 " 10 II. 10 " 16 24.0 97 44m. i 313 1 68 482 " 10 1 1. 2O " 16 24.0 97 54m. i 543 304 483 " 10 11.30 " 16 24.0 97 ih. 04m. I 755 194 490 " 10 2.07 P.M. 16 25.O IOI 3h. 4im. 4498! 268 494 " 10 3-30 " 16 24.0 97 5h. 04m. 6605 238 497 " ii II.O8 A.M. 16 28.0 114 7h. 35m. 10370 196 499 " 1 1 12.14 P.M. 16 25.0 IOI 8h. 4im. 12 157 224 503 " ii 1.24 " :6 25.0 IOI gh. 5im. 13442 196 506 " ii 3-II 16 24.0 97 nh. 38m. 16 532 190 508 " 12 9.36 A.M. 16 24.0 97 I4h. i6m. 20339 142 5io " 12 12.00 M. 16 23.0 93 i6h. 40111. 23717 130 512 " 12 3.OO P.M. 1 6 22.0 89 igh. 4Om. 27 717 176 Agitated surface of sand 517 " 13 IO.52 A.M. 16 24.0 97 22h. 4om. 31 749 135 layer at 4.28 P.M. 519 13 1.50 P.M. 16 2O. O 81 26h. iSm. 36428 116 522 13 4.44 P.M. 17 - 1 " 97 3Om. 530 127 524 14 IO.05 A.M. 17 25.0 IOI ih. 42m. 3001 103 526 14 12.55 P.M. 17 21. 85 .... 4h. 32m. 7 195 136 529 14 3.08 " 17 24.0 97 6h. 45m. 10413 164 533 " 16 9.25 A.M. 17 25.0 IOI gh. I4m. 14 128 148 535 " 16 11.32 " 17 25.0 IOI Iih. 2im. 17 561 150 539 " 16 2.40 P.M. 17 23.0 93 I4h. 2gm. 22 125 1 86 5)0 " 16 5-15 " 17 21.0 85 I7h 0401. 25 508 go 544 " 17 9.48 A.M. 17 2O. O 81 17(1. 2gm. 93 545 " 17 12.52 P.M. 18 2O. O 8l 24m. 724 88 549 17 3-30 " 18 24.0 97 3h. O2m. 4 501 148 550 " 17 4.31 18 24.0 97 4h. 03m. 6061 164 553 " 18 9.16 A.M. 18 25.0 IOI 5h. 07m. 7684 168 555 " 18 10.35 " 18 24.0 97 6h. 26m. 9650 168 557 " 18 1.05 P.M. 18 22.0 89 8h. 5&m. 13038 97 566 " 18 3-31 " 18 16.0 65 uh. 22m. 15928 130 567 " 18 4-34 " 18 18.0 73 I2h. 25m. 16933 90 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Jewell System. Rate of J 01 Collected. Filtration. 1 C | 1 Number R I S. i Period of Service Si nee Last u t? (J s of S . ^ s X Washing. ^ E S.U Remarks. 1 Date. Hour. Run. ol =<! Hours and Minutes. 1-^ rt 8 C II ^ a * % I=-3^ "U i J n 1895 571 Dec. 19 9-45 A.M. 18 18.0 73 .... 13)1. 3om. 18 114 116 Agitated surface of sand 572 " 19 11.48 18 18.0 73 .... I5h. 33m. 20 187 53 layer at 9.15 A.M. 573 19 12.50 P.M. 19 25.0 IOI . lom. 257 46 574 19 1. 00 " 19 25.0 IOI . . . .; 2om. 508 35 575 19 I.IO " 19 ,25.0 IOI ... 3om. 646 45 576 19 1.20 " 19 25.0 ioi .... 4om. 991 56 577 19 1.30 " 19 25.01 ioi .... 5om. 1331 71 578 19 1.40 " 19 25.0] ioi .... ih. oom. 1499 93 579 19 3- 9 " 9 24.0 g7 .... 2h. 3gm. 3841 98 58i 19 4.32 19 24.0 g7 3h. 52m 5 444 85 585 " 20 g 36 A.M. 19 25.0 ioi .... 5h. 03111. 7286 94 586 " 20 IO.OI " 19 24.0 97 5h. 28m. 7887 85 590 " 20 12. 02 P.M. 19 24.0 97 7h. 2gm. 10 798 84 592 " 2O 2.07 " 19 24.0 97 gh. 34m. 13628 144 595 " 20 3-59 " 19 20.0 81 . . . . nh. 26m. 16083 91 598 " 21 9.26 A.M. 19 24.0 97 I3h. ism. 18 234 124 Agitated surface of sand 599 " 21 3-57 I .M. 20 22.0 89 .... 2h. 58m. 4358 IO2 layer at 8.49 A.M. 604 " 21 g.24 A.M. 20 25.0 IOI .... 4h. 40111. 7010 Si 605 " 21 10.26 " 20 24.0 97 5h. 42m. 8486 24 611 " 21 12.30 P.M. 20 24.0 97 .... 7h. 46m. ii 366 62 615 " 21 3.28 " 20 22.0 8g .... ioh. 44m. 15400 1 08 620 ! 24 g.3& A.M. 2O 23.0 93 . ... I3h. o8m. 18 366] 70 Agitated surface of sand 626 24 12.37 P.M. 2O 2O. O Si .... i6h. ogm. 21 980 92 layer at 9.03 A.M. 630 24 3.18 21 24.0 97 . . . . ih. 42m. 2319 7 636 " 26 IO.O2 A.M. 21 25.0 IOI i 4h. 45m. 6 766 98 641 " 26 12. IO P.M. 21 23.O 93 6h. 4im. 9486 97 646 " 26 3-59 " 21 23.O 93 . . . . ioh. 3001. 14 799 468 652 " 27 IO.29 A - M - 21 20.0 81 .... I3h. 25m. 18 641 664 Agitated surface of sand 658 27 1.57 P.M. 22 24.0 97 i8m. 422 235 layer at 12.02 P.M. 659 " 27 2 26 " 22 25.0 IOI 47m. I 112 336 664 " 27 3-31 " 22 25.0 IOI ih. 52m. 2737 346 670 27 4-52 " 22 25.0 IOI 3h. I3m. 4 745 468 675 11 28 10.02 A.M. 22 25.O IOI 4h. 5gm. 7 312 855 681 " 28 H.5I " 22 24.O 97 6h. 48111. 9906 882 685 " 28 3.15 P.M. 22 21 .O 85 . ... gh. 53m. I39I4 702 Agitated surface of sand 689 " 30 9-54 A.M. 2 3 25.0 IOI 1701. 414 126 layer at 1.14 P.M. 690 " 30 10.24 " 23 23.0 93 47m-: 1145 302 693 " 30 1 1 . 09 " 2 3 24.0 97 ih. 32m. 2 igi 880 (Hp " 30 1.48 P.M. 23 25.0 IOI 4h. nm. 6 020 144 704 " 30 4.48 " 23 23.0 93 7h. O2m. 96(11 102 Agitated surface of sand 712 " 31 IO.54 A.M. 24 25.0 IOI I5m. 380 83 layer at 4.04 P.M. and 715 " 31 11.24 24 25-5 IOI 45i. 2035 55 Dec. 31, 9.49 A.M. 719 " 31 2.10 P.M. 24 24.0 97 3h. 3im. 4934 560 Agitated surface of sand i8g6 layer at 2.38 P.M. and 724 Jan. 2 9.04 A.M. 25 24.0 97 22m. 280 140 Jan. i, 3. 26 to 3.34 P.M. 725 " 2 g.27 " 25 25.0 IOI 45i 825 152 733 " 2 II 33 25 25.0 IOI 2h. 47m. 3 500 290 Agitated surface of sand 73g " 2 2.35 P.M. 25 14.0 57 5h. 4gm. 7 280 560 layer at 11.16 A.M. 742 " 2 3-59 " 26 25.0 IOI 15111. 273 184 744 " 2 4.14 26 23.0 93 3001. 713 240 750 " 3 1O.3O A.M. 26 21.0 85 3h. 25m. 4623 254 756 3 2 05 P.M. 26 22.0 89 6h. 48111 8803 432 Agitated surface of sand 761 4 10.57 A.M. 27 24.0 97 i8m. 425 364 layer at 12.02 p M. 764 4 11.48 " 27 23.5] gs ih. ogm. i 547 368 771 4 2 22 P.M. 27 20.0 81 3h. 43111. 4997 438 811 g 10.30 A.M. 28 25-0 IOI ih. 2om. 2 OI4 2IO Sterilized filter Jan. 8. 817 9 1.5(1 P.M. 28 23-5 95 4h. 46m. 6 914 2 4 8 [outlet 11.44 A.M. 822 10 11.35 A.M. 28 I6. 5 67 ioh. 37m. 14050 152 Shut inlet 11.32 A.M., 825 " 10 12.56 I .M. 29 25.0 IOI 46111. I OOO 1 66 [P.M. 829 " 10 1.52 " 29 25-0 IOI ih. 42m. 2 42O 288 [from 8.52 A.M. to 12. 20 839 1 ii 11.35 A.M. 29 22.0 89 7h. oim. I 705 152 Agitated surface of S.L. COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 53 TABLE No. 4. Continued. Jewell System. Rate of .-: Collected. Filtration. in 3 1 Number a J a -a Period of ServiceSince I ast w c 3 U . I Run. SJ u O S 5 X Washing. i ~S Remarks. 5 Date. Hour. u. ~ Hours and Minutes. T: * o 1 = "- !5 ~ ^ - tn ~ jj = vfi y, O i 3 a 1896 8 4 i Jan. 13 g.59 A.M. 30 24.0 g? 1511. 344 104 Agitated surface of sand 842 13 10.38 30 25.0 IOI 54m. 1318 I 7 6 layer from 9.44 A.M. to 845 13 1.58 P.M. 30 25.0 ioi 4h. 1401. 6 236 290 5.40 P.M. 848 " 13 4.10 " 30 25.0 ioi 6h. 26m. 9411 2O.| 855 14 11.52 A.M. 30 23.0; g3 loh. 57m. 1 15 722 160 Agitated surface of sand 800 14 2.05 P.M. 30 23.0 93 13)1. lom. 18 819 1 68 layer from 8.54 A.M. to 866i " 14 3.16 " 30 iig.o 77 I4h. 2im. 20 36. 140 9-57 A.M. 868 " 15 g.4g A.M. 31 ,24-0 97 I5m. 346 148 869) " 15 10. ig " 31 24.0 97 45m. 987 116 872 15 10.44 " 31 24.0 97 ih. lom. I 574 168 878 15 12.56 P.M. 31 24.0 97 3h. 22m. 4 671 226 - - -i 15 3-07 " 31 24.0 97 5h. 33m., 7658 124 888 " 16 IO.52 A.M. 31 24.0 97 gh. 43m. 13626 I2O 893 " 16 1.05 P.M. 31 :24.o 97 nh. 56m. 16 746 194 goo " 16 3.05 " 31 23.0 93 I3h. 5601. ig 626 230 Agitated surface of sand 912 " 17 11.23 A.M. 32 21.0 85 05m. 102 430 layer Jan. 17, 10.03 913 17 11.28 " 32 24.0 97 lom. 222 218 A.M. gi6 17 11.38 " 32 25.0 ioi 2om. 4 62 174 917 17 11.48 " 32 23.5 95 30m. 6 7 2 240 918 17 11.58 " 32 24.0 97 4om. go2 292 921 17 12 OS P.M. 32 23.5 95 5om. i 172 194 922 17 12.18 " 32 23.5 95 ih. oom. i 372 272 925 17 1. 06 " 32 23.0 93 ih. 48m. 2532 290 930 17 2.07 " 32 24.0 97 2h. 49^. 3952 332 932 7 3.00 " 32 23.0 93 3h. 42m. 5 192 186 936 17 4.00 " 32 24.0 97 4h. 42m. 6 602 232 942 " 7 5.03 " 32 24.0 97 5h. 45m. 8 142 298 949 " 18 1O.I2 A.M. 32 23.0 93 7h. 3gm. 10 922 188 953 " 18 i.ig P.M. 32 .23.5 95 I oh. 46m. 15342 I So 960 " 18 2.44 " 32 24.5 99 I2h. nm. 17342 igS 963 " 20 g.45 A.M. 33 22.5 91 07m. 1 60 306 I M " 20 IO.OO " 33 25.0 ioi 22m. 470 286 965 " 20 10.23 " 33 24.0 g- 45"i- I 010 282 974 " 20 4. 18 P.M. 33 24.0 97 6h. 4om. 9480 44 [layer at 1.39 P.M. 979 " 21 11.32 A.M. 33 23.0 93 loh. 24m. 14830 168 Agitated surface of sand 986 " 21 4.20 P.M. 33 23.5 95 15(1. I2m. 21 580 173 Agitated surface of sand <)<)2 " 22 9.26 A.M. 33 23.5 95 i6h. 58m. 24052 103 layer at 4.30 P.M. <)<)," " 22 2.21 P.M. 33 23.5 95 2ih. 53m. 3O 7OO 106 IK)2 " 23 10. II A.M. 34 23.5 95 36m. 824 88 008 " 2 3 3.45 P.M. 34 24.0 97 6h. lom. 8724 318 013 11 24 IO.I5 A.M. 34 23.5 95 gh. I4m. 12994 1 60 015 24 1.47 I -M. 3-1 23-5 95 I2h. 46m. 18 027 124 022 " 25 9.58 A.M. 34 23.0 93 I7h. lom. 24 097 128 Agitated surface of sand 025 25 2. 15 P.M. 35 25.0 IOI 25m. 266 106 layer at 9.00 A.M., and 034 27 IO. IO A.M. 35 25.0 IOI 4h. 5im. 7 186 688 from 9.07 A.M. to 11.32 1140 " 27 1. 09 P.M. 35 25.0 IOI ~h. som. II 266 I 196 A.M. "45 " 27 4 5 " 35 23.0 93 loh. 52m. 15 426 952 Agitated surface of sand 051 " 28 g.I5 A.M. 35 23.0 93 I3h. lom. 18 go6 500 layer at 3.10 P.M. "54 " 28 I.OO P.M. 36 23.0 93 ih. I2m. I 713 2 500 [layer at 3.42 P.M. ,, , " 28 4-35 " 36 24.0 97 4h. 34m. 6463 i 600 Agitated surface of sand 1,1,1, " 29 IO. 14 A.M. 36 23.5 95 6h. oim. 8583 3016 Agitated surface of sand 06g " 29 2.04 P.M. 37 23.0 93 ih. 24111. 2 02.) I 620 layer at 11.43 A.M. 070 " 2g 5.04 " 37 23-5 95 4h. 24m. i 6 174 675 075 " 30 11.03 A.M. 37 24.0 97 7h. I3m. 10 ig4 876 Agitated surface of sand 077 " 30 1.05 P.M. 37 23-5 95 gh. ism. 13 094 780 layer at 10.09 A.M. ,,-.i " 30 2.56 " 37 21 .O 85 nh. o6m. 15 534 804 084 " 31 10.55 A.M. 38 24.O 97 3h. 3om. 5 ooi 811 Agitated surface of sand 089 " 31 2.36 P.M. 38 20.0 81 7h. nm. 10031 910 layer at 11.29 A.M. 137 Feb. 5 IO.22 A.M. 39 25.O IOI 27m. 834 612 i |i " 5 11.50 " 39 24.0 97 ih. 55m. 2924 468 M5 5 3.08 P.M. 39 22 ."o 5h. ism.j 7534 600 54 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Jewell System. Rate of i Collected. Filtration. t | ~ 1 Number S. i s. Period of ServiceSince o ^ u u - 1 X, Date. Hour. of Run. !l " D ^ -< s X Last Washing. Hours and Minutes. It 6 Remarks. 3 s i a " J n iSg6 II5<> Feb. 5 5.08 I M. 39 22.0 Sg 7h. I3m. 10 144 8g2 1156 " 6 10. 13 A.M. 39 25.0 IOI 8h. 04171. ii 324 i 620 1 1 60 " 6 12. 14 P M. 39 23.5 95 . . . . ioh. D5 m - M 174 2 1^5 Agitated surface of sand 1164 " 6 3-13 " 40 24.5 99 40111. 1014 i 282 layer at 1.54 P.M. ii6g " 6 4.15 40 25.0 IOI ih. 42m. 2 394 i 196 1174 " 7 IO.I8 A.M. 40 24.0 97 4h. o8m. 5814 238 [layer at 11.17 A.M. 1178 " 7 1.32 P.M. 40 23.0 93 7h. iSm. 10 3U 480 Agitated surface of sand 1184 " 7 5.28 " 41 23-5 95 ih. 2Otn. i 814 i 200 Agitated surface of sand iigo 8 10 49 A.M. 4 25.0 IOI ih. 43 ". 2 337 1785 layer from 4.08 P.M. to 1192 " 8 2. 2O I .M. 41 19. o 77 5h. Mm. 6 727 500 5. 18 P.M. iigfi " 8 3.II " 41 22. O 89 6h. 05111 7827 i 900 1204 " o IO.22 A.M. 42 25.0 101 .... 3gm go8 675 1205 " 1. 01 I .M. 42 22. O 89 .... 3h. 14111 4278 616 Agitated surface of sand 1212 " 3.16 " 42 ig.o 77 5h. 2gm 7 138 516 layer at 1 1.47 A.M. 1217 " o 5.02 " 42 22. O Sq 7h. nm 9328 i 155 Agitated surface of sand 1222 " I IO.O9 A.M. 43 24.5 99 24m 578 4 T 5 layer at 3.20 P.M. 1225 " I I2.5O P.M. 43 24.0 97 .... 3h. oim 3978 735 Agitated surface of sand 1228 " I 3-13 " 43 23.0 93 5h. 24m 7 108 567 layer at 12.40 P.M. 1232 " II 5-12 " 43 24.0 97 7h. lom 9578 2 365 Agitated surface of sand 1241! " 12 3-18 " 45 18.0 73 ih. 4im 2 2O5 2 42O layer at 4.00 P.M. 1242 " 12 3-40 " 45 26.0 105 2h. 03m 2 6g5 2 360 1243 " 12 4-45 45 18.5 75 . . . .1 3h. o8m 4215 324 1249 13 9.48 A.M. 45 23.0 93 [ 4h. 36m 6 115 2Og 1252 13 12.24 I -M. 45 22.5 9 1 7h. I2tn g 675 317 1255 13 2.18 " 45 25.0 IOI gh. O2m. 12 igs 740 Agitated surface of sand I26l 13 4.50 " 45 23.0 93 iih. 34m. 15615 930 layer at i.oi P.M. 1266 M 10.24 A.M. 46 93 5om. I 178 234 I27O I.I7 I .M. 46 23.0 93 3h. 38m. 5013 I 110 Agitated surface of sand 1274 " M 3-19 " 46 23-5 95 4h- 53m. 6738 I IIO layer at I2.5g P.M. 1278 " 14 4.46 " 46 24.0 97 . . . .| 6h. 2om. 8818 940 1284 " 15 1288 15 10.15 A.M. 1.2g P.M. 46 46 23.0 23.0 93 93 .... 8h. igm. ... .! iih. 28m. 11468 790 15 688 i 097 Shut inlet 2.56 P.M., out 1292 " 15 3-O2 " 46 18.0 73 .... I3h. oim. I778S! 845 let 3.09 p.M. 1297 15 5.20 " 47 24.0 97 ih. 46m. 2 353 319 1303 17 IO. 12 A.M. 47 24.0 97 . . . . 3h. o8m. 4 163 I 100 1307 17 I.4U P.M. 47 24.0 g7 .... oh. 31 m. 8783 I 285 Agitated surface of sand I3II " 17 3.10 " 47 23 -! 93 .... 8h. oim. 10833 I 362 layer at 11.42 A.M. 1317 " 17 5.18 " 48 25.0 IOI : i8m. 394 5&g 1321 18 10.27 A M. 48 22.5 9 .... ih. 57m. 2 7M I 885 1325 " IS II. 5 S " 48 23.0 93 .... 3h. 26m. 4 794 I 250 1329 " 18 2.23 P.M. 48 22.0 89 .... sh. 5im. 7994 I 645 Agitated surface of sand 1332, " 18 4-52 48 13.0 52 8h. 2om. I 270 layer at 1.56 P.M. 1340 18 5.14 48 23.0 93 .... 8h. 42m. ii 804 1855 1344 19 10. IS A.M. 48 16.0 65 . ioh. oim. 13 4U 760 1348 19 n-35 " 48 22. o 89 .... iih. i6m. 15 974 50 Agitated surface of sand 1352 T 9 3.08 P.M. 49 22.5 gi 2om. 421 695 layer at 11.43 A.M. 1357 " 9 5.03 " 49 26.0 105 2h. ism. 3 on 2 170 1363 20 11.04 A.M. 49 24.0 97 3h. 05m. 4151 28 1367 " 20 I2.O5 I -M. 49 23.0 93 .... 40. o6m. 5 591 88 1372 " 2O I. Og " 49 23.! 93 5h. 07111. 6851 43 Agitated surface of sand 1376 " 20 2.IO " 49 23.0 93 .... 6h. o8m. 8 321 59 layer at 12.46 P.M. 1378 " 20 3 12 " 49 24.0 97 ... .1 7h. torn. 9771 206 1381 " 20 4.05 " 49 25.0 IOI .... 8h. 03m. 10031 575 1385 " 20 5.IO " 49 24.0 97 . . . .j oh. o8m. 12 64! i 3 go 1393 " 21 II.O4 A.M. 49 25.0 IOI .... ioh. 24m. 14 381 IO 1395 1398 " 21 " 21 12.45 P.M. 3-og " 49 49 21.0 85 23.0 93 I2h. osm. I4h. 26m. 16491 ig 621 120 [layer at 1.22 P.M. 4go Agitated surface of sand 1401 " 21 4-55 49 22-5 91 i6h. I2m. 22 Oil 560 Agitated surface of sand 1409 " 22 IO.24 A.M. 50 25.01 101 . . . 58m. I 331 65 layer at 5.10 P.M. 1410 " 22 1. 08 I .M. 50 21.0 85 ... 3h. 4201. 5 071 715 [layer at 1.28 P.M. 1413 " 22 3-05 " 5" 24.0 97 5h. 46m. 7781 i no Agitated surface of sand COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 55 TABLE No. 4. Continued. Jewell System. Rate of ,- Collected. Filtration. fc . C 2 1 Number . C d Service Since Last 5 ^S u u - a of Run. V . l- : I Washing. Hours and s| S.S Remarks. ^ Date. Hour. ! 111 o Minutes. lal i = C a 33 ~o.S d ;j ic_> (X <_> 5 x m 1896 1416 Feb. 22 4.58 P.M. 50 23.0 93 7h. 4901. 10441 700 1421 " 24 10.24 A.M. 50 23.0 93 - -- gh. 2501. 13 225 1 605 1424 24 1.20 I .M. 50 22. 5 91 - - I2h. 2im. 17235; 404 1427 24 3.26 5 24.5 gg I4h. 24111. 20 115 665 Agitated surface of sand 1432 24 5-19 " 5" 2 3-o g3 l6h. 1701. 22 844 6og layer at 2.15 P.M. 1438 1 25 10.30 A.M. 50 23.0 93 I7h. 58m. 25 195 960 1442 1 25 I.lS P.M. 51 23.5 95 i6m. 335 75 1446 1 25 3-12 " 51 22.5 91 .... 2h. 10m. 2965! 610 1451 " 25 5.02 51 23-5 95 4h. oom. 5515 450 > 157 " 26 IO. 2g A.M. 51 25.0 IOI -. . 5h. 47m. 7755i 420 1461 " 26 12. 08 P.M. 51 23.0 93 7h. 26m. 10 115 295 1467 " 26 3-15 " 51 24.5 99 loh. 33111. 14625 770 1470 " 26 5.18 51 20. 5 83 - I2h. 3&m. 17495 3 280 1477 " 27 IO.32 A.M. 51 24.0 97 - - - I3h. i6m. 18415: igi Agitated surface of sand I po 27 1.45 I -M. 52 26.0 105 .... 27m. 670 6g5 layer at 9.32 A.M. and 1484 27 3-02 " 52 30.0 112 - ih. 44m. 2 8lO 97 12. 18 P.M. 1489 27 5-12 " 52 25.5 103 3h. 54m. 6 430 I 215 1495 " 28 Tn c T o i C. Sterilized filter, Feb. 1497 " 28 10.42 A.M. 52 29.5 120 - - - - 5h. 54m. 9950 i 33 28. 1500 " 28 11.50 " 52 33-5 136 7h. 02m. 12050 i 820 1511 i * . C. 1513 " 29 io.3g A.M. 53 53 *3 9 27.0 log . . . . ! 4im. i 195 97 1517 29 1.38 P.M. 53 21.5 87 3h. 02111. 4885 2 555 1521 29 3-18 " 53 27.0 log . . . .; 4h. 42m. 7495 i 910 1524 : 29 3-38 " 53 27.0 log .... sh. O2m. 3 035 2 170 1527 29 5. II 53 27.0 log .... 6h. 35m. 10465 i 175 1532 Mar. 2 9-35 A.M. to 3.15 P.M. Z l i g&5 C. 1533 " 2 9.42 A.M. DJ 53 26.0 105 ..... 7h. 36m. 12085 3455 1537 " 2 10.25 53 26.0 105 .... I 8h. igm. 13185 1541 2 1.36 P.M. 53 25-5 103 nh. 3om. 18 165 1545 " 2 3-19 " 53 25-0 IOI I3h. I3m. 20785 I ?45 1550 " 2 5-10 " 53 25.0 loi I5h. 03m. 23 575 I 445 1553 " 2-3 3.l8 P.M. to 3-2O P.M. C -7 t 570 C. 1558 " 3 10.42 A.M. J J 53 23.0 g3 17(1. i6m. 26485 900 1564 3 12.58 P.M. 53 21 .0 85 igh. 22m. 29515 I 210 [layer at 2.09 P.M. 1566 3 3-13 " 53 22.5 9 2ih. 3im. 3 2 395 i 300 Agitated surface of sand 1568 " 3"4 3-2O I .M. to 3-2O I .M. 2? 8 I 465 C 1571 " 3 5.13 P.M. 53-54 53 43.0 21-5 IO4 85 23!!. 3im. 34935 i 885 1577 4 10.48 A.M. 54 27-5 in ih. 3om. 2 393 610 1581 4 I.OO I .M. 54 28.0 114 3h. 42111. 6 0031 880 1585 4 3-20 " 54 28.5 116 6h. O2m. 9853 i 320 159 4 5 05 " 54 28.0 114 7h. 47m. 12783 595 1596 5 IO-39 A.M. 54 28.0 114 gh. Sim. 16253 885 |(jOO 5 12.53 I -M. 54 27-5 III I2h. O5m. IQ 893 298 1603 " 4~5 3-2O P M. to 3.20 P.M. 27.8 112 815 r 1604 5-6 3. 2O 3. 2O 54 e i-C c 30. 3 6 to C. 1605 " 5 3.20 P.M. 3-4 3 j 54 28.0 114 . . . . I4h. 32111. 24013 )< = 1611 5 5-14 " 54 25.0 IOI . . . . i6h. 26m. 27063 725 1617 " 6 10-35 A.M. 54 23.0 93 i8h. I7tn. 29743 206 Shut inlet 10.33 A.M., 1620 " 6 II. 17 " 55 38-0 54 I8m. 53 * outlet 10.41 P.M. 1622 " 6 12.41 P.M. 55 54 .... ih. 42m. 3 656 S6 1626 " 6 3.20 " 55 36.0 146 . . . . j 4h. 2im. 9 526 57" 1627 6-7 3.2O P.M. to 3-2O P.M. c r 28 7 116 ! ..... C 1632 6 5.22 I .M. j j 55 * . / 34-5 140 i f,h. 2301. 13(06 550 1638 " 7 10.44 A.M. 55 25.0 IOI .... Sh. ism. 17206; 128 1642 7 12.54 I -M. 55 25.0 IOI loh. 25m. 2042(1 150 1647 7 3.2O " 55 25.0 IOI I2h. 5im. 24 046 197 1648 7-9 3 20 P.M. to 3.2O P.M. 55-56 23-5 95 C. 1651 7 5.19 P.M. 55 22-5 91 I4h. 5om. 26 g 3 6 108 1657 1 9 11.00 A.M. 25.0 IOI ih. 35m 2 427 5* 1662 9 12.52 P.M. 5f> 25-5 103 3 h. 2 7 m. 5 347 f i WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Jewell System. Collected. Number of Run. 56 57 5.6 56 57 57 57 57 57 57 57 57-5S 57 58 53 58 58-59 58 59 59 59 59 59 59 59-60 60 60 60 60 fco 60 60 60-6 1 61 61 61 61 61 61 61-62 62 62 62 62 63 63 63 63 63 63 63 64 64 64 64 64 64 64 64-6 65 65-6f Rate of iltration. i | i 2> Period of ervice Since Last Washing-. Hours and Minutes. *- C iy ^JU [X. 3 5 .2.8 | 305 95 no! 82 in 58 166 295 79 34 165 325 105 155 560 i&5 37 121 7OO 742 325 420 230 32O 410 510 131 64 49 72 39 169 M7 98 188 261 161 156 128 95 8q C7 Remarks. c i Is. i j ~i< o o ,_ - o. ?f 2 a 7. 1668 1669 1673 1679 1683 1687 1688 1694 1700 1703 1707 1711 1714 1720 1724 1728 1729 173 174 74 174 75 175 176 176 176 177 77 177 177 178 1786 1791 1792 1797 1 79? 1804 1811 iSis 181- l8ii 182; lS2< lSj( 183 184 184 184 184 185 1 86 1 86 187 187 187 187 188 188 188 188 189 Date. Hour. 1896 Mar. 9 " 9-10 9 10 " 10 " 10 " 10-11 " 10 " II " II " II " I I-I2 " II " 1 2 " 12 " 12 " 12-13 12 " 13 13 13 13 M 14 14 M M M 4 M " 16 " 16 " ie " 1 6 " 16 " 16 " 16 " 17 " 17 17 " 17 1? 1 17 ) " 17 ) " 18 " 18 ! " 18 i " 18 ) " 18 , " 18 3 " 19 3 " 19 9 19 t 9 19 ! 9 j 19 5 "20 5 " 20 I " 20 3.30 P.M. 3.30 P.M. 10 3. 10 P.M. 5.07 P.M. 10.23 A.M. 1-35 P.M. 3-10 " 3. IO P.M. to 3.2O P.M. 5.17 P.M. 10.23 A.M. I .30 P.M. 3 . 20 " 3.20 P.M. 10 3.26 P.M. 5.II P.M. 10. IS A M. 12.57 P M. 3.26 " 3.26 P.M. to 3. 15 P.M. 5.13 P.M. 10.32 A.M. I .12 P M. 3-15 " 5.03 " 9. 30 A.M. to 10.30 A.M. 10.30 A.M. 10.30 A.M. to I.OSP.M. I .08 P.M. 1. 08 P.M. to 3.15 P.M. 3.15 P M. 4.02 " 4.52 " 9.00 A.M. to 10.28 A.M. 10.28 A.M. 10.28 A.M. to 1. 12 r.M I . 12 P.M. I . 12 P.M to 3.15 P.M. 3.15 P.M. 5-05 " 9. 2O A.M. to IO.27 AM 10.27 A.M. IO.27 A.M. to I.I4 P.M I. [4 P.M. I . 14 P.M. 10 3.16 P.M. 3.18 P.M. 5-13 " IO.23 A M. 10.28 A.M. to I. 08 P.M I.oS P.M. I .08 P.M. 10 3.24 P M. 3.24 P.M. 5.02 " g.OOA M. to IO.4JA.M II .08 A.M. 12.27 P.M. 10.45 A.M. to 12.27 P-M 3.02 P.M. 12.27 P.M. to 3.02 P.M 5.O2 P.M. 3.02 P.M. to 5.02 P.M. g.OOA.M. " IO.25A.M 10.25 A.M IO.25 A.M. to 1. 08 P.M 5 - 4-4 4-5 3.0 4-5 5 - 5- 5 5.0 4.0 6.0 4.0 5-4 3-5 5-0 5.0 5.0 4.8 4.0 21 5 25.0 24.5 24.0 23.6 23.6 23-5 25.0 >4 6 IOI j. 6h. 0501. 9367 C. c. c. c. c. c. c. c. c. c. c. c. c. c. c. c. )C. >|C. c. > )C. 99 93 99 IOI 103 IOI 97 i 105 97 103 95 101 IOI IOI 100 97 87 IOI 99 97 95 95 95 IOI 7h. 42m. gh. 28m. 2h. oom. 3h. 3jin. ii 739 14307 2980 5400 jh. 42m. 7h. i8m. loh. O2m. i ih. 52m. 8670 ii 080 15 290 1 8 060 13)1. 43m. 53"i- 3h. 32111. 6h. 01 m. 7h. 48m. 04111. 2h. 44m. 4(1. 47m. 6h. 35111. 20 6go i 303 5 213 8903 ii 543 104 4274 7234 9854 Sh. 3201. 12 664 58m. I 428 24.5 25.0 25. u 99 IOI IOI 3h. 0501. 3h. 5201. 4h. 42m. 4 54t 57of 6 giC 23. c 24. c 25. c 24.- 24. c 24. c 23-: 23.? 24-1 24. c 24.1 24- 23. c 24. c 24. 24. 23. ( 24. 24. 23. 24. 24. 24. 24- 24. 19. 22. 24. 24. 22. 93 97 IOI 100 97 97 94 95 : 99 ! 9? 97 99 > 93 > 97 i gg 99 ) 97 > 99 99 > 95 3 97 > 97 97 5 99 > 97 3 77 2 go 7 too > 97 3 g2 , . . . 6h. 24111. 9 4&f 53m- i 39C .... 2h. 56m. . . . . 4h. 46m. 443: 702; . . . . 6h. 3Sm. 967; . . , 2h. 0401. 305. 4h. oSm. 6h. O3m. 49m 6134 67 8 784 6c i 197 i/ 22 3h. 2gm 5 19 J. 41 4: 7 3f 7 4 IV 5h. 45m 7h. 23m 844 1077 09 m ih. 28111 210, 26C 2 1 2O I 20C 53C 4h. O4m 5 840 115 j I 2O( 6h. O3m 8 510 22 . . . . i I OO( 58 48m I 220 8o< 8oc COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. TABLE No. 4. Continued. Jewell System. 157 1 a z j_ 1895 1901 1902 1907 1908 1913 1914 1919 1920 1922 1923 1931 1936 1937 1941 1942 1947 1948 1954 1955 J 959 1962 1964 1969 973 1983 1989 1993 1998 200 1 2005 2008 2OI2 2OI4 2OI5 2016 20I 7 2018 2Olg 2021 2023 2024 2025 2026 2O27 202S 2O29 2033 2035 2040 2043 2047 2050 2054 2057 2065 2076 2083 2099 2103 2106 2110 2121 Collected. Number Run. Rate of Filtration. b T3 a 1 Period of Service Since Last Washing. Hours and Minutes. ^ si ^ I.JU tL, .a t 3 U s.5 is M^ m Remarks. I Is L> I s - ]{I Date. Hour. 1896 Mar. 20 " 20 " 20 " 20 20 " 2 " 2 " 2 " 2 " 2 " 2 " 2 " 23 23 23 23 1 23 23 23 1 23 24 24 1 24 24 24 24-25 25 1 25 1 25 25 1 25 25 " 25 " 26 " 26 " 26 26 " 26 " 26 25-26 " 26 " 26 " 26 " 26 26 " 26 " 26 " 26 " 26 " 26 " 26 ;; 26 26 " 26 " 26-27 ;; 27 27 1 27 , 27 27 :: 2? 27 " 27-28 I.OS P.M. 1. 08 P.M. to 3.33 P.M. 3 33 i -M. 4-53 " 3-33 P.M. to 5.30 P.M. g.OO A.M. " 10.45 A.M. 10.45 A.M. 12.58 P.M. IO.45 A.M. to 12.58 P.M. 12.58 P.M. " 3.20 " 3.20 P.M. 3-20 P.M. to 5.00 P.M. g.OO A.M. " 10.25 A.M. 10.25 A.M. 10.25 A.M. to 12 M. 12.00 M. 12. OO M. to 3.OO P.M. 3.OO P.M. 5.16 " 3.00 P.M. to 5.30 P.M. g.OO A.M. " II. 30 A.M. 11.30 " " 2.30 P.M. 2.30 P.M. " 4.42 " 442 " " 8.30 " 8.30 " " II 30 " 11.30 " " 2. 30 A.M. 2.30 A.M. to 5.30 A.M. 5.30 " " 8.30 " 8.30 " " 11.30 " 11.30 " " 2.30 P.M. 2.30 P.M. " 5.30 " 5.30 " " 8.30 " 8.30 " " 11.30 " I 00 A.M. 1. 08 " 1.18 " 1-33 " 1.48 " 2.18 " 11.30 P.M. to 2.30 A.M. 2.48 A.M. 3.48 " 4.14 " 4.48 " 5-15 5-3 " 2.30 A.M. to 5.30 A.M. 5-49 A.M. 5.3O A.M. to 8.30 A.M. 8.30 " " 11.30 " II.3O " " 2.30 P.M. 2.30 P.M. " 5.30 " 5.30 " " S-30 " 8.30 " 11.30 " 11.30 " " 2.30 A.M. 2.30 A.M. " 5.30 " 5-30 " " 8.30 " 8.30 " " 11.30 " 11.30 " " 2.30 P.M. 2.30 P.M. " 5.30 " 5.30 " " 8.30 " 8.30 " " 11.30 " II.3O " " 2.3O A.M. 66 66-67 67 67 67 67-68 68 68 68 68-69 69 69 70 70 70 70 70-71 71 72 71-72 72 72-73 73 73-74 74 74-75 75 75-76 76 76-77 77 77-78 78 79 79 79 79 79 79 78-79 79 79 79 79 79 79 79 79 79-80 80-8 1 81-82 82-83 83 84 84 84-85 85 86 86 87 87-88 88 88-89 24.0 23.1 23.0 23-5 24-3 23.2 24.0 21.0 22.6 22.8 24.0 22.0 23 6 21.0 21.2 25.0 23.3 24.0 24.5 24.9 25.5 23.7 21.4 25-4 22.6 23.4 23.2 2 3 .8 24.9 2 5 .8 24.3 25.6 24.2 25.0 25.0 25.0 25.0 25.0 25.0 22.8 25.0 24.5 2 4 5 24.5 25-5 24-5 23.6 23.0 23.0 26.5 2 3.3 24.1 2I.I 24.7 20.1 23.2 23.6 2 4 .0 2 3 .8 21-5 24.2 23-5 23-5 97 93 93 95 98 94 97 85 9i 92 97 89 95 85 86 IOI 94 97 99 IOI 103 95 86 IO2 91 94 94 96 IOI 104 98 103 98 10 10 IO 10 10 10 92 IOI 99 99 99 103 99 95 93 93 107 94 97 89 IOO 89 94 96 97 96 87 98 95 95 24111. 593 600 I OOO I 000 I 000 I 2OO 495 415 465 860 895 I 905 785 440 405 700 I 250 800 475 i 245 I 230 179 So 270 132 74 306 i 030 495 48 158 405 178 107 156 113 171 229 420 345 215 600 435 460 415 232 221 482 124 171 355 700 520 805 330 477 485 575 415 128 i 650 150 H7 540 C. C. C. [layer at 11.25 A.M. Agitated surface of sand C. C. Agitated surface of sand layer at 3.05 P.M. C. C. C. Agitated surface of sand layer at 10.29 A.M. C. Agitated surface of sand layer at 3.30 P.M. C. [layer at 10.36 A.M. Agitated surface of sand [layer at 3.54 P.M. Agitated surface of sand [layer at 9.30 P.M. Agitated surface of sand [layer at 3.41 A.M. Agitated surface of sand [layer at 11.05 A.M. Agitated surface of sand [layer at 3.51 P.M. Agitated surface of sand [layer at 10.36 P.M. Agitated surface of sand This series of results on run No. 79 was used in obtaining the aver age bacteria for this run, but not for the day. Agitated surface of sand layer at 4.07 A.M. [layer at 7.47 P.M. Agitated surface of sand [at 1 1. 54 P.M. & 2. 1 1 A.M. Agitated surface of S.L. [layer at 6.15 A.M. Agitated surface of sand [layer at 11.33 A M - Agitated surface of sand [layer at 5.55 P.M. Agitated surface of sand Agitated surface of sand layer at 10.08 P.M. 05 m. ih. 25m. "3 2053 45m. 2h. 55m. I 085 4095 . . . . ih. 4im. 2371 ih. 47m. 2 5 4 3h. 2om. 5874 ih. 4om. 46m. 2 421 I 071 I2m. 2om. 3om. 45m. ih. oom. ih. 3om. 419 579 829 I 179 i 579 2339 2h. oom. 3h. oom. 3h. 23m. 3h. 57m. 4h. 25m. 4h. 39m. 3049 4529 5049 5789 6429 6789 4h. 58m, 7159 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Jewell System. .0 8 2; Collected. Number Run. Fill 1 a! ^ 3 11 CJ o a 6 5 D r- < 3 5 ul fe j c Period of "C % Washing. ^Jt Hours and ^ o Minutes. 1 g| i-iu !b !S u u . ^ e Remarks. Date. Hour. 2127 2131 2136 2139 2143 214(3 2148 2149 2150 2151 2152 2155 2158 2161 2165 2169 2173 2182 2185 2189 2192 2196 2199 2203 2206 22IO 2216 2220 2224 2229 2234 2237 2242 2247 225(1 2255 2262 2257 2271 2276 2281 2286 2289 2294 2299 2302 2306 2307 2308 2310 2322 2323 2324 2325 2326 2327 2328 2329 2331 2334 2338 2340 2341 1896 Mar. 28 " 28 " 28 " 28 " 28 " 28 " 28 " 28 " 28 " 28 " 28 " 28 " 28-29 " 29 " 29 29 29 29 " 29 " 29 " 29-30 " 3 " 3 " 30 3 " 3 " 31 " 3i " 31 April i " i " 2 " 2 2 " 3 3 4 4 4 6 6 6 7 7 :: ! :: 5 8 8 (< g 8 8 " 8 8 8 8 8 9 9 9 " 9 2.30 A.M. to 5.30 A.M. 5.30 " " 8.30 " 8.30 A.M. " 11.30 " II.3<> " " 2.30 I .M. 2.30 I .M. " 5.30 " 5.30 " " 8.30 " IO.25 I -M. 10.35 " 10.45 10-55 II.O5 " 8.30 I .M. to II.3O I .M. 11.30 " " 2 30 A.M. 89 89-90 90 90-91 91 91-92 92 92 92 92 92 92 92 23.8 2 4 .8 23-9 22.9 24-5 24.2 25.0 25.0 24-5 24-5 24-5 23-4 22.5 9 6 100 97 93 99 98 IOT 10 I 99 99 99 95 91 08 407 321 238 62 119 174 261 1 86 3 3 221 2-13 193 119 234 177 294 580 595 240 4<>5 125 256 476 58l 785 672 650 390 I 495 845 545 525 240 224 205 92 310 60 62 90 30 36 27 30 44 118 53 82 88 76 65 20 1 1 86 125 104 194 185 75 63 72 157 182 152 Agitated surface of sand layer at 3.41 A.M. Agitated surface of sand layer at 9.26 A.M. Agitated surface of sand layer at 4.11 I .M. C. Agitated surf, of sand C. layer at 10. n P.M. C. C. C. Agitated surface of sand layer at 6.17 A.M. Agitated surface of sand layer at 2.14 I .M. Agitated surface of sand layer at 9.31 P.M. Agitated surface of sand layer at 5.58 A.M. Agitated surface of sand layer at 2.01 I .M. Agitated surface of sand layer at 10.01 A.M. Agitated surface of sand layer at 3.05 P.M. [layer at 2.39 P.M. Agitated surface of sand [layer at 11.25 P.M. Agitated surface of sand [layer at 11.54 A.M. Agitated surface of sand Agitated surface of sand layer at 4.27 P.M. Agitated surface of sand layer at 1.09 P.M. Agitated surface of sand layer at 9.00 A.M. Agitated surface of sand layer at 2.09 I .M. Agitated surface of sand layer at 12.07 v.M. Agitated surface of sand layer at 11.42 A.M. C. C. C. C. Agitated surf, of sand layer at 10.17 A.M. 2h. 42tn. 3 826 2h. 52111. 4 126 3h. 02m. 437<> 3h. 13111. 455f 3h. 22m. 4 804 5.30 " " 8.30 " 8.30 " "11.30 " 11.30 " " 2.30 I .M. 2.30 I .M. " 5.30 " 5.30 " " 8.30 " 8.30 " 11.30 " 11.30 " " 2.30A.M. 2.3O A.M. " 5.30 " 5.30 " " 8.30 " 8.30 " " 11.30 " 11.30 " " 2.30 I .M. 2.3O I .M. " 5.30 " 9.15 A.M. " 11.30 A.M. 11.30 " " 2.30 " 2.30 P.M. " 5.30 I .M. 9. 15 A.M. "11.30 A.M. II.3O " " 2.30 I .M. 2.30 P.M. " 5 30 " 9.30 A.M. " I 1.30 A.M. II.3O " " 2.30 I .M. 2.30 I .M. " 5.30 " 9.20 A.M. " 11.30 A.M. 2.30 I .M. " 5.30 P.M. 9.30 A.M. " 11.30 A.M. 11.30 " " 2.30 P.M. 2.3O I .M. " 5.30 " 9.2O A.M. " II.3O A.M. 11.30 " " 2.30 I .M. 2.30 I .M. " 5.30 " 9.25 A.M. " 11.30 A.M. II.3O " " 2.30 I .M. 2.30 " " 5.30 " I I . OO A.M. II. TO " II. 2O " 9.2O A.M. to II.3O A.M. 11.30 " " 2.30 I .M. 3.32 I .M. 3-35 " 3.38 " 3-41 3-44 3-59 " 4.14 2.3O P.M. to 5.30 I .M. 9 2O A.M. " 11.30 A.M. 11.57 A.M. 12.32 P.M. I. O2 " 93 93^94 94 94 94-95 95 95-96 96 96 90-97 97 98 98 99-100 IOO-IOI 102-103 104 104 105 105-106 1 06 106-107 108 108-109 101) 109-1 10 I IO IIO-I 1 1 1 1 1 112 I 12 II2-II3 1 3 I 3 "3 113 113 114 114 114 114 114 114 114 113-114 114 14 114 114 24-1 23-9 23-9 25-1 23-5 27-5 22.7 24-6 23-8 23-7 24-4 25-1 24-2 23.3 24-0 23-7 23-2 23-1 24-0 24-0 24 3 24.7 24-3 23-7 25 8 23-2 23-9 25-2 22-9 24-8 24-4 24-3 25-5 24-5 24-0 24-3 24-7 22. O 24-0 :?.o 25.0 25.0 25-0 25.0 23-4 24.0 24.0 24.0 24.0 98 97 97 IOI 95 ii i 92 99 96 95 99 IOI 97 94 97 95 93 93 97 97 98 IOO 98 96 104 94 97 IOI 93 IOO 98 97 103 99 97 98 IOO 89 97 IOI IOI IOI IOI IOI 95 97 97 97 97 .. i 3h. 26m. 5 014 3h. 36m. 5 244 3h. 46m. 5 494 03111. 6 1 ( 6m. 131 09111. 211 12111. 291 15111.: 361 30111. 711 45111. I IOI 4h. 5&m. 7 121 5h. 3im. 7961 6h. Dim. 8 621 COMPOSITION OF OHIO RIVER U ATER AFTER PURIFICATION. 59 TABLE No. 4. Continued. Jewell System. I 1 . S Collected. Filtration. > c .i ~^ 5T^ Period of 1 i " 1 \- . i e. 5 & TJ Service Since fi -;:,_; U ^ g ""of"" Last 5 H S a " Remarks. Run. Washing. Sjgfc. y. Date. Hour. ul its "s Hours and ~a f u - ~ c Spr ii a? " ~23 "u X U 7. , 2 fb 03 1896 234J April 9 1 1 30 A M to 2 30 P M I 141 15 3 4 nc 102 C. 2346 9 2.30P.M. " 5.00 " 115 23.6 95 .... n>6 C. Agitated surf, of sand 2349 o 0.15 A.M. 115 24.0 97 .... 3h. 3im. 4983 62 layer at 4.32 P.M. 2351 0.45 U5 24.0 97 4h. oim. 5 683 33 2353! 1.15 "5 24.0 97 4h. 3im. 6 313 28 2356 o 2368 " o 1.45 US 25.0 ioi ; TT TO V t(l ? 1O I \f IIC m T m 5h. oim. 6973 39 C. 2374 " O 2 1O \f " C 1O " I 1 ft "II (17 C. 2380 " 9 20 A M " 1 1 30 A M 116 24 5 99 -j 9c - J^>^> 2382. 12.05 P.M. 116 24.0 97 .... 6h. oim. 8 765 42 layer at 9.00 A.M. 2384 12.35 116 1.4.0 97 .... 6h. 3im. 9454 2 3 2386 1.05 1 16 23.5 95 7h. oim. 10 118 31 2388] 1.35 " 116 23.0 95 .... 7 h. 3 im. 10 905 30 Shut inlet 1.34 P.M., 2389 2.01 117 24.0 97 .... O3m. 43 78 outlet 1.45 P.M. 2390 2.04 " 117 25.0 IOI 06 m. 123 So 2391 2.07 " "7 25.0 lot 09111. 193 57 2392 1 2.10 " 117 24.0 97 1 2 in . 263 32 2393 2.1 3 " 117 24.0 97 15111. 343 38 2394 2.28 " 117 24.0 97 3<>m. 73 31 2395 2.43 " 117 24.0 97 45m. i 093 37 2397 " TT -3O P M to 2 1O l> M 1 1 6 117 24.0 O7 J7 C. 2399 3.13 P.M. 117 24.0 V/ 97 ih. ism. 2 793 4 / 27 2405 3.58 " 117 24.0 97 2h. oom. 2903 24 2408 4.28 " "7 24.0 97 2h. 3om. 3 633 43 2410 " 4.58 " 2412 ; 2.30 P.M. to 5.30 P.M. 117 24.0 T T 7 o 1 . r, 97 97 3h. oom. 4363 38 18 C. 2415 " 3 9. 2O A.M. " II. 30 A. M.I 118 < ^ m 32 c. 2417! " 3 11.30 " " 2.30 P.M.! 118 -t *- j i 2.1 T Q8 j~ S3 2419 " 3 4-55 I .M. 118 ^4 - J v 23.0 93 6h. 4im. 9640 81 layer at 11.53 A.M. 2422 " 4 11.30 A.M. IIS 22. O 89 8h. 03m. II 617 13 2424 " 4 2.56 P.M. 119 22.0 8() 2h. 24m. ! 3 401 41 [layer at 4.51 P.M. 2425 ! " 4 5.00 " IK) 22.0 89 4h. 26111. 6i73 29 Agitated surface of sand 2428 " 5 10.45 A.M. 119 25.0 ioi 6h. 41 m. 7 630 20 [layer at i 26 P.M. 2429 " 5 2.43 P.M. 119 25. o : ioi loh. 3801. ! 13 409 20 Agitated surface of sand 2 43i: " 5 4-54 " "9 25. ; ioi I2h. 4/111. id i6n 20 Ag. surf. S. L. 4.53 P.M. 2434 6 10.40 A.M. 119 23.0 93 I5h. osm. 19960 15 Shut in!. 10.3,,, outl. 10.4^ A.M. 2 435 6 3-05 I .M. 20 25.0 ioi 4h. <x)in. 5884 14 Agitated surface of sand 2437 " 6 5.00 " 20 25.0 ioi 5h. 56m. 8714 5 layer at 2.36 P.M. 2440 " 7 10.40 A.M. 2o 24.5 99 8h. 0501. 2441 " 7 2.41 P.M. 21 24-5 99 ih. oom. I (121 44 2443 " 7 4-4" " 21 24.5 99 3h. osm. 4501 44 2446, 8 10.35 A.M. 21 24.5 99 5h. 3801. 7991 152 Agitated surface of sand 2 447 " 3 2.40 P.M. 21 25-0 IOI 9h. 33m. I407I 96 layer at 10.33 A.M. 2 449 8 5.15 21 25.0 IOI I2h. o6m. I794I 53 Agitated surface of sand 2453. " 2O IO.3O A.M. 22 25.0 101 .... ih. 07m. 1 676 4 layer at 3.58 P.M. 2454 " 20 11.55 " 22 25.0 IOI .... 2h. 3201. 3 73 8 2459, " 20 2.55 P.M. 22 25.0 101 5h. iSm. 7 696! 6 Agitated surface of sand 2460 20 5. If) " 22 25.0 101 .... 7h. 39m. ii 186 6 layer at 2.04 P.M. 2464 " 21 9 34 A.M. 22 25.0 IOI .... 8h. 27m. 12396! 6 2466 " 21 10.25 22 25.0 IOI .... 9 h. 18111. 1 3 filili 19 2469 " 21 12.41 P.M. 23 25.0 IOI .... 1 1 in. 247 24 2 473 " 21 1.50 " 23 25.0 IOI .... ih. 2om. 1967 670 2476 " 21 2.59 " 23 25.0 IOI .... 2h. 29m. 3667 17 2477 " 21 5.11 23 24.0 97 4h. 40111. 6857 21 Agitated surface of sand 2480 " 22 9.54 A.M. 23 25.0 IOI MI. 53m. 8657 24 layer at 4.39 P.M. 2482 " 22 10.48 " 23 25.0 101 6h. 47m. 9 987 19 2485 " 22 12.39 I .M. 23 25.0 IOI 8h. 38m. 12 727 II 2487 " 22 1.24 " 23 25.0 IOI 9(1. 2301.! 13 807 8 249O " 22 3.00 " 23 24.0 97 loh. 59m. 16 187 17 Agitated surface of sand 2 493 " 23 9.42 A.M. 23 24.0 97 13(1. oom. 20857 21 layer at 3.58 P.M. 2494 " 23 IO.23 " 23 24.0 97 14!). 50111. 21 887 20 Agitated surface of sand 2498 " 23 12.53 I - 1 . 24 23.0 IOI 28m. 675 32 layer at 10.46 A.M. 2500 " 23 2.03 " 24 24.0 97 ih. 38m. I 2 405 14 i6o WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Jewell System. Collected. Ra Fill teof b Period of c & to .0 3 z in Date. HOU , Number Run. Q. >f - 3 15^ 3 *^ U Jo. 5s 3 ^X |JU I J Last Washing. Hours and Minutes. 91 rS>3u g 0$ CO Remarks. 2503 2505 2509 2511 2514 2517 2519 2525 2527 1896 Apr. 23 " 23 " 24 24 " 24 24 " 24 " 25 " 25 3.06 P.M. 4.50 " 9.37 A.M. II 46 " i . 14 P.M. [2.49 " 4-44 " 12.45 " 2.57 " 124 124 124 124 124 124 124 125 125 25.0 24.0 25.0 25.0 25.0 25.O 25.0 25.0 25.0 101 97 IOI 101 IOI IOI IOI lor IOI 2h. 4im. 4h. 25m. 5 h. 4 2m. 7h. 49111. gh. I7m. loh. 42m. I2h. 47m. ih. O7m. 3h. igm. 3985 6515 8 865 11585 3 795 16235 9 55 i 773 5047 24 28 67 41 32 37 52 56 4 Agitated surface of sand layer at 10.06 A.M. [layer at 7.46 P.M. " 27-28 g.OO " " 3.OOA.M. " 28 25 6 " 28 " 28 2551 " 28-29 9.OO " " 3.OO A.M. 128 24.9 IOO Agitated surface of sand " 29 3.OOA.M. " g.OJ " 128 25.6 2561 " 29 9 . oo " " 3 . oo P. M . 128-129 25.3 IO2 [at 4.03 P.M. and 8.34 P.M. 2565 2566 2567 2568 2569 257 2571 2372 2573 2574 2575 2576 2577 2579 2580 2581 2582 2 5 S3a 2 5 83b 2584 2585 2586 2589 10.39 P M - 11.05 " 11.07 " 1 1 . 09 " II . 1 1 " 11.13 " 11.15 " 11.17 " II. 19 " I I . 21 " II .23 " 11.25 " 11.27 " 11.29 " 11.31 " 11.33 " 11.38 " 11.48 " I2.O3 A - M - 1.03 " 2.O3 " g.OO I .M. to 3.OO A.M. 129 129 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 130 24.6 22.0 28.0 26.0 J7.0 26.0 26.O 26.0 26.0 26.0 26.0 26.5 26.5 26.5 27.0 27.0 27.0 27-5 27.0 26.5 26.O 26.0 25.6 99 89 114 105 log 105 105 105 105 05 105 107 107 107 109 log 109 III log 107 105 105 I2h. 3001. O2m. O4tn. 06m. o8m. lorn. I2tn. 14111. 1 6m. i8m. 2O111. 22111. 24m. 26m. 28m. 30 in. 35"i. 45m. ih. oom. 2h. oom. 3h. oom. 18788 56 106 156 216 266 316 366 426 476 526 576 636 686 736 7g6 g26 I 1 86 i 5 g6 3 116 4676 102 139 86 86 70 49 58 38 36 39 24 44 33 "3 26 30 38 29 26 26 24 74 The series of results on run No. 130 was used in obtaining the aver age bacteria for the run but not for this day. 2590 2591 2592 2593 2594 2595 " 30 " 30 " 3 " 30 " 30 " 30 3.03 A.M. 4.03 " 5-03 " 6.03 " 7-03 " 8.03 " 130 130 130 130 130 130 26.0 26.0 26.0 26.0 26.0 26.0 26 o 105 105 105 105 i5 105 4h. oom. 5h. oom. 5h. 58111. 6h. sSm. 7h. 58111. Sh. 58111. 6 206 7 786 g 226 10 816 12 346 13956 52 23 27 53 8g 34 Agitated surface of sand layer at 4.44 A.M. 26OO " 30 12.57 P.M. 130 22.5 91 I3h. 5201. 21 406 23 [layer at 8.46 P.M. 2608 " 30 25.6 46 Agitated surface of sand 264O 65 Agitated surf, of sand layer 26 6 67 78 Ag surf of s I. at 1.14 A.M. 266O " 2 46 [layer at 3 07 P.M. 2666 78 28 i 2674 2675 2676 2677 2678 4 4 4 4 4 7.30 P.M. 17-32 " 7-34 " 7.36 " t7.3S " 34 134 "34 134 134 25.0 25.0 25.0 25.0 25.0 10 10 10 10 10 O2m. O(m. o6m. o8m. iom. 53 113 183 223 263 116 M9 132 61 65 From May 2-9, inclusive, the results of both single sam ples and ihose collected by the sampler were used to obtain the bacterial aver ages for days and for runs. COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 161 TABLE No. 4. Continued. Jewell System. Collected. R Fil ate of 1 5 u 01 i s z Date. Hour. Number of & ~ u c Ss U = $ C o = *I = 0.3 o a: I Last Washing Hours and Minutes. fl| ^JU ta v -. a g re ^ X Remarks. 267? 26Sc 2681 2682 2683 268.1 268? 2686 2687 2688 2689 2690 2691 1896 May 4 4 4 4 4 4 4 4 4 4 4 4 4 4 7.40 P.M. 7.42 " 7-44 " 7.46 " 7.48 " 750 " 7.52 7-54 " 7-56 " 7.58 " 8.03 " 8.13 " 8.28 " 3.15 P.M. to g.OO P M. 134 134 34 34 34 34 "34 134 134 134 134 134 34 133-134 26. c 26. c 26. c 26. c 26. c 27.0 27. c 27.0 26.= 27. c 27.0 27.0 27.0 105 105 i5 105 105 log log 109 107 109 109 109 109 I2m 14 m i6m 1 8m 2om 22111 24111 2III11 28m. 3om. 35m. 45m. ih. oom. 3". 3f>3 42? 473 S- . 583 <>33 68- 743 793 923 I 193 593 33 43 44 32 29 33 28 30 42 42 26 39 29 46 2695 2696 2697 2698 2699 2700 " 4 4 4 " 5 " 5 " 3-5 9.28 P.M. 1028 " 11.28 " 12.28 A.M. 1.28 " 2.28 " 34 134 134 134 134 134 27.0 27.0 27.0 27.0 27.0 27.0 26 5 log log log log log log 2h. oom. 3h. oom. 4)1. oom. 5h. oom. 6h. oom. 7h. oom. 3203 4823 6453 8053 9663 II 233 37 33 4 ) 47 38 go 2704 2703 2706 2707 2708 2709 :: \ " 5 5 5 5 " 5 3.28 A.M. 4.28 " 5.28 " 6.28 " 7.OO " 8.00 " 134 134 134 34 134 134 27.0 27-0 26.5 26.5 26.5 27.0 26 o log 109 107 107 107 109 8h. oom. gh. oom. gh. 58m. loh. 58m. nh. 58m. I2h. 5801. I28g3 14473 159 2 3 I7&53 19163 20773 96 86 69 28 29 29 37 Agitated surface of sand layer at 5.12 A.M. 27:4 2715 2716 2717 2719 2723 " 5 5 5 1 5 5 " 5 g.2S A.M. 10.28 " 11.28 " 12.28 P.M. 1.28 " g.OO A.M. to 3.OO P.M. 134 134 134 134 134 134-135 6.5 26.5 6.0 6.5 4-5 6 i 107 107 I5 107 99 I3h. s8m. I4h. 58m. I5h. 5601. i6h. 56m. I7h. 54m. 22 483 24 143 25633 27 53 28543 24 58 37 56 46 40 Agitated surface of sand layer at 11.19 A.M. Agitated surface of sand layer at 1.07 P.M. " S 2731 " 5- 6 g.OO " " 3.00 A M. 3 8 96 56 layer at 4.11 P.\ . " d 5 8 2736 " 6 " 6 9.35 A.M. 136 136 6.5 6 4 107 ih. 4om. 2763 25 layer at 6.35 A.H . Agitated surface >f sand 274.) 6 136 6 4 2746 2749 " 6 " 6 3 oo P.M. 3-OO P.M. IO g.OO P.M. 136 136 7-0 109 7 n. 03111. II 213 16 " 6-7 136 137 85 2753 2759 " 7 " 7 3.00 A.M. 3 OO A.M. to 9 OO A.M. 137 137 -:" - , IO<) igm. 5 9 So layer at 10.21 P.M. and I.C8 A.M. 2760 2765 " 7 g.OO A.M. g.OO A.M. 10 3.00 [ .M. 137 27.0 26 8 109 6h. igm. 10 iSg 32 38 [layer at 12.34 P.M. Agitated surface of sand 3.00 P.M. " 9 (xj " 46 277-1 " 7- 8 g.OO I .M. 138 138 26.5 107 o8m. ig2 57 layer at 6.o3 and 8.25 2780 2705 " 8 " 8 3.00 A.M. 3.OO A.M. to 9 OO A.M. 138 138 -:." log 6h. o8m. 9852 26 32 Agitated surface of sand 2786 2792 " 8 " 8 9.OO A.M. 9.OO A.M. tO 3.OO P.M. 138 138 27.0 ,- , log nh. 4101. 18732 8 18 layer at S. 19 A.M. 2793 2-<r " 8 " 8 3.00 P.M. 3.00 P.M. to 9.00 P.M. 138 i (8 i i ( i 25.0 1OI I7h. 41 m. 26 732 19 Agitated surface of sand layer at 3.04 P.M. 2803 " -g 2804 2808 " f9 3.00 A.M. 139 27.0 log gh. 5601. 15907 12 layer at 1.45 A.M. 2813 2818 2825 2830 9 9 " 1 1 " II g.OO A.M. 3.0O " 3.00 " 9.00 " i 3 g 140 MI 141 26.5 28.0 27.0 27.0 107 "4 log 109 I5h. 56111. 4h. l6m. 5h. 42m. nh. 4om. 25(>97 7 212 8955 18 2go 21 39 28 22 Agitated surface of sand layer at 9.31 A.M. [layer at 4.30 P.M. Agitated surface of san^ 162 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Jewell System. Rate of j % Collected. Filtration. JJ c _S b. Period of ^ "S 3 ^. A z Date. Hour. Number L 3 O C i a ^ t 5 * = I Service Since Last Washing. Hours and Minutes. ^^- Sll "su R1 iu Remarks. c |i Q. "* 1 -2 CQ </> u i _] fc. 1896 2857 May 12 3.OO A.M. 142 7.0 109 ih. 33m. 2 566 25 2867 " 12 9. oo " 142 7.0 109 7h. 33m. 12 196 54 2873 " 12 12.00 M. 142 7.0 109 loh. 3im. 1 6 796 22 Agitated surface of sand 2877 " 12 8.30 r.M. 143 3-0 93 4h. lorn. 6 711 19 layer at 10.26 A.M. 2881 " 13 2.OO. A*I. 143 7.0 .109. - . . , gh. jSm. 1 5 30 1- ^2 Agitated surface of sand 2885 13 - *?. od " 143,- 6. f. I0 > .> . I5h. 3$m. 25031, 12 layer at 1.38 A.M. 2891 13 l.OO Ml. M4> 7.0 H?9 } - 3h. i.2m. ?4S3P H 2896 13 7 .0<i " 144.;. 7-P id<) gh. 2im. 1491^ 13 Agitated surface of sand 2900 14 . . .. 3.00, A.M. M5 7.0 109 4801. 1258 JO layer at 5.15 r.M. 2905 14 - 9.06 V. M5- 7.0 109 6h, 48m. io Si8 4 2909 14 2.08 P.M. M5 6.0 105 iih. 54m. 19118 37 Agitated surface of sand 291-1 M 8.00 " M5 7.0 109 I7h. 46m. 28 748 So layer at 2.02 I .M. 2919 15 l.OO A.M. 146 7.0 109 j 4h. O2m. 6430 16 2923 15 S.oo " I4f> 7.0 109 i ih. O2m. 17640 16 2927 15 I l.OO " 146 7.0 109 I4h. oom. 23 340 52 Agitated surface of sand 2932 15 5.15 P.M. "47 6-5 107 3h. 14. 5 1 20 M layer at 9.09 A.M. 2961 5 I 1 . OO " 47 7.0 109 8h. 5 gm. 14 280 5 2970 " 16 5.00 A.M. 147 6-5 107 I4h. 57rn. 24 O2O 28 Agitated surface of sand 2981 " 16 IO.OO " 148 5.<> IOI 5 S,n. I 700 19 layer at 4.35 A.M. 2991 " 16 3.OO P.M. 148 6.0 105 gh. 58111. 9 610 15 2999 " 18 I.I7 " 49 25.0 ioi 1.4 05111. 215 1 08 3000 " 18 1.27 "49 25.0 ioi 1-5 i 5111. 435 4" 3002 " 18 3.00 " 149 24.8 100 I .( ih. 48111. 2805 91 3008 " 18 6.05 " 49 24-5 99 3-< 4h. 53111. 7 355 192 3010 " 18 9.00 " 149 25.0 IOI 3-3 7h. 4801. ii 695 3015 " 18 12. OO " 149 25.0 IOI 4.0 loh. 48111. 16145 34 3018 " 19 3.OO A.M. 149 25.0 IOI 5-4 I3h. 48m. 20655 57 3024 19 6.OO 149 25.0 101 6.0 i6h. 48m. , 25 155 5i 3027 19 8.30 " 49 25.0 IOI 7.0 igh. I3m. | 28 845 26 3032 19 12.OO M. M9 24-5 99 8.8 22h. 48111. 34 085 65 [layer at 2.22 P.M. 3036 19 3.OO I .M. 149 25.0 IOI 6.0 25h. 45m. 38 355 43 Agitated surface of sand 3041 19 6.00 " 149 25.0 IOI 8.2 28h. 45m. 142 885 37 [layer at 10.38 P.M. 3044 19 g.oo " 149 25-5 103 9.2 2ih. 45111. 47345 3 5 Agitated surface of sand 3050 " 20 l.OO A.M. 150 26.0 105 05111. 112 !9 2 D. Application of chemi- 3051 " 2O I. 10 " 150 25.0 101 1-3 15111. 372 99 D. cals unsatisfactory on 3053 " 20 3.00 " 150 25.0 ioi 1.9 2h. 05111. 3082 79 D. run No. 150; chem- 3057 " 2O 6.00 " 5 25. oj ioi 1.7 3im. 77 (>5 ical feed-pipe broken. 3060 " 2O 8.30 " 15" 25.0 IOI 2. I 3)1. oim. 4480 32 3069 " 2O 12.00 M. 151 25.0 IOI 3.3 6h. 31111. 973 57 3072 " 20 T.OO I .M. 151 25.0 101 5-4 gli. 3 1 in. 14 150 30 3077 " 2O O.oo " 51 25.0 H)l O.I I2h. 3im. 18620 56 3082 " 2O 9. oo " 5i 25.0 IOI 8.3 15!!. 31111. 23050 41 [layer at II. II I .M. 3o8( " 20 1 2 . OO " 151 25.0 IOI 5-9 i8h. 2gm. 27 260 48 Agitated surface of sand 308 c " 2 3.OO A.M. 151 24.0 97 9-3 2ih. 2gm. 31 760 62 Agitated surf.S.L.,4.3iA.M. 3093 " 2 6.00 " 151 23-5 95 g.( 24!!. 27111. 35940 73 Agitated surface of sand 395 " 2 7.54 " 152 25.5 103 2.3 05111. 109 231 layer at 6. 16 A.M. 3096 " 2 8.04 " 152 25.5 103 1.6 15111. 379 118 309 " 2 8.30 " 152 25.5 103 i.t 4im. 939 60 310 " 21 I2.OO M. 152 25. c IOI 2.7 4)1. 1 1 in. 6 28c 65 310 " 21 3.00 r.M. 152 24.5 99 7h. i im 10889 69 3" " 21 6.00 " 152 25.0 IOI 4-: i oh. inn I53I9 69 31 " 21 9.00 " 152 25. c IOI 6.1 I3h. iim. ig 759 47 3 " 21 12. OO " 152 25.0 IOI 8.1 i6h. urn. 24229 73 312 312 " 22 " 22 3.00 " 6. oo " 152 152 23.5 25.1 T95 101 g.f g.c igh. nm. 28 639 22h. ogm. 32 981, 72 gS [layer at 3.21 A.M. Agitated surface of sand 313 " 22 8.30 " 152 24. c 97 g.f 24h. 37m. 36 55C 88 Agitated surface of sand 3 3 ! " 22 10.24 " 53 25. c IOI i.f 05 m I3C 87 layer at 7.2ijA.M. 313 " 22 11.34 " 153 24. c 97 I. 1501. 37^ 99 3 3 7 " 22 12 00 M. 53 25. c IOI 2.( > ih. 4im. 2 581 66 3M 2 " 22 3.00 r.M. 153 24 ! 99 3-< ) 4h. 4im. 7031. 39 3M 315 IIS i " 22 " 22 . 22 6. oo " 9.00 " I2.OO " 153 53 1^3 25. 25. c 25. < 103 IOI 101 4- 6. 4- 7h. 41111. II 56t loh. 4im. 15 97C > 13(1. 39m. 1 20 341 61 41 98 [layer at 11.54 I .M. Agitated surface of sand COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 6.5 TABLE No. 4. Continued. Jewell System. Rate of C Collected. Filtration. B .S J ~t Ul u Period of u <* 3 Number c. o a "S ServiceSince Last Sf s ^fe S 3 Date. Hour. Run. !l |l| = Washing. Hours and Minutes. - * A V G Kc-m;.rks. .2 is i ."? si .JCJ 5^U X u S J m 1896 3164 May 23 g.iS A.M. 154 55- 222 6.0 o8m. 4^7 35 D. Run No. 154 was a 3165 " 23 9.28 " 54 55.0 222 6.0 iSrn. 957 23 F>. special run at the 3166: 23 9-3S " 54 55- 222 6.0 28111. I 447 26 D. request of Filter 3i&7 ! 23 10.08 " 54 2 4 6.0 5Sm. 2977 31 I). Company. 3168 23 10.38 " 54 50.0 2O 2 6.1 ih. 28m. 6617 27 D. 3175! 25 12.00 M. 55 29.5 120 2 . 7 ih. 4gm. 3322 74 3178: 25 2.OO P.M. 55 29.0 118 4.0 3h. 49111. 6 832 43 3182 25 6.OO " 155 3- 122 5- 7(1. 4gm. 14 132! 28 3185; 25 8.00 " 155 30.0 122 6.5 gh. 49 ^- 17932 21 3189 25 12. OO " 155 30.0 122 5-1 I3h. 47m 24842 34 Ag tated surface of sand 3192 26 2.00 A.M. 155 30.0 122 6.5 I5h. 47m. 28362 34 aver at 1 1.09 p M. 3 i 98 26 6.0O " 155 29.0 IlS 4.4 igh. 45m. 35 172 35 Agitated surface of sand 3200 26 7-39 " 156 29.0 US 1.8 05m. 133 "4 layer at 5.06 A.M. 3201 26 7-49 " 156 30.0 122 I. g I5 . 413 76 3203 " 26 8.30 " 156 30.0 122 2.1 56m. i 643 134 3209 26 IO.OO " 156 30.0 122 2.S 2h 26m. 4313 25 3213 " 2f> 2.00 P.M. 156 130.0 122 5.0 6h. 26m. 1 1 533 34 3216 26 4.00 " 156 ; 3 o.o 122 6.5 Sh. 26111. 15703 46 3222 " 26 S.co " 156 30.0 122 5-5 I2h. 24tn. 22 543 . 5 Agitated surface of sand 322 5 ! " 26 10.00 " 156 j3o.o 122 I4h. 24111. 26 143 52 layer at 6.20 P.M. 3229 27 2.OO A.M. 156 30.0 122 7-5 iSh. 22111. 32 883 46 Agitated surface of sand 3233; 27 5.1 " 57 25.0 I 01 1.8 i im. 130 81 layer at 11.58 P.M. 3234| 27 5.20 " 157 30.0 122 2. 1 2im. 49 73 3237 27 7.3 " 157 t3- 122 3-1 2h. 3im. 4480 46 324I 1 27 12.00 M. 157 3- 122 6.0 7h. oim. I 2 670 135 flayer at 2.14 P.M. 3246 27 3.00 P.M. 57 29.0 118 5-1 gh. 5gm. 1771 30 Agitated surface of sand 3256; " 27 6.OO " 57 30.0 122 6.9 I2h. 59111. 22 980 34 Agitated surface of sand 326i: " 27 9 20 " .58 30.0 122 2.0 05111. 182 206 layer at 7.13 P.M. 3262 " 27 9.3 " 158 30.0 122 2.0 15111. 422 152 3265 " 27 12 00 " 158 30.0 122 3-3 2h. 45111. 4982 57 3267 " 28 3.OO A.M. 158 J3O.O 122 7. 5h. 45 "- 10 342 51 Agitated surface of sand 3273 " 28 6.00 " 158 30.0 122 5-9 Sh. 43m. 15 702 47 layer at 3.11 A.M. 3276 28 7.30 " 158 30.0 122 7-3 loh. 13111. 18362 43 Ag tated surface of sand 3280 " 28 10.00 158 29.5 I 2O 7.0 I2h. 41111. 22 4O2 54 layer at 7.51 A.M. 3283 " 28 11.00 " 159 29.5 120 1.9 05111. 157 246 3284 " 28 1 1 . 20 " 159 29.0 I 2O 2.0 lem. 447 158 3295 " 28 2.OO P.M. 159 30.0 122 2.g 2h. 55111. 5 177 162 3298 28 4.OO " 159 30.0 122 4-2 4h. 55tn. 8737 300 3306 " 28 S . oo " 160 30.0 122 2.7 2h. 30111. 4497 423 I). 3310 28 9-35 " I()i 3. 122 2.O 05111. 193 668 1). 3311 " 28 9-45 1 61 130.0 122 2.O 15111. 443 560 1). 3312 28 9.58 ;; 161 30 . o 122 2 .O 28m. 833 650 t). Prescribed amount 3316 " 28 162 po.o 122 1. 9 05111. 157 460 I). <- of chemicals in- i i-jn " 28 11.25 " 162 30.0 122 2.0 13111. 337 345 1). sufficient. 3326 " 29 12.5(1 A.M. 163 30.0 122 2.0 05111. 165 715 i). 333" 29 1. 06 " 163 30.0 122 2.1 15111. 435 394 I). 3333 29 2.00 " 163 30.0 122 2.3 ih. 09111. 2 065 510 D.J 334 29 4.00 " 163 30.0 i 2 4.8 3h. ogm. 5625 3f>3 3342 29 4.35 " 163 30.0 122 5.7 3h. 44m. 6635 2/5 3346 29 5-59 " 164 2S.O "4 2.1 05 m. 293 151 3348 2g 6.09 " 164 24.0 97 2.2 15111. 473 149 29 7.30 " 164 25.O tor 2.2 ih. 36111. 2523 So 33"! 29 12.05 P.M. ,65 29-5 120 2-5 45111. 875 212 3364 29 2.OO " 1 66 29.5 120 3- r ih. oom. 1888 I2g ,368 29 6.00 " 169 3O.O 122 2.0 28m. 954 465 3375 29 S.lg " 171 30.0 122 2.1 1 5 m . 493 118 3380 3 12.27 A.M. 174 2O. O Si I .1 15111. 344 128 3384 3 2.OO " 74 20. Si 1-5 ih. 48111. 2184 222 3388 3 6.26 " 175 29-5 120 2.1 2bm. 720 415 3394 3 8.00 " 176 29.5 1 20 2.1| I4IT). 402 275 34 34 3 " 30 10.15 " 12.00 M. 77 178 30.0 30.0 122 122 2.2J 3om. 2.1 28m. 93 863 475 169 Agitated surface of sand 3434 June 2 3.50 P.M. 79 34-0 3.3! 2.1111. I .1"! 1 2*.) layer at 3.27 P.M. 1 6 4 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Jewell System. Rate of ti Collected. Filtration. t/5 j5 Ji Number s. jg. I enocl ol O/J 3 .; <J . X! E ^ Run. t. ~ lf| s l.ast Washing Hours and *| S Remarks Date. Hour. 3 ^1 Minutes. y.y. 2 a S .r. -- = g. J 1 -25 ^$ in u j S t. ; I S()6 I 3436 June 2 4.37 P.M. 179 34-5 MO 4.0 ih. 1 6m. 2 701 62O 3439 2 6.20 " 181 20. 5 83 2.1 IJIll. 410 191 3442 " 2 10.37 " 182 27.0 109 5.1 2li. 25m. 3 989 22 3447 3 3.30 A 11. 183 25.0 IOI 3.1 2h. 56111. 4 448 36 3451 3 6 oo " 184 25.0 TOI 2-7 3im. 745 44 345-1 3 9-^0 " 184 25.0 IOI . . . . 3h. 31111. 5 345 26 3458 3 T2.OO M. 184 25. o lot 6.7 6h. 26111. 434 25 3460 3 2.00 P.M. 184 20.0 8 1 9.5 Sh. 13111. 2 145 50 Agitated surface of sand 3464 3 4.30 " 185 24-5 99 2. 5 2)1. lorn. 3 356 70 layer at I 06 P.M. 3468 3 6.00 " 185 25.0 101 3.4 3)1. 40111. 5 526 144 3472 3 9 oo " 1 86 25.0 IOI 2.0 ill. 21 ID. 2 06 1 1 6 3478 3 12.00 " 1 86 25.0 101 4 .6 4)1. 2 1 m. 6451 36 [layer at 1.41 A.M. 3482 4 3.00 A.M. 1 86 25.0 ioi 8.5 7(1. 1401. 10741 36 Agitated surface of sand 34S4 4 3.30 " 186 21.0 85 7h. 44m. 11451 64 JShut inlet 3.27 P.M., out- 3487 4 6.00 " 187 25.0 IOI 2. I 2h. 07m. 3 198 43 let 3.38 P.M. 3 49- 4 9 oo 187 5h. 07111 . 7 ^88 JT 3496 4 10.37 " 187 25-5 103 7.0 fth. 44 "- 9 948 54 3498 4 12.02 P.M. 187 23.0 93 9.7 8b. (xjm. 11978: 38 3501 4 3-43 " .89 25.0 IOI 1-7 2im 851 24 3506 4 6.05 " 189 25.0 IOI 2.4 2h. 53m. 4611 1 1 6 3509 4 8.40 " 190 25.0 IOI 2.1 2ll. 12111. 3 422 61 3511 4 9-55 " 190 25.0 IOI 3.2 3)1. 27111. 5442 71 Shut inlet 9.53 P.M., out 3512 4 10.22 " 191 25.0 IOI 7 02111. 50 78 let 10.01 P.M. 3513 4 10.24 " 191 25.0 IOI .7 O4in. loo 44 3514 4 IO.26 " 191 27.5 III .7 o6m. 155 35 3515 4 10.28 " 191 27.5 1 11 7 08111. 210 33 3516 4 10.30 " 191 25.0 IOI 7 I om . 260 22 3517 4 10.32 " 191 25.0 IOI . 7 12111. 310 29 35i*> 4 10.34 " 191 27.0 C9 7 14111. 365 33 3519 4 10.36 " 191 25.0 01 7 16111. 415 23 3520 4 10.38 " 191 27.0 09 7 iSm. 470 27 3521 4 10.40 " 191 2J.O 01 .7 20111. 520 43 3522 4 10.42 " 191 27.0 09 .7 22II1. 575 10 3523 4 10.44 " 191 27.0 09 7 24111. 630 ; 15 3524 4 10.46 " 191 25.0 01 26m. 680 87 3525 4 10.48 " 191 25.0 ol .7 28m. 73; 99 3526 4 10.50 " 191 25.0 01 .7 30111. 780 II 3527 4 I0.;2 " 191 25.0 o; .7 32111. 830 114 35281 4 10-55 191 25. O; Ol .7 35111. 910 17 352 ) 4 1.05 191 25.0 oi 7 45111. I 160 16 3530 4 1. 20 " 191 25.0 ioi .i> ih. oom. I 540 25 353 4 1.50 " 191 23.0 ioi 2.O ih. 3i. 2350 52 353? 5 2. 2O A.M. 191 25.0 ioi 2.1 2)1. oom. 3 100 27 353 h 5 2.50 191 25.0 ioi 2. I 2ll. 30111. 3 870 46 3537 " 5 1 . 2O " 191 25.0! ioi 2.2 3h. oom. 4 620 33 3539 5 3-22 " 192 30.0 122 2-3 ih. O2m I 848 18 3543 3556 I I 6. co " 4-55 I .M. 192 197 29-5 35-0 1 20 142 3-9 2.5 3h. 40111. 35m. 6 528 I 134 94 Si 3559 5 IO.OO " 200 30.0 132 2.1 5"m. I 5 9 7 35f 6 2.30 A M. 2OI 25.0 101 2. I 2ll. I 2111. 3386 9 359 6 6 7.48 " 2O3 25.0 IOI 1.7 53"i- I 312 14 3624 6 10.55 " 203 25.0 101 2.f 4h oom. 5992 29 3629 " 6 1.55 I -M. 204 34-5 140 2-9 ih. 03111. 2016 14 3f>32 " 6 3-00 " 204 33-5 136 3-1 2h. oSm. 4116 12 3657 " 9 I2.5O " 2O5 25.C IOI 2. I ill. 54111. 2854 170 3661. 9 5.0O " 2O6 25. C IOI I .9 ih. 22m. 2074 39 3669 " o 11.15 A M. 2O7 25. t IOI I.( 56111. I 415 II . Agitated suiface of sand 3672 " I. CO P.M. 207 25. c IOI 2.C 2h. 4im 4045 9 layer at 10.06 A.M. 3&7 " o 330 " 2O7 25. L IOI 5.2 4)1. 1 1 in 6925 13 3682 " I IO.32 A.M. 207 25.( IOI 5-7 8h. 41111. 12945 14 3685 " I 1. 00 I 1 M. 208 25 .c IOI I .7 25111 644 7 3693 " II 3.42 " 208 25. c IOI 2.f 3h. O7m. 4 804 ib 3699 12 II. II A M. 209 25.0 IOI I .7 3im. 704 9 COMPOSITION OF Off TO RIVER WATER AFTER PURIFICATION. TABLE No. 4. Continued. Jewell System. 165 Rate of J Collected. Filtration. fc 1 Number S. f s. -d Period of Service Since Last be U j e Remarks. 3 Run. v ri O 5 5 S Washing. >h ^"c 7. Date. Hour. ol JjE Minutes. al !i i U ~ a cT A J ! u 1896 3703 June 12 1.38 P.M. 209 25.0 IOI 2-3 2h. sSm. 4474 12 3706 " 12 2.48 " 209 25.O IOI 2.8 4h. ogtn. 6 184 12 37 3 " 3 II. O2 A.M. 210 25.0 IOI 1-5 22m. 485 23 37 9j 3 I.OO P.M. 210 25.0 IOI 2.1 2h. 2Om. 3585 15 37251 3 2.55 " 210 25.0 IOI 2.7 4h. ism. 6465 51 3728 13 5.02 " 210 25.0 IOI 4-3 6n. I2m. 9335 7 17-37 " r 5 9 . (X) " 2 IO 6h. 4Om. IO I 28 1153 B. Collected before the J 1 J 1 3741 " 15 10.15 A.M. 211 25.0 IOI 1-7 5om. i 248 IO filter was in opera 3744 5 12.23 P.M. 211 25.0 IOI 2-7 2h. 5801. 4484 25 tion and after pe 3748 15 3.02 " 212 25.0 101 37m. i 005 19 riod of rest of 39 3754 15 4.31 " 212 25 o IOI 2. I 2h. o6m. 3355 II hours 30 minutes. 376o " 16 10.25 A.M. 213 38.0 154 3-7 3801. i 429 9 3764 " 16 12.38 P.M. 214 38.0 154 3-1 32111. i 27&I 5 3797 " 18 10.10 A.M. 216 25.0 IOI i .7 ih. 0701. i 7i8 ! 28 3802 3810 " 18 " 18 12.34 P.M. 2-49 " 2l6 216 25.0 25.0 IOI IOI 3- 5-0 3h. 3im. 5h. 46111. 5398 15 8678 8 3815 18 4-55 216 25.0 101 7-4 7h. 47m. 11678 28 [layer at 11.23 A M. 3819 " 19 10.00 A.M. 216 25.0 IOI 7-9 cjh. 22m. 14118 39 Agitated surface of sand 3825 " IO 12.56 P.M. 216 1 2h. 1 3m. 18346 lo Shut outlet 12. =;6 P.M. 3830 1 V 19 2.59 " 217 25.0 IOI 1.8 45m. I 072 69 3846 19 4.32 " 217 25.0 101 i.S 2h. i8m. 3422 "4 3860 " 20 11.33 A.M. 218 25.0 IOI 1.6 38m. 976 "7 3863 " 20 12.43 P.M. 218 25.5 103 7 ih. 48m. 2 7l6 69 3872 " 20 3-42 " 219 25.0 ioi 1.6 25m. 638 88 3875 " 20 4.38 " 219 25. o ioi 1.6 ih. 2im. 2038 IOI " 22 9-OO A.M. IO 69 B.Coll. before fikerwas in oper 3889 22 10.15 " - y 219 25.0 IOI ! 2.0 3h. 28m. 5258 74 ation and after rest of 39(1.3010. 3891 " 22 12.25 P.M. 5h. 3801. 8 538 300 A. Shut inlet 12.21 P.M., 3895 " 22 I. IS " 2 21 25.0 ioi 1.6 23m. 533 96 outlet 12.31 P.M. 3900 " 22 3.01 " 2 21 25. : 101 2. 1 2h. o6m. 3213 79 3904 " 22 5.cx) " 221 25.0 ioi 2.6 4h. O5m. 6243 97 3910 23 9.52 A.M. 22C 5h. 2501. 8275 Shut outlet 9.52 A.M. 3925 " 23 11.12 " 22 25 . o loi l.S 5Sm. 1438 216 3927 " 23 1.30 P.M. 22 25.0 IOI 2. I 3h. i6m. 49 8 393 " 23 3.20 " 22 25.0 ioi 1.7 5h. o6m. 7688 62 3934 " 23 4.42 " 22 6h. 2Sm. 9 759 77 Shut outlet 4.42 P.M. 3939 " 24 10. 17 A.M. 222 25.0 ioi 1.8 ih. 33m. 2389 840 3948 24 12-39 P-M. 222 3 1 1 . 55m. f, 946 470 Shut outlet 12.39 P.M. 395 " 24 34 " 223 25.0 ioi 1.6 38m. 793 385 3955 " 24 3-24 223 25.0 IOI 1.8 2h. 28m. 3 6 43 355 3965 " 24 4-5 223 25.9 IOI 2. I 3h. 55m. 5753 215 3979 " 25 9.50 A.M. 224 25.0 IOI 1.4 03 in. 93 45 3980 " 25 9-55 224 25.0 loi 1.4 08 m. 223 398 " 25 0.<X) " 224 26.0 105 1.4 I3m. 353 3982 " 25 0.05 224 25.0 ioi 1.4 i8m. 483 57 =,-; " 25 0. IO " 224 25 o ioi 1.5 2301. 613 2 5 3984 " 25 0.15 " 224 25.0! ioi 1.5 28m. 733 515 3985 " 25 0.20 " 224 25.0 ioi 1.6 33m. 863 i 750 19-7 " 25 0.25 " 224 25. oj ioi 1.6 38m. 983 362 vr- " 25 0.30 " 224 25.0 ioi 1.6 43m. 103 i 780 399 " 25 "35 224 25.0 ioi 1.6 48m. 233 215 3992 " 25 0.40 " 224 25.0 ioi 1.6 5301. 353 445 3993 " 25 0.45 22 4 25.0 ioi 1.6 58m. 473 95" 3994 " 25 0.50 " 224 26.0 105 1-7 ih. 0301. 613 249 3995 " 25 1.05 224 25.0! ioi ih. i8m. 983 420 399 6 " 25 1.20 " 22 4 25. o ! ioi 1.7 ih. 33m. 2373 99 3997 " 25 35 " 224 25.0 ioi 1.7 ih. 48m. 2753 395 3998 " 25 1.5" " 224 25.0 ioi 1.9 2h. 03m. 3163 900 3999 " 25 2.O5 P.M. 22 4 25.0 ioi 1.9 2h. iSm. 3523 365 4<J02 " 25 I.I5 " 22 5 25.0 ioi 1.7 46m. I 166 510 4<x>5 " 25 231 " 225 h om 3 060 250 Shut outlet 2.31 P.M. 4 cxx; " 25 " 25 3.16 " 4.25 226 226 25.0 ioi 1.4 1 3m. ih. 22m. 236 2091 600 173 Shut outlet 4.25 P.M. 1 66 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Jewell System. Rate of S 8 Collected. Filtration. 2 t Number % Is. 6 Period of Service Since S.S j 3 U ^ g Run. t . * I Last Washing. 1; Remarks. X Date. Hour b ~ < _ Hours and a u -! o c o ,_ -C Minutes. S M ja u c c 15 = S J | ^23 <J(J c/i o K ~ " B 1896 4013 June 25 5.00 P.M. 227 25.0 IOI 1-5 iSm. 465 347 4024 " 26 10.27 A.M. 227 25.0 101 1.8 2h. I4m. 3235 23 4O2S " 26 11.40 72*7 3h 27m. 5 204 6 1 Shut inlet 11.37 A.M. 4031 " 26 I.I4 P.M. 228 25.0 IOI 1.9 ih. i6m. I g2i 19 outlet 11.47 A.M. 4034 " 26 2-35 " 228 2h. 37m. 4 026 79 Outlet closed for wash. 4035 " 26 3-3" " 229 25.0 IOI 2.0 4om. I 005 9 4037 " 26 4.50 " 229 25.0 IOI 1.8 2h. oom. 2g45 29 [wasting at end of run. 4U42 " 27 0.2O A.M. Waste. Collected after 4044 " 27 IO.26 " 230 25.0 IOI 1.7 32m. 839 44 4048 1 27 12 IO P M 2TO 2h. i6m . 3 4 2 5 Shut inlet 12.06 P.M. 4052 " 27 1.58 " 231 25.0 IOI 1. 9 ih. 2gm. 2 194 315 outlet 12. 16 P.M. 4<>55 " 27 3.20 " 232 25.0 IOI ! 5 I5m. 441 39 4056 27 4.46 " 232 25.0 IOI i. 5 ih.- 4im. 2 551 41 4062 " 29 10. 16 A.M. 233 25.0 IOI 1-7 36m. 937 4 4064 " 29 11.49 " 233 25.0 IOI 2. I 2h. ogm. 3577 6 4068 " 29 1.28 P.M. 233 25.0 IOI 2.2 3h. 48m. 5787 9 4070 " 29 3.38 " 233 25.0 IOI 6.6 5h. 58m. 9 067 II 473 " 29 5.13 " 233 25.0 IOI 9.6 7h. 33m. II 267 9 Agitated surface of sand 4082 " 30 10. 14 A.M. 234 25.0 IOI 1.8 23m. 630 7 layer at 9.00 A.M. 4100 " 30 12.45 P.M. 234 23.5 95 2.6 2h. 54m. 4 270 5 4105 " 30 2.52 " 234 23.5 95 6.1 5h. oim. 7 290 3 4108 " 30 4.31 234 23.5 95 4.1 6h. 3im. 9338 4 [layer at 11.20 A.M. 4114 July IO.25 A.M. 234 23-5 95 o.o Sh. 53m. 12 730 * Agitated surface of sand 41 16 11.25 " 234 21.0 85 1-4 gh. som. 13942 Wasting i min., locu.ft. 4117 11.28 " 234 21 .O 85 1.4 gh. som. 13 942 4 " 100 " 4118 11.31 " 234 21. O 85 1.4 gh. som. 13942 7 " I5<J " 4H9 "-34 234 21.0 85 1.4 gh. Som. 13942 Opening outlet. 4120 it. 37 " 234 22.0 89 1-4 gh. 53111. 14010 4121 1 1 . 40 " 234 23-5 95 1-5 gh. 5501. 14060 4123 " I.I7 P.M. 234 23.0 93 10.7 ih. 32m. 16080 [layer at 2.11 P.M. 4125 2.08 " 234 15.0 61 II. 2 2h. 23m. 16890 . Agitated surface of sand 4126 2.14 ". 234 15-0 6 1 8.7 2h. 27m. 16 937 Starting to waste. 4127 2.16 " 234 16.0 65 g.o 2h. 27m. 16937 Wasting 2 min.. 45 cu.ft. 4128 2.18 " 234 18.0 73 9-5 2h. 2701. 16 937 4 " 95 " 4129 2.22 " 234 18.0 73 9-5 2h. 27m. 16 937 8 " 155 " 4130 2.27 " 234 17.0 69 g.6 2h. 27111. i f > 937 Opening outlet. 4135 4-34 235 23-5 95 I . 2 39 m. 951 4144 10.29 A.M. 235 23-5 95 2. I 3h. oim. 4 231 49 4148 11.30 " 235 23-5 95 4h. O2m. 5 701 3 I 2.OO M ih i->m 6406 1 1 c 4152 12.35 P.M. 235 236 23-5 95 I. 1 4". j- lll< urn. 325 M4 4153 1.04 " 236 23-5 95 I .2 40111. 965 in 4157 3.05 237 23-5 95 1.6 5om. 1 155 9 4165 " 3 10.14 A.M. 237 23-5 95 6.2 4h. 2gm. 6 2 g 5 4170 " 3 12.15 P.M. 3om. 8 065 2 Shut outlet 12.15 P.M. 4173 3 12.50 " 237 238 23-5 95 I. 2 V J 114 6 4174 3 12.55 " 238 23.0 93 I. 2 lorn. 244 10 4175 3 I.OO " 238 20. o 81 1.3 I5m. 344 I 4176 3 1.05 " 238 22.0 89 t-4 2om. 454 8 4 77 3 I.IO " 238 24.0 97 1.4 25111. 594 15 4179 3 1. 15 " 238 25.0 IOI 1-4 3om. 724 IO 4180 3 1.20 " 238 23.0 93 1-5 35m. 834 13 4181 3 1.25 " 238 24.0 97 1.6 4om. 954 : 8 4182 4183 3 3 I.3O " 1-35 " - 238 238 23.5 24.0 95 1.6 45m. 5om. I 064 7 I 194! 13 4184 3 1.40 " 238 23-5 95 1.8 55m. I 214 6 4185 " 3 1.45 238 23.5 95 1.8 ih. oom. I 344 1 1 4188 3 2.00 " 238 23.0 93 1.8 ih. ism. I 794 II 4189! 3 2.15 " 238 23-5 95 1.8 ih. 3om. 2 164 125 4 9 3 2.30 " 238 2 5 . 5 95 1.8 ih. 45m. 2 504 8 4191 3 2.45 2 3 8 23.5 95 1. 9 2h. oom. 2864 6 4192 3 3.00 " 238 23.5 95 2.0 2h. I5m. 3 204 24 4193 3 3-1.5 " 238 23.5 95 2.1 2h. 3om. 3 574 M * The bacterial results of July I were lost through melting of the culture medium. COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. TABLE No. 4. Continued. Jewell System. ,67 Rate of s {j Collected. Filtration. [i. y> .H . ~^ ^ Period of . " 3 1 Number s. o a g Last ^"1 ^ ^ Z E Run. OJ U 3 s " X Washing. :,"( O.U Remarks. . Date. Hour. oi lu = Hours and Minutes. 1*2 || c Is = 8. j 1 J!J cj X u s ,4 m 1896 4194 July 3 3.30 P.M. 238 23.5 95 2.2 2h. 45m. 3924 16 4229 6 3.27 " 239 23.5 95 37m. 960 281 42333 6 4.25 " 239 24.0 97 6.8J ih. 3501. 2 380 223 4234 " 6 5.00 " 239 25.0 IOI 2.3 2h. oSm. 3 130: 357 Agitated surface of sand 42343 6 5.03 " 239 25.0 IOI 2.5 2h. i im. 3 210 362 layer at 4.53 P.M. 4235 6 5.06 " 239 25.0 IOI 2.5 2h. 14111. 3 280 372 4236 6 5.09 " 239 25.0 IOI 2.7 2h. I7m. 3 340 399 4237 " 6 5.12 239 25.0 IOI 2.7j 2ll. 2Onl. 3420! 393 4238 6 5-15 " 239 25.0 IOI 2.8 2h. 23m. 3 5o 435 4246 7 IO.OO A.M. 239 25.0 IOI 5.0 3h. 38m. 5 680 997 [layer at 12.24 P.M. 4253 7 1. 00 P.M. 239 24-5 99 7.6 7h. 3&m. IO02O 341 Agitsted surface of sand 4257 7 2.43 " 239 25.0 IOI 2.0 Sh. 17111. 12 870 336 Agitated surface of sand 4258 " 7 4-59 " 240 23-5 95 3.3 ih. 34m. 2 364 247 layer at 2. 12 P.M. 4275 " 8 5.10 " 241 23.5 95 2.9 54 n. I 360 99 4280 9 10.25 A.M. 242 24.0 97 2.1 22m. 522 Ij 4283 9 12. 2O I .M. 242 23.0 93 4-7 2h. I7m. 3 192 o 4302 9 3.25 " 242 25.0 IOI 6.3 sh. 2om. 7 562 14 Agitated surface of sand layer 4315 4318 IO " 10 11.15 A.M. I.Og P.M. 243 243 25.0 21.5 IOI 87 4.9 2h. um 9.6 4h. O5m. 4 292 38 5 982 50 at 2.15 C.M. and 4.35 P.M. Agitated surface of sand 4321 " IO 3-M " =44 25.0 101 2-3 5Om. i 281 29 layer at 1.27 P.M. 4325 " 10 5.29 " 245 25.0 IOI 2.0 22111. 534 44 4329 " II 10.37 A.M. 245 24-5 99 4-1 2h. oom. 2 934; 39 4334 " II 1. 08 I .M. 246 25.0 IOI 2. I 3Sm. 920 3 4347 " II 3-14 " 247 23-5 95 1-9 17111. 394 3 4362 " II 5.08 " 247 19-5 79 2h. um. 344 52 4369 " 13 10.30 A.M. 248 25.0 IOI 1.9 13111. 363 22 437" 13 11.51 " 248 24.5 99 3.6 ih. 34tn. 2 383 25 4373 13 3.00 I M. 248 !25.o 101 5.1 4h. 43m. 6973 73 4378 13 5- 5 " 249 25.0 IOI 1.71 04m. I 80 58 4397 14 10.21 A.M. 249 25.0 IOI 4.2 ih. 4om. 2 040 5 [layer at 12.16 P.M. 4410 M 1.27 P.M. 249 25.0 IOI 5-1 4h. 44111. 7020 7 Agitated surface of sand 4423 14 3.29 " 249 25.0 ioi 5.8 6h. 44111 99 i 9 Agitated surface of sand 4426 14 5-21 " 250 25.0 ioi 1.9 14111. 423 2 layer at 3.21 P.M. 4430 " 5 10.31 A.M. 250 25.01 ioi 2.8 Ih. 55m. 4203 6 443 15 11.32 " 250 25.0 ioi 2.8 2h. 55m. 4 228 23 Agitated surface of sand 4432 15 n-33 " 250 25.0 ioi 2.8 2h. 56111. 4253 5 layer at 11.28 A.M. 4433 15 n-34 250 25.0 ioi 2.8 2h. 57m. 4278 2 4434 15 n.35 250 25.0 ioi 2.8 2h. 58m. 4293 5 1 4435 ID 11.36 " 250 25-5 103 2.8 2h. 5gm. 4318 4 i 443<> >5 11.37 " 250 26.0 105 2.8 3h. oom. 4348 9 i 4437 ID 11.38 " 250 25.0 IOI 2.8 3h. oim. 4373 II I 4438 5 11.39 " 250 25.0 IOI 2.8 3h. 02m. 4398 6 4439 " 15 11.40 " 250 25.0 IOI 2.8 3h. 03111. 4423 15 4440 15 11.57 " 250 25.0 IOI 3.0 3h. 2om. 5833 9 4441 5 12.58 P.M. 250 25-5 103 5.0 4h. 21 in. 6393 3 4444 15 2. II " 250 25.0 101 7-7 5h. 34"i 7233 3 4449 5 3.26 " 250 26.0 105 6.8 6h. 47111 9973 16 Agitated surface of sand 4453 " ID 5.2O " 251 25.0 101 2.2 27111 706 12 layer at 3.04 P.M. 4457 " 16 9.42 A.M. 251 25.0 IOI 2. 9 ih. mm i 846 9 4460 " 15 1 1. 06 " 25 25.0 IOI 5-7 2h. 43111 3936 8 4470 16 I.I4 P.M. 251 25.0 IOI 4.0 4h. 44111 6 906 5 Agitated surface of sand 4474 " 16 2.45 " 251 25.0 IOI 7-1 6h. ism 9 3<> 3 layer at 12.42 P.M. 4496 16 5.03 " 252 25.0 IOI 2.1 29111 582 2 4504 " 17 2.45 252 25.0 IOI 3-5 2h. o6m 3052 5 45"! " 18 lO.lg A.M. 253 25.0 IOI 2.O 05111 146 32 4512 " 18 IO.24 " 253 24-5 99 2.O lorn 256 9 4513 " 18 0.29 " 253 25.0 IOI 2. I 15111. 386 45M " 18 0.34 253 25. c 101 2.2 2om. 5f 4515 " IS o-39 " 253 25.0 IOI 2-3 25111. 636 12 4516 " 18 0.44 " 253 25.0 IOI 2.4 3om. 766 IO 45 7 " 18 0.49 " 253 25. c IOI 2.4 35" - 906 5 4518 " 18 0.54 " 253 25. c IOI 2-5 4001. I 006 8 4519 " 18 0.59 " 253 2; . 1 IOI 2-5 15111 I 136 8 1 68 WATER PURIFICATION AT LOUISVILLE. TABLE No 4. Continued. Jewell System. Ra teof tj Collected. Fill ation. c J5 ^ w Period of w c" 3 u Number Q. c u T3 Service Since ii *J u e of Run. OtJD X WaswU |$ S.| Remarks. x Hours and Date. Hour. O C lu = O Minutes. u 3) -5 4> C 3 !B " ii rt 3 Ur 1 * X 3*^ u s a " J 03 1896 4520 July 18 .04 A.M. 253 25.0 01 2-7 5om. 1276 5 4521 " 18 .og " 253 25.0 OI 2.8 55m. 1386 235 4522 " 18 .14 253 25.0 OI 3-0 ih. oom. 1 526 83 4523 " 18 .29 " 253 25.0 01 3-1 ih. ism. 2 CK)6 4526 " 18 44 253 25.0 01 4-1 ih. 3om. 2 286 S 4531 " 18 59 " 253 25.0 OI 4-7 ih. 45m. 2 676 5 4532 " 18 .14 P.M. 253 25.0 OI 5-5 2h. oom. 3056 5 4533 " 18 .44 " 253 25.0 01 7-4 2h. 3om. 3886 8 4534 " 18 .14 253 23.0 93 9.1 3h. oom. 4516 4 4535 " 18 .21 253 24.0 97 3-5 3h. 05111. 4616 4 Agitated surface of sand 4536 " IS .22 " 253 23 5 95 3-5 3h. o6m. 4 f 59 97 layer at 1.19 P.M. 4537 " 18 .23 " 253 15-0 61 3-5 3h. 07m. 4659 182 4538 " 18 .24 253 18.0 73 3-6 3h. 07m. 4<>59 62 Wasting I min., 2ocu. ft. 4539 " 18 25 253 18.0 73 3-6 3h. O7m. 4659 30 " 2 " 40 4540 " 18 .26 " 253 22.5 3-7 3h. o~m, 4659 27 " 3 " 65 " 4541 " 18 .27 " 253 21-5 87 3-7 3h. 0701. 4659 17 Opening outlet. 4542 " 18 .28 " 253 22.5 3-8 3h. o8m. 4691 12 4543 " 18 .2g " 253 23.0 93 3-8 3h. ogm. 4721 14 4544 " 18 .30 " 253 23-5 95 3-8 3h. lorn. 474f> 6 4545 " 18 44 253 25-5 103 4.6 3h. 24m. 5 196 17 4553 4554 " 18 " 18 .14 44 253 253 25-5 25.0 103 IOI 6.6 8.2 3h. 54m. 4h. 24m. 5846 6 596 I [layer at 3.19 P.M. 4556 " 18 3-14 " 253 25.0 IOI 9-7 4h. 54m. 7286 9 Agitated surface of sand 4557 18 3.24 4h. 59m. 7 296 157 Wasting 2 min., 45 cu. ft. 4559 " 18 3-44 " 253 253 25.0 IOI 7-7 5h. ism. 7 776 10 4560 " 18 4.14 " 253 22. 89 9-3 5h. 45m. 8496 . 42 4565 18 5-21 254 25.0 IOI 2.0 2im. 53 5 4571 " 20 11.15 A.M. 254 25.0 IOI 5.0 2h. 44m. 4 160 10 4573 " 20 1.32 P.M. 254 25.0 IOI 4-1 4n. 5gm. 7430 12 Agitated surface of sand 4578 " 20 3-52 " 255 26.5 107 2. I 1 6m. 421 3 layer at 1.20 P.M. 458o " 20 5.09 " 255 25.0 IOI 3-6 ih. 33m. 2361 II 4606 " 21 I2.OO M. 256 25-5 103 3-8 47111. I 180 18 [layer at 1.08 P.M. 4609 " 21 I. 2 I I .M. 256 25.0 IOI 2-7 2h. osm. 3070 26 Agitated surface of sand 4612 " 21 3.12 " 256 24.0 97 5-8 3h. 56m. 5850 44 [layer at 4.36 P.M. 4615 21 5.07 " 256 24.0 97 7-2 5h. 4gm. 8620 IOI Agitated surface of sand 4620 22 Il.Og A.M. 257 25.0 IOI 45m. I 495 M3 Shut inlet and outlet H.IOA.M. 4622 " 22 11.59 " 31 in. 93 87 Shut outlet 11.59 A.M. 4623 " 22 12.33 I -M. 259 34-0 138 3-4 I2I11. 354 32 4625 " 22 2.18 " 260 22. O 89 4.8 54111. i 44i 248 Agitated surface of sand 4631 " 22 4-45 " 26! 25.0 IOI 2.8 33m. 858 745 layer at 2.45 P.M. 463!- 2 3 11.23 A.M. 263 21.0 85 3.0 44111. 979 51 4644 !! 23 1.02 P.M. 263 22.0 89 2.0 2h. 2im. 2969 124 Agitated surface of sand 4646 3-45 " 264 2O. O 8 1 2.1 39111. 877 g6 layer at 12.16 P.M. 4647 " 2 3 4-54 " 264 20.0 81 9-9 2h. i8m. 2877 141 4<>79 ! 24 1.03 " 266 20.0 81 6.0 ih. 32m. I 737 60 4687 24 2.46 " 266 20.5 83 4-5 3h. I3m. 3 537 27 Agitated surface of sand 4688 24 3.18 " 266 22. O 89 8.0 3h. 45m. 4 197 126 layer at 1.54 P.M. 4693 24 5.02 " 267 20.5 83 . 1.8 33111. 657 [layer at 10.23 A.M. 4706 25 11.14 A.M. 267 22. O 89 3h. I3in. 3947 52 Agitated surface of sand 4709 25 1.07 P.M. 268 21-5 87 1.6 I 2111. 264 40 [layer at 1.56 P.M. 4713 25 3.21 " 268 20.5 83 5-4 2h. 15m. 2 824 Agitated surface of sand 4718 25 5-03 " 268 21. 85 5-0 3h. 4&m. 4614 70 Agitated surface of sand 4724 27 9.OO A. M. 268 3 2 layer at 4.10 P.M. 4728 " 27 11.48 " 269 20.5 83 2.2 2h. I7m. 2778 7 Agitated surfaceof sand 473 27 2. 02 P.M. 269 ?o.o 81 7-3 4h. 3im. 3424 38 layer at 11.05 A.M. 4735 27 3 . 1 6 " 269 21.0 85 4.1 5h. 43in. 7038 M Agitated surface of sand 4737 27 5.02 " 269 21.0 85 8.7 7h. 22m. 9078 12 layer at 2.47 P.M. 4770 " 28 11.26 A.M. 270 25.0 101 2.3 2 I 111 . 525 4 4782 " 28 1. 06 P.M. 270 23-5 95 2.8 ih. 58m. 2 765 8 Agitated surface of sand 4810 " 29 9-37 A.M. 271 24.O 97 1-7 04m. no 9 layer at 12.46 P.M. 4811 1 29 9.42 " 271 24.0 97 1.7 ogm. 250 6 4814 29 9-47 271 24.0 97 I4m. 350 3 4815 " 29 9-52 " 271 24.0 97 i.i igm. 470 17 COMPOSITION Of OHIO RIVER WATER AFTER PURIFICATION. 169 TABLE No. 4. Continued. Jewell System. Rate of j s Collected. Filtration. c jj ji Number R Is c Period of Service Since si- 3 o u . B Run. s u s^i I Washing. j v 2i s Remarks. 1 Date. Hour. fc l g *! Hours and TS-^ o I s ig.? 8 223 "CJ 1 S 3 ffi . 1896 4816 July 29 9.57 A.M. 271 23-5 95 2.0 24m. 1 sgo 10 4817 " 29 10. 02 " 271 23-5, 95 2.1 2901. 700 4 4820 1 29 10.07 271 23-5 95 2.2 34m. 820 3 4821 29 10.12 " 2/1 23-5 95 2.3 3gm. ; 930 3 4822 1 29 10.17 " 271 23 5 95 2.6 44m. i 060 6 4823 1 29 IO.22 " 271 23-5 95 2.8 4gm. i 1 60 12 4824 29 10.27 " 271 23-5 95 3-1 54m. i 2go 10 4825 1 29 10.32 " 271 23-5 95 3.5 5gm. i 400 5 4827 29 10.47 271 24.0 97 5.i ih. 14111. i 760 10 4828 29 11.02 " 271 22.5 91 7-o ih. 2gm. 2 120 IO 4832 29 1. 17 " 271 22.5 91 9.2 ih. 4401. 2440 4833 27 I ih. 57111. *? i<v\ 70 4835 2 9 " 29 1.47 " / * 271 23.0 93 2.3 2h. o6m. 2970 / V layer at 11.27 A.M. 4836! " 29 2.02 P.M. 271 23-5 95 2.5 2h. 2lm. 2 310 13 4838 29 2.17 " 271 23.0 93 2.7 2h. 36m. 3680 II 4839 29 2.32 " 271 23.0 93 3.0 2h. 5im. 4040 8 4841 29 2.47 271 24.0 97 3-5 3h. o6m. 4370 i 4842 29 1.02 " 271 24.5 99 3-8 3h. 2im. 4730 21 4845 29 1.34 " 271 23.5 95 5-0 3h. 53m. 553" 25 4848 29 2.02 " 271 23 5 95 5-7 4h. 2im. 6 1 60 IO 4850 29 2.32 " 271 22. 5 9 1 6.8 4h. 5im. 6840 10 4853 29 3 02 " 271 23.0 93 8.1 5h. 2im. 7 520 12 4859 : 29 3-34 271 22. 89 9-2 5h. 53m. 8 290 7 Agitated surface of sand 4864 1 29 5-3 J " 272 24.0 97 i-7 I5IT1. 343 14 layer at 3.49 P.M. Western Gravity System. 608 1095 Dec. 23 11.37 A.M. i IO.O Gi ih. o2m. 547 495 610 " 23 12.27 l -M- i 10 6 1 ih. 52m. i 018 544 614 23 3.22 " i IO.O 61 4h. 47m. 2698 328 622 24 9-55 A.M. 2 IO.O 6 1 13111. 1 66 910 623 24 10.25 2 IO.O 6 1 43m. 475 510 625 24 12.34 P.M. 2 IO.O 6 1 2h. 5201. i 700 220 632 24 3-25 " 2 IO.O 6 1 5h. 43m. 3 56 168 635 " 26 9-57 A.M. 2 IO.O Gi 8h. 32m. 5 281 580 643 " 26 12.21 P.M. 2 IO.O 61 loh. 56m. 6812 580 648 " 26 4.12 " 2 IO.O 61 I4h. 47m. 9193 288 653 " 27 10.35 A.M. 2 IO.O 61 I7h. 3om. 10 822 570 657 1 27 12.53 P.M. 2 IO.O 61 igh. 48m. ii 857 540 662 27 3 .I8 " 2 5.0 30 22h . I3m. 12 582 324 667 1 27 4. 23 " 3 18.0 no I5m. 232 592 668 " 27 4-44 3 18.0 no 31111. 544 480 677 " 28 10. II A.M. 3 17.0 103 2h. I2m. 2 309 873 680 " 28 11.47 " 3 8.0 49 3h. 48111. 3329 708 687 " 28 3.27 P.M. 3 3.0 18 7h. 2Srn. 4245 I 260 694 " 30 11.17 A.M. 4 6.0 36 3h. 55m. 2652 332 699 3 1.52 P.M. 5 14.0 85 iGm. 203 726 702 " 3 4-3S " 5 II. 67 3h. O2m. 2456 528 79 1 * T I O ^A A M o 1 5 m . 172 276 713 " 31 vO4 * "* 11.05 6 23.0 140 ih. 26m. I 141 224 720 31 2.18 P.M. 6 6.0 36 4h. 2401. 3 15 260 1896 726 Jan. 2 IO.O4 A.M. 7 9.0 55 15111. 150 2OO 728 " 2 10.34 " 7 8.5 52 45m. 425 ig2 734 2 11.36 " 7 II. O 67 ih. 47m. i 050 136 736 " 2 ; 1.51 P.M. 8 10. C 61 1 5m. 122 392 737 " 2 2.21 " 8 12.0 73 45111. 4/2 304 746 " 3 9.51 A.M. 9 7-5 46 1 5m. 93 250 751 1 3 10.41 9 9.0 55 ih. o5m. 55 208 757 1 3 2.O8 P.M. 10 8.0 49 13111. 119 .)!< 759 1 3 2.40 " 10 9-5 S 45m. 429 412 762 4 11.29 A.M. II 1 ) 55 19111. 165 95 767 1 4 12.0.) P.M. II 9 o "5 54m. 437 25 770 4 2 18 " II 8.5 52 3h. o8m. I 587 122 170 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Western Gravity System. Rate of 8 Collected. Filtration. w c o fr. A Number t jjs. Period of j =" . O ^ a Run. So; X Last Washing. PI s| Remarks. 5 Date. Hour. 61 1*1 Hours and Minutes. w rt 3 11 & u s a " .4 1 1896 777 Jan. 6 12.10 P.M. 12 14.0 85 ih. 27m. I 103 416 78i 6 3-33 " 12 8.0 49 4h. som. 2297 320 784 " 7 12.18 " 12 9.0 55 gh. 25m. 5 193 98 788 7 3-49 " 13 i i .0 67 I5m. 172 92 792 7 4.19 " 13 II. O 67 45111. 552 88 797 8 12.09 " 13 IO.O 61 5h. ism. 3512 134 800 8 2-33 " 13 IO.O 61 7h. 39111. 5 002 128 804 " 8 3-04 13 IO.O 61 8h. lorn. 5 288 190 808 " 9 IO.O9 A.M. 14 IO.O 61 14111. 123 76 813 9 10.39 " 14 ii. 5 70 44m. 443 52 818 9 2.O3 P.M. I 4 8.5 52 4h. o8m. 2 593 172 823 10 11-45 A.M. 15 12. 73 3h. I7m. 2275 97 830 " 10 1.56 P.M. 15 12. 73 5h. 28m. 3935 120 833 " ii IO-54 A.M. 16 12.5 76 2gm. 419 53 837 " ii 11.28 " 16 13.5 82 ih. 03111. 849 135 8523 14 11.07 " 17 II. O 67 I5m. 154 57 853 " 14 n-37 " 17 II. O 67 45m. 33 861 M 2.12 P.M. 17 12.0 73 3h. 2om. 2 321 94 864 14 3.09 " 17 ii. o! 67 4h. I7m. 2954 60 873 15 10.47 A.M. 17 14-0 85 7h. 55m. 5 554 130. 877 15 12.52 P.M. 17 13-0 79 i oh. oom. 7 222 go 883 " 15 3-14 " 17 8.0 49 I2h. 22m. 8793 164 889 " 16 10.58 A.M. 18 14.0 85 lorn. 89 48 894 " 16 I. Og P.M. 18 20. o 122 2h. 2om. 2679 ig2 901 " 16 3.08 " 18 18.0 no 4h. 2om. 4979 150 904 " 17 9-59 A.M. 19 15.0 91 o6m. 70 176 95 17 10.03 19 15.0 91 lorn. 130 124 906 " 17 IO.I3 " 19 24.0 146 2om. 320 132 907 17 10.23 " 19 25.0 152 3om. 550 141 908 17 10.33 " 19 25.0 152 4Om. 520 102 909 17 10.43 " 19 25.0 152 5om. i 080 I 3 2 910 17 10.53 " 19 25.0 152 ih. oom. i 376 102 926 " 17 I. II P.M. 19 2O. 122 3h. iSm.i 4 500 240 929 17 2.04 19 19.0 116 4h. Iim. 5 500 140 933 17 3-4 " 19 18.0 no 5h. nm. 6630 238 937 " 17 4.04 " 19 16.0 97 6h. nm. 7620 188 943 " J7 5-08 " 19 7-5 46 7h. ism. 8380 198 946 " 18 9.56 A.M. 20 25.0 152 17111. 460 144 951 " 18 IO.2O " 20 26.0 158 41111. 940 258 954 " 18 1.23 P.M. 20 27.0 164 3h. 44m. 5 860 256 959 " 18 2.41 " 20 24.0 146 5h. O2m. 7869 256 967 " 20 IO.34 A.M. 21 22. 134 i6m. 276 192 975 " 2O 4.25 P.M. 21 21.0 128 6h. o7m. 8 226 246 981 " 21 11.42 A.M. 22 25.0 152 2h. osm. 3 007 M7 987 " 21 4.25 P.M. 22 16.0 97 6h. 48m. 8857 202 994 " 22 9-35 A.M. 23 25.0 152 I5m. 265 136 995 " 22 10.05 23 27.0 164 45m. i 155 174 999 " 22 2 28 P.M. 23 25.0 152 5h. oSm. 7955 266 1003 " 2 3 IO.I6 A.M. 24 28.0 170 5om. 1345 73 1007 " 23 3-43 P.M. 24 21. O 128 6h. I7m. 9655 64 1012 " 2 4 IO.08 A.M. 25 25.O 152 igm. 409 49 1017 " 24 2.OO P.M. 25 25.0 152 4h. nm. 6 &4g 145 IO2O " 25 9.41 A.M. 26 28.5 173 I5m. 397 75 1023 " 25 10.07 " 26 25.0 152 ih. oim. 2467 65 1027 " 25 2.23 P.M. 26 20. O 122 4h. 57m. 6509 165 1031 " 27 9.44 A.M. 27 23.0 I4O I5m. 322 612 1035 " 27 10.14 27 23.O 140 45m. i 102 554 1041 " 27 I.I3 P.M. 27 13-0 79 3h. 44m. 4112 934 1046 " 27 4.25 " 28 21.0 128 34m. 689 504 1055 " 28 1. 08 29 I4.O 85 5om. I 015 i 586 1085 " 31 11.00 A.M. 30 2O. 122 5om. 870 278 1088 " 31 2.33 P.M. 31 13.0 79 ih. o6m. I IOI 82O 1097 Feb. i I2.I8 " 32 g.O 55 2h. i8m. i 644! 326 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. TABLE No. 4. Continued. Western Gravity System. Rate of S Collected. Filtration. u. Ji j; i Number g. 5 o- a ServiceSince V C* U ^ i of 15 u " V Last > ~ f n. oj Remarks. ^3 Run. v J O u - E Washing. !^6S, g Date. Hour. fc. 5 |J| Hours and Minutes. SS3 P P IS rr a * n U rt ^ t/> u S | .2 CO 1896 IIOO Feb. I 2.45 P.M. 33 18.0 no 4001. 621 137 IIO.| I 5.03 " 33 7.0 43 2h. 58111. 2 8ll 118 no " 3 IO.35 A.M. 34 17.0 103 3im. 442 240 mi 3 I.l6 P.M. 34 12. 73 3h. I2in. 3 132 632 1113 3 3-23 " 34 22. 134 5h. igm 5 362 560 "7 3 5.00 " 34 6.0 36 6h. 56m. 6742 i 124 122 4 10.22 A.M. 35 26.O 158 32m. 780 1 421 125 4 11.49 " 35 ig.O 116 ih. 5gm. 2 SlO 720 128 4 2.31 P.M. 35 3-5 21 4h. 4im. 4430 908 132 4 5.18 " 36 12. O 73 ih. 5im. I 926 900 I 3 8 5 IO.26 A.M. 36 I4.O 85 3h. 29111. 3466 600 142 5 11.53 " 36 4.0 25 4h. 56m. 4 168 960 146 5 3.14 P.M. 37 6.0 36 2h. 49111. 2732 324 51 5 5.09 " 38 15-5 94 41111. 604 308 57 " 6 IO. 22 A.M. 38 16.0 97 2h. 24111. 2544 , 2 . 162 " 6 12.24 P.M. 39 20. o 122 04111. 58 I 276 165 " 6 3 .I8 " 39 12. O 73 2h. 5801. 3 146 655 170 " 6 4.19 " 39 8.0 49 3h. 5gm. 3718 460 75 " 7 10.22 A.M. 40 18.0 IIO 5gm. i 252 800 179 7 1.37 P.M. 41 17.0 103 ih. I2m. 1639 I 600 1 88 " 8 10.37 A.M. 42 7.0 43 3h. i6m. 2 061 321 193 8 2.27 P.M. 44 16.0 97 32m. 641 825 199 8 4-53 " 45 4.0 25 ih. 05111. 403 860 206 " 10 10.35 A-M. 46 20.0 122 05 m. 82 252 209 " 10 1.58 P.M. 47 6.0 36 ih. 33m. I 266 272 213 " 10 3-24 " 48 20.0 122 26m. 493 2go 218 " IO 5.07 " 48 13-0 79 2h. ogm. i 573 672 223 " II IO.I8 A.M. 49 g.O 55 57m- 926 251 226 " II 12.57 P.M. 5" g.O 55 ih. 4om. 458 461 229 " 1 1 3- "9 " 5 8.0 49 5601. 983 298 233 " II 5.15 " 52 6.0 2Sm. 37S 950 238 " 12 10.26 A.M. 52 S.o 49 2h. ogm. i 118 2=0 Shut outlet 10.26 A.M. 244 " 12 4-53 I .M.. 56 10.5 64 ih. 22111. i 197 605 250 " 3 9.54 A.M. 57 19.0 116 33m. 632 1 06 253 13 12.30 P.M. 58 23.0 140 2im. 458 235 256 13 2.25 " 58 4.0 25 2h. i6m. 2 158 1045 259 3 4.41 " 59 8.5 52 ih. 4im. I g8l 425 267 M 10.30 A.M. 60 18.0 no ih. 05111. I 348 209 271 M 1.22 P.M. 60 7-o 43 3h. 57m. 3M8 37 275 M 3-3 " 61 20. o 122 45111- 946 372 1279 M 4 52 " 6 1 6.0 36 2li. 07 m 2 246 408 1285 " 15 10 20 A.M. 62 iS.o IIO ih. oim. I 500 390 1289 15 1.33 I .M. 63 17.0 103 o6m. 104 5>7 293 5 3.13 " 63 16.0 97 ih 4&m. 2 I 34 975 1298 15 5.24 " 63 9.0 55 3h. 57tn 3424 919 1308 17 1 17 10.19 A.M. 1.45 I -M. 64 65 22. O 23.0 134 140 ih. oom. 30111. 474 609 865 650 1312 "7 3-13 " 65 lt,.n 97 ih. 58111. 2409 732 3 3 7 3 .2S " 65 3-5 B2 2)1. 13111. 2659 I 097 1316 " 17 5. 4 65 6.0 36 3h. 59i. 3389 602 1322 18 IO.32 A.M. 66 20.5 135 ill. 06111. i 396 i 220 1326 " IS 12 05 P M. 66 7-0 43 2h. 39111. 2 766 8co 1330 " 18 2.3O " 67 22.0 34 41 m. 788 4^5 1335 " 18 5.OO " 67 g.O 55 3(1. urn. 3068 33 1338 " 18 5.08 " 67 7- 43 3h. 19111. 3 128 616 345 " 19 10.24 A - M - 68 20.0 122 ih. 08111. i 308 158 319 19 11.40 " 68 8-5 52 2h. 2401. 2498 221 1353 19 3.11 r M. 69 IS. 5 113 ih. 42111. 2 3IO 1305 355 9 4-55 69 5-o 3 3h. 26m. 3310 184 11 2" 1 1.09 A.M. 70 27.0 164 15111. 385 118 1373 " 20 I.I9 P.M. 7" 6.0 36 2h. 25111. 2405 240 1379 " 20 3.22 " 71 25-0 152 36111. 828 435 1383 " 30 5-05 " 71 -.M 49 2)1. Igm 2328! 371 172 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Western Gravity System. Ra tepf S Collected. Fill fc il > ^ - f. V- Period of v. ^ O Number a o a 3 Last s "i v i Remarks. a Run. re Washing. 7, Date. Hour. fc. ^ ul Hours and M inutes. 1-3 c = |l S c. JT y. 2(j "u & U s ~ h (3 1896 1391 Feb. 21 10.03 A.M. 72 24.0 146 3001. 613 89 1396 " 21 12.49 I -M. 7 2 4.0 25 3h. l6m. 2 513 49 1399 " 21 3.12 " 73 14.0 85 ih. 42m. 2085 310 1402 " 21 5.00 " 73 4.0 25 3h. 35m. 2805 308 1419 " 24 10. ig A.M. 74 23.0 140 ih. oim. I 310 212 1425 24 1.25 P.M. 74 4.0 25 4 h. o 7 m. 4 170 2f)2 1428 " 24 3.30 " 75 15.0 9 1 ih. 49111. 2 282 2<.g 1433 " 24 5.21 75 4.0 25 3h. 4Om. 3250 2fo 1439 " 25 10.35 A.M. 76 22.0 34 ih. 1 6m. i 749 I 495 1443 " 25 I. 21 P.M. 76 g.o 55 4h. O2m. 5 1 19 680 1447 " 25 3.15 " 77 25.0 152 42m. 949 605 1450 " 25 4-55 77 5-o 3" 2h. 22m. 2543 I 270 1458 " 26 IO.32 A.M. 78 20. o 122 o6m. 91 475 1462 " 26 12.15 r -M. 78 21.5 131 ih. 4 9 m. 2471 480 1466 " 26 3.12 " 78 15-0 9 T 4h. 46m. 5971 700 1468 " 26 5- II 78 2.O 12 6h. 45m. 6801 962 1476 " 27 IO.3O A.M. 79 25. ( 152 ih. 2om. i 82g 455 1481 " 2 7 1.48 P.M. 79 6.0 36 4 h. 3 8m. 5 519 630 1485 27 3. of) So 27-5 If) 7 48111. i 177 337 1490 27 5.15 " So 8.0 49 2h. 57m. 4037 48,, 1498 " 28 10.46 A.M. So 12. 73 4h. 58m. 5937 129 1503 " 28 3.26 P.M. Si 24.0 146 3 h. I 4 m. 4864 445 1505 " 28 4-57 " 81 14. < 85 4 h. 45m. 6774 485 1515 " 2 9 10.40 A.M. 81 4.1 25 6h. i6m. 7524 235 1518 29 1.40 P.M. 82 25.0 152 2h. 4Sm. 3880 35" 1522 2g 3-23 " 82 23.0 140 4h. 3im. 6330 905 1525 29 5.05 82 12.0 73 6h. ism. 8550 i 115 1534 Mar. 2 9.46 A.M. 82 24. 146 7h. 24m. 9 5oo 357 1538 " 2 IO.27 " 82 22. 134 8h. 0501. 104015 1542 2 1.39 P.M. S3 27.1 164 2fmi . 620 i 685 1546 2 3.21 S3 22.0 134 2h. o8m. 2 850 4 ooo 1551 2 5-13 " 83 15." g I 4h. oom. 5 380 i 735 1559 " 3 10.45 A.M. S3 25.< 152 6h. O2m. 7 720 i 280 1563 3 12.55 I .M. S3 20.1 122 8h. I2m. 10659 i 005 . 1572 3 5.16 " i6.( 97 ih. 57m. 2 512 goo 1578 4 IO.5I A.M. 84 23.0 140 4h. O2m. 5 202 i 175 1582 4 1.02 P.M. 84 23. < 140 6h. I3m. 8 212 610 1586 4 3.23 " 84 20. o 122 8h. 34m. II 122 328 1591 4 5.07 84 7-0 43 loh. iSm. 12 892 660 1597 5 10.42 A.M. 85 19.0 116 37m. 759 795 1601 5 12.57 I .M. 85 23.5 M3 2h. 52m. 3 7<>9 i 085 1606 5 3-24 " 85 23.0 140 5h. igm. 7 129 745 1610 5 5. II 85 17.0 103 7h. o6m. 9459 5f>5 1618 " 6 10.38 A.M. 85 19.0 116 gh. O3m. "499 39" 1623 6 12.44 I -M. 85 7.0 43 nh. ogm. 13449 237 1628 6 3.24 86 25.0 152 2h. 1401. 2909 5<X) 1631 6 5.19 " 86 23-0 140 4 h. ogm. 5619 245 1639 7 10.50 A.M. 86 20. o 122 6h. lorn. 8 159 174 1^45 7 3.14 P.M. 87 25.0 152 ih. 4im. 2 256 477 1652 7 5-22 " 87 12.0 73 3h. 4gm. 5 KJ& 345 1658 7 I I. Of) A.M. 87 12.0 73 6h. o^m. 7456 194 1663 9 12.55 I -M. 88 21.0 128 56m. i 204 401 1667 9 3-27 " 88 24.0 146 3h. 28m. 4 f 34 525 1674 9 5.10 " 88 8.0 49 5h. nm. 6414 5 5 Id.So o IO.26 A.M. 89 22.5 137 ih. O7rn. i 509 26=; 1684 " o I.3S P.M. Sg 23.0 140 4h. igm. 5 f 49 168 1689 3.12 " 89 19. o 116 5h. 53m. 7689 157 lG95 " o 5.19 " Sy 4.0 25 8h. oom. 9 179 210 1702 " i 11.25 A.M. 25-0 152 ih. 1401. i f)2g 1 86 1704 " i 1.33 I .M. 90 23.0 140 3)1. 22m. 4399 141 1708 " i 3-25 " go 20.0 122 5h. I4m. 6829 234 1715 " i 5. if- " 90 8.0 49 7 h. o 5 m. 8789 161 1721 " 2 IO.22 A.M. 91 23.0 140 ih. 0701. i 584 250 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 73 TABLE No. 4. Continued. Western Gravity System. Rate of J Collected. Filtration. i ~ 5 fc Number t c a Period of c Service Since V. tC u u - 1 Run. t . Washing. L. <* V &gu. if " Remarks. f. Date. Hour. o! g<| Hours and Minutes. ~| z II S a ? " -2,3(3 "CJ y> U s 2 00 1896 1725 Mar. 12 1. 00 P.M. 9 24.0 146 .... 3h. 45m. i 5 114 140 173 " 12 3-31 " 91 20.0 122 .... 6h. i6m. 8 404 146 1735 " 12 5-17 91 5-0 30 .... 8h. O2m.| 10014 112 1741 " 13 IO.29 A.M. 92 23-5 143 .... ih. O7m.i i 479 440 1745 13 I.I5 P.M. 92 22-5 137 3h. 53m. 5 317 213 1749 13 3.19 " 92 19.0 116 5U. 57m. 7 927 I2g 1754 13 5-04 92 8.0 49 7h. 42m. 9 507 152 1761 14 10.36 A.M. 93 23.0 140 .... ih. 24111. i 926 127 1767 14 I. 1 1 P.M. 93 22. O 134 .... 3h. 5gm. 5 426 128 1773 M 3.16 " 93 16.0 97 .... 6h. 04111. 7856 149 1780 14 4-54 " 93 6.0 36 .... 7h. 42m. 9056 158 1787 " 16 10.33 A.M. 94 22. 134 .... ih. 1401. I 736 131 1793 " 16 I.O7 P.M. 94 20. 122 3h. 58m. 5326 490 1799 " 16 3.20 " 94 4.0 25 .... 6h. oim. 7og6 2OO 1805 " 16 5.07 95 22.0 134 .... ih. 2om. i 824 187 1813 " 17 IO.32 A.M. 95 20.0 122 3h. ism. 4354 228 1819 17 I.I7 I .M. 96 24.0 146 4om. 953 157 1825 17 3-23 " 96 22-5 137 2h. 46111. 3 803 445 1829 17 5.09 " 96 8.0 49 4(1. 32m. 5 673 380 1837 18 1O.32 A.M. 97 22.0 134 ih. igm. i 879 730 1843 " 18 I. 1 6 I .M. 97 II .O 67 4h. 03111. 5 269 230 1850 " 18 3-27 " 98 23.0 140 ih. 03111. 1414 585 1856 " 18 5.05 98 22.0 134 2h. 4101. 3 544 535 1864 " 19 I0.5O A.M. 99 ig.o 116 I2m. 179 440 1872 19 1.22 P.M. 00 21.0 128 1 5m. 265 800 1876 19 3.05 " 00 6.0 36 ih. 58m. 1 985 580 1890 20 IO.3O A.M. 01 24.0 146 2im. 468 i ooo 1896 " 20 I. 12 P.M. 03 1.0 6 5im. 659 7OO 1898 " 20 I. 4 6 " 04 20.0 122 o6m. 137 500 4145 July 2 10.32 A.M. 07 14.5 88 55m. 756 130 4146 2 11.14 "7 ih. 37m. 1239: 68 Shut outlet 11.14 A.M. 4154 " 2 2.24 I .M. 09 i h iGm i 029! i 500 A. Shut outlet 2.24 P.M. 4158 2 3.07 " 10 14.0 85 i8m. 225 52 4166 3 IO.I8 A.M. 12 14.0 85 3 8 ih. O2m. 8gg; g 4178 " 3 I. 1 2 P.M. 12 3 h . 56m. 3 253 10 Shut outlet 1.14 P.M. 4187 3 1.56 " 13 15.0 91 nm. 166 50 4195 3 3.42 " 14 15.0 9 3-o lorn. 142 77 4199 3 5-00 " M 14.0 85 7.8 ih. 28m. i 142 62 4316 O 11.22 A.M. 15 15.0 91 ih. 5om. I 682 4 6 4319 1.22 P.M. 15 20. o 122 3-5 3h. som. 3632 131 4322 " o 3.18 " 15 15-0 91 6.8 5h. 46m. 5242 152 4330 " I IO.47 A.M. 16 13-5 82 3-3 ih. 32m. M37 118 (335 I I.I5 P.M. 1 6 16.0 97 4-4 4(1. oom. 3739 199 * 4338 " I 2.35 " 16 12. 73 2. I o o 5 ooo Wast. 3 min., 52 cu. ft. 4339 " I 2.40 " 16 14.5 88 2.1 o o 3000 " 7 " 112 434" " I 2.45 16 14.5 88 2.2 o I 300 " 12 " 182 4341 " I 2.50 " 16 14.5 88 2.2 o 364 " 17 " 252 4342 I 2.55 16 14.5 82 2.2 o O 392 " 22 " 312 4343 i 3.00 " 17 18.0 no I . I oom. o! 387 Opened outlet 3.00 P.M. 4344 i 3.05 " 17 15-5 94 1-5 05111. 5^1 234 4345 i 3.10 " 7 15-5 94 i-5 1 0111. 146 213 4348 " i 3.15 17 16.0 97 2-5 1 5m. 236 204 4349 " i 3.20 " 17 5-5 94 2.5 2om. 326 157 4350 " i 3.25 " 17 5-5 94 2-5 25111. 3g6 142 4351 " i 3.30 " 17 16.0 97 2.6 3001. 476 169 4352 i 3-35 17 II, .0 97 2.6 35m. 54<>, 235 4353 " i 3-4" " 17 16.0 97 2.6 4om. 616 216 4354 " i 3-45 J7 16.0 97 2-7 45m. 6g6 199 4355 " i 3.50 " 17 16.0 97 2-7 50111. 776 229 435<> i 3-55 17 H,.,, 97 2-7 55m. S 5 (, 207 4357 " i 4.00 " 17 16.0 97 2.7 ih. oom. 936 4358 " i 4."5 17 16.0 97 3-2 ih. 15111. I 176 186 435Sa " I 4-3" " i- ii,. ,, 97 3.6 ih. 3001. 1 )! 271 Wasting. 174 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Western Gravity System. Rate of ~ V Collected. Filtration. aj c .H i u - fc, Period of 5i ti JS 3 s Number Run. S. I 8 - . -d DC Last Washing. Hours and &fc CJ . E Remarks. Date. Hour. G its Minutes. III I! u ~ d 5 1 g j m 1896 4359 July II 4-45 P-M. 117 15.5 94 3-9 ih. 45m. 1646 4360 " II 5-OO " 117 15.0 gi 4.0 2h. oom. I 876 4361 " II 5.05 " 117 2h o^m I 953 176 4505 " 17 2.52 " 118 15.0 91 4-1 54m. 861 33 4524 " 18 11.34 A.M. 118 16.0 97 6h. o6m. 5 723 46 4547 " IS 1.57 P.M. ng 16 . o 97 2-5 22m. 340 36 4555 " 18 3-12 " Ijg 16.0 97 ih. 37m. I 540 61 4660 24 0.50 A.M. 1 20 IO.O 6 1 2. I o o 320 Wasting 33 min., 334 cu. ft. 4661 24 I.OO " 121 16.0 g? 2.5 osm. 61 158 4662 24 I.O5 " 121 15.0 91 2.6 lorn. 141 168 4663 1 24 I.IO " 121 15.0 91 2.6 I5m. 211 4664 24 I.I5 " 121 15.0 91 2.6 2om. 28l 315 4665 24 1. 2O " 121 13-0 79 2.1 25111. 351 63 4666 24 1.25 " 121 13-0 79 2.2 3Om. 421 268 4667 1 24 1.30 " 121 13-0 79 2.1 35m. 49! 4668 24 1-35 " 121 3-0 79 2.2 40m. 561 161 4669 24 1.40 " 121 14.0 85 2.5 45m. 611 352 4670 24 1.45 " 121 14-0 85 2.4 5om. 691 202 4671 24 1.50 " 121 14-0 85 2.4 55m. 751 446 4673 24 1.55 " 121 I4.o 85 2.5 ih. oom. 821 157 4 6 74 1 24 2.IO P.M. 121 14.0 85 2.6 ih. 1501. I 051 214 4675 24 2.25 " 121 14.0 85 2.6 ih. 3Om. I 261 165 4676 24 2.40 " 121 14.0 85 2.6 ill. 45m. I 481 178 4677 24 2-55 " 121 14.0 85 2.9 2h. oom. I 711 112 4680 24 I.IO " 121 14.5 88 2.9 2h. I 5 m. I gu 275 4681 1 24 1.25 " 121 15.0 gi 2-9 2h. 3om. 2 III 177 4683 24 1.40 " 121 15.0 gi 3-0 2h. 45m. 2 321 237 4684 24 1-55 " 121 15.0 gi 3h. oom. 2 541 299 4686 24 2. IO " 121 15.0 91 3-1 3(1. 15111. 2 761 498 4702 1 25 IO. IO A.M. 122 12. o 73 o O Wasting 70 min., 763 cu. ft. 4703 25 10.38 " 122 13.5 82 23111. ",OO "So 4707 1 25 I I.2g " 122 13.0 79 6.7 ih. 14111. 980 3041 4708 " 25 1. 01 " 122 13.0 79 7-2 2h. 46111. 2 I3O 240 4714 " 25 4.38 " 123 12.5 76 2.8 48111. 5 4 511 4883 3 11.20 " 124 14.0 85 3-5 2h. 15111. I 889 450 4888 31 2.O4 P.M. 124 14.0 85 5.1 4h. sgm. 43 9 104 4889 " 31 2.07 " 12 4 14.0 85 6.0 5h. O2m. 5189 1 66 4892 31 3.38 " 124 13.5 82 5-7 6h. 33m. 5639 136 Western Pressure System. 1895 607 Dec. 23 11.33 A.M. 18.0 128 58111. I 070 171 609 " 23 12.24 P-M. 20.0 142 :.... ill. 4gm. 1934 260 623 23 3.19 24.0 170 4h. 4401. 55 SO 172 621 24 9.50 A.M. 22. 156 7)1. 1301. 9 3 8 2 95 624 24 12.31 P.M. 21. 149 gh. 5401. I 2 960 So 631 ! 24 3-22 " 21. OJ 149 I2h. 45111. 16568 90 f>34 26 g.5I A.M. 20. o 142 15)1. 26m. 1 9 839 860 642 " 26 12. 17 P.M. 17.0 1 20 17)1. 52111. 22 596 130 647 " 26 4.09 " 28.O 199 2ih. 44m. 28 571 242 654 " 27 10.42 A.M. 28.0 199 24h. 37m. 33 239 360 656 27 12.5O P.M. 28.0 199 26h. 45m. 36582 480 663 27 3-25 " 22.0 156 2gh. 2Om. 40 150 268 669 " 27 4-47 " 2O. O 142 30)1. 42111. 41 631 324 676 " 28 lO.Og A.M. 19.5 138 31)1. 56m. 43 336 494 678 " 28 11.15 " 25.0 177 19111. 588 810 679 " 28 11.45 " 28.0 199 49111. I 460 I 170 686 " 28 3.25 P.M. 22.0 156 4)1. 29111. 6384 I 1 16 695 " 30 11.22 A.M. 3 20. 142 55m. 984 246 700 1 30 1.55 P.M. 3 21. 149 3)1. 2801. 4 280 460 703 30 4.43 " 3 23.0 163 6h. 16111. 8067 520 710 1 31 IO.O5 A.M. 4 21.0 149 15m. 303 380 714 3t II. 10 " 4 23.5 1 66 ill. 2om. 2 705 143 72i 3 2. 2O P.M. 4 22.5 1 60 4h. 3om. 5 228 280 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 75 TABLE No. 4. Continued. Western Pressure System. Ra r of w J Collected. Filtr uion. tb jj 5 . i. / _ Period of U 5? 3 1 Number 8. 0. o Service Since ^ t ^ b a of Run. Soi 5 si X \VasWng. j aS Remarks. z Date. Hour. ul < S 3 Hours and Minutes. T3-* O ^3 | 11 i !.? s J.35 i" JS U s -> OS 1896 727 Jan. 2 IO.I8 A.M. 5 15-5 IIO I5m. 192 236 729 " 2 10.48 " 5 15-5 1 10 45m. 672 121 735 " 2 "39 " 5 19-5 138 ih. 36111. i 632 I3O 738 " 2 2.28 P.M. 5 16.0 113 4h. 25m. 4752 54 6 m 747 3 10.04 A.M. 6 21. O 149 I5m. 216 1 60 752 3 10.46 " 6 20. 142 57m. I 046 140 758 3 2.13 P.M. 6 18.0 128 4h. 24m. 4876 208 763 4 II.3O A.M. 7 12.0 85 I5m. 194 65 768 4 12.07 P.M. 7 16.0 113 52m. 721 116 769 4 2.15 " 7 14-5 102 3h. oom. 2 671 120 778 6 12.14 " 7 14.0 99 gh. 5im. 8671 1 86 782 " 6 3.36 " 7 16.0 "3 I3h. I3m. II 812 228 787 7 3-35 8 20. o 142 I5m. 349 82 791 " 7 4-05 8 20. o 142 45m. 936 76 796 " 8 12.05 8 20. 142 5h. 25m. 6860 210 80 1 " 8 2.36 " 8 18.0 128 7h. 56m. 8584 188 805 8 3.10 " 8 18.0 128 Sh. 3om. 9 I7<J 314 807 9 IO.O5 A.M. 9 17.0 120 2im. 312 73 812 9 10.35 " 9 21.0 149 5im. go2 22 819 9 2.O6 P.M. 9 2O. O 142 4h. 22m. 5 152 140 824 IO 11.48 A.M. o loh . ogm. 1 1 692 118 831 " IO 2.00 P.M. 9 18.5 132 I2h. iim. 14 152 184 834 " II 10.56 A.M. 10 19-5 138 28m. - 457 130 838 " II 11.32 " IO 20.0 142 ih. 04m. I 137 177 852b 14 II.I8 " n 22-5 1 60 I5m. 283 42 854 14 11.48 " ii 21.0 149 45m. 88; 148 862 14 2.18 P.M. ii 20. 142 3h. ism. 3578 62 865 " 14 3.12 " ii 19. o "35 4h. ogm. 4582 IOO 874 " 15 10.49 A.M. n 22. 156 7h. 56m. 9013 136 876 15 I2.5O P.M. ii 22. 156 gh. 57m. ii 732 140 884 " 15 3.20 " n 24.0 170 I2h. 27m 14973 1 66 890 " 16 11.02 A.M. ii 21 .O 149 l6h. o8m 19993 14 895 " 16 I.I4 P.M. n 29.O 206 i8h. 2om. 23543 104 902 " 16 3.II A.M. n 27.0 igi 2oh. I7m. 26 92; 248 911 " 17 11.02 " 12 23.0 163 1501. 375 170 914 17 11.32 " 12 28.0 igg 45m i 107 927 17 I.l6 P.M. 12 27.0 191 2h. 2gm 3827 156 928 17 2. 02 " 12 26.0 184 3h. I5m. 5067 198 934 17 3.17 " 12 28.0 199 4h. 3om 7177 236 ~"~ 938 17 4.06 " 12 27.0 191 5h. igm. 8 597 214 944 " 17 5.13 " 12 25.0 177 6h. 26m. 10 197 34 95 " 18 10. 16 A.M. 12 3O.O 213 7h. 2gm ii 897 190 955 " 18 1.28 P.M. 12 33-5 224 loh. 4im. 7997 268 958 " 18 2.39 " 12 30.0 213 iih. 52m. 20347 274 966 " 20 IO.32 A.M. 12 23.0 163 I5h. 42m 27057 1 80 970 " 20 1.53 I .M. 13 27.0 igi 3om 630 212 976 " 20 4.28 13 29.0 206 3h. o$m 4960 260 980 " 21 11.37 A.M. 3 28.0 igg 6h. 24111 10340 177 988 " 21 4.30 P.M. 13 28.0 199 iih. I7m 18 700 211 993 " 22 9.29 A.M. 3 24. 170 I2h. i8m 20600 100 ooo " 22 2.33 P.M. 13 30. 213 I7h. 22m 2g 780 140 004 " 23 10.19 A.M. 13 29. 206 2ih. I4m 36 250 222 006 " 23 3.40 P.M. 13 2O. 142 26h. 35m 45 too 130 Oil " 24 IO.O5 A.M. M 25- 177 1701 35i 59 016 " 24 1.55 I -M. 14 30. 213 4h. 07m 7ogi 128 02 1 " 25 9.52 A.M. 14 25- 177 8h. 03m 14041 103 026 " 25 2. 2O P.M. M 24.0 170 I2h. 3im 20 411 170 036 " 27 10.17 A.M. 14 26.0 184 I5h. ogm 24151 635 042 " 27 I.l6 P.M. M 23.0 163 i8h. o8m 28531 836 047 " 27 4.3 " 15 25.0 177 39m 957 770 052 " 28 9.58 A.M. 15 26.0 184 2h. 04m 3017 1448 056 " 28 3.20 P.M. 15 16.0 "3 6h. 4om 8277 736 o6ia " 28 4.40 " 15 14.0 99 Sh. mi,,. ) l >7 444 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Western Pressure System. Rate of S g Collected. Filtration. 55 . B s Number S. c v _o a -a Period of Service Since >- U 1 Su < O VH| Last Washing. K d V Remarks. a Date. Hour. |.l Minutes. III V - i O s & 2 E Jj u 1896 06 1 b Tan 28 4 080 A. 062 " 28 4-45 " 15 15 14.0 99 Sh. 05m. 9604 5 ooo A. 086 " 3 1 1. 06 A.M. 16 20.5 146 .... 58m. i 219! 336 087 " 3 2.29 P.M. 16 14.0 99 .... 4h. 2im. 4 82g ! 613 098 Feb. 12.24 " 17 20. o 142 2h. oSm. 2 848 326 01 " 2.48 " 17 13.0 92 .... 4h. 32m. 5 508 102 05 " 5.07 17 16.0 113 .... 6h. Sim. 7 528 135 09 " 3 10.38 A.M. 18 25.0 177 38m. 874 477 12 3 1.20 P.M. 18 24-5 174 .... 3h. 2om. 4774 607 14 3 3.26 " 18 25.0 177 .... 5h. 26m. 7824 492 118 3 5.O2 " 18 21.5 152 7h. O2in. 10024 600 123 " 4 10.25 A.M. 19 24.0 170 39m. 880 126 4 11.52 " 19 25.0 177 2h. o6m. 2 920j I 200 129 4 2.36 P.M. 19 14.0 99 4h. 5om. 6640 990 133 4 5 22 " 19 14.0 99 7h. 36m. 9640 255 139 5 10.34 A.M. 20 22. 156 02 m. 33 404 143 5 11.58 " 20 26.0 . 184 ih. 26m. 2 143 522 147 5 3.17 P.M. 20 I8. 5 132 4h. 45111. 6553 i 376 152 158 5 6 5.12 " 10.29 A.M. 2O 21 ii. 5 29.0 82 206 6h. 4001. 57m. 8393 i6ig 442 980 159 6 12. 06 P.M. 21 20. o 142 2h. 34m. 3969 i 024 1 66 " 6 3.21 " 21 23.5 1 66 5h. 49111. 8619 2 040 171 " 6 4.20 " 21 22. O 156 6h. 48m. 9959 I 800 176 7 10.25 A.M. 22 26.0 184 53m. i 340 6OO 1 80 7 1.40 P.M. 22 23.0 163 4h. o8m. 53<> I 6OO 181 7 3-37 " 22 23.O 163 6h. O5m. 7650 700 189 " 8 10.45 A.M. 23 22.0 156 ih. 27m. 2054 194 8 2.31 P.M. 24 25.O 177 2gm. 744 768 197 " 8 4-56 " 24 2O. O 142 2h. 54m. 3524 660 205 " 10 10.27 A.M. 25 18.0 128 ih. 0301. 1336 1 66 210 " 10 2.02 P.M. 26 20. o 142 ih. 32m. 2005 200 214 " IO 3.26 " 26 12. O 85 2h. 56m. 3575! 49i 2ig " 10 5. H " 27 Ig.O 135 ih. 14111. i 510 i 250 224 " II 10.24 A M. 27 20.0 142 3h. oim. 2 540 376 227 " II 1.03 P.M. 28 2O. O 142 ih. 5im. 2653 151 230 " II 3.22 " 29 22. O 156 56m. I 365 605 234 " II 5.18 " 29 6.0 42 2h. 52m. 3 595 39 237 239 " 12 " 12 10.23 A.M. 1.20 P.M. 30 30 23.0 I4.O 163 99 ih. osm. 4h. 02m. i 426 62 4 97& &35 240 " 12 3-13 " 31 24.5 174 ih. ogm. i 522 252 245 " 12 4-57 " 31 21. O T 49 2h. 53m. 3 802 4g8 251 " 13 9.58 A.M. 32 21.0 149 23m. 532 5& 254 13 12.35 P.M. 32 23.0 163 3h. oom. 42121 117 257 13 2.2g " 32 22. O 156 4h. 54m. 6 782 505 260 13 4-45 " 32 18.0 128 7h. lorn. g652! 957 268 ! I4 10.35 A.M. 33 22.0 156 ih. I5m. 1 668 227 272 1.26 P.M. 33 21. O 149 4h. o6m. 5 288 132 276 " 14 3-33 " 33 ig.o 135 6h. ism. 6 898 209 278 14 4.56 " 33 I8. 5 132 7h. 36m. 8 488 296 286 15 10.24 A.M. 34 27.0 igi 55m. i 402 580 290 15 1.38 P.M. 35 24.0 170 o6m. 93 i 085 294 15 3.16 " 35 I4.O 99 ih. 44m. 2 273 I 290 299 15 5.27 " 35 Ig.O 135 3h. 55m. 4 683 446 305 17 IO.24 A.M. 36 25.0 177 5om. i 130 925 39 17 1.48 P.M. 36 23.O 163 4h. 1401. 5 910 i 025 17 3-35 " 37 24.0 170 o6m. 155 44i 315 17 5. ii " 37 22. 5 1 60 ih. 42m. 2355 il 323 " 18 10.38 A.M. 37 21.0 149 3h. 3gm. 4 965 i 160 327 " 18 12.07 P.M. 37 20-5 146 5h. oSm. 6 925 695 331 " 18 2.34 " 38 24.0 170 3im. 734 73 334 " 18 4-59 " 38 21.0 149 2h. 56m. 4 084 848 339 " 18 5- ii 38 21.0 149 3h. o8m. 4314 "9 346 i 19 10.27 A.M. 38 22.0 151 4h. 54m. 6 444 604 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. TABLK No. 4. Continued. Western Piessure System. 77 R teof - Collected. Fill ration. Ix. jS io . u u Period of J t* ^ 1 Number 0. JS. _ a Service Since Last s i s b e of . "^ u ^ Washing. s ^ a w Remarks. \ Date. Hour. Run. ul 1:1 ? Hours ind Minutes. las ll c f 2 a" 1.33 Uf/j (_> S j h. 03 1896 1350 Feb. ig 11.43 A.M. 38 20. 142 6h. lom. 8804 51300 A. 1354 " 19 3.13 P.M. 39 23-5 166 2h. 4gm. 3 980 595 1356 19 4-59 " 39 24.0 170 4h. 35m. 6540 442 1365 20 11.15 A.M. 40 25.0 177 I3m. 258 197 1374 " 2O 1.23 P.M. 40 24-0 170 2h. 2im. 3498 305 1380 " 20 3.25 " 40 22-5 1 60 4h. 23m. 6278 745 1384 " 2O 5.08 " 40 ig.o 135 Gh. o6m. 8 298 881 1392 " 21 10.07 A.M. 41 20.5 146 3Sm. 1658 77 1397 " 21 12.52 P.M. 41 23.0 163 3h. 23m. 4388 MS 1400 " 21 3-13 " 41 18.5 132 5h. 44m. 7 208 401 1403 " 21 5-01 42 25-5 1 80 ih. 24m. 2 103 53 1420 24 10.22 A.M. 42 23.0 163 3h. ism. 4490 187 1426 " 24 1.30 P.M. 42 22.0 6h. 23m. 9030 848 1429 " 24 3 31 " 43 24-5 174 56m. I 316 338 1434 24 5.24 43 23.0 163 2h. 4gm. 3956 615 1440 25 10.38 A.M. 43 25-0 177 4h. 33m. 6266 I 205 1444 25 1.23 P.M. 43 22.0 156 7h. i8m. 10066 I 510 1448 25 3.18 " 43 21.5 152 gh. 1 3m. 12 556 720 1449 " 25 4-53 " 43 16.0 "3 zoh. 4Sm. 14456 368 459 " 26 10-34 A.M. 44 21.0 149 ih. I2m. I 602 205 1463 " 26 12.2O P.M. 44 25-0 177 . .. 2h. 58m. 4272 595 1465 " 26 3.II " 44 25.0 177 5h. 4gm. 8 5 I2 650 1469 26 5-15 " 44 25-0 I So 7h. 53tn. II 602 645 1475 " 27 IO.28 A.M. 44 24-5 174 gh. 25m. 13 702 887 1482 27 1.50 I .M. 44 16.0 "3 I2h. 47111. 17 Sl2 618 1486 " 27 3-<-9 " 45 32.0 227 59111. 493 99" M9I " 27 5-25 45 26.0 184 3h. 15111. 5 213 212 499 " 28 10.48 A.M. 45 30.0 213 5h. oSm. 8033 3<J5 1504 " 28 3 29 P.M. 45 24.0 170 gh. 4gm. 15 123 710 1506 28 5.00 " 45 20. 142 Ilh. 20II1. 17243 583 1514 " 2g 10.40 A.M. 45 25-5 I So I2h. 48111. 19433 443 1519 29 1.42 P.M. 45 23.0 163 I5h. 50111. 23713 I 7<x) 1523 29 3-25 " 45 21.5 152 I7h. 33111. 25953 I 140 1526 1 29 5.08 " 45 21.0 149 igh. i6m. 28043 33" 1535 Mar. 2 9.50 A.M. 46 25-0 177 22m. 492 Sio 1539 2 10.30 " 46 25-5 1 80 ih. O2m. I 502 1543 " 2 I 42 P.M. 46 27.0 191 4h. I 4 m. 6 162 i 475 1547 2 3-22 " 46 24-5 174 5h. 54m. 8722 6000 1552 " 2 5.16 " 46 25.0 177 7h. 48m. II 462 720 1560 3 10.48 A.M. 46 25-5 i ., gh. som. 14342 I 600 1562 3 12.52 P.M. 46 25-5 1 80 nh. 54m. 17452 74" 1567 3 3-16 " 46 20.0 142 I4h. i8m. 20 982 410 1573 3 5.19 " 46 22.5 160 i6h. 2im. 23 922 442 1579 4 10.53 A.M. 46 20.il 142 iSh. 25m. 26772 940 1583 4 1.04 I .M. 46 22-5 1 60 2oh. 36111. 29 722 683 I ; " 7 4 3-24 " 47 24.5 174 3gm. 970 222 1592 4 5.oS " 47 24.5 74 2h. 23m. 4 260 605 1598 5 10.44 A.M. 47 27.0 191 4h. 2gm. 6180 I 040 1602 5 12.59 I .M. 47 25.0 177 6h. 44m. 9520 705 if,. ,7 5 3.26 " 47 23.O 163 gh. nm. 12 980 345 1609 5 5.07 " 47 25.0 177 loh. 52m. 15420 370 1619 6 10.40 A.M. 47 23.0 163 I2h. 55m. 1 8 140 231 1624 " 6 12.45 I .M. 47 23-5 1 66 I5h. oom. 20 960 105 1629 6 3.26 " 47 2f).O 184 I7h. 4im. 24 620 585 1630 6 5.19 " 47 26.0 184 igh. 34m. 27 460 385 1640 7 10.52 A.M. 48 28.0 igg 52rn. I 230 845 1643 7 12.59 I .M. 48 25.0 177 2h. sgm. 4 530 280 1646 7 3 .I6 " 48 25-5 1 80 5h. i6m. 76.00 435 if,; | 7 5-25 " 48 24.O 170 7h. 25m. 10760 320 1659 9 II. Og A.M. 48 22.5 160 gh. 3gm. 13930 212 1660 g 9. CO A.M. to 3.25 P.M. 48 23 -O 16-5 1664 9 12.58 I .M. 48 25-5 * v j 1 80 nh. 28m. 16310 235 1665 1.2S " 48 24.0 170 I3h. 55.11. 19710 ;; = WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Western Pressure System. > B 3 y. Collected. Number R;i FiHr s. kt; la u eof = - . - = = < a " 1 rt X 1 _] Period of Service Since Last Washing. Hours and Minutes. y. . u &D sy .= JU u. . <J . 8 n Remarks. Date. Hour. 1675 1681 1685 1690 1696 1701 1705 1709 1666 1716 1722 1726 I73T 1736 1742 1746 1750 1755 1762 1763 1768 1769 1774 1775 1781 1788 1789 1794 1795 1800 1801 1806 1814 1815 1820 1821 1826 1827 1828 1838 1839 1844 1845 1851 1865 1869 1877 1883 1891 1897 1903 1909 1915 1921 1924 1930 I93 1943 1949 1950 1960 1963 1966 1896 Mar. 9 " 10 " 10 " 10 " 10 " I " i " 9- 1 I " 12 " 12 " 12 " 12 " 13 13 13 ." 13 14 M M 4 M 14 " 14 " 16 " 16 " 16 " 16 16 " 16 " 16 " 7 " l ~ " T 7 " 17 J7 17 " 17 " IS " 18 " 18 " 18 " 18 " J 9 19 9 19 20 " 2O " 2O " 20 " 21 " 21 21 " 21 " 23 i 23 2 3 23 2 4 : 24 24 5.12 P.M. 1O.28 A.M. 1.40 I .M. 3.16 " 5.20 " IO.25 A.M. 1.36 I .M. 3-27 " 3.25 P.M. to 3.27 P.M. S.lg P.M. 10.25 A.M. 1.02 I .M. 3-33 " 5.20 " 10.31 A.M. I.I7 I .M. 3-21 " 5.06 " 10.40 A.M. 9.30 A.M. to 10.40 A.M. I.I3 I .M. 10.40 A.M. to I.I3 I .M. 1.43 P.M. " 3.18 " 3.18 I .M. 4-55 " 9-OO A.M. 10.37 " 10.37 A.M. to 1.20 I .M. I.2O I .M. 1. 2O I .M. to 3.22 I .M. 3.22 I .M. 5.09 " 9 22 I .M. to TO 3; A.M. 10.35 A M. 10-35 A.M. to 1.20 I .M. I. 2O I .M. 1. 2O I .M. to 3.25 I .M. 3.25 P.M. 5-07 " 10.34 A.M. 9.25 A.M. to 10.34 A.M. 1.19 P.M. 10.34 A.M. to I.I?) P.M. I.I9 I .M. " 3.30 " 9.30 A.M. " 10.53 A.M. 10.53 " " 12.20 P.M. 12. 2O P.M. " 3. II " 3-II " " 5.30 " 9.00 A.M. " 10.35 A.M. 10.35 " " I.I5 P.M. I.I5 P.M. " 3.37 " 3-37 " 5.30 " 10.48 A.M. 1.28 P.M. 3-27 " 5-07 " g.OO A.M. to 10.30 A.M. 10.30 " " 12.05 P.M. 12.05 P-M. " 3.05 " 3.05 " " 4.00 " 9-00 A.M. " 11.30 A.M. 11.30 " " 2.30 P.M. 2.30 P.M. " 5.30 " 48 49 49 49 49 49 49 49 49 49 49 49 49 49 50 50 50 50 50 50 50 50 50 50 50 50 5o 50 50 50 50 50 51 5i 51 51 51 51 5i 51 51 51 51 51 52 52 52 " 52 52-53 53-54 54 55 56 57 57 57 58 58 58-59 59 60 60-61 61 24.0 22.5 24.0 25.0 24.0 25.5 23-5 23.0 22.9 24.0 25.0 24.0 22.0 22. 2J,.O 23.O 24.0 22. O 23-5 22-5 23.0 23.0 23.2 24-0 24.0 21.6 23.0 23-3 23-0 23.7 24.0 23.0 23.4 23.5 23.2 23.0 23.2 22. 5 22.5 22.0 22 6 23.0 22. f) 21.0 25.0 21.2 21.8 iS.s 19-3 16.5 16.3 ig-5 iS.o 18.5 13-5 II. O 18.4 17.0 16.8 19.3 21.4 J9-7 n.- 170 1 60 170 177 170 i so 1 66 163 162 170 177 170. 156 156 170 163 170 156 1 66 1 60 163 164 164 170 170 131 163 165 163 168 170 ,63 1 66 1 66 164 163 164 1 60 . 160 156 1 60 163 156 M9 177 149 155 134 37 116 US 138 128 132 96 78 131 1 2O 119 137 152 140 98 I5h. 42m. ih. I2m. 4h. 24m. oh. oom. 8h. 04m. gh. 39111. I2h. I2tn. 14)1. O3m. 22 090 1618 6398 8 788 ii 778 1-5868 17438 19938 175 330 125 20J 225 199 165 182 C. C. C. C. C. c. C. c. c. I5h. 55" - I7h. 35m. 2oh. oSm. 22h. 39111. 24)1. 26111. ih. nm. 3h. 57111. 6h. oim. 7 h. 46,11. gh. 5om. 22 518 24738 28 418 32058 34588 I 693 5603 8423 10863 13 633 129 430 298 565 195 181 194 163 no 177 225 139 205 98 187 132 274 3f>5 38. 322 395 305 480 505 685 500 i go 118 185 420 765 425 280 250 325 450 97<- 780 I OOO 500 400 500 800 890 980 I 050 I 430 i "5 I 105 780 740 297 500 5 8c . . . . I2h. 23111. 17 173 I4h. 28m. l6h. osm. 20 083 22333 .... iSh. 17111. 25 243 .... 2ih. com. 29 043 23)1. O2m. ; 24h. 49111. 31 933 34 223 .... ill 19111. 1843 4h. O4tn. 5 663 6h. oqm. ~h. 5im. gh. 48111. 8 563 10853 13 543 I2h. 33m. 17 263 ih. 28m. iSm. 2h. 17111. 3h. 57m. i 580 335 2085 3275 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 179 TABLE No. 4. Continued. Western Pressure System. Collected. Ra Fill t< ol 8 5 a; Period of *" 1? 3 Serial Numbc Date. Hour. Number Run. u| H U I " <^t o U I = &.? S X - ServiceSince Last Washing Hours and Minutes. ^lo fc. tr^ 153 i a E ~ ^u M Remarks. 1896 6 1 62 208 12 6 80 28-? 6-3 6 i 16 o 06 66 1 1 g 825 009 J 6-7 " 26 t. * " 26 63 I 6 Z 16 " 26 "* " 26 fin !_i g 26 26 1 6 i 3 " 26 TCg 2059 2060 2061 2062 2063 " 27 27 27 27 27 2.30 A.M. 2-45 " 2-59 " 3.2O " 3-30 " 71 7 71 72 72 17.0 17.0 i6. 5 17.0 17.0 16 3 20 20 16 20 20 5h. lom. 5h. 25171. 5h. 3901. O2m. I2m. 5 202 5452 5 662 37 217 290 300 400 425 IS 9 E. E. E. The series of results on run No. 72 was used in 2of)8 2069 2070 2071 2072 2073 2074 " 27 27 1 27 ! 27 27 27 5-50 " 6.00 6. 10 " 6.50 " 7.20 " 8.20 " 72 72 72 72 72 19.5 19.0 18.5 17-5 17-5 16.5 38 35 32 24 24 16 26 33m. 43m. 53m- ih. 33m. 2h. 0301. 3h. 0301. 687 807 987 1667 2 217 3187 189 295 470 190 185 260 but not for the day. 2079 2080 " 27 27 9.20 A.M. IO.2O " 72 72 17.0 16.0 16 6 20 13 4h. 03m. 5h. 03m. 4237 5 151 305 405 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 27 11 27 .. 27 .. 27 27 " 27 27 27 " 27 : 27 27 27 " 27 I 1.30 A.M. 12. 10 I .M. 12.20 " 12.30 " I2.4O " 12.50 " I.OO " I.IO " 1. 20 * 1.30 " 1.40 " 1.50 " 2.OO " -2 72 72 72 72 72 72 72 72 72 72 72 72 16.0 16.0 15-5 15-5 15-5 15-5 15-5 15.5 15-5 15.0 15-0 15-0 15.0 15 8 3 13 IO IO IO 10 IO 10 IO 06 06 06 06 6h. I3m. 6h. 53111. 7h. 03111. 7h. I3m. 7h. 23m. 7h. 33m. 7h. 43m. 7h. 53m. 8h. 03m. 8h. I3m. 8h. 23m. 8h. 33m. 8h. 43m. 6417 6897 7057 7217 7367 7517 7681 7837 7987 8147 8297 8447 8607 342 302 257 3 7 266 35 370 408 360 270 309 285 345 685 26 u8 16 6 1 66 " 27 28 17 8 08 28 28 5 30 " " 8 30 " " 28 18 I 28 28 Tfi 9 28 5.30 " " 8 30 " 76 A > 28 4j 2(1 2 I ( ) 2 " 2q 2.30 A.M. " 5.30 " 77-78 15. q 12 7.12 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Western Pressure Systei 3 e z 1 2166 2170 2174 2175 2176 2177 2178 2179 2183 2186 2190 2193 2197 2200 2204 2208 2211 2217 2221 2225 2230 2235 2238 2243 2248 2251 2256 2259 2263 2268 2272 2278 2282 2287 2290 2295 2300 2303 2768 2775 2781 2787 2795 2799 2805 2806 28og a8ii 2815 2819 2822 2827 2831 2858 2868 2874 2876 2882 2886 2893 2897 2901 2906 Collected. Number Run. Rale of Filtration. i j Period of ServiceSince Last Washing. Hours and Minutes. e |.|L i| lA it 3 u u - !Tg F Remarks. S. ! U osS, o^S _0. N Date. Hour. 1896 IMar. 29 " 29 29 29 29 ; 29 29 29 : 29 29 1 29 29-30 30 30 30 1 30 1 30 1 3i : 3i 1 3i April i " I " 2 " 2 " 2 3 3 3 4 4 4 6 6 6 7 7 7 May 7 7 8 8 8 8 " 8-9 9 9 9 9 9 II " II " II " 12 " 12 " 12 " 12 13 13 13 13 14 14 5.30 A.M. to 8.30 A.M. 8.30 " " 11.30 " 11.30 " " 2.30 P.M. 3.26 P.M. 3-32 " 3-42 3-57 4.12 " 2.3O P.M. to 5.30 P.M. 5.30 " " 8.30 " 8.30 " "11.30 " 11.30 " " 2. 30 A.M. 2.30 A.M. " 5.30 " 5.30 " " 8.30 " 8.30 " " 11.30 " 11.30 " " 2.30 P.M. 2.30 P.M. " 5.30 " 9.30 A.M. " II. 30 A.M. 11.30 " " 2.30 P.M. 2.30 P.M. " 5.30 " 9.30 A.M. " II. 30 A.M. 11.30 " " 2.30 P.M. 2.30 P M. " 5.30 " 9.40 " " II.3O A.M 73 78-79 79 So 80 80 80 So 80 So 80-81 81 81-82 82 82-83 83 83 84 84 85 86 86-87 87 88 15-3 16.4 16.4 21.0 2O. O 17-0 I6. 5 16.0 l6.g 15-7 i6.g 17.1 17-3 15 fj 09 16 16 323 i 686 i 315 635 i 055 355 95 109 2475 715 i 056 2985 5950 1585 i 725 i 265 i 145 885 i 455 2545 I IIO 9f>5 i 025 630 i no 525 i 185 I 160 475 175 198 182 57 85 130 64 7i 94 28g 126 20 113 91 c. c. c. c. c. From May 7-9 inclu sive, the results of both single samples and those collected by the sampler were used to obtain the average bacteria for days and for runs. C. C. 49 42 20 16 13 J9 II 119 121 122 14111. 2om. 3Om. 45m. ih. oom. 304 404 564 804 1054 14.1 14.7 14-3 15.9 13.0 11.7 17.0 16.1 15-5 99 I0 3 IOO "3 92 83 120 "3 no 1 08 2.30 P.M. " 5.30 " 9.30 " " 11.30 A.M II. 3O A.M. " 2.30 P. M 2.30 P.M. " 5.30 " 9.30 " " II.3O A.M II. 30 A.M. " 2.30 P.M 2.30 P.M. " 5.30 " 9.30 " " 11.30 A.M 11.30 A.M. " 2.30 P.M 2.30 P.M. " 5.30 " 9.30 " " II.3O A.M 11.30 A.M. " 2.30 P.M 2.30 P.M. " 5.30 " 3.2O P.M. g.20 " 3.05 A.M. g.OO " g.oo A.M. to 3.30 P.M 3.00 P.M. " 9.00 " g.OO " " 3.00 " 3.OO A.M. 3.00 A.M. to 8.30 A.M g.OO A.M. 1. 2O P.M. 3-27 " IO.OO A.M. to 1.55 P.M. 3.OO P.M. 9.00 " 3.00 A.M. g.OO " 12.00 M. 8.30 P.M. 2.OO A.M. 8.00 " 1. 00 P.M. 7.00 P.M. 3-OO A.M. 9.00 " 89 go go 91 9 1 91-92 92 93 93 93 94 94 94 95 96 96 96 96-97 97 97 97 97-98 98 98 98 99 99 99 IOO IOO 101 101 IO2 103 102 103 103 i "4 14.7 17.0 13.0 15-8 16.0 16.0 14.6 19.2 18.4 16.6 18.9 18.8 17-7 23.5 23.5 24.0 22.0 22.6 22.2 22.2 23.0 23.3 22.5 23.5 23.0 23-4 23-5 22. O 24.0 1 4 . < 23.0 22.0 24.0 23-5 23.0 24.0 23-5 23.5 104 1 20 92 112 "3 113 103 136 131 "7 134 137 125 166 1 66 170 156 161 156 156 163 164 1 60 166 163 165 1 66 156 170 170 163 156 170 166 163 170 1 66 1 66 6h. 2om 0701 5h. 52m nh. 47m 8 870 128 7988 16378 I5h. O5m. 20023 253 171 346 225 230 363 287 1 80 219 242 278 219 93 115 171 152 156 132 4h. 56m. gh. i6m. nh. 23m. 6973 13033 15943 5h. 2im. nh. 2im. 5h. 03m. lib. O3m. ih. 5gm. loh. 2gm. 3h. 36m. gh. 36m. I4h. 36m. 3h. 57m. nh. 57m. 4h. 3om. 7497 5637 7 oio 15 280 2683 14 154 5 ooi 13 121 19731 5 5oi 16425 6443 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 181 TABLE No. 4. Continued. Western Pressure System. Rate of ~ g Collected. Filtration. il .a ^ ~^ > i- Period of u g> 3 w Number a I - a Service Since w s ~ u s z of Is Cos ~ < o X Washing. Hours and !1! V ? Remarks. Date. Hour. .s.E Minutes. si ii I u !*" 3 l u 1896 2gio May 14 2.0O P.M. 104 23.0 163 gh. 3om. 13813 131 2gi6 " T 4 8.00 " 105 24.0 170 3h. 34m. 5 igo 136 2g2o " 15 1. 00 A.M. 105 23.5 1 66 8h. 34m. 12 380 181 2924 15 8.00 " 1 06 23.0 163 3h. 4om. 4978 222 2928 15 11.00 " 106 26.0 184 6h. 4om. g268 237 2934 15 5.21 P.M. I Of) 18.0 128 I2h. iSm. 17 705 700! Wasting 2 min., 66 cu. ft. 2935 15 5.24 1 06 19.0 135 I2h. i8m. 17705 705 " 5 " 106 2936 15 5.27 " 1 06 22. O 156 I2h. i8m. 17705 216 8 " 166 2937 15 5.31 107 ig-S 138 02 m. 53 115 2938 15 5-33 " 107 24.0 179 04111. 113 151 2939 15 5-35 " 107 24.0 170 o6m. 153 170 2940 5 5-37 107 24.0 170 08 m. 203 134 2941 15 5-39 " 107 24.0 170 lorn. 253 122 2942 15 5.41 107 24.0 170 I2m. 2 93 102 2943 15 5-43 " 107 24.O I7O 1401. 333 71 2944 15 5-45 107 24-0 170 i6m. 383 54 2945 15 5-47 107 24.0 170 i8m. 443 105 2946 15 5-49 " 107 24-5 174 zom. 493 70 2947 15 5.51 107 24.5 174 22m. 523 68 2948 15 5-53 " 107 24-5 174 24m. 573 82 2949 15 5-55 " 107 24.5 174 26m. 623 70 2950 15 5-57 " 107 24.5 174 28m. 673 82 2951 15 5-59 " 107 24-5 174 3"m- 703 56 2952 15 6.04 107 24-5 174 35m. 843 66 2954 15 6.14 " 107 24-5 174 45m. I0 9 3 74 2955 15 6.29 " 107 24-5 174 ih. oom. M33 81 2956 15 7.29 " 107 24-5 174 2h. oom. 2863 65 2957 15 8.2g " 107 24.5 174 3h. oom. 4303 85 2958 15 9.29 " 107 24.0 174 4h. oom. 5733 99 2959 15 10.29 " 107 24-5 174 5h. oom. 7 "3 go 2962 15 11.00 " 107 25.0 177 5h. 3im. 7873 142 2964 " 16 12.29 A.M. 107 24.0 170 7h. oom. 10 033 f 5 2965 " 16 1.29 " 107 24-5 74 Sh. oom. n 623 400 2gf>6 " if) 2.29 " 107 23-5 1 66 gh. oom. I2gi3 i2g 2967 " 16 3.29 " 107 24.0 170 loh. oom. 14 293 71 2968 " If) 4- 2 g " IO7 1 1 h . oom . 15 813 200 2971 " 16 5.00 " 107 24.0 170 nh. 3im. 16703 122 2972 " 16 5-2g " 107 24.0 170 I2h. oom. 17263 log 2974 " 16 6.2g " 107 23-5 166 .... I3h. oom. 18753 108 2975 " 16 7.29 " 107 24.0 170 I4h. oom. 20133 go 2976 " 16 8.29 " 107 24.0 170 I5h. oom. 21 523 go 2979 " 16 9-29 " 107 22.5 160 . . . i6h. oom. 22983 81 2982 " 16 IO.OO " 107 23.0 163 .... i6h. 3im. 23 6g3 9 1 2983 " 16 10. 29 " 107 23.0 163 .... I7h. oom. 24 363 63 2984 16 11.29 " 107 23-5 166 i8h. oom. 25753 98 2985 " 16 12.29 ".M. 107 23.0 163 .... igh. oom. 27 173 161 2986 " 16 1.29 " 107 23.0 163 .... 2oh. oom. 28573 136 2987 " 16 2.29 " 107 23.0 163 .... 2ih. oom. 2 9 933 127 2993 " 16 3.00 " 107 22.5 160 .... 2lh. 3im. 30673 151 2994 " 16 3.29 " 107 23.0 163 .... 22h. oom. 31403 142 2998 " 18 12.00 M. 1 08 14-5 102 2.3 2h. 45m. 2 406 I 1 20 3001 " 18 2.48 P.M. 1 08 14.0 99 4.6 5h. 33m. 4826 429 3oog " 18 6.07 " 1 08 15.0 106 7.0 6h. 57m. 6086 265 3011 " 18 g.OO " 108 15.0 106 4.7 gh. som. 8706 3000 3016 " 18 [2.(X> " 1 08 5-5 1 10 7.0 I2h. <om. II 466 198 3019 " 19 3.00 A.M. 108 4-5 1 02 14-7 I 5 h. 5 om. 14096 185 3025 9 6.00 " 1 08 15-5 no 12.3 iSh. 5001. 16786 215 3028 19 8.30 " 108 16.0 113 10.0 2ih. 2om. I8gi6 95 3033 9 I2.O8 P.M. 1 08 14.5 102 20. 8 24h. 58m. 22046 208 3038 19 3-05 " 1 08 15.0 106 20. 8 27h. 55m. 24606 300 3042 19 6.00 " 1 08 14.0 99 18.5 3oh. som. 27086 539 3043 19 9.00 " 1 08 14.0 99 18.5 33h. 5om. 2gs86 420 3049 9 12. OO " 1 08 14.0 30.0 36h. 5om. 32073 582 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Western Pressure System. Ra e of S u Collected. Filtr ation. il .H jj ~~i; l/Tu Period of V. * a ji Number a Q. o ServiceSince i - ~ LJ v - 3 Y. Date. Hour. Run. ul !li ul X Last Washing. Hours and Minutes. *|^ *- tnS i> c Remarks. 5 ? ? ~ a. rf - JU "Cj in u Z *1 tu n 3054 1896 May 20 3.00 A.M. 108 15.0 1 06 27-7 3gh. som. 34726 68 3058 " 20 6.OO " 108 14.0 99 27-7 42 h. som. 37 246 235 3061 " 20 8.30 " 08 14.0 99 30.0 45h. 2om. 39 266 242 3063 " 20 9-43 " 09 15.0 106 2-3 o6m. 73 340 3065 " 20 9-53 " 09 14-5 102 2-3 i6m. 233 165 3070 " 20 12. OO M. 09 14.0 99 10.3 2h. 23m. 2 103 310 3073 " 20 3.0O P.M. 09 14.5 IO2 4-7 5h. 23111. 4813 190 3078 20 6.OO " 09 14.5 102 7-0 8h. 23m. 7353 159 3083 " 2O q.OO " 09 14-5 102 7.0 nh. 23m. 9883 146 3 >87 " 2O 12.00 " 09 14.0 99 ii. 6 I4h. 23m. 12 193 144 3090 " 21 3.00 A.M. 09 14.0 99 16.3 I7h. 23m. I4&93 "3 3094 " 21 6.00 " 09 14.0 99 18.5 2oh. 23m. 17093 139 3099 " 21 8.30 " 09 13.5 88 19.6 22h. 53m. 19 133 135 3102 " 21 12.00 M. 09 14.0 99 23.1 26h. 23m. 22083 162 3109 " 21 3.06 P.M. 09 14-5 102 27.8 29h. 29m. 24733 183 3"3 " 21 6.00 " 09 14.5 IO2 32.4 32h. 23m. 27223 99 3116 " 21 g.OO " 09 14.0 99 37-0 35h. 23m. 29 7^3 89 3"9 " 21 I2.OO " 09 14.0 99 41.6 38h. ism. 32253 141 3124 " 22 3.00 A.M. 9 14.0 99 55-4 4ih. ism. 34 793 105 3128 " 22 6.OO " 9 14.0 99 53-0 44h. I5m. 37243 146 3131 " 22 8.30 " 9 14.0 99 55-4 4&h. 45m. 39 293 67 3138 " 22 12. OO M. o 14-5 102 4-7 2h. 4om. 2 270 71 3M3 " 22 3.00 P.M. o 14.5 IO2 7.0 5h. 4om. 4840 101 3M9 " 22 6.OO " 14.0 99 7.0 Sh. 4om. 739 92 3152 " 22 9.OO " o 14.0 99 9-3 nh. 4om. 9860 69 3156 " 22 12.00 " o 14.0 99 7.0 I4h. 4om. 12 390 45 3158 " 23 3.00 A.M. o 13-5 96 9-3 I7h. 4om. 14940 39 3161 23 6.OO " 14.0 99 13.9 2oh. 4om. 17430 33 3163 " 23 8.30 " o 13-5 96 16.2 23h. lorn. 19 420 34 3170 " 23 12.00 M. 14.5 1 02 16.2 26h. 4om. 22330 72 3171 " 23 3.00 P.M. 14.0 99 18.5 2gh. 4om. 24 900 45 3176 " 25 12. (.15 " o 14.5 102 18.5 33h. 45m. 28 260 91 3179 " 25 2.0O " I) 14.5 1 02 20.8 35h. 4om. 29940 64 3183 " 25 6.OO " 14.0 99 30.1 3gh. 40 m. 32242 63 3186 11 25 8.00 " o 14.0 99 25.5 4ih. 4om. 33990 55 3190 ." 25 12. OO " o 14.0 99 32.4 45h. 4om. 38 410 45 3193 " 26 2.00 A.M. o 14.0 99 32.4 47h. 4om. 40 090 33 3199 " 26 6.00 " o 14.0 99 39-3 5ih. 4om. 43340 32 3204 " 26 8.30 " o 14.0 99 41.6 54h. lorn. 45 3f>o 369 3206 " 26 9.28 18.0 28 2-3 nm. 198 153 3207 " 26 fg.32 " 18.0 28 2-3 I5m. 278 138 3210 " 26 IO.OO 17.0 20 2-3 43m. 758 138 3214 " 26 2.0O P.M. 17.0 20 II. 6 4h. ,|3m. 4988 57 3217 " 26 4.00 " 17-5 24 ii. 6 6h. 43m. 7088 59 3223 " 26 8.00 " 17-5 24 20.8 loh. 43m. II 208 48 3226 " 26 IO.OO " 17-5 24 20.8 I2h. 43m. 13 308 68 323 " 27 2.00 A.M. 17.0 2O 27.8 i6h. 43m. 17488 62 3232 I. 27 4.00 " 17.5 24 30.1 i8h. 43m. 19 608 27 3238 " 27 7.30 " 7-5 24 32.4 22h. I3m. 23 268 66 3244 " 27 I2.O5 P- M . 17.5 24 39-3 26h. 48m. 27:878 165 3247 " 27 3. (JO " 17.0 20 43-9 2gh. 43m. 301818 36 3248 27 3.12 " 20. o 42 4-7 o 900 Wasting 2 min., 33 cu.ft. 3249 " 27 3.14 " 25.0 77 4-7 o 265 4 " 63 " 3250 .. 2? 3-16 " 20.0 42 4-7 o 210 6 " 103 " 3251 " 27 3.18 " 15-0 06 4-7 oim. 17 igo 3252 " 27 3-22 " 17.0 20 4-7 osm. 87 140 3253 " 27 3.32 " 18.0 28 7-0 1501. 267 119 3257 ;< 27 6.00 " 17-5 24 7.0 2h. 43m. 2917 37 3259 " 27 g.(X) " 17-5 24 9-3 5h. 43m. 6027 32 3266 " 27 12. 00 " 17-5 24 18.5 8h. 43m. 9 147 25 3268 " 28 3-00 A.M. 17-5 24 23-4 nh. 43m. 12 227 30 3274 " 28 6.00 " 17.0 20 30.1 I4h. 43m. 15 537 39 3277 " 28 7.30 " 2 I7JJ 24 37-0 i6h. I3m. 17077 "iOO COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. TABLE No. 4. Continued. Western Pressure System. I kale of j u Collected. Filtr at ion. Si C i. ti. Period of "C * ,3 > Number S. r ^ _ Service Si nee j; .= j u u 6 of ~: . USB, ! Waging. **$. G.S Remarks. S-. Date. Hour. Run. ui <o Hours and -c^ o U I < Minuu-s. | SS2 .! 5 3jg 5 i f. ~35 "5 X u i j n 1896 3281 May 28 0.05 A.M. 112 17.0 120 41.6 iSh. 48111.: 19 897 74 3286 " 28 2.14 P.M. 12 24.0 170 4-7 o o 43 Wasting 2 min., 48cu.ft. 3287 " 28 2.16 " 12 20.0 142 4.7 o 300 " 4 > 88 " 3288 "28 2.1$ " 12 2O. O 42 4.7 o o 282 6 " 108 " 3289 " 28 2.20 " 13 15.0 (.(> 4.7 O2m. 4 330 3290 3291 " 28 2.22 " " 28 2.30 " 13 I " 15-0 "" 4.7 04m. 7i 235 3293 " 28 2.00 " 13 19. o 35 j 4.7 ih. 42m. i 871 256 3299 " 28 4.OO " 113 18.0 28 9.3 3h. 42m. 4021 153 337 " 28 S.oo " 113 7-5 24 7.0 7h. 4201. 8091 165 3313 11 28 IO.OO 113 18.0 28 9.3 gh. 42.11. 10311 156 3335 ; ; ;f, " 29 " 29 2.35 A.M. 2. .IS 114 16.0 13 lorn. 122 152 3341 " 29 4.00 " "4 17.0 20 9.3 ih. 35m. I 652 705 3344 29 5-47 115 7-5 24 4-7i 05m. 77 325 3345 29 5-57 H5 17-5 24 4-7 I5 m - 247 4 3357 29 7.30 " 1 15 17-5 24 4.7 ih. 4801. i 847 297 3362 29 12.07 I -M. 116 17.0 20 ! 7.0 5 gm. 9 5 78 3365 " 29 2.03 " 116 15.0 06 | 7.0 2h. 55m. 2 765 69 3369 29 6.00 " US 5-5 i" 4.7 Igm. 439 168 3374 11 29 8.00 " 118 15-0 o<) 4.7 2h. igm. 2 269 34 3376 " 2Q 11.23 " 119 2h. 57111. 2 OI 7 86 3379 " 30 12.24 A.M. 120 17.0 20 4-7 28m. y 1 / 468 280 3385 11 30 2.29 " 121 17-5 24 4-7 iSm. 298 369 3392 3 7.20 " I2 3 17.0 1 2O 2.3 40111. 658 118 3403 " 30 12.24 I -M- 127 17.0 120 7.0 ogm. 87 57 3407 June I 12.00 M. 129 17.0 120 7.0 2h. 54m. 2933 67 34 Io " I 3.O<> P.M. 129 17.0 I 2O 4-7 5h. 54m. 5893 Si 34M " I 6.OO " I2g 17.0 120 13. g 8h. 54m. 9213 171 3416 I g.(x> " 132 20.0 142 4.7 04111. 207 27 3421 " I 12.OO " 34 20.0 42 4-7 2gm. 746 201 3423 2 4.OO A.M. 136 9-5 138 2-3 2301. 700 420 3426 " 2 6.45 " 138 20.0 142 4.7; 16111. 361 I8 7 3429 " 2 10.25 140 II.O 78 13.9 3 om. 492 3 gix> A. 343" " 2 II.O4 MI 16.0 "3 2.3 oim. 6 1 68 3433 " 2 12.00 M. 141 12.0 85 4-7 57111. 831 310 3437 " 2 4.39 P.M. 143 12. O 85 2.3 ih. i6m. i 034 5 ioo A 344 " 2 6.55 " 145 14.0 99 4-7 . iSm. 225 42 3443 2 10 40 " 148 14.0 99 4-7 35m. 561 41 3448 3 3.30 A.M. 153 14.0 99 4-7 04111. 263 41 3452 3 6. 20 " 156 14.0 99 II. 6 I5m 237 54 34 f >9 3 6.00 P.M. 158 14.0 99 4-7 41 m. 620 79 3473 3 9.00 " 59 14.0 99 4-7 ih. 48m. I 5l8 73 3474 3 9-37 " 1 59 14.0 99 4-7 2h. 25m. 2038 i ooo A. 3479 3 12. OO " 161 14.0 99 4-7 ogm. 144 20 3483 4 3.00 A.M. 162 14.1 99 Ih. o6m. 890 29 vt" 4 6.OO " 162 14.0 99 2-3 4h. o6m. 3 260 184 349 4 7-05 " ,63 14.1 99 2-3 36m. 456 29 3493 4 g.OO " 164 J7m 670 8 3497 4 10.40 " 164 14.0 99 2-3 2h. 27rn. 2059 5 35 4 1. 10 P.M. 164 3-5 g6 4-7 4h. 57m. 4079 87 3502 4 3-4^J " 165 17-5 24 2-3 ih. 46m. 1875 23 3507 4 6.28 " 165 17.0 20 2.3 4h. 28m. 4685 97 35 Io 4 8.45 " 1 66 17.0 20 4-7 ih. 4gm. I 8og 19 3533 4 12. OO " 167 17.0 20 2-3 2h. 07m. 2 197 174 354 5 3.25 A.M. 168 17.0 20 2-3 211. 57m. 2 965 72 3544 5 6.00 " 169 I6. 5 1 6 2-3 2h. i6m. 2 2g4 32S 3545 5 6.32 " 169 I6. 5 16 2-3 2h. 48m. 2 714 I 7o A. 3547 5 9.00 " 171 JI7.C 20 2.3 46m. j 647 21 3554 5 4.05 P.M. 175 21. < 49 4-7 ih. oom. i 248 89 3555 5 4.42 176 20.0 42 2.3 igm. 340 25 3560 5 IO.OO " 77 20. C 42 4-7 3h. i8m. 3857 49 3564 " 6 12.30 A.M. 177 20.0 42 4-9 o 3 X> Wasting 3 min., 7ocu. ft. i8 4 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Continued. Western Pressure System. Rs teof J S Collected. Fill ration. (I, . c S . u ^ i_ renoaoi V. ^ 3 u _- 1 a Date. Hour. Number Run. o. fcS c v _o a O 3 a Last Washing. Hours and Minutes. Jlj c !o B Remarks. 1 Is ^ * r^ 1" 3665 1896 June 6 2.23 A.M. 177 22. 56 4-7 o 400 Wasting 6 min., 135 cu. ft. 3566 " 6 2-35 " 178 2O. O 42 4-7 02 m. 31 37 3567 6 2-37 178 20. 42 4-7 0(111. 81 43 3568 " 6 2-39 " 178 18.0 28 4-7 o6m. 116 33 3569 6 2.41 178 20. 42 4-7 o8m. 156 5 3570 " 6 2-43 178 2O. O 42 4-7 lorn. 196 22 3571 " 6 2-45 178 20.0 42 4-7 I2tn. 236 18 3572 6 2-47 178 2O. O 42 4-7 1401. 276 51 3573 " 6 2-49 " 178 iS.O 28 4-7 i6m. 311 37 3574 " 6 2.51 178 20. 42 4-7 iSm. 351 19 3575 " 6 2-53 " 178 20. o 42 4-7 2Om. 391 28 3576 6 2-55 " 178 20.0 42 4-7 22111. 43 25 3577 " 6 2.57 178 2O. O 42 4-7 24111. 471 21 6 2.59 " 178 21 .0 49 4-7 26m. 516 4 3579 " 6 I .01 178 21 .O 49 4-7 28m. 561 67 358o 6 1 . 03 " 178 21. 49 4-7 30m. 60 1 27 3581 " 6 1. 08 " 178 20. 42 4-7 35m. 701 25 3582 " 6 1.18 " 178 2O. O 42 4-7 45"i. 911 39 3583 6 1-33 " 178 20.0 42 4-7 ill. oom. I 211 47 6 1 .43 " 178 ih. lom. I 426 800 A. 3587 6 2-33 179 20.0 42 2-3 35m. 665 26 3597 6 79 2O. O 42 4-7 5h. 53m. 68l 5 3625 6 10.58 " 179 20.0 42 4-7 5h. 4om. 850 31 3630 6 1.55 P.M. 183 2O. 42 2-3 o6m. 273 69 3633 6 3-04 " 183 19.5 38 4-7 ih. 1501. I 503 31 9 1.30 " 1 86 20.0 42 2-3 igm. 375 121 3670 10 I I. 2O A.M. 189 17-5 24 4-7 2h. 1401. I 374 2 9 3673 " 10 I .00 P M. 189 18.0 28 4-7 3h. 54m. 4184 10 4 3677 " 10 3.30 " 191 7-5 24 2.3 38m. 806 2g 3683 " ii 10.40 A.M. 193 18.0 28 2.3 ih. ism. 14^3 9 3686 " ii I. 2O P.M. 194 17-5 24 4-7 I2m. 198 19 3694 " 1 1 3-45 " 195 17.0 20 2-3 iqm. 366 4 3698 " 12 IO.2O A.M. 195 18.0 28 2-3 3)1. 24111. 3 596 12 3705 " 12 2.42 P.M. 196 17-5 24 4-7 35tn. 631 17 3712 " "3 IO. 14 A M. 197 20. o 42 4-7 ih. urn. I 470 16 3720 13 1.25 P.M. 198 2O. 42 2-3 2401. 5 O1 17 3726 J3 2.57 " 198 20. o 42 7-0 ih. 5&m. 2311 320 3729 13 5.03 " 199 18.0 28 2.3 ih. 43m. I 858 18 " 5 g.OO A.M. 199 118 B. 3742 " 15 IO.I8 " 199 18.0 28 2-3 3h. 28m. 3778 23 3745 15 12.25 f.M. 199 17.5 24 7.0 5h. 35m. 6 108 16 3749 " 15 3-05 " 2CO 18.0 28 4-7 som. 933 41 3755 5 4-33 " 2O I 17.0 20 2-3 43m. 790 27 3761 " 16 IO.29 A.M. 201 18.0 28 4-7 3h. 09111. 3 330 16 37^5 " 16 12.40 P.M. 2O [ 18.0 28 4-7 5h. 2om. 5 600 52 3768 " 16 3.29 " 202 18.0 28 2-3 2h. 09111. 2387 46 3773 " 16 4-32 " 203 .8.5 32 2-3 27m. 482 17 3777 " 17 IO.O8 A.M. 203 18.5 3 2 4-7 2h. 33111. 2882 24 3779 17 12.52 " 203 18.0 28 4-7 5 h. I 7 m. 5 792 21 3784 17 2.55 P.M. 2O4 17.5 24 2-3 56m. i in 37 379 2 17 4.22 " 204 18.0 28 2-3 2h. 23m. 2 701 37 379 s " 18 IO. 12 A.M. 205 18.0 28 4-7 ih. 05111. I 273 212 3803 " 18 12.37 P.M. 205 18.0 28 4-7 3(1. 3om. 39 2 3 I0 3 3811 " 18 2.52 " 205 18.0 28 4-7 5h. 45m. 6303 169 3820 " 9 IO.O4 A.M. 2O6 iS.o 28 4-7 58111. I 129 26 3831 19 3.07 P.M. 208 18.0 28 4-7 42m. 702 53 jg 17 " 19 4.25 " 208 2h. oom. 2 l6l 1 1 J U 4 / 3857 " 20 II.lS A.M. 20g 18.0 28 2-3 2h. I2m. 2341 103 3864 " 20 12.46 P.M. 2O9 18.0 28 4-7 3h. 4001. 3941 43 3871 " 20 3-25 " 210 18.0 28 2-3 ih. 36m. I 683 54 3876 " 2O 4.40 " 211 18.0 28 2.3 o6m. 103 93 3885 " 22 10.22 A.M. 211 18.5 32 2h. 1401. 2503 41 3890 " 22 1.23 P.M. 211 18.5 32 5h. igm. 5 923 91 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 185 TABLE No. 4. Continued. Western Pressure System. Rale of J - Collected. Fi tration. J Jj Number 1 JS. o Period of ServiceSinc . bi *- - o u . e Date. Hour. Run. (L.S fli X Washing. Minutes! ||| ks Remarks. % u A 2 ta 1896 38g6 June 23 2.22 P.M. 211 6h. iSm 7 044 3 3956 " 24 3.26 " 212 20. 125 2.3 ih. 36m I gig iSf 3966 24 4-54 " 212 20.0 142 3h. 04171 36g? 250 3g85 25 IO. l6 A.M. 212 18.0 128 4-7 4h. 56m. 5949 36c 4003 1 25 1.18 P.M. 212 2O. c 142 7-o 7h. 58m. 9469 35 4004 " 25 1-54 * 212 4007 " 25 3-i8 " 213 , i 300 i oot) 4014 " 25 5.00 " 214 18.0 128 4-7 ih. igm i 354i 63 4025 " 26 10.29 A.M. 214 iS.o 128 4-7 3h. iSm 3444 12 4032 " 26 I.I7 P.M. 214 iS.o 128 7.0 6h. o6m 6 34 42 4033 " 26 1.22 " 214 6h. 1 1 m 6481 cn 4036 " 26 3-30 " 215 18.0 128 4-7 ih. 3im 1643 JV Si 4038 " 26 4-54 2|6 18.0 128 4-7 I2rn I74l 59 4045 " 27 IO.3O A.M. 216 18.0 128 2-3 2h. i6m 2434 127 4051 1 27 1. 00 P.M. 2l6 18.0 128 4h. 4601 5058 4053 : 27 2.03 " 217 18.0 128 2-3 42m 4054 1 27 3-15 " 218 18.0 128 ogm 157 370 4057 27 4-51 218 iS.o 128 2-3 ih. 45m 1*57 4063 1 29 IO.I8 A.M. 219 18.0 128 2-3 ih. nm i 327 12 4065 1 29 12.30 P.M. 220 18.0 128 7.0 ogm 156 29 4069 29 1.32 " 220 i.S .1 128 2.3 ih. inn I 306 63 4071 29 3-40 " 221 18.0 128 2-3 30111 529 4083 1 30 IO.I6 A.M. 222 18.0 128 4-7 ih. 0701. I 220 32 4101 30 12.47 I -M. 223 18.0 128 4-7 22rn 356 34 4106 30 2.55 " 224 18.0 128 4-7 24111 sgs 32 4U5 July I IO.27 A.M. 225 18.0 128 4-7 ih. 2701. 1464 4124 I i.ig P.M. 2-26 18.5 132 2.3 35m. 630 4132 " I 3-19 " 227 17-5 124 2-3 48m. 830 4136 " I 4-37 228 18.0 128 2.3 38m. 655 4206 " 6 10.30 A.M. 229 17.0 1 20 7.0 ih. i8m. i 284 13 4211 6 12.43 P.M. 22g . 17.0 1 20 7-o 3h. 3im. 3 574 24 4239 6 5-19 " 22g 17-5 124 4-7 8h. 07111. 8334 108 4247 7 IO.OO A.M. 230 21.5 152 7.0 49111. i 265 22 4254 7 1. 00 P.M. 231 19. o 7-0 25m. 868 49 4256 7 3.00 " 232 18.0 128 23111. 567 19 4265 8 10.55 A.M. 233 iS.o 128 4-7 ih. 4801. 2 184 II 4268 8 12.45 P.M. 233 18.0 128 3h. 38m. 3984 33 4272 8 3-54 234 2h. oom. i 868 20 4281 9 10.30 A.M. 235 17.0 1 20 4-7 ih. 25m. i 375 6 4284 9 12.28 P.M. 235 17.0 1 20 9-3 3h. 23m. 3345 10 4303 g 3-27 " i h . 5 1 in . i 885 112 4310 9 5-13 " 237 16.5 116 7.0 ih. 13m. i 205 112 4368 13 IO. 19 " 238 17.0 1 20 7.0 ih. oSm. 1074 25 4372 13 12.00 " 238 17.0 120 4-7 2h. 4gm. 2784 9 8 4374 13 3-05 " 23g 17.0 1 20 4-7 ogm. 120 no 4380 M g.o4 A.M. 239 II. 78 o O 249 Wasting 4min., 55 cu. ft. i i- 1 14 9.07 " 239 ii.. 1 1 "3 o O 446 Wasting 7 min., 78 cu. ft. 4382 4 9.12 240 16.5 116 2-3 osm. 82 1 86 43*3 M 9-17 " 240 16.5 116 2-3 lorn. 72 71 43*4 14 9.22 " 240 16.5 116 2-3 15111. 262 59 4385 14 9-27 " 240 16.5 116 2-3 2om. 342 32 4387 M 9-32 240 17.0 120 2-3 25m. 442 45 4389 14 9-37 240 7-5 124 2.3 3om. 532 37 439 4 9-42 " 240 17-5 124 2.3 35m. 622 33 4391 14 9-47 240 17-5 124 2-3 40111. 702 14 4392 M 9-52 240 17.0 120 2.3 45111. 782 13 4393 14 9-57 240 17-5 124 2.3 5om. 872 15 4394 14 O.O2 " 240 7-5 124 2-3 55m. 952 7 4395 M 0.07 " 240 17.0 1 2O 2-3 ih. oom. i 032 13 I jip 14 O.22 " 240 [7.0 120 2.3 ih. 15m i 282! 5 4399 14 0.37 " 240 17.0 1 2O 2-3 ih. 3om. I 532 2 4400 M 0.52 " 240 [7.0 120 2-3 ih. 45m. i 782; 7 4401 14 1.07 - 1" 1 .-i . 2. ; 2ll. 1X)T1). 2 032 108 i86 WATER PURIFICATION AT LOUISVILLE. TABLE No. 4. Concluded. Western Pressure System. Ra r -1 -; o :ollected. Filtr uion. 1 S . 15 - V ,-. " Period of u ? 3 u Number o. & o ServiceSince i; - ^j CJ ^* 1 55 Run. fe ~ 5 li Last Washing Hours and |fc at* u c Remarks. Date. Hour. p If ^ Minutes. Ssl 1 C la o -JU <-> c/i u i a " J (i. CO 1896 4402 July 14 11.22 A.M. 240 17.0 1 20 2.3 2h. I5m. 2 292 4 4403 14 "37 " 240 17.0 1 20 4-7 2h. 3om. 2 502 8 4404 M 11.52 " 240 17.0 1 20 4-7 2h. 45m. 2 802 2 4405 M 12.07 P.M. 240 17.0 120 4-7 3h. oom. 3042 7 4406 M 12.22 " 240 17.0 I2O 7.0 3h. 1501. 3292 9 4407 14 12.37 " 240 17.0 120 9-3 3h. 3om. 3542 2 4408 14 12.52 240 16.5 116 9-3 3h. 45m. 3792 3 4413 M 1.07 240 16.5 116 9-3 4h. oom. 4042 I 4414 M 1.22 " 240 17.0 1 20 7.0 4h. ism. 4302 II 4415 M 1-37 " 240 17.0 1 20 7.0 4h. 3om. 4502 4 4416 M 1.52 " 240 17.0 1 20 9-3 4h. 45m. 4812 4417 M 2.07 240 17- 1 20 9-3 5h. oom. 5072 I 4418 14 2.22 " 240 17- 1 20 9-3 5h. I5m. 5322 4419 M 2.37 " 240 17. 1 20 u. 6 5h. 30111. 5 572 10 4420 M 2.52 " 240 17- 120 9-3 5h. 45m. 5 822 ii 4421 M 3.07 " 240 17. 1 2O 6h. oom. 6082 54 4445 15 2.19 " 241 16. "3 7-0 ih. o8m. i 070 7 445<> 15 3-19 " 241 16. 116 7.0 2h. o8m. 2 ogO u 4451 15 4-49 " 241 17- 120 7.0 3 h. 3 Sm. 3 OoO 15 4458 " 16 9.47 A.M. 242 17- 120 4-7 39m. 680 18 4461 " if II.08 242 17. I2 4 7.0 2h. oom. 2 080 4471 " if I.l6 P.M. 243 17- I 2O 4.6 22m. 380 31 4572 " 2( II. 21 A.M. 245 17- I 2O 7.0 2h. I7m. 2307 lO 4574 " 20 1.37 P.M. 245 17- 120 5-9 4h. 33m. 4607 31 4577 " 20 3-31 245 17- 120 9-3 Oh. 27m. f 5i7 20 4579 " 2( 5.06 " 245 17- I 2O 9-3 8h. 02m. 8 117 23 4585 " 2 9.05 A.M. 245 14. 99 9-3 o o 820 Wasting 5 min., 78 cu. ft. 458f " 2 9.10 " 246 17- 120 9-3 osm. 05 192 4587 " 2 9-15 246 17- 120 13-9 zom. 145 82 4588 " 2 9.20 " 246 17- 1 2O 7.0 I5m. 235 72 4589 " 2 9.25 " 246 17- 120 7.0 2om. 315 41 4590 " 2 9.30 " 246 17- 120 7.0 25111. 405 32 4592 " 2 9-35 246 17.0 1 2O 7.0 3om. 485 190 4593 " 2 9.40 " 246 16.5 116 7.0 35m. 575 207 4594 " 2 9-45 246 16.5 116 7-0 40111. 045 193 4595 " 2 9.50 " 246 16.5 116 7-o 45m. 735 171 4596 " 2 9-55 246 16.5 116 7-o 5om 815 225 4597 " 2 IO.OO " 246 ,6.5 116 4-7 55m 895 lOg 4598 " 2 10.15 " 246 iS.o 128 7-o ih. lom i 155 195 4599 " 2 10.30 " 246 18.0 128 7.0 ih. 25m i 425 203 4600 " 2 10.45 246 18.0 128 9-3 ih. 4om. 1085 133 4601 " 2 II. OO " 246 18.0 128 7-o ih. 55m i 965 154 460^ " 2 11.02 " 246 18.0 128 7-o ih. 57m. I 995 137 4&<x " 2 11.30 " 246 18.0 128 4-7 2h. 25m 2485 509 4605 " 2 12. OO " 246 18.0 128 4-7 2h. 55m 3035 498 4607 " 2 I2.3O P.M. 246 18.0 128 7.0 3h. 25m 3 555 2 99 4610 " 2 1.26 " 247 17.0 1 20 4-7 23m 370 357 4611 " 2 3-05 " 248 16.0 "3 7.0 I5m 227 292 461^ " 2 5.05 " 248 16.0 "3 7.0 2h. ism 2077 158 4621 " 2 11.14 A -M. 249 15-5 no 7-o 2h. ogm 2 OOI 040 4626 " 2 2.44 P.M. 251 14-5 1 02 7.0 4601 072 37 4630 " 2 4-03 " 251 14.0 99 16.2 2h. O5m. I 792 i 186 4729 " 27 11.58 A.M. 254 14.5 IO2 7.0 44m &44 145 473 27 1.56 P.M. 254 14.0 99 i 1.1 2h. 32m 2364 170 4734 " 27 3.IO " 254 14.0 99 18.5 3h. 40m. 3424 138 4768 " 28 1 1. Of) A.M. 255 15.0 1 06 5om. 723 207 4783 " 28 I. 10 P.M. 255 14.5 102 2h. 54m. 2 543 105 4793 " 28 3.07 " 256 14-5 102 I5m. 189 4831 " 29 11.17 A.M. 257 15.0 1 06 2h. I4m. 1973 4847 29 I-4I P.M. 259 14.0 99 29111. 399 101 4855 1 29 3.16 " 260 14.0 9, 37m. 517 173 COMPOSITION OF OHIO RIVER WATER *>, >H I- u u fc aj u C E a C rEl K (! " Jl spi[o uom &^ . > fS"S S t 3 ^U J3)i:.v\ 3I|) qo d, Ji 2 a (J is 1 _ .-. *rca o o a- o o n m i-i t i- w o !OM*1 c/}O -<or cnr^oc<i- >-!OOO! ooooooooooooooooo 9) Ol" u !/) 18 fc 5 E | E c 5 E E E S E | ^ : c _ g^ c_ c WATER PURIFICATION AT LOUISVILLE. UMIUIIUIIV O s Il & P-l O - N I- O -t- u "* w ^^i -- - o o c -T ui o N O N t 2 : en r>. c ) r-> -t en - ?) t^ o^ r- r O O O O O O 5 2 J3 5 5 S 2 o o o o o o o O o o 6 o o SESc5 : EEs! 1 ~" . J= J= JS J5 XI J= SS- S S - - - S 53 - - 5? S - ? , -<-""<^"""^c: "^^"^a:" \C wiO O ir oD r^- ~t r^. w c^Occc COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 189 q 3 - St o- o o o- o o O O CO -t N t O (O tO 3JKAY paiiddy josi I en O c O t ** OOOOOOOOOQO oooocoooooo . -=^= j:j= jz-=j=^j= ^: j- ^ jz J= .2 EEEEEEEEEEEE JT"5, to " ? C 1 r " Tr "\ m ^"Si % ~ 7 i Jf 1 ^ " "" "~ ~* " " " " " t * 1 " 3 "f " vS J= -= J= -= J= -c -c ^ J= J= x J= J= J= J= J= J= J= J= J= j= J= js j: x: j= j= j: j= js j= j= j= j= j= o-. * -t unn jo jjqum N h- NfO-1 w^Or^ O O O O O O O 190 WATER PURIFICATION AT LOUISVILLE. /-S W C L. ^ cC a 5 JMI> [ in spi|<>s p.ipuodsng jo lunotuy 33i.M3.iv paivumsy SE6EEBSSEBSS6ES6EEBEBE6EE6EEE8BE6BSBE r C >n " f> 3 c-i O n i- O " ir. 01 O i^ O O O O O f> in O O O Cl O M o O O O O O O COf k - l tOt^^O"^t-it^>-(Oc r ^ | - i nOc^Oe^OTt-C lO > ?f 1 O>-(OOcoOOt^OOt^ S E E E t E E S ^ 5 _ S | j. d_| . S S || | B_ S | J S ^ E^ | ^ _ J S E^ H_ E : x .c js x: x J= js COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 191 ADU3IDIB3 IEIJ31DKH S3E43AV S.J si! 12* 3d SUO|1EOUOHI!1V : O O O r^ ui o OOOOO-OOt?OOO>-i- - OOO>- "OO~>- "OOOOOOi- OO oooooooooooooooooooooooooooooooooooo s.. s . . s.. s.. x. . s. s . a < " fc " " <; "" BU " " <"" fc" -<" - O O tn t^ O O O S.S.-S..S..S..S..S. S. - : 192 WATER PURIFICATION AT LOUISVILLE. o o c> S c* o o S*& o 5 5 Q p o <? p o"? O^WM I - I ^ > I- i O^ M O * I O O ^ <N HH oooooooooooooooooooooooooooooooooooo IM -rtciino n OO .? O O O "tO O t^%O O 88: 88888888 s s : . s a ...... s a a a ........ a . . . ssss, s a a a s s a a CT> O O O M r^-co o^ o -* M en Tf- u-no COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION, 1 93 una jo jaquin U| spnos papusdsns jo Hinooiy 33KJ3AV p3iiuns i B- c* oco o* co O O O O O > O N o o mo - ) O t t O 3" - O -TO " ~t -t ooooooooooooooooooooooooooooooooooo u S I^A\ SZZS-RXS ffRRS^&SJTR^Is^ P 3J31|!J CO -f "") C^ TtO - sjaas-gs^sssgis^ p3i|ddy ^ N w r^- o O o HHH^I|o|H X Q ^- "o a 1 E E S E E E E co ~t in c* \o O O |E|ESEgEEEEEEE 1 1 (|E E S E^E^ E_E| E| E| |E E S E E E -o </) T CO t^ c< - O O a a. e 2 S. iiinis 7. Z Z 7. Z Z Z SSS. Z Z 5SSS-. S 3 I sJJslff-r COM r^ u-i"t-T-l-cnO *n1-^ o S T^Nt-mmOO-COTtN-CJN M S o 0- >> M Mjg. . S . S E S S S 7. 7. 7 7 . ZZ - 7.7.7.-. . I r^ o -t *- o oo en w ir> t^ T w o w fo *-" u^O 1 - 1 OOO-r^ r^r^c^ 1 en ^H co Tf- in in "t cn^-rfc. 1^ n m 0> C-CO 1- JH N & ^^^,00 ^^OOOOCOOOMNNmt-KnTt s OVX S uny jo i-miimx sssaHas MMMMNNC<CINCIMN S. Z7.7.- . ~. Z ~. . - r-b*- 194 WATER PURIFICATION AT LOUISVILLE. a| J3)E M p3l[ddy JO! jo tuns aqi ipiqM aSviuai. -ttfin^-t-t-tt-t-tf-t-ttt -- ------------ ooooooooooooooooooooooooooooooooo paijddv E E S E E S E E E E E E E E E E E E E E E E E E E ^ 1 Owt- OO - O s a- s- - s- s- s- - a- s - s- sss- s- a - *>!>, U s- a- a- - a- a- a- - a- a : a ; 7. -. a ; s : a v rt::--:---r-- B - . - - - - s <* i O -H o s ^ COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 195 o o q q q q S? ?2 S 2 ? S ~ S CT S 2* g AI ui spi[<>s papuadcng Euiraniv 10 aiKi|d]ns dy jo lunomy 3EJ3AV 8225522223252522952 oooocooooooooooooo 82S2252525S55SS55 OOOoOOOOOOOOooOC O r-I^O in O> in 00 OO t~> o>oo CT>mNoDODQO O>Ot WAV PSilddV J s pai|ddv a>o I- I- 0>c *T-t-t-r-i--j--r-r TO TT-t-tTc^T-tTf *to 555 TOO TTTTT^T ooooooooooooooooooooooooooooooooooooo s-. j: ji j: xi J= .e J5 j= js f ! jo J= o J3 ^= J= -C J3 J3 J= j= j3 j: a s a a 8 a^ a || a a a a a a a a a | a B a a | a a | a SB a a a e ,c.c.c.e.e.r.e.c.c.e.e.cx!.c a a . s . < ^ " < *nMnwmwCMniHOa OMmin a> o i- !- i-O r** r^oo co O^C^O M tneoT s s . , . as,,, s s , s , . , . s . : 5ss ::::::: . : " 196 WATER PURIFICATION AT LOUISVILLE. t t -a- t t T J31KM pai[ dd V jo s ! EA\ sisK.tt pun i|s*A\ SMJ uing 3t[i qoiqM sainuMJSc! O O -to r^ t-tt-tO t- ocooooooooooooooooooooooo -oooooooo - *" Q3 . ^ ^ * . . E E E E E SEE E E = E E SEE E S E E : E E E : SEE r< r. M r< N N CT N N N N M N N N (-1 N N flN - tl fl - Cl SEES C-, O i" C- || E E E E_E | S _S E E E i- o r-. - 111 5 E E r^ N o ;_; ; "== ==5-=. .c j: x e g -= J5 X JS J3 J3 ^= - to" ,- -c : .n x EEEEE^EEIEEEE^EEEEEEEEEEEEE ; :iEE ; ; ;EE S S S S 2 S b -i d i b " < mo -t o ISc^^^tM !?>N S , . S S S . . S S . . . S ; * < pi c;""^c;"-"< ^OO-tOunnf^i^-O I coo -to incoo -tN COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 197 O <fl O i UOI[|!W J3d SlJHd -J91KM Aifl ui spi|os pipusdsns junouiy 33EJ3AV paiEUins o en o o -r-r-t-o cg o"o g"g~g 2 g g g g g ~ g g gg 8888888888 OJ QOOOOOOulOOOOOOOOO^pOO TO O C^OD O * ~ ft o C C^inin tC". paaailijun | I OOOOOOOOOOOO oooooooooooooo mo o O O "-i t O - E E E E as*57? sss?? ! , . , (j ** ff o ...... x . . . . ssa. sa. a. . xx. x. xxx a. < <~ t < iC" "<" s s a . a a, < fc < " al < I > O t 1-1 un jojsquinx o tntnmS^mc^mmn-r^r j ?? WATER PURIFICATION AT LOUISVILLE. uny jo jsqumN ir>o r-cooow^-tuiot--ooc?0"0 m f >no r,Qoo^ONm-*moi^oooo * * 1- * ^- ^OOOvOO^OOO^OO 1^ COCO *> yijj d IB.uaiDcg aSuaaAy ^H^IAO "4JSIS"^S "^^ ^ iiiJiiSiiisSlI ,^v ^sHglllE^^le, ?-s " S EII HcS SE^s H Hf as, !S 1 NNtn w \o 1- en * TJ- ^f o >-< M H H m 0.1 uinmiXBM S So^RoS ^8 Ro S ^-25< ^8 S :KSR ooo8 c S C.|C^ nM tni-< M 1H 1-1 HI tS II OQOOOOOOOOOOOOOOOOOCOOOOOQOOQOOOOOOOO oooooooooooooocoooooooooooooooooooooo uol[ J 3 A1>[ W d SIJEJ -J31EAA ui spijos pspusdsns jiy 33BJ3AV psicmnsg U01|EO J3d SUIEJJ) EUliuniv J03lt|d|ns oooopoooooooogopogooooooooooooooooooo .- -SJHOH tr asd 3JD V - g^s^K Sf-S 8"S ?! 8 S"8;g;?. < 8^.SSS ; 8>SS.:&SS;SiS:8 3 SS^S "^ S S o t^co i^ cno 01-iOt^Np-iooMinir.cncnNtn^N -*oo to co a- >D o o o>co M o CT.CO N ""5 * ; J3d jsaj oiqno J31KA,\ J3)EA\ psiiddy J<> si 3JSBA\ P"K MSBM 3M1 5 sq] tpiqA\ 3 Si!:uMJ3d n*^^ I -.Nnn-*-*oocojongp.ooeor.a>^ < >Mo, | J P3J311HUA MNOriMONrJ-fMNC-iciMOOOOOOOOOtOOOOOOOOOOOOO ^ & -p3J3)|IJ 1 qsE A \ coOOI-OOOOt^OOOOOOOOOOOOOOONOuiOOOOOOOOOO o ^c ^ o i^ *i- & p3J31[IJ Ijllllllllf Ifllf jlf Hlf II K 5 CS RSS o | a a pai|ddv !!!! S?!!=f !!!!!!!!!!!!!!! !!!!!!!!!!; S i SEESE^S^EEeE^SSEESEeEEeeESSSSESaSISEe MM-^-MHiO T-^-w5wrONentnr)cncn^ttnOOtnOOmOt*->OOOtnOcloOO 1 T3 a 1 BeBSBSEBEIE^ieiaEIEBSEHESSSS^aSSBSBBe B s SIEI^EEEI^EIE^SS^EEI^EEEEEEEEESEESSSS H * g ?,, -^5 5 ) S.-^ ) S S. o "^- o ^ S. S.S "^o "S S S m Si ^- S S. S So S. S S Si tn cn I liig.ip.iiisiiitis.sii^i^is.i.iiii.isiiiiii a O 5 > -^ 1 -g,-^.f,-g g S. ^.S "S. o ^. M S.S ">o "mw w m cn ^- cn S. S So S S S S cn tn s s . . s s. s s- s,.. s. ss s. s- a. , s 3 a < el -i^< s< < e^<e; < < - ^ (C "3 d> en in w a* w M cJ^ d^ o CT* w d o t-I d> o o^ w en o IH ui M ^i- M r^ w ri M rA w o cf i-! ^ M w s o ^ji C Q "fe S S S S . . S S - S S. , , S. SS S, S. S,, a ri < s. <t< oi < p;<fc ^s<< H.tnm5rfMO M ^^l^OC )mDC)HlkhHlNncn t ClO tr lOClCO^-ni-irJ-^-O J i N^- tc M s O*" " " N " l1MMi-l>-<MIHlHNMMnMC ^ja u, Q "& S unH jo jaqiun>i ^5i2asXKSK SR S>aSSSS5SSSg5 1 R.!tR5SK RP:?.R55 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 199 * un H J J3qmnj^ <S S W M S i?S <? o> oooo^S c- o &SS ooo o o v S o o S En2^?-^ > 2 -r < ;2 1 ADua, Hlj ^i [vijsi nrn .i3i:j,.\ y ^^g^gg-g^ggogggc; go o ^S^igggggggggggggg; . 3 B B A V in& &TsTSTHSgSSslSipSS^*^" 1 ^ 5 -^^** U fc | -ninmiu.jv S 8 &"8 5,!^S5o S K^ S> 8- :J?N^ S3 ?> $?? s? =s l& ^SISZZZZZSS^ Z ZS ? :S?S -^^^"-go-rcoc^ || 0000022888222222222222222222222222222 ocoogocoooocooooccoocgcocoocooooooSOQ *^ n^TTT^NNmmtntO^cn-l-mNNtNN^mT^NmCHN^^K.^-.^-^ u i > 1 1 "IN IH jad sjJEd -JlE,\v ji spi|os papuadsns o^o^oogooooooooojoggggooooogoopc^ooooo^oc V V K .a raiiraivi^USlptS ?S^S > P.S2 ?-g;SSSoS2o 1 M8f;5?>?;SSo-"i?SS > <2S8<ST p.ii|.li V JO JUnoUlV 33BJ3AV re 35 ; jad suony uojnjW ^-^-----?s?^o^s^s&^s;;^S5^^^ i! - <- s d,rwn D ??!SJ?5?S??J?S ^^?^^^??5??S 1 ^85????^??^??^???^^^ jajEA^ pailddy J ! OOCOOODOT r-0 00-0 UIUIUIO TO OOClt^OMOOON O> t^O I- r- TOO T Jo SSt am qoiqM a3EiuaDJ3j " S | P "" H " in ooooooooooooooooooooooooooooooooooooo 1 ^ -PI.U S S? 2 S 1 i AV HS5ISSItSftss K&SilS HH s^HHs^^T? a ^ poH!J ^^5l i l||s|^s|l58|||H|||||||||||-s|| K 2 S S C 1 1 % o pailddy !!!!?!!?!!!|f!|!!l!!r-!!!!!!!Efo!!f!! <n o = 55 ! t !l!l8 e l8 E l8 S 8 e 8 E l8 E 8 S |IIIIl8 e l3 E IIIIII!IIIIII w 1 s 1 ^r^^^^rr^^?;" _i n < c .fl EEEEEEEEEEEEEESEEESESEEEEEESEEEEEEEEE 3 * jj | SB 686188668688866886 ||B8J 88868 ijESii en m ^OT^T^^OOOr-^^OO^T-- ^TTTTTT^OI-Or-r-^co | IIIIIIIIIIIIIIIIIIIIIIIIIIIIIITIIIIII a O Si fl-S "S> m^-o S. SS S. S. S.fLrc ^.^5 "S "S " 5 "T T T T T -i-"Sio i- t~ i^ t- <-i ?. C S5 . 53 . S3 . S . 53 . S3 . S S . 58 55 J3 . . 55 . - S3 53 S3 S3 55 . S3 55 . . . . S S O I SJSSSlfSJJJfSSJSISSJIIlflSSiSSSSSRlJ HI ^ ,, r "rrrrr? . . Q "s ^ 55.. S. S3. S. S3. 53. J3 S3 . J3S33.. 55.. SSSSSSS- S3S.--. S 1 X ^>OTtncc t^N TO TO JTO O O TmOcc T-j^-O S c . t g ^o cTo S o^S Sl" u ^^Sfrpfr^^^^^?^^^?,^?,?,?,--"^"" " O "s" ~ !" " un M , OJ aqa,n N Sff^fffSi?S5 > 83.8.83:S < 8>S;?.8;8 5 S JJS 1 ^ S3 8*2 ~ 2 ??^S WATER PURIFICATION AT LOUISVILLE. ooocoooooooooo oooooooooo^ooogoooooo OOOOOOTOOOOOOOOOOOOOOOOOOOOOOOOOOOO OOOOOOOOOOOOOOOOOOOOl^OOOOOOOOO^OOOO : J3 j= js . ^ s." < s. < ^ < ^ < ^ < - < l";---,-----^.---------- ----"--"-- S 5 S S S ; SS S; S S : a ; ?: ; S ; 7. - S ; ; : : ; S- 2 S S ; : ; t < t < i <<i. <^ < - < fc < - < 1 -< COAf/ OSJTWN OF OHIO RIVF.R IVATl-.R AFTER PURIFICATION. 201 SUOJIK J UOI[|I W OOOOOOOOOOOOOOOOOOOOOOOOOOOCOOOOO oooooooo^-ooooooooooooooooooooooo l ^> ^=J EEEESEEESEEEEEEEEEEEEEEEEEEEEEE E | E E E E E E^ E E E ^ | E E c^ _ E | E II | I E E^ E I * ~. S ^ < ^ 8 8 8 E WATER PURIFICATION AT LOUISVILLE. C | O ^3 ^ ad suo[[Ky uoiniW O-.no> OOOOOOOOQOQPOQQQOOOOOOOOQOOOOOOOOOO . o o o o o o o o o o o > o o o o o OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOH-OOO : js j= j= J= j; j= -c j= js 43 j: j= j: 43 J= ^ O O w en en ny jo asquin>j COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 203 ,5 E o * A * -OTO-TUT J.I 11111..1UV 03KJJAV pOJBUljJSa * ~ OOOCOOOOOOOOOOCCOOOOO^-COOdOOOOOOOOc . O ct O O O ""> O C O O f> "- C ^ O C I - i i - O O O f- O O fiO OOC^OOOOO -O T i ~ O O EEEEEEEEEEEEEEEEEEEEEEEEEE agaBaa^aBBiBBBSBEEBB^BBEsgBBBEsgEBSBBBE 5 as., ssss. s. s WATER PURIFICATION AT LOUISVILLE. coooocooccooo oocc OOOOOOOOOO N 03 <N CN OCOc^OOOOOOOCO OOCQOOCOCOOOOCOCCOOOOOC OOOQOC-C>Ot- o -J-rooo O^eOOO O^C) f^OO O ^?o ^"I-H m O c^iOc/: OO r->oC -1-C)rrj ci MC/JO n M ^c ESESEEcEEEEEEESeEESEEEEEEEEEEEEEEESS E E E E E E E E E E E E B | 6eESE E S |,P,.HIR S l.| S l EE |S HO OO i-> t-i Oir -ft>c6^ TeJc>i^c\>Ai^ cf -i-cn^oo^c v - - d^o-i-6 i 6 c^ o* 6 d^ o^ ey- NC^7om1-Tf t^r-COOC>O It L7 ^", O J- ^- M CO C^ "T ^O r- O ( S -T^Mririr? COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 205 K. tt) 10 c o i- -1- -too O C* O* c OOOQOOOOOOOOOOOOOOO oooooooooooooooo 3 o - C> -t-o \:.\v pon<)dv jo i t .iiii saEiiio ij I s A\ *332SH1fS8R >oomMOOi/ S^?,2SS^ f. ?t^toc^ri -I- f-> oo o mco M oco M o M I^CO ^1 i^OT O P3J311IJ - 0- > Cl u- O-t-r-^coui^ci .nco m 00 co j - - o ,3,,dd V ,$,3, ?. . ^ 2, 5 "R S $ 3 Orir)Ot^OO"^ c> tn -t c4 o o M = >, Q SEEcEEESEESEEEc OfiOQ-OO^O-clgOOO EEEEEEES -fOJcioO-t-O -= -C X E E E E E E E ^ o o O O co 5 f> o o o fi o J= J3 E E E E || | E E | S E E | E EEEEEEEE Q^ i^ r^. o O *C -f co 1 1 E | E E | S s | S_E E| E II S E 5 E E S_ E S6SSSESS O inr-.-r-tr-.-tr-*. E E E E E E LD w -Tco O Q*O S-g = -S -5 g -g -S -g -S | cu i iliiiiiiiiiiiii Illlllll Illlllll 2i j=J=j=j=J=j:j3j=j3^^j=j=j=j: J=x-cj=x:j3j:j= J3 J5 JS J3 -C M fl W c^l N IN M O W N s -- a- a aa sa- s - - - as- as- " : ; - S - S - S S - d = -f C) Cl u^ 1-1 - ct w O O *r> w N N Cl c>c>o t^Mo- i o cJoc; N t iw-fN O ri m M w e^i o N ^S- ^ c! 4d l?l W jf jj rj. tjco o^ 0^ M m Tt t ui -no t^r^r-oDto co o"0 - M r, n f H 0>c .0 a MI - b! c a J?*ffMSfr8?Jaa S S : S S ; S p (nM^C 1 -- in co O s : a : s : s 5 i 5?cT !??? c2>?>" M t<1T --* in>n r^t^r^cooooo OOMM_11-1NC< s o . c M -^l uny jo JsqmnN g-jyor-.* o>o M cji^^^o^a <* s. N ?^^. .j.,2.20 n-i r>o o o >r> O ^ 206 WATER PURIFICATION AT LOUISVILLE. : t t ^ I O ^O t *t ~ COO^C^OOOOOO OOCOOOOOOOOOOOOOOOOmoOOOOO ,e.= .n.c.c.c.n.e.c.i:. SEEEEEEEEEEEEEEESEEESEE COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 207 _ SfOS .2 I J Hl SMK;) " !II!I\: o o c o o"o 6"o"g ooooooooo O C- NO una jo jaqmn N E E E S E E E S E J 1 E E S E E E E E E E E E E S E E E S E E E S 0- O O O O O - O O -H 2o8 WATER PURIFICATION AT LOUISVILLE. 8.8 1- * Tt-0 m T w (N O O OOO ^OQf < iOO*"fOro-t-o-f-t-r^r^o< >O o M P!l*lV jo si I ^ N ^ psjailtlUjI OOOOOOOOOOOOOOOOOOOOOOOOOOGOOOOOOO pa-isiM : E B E S E | j E S S E j E J E j E E E E S E E E E E | E E | E E S E E E E S . . S S 2 S S S S . S S S S S - . S S S . . S . S . S S S <" c:^5;<"" -s:^c:-<t;<s.<(""t.<:- < <&.< *S 5" CJ ^ >( ^. Tn f o^S, 8 ^ m ?> < 2 ct?iKM V S o^^riac^SNONOc;c^mNN OMOw^rtOOod i dod^d^d^wc^MeJd^d^do s Hto^^M^ovc*>woo^ S o o o COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 209 "> 3 S "o ll W J- d MJB.t -J.HKA\ JSAia ill sp.ios papuailsn;; jo lunuuiy 33iua.\v iiajiiuiiis J3d ij o;qn D oic^v pjiiddy J si JO U1IIS 3111 1|DI11A\ 3JSBJU5DJ3J ooooooooooooooooocooooooooooooooooooo OOOOOO --r~inMOto-TOO T r^ O c^ r^ O C O O * >-" en t^ -1- -t -t -f ir ir, >r ( iti OOOO I^r^r^WDcocrj OOOOC MMMMCinne4C4c*C4C4cinnc4cictMC4cir4Mcic4c-iNC4 ci ci ci N n f> I O i- OOOO r*^OO WATER PURIFICATION AT LOUISVILLE. -r^OO O i O oo "o" r^ C -TOCOOOOOO O C> w Cs . <1> C i. 0- U C ij >- m < V ^2 oooooooooooooooooooocoooooooooooooo EEEEeESSaeESEEEESEEEESEBEEEEESESESe c c c " c c n c c i:: c r: c S c c S : S c c " c S"S" G c 5>"S. SSSSSSSS. S,. SS. . SSS. SSSSS... S. S. . S. ". . ssassss. s., s s . . sss. COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. .<3U.ll3l.ua [IMJJIOL-a 331M3AV ID >3 * % 0) s * s (fl <o t z - 1 ^ < H o UO|[| !W J3d S1JEJ -JilBM jaAiN u| spiiog pspuadsng jo lunotuy aaujiAy pan-iims uoiiL-9 JM! suiwjy Euminiv J" ajiiijclins pai[ddv 0- C.O oooooooooooooooooooooooooooooooooo R|R| EEEEEEEEEEEEEEEEEEEEEEEEEEEEEESSEE O -t <r> T O Oco M O w -fr-O mu-iM O >^> N "". -n ^c -fc^O M Mu->ir>o -T-tO) OO"-CMOOOO^COMOc^t--tO"~>OOi-OOOc^wOOOOOO g fS ZT^"!?^"^ O"M 1 S H BIBBeSBI^BBS^SSeSESHBSSISSISBBlSSe na joJ3qmn\ Presc Presc WATER PURIFICATION AT LOUISVILLE. W c h-I i- 2 0) *^ +-* H v 8. a u y 33EJSAV panmnis}! < o -1- -t r^ -f^ OOOOOOMO"O -c^-f-l-C O OO -fmino O n "OOOOOOOOOOCO OOOOO . 01 5 . H> O >- 2 o. w c J u a o ^ s 5 COMPOSITION OF OHIO RIVER WATER AFTER PURIFICATION. 213 S.S 100-OOt^r^OQOOOQOOOOOOOOOOOOOOOOOOOOOOO pailddy J liinouiv | d SJno H 1 oooooooooooooooooooooooooooooooooooooo 2 > 8 8 8 "8 8 j; xi .c js O O -1-r>iDOO ss. s s s . . s s . . s s - - ss ...... s. s. ss b < a. " 4 i. o. " < " fc " ^ fc " " < fc <; - 2I 4 WATER PURIFICATION AT LOUISVILLE. O in P 2 ZH a W C H <D ^ ooooo^oo^oooooooooooogooooooo . ssssss- s . s . ?. s : s ::: s s - a j ; ; " 1- t t t -t r^co OO t- 1 SUMMARY AND DISCUSSJOiV OF DATA OF 1895-96. CHAPTER IX. SUMMARY OF THE PRINCIPAL DATA UPON THE EFFICIENCY AND ELEMENTS OF COST OF PURIFICATION, BY THE RESPECTIVE SYSTEMS, OF THE OHIO RIVER WATER, DI VIDED INTO TWENTY PERIODS, ACCORDING TO THE CHARACTER OF THE UNPURIFIED WATER ; TOGETHER WITH A DISCUSSION OF SOME OF THE MORE IMPORTANT FEATURES. BEFORE the presentation of a summary of the principal data obtained during 1895-6 upon the efficiency and the elements of cost of the purification of the Ohio River water, in twenty periods, according to the character of the unpurified water, there will be given as a matter of record some tabu lations showing the character of the purified water by days. For a detailed account of the composition of the Ohio River water by days, and of the amount of sulphate of alumina ap plied to the river water, reference is made to Chapters I and II, respectively. The ques tion of the decomposition of the sulphate of alumina and its removal from the water was discussed carefully in Chapter III. Table No. i. The first set of> tables in this chapter con tains a daily statement of the appearance of the water after purification by the respective systems. As already explained the appear ance of the filtered water is designated by five degrees of clearness, which may be described briefly as follows: Degree No. i signifies a brilliant water. Degree No. 2 signifies a clear water. Degree No. 3 signifies a slightly turbid water. Degree No. 4 signifies a turbid water. Degree No. 5 signifies a very turbid water. The first three degrees of clearness refer in each case to an appearance of the water which is satisfactory. It is doubtful if the con sumers would distinguish between these three degrees of clearness unless their attention were directed to the matter. Degrees Nos. 4 and 5 would be noted by the consumers, but it is to be stated that the adjectives used above have only a comparative value in rela tion to a brilliant water. In both cases the turbiditv would be very slight when compared with the river water before purification. Degree No. 4 refers to an appearance which would not be unsatisfactory for short periods if the water were of a proper charac ter in all other particulars. Degree No. 5 was objectionable both in its direct and in direct bearings, and was seldom noted for periods of long duration. In the second set of tables arc recorded the percentages of removal from the river water, by the respective systems, of the carbon aceous and nitrogenous organic matter, as indicated by the oxygen consumed and the nitrogen in the form of albuminoid ammonia, respectively. As a matter of convenience the total amount of nitrogen in the form of al buminoid ammonia in the river water, and the percentage of the total amount which was found to be undissolved in the water, are given. It will be obsereved that the total amount of nitrogeneous organic matter in the filtered water was less than the amount dissolved in the river water before purification. Table No. .?. The third set of tables contains a record of the daily average number of bacteria per cubic 2l6 WATER PURIFICATION AT LOUISVILLE. centimeter in the Ohio River water before and after purification by the respective sys tems, and also the daily average bacterial efficiency of each system. Bacterial efficiency means the percentage which the difference in the numbers of bacteria in the water before and after purification is of the number of bac teria originally present in the river water. In the next set of tables are presented the principal data obtained during the investiga tions with regard to the efficiency and cost of purification by this method. As stated at the outset of this chapter the results are divided into twenty periods, according to the charac ter of the unpurified river water. This was necessitated by the marked variations in the composition of the river water, affecting both the efficiency and the cost of purification; and also by the varying conditions under which the respective systems were operated. The official investigations of the several systems of purification by the method in ques tion began on Oct. 21, 1895. For a number of weeks after that time the investigations were less exhaustive than they were during the later portion of the period when the laboratory work had been more fully planned to meet the requirements of the problem. The Warren System began operations during the latter half of September. On October 21 the operators of this system contemplated some modifications in its construction. Ow ing to the remarkably low stage of the river at that time it was desirable for the Water Company to obtain data with this character of water. Accordingly they were officially requested to postpone their changes for a short time and operate the system with varied amounts of sulphate of alumina from day to day. This the} consented to do, but protested against the merits of their system being judged from operations preceding their contemplated changes. Operation of the Jewell System began early in July, 1895, and was said to have been continued nearly every day up to October 21, the commencement of the official tests. The installation of the Western Systems did not begin until early in November, and it was not until December 23 that they were ready for regular official inspection. Explana tion has already been given of the lengthy delays in these systems after about April i. From March 24 to 30 and April 27 to June 6 all systems were requested by the Water Company to be operated twenty-four hours per day, excepting Sundays during the latter period. Otherwise the systems were operated about 8.5 hours per day (irom 9 A.M. until 5.30 P.M.). A brief account is next given of the several periods into which the investigations are di vided according to the grade of the river water and other conditions of operation. Period No. i. This period extended from Oct. 21 to Xov. 25, 1895. It represents the last portion of the most severe and extended drought which had been noted for many years. With the low stage of the river there was an absence, comparatively speaking, of suspended organic and mineral matters in the river water; the amounts of dissolved organic and mineral matter were abnormally high; and the bacteria, while comparatively few in number, contained an unusually large propor tion of species coming from the sewage of cities situated farther up in the valley. The Jewell System was the only one that was in service during the full period. The Warren System was in service for a consider able portion of the time, but under the con ditions stated above; while the installation of the Western Systems was not completed. Period No. 2. This period extended from Nov. 25 to Dec. 24. During this time light rains fell. The water varied somewhat in character. At times the indications of sewage pollution were more marked than during the first period. The water became more muddy, and the chlorine and alkalinity decreased in a marked degree toward the end, although this did not follow in the case of the other soluble constituents. In fact there was an increase in some of them, notably in the ni trogen as free ammonia. With the rains the number of bacteria increased considerably. The Warren and Jewell systems were in SUMMARY AND DISCUSSION OF DATA OF 1895-96. 217 regular service during this period, hut the Western Systems did not begin operation un til it was practically ended. Period No. j. This period extended from Dec. 24, 1895, to Jan. 13, 1896. It repre sents the rising, maximum and falling stage of the river after the first heavy storm of the season. From this time to the close of the tests the chief variation in the river water was the amount of suspended matter which it con tained. These amounts will he noted in the tables. All four systems were in service. Period No. 4. This period extended from Jan. 13 to 27, 1896. It represents a fairly uniform grade of the river water from the fall of the preceding rise to the beginning of the next subsequent one. All four systems were in service during the greater part of the time. The sand layer of the Warren System was changed on Jan. 23. Period No. 5. This period extended from Jan. 27 to Feb. 6, and represents a rising stage of the river and increasing amounts of suspended matter in the river water. For unavoidable reasons the Western Systems were out of service on Jan. 29 and 3- On Feb. i the scope of the investigations was enlarged. The sand layer of the Jewell System was changed on this date. Period No. 6. This period extended from Feb. 6 to 13, and represents the height of a rise in the river when of course the suspended matter in the water was unusually high in amount. From Feb. 8 until about April i lime was applied to the river water in the case of the Jewell System. All four systems were in service, but the Warren System was delayed from time to time by changes in the devices for the intro duction of wash-water beneath the sand layer. Period No. J. This period extended from Feb. 13 to 27. and represents a falling stage of the river with decreasing amounts of sus pended matter in the river water. The most noteworthy features in the op eration of the several systems, speaking in general terms, were the irregular results along several lines, particularly those of bac terial efficiency and application of chemicals. Period No. 8. This period extended from Feb. 27 to March 20, and represents com paratively clear water between successive rises. All of the systems were quite regularly in service. On March 16 the operators of the respective systems were officially asked to comply with certain requests, leading to more regular and more efficient results of purifica tion. On Feb. 29 a new and separate pipe for the supply of river water was connected with the Western Systems. Period No. 9. This period included March 20 and 21, and represents very muddy water at the beginning of an extended freshet. The Western Gravity System went out of service owing to its failure to purify enough water to serve for washing its own sand layer. ; Period No. 10. This period extended from March 23 to 30, and represents a muddy con dition of the water and a high stage of the river. The suspended matter for the most part had a red color, however, and was much coarser than was noted under other condi tions. From March 24, 9.00 A.M., to March 30, 5.30 P.M., the systems were operated con tinuously, with the exception of the Western Gravity System, which was not operated at all. Period No. //.This period extended from March 30 to April 7, and represents a muddy water and falling stage of the river. Rains caused the water to vary considerably in char acter. All systems except the Western Gravity System were in service. On April 3 the representatives of the re spective systems were officially requested to get in readiness to operate, upon 48 hours notice, their systems night and day for such periods as the Water Company deemed ad visable. Period No. 12. This period extended from April 7 to 27, and represents a falling stage of the river and comparatively clear water. The end of this period marked Ihe beginning of a six wrecks period of continuous operation during each week from 9.00 A.M. on Monday to 4 P.M. on Saturday. This period was chiefly characterized by re pairs, following the official communications of March 16 and April 3, as noted above. 2l8 WATER PURIFICATION AT LOUISVILLE. In the Warren System the sand layer was changed. This caused a delay from April 13 to 20. The Jewell System was in regular ser vice. Neither of the Western Systems was operated at all during this period. Period A r c>. /j. This period extended from April 27 to May 18, and represents a period of comparatively clear water in the middle of a protracted drought. It also represents the first half of the six weeks period of continu ous operation. The Warren and Jewell systems were regu larly in service. From May 7 until the close of the period the Western Pressure System was in regular operation. At about the same date the repairs of the Western Gravity System were also completed. It was operated un officially on several occasions, but it was not put in official operation until after the Water Company requested an official explanation of the reason of its withdrawal from the tests. This request was made during the last week in June. During the intervening period this system was left out of consideration by all parties so far as active operations were con cerned. Period No. /-/. This period extended from May- 18 to 28, during the time of continuous operation, and represented the last portion of the comparatively clear water, before the end of the long drought. During this period the conditions of operation, with regard to rate of filtration and amount of applied sulphate of alumina, were prescribed by the Water Com pany as shown in table No. 4 of the last chap ter. No unusual delays occurred in the case of the Warren, Jewell, and Western Pressure systems, except as occasioned by the pre scribed conditions. Period A r o. 75. This period extended from May 28 to June 3, and represents a rapidly rising stage of river when the suspended mat ter was in part exceedingly fine, as noted in Chapter T. Great difficulty in securing coagula tion, even with large amounts of sulphate of alumina, was experienced.. Conditions of continuous operations, as above outlined, were prescribed by the Water Company so far as it was practicable. Period A T o. 16. This period extended from June 3 to 9, and represents the last of the con tinuous operations for 24 hours per day; and also the last of the prescribed conditions. The water was muddy but rather variable in character. The Warren, Jewell, and Western Pressure systems were in operation without any seri ous delays. Period No. if. This period extended from June 9 to July i, and represents three con secutive minor rises of the river. The period closed with the beginning of a marked rise which caused the water to become very muddy. The Warren, Tewell, and Western Pressure systems were regularly in operation with the exception of three days in the case of the Warren System. This delay was caused by repairs of a break in the agitator machinery. Period No. 18. This period extended from July i to 6. and represents very rapidly rising and falling stages of the river. Heavy rains fell on the local watershed and the water be came very muddy. The rise quickly subsided, and the period ended at the beginning of a minor rise. During this period the remodeled Western Gravity System was put in operation accord ing to a proposition offered by the Western Filter Company. It was agreed that for the balance of the investigations the pressure sys tem was to be operated on the first four days of each week, and the gravity on the last two days. This proposition was made in reply to a request from the Water Company for an official explanation of the fact that at that time the Western Gravity System had been withdrawn from the tests for a period of more than three months. The communication re ceived from the Western Filter Company, under date of June 26, 1896, is as follows: " The difficulties experienced in the earlier part of the filter tests occasioned by running both our filters on the same main with the other filters, which we hoped to remedy by the changes made in April, have been but par tially removed. We find now after several unofficial trial runs that, owing to wide varia tions in the pressure due to changes in ve locity in a 4-inch main, brought about by opening and closing either outlet, it is liable to impair seriously the results of our work to SUMMARY AND DISCUSSION OF DATA OF 1895-90. 219 run both filters at the same time. We have therefore continued the service of our pres sure filter, beliving that we obtained better results, at least mechanically, from that por tion of our plant. " \Ve are prepared, however, if it be desir able for the information of the Water Com pany, to run our gravity filter at such inter vals and for such periods as may be deemed advisable, discontinuing the service of our pressure filter during such periods." Operations of all systems were suspended on July 4, and on July 4 and 5 the sand layer of the Jewell System was changed. Period No. 19. This period extended from July 6 to 22, and represents a fairly uniform stage of the river with comparatively muddy water. Occasional rains on the local water shed caused several minor rises, but none of any large amount or extended duration. The Jewell Svstem was operated with higher amounts of sulphate of alumina than the condition of the water warranted, and consequently the effluent of this system was frequently acid. The Warren System was in regular opera tion, and the Western Systems were operated under the arrangement outlined above. Period No. 20. This period extended from July 22 to the close of the investigations on Aug. i. It represents the most marked rise noted during the investigations, and through out this period the river water contained sus pended solids ranging from 805 to 3347 parts per million. Great difficulty was experienced by all the systems in handling this water, owing to the frequent washings of the sand layer necessi tated by the large amounts of mud in the water after the short subsidence and before filtration. Relatively high amounts of sul phate of alumina were used with the view of securing satisfactory coagulation. During this period the operation of the sys tems was delayed in a number of instances in order that the Water Company might make a number of tests and observations of an engin eering nature. A very slight excess of sul phate of alumina above the amount capable of decomposition by the river water was ap plied during the majority of the period in the case of the Jewell System. Complications of a greater or less degree arose in the case of the \Varreu System, beginning about July 22, from the passage of sludge from the settling basin on to the filter. This was remedied on July 27 by cleaning the settling basin. Just how far this complication affected the results of this system is difficult to say. But it doubtless caused the sand layer to be washed at abnormally frequent intervals and caused a greater or less increase in the amount of applied sulphate of alumina, and something of a decrease in bacterial effi ciency. This table comprises all the qantitative and the leading qualitative data, arranged and compiled for each of the twenty periods into which, as has already been explained, the re sults of the investigations are divided. Maxi mum and minimum results during the periods were obtained by inspection of the records presented in fable Xo. 5 of Chapter VIII. The exact significance of all expressions not explained here was presented in Chapter VIII. The data presented in this table, and the methods of averaging employed, are as follows: Periods of Time. The average length of time per run included in the periods of " operation," " service," and " wash " are ex pressed in hours and hundredths of hours. These results were obtained in each case by dividing the total respective times by the number of runs included in the period. Quantities of Water. The average quanti ties of water per run are given in cubic feet. These were all computed by dividing the re spective total quantities for the period by the number of runs included in the period. In all computations for this table the actual quanti ties registered by the meters were used. Percentage which ///( Sinn of the Wash and Waste Water was of the Allied Water. These results were obtained by dividing the total quantity of wash and waste water for the period by the total quantity of applied water. Actual Rate of Filtration. The average actual rate of filtration in cubic feet per min- WATER PURIFICATION AT LOUISVILLE. ute was determined by dividing the total quantity of water registered by the liltered- water meter by the total period of service. These results are also given in million gal lons per acre per twenty-four hours by multi plication of the rates in cubic feet per minute by the proper constants. Average Net Kate of Filtration. These re sults were obtained by dividing the difference between the quantity of applied water and the quantity of wash and waste water by the period of operation, using in all cases the totals for the period. The rates in million gallons per acre per twenty-four hours were obtained from these rates by the use of the proper constants. A r c/ Quantity of Filtered Jl atcr per Run in Million Gallons per Acre. These results were obtained by deducting the quantity of wash and waste water from the quantity of the ap plied water, and multiplying the results by the proper constant value of i cubic foot in mil lion gallons per acre. Averages were ob tained by using total quantities for the period. Estimated Suspended Solids in River ll ater. Under this head are given the maximum and the minimum average amounts of suspended solids estimated for the different runs, and the average amount for the entire period. The averages were obtained by multiplying the average solids for each run by the quan tity of applied water on that run, and dividing the sum of these products by the total quan tity of applied water for the period. Grains of Applied Sulphate of Alumina per Gallon of Applied Water. The maximum and minimum amounts averaged for any run during the period are given, and also the weighted averages for the period. The latter were obtained in the same manner as the average suspended solids. Average Grains of Applied Sulphate of Alumina per Gallon of Net Filtered IVatcr. These results were obtained by dividing the amounts of sulphate of alumina per gallon of applied water by the percentages which the net filtered water was of the applied water, using in both cases averages for the period. Degree of Clearness. The maximum, mini mum and average degrees of clearness are given. The average degree given in each case is the sum of degrees recorded as aver ages for each run divided by the number of runs. Bacteria per Cubic Centimeter in River ll ater. The maximum and minimum aver age numbers per run. and the average number for the period, of the bacteria in the river water are given. The averages were obtained in the same manner as the average amounts of suspended solids. Average, Maximum and Minimum Numbers of Bacteria per Cubic Centimeter in the Filtered ll ater. These results are actual averages of the observations recorded as maximum and minimum numbers of bacteria for the several runs. \Yhere the number of observations was less than one half of the number of runs for the period no average is given. Average Numbers of Bacteria per Cubic Cen timeter in Filtered ll ater. The averages per run which were the maximum and minimum during the period are given, and also the average number for the periods. The aver ages were obtained by multiplying the aver age for each run by the quantity of filtered water on that run, and dividing the sum of these products by the total quantity of fil tered water for the period. Average Bacterial Efficiencies. The average efficiencies per run, which were the maximum and minimum for the period, are given, and also the averages for the periods. The latter were obtained by dividing the difference be tween the average numbers of bacteria in the river water and in the effluent by the average numbrr in the river water, these averages having been determined as described above. In this table are presented the total periods of time devoted to operation,-service and pre paring the filters for filtration; the total quan tities of water recorded by the meters; and averages of each of these periods and quan tities per run, obtained by dividing the respec tive total amounts by the number of runs. The records of the runs not included in aver ages are omitted from this as well as all other tables. There are presented also in this table the following averages: Average Actual Rate of Filtration. These results were obtained by dividing the total SUMMARY AND DISCUSSION OF DATA OJ< 1895-96. quantities of filtered water by tbe total periods of operation, rates in cubic feet per minute being transferred into million gallons per acre per twenty-four hours by the use of the proper constants. Average Grains of Sulphate of Alumina. The average amounts of sulphate of alumina per gallon of filtered water were obtained in each case by multiplying the average amount for each run by the total quantity of applied water on that run, and dividing the sum of these products by the total quantity of ap plied water. The average amounts per gal lon of net filtered water were obtained by di viding the respective amounts per gallon of applied water by the percentages which the net filtered water were of the applied water. Average Bacterial Efficiencies. These re sults were obtained in the same manner as were the average amounts of sulphate of alu mina used per gallon of applied water. Table No. 6. In this table are given summaries of the leading results obtained from the entire in vestigation and from certain portions thereof. Summary No. i includes the data obtained during the entire investigation (excluding those runs not included in averages). Summary No. 2 includes all the data given in Summary No. i, except those obtained during the periods when the operations were under conditions prescribed by the Water Company in regard to rates of filtration and amounts of sulphate of alumina applied (Periods 14, 15 and 16). Summary No. 3 includes all the data used in Summary No. 2 except those obtained during Period No. i, when the operations of the Warren System were under protest of the Cumberland Manufacturing Company, but were continued at the request of the Water Company. Summary No. 4 includes the data from all the perio ds when the Warren, Jewell and Western Pressure systems were in service, ex cept those when the conditions of operation were prescribed as noted above. Summary No. 5 includes those periods when all of the systems were in operation, ex cept Periods 14, 15 and 16. In this table the same expressions which have been used throughout are employed, and reference is made to the explanation of Table No. 5 of Chapter VIII, where the exact significance of the several expressions is ex plained. The data presented and the methods of computation employed are as follows: The periods of service and of wash are ex pressed in percentages of the period of opera tion. The quantities of water used for washing, the quantities of filtered water wasted, the quantities of unfiltered water wasted, and the quantities of effluent are, respectively, ex pressed in percentages which they were of the corresponding quantities of water applied to the respective systems. It is to be noted that in all cases in this table the quantity of filtered water (effluent) is the difference be tween the applied water and the waste water, and not the quantity measured by the meter. Average actual rates are given, the rates in cubic feet per minute being obtained by dividing the total quantity of effluent by the total period of service. The rates in million gallons per acre per twenty-four hours were obtained by the usual method of transference from comparative tables. The average net rates were obtained in the same manner as the actual rates, except that the net filtered water and net period of opera tion were used. The average amounts of sulphate of alu mina per gallon of applied water were ob tained in each case by multiplying the average amounts per run by the quantity of applied water on that run, and dividing the sum of the products by the total quantity of applied water. The amounts per gallon of net filtered water were obtained by dividing the amounts per gallon of applied water by the per centages which the net filtered water was of the applied water. The average bacterial efficiencies were cal culated in the usual manner of obtaining efficiencies, using for average numbers of bacteria results obtained in the same manner as the average amounts of sulphate of alu mina. 222 WATER PURIFICATION AT LOUISVILLE. TAIH.E No. 1. SUMMARY BY DAYS OF THE API KARANCK OF THK EFFLUENTS OF THE RESPECTIVE SYSTEMS, Expicssed in Decrees of Clearness. Warren ~ . Jewell ? . . 2 October Warren 3 Jewell November Wesiern Gravity Warren 4 lewell December . . 2 Warren ] Sc/> Jewell 3 5 Western Gravity * Western Pressure Warren 2 . . February Jewell Western Gravity . . . 3 4 7 2 4 3 2 5 2 2 3 4 4 2 2 2 5 2 I 3 3 i 2 3 2 . . I . . T ? T 3 > ? 2 3 Jewell 3 q o o 1 ^ ? 2 3 March , Western Pressure . Warren 2 .. 2 * 2 ; I 2 2 I 1 2 2 2 3 4 4 ? 4 2 2 3 2 2 2 4 2 2 lewell . . 3 , 2 2 I April . Warren 7 7 ? T i ? ? I ? 2 2 Jew-11 ? 7 4 I May Western Pressure . Warren 2 2 4 2 3 2 2 1 2 3 2 2 I I 2 2 2 5 2 lewell ; I 2 June Western Pressure . Warren 2 I 3 3 2 2 2 3 3 2 2 2 2 4 2 2 2 2 2 2 2 2 2 2 2 : 2 2 2 2 2 2 2 2 5 2 5 2 4 2 2 3 2 7 2 3 2 2 2 . . July a Western Pressure . 3 3 3 2 2 2 2 2 2 2 2 3 3 2 2 2 . . SUMMARY AND DISCUSSION OF DATA OF 1895-96. TABLE No. 2. 223 DAILY RESULTS OF THE DETERMINATION OF ORGANIC MATTER IN THE OHIO RIVER WATER, EXPRESSED IN PARTS PER MILLION OF NITROGEN AS ALHUMINOID AMMONIA, AND OF OXYGEN CONSUMED, RESPECTIVELY, TOGETHER WITH THE PERCENTAGES OF REMOVAL OF ORGANIC MATTER BY THE RESPECTIVE SYSTEMS OF PURIFICATION. Nitrogen as Wa in River Percentape For ieinova] b) the Respect ve Systems Total Oxygen Percent Oxy K en Co alter Exprc ssed as Date. Total. Percentage which was in River Water. Warren. Jewell. Western Western Suasion Grav.ty. 66 2$ ,, ,- l( T 2SS ifi ,, 2 6 2 6 ifi " 26 71 6 " 28 216 ^8 " * ^6 26 Nov T df) 62 48 20 6 " 8 12 22 I 8 I " 1 6 28 18 18 26 2 8 " 26 .246 ^ 8 27 232 26 11 28 36 28 Dec i .216 ^6 2 6 2 8 18 .216 2.6 6 48 2 8 " 7 8 .234 46 63 67 .184 56 " ii 65 " T 3 .158 24 50 48 224 WATER PURIFICATION AT LOUISVILLE. TABLE No. 2. Continued. Wa in River I crcentape of Organic r Mailer Expr m of Album essed as Nitr o"en in the" Oxygen Sysums if Organic N Oxygen Co utter Exprt ssecl as Date. Percenta R e in River Suspension Gravity. Gravity. 1895 Dec. 14 36 3- 5 28 " 16 188 " 17 48 " 18 .228 60 4 s " 26 " 28 Av. 27-30 .388 6l 75 79 79 79 6.0 70 77 77 73 1896 2 " 3 1.187 1.187 83 S3 89 89 89 93 93 92 92 12.3 ii. 8 ......... 88 87 84 84 89 87 89 86 90 86 Av. 4& 6 .6=3 76 82 88 88 7 8 83 82 Ian. 7 6 5 83 Av. 8,9,10 Jan. it \ -423 ( .261 48 45 74 72 68 63 73 65 72 6 1 5-4 5-8 5-3 65 S3 Si 67 72 72 65 69 75 67 72 74 Av. 14, 15 Jan. 16 ] -209 12 86 62 62 54 4-1 4-1 6 1 73 7 6 63 76 61 66 58 " T 7 " 18 6 1 " 22 135 7 56 50 47 47 3-1 61 55 45 52 " 24 " 25 " 20 Av. 27, 28 j .369 70 Si Si 81 70 74 76 5-5 6.6 78 77 82 69 64 78 68 75 74 " 3 " 3 1 Feb. i .365 8 Sq 6.6 85 82 So " 2 " 3 67 8-3 go 83 83 83 85 82 82 84 81 81 5 6 7 " 8 439 .417 593 Si 81 87 Si 86 88 93 84 86 93 85 83 90 84 82 90 8.4 8.2 ii. 4 82 88 91 86 84 90 85 83 84 87 83 So 87 " 9 " 10 577 88 94 92 94 93 12.8 fi Q 91 88 84 91 88 91 87 76 88 88 ! T 3 .215 60 78 33 80 So 8 i 76 78 4-6 7 6 78 85 70 74 78 " 15 353 -8 91 84 87 . 85 8.7 87 84 87 85 SUMMARY AND DISCUSSJON OF DATA Ol<~ 1895-96. 225 TABLE No. 2. Continued. Nitrogen a raon Albuminoi ter. Percentage Ki Removal bj nn of Albui the Respei linoid A mm live System Total Ox v gen I , n nil ... Krm v Oxy K cn C by the Res ^ective ssed as Date. Total. rcrccntiiK iSuspTnsio Warren. Jewell. Western Gravity. Western Pressure. in River Water. Warren. Jewell. Western Gravity. Western 1896 Feb. 16 " I? " 18 .292 .366 75 86 95 82 4 84 7-5 88 S3 85 88 7 " 19 " 20 .484 .232 88 95 84 94 86 92 84 II. 2 6 -\ 92 86 91 8q 92 79 " 21 22 .256 .288 81 76 92 87 86 82 85 87 7.2 6 o 90 87 90 So S7 86 " 23 " 24 " 25 .162 152 69 64 79 71 74 67 78 78 4-3 3 s 84 82 74 63 79 72 " 26 " 27 .146 .200 71 70 82 73 78 78 75 3-3 79 82 64 82 76 76 " 28 " 29 Mar. I .102 .088 35 43 57 73 51 55 i* j 68 3-0 2.8 70 64 57 54 ! I " 2 3 .084 .118 50 53 64 68 5" 68 52 52 2.6 73 Si 42 63 65 62 4 5 .088 . IOO 57 48 61 74 50 60 6 4 66 2-5 72 62 44 60 64 " 6 7 8 .184 . IO2 36 46 76 78 45 7i i f 1.8 2.5 76 56 68 i " i" 9 10 .076 .OgO 53 51 68 62 68 69 63 66 2.3 61 70 68 65 7" " n " 12 .096 .116 46 38 73 74 65 69 65 j 4 8 i i. 9 79 63 6^ 6 3 * " 3 M " 15 .094 .108 49 54 68 65 64 67 {64 I" 2. I 2-3 67 65 52 65 { j, " 16 " 7 . no 130 42 58 75 77 60 68 53 67 ( 2 . 5 72 60 60 64 ( i- 19 " 20 " 21 .132 .238 .700 1 .046 50 73 90 74 79 95 74 77 93 i" 95 \ I* 3-0 5-0 14.2 17 R 73 82 94 73 So 89 { 91 i" i* " 22 " 2 3 .580 86 88 " 2 4 .476 86 9 2 " 25 .414 86 8 i " 26 .462 85 89 8 Q So " 27 344 78 87 83 86 So Sj 88 " 28 454 84 88 80 8 3 80 80 " 29 374 79 S5 82 86 88 88 88 " 3 .852 87 95 " 3 1 .932 April I 1 .032 96 06 TC: 8 2 .960 93 97 96 15 6 06 3 1.080 93 96 96 18 4 06 06 4 .620 84 94 5 6 .500 83 90 8 6 7 .380 81 90 " 8 .370 75 85 85 6 7 88 88 9 .380 79 88 88 " 10 .312 74 85 82 6 -* 86 84 " n .286 72 78 82 6 5 83 86 " 12 " T 3 .270 68 80 14 .166 53 66 60 " 15 .208 62 69 78 " 16 .164 50 66 T 6 6n " 17 172 51 63 18 .206 52 64 " T 9 WATER PURIFICATION AT LOUISVILLE. TABLE No. 2. Continued. Nitrogen as Wa in River ter. Percentage of Organ! the F Removal bv Matter Kx| jrm of Albu the Respect ressed as N Systems Total Oxygen Percent a Systems ge Removal of Organic isOxygenC by the Resp Matter Expr ective essed Date. Total. Percentage which was Suspension Warren. Jewell. Western Gravity. Western Pressure. in River Water. Warren. Jewell. Western Gravuy. Western Pressure 1896 198 81 188 66 66 go 56 61 81 118 65 6s i 16 2 8 *8 " 26 68 68 " 28 67 6a 64 58 67 67 65 60 1 68 48 67 68 60 188 60 67 i 6 65 i 6 60 6 .188 68 69 68 68 .164 61 63 " 8 62 fie 60 63 58 68 .286 66 84 Si 7,S go 68 2-6 65 72 fil 66 65 76 7S 62 182 60 76 75 i 6 67 " if) 60 60 68 75 " 18 cS 52 " 9 .158 59 56 ,s 61 35 228 62 73 68 68 63 58 6S fie i $ 66 55 " 22 .228 61 70 66 " 23 6c* 68* " 25 .166 58 cfi " 26 .158 58 58 ( ( " 27 .186 58 69 66 \n 67 67 l?3 " 28 .566 79 86 85 8 4 83 83 ( " 2 9 .602 78 87 87 ST g7 81 " 3 .502 76 86 88 7 8 8=1 37 " 3 1 June i .446 76 84 86f 83 86f .576 89 87 8 o 88 86 84 " 3 67 86 S 5 ^i c R 8d 84 84 " 4 .324 69 80 Si Si 83 83 83 " 5 .274 54 74 ( 76 81 1 , 6 48 74 67 1 ; " 7 8 " 9 .404 So 83 85 ( 6 S Si 84 ( " 10 .340 7 2 Si 82 J82 6 i 80 84 >7 " it .282 68 r 76 ( " 12 .250 65 70 68 76 " 3 .232 <J3 7 2 76 1 67 7<4 " 14 " 15 364 74 79 So =; 6 " 16 .248 64 69 67 " 7 .260 65 69 i 6 65 73^ " 18 .288 67 72 76 64 78 " T 9 .392 72 83 81 ( 7 8 1 " 20 .268 66 74 75 J8, 4-3 f)0 65 j 73 Average for 22, 23, and 25. f Average May 30 and June I. Average 16, 17, and 18. | Average 12, 13, and 15. SUMMARY AND DISCUSSION OF DATA OF 1895-96. TAULE No. 2. Concluded. 22-J Date. Nitrogen as Albuminoid Water. Percentage Removal by the Respective Systems of Organic Matter Kxpressed as Nitrogen in the Form of Albuminoid Ammonia. Total Consumed in River Water. Percentage Kemova by the Respective Systems of Organic Matt, r Kxpressed as Oxygen Consumed. Total. Percentage Suspension Gr.mty. I ressure. ^ 1896 June 21 " 22 " 23 " 24 " 25 26 " 27 " 28 " 29 " 30 July I 2 " 3 4 5 " 6 7 " 8 9 10 " ii " 12 " 13 M " 15 " 16 " 17 " 18 " 19 20 " 21 " 22 " 23 " 24 " 25 " 26 " 27 " 28 " 2 9 " 3 " 3i .222 .224 .460 .326 304 374 53 55 77 65 62 67 67 65 So 67 62 79 4.0 3-7 5-1 6 o 63 70 73 67 f 5 71 (o i i: 78 So 1 80 J79 5-1 5-5 7i 78 442 .356 .628 .964 .640 71 72 82 89 83 84 82 9 93 90 1 7-4 5.6 10.7 15.0 12. I 86 So 9 9 1 90 79 88 92 89 73 88 9i 89 86* 83*. 92 89 90 90 .258 .398 432 378 .346 .224 57 74 75 72 63 52 69 83 83 80 79 73 74 83 i" 81 73 1" {,0 4.8 7-o 6.8 6.6 5.8 4-,4 73 84 82 83 79 77 81 84 ! 88 82 i " i" "P" J76 .382 .310 .420 -378 394 504 72 66 70 76 79 81 80 77 81 84 1" 84 82 84 84 {,, i" i* 7.0 6.6 8.3 7-0 7.0 8.7 S3 82 84 86 i < 87 86 89 87 (- t .; j 86 | 86 i -318 494 .824 1 .360 2.400 i . 320 70 78 89 90 95 91 81 85 90 91 96 93 84 87 92 94 97 96 jji i 6. s 7-7 14.9 24.8 35.8 22.6 82 84 90 93 96 92 86 87 92 95 97 95 j ,5 | !" {95 i. 200 I. 120 .880 I.2OO .470 89 89 85 90 75 93 94 92 !" 95 94 92 | 21.7 17.8 T 9-7 23.4 13-6 93 92 9i ),3 95 93 91 1" 84 9 o Average June 30, July I. 228 WATER PURIFICATION AT LOUISVILLE. TABLE No. 3. AVERAGE DAILY NUMBER OF BACTERIA PER CUBIC CENTIMETER, IN THE OHIO RIVER WATER AND IN THE SEVERAL EFFLUENTS, TOGETHER WITH THE AVERAGE BACTERIAL EFFICIENCY OF THE RESPECTIVE SYSTEMS OF PURIFICATION. SUMMARY AND DISCUSSION OF DATA OF 1895-96. TABLE No. 3. Continued. 229 1895 Dec. 27 Jan. Mar. Bacteria pel ( ubi ( , n i meter. Eff uents of the R jspective Syste ms. River Water. \Vcst. in Western Western W.- i 12 OOO 35 7oo 779 i 169 410 813 501 947 358 897 93-5 76.7 96.6 97-7 95-8 97-3 97.0 97-5 12 OOO 13 ooo 328 375 3" 233 529 253 409 263 97-3 97.1 97-4 98.2 95-6 98.1 96.6 97-9 10 700 14300 8600 797 1 66 151 261 343 390 245 325 Si 258 169 TOO 92.6 98.8 98.2 97.6 97.6 95-5 97-7 97-7 99. i 97.6 98.8 98.8 368 9" 6 98 7 97 S 98 i 4 loo 59 98 6 96 . 3 I goo I 800 2 500 28 94 36 229 202 152 100 108 94 78 151 153 98-5 94.8 98.6 87.9 88.8 93-9 94-7 94-u 96.2 95-9 91.6 93-9 38 98 o 89 8 800 3000 3 200 6 500 73< 25 50 73 98 60 156 156 181 262 189 6 1 128 130 159 228 88 M7 122 219 244 96.9 98.3 97-7 98.5 99.2 80.5 94.8 94-3 96 . o 97-4 92.4 95-7 95-9 97.6 96.9 89.0 95-1 96.2 96.6 96.6 6 400 3000 2 6oO 119 5i 34 229 170 104 219 174 192 68 218 194 120 T76 98. i 98.3 98.7 96.4 94-3 96 . o 96.6 94-2 92.6 qS.6 96.6 93-5 95-4 96 3 3 600 06 06 i 7 200 60 "7 102 136 99-2 98.4 98.6 98.1 10600 14700 1 8 200 J56 470 525 945 I 533 651 i 56 747 876 98.5 96.8 91.1 89.6 93-9 89.2 93.0 94-0 23 400 72 820 14300 21 3OO 57 56 860 549 475 188 99.6 94.0 96.2 96.7 62 200 197 55 ooo 8ie 98 5 71 ooo 30 800 55000 29 800 583 220 362 365 643 I 563 639 i 395 548 805 i 200 669 686 I 4 f>i 967 7M 99.2 99-3 99-3 98.8 99.1 94-9 98.8 95-3 99-2 97-4 97.8 97-8 99-o 95-3 98.2 97.6 14400 19 800 279 740 37i 527 180 98.1 94-9 97-4 9 6 -3 98 i 28 ooo 98 5 98.7 14 800 ii goo 10 800 430 589 139 549 848 763 453 340 950 409 216 850 97-1 95-1 98.7 96.9 92-9 92.9 96.9 97-1 91.2 97-2 98.2 92.1 21 800 14400 20 700 17400 10 700 20 ooo 107 80 45 210 7 6 I 079 I 581 919 341 318 789 647 467 291 189 638 710 547 532 169 99-5 99.4 99-3 98.8 99-3 95-1 89.1 95-6 gS.o 97.0 96.4 95-5 97-7 98-3 98.2 97-1 95-1 97-4 96.9 98-4 15 ooo 16 200 25 200 4 loo 4500 14 800 73 71 202 65 156 178 821 M5I 1941 768 I 575 I 601 236 I 012 6 54 476 353 f>5i 497 951 574 676 533 903 99-5 99.6 99.2 98.4 96-5 98.8 94-5 91.0 92-3 81.3 65.0 89.2 98.4 93-8 97-4 88.4 92.2 95.6 96.7 94.1 97-7 83.5 88.2 93-9 44600 33400 29 800 18000 12 2OO 10 500 367 302 548 3/1 207 1 80 2 148 1324 851 718 299 I 4 6 i 944 i 062 693 797 343 332 2 251 798 6ia 615 326 470 99-2 99.1 98.2 97-9 98.3 98.3 95-2 96.0 97-2 96.0 97.6 98.6 95- ( 96.8 97-7 95-6 97-2 96.8 95 o 97.6 97-9 96.6 97-3 95-5 23 WATER PURIFICATION AT LOUISVILLE. TABLE No. 3. Continued. Date. Bacteria per Culiic Centimeter. Bacteria! Efficiency of the Respective Systems. River Water. Effluents nf the Respective Systems. Warren. Jewel.. Western Western Gravity. Pressure. Warren. Jewell. Western Gravity. Western Pressure. IS 9 6 Mar. 8 9 10 " IT L" I2 " 13 M 15 " 1 6 " 17 IS " 19 " 20 " 23 1 ~ 4 -5 " 26 " 2 " 29 3" " 3i April i 3 4 5 6 8 9 IU " 1 1 " 13 M " 5 " if> " T 7 " 1 8 " 19 20 " 21 " 23 24 25 26 " 27 " 28 " 29 " 30 May I " 3 4 5 6 7 8 9 [0 " II 12 " 13 14 15 " If) " 17 18 9 14 ooo II 500 7700 I 300 I OOO 2 IOO 190 301 156 293 161 -33 137 121 251 547 203 409 2OO I SO If)2 233 140 239 221 If) 9 372 162 !59 98.6 99-O 97.4 98.8 9/.I 98.4 98 . 6 97.8 97-3 95-0 98.7 98.3 97.1 98.3 97-7 98.6 97-9 98.8 98.3 98.1 97.8 96.7 98.5 98.7 6 600 6 300 9 700 4400 46 700 57 200 76 31 49 122 137 2f)0 172 86 35 7" 850 928 252 3<)2 520 606 733 349 37" 441 Soo 55" I 087 99-5 99.8 99.8 99.6 99-7 99-5 99.0 99-5 99-8 98 .2 98.4 98.5 98. 1 97-4 98.2 98.4 97-9 97-7 97.8 97-7 98. 8 98.1 30 500 37300 46 ooo 47 900 31 900 34 i"" 49200 25 7<->o 26 700 39 600 27 500 31 oco 27 ooo If, 4 4 103 136 141 in 123 125 3^9 263 155 82 116 844 321 182 53i 466 ,64 379 679 S45 638 223 20 1 7i 935 368 404 256 294 537 2 221 I 712 I 62S I 023 755 940 185 99-5 99 -S 99.8 99-7 99.6 99-7 99.8 99-5 98.6 99-3 99-4 99-7 99 ( 97-2 99 1 99.6 98.9 98.5 99-5 99-2 97-4 96.8 98.4 99.2 99 3 99 7 96.9 99.0 99.1 99-5 99. 1 98.4 95-5 93-3 93-9 97-4 97-3 97-0 99-3 1 8 700 18 500 13 Soo 21 OOO 7300 13 ooo 34 40 105 nS 42 25 3i 64 129 164 40 4i 9 1 76 99.8 99. S 99.8 99-7 99-5 99.6 99 4 99.8 96-5 99-7 9 600 7 5oo 3 7 i 700 3 ooo 5 9 28 QO 6 , 33 98 3 | 4 Soo 4 ooo 5 5"0 4400 7 7 8 300 34 82 1 66 no 348 309 126 16 23 46 48 99-3 98.0 97.0 97-5 95-5 96.3 99-9 99-7 99-7 99-5 99.4 99 4 6 500 6 100 7 loo 4 100 5400 7 loo 401 605 I3S 94 104 157 62 1 60 64 42 6 1 78 93-8 90. i 98.1 97-7 98.1 97-8 99.0 97-4 99.1 99.0 98.9 98.9 i 7 ooo 6 ooo 5 ooo 5 200 4000 5 900 55 57 S7 76 50 49 52 45 48 34 17 39 99.2 99.1 98-3 9S.5- 98.8 99-2 99-3 99-2 99.0 99-3 99.6 99-3 145 20 1 227 97-2 95.0 96.2 7300 5 9"0 3 600 6 500 6 400 7 300 44 61 12 37 25 10 32 16 10 37 27 17 176 153 167 in i 7 99-4 99.0 99-7 99.4 99.6 99.9 99.6 99-7 99-7 99.4 99.6 99-8 96.8 97.0 95-8 97-4 98-3 98.4 6 600 4300 1 20 81 75 93 701 324 98.2 98. i 98.9 97.8 89.4 92-5 SUMMARY AND DISCUSSION OF DATA OF 1895-96. TABLE No. 3. Concluded. 23 Dale. Mucteria per Cubic Centimeter. Bacterial Efficiency of the Respective Systems. River Water. KfHuents of Ihe Respective Systems Western Western Jewe.1 S Western lS(/> May 20 " 21 " 22 " 23 " 24 " 25 " 26 " 27 " 2S " 29 " 30 " 31 June I 2 " 3 4 " 5 " 6 " 7 S " 9 " 10 " it " 12 " 13 " >4 " 15 16 " "7 " 18 " 19 " 20 " 21 " 22 " 23 " 24 " 25 " 26 " 27 " 28 " 2g " 30 July I 2 3 4 5 6 " 7 " 8 " 9 10 " II " 12 " 13 M 15 16 " 17 " IS 9 " 20 " 21 " 22 " 23 " 24 " 25 " 26 " 27 " 28 " 29 " 30 " 31 4800 6 100 5 100 6 too 66 70 5" 73 73 70 2 s 184 124 60 59 98.6 98.9 99-0 98.4 J - 98.8 08 6 96.2 98.0 98.8 99.0 I 900 I 800 4 loo 23400 26 too 19 700 4 39 130 IO2 179 92 59 50 S4 194 245 322 94 82 119 261 150 57 97.8 97.8 96.8 99-6 99-3 99-5 Of) 95.1 95-4 97-1 98.9 99-4 99-7 gg O Q9. 2 99. 1 f.Q | 18 800 15 5oo 13 800 8 500 6 400 4900 6 1 101 45 36 28 25 165 93 60 93 36 44 99-7 99-3 99-7 99.6 99.6 99-5 99- 1 99-4 99.6 98.9 99-4 99.1 208 53 45 28 IS 98.7 99.6 99-5 99.6 99.6 ii 300 10 700 6600 6 too 13400 165 49 42 40 188 104 II TO II 4" 121 54 1 1 14 93 98.5 99-5 99-4 99-3 98.6 99-2 99-9 99.8 98.9 99-5 99.8 99.8 99-3 99-7 13500 8400 II OOO 10600 18000 10 500 132 3<> 74 81 73 86 16 7 27 33 30 161 30 73 99.0 99- 6 99-3 99.2 99.6 99.2 99-9 99-9 99-8 99. f> 99-7 08 u 20 53 Si 99.8 99-7 99-2 99.8 99-3 7 7< 8000 S 300 7 500 6 (xx> 10 800 So 141 234 52 86 118 453 550 37 56 (,8.9 98.9 99-3 ( j8 5 2lS 3"5 51 97.2 99-3 94-5 97-4 95-9 99.2 97-7 13300 10900 37 8 5 43 33 99-7 99-7 99-7 99-9 ! 24 200 12 OOO 95 80 72 16 83 42 99.6 99-3 99-7 99-9 99-7 99.6 7400 5 BOO 6 700 9 200 IOOOO 9 6<x> 64 S2 69 53 71 40 353 480 99 10 40 24 48 3" 21 60 99- 1 93.5 99.0 99-4 99-3 99-6 95-1 91-3 98.5 99-9 99.6 99-7 99-3 99-4 99-7 99-4 no 208 98.9 97.8 7 7<x> IO IOO 8 300 5 5<*> 5 /oo 9 9<x> 64 33 30 3" ii 23 44 6 8 II 5 34 78 25 1 1 18 99-2 99-7 99-6 99-4 99-8 99-8 99-4 99-9 99-9 99. S 99-9 99-7 99.0 98.2 99-9 99-7 33 48 99-4 99-5 7 ooo 17 100 33800 27 100 31 O xj 17 300 24 "3 752 597 I 327 433 9 47 251 103 71 54 22 204 623 99-7 99-3 97-8 97-8 95-7 97-5 99-9 99-7 99-3 99.6 99.8 99-7 99-7 98.8 98.2 234 284 99-2 98.4 I 7 8<X) 24500 9500 1 2 OOO 6800 306 60 52 12 33 19 6 15 151 156 , 137 98.3 99-8 99-5 99-9 99-9 T) 9 .r, 99. S 99-9 99-2 99-4 98.6 f 2 3 2 WATER PURIFICATION AT LOUISVILLE. TABLE SUMMARIES OF Warren NumlxT <->f period 1 2 3 4 5 6 7 8 ( T);lte )ct.2I, 95 Nov. 25 Dec. 26 an. 1 3. 96 Jan. 27 Feb. 6 Feb. 13 Feb. 27 X " " | Hour O.OOA.M. 9-45 A.M. I.I7 A.M. 2.75 P.M. I.O7 A.M. 2. 36 P. M. 1. 1 I P. M. 5 P.M. T , , Date Kndcd , , fov.25, 95 Dec. 26 11.13, 96 Jan. 27 Feb. 6 Feb. 13 Feb. 27 Mar. 20 ( Hour 9.45 A.M. I.I7A.M. 2.15 P.M. [1. 07 A.M. 2.36 P.M. 2. II P.M. .. 5O P.M. l.OO A.M. Runs inch ded in period II4 I S ^1 325O 51-65 66-78 79-84 85-100 IOI-I35 i Maximum 12. OS 23.32 12.33 7.18 8.70 6.78 8.35 6.07 Minimum 4.83 6. 5 S 3.60 3.62 3.18 3-97 3-45 2-43 Average 8.87 11.22 5.98 5-10 5-82 5.27 5.72 4.32 ( Maximum 11.83 22.87 12.08 6.33 8.03 5.80 8.07 5.78 Period of service. (I lours. )..} Minimum 4.58 6.03 3.28 3-07 2.62 3-35 2.87 2.15 ( Average 8.58 IO.72 5.50 4.50 5-15 4-59 5-25 3-97 j Maximum 0.42 0.78 1.03 0.73 0.92 0.98 o.6S o.55 Period of wash. (Hours.).. . -] Minimum 0.17 0.15 0.25 o.53 0.57 0.58 0.28 0.22 ( Average 0.28 O.5O 0.48 0.60 0.67 o.6S 0.47 0-35 Ouantitv of applied water. \ ^i " 1 -"""" 1 (Cul)ic feet.) 16 970 7 599 32 IO2 8791 16480 3 597 7704 3285 Soil 2771 6082 3683 8719 3842 8401 ( Average n 816 14944 6 685 4 749 5 3<>7 4 788 (JIIQ 5 379 ( L )uantitv of filtered water. \ M lx . lmum (Cubic feet i 1 6 640 7359 31 048 8 126 15 853 3 139 6472 2 798 7487 2092 3258 8491 3614 8 182 2885 ( Average ii 457 14 OO7 6215 4 if>5 4 745 4 286 5 735 5171 Quantity of wash water. \ ^i- ,^ , ", 1 741 283 54 297 723 271 479 244 586 432 572 437 5f>i 295 i 286 421 ( Cubicfcet -) (Average 360 376 374 298 497 485 446 509 Quantity of filtered waste \ \^ o 469 o 453 o 440 198 466 219 407 234 493 202 O water. (Cubic feet.) | Averagl . o 221 M4 258 253 278 M5 IO Quantity of unfdtered waste \ J^*-^ water (Cubic feet ) . . ) 400 250 450 I So 300 i So 400 i So 250 i So 250 180 300 1 80 300 If, 7 181 312 2O6 198 218 197 2 2O 219 Percentage which wash and ( Maximum IO 10 26 31 24 26 32 waste water was of applied i Minimum 4 2 4 13 12 17 9 S water . . ( Average fa ii 16 IS 21 13 13 26.8 25.5 23.6 17. i If). 8 16.3 23.2 25.0 Actual Cubic feet per min. \ Minimum 19.8 19.2 14.6 4-7 13.3 I5-I 1 6. 8 17.6 rate of ( Average 22.3 21. 8 18.9 15.4 15.4 IS- f lS.2 21-7 filtra- ( Maximun tion Mil -8 als P eracre ) Minimum ]>er 24 hours. . . | Aye c 147 109 122 155 I if) 132 144 89 114 104 89 93 IO2 Si 93 99 92 95 MI 102 no 152 107 131 Ave. net rate j Cubic feet per minute 2O.9 20.0 16.6 13.0 12.5 12. 15.5 IS. I of filtration, j Mil. gals, per acre per 24 hr "5 121 101 79 76 73 94 no Net quantity of filtered water ( Maximun 68 131 66 27 29 21 3i 32 per run, in mil. gals, per J Minimum 30 33 12 n 8 12 12 IO acre ( Average 46 59 2f) 17 18 16 22 20 Average estimated suspended ( Maximun 22 5" 870 IOO 461 967 486 210 solids in river water. (Parts ! Minimum 4 15 IOO 20 200 550 I 3 6 22 per million ) . ( Average 13 27 23O 410 320 73 290 40 Grains of applied sulphate of ( Maximun 1.34 1-77 5-25 6.08 6.83 4.87 6.90 5.40 alumina per gallon of ap- -! Minimum 0.48 0.75 I.9I 3.05 2.80 2.20 2.19 I.9I plied water. f Average 0.84 1.17 3-79 3.61 3-99 3-7 3.66 3-3^ Average grains of applied sulphate of alu mina per gallon of net filtered water . . 0.89 1.25 4.26 4.31 4.87 4.69 4.21 3-S7 Degree of clearness of fil- J ^. . tercd water.. . . . 1 , 2 I 4 I 5 2 2 2 3 i 3 i 4 I 3 i ( Average , . . . . ( Maximun Bacteria per cubic centimeter J -, . 1.4 37 1 06 2.5 9 200 2 COO 3-3 27700 I 8()0 2.O 10 300 Soo 2.O 8 1 ooo IO f)OO 2.O 55 ooo 14400 2. 2 21 SOO 4 IOO 1-9 44 ooo 4500 in liver water } Average 168 4 7oo IO IOO 4 600 34 400 33 800 1 6 ooo 17 900 Average maximum number of bacteria pe cubic centimeter in filtered water 96 479 559 84 474 432 238 365 Average minimum number of bacteria pe - cubic centimeter in filtered water 20 183 157 46 136 252 55 59 I Maximum Bacteria per cubic centimeter \ Mi , limum in filtered water. . . . ) , 99 12 So 928 IO 127 25 57" 54 534 238 219 735 29 ( Average 42 349 328 72 290 350 121 224 ( Maximun 91.7 98.2 99.9 99.2 99.8 99-4 99-7 99.8 Average Bacterial efficiency... < Minimum 60.9 86. 2 91.3 94.2 96.9 98.3 98.5 95-0 ( Average 75-0 92.6 96.8 98.4 99-1 99.0 99-2 98.8 SUMMARY AND DISCUSSION OF DATA OF 1895-96. 233 NO. 4. RESULTS BY PERIODS. System. 9 10 11 12 13 14 15 16 17 18 19 20 Mar. 20 Mar. 23 Mar. 30 Apr. 7 Apr. 27 May 18 May 28 Junes June 9 June 30 July 6 July 21 9.00 A.M. 9. 30 A.M. 4.23 P.M. 12. 37 P.M. g.OOA.M. II.I4A.M. 2.17 P.M. 3-30 P.M. 9-34 A.M. 3.48 P.M. 2.21 P.M. 4.41 P.M. Mar. 23 Mar. 30 Apr. 7 Apr. 27 May 18 May 28 June 3 June g June 30 July 6 July 21 July 31 9. 30 A.M. 4.23 P.M. 12.37 P.M. g.OOA.M. 11.14 A.M. 2.17 P.M. 3- 30 P.M. 9- 34 A.M. 3 48 P.M. 2.21 P.M. 4.41 P.M. g.ooA.M. 136-141 142-178 179-195 196-205 206-240 241-253 254-275 276-284 285-302 303-310 311-329 330-346 ;.- - 5.60 8. 3 2 1 1. 80 15-38 21.40 11.23 10.88 10 32 5-23 7.63 7-25 2.48 2.42 2.30 2.13 7-45 9.20 2.08 5.92 4-43 2.30 3-73 1.22 2.92 4.02 3-95 8.57 11.27 15.15 4-47 7.91 6.77 3-33 5.46 3 71 2-77 5-28 8.00 11.40 14.92 20.90 10.67 10.40 9.87 4.80 7.22 6.90 1.98 2.08 2.03 1.83 6.88 8.78 1.67 5.52 4.07 1.48 3-33 0.65 2-53 3-7" 3.65 8.20 10.80 14.67 3-97 7-43 6.34 2.90 5.08 3.28 0.50 o.55 o.35 0.47 0.68 o. 60 o. 60 0.58 0.62 0.57 0.47 0.55 0.30 0.25 0.27 0.28 0-35 0-35 0.35 0.40 0-33 o-37 0.33 0-35 "37 0.32 0.30 o.37 0.47 0.48 0.50 0.48 0.43 0.43 0.38 0-43 3426 5889 8515 15 366 18 164 21 296 12 607 10 865 13 326 6496 10629 9273 2569 2 256 2 266 2 IlS S 240 IO060 2085 7 154 5 5i8 2337 4451 I 471 2991 4 112 4016 10015 13394 16044 5049 8737 8617 4188 6876 3981 3248 5 74 8086 15378 18 283 20485 12 786 10927 13473 6 594 9826 9382 2 129 2 O28 2 117 I 959 8343 8622 2 131 7292 5549 2 030 4500 6l 7 2737 3952 3857 10016 13485 15540 4672 8726 8 590 3966 6890 3871 517 556 521 650 I 073 1389 884 728 683 670 646 878 429 353 4OO 421 4f 3 611 557 617 506 473 515 424 468 440 450 5i8 684 985 707 672 569 5f>3 572 622 o o o o o o O o o O o o o o b o o o o o o -131 228 430 176 404 150 298 44 61 386 44 412 158 167 158 44 35 44 35 44 35 44 26 44 239 177 189 118 61 70 68 44 44 127 43 158 37 28 27 28 it 9 33 10 II 41 15 79 18 II 8 4 4 4 6 6 4 10 6 8 24 15 1 6 6 6 7 1 6 8 7 17 9 19 19.5 18.9 18.4 25.0 21.8 19.7 23.2 22.7 23.0 23-3 23-3 23.1 17-3 16.2 1 6. 8 17.4 19.7 16.1 15.6 16.4 21. 22.2 22.1 13-1 17.9 17.8 17.6 20. 6 20.8 17-7 19.7 19.6 22.6 22-7 22.6 19.7 118 "4 in 152 132 119 141 137 140 142 142 141 105 98 102 105 119 98 95 99 127 134 134 So 108 108 107 125 126 107 119 119 137 137 137 119 13.0 12.7 14.2 18.3 18.5 17.1 15.9 16.9 19.7 17-4 ig.I 14-3 79 77 86 ill 112 104 96 102 119 105 116 87 12 22 33 61 73 84 49 43 53 24 34 35 6 7 7 6 31 35 6 27 20 7 16 5 10 15 M 39 52 52 18 31 33 15 26 13 I 276 660 I 131 370 1 80 200 829 459 582 1674 637 3347 993 338 400 70 57 38 540 1 60 210 720 190 i 050 I 130 450 800 1 80 100 90 680 290 295 I 090 440 i 740 6.60 9.08 6.72 1.92 2.68 1. 80 5.90 5-33 4.67 6.07 5-75 9.62 5-21 2. 7 8 3-25 o.73 0.49 0.83 3.16 2.32 I.IO 3-37 2.29 3-36 5.97 4-43 5.<>7 1.33 1.41 1-33 4-52 4.03 2.64 4.61 3.02 6.27 7.85 5-22 6.03 1.42 1.50 1.43 5.50 4.38 2.84 5-55 3-32 7-75 2 5 3 3 3 2 2 2 2 3 4 5 2 2 2 I I I 2 I I i 2 2 2.0 2.8 2.6 2.O 2.1 1.6 2.0 i-7 1-7 2.O 2.2 3-1 60 100 55400 42700 17 800 7400 10 500 28 700 14000 13900 24 2OO 17 loo 34 too 41 6OO 25 700 19 200 4OOO 3 700 i 500 8 200 4 900 6 500 9500 5500 9500 54600 40300 27700 9500 5 600 4500 20 IOO 8600 9500 15 ooo 8 900 22 IOO 3OQ 211 203 165 140 112 55 163 8g 604 I >4 78 III 62 73 29 23 45 3<> 225 329 245 635 295 i 075 116 293 87 298 105 95 i 475 93 5" 35 23 ii 32 40 22 20 4<) 17 13 181 120 M5 113 137 57 I Of. 4 1 06 70 47 354 99.8 99.9 99-8 99.8 99.8 99.4 99.8 998 99-7 99-7 99.8 99-9 99.4 95-4 99-7 96.3 85.5 97-4 98.7 99.8 95-3 99.0 98.6 94-7 99-7 99-7 99-5 98.8 97-5 98.7 99-5 99 5 98.9 99-5 99-5 98.4 234 WATER PURIFICATION AT LOUISVILLE. TABLE No SUMMARIES OF RESULTS Jewel Number of period 1 2 3 4 5 6 7 8 Feb. 29 ( Date Jieean. . . . -< , T 3ct.2i, 9 5 Nov. 21 j Dec. 24 an. 13, 96 Jan. 25 Feb. 6 Feb. ii j I lour 2. 03 P.M. i. 38 A.M.: 1.36 P.M. 9.44A.M. 2.00 P.M. 2.33 P.M. ) 45A.M. 9.58 A.M. Ended . . . . * P^ te Vov.21/95 Dec. 24 Jan. 13, 96 Jan. 25 Feb. 6 Feb. ii Feb. 29 Mar. 20 / Hour [1.38 A.M. 1.36 P.M. 9.44 A.M. 2.00 P.M. 2.33 P.M. 9.45 A.M. 9-58 A.M. 19.37 A.M. Runs included in period I 12 1320 21-29 3034 35-39 40-42 43-52 53-64 I Maximum 21. IO 28.83 16.30 23-55 14-58 9-63 20 55 24.17 Period of operation. (Hours.) ; Minimum 5-20 16.60 6-97 16.08 8-55 8.33 7-95 6-73 ( Average I3-I7 21-97 11.00 18.80 II-35 9.02 13-52 11.84 ( Maximum 20.78 26.10 15-65 23.12 14.23 9-35 20.07 23.80 Period of service. (Hours.). . -J Minimum 4.85 15-93 6.57 15-57 8.22 7.80 7-45 6.30 ( Average 12.82 21.40 10.38 18.38 10.87 8.45 12-95 11-54 ( Maximum O.5O 0-95 1.03 o 52 0.83 0.87 0.83 0.43 Period of wash. (Hours.). i Minimum 0.27 o-33 o 37 0.32 0-33 0.28 0.37 0.25 ( Average 0-35 o 57 0.62 0.42 0.48 0.57 0.57 0.30 Quantity of applied water. j Minimum (Cubic feet. ).. 1 . 36355 7896 19 889 40391 21 730 30524 21 089 7932 13673 354io 22 289 26 464 2O 6O7 II 074 15299 13 280 10 225 ii 509 22 575 10 224 16 523 35421 8 790 17404 ( Average Quantity of filtered water. i ;* lximum (Cubic feet.)... . 34677 7781 19363 38352 20488 29 701 21 335 7947 13643 34946 22 351 26442 20433 I07I5 15 O62 13 127 IO 220 II 319 22 950 10237 16770 35292 8 840 17634 ( Average Quantity of wash water. Maximum 1 (Cubic feet) j Minimum 731 259 519 856 534 676 i 025 467 665 45 S 432 443 617 383 447 538 49 517 810 505 579 844 442 559 ( Average ... / m i i Maximum Quantity <>i filtered waste \ ,,. . v .,, , . , . J. Minimum water. (Cubic feet ) . . o o 443 o 301 o 310 9 403 308 97 o ( Average o o 122 60 136 134 66 8 Quantity of unfiltered waste \ ,, . x . ,r> i c L \ { Minimum water. (Cubic teet.) ) . ( Average 235 32 104 214 32 IOO 32 32 32 32 32 32 32 3 2 32 32 32 32 32 32 214 31 Percentage which wash and j Maximum 8 3 II 4 9 9 7 IO waste water was of applied -1 Minimum 2 2 3 i 2 4 2 2 water ( Average 3 3 ft 2 4 6 4 3 (I Maximum 27.8 24.5 i 23.1 25.2 24.0 23-4 29.0 30.1 Cubic feet per min. -! Minimum 22.9 21.5 20. i 23.4 21.2 21.5 14.0 23-4 ( Average 25-3 23.1 21.9 24.0 23.1 23-3 21.6 25.5 .,., , - - . ( Maximum 112 99 93 I O2 97 95 118 122 L -< Minimum 93 87 Si 95 86 87 57 95 i per 24 hours. . . ( Average 102 93 89 97 93 94 87 103 Ave. net rate ( Cubic feet per minute 24.4 22.5 19-5 24.1 21.6 22.5 19.6 24.5 of filtration j Mil. gals, per acre per 24 hr. 99 91 79 97 87 91 79 99 Net quantity of filtered waste j M - X ^per run, mil. gals, per acrel . " " 99 20 54 105 59 82 57 20 36 97 60 73 56 28 41 35 27 30 62 27 44 96 22 47 Average estimated suspended ( Maximum 25 35 850 60 : 460 970 430 2IO solids in river water. (Parts -| Minimum 7 15 80 20 250 580 120 50 per million) . / Average 16 26 345 35 320 730 290 70 drains of applied sulphate ( Maximum 2 41 1.26 4.42 1. 12 2.32 2.39 4-82 1-55 of alumina per gallon of-] Minimum 0.40 0.48 1-25 0.83 i. 21 2.16 1.36 0.65 applied water f Average O.6S 0.87 2.35 0.96 1.72 2.25 2.78 i. 06 Average grains of applied sulphate of alumina per gallon of net filtered water. 0.70 0.90 2.50 0.98 1.80 2.40 2.90 i. ii Degree of clearness of filtered ( Mam water. ] Minimum 3 i 1.6 3 i 1.6 5 2 3-2 3 5 2 2 2.6 3.7 4 2 3-0 5 2 2.9 5 2 ( Average Bacteria per cubic centimeter ( Maximum in river water j Minimum 675 126 S 700 2 IOO 4400 27 6OO I 8OO 9700 6 800 54 loo i 300 10 900 3 800 22 900 41 200 14400 33 loo 21 800 4700 15 6OO 34400 9400 1 8 500 ( Average Average maximum number of bacteria pei cubic centimeter in filtered water 473 611 550 315 i 779 i 446 2002 779 Average minimum number of bacteria pci cubic centimeter in filtered water 56 115 162 105 912 418 245 124 Bacteria per cubic centimeter ( Maximum in filtered water. . . 1 Minimum 367 26 154 659 77 271 546 202 297 248 i (.4 186 2372 688 1088 i 346 740 960 I 600 504 I 015 1645 35 533 ( Average 83.1 97-3 98.2 96.3 98.0 98 i 97-4 99-6 Average bacterial efficiency. . < Minimum o.o 83-5 88.8 86.4 86.0 94-9 90.6 94-7 ( Average 59-o 93-8 96.9 95.1 95-3 97-1 93 5 97-1 SUMMARY AND DISCUSSION OF DATA OF 1895-96. 235 4. Continued. BY PERIODS. System. 9 10 11 12 13 14 15 16 17 18 19 20 Mar. 20 Mar. 21 Mar. 30 Apr. 7 Apr. 27 May 18 May 28 June 3 i June 9 July i July 6 July 22 9-37 A.M. 5.08 P.M. 10. 30A.M. 9.23 A.M. 9.25 A.M. 1. 12 P.M. II.O5 A.M. 2.20 P.M. 10.56 A.M. 3-55 P.M. 2.20 P.M. I0.24A.M. Mar. 21 Mar. 30 Apr. 7 Apr. 27 May 18 May 28 June 3 June 9 July I July 6 i July 22 July 30 5.08 P.M. 10.30 A.M. 9.23 A.M. 9.25 A.M. 1. 12 P.M. II.O5 A.M. 2.20 P.M. 10-56 A.M. 3.55 P.M. 2.20 P.M. 10.24 A.M. 11.37 A.M. 65-69 70-96 97-1 1 1 112-125 126-148! 149-158 159-184 185-204 205-234 235-238 239-256 257-272 3-5S 8.83 6. 02 16.92 21.00 34-77 8.55 8.68 13.68 7-OO 8.97 8.52 2-73 3-17 1.08 3-90 5-57 13-77 0-93 1. 08 1-43 1.8 5 1.98 0.72 3.20 5.65 3.86 9-79 16.90 21.63 2.65 3-55 5.02 4.25 5.55 3.48 3-i8 8.53 5-77 16.62 20.70 34-48 8.30 8.45 13.15 6.50 8.57 7.98 2.38 2.63 0.15 3-67 5-40 13-50 0.63 0.82 i.iS 1-43 1.28 0.52 2.85 5-35 3-53 9.52 16.62 21-35 2-43 3.30 4.70 3.85 5-13 3.10 0.43 0.70 0.93 0.32 0.48 0-35 0.30 0.35 0.77 0.50 0.38 0.60 0.27 0.18 0.23 0.20 0.17 O.2O 0.17 0.20 o. 20 0.23 0.27 o. 20 o.35 0.30 o-33 0.27 0.28 0.28 0.22 0.25 0.32 0.40 0.42 0.38 4437 12 iSl 7814 23 g2 3I3I9 51 286 12 533 12540 1 18448 9046 12 274 12274 3288 3903 776 5158 8623 21 OI2 I 032 2 637 j 2 380 2 062 2 316 858 3948 7427 5050 13641 25 565 33604 4 220 5627 7120 5553 7699 4024 4434 12513 8015 24483 32 264 51483 12 229 12 463 17699 8971 12396 9903 3350 3770 183 5286 8817 20795 I 041 I 806 2 Ogl i 860 i 888 855 3989 7659 5094 13927 26 225 34174 3904 5 443 7 it>5 5360 7497 4038 625 617 745 1073 874 732 799 Soi i 996 i 443 i 105 I 078 469 395 464 427 . 398 508 469 486 515 958 582 376 555 5 598 636 580 606 59 587 756 i 151 860 772 12 245 592 176 93 206 107 763 81 417 H3 O o o o o o o 2 17 75 o 9 12 15 19 78 43 73 10 114 274 O 214 187 198 O o o o o o o o 23 o 15 o o 7 o 20 37 18 22 142 8 6 3 47 20 3i 51 50 63 II 5 7 3 I I 6 5 5 16 6 10 15 7 3 5 2 2 15 II 12 21 12 20 23.8 25.2 24.9 25.3 27.2 29.6 34-5 35-2 37-0 23.8 25.0 30.9 22.6 22.7 20.3 23.2 25.1 24.4 20.5 24.5 22.4 21.6 23.1 iS.a 23.3 23.8 24.0 24.4 26.3 26.7 26.8 27.5 25.2 23.2 24.2 21.8 96 102 101 102 110 1 2O 140 M3 150 96 IOI 126 91 92 82 94 IOI 99 83 99 9i 87 93 74 94 96 97 99 1 06 108 108 in 102 94 98 88 17.5 20. 2 19.0 22.2 24.6 25-4 22.6 23.5 20.7 17.1 20. i 15-5 71 82 77 90 too 103 91 95 84 69 81 63 105 32 20 65 86 141 32 33 45 21 3i 24 7 8 O 13 22 57 2 5 5 3 3 I 9 19 12 3 6 70 9i 10 14 17 12 19 9 I 280 730 I 130 350 igO 130 830 460 590 I 700 690 3400 990 340 490 70 60 40 400 1 60 1 80 I OOO 190 I 200 I 130 450 850 1 60 IOO 80 640 300 340 I 310 450 I 860 5-40 5-23 6.32 2.23 3.66 1.84 6.92 7.70 7.61 7-45 7-17 12.62 2.88 2-39 3-1 0.96 I-I3 0.56 1.46 3-52 1.29 5-30 4.26 5.76 4.17 3-44 4-36 1-34 1.76 1.26 4.76 4.96 4.29 6.35 5.65 8.58 4.91 3-70 5.02 1.41 1. 80 1.29 5.6o 5.58 5.00 8.14 6.58 10.72 3 4 3 2 2 2 4 2 2 3 2 3 3 2 2 I I I I I I i I 2 3-o 2.6 2.6 1.4 1.2 1.4 2.7 1.8 1-3 i-7 1.6 1.2 60 100 53 ooo 42 700 I9OOO 8300 6 2OO 32 500 16 700 18 ooo 24 200 17 100 37300 41 600 25 900 19400 3 too 3700 I 800 8 200 4300 6 ooo 12 2OO 5 100 9500 55300 40 100 cgr 28 100 8000 5700 684 4300 i8 19 300 8 800 9300 219 18 ooo 98 9300 122 22 IOO 301 244 449 264 123 27 oc 1 D -j-i 68 29 26 1905 I 250 1495 - / I6 4 * J 1 60 Jj 9 2 475 107 655 127 409 745 440 103 32 8 12 38 22 7 5 4 3 6 968 4l6 522 50 48 73 IgO 47 9i 43 62 86 99-3 99-7 99-8 99-8 99-7 98.8 99.8 99-9 99-9 99-9 99.9 99-9 96.6 97-7 95.0 98.2 97.6 94-9 97.6 98.9 91.0 99-5 94-1 97.8 98-3 99.0 98.1 99-4 99.2 98.3 99.0 99-5 99-o 99.8 99 3 99.6 236 WATER PURIFICATION AT LOUISVILLE. TABLE SUMMARIES OF Western Gravity 1 2 3 4 5 6 7 8 Beean \ Date lee. 24, 95 _ 9.42A.M. fan. 14, 96 O.52 A.M. 2-16 23.13 2.38 8.C2 22.75 1.88 7.70 0.50 0.17 0.32 13036 I 564 4544 12 679 982 4267 616 162 419 242 37 101 400 5 169 48 5 15 II. 6 7-9 9.2 72 48 57 8.0 49 51 4 17 870 80 290 fan. 14, 96 IO.52 A.M. Jan. 27 9.29 A.M. 17-26 13.93 6.72 8.37 13.40 6.42 8.07 0.53 O.20 O.3O II 766 73" 9969 ii 534 6989 9632 864 350 448 215 74 129 347 o 208 12 7 8 24.7 11.4 19.9 152 7 123 18.3 "3 45 28 39 50 17 27 1.65 Jan. 27 9.29A.M. Feb. 7 9.23 A.M. 27-39 7.90 3-28 4-95 7.70 2.98 4.67 0.32 0.20 0.28 : 7098 2305 4058 6982 2 009 3 75 839 269 466 74 40 103 600 50 250 50 7 20 I5.I 10.7 13.2 94 64 8 1 II. 68 28 6 M 460 320 2.79 Feb. 7 9.23 A.M. Feb. ii 9.21 A.M. 40-48 3.83 1.88 2.82 3-57 1.62 2.52 0.40 0.27 0.30 3375 i 240 2257 2748 95i i 805 503 256 404 227 109 151 400 100 301 66 23 38 15.9 5.8 ii. 7 98 3" 72 8.3 5i 10 2 6 970 640 780 Feb. ii 9.21 A.M. Feb. 28 12.12 P.M. 49-80 7.07 0.50 3-68 6.80 0.25 3-43 0.38 0.17 0.25 6967 258 3 206 6812 79 2 951 620 97 425 179 3i 104 730 o 54 146 9 21 19.2 5-3 14-4 118 33 89 ii. 5 7i 27 o ii 560 no 280 9.33 0.78 1.94 2.46 3 i 1-5 28 ooo 4 100 16 200 652 3" i 137 69 541 99-5 93-0 96.7 Feb. 28 12.12 P.M. Mar. 20 10.09 A.M. Si-ioo 11.77 2. 4 G 7.30 11.47 2.23 7.07 0.32 o.iS 0.23 13798 2 552 8 502 13 500 2087 8236 613 402 527 2lS 41 9 8 42O 70 178 42 5 9 20.3 13.9 19-5 126 85 J2I 17-7 109 56 6 33 210 4 6 7 1.42 0.59 0.78 0.86 2 I 1-5 39700 7700 22 OOO 735 285 1941 140 520 98.9 95-1 97.6 ( Hour Knded i Date led \ Hour Runs included in period ( Maximum Period of operation. (Hours. )< Minimum ( Average Maximum ( Average Period of wash. (Hours.). . . -! Minimum ( Average Quantity of applied water. j ^^ (Cubic feet.) 1 * Quantity of filtered water. ( JJ^ (Cubic feet.) ... . . ) . Quantity of wash water. ( w" (Cubic feet.) (Average Quantity of filtered waste ( M^ni mum water. (Cubic feet. ) | Average Quantity of unfiltered waste \ Maximum w.iter. (Cubic feet.) } Average Percentage which wash and ( Maximum waste water was of applied -| Minimum [" ( Maximum Actual Cubic feet per min. < Minimum rate of ( Average filtra- ,-., ( Maximum "ysar >F ( Average Ave. net rate ( Cubic feet per minute of filtration, j Mil. gals, per acre per 24 hr Net quantity of filtered waste j Maximum per run, in mil. gals, per-! Minimum Average estimated suspended I Maximum solids in river water. (Parts -1 Minimum per million.) ( Average Grains of applied sulphate of ( Maximum alumina per gallon of ap- -! Minimum plied water ( Average 2.67 3.14 3 i i-7 32 400 i 800 8 200 396 214 7S 3 81 302 98.8 93-2 96-3 1.07 1.16 2 I 1.6 7 300 i 900 4 800 195 - 98 228 68 148 98.6 92.6 96.9 1.86 3-32 4 I 2.2 8l OOO 12000 37 600 880 386 1586 127 679 99.0 92.7 98.2 3.22 5.20 3 i 55 ooo 14400 34500 Average grains of applied sulphate of alu mina per gallon of net filtered water . . . Degree of clearness of fil- { Ca tered water (Average Bacteria per cubic centimeter (M-imum Average maximum number of bacteria per cubic centimeter in filtered water Average minimum number of bacteria pci Bacteria per cubic centimeter ( J5im in filtered water (Average (Maximum Average >acterial efficiency. . ) Minimum ( Average i 600 252 679 99-3 95-9 98.0 SUMMARY AND DISCUSSION OF DATA OF 1895-96. 237 No. 4. Continued. RESULTS BY PERIODS. System. 9 i 10 11 12 13 14 15 16 17 18 19 20 July 2 July 10 July 24 9.37 A.M. 9.32 A.M. 10.55 A.M. July 2 July 10 July 24 July 31 9.32 A.M. 10.55 A.M. 5.17 P.M. 107-114 115-120 I2I-I23 4.48 8.12 6.23 O.78 1.25 1.78 1.25 2.05 4.70 3-75 3-97 7-37 3.30 o 53 0.80 1.13 I.OI 57 4.08 2.53 o 58 0.60 0.75 2.73 0.17 0.40 0.43 0.23 0.28 0.48 0.62 1.22 i 133 3 537 7023 47 88 616 935 i 248 I 014 996 I 564 4067 2 920 i 081 3 2 53 6 753 2981 280 613 i 005 715 580 I 262 3 7 6 2 O27 784 892 93 1298 47 727 631 <68 60 1 806 901 322 360 393 I 672 179 196 61 261 284 688 660 100 240 240 40 130 287 52 102 200 93 5 9 1 33 10 46 58 29 6 1 14.1 5-7 14.2 8 8 12.8 14.8 ii.7 12 5 13.5 15.3 13.3 89 IOI 91 83 96 75 87 86 i 6 5 3 IO. 2 5.1 65 26 8 6 6 640 J 5 8 9 5 J 8 , 8 58 169 71 216 511 7 16 ft? 08 1 98 6 98 9 1 WATER PURIFICATION AT LOUISVILLE. TABLE No. SUMMARIES OF Western Pressure 1 2 3 4 5 6 7 8 ( Date Dej.23, 95 I0.35A.M an. 14, 96 11.03 A.M I -10 32.72 6-55 12 02 32.38 6.13 11.67 0.42 0.22 0.35 43978 6 006 13 690 43 793 (> 7-15 13 550 760 493 653 185 Si 43 o o o 12 2 6 22.5 14.8 io-4 1 60 105 138 17.9 127 209 32 62 870 100 320 Jan. 14, 96 I I . O3 A . M . Jan. 27 3 51 I -M. 11-14 28.47 18.55 22.92 28.08 18.25 22.51 0.55 0.30 0.38 47 172 31 213 35382 4"939 31 M7 35 156 6Si 638 663 309 176 225 o o 3 2 28.4 22.3 27.8 2O2 I5S 197 25.2 I 7 8 225 142 168 TOO 20 42 I 37 0.87 i. 06 i. 08 2 I I.I 7 ooo 2 300 4 800 421 82 322 116 206 96.8 95-o 95-7 Jan. 27 3.51 P.M. Feb. 7. 9.32 A M. 15-21 8.73 7.25 S.oo 8.27 6.97 7.63 047 0.28 0-37 II 660 7656 9 609 1 1 479 7431 9395 790 471 626 .280 166 214 o o o 12 7 9 24.0 16.4 20.5 170 "5 146 18.3 130 53 33 42 464 244 340 2.64 1-39 2.04 2.24 5 i 2.4 71 ooo 12 4OO 39000 I 087 428 I 461 188 760 99.1 93.2 98.1 Feb. 7 9.32 A.M. Feb. II 11.32 A.M 22-27 8.27 2.72 4-35 7-97 2-55 4.07 0.32 0.27 0.28 9 722 3 263 5 128 9536 3081 4938 662 431 520 269 Mi 190 o o o 21 9 14 26.5 18.4 20.2 188 131 143 16.9 119 43 13 21 967 636 750 3.08 0.71 3.22 3-75 3 i 2.0 55 ooo 14400 32 500 i 027 459 967 166 68 1 98.9 95-8 97-9 Feb. II 11.32 A.M. Feb. 27 2.10 P.M. 28-44 13.12 3.23 6-55 12.78 3.02 6.25 0.38 0.22 0.30 17997 3896 8368 17819 3739 8 219 832 375 602 271 63 149 o o 17 5 9 23-5 19.9 21.9 1 66 141 155 19.4 138 S3 16 39 500 126 290 4-23 i. 06 2.25 2.48 5 i 2-4 28000 10 700 16 600 829 257 975 151 544 99-2 9i-3 96.7 Feb. 27 2.IOT.M. Mar. 20 9. 20 A.M. 45-52 25.42 8.50 19.40 25-17 S.oo 19.05 0.50 0.25 o.35 35056 10445 26859 34825 10 001 26 642 898 606 749 444 101 218 o o 12 2 4 24.2 20.8 23-3 171 148 I6 5 22.2 157 166 44 125 210 40 70 1-44 0.69 0.84 0.86 3 i 1.8 37 500 8 700 17 800 i 549 221 I 382 233 556 98.2 91.8 96.9 i Date 1 Hour ( Maximum Period of operation. (Hours. ) -j Minimum ( Average j Maximum Period of seivice. (Hours.). . ) Minimum ( Average { Maximum Period of wa-h. (Hours ).. /Minimum ( Average Quantity of applied water, j ^~ ( Average Quantity of filtered water. \ M . lximum ,f-*< i c j. \ \ Minimum (Cubic feet.) ) .Uvn- Quantity of wash water. ),,." (Cubic feet ) ( Average Quantity of filtered waste : J water. (Cubic feet.). .. , j Minimum ( Average r ni. . ( Maximum Quantity of unaltered waste),.. . water (Cubic feet ) ( Average Percentage which wash and ( Maximum waste water was of applied -j Minimum f 1 Maximum Actual ! Cubic feet per min. ! Minimum rate of! (Average f. ltra - | Mil. gals, per acre ( ^iximum tlon per 24 hours.... i" mmlum (_ ( Average of nitration j Mil. gals per acre per 24 hr. re,. . ( Maximum Net quantity of filtered waste W, inimum per run, in mil. gals, per acre | Avenl e solids in river water. (Parts ] Minimum Grains of applied sulphate f Maximum of alumina per gallon of ! Minimum 2.67 2.84 2 I 1.6 35 700 i 800 9 200 459 183 i 032 107 287 98.8 93-9 96.9 Average grains of applied sulphate of alu- Degree of c.earness of filtered \*g~ ( Average , . ( Maximum Bacteria per cubic centimeter] Minimum ( Average Average maximum number of bacteria per Average minimum number of bacteria per , . ( Maximum Bacteria per cubic centimeter ( Mimmum in river water | Average 1 Maximum Average bacterial efficiency. . . < Minimum ( Average SUMMARY AND DISCUSSION OF DATA OF 1895-96. 239 4. Concluded. RESULTS BY PERIODS. S y s te m. 9 10 11 12 13 14 May 18 9.15 A.M. May 28 12. 18 P.M. 108-112 54.62 20.87 39-97 54.67 20.63 39-65 o.53 0.23 0.32 45 180 21 446 35407 45752 21 786 35 563 783 630 671 329 107 165 o o 4 2 2 17-6 14.0 14.9 125 99 1 06 14.5 102 215 104 1 68 150 50 90 1.87 0.51 1 .16 i.ig 3 I 2.0 6800 2 5OO 3900 875 44 5" 75 195 gS-3 90.5 95.0 15 16 17 18 19 20 Mar. 20 9.20 A.M. Mar. 23 9.19 A.M. 53-57 4-65 2.08 3.28 4-33 1-75 2-95 0.40 0.25 0.33 3787 2 276 3 069 3 595 2043 2873 763 625 696 294 128 196 o o 44 22 29 19.4 13.5 16.2 137 96 "4 II. I 79 14 6 II I 276 993 I 140 Mar. 23 9.19 A M. Mar. 30 9. ia A.M. 58-82 9.03 3-33 5-95 S.73 3.00 5.67 o.33 0.20 0.28 8797 3230 5566 8907 3070 5 577 962 625 737 193 58 124 o o 31 IO 15 19.2 14.6 16.4 136 103 n5 3-3 94 38 1 1 23 660 333 450 4.96 I of) Mar. 30 9.14 A.M. May 7 9.00 A.M. 83-94 8.52 3-28 5-55 8.27 2.93 5-27 0.38 0.23 0.28 8746 2307 5025 8875 2 O62 5 oio 764 656 706 245 86 156 o o 40 9 17 18.5 11.7 15.9 132 83 112 12.5 88 38 7 20 I 131 352 780 5 .I6 May 7 g.oo A.M. May 18 9.15 A.M. 95-107 22.58 1 1 . go 13-94 22.37 1 1. 60 13.62 o 40 0.22 O.32 3M94 16048 19 223 3IS07 16053 18 883 79 481 657 270 68 168 o o o 6 3 4 24.2 22.1 23.1 I? 57 164 22. I 157 I 4 8 73 89 185 70 130 1.97 0.68 i. 06 I. 10 4 I 2.2 7900 3 600 5 800 248 45 267 109 1 80 98.3 95-4 96.9 May 28 12. 18 P.M June 3 5.19 P.M. 113-157 14.12 0.70 2.09 13.00 0.52 1.82 1.40 0.15 0.27 13454 539 I 906 I3>59 464 i 764 I 257 310 587 i 255 44 134 o 218 8 38 20.8 12.6 16.2 148 89 114 9-5 67 57 o 6 829 400 600 10.70 1.91 4.41 7.11 5 2 4-o 30600 7900 23 900 June 3 5.19 P.M. June 9 g 45 A.M. 158-183 7-37 0.80 2.62 7.08 0.65 2.40 0-53 0.15 O.22 8 224 S2 3 2555 S 215 750 2448 857 380 ?22 2*6 45 100 o o o 63 6 24 20. 6 13.0 T7-3 147 92 122 12.4 88 37 i 10 459 1 60 280 7-48 2.43 4.06 5-35 3 2 2-3 18 goo 4300 8 300 June g 9.45 A.M. July i g.oo A.M. 184-224 8.90 0.77 3-32 8.57 0.57 3.05 0.45 0.15 0.27 IO II I 797 3458 10 175 713 3369 794 324 577 253 53 121 O O 84 7 20 21.0 I6. 7 18.4 149 118 131 13-9 98 45 i 3 582 200 320 8.58 1.52 4.84 5-12 3 2 2.1 21 200 6OOO 10 400 July i g.OO A.M. July 6 g.I2 A.M. 225-228 3-73 43 2. 10 3.12 0.92 I .60 0-55 0.38 0.50 3513 I 251 1909 3304 935 1635 798 735 766 296 79 234 o o 81 30 32 17.2 16.8 17.0 122 Jig 120 7-3 5 12 I 4 870 870 870 5.58 4-53 4.91 IO.2O 3 3 3.0 July 6 9.12 A.M. July 22 9.05A.M. 229-248 8.52 1.48 3-77 8.27 I .22 3-37 0.60 0.23 0.40 8553 1578 3688 8359 i 281 345 93 6 564 744 374 53 175 o o o 54 IO 25 2O. 2 15 4 17.1 143 109 121 12.2 86 13 2 13 560 22O 440 7-55 2.63 4.62 6.18 2 2 2.O 25400 5 ooo 8 loo 130 2 7 July 22 9.05 A.M. July 2g 3-43 P.M. 249-260 4.40 0.92 2.43 4.18 0.73 2.03 1.30 0.17 0.40 4 261 647 i 946 3684 617 i 777 83" 484 658 742 27 77 O 150 17 43 15-6 J3-3 14.6 in 94 103 7.6 53 17 o 5 2 170 I 170 I 530 9.69 3.20 5-50 10.00 3 2 2-5 33300 9500 22600 3.16 4-45 3 3 3.0 60 100 41 600 56 300 Sio 593 i 153 450 773 99-3 98.0 98.6 3-23 3.80 4 2 3.0 55400 25 400 4O6OO I 168 419 3276 196 726 99.6 92.4 98.2 3.46 4.17 5 2 3-2 39600 1 8 500 26 ooo 967 614 2 545 76 776 99.6 91.4 97-o 470 27 191 99-9 97-5 99-2 448 9 72 99-9 96.4 99.1 I OOO 8 76 99-9 94.6 99-3 357 8 58 99-9 97-5 99-3 646 76 314 99-4 98.1 98.6 240 WATER PURIFICATION AT LOUISVILLE. TABLE No. 5. GRAND TOTALS AND AVERAGES FOR THE ENTIRE INVESTIGATIONS. Warren Western Western System. System. System. Total per ods in days ion 89.76 83.76 6.00 347 334 2473 5iS 2404357 176285 17 292 49611 6h. 27m. 6h. oim. 26m. 7405 7 124 528 5i 149 19.9 114 2.70 3.00 96.7 90.40 86.68 3.72 272 260 3 f 6 1 073 3077341 162 997 9739 4387 Sh. aim. Sh. oom. 2im. II 808 n 831 627 37 18 24.7 100 2.49 2.65 96.0 25.81 24-15 1.66 124 122 565 207 526 112 6077! 17442 22 422 5h. 05m. 4h 45m. 2Om. 4633 4312 498 143 ISS I5-I 92 2.90 3-53 97-4 65.99 62.72 3-27 261 260 i 773 994 i 739628 164658 38371 6h. osm. 5h. 47i- i8m. 6823 6 691 633 148 o 21.7 154 2.41 2.72 97.3 ~, Service . ( Wash Total number of runs i Total quantities of water by meter, in .... f Annlied | Filtered <j Wash Periods of time .... Average P er run Quantities of water I Average actual rate ... Average grains of sul phate of alumina. ... Average bacterial effic U fill ed ( Wash 1 Wash ( Million gallons per acre per 24 hours . j Per gallon of applied water ~\ Per gallon of net filtered water SUMMARY AND DISCUSSION OF DATA OF 1895-96. o < S3 5 = II <a II &S ji 1 ! :.= , : I^S j :| n u A i . rt it . rt p . rt o~ c ^ o a : o ; c a. = (j U 1 4> *-"*- ^ S D V C : : EC c OJ 1> < = c Z -.: Z = -a ^ V - - 7) !^ 242 WATER PURIFICATION AT LOUISVILLE. OUTLINE OF THE METHODS FOLLOW KD IN THE DISCUSSION OF THE RESULTS OF THE INVESTIGATIONS. At the outset of this discussion the fact is to be recorded that the amount of strictly comparable data forms only a small propor tion of those presented in the foregoing tables. This was due to conditions which unavoidably caused results to be influenced by more than one varying factor at the same time. To keep conditions parallel with re gard to certain important factors was imprac ticable, owing to the arrangements under which the tests were conducted. Neverthe less, considerable light was obtained upon those laws which appear to control, practi cally speaking, the efficiency and elements of cost of purification by this method. This is especially true of the laws when taken as a whole. When single laws or principles are considered, it will be found that in many cases they are intimately associated with others, and the data lead to valuable and practical suggestions, rather than to well-de fined and specific conclusions. This was not true in all cases, however, as much definite information of practical significance was ob tained. The discussion is presented under five main sections, as follows: 1. The quality of the Ohio River water after purification, with reference to the re spective systems. 2. Prominent factors, which influenced the qualitative efficiency of purification in the case of the respective systems. 3. Prominent factors which influenced the elements of cost of purification in the case of the respective systems. 4. Comparison of the elements of cost of purification, by the respective systems, of twenty-five million gallons of Ohio River water daily, based on the foregoing results. 5. General conclusions. The discussions and conclusions in this chapter relate solely to the information ob tained up to August i, 1896. In 1897 addi tional light was obtained on a number of important points connected with this general method of purification, as is stated in Chap ter XV. For a complete understanding of the practical significance of these tests it is necessary to study, together, both Chap ters IX and XV, but it must be borne in mind that they refer to distinctly separate data, which were obtained under different condi tions. SECTION No. i. THE QUALITY OF THE Onio RIVER WATER AFTER PURIFICATION, WITH REFERENCE TO THE EFFICIENCY OF THE RESPECTIVE SYSTEMS. For the sake of explicitness the quality of the water after purification is discussed with the following points in view: A, Physical character. B. Chemical character. C . Biological character. It will be remembered that this method was followed in Chapter 1. where the composition of the Ohio River water before purification was described; and reference is made here to that chapter for detailed data for comparison with those presented in this chapter and the preceding one. Appearance. As a rule the appearance of the effluents of the respective systems was satisfactory with regard to freedom from turbidity. Each of the effluents was turbid at times, but the operation of the systems was usually modified promptly, so as to correct a failure in appearance. In the case of the Western Systems, but not in the Warren or lewell systems, the effluents were usually turbid immediately after washing the filters, and it was the cus tom to waste the effluent until it became clear. Tn the Warren and Western Pressure systems, the effluent usually became turbid after the loss of head had reached a certain but varying amount, and in these systems it was the tur bidity of the effluent which determined the time of washing the filter. This was true in a great many instances of the Jewell System, but by no means uniformly so. The com position of the river water with its minute particles of clay, and the degree of coagula- SUMMARY AND DISCUSSION OF DATA OF 1895-96. 243 tion of the river water with reference to the actual conditions of nitration, were important factors associated with the appearance of the effluent, as will be evident from the following portions of this chapter. The foregoing summary of the data on tlu appearance of the several effluents, in Table No. i, shows that the effluent of the Jewell System was the most satisfactory in this re spect. Color. As the river water itself, inde pendent of its suspended matters, is practi cally colorless all of the effluents were naturally satisfactory with regard to color. Whenever they showed a noticeable color it was not due to dissolved coloring matter in the filtered water, but to a turbidity which has been referred to under " Appearance." Taste. The taste of the several effluents was satisfactory, although it differed some what from that of the river water, owing to the varying amounts and kinds of suspended matters in the latter. Odor. The slight musty, aromatic or vegetable odor of the river water was sub stantially unchanged by the purification of the water by this method. In practically no case was the odor objectionable, or more in tense than would be expected from a surface water. Chemical Character of the Effluents. Organic Matter. \Yith the possible excep tion of those abnormal conditions when the appearance of the effluent failed to be satis factory, the organic matter in the river water was reduced to a satisfactory degree by each of the systems of purification. The summary of the data upon this point, presented in Table No. 2 of this chapter, shows the percentages of removal. It will be noted that practically all of the suspended organic matter, and a cer tain amount of dissolved organic matter which was dependent upon the quantity of applied sulphate of alumina, were removed from the water in the case of each of the re spective systems of purification. Dissolved Oxygen. The amount of free at mospheric oxygen dissolved in the water was substantially unchanged by treatment bv the Jewell and Western Gravity systems. There was a slight increase as a rule in the case of the Warren System, due to passage of the effluent over the weir by which the rate of filtration was regulated. In the Western Pressure System there was apparently no change until the warm weather of June and July, when there was a reduction in the amount of oxygen dissolved in the water. The results of determinations of the amounts of oxygen dissolved in the river water and in the effluents of the respective systems of puri fication are given in the following table: PERCENTAGES WHICH THE FREE OXYGEN DISSOLVED IN THE OHIO RIVER WATER, AND IN THE EFFLUENTS OF THE RESPEC TIVE SYSTEMS OF PURIFICATION, WAS OF THAT NECESSARY FOR SATURATION AT THE ACTUAL TEMPERATURE. Temper- F.fflvj ems. Deg re c. Water. Warren. Jewell. Western Grav.iy. Western 1895 Dec. 3 7 s Si 78 76 So 6 So 82 78 87 86 82 1896 86 88 84 " T 7 98 Feb. 10 5-9 97 " 26 Mar. 4 3-4 100 IOO 92 92 99 " ii 6-5 9 88 5.2 98 May 6 23.1 85 96 85 83 24.5 " 2g 24.7 71 87 76 85 87 80 88 So " IS 25.3 78 87 78 68 26.S So 85 67 July 9 " 18 25-5 25-6 72 71 78 83 76 77 ""76 60 Undecomposed Sulphate of Alumina. The question of the passage of undecomposed sul phate of alumina was presented and discussed with care in Chapter III. It may be again stated that with the skill and care requisite for the efficient and economical operation of a system of purification by this method there is no occasion for the passage into the efflu ent of undecomposed chemical, applied for the purpose of coagulation, so far as can be judged from the quality of the Ohio River water met with during these investigations. The only instance where the effluent was acid, due to an excess of sulphate of alumina, for 244 WATER PURIFICATION AT LOUISVILLE. several days in succession was in the case of the Jewell System during July. This was due solely to carelessness on the part of the operators of the system. In all of the systems the lack of adequate provisions for subsidence made the possibilities of this occurrence much greater than should be per mitted in practice. Carbon Dioxide. From a practical point of view the amount of carbon dioxide, more familiarly known as carbonic acid gas and dis solved in water in the form of free carbonic acid, is of considerable significance by virtue of the part which it plays in the corrosion of "iron pipe, tanks and boilers. Corrosion of uncoated iron receptacles for water by the joint action of carbonic acid and free at mospheric oxygen gas dissolved in the water is substantially as follows: Carbonic acid attacks the iron, when in completely protected by paint or other prep arations, and forms what appears to be the ferrous carbonate of iron. The oxygen in the water changes this com pound, formed by the action of the carbonic acid, into the insoluble ferric hydrate of iron, and at the same time liberates carbonic acid gas. This carbonic acid attacks more iron and the action goes on by a repetition of this process, aided by such additional amounts of carbonic acid and oxygen as the water brings to the attacked surface. The rate at which the iron surface is corroded depends upon a series of factors, the relative importance of which is not accurately known. There is no conclusive proof, however, so far as is known, that the-action is a self-limited one in the case of pipes and tanks. In boilers the high tem perature drives off these gases, and corrosion .appears to be more irregular and less marked, and would be located at the water line. By means of this process there is formed in the iron a depression of greater or less size, according to the period of exposure and other conditions. In this depression, and reaching out from it, is a formation which is called a tubercle. These tubercles have been found to consist principally of iron hydrate or oxide, together with a little silica, lime, magnesia and carbonic acid, and such compounds from the water as are coagulated by the iron hy drate. It will be noted that this action is an illustration of the principles employed in the preliminary treatment of water by the Ander son process of purification, with which you are familiar in a general way. This corroding action is possessed by the Ohio River water before purification, by virtue of the carbonic acid and oxygen dis solved in it. Whether or not the other sub stances in the river water influence to an ap preciable degree (practically speaking) the corrosion of iron is not now accurately known. But concerning such ingredients, if any are present they are in solution and there are strong reasons for the belief that they would be substantially unaffected by the treat ment in question. The corrosion of iron by the river water was shown by an inspection of the water-pipes at the laboratory and pumping station. In the Ohio River water after purification by this method, this corrod ing action is apparently increased, practically speaking, by an amount proportional to the composition and quantity of alum or sulphate of alumina added to the water to effect coagulation. This increased corroding action is indicated by the following experiment: On July 8, 1896, two glass flasks contain ing a considerable quantity of cast-iron bor ings were filled with river water and Jewell effluent, respectively. The amount of oxy gen dissolved in the water of the two flasks was practically the same about 75 per cent, of that necessary for saturation. The effluent of course contained more carbonic acid, due to the decomposition of the applied sulphate of alumina, which on that day averaged 4.67 grains per gallon. The flasks and their con tents were allowed to stand forty-eight hours with occasional stirring. Analyses were then made with the following results: PARTS PER MILLION. Sample. Carbonic Acid Before Dissolved Iron After River water 8.70 Jewell effluent 40.9 21.30 The above experiment was repeated with like results, and serves to show the increased corroding action of the water after purifica tion. These results, however, must not be taken as a basis for computation of the rate of corrosion in actual practice, because the vari- SUMMARY AND DISCUSSION OF DATA OF 1895-96. ous conditions affecting this action were not sufficiently parallel to yield data for any other purpose than that for which the experiment was made. To what extent steam boilers, and cast iron, wrought iron or other metal used for dis tributing and service pipes or fittings, in the case of this water, are, or may be, effectually protected from corrosion by a suitable sur face coating, is a matter which the writer has not investigated, and upon which he has no opinion to express at this time. In Chapter I it was shown that during June and July, 1896, the Ohio River water contained from 21.1 to 30.8 parts per million of carbonic acid gas by weight. At some seasons of the year the water doubtless con tains much more than the above quantity of carbonic acid. The evidence indicates that in Nov., 1895, it was at least 75 parts. A con siderable portion of the carbonic acid, and at times perhaps all of it, is engaged in holding the carbonates of calcium and magnesium in solution in the form of bicarbonates. The bicarbonates are not. stable compounds, rel atively speaking, and there is substantial proof that they give up their carbonic acid to facilitate the corroding action in question. With regard to the relative rates of corroding action by free carbonic acid gas and partially engaged carbonic acid in the form of bicar bonates, there are no available data to lead to a satisfactory expression of opinion. The amount of carbonic acid gas liberated by the decomposition of alum or sulphate of alumina, is capable of both approximate deter mination and estimation. The latter requires, however, an exact knowledge of the amount and composition of the applied alum or sul phate of alumina. From the data presented in Chapters I and II the amount of carbonic acid liberated in the water by the decomposi tion of the applied chemical during these tests may be estimated with sufficient close ness for practical purposes. The amount of liberated carbonic acid gas per unit quantity of applied chemical varied somewhat in the several lots which were used, owing- to the different percentages of sul phuric acid. But taking a chemical of average composition, and assuming that the chemical united wholly with the alkaline compounds, and not with organic or suspended matters, the amount of liberated carbonic acid gas may be adequately shown as follows: In the case of the potash alum used by the Western Company each grain per gallon would liberate about 2.5 parts per million of carbonic acid gas by weight. With the several different lots of sulphate of alumina, the parts per million by weight of liberated carbonic acid gas would range from 3.6 to 4.0, and average about 3.7, for each grain per gallon of applied chemicals. These figures refer solely to the liberation of chemi cally combined carbonic acid gas. In addi tion thereto, in the case of bicarbonates an equal amount of carbonic acid, partly en gaged by holding calcium carbonate in solu tion, would also be set free-. It is very ques tionable, however, whether this last action would affect corrosion appreciably, if at all. From the above statements, together with the foregoing tabulations in detail of the amounts of chemical applied by the respec tive systems, correct information may be ob tained as to the average quantity of carbonic acid gas liberated in each case, for runs, days or periods. Some observations worthy of mention were made upon the wrought-iron reservoir used for the storage of filtered water for use in washing the filters. Throughout the in vestigations, this iron reservoir, which was not protected on its inner surface by a coating of paint, tar or other material, was practically filled with filtered water. In fact its use for this purpose began early in July, 1895. From the close of the tests on Aug. r, 1896, the reservoir remained full of filtered water, in an undisturbed condition, until Oct. 17, when one of the systems was operated for a few hours. It then remained undisturbed for another month, when it was drained. Several days after draining, the inner surface of the iron was examined and found to be corroded to a considerable degree. Tubercles were found ranging in size from that of "a pin-head to about 0.4 inch in height, and i inch in di ameter as a maximum. Their size was very variable. It is of course certain that the cor roding action was increased somewhat by the acid effluent of the Jewell System during a number of successive days in July. This 246 WATER PURIFICATION AT LOUISVILLE. inexcusable acidity was perhaps not the chief factor, however, as the effluents regularly had considerable corroding action, as was shown by the iron in the effluents which stood over Sunday in the iron outlet pipes. A more exhaustive study of this subject was made in 1897, ail( m passing it may be noted, the corroding action of the undecom- posed chemical in the effluent even at rare intervals was of great significance, as it ac celerated the action of the carbonic acid. The suspended matter in the river water forms a partial protective coating to the metal, and this explains for the most part the results of the experiment on July 8, 1896. Further more, the more extended data of 1897 showed that the evidence obtained in 1896 indicated an abnormally high percentage increase of carbonic acid after purification. Additional information upon this subject is given in Chapter XV. Analysis of one of the tubercles mentioned above showed it to be composed very largely of iron in the ferric oxide state, with a small amount of calcium carbonate (lime). The percentage composition was found to be as follows: Water 1 1.42 Silica (SiO 2 ) 0.19 Oxide of iron (Fe 2 O 3 ), by difference. . . 85.1 1 Alumina (A1 2 O 3 ) Trace Lime (CaO) 1.77 Magnesia (MgO) 0.03 Sulphuric acid (SO :i ) 0.08 Carbonic acid (CO 2 ), by estimation. . . 1.40 Organic matter Trace In concluding this account of the increased corroding action of the water after purifica tion, due to increased amounts of carbonic acid gas proportional to the quantity of alum or sulphate of alumina added to the water, it may be stated that the adoption of this method of purification would call for especial care in coating the inner surface of pipes, and for all feasible means of keeping the amount of applied sulphate of alumina at a minimum. So far as experience teaches us. the corrod ing action of this water before and after puri fication, on lead, would not give trouble, be cause it quickly forms a coating by itself which protects the lead from further action. It may also be added, that the carbonic acid gas may be removed from water by lime water or caustic soda, with subsequent sub sidence or filtration. It is not probable, how ever, that such steps would ever be necessary. Passage of Lime from the Form of Carbon ates to that of Sulphate. It has been explained in Chapter 111 that the alkalinity of this water was produced, for the most part, if not wholly, by the carbonates and bicarbonates of lime and magnesia, respectively, and that it was reduced by an amount approxi mately proportional to the quantity of alum or sulphate of alumina added to it. With sulphate of alumina of average composition the alkalinity has been found by actual tests to be reduced about 8.1 parts per million for i grain of this chemical added to i gallon of ordinary river water. This means practi cally, since the evidence indicates that the lime is more abundant than magnesia, that this amount of lime and magnesia, but principally lime, is converted from the form of carbonate or bicarbonate to that of sulphate. That is to say, the permanent hardness or incrusting constituents is increased by about 8.1 parts per million, according to the conventional method of expressing permanent hardness in terms of calcium carbonate. The actual weight of the compounds increasing the in- crusting constituents would be more than this, because calcium sulphate weighs 1.37 times as much as an equivalent amount of calcium carbonate. With potash alum, such as was used by the Western Company, the application of i grain per gallon was found to reduce the alkalinity, and increase the incrusting constituents about 4.5 parts per million. The data presented in Chapter I show that the incrusting constituents of the Ohio River water ranged from 30 to 51 parts per million. when tested during this period. With the above data on the increase of incrusting con stituents due to the application of alum or sul phate of alumina, and the foregoing records of the amounts of these chemicals employed by the respective systems, a correct idea may be obtained as to the increased incrusting constituents of the several effluents. It is the amount of incrusting constituents of a water, due to the chlorides, nitrates and SUMMARY AND DISCUSSION OF DATA OF 1895-96. 247 sulphates of lime and magnesia, which chiefly determines its fitness for boiler use. When proper care is taken of boilers, it appears that the Ohio River water does not give serious trouble except during low water in the fall, by the formation of boiler scale; although the suspended matter in the water forms a sludge, which requires frequent flushing of the boil ers, and occasional removal by manual labor. With the probable exception of magnesium chloride, due to its tendency to decomposi tion and formation of hydrochloric acid, there is no more objectionable ingredient of water for boilers than sulphate of lime. This com pound, which is formed by the addition of alum or sulphate alumina to water, as ex plained above, and which is soluble at ordi nary temperature, produces at boiler tempera tures a fine hard scale, in which practically all of the suspended matters of the water be come embodied, when those matters consist of fine clay. In the case of heavy mud, these incrustations are attached to the sludge. Unless removed, the scale formed in this man ner eventually causes a marked waste in the consumption of fuel by retarding the trans mission of heat to the water; and it is com pletely removed with great difficulty. Such a scale was found in Boiler No. 3 at the pumping station of this Company, as you have been advised. This boiler was said to have been filled with the effluent of the Warren and Jewell systems on July 7. During the next run of five weeks, muddy river water was introduced to replace the steam which was not condensed and re turned from the engine to the boiler. The boiler was carefully examined after one sub sequent run to this one was made, without cleaning during the interval of rest. On ex amination the tubes and plates were found to be covered with a hard rough incrustation such as above described. This was especially noticeable on the iron plates around the fire box. In places there were evidences of cor rosion. A portion of this incrustation was removed and analyzed, with results which show the following percentage composition: Water, with organic and volatile mat ters ifi-5 2 Silica (SiO 2 ) 19.65 Oxide of iron (Fe 2 O 3 ) 4.30 Alumina (A1 2 O 3 ) 9.66 Lime (CaO) 37-97 Magnesia (MgO) 0.60 Soda (Na 2 O) Undetermined Potash (K 2 O) Undetermined Chlorine (Cl) Trace Nitric acid (N 2 O 5 ) Trace Carbonic acid (CO 2 ) Trace Sulphuric acid (SO 3 ) 1 1.87 The alumina which was found in the in crustation came from the silicates (clay) of the river water subsequently added to the boiler, and not from the chemical applied in the course of purification. At the time when the boiler was said to have been filled with the effluent, there were about four grains per gallon of sulphate of alumina being added to the river water on an average. This practically doubled the in- crusting constituents of the water, and added to the effluent about 44 parts per million of calcium sulphate by weight. This sulphate was soluble as it entered the boiler, but the high temperature caused it to be insoluble, with the result that a very hard scale was formed, which included a large portion of the suspended matter of the water subsequently added to the boiler. In fact the analyses show that less than 20 per cent, of the incrustation was composed of sulphate of lime. The above experience shows that all rea sonable steps should be taken to keep the amount of applied sulphate of alumina to a minimum. In this connection, however, it is to be stated that the amount of sulphate of alumina added to the river water on the date when the boiler was said to have been filled, was about 50 per cent, greater than the aver age amount employed during these tests. Furthermore, the mud, silt and clay in the water subsequently put into the boiler, added very materially to the incrustation, as shown by the results of the chemical analyses. By the use of soda, it is possible to remove the sulphates of lime and magnesia from the water; and trisodium phosphate will also serve this purpose, should manufacturing -es tablishments choose to remove these ingredi ents before the water enters the boilers. WATER PURIFICATION AT LOUISVILLE. Further discussions of the incrusting power of the Ohio River water, before and after puri fication, with additional data, will be found in Chapter XV. At this point, it may be briefly noted that when the river water is muddy and requires the largest quantities of coagu lant, the incrusting constituents naturally present in the water are so low in amount that the total incrusting power of the efflu ent would be much less than that of the natural river water during the fall months. Biological Character of the Effluents. Microscopical Organisms. The tables in Chapter VIII show that practically no diatoms, algae or other microorganisms, which may be readily recognized by the aid of the microscope, were present in the efflu ents. This would be naturally expected un der the circumstances, owing to their greater size when compared with the bacteria. It is to be noted, however, that very few organ isms of this nature were found in the river water, owing to unfavorable natural condi tions existing there. In this connection there arises a question of much practical significance, as was pointed out in a communication addressed to you on July IT, 1896; that is, the conditions un der which the growth of microorganisms, notably alga?, in the effluents could be pre vented during the period when the water is stored prior to distribution. In the case of all the effluents the conditions for growth of algae would be favorable in the presence of sunlight; and should these forms once be come established in the distributing reservoir the probability of the production of objec tionable tastes and odors in the effluent, no matter how satisfactory was its character as it left the filters, would be a very serious state of affairs. There are no specific data to offer upon this subject. Bacteria. The removal of bacteria from a water which at times shows such marked proof of sewage pollution as is the case with the Ohio River, is a very important matter. This is particularly so in view of the rapidly increasing population in the Ohio River valley, and the set of data upon this System. Bacterial Kfficie icy. Jewell - Western Pressure 97-3 point was made as complete as practicable. Comprehensive summaries of these data have already been presented in this chapter. The bacterial efficiency of the respective systems, as shown by the total averages, was as fol lows: The above results are not directly compara ble, because the length of service and the condition of the river water during service were quite unlike, as shown by the data of each of the twenty periods of different grades of river water. Excluding the Western Gravity System on the grounds of failure to purify enough water, when the river water was in a muddy condi tion, to wash its own sand layer, and taking the averages of all those periods in which the remaining systems were in operation without any prescriptions from this Company, the fol lowing bacterial efficiencies are obtained: Syste Warren Jewell Western Pressure. 98.5 97-9 97-4 The above figures show the relative effi ciency which the systems possessed in the re moval of bacteria from the river water. Dur ing the early part of the tests the bacterial efficiency was irregular and unsatisfactory at times in the case of all the systems, but least so in the case of the Warren. This was due in part to limitations of the devices em ployed in the respective systems, and in part to a lack of care and skill in adapting the operation of the system to meet the re quirements of the rapidly varying character of the river water. In February and March the bacterial efficiencies of the systems, speak ing in general terms, were so unsatisfactory that an official communication was addressed on March 16 to the operators of the systems. The request was made among others that they should keep the bacterial efficiency of their systems above 97 at all times when the num ber of bacteria per cubic centimeter in the SUMMARY AND DISCUSSION OF DATA OF 1895-96. 249 river water exceeded 7000, and when the bacteria in the river water were less than this number there should not be more than 200 per cubic centimeter in the effluents. Following this official request for greater uniformity in bacterial efficiency, the applica tion of chemicals and the rate of filtration, a number of changes and improvements were made in the systems. As a rule the removal of bacteria from that time to the close of the test was satisfactory, provided we disregard the amount of chemi cals employed to effect the purification. There was one prominent point of much practical value learned in connection witli the bacterial efficiency of the systems. The opinion has generally prevailed that the qual ity of the effluent of a filter of the type em ployed in these tests would not be satisfactory immediately after washing the sand layer, and for some minutes it would be necessary to waste the filtered water. The satisfactory bacterial results obtained from the Warren and Jewell systems, in which the sand layer was quite thoroughly washed as a rule, show clearly that the unsatisfactory quality of the filtered water just after washing is not an in herent feature of this type of filters under the existing conditions, but a consequential one, arising from incomplete washing of the sand layer, and other factors. Inspection of the results showing the aver age bacterial efficiency of the systems indi cates them to be fairly satisfactory when com pared with available data upon the efficiency of filters of the English type. Such compari sons of data, however, require the careful con sideration of several facts. In the first place, the growths of harmless bacteria generally recognized to prevail to a greater or less de gree in the underdrains and lower portions of filters of the English type, did not become established to any marked degree in the cor responding portions of these filters of the American type, owing evidently to the wash ing of the sand layer at frequent intervals. This was especially true of the filters of the Warren and Jewell systems, in which the thor oughness of washing was enhanced by the accompanying agitation of the sand. Another fact bearing directly upon this point is, that if any slight growth of bacteria within the lower portions of a filter of the American type should take place, the rate of filtration would cause the bacteria to be diluted in the effluent to about fifty times the extent that would be the case in English tillers. These facts, together with the results of numerous comparative observations of the species of bacteria in the water before and after purifi cation, show that, in order to insure to the consumers of the same water the same protec tion from disease germs, the bacterial effi ciency by this method of purification must be somewhat higher than in the case of English filters. Several years ago it was demonstrated by a series of separate investigations in different parts of the world, that English filters were not " germ-proof," but with skill and care in their construction and operation, they could be made very nearly so. It follows from the experience gained in these tests, as stated above, that skill and care are apparently more necessary in the case of the method of puri fication as practiced and investigated in these tests, than in the case of the English filters. With more adequate provision for subsidence, the necessary skill and attention in operation would be materially reduced. This fact of not being " germ-proof " was confirmed by the marked similarity from time to time of the bacteria in the water before and after purifi cation. Evidence upon this point was also obtained from the artificial application of bac teria, which in their life-history in water re semble the bacillus generally recognized as the specific germ of typhoid fever. Bacillus Prodigiosiis. This germ was ap plied to the filters of each of the systems on a number of different occasions during the month of July, 1896. It was not considered prudent to apply these germs to the filters at an earlier period. The reason of this was to guard against the possibility of having any bacterial growth within the filter, if such should occur, attributed to any actions of the Water Company. In passing, it may be added that the introduction of Bacillus prodigiosns requires the addition to the water of a certain, but very small, amount of organic matter in which the germs were grown. When these germs were applied to the filters it was done very cautiously, and 2 5 WATER PURIFICATION AT LOUISVILLE. with scarcely an appreciable addition to the amount of organic matter naturally present in the water. The method consisted of applying to the water above the sand layer a measured quantity of a solution which contained mil lions of this germ to each cubic centimeter. These applications were made every five min utes for several hours, and the applied germs distributed in the water as uniformly as prac ticable. During these applications, samples of the effluent were collected at frequent in tervals for analysis. The results of these applications of Bacillus prodigiosns to the several filters are summa rized in the next table. They show clearly that under the existing conditions an appre ciable number of these germs was able to pass through the filters. Owing to the lim ited amount of data, all of which were ob tained under a narrow range of conditions, they are of little value in the consideration of the absence of- bacterial growths within the filters. SUMMARY OF THE RESULTS OF THE APPLI CATION OF BACILLUS PRODIGIOSUS TO Number of Applied Number Number of B. pro JiRiosus B prodigi- of Average rubicren- HReof Effluent Rvmoval Filtered Analyzed M ax. Min. Average. WARREN FILTER. July 9 " 16 465 3470 20 IS IS 1 6 o o 3 7 99.4 99.8 " 24 800 720 0.7 99.9 " 28 150 28 : 4 o 0.2 99.9 JEWELL FILTER. July 3 170 23 3 o 0.4 99.8 " 18 I 170 35 4 o I 99.9 " 29 350 29 3 I 99-7 WESTERN GRAVITY FILTER. July II 975 21 6 o 2 99.8 " 24 940 23 29 o 10 98.9 WESTERN PRESSURE FILTER. July 14 450 33 19 o 3 99-3 " 21 480 22 27 o 7 98.6 SECTION No. 2. PROMINENT FACTORS WHICH INFLUENCED THE QUALITY OF THE OHIO RIVER WATER AFTER PURIFICATION IN THE CASE OF THE RESPECTIVE SYSTEMS. The following pages contain a brief discus sion of the influence of the factors which from a practical point of view appeared to be of the most importance in the purification of this water. As a rule these factors affected both the quality of the effluents and the cost of treatment. Their influence upon the quality of the effluents is considered in this section, and their relation to the cost of treatment is presented in the next section of this chap ter. In some cases the factors were common to all the systems, but in others they differed to a marked degree in the respective systems. An outline and comparison of the factors af fecting the systems differently are given. Composition of the River Water. The quality of the effluents in all cases was affected by the composition of the river water, on account of the widely varying amounts of chemical which were added to the water in its various stages, as shown in the foregoing summary of results by periods. Owing to the large quantity of suspended matter frequently present in the river water, it was necessary at such times in the case of all the systems, to add to the water comparatively large amounts of alum or sulphate of alumina. As already explained, such additions caused an increase in the corroding action of the effluents, due to the increased amounts of carbonic acid gas set free; and also made the water less desir able for boiler use. owing to the compara tively large amount of lime changed from carbonate into the form of sulphate. Large amounts of alum or sulphate of alumina did not necessarily affect the quality of the efflu ents in other respects, provided the quantity was kept below that which could be decom posed by the carbonates and bicarbonates (alkalinity) of the river water. One observation of much practical signifi cance, which was repeatedly noted, may be recorded here. This observation refers to the relation of the amounts of sulphate of alumina necessary to produce a perfectly clear effluent, and that necessary to give a satisfactory efficiency to the system in the removal of bacteria. Early in the investigations it was noted that as the composition of the river water varied from time to time, the amounts of sulphate of alumina necessary to produce the two SUMMARY AND DISCUSSION OF DATA OF 1895-96. 2 5 above-named results, respectively, were far from parallel. The most marked examples of this were observed in March and May, iXgf). During the latter days of March, the suspended mat ter in the river water was comparatively large in size, and of a red color; and the effluents in many cases were clear, even brilliant, with out a satisfactory removal of bacteria. Dur ing the latter part of May, however, the rains which fell after a long period of drought pro duced such a character of the water in the river, that the suspended matters were very light and minute. In some cases, the sus pended particles were finer than bacteria and not more than o.ooooi inch in diameter as measured under the microscope. Under these circumstances the appearance of the effluent became unsatisfactory in less than one-half hour after the niters were put in service after a thorough washing, and when the number of bacteria in the effluent remained normal. But the most interesting observation made upon these conditions was that the effluent became unsatisfactory in ap pearance before there was a perceptible in crease in the acting head necessary to pro duce filtration at the given rate. These effects were evidently due to an inadequate degree of coagulation of the water as it entered the sand layer to be filtered at the given rate. I Application of Alum or Sulphate of Alumina. Preparation of Solutions. This question was referred to in general terms in Chapter II. In the case of the solutions of sulphate of alumina, with a very few exceptions, it is not probable that the quality of the effluents was affected by irregularities in their prepara tion. In all cases, however, it would be de sirable in the operation of a large system to employ more careful and systematic methods than were noted at times. The point to which attention is especially invited at this time is in connection with the first device used by the Western Company for the preparation of solutions of potash alum. In winter weather, when the temperature of the water was low, the strength of the solu tions, obtained by the passage of a current of water through an alum tank placed on a by pass on the main water-pipe, was so irregular that the method may be pronounced a failure, so far as its application for the coagulation of the rapidly changing Ohio River water is concerned. The use of this device was dis continued by the Western Company on April <j. Uniforinitv of Application of Solutions. With regard to the regular application of so lutions of chemicals in suitable amounts to effect proper coagulation of the varying river water, all devices gave the appearance of be ing crude, so far as their part in the produc tion of a satisfactory effluent was concerned. With a water of a certain grade the applica tion of chemicals by the Warren device was fairly satisfactory on the whole. With the Jewell and second Western devices satisfac tory results could be obtained by giving them close attention. Such attention, however, was not regularly given to them, especially during the earlier part of the investigation. In its use during these tests the first West ern device was a failure, even when an attend ant stood over it practically all the time. Not only did the operators lose control of the rate of application of the solution, but, as noted above, their inability to control the rate of ap plication was increased by the widely varying strength of the chemical solutions. In warmer weather when the water would dis solve greater quantities of alum the operation of the device would be still more unsatisfac tory. Application of Lime. Lime was applied to the river water only by the Jewell System. Under the existing cir cumstances its application was unnecessary. The object of its application was apparently to guard against the passage of undecom- posed sulphate of alumina into the effluent, and perhaps to facilitate the coagulation of the water by the regular chemicals. The trial of the application of lime by the Jewe l System was followed by disastrous results so far as the quality of the effluent was con cerned. This was due chiefly to the manner of application of the lime. At times the lime and sulphate of alumina must have reached the river water alternately, and produced of WATER PURIFICATION AT LOUISVILLE. course an effluent of unsatisfactory character. At other times the sulphate of alumina was decomposed in the feed pipe leading from the pump to the main inlet water-pipe. After the adoption of the separate feed pipes, shortly before the abandonment of the use of lime, the above difficulties were removed to a greater or less degree. In this connection reference is made to some comparative ex periments upon the coagulation of the Ohio River water by aluminum hydrate under dif ferent conditions made by the Water Com pany, and recorded in Chapter XII. It is not to be inferred from these com ments that the use of lime would not be ad vantageous for some conditions and for some waters. The above criticism refers only to the fact that its application during these tests was unnecessary, and that its manner of applica tion in the Jewell System was not well ad vised. For all the varying conditions of the river water there was doubtless a certain amount of sulphate of alumina, which, by virtue of its coagulating power, was best adapted for the purification of the water by each of the sys tems. To define this optimum amount for any given conditions is impossible with the available data. Inspection of the records and summaries shows that all of the systems were operated far from this mark at times. This was more noticeable during the early than during the later part of the investiga tions, when the operators had some experi ence to guide them. It will readily be seen that the liberal use of chemical by the War ren and Jewell systems was reversed at times during the tests. Comparing the effect of the actual quantities used upon the quality of the effluent with that of the optimum, it may be stated that in numerous cases the actual quantity appeared to be below the optimum, and showed its effect by high bacteria in the effluent, or unsatisfactory appearance, or both. At times the actual quantity was in excess of the optimum, and affected the ef fluent by an unnecessary increase in its cor roding action and capacity to form scale in boilers, as already discussed in this chapter. Comments along this line will be found in the next section, under the cost of applied chem icals. For further information in this con nection reference is made to the detailed data in foregoing tables. Provision for the Removal of Suspended Matter from the River Water by Sedimentation. At times of flood the Ohio River water contains large quantities of heavy mud, and experience indicates that in the neighbor hood of 75 per cent, of this mud on an aver age may be removed economically by plain subsidence without the use of coagulating chemicals. At other times, especially during the spring and early summer, the river water contains large quantities of very finely di vided suspended matter, which would require days and perhaps weeks for the removal of the bulk of it by plain subsidence. The evi dence presented in Chapter IV, however, in dicates that it could be removed after relatively short subsidence, following the ap plication of comparatively small amounts of a coagulating chemical. In all cases not only were provisions for plain subsidence entirely lacking, but the pe riod for coagulation and subsidence was far too short. Turning to the summaries of re sults it will be noted in the case of all the sys tems that the increase in applied chemical, due to muddy water, reached at times as high as 8 grains per gallon. The effect of the ap plication of large quantities of chemical, with reference to corrosion and incrustation, has been referred to above. With plain subsid ence followed by longer periods of coagula tion and subsidence these high amounts of coagulant could be reduced materially. There are also indications that at times it would be advisable to make more than one application of coagulant. The Warren Sys tem was superior to the others with regard to provisions for the removal of suspended mat ter from the river water by subsidence, but even in this case the provisions were wholly inadequate to prevent the use of excessive amounts of an expensive chemical, and their attending effects. In fact this was the weak est feature of all these svstems. SUMMARY AND DISCUSSION OF DATA OF 1895-96. 253 Provisions for Cleaning. All of the systems were very weak in provisions for ready and economical cleaning of the compartments in which sedimentation took place. How far this affected the character of the effluent it is difficult to estimate, as apparently these com partments were cleaned in the case of each of the respective systems at as frequent intervals as was thought necessary. It is probable, however, that the difficulties of cleaning, due to lack of provision for the ready and econom ical performance of this operation, often led to delay in cleaning, and to the consequent passage of sludge from the basin or chamber upon the filter. Provisions for Inspection. The need for regular and systematic inspection of the in terior of the compartments in which sedimen tation took place, and of the contents of these compartments, seemed to have been almost wholly ignored in these systems. This was especially true in the case of the Western Sys tems, where the arrangements were such as to necessitate the draining of the settling chamber, to examine its contents. The fail ure to inspect the condition of the contents of these compartments, with regard to the amount or sludge, undoubtedly led also to delays in cleaning the compartments, with the resultant effect of the passage of sludge from the basin or chamber upon the filter. The most notable example of this was found in the Warren System on July 22-27. Degree of Coagulation of the Partially Sub sided Water as it entered the Sand Layer. The degree of coagulation of the water as it entered the sand layer was found to be the most important feature of satisfactory and efficient purification by this type of filter. To an experienced observer the proper degree could be told with considerable accuracy by the size of the Hakes or masses of coagulated material and the rapidity with which they subsided. The coagulation had to be thor ough; that is, the volume of hydrate had to be practically sufficient to envelop all sus pended matters or furnish enough gelatinous surface to which the particles could adhere. It is believed that just after washing this is absolutely true, but after a filter had been op erated for a time and flakes of coagula had accumulated in the sand there might be per missible a slight departure from thorough coagulation. Thorough coagulation of the water above the sand, or very nearly thor ough, was absolutely essential in order to re move the fine clay particles; and during the greater part of the year the removal of bac teria was satisfactory when the effluent was free from turbidity. At times during the winter, however, when the suspended matter was very coarse the- bacterial efficiency required especial atten tion. Taking everything into consideration it may be safely concluded that the volume of available hydrate in the water as it entered the sand layer was the most important fea ture of successful filtration; and that when the amount of hydrate present departed materi ally from that necessary for complete coagu lation a uniformly satisfactory quality of the effluent could not be expected. Sand Layers of the Several Filters. Very little information upon the relative value of each of the more important features of the sand layers is available, as was stated to you July n, 1896. This is explained by the fact that a series of factors unavoidably worked together to disguise the influence of any particular factor. The two principal factors which affected the efficacy of the sand layers in purifying the water were the degree of coagulation of the water and the rate of fil tration. Both of these factors are referred to in their proper order. All things considered the sand layers of the Jewell System were the most efficient in pro ducing- an effluent of satisfactory character. A brief outline of the more important points as the systems stood at the close of the tests is as follows: Thickness. The thickness of the sand lay ers in the respective systems was as follows: Warren Jewell Western Gravity. . Western Pressure. Inchr Z7.O 30.5 31.0 49-5 2 54 WATER PURIFICATION AT LOUISVILLE. Although the Jewell sand layer gave the best results it is still an open question what the thickness of the sand layer should be to give the best results under favorable circum stances with regard to other conditions. The Western sand layers did not appear to advantage, owing, it seems, to failure to wash them satisfactorily and at times to the degree of coagulation with reference to the rate of filtration. Size of Grain. A comparison of the sizes of the grains in the several sand layers, as shown by their effective sizes (10 per cent. of the material finer than the diameter given in millimeters), is as follows: System. Effective Size. Warren o. 51 Jewell 0.43 0.43 Western Pressure 0.44 The Warren material was apparently too coarse to give uniformly the most efficient re sults under the existing conditions. The fine material of the two Western systems was mixed in both cases with very coarse sand, giving the resultant layer a rather higher ef fective size. These mixed sands were, how ever, as fine in effective size as that of the Jewell, but their use wa s handicapped by fac tors mentioned under Thickness of Sand Layer. Composition. The available evidence, so far as it goes, points to as satisfactory results under suitable conditions from natural sands, of uniform and proper size, as from the more expensive crushed quartz. Rate of Filtration. Unless some abnormal and disturbing fea tures appeared, the rates of filtration were maintained so as to give as nearly as possible the contract capacity of 250,000 gallons per 24 hours for each system. At times of very muddy water none of the systems reached the ( prescribed amount, owing to the incomplete preliminary subsidence. The rates differed somewhat in the different systems, due for the most part to the different areas of the sand layers in relation to the contract rate. As a rule the rates were held fairly uniform under the same conditions of river water, the range of variation under parallel conditions being usually less than 10 per cent. The widely varying factors which influenced the results at different times in addition to changes in rate prevent the drawing of any specific conclusions. In the following table are given the rates in million gallons per acre daily, the amounts of suspended solids in the river water in parts per million, the amounts of sulphate of alu mina used, in grains per gallon, and the bac terial efficiencies of the several systems, for those runs on which the unit rates were the maximum and the minimum, and also the average of these quantities for the entire in- TABLE OF LEADING RESULTS WITH MAXIMUM AND MINIMUM UNIT RATES, AND AVERAGE RESULTS FOR THE ENTIRE INVESTIGATION. WAKRKN JEWELL. Min. Max. Aver. Min. Max. Aver. Rate 7 80 1677 7.92 97-5 155 36 0.98 95.2 120 2.70 96.7 57 2OO 2.01 95-7 150 220 5-50 99.9 100 2.49 96.0 Sulph. of alu mina Bacterial effi ciency WKSTKRN GRAVITY. WBSTBRN PKESSUHE. Min. Max. Aver. Min. Max. Aver. Rate 33 420 9-33 99.0 152 17 o. 71 92.6 96 83 202 154 Sulph. of alu mina Bacterial effi ciency 2.90 97-4 91.4 1.24 96.8 2.41 97-3 As stated above, the wide variations in the controlling factors make it practically impos sible to draw any definite conclusions from the results in regard to the relative efficiencies of different unit rates. That within the ranges employed, the lower rates did not give uniformly any better results than the higher ones, and that the maximum limits of safe rates were not reached in this work is indi cated by the results of the last table and clearly shown by the next table. In this table are given averages of the leading results of the Warren and Jewell systems, omitting Periods Nos. i and 2. These averages have been arranged with regard to the unit rates in order to allow comparison of the results from this standpoint. In the case of the Western Systems the degree of coagulation SUMMARY AND DISCUSSION OF DATA OF 1895-96. 255 and other factors were too irregular for these systems to be given a place in this table. The headings and results have a and explained before. been used LEADING AVERAGE RESULTS OF THE WARREN AND JEWELL SYSTEMS, ARRANGED ACCORDING TO UNIT RATES OF FILTRATION. Warren System. Rates included in averages 80-89 90-99 46 IOO-IO9 69 IIO-II9 120-129 130-139 58 140-149 Rates in million gallons per acre per 24 hours .... 86 5 Amounts of suspended solids in river 526 258 Amounts of applied sulphate of alu- 4 46 4 IS <! 87 2.8o Amounts of water filtered per run in 2 868 5 660 Percentages which the water used for washing and wasted was of the Bacterial Efficiencies 98.0 98.4 98.2 99.1 97-9 98.9 98-5 Jewell System. Rates included in averages 80-89 90-99 100-109 I IO-II8 119-130 138-143 7 6 18 8 Rates in million gallons per acre per 85 o Amounts of suspended solids in river water in parts per million 1310 59i 220 476 458 278 Amounts of applied sulphate of alu mina in grains per gallon 5.80 3-9 2.88 3-85 3-95 5.06 Amounts of water filtered per run in Percentages which the water used for washing and wasted was of the 18 16 Bacterial efficiencies 97-3 ; 98.0 97.6 98.7 98.6 99.2 These data show conclusively that within these limits in rate of filtration and under the given conditions there were other factors of more practical significance than the rates of flow through the sand. Unquestionably, the most prominent of these factors was the de gree of coagulation of the water as it entered the sand layer. How high it would be possible or practic able to carry the rate of nitration in efficient and economical purification by this method cannot be stated at present; and it would re quire for its proper solution conditions where the bulk of the suspended matters was removed by subsidence. Some data showing the in- advisability of employing rates of less than 100 million gallons per acre daily were ob tained in 1897, and are recorded in Chap ter XV. Regulation and Control of Operation of tlic Pilfers. While considerable care was necessary in the operation of the niters in order to secure a satisfactory effluent it was very seldom that moderate irregularities in the rate of filtration made their influence apparent. The pres ence of certain irregularities in the pressure or acting head upon the Western pressure filter was found to be a very disturbing and disastrous element, and as a rule caused fil tration to be stopped and the filter washed. Loss of Head. As the frictional resistance to the passage of water of the sand layers and their accumu lation increased, the necessary acting head, or loss of head, increased and approached tow ards the maximum available head. The sev eral filters behaved quite differently in this respect. The Warren filter contained a coarse sand layer and almost without exception the appearance and quality of the effluent became unsatisfactory when the loss of head was much less than the total available head. As a rule the filter of the Warren System was washed when the loss of head was less than 3 feet. WATER PURIFICATION AT LOUISVILLE. During a large part of the time the Jewell System employed the full available head (12-14 feet) ou their filter before the sand layer was washed, although it is to be noted that the last two feet of head (from 10 to 12 feet) were not of great value as the resistance increased very rapidly above 10 feet. This was not always possible, however, especially when the coagulation of the water as it en tered the filter from the settling chamber was incomplete. This factor was also a very im portant one in connection with the employ ment of surface agitation during the runs. From the detailed data in Chapter VIII, it will be noted that in many instances this op eration was successfully used in the Jewell System. In the Western Gravity System the filter was washed because the loss of head reached the maximum in the case of the Western gravity filter (A), and in very few cases did the quality or appearance of the effluent de teriorate as the loss of head increased. The opposite of this was uniformly true in the case of the Western gravity filter (B). Frequent breaks " were noted in the Western pressure filter when the loss of head was a small fraction of the available head, which was equal approximately to the pres sure in the force main to the reservoir. These breaks, evidently due for the most part to ir regularities in pressure, causing the water to pass through some channel or place of les sened resistance at an abnormal rate, caused, in turn, the effluent to become turbid, and necessitated washing the filter. Previous to June i, this filter was operated with widely varying maximum acting ! heads, the appar ent custom being to wash at from 30 to 50 feet loss of head, unless the effluent showed deterioration. After this date the practice was changed. 20 feet being- apparently adopt ed as a maximum. It was seldom, however, that more than 15 feet of acting head were employed, and as a rule it was found advisa ble to wash the filter when the loss of head was less than 10 feet. With a water which contained such a large amount of very finely divided suspended mat ter, as was the case during the last days of May, all the filters gave a poor effluent before there was any marked increase in the loss of head, as already noted. There were no well-marked indications that there was any difference in the action of the negative head of the Jewell filter, as com pared with the positive head of the other fil ters. By some the belief has been held that just after washing filters of this type, the effluent is of an unsatisfactory character, and that it is necessary to waste the first portion of the fil tered water. One of the most important points learned in these tests was that such was not necessarily the case, provided the sand layer was thoroughly washed and the applied water sufficiently coagulated. As a rule the sand layers in the Warren and Jewell systems were washed quite thoroughly, but more especially and uniformly so in the case of the Warren; and it was very rare during the later part of the tests that the first por tion of these effluents after washing was in ferior in appearance or character to the re maining portions. In connection with this important and practical point it is to be noted that both of the above systems employed con stant agitation of the sand layer during wash ing. The sand layers of both of these systems were usually washed until the filtered water, which was pumped into them from below the sand layer, was comparatively clear as it over flowed from the filter above the sand. All of the systems regularly employed fil tered water for washing the sand layer during these tests, except in the case of the Western Systems from June 24 to July 27, inclusive, when river water was used by these two sys tems. From the fact that the operators of the Western Systems abandoned the use of unfil- tered water for washing it may be inferred that its use was not wholly satisfactory. One reason why the Western Systems tried the use of unfiltered water for washing, ap parently, was that it was their uniform custom to waste the first portion of effluent after washing the sand layer. It seems very clear that this was required partly by incomplete washing and partly by incomplete coagula tion of the applied water. Several factors af fected the thoroughness of washing the sand layers of these filters. The most important one was the absence of accompanying agita- SUMMARY AND DISCUSSION OF DATA OF .895-96. 2 57 tion. Another important one was the unsat isfactory device used for the distribution of the wash-water beneath the sand layer. As already described in Chapter V, this was ac complished by means of slotted brass tubes, which also served for the collection of the effluent when the niters were in service. These slots were wedge shaped, with the smallest width on the outside forming the opening. It was expected that by this ar rangement the openings would not clog up, but that any matters small enough to lodge in them would pass through. This was in a measure found to be true at first, but as comparatively large numbers of sand grains passed through the slots the thin edges soon became worn, resulting in the passage through of more grains or in their lodgment in the opening. When the flow of water was reversed dur ing washing, some of the sand grains in the tubes were doubtless forced into the beveled openings and obstructed the Mow. When these strainer tubes were removed about the first of April many of them were found to be one-third full of sand. This state of affairs naturally caused the sand layer to be incom pletely washed, especially in the absence of agitation. The ball nozzles which were used for the distribution of wash-water in the West ern gravity filter (B), after the above ex perience with the slotted brass tubes, did not at the outset possess the action for which they were designed. This filter in its modi fied form was not operated long enough to demonstrate the practicability of these de vices. Sand samples from the various filters were collected from time to time and analyzed chemically and bacterially with the following results: RESULTS OF CHEMICAL AND BACTERIAL ANALYSES OF SANDS OF THE RESPECTIVE SYSTEMS. Nitrogen in the form of Albuminoid Ammonia expressed in Parts per Million by weight of dry Sand. Bacteria expressed in Numbers per Gram of dry Sand. Depth below Surface. 0-0.25 Inch. i Inch. Nitro-i Bac- gen. jteria. 3 Inches. 6 Inches. 12 Inches. 24! flics. Bot of Layer. Date. System. Nitru- gen. Bacteria. Nitre- (fen. ,-. Nii.ro Ren. Bac- Nitro Bac- Nitro- en. Bac- Nitro-.j, . ri gen. ! Before Washing. 1896 Jan. 20: Warren. . May 22 July 17 Jan. 8i Jewel Feb. 28 July 81 Jan. i6| Weste July 251 Jan. 12! Weste June 13 July 14 i Gravity .... 26.0 So.c 38.8 365.0 45.0 73-3 171.5 65 too 44 Soo 20. o 12.0 9.0 11. 8 6.0 31 5o 7 .6 S.o 35500 3-4 12 600 1. 9 56.8 21 .6 17.6 9-6 6.0 1 26 (XX) ;;;, 6.3 21 OOO 2..1 14 700 2-7 452 ooo i Pressure . . . 100.8 114.2 870000 6.1 14-4 9 600 8-5 9.6 n r> IIIOOO 13-4 II .2 21 IOO 80.5 23.6 IIO.O 43 7" t vfter Was . . . . 16.3 26 ioo .... thing- M.7 27300 I5.2i 46700 14 Soo " V oo 2.0 4 13.4* # 28 S* * | 2.4 2 1* # * 4-1 1 1 " I 800 Gravity 2-;. 8 n 5 5 s Pressure ... 1 1 . <; TO ?OO 8.8 5 &>-> 0-0 After steaming the sand. Sand was practically sterile. WATER PURIFICATION AT LOUISVILLE. Relation of Proper Attention to the Efficiency of Purification of the Ohio River Water by this Method. J his subject was touched upon in Chapter VII, where the manner of operation of the several systems was outlined. In connection with the factors which exerted an appreci able influence on the quality of the filtered water, this one must be clearly borne in mind. The impression which some people have that large systems of water purification by this method will at all times yield an effluent of satisfactory character with merely nominal attendance is wholly incorrect so far as the Ohio River water is concerned. In the first place both efficiency and economy require that very close attention be given to the ap plication of chemicals. Setting aside the question of cost, any excess of the chemical above the optimum is attended with increased amounts of corroding and incrusting con stituents in the filtered water, and, under un skilled supervision, chemical in excess of the amount whidh the water will decompose might be added at times, resulting of course in the inadmissible presence of imdecomposed chemical in the effluent. On the other hand a reduction of the amount of chemical by a small percentage below the optimum would cause immediate deterioration in the charac ter of the effluent, a deterioration which at times could not be determined for several days, as in the case of the Ohio River water a clear effluent is not necessarily a pure efflu ent, especially during the winter months. It will be seen, therefore, that the efficiency of filtration requires very close adherence to the optimum amount of chemical treatment, a problem which is very difficult of solution in the present state of the art, and the solu tion of which by unskilled hands is absolutely out of the question. Among the many other different factors which affect the quality of the effluent and which are to a greater or less degree dependent on the character of the at tendants for their efficiency may be men tioned the following: The decision to wash the sand layer, the process of washing, the opera tion of the many mechanical devices, and finally the general supervision and systematic methods of procedure without which no sys tem of this kind can be successful. With adequate facilities for the removal of the coarse matter by plain subsidence the amount of attention for successful operation would be largely reduced, because it would make the water just prior to its filtration much more uniform in character. SECTION No. 3. PROMINENT FACTORS WHICH INFLUENCED THE ELEMENTS OF COST OF PURIFICA TION OF THE OHIO RIVER WATER IN THE CASE OF THE RESPECTIVE SYSTEMS. The factors which will be considered in re lation to the elements of cost are substan tially the same as those noted in the last sec tion with regard to the quality of the efflu ents. In many respects the t\vo sections should be considered side by side. In the following pages of this chapter com parisons are made of the different factors as they were found at times of fairly clear and muddiest water and, so far as possible, under normal conditions, respectively. For this purpose averages are used as fol lows: For fairly clear water, averages during Period No. 13; for muddiest water, averages during Period No. 20, excluding those runs which were affected by the period of sub sidence over night or Sunday ; for normal water in most cases, averages for the entire investigation. When the Western Gravity System is referred to, Periods Nos. 8 and 9 are used in place of Periods Nos. 13 and 20. It must be borne in mind that the averages as obtained above do not represent the con ditions as they would exist under the actual times of muddiest water in the Ohio River, but they are taken as the best figures obtained during the investigation, and as sufficiently marked to illustrate the points under discus sion. Composition of the River Water. Inspection of the records for the several periods shows clearly that the amount of sus pended matter in the water exerted a marked influence upon the cost for the chemical. The extreme quantities of sulphate of alumina ap plied to the river water were 12.60 and 0.40 SUMMARY AND DISCUSSION OF DATA OF 1895-96. 2 S9 grains per gallon of applied water. At 1.5 cents per pound this would make on this basis the daily cost for the chemical in the operation of a plant of 25,000,000 gallons daily capacity range from $678 to $21 by these systems. The above figures, however, do not show the full influence exerted by the composition of the river water upon the cost of chemical in purification by this system. When the river water was comparatively clear the aver age percentages of applied water which was wasted and used for washing the sand layers in the Warren, Jewell, Western Gravity and Western Pressure systems were 6, 2, 9, and 4, respectively. When the water was in its- muddiest condition these average percentages became 34, 25, 99, and 58, respectively. In view of the fact that the evidence indi cates that the Western Gravity System (A) was unable at times of muddiest river water to purify enough water to wash its own sand layer properly, and that the data in regard to the Western Gravity System (B) are not sufficiently extended to permit the drawing of any comparisons, these systems will be omitted from further comparisons at this time. As will be seen by inspection of Table Xo. 4, and of the tables in Chapter VIII, where the full records are presented, the amounts of wash and waste water in the case of the Jewell and Western Pressure systems, exceeded at times TOO per cent, of the filtered water. The contiguous records indicate, how ever, that these results were abnormal and not likely to occur under regular conditions of practice. Figuring the amount of applied chemical upon the average quantity of net purified water, the following range of daily cost for the chemical for the purification of 25,000,000 gallons of Ohio River water is obtained: System. Daily Net Capacity in Gallons. By Contract. With Clear Water. With Muddiest Water. Warren 250 ooo 250000 250000 2OO OOO 266 ooo 238 ooo 1 2O OOO 160 ooo 64 ooo Western Pressure System. Daily Cost. Minimum. Maximum. Warren $So 06 *547 584 filfl Jewell Western Pressure... From the above-stated increased per centages of water wasted and used for wash ing at times of muddy river water, it follows that the appliances and their operation, for supplying this increased amount of water, would be factors in the cost of purification. It is to be noted, however, that under suitable arrangements, water which was wasted either before or after purification, might be purified subsequently. Water which was used for washing the sand layers, however, would probably have to be discharged into the sewer. The composition of the river water, with regard to its suspended matter, influenced the cost of purification by reducing the net capacity of the respective systems. This is shown by the following table, in which the actual average net capacities of each system as operated, is recorded in gallons per 24 hours at times of clear and muddiest water. The above data indicate the size of the re serve portion of the respective systems which would be necessary in order to obtain the full quantity of purified water when the river was in the muddiest condition, as shown by these tests. They do not indicate an adequate re serve portion for all conditions, because the Ohio River water contains at times more mud than was the case in any instance dur ing the tests of these filters. Kind of Chemical. Since sulphate of alu mina on an average contains about 60 per cent, more available alumina than potash alum, and the two chemicals are of approxi mately equal cost, the former is cheaper than the latter in the ratio of about 16 to 10. Preparation of Solutions. On the grounds of cost alone the preparation of chemical so lutions, in the process of application, should receive fully as much attention as was the case in these tests. In sojne instances econ omy demands more care in this particular than was regularly given to it in each of the systems. The chief point in this connection, WATER PURIFICATION AT LOUISVILLE. however, is to record the failure of the first Western device to yield solutions of even ap proximately uniform strength, such as econ omy demands in the treatment of the Ohio River water in its rapidly changing condi tions. Method ami Uniformity of Application. ir regularities in the application of the chemical were frequently so marked that they would affect the cost of operations on a large scale. They were more noticeable during the early part of the tests, before the operators of the respective systems were cautioned on this point in an official communication dated March 16, 1896, in which among other points their attention was called to such irregulari ties. The \Yarren device was more satisfactory than the others, all things considered, because it was most nearly automatic. It had several crude features, however, and it is by no means clear that its use would be thoroughly satis factory in a large system. The Jewell and second Western devices were satisfactory provided they received suf ficient attention. At different times during- the same day afid with the same water the rate of application varied several fold. This means that if the minimum rate of application was sufficient for its purpose the higher rates caused a waste of chemical equal to their excess over the minimum. It may be noted that the use of sufficient chemical to insure proper coagulation, and a sufficient amount of aluminum hydrate in the water as it passes onto the sand layer, are absolutely essential for the success of this method of purification, and that the use of an insufficient amount of chemical is out of the question. To do this on a large scale with the devices submitted for investigation would be less easy than would be thought at first to be the case. The first device of the Western Company for the application of chemical was a failure as operated at the beginning of these tests; and it was abandoned shortly after the official communication of the Water Com pany on March 16. as mentioned above. The cost of power for the application of the chemical depends of course on the.metho l of application used and the strength of solu tions, but with anv of these devices it would be comparatively insignificant. With the Warren device the power would be practically only that required to lift the chemicals and water to the mixing tanks. The first Western device required substantially no increase in power over that required for handling the river water. The Jewell and the second West ern devices both involved the pumping of the solution against the full pressure in the mains. Assuming an average percentage strength of one per cent, and an average amount of chemical of 2.50 grains per gallon, and that the necessary water and the chemical were de livered on the level of the main house tloor, the approximate amount of power required on a basis of 25,000,000 gallons per 24 hours would be in each case as follows: Warren System 0.4 H.P. Jewell System 4.0 H.P. Western Press. System. 4.0 H.P. (2(1 device) In the case of the Warren device a very small amount of power was used in turning the propeller wheel which operated the chemical pump. The power thus used was probably only a very small fraction of the total power used in the application of the so lutions, and was furnished by the velocity pressure in the water supply pipe. The principal element of cost of purifica tion of the Ohio River water by this method would be the sulphate of alumina used for coagulation. That is to say. this element would exceed any other one. including the interest on the cost of construction of the system. The cost for sulphate of alumina would of course be proportional to the amount, and the amount used would depend upon a series of factors, of which the follow ing are the most important: 1. Composition of river water, with regard to the quantity and character of suspended matter. 2. The optimum quantity of coagulant under the given conditions. The chief vari able factor affecting this was the composition of the river water, and the chief fixed factor was the period of coagulation and sedimenta- SUMMARY AND DISCUSSION OF DATA OF 1895-96. tion. In practice plain sedimentation should precede coagulation, and reduce the required amount of coagulant. 3. Relation of the actual to the optimum quantity of applied chemical necessary to secure the proper degree of coagulation in the water during and after subsidence. A brief discussion of the several factors will be found in this chapter in their logical or der. Application of Lime. The application of lime as tried in an ex perimental way by the Jewell System in creased, apparently, the cost of proper treat ment, disregarding the cost of the lime itself, because it seemed to diminish the coagulating power of the resulting aluminum hydrate. Under other and better conditions of ap plication this might not be the case. In this connection, see the results of comparative ex periments recorded in Chapter XII, on the coagulating power of aluminum hydrate pre pared in several different ways. Provisions for Coagulation and Sedimentation. In relation to the quality of the effluent it was indicated that in this respect all of the sys tems were very weak, although the Warren was least so. When it comes to a question of cost this weakness of all the systems in their present form would make their adoption ex pensive to an unnecessary degree, and there fore of questionable admissibility. The merely nominal period of subsidence with coagulation in the Jewell and Western sys tems, and for one hour or less in the \Varren System, and with no plain subsidence in any case to remove coarse matter, materially in creases the cost of purification as follows: 1. It increases the cost for chemical. 2. It necessitates a reserve portion of the system with all appurtenances to handle the water when in its muddiest condition. 3. It necessitates the waste of an unusually large amount of filtered water for the purpose of washing the sand layers. 4. The increase of water wasted as indi cated above increases the amount of water to be pumped, and therefore the aggregate cost of pumping. 5. The use from time to time of a compara tively large reserve portion of the system would require the constant employment of a full set of trained attendants, capable of op erating the entire system. This would be necessary because the freshets in the Ohio River frequently appear in a most irregular and unexpected manner. Inspection of the records of the freshets in the Ohio River during the past thirty-five years, presented in Chapter 1, shows that during many years the river was in a state of flood for longer and more frequent periods than was the case during these investigations. This means that in many cases the river water contained more suspended matter than was encountered during these tests, and there fore the five factors of cost stated above would be correspondingly increased. In this connection it is not to be forgotten that the Western Gravity System as first de signed was voluntarily taken out of service by the Western Filter Company on March 21, because it w r as unable to yield enough filtered water to serve for wash-water. The investigations demonstrate conclu sively that economy demands plain subsidence suplemented by a considerable period of coagulation and subsidence, followed at times by further application of chemicals to effect coagulation for filtration, in the case that high rates of filtration should be employed. With regard to the best manner of carrying into practice such an improvement there are very few specific data to serve as a guide, as was first pointed out to you in a general way in a preliminary report dated July 11. 1896. The closing pages of Chapter IV contain the only evidence to offer upon the subject which was obtained in 1896. Much additional evidence along this line was obtained in 1897, and is recorded in Chapter XV. Provisions for Cleaning the Settling Basin or Chamber. This matter was not economically handled in the case of any of the systems. Except in the case of the Jewell it was almost ignored, practically speaking, and in this sys tem the provisions were inadequate for economical use. Provisions for Inspection of the Condition of the Settling Basin or Chamber. This factor was apparently lost sight of in all of the svs- 262 WATER PURIFICATION AT LOUISVILLE. terns, and the failure to determine the con dition of the contents of these compartments and if necessary remove the accumulation of sludge, undoubtedly led at times to the pas sage upon the filters of mud which should have been held in the settling basins. The effect of this passage of mud on to the filters on the cost of operation was to decrease the length of runs and therefore to increase the percentage of water wasted and used for washing. The most notable occurrence of this kind took place in the Warren System on July 22 to 27. Degree of Coagulation of the Water as it Entered the Sand La\cr. One of the most clearly established points in connection with these tests was the abso lute necessity of thorough coagulation of the water as it entered the sand layer. With dif ferent characters of water the required de gree of coagulation varied somewhat, but it was made perfectly clear that with the high rates of filtration employed in these filters of the American type a high degree of coagula tion is very essential. In the adoption of a system of 25.000,000 gallons capacity, to last for many years, it is very questionable whether wooden structures as employed in the Warren, Jewell, and West ern Gravity systems would be advisable in preference to metal. In all of the systems it would be very desirable and probably prac ticable to make the more important parts of the filters, such as strainer systems, more readily accessible. The unsatisfactory results from the use of wood was illustrated by the foul and slimy deposits upon the walls of the filtered-water compartment beneath the sand layer of the Warren filter when that system was removed after the close of the tests. This particular construction could and should be improved. Tn the filters of both the Jewell and the Western systems, but especially in the case of the latter filters, there was a considerable stick ing together of the sand grains at the bottom of the sand lavers. This segregation of the sand grains was evidently associated with the use of cement, and it is possible that this might lead to serious difficulty in time. Sand Layers of the Several Filters. The more important data upon the several sand layers, as they appeared at the close of the tests, are reported in the foregoing section in relation to the quality of the effluent. As stated there, the various data were so compli cated by a series of factors that it is impos sible to draw conclusions with regard to sev eral points of great practical significance. The more important points are noted in turn as follows: Thickness. While the Jewell filter was able to yield the most economical results there is no proof that it was the best one which could be adopted with regard to thick ness. Size of Sand Grain. In the case of the Jewell sand layer the most economical results were obtained. But in the case of size of grain as well as thickness of layer there are no data to show what would be the most eco nomical conditions to adopt in practice. Tak ing everything into consideration, especially the frequency of tiny flakes of aluminum hy- drate in the Jewell effluent, it is probable that a finer size of grain would be more advan tageous. Composition of Sand. It appears from the available evidence that as satisfactory results may be obtained under suitable conditions from natural sand layers as from those made from the more expensive crushed quartz. The location of the sand layer in the upper part of the filter, with a depth of three feet or less of water above the sand, such as was the case in the Jewell and Western gravity (B) filters, is a marked step in advance. By the use of a trap, and a suitable location of the point of discharge of the effluent, the total available head may be undiminished and at the same time the following economical ad vantages may be insured: i . The difficult and in a measure impossi ble task of satisfactorily filtering all of the SUMMARY AND DISCUSSION OF DATA OF 1895-96. 26.1 water remaining above the sand just prior to washing is readily removed under normal conditions. 2. The quantity of chemically treated water, which it is necessary to remove before washing and to either waste or pump again, is materially reduced. The second weakness \vas most noticeable in the Western gravity filter (A). There is no evidence, however, to indicate that the use of a negative head (suction) has any advantage other than those stated above. Loss of Head. Initial. The initial loss of head is an in fluential factor in the cost of operation of a system of purification, in that the available head is reduced during filtration by the amount of the initial loss of head. It is also indirectly connected with the successful prac tice of surface agitation. The initial loss of head was determined mainly in the case of these systems by the resistance of the several sand layers. In all cases the strainer systems when clean offered apparently no measurable resistance to the flow at the contract rate. In the case of the two Western systems, the pres ence of sand in the strainer tubes as above noted probably increased the initial loss of head to a greater or less degree. Maximum. The maximum loss of head (maximum acting head) which can be utilized is an important factor in the cost of opera tion, in that it influences the length of runs (period between washes) and thus affects the relation between the actual and net rates of filtration. In this respect the respective filters be haved very differently. The Warren and the Western pressure filters were practically al ways washed because of the failure in charac ter of the effluent and not because the resist ance of the sand layers required greater available head than the construction allowed. The reverse was true, except in cases of pe culiar conditions of the river water, with the Jewell filter. This matter was discussed in .the preceding section of this chapter, in re gard to its effect on the quality of the efflu ent. A very noticeable point in this connec tion was mentioned there, that is, that during Loss of Head in >et. j Initial. Maxii num. Normal. Extreme. Jewell I 6 13 6 65 4 the last of May the composition of the river water was such that, with the amounts of chemicals used, none of the systems was able to obtain a degree of coagulation of the water as it entered the sand layer, sufficient to clog appreciably the layer before the character of the effluent failed. Almost without exception during this period the filters were washed without any appreciable increase in loss of head over the initial. A comparative idea of the relative signifi cance of these factors under the contract rate of 250,000 gallons per twenty-four hours is indicated in the following table: Pressure System. One of the economical advantages claimed for pressure systems as compared with grav ity systems is that they would lessen the cost of purification by removing the necessity of a secondary pumping of the water if such should be the case with gravity systems. This would depend upon local conditions, and under some circumstances it might be true. But the experience obtained during these tests indicates clearly that, with the muddy Ohio River water at Louisville, a single pumping of the river water to and through a pressure system is out of the question on the grounds of unnecessary cost. The excessive cost attending the use of a pressure system would be caused by the large closed compart ments which it would be necessary to insert between the main (primary) pumps and the pressure filters, in order to secure, under the existing conditions, the removal of mud, etc., which the economical treatment of this water before its filtration demands. It may occur to some that a combined pres sure and gravity system might be desirable for the purification of this water supply. That is to say, pump the water from the river 264 WATER PURIFICATION AT LOUISVILLE. to an elevated open subsidence basin of ad equate size, and then allow the propeny treated water to flow through pressure niters on its way from the subsidence basin to the consumers. From the experience obtained with the Western pressure filter the adoption of this scheme would lead to serious difficul ties, owing to irregularities in operation aris ing from variation in the rate of consumption at different hours (and minutes) of the day, such as the ordinary necessities of the con sumers demand. Unusual rates of consumption such as might occur in putting out large fires, etc., would increase these difficulties. That irregu larities such as would occur in this way were a very serious matter in the operation of the Western System was shown conclusively, and is so indicated by the official communication received from the Western Filter Company on June 26, in explanation of the withdrawal from these tests of their gravity system for a period of three months. This letter was given in an earlier part of this chapter, in the de scription of Period No. 18. In passing it may be mentioned that all experiences in water fil tration, with which the writer is familiar, point clearly to the advisability, if not to the necessity, of placing a reservoir, not neces sarily large, between the filters and the dis tributing system in order to maintain as uni form pressure as possible in the pipes and to guard against irregularities in the operation of the filters, with their attending difficulties. Rate of Filtration. The rate of filtration is a very prominent factor in the cost of construction of a large system. It also affects the cost of operation. The available data can only be regarded as suggestive upon this point. The fact that the Western pressure filter, however, yielded for comparatively long periods at a time an efflu ent which compared favorably with the others in character, at a much higher rate per unit of sand surface, is a matter which cannot be ignored or considered lightly. This is especially true when it is remembered that the Western pressure filter was operated in the face of many complications, including irregu lar coagulation, a faulty strainer system, and Filter. Rate of Filtration. Normal Clear Water. Muddiest Water. Cubic Feet per Minuie. Million Gallons per Acre per 24 Hours. Cubic Feet per Million Gallons per Acre per 20.8 26.3 23.1 126 1 06 164 20.1 22.2 "3-7 122 yo 97 Jewell Western Pressure. . absence of agitation of sand layer to secure complete washing, etc. The actual average rates of filtration em ployed in the case of the several filters with fairly clear and muddiest water, respectively, are as follows: The contract rate of 250,000 gallons per twenty-four hours is equivalent to 23.21 cubic feet per minute. Owing to different areas of filtering surface this rate per unit of area dif fered for the several filters, as follows: Filter. Area in Square Feet. Contract Rate in Million Gallons per Acre per 34 Hours. Warren 141 Jewell 94 Western Pressure. . 66.20 57 That comparatively wide ranges in unit rates did not affect the cost for chemicals by a corresponding amount will be seen by an in spection of the tables given in connection with the discussion of the effect of rate of fil tration on the quality of the effluent in the preceding section. Some further investiga tions along this line were made by the Water Company during 1897, and are recorded in Chapter XV. Wasliing the Sand Layer. The washing of the sand layer was a con siderable item in the cost of operation of the respective systems, for two reasons: 1. The net rate of purification was reduced as the frequency of washing increased, because of the increased percentage of water wasted and used for washing. 2. The cost of the operation of washing was a considerable item, and this was also in creased in total expense as the number of washes increased. The chief factors affecting the frequency SUMMARY AND DISCUSSION OF DATA OF 1895-96. of washing have been discussed in the pre ceding pages. It was mainly determined by the relation between the degree of coagula tion of the water as it entered the sand layer, type of sand layer, rate of filtration, character of filtered water, and available acting head. The relation of the agitation of the surface of the sand layer to the frequency of washing is discussed beyond. The cost of the operation of washing was dependent upon the amount of water used, the pressure at which it was delivered, and the cost of agitation of the sand, if employed. Amount of Water Used. The comparative amounts were determined entirely by the method of washing employed. In the case of the Warren, Jewell, and Western Pressure systems, the average quantities for each wash during the entire investigation were 528, 627, and 633 cubic feet, respectively. Mctliod of ll its/iing. All indications point to the conclusion that the most economical method of washing is to carry the process to a point where all detachable materials are re moved from the sand, but no further. This means that it seems best to wash the sand until the wash-water after passage through it is practically clear. The amount required depends principally upon the amount and character of the matter accumulated on the sand grains, and upon the relative efficiency of equal quantities of water. The latter factor is dependent upon the dis tribution of the water throughout the sand layer and upon the agitation of the sand layer, notably the rubbing together of the sand grains by the agitator teeth. During the latter part of these tests the Warren filter was uniformly washed to a sat isfactory degree. This was the case as a rule in the Jc\vell filter, but not uniformly so. The Western pressure filter was almost never washed as thoroughly as it should have been. This was probably the chief reason why it was necessary in this system to waste the first por tion of the effluent after washing, owing to its unsatisfactory character. The amount wasted was usually only a small percentage of the total water filtered on the run. but during muddy conditions of the river water the amount of filtered water wasted became pro portionately large and at times exceeded in amount the quantity of satisfactory effluent obtained. Distribution of W atcr throughout the Sand La\er. The distribution of the water during washing was affected by the agitation of the sand layer and in turn reduced the cost of agitation. The main factor affecting the dis tribution was the system used for this pur pose. The Warren distributing system was handi capped for a time during the early part of the tests by having an undue portion of the wash- water deflected from the central well through a small area of the strainer system and sand layer. This was apparently remedied to a large extent by the changes made on Feb ruary 12, 1896. The distributing system of the Jewell filter was apparently quite satisfactory. The most notable points about this device were the re striction of the neck of the strainer cups, and the small deflector in the cups just above the neck. By the first arrangement the greatest resistance to the passage of the water was met at this point, thus causing a distribution of the water throughout the entire system. The small casting which was placed in the cups just above the neck served to break and de flect the stream of water just before it entered the sand layer. The distributing systems of the Western pressure and Western gravity (A) filters were handicapped by the presence of sand in the tubes, as was noted in the last section. It is difficult to determine how far these dis tributing systems affected the quantity of water used, as these filters were never washed thoroughly. In connection with the distributing systems of the Western filters, it is to be noted that when they were clean there was less restric tion to the passage of the water through the distributing system than in the supply pipe. Such an arrangement naturally involved the passage of the water most rapidly through those portions of the system nearest the con nection with the supply pipe. The accumu lations of sand in the strainer tubes reduced the total outlet area and therefore increased the resistance of the tubes. These accumula- 266 WATER PURIFICATION AT LOUISVILLE. tions, furthermore, increased the tendency of the water to pass into comparatively small portions of the sand layer. The ball-nozzle system of the Western gravity filter (E) \vas not in use for a suf ficient length of time to determine its relative efficiency. Observations during and after construction, however, indicated unequal dis tribution of the water at different parts of the sand layer. The mechanical agitation of the sand layer during washing greatly aided in distributing the wash-water and increased the relative efficiency of equal quantities of water. The Jewell and Warren systems used mechanical devices for agitating the sand throughout the tests. The current of water was relied upon for agitation in both the Western filters. In this connection it is to be noted that a modi fication in the mechanical agitators whereby the sand would be floated to a less degree and the grains rubbed together more than was the practice in these tests suggests an economical advance, as equally satisfactory- results might be obtained with the use of less wash-water. Pressure of Wash-mater. The pressure under which the water was delivered at the inlet of the distributing systems is the second factor in the cost of the operation of washing. This was widely different in the case of the several filters at times during the tests, on account of changes in other factors. The ef ficiency of the various . pressures used de pended largely upon the amount and dis tribution of the resistances of the distributing systems. The increased resistance of the dis tributing systems of the Western filters, due to the presence of sand in the tubes, was clearly shown by the increases made from time to time in the pressure employed. The use of mechanical agitators greatly decreased -the pressure required, as was shown by the in crease in pressure used by the Jewell filter, as the efficiency of the agitator decreased. Owing to these and other factors it is difficult to estimate the pressure necessarv under nor mal conditions of operation. The following table gives the pressures which were used at the close of the tests, in pounds, of the water at the inlet to the several distributing sys tems: 2.O 7-5 5-9 Agitation of the Sand Layer. in the West ern pressure filter the sand layer was never agitated except by the current of wash-water unless the removal of sand to repair the strainer system be so regarded. Agitation was regularly employed in the Warren and Jewell sand layers. All indications point to a decided advantage in the constant agitation during washing. The Warren agitator was changed and re paired several times, and during the later part of the tests appeared to fulfill its purpose. Considerable difficulty was experienced with the Jewell agitator on acount of its stopping and refusing to work at frequent intervals. At first this was thought to be due to the use of too small an engine (nominally 5 H.P.) or, perhaps, to some obstructions of the strainer system. Careful inspection, however, led to the conclusion that it was caused by a binding of the gears, due to a warping yf the timbers upon which the agitator rested. This could be readily remedied by a more careful con struction. The devices for agitation could and should be improved in simplicity of construc tion, and in both the Warren and Jewell sys tems the devices were too weak for their pur pose. As noted above, an improvement in these devices whereby the sand grains would be rubbed together more energetically would probably result in a saving of wash-water. Pou<cr Used for Agitation. At the close of the tests, with the pressure and quantities of wash-water then in use, the power required to operate the agitators of the Warren and Jewell systems was as follows: Warrer Jewell. The maximum power required occurred in the Warren at the time of lifting the rakes and in the Jewell at the time of forcing the rakes into the sand and starting them in mo tion. SUMMARY AND DISCUSSION OF DATA Ol< 1895-96. 267 Surface Agitation. In the case of the Jewell System it was found that with certain conditions of river water, and of its coagulation, the resisting layer of mud on the surface of the sand could he broken up and filtration then continued without washing the filter or injuring the character of the effluent. This operation, when successful, reduced the resistance of the sand layer and so lengthened the run. It was therefore an element of more or less magni tude in the consideration of cost, in that it decreased the frequency of washing. The success of surface agitation was very closely dependent on the degree of coagulation of the water as it entered the sand layer, and on the character of the sand. In these tests the use of surface agitation at times of very muddy water, or when the river water contained large amounts of fine clay, was not as a rule attempted, and when tried was not successful. The cause of the failure seems clearly to have been the incomplete coagula tion of the water at these times. Relation of Proper Attention and Supervision to the Econoin\ of Purification of the Ohio River Water by this Method. This subject has been referred to in Chap ter VII, and again in this chapter in relation to the efficiency of filtration. The most marked effect of proper supervision of the op eration of these systems was on the cost of treatment. As has already been presented, there is at all times a certain optimum amount of chemical, below which satisfactory results cannot be obtained, and above which all chemical used is practically wasted. In the light of our present knowledge the deter mination of this optimum amount, for such a rapidly and widely varying character of water as that of the Ohio River, is a very difficult problem. At times the optimum amount could be very clearly determined by one thor oughly familiar with the methods of proced ure, while at other times, especially with water which would give a clear effluent con taining large numbers of bacteria, the de cision required judgment based on extended experience. A system of purification of the Ohio River water is clearly one of combinations of methods and devices, which experience has demonstrated cannot be handled economi cally by unskilled labor. It would be easv for untrained attendants to waste many thou sands of dollars annually by the needless use of the chemical. A comparison of the average daily cost of the chemical used during each of the periods by the Warren, Jewell, and Western pressure systems, on a net basis of 25 million gallons daily, is represented in the following table and shows this point. DAILY COST FOR CHEMICAL. Period. Warren System. Jewell System. Western Pressure System. I V $37.00 2 67 . oo 49 .00 3 22g.OO 134.00 f 1 5 2 . oo 4 231 .OO 52.00 57.00 c. 26 1 . OO 225.00 1 20.00 6 252.CO 1 29 . oo 201 .00 7 226.OO 1 5 5 . oo 134.00 8 208.00 60.00 46.00 9 423 oo 264.00 239.00 280.00 I 99 . oo 204.00 I 324.00 76 .co 269.00 76 .00 224.00 3 80.00 96.00 59.00 4 76.00 69.00 64.00 5 295.00 300.00 395.00 6 235.00 300 . oo 288.00 7 152.00 267.00 275.00 8 2Q7.OO 437-0 548.00 9 1 79 . oo 354.00 335.00 20 416.00 575-oo 536.00 In less skilled hands these variations would probably have been more marked. With an adequate employment to its economical limits of subsidence both with and without coagu lation, the necessity for variation in the amount of applied chemical would be much less, and the opportunity for departure from the optimum amount would be reduced materially. Nevertheless, there would be no condition where this river water could be economically purified except by skilled labor and supervision. Jt is true that the above data are compli cated by other factors, some of which vary from time to time, but they serve to illustrate the point in question to a very considerable degree. WATER PURIFICATION AT LOUISVILLE. SECTION No. 4. COMPARISON OF THE ELEMENTS OF COST OF PURIFICATION OF TWENTY-FIVE MIL LION GALLONS OF OHIO RIVER WATER DAILY BY THE RESPECTIVE SYSTEMS, BASED ON THE RESULTS OF THESE IN VESTIGATIONS. in the following pages are given the ele ments of probable cost, so far as it is feasible, of the purification of 25 million gallons of Ohio River water daily by each of the sys tems representing the method in which co agulation and partial sedimentation by alu minum hydrate formed from sulphate of alumina, and subsequent rapid filtration, were employed. As a matter of convenience the elements of cost are subdivided into those of construction and those of operation. All of these estimates are based on the results ob tained from these investigations, upon the op eration of small test systems contracted to purify 250,000 gallons per twenty-four hours. From the nature of the existing conditions at this time, and in the absence of definite knowledge as to the cost of the various de vices, these estimates of necessity deal for the most part with elements of cost. They are so arranged that when the exact cost of the several devices is known, the aggregate cost may be readily computed. Wherever the ex isting conditions permit of it, actual estimates of cost are given. The following data are summaries of the principal elements of cost of the respective svstems as demonstrated in the previously de scribed investigations. In so far as possible the several amounts have been determined and estimated on the basis of a normal or slightly muddy river water containing about 100 parts per million of suspended matter, and also for a fairly muddy river water containing about 1800 parts. The majority of the com parisons call for normal and maximum fig ures, but in some cases representative aver ages for these tests are required. In the latter instance the data presented in Table No. 5 are used. Under the conditions of proper preliminary treatment before filtration by sub sidence with and without the aid of coagu lants, it might be expected that the water which reached the niters would compare with the normal river water as used in these sum maries. On the other hand, the figures given for the muddiest water do not represent the extremes which would be obtained under con ditions of actual muddiest water in the Ohio River. Such conditions are. however, un usual and of comparatively short duration, and for these data reference is made to the tables of individual runs which were pre sented in Chapter Yll I. As a rule the figures given as maximum in the following sum maries are averages for Period No. 20, ex cluding those runs which were affected by the period of subsidence over night, or were otherwise abnormal. It will be seen by refer ence to Chapter VIII that on July 24, 1896, the amount of suspended matter in the river water was 3347 parts per million, or nearly double the amount which is considered as the average muddiest water. (The maximum amount of suspended solids found during the entire investigations was 5311 parts per mil lion on March 6, 1897.) These excessively high amounts would probably never reach the filter in practice, where proper provision for preliminary plain subsidence was made; and, with the use of coagulation and subsidence as mentioned above, the water reaching the filters would probably not be excessively muddy. The Western gravity filter (A) does not appear in these comparisons because it was found to be unable to purify at all times enough water to wash its own sand layer, and its operation was discontinued by the West ern Filter Company. The Western gravity filter (B) was not operated long enough to yield adequate data, but there are no indications of its being com parable to the Warren or Jewell filters. As will be seen on examinations of the tables in Chapter VI 1 1, the percentage which the wash and waste water was of the applied, exceeded 100, three times in the case of the Western pressure filter and once in the case of the Jewell filter. As the contiguous results indi cate that these percentages were abnormal, they have not been considered as liable to i occur in regular practice. SUMMARY AND DISCUSSION OF DATA OF 1895-96. Summaries of Elements of Cost. 269 NOKMAI. PERIOD OK SKKYICK OK THE FILTERS OK THE RESPECTIVE SYSTKMS BETWEEN WASIIKS. HOURS AND MI.NUTHS. System. Ordinary Water. Muddiest Water. Warren ioh. 48111. 2h. oSm. Jewell ifih. 38111. 2h. 17111. Western Pressure 1311,38111. ill. 3801. NORMAL PERIOD REQUIRED KOR WASHING THE SAND LAYERS OK THE RESPECTIVE SYSTEMS. MINUTES. System. Average. Warren 26m. NORMAL PERIOD USKD FOR WASTING UNSATISFACTORY FILTERED WATER AFTER WASHING THE SAND LAYERS OK THK. RESPECTIVE SYSTEMS. MINIFIES. System. Ordinary Water. Muddiest Water. Warren o o Jewell o o Western Pressure 14 NORMAL QUANTITY OK RIVER WATER APPLIED TO THE RESPECTIVE SYSTEMS, IN GALLONS PER 24 HOURS. System. Ordinary Water. Muddiest Water. Warren .... 206 ooo 182 oco Jewell 271000 213000 Western Pressure 248500 152000 NORMAL QUANTITY OF COAGULATED AND PARTIALLY CLARIFIED WATER WASTED BY THE RESPECTIVE SYSTEMS PRIOR TO WASHING THE SAND LAYERS, IN GALLONS PER 24 HOURS. System. Ordinary Water. Muddiest Water. Warren 960 14 ooo Jewell o 4 ono Western Pressure o o NORMAL QUANTITY OK FILTERED WATER USED IN WASHING THE SAND LAYERS OF THE RESPECTIVE SYSTEMS, IN GAL LONS PER 24 HOURS. System. Ordinary Water. Muddiest Water. Warren ... 10 800 47 200 Jewell 6 2oo 49 200 Western Pressure. ... 8 500 71 700 NORMAL QUANTITY OK FILTERED WAFER WASTED BY THE RESPECTIVE SYSTEMS. OWING TO UNSATISFACTORY AP PEARANCE, IN GALLONS PER 24 HOURS. System. Ordinary Water. Muddiest Water. Warren o o Jewell o 300 Western Pressure 2 200 16 300 NORMAL NET QUANTITY OK FILTERED WATER (EXCLUSIVE OK WASH WATER AND WASTE WATER) YIELDED HY THE RESPECTIVE SYSTEMS, IN GALLONS PER 24 HOURS. System. Ordinary Water. Muddiest Water. Warren 200 ooo 1 20 ooo Jewell 266 ooo if)O ooo Western Pressure 238 ooo 64 ooo NORMAL NET QUANTITY OF FILTERED WATER (EXCLUSIVE OF WASH WATER AND WASTE WATER) YIELDED BY THE RESPECTIVE SYSTEMS, IN MILIION GALLONS PER 24 HOURS PER ACRE OK FILTERING SURFACE. NORMAL PERCENTAGE WHICH THE SUM OK THE WASH WATER AND WASTE WATER FORMED OK THE RIVER WATER APPLIED TO THE RESPECTIVE SYSTEMS. System. Ordinary Water. Muddiest Water. Warren 6 34 Jewell 2 25 Western Pressure 4 58 NORMAL RATE AND PRESSURE AT WHICH THE FILTERED WATER WAS SUPPLIED FOR WASHING THE SAND LAYERS OF THE RESPECTIVE SYSTEMS. if -,,.. ;., r.,11^,^ Pressure in Pounds per System. ncrMinuu Square Inch at the Bottom of the Sand Layer. Warren 460 2 . o Jewell 530 7.5 Western Pressure 650 5.9 NORMAL QUANTITY OF APPLIED SULPHATE OK ALUMINA IN GRAINS PER GALLON OF RIVER WATER SUPPLIED TO THE RESPECTIVE SYSTEMS. System. Ordinary Water. Muddiest Water. Warren 1.41 6.77 Jewell 1.76 8.74 Western Pressure 1 .06 5-27 NORMAL QUANTITY OK APPLIED SULPHATE OF ALUMINA IN GRAINS PER GALLON OF NET FILTERED WATER YIELDED BY THE RESPECTIVE SYSTEMS. JO. 20 3.00 10.86 2.65 I 2 . 60 2.72 System. Warien Jewell Western Pressure Using the foregoing data as a basis of com putation, the following principal elements of cost of installation and operation of a system of 25 million gallons daily capacity are pre sented, with actual estimates of cost wherever it is feasible. NUMBER OF THE RESPECTIVE UNIT SYSTEMS WHICH WOULD BE NECESSARY TO FURNISH 25 MILLION GALLONS DAILY OF PURIFIED WATER AT TIMES OF ORDINARY RIVER WATER. Warren System 1 24 Jewell System 94 Western Pressure System 105 NUMBER OF THE RESPECTIVE UNIT SYSTEMS WHICH IT- WOULD BE NECESSARY TO HOLD is RESERVE IN ORDER TO SUPPLY 25 MILLION GALLONS DAILY OK PURIFIED WATER AT TIMES OF MUDDIEST RIVER WATER. Warren System 85 Jewell System 62 Western Pressure System 287 TOTAL NUMBER OF THE RESPECTIVE UNIT SYSTEMS WHICH WOULD BE REQUIRED TO SUPPLY UNIFORMLY 25 MILLION- GALLONS DAILY OF PURIFIED OHIO RIVER WATER. Warren System 208 Jewell System 156 Western Pressure System 391 RATE IN CUBIC- FEET PER MINUTE AT WHICH THE WASH WATER WOULD HAVE TO BE SUPPLIED TO WASH THE SAND LAYERS OK A 25 MILLION GALLON PLANT. Warren System 2650 312 Jewell System. 1775 177 Western Pressure System 8 560 219 WATER PURIFICATION AT LOUISVILLE. The head against which the pump would have to operate in furnishing the wash-water would depend largely upon the relative loca tion of the pump and the different niters with reference to the source of supply, but it would be such that the available pressure on the bottom of the sand layer would be about 5, JO, and 8 pounds for the Warren, Jewell, and Western Pressure systems, respectively. There would be required, in the case of the Warren and Jewell systems, engines to fur nish the power necessary to operate the agi tating machinery when the sand layers were being washed. On the basis of the above data, the maximum and average amounts of power required for this purpose would be as follows: Maximum. Average. Warren System 383 H.P 45 H. P. Jewell System 2So " 25 In addition to the above elements of cost of installation of a 25 million-gallon plant there would be the cost of suitable preparation of the grounds upon which to locate the system, and also the buildings to house the niters. The area occupied by the total number of required unit systems on the above basis would be as follows: With regard to the cost of application of the sulphate of alumina, the very small amount of power required depends largely upon the location and arrangement of the system, strength of solution used, etc., in the case of all of the systems. Another important factor connected with the cost of operation of such a system of puri fication, and also with the installation, is the extra pumping of the supply of water. To a great extent this factor is dependent upon other details of construction and would prob ably exceed somewhat the minimum. A very close idea can be obtained of the importance and significance of this factor by considering simply the difference in level of the water above and below the sand layers of the several filters (loss of head). Taking these figures as presented in foregoing tables, 4.5, 13.6, and 65.4 feet for the Warren, Jewell, and Western Pressure systems, respectively, and adding to the total net capacity of 25 million gallons daily the average percent ages of water wasted and used for washing the sand layers, 6, 2, and 4 per cent, for ordinary water, and 34, 25, and 58 per cent, for mud diest water for the three systems in the order above given, the following amounts of power required are obtained: In regard to the Western Pressure System it is only fair to state that with the normal maximum loss of head of 20 feet these figures would be reduced to about 30 per cent, of the figures given. The question of the insertion of pressure filters in the direct line of pipe from the main pumps to the reservoir or from the reservoir to the city has already been shown to be out of the question; and it is only necessary to add that the cost of in creased pumping is represented by the above figures no matter where the filters are located. The principal cost of operation would be that of the sulphate of alumina required for coagulation of the river water. Using the present quotations of 1.5 cents per pound for commercial sulphate of alumina delivered in carload lots free on board cars at Louisville, the estimates of cost based on the above data are as follows: DAILY COST OF SULPHATE OF ALUMINA IN THE PURIFICA TION OF 25 MILLION GALLONS OF OHIO RIVF.R WATER BY THE RESPECTIVE SYSTEMS. Warren System Jewel] System Western Pressure Svsteni 547 584 676 142 146 FICATION OF 2^ MILLION- GALLONS OF OHIO RIVER WATER DAILY HY THE RESPECI m<: SYSTEMS, RASED ON THE QUANTITY OF WATER IN GALLONS PER DAY, WHICH WOULD BE WASTED AND USED FOR WASHING THE SAND LAYERS OF THE RESPECTIVE SYSTEMS IN THE PURIFICATION OF 25 MILLION GALLONS DAILY OF THE OHIO RIVER WATER. Wa Experience during these tests showed clearly enough that to purify 25 million gallons of the Ohio River water daily in SUMMARY AND DISCUSSION OF DATA OF 1895-96. 271 all its varying stages and conditions, -without wasting sulphate of alumina and at the same time giving a purified water of a satisfactory character, was absolutely out of the question in the absence of constant care and skillful supervision. The attention given to each sys tem of a rated capacity of 250,000 gallons per twenty-four hours was of course several times greater than would be necessary in a portion of corresponding si/e in a system having a capacity of 25 million gallons daily. The amount and scope of necessary analytical work would also be much modified in actual practice, especially after a large system had been in operation for a sufficient period for the formulation of a practical and systematic method of procedure. The best idea which can be given you at this time as to the cost of the necessary atten tion for the operation of a system to purify 25 million gallons of the Ohio River water daily is afforded by the statement that it would certainly not be less than that for the proper operation and maintenance of your present pumping station, which I understand is 14,000 dollars per annum. SECTION No. 5. GKNKRAL CONCLUSIONS. The practical results of these tests, as ap plied to the problem of purifying the Ohio River water for the supply for the city of Louisville may be summed up in the follow ing manner, in which reference is made to the general applicability of the method investi gated and to the relative merits and demerits of the respective systems. Applicability of the Method to the Clarification and Purification of the Ohio River Water. These tests and investigations have proved conclusively that the general method em bodying subsidence, coagulation and filtra tion is most suitable for the proper and economical purification of the Ohio River water at this city. With regard to the use of coagulants it may be stated in unqualified terms that their use is imperative for this water, because for at least six to ten weeks in the spring and early summer the Ohio River water contains such large quantities of fine clay particles, many of which are smaller than bacteria, that clarification and purification without coagulation would be impracticable if not impossible. While this general method, which was fun damentally adopted by each of these systems, is the most suitable one, in the light of our present knowledge concerning the science and art of water purification, yet in no case did the systems tested carry out these prin ciples in a manner demanded by the economi cal and efficient purification of this water. Expressed in briefest terms, the reason of this was that they failed to remove the suspended matter sufficiently before the water reached the sand layer. With regard to the relative advantages of American and English filters for the purification of the water, after its par tial clarification bv subsidence, aided at times by coagulation, no data were obtained at this time, although in Chapter XVI the question is referred to briefly in reference to an earlier series of tests with English filters, made by this Company. The genera! defects, with their practical significance, will next be pointed out by a full comparative summary of the principal fea tures and devices of the several systems. After this is presented in brief the quality of the fil tered water; and, at the end, the final con clusion from this portion of the investigation. General Defect of all Systems, i^ itJi its Results in the Application of this Method of Purifi cation to the Ohio River Water. In this connection it is to be clearly borne in mind that the Ohio River water possesses a marked variability, both as to character and amount of suspended matter contained in it, and at times the amounts are extraordinarily large. This water, it may be fairly said, is a much more difficult one to purify than those waters concerning which data upon purifica tion are available, and which have been treated on a large scale by American filters. In justice to the several filter companies it is to be stated that they entered these tests with systems which represented their best devices based upon their general information and ex perience when arrangements were made for these tests, and not with devices designed to meet the specific requirements of this case. 272 WATER PURIFICATION AT LOUISVILLE. Of the defects possessed by the systems in these tests there is one which causes all others to drop into almost complete insignificance. As stated above, this great defect was the fail ure to remove suspended matter sufficiently from the water as it reached the sand layer of the filter, in each case. This would pro duce the following effects upon the process: 1. It would increase to an excessive degree the cost of a chemical to serve as a coagulant, which is the principal item of expense in this method of purification of this water. 2. It would necessitate a reserve portion of the system with all the appurtenances to handle the water when in its muddiest con dition. In the best systems this reserve por tion would have to be from no to 80 per cent, of the system regularly in use. 3. It would necessitate at times of muddy water the waste of an unusually large amount of filtered water for the purpose of washing the sand layers. When the river water is in its muddiest condition this percentage in the case of the best system would average from 25 to 35 per cent., and might for short inter vals reach nearly double the average. 4. It would necessitate, by virtue of the water thus disposed of, an increase in the nor mal pumping appliances, and, therefore, the aggregate cost of pumping. 5. Owing to the wide variations in the char acter of the water as it reached the sand layer, it would make very difficult the task of oper ating the systems so as to secure efficient purification at the least possible cost. 6. It would necessitate regularly a large set of trained attendants to operate the reserve portion of the system, beside those regularly engaged in operating the portion of the sys tem regularly employed. 7. It would increase certain undesirable features of the filtered water with reference to its corroding and incrusting powers. This defect was so great in the case of the W estern gravity filter (A) that when the river was very muddy it was unable to yield enough filtered water to wash its own sand layer, as already stated. For this reason this filter will not be mentioned further. With regard to the Western gravity filter (B), it was not operated long enough to al low adequate data to be secured, but gave no indications of being comparable to the Warren or Jewell filters. It will not be men tioned again in this connection. The systems were not perfect in other re spects, but none of the remaining weaknesses were of such vital importance as was the one above. Herewith is presented a comparison of cor responding devices of the respective systems with regard to their applicability in treating the Ohio River water successfully by the method of purification under consideration. ritiin Subsidence. All of the systems were totally lacking in this very essential requisite for the most economical and efficient clarifi cation and purification of this water. Kind of Chemical Used. Sulphate of alu mina was the principal coagulating chemical used in these tests. So far as could be learned at this time its use was satisfactory for the re quired purpose. Potash alum was used in the Western Sys tems only because of its physical characters, and was abandoned after an improvement was made in the device for the application of coagulants. It is too expensive for regular use. Lime, electrolytically decomposed salt and metallic iron were tried experimentally in the Jewell System, but were abandoned. Preparation of Chemical Soli/lions. The first Western device was a failure. In all other cases the addition of known weight of chemicals to known volumes of water was sat isfactory when it received sufficient care and attention. Application of Coagulant. The first West ern device was a failure. Experience showed that the Warren device was most nearly auto matic and 011 the whole did the best work under these conditions. Satisfactory results were obtained from the Jewell and second Western devices when they received sufficient attention and regulation. It is quite probable that in practice the most satisfactory results could be obtained by gravity discharge of the solutions. The use of iron pipes, fittings and pumps SUMMARY AND DISCUSSION OF DATA OF 1895-96. 273 to handle solutions of sulphate of alumina is not admissible. Brass and aluminum bronze were not attacked. Quantity of Coagulant Used. The sum maries already presented show that the grains of sulphate of alumina used per gallon of net filtered water in the case of the several sys tems were as fojlows: Warren. Jewel.. " 10.86 12.60 i . 10 2.72 The available information indicates that the river water during these tests was somewhat easier to purify by these systems than would be the average water year by year. There fore it is concluded that in no case would any of these systems treat the water with less than an annual average of at least 3 grains per gal lon of ordinary sulphate of alumina. The Period of Coagulation. The effective period of coagulation in minutes at the con tract rate, including the settling basins and the compartments in the niters above the sand layers, was in each case as follows: In no case were provisions made to allow a division to be made in the application of the coagulant to allow favorable conditions for coagulation and subsidence, and of coagula tion and nitration. It appears that at times this will be necessary. None of the above periods with a single point of application of coagulant would be advisable in practice. At times they ought to be much longer. In this connection it may be noted that the value of secondary applica tion of coagulant was appreciated by the op erators of the Warren System, as shown by its trial of July 22, under the conditions which were available. Coagulation and Sedimentation. As noted above, coagulation and sedimentation, inde pendent of coagulation and filtration, would be a great benefit at times, but was not pro vided in any of the systems, although its im portance was recognized by the operators of the Warren System. Inspection and Cleaning of Settling Basins. No adequate arrangements in this particular were made in any of the systems, although the Jewell was superior to the others. Coagulation of Water on Sand Layer. This is a point of great practical importance and depends upon t he quantity of coagulant and provisions for coagulation and sedimentation. The latter points are mentioned above. Structure and Type of Filter. The use of wood in a permanent plant would not be ad visable, although for experimental purposes wood suffices. In this respect the Western pressure filter was superior. The disadvan tage of wood was shown by the foul odors in the filtered water compartment at the bottom of the Warren filter. Compared with pressure systems the grav ity filters were found to be more practicable for the purification of this class of water under ordinary circumstances. The location of the sand layer near the top of the filtered tank, and the use of a negative pressure, as in the case of the Jewell filter, was a distinct advantage in that it reduced the wasting of coagulated but unfiltered water above the sand layer at times of wasting and similar operations. In other respects no ad vantages of a negative pressure were noted. In practice all important parts should be made as accessible as possible, and in this re spect several modifications in all the filters could be made to advantage. Sice of Filters. All of the filters were built to purify 250,000 gallons per twenty-four hours, and this size, and, so far as our knowl edge goes, this is the prevailing one in prac tice. On a large scale the cost of construc tion and of operation with regard to attendants could be materially reduced by increasing the size of the filters. The limit in size, apparently, would be determined by the arrangements for successful agitation. In this connection it is said that the Jewell Company is now building large filters. Sand Layer. The data upon this point are so obscured by other factors that it is difficult to compare them fairly. The indications are that the Warren sand layer was too coarse and that the greater frictional resistance of the Western pressure sand layer made other operations much more nearly satisfactory mdcr the existing conditions than would have )een the case had a coarser sand been em- 2 74 WATER PURIFICATION A 7^ LOUISVILLE. ployed. This observation is based upon the comparative freedom from fine particles of aluminum hydrate in the effluent of this filter in the presence of irregular coagulation of the applied water (see Chapter III). Whether it would be better to use a greater thickness of layer or finer sand, to secure increased fric- tional resistance, is not plain. The latter would probably be advisable, as it would not increase the cost of construction. The sand layer of the Jewell filter gave the best results under the existing conditions, but in the opinion of the writer it would be better to use an equal depth of finer sand. There were no indications that crushed quartz was distinctly superior to the cheaper natural sand. Filtered-water Exits. To secure a uniform and regular rate of flow of water through the sand layer, the exit area for the filtered water and the inlet area for the wash-water, at the bottom of the sand layer, should apparently be less than that of the main pipe beneath them. In respect to this condition the Jewell filter alone fulfilled it. It would seem advisable, however, to decrease the distance between the strainer cups to secure more uniform flow in the lower portion of the sand layer. The Western exit devices were very poor, because in the slotted tubes sand accumulated in a short time. So far as could be noted, the exits of the Warren filter served their purpose fairly well, but the varying space occupied beneath them by the supporting frame was undesirable. In no case were these portions accessible without removing the sand. Loss of Head. The indications were that about TO feet of maximum available head, as ordinarily utilized in the Jewell filter, was best. Amounts above this, as in the Western pres sure filter, could be used too seldom to be advisable. In the Warren filter not more than 4 feet were used, owing chiefly to the coarse ness of the sand layer. With regard to pressure filters and negative head, see foregoing remarks on types of fil ters. Rate of Filtration. There are no indica tions that it would be advisable to employ rates of less than 100 million gallons per acre daily, and it is quite possible that this limit could be safely raised. The data, however, are too complicated by other factors to make this a decisive conclusion. But it is probable in view of the results from the Western pres sure filter that in practice under favorable conditions the plant could be operated so as to make increased (uniform) rates in a measure meet increased demands for filtered water. Regulation and Control. This is an im portant point both with regard to necessity of uniform rate to give satisfactory results and also in respect to cost of operation. The automatic controller of the Jewell filter was very crude, but a step in the right direc tion. }Vashing the Sand La\cr. Thorough wash ing of the sand layer is very important. To se cure this it is necessary to distribute the wash- water uniformly under the sand layer. In this respect the Jewell filter was the most satisfac tory, though as mentioned before a smaller distance between the cups seems desirable. Agitation of the sand layer during washing was an advantage as shown by the operation of the Warren and Jewell filters. Of the two agitating devices, that of the Jewell filter was less cumbersome and did not move the sand from the center toward the periphery. It worked poorly at times, apparently due to a binding of the gears occasioned by the warp ing of the partly submerged timbers upon which the agitator rested. Both the Warren and Jewell devices lacked simplicity of detail and were too weak for the purpose. These defects could and should be remedied. Surface Agitation. This procedure to re lieve clogging was used in the Jewell filter and was a decided step in advance. Its suc cess is associated closely with the degree of coagulation of the water entering the sand layer, the character of the sand layer and the arrangement of the tank containing the sand layer. The successful employment of this method could probably be extended by a modification of the above factors. Steaming. This did not seem to be neces sary during these tests, although it might be the case in some instances. This disadvantage of it is that it makes the organic matter on the sand serve as a better food for micro organisms. SUMMARY AND DISCUSSION OF DATA OF 1895-96. 275 Quality of the Filtered Ohio River Water. With proper attention to the operation of the systems, and an adequate degree of coagulation of the water as it entered the sand layer, these systems could produce a quality of filtered water which would be thor oughly satisfactory under all ordinary con ditions with regard to appearance and sani tary character. From an industrial standpoint, the filtered water would have a greater corroding action upon uncoated iron receptacles but not upon lead pipe; and it would contain more incrust- ing constituents when used in steam boilers. Concerning this last point the total quantities would not be excessive, compared with aver age Western waters, and the removal of the suspended matters would largely if not wholly offset the added sulphate of lime. Owing to inherent qualities of the Ohio River water, the storage of the effluent in open reservoirs in this climate would require very careful consideration, and the period could not be a long one, owing to conditions favor ing growths of alga?, etc. Final Conclusions. In all these systems the provision for sub sidence, both with and without coagulation, was thoroughly inadequate in each case; but with regard to filtration proper the Jewell fil ter was the most satisfactory. 2 7 6 WATER PURIFICATION AT LOUISVILLE. CHAPTER X. DESCRIPTION OF THE HARRIS MAGNETO- ELECTRIC SYSTEM OF PURIFICATION, AND A RECORD OF THE RESULTS ACCOMPLISHED THEREWITH. THIS system consisted essentially of a series of large, iron-covered tanks, and a set of electrical and magnetic appliances. Accord ing to the terms of the contract this experi mental system was to have a capacity of 250,- ooo gallons per twenty-four hours. A brief general description of the system is as fol lows : On the inlet water-pipe there was a small iron cylinder with a porcelain lining. As the water passed through this cylinder, called a spark drum, it met the discharge of an electric current of high voltage. From this cylinder the water passed in succession through three large, round iron tanks with conical bottoms. Each of these tanks contained a lining for the purpose of insulation. The water entered each of these tanks, in turn, at the side about two feet from the top. In the upper portion of the tanks were electrodes between which the water flowed as it passed out of the tanks at the top. On the top of each of the tanks was a set of electro-magnets. The outlet pipe con nected with an opening in the cover and be tween the magnets. The three tanks were similar in construction, and the outlet pipes from the first two tanks entered the second and third tanks, respectively. The fundamental principles upon which this system was based were never accurately explained to me. Electro-chemical action was considered to be an important factor in connection with the destruction of the bac teria and organic matter in the water. It was intended tha^all suspended matter would be repelled by the action of the magnets situated at the top of the three tanks; and the mag nets were to force the suspended matters, including the bacteria, to the bottom of the tanks, where pipes leading to the sewer were provided. There will next be presented a more de tailed description of these devices and the ac companying electrical machines and appli ances. Before doing so, however, it is to be recorded that, owing to delays in the prepara tion of castings, etc., the construction of this system was not begun until March 27, 1896. No official attention from the laboratory was given to the system until June 24. A large portion of the intervening period of three months was occupied in improvements, especially with regard to an insulating lining for the three large iron tanks, as will be ex plained beyond. The spark drum, at the beginning of the system, was a cast-iron cylinder of a special design. It was 18 inches long and 10 inches in diameter. Near each end on opposite sides, a branch was taken off to connect with the inlet and outlet water-pipes, respectively. The cylinder and branches were one casting and were all lined with porcelain. The ends were closed with caps which were bolted on to the drum. At the center of each end there was a stuffing box, through which there were passed, respectively, the two pole pieces of the high voltage circuit from a Ruhmkorff coil. When the system was in operation these pole pieces were said to be 3 inches apart. Iron Tanks containing the Electrodes and Electro-magnets. These tanks, three in number, were made of cast iron, i inch in thickness. The upper half of each tank was cylindrical in form, and the lower half was in the form of a cone with the apex at the bottom. HARRIS MAGNETO-ELECTRIC SYSTEM OF PURIFICATION. 277 The inside dimensions were as follows: Di ameter of cylinder, 35.5 inches; depth of cylinder, 36 inches; depth of cone, 36 inches; and diameter of opening at the apex of the cone (bottom of the tank), 3 inches. This opening at the bottom of the tank connected with a pipe which led to the sewer. Brass covers closed the top of the tanks, and sup ported the electro-magnets in the manner de scribed below. The tanks were placed on suitable pedestals. The lining of the tanks was originally of cement. This did not give satisfactory insu lation and at the time that the system was ex amined officially the tanks were lined with soft rubber sheets. The inlet water pipes, 3 inches in diameter, entered at the side of the tanks, 2 feet from the top. The opening for the inlet pipe was lined with porcelain. A porcelain hood, 0.625 inch thick, 3 inches in diameter and 4 inches long, was provided at the inlet opening to break the currrent of the water. At the apex of the conical bottom of the tanks there was a 3-inch opening which connected with a 3- inch pipe leading to the sewer. Plug valves controlled the flow through these blow-off pipes. The main exits from the tanks were openings in the covers; and into cast-iron chambers on these covers were connected the outlet water pipes, 3 inches in diameter. Electro-magnets. On the brass casting, i inch in thickness, which formed the cover of each tank, there rested a set of 5 electro magnets. In the center there was a large one, 15 inches in diameter, with a core of 12 inches. Four small magnets, each 8 inches in diameter with a core of 6 inches, surrounded the central one. The core of the large central magnet passed through an opening in the brass cover and was fastened on the under side to an iron disc if> inches in diameter and r inch thick, which formed the negative pole. The cores of the four outer magnets passed through the brass plate and connected with an iron ring on the under side, which formed the positive pole. This ring was 10 inches wide and i inch thick. Between the disc and the ring was a circular opening 0.25 inch wide and 16 inches in diameter. Communication with the outlet recess on the top of the cover was obtained by a number of small holes drilled through the brass cover just above this circular opening. The magnets were connected at the upper end by a cast-iron cross about 6 inches thick. With a full current from the generator (45 amperes) the lifting force of each set of (5) magnets was said to be about 6 tons. Electrodes. The size and arrangement of the electrodes were changed a number of times during the period covered by the pre liminary trials of the system. On the date of the official examination the positive electrode consisted of a series of pressed carbon plates. The plates were o.25-inch in thickness and 12 inches in width. They were placed in a parallel and vertical position, and suitable in sulation and support were provided to keep them about i inch apart. The lengths of the carbon plates varied with the length of the parallel chords which they formed with the periphery of the tank, respectively. The total area of these plates (one side) was about 96,- ooo square inches. The top of the carbon plates was about 6 inches below the brass top of the tanks. From the bottom of the plates to the plane in which the water entered the tanks the distance was 6 inches. The negative pole was placed near the bot tom of the tanks. It was a small sheet of metallic aluminum about 0.06 inch in thick ness, and about 150 square inches in area. Suitable openings in the tanks were pro vided for the connection of the wires to the electrodes. Piping.- The inlet and outlet pipes of the spark drum were 4 inches in diameter. With this exception all the piping, including inlet, and blow-off pipes of the respective tanks, was 3 inches in diameter. The outlet pipe from the third and last tank led to a con denser where the exhaust steam from the en gine which operated the generator was con densed. All of the tanks and also the spark drum were closed compartments. The rate of flow of water through the system was controlled by a valve on the main inlet pipe which con tained the river water under about fio pounds pressure. Suitable valves were also provided on the inlet and outlet pipes of each tank. The blow-off pipes at the bottom of each WATER PURIFICATION AT LOUISVILLE. tank were 3 inches in diameter and connected directly with the sewer. The flow through these pipes was controlled by 3-inch plug valves. Engine. A simple stationary engine was used to drive the generator. Its principal dimensions were as follows: Diameter of steam cylinder, 9.25 inches; length of stroke, 8.75 inches; and cut-off, 70 per cent. The fly-wheel and driving-pulley were combined, and had a diameter of 4 feet and a rim 12.5 inches in width. Its weight was about 1200 pounds. From the engine the power was conveyed to the generator a distance of about 20 feet, by means of a leather belt 6 inches wide. Dynamo-generator. The generator was a compound-wound, bi-polar machine. It was wound to generate a direct current of 220 volts and 45 amperes, at a speed of 1 125 revo lutions per minute. The driving pulley was 8.75 inches in diameter and 6 inches wide. A rheostat was provided to regulate the inten sity of the field by regulating the amount of current passing through the shunt winding. It was, however, seldom used. Electric Circuits. At the switch-board the main circuit was divided into three principal sub-circuits. The first of these sub-circuits passed directly to the electro-magnets situ ated on top of the tanks; the second led to the electrodes within the tanks; and the third passed through an interrupter to a Ruhm- korff coil from which the induced current of high voltage passed to the spark drum. All of these circuits were arranged in parallel on the main circuit. A fourth sub-circuit was also taken off to a small electric motor which turned the interrupter on the third circuit. Resistance coils were used to control the electric current. They were made of 50 coils of No. 14 galvanized iron wire, about 4000 feet in all being used. Connections were made so that any number of coils could be used as desired. Results Accomplished by the Harris Magneto- electric System. This system was in official operation only for one hour, from 4.00 to 5.00 P.M. on June 24, 1896. The record of its operation, with the results of analyses of the river water before and after passage through the system, is as follows: The rate at which the water passed through the full system was gradually increased until at 4.10 P.M. it had reached 23.5 cubic feet per minute, equivalent to 254,000 gallons per 24 hours. For ten minutes this rate was held approximately constant. At the end of this time, 4.20 P.M., samples of water, the an alyses of which appear in the next tables, were collected as follows: Bacterial sample Xo. 3959 was taken from the water as it left the spark drum. Bacterial sample No. 3960 was taken from the water as it left the last tank. Chemical sample No. 671 was taken from the water which was " blown off " at the bot tom of the tanks. Chemical sample No. 672 was taken from the water as it left the last tank. For the next ten minutes the average rate of flow of the water through the entire sys tem was 16.5 cubic feet per minute, equiv alent to 178,000 gallons per 24 hours. At the end of this time, 4.30 P.M., the fol lowing samples were collected for analysis: Bacterial sample No. 3961 was taken from the water as it left the spark drum. Bacterial sample No. 3962 and chemical sample No. 673 were collected from the water which had passed through the entire system. During the next period of fifteen minutes there was maintained an average rate of flow of 12.5 cubic feet per minute, equivalent to 105,000 gallons per 24 hours. At 4.45 P.M. samples corresponding to those noted above were taken as follows: Bacterial sample No. 3968 was collected from the water after passage through the spark drum. Bacterial sample No. 3969 and chemical sample No. 674 were collected from the water after passage through the entire system. From 4.45 to 5.00 P.M. the average rate of flow of water through the system was 6.2 cubic feet per minute, equivalent to 67,000 gallons per 24 hours. The following samples were collected at 5.00 P.M. which was the end of the test of this system. HARRIS MAGNETO-ELECTRIC SYSTEM OF PURIFICATION. 279 purification of the water after its passage through the system. This system was never put in official opera tion after this date. Various portions of it, however, were utilized in the devices which were operated by the Harris Company during the following month. RESULTS OF BACTERIAL ANALYSES OF SAMPLES DESCRIBED ABOVE. Bacterial sample No. 3971 was collected from the water after passage through the spark drum. Bacterial sample No. 3970 and chemical sample Xo. 675 were collected from the water after passage through the entire system. At 4.44 P.M. chemical and bacterial samples of the river water, having the following num bers, respectively, 679 and 3963, were col lected for analysis. The electric current, during the period from 4.00 to 5.00 P.M., June 24, had an aver age voltage of 206 and an amperage of 20 to 21. as it left the generator on its way to the full system which was in use at this time. In the next table are presented the results of analyses of the several samples of water de scribed above. They show no appreciable RESULTS OF CHEMICAL ANALYSES OF SAMPLES DESCRIBED ABOVE. (Parts per Million.) Number of Sample. Source of Sample. Collected June 24. Hour. Bacteria per Cub. Centimeter. 3959 Spark Drum 4.2O P.M. 10 400 3960 Effluent 4.20 II 800 3961 Spark Drum 4 30 7 600 3962 Effluent 4.30 10 100 3963 River 4.44 II fioo 3968 Spark Drum 4 45 13 100 3969 Effluent 4-45 8 800 3970 Effluent 5.00 9 Soo 3971 Spark Drum 5.00 i Nitr ogen e 1 ." inni,! a Evaporation. after Ignition. 1 V u * in - Q 3 "^ 5 s 4) V d < 1 V u 2 S S c J2 rt u 4J c >1 o X Date. 1896 Hour. |l< 1 g u i S " Si 1 1 S -5 1 ! 3 c 1 i M c - !/> H O o H j & C fc U H Q a < Q A 670* June. 24 4.44 r-M. .23 5-4 .334 170 164 .022 .003 .6 28.0 524, 326 1 98 450 304 146 73-0 O.O 2.10 &7it 24 4 .lS .15 6-4 .414 312 I O2 .O3f) .001 . > 27-5 59* 397 199 498 371 127 70.0 0.0 8.30 672t 24 4 20 .15 7-0 .404 308 096 -034 .003 -7 27.6 <;S6 383. 203 4871 35 136 69.8 o.o 9.50 671* 24 4.30 .16 5-7 .346 248, 098 .032 .003 .6 27-5 527 324. 203 443 304 139 70.0 o.o 2.50 674* 24 4-45 .18 5-9 -348 250! 098 .034 .004 -5 27. 1 521 330, 191 447 308 139 -0.3 o.o 2.20 675^ 24 5.00 1. . . . .23 --.- 370 272 098 .038 .004 .7 27-5 514 310 204 427 276 151 70.3 o.o O. IO t " Blow-off " at bottom of tanks. J Effluent. 280 WATER PURIFICATION AT LOUISVILLE. CHAPTER Xi. DESCRIPTION OF THE DEVICES OPERATED BY THE HARRIS COMPANY IN JULY, AND A RECORD OF THE RESULTS ACCOMPLISHED THEREWITH. DURING the month of July, 1896, a num ber of devices, more or less alike, were oper ated by the Harris Company with the view to purifying- the Ohio River water. Various portions of the original system were utilized in the several devices, as will appear in the descriptions beyond where they are taken up in turn. The devices of July appeared to be based mainly, as I understand the matter, on the results of some experiments made in a small glass jar during the last week in June. These experiments may be summarized briefly as follows: A glass jar, of about one gallon capacity, was rilled about three-fourths full of river water, and in the water were placed two cir cular strips of aluminum sheet. The thick ness of the aluminum sheets was about 0.06 inch. The two strips were separated from each other by suitable blocks of an insulating material, about 0.125 mcn thick. The cross section of the electrolyte (equal to the area of one side of one of the strips) was about 30 square inches. Through these electrodes there was passed a current of electricity from the generator. A considerable quantity, of gas, practically all of which was hydrogen, was set free at the nega tive pole by the action of the current. There was formed a white gelatinous substance which appeared for the most part, if not wholly, at the positive pole. It was found that this substance was aluminum hydrate. This is the same compound that is formed by the decomposition of alum or sulphate of alumina by lime, as has been explained in pre ceding chapters. The aluminum hydrate coagulated the sus pended matters in the water, in a similar man ner as when sulphate of alumina was added to the water. Instead of the coagulated masses subsiding at the bottom, as in the application of sulphate of aiumina, the greater part of them were carried to the surface by the rising currents of hydrogen gas. When the electric current was turned off a portion of the matters suspended through out the water settled to the bottom of the jar, while some of them joined the thick scum which formed on the surface of the water. At the end of a few minutes, five or less, the main bulk of the water became quite clear, with the exception of a few scattering particles of aluminum hydrate. With regard to the length of time necessary to coagulate and clarify the water, this de pends upon the strength (amperage) of the electric current. The reason of this lies in the fact that it is the strength of the current which determines the rate of conversion of metallic aluminum into the form of aluminum hydrate, disregarding any secondary solvent action of the initial compounds. In the ex periments which received official attention the current was applied for 5 and 10 minutes, respectively. It was estimated that the amounts of aluminum which were converted to aluminum hydrate were about 7 and 17 grains per gallon, respectively. By the aid of a siphon portions of the clari fied water were removed from the jar for analysis. The bacterial results showed that the numbers of bacteria in the river water ranged from 8 100 to n 600, while in the clarified water from the jar the numbers were 6, 68, and 120 per cubic centimeter, respect ively. Disregarding the scattering particles of aluminum hydrate, the chemical results DEI ICES OPERATED B Y JJJE HARRIS COAJPAAY JX JULY. 281 showed the removal of all the suspended or ganic and mineral matters present in the river water; and, further, that there had been an appreciable reduction in the organic matter which was dissolved in the water. STATUS OF THE SITUATION ON JULY i, WITH REGARD TO THE MAGNETO-ELECTRIC SYSTEM AND DEVICES. The magneto-electric system having been abandoned, practically speaking, by the Har ris Company after the official test of one hour on June 24, there were operated during July several devices in which use was made of the principles illustrated by the jar experiments described above. It is to be recorded here that aluminum electrodes were known to have been employed by others for the purification of certain waters at a date earlier than that of these experiments. But it is also to be stated that the Harris Com pany claimed that their magnets would sup plement and increase the action of the elec trodes; and, further, that the magnets would facilitate the clarification of the coagulated water, and, perhaps, do away with the neces sity of subsequent filtration through sand. The devices operated in July will be de scribed in turn, together with the results which they accomplished, respectively. DEVICE No. i. The first device was offered by the Harris Company for official examination on July 9. It consisted essentially of a closed iron tank lined with porcelain, which contained a set of aluminum electrodes. At the bottom of the tank was placed a set of magnets. After treatment in this tank the water passed to the top of a small stand pipe, through which it flowed from top to bottom, and thence to the sewer. The Tank in which the Wafer wax Treated. The tank in which the water was subjected to electrolytic action was a small cast-iron cylinder surmounted by a brass dome. It was lined with porcelain. The principal inside di mensions were: Diameter of cylinder and base of dome, 1.71 feet; height of cylinder, 2 feet; and height of dome, i foot. The water entered this tank at the top and passed out at the bottom, from which point it was conveyed to the top of the stand pipe. Electrodes. Aluminum sheets, arranged in the form of a manifold, composed the elec trodes which were placed within the porce lain-lined tank. They were made of sheets of about 0.06 inch thick, which were held to gether by hard rubber bolts, the desired dis tance between the plates, 1.75 inches, being maintained by the use of hard rubber separa tors. Alternate sheets of metallic aluminum in the manifold formed the positive and negative poles, respectively. The total area of active electrode surface (anodes) was about 2550 square inches. Magnets. The magnets were similar in their arrangement to those of the three tanks of the original system, described in the last chapter, except that they were placed on the bottom instead of the top of the tank. There were five magnets in the set. The central one, forming the negative pole, was 8 inches in diameter. The other pole was formed by four smaller magnets of a diameter of 4 inches. The latter were placed around, and connected with, a ring which surrounded the central magnet. Stand Pipe. The stand pipe, into which the water passed after treatment in the iron tank, consisted of a single 1 2-foot length of iron pipe. Its diameter was 20 inches, and both ends were closed by caps. At four equi distant points in the stand pipe there were placed tin cones each 12 inches high and 20 inches in diameter. At the apex of each cone was an opening i inch in diameter, through which the water flowed downward in its pas sage through the stand pipe. The cones were all placed with the apex upward. At the side of the tanks were four openings, one above the base of each of the respective cones. These openings connected with blow-off pipes lead ing to the sewer. The main exit was about 3 feet from the bottom. Piping. The inlet and outlet water pipes of the tank and stand pipe were i inch in diameter. River water was supplied to the device under a pres- 282 WATER PURIFICATION AT LOUISVILLE. sure of about 60 pounds. The blow-off pipes leading to the sewer were 0.50 inch in diam eter. Electrical Machines and Appliances. The engine and generator were the same that were used in the system described in the preceding chapter. The circuits were similar except that the interrupter and Ruhmkorff coil were not used. This device was in operation, officially, from 2.30 P.M. to 5.00 P.M. on July 9. In that time 514 cubic feet of water passed through it. The rate of flow ranged from 3 to 4 cubic feet per minute, and averaged 3.43 cubic feet per minute, which is equivalent to 37,000 gal lons per 24 hours. The amount of metallic aluminum, which was converted electrolytically into aluminum hydrate, was estimated to be equivalent to about o. 10 grain per gallon of water treated. Observation on the electric current em ployed in the operation of this device showed that the amperage averaged 18.8 and the voltage 211. The average current was equiv alent to .0122 ampere-hour per gallon and the power was 144 electric H.P. per million gal lons of water treated per 24 hours. Samples of water for chemical and bacterial analyses were collected as the water left the stand pipe at 3.30 P.M. and 4.30 P.M., after the device had been in operation i and 2 hours, respectively. CHEMICAL RESULTS DEVICE No. 1. Analyses of Samples Described Below. (Parts per Million.) Nitr 3K en R< sinu. Fixe d Res due oi 1 as rt Ev; porat on. afte Igni ion. 1 1 1 A mr.m, E gl s u z Date. Hour. 2% S T3 * | g | | > | > | > 7. h U o $ 3 C- 5 u H 1 Q H Q | Q c 71 1 July 9 9.30 A.M. 26.0 .25 6.6 .378 .272 . 106 .030 .002 .6 6.9 sSq 47" IIQ 523 436 87 56.0 o.o 30.70 719 " 9 Av. 2 samples. 15 6.4 .404 .302 .102 .062 .006 9 7-1 594 475 Ilq 521 435 S6 54.0 o.o 22.00 The two corresponding chemical samples of the treated water were mixed together and analyzed as sample No. 719. The results of the analysis of this sample and that of the river water on that day, sample Xo. 714, col lected at 9.30 A.M., are given above. The results of bacterial analysis of these samples and of the river water on that day are as follows: DEVICE No. 2. BACTERIAL RESULTS DEVICE No. 1. Col ected. Hacieria Source of Sample. Date. 1896. Hour. per Cubic River July 9 9.30 A.M. 8 600 River Treated water " 9 4-3" " 8 700 No substantial purification of the water by this device is shown by the above results. After this test of July 9 modifications and ad ditions to the device were made. On July 1 6 a second device was offered for official inspection. It was simply an elabora tion and extension of the first device. There were added two duplicates, practically speak ing, of the porcelain-lined iron tank contain ing electrodes and magnets as above de scribed. The original one was also used, making three in all. The total area of active electrode surface (anodes) was about 4000 square inches, and the distance between the plates averaged 1.75 inches. To facilitate clarification of the water after the treatment in these tanks there were used, in addition to the stand pipe, the three large iron tanks employed in the original system. The flow of water through the stand pipe was reversed ; that is to say, it entered at the bottom and passed out at the top. In the passage of the water through each DEVICES OPERATED BY THE HARRIS COMPANY IN JULY. 283 of these seven closed vessels successively, it was first treated in the three porcelain-lined tanks containing the electrodes and magnets; and thence it passed through the stand pipe and the three large iron tanks for clarification by subsidence. The cubical capacity of all the vessels was about 126 cubic feet, as fol lows: each of the three porcelain-lined tanks, 5.5 cubic feet; stand pipe, 26 cubic feet; each of the three iron tanks of the original sys tem, 28 cubic feet; and piping (i and 3 inches in diameter), 5 cubic feet. Results Accomplished by Device No. 2. The device was put in operation on the above-mentioned date, but only for 26 min utes after the water appeared at the outlet pipe. At the commencement of the test all the tanks and the stand pipe were drained. Water was applied at 11.52 A.M., and first ap peared at the outlet at 12.39 P - M - The opera tion of the device was stopped at 1.05 P.M. The period of operation was too short to yield any decisive information, except that there was no improvement in the appearance of the water after treatment at the average rate of 4 cubic feet per minute. Samples of water, before and after treat ment, were collected for analysis, and the re sults are presented below as a matter of record. It will be noted that the numbers of bacteria in the water increased during the pas sage through the device. The reason of this appeared to be that there were accumulations on the walls of the three rubber-lined tanks through which the water last passed. BACTERIAL RESULTS DEVICE No. 2. Source of Sample. Col ected. Bacteria per Cubic Date. 1896. Hour. M. M. River . Ji : ly 16 16 16 16 16 16 16 16 9.30 A 12.45 P 12.50 12.50 12.55 1 .00 1.05 5.00 6 Soo 22 900 Treated wa < Treated wa River . . . 5 600 13700 9 6co 10 200 5000 Treated wa Treated wa Treated wa River With regard to the following results of chemical analyses, sample Xo. 740 was col lected from the river water at 12.50 P.M.; while sample No. 741 represents a mixture of three equal portions of the treated water, collected at 12.45 r - M -.. 12.55 p -M-> and 1.05 P.M., respectively. CHEMICAL RESULTS DEVICE "No. 2. Analyses of Samples Described Above. (Parts per Million.) Nitr Collected Fixed Residue e as Evaporation. after Ignition. 1 z E Kg c Ammonia. i si Jj ^ 3 z Date. Hour. i M c g S2 u *D 1 13 ^ S - 1896. g-2 g a -3 , H 3 S 55 ?; "5 5 i g O. 1 i 1 ^ (/I fi O O H in Q t H 1 Q h 3 Q < Q i 740 Til July 16 " 16 12.50 I .M. Av. 3 samples 27 M 7.0 6.6 .3^2 370 .214 .260 .748 .11 .030 .044 .000 .012 .6 4 2.2 540 2 . 5 490 404 136 480 37o> 120 429 404 76 52.7 366 63 52.3 o.o o.o 15-40 15.30 DEVICE No. 3. The third device was offered for official in spection on July 1 8. It comprised the seven closed vessels described as the second device, and in addition there was a small filter of sand. Concerning the portions of the device common to the second one. it is to be stated that the second and third porcelain-lined tanks were used only as settling chambers. Alu minum electrodes, the anode area of which was about 2300 square inches, were placed only in the first porcelain-lined tank where the river water entered the device. The direct use of the magnets was also abandoned, and they were used only as resistance coils to control the current and prevent the passage of a greater current than the fittings were designed for. The water passed through the seven closed vessels in the same order as in the second device. When the water reached the last tank a portion of it was passed through the filter. 284 WATER PURIFICATION AT LOUISVILLE. Filter. The filter \vas made by filling with sand a galvanized iron tank, which was 2 feet in di ameter and 3 feet deep. The outlet of this tank was a common o. 5-inch tap. On the bottom of the tank were placed three pieces of slotted brass pipe, which were to serve as strainers and allow the water to pass to the outlet tap while the sand was kept in place. These tubes were 1.5 inches in diameter on the inside. The length of one was 10 inches and of each of the other two 8 inches. They were all screwed into a cross which connected with the outlet tap. In each tube there were 5 rows of circumferential slots which aver aged 0.024 inch in width and 0.72 inch in length. These slots were spaced about 0.125 inch from center to center. Sand Used. Coarse sand of an effective size of 0.56 millimeter was placed on the bot tom of the tank to a depth of 6 inches, sur rounding and covering the slotted tubes. Above this layer of sand were 24 inches of a somewhat liner sand having an effective size of 0.51 millimeter. The surface of the sand was about (> inches below the top of the tank. Resistance of Strainers and Sand Layer. \Vith the sand entirely removed from the tank, the maximum rate at which the strainer tubes and the outlet tap allowed water to flow was 0.37 cubic feet per minute, equivalent to 55 million gallons per 24 hours per acre of tank area. \Yith the sand in place, and the surface free from accumulations of suspended matter, this rate was 0.28 cubic foot per min ute, equivalent to 42 million gallons per acre per 24 hours. Piping. Suitable piping was provided to carry a portion of the water, as it passed through the outlet of the last tank, to the filter. Connections were made with the out let pipe so that purified water could be forced up through the sand for the purpose of wash ing. The sand was stirred by hand during the process of washing. This device was operated during four hours on July 18. At 10.05 A.M. operation was be gun, with the tanks and filter empty. River water was applied to the device at a rate of about 3.0 cubic feet per minute. The full set of (seven) tanks was filled with water at 10.47 A - M - ^t triat t 1116 water from the outlet of the last tank was applied to the filter. The rate of Mow of water through the tanks was then increased to about 5 cubic feet per minute. For the greater part of the time the water passed through the filter at about the maximum rate; and observations showed this to be from o. 16 to O._M cubic foot per minute, equivalent to 24 and 31 million gallons per acre per 24 hours, respectively. After 12.05 [VM - tne water passed through the tanks at an average rate of about 1.3 cubic feet per minute, equiv alent to 14,000 gallons per 24 hours. The op eration of the device was stopped at 2.05 P.M. After the tanks were filled the blow-off pipes at the bottom of the tanks were opened for a few seconds every fifteen minutes, to re move suspended matters which had settled to the bottom. Of the 630 cubic feet of river water w.hich were applied to the device from 10.05 A - M - to 2 -5 P -M., 20 cubic feet were dis posed of in this manner; 484 cubic feet passed through the tanks; and the remaining 126 cubic feet were retained in the tanks and the piping system. During the operation of the device the voltage and amperage of the main circuit averaged 210 and 29, respectively. The voltage on the electrodes was 52. An opportunity to obtain adequate obser vations on the amount of metallic aluminum used was not afforded. However, taking the results of later experiments as a basis, it is estimated that the amount of metallic alu minum converted electrolytically into alu minum hydrate averaged about 0.15 grain per gallon of applied water. The amount of elec tric current was .0129 ampere-hours per gal lon up to 12.05 P.M., and .0497 ampere j hours per gallon after that time. The corresponding amounts of electric power were 152 and 583 electric H.P. per million gallons of water per 24 hours, respectively. After the operation of this device was stopped at 2.05 P.M., the filter was washed. River water without previous treatment was then allowed to flow through- it for thirty-five DEVICES OPERATED BY THE HARRIS COMPANY IN JULY. minutes (2.30-3.05 P.M.) at a maximum rate of ab out 30 million gallons per acre per 24 hours. Samples of water before and after treatment by this device were collected and analyzed with the results indicated in table opposite. It will be noted in connection with the above results that the number of bacteria in the water decreased in passage through the tanks, but increased after filtration. The ex planation of this is that the filter was not op erated long enough to wash out the bacteria originally contained in the sand. In the next table are presented the results of analyses of chemical samples on this date. Sample Xo. 744 was collected from the river water at 9.30 A.M. Samples Xos. 745 and 747 were collected at 11.55 A.M. and 2.00 P.M., respectively, from the water as it left the outlet of the last tank. Samples Xos. 746 and 748 were collected at 11.55 A - M - anf l 2 - I .M., respectively, from the water after it passed through all the tanks and through the filter. BACTERIAL RESULTS DEVICE N o. 3. Source of Sample. Hour of Bacteria per Cubic 600 Treated water from outlet of last Treated wa e after filtration. . . . Treated wa e from outlet of last tank 12.00 " 7 100 Treated wa e after filtration. . . . Treated wa e from outlet of last 12.05 " 9 loo Treated wa e after filtration. . . . Treated wa e from outlet of last tank 2.OO " 13000 Treated water after filtration .... 2.O5 " 8 400 River water after filtration with out preliminary treatment Treated water from outlet of last tank (12.00 M.) after 5 hours 3.04 " 251 Treated water from outlet of last tank (2.00 P.M.) after 3 hours Sample Xo. 752 was collected at 3.04 P.M. and represents river water after filtration without preliminary treatment in the tanks. CHEMICAL RESULTS DEVICE No. 3. Analyses of Samples Described Above. (Parts per Million.) Collected Nitrogen Residue on Fixed Residue . after Ignition. - <; u C Ammonia . 5. g -- 7, Date. U S3 M ~ si u a s s -a a fr a a a 1896. 8.3 u & _ o > V Z z c _: 5 3 _ 5 -^ 3 - g 3 i| .iS ^ g n % - * . ^ en </> (" u c H a o H yi a H <fi a < a ~ 744 July 8 9.30 A.M. 25-7 .19 8.71-504 .406 .098 .038 .OOI -5 4.0 751 637 116 668 594 74 6O.3 o.o 33-30 745 8 11.55 " .10 6.3-352 .264 .oSS .062 .004 5 4.1 496 387 109 414 71 57-6 o.o 29. 10 74 8 11.55 " 13 0.8 .082 .OOO .082 .060 .004 4 4.0 107 107 66 66 57-9 o.o .07 747 8 2.0O P.M. .07 I .2 .128 .036 .092 .146. 004 -4 4.0 172 69 103 M4 69 65 59-0 0.6 4. 16 748 8 2.00 " .08 O.g .092 .000 .092 .098 .004 4 2.5 98 o 98 66 66 59-0 o.o .05 758 8 3.04 " .! - 1-9 .178 .030^007 4 4.0 127 IOI 6o.O 0.0 9.20 DEVICE No. 4. The fourth device presented for official in spection was operated continuously from 9.50 A.M. to 3.25 P.M. July 23, and from 6.30 P.M. July 23, to 9.00 A.M. July 24. It was intended to operate this device continuously up to August i , but the Harris Company de cided to make a further modification after the operation as above noted. This device com prised all of the seven closed vessels which had been used before, and the small sand fil ter. The first small porcelain-lined tank alone contained electrodes of metallic alu minum, the active area (anode surface) of which was about 2300 square inches. The set of magnets in this tank was also used. On top of this tank was set the 2O-inch stand pipe. The cover of the tank and bottom of the stand pipe were removed, as were also the tin cones in the pipe. Water to be treated entered at the bottom of the small ta^nk, and passed up through the stand pipe, from the top of which it passed to and 286 WATER PURIFICATION AT LOUISVILLE. through the series of two small tanks, and three large ones, for the purpose of sedimentation. The records of the operation of this device are presented in the next tab.es. For convenience in presentation, two tables are given, the first giving the record of the device exclusive of the sand filter, and the second, the records of operation of the filter. Following these two tables are presented the results of analyses of samples of river wa er before and after treatment by this device. Alinnin: in Used. Satisfactory observations of the amount of metallic aluminum used in the treatment of the water were not obtained. The river water was very muddy at this time, and examination of the porcelain-lined tank at the close of the runs showed that it was filled wi h a very thick deposit of mud, which had subsided during the passage of the water through this tank and the stand pipe placed on the top of it. Another observation of im portance was that, at the bottom of the porce lain-lined tank, there was found a large num ber of scales of aluminum oxide, which had been formed on and removed from the metal lic e ectrocles by the action of the electric cur rent. RECORDS OF OPERATION DEVICE No. 4. Number of Run .... 1 2 July 23, 9.50 A.M. Jul v 23, 6. 20 P.M. Ended " 23, 3.25 P.M. 24, 9.00 A.M. Period of service. . . 5 hrs. 35 min. 14 hrs. 30 min. Period of tilling o " 35 " o " 15 " Cubic feet of water t recited * 2 275 4 464 Cubic feet of water O o Average rate of treatment Cubic feet per min. 6.77 5-35 Gallons per 24 hrs.. 72 goo 57 600 Average voltage (entire system). . . 206 214 Average amperage (entire system). . . 29.6 26.0 Average electric ampere hours per gallonf .0098 .0108 A v e rage electric ii. i . per mil. gals. per 24 hoursf. . . . i i . 129 RECORDS OF OPERATION OF THE FILTER DEVICE No. 4. Number of Run 1 2 3 July 23, 9.50 A.M. " 23, 2.OO P.M. 5 hours 10 minutes. 4 " 53 17 45.6 7-5 O.I2 ITS July 23, 2.00 P.M. " 24, 2.50 A.M. 9 hours 50 minutes. 9 30 20 " 87.8 10.2 0.15 22 July 24, 2.50 A.M. " 24, 9.00 " 6 hours 10 minutes. Ended Period of operation 67.8 Average rate : 0.18 27 Million gallons per acre per 24 hours RESULTS OF BACTERIAL ANALYSES OF SAM PLES COLLECTED JULY 23, 12.00 M. DEVICE No. 4. Source ofSa-Me. ! Bacteria per Cubic Centimeter. Treated water from outlet of stand pipe. Treated water from outlet of last tank. Treated water after filtration 23 200 29 500 4 :>"" Samples of water before and after treatment by this device were collected and analyzed with the results presented beyond. Of the chemi cal samples, No. 763 was collected from the river water before treatment, No. 768 was col lected from the water after it had passed through the first small tank and stand pipe extension thereof, No. 769 was taken from the water after it had passed through the en tire system of tanks, just before it was turned into the filter, and No. 779 was taken from the effluent from the sand filter. After the run ending July 24, 9.00 A.M., the stand pipe was removed from the top of the porcelain-lined tank, and p aced on the floor again, owing to the complications in the electrolytic cell arising from the sedimenta tion which took place when in the first-men tioned position. The remaining operations are described as those of Device No. 5. DEVICES OPERATED BY THE HARRIS COMPANY IN JULY. 287 CHEMICAL RESULTS DEVICE No. 4. Analyses of Samples Described Above. (Parts per Million.) Z i Date. 763 768 769 :: July 23 9-30A.M " 23 12. OO M. " 23 12. OO " " 23JI2.00 " 25.1 .37,24 -23 24-5 . .23J24.0 1-331 3-7 1.360 1.228 .132 .056 1.280 1. 168 . 112 .026 1.360,1 .1541 . 206^.080 .2o6|. 1... .[.038 DEVICE No. 5. The fifth and last device was presented for official inspection on July 27. Five runs were made from July 27 to August i, the device being operated day and night so far as was feasible. In this device, as in the preceding, all seven of the closed vessels were used as well as the filter. In all of the three small tanks were sets of aluminum electrodes and the magnets on these tanks were also used. The electrode manifold in the first tank was somewhat larger than in the second and third, the current (circuits in parallel) dividing as follows: Electrodes Number i, 7.6 amperes; electrodes Number 2, 5.8 amperes; elec trodes Number 3, 6.0 amperes. The total active area of the electrodes (anode surface) was about 4000 square inches. The river water entered the first small tank at the bottom, passed through it and thence upward through the stand pipe. This pipe was set on the ground as when first used, the top of the small tank and the bottom of the pipe being closed. The tin cones were used in this pipe as in the third device, and piping was provided to flush out the sediment at the base of the cones and the bottom of the pipe. From the stand pipe the water passed suc cessively through the second and third small tanks, thence through the three large tanks (used as settling chambers) and finally a por tion of it passed through the filter. From July 30, 5.50 P.M., to July 31, 4.00 P.M., the water filtered was taken from the third small tank, instead of from the third large tank which was the last one of the series. The records of operation of this device and the results accomplished therewith are pre sented in the following tables. As before, the records of operation of the device exclu sive of the filter, and of the filter itself, are presented separately. Following these tables are presented observations on the amount of aluminum used and the efficiency of the elec tric generating plant. RECORDS OF OPERATION DEVICE No. 5. Records of Operation of Device Exclusive of Filter. 1 2 3 4 5 Began j July 27, July 28, July 29, July 30, July 31, Ended j 7-47 P.M. July 28, J.20 P.M. July 29, 12.45 r - M - July 30, 5.30 P.M July 31, Aug. i. 9.OO A.M. 6.00 A.M. g.OO A.M. i8h. 3701. 9 -13 A.M. Period of filling, " . ... Cubic feet of water wasted ... 176 Average rate of treatment. Cubic feet per minute 3.28 I.I? I.I5 I .20 12 qoo 4.68 50 400 198 182 23 6 28.8 29. 1 31 . 1 Average electric horse-power per million callous per 24 hours f. . 186 605 61? 601 150 After syste itchboard readings of entii WATER PURIFICATION AT LOUISVILLE. Records of Operation of Filter. 1 Average Rate. hos Hours and Minutes. Cubic Feet. JiS Began. Ended. Cubic Ft. Gallons per Min. per Acre i July 27, 7.47 P.M. July 27 11.18 P.M. 4h. iSn . 4!]. 05111. " 13111. 21 .O 3-3 O. II 16 2 27, ii. 18 " 28 4.30A.M. 5h. I2n . 4h. 57m. 15111. 1Q.6 7-3 0.07 II 3 28, 4.30A.M. 28 5.50 " ih. 2om. ih. oom. 2Orn. 6.0 17-3 0.18 27 4 23, 5.50 " 28 8.00 " 2h. 32n . 2h. 15111. lym. 20.2 12. I 0.15 22 28 8 22 " 28 6 29, 12.25 l -M. 30 5.50 P.M. 2oh. 5811 . 20h. 38111. 20111. 255-8 10.5 0.21 3 7 30, 5-5" " 30 8.40 2h. som. 2h. 40111. loin. 20. O IO.O O. 12 18 8 30, 8.40 " 31 1. 25 A.M. 4h. 45tn. 4(1. 33111. 12m 35-" 12.0 0.13 19 9 31, 1.25 A.M. 31 6.20 " 3h. 5511 . 3". 45" . lorn. 2S.O 13-5 O. II 1 6 10 31, 5-20 " 31 3.50 P.M. 6h. 54 m. 61, . 47111. oym. 54-5 I 4 .8 0.13 9 II 31, 3.^0 I .M. 31 4.00 " oh. 3211 . oh. 2om. I2m. 4-4 13-3 O.22 33 12 3", 4-00 " Aug. I 9.43 A.M. I7h. I3in. 202.9 0.20 30 During runs numbers 7 to n, inclusive, the water to be filtered was taken from the third small porcelain-lined tank. During all the other runs it was taken from the last large tank. Results of Analyses Device No. 5. In the following tables are presented re sults of frequent analyses of the water before and after treatment by this device. Samples were collected from the water before treat ment, after it had passed through the first tank, after it left the last tank, and after it had passed through the filter. The first table shows the results of the bac terial analyses of the river water before treat ment, during the periods when this device was in operation. In the second table are given the results of bacterial analyses of the treated water after it had passed through the first tank. The third and fourth tables contain the re sults of bacterial analyses of the treated water from t he last tank and of the effluent from the filter, respectively. Following these are tables containing the results of the corresponding chemical anal yses, the time and place of collection being given by reference to the corresponding bac terial samples. When two or more bacterial numbers are given it is to be understood that portions collected at the same time and place as these bacterial samples were mixed and the analvsis made of the mixture. BACTERIAL RESULTS RIVER WATER. Number Colle cted. Bacteria per Cubic Sample. Date. 1896. Hour. Centimeter. 4726 July 27 g.30A.M. 16 ooo 473 27 5.00 P.M. 19 600 4741 27 9.00 " 14 400 4745 28 3.00A.M. 18 500 4750 28 9-30 " 24 500 4796 28 6.OO P.M. 12 400 4800 28 12.00 " 8 600 4804 29 6.0O A.M. 10 700 4812 4861 29 9.30 " 5 OO P.M. 9500 4866 29 30 9.30 A.M. 17300 4871 30 5.OO P.M. 6 800 4878 3i 9.30 A.M. 6 800 4881 3i 10.45 " 7900 4891 3i 3-35 P.M. 5 800 4898 3i 4.30 " 8 600 4899 3i 4-30 " 9 600 BACTERIAL RESULTS DEVICE No. 5. Effluent from first tank. Collected. Number Bacteria Sample. Date. 1896. Hour. Rate. per Cubic Centimeter. 4738 July 27 9.00 P M. 2.6 12 7OO 4742 28 3-00 A M. 2.5 12 900 4747 28 g.OO * 4-4 15 ooo 4797 28 6.OO P M. .0 2 900 4801 28 12. OO * .0 3900 4805 29 6.00 A M. .0 5700 4856 29 3.05 P M. 3 5 800 48663 3" g.OO A M. .2 5 800 4872 ; 30 5.30 P M. .0 i 500 4875 ; 31 8.OO A M. . I 3 600 4879 31 IO.2O " .2 5 600 4884 31 12.05 P .M. .2 2 800 4894 31 4.IO 4 .2 IO 100 4900 31 5.0O 3.8 4 600 4904 Aug. i 9.00A.M. . 4-0 2 OOO 1 Rate of treatment in cubic feet per DEVICES OPERATED BY THE HARRIS COMPANY IN JULY. 289 BACTERIAL RESULTS-DEVICE No. 5. Effluent from last tank. Number Sample. Collected. Rate.* Bacteria per Cubic Centimeter. Date. 1896. Hou , 473 July 27 g.OO P.M. 2.6 I 800 4743 " 28 3.OO A.M. 2.5 5 300 4748 " 28 9.00 " 4.4 15 ooo 4798 " 28 6.OO P.M. i-o 215 4802 28 12. OO " 10 192 4806 " 29 6.00A.M. i.o 252 4857 29 3.05 P.M. i . 3 700 48&7a " 30 g.OO A.M. 1.2 900 4876 " 30 g.OO " 1.2 900 48833 " 3 11.40 " I 2 310 4885 11 31 I2.O5 P-M. 1.2 25O 4895 " 31 4.10 " 4-2 7 800 4901 " 31 5.00 " 3-8 2 OOO 4905 Aug. I g.OO A.M. 4.0 3 5o ubic feet pe BACTERIAL RESULTS DEVICE No. 5. Effluent from Filter. Collected. Number Bacteria per of Sample. Date Hour. Rate.* Cubic 1896. 474 July 27 g.OO P.M. 0.26 132 4744 28 3.OO A.M. o . 09 So 4749 28 9.00 " 0. II I 800 4799 " 28 6.00 P.M. O.2O 262 4803 28 12 00 " 0.20 399 4807 " 29 G.OO A.M. 0.2O 188 4858 29 3.05 P.M. 0.25 171 4863 29 5.12 " O.24 700 4868a 30 g.OO A.M. o. 14 435 4873 30 5.30 P.M. 0.14 130 4877 31 S.OO A.M. 0.15 173 4880 31 10.20 " 0.13 300 4886 31 12.05 P.M. 0.13 505 4896 " 31 4.10 " O. 21 I 500 4902 31 5-00 " 0.24 700 4906 Aug. I 9.00 \. M. 0.23 700 Cubic feet per minute. CHEMICAL RESULTS DEVICE No. (Parts per Million.) < o Nitrogen Residue on Fixed Residue * ^ t e Evaporation. after Ignition. 1 a Date of Ku Collection. S. G c asAlbuminoid . 3 o jj "? h I u o S t- If Q| ^ e " E < z. X .-. h a. 3 i h c a c M Q i 781 July 27 27 18.8 I.I20 I .OO2 .118 .124 .019 .76.5 1496 1387 109 1356 1286 70 55-1 0.0 70.80 782 27-28 3i 13.2 .740 .622 .118 .092 .018 .6 6.5 1061 952 109 950 880 70 55.2 o.o 20.00 783 787* " 27-28 " 28-29 30 3-5 .I 5 8 .040 .118 .044 .on .6 g 6.5 166 57 [09 124 54 70 51.0 0.0 2.OO 788 789 " 28-29 " 28-29 .10 .12 2-4 2.1 .102 .080 .012 .090 .216 .144 .033 . IOO i: 6.0 4-0 "5 87 26 89 94 62 31 63 46.8 49-0 o.o o.o .70 . 12 794 " 29-30 .11 ii. 8 .630 530 . IOO . 2O2 .023 .8 T.I 926 826 IOO 841 774 67 56.0 0.0 34.50 795" 796" " 29-30 .11 3.0 .150 050 .100 .194 .032 9 3.0 194 94 IOO 163 96 67,52.2 0.0 6.50 799 1 " 31 13 10.2 .680 .5881.092 .248 .012 7 628 536 92 560 493 67 52.6 o.o 28.70 31 .07 2.0 .088 .ooo|.o88 . 140 333 .6 1.6 92 93 68 o (,- 53-J 0.0 .05 3 1 13 2.3 .no .018 .092 .208 .020 .6 ! i 130 3S IOI 34 7 51-1 o.o .70 31, Aug. i .. .27 11.3 .540 .418 . 122 .042 .006 .8 4.0 699 576 123 634 54490 62.0 o.o 25.30 805" 3 i I .27 ii. 9 .510 -.388 . 122 .040 .006 8 1-7 6^8 535 123 583 4939043.0 0.0 24.50 3 p " i .27 4.3 .198 .076 .122 .026J.OP4 .s ; - 204 -i 123 163 73 90 50.0 o.o 9.00 4799,4803,4807. 4856. 4806,1. "4857,4867,1. 4858.4868,1. 4879. "4080. "4883,1. i> 4900, 4904. < 4901, 4905. "4902,4906. SUMMARY OF ANALYTICAL RESULTS DEVICE No. 5. The degree of purification accomplished by the fifth device, at different points in the pas sage of the water through it, is indicated by the following tables. The removal of organic matter as indicated by the nitrogen in the form of albuminoid ammonia and oxygen consumed is first presented. Following this is given a summary of the bacterial analyses of the water at different points on its passage through the device. PERCENTAGE REMOVAL OF ORGANIC MATTER FROM THE RIVER WATER AT DIFFERENT POINTS ON ITS PASSAGE THROUGH THE DEVICE. Percentage Removal of Nitrogen as Albuminoid Ammonia. Point of Collection of Sample. First Tank. Last Tank. Filter. July 28 O 34 93 29 63 88 9 1 30 48 88 <H " 3i 77 Si Aug. i o o 58 Percentage Removal of Oxygen Consumed. July 28 o " 29 64 " 3 " 31 Aug. i 290 WATER PURIFICATION AT LOUISVILLE. SUMMARY OF BACTERIAL RESULTS. BACTERIA PER CUBIC CENTIMETER IN THE RIVER WATER BEFORE TREATMENT AND AI DIFFERENT POINTS ON ITS PASSAGE TH ROUGH THE DEVICE. Date. 1896. Ju y 27 Ji Iy 2 8 uly 28 July 28 July Julyag July 2 , Hour... River First Tank Last Tank Filter... . V 12 I M j 400 7OO 800 132 iS 500 24 500 1 2 900 1 5 ooo 5 300 1 5 ooo So i Soo 12 400 2 900 215 262 8 600 3 900 192 399 10 700 5 7 252 188 9500 5 Soo 700 171 Date. 1896. July 30 July 3 o Julys Julys Julys : July 3 Julys- Auif. Hour.. River.. First 9 A 17 .M. 3OO 5 P.M. 6800 SA.M 6 Soo I O A. .1 6 So . 12 M 3 7 900 4 P.M. 5 P.M. 9 100 9 100 (JA.M. 6 Suo tank. Last 5 Soo 1 500 3 600 5 60 3 2 SOC 10 100 4 600 2 000 tank. Filter. . 900 435 130 I 200 73 3 250 7 Soo 2 ooo 31 505: I 500, 700 3 500 70 J AVERAGE BACTERIA IN RIVER WATER AND AT DIFFERENT POINTS IN ITS PASSAGE THROUGH THE DEVICE, AND THE AVERAGE PERCENTAGE REMOVAL AT T River. HESE POINTS. First Tank. 6 320 40 T L ant ! Fi f. Average Average bacter percen a per c. c tage removal II 300 3 ooo 500 76 96 From July 27, 7.47 P.M., to July 30, 9.00 A.M., 37 440 gallons of water were treated and the weight of the aluminum electrodes decreased 18 620 grains, equal to 0.5 grain per gallon. From July 30, 9.00 A.M., to July 31. 12.07 P.M., 9465 gallons of water were treated and the weight of the aluminum electrodes de creased 7490 grains, equal to 0.78 grain per gallon. From July 31. 12.07 i -M-> to Aug. i, 9.43 A.M., 34 500 gallons of water were treated but there was no reduction in weight of elec trodes, owing to a failure in the electrical con nections. It is to be stated that, of the amount of alu minum noted above, a considerable portion passed into the oxide state; scaled off the electrodes, fell to the bottom of the tanks and was of no aid in the purification of the water. Further, it was impracticable, under the ex isting circumstances, to determine how much of the aluminum which was removed from the electrodes by the electric treatment was actually utilized in the purification of the water. Arrangements were made to study this point, but after some work had been done it was seen that no results of practical value were being obtained, because it was not known how much aluminum in the form of available hydrate had actually been added to any sample of water for analysis. This work was stopped and taken up under more favorable conditions after Aug. i and is described in the next chapter. Efficiency of the Electric (Generating Plant of the Harris Magneto-electric ll aicr Purification S \stcin. Three tests were made of the relation be tween the steam and electrical power, show ing the average combined efficiency of the generating plant to be 21.8 per cent. STATUS OF THE SITUATION ON AUGUST i, WITH REGARD TO THE MAGNETO-ELEC TRIC SYSTEM AND DEVICE. The more important features of the situa tion on Aug. i, when the investigations upon the other systems of purification were brought to a close, may be briefly summa rized as follows: 1. The original system was a complete fail ure at the time of official examination; and, practically speaking, it was abandoned there after by the Harris Company. 2. It is possible to purify the Ohio River water to a greater or less degree by the appli- :ation of aluminum hydrate prepared electro- lytically from metallic aluminum. With other ivaters purification in this or a similar manner had been previously claimed by other per- ions. 3. The available data, so far as they went, indicated that the direct effect of the electric action and the magnetic action was of little r no practical value. 4. It is possible to purify the Ohio River DEVICES OPERATED BY THE HARRIS COMPANY IN JULY. 291 water to a satisfactory degree with electroly- tically formed aluminum hydrate by the em ployment of subsequent sedimentation and filtration. Further, no lime in the river water is required with this process, and there is no increase in the corroding and incrusting con stituents of the filtered water. 5. The available data, so far as they went, indicated that the manner of purification last stated would be costly to an excessive and un reasonable .degree if applied to the water supply of a large city. It appeared, further, that in continuous service it might allow per iods of marked irregularity in efficiency. 6. Viewed from a practical point, the data which it was feasible for the Water Company to obtain, under the existing circumstances, were very limited and unsatisfactory. This was due chiefly to the poor and ill-advised construction of the devices as they were as sembled together, and, further, to the fact that the total period, during which the de vices were operated for official inspection, was less than 119 hours. 7. The situation may be comprehensively described by the statement that this method of purification had not advanced beyond the experimental stage, where laboratory experi ments are best adapted to show the practica bility and merits of the method. From the above statements it is plain that at the close of the regular investigations the situation with regard to the magneto-electric devices was quite unsatisfactory, from the Water Company s point of view, as well as from thai of the Harris Company. The indications were that the method was not a commercial success, so far as the purifi cation of the Ohio River water is concerned. But, owing to the unsatisfactory manner of the construction and operation of the several devices, the very limited evidence was ob scured by a variety of complications; and there was really obtained very little informa tion of positive value. A conference was held at this time by the officers of the Water Company to discuss the situation. The advisability of continuing the investigation with the makeshift devices offered for inspection was out of question. The real problem was to decide whether it would be wiser for the interests of the Water Company to dismiss the subject as it stood, or to have the fundamental principles of the method investigated by the experts who were regularly employed by the Water Company at that time. It was decided to pursue the latter course. The Chief Chemist and Bac teriologist was instructed to retain the ser vices of the assistant analysts for two or three weeks or so. An arrangement was also made with the Harris Company to make use of their dynamo and other electrical appliances; and the services of the man who had operated the devices were also availed of. The results of these investigations by the Water Company are presented in the next chapter. 292 WATER PURIFICATION AT LOUISVILLE. CHAPTER Xii. INVESTIGATIONS BY THE WATER COMPANY IN AUGUST INTO THE PRACTICABILITY AND ECONOMY OF THE DEVICES OPERATED BY THE HARRIS COMPANY. IN this chapter there are recorded the re sults of the various experiments, which were made by the Water Company under my direc tion during August, 1896, with the view to learning the practicability and economy of the principles upon which the magneto-electric treatment of water was claimed to be based. The principal points of importance, to which attention was directed, may be out lined as follows: 1. The direct and indirect effect of the ap plication of electricity upon the bacteria and organic matter in the Ohio River water. 2. The direct and indirect effect of the ap plication of electricity in the purification of the Ohio River water through the formation of aluminum hydrate, from metallic aluminum electrodes. 3. A comparison of the efficiency, in the purification of the Ohio River water bv coagulation and sedimentation, of aluminum hydrate formed electrolytically from the me tallic aluminum, and of aluminum hydrate formed from the decomposition of sulphate of alumina by lime. 4. The effect of the action of the electro magnets upon the constituents of the Ohio River water, and upon the rate at which alu minum hydrate is formed from aluminum electrodes. 5. Rate at which aluminum hydrate is formed electrolytically from metallic alu minum. 6. Regularity during continuous service of the rate at which aluminum hydrate is formed electrolytically from metallic alu minum. 7. Amount of metallic aluminum which is wasted in the electrolytic production of alu minum hydrate. 8. A comparison of the cost of aluminum hydrate formed by the application of sulphate of alumina, and by the action of electricity on metallic aluminum, respectively. 9. Observations on the amount of power which would be required to produce alu minum hydrate in a large scale by means of electricity. The experiments which were made on the several problems will be described in order. THE DIRECT AND INDIRECT EFFECT OF THE APPLICATION OF ELECTRICITY UPON THE BACTERIA AND ORGANIC MATTER IN THE OHIO RIVER WATER. Experiment No. i. This experiment was made to learn the effect of a current of high voltage, such as used in the " spark drum " of the original system (Chapter X). River water was treated with an electric current of high voltage for one hour in the porcelain-lined " spark drum " with the poles 2.56 inches apart. The voltage of the current as it left the Ruhmkorff coil used could not be accurately measured, but was estimated by the representative of the Harris Company to be about 200.000. with an amperage equiva lent to a very small fraction of one unit. This experiment was performed in duplicate. The temperature of the river water increased about 2 C. during this treatment for one hour. At the beginning of the experiments the temperature of the river water was 31 C. The average results of bacterial analyses of a INVESTIGATIONS OF THE PRINCIPLES OF THE HARRIS DEVICES. 29,5 corresponding set of samples in the two e periments are as follows: REMOVAL OF BACTERIA IN OHIO RIVER WATER BY ELECTRIC TREATMENT IN THE "SPARK DRUM." Length of Bacteria Length of of River Water. per Cubic Centi of River Water. per Cubic Centi D age 1 Minutes. me er. Minutes. meter. 3 23 600 M 3" 1 6 goo 38 5 20 500 25 45 17 2OO 37 10 2O f)OO 25 60 14 loo 48 The carbonaceous organic matter in the river water was reduced by this treatment, as indicated by the " oxygen consumed," from 9.8 to 7.9 and 7.3 parts per million in 15 and 60 minutes, respectively. No effect was ap parently produced upon the nitrogenous or ganic matter in the river water. Experiment No. 2. This experiment was made to learn the effect of currents of comparatively high density, and to study the effect of amount of current, it was performed in duplicate with small carbon electrodes using as high amper age and low voltage as it was practicable to obtain under the circumstances. River water was treated in a glass jar for 10 minutes, when the temperature of the water became so high that it was necessary to stop the experiment in each case. The average amperage and voltage of the current in these experiments were 10.5 and 40, respectively, and the density of the current was about 90 amperes per square foot of cross-section of electrolyte. The average results of bacterial analyses of a corresponding set of samples in the two experiments were as follows: REMOVAL OF BACTERIA- IN OHIO RIVER WATER BY ELECRIC TREATMENT, AS ABOVE STATED. Length of Treat ment of River Water. Minutes. Ampere-hours per Gallon. Bacteria per Cubic Centimeter. JuSST I 5 10 O o. 23 0.70 I-I7 2.33 28 500 22 2OO 13 30O 7 7" 2 700 22 53 73 Qi The average temperature at the close of the tests was 53 C., which was so high that by actual experiment it was found to be the explanation for the removal of 70 per cent, of the bacteria originally present in the river water. There was a noticeable disintegration or change in the nitrogenous organic matter, but there was practically no purification by the direct application of electricity. The change in carbonaceous organic matter could not be determined owing to the presence of carbon from the electrodes. No apparent effect directly or indirectly on the rate of sedimentation of the river water was noted after these electric treatments with carbon electrodes. It may be added here that experiments in 1897 bore out this fact that, directly, the elec tric current had no practical effect in the purification of the Ohio River water. THE DIRECT AND INDIRECT EFFECT OF THE APPLICATION OF ELECTRICITY IN THE PURIFICATION OF THE OHIO RIVER WATER THROUGH THE FORMATION OF ALUMINUM HYDRATE FROM METALLIC ALUMINUM ELECTRODES. In this connection experiments similar to the one last described were made except that metallic aluminum electrodes were used in place of carbon ones. River water was treated in a glass jar for 10 minutes with a current having an average amperage and voltage of 6.3 and 97, respectively. At the conclusion of the experiment the contents of the jar were well mixed, and a portion re moved for analysis. Sedimentation rapidly took place in the re maining portion in the jar. The supernatant liquid was clear and of a satisfactory quality with regard to organic matter and bacteria. About 0.75 gallon was used in this experi ment, and the amount of treatment was 1.05 ampere hours. Assuming full rate of decom position, the amount of aluminum added to the water was therefore about n.o grains. This indirect purification of the river water by means of the production of aluminum hy drate was very marked and will be more clearly presented in the next section. For the purpose of learning the direct effect of this electrical treatment, independent of the subsiding action, analyses were made of 294 WA IER PURIFICATION AT LOUISVILLE. the unsettled portion of the treated water. Here it may be stated that analyses of the gas, which was liberated quite rapidly from the negative pole, showed ii to he composed very largely, if not wholly, of hydrogen. The number of bacteria in the river water before and after this treatment were found to be 3800 and 4200 per cubic centimeter, respect ively. The results of chemical analyses of the water befc re and after treatment without sub sidence, represented by samples Nos. 807 and 808, respectively, are presented in the next table. It will be noted that the oxygen con sumed (carbonaceous matter) was reduced from 9.7 to 8.3 parts. With regard to the nitrogen as free ammonia and albuminoid ammonia, the results show that a certain change was effected in the nitrogenous or ganic matter. Liut as there was no marked increase in nitrogen in the oxidized form of nitrites or nitrates, this condition cannot be regarded as one of purification, it was simply an initial step in that direction. This experiment, which was duplicated in its most important parts, leads to the conclu sion that the direct action of the electricity applied to the river water by means of alu minum electrodes, without subsidence, ef fected practically no purification. RESULTS OF CHEMICAL ANALYSES OF SAMPLES AS DESCRIBED ABOVE. (Parts per Million.) Collected Nitrogen Residue on Fixed Residue ^ 6 as Kvapoiation. after Ignition. B - u 8 Ammonia o .,. 3 1 Date. Hour. U c i rt C a Z U | u a | g & 3 1 ,i j ^. y. | 3 Q. | - g 1 = c . I H U o H I a a h 1 C ^ 1 C < 5 i 807 AlIK- 7 298 6 ^ 1 808 7 .03 8 3 . 520 .440 .080 .224 .007 .6 5-o 839 737 102 686 620 66 48.0 0.2 23.OO A COMPARISON OF THE EFFICIENCY, IN THE PURIFICATION OF THE OHIO RIVER WATER BY COAGULATION AND SEDI MENTATION, OF ALUMINUM HYDRATE FORMED ELECTROLYTICALLY FROM ME TALLIC ALUMINUM, AND OF ALUMINUM HYDRATE FORMED FROM THE DECOM POSITION OF SULPHATE OF ALUMINA BY LIME. Preliminary experiments indicated that aluminum hydrate, formed from the decom position of sulphate of alumina, was some what more effective in the purification of the river water than was the case with the same amount of aluminum hydrate when pre pared electrolytically from metallic aluminum. The reason of this appeared to be that the particles of aluminum hydrate in the first instance were smaller and more numer ous than in the case where this compound was formed clectrolytically. This explanation of the observations seemed plausible by virtue of the fart that electrolytically formed aluminum hydrate, being present in larger particles at the beginning, could produce fewer centers of co agulation; and that, since the purification was brought about by coagulation and sedi mentation, the finer particles of the* precipi tate formed from sulphate of alumina gave a better opportunity for the cumulative, en veloping and coagulating action of the alu minum hydrate. Three sets of carefully prepared experi- ments were made to obtain accurate data upon this point, and to learn the correctness of the theory presented above. A descrip tion of each set of these experiments is next presented, followed by a summary of the practical results obtained therefrom and a record of the results of analyses. Experiment No. T. On August 1 1 three separate portions of the same sample of river water were treated, re spectively, with 0.88 grain of aluminum hy drate prepared in the following manner: In the first portion this amount of alumi- INVESTIGATIONS OF THE PRINCIPLES OF THE H. DEVICES. 295 num hydrate was prepared electrolytically in the water from metallic aluminum. In the second portion the given amount of aluminum hydrate was obtained by adding a calculated quantity of sulphate of alumina, which was decomposed in the water by the lime and magnesia present there. in the third portion there was added the stated amount of aluminum hydrate in the form of a gelatinous mass, prepared from suit able chemicals in the laboratory. This amount of aluminum hydrate (1.08 grains per gallon) corresponds to 0.31 grain of metallic aluminum per gallon of water. The experiments were made with about 2 gallons of water. At the conclusion of the treatment the water was well mixed and allowed to remain undisturbed for 24 hours. Samples of the clarified water were removed by means of a siphon and subjected to chemical and bacterial analyses with results which are pre sented beyond, together with those of the un treated water. Experiment No. 2. On August 14 the above experiment was duplicated with the river water of that day, with the exception that the amount of alu minum hydrate added in each case was 0.51 grain per gallon, equivalent to 0.18 grain of metallic aluminum per gallon of water. Chemical and bacterial analyses were made of the river water after subsidence for 24 hours. Determinations were also made of the number of bacteria and amount of sus pended matter in the clarified water after sedimentation for 5 and 20 hours, respec tively. Experiment No. j. On August 18 the above experiment was duplicated, for the most part, with the river water of that day. The amount of aluminum hydrate, and its equivalent of metallic alu minum, were 0.14 and 0.05 grain per gallon, respectively. In this experiment an important modifica tion in the manner of application of alumi num hydrate to the third portion of the water was made, with the view to demonstrating more positively the accuracy of the above stated theory in explanation of the varying results obtained. Instead of adding to the third portion the stated amount of aluminum hydrate in the form of a gelatinous mass, the following procedure was adopted: The calculated amount of sulphate of alu mina (the same as was added to the second portion) was added to 10 per cent of the pre scribed volume of water. After standing 15 minutes, during which the sulphate of alu mina was decomposed by the lime and mag nesia into aluminum hydrate, the remaining quantity of untreated river water was added. In this manner the prescribed amount of alu minum hydrate was added in the form of larger initial particles, as it was mixed with the full quantity of water, than in the case of the second portion. Samples of water were collected for analy sis after the plan stated in the last experiment. In the following tables are summarized the principal analytical results obtained from each portion of water treated in these three sets of experiments. The first table contains the percentages of removal from the water, after 24 hours subsi dence, of organic matter, as indicated by al buminoid ammonia and oxygen consumed, respectively; of the suspended matter, and of the bacteria. The second table contains the results of the determination of the respective amounts of suspended matter in the several treated waters of experiments Nos. 2 and 3, atfer sub sidence for different periods. The third table contains the number of bac teria found in the several portions of treated water in experiments Nos. 2 and 3, after dif ferent periods of subsidence. SUMMARY OF ANALYTICAL RESULTS FROM THE LAST THREE EXPERIMENTS, SHOW ING THE PERCENTAGES OF REMOVAL OF ORGANIC MATTER, SUSPENDED MATTER, AND BACTERIA, AFTER 24 HOURS SUB SIDENCE. Percentage Removal of Albuminoid Ammonia. Number of Por ion of Water Trea ted. Kxperiment. Firs!. Second. Third. I 2 3 f>4 83 43 80 93 69 55 80 58 296 WATER PURIFICATION AT LOUISVILLE. Percentage Removal of Oxygen Consumed. Portion of Water Treated. Number of Experiment. First. Second. Third. I 66 87 54 2 80 92 73 3 52 67 65 Percentage Removal of Suspended Matter of the River Water. i 93 IOO 78 2 96 100 93 3 78 96 88 Percentage Removal of Bacteria. i 37 86 37 2 74 94 57 3 39 88 73 SUMMARY OF ANALYTICAL RESULTS FROM EXPERIMENTS NOS. 3 AND 3, SHOWING THE AMOUNTS OF SUSPENDED MATTER IN THE WATER AFTER DIFFERENT PERIODS OF SUBSIDENCE. Experiment No. 2. Period of Subsidence. Hours. Parts per Million of Suspended Solids. First Portion. Second Portion. Third Portion. 5 20 I 397 136 60 i 397 5 4 I 397 199 99 24 52 o 99 Experiment No. 3. Period of Parts per Million. Suspended Solids. Hours. First Portion. Second Portion. Third Portion. 243 243 243 6 104 33 83 20 56 13 33 24 54 10 33 SUMMARY OF ANALYTICAL RESULTS FROM EXPERIMENTS NOS. 2 AND !!, SHOWING THE NUMBERS OF BACTERIA IN THE WATER AFTER DIFFERENT PERIODS OF SUBSI DENCE. Experiment No. 2. Bacteria per Cubic Centimeter in River Water Period of after Treatment. Hours First Portion. Second Portion. Third Portion. O I 7 600 17600 I - 600 5 2O 700 900 T7400 20 5900 600 7700 24 4500 I 2OO 8 loo Experiment No. 3. o 25 600 25 600 25 600 6 18 200 10 800 1 8 900 20 16 700 3 500 7 ooo 24 15 600 ; _< i. < 7 ooo On the next page are given the results of chemical analyses of the original river water and of the several portions of treated water in each experiment after subsidence for 24 hours. As already stated, samples of the treated water were removed from the bottles by means- of a siphon, without disturbance of the solid matters which had settled to the bottom. Attention is especially called to the fact that the electrolytically produced aluminum hy drate, as well as that prepared in the labora tory, did not decrease the alkalinity of the water. This means that no carbonic acid gas was set free, and that no lime passed into the form of sulphate, because the formation of this compound in these cases was independ ent of the lime in the water. The samples of water for chemical- analysis were numbered as follows: Treated Water after 24 Hours Subsidence. Original Water. First Portion. Second Third 809 810 Sn 812 813 814 815 816 . 819 820 821 THE EFFECT OF THE ACTION OF THE ELEC TRO-MAGNETS UPON THE CONSTITUENTS OF THE OHIO RIVER WATER AND UPON THE RATE AT \VHICH ALUMINUM HY DRATE is FORMED FROM METALLIC ALUMINUM. So far as could be learned from a series of experiments directed to this point the electro magnets produced no appreciable effect in the purification of the Ohio River water, either directly or indirectly, by facilitating subsi dence or increasing the rate of formation of aluminum hydrate. RATE AT WHICH ALUMINUM HYDRATE is FORMED ELECTROLYTICALLY FROM METALLIC ALUMINUM. One of the most important points in con nection with the treatment of water by elec trical devices is information upon the cost of the production of aluminum hydrate. Tt was definitely known that the rate of forma- INVESTIGATIONS OF THE PRINCIPLES OF THE HARRIS DEVICES. 297 RESULTS OF CHEMICAL ANALYSES OF SAMPLES DESCRIBED AMOVE. (Parts per Million.) i - z Collected. u ^ \ \ u ~t f i 2 -6 1 U o Nitrogen 1 R t Ev; poralK , , Fixed Residue after I S n,ti,,i,. i 3 < -a a c Albuminoid i - B "< i X iz Date. 1896. " h 1 T3 1 e H i c c * J5 s < a d o 3 U if 809 810 515 813 814 815 816 818 819 820 -i Aug. I I i i iS :: 18 .30 .u .09 .23 . its 9.8 3.3 1-3 4-5 19.7 3-9 1.6 4-4 6.6 3.2 2.2 - - 430 1=6 084 "94 I 040 1 68 074 200 2U> 142 050 I 10 .292 .018 .000 .056 932 .060 .000 .002 ,?6 .035 .O .O >2 i .138 .032 .138 .030 084 .022 138 .020 loS 046 108 .016 074 .022 loS .020 104 .026 104 .010 080 .010 086 .010 ooo 003 ooo ooo 004 002 OO2 OOO ooo O02 CO 2 OOI 7 .6 .6 .6 7 7 .6 .6 7 .6 .6 .6 5-3 5-4 5-3 5-3 6.1 5-5 6.1 6.1 5-9 5-7 5-7 5-9 520 158 163 219 I 524 T.8S 1 60 235 380 193 141 161 389 27 88 I3SS 52 o 99 241 54 10 3 131 131 163 131 136 136 1 60 136 139 139 131 131 455 112 IlS 169 I 4 .8 150 119 193 33 54 106 116 375 32 o 89 i 324 50 o 99 241 64 u 21 80-5.4 8075.0 11821.5 8075.4 9474-0 9473-0 11946.2 9474-0 90 74 . o 9072.0 9566.0 95 66. o 0.0 o.o o.o o.o o.o o.o o.o o.o o.o 0.0 0.0 o.o 14 40 2.60 0.07 6.50 46.00 6.00 0.07 5-00 12. OO 4 oo 0.80 1.48 .03 .18 .IS .16 .16 tion of this compound would be proportional to the strength (amperage) of the current, pro vided no irregularities occurred. But ade quate data were lacking, both with regard to the rate of formation and to the likelihood of irregularities such as would affect the cost and efficiency of the treatment. In the course of 54 experiments, which were made during the first three weeks in August, attention was directed to this point so far as was practicable. The experiments were made in glass jars of one gallon capac ity and in the porcelain-lined tanks described in the last chapter. Early in the work it was learned that the quality of the water appar ently exerted little or no influence upon the rate of formation of aluminum hydrate, or the influence of other factors disguised its effect. In a majority of cases river water, from which the suspended matters had been removed by filtration, was employed. The amount of aluminum hydrate, which was formed under certain recorded conditions, was determined both by weighing the amount of hydrate formed and by noting the decrease in the weight of the metallic aluminum electrodes. Several important observations were made as to the manner in which the aluminum hy drate was formed. When both electrodes were of metallic aluminum the electric cur rent caused the hydrate to appear at both poles, but for the most part at the positive pole. A considerable quantity of gas was liberated at the negative pole. Gas was also set free at the positive pole, but less uniformly and only about .20 per cent, of the volume of that which came from the negative pole. Analyses of the gas from the negative pole showed it to be practically all hydrogen. If there were any oxygen (in excess of that com ing from the atmosphere), ozone, or hydro gen peroxide present in the water or gas, the quantities were so small that they escaped de tection. Upon shutting off the electric cur rent and allowing the electrodes to remain in the treated water it was repeatedly noted that aluminum hydrate continued to appear for 20 seconds or more after the current ceased. This observation in connection with other information indicated that the electric current does not form the hydrate directly from the metallic aluminum, but it prepared the water by partial decomposition so that certain constituents of the water were able to produce the aluminum hydrate. When an electric current had passed through the aluminum electrodes for several hours it was found that the positive poles be came coated with aluminum oxide (aluminum rust), and that from time to time this coating fell off in the form of scales of a considerable size and number. At the negative poles there appeared a fine black coating, which was found to be composed of minute particles of aluminum in a (spongy) metallic state. From an economical and practical point of view these observations have considerable significance, as will be shown in following sections. In the next table are recorded all the re- WATER PURIFICATION AT LOUISVILLE. suits of determinations of the rate of forma tion of aluminum hydrate, electrolytically. These results are expressed in the number of grains and grams of aluminum decomposed from the metal electrodes in one hour for each ampere of electric current, according to the conventional method. The weights of aluminum hydrate which these amounts of metal would form may be obtained by multi plying the respective figures by 2.85. The results were obtained from small electrodes (of bright metal at the beginning of each ex periment) during short periods, as a rule, in order that the maximum results, free from ab normal irregularities, might be obtained. The irregularities have already been referred to and will be mentioned again beyond. To make a comprehensive study of them on a small scale was impossible. It will be noted that a variety of combinations of electrodes were tried. The experiments whose numbers do not appear in this table are referred to the foregoing portion of this chapter. The tabu lations are self-explanatory, except that it is to be staled that chemical symbols are used as abbreviations of the substances used as electrodes. Thus, Al means aluminum; C, carbon; and Pt, platinum. SUMMARY OF RESULTS SHOWING THE RATE OF ELECTROLYTIC DECOMPOSITION OF METALLIC ALUMINUM. c s Cur Cu rrenton Electrod es. g Rat .of S S ^ f c n a Decompt Metallic i Uilmlnurn C o 2 S Elect odes. Mag rets. ^ :re-Hour. a *"5 c $ Amperes Volts. .E bj ** i ~ " 1 ^IT; a E a a ^ . I| |j. t? a I = 1 c be a g u sj c 1 z Q < y. > ^ S < "- S < H O O I oh. lorn. 300 Al. Al. " 1.2 0.9 I.OO 134 131 133 0.17 12.3 0.80 2 oh. lorn. 300 o 3-4 o.S 1.23 228 223 224 0.20 6.0 0-39 3 oh. 05111. 300 o 9.0 4.1 5.65 95 72 84 0.47 14.2 0.92 6 oh. osm. 300 o o 10.0 2.0 5.20 101 83 91 0.43 9.2 0.60 60 oh. 05111. 300 o o S.o 6.0 6.79 93 83 88 0.57 8.0 0.52 7 oh. O5m. 300 o o S.i 5.6 6.55 98 Si 90 o.55 8.9 0.58 7" oh. 05111. 300 o 8.3 6.0 6.86 93 So 87 0-57 9.0 0.58 15 oh. lorn. 200 o o 7- 2 4.1 5.08 113 98 07 0.85 7-2 0.47 16 oh. lorn. 200 o 6.2 3-9 4.76 in 98 05 0.79 9.0 0-59 17 oh. lorn. 200 o o 5.6 5-o 5.io 108 105 06 0.85 9-2 0.60 22 oh. Tom. 200 o o 5-7 4.1 4-75 112 IO2 07 0.79 9-2 0.60 37 oh. 0501. 2OO o o 6.5 5-73 107 103 05 0.48 8.3 0-54 39 ih. oorn. 2OO o o 8.3 s .s 5.20 113 95 05 5.20 8-7 0-57 ih. oom. 200 o S.I 3-8 5-27 1 II 85 03 5.30 6.2 0.41 42 4h. 30111. 2OO 1 1. o 4-3 i-7 2.33 131 114 25 10.50 7-6 0-49 48 oh. 05111. IOO Pt.* Al. o o 9.8 8.7 9-44 231 230 31 0-79 3-9 0.25 49 oh. 0301. IOO Al. Pt. o o 10.5 8.4 9.42 210 207 09 0.47 5-2 o-34 50 oh. lorn. IOO " C. o o 3-9 3-6 3-83 213 212 12 0.64 6.7 0.44 51 oh. 05111. IOO C.* Al. o 6.4 5.0 6. 02 215 213 14 0.50 4-2 0.37 52 3h. oom. 2OO Al. 7.2 222 8.3 3-5 5.30 112 85 92 15.90 8.0 o. 52 53 2h. oom. 200 Ironf 7-2 222 2-3 2.1 2.18 215 189 2OO 4-36 8.6 0.56 carl on or platinum) " "not num hj ndersu drale i n those cases v ,here m tallic al u,,,inu,n formed of elm tivepol only (the results, e being jut with ment was made. This pipe led to the In the consideration of the results in the foregoing table it is to be borne in mind that the amounts of metal decomposed include the oxide as well as the hydrate of aluminum; as the oxide came off irregularly in films or scales, it is probable that the h .ghest results are associated with this factor. Concerning the low results, the utilization of electric cur rent in the formation of oxygen at the posi tive pole appears to be the explanation of them. Taking the average of all these experi ments, it is found that rate of decomposition of aluminum was equal to 8.16 grains or 0.53 gram per ampere-hour. The entire current used in this set of experiments was 50.42 ampere-hours, and the to tal amount of metal decomposed was 395.3 grains or 25.67 grams. Computing the rate of decomposition on this basis, the rate of 7.84 grains or 0.509 gram per ampere-hour is obtained. This is practi cally an exact agreement with the results of INVESTIGATIONS OF THE PRINCIPLES OF THE HARRIS DEVICES. 299 Watt (" Electro-Deposition of Metals," page 548), who states that the rate of decomposi tion of aluminum is 7.92 grains or 0.514 gram per ampere-hour. This subject is taken up more at length in Chapter XV, Section No. 4, where an ex planation of the difference between the above rate of decomposition and that indicated by the electro-chemical equivalent of aluminum is offered. REGULARITY DURING CONTINUOUS SERVICE OF THE RATE AT WHICH ALUMINUM HYDRATE is FORMED ELECTROLYTI- CALLY FROM METALLIC ALUMINUM. When the metallic aluminum electrodes were bright and clean the rate of formation of aluminum hydrate per ampere-hour, with other conditions equal, seemed to be fairly uniform. After a time, however, the positive pole became coated with oxide of aluminum, and the rate of formation of the hydrate de creased. In time this coating of oxide be came so thick that it fell off, and the rate temporarily increased. On a large scale of operation this would doubtless be an important matter for consid eration. From the available evidence, with runs of not more than five hours, variations of 20 per cent, in the formation of aluminum hydrate per ampere-hour were noticed, when the conditions other than the coating on the electrodes were apparently parallel. AMOUNT OF METALLIC ALUMINUM WHICH is WASTED IN THE ELECTROLYTIC PRO DUCTION OF ALUMINUM HYDRATE. Upon passing electricity through alu minum electrodes this metal passes into three different forms: 1. Aluminum hydrate which appears at both poles, but for the most part at the posi tive pole. It is this form which serves in the purification of water as has been described and discussed. 2. Some of the metal passes into the solu tion and is deposited at the negative pole. This is probably deposited in a spongy metal lic state somewhat similar to that noted under some conditions with other metals in electro plating. The amount of the metal which passes into this state is very small as far as could be learned. 3. After the positive pole has been in ser vice for some time it becomes coated with a layer of aluminum oxide, which, as stated above, gradually becomes thicker until it scales and falls off. The aluminum oxide serves no purpose in water purification and is a waste product which would be expensive in a large plant. The results of the experiments show that the amount of metallic aluminum which was wasted by passage into the oxide state ranged from 25 to 40 per cent, of that which was converted into the form of aluminum hydrate. It may also be added that this was borne out by the experiences with the devices operated in July. In two instances during the last week in July where the periods of observed opera tion were 5 and 20 hours, the percentages of metallic aluminum which were wasted were estimated to be 51 and 53. respectively. A COMPARISON OF THE COST OF ALUMINUM HYDRATE FORMED BY THE APPLICATION OF SULPHATE OF ALUMINA TO \VATER, AND BY THE ACTION OF ELECTRICITY ON METALLIC ALUMINA, RESPECTIVELY. It has already been stated that of the two methods of production of aluminum hydrate the electrolytic one furnished a product which was less efficient in the coagulation and sedi mentation of the Ohio River water, and which was associated with a wastage of 25 to 40 per cent, of the metallic aluminum from which it was prepared. Aside from the question of cost of power to produce aluminum hydrate electrolytically, which is referred to in the next section, there is a marked difference in the cost of the commercial product used in the two methods of preparation of aluminum hy drate. Current quotations (Aug., 1896) give the cost of metallic aluminum and sulphate of alu mina as 56 and 1.5 cents per pound, respec tively. To produce 100 pounds of aluminum hydrate by the two methods it would require 35 pounds of metallic aluminum and 420 pounds of sulphate of alumina of average composition, respectively. The cost of these two commercial products, on the above basis, would be $19.60 and $6.30, respectively. The 300 WATER PURIFICATION AT LOUISVILLE. facts show conclusively that electrolytically prepared aluminum hydrate is more than three times as expensive as that prepared from sulphate of alumina, independent of the wastage of metallic aluminum and the cost for electric power. OBSERVATIONS ON THE AMOUNT OF POWER WHICH WOULD BE REQUIRED TO PRO DUCE ALUMINUM HYDRATE ON A LARGE SCALE BY MEANS OF ELECTRICITY. In the purification of the Ohio River water by the aid of aluminum hydrate prepared by the two methods under consideration, the first point of practical importance which suggests itself is this one: The electric method would require boilers, engines, and a generating plant large enough to supply the needful amount of aluminum hydrate when the water is muddiest, while in the other method such a condition of the water would require in ad dition to the normal appliances simply a stronger solution of sulphate of alumina and a more rapid addition of the chemical solu tion. To estimate the amount of electric power necessary for the formation of aluminum hy drate from metallic aluminum in the purifica tion of the Ohio River water, the necessary quantity (amperage) and electric-motive force (potential) of the current must be known. The formation of hydrate is proportional to the amount (amperage) of current and. data on this point are presented in the foregoing tables. \Yith regard to the voltage (corresponding to pressure in hydraulics) necessary in a plant on a large scale, this cannot be stated from the evidence available at this time. It would doubtless be much less than 210, however, which was the voltage supplied to the devices operated by the Harris Company. As a matter of convenience and record, it may be stated that to treat one million gal lons of water in 24 hours with electrolytically prepared aluminum hydrate equivalent to one grain per gallon of sulphate of alumina, the number of horse-power of electricity actually- used, according to these tests, would be about one-half the potential of the current. CONCLUSIONS. From these investigations made in August upon the purification of the Ohio River water by electricity the following conclusions may be drawn: 1. The direct application of electricity and electro-magnets, as used in these devices, produced no substantial purification of the Ohio River water. 2. The electrolytic formation of aluminum hydrate in the Ohio River water enabled sub stantial purification to be effected, provided sedimentation and filtration were subse quently employed, as was similarly done by the other systems investigated in these tests and whic h used aluminum hydrate obtained from the application of sulphate of alumina. 3. The use of electrolytically prepared alu minum hydrate has the advantage of the ap plication of sulphate of alumina in the follow ing points, as was indicated to be the case by the operation of the devices of the Harris Company in July, as stated at the close of Chapter XL A. No lime in the river water is required for the successful application of the process. B. There is no opportunity of getting dis solved chemicals into the filtered water. C. There is liberated no carbonic acid gas, which in time might injure boilers and dis tributing pipes. /.). There is dissolved in the filtered water no additional sulphate of lime, which at times might give annoyance and trouble when the water is used in boilers. 4. For the purification of the Ohio River water on a large scale the electric method, as compared with the method in which sulphate of alumina is used, is not a commercial suc cess because of its excessive cost. During 1897 some further investigations under different conditions, in connection with the use of metallic aluminum electrolytically decomposed for the coagulation of the Ohio River water, were made by the Water Com pany. The results of these investigations are recorded in Chapter XV, but it may be noted here that the conclusions given in this chapter were confirmed. THE MARK AND B ROW NELL ELECTROLYT1CAL DEVICES. CHAPTER Xlil. DESCRIPTION OF THE MARK AND BROWNELL ELECTROLYTICAL DEVICES,* AND A RECORD OF THE RESULTS ACCOMPLISHED THEREWITH. AFTER devoting the early part of Decem ber, 1896, largely to investigations on the in creased corroding and incrusting power of water which had been treated with sulphate of alumina, attention was again directed to methods of preparing coagulants by electro lytic means, whereby the use of commercial sulphates might be avoided. Several Louisville gentlemen called the at tention of the Water Company at this time to some electrolytical experiments upon water purification, which had been in progress for several weeks. At their request, the directors and officers of the Water Company, on Dec. 21, 1896, inspected the operation of two small experimental devices along this line, which were operated by Profs. Palmer and Browne!! in their laboratory at the Louisville Manual Training High School. Among the other gentlemen interested in these experiments was Prof. Mark, who, with Prof. Brownell, had been retained in consultation upon elec trical matters by the Harris Company in July. 1896; and in this way they were introduced to the electrical and electrolytical treatment of the Ohio River water. The object of these devices, and of their operation, was to indicate the relative merits of the application of electricity to electrodes composed of metallic iron and metallic alu minum, respectively, in the preliminary treat ment of the Ohio River water, in order to se cure coagulation in a system of purification, of which filtration through sand is the last step. A general idea of these two devices, which were intended to be duplicates, with the ex ception of the metal used for the electrodes, may be obtained from the following outline: Tap water (which was equivalent to Ohio River water from which a portion of the sus pended matter had been removed by passage through the Crescent Hill Reservoir and some of the distributing pipes) was led through a meter to two glass cylinders, each about 4 inches in diameter, and i.^ to J.o feet in height. On the top of each cylinder there was fastened a suitable cover, in which were per forations for the glass inlet and outlet pipes, respectively. The glass cylinders of the two devices contained, respectively, metallic iron and metallic aluminum electrodes, which were fairly similar in shape and size. In each case the electrodes were made of sheets fastened together, about 0.5 inch apart, with suitable insulating materials. The electric current was supplied by a dynamo generator. The water was admitted to the cylinders at the top, and was drawn off at the bottom. A second perfor ation in each of the covers contained a glass tube, to which a rubber tube was attached. Each rubber tube was controlled by a clamp, so that from time to time the gas, which ac cumulated at the top of the cylinders, could be removed. From the glass cylinders of the two devices, the water passed to t he respective fil ters. Each filter was made by putting a layer of fine sand in a metal cylinder, about 6 inches in diameter, and 2 to 3 feet high. It was said that each filter contained about i foot of fine sand, beneath which were several inches of coarse sand, and a metal strainer, which was placed above the outlet at the bottom of the cylinders. All the water from the electro lytic cells (glass cylinders) passed on to the filters, but in each case a considerable portion of the treated water overflowed, owing to the relatively small size of the filters. Including the " Palmer and Brownell Water Purifier," as explained beyond. 3 02 WATER PURIFICATION AT LOUISVILLE. The experiments were begun before the ar rival of the party from the Water Company. The writer understood that the electrodes had been weighed, and that, in the case of the iron device, 65 cubic feet of water had been treated after weighing, but before our arrival. It was said that the operation of the aluminum device, which was in service when we arrived, began about 1.30 P.M. This device was op erated until about 4.10 P.M. Before this time chemical and bacterial samples for analyses were collected of the tap water before treat ment, of the electrolytically treated water as it overflowed from the top of the filter, and of the effluent as it passed from the outlet at the bottom of the filter. The time required to collect a gallon of this effluent was 30 min utes. It was said that the quantity of water passed through this electrolytic cell from 1.30 P.M. to 4.10 P.M., was 30 cubic feet. The electric current varied a little while we were present, but the amperage and voltage aver aged 3.6 and 32, respectively. From 4.20 P.M. to 4.50 P.M., about 7 cubic feet of water were passed through the iron device. The amperage and voltage averaged 5.4 and 27, respectively. Samples of water for analysis, corresponding to those taken from the aluminum device, were collected, with the exception of the tap water. The flow of water through the filter of this device was much faster than it was in the case of the other filter. This filter was said to be less satisfactory in its construction than the other one, but it was apparently much less clogged at the surface. It required 7 minutes to fill a gallon bottle with this effluent. In the case of each electrolytic cell, the for mation of gas was only slightly in excess of the amount which was absorbed by the flow ing stream of water, and only a very small quantity accumulated at the top of the cell. The samples were taken to the laboratory of the Water Company and analyzed. The re sults of the bacterial analyses are given in the following table, in which reference by serial numbers is made to the results of the chemi cal analyses, as stated in the second table. RESULTS OF BACTERIAL ANALYSES. Source of Samp le Number of Chemical Sample.* Number of Haclerial Sample. Bacteria per Cubic Centimeter. 826 Electrolytically ( num) treated before filtratio The same, but ilumi- watcr, after 827 828 4969 6 500 216 Electrolytically treated water, (iron) before The same, but after 830 *See lie following table. RESULTS OF CHEMICAL ANALYSES. (Parts per Million.) jg Nitrogen Residue on Fixed Residue . (j s as .a Evaporation. after Ignition. i u 1 Ammonia. i i < u Date. Corresponding 5 ~ 1 S rt w ^ i: u o D n OJ V i8q6. Bacterial E So fe 2 1 .1 V y. y - a 1 3 | 1 | 1 c i/) H O U o H x a fe 5 H Q f~ C5 < Q 8-6 Dec 2 4968 ,,| "8 827 4 6 O88 J 828 R-0 " 2 4970 I (. . . 2.5 .104 .OOO .104 .080 .014 0.84.9 137 o 137 103 o 103 70.3 O. I 830 " 2 4972 2 2.6 .112 .000 . 112 .076 .016 0.84.9 145 145 no O no 69.3 o 0.9 The above data indicate that the amounts of electric power used in the iron cell and the aluminum cell were at the rate of 197 and 75 E.H.P. per million gallons per 24 hours, respectively. From the loss in weight of the iron and aluminum electrodes, it was stated that the consumption of these two metals was at the rate of 86 and 60.7 pounds per million gallons, respectively. It is estimated from the above data, that the amount of electric current applied to the water in its passage through the aluminum and iron devices, was equal to 0.042 and 0.070 ampere-hour per gallon, respectively. THE MARK AND BROWNELL ELECTRO LY TIC A L DEVICES. 33 The analytical results show that with the aluminum device, the water was well purified, but in the case of t he iron device, due in part, apparently, to the construction of the filter, the effluent contained some iron, and a high number of bacteria. The greater part, if not all, of this iron apparently came from the silt in the tap water, and not from the electrodes. Taking the results of these experiments in general terms, it may be stated that they were on too small a scale, and of too short dura tion, to convey any practical specific informa tion on the purification of the Ohio River water, other than that iron, a comparatively cheap metal, may be electrolytically decom posed in a manner similar to aluminum with the formation of a gelatinous hydrate, capable of coagulating the mud, silt, and clay in water. PLANS FOR TESTING THE TREATMENT OF THE OHIO RIVER WATER ELECTRO LYTICALLY WITH THE USE OF IRON ELECTRODES. Arrangements were completed on Jan. i, 1897, whereby the Water Company should te.st the practicability of an experimental sys tem of water purification, the electrical appli ances of which were to be designed by, and constructed under the supervision of, Profs. Mark and Brownell. This experimental sys tem included a set of appliances which would enable 250,000 gallons of river water to be treated electrolytically with iron electrodes, in order to secure coagulation, preparatory to filtration. As a matter of convenience, it was arranged with the O. H. Je\vell Filter Company to make use of their test filter, which at that time was the only one remain ing at the pumping station. Comparing this process with those systems of purification which were tested during the preceding year, it will be noted that with the exception of the magneto-electric system, the general principles of the methods of pro cedure were substantially the same, in that they consisted of the following: 1. Treatment of the river water with a co agulating chemical. 2. Partial clarification of, and removal of suspended matter from, the treated water by subsidence in a settling chamber. 3. Rapid filtration of the coagulated and partially subsided water through a sand layer. The most marked feature of difference in the Mark and Brownell devices, from the or iginal Jewell System, was the kind and method of formation of the coagulating chemical. The electricity by itself does prac tically nothing in the purification of the river water. Its action is almost wholly, if not com pletely, an indirect one, in preparing a coagu lating chemical. This was clearly demon strated in the devices operated in July, 1896, by the Harris Company, and investigated further by the Water Company in August, 1896, as stated in the two preceding chapters. The preparation of hydrate of aluminum by the electrolytic decomposition of metallic alu minum, was found to possess an advantage as a coagulant when compared with hydrate of alumina prepared from the decomposition of sulphate of alumina by the lime which is naturally present in the river water, in that there are added to the purified water no chemical properties to corrode iron vessels, or to incrust steam boilers. Hydrate of iron possesses gelatinous and coagulating proper ties somewhat similar to those of the hydrate of alumina. [Metallic iron is much cheaper than metallic aluminum, and the indications were that similar advantages in the coagula tion of the Ohio River water might be ob tained electrolytically in this manner, at a reduced cost. To ascertain the efficiency of this process, and to obtain data indicating the cost of installation and of operation of such a system for the purification of the water sup ply of this city, was the object in testing the devices of Profs. Mark and Brownell. Ar rangements were made whereby Profs. Mark and Brownell were engaged by the Water Company to design and superintend the con struction of the necessary electrical devices. These devices in connection with the Jewell filter, were to be operated and tested by the Water Company under the direction of the Chief Chemist and Bacteriologist, but Profs. Mark and Brownell were to inspect daily the devices designed by them, and make such recommendations as seemed advisable. The time from Jan. i to Feb. 12 was oc cupied in constructing electrodes, electro lytic cells, and a dynamo generator, especially 34 WATER PURIFICATION AT LOUISVILLE. adapted to this class of work; in making suit able piping and electrical connections; and in installing an engine to operate the generator. During this period, considerable attention was given to the practical significance of sev eral features of the process, so far as the avail able appliances permitted. These special in vestigations made by the Water Company during this unavoidable delay in the regular work, are recorded in Chapter XV r . CONSTRUCTION OF THE MARK AND BROWNELL DEVICES. The devices consisted of an engine, dynamo generator, two electrolytic cells and two sets of electrodes, in addition to the necessary pip ing, cables and wiring, to give them the required water, steam and electrical connec tions. The electrolytic cells were in duplicate in order to test two forms of electrodes. One set was made of wrought-iron plates, designed by Prof. Brownell, and the other designed by Prof. .Mark, was made of cast-iron pipes placed within each other. Each cell with its set of electrodes was tested separately, in con nection with the Jewell filter. It may be added here that these electrolytic cells, and the electrodes which were made ac cording to the plans of Prof. Brownell, were found, after the work had been begun, to repre sent " The Palmer and Brownell Water Puri fier Patent applied for." So far as is known, an application for a patent for the devices de signed by Prof. Mark was not made. The tests of the Water Company at this time, therefore, included, but extended beyond, that of the " Palmer and Brownell Water Purifier/ which was mentioned in the intro duction to this report. For this reason the more comprehensive expression of " Mark and Brownell Devices " is used as the title of this chapter. The principal details of the construction of these devices were as follows: Engine. The engine was a new, horizontal, fly-wheel machine of the Atlas make. Its principal dimensions were as follows: Di ameter of steam cylinder, 11 inches; length of stroke, 14 inches; diameter of fly wheel, 4 feet; and diameter of driving wheel, 4 feet. The speed was regulated by a fly-wheel cen trifugal governor. It was connected to the dynamo generator by a lo-inch leather belt. Dynamo-generator. The dynamo-genera tor, driven by the Atlas engine, was made by James Clark, Jr., & Co., of Louisville. It was a direct, four-pole, compound machine, rated at 50 volts and 400 amperes at a speed of 800 revolutions per minute. The current was regulated with the aid of a field rheostat, and could be controlled at practically any desired amperage within the range of the machine. Electrolytic Cells. The two duplicate cells were made of wrought-iron plates, 0.25 inch thick. The main body of the cells was cylin drical in form, 2.5 feet, in diameter, and 6.0 feet high. At the top, the cells were capped by a dome-shaped cover, and the bottom of each was in the shape of a cone. The shell was riveted together with o.5-inch rivets. A flange was riveted to the upper edge of the plate, to form a connection with the cover, which was bolted to it. The cover was a dome-shaped iron casting, i.o inch thick, 30 inches in diameter, and 12 inches high on the inside. On the bottom of the cover was a suitable flange, to allow the cover to be bolted to the shell. In the center of the cover at the top, there was a shoulder suitably tapped for connection with the inlet water pipe. A 0.5- inch pipe with a valve was tapped into the cover, to enable the operator to blow off the gas which accumulated at that point when the device was in operation. A sight gauge was connected to the cover, also, to allow the quantity of accumulated gas to be noted. The conical bottom of the cell was of wrought iron, 27 inches in diameter and 12 inches deep on the inside. The upper portion of the cone was capped with a short cylin drical shoulder, by which the conical bot tom was riveted to the main shell. At the apex of the cone (the bottom of the cell) there was a flange for connection with the 3-inch blow-off pipe. The 4-inch outlet pipe was connected with the shell by means of a flange joint at a point (> inches above the bottom of the cylindrical portion of the cell, and extended into the cell about 8 inches. The inner surface of the cell con taining the Brownell electrodes was covered with a heavy coat of asphaltum paint. The inner surface of the other cell acted as part THE MARK AND BROW NELL ELECTROLYT1CAL DEVJCES. 3S of the negative electrode, and was not in sulated. Brov. iicll Electrodes. In one of the electro lytic cells the manifold of wrought-iron plates, designed by Prof. Brownell, was placed. The other cell contained the Mark electrodes, which are described in the following sec tion: The Brownell electrodes were made of 28 wrought-iron plates, each 0.25 inch thick, and 50.375 inches long. The manifold was ar ranged in two sections of 9 plates each, and one section of 10 plates. Of these sections, the middle one was made of 10 plates of prac tically the same width, 24 inches. But the width of the plates in each of the two outer sections ranged from 22 to 10 inches, in order to keep the distance between the inner per iphery of the cell and the edge of the plates approximately uniform. The plates of each section were fastened together by six 0.75- inch bolts, which were covered with hard rub ber tubes to secure insulation. On these hard rubber tubes were placed hard rubber wash ers, 0.5 inch thick, to serve as insulating dis tance pieces in keeping the plates separated from each other. All three sections of the manifold were supported on a frame of oak, made of four 2- by 4-inch pieces. This frame itself rested on an angle iron at the bottom of the cell, and had suitable appliances at tached to it to aid in lifting the electrodes out of the cell. The space between the sec tions of the manifold was 2 inches. The total area of one side of these plates was 27,400 square inches, and the cross- section of the electrolyte (the water between the plates) was 22,400 square inches. On Feb. 26, 1897,^11 annular wooden frame was put into the cell just above the top of the electrodes, to fill the space between the elec trodes and the wall of the cell, and the sup porting frame at the bottom was also filled in to close the corresponding space there. Mark Electrodes. These consisted of ten pieces of cast-iron water-pipes, and of the wall of the electrolytic cell in which they were placed. They were 42 inches long, and i, 3, 6, 9, 12, 15, 18, 21, 24, and 27 inches in diameter, respectively. At the bottom of the cell was a frame of 2- by 4-inch oak pieces, resting on an angle iron, and the electrodes were placed on this frame. The electrodes were not fas tened together, but were separated from each other by insulating distance pieces of vulcan ized fiber. The distance between the several pieces of pipe ranged from 0.5 to 1.5 inches, and averaged about 0.93 inch. With regard to the thickness of the elec trodes, there was also considerable variation, but the average thickness was about 0.5 inch. The total superficial area of one side of these pipes was 20,200 square inches, and the cross-section of the electrolyte was 17,900 square inches. Electrical Connections. From the genera tor, the current was led to a switch-board, where an automatic circuit breaker and a suitable ammeter and voltmeter for measur ing the current and potential were provided. The potential of the dynamo was regulated by means of a rheostat on the shunt of the mag net circuit. Cables conveyed the current to the two cells. At no time were both cells used at the same time, so the one set of main connections served for the two. In the case of each cell there were two openings near the top. There were placed in these openings wooden plugs, through which iron bars with binding posts at each end were driven. These bars formed part of the electrical circuit, and the wooden plugs were relied upon for in sulating them from the cell. To the outer binding posts the cables connecting with the switch-board were attached, and to the inner binding posts were fastened the cables con necting the anodes (positive plates) and cathodes (negative plates), respectively. In the case of the Brownell electrodes, the plates were connected with the cable by means of a brass lug bolted to each plate. The corners of the alternate plates on each upper side were cut off, to give ample space, and the two cables were passed through and soldered into the holes in the brass lugs, which were bolted on to the uncut corners of the plates. In this way every other plate was made positive or negative, as the case might be. To arrest electrolytic action on the brass lugs and the copper cable, they were coated with insulating paint. In the case of the Mark electrodes, the two cables attached to the inner binding posts of 306 WATER PURIFICATION AT LOUISVILLE. the cell, were connected with alternate pipes by means of brass lugs, in a manner quite similar to that used on the Brownell elec trodes. The outer wall of the cell was one of the plates forming the cathode (negative pole). Piping. Each of these electrolytic cells was provided with 4-inch piping and fittings, so that the river water entered the cells at the top, and left them at the side near the bot tom, and passed to the settling basin below the Jewell filter, which was operated as usual, except that the application of sulphate of alu mina was omitted. At the conical bottom of these cells, there was a 3-inch blow-off pipe leading to the sewer. All of these pipes were provided with valves and meters wherever necessary, and the only remaining points worthy of mention are the devices at the inlets of the cells, to facilitate the distribution of the river water between the electrodes. In the cell containing the Brown- ell electrodes, there was attached to the inlet pipe a cast-iron dome, 24 inches in diameter, and 8 inches high. It was screwed on to a nipple which was connected to a flange on the cover of the cell. The plate which formed the base of the dome was slotted to cor respond with the water spaces between the electrodes, and in setting the electrodes in place, care was taken to have the water spaces and the slots in the dome come opposite each other. The base of the dome was 3.5 inches above the top of the electrodes. A similar dome was attached to the inlet pipe in the cover to the other cell, containing the Mark electrodes. In this case, however, the base of the dome, which was 12 inches above the top of the electrodes, was perforated with forty o. 5-inch holes, arranged uniformly over the plate. Operation of the Mark and Brownell Devices Brownell Electrodes. These devices and appurtenances were assembled together ready for operation on Feb. 12, 1897. On the afternoon of Feb. 13, and the whole day of Feb. 15, the devices were in preliminary operation, and the Jewell filter was washed five times with filtered water to put it in normal condition, following a long- period of rest. Official operation began on Feb. 1 6. The rate of electrolytic treatment, and of filtration, was to be 250,000 gallons per 24 hours, or 23.2 cubic feet per minute. A brief general account of the operations will now be given, followed by a summary of the principal data, including the results of an alyses of the river water before and after treatment. The river water was fairly muddy at this time; while the filtered water was not of sat isfactory appearance or quality, and showed marked signs of insufficient coagulation to secure satisfactory results under these condi tions. The average voltage and amperage on Feb. 1 6, were 37 and 450, respectively; and on Feb. 17, these figures were 32 and 395. On the evening of Feb. 17, it was necessary to stop operations, in order to make some re pairs and modifications on the dynamo gen erator. The dynamo generator was not ready for operation again until Feb. 22, and the tests were suspended during the intervening time. Mud, iron hydrate formed from the electro lytic decomposition of the electrodes, and a limited amount of other suspende d matter, such as silt, clay, and bacteria, accumulated in the conical bottom of the cell beneath the outlet pipe during operation, and, at Prof. Brownell s request, about 2 cubic feet of this liquid and solid material were blown off into the sewer once an hour. The accumulation of gas at the top of the cell was also blown off hourly, to prevent the electrodes from be coming uncovered and consequently out of active service. Operations were resumed on Feb. 22, at 12.38 P.M., following the repairs of the dy namo. The filter was washed, and the water and mud removed from the settling chamber of the Jewell System. The rate of treatment and of filtration was 23.2 cubic feet per min ute, as provided for in the plans and designs. But the filtered water was very turbid and un satisfactory in appearance, and at 1.55 P.M., operations were stopped, and the filter washed again. The devices were then started at the rate of 15 cubic feet per minute. As the fil tered water did not improve materially in ap pearance, the rate was reduced to about 12 cubic feet at 4.00 P.M., and at 5.00 P.M. it was THE MARK AND BROW NELL ELECTROLYTICAL DEVICES. 37 further reduced to about 10 cubic feet per minute. During the afternoon, the current was approximately constant, and the voltage and amperage averaged 27 and 402, respect ively. At no time was the filtered water free from a distinct muddy appearance, although the removal of coarse suspended particles caused it to look much better than the un treated river water. On the morning of Feb. 23, the devices were operated for a short period at the rate of 13 cubic feet per minute, but the electro lytic treatment, under the existing conditions, was insufficient to give satisfactory results. The river water, however, was exceedingly muddy at this time, which was the early por tion of a period of a very heavy flood in the Ohio River. During the balance of this day, and during the following day, Feb. 24, the devices were operated at the rate of 4.9 and 9.8 cubic feet per minute, respectively. On both days the quantity of electric current was as great as the appliances would allow (400 amperes), but the filtered water was decidedly muddy. i As it was clearly evident that these devices were unable to purify the Ohio River water when it was in the existing condition, and further tests without modifications would be a waste of time and money, the writer re ported the condition of affairs to President Long, and requested instructions in the prem ises, as follows: (Copy.) FEB. 25, 1897. Mr. Chas. R. Long, President Louisville Water Company, Louisville, Kentucky. DEAR SIR: Since Monday noon of this week, we have operated daily the electrical devices designed by Profs. Mark and Brown- ell, with the view to purifying the present, very muddy, water of the Ohio River, at rates ranging from 23.2 (contract) to 5 cubic feet per minute. At no time have we obtained an effluent after nitration which could be proper ly called purified, or which could be compared with our earlier results during this series of tests. So far as my knowledge goes, whatever suggestions that may have been made by Profs. Mark and Brownell, have been fol lowed out in the operation (A these devices, but thus far I have received no formal notice as to their wishes in this matter. In view of the fact that the results obtained from these devices last week were not satis factory with regard to the quality of the fil tered water, and that the results of this week appear to be of no practical value to this Com pany, either with regard to capacity to handle this water, or with regard to its satisfactory purification, I have stopped the operation of these devices. I hereby notify you officially of the present conditions, and respectfully re quest instructions in the premises. Very respectfully, [Signed] GEORGE W. FULLER, Chief Chemist and Bacteriologist. On Feb. 25, the makers of the dynamo made some changes in the machine, as recom mended by Prof. Brownell. Unofficial tests showed that on this, and several following days, there was a marked reduction in the maximum amount of current which the dy namo could put through the electrolytic cell. On preceding days, 500 amperes could be ob tained at a potential of about 40 volts. It was now found that conditions had changed so, that with 55 volts, the maximum potential of the machine, only about 375 amperes of current could be obtained. The devices were carefully examined by Profs. Mark and Brownell. This marked reduction of 45 per cent, in the maximum amount of current which could be passed through the electro lytic cell, when the dynamo was operated at its maximum potential, means practically a similar increase in the cost of electrolytic treatment under the conditions stated, and is a factor which is given consideration beyond in the discussion of these results. At the request of Prof. Brownell, there was placed on Feb. 26 an annular wooden frame both at the top and the bottom of the elec trolytic cell, in order to reduce the space through which the river water could pass without treatment. Profs. Mark and Brownell examined the devices on Feb. 27, and requested that an un official run be made with the maximum elec tric current, and a rate of flow of water of IO cubic feet per minute. Their request was 3 o8 WATER PURIFICATION AT LOUISVILLE. complied with. But in spite of the fact that the river water contained only about one-fifth as much mud as was the case earlier in the week, it was not possible, even at this low rate of treatment, to obtain a filtered water which was not muddy in . appearance. The Brownell electrodes were not used again, ex cept in some special comparative experiments, as requested by President Long. The results of analyses of the water before and after treat ment, with summaries of the principal data and discussions thereon, are presented be yond. The devices were not operated again until March 5. when the second electrolytic cell, containing the Mark electrodes, was used. The rate of treatment, and of filtration, was 10 cubic feet per minute on March 5, and on March 6 the rate was 5 cubic feet. On both days the dynamo was run at the maximum output which could be safely handled. The filtered water, however, was unsatisfactory in appearance, and the greater water space be tween these electrodes made it impossible, with the available appliances, to secure as much coagulation of the river water as in the case of the Brownell cell, on account of the decrease in current, due to the increased re sistance which the current met in its passage through the Mark cell. As i rofs. Mark and Brownell formally an nounced at this time that they had no further modifications or suggestions to make, these devices in their present form were not regu larly operated again. Sl MMARY AND DlSCl SSlON OF THE RESULTS ACCOMPLISHED BY THE MARK AND BROWNELL DEVICES. There were nine official runs made with the devices during the period from Feb. 16 to March 6. Seven of these runs were with the Brownell electrodes, and the last two were with the Mark electrodes. In the case of the latter, very little information was obtained, so the discussion will be confined to the re sults obtained from operations with the Brownell electrodes. To obtain more light upon the work accomplished by these devices, President Long requested, first, that a com parison be made of the results accomplished by the electrolytical devices with the Brownell electrodes in connection with the Jewell filter and those obtained with the same river water by this filter without any treat ment whatever; and, second, that a set of aluminum electrodes be made to duplb cate the iron electrodes designed by Prof. Brownell, and then compare the results obtained with the respective electrodes, in connection with the Jewell filter, in purifying the same river water. Owing to unavoidable delays in securing aluminum sheets, these comparative tests were not completed until April 4. The results obtained from them are recorded just beyond this discussion, and are followed by the detailed results of analyses of samples connected with the tests of these de vices. The following set of tables contain a sum mary of the principal data obtained with reference to the rate of treatment and filtra tion, the amount of electrolytic treatment em ployed, the amount of electric power used, and the degree of purification after the treated river water had passed through the Jewell fil ter, and also the results of bacterial and chem ical analyses of the individual samples of the water before and after treatment. BACTERIA PER CUBIC CENTIMETER IN THE OHIO RIVER WATER TREATED BY THE MARK AND BROWNELL DEVICES AND THE JEWELL FILTER. Number of Sample. Date. 1897. Hour. Bacleria per Cubic 4973 Feb. 16 9.30 A.M. 22 60O 4975 if> I.3O P.M. 25 TOO 4979 16 4-30 " 34000 497ga 17 9.30 A.M. 22 IOO 4982 17 12.30 P.M. 24 900 4985 17 5-OO " 24 200 4986 22 9.30 A.M. 20 500 4989 22 2.35 r.M. 60 400 4992 22 5.OO " 59500 4993 23 9.30 A.M. 63400 4996 23 12.30 P.M. 59700 4999 23 5.OO " 56 500 5000 24 9.30 A.M. 41 ooo 5003 24 12.30 P.M. 1 6 700 5006 24 4-3 " 36500 5039 Mir. 5 9.30 A.M. 37900 5041 5 4.30 P.M. 28 700 . 5056:1 6 11.30 A.M. 31 500 5058 6 2.OO P.M. 4J ooo 5061 6 5.00 " 47 ooo THE MARK AND BROWN ELL ELECTROLYTICAL DEVICES. 39 SUMMARY OF RESULTS ACCOMPLISHED BY THE MARK AND BROWNELL DEVICES. - et Began. Electric Current. Quantities of Water. Cubic Feet. Average Rates of Filtration. \ Date. 897- Hour. II. I- pc. Million Ampere Hours Gallons per per Gallon. Filtered. Wash. ! Million Cubic Feet Gallons per per Minute. , Acre per X 1 24 Hours. Brownell Electrodes. I Feb. 16 11.44 A.M. 93 0.044 5383 820 22.6 92 3 h ,6m. 2 " 16 4.05 P.M. 73 O.O4I 6344 595 22.5 91 4h 42m. 3 " 17 1.22 " 6 5 O.O40 6273 9-13 23.0 93 4 h 33m. 4 " 22 2.22 " 9 7 0.064 2574 542 14.0 57 31 O2m. 5 " 2 3 9.52 A.M. 121 O.070 626 381 13.0 52 oh 48m. 6 " 23 11.05 " 28 3 0.181 I 902 455 4-9 20 6h 2511). 7 " 24 9-52 " 175 0.091 4 735 674 9.8 40 Sh 05m. Mark Electrodes. \f h - 4.2O P.M. IQ A o ooo 711 IO. I 41 ih T "km 2 " 6 1.27 " iy~t 384 \j . vjv/y 0.197 I 08 1 5-i 2O 3 h 33m. Estimated Aver Nitrogen a i Albuminoid Ammonia. Oxygen Consumed. Average Bacteria sl age Suspended DeRree of Pa r:s per Million. Parts per Million. per Cubic Centimeter ; Average 12 3 7. Solids in Ri.er Clearness Water. Harts of Effluent, per Million. River Water. *""" Reeved . wile?. Affluent. ,? River Water. Effluent. Bacterial Efficiency Brownell Electrodes. I 254 3 .326 . 116 64 6.4 2.0 69 23 800 5400 77-3 3 580 86 7 3 308 3 .380 .114 7 6.8 2.0 71 22 300 3 78o 83.0 1 474 4 -542 .182 66 9- 1 3-9 57 60000 8 ooo 86.7 4372 4 3-084 .110 96 50.3 2.3 95 63 400 4400 93.1 6 4372 4 3.084 .110 96 50-3 . =-4 95 58 100 4 200 92.8 1 3604 4 2.360 : .084 96 48.9 2.5 : 95 31 4CO 4 260 86.5 Mark Electrodes. i 205 5 .762 .260 66 12.0 4-2 65 28 700 7300 74-5 2 53" 5 4.900 . i-r 96 65 i) 3-0 ): 39 Soo 22 100 44-6 BACTERIA PER CUBIC CENTIMETER IN THE OHIO RIVER WATER AFTER TREATMENT BY THE MARK AND BROWNELL DEVICES AND THE JEWELL FILTER. Rate of 41 5! Collected. Filtration. ! | n Number s. JSL r> Period of ServiceSince Last 2-jL u 8 2; Run. S.J 6 3 X Washing. Hours and |lu *i Remarks. Date. Hour. u 1 c i.I "a Minutes. u V- .3 V = g 1897. s 3S. f a 3 -J(J 5u X u s J n Brownell Electrodes. 4974 Feb. 16 12.30 P.M. I 23-5 95 4-7 oh. 46m. I 241 7 200 4976 1 6 1.30 " I 23-5 95 5-2 Ih. 44m. 2 541 5 200 4977 l f > 2.30 " I 23.5 95 2h. 44111. 3 881 3 800 49781 16 4.30 " 2 23-5 95 3-5 oh. 2501. 588 4 400 4980 17 IO 2O A.M. 2 23-5 95 8.2 2h. 25m. 3 244 3 090 4981 17 1 2 . 30 " 2 23-5 95 8-4 4h. 35111. 6 068 3 250 4983 I/ 2.30 P.M. 3 23-5 95 4-4 ih. o8m. i 584 4 120 4984 " 17 5.00 " 3 23.0 93 7-2 3h. 3601. 4964 3160 4987 " 22 i-3 " 3 23.0 93 7-7 4h. oSm. 5738 4050 " 22 2-35 4 20.0 Si 2.8 oh. 13111. 285 > Km 499 " 22 3.30 " 4 16.0 65 2.7 ih. oSm. I 175 10 2(KI 499 i " 22 5.00 " 4 12.0 48 2.0 2h. 35m. 3 385 5 830 4994 " 23 10-30 A.M. 5 IO.O 41 oh. 38111. 380 4400 4995 " 23 12.30 P.M. 6 5-o 20 0.4 ih. 25m. 424 4 700 4997 " 23 2.30 " 6 5.0 20 0.4 3h. 2501. 980 4 030 4998 " 23 5.00 " 6 5-0 2O 0-5 5h. 55m. i 295 3 890 5001 " 24 IO.3O A.M. 7 IO.O 41 0-7 ih. O5m. 645 3 050 5002 " 24 I2.3O P.M. 7 9-5 40 1-3 3h. 05111. I 815 6000 5004 " 24 2-30 " 7 IO.O 41 i. g 5h. osm. 3 005 4 400 5005 11 24 4.30 " 7 IO.O 41 2.8 7h. 05m. 41551 3500 Mark Electrodes. 5056 Mar. 5 5.30 P.M. I IO.O 41 1.6 ih. rom. 7U 7300 5<>57 6 2.00 " 2 5.0 20 0.5 oh. 33m. 170 12 7OO 5059 " 6 3-30 " 2 5.0 20 0-7 2h. 03111. 630 29 500 5060 " 6 5.00 " 2 5.0 20 0.9 3h. 33m. 1093 24OOO 3io W A TEK PURIFICATION A T LOUISVILLE. RESULTS OF CHEMICAL ANALYSES OF THE OHIO RIVER WATER BEFORE AND AFTER TREATMENT BY THE MARK AND BROWNELL DEVICES AND THE JEWELL FILTER. (Parts per Million.) 1 y. Collected. Date. ,897. ga H C ^ I O o Nitrogen K aporat 1 Fixed Residue after Ignition. ft - o c as Albumin H s = l z a T: -, 1 3 * 831 833 835 :i 837 S39 5 848" 850 832^ 834 9 836,; 836/1 836" 838^" 838" 840<; 18 840* "> 840" 8511," 85 1/-" 85 if" 85 1" Fel> if) " 7 22 23 24 Mar. -. 6 Feb. 16 " 17 22 22 22 22 23 23 23 24 24 24 24 Mar. 6 6 6 6 5-0 6.0 7.0 8.0 S 5 10. O IO.O 5-0 6.0 7.0 .3 3 5 23 .18 .23 23 .24 6.4 6.S 9.1 50.3 48.9 12. 6 5 .0 2.0 .326 .380 542 3.084 2.360 .762 4.900 .116 .114 .204 .262 .402 3.004 2.2S6 .654 4-754 Lfflue .122 .118 .140 .080 .074 .11,8 .146 nt w Rive .062 .048 .028 .098 .060 .039 .090 th B - Water. .024 i.o .028 0.7 .015 1.5 .012 1.2 .015 1.4 005 i.S .007 1.4 rownell E .012 0.9 .Ol6 0.7 4-4 4-5 4-4 4-1 4.4 3-1 3-3 .led 4.6 376 428 620 4 5oi 3 732 892 5431 rode "4 292 1 86 284 158 182 134 152 132 184 163 163 des. 309 273 218 257 254 308 474 4372 3604 761 5 3ii s. 22 2O 46 29 28 31 2O 332 388 5I 4223 3537 820 5 165 89 240 300 447 4 128 3445 726 5076 92 88 1 14 95 92 94 89 49.0 50.5 6l. 5 62.5 56.0 64.3 62.2 49-5 62.0 62.0 54-0 62.8 o o o o o o o o o o o o o o o o o o 16.0 26.0 60.0 340.0 340.0 34-o 116.0 0.7 0.8 24.0 21. 7.6 19.0 56.0 64.0 45-0 58.0 2-5 5-5 5-0 4-5 13.0 9.6 4.6 9.2 .058 186 146 40 138 29 53 5 23 4 56 35 35 189 153 98 137 146 1 1 . I 4 6 I 4 6 t29 i 2 129 129 128 128 128 128 1 2O I 21 130 1 2O 91 94 56 56 44 22 43 01 27 97 47 28 27 72 37 88 20 I So 140 42 130 27 1 32 5 55 36 35 183 148 99 131 114 114 114 114 .95 95 95 95 9 2 92 92 92 89 89 89 89 S.o C r "I ^ .182 .042 .140 ."34 .OlS 1-5 4-3 4 9.0 " A .22 2.^ .1 IO .030 .080 .070 .014 1.2 4.0 2 S i 4 10.0 5 .22 2.5 .084 .010 Efflu .074 .062 ent with .016 Mark 1-3 Ele 4.6 ctro .... 5 23 3.0 .198 .052 .146 .094 .008 i . i 3-0 "4994,4995,4996. 5001. " 5002. "> 5004. " 4980, 4 l^rx Vf 5 8 %. *V 5 9 o%. " va . ^49B8. 49,. o, 4991. 14 4994 15 4995- " 499 6 Decomposition of the Browncll Iron Electrodes Loss of Metal. During the period covered by the official tests of the Brownell electrodes the total quantity of water treated was 29 965 cubic feet; the duration of treatment for this total quantity of water was 33.25 hours, and the average number of amperes of current was 401. The difference in the weights of these elec trodes, at the beginning and at the end of these tests, showed a loss of 17.5 pounds. This indicates an average decomposition of 0.59 gram (9.02 grains) of metallic iron per am pere hour. Carefully conducted experiments made on a small scale with bright iron, free from all rust, showed that the rate of decom position was rather variable, but, on an aver age, it approached the theoretical rate of 1.05 grams (16.17 grains) per ampere hour. These small experiments also showed, however, that the rate of decomposition diminished as the electrodes continued in service. The reason of this appeared to be associated with coat ings of rust on the surface of the electrodes, especially on the anode (positive pole). It will be noted that in these official tests the actual average rate of decomposition of iron was only 56 per cent, of the theoretical rate. This question of the rate of the electrolytical decomposition of metallic iron is one of great practical importance in this connection, be cause it is a controlling factor in determining the size of power plant which would be re quired in the application of this process. Further investigations were made along this line after the close of the work on the Mark and Brownell devices, and the results are re corded in Chapter XV. THE MARK AND BROWNELL ELECTROLYT1CAL DEVICES. Loss of Electrolytically Formed Hydrate of Iron in the Broii iicll Cell due to the Arrange ment of its Outlet Pipe. From the foregoing description of the elec trolytic cell, it will be recalled that the outlet water-pipe was not placed at the bottom of the cell, but at the side, about 6 inches above the apex of the conical bottom. The portion of the cell beneath the outlet pipe had a ca pacity of about 2.8 cubic feet, equal to 21 gal lons. At Prof. Brownell s recommendation the liquid and solid materials at the bottom of the cell were blown into the sewer about once an hour. The quantity removed each time was about 2 cubic feet on an average. It was very soon noticed that the liquid, which was removed in this way from time to time, subsided very quickly, and evidently contained a large amount of iron hydrate. On Feb. 23, when the river water was very muddy and contained 4372 parts per million of suspended matter, a test was made to learn the amount of iron hydrate which passed to the sewer through the blow-off pipe. Samples of the semi-liquid matter, which was blown off into a cask before passing to the sewer, were carefully collected after thor ough mixing, and corresponding samples of the untreated river water were also taken. The experiment was continued during the afternoon of this day. The results of the analyses of the several samples of river water and blow-off water, respectively, indicated that the iron coming from the electrodes and leaving the cell in this manner, was equiva lent to 0.52 gram (7.94 grains) of metallic iron decomposed per ampere-hour during this test. It is impossible to state accurately what the rate of decomposition of iron was during the afternoon of Feb. 23, but it is probable that it was not far from the average rate for all the runs, as stated in the last section. This would mean that 84 per cent, of the iron which came from the electrodes and left the cell, passed to the sewer with a portion of the heaviest mud, before there was an opportun ity for it to coagulate the finer particles in the water, and prepare it for efficient filtra tion. Owing to the very muddy condition of the river water on this date, it is probable that the above experiment exaggerates the prac tical significance of this point, with reference to ordinary river water. Nevertheless, it throws much light on the actual conditions at that time, and illustrates a weakness in the arrangement of the cell. Passages in the Cell containing the Brownell Electrodes, tliroitgh which the River IVater could pass with little or no direct Elcctro- l\tical Treatment. In the electrolytic cell containing the Brownell electrodes, the regular spaces be tween the electrodes for the passage of the water to be treated were 0.5 inch. There were also two spaces 2 inches in width, be tween the sections in which the plates were arranged; and in addition to this, there was an annular space, ranging from 3 to 3.5 inches in width, between the edge of the electrodes and the wall of the cell. The quantity of elec tric current which can pass through a j-inch water space is only one-fourth as much as can pass through a half-inch water space, other conditions being equal; and there was practically no current passing through the an nular space between the cell and the electrodes. The percentages of the total sectional area of the cell which was occupied by the electrodes and the water spaces, respectively, together with the percentages which the treatment of the water passing through each of the sec tions was of the maximum, are as follows: These figures show that, originally, there was an opportunity for fully one-half of the water to pass through the cell without direct electrolytic treatment. On Feb. 26, an annular wooden frame, 3 inches wide, was placed in the cell at both the top and the bottom of the electrodes, and which reduced the outer annular water space, in which there was no treatment, from 31 to about 6 per cent, of the total sectional area of the cell. This corrected in part the defect of having some of the water pass through the cell without coming at once in intimate con tact with the coagulant during its initial for- 312 WATER PURIFICATION AT LOUISVILLE. mation. All the official rims with this cell, however, were made before the frames were put in place. But on the following day the cell was used in its modified form, on an un official run, and the devices failed to purify the water satisfactorily. With the conditions which would be ob tained upon a large scale of operation, it is very probable that this defect in the arrange ment of the cell would reduce the efficiency of the treatment, owing to the lack of uniform distribution of the solid hydrate in the water. In this experimental cell, however, it is doubt ful whether the point in question exerted any influence, because the piping was so arranged that after leaving the cell, the water passed through several elbows, valves, and meters, all of which aided in mixing well together the treated and untreated portions of the water. Variations in the Resistance of Ohio River Water to the Passage of an Electric Current. As already stated in the description on the operation of these devices, it was found that up to, and including, Feb. 24, it was possible to pass, for a short time, 500 amperes of cur rent at a potential of 40 volts through the cell containing the Brownell electrodes. But on several days following this date the maxi mum potential (55 volts) of the dynamo could give only about 375 amperes. On ac count of the fact that it is the amperage of the current which controls the amount of electrolytic decomposition, this reduction of about 45 per cent, in the efficiency of the plant, with substantially the same expenditure of power, is a very serious problem from a practical point of view. Considerable study was given to the matter, and it was found that it was caused during this flood by the dilution of the dissolved chemical compounds in the river water. So far as could be learned, the mud and other suspended matters in the river water exerted practically no influence on its conductivity. That is to say, with small elec trodes placed in two-gallon jars, it apparently made no difference in the resistance of the same river water, whether the suspended mat ters were present or whether they were re moved by the passage of the water through filter paper or a Pasteur filter. The results of further studies along this line are recorded and discussed in Chapter XV. In the following table are presented com parative results, showing the specific re sistance of the Ohio River water, expressed in the conventional form, during the period of flood from Feb. 22 to March 6. They serve to show the variations which were en countered in the river water; and they also indicate the range in power which would be required to do the same amount of efficient work, other thing s being equal. As will be seen by the comparison of these results with small electrodes (150 centimeters square) with the results referred to in Chapter XV, and which were obtained from the use of large electrodes, these figures can be safely used in estimating the required amount of electro motive force for different services in which the Ohio River water is used as an electro lyte. SPECIFIC RESISTANCE OF THE OHIO RIVER WATER, EXPRESSED IN OHMS PER CENTI METER CUBE. Number of Correspond ing Chemical Sample.* Date. 1897. Specific Resistance. 835 February 22 7 300 837 23 5 9 839 24 8 700 841 25 II 600 842 26 13 200 843 27 16 700 844 March I 12 300 845 2 14900 846 3 11 250 847 4 14 ooo 848 5 7900 850 6 7 200 * The chemical composition of these samples may be seen by reference to Chapter I. In the foregoing summary of results, the electric horse-power used during each run is given. The relation between the electric power and the actual steam power used, both with regard to these conditions and the prac tical conditions on a large scale, are discussed in Chapter XV. As a more accurate idea of the cost of generating electric power, it may be stated here, that these generating appli ances, under the most favorable conditions, yielded about 80 per cent, of the power con- THE MARK AND BROW NELL ELECTROLYTICAL DEVICES. tained in the steam which was used. On a large scale it may be reasonably expected that this efficiency could be maintained or slightly increased. General Status of this 1 rocess at the Close of these Tests. A . Broivndl Electrodes. The summary of analytical results obtained in connection with the Brownell electrodes, already presented, shows that at no time was the filtered water satisfactory, either with re gard to appearance or to the number of bac teria contained in it. The amount of organic matter in the filtered water, as indicated by the nitrogen in the form of albuminoid am monia, and by the oxygen consumed, was several times as great as was normally present in the filtered water during the previous year. Owing to the very muddy condition of the river water, however, and the consequently large amount of organic matter which it con tained, the percentages of removal of organic matter were high. B. Mark Electrodes. The Mark electrodes, composed of circular cast-iron pipes, were placed in a duplicate cell, and were used in connection with the same generating appliances as in the case of the Brownell electrodes, for two runs, which were made on March 5 and 6, respectively. At this period the river water was very muddy, and on March 6 it was in the mud diest condition which existed during the en tire investigations. Very little information w r as obtained from the results of these two tests, other than that the filtered water was muddy, showing that the devices were wholly inadequate to coagulate the water properly, even at one-fifth of the regular rate of treat ment and of filtration. Considering these electrodes in general terms, however, they seemed to possess some advantages over the Brownell electrodes in arrangement for operation on a large scale, in that the water was uniformly treated. I>ut in these experimental cells, the outlet pipe leading to the Jewell settling chamber was so arranged that it is doubtful whether any such advantage existed here. It is also possible that the cost of construction of complete sets of electrodes on a large scale would be less in the case of the Mark electrodes. These latter electrodes possessed a disadvantage when compared with the former, in that the water space between the individual electrodes was twice as great, and, accordingly, the amount of power required would be twice as great, other things being equal. It seems hardly necessary to state that these poor results were caused by inadequate preparation of the river water before its pas sage through the Jewell filter; and, further, that in the absence of sufficient coagulation these results cannot be taken as a measure of the merits and practicability of the general method of water purification, in which the electrolytic treatment is a preliminary step preceding filtration. These results refer only to a particular set of devices, possessing a number of weaknesses, which might be remedied in a large measure by practical means. In this connection, it must not be over looked that, as already stated, the electrolytic treatment of water adds to it no sulphuric acid to combine with lime and form incrustations in steam boilers; nor does it liberate in the water carbonic acid gas to increase the cor roding action of the water on wrought-iron receptacles. This advantage of the electro lytic formation of coagulating chemicals (either aluminum hydrate or iron hydrate), over the decomposition of the commercial sulphates of these metals by the lime dissolved in the water, is a matter of importance. An other advantage of the electrolytic treatment is that it is independent of the amount of lime dissolved in the river water, and the possibil ity of undecomposed sulphates passing into the filtered water is obviated. A further con sideration of this process in its various phases will be found in Chapter XV. COMPARISON OF THE QUALITY OF TIIK OHIO RIVF.R WATER AFTER FILTRATION FOL LOWING ELECTROLYTIC TREATMENT IN THE BROWNELL CELL, AND AFTER FiL- TUATlOX WITHOUT ANY PRELIMINARY TREATMENT. As requested by President Long, two runs were made with the Jewell filter on March i i . The first run was made at the regular rate of WATER PURIFICATION AT LOUISVILLE. 23.5 cubic feet per minute, and the river water before filtration was treated with the maxi mum electric current (400 amperes) in the Brownel! electrolytic cell. The second run was made at the same rate, but the river water received no preliminary coagulating treat ment whatever. In each case the settling basin of the Jewell System was drained and cleaned, and the filter thoroughly washed be fore filtration was begun. Both runs were continued until the filter became clogged so that it would not allow the passage of water at the prescribed rate. A comparison of the result obtained from these two runs is shown in the following table: BACTERIA PER CUBIC CENTIMETER IN THE OHIO RIVER WATER TREATED DURING THE COMPARATIVE TESTS DESCRIBED ABOVE. * Sample" Date. Hou, Bacteria per Cubic 1897 5081 March 11 9.30 A.M. 38 IOO 508.) " ii 1. 00 P.M. 41 500 5087 " ii 3-25 " 36 2OO COMPARATIVE SUMMARY OF RESULTS ACCOMPLISHED BY THE JEWELL FILTER WITH (RUN NO. i) AND WITHOUT (RUN NO. 2) PRELIMINARY TREATMENT BY THE MARK AND BROWNELL DEVICES. (Brownell Electrodes.) Bej an. Electric Current. Quantities Cubic of Water. Feet. Average Filtr Rates of Mum her of Run. H.P. per Million Ampere Million Service. Hours and 1897. Gallons per 24 Hours. Gallon. per Minute. A Hour" * 4 March n 85 80 Estimated Average Suspended Solids in Oelrness! Nitro^as Albumino ts per Mill dAonta. % gen Consul ts per Mill ned. Average E Cubic C acteria per River Water. Parts per Million. Effluent. River Water. Effluent. Per Cent. Removed. River Water. Effluent. Per Cent. Removed. River Water. Effluent. I 751 I 751 Muddy. 2.400 2.400 -374 .566 34 76 24.6 24.6 6-4 9.0 74 63 39800 36 20O 13 IOO 2O 2OO 67.0- 44.0 NUMBER OF BACTERIA IN THE EFFLUENT OF THE JEWELL FILTER WITH AND WITHOUT ELECTROLYTIC TREATMENT. Rate of ! 8* 8 ! Collected. Filtration. jj 15 i Number S. i& ,,- Period of ServiceSince t c . u u . s of Run. C . 5| 1 Washing. ^1l Is Remarks. X Date. j- - "^ o ^ - Hours and T3 ^ o ,5-S 1897. Hour. u c o U X Minutes. ^ i; c c Is 38.? ^ JU 0(J Ji o s J CO Effluent with Electrolytic Treatment. 5082 March n 12.50 I .M. I 23-5 95 IO.O oh. 55m. i 287 12 7OO 5083 " ii I.OO " I 23-5 95 10.8 ih. osm. i 5i7 13500 Effluent with No Coagulating Treatment. 5085 March n 3.15 I .M. 2 23.0 93 5.6 ih. 02m. 1386 17700 5086 ii 3-25 " 2 23.0 93 7-9 ih. I2tn. i 636 19 700 5088 " ii 3-55 " 2 23.0 93 II. O ill. 42111. 2 197 23 2OO THE MARK AND BROWN ELL ELECTROLYTICAL DEVICES. 3 5 RESULTS OF CHEMICAL ANALYSES OF THE OHIO RIVER WATER BEFORE AND AFTER FILTRATION THROUGH THE JEWELL FILTER WITH AND WITHOUT ELECTROLYTIC TREATMENT. (Parts per Million.) ; a Nitrogen Residue 90 Fixed Residue . U :: B 3 as Rvaporat " alter Ignition. i a P g Ammonia. i d u 6 < Corresponding r! M o -. B ^ 1 d V D _ E .- * Date. Bacterial Num. a a tj s u < .- - j j c JJ . ** .2 .897. bers or Hour of B -, >, 2 i = , -| i! z 5 a , 2 1 ! 0! , c y, S- Q . o f- 3 a Q U H 1 a h 3 la < Q ~ ssi 1 Mar. ii 9.30 A.M. II.O .22 24.6 2.400 2.274 .126 .042 .002 : . I 5-1 1881 1751 I )0 1766 1664 102 61.0 O 96.0 8s6- j " 1 1 5082, 5083 II. O 5 6.4 374 .248 .126 036 .003 1 i 5-0 587 547 I3! 537 435 102 60.7 O 21.0 57" " ii 5085, 5086 II. 5 9.0 5.66 .440 .126 .034 .005 1-3 5 ^ 763 633 130 704 602 IC2 03.1 O 28.0 A COMPARISON OF THE EFFICIENCY IN THE ELECTROLYTIC TREATMENT OF WATER BEFORE FILTRATION, OF THE BROWNELL ELECTRODES AND OF ALUMINUM ELEC TRODES OF THE SAME SlZE AND AR RANGEMENT. At President Long s request, a set of alu minum electrodes was made to duplicate as nearly as possible the iron electrodes devised by Prof. Brownell. Owing to delays in secur ing the aluminum plates, it was not until April 2 that these comparative tests could be made. These aluminum electrodes were placed in the cell which formerly contained the Mark electrodes. From April 2, at 3.41 P.M., to April 4, at 6.30 A.M., nine runs were made in connection with the Jewell Sys tem. The first five were made with the iron (Brownell) electrodes, and in the last four runs the aluminum electrodes were used. The rate of treatment and filtration was kept as nearly as possible to the regular rate of 23.5 cubic feet per minute, and the conditions of operation, other than the amount of current used, were as nearly the same as possible. In the case of each cell, no sediment was blown off at the bottom. It was found that a current of 100 amperes on the aluminum electrodes was sufficient to secure a perfectly clear effluent, while with 450 amperes of current on the iron electrodes the effluent was not clear. The iron elec trodes, however, were somewhat covered with rust from their earlier use, while the alu minum electrodes were new and bright. It was found that the iron electrodes, following the long period of disuse, gave more efficient coagulation as they continued in service. \Yhile these experiments were continued suf ficiently to serve their general purpose, yet they were of too short duration to allow satis factory determinations of the amount of metal used, owing to complications from mud, rust and water. The principal data of these comparative tests are presented in the follow ing summary and results of analyses: BACTERIA I>ER CUHIC CENTIMETER IN THE OHIO RIVER WATER TREATED DURING THE COMPARATIVE TESTS DESCRIBED ABOVE. Serial Number. Date. 1897. Hour. Bacteria per Cubic Centimett 5226 April 2 5.30 P.M. 3 700 5228 2 8.00 " 3420 5230 3 12.30 A.M. 2940 5232 3 3-30 " 3500 5234 3 9.30 " 3 150 5243 3 3.30 P.M. 4 200 5249 3 9.30 " 6 400 5252 4 12.30 A.M. 5 100 5259 4 5.30 " 5 7oo 3 i6 WATER PURIFICATION AT LOUISVILLE. COMPARATIVE SUMMARY OF RESULTS ACCOMPLISHED BY THE JEWELL FILTER AND MARK AND BROWNELL DEVICES WITH IRON ELECTRODES (RUNS NOS. 1 TO 5) AND ALUMINUM ELECTRODES (RUNS NOS. 6 TO 9). Began. Electric Current. Quantities of Water. ~ Cubic Feet. Average Rates of Filtration. Periods of of Run. H P per Cubic Feet Million Gallons Hours and Date. 1897. Hour. Million Gallons per 24 Hours. per Gallon. Filtered. Wash. per Minute. per Acre per I April 2 4.28 P.M. 95 0.047 5 214 518 23-5 95 3h. 42m. 2 " > 8.33 " 78 0.042 7265 603 23.6 95 5h. o8m. 3 " 3 2.07 A.M. 65 0.041 5902 602 21. 8 88 4h. 3im. 4 " 3 7.07 35 0.029 4970 568 23.2 94 3h. 34m. 6 " 3 3.36 P.M. 24 0.024 4092 514 23.2 94 2h. 55m. 7 " 3 6.57 " 20 O.O2O 4552 541 22.5 91 3h. 2301. 9 " 4 2.44 A.M. 6 O.OO8 4974 21 .2 86 3h. 54m. . Estimated Average Sus- Degree of Nitrogen as Albuminoid Ammonia. Parts per Million. Oxygen Consumed. Parts per Million. Average Bacteria per Cubic Centimeter. Average Bat z in River Water- Parts per Million. of Effluent. River Water. Effluent. Per Cent. Removed. River Water. Effluent. Per Cent. Removed. River Water. Effluent. Efficiency. i 213 3 .248 075 71 4-5 .2 73 3300 3" 90.4 2 213 .248 075 71 4-5 .1 76 2 900 133 93-3 3 213 3 .248 .075 7 4-5 .0 78 3500 154 95.6 4 213 3 .248 .075 71 4-5 . I 76 3 Io 240 92-3 5 200 3 .248 ."75 71 4-5 .0 78 3700 207 94-4 6 200 i .208 .050 76 4.8 o.S 83 5300 94 98.2 7 2OO i .208 .050 76 4.8 o.S 83 5700 IOO 98.2 8 2OO i .208 .050 7 6 4.8 0.8 83 5700 126 97.8 9 2OO i .208 050 76 4.8 0.9 Si 5400 138 97.6 RESULTS OF CHEMICAL ANALYSES OF THE OHIO RIVER WATER BEFORE AND AFTER TREAT MENT BY THE MARK AND BROWNELL DEVICES AND THE JEWELL FILTER WITH IRON ELECTRODES AND ALUMINUM ELECTRODES, RESPECTIVELY. (Parts per Million.) Collected. \ i Nitrogen Residue on Evaporation. Fixed Residue after Ignition. a i J3 g S u c e 8 ..; <- 3 Corresponding o : . C , , , -.; 91 -_: V S Date. t i aj U < .ti - . s 8 Jfc 5 73 1897. Hour of Collection. IQ si - ;. 2 Ji i"3 . ~ a ; 3 t. 1 | | <J! H Q u o H a P b. U - Q - /; 5 < a River Water. 875 Apr. 2, 3 5.30 P.M., 8.00 P.M. 12. 30 A.M., 9.30A.M. 10.5 .16 4-5 .248 .166 .082 .O2O .OO2 1.2 4-5 "2? 213 112 229 2O9 9 39-2 o 8.8 878 " 3,4- 3.30 P.M., 9.30 P.M. 12. 30 A.M., 5. 30 A.M. II. .20 4.8 .208 .138 .070 .014 .OOI 0.8 2.8 3 200 IO9 2Si|i93 88 43.8 o 12.8 Effluent with Iron Electr odes i. rr> TOI 3 Tfi .) T3 93 o J 96 o 876 877 " 2, 3 " 3 5227,5229,5231, 5233 5235, 5241 II .0 3 3 .32 .20 .2 .O .076 .074 .014 .016 .001 .000 I.O 0.9 3-o 3- 12 08 93 92 . . . 42.0 42.0 o o I .0 0.5 Effluent with Aluminum Electrodes. 8 79 a 87 9 b 8790 8791! 879 Apr. 3 " 3 " 4 " 4 " 3, 4 5=45 5247 5253 5258 5245,5247,5253, 5258 11.4 06 90 o ... o 8 " 1 7 ... no 85 (1 06 84 .IO 0.8 .050 .000 .050 .016 .002 i-3 2.9 03 o l<>3 83 8344.1 o.o THE MARK AND BROWN ELL ELECTROLYTICAL DEVICES. NUMBER OF BACTERIA IN THE EFFLUENT OF THE JEWELL FILTER FOLLOWING TREAT MENT IN THE MARK AND BROWNELL DEVICES WITH IRON ELECTRODES AND ALUMINUM ELECTRODES, RESPECTIVELY. Rate of S j Collected. Filtration. c J 3 Number S. 1*. i Period of Service Since Last aj c . u ^ S 3 Date. 1897. Hour. Run. k = O u 3 3 Washing. Hours and Minutes. |f| rt a Remarks. u fsi ~ D. cT i *JCJ rt ^ en (J s 2 K Effluent with Iron Electrodes. 5227 April 2 5.30 P.M. I 24.0 97 3-5 ih. O2m. 1499 385 522O 8.00 " I 24. o 97 3h. 32m. 250 5231 " 3 12.30 A.M. 2 24.0 97 7-3 3h. 57m. 5 555 193 5233 " 3 3.30 " 3 24.0 97 3-9 ih. 23m. i 710 154 5235 3 9.30 " 4 23-5 95 5-3 2h. 23m. 3378 267 5236 3 10.00 " 4 23-5 95 7-5 2h. 53m. 4088 239 5237 " 3 10.30 " 4 22.0 89 9-7 3h. 23m. 4768 214 5238 " 3 12. OO M. 5 23-5 95 3-5 oh. 48m. i 148 219 5239 " 3 12.30 P.M. 5 23-5 95 3-9 ih. iSm. i 868 198 5240 " 3 i. oo " 5 23.5 95 4-3 ih. 48m. 2538 181 5241 "3 1.30 " 5 23-5 95 5.3 2h. iSm. 3 208 222 5242 " 3 2.00 " 5 20.0 Si 8.6 2h. 4Sm. 3928 217 Effluent with Aluminum Electrodes. 5244 April 3 4.0O P.M. 6 24.0 97 3-1 oh. 24m. 588 107 5245 " 3 4.30 " 6 23.0 93 3-4 oh. 54m. i 238 85 5246 " 3 5.30 " 6 23.5 95 5-7 ih. 54m. 2 7lS 91 5247 " 3 8.00 " 7 23-5 95 3-8 ih. 03111. I 594 106 5248 " 3 9.00 " 7 23-5 95 5 7 2h. 03m. 2976 95 5250 " 3 10.00 " 7 23-5 95 10.9 3h. 03111. 4306 120 5251 3 11.30 " 8 23-5 95 3.0 oh. 45m. 1074 159 5253 " 4 12.30 A.M, 8 23.5 95 4.0 ih. 45m. 2474 no 5254 " 4 1.30 " 8 23.5 95 7.0 2h. 45m. 3834 I 1O 5255 " 4 2.0O " 8 23.0 93 9.8 3h. ism. 4544 139 5256 " 4 3-30 " 9 23-5 95 oh. 4601. 812 199 5257 " 4 4.30 " 9 23-5 95 3-9 ih. 46m. 2342 93 i 5258 " 4 5.30 " 9 23-5 95 6.5 2h. 46111. 3692 121 WATER PURIFICATION AT LOUISVILLE. CHAPTER XIV. DESCRIPTION OF THE MACDOUGALL POLAR ITE SYSTEM OF PURIFICATION, AND A RECORD OF THE RESULTS ACCOMPLISHED THEREWITH. ON March 3, 1897, just as the tests of the Mark and Brownell devices were being com pleted, arrangements were made whereby trie efficiency and cost of operation of the Mac- Dougall Polarite System should be investi gated, with reference to the purification of the water supply of this Company. In order that the results accomplished by this system might be comparable with those of the fore going tests, the rate of treatment was ar ranged to be 250,000 gallons per 24 hours. In brief, this system, known abroad both as the International System and the Howatson System, was represented to consist of a double nitration of river water, without the use of coagulating chemicals, obtained either elec- trolytically or from commercial chemical products. The first filtration was to be through a layer of sand, with the view to re moving the coarser matters suspended in the river water; and the second filtration was to be through a layer of a special material, called polarite, which is described below. This sys tem of water purification has never been tried in this country, but it is said to be in success ful operation in purifying turbid or muddy waters in several places in England, Egypt, and India. In order to guard against delays, it was mutually agreed that for the first (sand) filter use should be made in an undisturbed condi tion of the Jewell filter, which was then at the disposal of the Water Company. Owing to the fact that it was necessary to send to Eng land for the polarite, this system, however, was not ready to be tested until May 10. With the exception of Sundays, and sev eral unavoidable periods of delay, this system was operated night and day from May 10 to June 12, inclusive. During the remaining time, from the close of the tests of the Mark and Brownell devices until the end of the in vestigations, attention was directed to some plans and devices of the Water Company, as are described in Chapter XV. When the MacDougall Polarite System was being constructed, it was found that a separate tank with baffle plates was to be sub stituted for the settling basin under the Jewell filter. On May 28, a " clay extractor," con sisting of an iron tank with two compart ments, each filled with about 14 feet in depth of coarse coke, was substituted for this tank containing baffle plates. The settling basin and filter of the Jewell System have been fully described in the foregoing chapters, and a plan and section have also been presented. A correct understanding of the other devices used in connection with the polarite system may be obtained from the following descrip tion. Meters and gauges were provided wherever necessary, in order to secure meas urements of the quantities of water which were treated, and also of the resistance which the water met as it passed through the several layers of filtering material. IRON TANK WITH BAFFLE PLATES. This tank, placed just outside of the Jewell house, was cylindrical in form, 3 feet in di ameter, 16 feet high, 113 cubic feet in capac ity and made of boiler iron, 0.19 inch in thickness. It was put together with o.5-inch rivets. The bottom of the tank was conical in form, 2 feet high and tapered at the lowest point to an apex 3 inches in diameter, where a blow-off pipe leading to the sewer was con nected. The inlet pipe, leading from the river water main, was 4 inches in diameter, and en- THE MACDOUGALL FOLAR1TE SYSTEM OF PURIFICATION. tercel the tank by a flanged joint about 5.5 inches above the point where the conical bot tom with a shoulder was riveted to the main cylinder. In this main cylinder were six baffle plates, placed at 0.25, 2.25, 4.25, 6.25, 8.25 and 10.25 feet, respectively, from the center of the inlet pipe. These baffle plates were circular in form, and riveted to the shell of the tank. In order to provide a passage for the water, a small segment of each plate was cut away, leaving a maximum perpendicular distance between the edge of the plates and the adjoining shell, of about 4 inches. The lowest plate was cut away on the side diam etrically opposite the point where the inlet pipe entered; and the openings through the remaining baffle plates were arranged alter nately with reference to the inlet pipe and the opening in the first plate. Perpendicular to the upper baffle plate, which was 3 feet from the top of the tank, two iron plates were riveted to the baffle plate and to the shell of the tank. Each of these partition plates ex tended to within 12 inches of the top of the tank, and they were each 22 inches in length. In the center of the tank, and parallel to these two partitions, which were in the same verti cal plane as the cut edge of the lower baffle plate, was another partition, 2.83 feet in height, extending from the top of the tank to within 14 inches of the upper baffle plate. Across the bottom of this central partition, and extending to the two outer partitions, was a false bottom, made of a screen with meshes of about 0.25 linear inch. In the inner compartment, formed by the two outer ver tical partitions and the shell of the tank, the upper baffle plate and the screen, was placed excelsior, which rested on the baffle plate. The superficial area of the excelsior was 7.85 square feet. It was removed on May IT. When the water, in its upward flow through the tank, reached the last baffle plate, it passed through the normal opening and t he space above it, bounded by the shell of the tank and one of the outer partitions; and thence it flowed over this partition into the central compartment. As described above, this compartment was divided into halves by a central partition, which extended to the ex celsior compartment at the bottom. From one side to the other, the water passed by flowing through the excelsior and underneath the central partition. Thence rising in the other half of the central compartment, the water overflowed the outer partition, into a compartment between this partition and the shell of the tank. The outlet pipe, 6 inches in diameter, connected with the tank by a flange joint at the center of this compartment, and 24 inches below the top of the tank. From the tank, the outlet extended directly into the open compartment above the sand of the Jewell filter, ending in an elbow and a 6-inch nipple turned down. No arrangement was provided to break the stream of water as it entered the filter. The lower end of the nipple forming the outlet was on a level with the staves of the outer tank, and 35 inches above the sand. It was 12 inches from the edge of the outer tank to the center of the nipple. At the contract rate of flow, 250,000 gallons per 24 hours, the vertical velocity in the iron tank was 3.28 lineal feet per minute, or 1.2 per cent, of the velocity in a 4-inch pipe. This tank was used from May 10 to May 19, inclusive. On May n, the excelsior was removed from the inner compartment; but no further changes were made. The sys tem was not in operation from May 19 to 28. CLAY EXTRACTOR. This device consisted of a rectangular iron tank, 6 by 3 feet, and 16 feet high. It was made of o.ig-inch plates of boiler iron, riveted together, and was divided into two equal com partments by a central partition. At points 1.75 and i.o feet from the top and bottom, respectively, there were angle irons riveted to the wall of the tank, and upon which rested screens of about o/>25-inch mesh. The space between the screens in each compartment, 14.3 feet in height, was filled with crushed coke. The diameter of the pieces of coke ranged from 0.5 to 2 inches, and averaged more than i.o inch. A stay bolt, 0.5 inch in diameter, was passed through the tank from end to end, and through the central partition about 5 feet from the top. Midway in each of the compartments at the bottom of the tank, an iron plate, i foot high and 3 feet long, was placed beneath the above-men tioned screens, parallel to the central parti tion. These plates relieved the angle irons 3 20 WATER PURIFICATION AT LOUISVILLE. of some of the weight of the layer of coke above them. In the corners of each of the two compartments at the bottom of the tank, where the river water entered and the wash- water left it, curved iron plates were arranged, to guard against accumulations of mud and other deposits. At the center of each side of the two lower compartments, there were openings, to which 4-inch flanges were at tached. The two 4-inch pipes entering the tank at ;he front by means of these flanges were used to admit the river water to the tank; an 1 the two corresponding pipes at the rear conducted the wash-water to the sewer. At the center, on the front side of each of the two open compartments above the coke at the top of the tank, there were openings, to which the two 6-inch outlet pipes were con nected by flange joints. The two inlet river- water pipes, the outlet river-water pipes, and the two wash-water outlet pipes, were in each case branches of a single main pipe of the same diameter; and on the first and last pair of pipes gate valves were provided, with the view to using the two compartments either separately or together. In practice, however, both compartments were used together, as the central partition was unable to prevent the water from passing from one to the other. It was arranged that, when the coke should become clogged and require cleaning, the water could be drained out through the outlet leading to the sewer; and connections were provided so that the effluent of the Jewell fil ter could be pumped through a 2-inch pipe to the open compartment at the top of the tank, and then flow by gravity to the wash- water outlet pipe at the bottom. A 10- by 18- inch hand-hole, with its center 1.5 feet from the bottom and the side of the tank, was placed at each end of the tank, with the view to taking out some of the clogged coke at the bottom, if it should become too much clogged to be cleaned by the above-stated method of washing. The main inlet and outlet pipes were the same as were used with the iron tank con taining baffle plates. On the inlet there was a meter and a pressure gauge, to show the head required to force the water up through the clay extractor. This new device was com pleted May 28. POLARITE FILTER. The polarite filter consisted originally of a layer of polarite placed between layers of sand, with all the filtering material resting upon the underdrains in a large open wooden tank. This tank was rectangular in form, 23 feet long, 10.2 feet wide and 7.3 feet deep, as shown by inside measurements. It was made of 2-inch smooth pine planks, fastened to sup ports as follows: The floor was supported by eight pieces of timber, 10 by 3 inches, which extended about 5 feet beyond each side of the tank. These timbers were connected to gether at each end by planks. On each side of the tank there were eight upright pieces of timber of the same size as those at the bottom, and in each case a timber 6 by 3 inches ex tended from the end of the bottom supports to about midway on the upright supports. The floor and sides were laid first, and spiked to the supports mentioned above. The planks at the end were fitted into a shallow vertical groove, which was cut on the inside of the side planks, and the side planks were spiked to those on the end. Midway on each end, there was an upright oak timber, 6 by 6 inches, to which the planks were also spiked. Two stay- rods, 0.75 inch in diameter, passed length wise through the tank, and were fastened at the ends to these oak timbers. On the bottom of the tank, there were placed at right angles to each other, two layers of 3-inch tiles, with the long dimension horizontal, and the ends about 0.5 inch apart at the joints. The space between the tiles was filled with coarse gravel. In addition to the tiles at the bottom of the filter, there were laid on one side, both ends and across the center from end to end, rectangular troughs of wire, with o. 5-inch meshes. These troughs were 4 inches square in section, open at the bottom, and were intended to serve as an aid to the tiles, in conducting the filtered water to the end of the tank, where the outlet pipe was placed. The slope of the bottom of the tank toward the outlet, was about i inch in its length of 23 feet. At each of the four cor ners, on the inside of the tank, there was con structed a stand pipe, made of 3-inch hard tile, with cemented joints. These pipes ex tended above the level of the water which THE MACDOUGALL POLAR! TE SYSTEM OF PURIFICATION. .1 stood upon the sand when the filter was in operation, and were designed to act as air vents. On the top of the tile drains (and the wire troughs) the following layers of filtering material were placed, successively: 6 inches of coarse gravel; 3 inches of coarse sand: 6 inches of fine sand; 20 inches of polarite; and 6 inches of fine sand. The original depth of filtering material in the polarite filter, not counting the coarse sand and gravel at the bottom, was 32 inches, and the area of filter ing surface was 234.8 square feet, a little more than double that of the Jewell filter (115.8 square feet). The proposed area of the polarite filter was determined by Mr. Mac- Dougall, on the assumption that the polarite filter could be operated with satisfactory re sults at one-half of the rate of filtration em ployed in the case of the Jewell filter, or about 50 million gallons per acre per 24 hours. A rectangular wooden trough, 6 inches wide and 6.5 inches deep, extended completely around the inside of the tank, 4.5 feet from the bottom. The 6-inch inlet pipe discharged the effluent of the Jewell filter into this trough at the northwest corner of the tank. At the southwest corner of the tank, a 6-inch pipe led the wash-water from this trough to the sewer. The main outlet pipe was 6 inches in diameter, placed as near as practicable to the bottom of the tank, on the end near the northeast corner. The effluent of the polarite filter discharged into the wrought-iron reser voir, in order to store enough water for wash- water during operation. Arrangements were made whereby this effluent could be pumped under pressure through a 3-inch pipe which entered the polarite filter at the bottom, about midway on the north side. Valves, meters, and gauges were inserted wherever conven ience required. The sand in the polarite filter was taken from several sources. A part of it came from lots which remained at the pumping station from the preceding year, and a part was taken from the bed of the Ohio River. All of these materials were washed carefully in a wheelbarrow with filtered water from a hose. Analyses of the more important filtering ma terials are given beyond. Changes in the Polarite filter. A number of changes were made in the polarite filter as follows: 1. On May 11, it was found that the joints of the air-vents (stand pipes at the corners) leaked, and allowed some of the water to reach the bottom of the filter without passing through the filtering materials. The two upper sections of each vent were removed, and the lower sections were filled with filter ing materials, to correspond with the main filter layers. At the same time there were added above the filtering materials just mentioned, 5 inches of coarse coke and 5 inches of fine coke. This increased the depth of the filtering materials to 42 inches, and owing to the fact that the coke reached the sides of the distributing trough, which was placed around the tank on the inside, the surface area was reduced to about 200 square feet. 2. On May 14, there were placed on the top of the above-mentioned coke, 2 inches of fine sand. This increased the depth to 44 inches, but the surface area was unchanged. 3. On May 17, air vents were placed in the northwest and southeast corners of the filter, as follows: Three pieces of pipe, 1.5 inches in diameter, were put through the side of the tank at 12, 40, and 48 inches from the bottom, respectively; and by the aid of an el bow, each pipe was extended above the water level. The object of this was to take out ac cumulations of air at different levels within the filter. 4. During the period from May 19 to 28, inclusive, this system was out of service, and a number of changes were made. The princi pal one was the substitution of the clay ex tractor, for the small upright settling tank, as already described. The upper layer of sand and both layers of coke were removed. After the removal of these materials, the lower lay ers were washed by pumping filtered water into the bottom of the filter for about 3 hours. During this time about 3,600 cubic feet of wash-water were used. The surface of the 6-inch layer of fine sand, resting upon the polarite, was leveled, and upon it were placed 3 inches in depth of fine coke, which had been 3 22 WATER PURIFICATION AT LOUISVILLE. carefully washed with filtered water from a hose. Above this coke layer, 3 inches of fine sand, washed in the same manner as the coke, were placed. This left the surface of the filter below the bottom of the distributing trough, and consequently the area was restored to 234.8 square feet. The depth of the filtering material was 38 inches. The leading features concerning the sev eral filtering materials are as follows: Polaritc. Polarite is the trade name of a hard, black, porous and magnetic iron sub stance, which does not rust or dissolve in water. It is understood that it is prepared from a suitable natural ore by a patented pro cess, and by crushing and screening, any de sired size of grain may be obtained. By vir tue of its numerous and minute pores, in which atmospheric oxygen may be occluded, polarite is claimed to possess a powerful ac tion in the destruction of organic matter by oxidation. The chemical composition of the polarite used in these tests was as follows: PERCENTAGE COMPOSITION OF POLARITE. Silica (SiO.j) 22.65 Magnetic oxide of iron (Fe 3 O 4 ) 49. 17 Alumina (Al O I I lvith a trace f man g anese I2 ,(, 31 \ and phosphoric acid. f Lime (CaO) 0.85 Baryta (BaO) 0.02 Magnesia (MgO) 7.00 Undetermined, chiefly water and carbon 8.15 It was found that the nitrogen in the form of albuminoid ammonia in the polarite was 5.4 parts per million. The grains of polarite were very coarse for a filtering material, as shown by the following mechanical analysis: MECHANICAL ANALYSIS OF POLARITE. Effec giv 12. 5. 3- :nt. by we 100. 93-4 50.1 I5.I 1.6 0.4 1.78 Sand. There were three lots of sand used in the polarite filter; one lot was coarse, and the other two were fine sand. The coarse sand resting upon the coarse gravel at the bottom of the filter, and the fine sand at the surface, were taken from the river bed; while the fine sand forming the layer just beneath the porlarite was obtained at the pumping sta tion, where it had been left by some of the fil ter companies from the preceding year. All of this material, as already stated, was washed with filtered water from a hose in iron wheel barrows. The amount of organic matter re maining on the sand after washing is indi cated by the nitrogen in the form of albu minoid ammonia, as follows: Number Nitroeen as of Source. Location in Filter. Ammoim Sand. ; Harts per Million. 1 6 Riverbed. Laver above gravel. 9.2 17 Pumping station. Layer between gravel and polarite. 6.S 18 River bed. Layer at surface. 17.4 The efficient size of these filtering materials, as shown by mechanical analyses, was as fol lows: MECHANICAL ANALYSES OF SAND USED IN THE POLARITE FILTER. Finer than 3.90 millimet " " 2 . 04 " 0.93 " 0.46 " " 0.316 " 0.182 Effective si/.e (ten per cent, fi than given diameter in millime ters) 07 8 89.2 67 6 2 8 O.I 0.1 Coke. The coke was of the ordinary com mercial variety. The coarse coke was of the grade commonly called nut size, and it con tained when new 40.0 parts per million of nitrogen in the form of albuminoid ammonia. In the case of the fine coke when new the nitrogen in this form amounted to 49.8 parts. For this line of work the amount of lime and iron contained in coke is of some significance. The amounts of these constituents in the two grades were as follows: Fine Coke. 17.43 per cent. 0.95 " " 1.83 " " The size of the finer grade is indicated by the results of the following mechanical an alyses: THE MACDOUGALL POLARITE SYSTEM OE PURIFICATION. 323 MECHANICAL ANALYSES OF FINE COKE USED IN THE POLARITE FILTER. 100.0 91.3 83.2 67.6 -44-7 27.6 18.7 Finer tlian 12.0 millimeters. 5-89 3-9 2.04 0.93 0.46 0.316 O.IS2 0.105 3-3 Effective size (ten per cent, finer than given diameter in millimeters) 0.21 Operation of the Polarite System. The operation of the polarite system was divided into two periods, namely: From May 10 to 19, and from May 28 to June 12, 1897, inclusive. During the intervening time be tween the two periods changes were made in the filter, and the clay extractor was substi tuted for the small iron settling tank, as has been described. The general features of the operation were arranged by the Water Company to be as fol lows: The operation of this system was under the direction of Mr. MacDougall, or his represen tative. The required rate of treatment was 250,000 gallons per 24 hours, equivalent to 23.2 cubic feet per minute; the system was under operation as continuously as practic able from 6.00 A.M. on Monday until 6.00 P.M. on Saturday, during each week. It was also understood by the Water Company that no chemical coagulants would be used in connec tion with this system. A record of the most important points con nected with the operation of this system is as follows: Period No. i. The first period of operation extended from May 10 to May 19, when the system was shut down for alterations and repairs for a con siderable length of time. The river water was fairly uniform in character for the first four days, ranging from 171 to 260 parts per million of suspended solids. On May 14 and 15 a slight rise caused the suspended matter to increase to 1,260 parts per million. For the remaining three days the solids averaged 486 parts per million. The operation of the system began on May 10 at 9.10 A.M. At 3.45 A.M. on May 11 the system was stopped in order to remedy leaks in the air vents of the polarite filter, and to add to the filter 10 inches of coke, as already stated. Operations were resumed on May 12, at 9.04 A.M., but the rate of filtration was re duced to 12.0 cubic feet per minute, which was about one-half of the normal rate. This re duced rate was held until May 13, at 9.00 A.M., when it was increased to 18.0 cubic feet per minute (about three-quarters of the normal rate), which was held until May 14, at 9.44 A.M. Up to this time the efiluent was never clear in appearance as it left the polarite filter; although the water was clearer at this point than it was when it left the Jewell filter. The system was stopped on May 14, at 9.44 A.M., and 2 inches of fine sand were added to the surface of the polarite filter, after it had been washed for several hours by pump ing filtered water through from below, at the rate of 12 to 14 cubic feet per minute. Dur ing this time 2,083 cubic feet of wash-water were used. On May 14, at 7.46 P.M., the operation of the system was resumed; and from that time until May 15, at 9.00 A.M., a solution of sul phate of alumina was applied to the water as it left the Jewell filter on its way to the polar ite filter. During this period the rate of ap plication of sulphate varied. It ranged from 1 1.09 to i.oo grains per gallon, and averaged 4.39 grains. The rate of filtration was 13 cubic feet per minute from t he last resump tion of operation until May 15, at 6.00 P.M., when the system was stopped from Saturday night until Monday morning. During the greater portion of the night of May 14, when sulphate of alumina was applied to the water as it entered the polarite filter, the effluent of this filter was clear. At all other times it possessed a decided turbidity. Whenever the system was out of service it was repeatedly noted that many air bubbles, some of which were quite large, appeared on the surface of the water on the polarite filter, near the edge of the tank. On the morning of May 17, the water in this filter was allowed to drain out, and it was found that there was a scum of aluminum hydrate, clay, etc., about 3 2 4 WATER PURIFICATION AT LOUISVILLE. 0.25 inch thick deposited upon the surface of the filter. There was also seen a large num ber of holes around the edges of the filter, some of which ranged from 3 to 8 inches in width and depth. Air vents, as previously de scribed, were inserted at this time, and about three wheelbarrows of fine sand were added, to fill the holes at the sides. The thick scum cracked in places, and the entire surface of the polarite filter was raked. Filtered water was pumped into this filter from below, until the gauge showed that the water level was 32 inches above the bottom. The Jewell filter was put in operation at 13 cubic feet per min ute on May 17, 11.39 A.M.; and, above the point just stated, the polarite filter was filled from the top. This caused a good many bub bles to appear at the surface, especially along the edge of the filter. From the beginning of the operation of the system on this date, until 2.28 P.M., sulphate of alumina was applied to the water as it left the Jewell filter at an aver age rate of 1.06 grains per gallon. At 2.00 P.M. the rate of filtration was increased to 18 cubic feet per minute; at 4.00 P.M. it was in creased to 23.5 cubic feet; and at 4.45 P.M. it was decreased to 18 cubic feet. The system was out of operation from 5.00 to 5.21 P.M. on this date, in order to allow air to escape from the polarite filter. The surface of this filter was also raked, and the water above the sand drained off to a depth of i foot at this time. From May 17, at 5.21 P.M., until May 19, at 8.45 A.M. (the close of the first period of operation) this system was operated at a rate of about 18 cubic feet per minute, with out any special features of importance. The effluent of the polarite filter was at no time, during this portion of the period, free from a decided turbidity. Independent of repairs, and of the wash- water pump, this system required the atten tion of one regular attendant to control the rate at which the river water entered the set tling tank; to regulate the outlet valves of the Jewell and polarite filters, so that the depth of water upon them was approximately constant; to apply sulphate of alumina solu tions; and to agitate and wash the Jewell filter. During this period, from May 10 to 19, there were available by arrangement, 7.48 days, of 24 hours, in which to operate the sys tem. Of this time 5.73 days were devoted to regular operations, and the balance of 1.75 days (23 per cent, of the period) to re pairs and changes. Of the 5.73 days de voted to regular operations, 3.8 per cent, of the time was occupied in washing and agita ting the Jewell filter. The quantity of water passed through the Jewell filter during this period was 120,865 cubic feet, and the quantity of effluent of the polarite filter was practically the same. This made the actual rates of treatment of water by the two filters, expressed in different forms, as follows: AVERAGE ACTUAL RATES OF FILTRATION IN THE POLARITE SYSTEM. Jewell Pobrite Filter. Filter. Gallons per 24 hours 164 ooo 164 ooo Cubic feet per minute 15-2 15.2 Million gallons per acre per 24 hours 62 *3o.5-f36.o * Original area, f Modified area. Of the total quantity of water treated, 120,865 cubic feet, sulphate of alumina was applied to 11,919 cubic feet (9.9 per cent, of total quantity) at an average rate of 3.80 grains per gallon. This amount of sulphate of alumina was equivalent to 0.37 grain per gallon of the total quantity of water treated during the period. The accumulation of sediment, etc., on the surface of the Jewell filter was removed in a manner somewhat different than was pre viously the case, in that the filter was washed less frequently, and surface agitation was em ployed more frequently, with the modification that, at the close of agitation, the water above the top of the staves of the inside tank was al lowed to pass to the sewer. The average quantity of unfiltered water thus wasted at the time of each agitation was about 200 cubic feet. A record of the washes of the filter is as fol lows: WASHING OF THE JEWELL FILTER PERIOD No. 1. "o " Date of Washing. tt |l i[| .g! II Day. Hour *O U5 5 2 |||l i^ z a, a ?, I* May 10 5.2O P.M. 2om. 750 i953 2 " ii 4.30 " 2om. 739 11815 3 " 4 10.47 A.M. lorn. 529 40720 6 4 " 18 4.25 P.M. iSm. 644 41 938 14 5t 19 15439 4 Washed by Water Company to prevent stopping at night, in the absence of instructions to the operator from Mr. MacDougall. f Washed by Water Company, preparatory to use with Water Company s devices. THE MACDOUGALL POLARITR SYSTEM OF PURIFICATION. 325 As the operation of this system progressed, greater, as the agitations wtihout washing f TiMvpll filtpr \vn \\-nsliprl Ipse frpnufiitlv iiirrenspd the Jewell filter was washed less frequently, and agitated more often, as seen by the next table, in which the principal data are recorded. It will be noted that the loss of head following an agitation gradually became increased. The polarite filter did not become clogged except by the air which was trapped in its pores at times. The bulk of this air rose to the surface when the outlet was closed. AGITATIONS OF THE JEWELL FILTER PERIOD No. 1. Date of Agitation. Period of Quantity of Filtered Water. Cubic Feet. Initial Loss of Head in Feet following Agitation. Day. Hour. Agitation. Minutes. Since Last Washing. Since Last. Agitation. Agitations -Rate 23.5 Cubic Feet per Minute. 1897 2 3 3 13 9.05 " 1.14 r.M. 2f>m. 23m. 16674 20665 24 869 4755 3991 4.0 3-9 5 6 13 4 10.09 " 4.07 A.M. 24m. 23111. 39J25 45095 4256 5970 5-2 6.0 8 15 4.49 A M. 26m. 6864 3 Sio 3-i J ao P M 16 157 C > M 3 4 15 1 6 17 17 18 18 S.I 9 11.40 2.45 A.M. 2 1 m . 22111. 23899 27038 30015 2 770 . 3 137 2979 5-7 6-5 6.7 17 18 18 18 8.00 23m. 34 793 36 ->i6 1484 7-0 18 I 728 20 18 18 2.00 P.M. 2om. 39 9 4 970 7-i 2 6 22 23 18 19 11.39 3-34 A.M. 22m. 23m. 6 926 10691 3 299 3 705 2.6 2-7 Period No. 2. This period, following repairs and changes which have been enumerated, extended from May 28, 12.35 P.M., until June 12, 6.00 P.M., the close of the tests of this system. During this time the system was operated continuously, excepting from Saturday nights until the following Monday mornings, at ap proximately the full rate. The exact rate called for in the contract was 23.2 cubic feet per minute; but it was the general custom throughout all these tests to maintain the rate as nearly to 23.5 cubic feet as practicable. There were no delays, changes or repairs of importance during this period. The river water contained very little mud at the beginning of this period, but mostly very fine clay particles, averaging in amount about 90 parts per million of suspended mat ter for the first two days: and the effluent of the polarite filter was decidedly turbid. On May 31 and June i, sulphate of alu mina was applied to the water as it flowed to the Jewell filter, as follows: May 31, 7.51 to 8.45 A.M., 0.60 grain per gallon; May 31, 9.00 A.M. to 3.00 P.M., 0.24 grain; May 31, 5.00 to 9.00 P.M., 0.21 grain; May 31, 11.30 P.M., to June i, 3.05 P.M., 0.18 grain per gal lon. During those portions of the time be tween May 31, 7.51 A.M., and June 1,3.05 P.M., which are included above, there was no application of chemicals. From June i 3.05 P.M. to 5.20 P.M., sulphate of alumina at the rate of o. 18 grain per gallon was applied to the water as it left the Jewell filter on its way to the polarite filter. After this there was no further application of chemicals. In the absence of rains in the Ohio Valley, the river water rapidly became clearer, and during the remaining portion of this test the suspended solids averaged only about 30 parts per million. The effect of this is shown very 3 26 WATER PURIFICATION AT LOUISVILLE. clearly by the fact that from May 31 till June 8 the Jewell filter was neither washed nor agi tated during continuous operation at the nor mal rate. During this time, without any ad ditional application of chemicals (a iarg;e por tion of the above-stated applications remain ing in and upon the filters), the effluent was clear or slightly turbid. At no time, however, was it bright and perfectly free from visible suspended matter. In addition to the washing at the beginning of this period the Jewell filter was washed on May 31, 7.19 A.M., and on June n, 8.53 A.M., the quantities of effluent of the polarite filter used for wash-water being 543 and 644 cubic feet, respectively. The quantity of water filtered "between the last two washings was 315,199 cubic feet. The Jewell filter was agitated, and about 200 cubic feet of unlii- tered water drained from the surface of the sand, eight times during the last three days of the eleven days between washings, as fol lows: AGITATION OF JEWELL FILTER PERIOD No. 2. Quantity of Filtered Water. Date of Agitation. Cubic Feet. of Agi- Period of Agitation. Date. Since Last Since Last ,897. Washing. Agitation. 1 "26 : 9 2.00 " 31 m 260611 33751 27 | 10 6.35A.M. 3om 282 914 22 303 28 IO 4 . 51 P.M. 3 nil 296558 12644 29 10 9-30 " 2401 302 358 5 Soi 30 10 12. OO " 24111 395 204 2845 3i II 3.OOA.M. 25111 308 735 3 531 32 II 7.17 " 34 m 313 844 5 109 The data on loss of head before and after each agitation are not complete, but it may be stated that the latter agitations did very little in decreasing the frictional resistance of the sand layer, and that the filter was exceed ingly dirty on June 1 1, when it was absolutely necessary to wash the filter. During this period the polarite filter was not washed, raked, or scraped, and at the close of June 12, the loss of head was 21 inches, with the normal rate of filtration. It was also found that the clay extractor did not become clogged during service. Boih compartments were regularly used, although neither of them was washed. The acting head necessary to force the water through this de vice remained 1.9 feet during this period. From May 28 to June 12, there were avail able by arrangement 12.22 days of 24 hours, in which to operate this system. Of this time, 1 2.07 days were devoted to regular opera tions, and the balance of 0.15 day was de voted to unavoidable delays, largely in mak ing preparations to apply sulphate of alu mina. Of the 12.07 days devoted to regular operations, 1.6 per cent, of the time was oc cupied in washing and agitating the Jewell filter. The quantity of water passed through the Jewell filter during this period was 397,355 cubic feet; and the quantity of effluent of the polarite filter was substantially the same. About 399,000 cubic feet of water passed through the clay extractor, the increase here (1645 cubic feet, equal to 0.4 per cent, of total) being due to the water which was drained from the surface of the Jewell filter. This made the actual rates of treatment of water by the three devices, expressed in dif ferent forms, as follows: AVERAGE ACTUAL RATES OF FILTRATION IN THE POLARITE SYSTEM. Gallons per 24 hours. .. Cubic feet per minute. . Million gallons per acre per 24 hours 246 ooo 22.9 245 ooo 22. S 245 ooo 22.8 Of the total quantity of water treated, 397,355 cubic feet, sulphate of alumina was applied to 40,328 cubic feet (10 per cent, of total quantity) at an average rate of 0.25 grain per gallon. This amount of sulphate of alu mina was equivalent to 0.026 grain per gallon of the total quantity of water treated during this period. QUALITY OF THE OHIO RIVER WATER AFTER TREATMENT BY THE POLARITE SYSTEM. In the next two tables are presented sum maries of analytical results, showing the bac terial efficiency of the system, and also the percentage of removal of organic matter from the river water by it. Beyond these tables are the results of individual bacterial analyses of the river water, and of the water after it passed through the several devices compris- THE MACDOUGALL POLAR1TE SYSTEM OF PURIFICATION. ing the system; and also the results of chemi cal analyses of the water at different stages of treatment. For further information con cerning- the composition of the river water during this period, in addition to that shown by the summaries of results, reference is made to Chapter I. The summaries of analytical results are di vided into two periods, which are the same as have been used in the foregoing pages con cerning the operation of this system. It will be noted at once from the summaries that the river water was quite muddy during the first period: while during the second period it was very clear, comparatively speaking. During the two periods, the suspended solids in the river water averaged 004 and 40 parts per million, respectively. As was reported to you at the time of oc currence, the collection of samples for anal ysis, during the first day of operation of this system, was informally protested by Mr. Mac- Dougall. on the ground that this system was not yet ready for normal work. No official notice was received as to the time when this protest was withdrawn, if at all, and no further mention of the subject was made to the writer. Owing to the fact that so many changes were made during the first period, in an effort to adapt the system to the purification of the Ohio River water, it does not seem fair to give the results obtained during this portion of the operation serious consideration. Concerning the results obtained during the second period, following several changes, it will be seen that the system accomplished no substantial bacterial purification of the water until sulphate of alumina was applied to the water, as it flowed to the Jewell filter. It further appears that the aluminum hydrate, deposited upon and within the Jewell filter, and which was not removed to a marked de gree by the surface agitations of the sand, or by the single short washing which this filter received after the application of sulphate of alumina ceased, was clearly a more important factor in the purification of the river water, than was the polarite. The coke in the clay extractor contained lime, metallic iron and organic matter. The latter served as a food for the river water bac teria; and while there was no increase after the second day of operation in the albu minoid ammonia in the water as it passed through the extractor, the growths of bacteria continued until nearly the close of operations. It was especially noticeable that the growths were greatest for a short time following a period of rest, such as the first few hours of operation on Monday mornings. The lime and metallic iron in the coke were steadily removed by the carbonic acid contained in the river water. This explains the usual increase of about -2 parts in the alka linity of the water during passage through the clay extractor. The still greater increase in alkalinity of the water, during passage through the polarite filter, is similarly ex plained by the coke layers in this filter. It was found, from an average of three sets of analyses on June 4, that the river water was practically saturated with atmospheric oxygen at that time, and that the effluent of the polarite filter contained about 90 per cent, of that necessary for saturation at the actual temperature (21 C). Contrary to experience sometimes met with effluents of filters containing coke, the oxygen consumed results were apparently not affected, practical} speaking, by the iron and organic matter contained in the coke. In all cases, the nitrogen in the form of albu minoid ammonia, and the residue on evapora tion, were determined in the water as col lected, and also after it had been passed through a Pasteur filter. The difference in the corresponding results is recorded in each case as the amount in suspension. During the last few days of the tests, the effluents were so clear that the amount of suspended matter could not be satisfactorily weighed, al though it was generally visible in minute quantities. With regard to the nitrogen in the form of albuminoid ammonia, which was suspended in the effluents at this time, the an alyses indicate that small amounts were pres ent, although they approached so nearly to the limits of accuracy of the method of an alysis that full weight cannot be given to them. In conclusion, it may be stated that the evidence obtained during these tests, taken as a whole, shows that the MacDougall Polarite WAFER PURIFICATION AT LOUISVILLE. System (which was investigated with the view to dispensing with the use of coagulating chemicals) was not applicable, in the form in SUMMARY OF THE AVERAGE RESULTS BY DAYS OF BACTERIAL ANALYSES OF THE OHIO RIVER WATER BEFORE AXI) AFTER TREATMENT BY THE POLARITE SYSTEM. i J eriod No. 1 . Bacte.iap cr Cubic C , ptim ,. K , r Bacterial Efficiency. D; tc. 1897. River Polarile [ewell I nlarite Water. Effluent. Effluent. Kilter. Filter. May 10 15 300 7 500 14 900 51.0 2 . 6 " II 1 1 Soo 3 600 9500 69.5 19.6 " 12 31 ooo 9 700 30 ooo 68.7 3-2 " 13 36 200 16000 34 9"" 55-7 3-6 " M 52400 38 200 72 200 27.0 -38.7 " 5 49400 19 5oo 22 70O 60. 5 54.0 " 17 22 6OO 9 ooo 21 IOO 60 2 7-1 " 18 16 500 13900 14 900 15-7 9-7 " I9 12 100 17 ooo 10 300 - 45-o 14.9 which it was tested by this Company, to the efficient and economical purification of the Ohio River water. Period No. 2. Bacter , per Cu bic Centimeter. Bacterial Efficiency. Date. u- j j s 1897. ^ V c <u 5 u ! u -S II 15E is xS "35 5 .~ ^ "c *" W fe "o ^ X W A 0. u A a, May 28 7900 8 300 5 360 3 645 - 5 32 54 29 7 700 7 600 5 100 3 77 i 34 51 " 31 2 560 10 300 3 280 2 040 303 28 20 June i 2 2OO 2 8lO 2470 I 270 28 12 42 2 i 910 770 i 157 716 7 37 63 3 i 39" 530 388 2^0 10 72 82 4 97" 490 194 i So - 53 80 81 5 710 190 158 116 -67 78 84 7 770 1 60 222 192 - 50 71 75 S : 640 260 IlS 81 - 97 82 87 9 490 . 9 1 o 78 44 - 86 84 91 10 460 700 55 45 - 52 88 90 " n 470 547 48 36 - 16 90 92 12 590 590 40 36 93 94 SUMMARY OF THE RESULTS OF CHEMICAL ANALYSES OF THE OHIO RIVER WATER, BEFORE AND AFTER TREATMENT BY THE POLARITE SYSTEM, SHOWING THE REMOVAL OF OR GANIC MATTER, AS INDICATED BY THE NITROGEN AS ALBUMINOID AMMONIA, AND BY OXYGEN CONSUMED. (Results in Parts per Million.) Period No. 1 . . . , Removal Date. >gen 1897. Water. Effluent. Effluent. Effluent. Effluent. Water. Effluent. Effluent. Filter. Fiiter. May 10 )_ .282 .112 .108 60 62 4-5 2.0 1.8 56 60 " 12 .348 .! 7 6 .140 51 60 5-3 2-7 2.4 49 55 " 13 .406 .232 .160 43 60 6.0 3-5 1.4 42 77 14 .884 334 .no 62 8S 17.2 6.8 2.4 60 86 15 1.042 396 .116 62 89 I6. 3 6.0 2.0 63 88 " 17) ;; is 448 .260 .130 42 71 10.4 4.9 2.9 53 62 " 19) Period No. 2. Date. u . Uj ^ u ,897. v v 3 - U 4; - = u S = = - 5 V. V- S s -5 = . - >,- "B E 5 S 3 ^> E f E "HE ^c "c E s> *^E c E U Jj ww ^ W H -w P-W a w B-W w May 28 / " 29 i .146 .152 .108 .096 -4 26 34 3-5 3-2 2.7 2.5 9 23 29 31 r_ June i ) .138 .144 .092 .098 -4 35 29 2.8 2.6 2.3 1.8 7 18 36 2 | 3 i .166 .148 .106 .102 n 36 39 3-1 2.7 2.3 I.g 13 26 39 J| .180 .166 .112 .096 8 38 47 3-1 2.4 1.9 1.9 23 39 39 8 f .186 .144 .126 .104 23 32 44 2.4 2.3 1.8 1-4 4 25 42 " 9 ( 1 10 ) .164 .132 .106 .078 20 35 52 2.2 2.0 1.6 1-5 9 27 32 " 12 \ .176 .122 .094 .082 31 47 53 2.2 2.1 1.6 i.3 5 2 7 41 THE MACDOUGALL POLARITR SYSTEM OF PURIFICATION. O r~ m Q "i O O OOOOOOOO-r>OO"im OOCOTOOOO " c^co t <r -ro fi O o T T c* o -fj q q o i^ m r^ r-i M q o -rco o -r w m -r tn ./> ct o q q r-io - -do d oor^oo>o M-C-T--OOOOOOO Euiuiniv paAios^iQ I o o o o o o o cTo 6~o" bo^ooooooo o~o"^ o o o o o o u o c c papuadsns | T - -t -t o o - CO T OQQQOOO oo J; ooooooo oooobo^o -. ooooooooooooo ^- oocoooooooooo o o - o O O O O O O 3 O"NN"OOOOOOO o 2 ocTcTooJooooooo - o o o c > C I I I ! I u M ~t r^ s - < o o< o o 329 33 WATER PURIFICATION AT LOUISVILLE. BACTERIA PER CUBIC CENTIMETER IN THE OHIO RIVER WATER 1 BEFORE AND AFTER TREATMENT BY THE POLARITE SYSTEM. Date. Hour. Rate of Fi tration. Cubic Feet per Minute. Bacteria per Cubic Centimeter. Remarks. River Water. Settling Pipe Effluent. Jewell Effluent. Polarite Effluent. 1897 May 10 " 10 " 10 " 10 " 10 " 10 10 " ii " ii " 12 " 12 " 12 " 12 " 12 " 12 " 12 " 12 " 13 13 13 13 13 13 13 13 13 13 3 M 14 14 14 14 14 15 15 15 15 15 15 15 15 15 17 17 17 17 " IS " 18 " 18 " 18 " 18 " 18 " 18 " 18 " 18 " 18 " 19 19 IQ 11.00 A.M. I.I5 P.M. 3-4 " 5.00 " S.io " 9.U5 " 11.00 " 23-5 23.5 23-5 23-5 23.5 23-5 23-5 5 800 6 400 5 800 II OOO 10450 8 300 4500 3490 3 (>9 21 IOO 1 8 200 15 ioo 15 800 13 too 8 900 12 22O 8550 10 500 * Application of sulphate of alu mina began at 7.46 P.M. f Application of sulphate of alu mina stopped at 9.00 A.M. 17 ooo 13 700 3.00 " 10.30 " 11.30 " 12.30 P.M. 2.40 " 4-45 " 5.00 " 9.00 " II.OO " I.OO A.M. 3.00 " 5.00 " 8.00 " g.oo " 10.00 " 12. OO M. 2.OO P.M. 4.00 " 9.00 " 1 1 . OO " I.OO A.M. 3.00 " 5.00 " 9.00 " g.OO P.M. II.OO " I.OO A.M. 3.00 " 5-3" " 7.OO " g.OO " II.OO " I.OO P.M. 3.OO " 5.30 " 3.30 " 5.05 g.OO " II.OO " I.OO A.M. 3.30 " 5.00 " 9 oo " 11.00 " I.OO P.M. 3.00 " 5.30 9.00 " II.OO " 10. ii Soo 45 7oo 46 500 43 ioo 42 900 12. O 12.0 12 700 19050 34700 38600 12. O "5 12.0 12.0 12. O 2.0 2.0 7300 5 250 4 190 3 no 27950 19 200 19 500 36 ooo 25 ooo 17500 12 050 18 950 39000 60000 21 700 24 ioo 33000 38 7OO 39700 35 300 8.0 8.0 8.0 S.o 8.0 8 o 8.0 8.0 8.0 8.0 3.0 3-0 3.0 3-0 3-0 3-o 3-0 3-o 4.0 3-o 3-5 8.0 15 900 7 700 it 500 20 80O 19300 15 ioo 20 500 18 200 81 500 64 500 21 900 22 6OO 17400 19 800 24000 37 ooo 31 700 13 800 13 750 8 500 9 500 9400 48 ooo 37 200 21 5OO 54100 39 ooo 18 900 38500 45 ioo 255 ooo 64 200 * 1 7 ooo 13 200 15 ooo 15 ooo 32 400 19500 fi3 500 32 500 24500 33750 18000 34500 27 200 12 5OO IO IOO 12 5OO II 400 16 900 1 8 500 30500 13900 9 1 20 10640 17 200 8 900 9700 8 400 12 800 36 500 45 200 1 36 Soo 49 ioo 62 700 6 1 ooo 57 ioo 63900 59 700 64 500 24 500 26400 8.0 8.0 8.0 8.0 S.o 8.0 8.0 1 8 800 8 900 8 700 9300 ii 500 8 600 21 gOO 1 8 900 14 800 15 ooo 27 300 8 ooo 3900 4 800 27 800 18 500 16 ioo 21 200 8.0 8.0 8.5 8.0 8.0 8 o 19 500 10 710 9 200 3.00 " 5.00 " 8.0 8.0 12 IOO 1 As a matter of convenience, the numbers of the bacterial samples are omitted. Samples Nos. 5809 to 5962 were collected from May 10 to 19, and samples Nos. 6150 to 6472 from May 28 to June 12. THE MAC DOUG ALL FOLAKITF. SYSTEM OF PURIFICATION. 33 BACTERIA PER CUBIC CENTIMETER IN TIIF, OHIO RIVER WATER BEFORE AXD AFTER TREATMENT BY THE POLARITE SYSTEM. Continued. Rate of Filtration. B icteria per Cu bic Centimes Dale. Hour. Cubic Feet per Minute. River Water. Clay Extractor F.fflui:nt. Jewell Effluent. Polarite Effluent. Remarks. 1897 " 28 6 450 4400 " 28 " 29 29 g.oo 1. 00 A.M. 3 oo " 23.5 23-5 23-5 8 500 IO IOO 6 loo 8 300 5 600 9 200 3 660 3 750 3900 4 900 2 890 3 MO 3 1 20 3 120 29 29 29 1 3 1 3 1 3 1 3 1 3 une g.oo " 12.00 M. 5.OO I .M. g.OO A.M. 11.00 " 5.OO P.M. 8.00 " 11.00 " 23.5 23-5 23.0 23-5 23-5 23-5 23-5 23.5 6 700 8000 2 890 2 420 2010 2 930 7400 7 100 8 400 21 OOO 21 500 2970 3 200 4 ioo 5400 6 500 6950 *5 120 3 19 2 400 2 ggO 2 720 3 220 4950 5 too 3750 2 150 I 440 1 580 I 270 * Application of sulphate of alu mina began at 7.51 A.M. 5.00 " 23-5 3 Coo 3710 242O 1 350 i 935 " II. OO " 4.0O P.M. 23.5 23.5 I 580 i 850 I 650 2 520 2 720 fl950 i 750 t 640 i 695 f Began application of sulphate o: " 11.00 " 230 i 750 I 350 I 890 I 120 stopped application at Jewel filter ;; 5.00 " 23-5 i 290 I 690 I 320 870 850 \ Application of sulphate of alu ;; 11.30 " 23.5 2 IIO 2390 I 040 I 540 650 595 " 5.00 " 23-5 2 290 2 04O 640 715 302 489 II. OO P.M. 23-5 1950 I 310 I 840 529 610 45i 258 3 5.00 " 23-5 I 360 I 250 43 660 281 385 3 11.00 " 23.5 I 510 1970 339 286 306 222 3 5.00 " 8 oo " 23-5 I 530 I 140 301 254 I 7 8 173 3 11.00 " 23-5 I I/O I O20 219 2OO 286 4 5.00 " 8 30 " 23-5 I O6O 930 I 580 242 149 191 4 12.00 M. 23-5 990 950 184 319 125 4 5.00 " 8.00 " 23-5 970 I 57 161 151 129 146 4 11.00 " 23-5 8So i 740 172 99 log 5 5.00 " 8.00 " 23-5 860 3 i8d M5 154 87 in 5 II. OO " 23.5 650 710 157 202 95 153 5 4.30 " 23-5 610 840 i 580 167 415 143 399 7 II. OO " 23-5 790 890 237 194 223 181 7 5.00 " 8.00 " 23-5 670 870 190 152 103 141 7 " 8 II. OO " 23-5 660 I 420 850 144 117 1 08 108 8 8 8 8 5.00 " II. OO " 4.30 I .M. 8.00 " 23.5 23-5 23-5 780 550 610 910 1970 i 740 93 163 Si 23 118 81 53 87 g Agitated and drained water frorr surface of sand layer at 1.25 P.M 8 9 n.oo " 2.OO A.M. 23.5 23-5 610 i 090 i 420 96 97 74 53 WATER PURIFICATION AT LOUISVILLE. BACTERIA PER CUBIC CENTIMETER IN THE OHIO RIVER WATER BEFORE AND AFTER TREATMENT BY THE POLARITE SYSTEM. Concluded. Rate of Filtration. Ba cteria per Cu bic Centimete r. Dale. Hour. Cubic Feet per Minute. River Water. B.2S* Effluent. Jewell Effluent. Polarite Effluent. Remarks. IS 97 June 9 5.00 A.M. 23-5 570 I 310 104 53 9 II. OO " 23-5 39 540 88 * 8 4 37 53 * Agitated and drained water from 9 5.00 " 23-5 500 640 55 53 surface of sand layer at 2 oo P.M. 9 I I . OO " 23-5 500 510 54O 54 61 27 33 " o 5.00 " S oo " 23-5 510 620 580 55 t59 53 f Agitated and drained water from " II.OO " 23o 450 97 49 41 surface of sand layer at 6.35 A.M. " o 5-30 " 23-5 370 950 *59 47 surface of sand layer at4.$i I -M. 3 Agitated and drained water from " o II.OO " 23-5 520 460 60 lUc 39 surface of sand layer at 9.30 P.M. I 5.00 " 8 30 " 230 530 300 154 27 surface of sand layer at 12.00 A.M. " I II.OO " 2j.5 390 I 090 t5i 30 41 "Agitated and drained water from surface of sand layer at 3.00 A M. " I " I " I " 2 5.00 " 8.00 " II.OO " 5.OO A.M. 23-5 23-5 23-5 23-5 53 410 390 370 300 420 370 47 45 37 41 40 32 52 43 26 * Agitated and drained water from surface of sand layer at 7.17 A.M. f Washed Jewell filter from below at 8.53 A.M., using 644 cubic feet of wash water. ,, 38 28 " 2 4.30 " 23-5 790 490 53 48 INVESTIGATIONS Ol- THE WATER COMPANY FROM APRIL TO JULY, 18U7. 333 CHAPTER XV. INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 1897, AND A RECORD AND DISCUSSION OF THE RESULTS ACCOMPLISHED IN CONNECTION THEREWITH. TIIK status of the problem of the purifica tion of the Ohio River water, for the water supply of the city of Louisville, on March 10, 1897, when it was decided that the Water Company should conduct further investiga tions along this line, and independent of the companies previously associated in the tests, may he briefly stated as follows: i. The Warren, Jewell, Western Gravity, and Western Pressure Systems, in the form in whic h they were tested, were not capable of purifying the Ohio River water in a practi cable manner, for the following principal rea sons: a. The absence of suitable provisions to al low the suspended matters to subside by gravity caused the use of coagulating chemi cals in amounts which made their use expen sive to an excessive degree, and which were objectionable with regard to the quality of the filtered water. b. The absence of suitable settling basins to allow subsidence (sedimentation) to be employed to its economical limits caused, at times of muddy water in the river, so much suspended matter to pass on to the filter, that in at least one case the filter was unable to purify enough water to wash it properly; and in the case of the best filters, it would be necessary in practice to provide reserve fil ters, with an area of 65 to 75 per cent, of those normally in service. The cost of construction and of operation at irregular, but unknown, intervals, of such a large reserve portion of a plant, would increase the cost of purification very materially. r. The absence of suitable settling basins to allow subsidence to be employed to its economical limits, would cause, at times, in practice with the best available filters, the wasting of from 25 to 35, and occasionally even a greater, per cent, of filtered water, in order to wash the filters properly. This large waste of water would call for the use of more pumping machinery, and would thus increase the cost of pumping the water which actually reached the consumers. Furthermore, the waste of filtered water, which would be used in washing the filters, and thence run into the sewer, would mean practically a correspond ing waste of coagulating chemicals. (/. It was demonstrated conclusively that, owing to the very frequent and marked changes in the composition of the water in the Ohio River, it was difficult to give to the systems of purification suitable attention to guard against, on the one hand, an imper fectly purified effluent, and, on the other hand, an unnecessarily large application of coagulating chemicals, with its needless in crease in the cost of chemical, and in certain objectionable results in consequence thereof. With a proper employment of subsidence, it would be much easier to operate a large system of purification satisfactorily, inde pendent of the consideration of a large and needless reserve portion of the system. 2. The Harris Magneto-electric System was a complete failure. 3. The use of an electrolytic action upon metallic aluminum plates, such as was tried by the Harris Company in July, 1896, and inves tigated further by the Water Company, in August, 1896, with a view to its substitution for sulphate of alumina, as a means of-secur- ing a coagulating chemical, did not give prom ise of practicability, owing to its cost. This method possessed certain advantages over 334 WATER PURIFICATION AT LOUISVILLE. sulphate of alumina, however, as was pointed out in foregoing chapters. 4. The suhstitution of iron plates for alu minum plates, in the electrolytic method, re duced the cost, hut the results obtained in connection with the Mark and Brownell de vices were so inadequate that the practicabil ity of this method was an open question at this time. > The results of laboratory experiments indicated the advisability of considering other coagulating chemicals than sulphate of alu mina, notably the iron compounds. In cor.nection with the importance of the last mentioned point, it may be added that the results of the experiments in 1884, and of more recent observations, indicated, but did not prove conclusively, that, after the removal of coarse particles by plain subsidence, the Ohio River water would require, at times, coagulation and further subsidence before economical filtration, either by the English or American type of filters, could be adopted. As was stated in the last chapter, it was de cided shortly before this time to test the polarite system, in which it was claimed that no use was made of chemicals. It was learned, however, before these investigations were half completed, that this last system was im practicable. OBJECTS OF THE INVESTIGATIONS. The objects of the investigations were to obtain practical information, so far as was possible on an experimental scale, and with appliances which could be made promptly available, on the following principal points: 1. The removal of mud, silt, and clay from the river water by plain subsidence (i.e.. un aided by coagulating chemicals). 2. The relative economy, advantages, and disadvantages of all available coagulating chemicals, prepared by various methods (chemical and electrolytical). 3. The most economical and efficient ap plication of coagulating chemicals to aid in the removal of the bulk of various suspended matters by subsidence. 4. The most economical and efficient ap plication of coagulating chemicals, to aid in the rapid filtration of the water, follow ing a partial purification by preliminary sub sidence. 5. The consideration of the best method of grouping together the available information upon the foregoing [joints, with the view to determining the system of purification which would give at a minimum cost, an effluent of the best quality from a practical point of view. PLAN OK PRESENTATION OF THE RESULTS OF Til KSE 1 NVKSTIGATIONS. There were two principal lines of evidence obtained during these investigations. The major portion came from experiments made with a series of devices, including settling basins, electrical and other appliances, for the preparation and application of different co agulating chemicals, and also the Je\vell filter. Additional data of importance were also ob tained from laboratory experiments, and from the separate operation of portions of the above-mentioned series of devices, notably those for the electrolytic production of co agulating chemicals. The plan adopted for the presentation and discussion of the results of these investiga tions is as follows: A description is first given of the devices which were employed. The conditions under which these devices were operated are next described. A detailed record of the results of analyses is then presented, followed by a summary of all the principal data obtained from the operation of these devices. The remaining portion of the chapter, con taining a discussion upon various points di rectly related to the above-stated objects of the investigations, is of most importance. In it use is made of such portions of the sum mary which precedes it as bear upon the point in q-uestion; and the results of the laboratory and other special experiments are recorded in their appropriate place in the discussion. It will be noted that, at the close of each section of the discussion, conclusions are drawn, so far as the available information upon the point in question will permit. These conclusions, with those from other portions of the report, are grouped together as a mat ter of convenience in Chapter XVI. INVESTIGATIONS OF THE WATKR COMPANY FROM APRIL TO JULY, 1897. 335 DESCRIPTION OF THE DEVICES ARRANGED AND OPERATED BY THE WATER COMPANY. For the purpose of carrying on investiga tions along the several lines referred to above, use was made in so far as possible of appar atus and devices available at the pumping station. Supplementary devices were de signed of such style and construction as would be convenient in carrying on this work without prolonged delays for construction, and at the same time conform to the arrange ment of devices already at hand. In order to make the results of operation comparable with those of earlier work, the system was arranged to be operated on a basis of 250,000 gallons per 24 hours (23.2 cubic feet per minute), as was the case with the other systems investigated. Except for the purpose of comparing the effects of different rates of filtration, and in some cases of in creasing the amount of electrolytic treatment, this rate was maintained as nearly as possible. The full system of devices, which were in use wholly or in part during these investiga tions, included the following: 1. Two circular wooden tanks of about 4000 and 1000 cubic feet capacity, respect ively. These were used as settling (sub sidence) basins. 2. Four wooden tanks of about 400 gallons capacity each, for the preparation of chemi cal solutions. 3. One electric generating plant, consist ing of a 5O-H.P. steam-engine, and a 2O-kilo- watt dynamo. (See Chapter XIII.) 4. Four electrolytic cells for the prepara tion of coagulants electrolytically. (See Chapters X and XIII.) 5. Four sets of metal electrodes, two of iron and two of aluminum. 6. The settling chamber and filter of the Jewell System of purification. (See Chapters IV and V.) 7. Necessary pumps, piping, valves, and meters, to allow the desired operations and observations. This system of devices, or portions thereof, was arranged at various times in a manner to permit several different ways of operation, as indicated below: 1. River water was admitted directly to the inlet of the Jewell settling chamber; and various coagulants were applied at the inlet or the outlet of this chamber, as desired. 2. River water was admitted at the bottom of the large basin (basin No. i), removed from the top of this basin, and pumped into the bottom of basin No. 2. Thence it was re moved from the top of basin No. 2, and pumped into the Jewell settling chamber. Different coagulants were applied at the inlet to basin No. 2, and the inlet to the Jewell settling chamber, or at only one of these places, as desired. 3. In this case the use of basin No. 2 was omitted, and the water passed from basin No. i directly to the Jewell settling chamber. Coagulants were applied at the inlet to basin No. i (chemicals only), and at the inlet or outlet of the Jewell settling chamber, as de sired. A more detailed description of the various devices is next presented. Settling Basins. Basin No. i. This was a circular pine tank, placed in the house formerly occupied by the Western Systems, and had the follow ing inside dimensions: Diameter at the bot tom, 17.0 feet; diameter at the top, 10.33 feet; and height, 19.0 feet. Its total capacity, and average working capacity were 4110 and 4000 cubic feet, respectively. The staves and the bottom of the tank were 3 inches in thickness. It was held together by 13 iron bands, varying in width from 4 inches to 2 inches, and each 0.125 inch thick. The lowest band was at the bottom of the tank, and above it the successive bands were placed at distances apart increasing from i foot at the bottom to 3 feet at the top. The 4-inch inlet pipe entered at the side. about 2.5 feet above the bottom, and extended into the tank a distance of about 8 feet. The flow of water was regulated by a gate valve placed on the inlet pipe outside the tank, sup plemented by a single-seated check valve, op erated by a float on the inside of the tank. In front of the mouth of the inlet pipe, which lay horizontally, was a small baffle plate, distant about 3 inches, arranged with the view to 136 WATER PURIFICATION AT LOUISVILLE. breaking the current of water. The 4-inch outlet pipe was connected to the tank about J 5 inches from the top, and on the same side as the inlet pipe. Connections were made for draining into the sewer through an 8-inch opening in the bottom of the tank at one side. During op eration this opening to the sewer was closed by a plug. The bottom of the tank was nearly level, and it was necessary to use a stream of water from a hose, and a broom, to remove the accumulation of sediment. Applications of aniline dyes, and of caustic soda, were made to the water as it entered the basin, and tests were made to note the appear ance of the chemicals at the outlet. The re sults of these tests, to learn the way in which the water passed through this basin, are sum marized just beyond, together with cor responding results from the other settling basins. Basin No. 2. This basin was made by re pairing the tank formerly used as the West ern gravity filter. (See Chapter V.) The inside dimensions were as follows: Diameter at the bottom, 10.0 feet; diameter at the top, 9.5 feet; and height, 14.0 feet. Its total ca pacity, and average working capacity were 1045 an d looo cubic feet, respectively. The 4-inch inlet pipe was connected to the side of the tank by a flange joint, 2.0 feet above the bottom. A 4-inch pipe, which was connected to the tank on the opposite side from the inlet, and 2.5 feet below the top, formed the outlet. There were no baffle plates or other devices in this tank to assist in making uniform the displacement of water. A 3-inch pipe was connected to an opening in the bottom, to allow drainage to the sewer. The sediment on the bottom was removed by a stream of water from a hose and by a broom. The results of the application of chemicals to the water, in ordej to learn the manner in which the water passed through this basin, are summarized beyond, with corresponding results from the other basins. Jewell Settling Chamber. This chamber has already been described in full in Chapter IV. It was a closed compartment cylindri cal in form, 6.79 feet high and 13.5 feet in diameter, inside dimensions, having a capac ity of 879 cubic feet. It was used without modifications in all operations up to June 19. On this date wooden boxes were inserted to cover the portions of the inlet and outlet pipes lying within the chamber. The end of the inlet pipe was enclosed in a box, so that the water entered the chamber at the bottom instead of 1.55 feet above the bottom as for merly, as the sides and front of the box were 3 inches above the floor of the chamber. By means of wooden framing the outlet pipe (at the top of the chamber) was closed in, leaving an opening 6 inches wide and 24 inches long, about diametrically opposite to the inlet pipe. The water passed through this opening into the portion enclosed by the wooden parti tions, and thence to the filter, by means of the regular central pipe which served as an out let to the settling chamber. The results of the application of chemicals to the water (after the above stated modifica tion was made), in order to learn the manner in which the water passed through the cham ber, arc summarized in the next table, to gether with corresponding results from basins Nos. i and 2. TABLE SHOWING A SUMMARY OF THE RESULTS OF THE TESTS OF THE DISPLACEMENT OF WATER ON ITS PASSAGE THROUGH SETTLING BASINS NOS. 1 AND 2. AND THE JEWELL SETTLING CHAMBER. Normal capacity in cubic feet 4 ooo I coo 879 Time of filling at the regular rale (23.2 cubic feet per minute) ! 2h. 52m. 43m. 3Sm. Time elapsing between applirationj of chemicals at the inlet and first appearance at the outlet 3gm. 4m. Time elapsing between application | of chenvcals at the inlet and max-; imum appearance at the outlet.. .. ih. lorn. lorn. Percentage which chemical appearing at the end of "time of filling" was, of maximum which appeared .... 24 33 Time elapsing between application of chemicals at the inlet and appearance of 50 per cent, at the outlet 2h. 4gm. 3Om. Application of Chemical Solutions to Secure Coagulation. For preparing chemical solutions four wooden tanks, each of about 400 gallons capacity, were used. They were arranged in INVESTIGATIONS OF Till . M ATKR COMPAXY FROM APRIL TO JULY, M>7. 337 pairs, one above the other, and the corre sponding upper and lower tanks were used for the same kind of chemicals. Solutions were prepared in the upper tanks, and ad mitted to the corresponding lower tanks as required, the lower tanks serving for pump wells and storage basins. Pumps. In addition to the small pump used by the Jewell Company in 1895-96 (de scribed in Chapter 11), a small duplex pump, having the following principal dimensions, was placed in operation: Diameter of steam cylinder. 1.125 inches; diameter of water cyl inder, 2 inches; length of stroke, 2.75 inches. Piping. Each pump was provided with two suction pipes, 0.75 inch in diameter, which reached to within i inch of the bottom of the lower tanks. By this arrangement the pumps could be supplied with solution from either tank. For delivery pipes, use was made of such small piping as was available at the pumping station. In the case of the old (simplex) pump, the o.75-inch heavy lead pipe used in the Jewell System was utilized. It was connected to the inlet pipe of the set tling chamber, at a point about 6 feet from the chamber, or lowered about 3 feet into the pipe leading from the top of the settling chamber to the filter, as it was desired to ap ply the chemicals before or after the water had passed through the settling chamber, re spectively. The delivery pipe from the new (duplex) pump was connected to the inlet to basin No. i. or to the inlet to basin No. 2, as desired. It was mainly made up of 0.5- inch iron pipe, but for connection to the inlet to basin No. i about 40 feet of o.75-inch iron pipe were also used. Kinds of Chemicals. In connection with the operation of the devices under considera tion five different chemicals, as follows, were used at various times: 1. Sulphate of alumina (two different lots). 2. Persulphate of iron. 3. Potash alum. 4. Protosulphate of iron (copperas). 5. Caustic soda. A number of other chemicals were used in laboratory experiments, as will appear in a subsequent portion of this chapter. The composition of the above chemicals is given in the next section. No special features are to be noted here, except in the case of the persulphate of iron. In making solutions (from 0.3 to 0.9 per cent., accord ing to the quantity to be applied) of this chemical in filtered water, it was difficult to dissolve it. It was found, however, that a solution could be made best by adding, suc cessively, small quantities of water to the sub stance at the outset, and decanting the solu tion into another tank. Hot water could not be used, because it decomposed the chemical. and with cold water there was also a tend ency for the iron to form the sticky hydrate, which retarded the solution of the portion covered by it. This chemical contained considerable material which was completely insoluble (see the analysis below), and it was necessary to remove this by straining through cloth, because it cut the fittings of the pumps and meters. There was also enough free sulphuric acid (2.73 per cent.) in the persulphate of iron to corrode the brass and iron pipes and fittings very rapidly. It is possible, however, to obtain a commercial product of this kind which does not possess these disadvantages, except perhaps a small amount of insoluble matter. Tn the case of the potash alum, which con tained a small quantity of ammonia, all quan tities are calculated, as was the case in 1896, as sulphate of alumina, on the basis that six teen parts of the former equal ten parts of the latter. Composition of Chemicals. The two lots of sulphate of alumina, and the lot of persul phate of iron, were analyzed with the follow ing results: PERCENTAGE COMPOSITION OF COAGULATING CHEMICALS. Sulphate o f Alumina. Petsul- I.o, No., Mailer insoluble in waler Available alumina (A1>O,) Sulphuric acid (SO,) 0.22 15.96 34 26 0.06 19.64 36.85 5-56 o.oo 55.08 Oxkle of iron (FejOi) o.oo Trace I ime (CaO) 0.56 0.06 o . oo Water (H?O) 49.00 43 39 4.46 The potash alum was of high grade and had about 10.7 and 34.0 per cent, of alumina and sulphuric acid, respectively. The copperas crystals were also of a good quality, and contained 57.55 and 28.78 per 338 WATER^P UNIFICATION AT LOUISVILLE. cent, of iron oxide and sulphuric acid, re spectively. Analyses of the caustic soda showed that it contained 73.4 per cent, of sodium oxide. Devices for the Application of Electrolytic Treatment to Secure Coagulation. Two electrolytically prepared coagulants (hydrate of aluminum and hydrate of iron) were used at various times. Each was pre pared by the electrolytic decomposition of electrodes of the respective metal. For the purpose of preparing these hydrates, elec trodes in the form of manifolds of plates were placed in closed iron cells, and the river water and electric current passed through them. The following devices, all of which were placed in the house formerly occupied by the Warren System, were used: Generating Plant. The electric generating plant used in connection with the Mark and Brownell devices was employed for genera ting the necessary electric current. The dynamo was rated at 400 amperes and 50 volts, but could be operated as high as 450 amperes safely for short periods. For a full description of this engine and dynamo see Chapter XIII. As already described, the electric current was regulated by means of a field rheostat. In order to maintain the re sistance necessary to balance the potential of the machine, when operating with low am perage, a large rheostat was inserted on the main circuit. At the close of these investi gations, Aug. i, 1897, tests were made of the engine and dynamo, with the following re sults: RESULT OF TESTS OF ELECTRIC GENERATING PLANT. ENGINE TEST. Kffic Indi 50 37-5 25.0 12 .O DYNAMO TKST. 95 per cent. Q3 85 " 70 Urake Horse Po Owing to the fact that the engine (50 I.H.P.) was a much larger machine than the dynamo (28 E.H.P.), and was accordingly at all times operated considerably below the point of maximum efficiency; and, further, owing to the fact that at times the entire plant was too large to give economically the small amount of electric power required for the treatment of fairly clear water, the com bined efficiency of the engine and dynamo was very low in some cases. In practice, however, a generating plant would be ar ranged in several units, and such portions of it used as would furnish the electric current most economically. Under such circum stances the above tests show that 80 per cent. of the indicated horse power at the engine could be obtained as electric power. This agrees fairly well with the results of good I modern practice. Electrolytic Cells. Four electrolytical cells, j with covers, were used in this work. Cells Nos. i and 2 were those of the Mark and Brownell devices, and have been fully de scribed in Chapter XIII. Cells Nos. 3 and 4 were two of the large cells of the Harris Company, with whom arrangements for their use were made. These cells have been de scribed in Chapter X. Several changes in all of these cells were made, as follows: Changes in Cells Nos. i and 2. These du plicate cells were 30 inches in diameter, had a capacity of 35.2 cubic feet, and were not insulated on the inner walls. The special distributing devices, attached to the inner side of the dome at the top, were removed, and, in each case, the 4-inch inlet pipe was connected by a flange joint to the center of the dome-shaped cover. The original out- lets on the side were closed by plugs, and the 3-inch opening at the apex of the conical bottom in each cell was used as an outlet. Changes in Cells Nos. j and 4. These du plicate cells were removed from the Harris house to a position by the side of cells Nos. i and 2. They were 35.5 inches in diameter, and had a capacity of 28 cubic feet. The in sulating rubber linings, and the covers of a special casting, with magnets resting upon them, were removed. A slightly arched and circular iron plate, to which the 4-inch inlet INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 1897. 339 pipe was attached in the center, served as a cover in each case. A wooden frame was built in the bottom, so that the electrodes when placed upon it reached within 0.25 inch of the flange to which the cover was attached. There was tapped in each cover a hole, to which a o.25-inch pipe, with a pet-cock, was attached, in order to allow the escape from time to time of accumulated gases. The 3- inch opening in the bottom of the cone was connected with a 4-inch pipe, which served as an outlet and waste pipe. Electrodes. New electrodes, numbered the same as the cells in which they were placed, were constructed, and used exclusively in the tests described in this chapter. Electrodes Nos. i and 3 were each made of a manifold of wrought-iron plates o. 125-inch thick. Electrodes Nos. 2 and 4 were made of alu minum, to duplicate Nos. i and 3. In the method of construction these four electrodes were identical. The "size of the cells used necessitated slightly different forms, however. Each manifold contained fifty-six plates. Size of Plates. Electrodes Nos. I and 2 were made of plates 50 inches long. In or der to fit the cell the widths of the plates varied from 24 to 12 inches, averaging 20.4 inches. The total area of one side of the plates was 56,000 square inches, and the cross section of the electrolyte was 55,400 square inches. When new, these sets of plates weighed 1780 and 674 pounds, respectively. Electrodes Nos. 3 and 4 were made of plates 36 inches long. The widths varied from 30 to 22 inches, and averaged 27.4 inches. The total area of one side of these plates was 56,200 square inches, and the cross section of the electrolyte was 55.500 square inches. When new, these sets of plates weighed 1824 and 682 pounds, respectively. Formation of Manifolds. All of the four manifolds were formed alike. The plates were held together by six i-inch iron bolts, and the desired distance between the plates, 0.25 inch, was maintained by the use of wash ers, or separators. These bolts were set as far to the edge of the sets as the width of the outer plates would allow. The plates were so thin, however, that they buckled badly, and it was necessary to insert many small pieces of insulating material between them at different places. In order to relieve this diffi culty to some extent, o. 5-inch bolts with separators. were placed in the corners of the electrodes, at the same time that the other changes were made, and which were com pleted May 30. One of the upper corners of each plate was cut off, and the plates arranged in the mani fold so that the cut and uncut corners came alternately on each side. To the uncut cor ners on each upper side brass lugs were riveted. The cables carrying the electric current were soldered to these lugs. Extensions of six of the plates of each set, with openings which coincided along the cen ter line, were arranged as lifting lugs to aid in handling the electrodes. When placed in the cells the electrodes rested upon a wooden framework arranged in the bottom. The space between the edge of the plates and the uninsulated inner wall of the cell ranged from 0.5 to 3 inches, and averaged about 2 inches. The outer portion of the frame work at the bottom was solid, so that there was no opportunity for water to pass down ward through a cell, except through the spaces between the plates of the electrodes. Insulation of Electrodes. Owing to inabil ity to secure, without long delay, hard rubber fittings for the insulation of the electrodes, it was decided to proceed as follows: The iron bolts by which the manifold of plates was held together were covered with steam hose, and circular separators of vul canized fiber 0.25 inch thick were placed on the bolts between each pair of adjoining plates. This method of insulation proved to be a failure, owing to a certain, but not ac curately known, portion of the current pass ing through the cell on the fittings. This portion of the current was consequently wasted, so far as the treatment of the water is concerned, and the greater part of the re sults obtained with these devices during the month of April, is of very little or no value. The cause of this failure was the presence of small particles of metal in the hose or the fiber, or both. An arc was probably formed between the plates, and the metallic particles and molten metal were deposited in the inter vening space. Repeated attempts, with only 34 WATER PURIFICATION AT LOUISVILLE. partial success, were made to remedy these difficulties by the liberal use of insulating tape and paint on the hose, and the removal and repair of fibers showing evidence of metallic particles. Wooden bolts were also substituted for the iron ones. But, as the electrodes continued in service, it was found that the vulcanized fiber separators absorbed water so that they swelled to a degree that caused the electrodes to lose their original form, and stripped off the heads of the wooden bolts. During the first week in May it was decided to abandon the original insu lating appliances, and procure an entire set of hard rubber fittings to cover the iron bolts, and to separate the plates. On the large bolts the new separators were 3 inches in diameter, and on the small bolts. 2 inches. These changes, which proved to be thor oughly satisfactory, were completed on May 30. In the modified form the fittings weighed about the same for each set, 84 pounds. Electrical Connections. The main electrical circuit was the same, for the most part, as in the case of those devices described in Chapter XIII. Near the switchboard there was placed a rheostat, by which the current pass ing through the cells could be more satisfac torily regulated. The circuit was arranged so that any, or all. of the cells could be con nected at once, and the direction of the cur rent through any of the cells could be promptly reversed. Connections with the electrodes in cells Xos. I and 2 were made through two open ings filled with wooden plugs, through which iron binding posts were driven. The main circuit was connected to the outer binding posts, and to the inner binding posts the cables attached to the lugs riveted to alter nate plates were connected. Similar arrange ments were made for connecting the main circuit with the electrodes in cells Xos. 3 and 4. except that the connection through the openings in the cells were brass binding posts, placed in hard rubber stuffing boxes. Modification of Electrode No. I. In order to give greater treatment to the water than the original form allowed, electrode Xo. i was changed, on July 9, into two electrodes in series. This was accomplished by divid ing the set in halves, electrically, and con necting one-half of the plates on one side to half of the plates on the diagonally opposite side. Xo changes were made other than in the wiring, as described. r/ping. As the iron electrodes (Xos. I and 3) were never used in connection with the filter at the same time as the aluminum electrodes (X T os. 2 and 4), the inlet and out let pipes of each pair of cells (Xos. i and 2 and Xos. 3 and 4) were branches of the same main pipe, respectively. The inlet and outlet pipes were 4 inches in diameter, except that the outlet pipes were reduced to 3 inches at the connections with the cells. Further, the outlet pipe of each cell was itself branched, so that by suitable valves and pipe connections it could also serve as a waste pipe to the sewer. When basin Xo. T was in service the water as it left that basin could be pumped through either cell Xo. 3 or cell X T o. 4, on its way to basin Xo. 2. Similarly, as the water was pumped from basin Xo. 2 to the Jewell set tling chamber, it could pass through either cell Xo. i or cell Xo. 2. Cells Xos. 3 and 4 could not be used when basins Xos. i and 2 were out of service. At such times river water could be taken from the old Warren inlet pipe (after slight changes) and passed through either cell Xo. T or cell Xo. 2. on its way to the Jewell settling chamber. Dur ing the last of the tests, after July 9, the piping was arranged so that the water could be pumped from basin Xo. I through cither cell Xo. 3 or 4. and then through Xo. i or 2 to the Jewell settling chamber. A further modification at this time, as stated above, al lowed the passage of river water from the old Warren inlet directly through cell Xo. I or Xo. 2, and to the Jewell settling chamber. When chemical solutions were applied to the water prior to filtration, by-passes and valves made the electrical appliances inde pendent of the other devices. At such times special experiments were usually made with these appliances, and water for that pur pose was taken from the main through the old Warren inlet pipe, located by the side of cell Xo. i. INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 1M7. 341 The Jewell Filter. It was arranged with the O. If. Jewell Fil ter Company to make use of the Jewell filter, which was the only one remaining at the pumping station. Xo modifications were made in this filter, which has been fully de scribed in Chapter V. Pipes, Valves, Pumps, and Meters. From the foregoing account of the ways in which the several devices could be connected, a general idea may be obtained as to the ar rangement of the piping. Further details are not of importance. But it may be recalled here that a majority of the piping was 4 inches in diameter. A small portion was 5 inches, and for a short distance from the out lets of the electrolytical cells the diameter was 3 inches. Suitable valves, meters, and gauges were placed where convenience re quired. Owing to the fact that the elevation of the Jewell filter was very nearly as high as that of basins Xos. i and 2, and that the water passed through some 350 feet of old pipe, with quite a number of turns, valves, meters, and electrodes, it was necessary to set up pumps on the pipe between basins Xos. i and 2 and basin Xo. 2 and the Jewell System. These two pumps had capacities of about 250,000 and 400,000 gallons per 24 hours, under the pressure used, respectively, and were ones which the Water Company had on hand at the time. Adaptation to Existing Conditions in the Con struction and Arrangement of These Devices. In the consideration of the construction and arrangement of the devices which have been described in the foregoing pages, it must be borne clearly in mind that they were designed to enable as much practical infor mation as possible to be obtained in connec tion with other appliances at hand, and were not intended to be illustrative of the best forms for adoption in practice. They were arranged to yield data, with a minimum ex penditure of time and money, which would show the lines which it would be most prac ticable to follow on a large scale. There were many features in the devices which cannot be taken as models of good practice, although they served their purpose in this work. Thus, at the outset of these tests, it was known that the settling basins were all far too small to give the most eco nomical and efficient results; the electrical appliances were not well arranged to meet the requirements of all kinds of river water. The question of closed electrolytical cells as compared with open channels or conduits, the thickness of metal plates, the water space between the plates, and the manner of fas tening together and insulating them, were all open questions: and the desirability of test ing filters with different depths and sizes of sand was unquestioned. In the discussion following the results of these tests, mention will be made in several instances of methods for securing practicable results from funda mental principles established by these tests. DESCRIPTION OF THE CONDITIONS AND METHODS OF OPERATION OF THE DE VICES ARRANGED BY THE WATER COM PANY. The principal features concerning the gen eral operation of these devices during the several periods from April 5 to July 24. in clusive, 1897, are as follows: Composition of the River \\~atcr. It is dur ing the period of the year covered by these tests that the Ohio River water contains the largest amount of very minute clay particles, which, although less in total weight than the heavy mud of the winter freshets, make the water most difficult to clarify and to purify economically. For further reference to the composition of the river water during the spring and early summer, in addition to com ments beyond, see Chapter I. tntcrrnption of Tests. These investigations were not continuous, owing to the fact that the polarite system was tested during this pe riod, according to earlier arrangements. This caused the regular operations of the devices of the Water Company to be suspended from May 10 to 19. and from May 28 to June 12. On account of the abnormal clearness of the 34 2 WATER PURIFICATION AT LOUISVILLE. river water, the operation of all the devices was not resumed after the close of the polar- ite tests until June 19, except the electrical appliances, which were tested almost con stantly from May 30 to June 30. Xo other abnormal interruptions occurred, except from May 3 to 5, when it was impossible to obtain river water which had not settled for several days in the force mains. Different Provisions for Subsidence. Three settling basins, already described, were used in different ways (luring- these tests, accord ing to the amount and character of the sus pended matter in the river water and the na ture of the point under investigation. The Jewell settling chamber was used without exception. Basins Nos. i and 2 were both used on 119 different runs, as follows: April 1 I to May 28, 73 runs: June 20 to 22, 5 runs; June 24 to July 8, 41 runs. From July 14 to 15. and TO to 19. basin Xo. i was used with out basin Xo. 2. Comparison of Different Coagulants. A comparison was made of the efficiency and economy of hydrate of iron, obtained electro- lytically, and from commercial sulphates; and of hydrate of aluminum prepared electrolytic- ally, and from commercial sulphates. So far as practicable the several coagulants were ap plied to practically the same water, so as to obtain comparable results. From earlier statements, it will be recalled that fault} insulation of electrodes during April caused the electrolytic results of that month to be of uncertain value. The re modeled electrodes were tested, independent of the filter, from May 30 to June 20. Quantity of Coagulants. For obvious reasons the investigations were conducted with the view to determining the minimum quantity of coagulants which would yield an effluent of satisfactory, purity. In doing so it was necessary of course, at times, to estab lish definitely that certain quantities were in sufficient. At such times (usually short rnns with the filter) the effluent was unsatisfac tory; and, in a measure, the results were negative, although they possessed a positive value. Application of Coagulants. In studying the optimum method of application of the differ ent coagulating chemicals, they were applied so as to give a range of conditions with re gard both to the period of coagulation and subsidence, and to the period of coagulation prior to filtration, in the case of waters of dif ferent character. This range was limited by the capacity and facilities of the settling ba sins, already described. filtration. The Jewell filter was operated, in general terms, in a manner similar to that described in Chapter VII, except that there was no controller on the outlet pipe, and the above-mentioned conditions of operation called for some changes at times in the rate of filtration, and frequency of washing, as noted below. Rate of Filtration. As a rule the rate of fil tration was kept as nearly as possible at 250,000 gallons per 24 hours, or 94 million gallons per acre daily. This is equivalent to 23.2 cubic feet per minute, but it was the cus tom to adjust the valves to give 23.5 cubic feet. Owing to the fact that the electrical appliances were too small to furnish sufficient electric current to treat the water properly when in a very turbid condition, it w 7 as neces sary at times to reduce the rate of flow of water through the electrolytic cells, and, con sequently, the rate of filtration. On sev eral occasions the rate of filtration was re duced for this reason to about 16 cubic feet per minute. Early in these tests it was found that a larger amount of coagulating chemicals was required just after washing the filter than was the case during the major portion of the run, providing the same rate was maintained. With the view to reducing the amount of coagulating chemicals to a point sufficient for the latter portion of a run, the rate of filtra tion was reduced several times to about one- half the normal, for a short period just after was hing the filter. Length of Runs. While the regular custom of allowing a run to continue until the avail able head was exhausted or the quality of the effluent failed prevailed for the most part, there were a number of occasions when a comparison of coagulants required runs of only about 1000 cubic feet of effluent. These short runs served the special purpose for which they were made, and, therefore, are placed in the records, although they were INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 18H7. 343 abnormal so far as length of run is con cerned. Washing tlic Filter. Surface agitation of the sand layer was employed whenever it was practicable. In all cases of washing, it was carried to a point where the wash-water flow ing to the sewer was comparatively clear. The filter was always washed after short special runs, regardless of their length. Delays in Operation. In addition to the short delays incidental to such work, there were two sources of extended delays. The first occurred several times in April, when re pairs of the electrodes were necessary. A far greater cause for delay was the change of water in all the settling basins in service which held treated water, when there was a change either in the rate or kind of treatment to coagulate the water. It is estimated that this necessary cause of delay covered about .28 per cent, of the time devoted to the actual tests. Records of Operation, and Samples for An alysis. In this respect the same general plan which was adopted in the tests of 1895-96 was followed. As a matter of convenience, and for the sake of clearness, the records of operation, with summaries of analytical results, are pre sented in the next section by runs listed in serial number, rather than by days. The general operations are divided into three periods, viz: Period No. I, which extended from April 5 to May 10, the beginning of the tests of the polarite system. Period No. 2, which covered the time oc cupied by the changes made in the polarite system. Period No. 3, which extended from the close of the tests of the polarite system until the conclusion of the experimental work. In order to facilitate a more thorough un derstanding of the conditions of operation, and the summary of results beyond, the fol lowing outline of the important special feat ures of each period is presented. Period No. I. This period extended from the beginning of operations with these devices on April 5 to the time when the polarite system was ready for operations on May 9. From April 5 to 8 operations were continuous, day and night. On April 8, 9 and 10, operations were from 7.00 A.M. to 6.00 P.M. on each day. Con tinuous day and night operations were begun again on April u, and continued through out the period. From April if>, 2.00 P.M., to April 20, 4.00 P.M., and from April 24, 2.44 P.M., to April 26, 6.20 A.M., operations were suspended, to allow work in repairing the insulation of the electrodes. The system was closed down on May i, and from May 3 to 5, inclusive, attention was given to labora tory experiments, as the main pumping en gines were not in service, and it was not pos sible to obtain river water which had not been affected by a varying period of subsidence in the reservoir and pipes. Operations were be gun again on May 6, and were continued till 5.12 A.M. on May 9, when the filter was put in shape for use with the polarite system. The river water at the beginning of this period was about of a normal character, the suspended solids averaging about 350 parts per million. A slight rise increased the sus pended matter on April 9 to about 840 parts per million. From this date to May i the water gradually became clearer, the sus pended solids on the latter date averaging only 77 parts. From May 6 to 9 the sus pended solids ranged from 453 to 301 parts per million. During this period 66 (Nos. i to 66) runs were made. From April 5 to 10, including the first 14 runs, operations were with the original Jewell System, in an unmodified form. Attention was devoted to a compari son of the efficiency of the hydrates of iron and aluminum obtained from persulphate of iron and sulphate of alumina, respectively. The first 6 runs were without agitation of the surface of the filter. After this the use of surface agitation was made a regular feature in all runs continued to their normal length, providing the effluent remained satisfactory. On April n the full system of devices ar ranged by the Water Company was put in service. This system included the three set tling basins and the Jewell filter, together with the various devices for preparing and applying the several coagulants, all of which have been described above. 344 WATER PURIFICATION AT LOUISVILLE. Coagulants were applied in equal amounts at the inlets to basin No. 2 and the Jewell settling chamber, respectively, except on runs Nos. 53, 54, 65 and 66. On run No. 53 the entire chemical was applied at the in let to the Jewell settling chamber, and on runs Nos. 54, 65 and 66 at the inlet to basin No. 2. Four differently prepared coagulants were used: Klectrolytically decomposed iron, elec- trolytically decomposed aluminum, sulphate of alumina and persulphate of iron. Ex planation has already been presented of the difficulties met with in the insulation of the electrodes, and it will only be noted here that considerable uncertainty is attached to the electrolytic work during this period. There were two leading points under con sideration throughout this period. 1. Comparison of the efficiency of the four coagulants used. 2. Determination of the minimum coagu lant which could be used with safety under normal conditions of operation of the system used. Practically all of the runs were intended to throw light on the first point, and, in so far as possible, all coagulants were used on the same character of water. Of the 66 runs, 25 were made with sulphate of alumina, 10 with persulphate of iron, 20 with electrolytically decomposed aluminum, and 1 1 with electro lytically decomposed iron. In regard to the second point, the rapidly changing character of the river water, and the difficulties experienced with the electrodes, interfered to a large extent with this line of investigation. Much information can be gained, however, by a comparative study of consecutive runs. Operations were regular throughout this period, the rate of 250,000 gallons per 24 hours (23.2 cubic feet per minute) being maintained as closely as possible, except when the necessity of greater treatment than the capacity of the electrolytic plant would allow required a reduction in the rate. Period No. 2. This period extended from May 19 to May 26, inclusive, and included runs Nos. 67 to 87. Operations were continuous during the day and night, except for a short delay from 4.58 A.M. to 9.42 P.M. on May 21, to examine the strainer system of the Jewell filter. The river water contained about 280 parts per million of fine suspended matter at the beginning of the period. Absence of rains caused the water to become clearer during the latter part of the period, the suspended matter decreasing to 100 parts per million. Attention was devoted solely to the deter mination of the safe minimum amount of co agulant for filtration, in connection with sub sidence; and for this purpose the full system of three settling basins and the filter was em ployed. Sulphate of alumina alone was used. It was applied in all cases at the inlet to basin No. 2. By the use of these basins, together with the chemical treatment at basin No. 2, the amount of suspended matter in the water at the top of the filter was usually kept below TOO parts per million. As it was found that the character and the rate of clearing of the first water filtered after washing was usually the controlling feature in the consideration of the minimum amount of coagulant, several runs were stopped after the effluent had reached a fairly normal or constant character. This procedure caused a number of the runs to be very short, the ob ject of several being accomplished with 1000 cubic feet of effluent, or less. The principal modification of the operation was the successive use of several rates, 9, 12, 1 8, and 24 cubic feet per minute, in order to compare the net amounts of coagulants needed for filtration at the several rates. Period No. j. This period extended from June 19 to July 24, when regular operations were finally sus pended. In all 98 runs were made, Nos. 88 to 185, inclusive. Operations with the filter were suspended 011 July 4 and 5, and from July 10, at 5.45 P.M., to July 14, at 8.01 P.M. Attention was directed during the latter pe riod to special laboratory tests. Aside from these delays, and some minor ones incidental to the methods of operation of the system as arranged, operations were continuous during the day and night. INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 1W7. 345 Two minor rises of the river occurred dur ing this period, causing considerable varia tions in the amounts of suspended matter. The range was from 66 to 711 parts per mil lion, with an average of 320 parts. Investigations were mainly along the fol lowing lines: i. Comparison of efficiencies of sulphate of alumina and electrolytically decomposed iron, and a determination of their relative effi ciency. Other coagulants, including persul phate of iron, in which the free acid had been neutralized with caustic soda, copperas (pro- tosulphate of iron), alone and with caustic soda, and electrolytically decomposed alumi num, were also tried for short periods. The bulk of the work, however, was with the first two coagulants, 62 runs having been made with sulphate of alumina, and 26 with elec trolytically decomposed iron. It is to be noted that the uncertainties at tached to the early electrolytic work were re moved during the last of May and early part of June by the insertion df new insulating ma terials. On July 9 electrode No. i was re modeled to give double the treatment which was previously available. 2. Comparison of efficiencies of various pe riods of coagulation preceding nitration, and determination of minimum coagulant allow able with each. For this purpose the place of application of the last close of chemicals was changed from time to time to the follow ing points: Inlet to basin No. i; inlet to basin No. 2; inlet to the Jewell settling chamber; and the outlet of the Jewell settling chamber (top of filter). For the effective period of coagulation in the several basins reference is made to the description of these basins in the early part of this chapter. 3. Investigations of the rate of clearing of the effluent following a washing of the filter. This point was studied more or less through out the period under different rates of filtra tion, and with different amounts of coagulant. Several runs were made specially for this point, in which the filter was washed after the first 1000 cubic feet or so of water had been filtered. In order to carry on these studies, use was made of basins Nos. i and 2, and the Jewell settling chamber and filter, together with the devices necessary to prepare and apply the various coagulants. As the river water changed in composition several times, it was necessary to modify the arrangement from time to time during the period. River water was admitted directly to the Jewell settling chamber without any preliminary subsidence on runs Nos. 88 to 91, 97 to 105, 147 to 156, 164 to 167, and 174 to 185, inclusive. Basin No. i only was used on runs Nos. 157 to 163, and 1 68 to 173, inclusive. On the other runs, Nos. 92 to 96 and 106 to 146, inclusive, ba sins Nos. i and 2 were used. The principal points of significance in con nection with the various runs are given in serial order in the following list. After the list is the section containing the several tables showing the results of the operation of these devices. Notes on Special Features of these Runs. Nos. i to 7. Comparison of efficiency of per sulphate of iron and sulphate of alumina, using the original Jewell System, without agitation of surface of sand layer. Nos. 8 to 14. Same as Nos. i to 7, but the surface of the sand layer was agitated whenever practicable. \o. 15. First run with new devices, includ ing three settling basins, filter, and de vices for preparing and applying coagu lants. Iron plates of electrodes were new and bright. No. 1 8. The filter was run dry; that is, the sand layer was so heavily loaded with suspended matter that surface agitation neither affected the rate of filtration or the character of effluent. No. 19. Amount of applied chemicals was re duced during run, resulting shortly in failure in the character of the effluent. No. 20. First run with aluminum electrodes. Plates were new and bright. Nos. 20 to 23. Using aluminum electrodes. Effluent failed suddenly after from 2 to 4 hours filtration, apparently due to ac cumulations of gas within the sand layer. No. 24. Flectrode No. 3 found touching wall of the cell after a short run. This was remedied, and run continued. No. 26. Chemical pump broke down, and 346 WATER PURIFICATION AT LOUISVILLE.. lack of coagulant caused an early failure of the effluent. No. 27. Low voltage on cell No. 3 indicated a leakage of electric current. April 1 6. Repaired insulations of bolts and electrodes. No. 32. Amount of treatment varied three times during the run. Nos. 32 to 34. Resistance of electrodes Nos. i and 3 steadily falling, indicating stead ily increasing leakage of electric current. No. 42. Electrode No. 3 found short cir cuited after this run. April 25, 26. Removed iron bolts from elec trodes and put in wooden ones. Cleaned separators. No. 44. First run with repaired electrodes. Plates had been exposed to the air, and were therefore heavily rusted. No. 54. Period of service was shortened by closing work for the day. No. 70. This is the only run recorded in which the hydrate came through the sand layer into the filtered water so as to be plainly visible. May 30. Iron electrodes reassembled with iron bolts and hard rubber insulators, and put in service to study the rate of decomposition of the metal. June 3. Aluminum electrodes reassembled with iron bolts and hard rubber insu lators, and put in service to study the rate of decomposition of the metal. No. 91. Several changes in the rate of treat ment and filtration were made during this run. No. 112. Failed to drain settling basins after No. in. This run was therefore af fected by previous rate of treatment. No. 127. Shortened the run to work with copperas. Nos. 133 to 135. Used old aluminum elec trodes, which were very heavily coated with oxide. Direction of electric cur rent was reversed several times. Nos. 138, 139. Application of caustic soda re moved accumulations of organic matter in the sand layer, and carried them into the effluent. No. 147. Began to use remodeled electrode No. i. No. 149. Failure to supply coagulant caused an early failure of the effluent. No. 156. Period of service was shortened by closing work for the day. RESULTS ACCOMPLISHED BY THE DEVICES ARRANGED AND OPERATED BY THE WATER COMPANY. The results of chemical analyses of the river water after treatment by these methods and devices are presented in a series of tables, in which all of the data obtained with the use of each coagulant is presented separately. With regard to the collection and notation of samples, methods of analysis, and significance of results, reference is made to Chapter VIII, where the corresponding data of 1895-96 were presented. For detailed information concerning the composition of the untreated river water, the tables of analyses in Chap ter I may be consulted; and the amounts of suspended matter, and numbers of bacteria which on numerous occasions were deter mined in samples collected as the water left the several settling basins, are recorded in a subsequent portion of this section. With regard to features in the chemical composition of the effluents, as shown by special analyses, a full account of these mat ters will be found in the discussion of results, which is the closing portion of this chapter. At this place it may be briefly stated that in the case of sulphate of alumina, persulphate of iron, and protosulphate of iron, the in crease in carbonic acid and sulphate of lime (incrusting constituents) in the effluent, was proportional to the decrease in alkalinity of the river water by the respective treatments. With sulphate of alumina, and persulphate of iron, none of the applied chemicals in an un- decomposed form appeared in the filtered water, and there was no diminution in the oxygen dissolved in the water. But when copperas (protosulphate of iron) was used, the carbonic acid in the water retarded the oxidation of the ferrous compounds by the dissolved oxygen, so that when this chemical alone was used, some of the iron passed into the effluent. The use of caustic soda in con nection with copperas changed the carbonic INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 18H7. 347 acid to carbonate of soda, and, under these conditions, copperas could be satisfactorily applied as a coagulant, provided the amount applied was not in excess of that capable of oxidation by the constituents of the water. It caused a reduction for each grain per gal lon of about 0.5 part of dissolved oxygen, which was required in order to convert the iron into a completely insoluble form. In passing, it may be noted, that the application of caustic soda caused a marked removal of the organic matter accumulated on the sand grains of the filter, as shown by the analyses. In those cases where common alum crystals were used as a coagulant the normal free am monia in the effluents is estimated because the coagulant contained some ammonia. Concerning the special chemical features of the effluent obtained with electrolytic treat ment, there were no additional carbonic acid, incrusting constituents, or dissolved metal in the water as it left the filter under ordinary conditions. With aluminum electrodes this was true under all conditions; but with iron electrodes the decomposed iron had to be oxi dized by dissolved oxygen, in order to convert it into a completely insoluble form. As in the case of copperas, therefore, the electro lytic iron treatment could not be safely em ployed beyond a certain amount, which was limited by the oxygen dissolved in the water. Otherwise, dissolved iron would pass into the filtered water. Microscopical analyses were not made regularly, but from occasional examinations of the effluents, it may be stated that, as was the case during 1895-96, the effluents were practically free from microscopical organ isms. The results of the bacterial analyses of the effluent with each coagulant are presented in the set of tables following the chemical re sults, and are given in the same form as was used and explained in Chapter VIII. There are no special features worthy of comment in this connection, except perhaps to point out the fact that, repeatedly, turbid effluents were found to give a fairly satisfactory bac terial efficiency. In no case, however, was an admissible bacterial efficiency obtained when the coagulation of the water passing on to the sand layer was lacking to a marked degree. These analytical results are summarized by runs, together with a record of the kind, method of application, and quantity of co agulating teatment, and of quantities of water treated, with corresponding periods of time, at the close of this section. A full explana tion of this summary precedes it. WATER PURIF1CTA1ON AT LOUISVILLE. H a I -pspuadsns O O O HI -1- 1 o -to O O O O ->OO t^OOOCO OOC ooo -ooooooooooooooooooooooco Q O O * n ~ O O O~O O o o o o o o .000 OT)OOOOO"ifOOffw">O"~>OOOu"r>Ow^O i O O O O O CC |j cf-^ o r^ o HI 8o2g5o? Oc2 1 ^2^o > 1 S 2 > 2?l"ooo o8 UOJI D Q .. imn| y p3A[OSSIQ OOOOOOOOO 000 ^ OOOOOOO , OOOOOOOOOOO 7. < < O as h Xiiui|rji|v m^wm-l-cr. (X, SI C U ! c P^O^Q < 8^s?8.sss^ ^ SBs > -C?KS& >, j^RS sa ^ss g; < { papusdsns OOOOOOOOO W OO < OOOOOOO < -OOOOO OOOO a. ^ u u : <$ E 1 wi - 3 SSS ^ ^-cor.0^,. H Si & c , p,0...a a 1-1 "^ H H < c o - papusdsns y u : OOOOOOOOO O^O ^ OOOOOOO ^ -OOOOO OOOO Sj 1 I-:ox E " ^ u ONOM-OOI^C-O ONO |5 t C-0 MOD ^ 0-t-OT.O "ICO N - in H m & fA b 53]BJll>i > W "" S J U > "T" 1 Id T ,,. u,, V3 ,, H ^ id z - - w ooooooooo Q i^o 1-1 * H 2? o"o 0^020 ^O o oo ooooooo o P 1 fc s Sc is i P3AIOSS1CI OOOOOOOCO ^ o~" O^OOOO^Co -Z-S -COOOOOOOOO B J ! H fe : _LJ u, c. Z w papusdsns Ij ^ O td u H c (d s" 1 a! j t u. i ^ iwiox ^* "*" ~~ f" 1 D ooooboo o"o 3 o S 11 WQ o o oo o^o o" U o~ o" o" o" o"o S" o o o H D > D u p *" JUDu:>a xxo 6 d d d d d d 6 d C 4 - b o d d d d >- S 6 >-i c ~ >- - i H "" OOOO-niiOO [jj Ocii-< ^ : ] o M 2 o o o O frj JH-r-i-iOi- O -tt- OO HUES ,) JO 33JHJ U Id en > J ?n?E!^ X Z O o OF CHEMICAL ANfl C U E II U u > (, > S, S!!S" W 3 ^o ^> o t^ X ^g- g ""> U ESULTS or^ocii-," O >. O nt^Mtn 02 "a. 3 3 D c.: : : j?jj 3D a. : - - - c - - 3 - - 1-1 w "* S. IS "* oJ < o; J. 1X .-MS silK!^!? if? Is^ltssl I^IIOlss 1 ^ 35 W A TER PURIFICATION AT LOUISVILLE. RESULTS OF BACTERIAL ANALYSES OF THE EFFLUENT OF THE JEWELL FILTER WITH SULPHATE OF ALUMINA. Rate of j Collected. Filtration. c .y _ " * iT~ Period of u b 3 1 Number a J - & ServiceSinci I ast S" ^ fe e Run. 8 a (3 b i Washing. ^K . o." Remarks __ p X Bute. j Hour. t! :l o Hours and Minutes. t~ v-- c Is 8.? i =3 5 Su c/i u s ii tL. CQ 1897 5260 April 5 12.30 P.M. i 24.0 97 3-0 oh. 57m. I 442 2 7 8 5262 5 4.00 " 2 24.0 97 2.8 oh. 45m. I 068 219 5263 " 5 4-30 " 2 24.0 97 3.0 ill. 15111. I SoS 290 52^4 " 5 5-00 " 2 24.0 97 3.3 ih. 4501. 2478 3 I 5265 5 5-30 " 2 23.5 95 3.8 2h. ism. 3178 292 5267 5 S.oo " 3 23.5 95 3.1 oh. 06111. 154 35 5268 5 <).30 " 3 23.5 95 3.6 ih. 36m. 2 284 620 5270 " 5 11.30 " 3 23-5 ! 95 6.5 3h. 36m. 5009 4IO 5284 6 11.00 P.M. S 24.0 97 5 .0 ih. i6m. 8853 78 5285 6 11.30 " 8 23.5 95 6.8 ih. 46111 2423 42 Agitated surface of sand layer at 5287 7 I2.3O P.M. 8 23.5 95 3.9 2h. 41111. 3853 185 11.49 I .M- 5288 7 1.30 " 8 23.5 95 4.8 3h. 41111. 5 3 )3 350 5289 7 3.OO " 8 23.5 95 7. i 5h. urn. 7303 322 5290 " 7 5 . oo " 9 23.5 ! 95 3.0 oh. 25m. 643 545 5291 7 6.00 " 9 23.5 95 4.0 ih. 25m. 2053 i 250 Closed outlet at 6.20 A.M. 53M " 8 12.00 M. 12 23.5 95 5.5 ih. 35m. 2 248 IOO 53if> 8 3.00 P.M. 13 23.5 95 3.8 ih.45m. 2 619 780 5318 8 4.OO " 13 23.0 93 9.1 2h. 45m. 3 895 355 5320 " 8 4.21 " 13 23.5 95 . . . . 3h. 0401. 4379 5 78o|Agitated surface of sand layer at 5321 8 4.23 " 13 23.5 95 3h. o6m. 4429 2 500 4.15 P.M. 5322 8 4.25 " 13 23.5 95 3h. oSm. 4469 950 5323 " 8 4.27 " 13 23.5 95 3h. loin. 4519 405 5324 " 8 4.29 " 13 23.5 95 3h. I2m. 4 569 337 5325 " 9 0.00 A.M. 13 23.5 95 5-3 5h. i6m. 6246 179 5327 9 I . OO " 13 23-5 95 6.0 6h. i6m. 8429 870 5329 9 1.50 " 13 23.5 95 7h. o6m. 9 749 780 Agitated surface of sand layer at 5330 9 1.57 " 13 23-5 95 7h. o8m. 9799 i Soo 11.50 A.M. 5331 9 1.59 " 13 23.5 95 7h. lom. 9839 2 510 5332 9 2. OI P.M. 13 23-5 95 7h. I2m. 9889 2 ISO 5333 9 2.03 " 13 23.5 95 7h. 1401. 9929 I 890 5334 9 2.06 " 13 230 95 7h. I7m. 9999 I 680 5335 9 2.09 " 13 23.5 95 7h. 2om. 10066 I 590 5336 9 2.12 " 13 23-5 95 7h. 23111. 10 131 I 2OO 5337 9 2.15 " 13 23-5 95 7h. 26m. 10 200 I 170 539S 13 g.OO A.M. 19 23-5 95 s s 2h. 57m. 3 756 I 275 5399 13 O.OO " 19 24.0 97 4.1 3h. 57m. 5 176 9910! 5400 13 1. 00 " 19 23-5 95 4-9 4h. 57m. 6 556 410 5401 13 2.00 M. 19 23.0 93 5-2 5h. 57m. 7916 1650 5402 13 I.OO P.M. 19 23.0 93 6.4 6h. 57111. 9306 2 690 543 13 2.00 " 19 23.0 93 6.9 7h. 57m. 10 676 I 27O 5405 13 3.00 " 19 23.0 93 7-0 8h. 57m. ii 986 I 350 5435 15 9,30 A.M. 26 23.5 95 3-5 2h. i6m. 3241 59 543f> 15 0.30 " 26 23-5 95 4-0 3h. i6m. 4 641 i 020 5437 15 1 . 3O " 26 23.0 93 4.6 4h. oim. 5 691! i 080 5443 15 0.30 P.M. 28 23.5 95 2-3 oh. 45111. 955 204 5444 " 5 I. 3 " 28 23.5 95 3-5 ih. 45m. 2345 142 5445 " 16 2.3O A.M. 28 23-5 95 4.0 2h. 45m. 3755 471 5447 " 16 3.00 " 28 23-5 95 3-1 ih. 38m. 2329 269 5448 " 16 5-00 " 30 23-5 95 3-i oh. 38111. i 031 68 5449 " 16 6.00 " 30 23.0 93 3-0 ih. 38111. 2 401 49 5499 " 22 9-3 " 36 23.5 95 3-o ih. 33m. 2 099 61 5500 " 22 10.30 " 36 23-5 95 3-4 2h. 33111. 3499 i 450 5501 " 22 11.30 " 36 23-5 95 3.8 3h. 33111. 4959 37 5502 " 22 I2.3O P.M. 36 23 5 95 4.0 4h. 33m. 6349 68 553 " 22 I. 3 " 36 23-5 95 4-3 5h. 33m. 7659 55 5504 " 22 2.30 " 36 23-5 95 4.8 6h. 33m. S 859 165 5509 " 22 3-30 " 36 23.0 93 5-0 7h. 33111. 10269 335 5533 " 24 g.OO A.M. 43 23-5 95 3-7 2h. 40m. 3 521 24 5534 24 10.00 " 43 23-5 95 3-9 3h. 4om. 4 881 26 5535 " 24 I I.OO " 43 23.0 93 4-2 4h. 40111. 6 221 26 5536 24 12. OO M. 43 23-5 95 4-9 5h. 4om. 7 611 21 5537 " 24 I.OO P.M. 43 23-5 95 5-0 6h. 4om.i 9011 59 INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, IS .il. 351 RESULTS OF BACTERIAL ANALYSES WITH SULPHATE OF -ALUMINA. Continued. 1 R-il .. I S 1 Collected. Filtration. t IS 1 Number s. JsL d Period of Service Since at-- u u . z of Run. I,; 3 t; 1 X Last SsarTd m S Remarks. Date. Hour. o c c< i "3 Minutes. u tr2 aj C E la = g. J 3 = JO |(5 u s >s u. 03 1897 5538 April 24 2.30 P.M. 43 23-5 95 6.0 8h. iom. II OOI 242 Closed outlet at 2-44 P.M. 5589 " 29 4.00 A.M. 50 23.0 93 oh. 3gm. 764 43 5591 " 29 5.00 " 50 23.0 93 ih. 39m. 2 024 44 5592 29 6.00 " 50 23-0 93 3.0 2h. 39111. 3424 23 5593 " 29 9.00 50 23-5 95 4-5 5h. sgm. 8 664 24 5595 " 29 10.00 " 50 23-5 95 5-2 6h. sgm. 9 044 22 5596 29 II.OO " 50 23-5 95 6.2 7h. 3gm. 10454 45 5600 " 29 12. OO M. 50 23-5 95 7-5 8h. 3gm. 11907 107 Agitated surface of sand layer at 5601 " 29 1. 00 I .M. 50 23-5 95 6.0 gh. 37m. 13397 H7 12.24 ! - M - 5602! " 29 2.OO " 50 23.0 93 6.6 toh. 37m. i 14747 173 5603 29 3.OO " 50 23-5 95 7-7 iih. 35111. 16 120 510 Closed outlet at 3.08 P.M. 5762 May 7 9.00 " 59 23-5 95 2-7 oh. 48m. I 158 46 5763 7 10.00 " 59 23-5 95 3.0 in. 48m. 2588 M9 7 II.OO " 59 23-5 95 3.2 2h. 48171. 3968 71 5768 7 12.00 " 59 23-5 95 3.6 3h. 4Sm. 5318 123 5773 8 4.00 A.M. 60 23-5 95 2.6 oh. 57m. I 369 .89 5774 8 5.00 " 60 23 . 5 95 3.0 ih. 57m. 2 769 58 578o " 8 9.30 " 61 23-5 95 3.5 2h. 3om. 3510 76 578i 5782 8 8 IO.OO II.OO 6 1 61 23-5 23-5 95 95 3-8 3-9 3h. oom. 4h. oom. 4220 311 5660 116 Closed outlet at 11.04 A.M. 5783 8 I2.OO M. 62 23-5 95 3-0 oh. 33m. 747 18 5784 8 1. 00 P.M. 62 23-5 95 Z.6 ih. 33111. 2 197 89 5785 5786 8 8 2.OO " 2.3O 62 62 23-5 23.0 95 93 3-4 3-7 2h. 33m. 3h. 0301. 3577 "o 4 167 129 Closed outlet at 2.35 P.M. 5787 8 3-05 " 63 23-5 95 oh. 05m. 119 122 5788 " 8 3.08 " 63 23-5 95 oh. iom. 239 37 5789 8 3-13 " 63 23-5 95 oh. ism. 359 57 5794 8 4-3" " 63 23.0 93 2.6 ih. 32m. 2 089 .22 5795 8 5-3" " 63 23-5 95 3-0 2h. 32m. 3 54< 105 b796 " 8 8.(X> " 64 23.5 95 2.8 oh. 2Sm. f 7 1 191 5797 8 9.00 64 23-5 95 3-4 ih. 28m. 2 051 142 5799 8 II.OO " 65 23-5 95 2 5 oh. 27m. 735 189 5800 " 8 12. OO " 65 23-5 95 2-7 ih. 27m. 2055 74 5801 9 1. 00 A.M. 65 23-5 95 2-9 2h. 27m. 3445 142 5802 9 1.30 " 65 23.? 95 3.0 2h. 57m. 4 I5 92 Closed outlet at 1.43 A.M. 5803 9 3-3" " 66 23-5 95 2-5 oh. oim. 23 4i 5804 9 4-3" " 66 23-5 95 2.7 ih. oim. I 433 45 5964 19 II.oo I .M. 67 23.5 95 2.1 oh. 56m. t2 4 6 272 5968 19 12. OO " f 7 23.5 95 2.8 ih. 56m 2 606 196 5969 " 20 I.(X) A.M. f>7 23.5 95 2-9 2h. 56m. 3966 490 5971 " 20 3-OO 68 12. 48 0.8 oh. 33m. 381 274 5972 " 20 4-00 " 68 12.0 48 0.9 ih. 33m. I 181 M4 5973 " 20 5.00 " 68 12.0 48 1.0 2h. 33m. I 821 139 5974 2O 6.00 " 68 12.0 48 I .2 3". 33m- 2 621 I 080 5976 " 2O 9.00 " 68 12. O 48 i.S 6h. 33m. 4891 3900 5977 " 20 II.OO " 68 12. O 48 1.9 8h. 33m. 6381 244 5978 " 20 1. 00 P.M. 68 12.0 48 2.2 loh. 33m. 7841 308 5980 " 20 3.OO 68 12.0 48 2.8 I2h. 33m. 9381 395 " 20 5.30 " 68 12. 48 3-1 I 4 h. 53tn. II 26l 165 5985 ,, 20 8.00 " 68 12.0 48 4.1 i?h. 33m. 13 221 230 ,, 20 9.00 " 68 12.0 48 4-4 l8h. 33m. I395I 343 jggE ,, 20 II.OO " 69 23-5 95 2.5 oh. 48m. I 080 401 5989 ., 20 12. 00 " 69 23.5 95 2.6 ih. 48m. 2460 4500 5990 " 21 1. 00 A.M. 69 23.5 95 2.8 2h. 48m. 3810 307 5991 " 21 2.OO " 69 23-5 95 3-0 3(1. 48m. 5 180 635 5993 " 21 3-oo " 69 23-5 95 3-2 4h. 48m. 6680 I 760 5994 " 21 4.(X> " 69 23-5 95 3-5 5h. 48m. 7950 215 5995 " 21 4-3" " 69 23-5 95 3-7 6h. l8m. 86 3 c 375 Closed outlet at 4.41 A.M. 5997 " 21 I0.3 J P.M. 7 12.0 48 I.O oh. 3801. 480 417 5998 " 21 12. (X) " 70 12. P 1.0 2h. o8m. I 480 397 5999 " 22 I.OO A.M. 7" 12. r 1. 1 3h. o8m. t 400 259 6000 " 22 2.OO " 70 12.0 48 I .2 4h. o8m. 3 u" 460 6002 " 22 3.OO " 70 12.0 48 1-5 5h. o8m. 3860 257 6006 " 22 5.OO " 70 12. r 1.8 7h. o8m. 5 380 423 352 WATER PURIFICATION AT LOUISVILLE. RESULTS OF BACTERIAL ANALYSES WITH SULPHATE OF ALUMINA. Conti Rate of {J % Collected. Filtration. ^ -H |~u Period of ^ a .3 OJ Number g. a T3 S . rviceSince i-- j u a P x Run. a v t :j -3 j C 3 ^ < I Washing. &" Hours and -a = o S.U Remarks. Date. Hour. U I Minutes. 3.0 <D - ~ II ^ a? 1 =33 ^U I/I O S - it n 1897 6008 May 22 9.00 A.M. 70 I2.O 4S 3-2 nh. o8m. 7 943 241 6009 " 22 I2.OO M. 70 12.0 48 5-5 I4h. oSm. I 015 380 6011 " 22 3.00 P.M. 70 12.0 .(S 8.2 I7h. oSm. 12 303 362 6012 " 22 5-30 " 70 12.0 48 il. 7 igh. 38111. 14203 546 Agitated surface of sand layer at 6013 " 22 5-45 70 12.0 48 3-o igh. 4gm. 14403 635 5.30 P.M. 6014 " 22 7.00 " 70 12.0 48 3-2 2ih. 0401. 15 543 472 6016 " 22 9.00 " 70 12.0 48 3-9 23!!. 04m. 17 013 498 6020 " 22 11.00 " 70 12.0 48 4.0 25h. 04111. 18 513 578 6021 " 23 I.OO A.M. 7O 12.0 48 4-5 27h. O4m. 20013 497 6023 " 23 3.00 " 70 12.0 48 5-1 2i)h. 04tn. 21 543 262 6024 " 23 6.00 " 70 12.0 48* 6.0 32h. 04111. 23643 Sio 6025 23 7.00 " 70 12.0 48 6.1 33h. 04111. 24433 750 f)027 23 9.00 " 70 12.0 48 6 5 35h. 04m. 25643 995 6028 23 12.30 P.M. 70 12. O 48 S.i 39h. 04111. 28 223 I 270 6030 23 3.00 " 70 12.0 48 9. i 41 h. 04111. 30093 610 603 1 23 4.30 " 70 12.0 48 10. 1 42h 34111. 31 203 I OOO Agitated surface of sand layer at 6032 " 23 6.30 " 71 23-5 95 2.0, oh. 3om. 720 2 290 4.30 P.M. Closed outlet at 4.45P.M. 6033 " 23 7.00 " 71 23.5 95 2.2 ih. oom. I 410 I 890 Closed outlet at 7.11 P.M. 6034 23 9.00 " 72 23-5 95 2.1, oh. 43111. 935 I 020 6036 " 23 10.00 " 72 23-5 95 2.3 ih. 43111. 2 275 26l 6037 " 24 12.30 A.M. 72 23-5 95 2.8 4h. 1301. 5645 242 6038 " 24 2.01) " 72 23-5 95 3.4 sh. 43m- 7635 4 lS 6040 24 3.00 " 72 23.5 95 3.6 6h. 43111. 9005 725 6041 24 4.00 " 72 23-5 95 3.9 7h. 43.11. 10 565 460 6042 24 6.00 " 72 23-5 95 4.8: gh. 43m. 13075 185 6044 24 9.00 " 72 23-5 95 6. Sj I2h. 43m. 17 135 235 Closed outlet at 9.10 A.M. 48 ... i I 26o 6047 24 24 9-3 r.M. 10.00 " I J 73 2.0 48 I . i ih. lorn. 480 2 250 " 10.24 t - M - 6048 " 25 I.OO A.M. 74 2.0 48 i.o oh. 32111. 362 4 5o 6049 " 25 2.00 " 74 2.0 48 1 .0 ih. 32111. I 062 2 95" Closed outlet at 2.18 A.M. 6051 " 25 5.oo " 75 2.O 48 I.O oh. 52m. 479 54" 6056 " 25 6.00 75 2.0 48 I.I ih. 52m. i 219 235 6058 " 25 9.00 " 75 2.0 48 I.f) 4h. 52m. 3 559 260 0059 25 II. oo " 75 2.0 48 ^ 1.8 6h. 5201. 5 159 125 6060 " 25 I.OO P.M. 75 2.0 48 2.0 8h. 52m. 6649 263 6062 " 25 3.00 " 75 2.0 48 2.8 loh. 5201. 8 IK) 375 6063 " 25 5.3 " 75 2.0 48 3-3 I3h. 22m. 9 949 342 6064 " 25 7.30 " 75 12.0 48 4.2 I5h. 22m. II 449 495 Closed outlet at 7.30 P.M. 6066 " 25 0. 10 " 76 12. O 48 .... oh. ism. 174 595 6067 " 25 0.15 " 76 12.0 48 .... oh. 2om. 234 54" 6068 " 25 o. 20 " 76 12. O 48 .... oh. 25m. 295 520 6069 25 0.25 76 12. O 48 .... oh. 3om. 354 391 6070 " 25 2.00 " 77 9.0 34 oh. 2901. 284 420 6071 " 26 2.OO A.M. 78 23-5 95 2.1 oh. 45m. 996 223 6072, " 26 2.36 " 79 23-5 95 2.8 oh. 05111. 88 419 6073 11 26 2.41 " 79 23-5 95 oh. lorn. 208 435 6074 " 26 2.46 " 79 23-5 95 oh. 15111. 308 352 6075 " 26 2.51 " 79 23-5 95 30 oh. 2om. 418 315 6077 " 26 3.15 79 23.5 95 oh. 44111. 95S 109 Closed outlet at 3.25 A.M. 6078 " 26 5-45 So 23-5 95 oh. iSm. 393 497 6079 " 26 6.00 " 80 23-5 95 2.1 oh. 33m. 713 STO 6081 " 26 9.00 " 8 23-5 95 2.2 oh. 3om. 697 301 6082 " 26 10.00 " 8 23.5 95 2.6 Ih. 3Om. 2 067 35(> 6083 " 2& I I.OO " 8 23.5 95 2.8 2h. 3om. 3477 339 6084 " 26 12.00 M. S 23-5 95 3-o 3h. 3m. 4847 268 6085 " 26 2.00 P.M. 8 23.0 93 3-5 5h. 30m. 7597 229 6087 " 26 3-30 " 8 23.0 93 3-8 7h. oom. 9657 235 6088 " 26 5-3 " 8 23-5 95 4.8 9h. oom.] 12 497 319 6089 " 26 7.00 " 8 23.5 95 5-4 i oh. 3om. M397 216 6090 6095 " 26 " 26 8.30 " 10.52 " 8 8 23-5 23.0 95 93 6.4 I2h. oom. I4h. 22m. 16517 19597 304 338 Agitated surface of sand layer at 6096 " 26 11.05 " 8 23.5 95 I4h. 3im. 19747 432 10.53 i -M. 6097 " 26 II. 10 " 8 23-5 95 5.0 I4h. 36m. 19847 263 6098 " 27 I.OO A.M. 8 23-5 95 6.0 i6h. 26m. 22327 311 INVESTIGATIONS OF THE WATKR COMPANY FROM APRIL TO JULY, 1807. RESULTS OF BACTERIAL AN ALYSES WITH SULPHATE OF ALUMINA. Continued. Rat eof *j - C ollccted. Filtr. fc, Ji 2 ^ u e 4j Period of u ^ T3 Number 8. = a . i Last 3-= | < i$ Remarks 3 X Hate. Hour. Run. ?i 5 Ss - o 5 ..E Washing. Hours and Minutes. ill II P C ^s ^ a cT $ ^JO su tn u s j b. o 1897 6100 May 27 3.00 A.M. Si 23.5 95 7-o i8h. 26m. 25057 382 6101 " 27 4.05 " 82 23-5 95 oh. o7m. 145 522 6l(j2 " 27 4.10 " 82 23-5 95 Oh. I2IT1. 245 575 6103 " 27 4.22 " 82 23-5 95 oh. 24m. 545 499 6104 " 27 4-43 82 23-5 95 oh. 45m. I 045 370 Closed outlet at 4.44 A.M. 6105 " 27 5-23 83 12.0 48 oh. lorn. 128 411 6106 27 5-34 " 83 12.0 48 oh. 21 m. 235 298 6107 " 27 5-54 S3 12." 48 oh. 4im. 498 453 6109 27 9-05 84 12.0 48 oh. lorn. 150 590 6110 " 27 9.18 84 12.0 48 oh. 23m. 2gS 382 6jn " 27 9-35 " 84 12. 48 oh. 4om. 498 272 6112 27 10.16 " 84 12.0 48 ih. 2im. 998 197 1113 " 27 10.24 " 84 23-5 95 ih. 29111. i 168 193 6114 " 27 II. OO " 84 23.5 95 2h. 05111. 2018 227 6115 " 27 12. OO M. 84 23.5 95 3h. osm. 3388 215 6116 " 27 1. 00 P.M. 84 23-5 95 4h. 0501. 4788 161 6117 " 27 2.00 " 84 23.5 95 5h. osm. 6188 305 6119 27 3-00 " 84 23-5 95 5- ) 6h. osm. 7558 167 6120 " 27 4.00 " 84 12.5 50 3-7 7h. osm. 8368 162 6121 27 5.00 84 12.0 48 4-3 8h . osm. 9098 95 6122 27 6. oo " 84 12.0 48 5-" gh. 05111. 9838 164 Closed outlet at 6.00 P.M. 6123 27 8.14 " 85 12.0 48 Oh. I2IT1. 150 529 6124 27 8.22 " 85 12. O 48 oh. 2om. 250 342 6125 " 27 8-43 " 85 12.0 48 oh. 4im. 530 302 6126 27 9- 5 " 85 12. O 48 Ih. I3m. 950 233 6127 27 9-25 85 23-5 95 ih. 23m. i 150 269 6128 27 9.30 " 85 23.5 95 ih. 28m. i 250 371 6129 " 27 9-35 85 23.5 95 Ih. 33m. i 380 337 Closed outlet at 9.45 P.M. 6131 " 27 0.17 86 23.5 95 oh. o6m. 138 480 6132 " 27 O.23 " 86 23.5 95 oh. I2m. 258 295 6133 " 27 0.32 " 86 23-5 95 oh. 2im. 498 3" " " " 10.46 P.M. 6134 " 28 2.43 A.M. 87 23.5 95 oh. o8m. 145 228 6135 28 2. 4 S " 87 23.5 95 oh. 1301. 245 263 6136 " 28 2.58 " 87 23-5 95 oh. 23m. 495 H7 6138 28 3.00 " 87 23-5 95 2.5 2h. 25m. 3225 89 6142 " 28 4-30 " 87 23.5 95 2.8 3h. 55m. 5305 231 6143 " 28 6. c " 87 23-5 95 3-5 5h. 2501. 7335 104 6144 " 28 8.00 " 87 23.5 95 4.0 7h. 25m. 0065 510 6146 " 28 9.00 " 87 23.0 93 4-7 8h. 25m. 1435 188 6147 " 28 10.00 " 87 23.0 93 5-4 gh. 25m. 2805 222 6148 28 II. OO " 87 23-5 95 6.4 loh. 25m. 4185 97 6149 " 28 11.30 " 87 2 3-5 95 6.9 loh. 55m. 4885 254 Closed outlet at 11.34 A.M. 6520 June 20 4.53 P.M. 92 23.0 93 oh. osm. 114 39 6 6521 " 20 4.58 " 92 20.0 81 oh. lorn. 211 4OO 6522 " 20 5.03 " 92 24.0 97 oh. ism. 309 388 6523 " 20 5.13 " 92 23-5 95 oh. 25m. 564 295 * 6524 " 2O 6. oo " 92 23.5 95 3-0 ih. I2m. I 574 357 6525 " 2O 7.00 " 92 23-5 95 3-8 2h. I2m. 2904 57 6526 " 20 S.oo " 92 23-5 95 4-7 3h. I2m. 4254 247 Closed outlet at S.io P.M. 6527 " 20 1.45 " 93 23-5 95 oh. osm. in 198 6528 " 20 1.53 " 93 23.5 95 oh. I3m. 3" 153 6529 " 20 2.OO " 93 23-5 95 2.2 oh. 2om. 501 130 6530 " 21 2.10 A.M. 93 23.5 95 oh. 3om. 701 184 6532 " 21 2.30 " 93 23-5 95 2-5 oh. sotn. I 161 117 6533 " 21 1.30 " 93 23--) 95 3-o ih. som. i 541 138 6534 " 21 2.30 " 93 23.5 95 3-7 2h. som. 3861 345 6536 " 21 3.30 " 93 23-5 95 4-7 3h. som. 5 231 945 Closed outlet at 3.45 A.M. = >7 " 21 5.2" " 94 23-5 95 oh. 07m. 184 "5 - i " 21 5.30 " 94 23.5 95 2-3 oh. I7m. 434 138 6539 " 21 6.0" " 94 23-5 95 2-5 oh. 47m. I 214 76 (,= |i 21 9.00 " 94 22.0 3-0 3h. 47m. 5 304 94 6542 " 21 IO.OO " 94 L d.o 105 4.0 4h. 47m. 6654 155 6543 " 21 11.00 " 94 23-5 95 4.0 5h. 47m. 8054 425 6544 " 21 12. OO " 94 23.5 95 4-5 6h. 47m. 9444 395 354 WATER PURIFICATION AT LOUISVILLE. RESULTS OF BACTERIAL ANALYSES WITH SULPHATE OF ALUMINA. Continued. Rate <>i ~ Collected. Filtration. a b c IS J3 Number a o a a Period of ServiceSince o3 ^ u u . G jjj is : X Last Washing. j ks Remarks. if Date. Hour. ^ 3 lul Hours and Minutes. |s| il U = a ? 2 b cc 1897 6545 June 21 I.OO P.M. 94 23-5 95 4.8 7h. 47m. 10854 315 6546 " 21 2.00 " 94 23-5 95 5-1 8h. 47m. 12254 6548 " 21 3-oo " 94 23-5 95 5-7 gh. 47m. 13644 287 6549 " 21 4.30 " 94 23-5 95 6.0 Iih. I7m. 15 754 6550 " 21 5.00 " 94 23-5 95 6-3 nh. 47m. 16424 328 6551 " 21 8.03 " 95 23-5 95 2.2 oh. O5m. 126 162 6552 " 21 8. 10 " 95 23-5 95 oh. I2m. 266 156 6554 " 21 8-30 " 95 23-5 95 2.8 oh. 32m. 736 126 6555 " 21 8.40 " 95 23-5 95 oh. 42m. 986 105 Closed outlet at 8.41 P.M. 6556 " 22 12.52 A.M. 96 23-5 95 oh. osm. 103 198 6557 " 22 I.OO " 96 23-5 95 2-5 oh. I3m. 343 218 6558 " 22 I. 10 " 96 23.5 95 oh. 23m. 53 253 6559 " 22 1.20 " 96 23-5 95 oh. 33m. 733 290 6560 " 22 1.30 " 96 23-5 95 2.6 oh. 43m. I 023 3M Closed outlet at 1.34 A.M. 6562 " 22 4-45 97 23-5 95 oh. 05m. 95 39 6563 " 22 4.53 97 23-5 95 oh. I3m. 275 495 6564 " 22 5.00 " 97 23-5 95 2.5 oh. 2om. 475 415 6565 " 22 5-3 " 97 23-5 95 2.5 oh. som. I 165 157 6566 " 22 6.00 " 97 23-5 95 2.6 ih. 2om. i 865 305 Closed outlet at 6.11 A.M. 6568 " 22 9. 10 " 98 24.0 97 oh. osm. 140 146 6569 " 22 9.15 " 98 23-5 95 oh. lorn. 260 Si 6570 " 22 10.00 " 98 24.0 97 2.7 oh. 55m. I 380 79 6571 " 22 11.00 " 98 24.0 97 3-o ih. 55m. 2 82O 128 6572 " 22 12. OO " 98 23-5 95 3-2 2h. 55m. 4240 149 6573 " 22 I.OO P.M. 98 23-5 95 3-6 3h. 55m. 5640 194 6574 " 22 1. 10 " 98 23-5 95 4h. osm. 5870 241 Closed outlet at 1.12 P.M. 6575 " 22 3-13 " 99 oh. osm. 148 35 6576 " 22 3.18 " 99 20. o 81 oh. lorn. 253 265 >577 " 22 3-23 " 99 23-5 95 oh. ism. 378 33 6579 " 22 3-38 " 99 23-5 95 oh. 3001. 738 189 6580 " 22 4.00 " 99 23-5 95 2.6 oh. 4401. 1068 233 6581 " 22 5-3 " 99 23-5 95 4.0 2h. 14111. 3098 82 6582 " 23 1.30 A.M. 101 23-5 95 2-5 oh. o6m. 168 269 6583 " 23 1-35 " 101 23-5 95 oh. nm. 248 265 6584 ! 2 3 1.40 " 101 24.0 97 oh. i6m. 368 255 6585 1.50 " 101 23-5 95 oh. 26m. 578 1 80 6586 " 23 2.00 " 101 23-5 95 2-7 oh. 3601. 888 173 Closed outlet at 2.07 A.M. 6588 " 23 4.17 102 23-5 95 oh. osm. 119 297 6589 " 23 4-23 " 102 20.0 81 oh. nm. 239 415 6590 " 23 4.30 " 102 23-5 95 2. I oh. i8m. 419 274 6591 : 23 5.00 " IO2 23-5 95 2.4 oh. 4Sm. 1099 295 6592 " 23 5-30 " 102 23-5 95 2.6 ih. i8m. 1789 272 "593 " 23 6.00 " 102 23-5 95 2-7 ih. 4sm. 2469 268 G 594 " 23 8.00- " 102 23-5 95 3-0 3h. lorn. 4379 261 Closed outlet at 8.10 A.M. 6595 ;< 23 2.40 P.M. 103 oh. osm. I2O TC7 6596 " 23 2-45 " 103 24.0 97 oh. lorn. 240 136 6597 " 23 2.50 " 103 23-5 95 oh. ism. 370 142 6599 " 23 3.00 " 103 23-5 95 2-7 oh. 25m. 630 161 6600 " 23 3.30 103 23-5 95 2.8 oh. 55m. I 360 187 6601 " 23 4-35 103 23-5 95 2h. oom. 2 860 272 Closed outlet at 4.35 P.M. 6602 " 23 6.00 " 104 23-5 95 2.0 oh. i6m. 374 153 6603 " 23 8.00 " 104 23-5 95 2-9 2h. I4m. 3274 45 6605 23 8.30 " 104 23-.S 95 2.9 2h. 44m. 4014 41 6606 23 9.30 " 104 23 5 95 3-1 3h. 44m. 5434 86 6607 23 10.30 " 104 23-5 95 3-3 4h. 44m. 6854 181 6608 23 11.30 " 104 23-5 95 3-5 5h. 44m. 8 284 212 Closed outlet at 11.35 P.M. 6628 24 2.13 " 107 20.0 81 oh. osm. 133 230 6629 24 2.18 " 107 21.0 85 oh. lorn. 273 l6q 6630 24 2.23 " 107 21. O 85 oh. ism. 378 147 6632 24 3 . oo " 107 2O. O 81 2.0 oh. 52m. I 168 151 6633 24 4.00 " 107 20.0 81 2.2 ih. 52m. 2388 825 Closed outlet at 4.10 P.M. 6637 24 9.11 109 18.0 73 oh. osm. 76 605 6638 24 9.16 " 109 18.0 73 oh. lorn. 166 625 6639 " 24 9-3 " 109 18.0 73 oh. 24111. 406 580 Closed outlet at 9.34 P.M. INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 1807. 355 RESULTS OF BACTERIAL AN ALYSES WITH SULPHATE OF ALUMINA. Continued. k.,i, ol ~ 8 Collected. Filtration. ii. in . - t Number I i& a Period of u ** 3 (j . E iz Date. Hour of i 5 un. ; 2 Hi K Washing. Hours and Minutes. ^|^ " *".5 t| Remarks. 2 || = 0.3" Z -23 rtW & u S, J ta n 1897 6687 June 26 9.30 P.M. 116 23-5 95 2.2 oh. o6m. 145 127 6688 " 26 9-35 " 116 23-5 95 oh. nm. 245 73 6689 " 26 9-45 " 116 23.5 95 oh. i6m. 555 81 6690 " 26 o.oo " 116 23.5 95 2-3 oh. 3&m. 865 39 6691 " 26 O.2O " 116 23.5 95 oh. 56111. i 295 89 Closed outlet at 10.21 P.M. 6692 " 27 2.20 A.M. "7 23.5 95 oh. O7m. 207 86 6693 " 27 2.30 " "7 23-5 95 2.2 oh. I7m. 447 51 6694 " 27 2.45 "7 23.5 95 oh. 32m. 745 77 6695 " 27 1. 00 " "7 23.5 95 2.0 oh. 47m. i 147 99 Closed outlet at 1.04 A.M. 6697 " 27 3.00 " nS 23.5 95 2.O oh. O5m. 1=6 580 6698 " 27 3.15 " 118 23.5 95 oh. 2Om. 456 3800 Closed outlet at 3.15 A.M. 6726 " 28 3.50 124 ; 23.5 95 oh. 05111. 122 6727 " 28 3-55 " 124 23.5 95 oh. lorn. 242 <>--- 6729 " 28 " 28 4.00 " 4-30 " 124 124 23-5 23-5 95 95 2-5 oh. i5m. oh. 45m. 342 I 052 880 Closed outlet at 4.30 A.M. 6730 " 28 8.00 " 125 23.5 95 2.5 oh. 54m. I 263 29 6732 " 28 9.00 " 125 23-5 95 2.3^ ih. 54tn. 2663 26 6733 " 28 0.30 " 125 23.5 95 3.0 1 3h. 24m. 4763 20 6734 " 28 2.00 M. 125 23.5 95 3.7 4h- 54m. 6853 32 6735 " 28 2. GO P.M. 125 23.5 95 3.9 6h. 54tn. 9653 39 6737 " 29 0.47 126 23-5 95 oh. osm. 152 372 6738 " 29 0.51 126 23.5 95 oh. urn. 252 163 6739 " 29 1.02 " 126 23.5 95 oh. 22m. 522 204 6740 " 29 1.24 " 126 23-5 95 oh. 44m. 992 256 6741 " 30 2.30 A.M. 126 23-5 95 2.6 ih. 5001. 2 532 269 6742 " 3 I 30 " 126 23.5 95 2-7 2h. 5om. 3882 i 250 >7t i " 3 2.30 " 126 23.5 95 3-o 3h. som. 5 262 9i 6745 " 3 3-30 " 126 23-5 95 3-4 4h. 50tn. 6682 49 6746 " 30 4.30 " 126 23-5 95 3-5 5h. som. 8 112 92 6747 " 30 5.30 " 126 23-5 95 3-7 6h. 5om. 9492 94 6749 " 3 9-00 " 127 23-5 95 2. I oh. 5im. I 232 127 6750 " 3 9.30 " 127 23.5 95 ih. 2im. I 891 26 Closed outlet at 9.35 A.M. 6763 " 30 8.IO P.M. 130 23-5 95 oh. o6m. 150 119 6764 " 3 8.15 " 130 23-5 95 oh. Iim. 250 69 6765 " 3 8.25 " 130 23-5 95 oh. 2im. 500 59 6766 3 8.45 " 130 23.5 95 oh. 4lm. I OOO 84 6771 " 3 IO.OO " 130 23.5 95 2.6 ih. 5&m. 2750 61 6772 " 3 11.00 " 130 23-5 95 2.8 2h. 56m. 4130 34 6773 30 12. OO " 130 23.5 95 2-9 3h. s6m. 5470 too 6774 July i 1. 00 A.M. 130 23.5 95 3-2 4h. 56111. 6870 84 ^775 i 2.OO i ^o nc 8 270 6776 1 I 3.OO i j\j 130 23. 5 OO ys nc 3 -4 37 6h. 56m. 9610 6778 " I 5.41 " 131 * J J 23-5 VD 95 I oh. o6m. 142 81 6779 " I 5-47 " 131 23-5 95 oh. I2m. 242 39 6780 " I 5-57 " 131 23-5 95 oh. 22m. 492 8 6784 " I 6.19 " 131 23.5 95 oh. 44m. 992 68 6785 " I 8.00 " 131 23-5 95 2-7 2h. 25111. 3 262 84 <7-7 " I 9.00 " 131 23-5 95 3-o 3h. 25m. 4662 23 <>:-- " I 10.00 " 131 23-5 95 3-2 4h. 25m. 6042 25 6789 1 1 1 .00 " 1 1 1 2T C QC 37 5h. 25m. 7 jio 6824 " 2 5.00 P.M. 1 J * 136 *-j 3 23-5 V3 95 / 2.O oh. 04m. / <*+* IO9 67 6825 " 2 5.30 " 136 23.5 95 oh. 34m. 809 6826 " 2 7.30 " 136 | 23.5 95 29 2h. 34m. 3519 20 6827 " 2 8.30 " 136 23.5 95 3-7 3h. 3 4 rn. 4869 18 6832 " 2 9-30 " 136 23.5 95 4-3 4h. 34m. 6259 57 f..s 3; , " 2 10.30 " 136 23-5 95 4-5 5h. 34m. 7649 61 1,831 " 2 11.30 " 136 23.5 95 4.6 6h. 34m. 8999 22 6835 " 3 1. 00 A.M. 136 23.5 95 5-o 8h. 04111. II079 21 6837 3 3.02 " 137 23-5 95 oh. 07111. 142 99 (,S;, 3 3-05 " 37 23-5 95 oh. lorn. 242 87 6839 3 3-22 " 137 23.5 95 oh. 27m. 602 66 ,-.( ! 3 4.00 " 137 23-5 95 ih. osm. I 502 26 6841 3 5-00 " 37 ; J3-5 95 3 o 2h. osm. 2 842 3i 6845 3 6.00 " 137 23-5 95 3-8 3h. osm. 4252 25 35 WA TEK P URIFICA TION AT LO UIS VI L L E. RESULTS OF BACTERIAL ANALYSES WITH SULPHATE OF ALUMINA. Continued. Rate of j i. Collected. Filtrat on. fc i/i u - Number 8. i Period of i.E . U . Run. K Last Hou S rsTn R d g|| || j Remarks. x Date. Hour. =<J u c O ^-C Minutes. 2 ".35 i: 3 =i = a? i ^JU |(J J) u 5 -> U. : 03 1897 6846 July 3 8.OO A.M. 137 23-5 95 4 .8 5h. 5 m. 7 022 39 6848 3 g.OO " 137 23- 5 : 95 5.5 6h. 0501. 8412 72 6859 " 6 1.30 P.M. 140 230 95 2.2 oh. lom. 272 291 6860 6 1-45 " 140 23-5 95 2-5 oh. 25m. 652 98 6861 6 2.3O " 140 24.0 97 2-5 ih. lom. i 692 47 6863 " 6 4.00 " 140 23-5 95 2-9 2h. 4om. 3 792 62 6865 6 5-30 " 140 23-5 95 3-5 4h. lom. 5912 41 6866 6 8.30 " 140 23-5 95 4.8 7h. lom. 10 192 49 (,S()7 " 6 9.00 " 140 23-5 95 5.0 7h. 40m. 10 902 33 6870 " 6 10.00 " 140 23.5 95 5.6 Sh. 40111. 12282 38 6871 6 I I . OO " 140 23.5 95 6.oj gh. 40111. 13622 37 6872 " 6 12.00 " 140 23.5 95 6.9 loh. 40tn. 14972 31 6873 " 7 1. 00 A.M. 140 23-5 95 7.5 nh. 40m. 16 322 92 6874 6876 7 7 2.OO " 3.00 " 140 23.5 140 23.5 95 95 7-S i2h. 4om. 8.3 I3h. 4om. 17 602 18 912 126 49 Agitated surface of sand layer at 6877 7 4.00 " 140 23.5 95 5.2 i4h. 32m. 20022 58 3.37 A.M. 6878 " 7 5.OO " 140 23.5 95 5.7 ish. 3201. 21 3,2 in 6879 7 6.00 " 140 23.5 95 6.6 i6h. 32m. 22 522 143 6881 7 9.00 " 140 23.5 95 7.6 I9h. 28m. 26 592 405 .Closed outlet at 9.03 A.M. 6882 7 10.24 " 141 23.5 95 oh. osm. 146 33 6883 7 10.30 " 141 23.5 95 2.1 oh. urn. 286 295 6885 11.00 " 141 23.5 95 2-3 oh. 41 m. I OO8 200 Closed outlet at 11.13 A - M - (,ssi " 7 11.46 142 23.; 95 2-3 oh. o6m. I 4 8 195 6887 7 11.51 " I 4 2 23.5 95 oh . 1 1 m . 2 5 8 I 5 6 68 SS 7 12.00 M. 142 23.5 95 2. I oh. 2om. 478 156 6889 7 1. 00 P.M. 142 23-5 95 2-9 ih. 2om. 928 61 6890 7 2.OO " 142 23-5| 95 3-7 2h. 2om. 3313 98 6892 7 2.30 " 142 23.0 93 3-8 2h. som 4008 144 Closed outlet at 2.39 P.M. 6893 7 3-5() " 143 23.5 95 2. I oh. o6m 148 320 6894 7 4.OO " 43 23-5 95 2.1 oh. lom 238 305 6897 7 4.11 " U3 ; 23.5 95 oh. 24111 498 282 6898 7 4,3" " 143 23.5 95 2. 2 oh. 4om 948 259 6899 7 5-30 " 143 i 23.5; 95 2-3 ih. 4om 1663 191 6901 7 9-32 " M4 23-5 95 oh. o6m 141 310 6902 7 9-36 " M4 23-5 95 oh. lom 241 242 690; 7 9.48 M4 23-5i 95 ! oh. 22m 471 625 6905 7 10.30 " 144 23-5 95 2.7 ih. 04111 1481 133 6gof 7 I I . OO " M4 23-5 95 2-9 ih. 34111 2 171 192 6907 " 7 11.30 " 144 23.5 95 2h. 04111 2S7I 182 Closed outlet at 11.35 P.M. 6qoS " 8 12.54 A.M. 145 23-5 95 oh. 07111 145 261 6909 " 8 I . OO " M5 23-5 95 2.0 oh. 13111 275 198 6910 S 1 . 30 " M5 23-5 95 2.1 oh. 43m. 965 174 6qi " 8 2.00 " M5 23-5 95 2.1 ih. 13111. i 535 82 69 1 . " 8 4.00 " 145 23.5 95 2.2 3h. 1301. 4375 61 691; S 5.00 " 145 23-5 95 2-7 4h. 13111. 5725 68 691 " S 6.00 " M5 23-5 95 2.g 5h. 13111. 755 69 691 " 8 9.00 " M5 23-5 95 3-6 8h. I3m it 245 147 Closed outlet at 9.15 A.M. 692 " S i 10.03 " 146 23-5 95 oh. 06 m 151 245 692 8 10.12 " 146 23-5 95 oh. 1 5m 37 260 692 " S 11.30 146 23-5 95 2.2 ih. 33m 2 igl 138 692 " 8 I.OO P.M. 146 23-5 95 2-7 3h. 03111 4281 152 692. 8 2.30 " 146 23-5 95 2-? 4h. 3301 6321 MS 694 9 4.52 149 23-5 95 oh. 07m 251 305 694 9 5.00 " 149 23-5 95 2.2 oh. ism 381 127 694 9 8.00 " 149 23-5 95 2.8 3h. ism 4751 56 694 9 9.00 " 149 23-5 95 2.r 4h. I5m 6171 61 6g66| " 10 4.40A.M. 153 23.5 95 oh. osm M<>! U7 696- " TO 4.45 153 23-5 95 oh. jom 24f 67 6 9 6c " 10 5.30 " 153 23-5 95 2.: oh. 46111 i 036 64 6g7( ) 10 6.00 " 153 23-5 95 2.-1 ih. i6m I 746 60 697 " 10 S.oo " 153 23-5 95 2.1 3h. i6m 4616 295 697: " 10 9.00 " 153 23.5 95 3-( 4h. i6m 6036 70 697 " 10 10.00 " 153 23.5 95 3-c 5)1. 1 6m 7436 72 Closed outlet at 10.05 A.M. 698 10 5.24 P.M. 156 23-5 95 oh. o6m I5C 171 INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 18H7. 357 RESULTS OF HACTERIAL ANALYSES WITH SULPHATE OF ALUMINA. Continued. ( Elected. s a 7. Date. Hour. 1 6986 1897 July 10 5.32 P.M. 6987 " 10 5.40 " 6992 ;; M 8.55 " 6993 8.59 " 6994 " T 4 9.10 " 6995 6998 :: 11 9.30 " O.GO 6999 !4 1.23 " 7000 " 14 1.27 " 7003 " M I. 3 8 " 7004 " M 2.OO " 7006 " 15 4.49 A.M. 7007 15 4-53 " 7010 15 5.04 7011 J5 5-30 " 7012 15 6.00 " 7013 15 S.oo " 7015 15 9.00 " 7018 15 0.06 " 7019 15 o.io 7022 15 0.30 " 7023 15 1.30 7024 15 2.30 P.M. 7025 15 2.52 7026 15 3.OO 7030 15 3-30 " 73 15 4.30 " 7032 15 8.00 " 7036 15 9.00 " 737 15 o.oo " 7038 15 1.16 " 7039 15 i . 20 " 7040 15 1.30 7043 15 2.OO 7044 " 16 I.OO A.M. 7046 " 16 3-5f> " 7047 " 16 4.00 7<>49 " 16 5.00 7050 " 16 6.00 " 7052 1 16 9.00 " 753 " 16 10.17 " 7054 " 16 10.22 " 7055 " 16 11.00 " 7"57 " if. 12. OO M. 7058 " 16 I.OO P.M. 7061 16 9.58 " 7062 " 16 IO.O2 " 7064 " 16 10.12 " 7065 " If) 11.00 " 7066 7067 " 6 " 17 12.00 " I.OO A.M. 7068 17 2.00 " 7070 17 4.14 " 7071 17 4.18 " 7072 17 4.30 " 7"75 17 S.oo " 7076 17 6. oo " 777 17 8.00 " 7079 17 9.00 " 7080 " r ~ IO.OO " 7081 " 17 II. OO " 7082 7 12.00 M. ,-083 17 I.OO P.M. 7084 17 2.OO " Rate of S s Filtration. 1 . ~ lumber V C V o. _o a i -o Period of .S u . Last > S a." Remarks. Run. Si v G 5 3 I [1.5 j-< o : _ H W ouVs^d r* . [i, -c!-~ y 8 Minutes. V 3^ S ^ IS S : cT ~,J;j "u u S j J e I 5 6 23.5 95 oh. 14111. 300 192 156 23-5 95 oh. 22m. 500 72 157 23-5 95 oh. o6m. 150 193 157 23.5 95 oh. lom. 250 251 157 23-5 95 oh. 2im. 500 340 157 23.5 95 ; 2.2 oh. 4im. 96O l82 157 23.5 95 ih. urn. i f 55 143 Closed outlet at 10.01 P.M. 158 23.5 95 .... oh. o6m. 148 244 158 23-5 95 oh. lom. 248 225 158 23-5 95 oh. 2im. 498 217 I 5 S 23-5 95 2-3 oh. 43m. 998 229 Closed outlet at 12.03 A.M. 59 23-5 95 --- oh. lom. 254, 260 159 23-5 95 --. oh. 14111. 354 274 159 23.5 95 oh. 25m. 5M 177 159 23-5 95 2.2 oh. 5im. i 294 230 159 23. 5 95 2.4 ih. aim. i 884 M5 159 23-5 95 2.7 3h. 2irn. 4 634 96 59 23-5 95 3-3 4 h - 2irn. 5984 138 Closed outlet at 9.18 A.M. 1 60 23.5 95 oh. o()in. 150 122 1 60 23 -5 95 oh. lom. 250 84 1 60 24.0 97 2-3 oh. 30111. 730 74 1 60 23-5 95 3.0 ih. 3om. 2 HO 80 1 60 23-5 95 3-7 2h. 30111. 347 88 161 23.5 95 .... oil. 06111. 43 96 161 23-5 95 2.3 oh. 14111. 363 86 161 23-5 95 2.6 oh. 44m. i 063 177 161 23-5 95 3.2 ih. 44111. 2573 88 If.2 23-5 95 2.2 ih. 30111. 2059 1 66 162 23-5 <_>5 2.6 2h. 30111. 3419 3! -.62 23-5 95 3.0 3h. 30111. 4799 73 Closed outlet at 10.00 P.M. 163 23-? 95 oh. o6m. 149 59 1 63 23.5 95 oh. torn. 249 38 163 23-5 95 j. i oh. 2om. 499 34 163 23-5 95 oh. 5orn. : I 159 48 163 23-5 95 2.,) ih. 50m. ; 2 539 43 164 23-5 95 oh. o6m. 148 Si \ 164 23.5 95 2. I oh. torn. 232 92 164 23-5 95 2.2 ih. loin. I 628 49 164 24.0 97 2.5 2h. lom. 3078 65 165 23-5 95 2.3 oh. 46111. i 098 128 Closed outlet at 9.04 A.M. 1 66 23-5 95 oh. 0501. 151 55 1 66 23-5 95 oh. lom. ] 251 57 1 66 23-5 95 2-4 oh. 48m. i 161 37 1 66 23-5 95 2.6 ih. 48m. : 2 561 3" 166 23-5 95 2-9 2h. 48111. 3 991 98 Closed outlet at 1.12 P.M. 167 23-5 95 oh. o6m. | 142 63 167 23-5 95 oh. lom. i 242 5 167 23-5 95 oh. 2om ! 492 55 167 23-5 95 2.3 ih. oSm. i i 572 37 167 23-5 95 2.9 2ll. 08 111. 3 022 26 167 23-5 95 3.0 3h. o8m. 4 352 33 167 23-5 95 3- 2 4h. o8m. 5 582 25 168 23-5 95 oh. 07111. 146 23 168 23-5 95 oh. urn. 246 17 168 23.5 95 2.2 oh. 23m. 536 14 168 23-5 95 2-3 oh. 5301. I 236 u 1 68 23-5 95 2.51 ih. 53111. 2666 27 1 68 23.5 95 2-9 3h. 53m. 5456. 7 168 23-5 95 3.1 4h. 53111 6876 31 1 68 23.0 93 3.6 sh. 53m. 8266 22 168 23-5 95 3.8 6h. 5301. 9 666 19 168 23-5 95 3-9 7h. 53m. 1 1 086 37 if.8 23.0 93 4.0 8h. 5301. 12456 49 1 68 23-5 95 4-4 9h. 53111. 13856 72 358 WATER PURIFICATION AT LOUISVILLE. RESULTS OF BACTERIAL ANALYSES WITH SULPHATE OF ALUMINA. Continued. Rate of j 8 Collected. Filtration. EL. c . u t/nr Period of " <*> 3 V Number 5. C P. _^ Service Since i-s*j u ^ 1 5 " " a Washmf?. j| 11 i! Remarks. X , Date. Hour. fe 5 c < c Hours and -ol* u .- u C o ,_ X Minutes. L. w!5 V C & Is U |c.S S z36 9? 1897 7090 July 17 5.13 I .M. 169 23.5 95 oh. 0701. 153 12 7cgi 17 5.30 " 169 23.5 95 2.1 oh. 22m. 563 23 7094 i? 7.30 " 169 23-5 95 2.g 2h. 22tn. 3433 35 7095 T 7 8.30 " i6g 23.5 95 3.1 3h. 22m. 4763 14 7097 17 9.30 " 169 23.5 95 3.7! 4h. 22m. 6 113 19 7098 17 II.OO " i6g 23.5 95 4.6 sh. 52m. S 253 7 7099 17 12.00 " 169 23.5 95 4-g 6h. 52m. 9693 13 7102 18 I.OO A.M. 169 23-5 95 5.3 -h. 52m. n 403 30 7103 " 18 2.00 " i6g 23.5 95 5.6 Sh. 52m. 12 503 32 7105 18 3.00 " 169 23.5 95 5.8 gh. 52111. 13 Sg3 44 Closed outlet at 3.07 A.M. 7106 18 4-50 " 170 23.5 95 oh. osm. 134 36 7107 " 18 5.00 " 170 23.5 95 2. I oh. 15111. 374 17 7108 " 18 5.30 " 170 23.5 95 2.1 oh. 45m. i 054 20 7109 " 18 6.00 " 170 : 23.5 95 2.2 ih. 15111. i 744 13 7113 " 18 g.oo " i/o 23.5 95 2.4 4h. ism.; 5 954 29 7116 18 11.00 " 170 23.5 95 3.7 6h. ism. 8 754 ii 7117 " iS I.OO P.M. 170 23.5 95 4. i Sh. 15111. 11-544 14 7119 " 18 3.00 " 170 23.5 95 4.8 ioh. ism. 14 344 26 7120 " IS 4.00 " 170 23.5 95 5.0 iih. ism. 15 754 Si Closed outlet at 4.06 P.M. 7121 " 18 5.22 " 171 23.5 95 .... oh. o6m. 150 15 7122 " 18 5.30 " 171 23.5 95 2.1 oh. 1401. 340 19 7125 " is 8.00 " 171 23.5 95 2.8 2h. 44m. 3 910 18 7126 " 18 9.00 " 171 23-5 95 3h. 44m. _5 270 4 7128 " IS 10.00 " 171 23-5 95 3.8; 4h. 44m. 6 610 5 7129 18 II.OO " 171 23.5 95 4.2 sh. 44m. 8030 2 7130 18 12. OO " 171 23.5 95 4.8 6h. 44m. 9 450 3 7133 19 I.OO A M. 171 23.5 95 5-2 7h. 44m. 10 890 21 7134 19 2.OO " 171 23.5 95 5.7 Sh. 44tn. 12 320 26 7136 19 3.00 " 171 23-5 95 5.9 gh. 4 4 m. 13 730 36 71.37 19 4.OO " 171 23.5 95 6.2 ioh. 44111. 15 130 124 7138 19 8.56 " 172 : 23.5 95 . . . . oh. 06111. 151 14 7140 ig g.oo_ " 1/2 23.5 95 2.2: oh. lorn. 286 21 7141 19 10.00 " 172 i 23.5 95 2.4 ih. torn. 1686 2 7M4 19 12. OO M. 172 23.5 95 2.9: 3(1. torn. 4 526 7 7M5 19 2.00 P.M. 172 23.5 95 3.4 sh. ion). 7 326[ 8 7147^ 19 3-30 " 172 23-5 95 3.6 6h. 40111. 9 456 5 7148 19 5.00 " 172 23-5 95 4.0 Sh. lorn. II 536 18 7M9 19 7.30 " 172 23.5 95 4.7 ioh. 40111. 15 066 in Closed outlet at 7.49 P.M. 7i5i 19 9.09 " 1/3 23-5 95 ... oh. o6m. 150 52 7152 ig 9-13 173 23-5 95 . . . . oh. lorn. 250 26 7155 " 19 9.30 " 173 23.5 95 2. I oh. 27m. 650 43 7156 19 10. OO " 173 23-5 95 2. I oh. 57m. i 370 32 7 57 19 10.30 " 173 23.5 95 2.2 ih. 27m. 2 070 49 7158 19 12. OO " 173 23-5 95 2.3 2h. 57111. 4 i So 36 7161 ; 20 I.OO A.M. 173 23.5 95 2.4 3h. 57m. 5580 21 7162 " 2O 3.00 " 173 23-5 95 2.7 sh. 57111. 8390 30 7164 " 20 5.00 " 173 23.5 95 2.9 7h. 57m. 10 090 27 7167 " 20 6.00 " 173 23-5 95 9 h. 57m. 12 5OO 31 7168 " 2O 8.00 " 173 23-5 95 3-5 ioh. 57m. 15 300 23 7170 " 20 IO.CO " 173 23-5 95 3-7 I2h. 57m. 18 ioo 3 1 7171 " 20 12. OO M. 173 23-5 95 3-9 I4h. 57m. 20950 25 7172 " 20 2.00 P.M. 173 23.5 95 4- 1 i6h. 57m. 23 780 28 7174 " 20 3-30 " 173 23-5 95 4-3 i8h. 37m. 25900 31 7177 " 20 5.00 " 173 23-5 95 4.8 igh. 57m. 28 990 33 7178 " 20 8.00 " 173 23-5 95 5-0 22h. 57m. 32 ioo 3 Closed outlet at 8.04 P.M. 7180 " 20 9.15 " 174 23-5 95 oh. o6m. 153 21 7181 " 20 9.19 " 174 23-5 95 oh. lorn. 253 27 7182 " 2O 9.30 " 174 23-5 95 2.2 oh. 2im. 493 21 7183 " 20 10.00 " 174 23-5 95 2.2 oh. 51 m. i 173 7i85 " 20 II.OO " 174 23-5 95 2-5 ih. sim. 2573 41 7186 " 2O 12.00 " 174 23-5 95 2.8, 2h. 5im. 3953 39 7187 " 21 I.OO A.M. 174 23-5 95 3.0 3h. Sim. 5373 25 7188 " 21 2.OO " 174 23.5 95 3.2; 4h. Sim. 6612 16 7190 " 21 3.30 " 174 23-5 95 4. i| 6h. 2im. 8613 12 INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 1897. 359 RESULTS OF BACTERIAL ANALYSES WITH SULPHATE OF ALUMINA.- Concluded. Rat ; of 5 s :ollccted. Kiltra lion. Ji lo k. tr. u Period of u "* 3 1 Number 0. 1 - a ServiceSince ii-- J u s of M "rt v "> S Last > " 1J V M Remarks Run. flj V C 5j - X Washing. J>^ i S. Date. Hour. u. - *x Hours and Minutes. ?SJ5 Is ^ a ct i ~u rt^J X u S j n iSg? 7191 July 2 5.00 A.M. 1/4 23.5 95 4-7 7h. 5im. 10 693 21 7192 " 2 6.00 174 23-5 95 5-2 8h. sim. 12 143 14 7195 " 2 9.00 " 174 23-5 95 6.8 nh. 5im. 16293 48 Agitated surface of sand layer 7196 " 2 9.51 " 174 23-5 95 I2h. 4om. 17443 38 at 9.39 A.M. 7197 " 2 11.30 " 174 23-5 95 5.8 14(1. igm. 19763 31 7198 " 2 I.I7 P.M. 175 23.5 95 oh. o6m. 55 73 7199 " 2 2.OO " 175 23.5 95 2-7 oh. 4gm. 205 24 7202 " 2 3.30 " 175 23.5 95 3-o 2h. igm. 3335 26 7215 " 2 5.00 " 175 23-5 95 4-1 3h. 49m. 5425 27 7216 " 2 8.00 " 175 23-5 95 4-9 6h. 4gm. 9635 25 7218 " 2 10.55 " 176 23-5 95 oh. 0501. 145 58 7219 " 2 II. OO " 176 23-5 95 2.1 oh. torn. 300 48 7221 " 2 11.30 " 176 23.5 95 2.2 oh. 4om. 965 51 7222 " 2 12.30 A.M. 176 23-5 95 2-5 ih. 4om. 2 505 57 7223 " 2 1.30 " 176 23.5 95 2-7 2h. 4Om. 3775 33 7224 " 2 2.30 " 176 23.5 95 3-1 3h. 4om. 5 175 41 7226 " 2 3.30 " 176 23.5 95 3-4 4h. 40m. 6595 45 7227 " 2 5-00 " 176 23-5 95 4.0 6h. lom. 8715 49 7228 " 2 8.00 " 176 23-5 95 5-3 gh. lom. 12 825 67 7230 " 2 g.oo " 176 23-5 95 5-8 loh. lom. 14225 86 7231 " 2 o.oo " 176 23-5 95 6.4 nh. lom. 15635 88 7232 " 2 I.OO " 176 23-5 95 6.7 I2h. lom. 17035 94 7233 2 2.OO M. 176 23.0 93 72 I3h. lom. 18435 73 Closed outlet at 12.00 M. 72433 " 2 1. 08 P.M. 179 23.5 95 oh. o6m. 55 170 7244 " 2 I. 12 " 179 23.5 95 oh. lom. 255 162 7245 " 2 1.24 " 179 23-5 95 oh. 22m. 505 134 7247 " 2 1-45 179 23.5 95 oh. 43m. i 001; 107 Closed outlet at 11.45 [ - M - 7248 " 23 2.51 A.M. 1 80 23-5 95 oh. o6m. 148 in 7249 " 23 2.56 " 1 80 23-5 95 oh. nm. 248 152 7251 " 23 1.30 " i So 23-5 95 2.2 oh. 45m. i 048 Sg 7252 " 23 2.30 " 180 23.5 95 2.5 ih. 45m. 2438 61 7254 23 4.00 " 1 80 23.5 95 2-7 3h. ism. 4638 68 7255 " 23 5.30 " 1 80 23.5 95 3-5 4h. 45m. 6808 77 7265 23 8.24 P.M. 182 23-5 95 oh. o6m. 150 283 7266 23 8.28 " 182 23-5 95 oh. lom. 250 273 7267 " 23 8-39 " 182 23-5 95 oh. 2im. 500 136 7271 " 23 g.oo " 182 23-5 95 oh. 42m. I OOO 188 Closed outlet at 9.00 P.M. 7272 1 23 0.18 " 183 23-5 95 oh. lom. 144 223 7273 " 23 0.22 " 183 23-5 95 oh. 1401. 244 258 7274 " 23 0.33 " 183 23.5 95 oh. 25m. 494 275 7275 " 23 I.OO " 183 23-5 95 2-3 oh. 52m. I 164 231 7277 23 2.OO " 183 23-5 95 2-4 ih. 52m. 2 554 193 7278 24 I.OO A M. 183 23-5 95 2-7 2h. 52m. 4054 119 . 7279 24 2.OO " 183 23.5 95 2.8 3h. 52m. 5454 82 Closed outlet at 2.00 A.M. 7281 24 3-05 " 184 23-5 95 oh. osm. 47 208 7282 " 24 3.09 " 184 23-5 95 oh. ogm. 247 200 7283 " 24 3.20 " 184 23-5 95 oh. 2om. 497 138 7285 " 24 3-45 184 23-5 95 oh. 45m. I 067 92 Closed outlet at 3.45 A.M. 7286 " 24 4.50 " 185 23-5 95 oh. o6m. 146 IO2 7287 " 24 4-54 185 23-5 95 oh. lom. 246 126 7289 24 5.30 " 185 23-5 95 2.2 oh. 46m. I 096 91 7290 24 6.1X3 " 185 23-5 95 2.5 ih. i6m. i 836 82 7291 " 24 8.00 " 185 23-5 95 2.8 3h. l6m. 4696 133 7293 24 g.oo " 185 23-5 95 3-o 4h. i6m. 6 116 I6 7 7294 24 10.00 " 185 23-5 95 3-2 5h. l6m. 7 546 253 Closed outlet at 10.02 A.M RESULTS OF BACTERIAL ANALYSES WITH PERSULPHATE OF IRON. 1897 5271 April 6 3.30 A.M. 4 23-5 95 3-3 oh. 43m. i ug 141 5272 " 6 4-30 " 4 23-5 95 4.0 ih. 43m. 2469 192 5273 6 5.30 " 4 23-5 95 5-8 2h. 43m. 3 755 89 5275 6 10.30 " 5 23-5 95 4.0 2h. osm. 2 760 247 3 6 WATER PURIFICATION AT LOUISVILLE. RESULTS OF BACTERIAL ANALYSES WITH PERSULPHATE OF IRON. Continued. Rat eof u S C ollected. Filtr, tion. il ~ U in u Period of V. &* a <u Number 6. 0. n ServiceSince ii J ^ V- a Run. u O u 5 S3 Last Washing. Hours and IK Q Remarks. "3 * Date. Hour. o a |^S "3 Minutes. Sil si "2 S ~ a. 3" 1 "i-< U n U w u S E n 1897 5276 April 6 - o oo M f 23 5 95 840 5277 6 2.00 P.M. 6 23.5 95 3.1 ih. i8m. 1 714 igi 5278 " 6 3.00 " 6 23-5 95 5-6 2h. i8m. 3 54 172 5280 6 3-30 " 6 23-5 95 7-7 2h. 4801. 3814 172 Closed outlet at 3.42 P.M. 5281 6 5.OO " 7 23o 95 3-1 ih. oom. i 382 228 5283 6 5.30 " 7 23 5 95 3-7 ih. 3om. 2082 319 5293 7 10.00 A.M. 10 23-5 95 5-2 2h. oom. 2837 168 5294 7 IO.3O " IO 23-5 95 7-5 2h. 3001. 3517 177 Agitated surface of sand layer at 5295 7 10.37 " 10 23-5 95 2h. 35m. 3617 397 10.32 A.M. 5296 7 10.40 " 10 23-5 95 2h. 38m. 3688 415 5297 7 10.43 " IO 23.5 95 2h. 4im. 375 6 239 5298 10.46 " IO 23-5 95 2h. 44111. 3824 198 5299 7 10.49 " 10 23-5 95 2h. 47m. 3892 182 5300 7 10.52 " 10 23-5 95 2h. 5om. 3963 159 5301 7 10.55 IO 23-5 95 .... 2h. 53m. 4035 177 5302 7 11.30 " 10 23-5 95 5-2 3h. 28m. 49 7 1 66 5303 7 1. 00 P.M. 10 23-5 95 6.6 4h. 58m. 7027 M5 5304 7 2.00 " IO 23-5 95 5h. 58m. 8472 142 5305 7 2.O6 " 10 23-5 95 6h. oim. 8540 260 5306 7 2.Og 10 23-5 95 6h. 04m. 8612 810 5307 7 2.12 " IO 23-5 95 6h. o7m. 8679 211 5308 7 2.15 " 10 23-5 95 Gh. lorn. 8749 1 69 5309 " 7 2.18 IO 23-5 95 6h. I 3 m. 8817 126 7 2.21 " IO 23-5 95 6h. i6m. 8887 141 [4-48 P.M. 5311 7 3.00 " IO 23-5 95 7.0 6h. 55m. 9 747 1 60 Agitated surface of sand layer at 5313 7 4-50 " 10 23-5 95 7-9 8h. 42in. 12 117 181 Closed outlet at 5.00 P.M. 5338 9 3-3" " 14 23-5 95 3-5 oh. 35m. 836 137 5340 9 5.00 " 14 23-5 95 6.2 2h. 05m 2848 28g 5341 9 5.25 14 23.5 95 6.8 2h. 3om. 3278 187 5342 9 IO.3O A.M. 14 23-5 95 6.0 5h. 49111. 6862 232 5344 9 I2.OO M. 14 23-5 95 6.5 6h. 1 5m. 8643 39 1 12 7.30 P.M. 18 23-0 93 4-5 2h. 05m. 3093 I 375 5382 " 12 8.30 " 18 23.5 95 5-1 3h. osm. 4413 191 5334 " 12 9-30 " 18 23-5 95 5-9 4h. osm. 5922 88 5385 " 12 10.30 " 18 23.5 95 6.6 5h. osm. 6813 1 20 5386 " 12 11.30 " 18 23.5 95 6.9 5h. 4im. 7983 195 5387 " 13 12.30 A.M. 18 23-5 95 7-7 6h. 4im. 9413 183 Agitated surface of sand layer at 5388 13 1.30 " 18 23-5 95 6.5 7h. 3om. 10693 IOO 12.53 A.M. 5389 13 2.30 " 18 23-5 95 6-7 8h. 35m. 12043 470 5391 13 3.21 18 23-5 95 gh. 23m. 13 123 425 Agitated surface of sand layer at 5392 13 3-24 18 23.5 95 gh. 26m. 13 193 gio 3.18 A.M. 5393 13 3-27 " 18 23-5 95 gh. 2gm. 13273 256 5394 " 13 3.30 " 18 23-5 95 7-7 gh. 32m. 13353 202 [4.12 A.M. and 4.59 A.M. 5395 13 4-3 " 18 23-5 95 7-9 I oh. 30111. 14653 147 Agitated surface of sand layer at 13 5.30 " 18 23-5 95 nh. 27m. 15853 287 Closed outlet at 5.46 A.M. 5425 M 10.30 P.M. 25 23-5 95 2-5 oh. 3om. 665 ig2 5426 M 11.30 " 25 23-5 95 4.0 ih. 3om. 2075 310 5427 15 12.30 A.M. 25 23-5 95 4-4 2h. 30111. 3475 120 5428 15 1.30 " 25 23-5 95 4.8 3h. 30111. 4865 47 5429 15 2.30 " 25 23-5 95 5-5 4h. 3om. 6255 78 5431 15 3-30 " 25 23-5 95 6.8 5h. 3om. 7645 7 5432 15 4.30 " 25 23-5 95 7-o Gh. 3om. 9035 6g Agitated surface of sand layer at 5433 " 15 5.30 " 25 23-5 95 7.8 7h. 25m. 10305 264 4-37 A.M. 5451 " 16 g.oo " 31 23-5 95 3-0 ih. oom. I 392 58 5452 " 16 10. OO " 31 23-5 95 3-5 2h. oom. 2 782 98 5453 " 16 1 1 . OO " 31 23-5 95 4.0 3h. oom. 4 182 T 59 5454 " 16 12. OO M. 31 23-5 95 4-4 3h. 35m. 4932 ng 5455 " 16 1. 00 P.M. 31 23-5 95 5-0 4h. 35m- 6 322 239 5459 " 16 2.00 " 31 23-5 95 5-7 5h. 35111. 7 732 310 Closed outlet at 2.13 P.M. 5489 " 21 8.00 " 35 23-5 95 3-9 ih. 23m. 2135 183 5491 " 21 10.30 " 35 23.0 93 4-5 3h. 26m. 4775 71 5492 " 21 11.30 " 35 23.0 93 5-o 4h. 26m. 5 635 log 5493 " 22 12.30 A.M. 35 23-5 95 6.4 5h. 26m. 7425 Go 5494 " 22 1.30 " 35 23-5 95 8.0 6h. 26m. 8775 114 INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 1M7. 361 RESULTS OF BACTERIAL ANALYSES WITH PERSULPHATE OF Collected. K;iti : Filtration. 1 1 15 1 Number 6 !& c Period of Service Since s|L- u u - 8 y. Run. v . fc 3 151 I Washing. Hours and ll .5 g Remarks. Date. Hour. ^K o Minutes. ii % & t c 11 3&JT i %2>> iju in <_> H j (L, 00 1897 5496 April 22 3.00 A.M. 35 23.0 93 S.8 yh. 56111. 10 805 II? Agitated surface of sand layer al 5497 " 22 4.00 " 35 23.5 95 8.o ; 8h. 5401. 12 115 495 3.02A.M. 5605 " 29 7.00 I .M. 51 23.5 95 2.7| oh. i6m. 384 45! 5606, " 29 9.00 " 51 23-5 95 3-9 2h. i6m. 3 174 43 5607 " 29 IO.OO 51 23-5 95 4-i 3!). 1 6m. 4564 29 5609 " 29 11.00 " 5i 23-5 95 5-6 4h. i6m. 5924 47 5610 29 12.00 " 5i 23-5 95 6.4 5h. i6m. 7304 32 5611 " 3 I.OO A.M. 51 23-5 95 8-3 6h. i6m. 8694 44 Agitated surface of sand layer at 5612 " 30 2.OO " 51 23-5 95 5-8 7h. iim. gg64 29 i . 1 9 A . M . 5614 " 3 3.00 " 5i 23-5 95 8.0 Sh. Tim. ii 3M 28 = !,,= " 3" 3-5<) " 5 23.0 93 8.8 gh. o6m. 12 544 19 56l6 " 30 4.07 51 23.5 95 8.0 9h. O9m. 12 624 47 = 17 " 3 4.10 " 51 23-5 95 8.1 oh. 1 2m. 12674 28 5618 " 30 4.13 " 5i 23-5 95 8.2 ()h. 15111. 12 734 1 1 5619 30 4.16 " 5i 23.5 95 8.3 9h. iSm. 12 814 16 562o| 30 4.30 " 51 21.0 89 9.3 gh. 32m. 13 144 33 Closed outlet at 4.49 A.M. 6976 July o 12. l8 I .M. 154 23-5 95 oh. o6m. 150 99, 6977 " o 12.26 " 154 230 95 oh. 14111. 342 186 Closed outlet at 12.26 P.M. 6978 o 2.15 " 155 23.5 95 oh. 07111. 167 97 6980 " o 2.3O " 155 23-5 95 2.2 oh. 22m. 497 "4 6982 " o 2.45 " 155 23.5 95 oh. 37m. 807 7"| 6983 o 3.00 " 155 23.5 95 oh. 52m. I 187 54 1 6984 o 4.1x5 " 155 23.5 95 3-5 ih. 52m. 2 537 58 Closed outlet at 4.04 P.M. RESULTS OF BACTERIAL AN ALYSES WITH COPPERAS. 1897 6720 June 28 12.50 A.M. 23 23-5 95 oh. o6m. 141 6721 " 28 12-55 23 23-5 95 oh. nm. 251 6722 " 28 I . OO 23 23-5 95 2.8 oh. i6m. 371 6723 " 28 I. 10 " 23 23-5 95 oh. 26m. 581 6751 " 30 11.56 28 23-5 95 oh. o6m. 154 34 6752 " 30 12. OO M. 28 23 5 95 2.0 oh. lorn. 254 4 6756 " 30 12.30 I .M. 28 24.0 97 2.6 oh. 4001. 964 121 6757 " 3 I.OO 28 23-0 93 2-9 ih. lorn. 1644 170 6758 " 30 1.30 " 28 23.0 93 3-2 ih. 40111. 2324 229 6759 30 2.0(> " 28 22.0 89 2h. lorn. 2934 147 Closed outlet at 2.00 P.M. 6761 " 3" 4.05 29 23-5 95 2-7 oh. ogm. 220 254 6762 " 3 4.30 " 29 23-5 95 2-7 oh. 34111. 800 188 6849 Ju y 3 1.40 " 23-5 95 oh. o7m. 204 i 700 Samples 6849-6857 with cop 6853 3 2.00 " 38 23-5 95 3-4 oh. 27m. 734 219 peras and caustic soda. 6854 3 2.15 3 8 23-5 95 oh. 42m. I 074 "3 Closed outlet at 2.16 P.M. 6855 3 3-30 " 39 23.5 95 2.8 Oh. 22111. 577 o 6856 3 4.00 " 39 23-5 95 2.9 oh. 5201. I 247 12 6857 3 4-3 " 39 23-5 95 2.g : ih. 22m. i 847 IOi RESULTS OF BACTERIAL ANALYSES WITH ELECTROLYTICALLY PREPARED HYDRATE OF ALUMINUM. 1897 5406 April 13 5.30 P.M. 20 23.0 93 2.5 oh. 04m. no 325 5407 13 8.00 " 20 23.5 95 4.0 2h. 34m. 4580 2 III Closed outlet at S.oS P.M. 5409 14 I.OO A.M. 21 23-5 95 4-2 ih. nm. i 599 670^ 5410 M 2.00 " 21 23-5 95 5-8 2h. nm. 2959 I 100 Closed outlet at 2.10 A.M. 5411 M 3.00 " 22 23-5 95 2-9 oh. 23m. 521 369 5413 M 4.00 " 22 23-5 95 4-7 ih. 23m. ign 236 54M 14 5.OO " 22 23-5 95 5-8 2h. 23m. 32gi i 290 Closed outlet at 5. 10 A.M. 5415 14 6.00 " 23 22. ( 89 3-1 oh. 27m. 6i<] 387 55ic " 22 8.30 P.M. 37 23-5 95 I- 2h. 2om. 3 35 242 Closed outlet at 8.32 P.M. 55" " 22 9-30 " 38 23-S 95 3-5 oh. 3om. 721 i 125 5513 " 22 I0.3O " 38 23-5 95 3-3 ih. 3om. 2 III 99 55M " 22 11.30 " 38 23- = 95 S- r 2h. 3om. 35" 173 55i? " 23 I.3O A.M. 39 23-! 95 2-7 oh. 48m. I 122 "7 Closed outlet at 1.35 A M. 362 WATER PURIFICATION AT LOUISVILLE. RESULTS OF BACTERIAL ANALYSES WITH ELECTROLYTICALLY PREPARED HYDRATE OF ALUMIN UM. Continue J. Rate of J .- Collected. Filtration. J! u. 5i J .o her SL jjS._ ~o Period of s.s O Nun a of Ru V Washing 0-& 8.5 Remarks. 7, Hours and Dale. Hour. o i o J.3C Minutes. s = c S = 0.0 o - 23 S;j M U S " u. 1897 5517 April 23 4.30A.M. 4( ) j 23.5 95 3.0 oh. 56m. I 325 217 5518 " 23 5-3*^ 4* OT C flK 4.2 ill. 58m. 2 085 5520 " 23 10.30 " 41 23.5 95 2.8 oh. 50111. i 157 142 5521 " 23 11.30 " 4 23-5 95 3. 7 ih. 5om. 2557 148 5522 " 23 12.30 P.M. 41 23.5 95 4.8: 2h. 50111. 3917 107 5523 " 23 1-3 " 41 23.5 . 95 7.5 3h. 50111. 5 257 179 Closed outlet at 1.45 P.M. 5544 " 26 9.30A.M. 44 23.5 95 3.3! ih. 14111. i 722 380 5545 " 26 10.30 " 4 1 24.0 97 4.2 2ll. O2I11. 2 852 254 5546 26 1 1 30 " 4 4 23.5 95 6. 5 3h. 02111. 4232 247 5547 26 12.30 P.M. 44 23.5 95 8.7! 4h. oam. 5 602 232 Agitated surface of sand layer at 5548 " 26 1.30 " 44 23.5 95 0.9 5h. oom. 6 962 485 12.40 P.M. 5550 26 5.30 " 4 > 23.5 95 2.7 2h. 2IH1. 2798 145 5551 26 6.30 " 4 ) 23.5 95 3.0 3h. 21111. 4 288 595 5552 1 26 7.30 1 45 , 23.5 95 4.9 4li. 2irn. 5 608 6 838 400 5553 " 26 5554 " 26 "J J 4 9-30 " 4 ) -JO V 3 ) 23.5 95 7-5 5^ 2 1 in. 5.3 Oh Kjm. 8 168 2 33 07 9.04 P.M. 5559 " 26 10.30 " 4 > 23.5 93 S.o 7)1. igm. 9538 141 Closed outlet at 10.35 P.M. 5571 " 28 9.00 A.M. 4 3 24.0 97 7.2 2h. 56m. 4093 169 Agitated surface of sand layer at 5572 " 28 10.00 " 48 24.0 97 4.5 3h. 56m. 5423 201 9.44 A.M. 5573 " 28 i i.oo " 4 3 23.5 95 6.2 4h. 54111. 6930 435 5577 28 12.00 " 4 3 23.5 95 7.0 5h. 54111. 8 320 455 5578 " 28 i.oo P.M. 4 5 23.0 93 7.6 6h. 54m. 9680 995 5583 " 28 4.00 " 4 ) 23.0 93 2.6 oh. 32m. 598 237 5584 28 5-3 ) " 4 ) 23.5 95 3.O 2ll. 02111. 2 678 199 5585 " 28 8.3 -> " 49 ! 23.5 95 6.4 5h. 02m. 6868 319 [9.52 P.M. 5587 " 28 9-3 " 49 23.5 95 S.o 6h. oam. 8 228 453 Agitated surface of sand layer at 5588 28 10.30 " 49 23.5 95 4.8 6h. 59m. 8848 585 Closed outlet at 10.43 P M . 5622 " 30 9.30 A.M. 5 2 23.5 95 2.4 oh. 0501. 147 n5 5623 30 10.30 " 5 2 23.5 95 3.0 ih. 05111. i 557 no 5624 30 11.30 " 5 2 23.5 95 4.8 2h. osm. 2957 177 5625 " 30 12.30 P.M. 5 2 23.5 95 5-3 3h. 05111. 4357 47 [2.06 P.M. 5626 " 30 1.30 " 5 2 23.5 95 7.1 4h. osm. 5 747 195 Agitated surface of sand layer at 5627 " 30 2.30 " 5 2 23.5 95 5.6 5)1. 03m. 7"47 253 [4.10 P.M. 5632 " 30 4.00 " 5 2 23.5 95 7.7! oh. 33m. 9 087 410 Agitated surface of sand layer at 5633 " 30 5.00 " 5 2 23.5 95 7.9 "h. 31111. 10 297 333 Agit.surf.of sand layer at$.O4P.M. 5634 " 30 5.30 " 5 2 23.0 93 8.0 7h. 59111. n 067 429 Closed outlet at 5.44 P.M. 5635 " 30 S.oo " 5 3 23.5 95 2.4 oh. 3im. 698 141 5640 " 30 9.00 " 53 23.5 95 3.0 ill. 31111. 2 258] 158 5"4i 30 10.00 " 53 23.5 95 4-1 2h. 3im. 3 538 95 5642 3" 11.00 " 5 3 23.5, 95 5-7 3h. 31111. 4 908 90 5643 " 30 12.00 " 53 : 23.5 95 S.I 4h. 24111. 6248 83 Agitated surface of sand layer at 5644 May I.OO A.M. 5 3 23.5 95 6. 3 5h. 24111. 7478 87 12.17 A.M. 5645 " 2.00 " 5 3 23.5 95 7-3 Oh. 24111. 8818 164 5646 i 3 ..oo " 5 3 23.5 95 7h. I 9 m. 0088 321 Agitated surface of sand layer at 5648 4.00 " 5 3 23.5 95 7-2 8h. igm. 1568 143 12.55 A.M. 5649 5.00 " 5 3 23.5 95 9h. I4m. 2678 429 Agitated surface of sand layer at 5650 5.03 " 5 3 23.0 93 9!]. 1 7m. 2 768 243 4.55 A.M. 5651 " 5-of) " 5 3 23.5 95 9h. 2om. 2838 179 5653 " 9.00 " 5 4 23.5 95 2.8 oh. 45m. I 031 121 5654 " 10.00 " 5 4 23.5 95 3-2 ih. 45m. 2441 93 5055 " 11.00 " 5 4- 23.5 95 3-5 2h. 45m. 3SSl 95 5656 " 12. OO " 5 t 23.5 95 3.8 3h. 45m. 5 251 92 5657 " I.OO P.M. 5 4 23.0 93 5-0 4h. 45m. 6 601 IOO 5658 2.00 " 5 ^ 23.5 95 6.0 5h. 45m. 7981 69 5663 3.0 " 5 \ 23.5 95 7-7 6h. 42m. 93Si 121 Agitated surface of sand layer at 5738 " 7 I.OO A.M. 5 i 23.5 95 4-5 ih. iSm. i 528 130 3.08 P.M. 5739 7 4.30 " 5 > 23.5 95 2.5 oh. 23m. 611 114 - Oi 5744 7 5.00 " 5 23.5 95 2.6 oh. 53111. I 331 I2g 5745 7 6.00 " 5 > 23.5 95 4-4 ih. 53m. 2 761 92 5746 7 9.30 " 5 7 23.5 95 2.7 oh. 28m. 577 252 5751 7 10.30 " 5 7 23.5 95 3-1 ih. 28m. 1947 214 INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 1897. 363 RESULTS OF BACTERIAL ANALYSES WITH ELECTROLYTICALLY PREPARED HYDRATE OF ALUMINA. Concluded. Collected. Rate of Filtration. i S i Number R I k . = Period of Service Sine 5 = . 3 U j e a 7. 1 Date. Hour. of Run. U. g - .? - Was hmif. H-.urs and Minutes. ii A & Remarks. 3 "5 ~ D. n . c U S. _; 1 1897 5752 May 7 11.30 A.M. 57 23-5 95 4-1 2h. 28m 3287 319 5753 7 12.30 P.M. 57 23.5 95 5-5 3"- 2Sm 4 537 28g Closed outlet at 12.33 - M - 5754 7 2.30 " 58 23.5 gs 2.2 oh. I7m 3<)7 5755 7 3.30 " 58 23.5 95 3.2 ih. 17111 i 797 139 5760 7 4.30 " 58 23.5 g5 4.7 2h. I7m 3 187 i 175 6806 July 2 4.45 A.M. 133 23.5 95 .... oh. o6m. 150 170 6807 2 4-50 " 133 23.5 95 oh. nm. 250 89 6808 " 2 5.00 " 133 23.5 95 2. i oh. 2im 480 136 68og " 2 5.30 " 133 23.5 95 2.2 oh. 5im. i 170 60 6810 " 2 6.00 " 133 23-5 95 2.3 ih. 2im. i 830 52 6815 " 2 g.oo " 133 23.0 93 4.0 4h. 1301. 5830 127 6816 " 2 10.30 " 134 23-5 95 2.2i oh. ogm. 214 112 6817 " 2 11.30 134 23-5 95 2.7 ih. ogm. i 594 73 6821 " 2 1. 00 P.M. 134 23-5 95 3.3 2h. 39111. 3 704 126 6823 " 2 3-53 " 135 ! 23.5 95 oh. 13111. 311 ! 147 Closed outlet at 3.53 P.M. RESULTS OF BACTERIAL ANALYSES WITH ELECTROLYTICA1.LY PREPARED HYDRATE OF IRON. I8g7 | 5346 Apri ii 3.00 A.M. 15 23.0 93 2.8 oh. 50111. i 153 4000! 5347 " ii 4.30 P.M. 15 23-0, 93 3.9 2h. 20111. 3123 5410 5350 " ii 5.30 " 15 23.0 93 4.2 3h. 20111. 4403 7900 5351 " 1 1 6.3O " 15 23-0 93 ! 5.0 4h. 2om. 5723 6000 5352 " ii 7.30 " 15 24.0 97 6.0 5h. 20111. 7 123 9800 5353 " ii 8.30 " 15 22. o 8g 7.3 6h. 2om.| 8353 24000 5355 " ii 9-3 " 15 22.0 89 S.o 7h. igm. 9663 21 ooo Agitated surface of sand layer 5356 " n 9-43 " 15 23.5; 95 6.6 7h. 32111. 10 073 6 650! at 9.37 I -M. 5357 " ii 9.46 " 15 23.5 95 7h. 35m. 10143 44500 5359 " " 9-49 " 15 23.5 95 7h. 3Sm. 10203 10000 5360 ii 9.52 " 15 23.5 95 7-i: 7h- 4im- 10 263 12 COO 5364 " ii 11.30 " 16 18.0 73 2.0 oh. 38m. 662 2 72O 5365 " 12 I2.3O A.M. 16 18.0 73 2.7, ih. 38m. I 712 2 goo 5366 " 2 1.30 " 16 18.0 73 3.2 2h. 38m. 2 772 i 680 5367 2 2.30 " 16 18.0 73 4.0 3h. 38m. 3 ,M)2 2 2OO 5369 2 3.30 " 16 17.0 6g 5.0 4)1. 38m. 4842 2350 5370 2 5.00 " 17 18.0 73 2. i oh. som. 892 I 600 5371 " 2 6.OO " 17 18.0 73 3-0 ih. 5om. I gg2 2030 5373 " 2 g.oo " 17 17-5 71 5-0 4h. som. 4g32 15 200 5374 " 2 IO.OO " 17 16.0 65 5-9 5h. som. 5 go2 12 600 5375 " 2 II. OO " 17 18.5 75 7.0 6h. som. 6 882 13 500 5376 " 2 12.00 M. 17 14.0 57 6.0 7h. 5 om. 7882 12 700 5377 2 1. 00 P.M. 17 18.0 73 7-o 8h. som. 8782 4640 5378 " 2 2.00 " 17 16.5 67 8.0 gh. som. g 782 6 400 5379 5417 " 2 " 4 3.00 " 1.30 " ! 7 17.0 69 8.2 loh. soin. ... nh r c m _ 10 762 9 450 355 ^ 7I 54i8 4 5.00 " 25 23-5 95 3-9 ih. 05 m. i 715 42g 5420 4 5-30 " 25 23.0 93 4.0 ih. 35m. 2375 755 5461 " o 8.30 " 32 23-5 95 5-8 3h. 56m. 5400 i 240 5462 " o 9.30 " 32 23-5 95 7-o 4h. 56m. 6776 i 450 5468 " o 12. OO " 33 23-5 95 3-0 ih. oom. I 404 449 5469 1 1. 00 A.M. 33 23-5 95 4-2 2h..ootn. 2744 397 5470 2.OO " 33 20. o 81 5-9 3h. oom. 3994 295 5471 " 3.OO " 33 20. o 81 6.8 4h. oom. 5 154 267 5472a 4.00 " 33 20. O 81 7.5 5h. oom. 6 324 347 5474 g.oo " 34 18.0 73 3-5 3h. 25m. 3497 167 5475 10.00 " 34 18.0 73 4.7; 4h. 25m. 4357 200 5476 11 II. OO " 34 18.0 73 5.0 5h- 25m. 5647 139 5477 12.00 M. 34 18.0 73 6.0 6h. 25111. 6697 17- 5478 5479 5480 I.OO P.M. 2.OO 3.00 " 34 34 34 18.0 18.0 18.0 73 73 73 7.o| 7h. 25m. 5.6| 8h. 25111.; 7.81 gh. 25111. 7747 8777 9823 217 [at 1.35 I .M. 138 Agitated surface of sand layer 184 Agitated surface of sand layer 5481 3.06 34 18.0 73 gh. 3im. 99 7 420 at 3.02 P.M. 3"4 WATER PURIFICATION A 7 LOUISV/LLh. RESULTS OF BACTERIAL ANALYSES WITH ELECTROLYTICALLY PREPARED HYDRATE OF IRON. Continue,/. 1 Rate of ^ Collected. Filtration. .5 . J T- 1 t-riod of ^ ti .0 Number a 0. a ServiceSince w.E ^ u E 3 R un j . ^ : - ^ W shine " ~ *J Remarks ^ [i. " * r ~ Hours and -c^u .-- Date. Hour. u i o ,_ I o Minutes. , tyiS = ~ fi = 0.? I ^25 5<J y> <j 2 >J fc B jSg? 5482 April 21 3.09 I .M. 34 iS.o 73 . gh. 34m. 9967 212 5483 " 21 3.12 " 34 iS.o 73 , gh. 37m. 10017 192 5488 21 4.OO " 34 iS.o 73 i 6.8 ioh. igm. 10867 232 Agitated surface of sand layer at 5560 27 8.30 " 46 23-5 95 ! 3-3 2h. 29111. 3477 420 3-49 i -M- 5565 2- 9.30 " 46 23.5 95 3-9 3h. 29111. 4807 740 5566 " 20 I.OO A.M. 47 23.5 95 2.8 ih. oom. i 347 248 5567 " 28 2.OO 47 23.5 95 3-1, 2h. oom. 2 707 79 5568 " 28 3.00 " 47 23.5 95 4.1 jh. oom. 4037 202 6478 June 19 0.05 " 88 23.5 95 <)h - 5"i- 131 387 6479 " 19 o. I o " 88 20 . o 81 . . . . on. lom. 231 292 6480 " 19 0.15 " 88 20.0 81 oh. 15111. 331 427 6481 " ig 0.30 " 88 i 22.0 89 . . . . oh. 30111. 661 237 6482 19 1 -45 88 34-5 99 oh. 45m. i 031 295 6483 " 19 I.oo " 88 24.5 99 . . . . ih. oom. I 401 3O2 6484 " ig 1.40 " 89 10.0 40 oh. osm.: 82 251 6485 19 T -45 89 10. 40 oh. lom. 132 299 6486 19 1.50 " Sg 12.0 48 oh. ism. 192 249 6487 19 .55 89 I2.O 48 oh. 20111. 252 26l 6488 19 2.00 I .M. Sg 18.5 75 oh. 3om. 437 301 6489 " 19 2.15 89 18.5 75 . . . . oh. 4om. 622 235 6490 19 2.32 " 89 23.0 g3 .... < oh. 5701. i 012 263 6491 19 1.30 " Sg 24.0 97 i.o ih. 55m. 2 442 97 6492 19 2.30 " Sg 23.5 95 1.5 2h. 55m. 3 842 142 6494 " 19 3.3 > 89 23-5 95 1.8 3h. 55m. "! 272 99 C>495 19 4.30 " Sg 230 95 2.0 4h. 55m. 6662 260 6496 19 5.30 8g 23-5 95 2.1 5h. 55111. S 062 141 f 497 ig 6.30 " 8g 23-5 95 2. i ( 6h. 55m. 9352 154 6498 19 S.oo " Sg 23-5 95 2.3 Sh. 2501. II 402 145 6500 " 19 9-30 " Sg 23-5 95 2.4 gh. 55m. 13 222 201 6501 i g 0.09 Sg 23-5 95 ioh. 34111. 14 152 250 Agitated surface of sand layer at 6502 19 O.2I Sg 23-5 95 6-5 ioh. 39m. 14 252 ig2 lo.og r.M. 6503 19 i-57 " go 23-5! 95 oh. osm. 71 342 6504 20 2.05 A.M. go 23-5, 95 2.2 oh. 13111. 381 278 6505 " 20 2.2g " go 23-5, 95 oh. 3701. Sn 292 6506 20 1 . 30 " go 23-5 95 2.6 ih. oim. i 391 223 6507 20 2.OO " do 23.5 95 2.6 ih. jim.j 2 277 183 6508 2O 4.21 " gr 12.0 48 1 .0 oh. 05m. fig 295 6509 20 4.26 " gl 12.0 48 i.o oh. torn. 129 377 6510 20 4.30 " gl 12. o 48 o.g oh. 14111. 189 380 6511 " 20 4.45 " gl 12.0 48 oh. 2gm. 359 281 6513 " 20 5.00 " gi 18.5 75 1.6 oh. 44111. 609 280 6514 " 20 5.30 " gl 20. o Si 1.7 ih. 14111. I 229 274 6555 20 6.00 " gi 20. o 8 1 2.0 ih 44in. i 739 271 6516 " 20 6.30 " gl 15.0 61 1.7 2h. 1 4m. 2 25g 205 6517 " 2O 7. (30 " 91 15.0 61 1.7 2h. 44111. 2 749 219 Closed outlet at 7.00 A.M. 6609 24 12.50 " 05 23.5 95 oh. O5m. 107 470 6610 24 12.55 " 05 23.5 95 oh. lom. 227 450 66 1 1 " 24 I.OO " 05 23.5 95 2-3 oh. 15111. 347 565 6612 24 I. 10 " 05 23.5 95 oh. 25m. 577 360 6613 24 1.20 " 05 23.5 9 oh. 35m. 807 268 6614 24 1.29 " 05 23.5 9 . oh. 44m. 1047 385 Closed outlet at 1.29 A.M. 6616 24 4.40 " 06 20. o 8 oh. o6m. 90 3 5 6617 24 4-45 06 20. o 8 oh. nm. 190 485 6618 24 4.50 " 06 20.0 S oh. 1 6m. 290 535 6619 24 5.00 " 06 2O. O 8 i.g oh. 26m. 55 390 6620 24 5.30 " 06 2O. O 8 oh. 56111. i ogo 435 6621 24 6.00 " 06 20.0 S 2.0 ih. 26m. i 710 465 6623 24 9.00 " 06 20.5 s 3-o 4h. 26m. 5 330 198 6624 24 IO.OO " 06 20. O 8 3-7 5h. 26111. 6 480 258 6625 24 I I.OO " 06 2O. O S 4-5 6h. 26m. 7 820 229 6626 24 I2.OO M. 06 20.0 8 4-7 7h. 26m. S goS 395 6627 24 12.30 P.M. 06 19. o 77 4.8 7h. 5601. 9480 475 Closed outlet at 12.30 P.M. INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 18!>7. 365 RESULTS OF BACTERIAL ANALYSES WITH ELECTROLYT1CALLY PREPARED HYDRATE OF IRON .Continued. Rate of w Collected. Filtr ition. ;| fc X umber I o a I enoa 01 ".s j u u . g ^ o 3 . L ? st - a ~ i) v 2i Remarks. Run U v O o 1 I Washmg * *~ u. " s z Date. Hour. "ol c< Hours and o Minutes. : ^i ii J! s u .-a* 8 n 1897 6634 June 24 5-45 l -M. os iS.o 73 1.7 oh. osm. M5 59 6635 " 24 7.30 " 08 18.0 73 2.0 ih. 5001. 2 065 625 Closed outlet at 7.50 P.M. 6640 " 24 11.55 10 23.5 95 .... oh. 06111. IS i 495 6641 " 24 12.00 " 10 23.5 95 .... oh. nm. 2Sl 1680 6642 " 25 12.15 A.M. i > 23-5 95 |. ... oh. 26m. 581 2 750 Closed outlet at 12.32 A.M. 6644 " 25 3.12 " i 18.0 73 oh. osm. 105 695 6645 " 25 3.20 " i i 18.0 73 .... oh. 1301. 285 710 6646 " 25 3.30 " I IS.,, 73 1.5 oh. 23m. 485 545 6647 " 25 4.00 " I iS.o 73 i .5 oh. 53m. I 025 575 66 4 s " 25 5.00 " I iS.o 73 1.7 ih. 53171. 2 065 260 6649 " 25 6.00 " I I S . o 73 2.0 2h. 53m. 3035 244 6650 11 25 8.00 " i 1 8 . o 73 ; 2.6 4h. 53m. 5 275 1 66 6652 " 25 9.00 " i iS.o 73 3.0 sh. 47171. 6445 152 6653 " 25 10.00 " i 18.^ 73 3-5 f>h. 47m. 7435 86 6654 " 25 11.00 " i 18.0 73 4.0 7h. 47m. 8515 in " 25 12. OO i 18.0 73 4-5 8h. 47111. 9625 131 6656 " 25 I.OO I .M. i 1 18 o 73 4 S cjh. 47111. 10 745 123 6657 " 25 2.00 " i iS.o 73 5-2 loh. 47111. n 835 107 6659 " 25 3.00 " i 18.0 73 5-7 nh. 39 m - 2935 93 6660 " 25 4.30 " i 1 19.0 77 6.0 I3h. 09111. 14455 181 6661 " 25 7-3" " i 18.0 73 7.2 i6h. 09111. 17 815 151 6663 " 25 9.00 " I 18.0 73 7.8 i7h. 39m. 19365 192 6664 " 25 i ii. oo I IS.o 73 9.0 icjh. 39m.l 21 455 179 6665 26 12.30 A.M. I 18.0 73 9.5 2ih. ogm. 23055 127 Agitated surface of sand layer at 6666 " 26 12.45 I 18.0 73 6.2 2ih. 2om. 23 245 385 12.36 A.M. 6667 "26 12.50 " I iS.o 73 , 6.3 2ih. 25m. 23 355 177 666S 26 i.oo " i iS.o 73 6.5 2 ih. 35m. 23535 194 6670 " 26 3.00 " I iS.o 73 7-8 23h. 35m. 25585 137 6671 "26 5.00 " [ 18.0 73 S.S 25h. 35mJ 27695 142 6672 26 6. ex, " i 18.0 73 , 9.0 26h. 35111. 28 725 K)6 Closed outlet at 6.00 A.M. 6673 "26 7.38 " 12 23-5 95 oh. 44111. i 053 212! 6675 "26 11.22 " 13 23-5 95 oh. o6m. 138 295! 6676 26 11.27 " 13 23.5 95 oh. inn. 2SS 258 6677 " 26 11.37 " 13 23.5 95 . oh. 21 m. 498 281 6678 " 26 ; 12.00 M. 13 23-5 95 . . . . oh. 44111. i ooS 275 Closed outlet at 12.13 P.M. 6680 26 , 3.19 P.M. 14 18.0 73 oh. 07m. 150 395 6681 " 26 3.25 " 14 18.0 73 oh. 13111. 250 410 6682 " 26 3.38 14 18.0 73 oh. 26171. 500 335 Closed outlet at 3.40 P.M. 66,S ; 26 5.40 15 23-5 95 oh. o6m. 150 315 6684 " 26 5.50 " 15 23 5 95 oh. ifim. 370 295 6685 " 26 7.30 " 15 23-5 95 2-5 ih. 56m. 2690 340 ; Closed outlet at 7.33 P.M. 6699 " 27 5.30A.M. 19 23-5 95 2.0 oh. osm. 1 16 i 020 6700 27 5.40 " 19 23-5 95 oh. ism. 326 690 6701 27 6.00 " 19 23-5 95 oh. 35m. 796 385 Closed outlet at 6.00 A.M. 6703 " 27 9.00 " 20 23-5 95 3-0 ih. oom. 1383 136 Closed outlet at 9.05 A.M. 6704 " 27 11.33 " 21 18.0 73 oh. osm. 148 260 0705 " 27 11.39 " 21 18.0 73 oh. 14111. 248 335 6706 " 27 12.00 M. 21 18.5 75 i . 7 oh. 3501. 5781 320 6707 " 27 I2.3O I .M. 21 17.0 69 i. 9 ih. osm. i 278 146 6708 6709 27 i.oo " 27 1.30 " 21 21 18.0 19.0 73 77 2. I ih. 35m. 2h. osm. i 718; 189 2366 181 Closed outlet at 1.30 P.M. 6710 " 27 2.13 " 22 18.0 73 oh. o8m. 152 223 6711 " 27 2.18 " 22 18.0 73 oh. I3m. 252 272 6712 " 27 2.32 " 22 18.0 73 oh. 27m. 502 198 6713 " 27 3.00 22 17-5 71 oh. ssm. I 012 187 6715 " 27 4.^0 " 22 18.5 75 2.4 ih. som. 2062 I ig 6716 " 27 5.30 " 22 19.0 77 3-3 3h. 2om. 3722 6717 27 S.oo " 22 18.0 73 44 5h. 5om. f>442, 6718 " 27 9.00 22 18.0 73 6h. 5001. 7494 218 Closed outlet at 9.00 P.M. "794 July i 5. m " 32 23-5 95 2.0 oh. o6m. 158 84 6795 " i 5.30 " 32 23-5 95 2.2 oh. 36m. 938 92 366 WATER PURIFICATION AT LOUISVILLE. RESULTS OF BACTERIAL ANALYSESWITH ELECTROLYTICALLY PREPARED HYDRATE OF IRON. Concluded. Collected. Rate of j u Filtr ition. C ~ 1 Number o- |s. ^ Period of * jy. Service Since tJ c 3 (J j B of - 5 \V 1 > - ^ V w Remarks. Run. V V (i" b 1 I Ho urs and *^ 1 ?s J Dale. Hour. 3 ^i o Minutes. "s J E i ; -is i= a 1 ^25 %u * u s J I &. 03 1897 6796 July S.OO P.M. 132 23.5 95 2-9 3h. o6m. 4 308 39 6798 9.00 ^ 132 23.5 95 3-4 4h. o6m. 5 678 49 67(1(1 " 5 h. o6m. u / yy 6800 1 I . OO " 132 23 ;> 132 , 23.5 95 4* 3 4-7 oh. o6m. 8 448 94 6801 I2.OO M. 132 i 23.5 95 5-0 7h. 06111. 9 858 87 6927 4-39 A.M. 147 i 23.5 95 oh. 0501. 140 420 6928 4-44 147 23.5 95 oh. lorn. 240 360 6929 " 4.55 147 23.5 95 oh. 21 m. 490 325 6932 9 5.20 " T47 23.5 95 oh. 46m. i 090 244 f 933 9 6.30 " 147 20.0 Si 2. 2 ih. 56m. 2 510 154 6934 9 S.oo " 147 iS.o 73 24 3(1. 26m. 4 1 60 259 Closed outlet at 8.14 A.M. 6935 9 8.55 " 148 18.5 75 oh. oSm. i 152 226 6936 " 9 9.02 148 18.0 73 oh. ism. 302 212 f 937 9 148 18.0 73 7 oh. 43m. i 908 I 7 6 6940 9 IO.OO " 148 iS.o 73 i-9 ih . 13m. 1338 124 6941 9 I I . OO " 148 iS.o 73 2. I 2h. 13111. 2 41^ 88 942 9 12. OO M. 148 iS.c 73 3-0 3 h. 13111. 3488 S3 6943 9 1. 00 P.M. 148 iS.o 73 3-5 4h. 13111. 4 558 96 6952 9 ".39 " 150 20.0 Si oh . 08 in . 153 315 f>953 9 11.41 150 20. o Si oh. 13111. 253 340 6954 9 12.00 " 150 2(3.0 Si oh. 2gm. 443 555 jClosed outlet at 12.00 P.M. 6956 10 I.I9 A.M. 151 iS.o 73 oh. 08111. 147 330 6957 10 1.24 " 151 iS.o 73 oh. 13111. 247 430 6959: 10 I. 3 S " 151 iS.o 73 oh. 2/m. 497 340 6960 " i o 2.OO 151 18.0 73 oh. 49111. 907 173 Closed outlet at 2.00 A.M. 6961 10 2.40 " 152 18.0 73 oh. I5m. 251 275 6962 " 10 2.45 152 , iS.o 73 oh. 2om. 351 275 6963 " 10 3.oo " 152 18.0 73 ! oh. 35111. 601 269 Closed outlet at 3.10 A.M. 7234 22 1.20 P.M. 177 23.0 93 .... oh. oom. ; 154 151 7235 " 22 1.30 " 177 i 23.0 93 2. I oh. 1 6m. 394 174 7236 22 2.00 " 177 23.0 93 2. I oh. 46m. 1084 202 Closed outlet at 2.00 P.M. 7238 " 22 3.09 " 178 23.0 93 oh. o6m. 149 139 7239 " 22 3.24 178 23.0 93 oh. 2im. 499 138 7-1 " 22 5.00 " 178 23-5 95 2-3 ih. 57m. 3449 63 7242 " 22 S.oo " 178 23.5 95 3-7 4h. 57m. 6899 38 7257 " 23 9.58 A.M. 181 23.5 95 oh. o6m. 146 165 7258 23 IO.I3 " 181 23 5 95 oh. 2im. 496 139 7259 23 11.00 " 181 23-5 95 2.2 ih. oSm. i 596 128 7260 " 23 I2.IX3 M. 181 23.5 95 2-7 2h. oSm. 3 > O6 in 7261 23 I.OO P.M. 181 23-5 95 3-0 3h. o8m. 4386 116 7262 23 2.00 " 181 23.5 95 3-6 4h. oSm. 5806 90 7263 " 23 3.00 " 181 23.0 93 4.0 5h. o8m. 7 206 163 Closed outlet at 3.10 P.M. INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 1807. 367 Summary of Results, showing the Amount of Suspended Matter and Number of Bacteria in the River Water as it passed througli the several Settling Basins. In this table are given all of the determina tions of the amount of suspended matter and number of bacteria in the effluents from the several settling basins, together with corre sponding determinations of the river water. It will be seen that in several instances the determinations in the case of the effluents gave larger results than were found with the river or with the effluent from preceding ba sins. This is mainly accounted for by the fact that it was not found practicable to col lect samples from the same water as it passed through the system, but all samples from the settling basins were taken at the same time, and are tabulated with samples of river water collected from one to three hours earlier. The samples do not, therefore, represent the actual condition of the same water as it passed through the system, but the condition of the various effluents at the hour given. All of the columns given are explained by their headings, except that headed " Treat ment." Under this heading and subheading " Kind," two letters are given. The first let ter refers to the kind of coagulant used, and the second to the place of application. These letters refer to coagulants and places as fol lows: Kind of Coagulant. A. Hydrate of alumina from sulphate of alumina. B. Hydrate of iron from persulphate of iron. C. Hydrate of alumina prepared electro- lytically from aluminum. I). Hydrate of iron prepared electrolyti- cally from iron. E. Hydrate of iron from protosulphate of iron. F. E. with caustic soda. The soda was applied at basin No. 2, and the copperas at the Jewell settling chamber. Place of Application. A. Basin No. 2 and Jewell settling cham ber, in equal amounts. B. Jewell settling chamber. C. Top of filter. D. Basin No. I and Jewell settling cham ber. E. Basin No. I and top of filter. F. Basin No. i. G. Basin No. 2. Coagulants were always applied at the inlet pipes. Under subheading " Amount " the total amounts of chemicals used are given in grains per gallon in the case of commercial chemicals and in ampere hours per gallon in the case of electrolytic treatment; and, in the case of application at 1) or E, the separate amounts are given as foot-notes in the order in which they were applied, the upper one being first and the lower second. Where a settling basin was not in use the fact is so recorded in the column for that basin. 3 68 WATER PURIFICATION AT LOUISVILLE. AMOUNT OF SUSPENDED MATTER AND NUMBER OF BACTERIA IN THE RIVER WATER AS IT PASSED THROUGH THE SEVERAL SETTLING-BASINS. Date 1897. Trea men, RiVL r Water. F.frl Basi nent of i No. i. Efflu Basin ent of No. 2. Effluent Settling of Jewell Chamber. Day. Hour. c 3 B! 6 P /. 2 p 1 <. 1 1 is. at; 3p! 3 U u - g - C_j c v.2 T3 -b 3 . 5 U u ^O 03 s. 2 x "A ~v i- !5 D u c u _ .i ^3 , pa ""i T3 u f g, c^t y; 2 - .5 U ^ ! u<j n A ril I) A 084 7 8 4 47<J 14 15 16 " 20 21 " 22 24 " 26 27 " 28 10.30 " 10.00 " 1. 00 " 9.30 4.00 " 3.30 " 2.30 " 9.30 " 9.00 " 25 28 31 32 34 35 43 45 46 48 B-A A-A B-A D-A D-A B-A A-A C-A D-A C A 2.83 3 oi 3-37 .080 .136 i. Si 1.44 .030 .078 322 347 407 205 183 231 196 184 1 60 38 TOO 28 700 37 ooo 1 6 800 12 6OO 14 600 9400 7 4<)<) II 700 185 216 268 151 192 295 133 107 "5 47900 34 500 29 500 1 8 700 10 700 12 IOO 4 ooo 3900 II 400 225 237 1 66 1 60 120 131 1 06 71 I 19 82 24 200 31 700 14 ioo 19400 7900 5 600 4 200 2 400 I 2 900 87 129 73 128 97 95 80 50 Si ii 200 ii 500 3 50<) ii 800 7300 3 900 2 900 i Soo IO 2OO " 28 29 3<> 3" May I ! 7 4.00 P.M. II. OO A.M. 3.00 I .M. 9. oo " 3-OO " 5.OO A.M. 10.30 " 49 50 52 53 54 56 57 58 C-A A-A C-A C-B C-G C-A C-A C-A .014 1.17 .019 .019 .019 .048 .040 .028 133 136 77 77 77 453 453 4^3 19 500 14 800 IO 2OO 9700 7900 51 500 31 ioo 107 127 76 68 59 181 175 1/5 22 4OO 1 5 Soo R IOO 8 Soo 7 Soo 35 900 19 600 9 2 116 47 67 74 38 212 23 ioo S 200 6 300 9 ioo 6 ioo 36 200 2 OOO 2 IOO 76 S3 34 37 47 104 115 14 900 5 800 3 Soo 4300 4 600 33 ioo 7 100 8 Soo 7 8 " 8 8 9 19 20 " 22 " 22 IO.OO " 4.00 A.M. 9.30 " 4.30 P.M. 4.30 A.M. 11.00 I .M. 5.30 " 3.OO A.M. 9.OO I .M. 59 60 6 1 63 66 67 68 70 70 A-A A-A A-A A-A A-G G-A G-A G-A G-A G A 2.47 3.02 3.6o 3-6? 3-75 1.96 1.92 1. 12 1. 12 30 3 3 30 3" 273 277 263 255 28 500 19 loo 24 2OO 21 IOO 13 ioo 13 400 10 500 -97" 7400 172 168 157 148 no 201 195 M4 138 27 900 25 500 23 200 20 200 14 200 1 3 600 9 200 I I OOO IO OOO 153 147 88 83 "5 184 170 139 128 5 ioo 0650 O IOO 9400 10 800 8 900 8 Soo 9900 5 Soo 94 84 SS 72 84 122 99 122 90 12 OOO 8 500 6 ioo 6 200 9300 6 200 5400 6 400 3900 25 26 28 5.00 A.M. 8.30 P.M. 3.00 A.M. 75 81 87 G--A G-A G-A G C o.So o 92 1.26 "3 IOO 98 9300 18 200 6 700 1 08 86 92 8 800 15 700 6 900 79 77 102 6 300 IO IOO 6 too 55 75 85 5 500 8 400 5 200 G B G G lS4 166 : .:::::: ,, 08 G B I 46 ^9S Not Not Q8 G B I 46 .. i , 165 G C .. i< . >i ,, G B 368 82 G C 368 28O G B I 88 .1 .. ,i .1 ii D B ,, ii ii 278 c " O6 D B Tfil 06 D B 605 581 (| D A 616 V. D A c8s 64 XI D A n n D A 158 " 26 " "6 12.00 M. 13 D-B D A 0.038 425 413 426 }Sl 237 28l " 26 ir. " 16 A A " 27 I2.3O A.M. 17 18 A-A A A 2.00 622 673 583 620 145 6 " D A of 586 567 D A J 65 " 28 . A A 94 26 A A 408 389 A A 2 OO M E A 6 88 7 IOO 30 July i " i g.OO P.M. 6.IO A.M. 11.30 " 30 31 31 A-A A-A A A 2.98 2.52 2.52 308 297 279 6 ioo 4 900 4900 297 280 256 4 700 4 600 250 253 248 4 200 144 112 "5 2 600 I 300 UNIVE INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, /-V//7. 369 SUSPENDED MATTER AND BACTERIA. Concluded. 1). ill- Trea Riv r Water. Effl uent of Eft uent of Kffluei t of Jewell Has n No. i Bas n .So. -2. Setilin -Chamber. = S II 3 U ^ 1 o 1 II u u . II 1 Day. Hour. U ^ -c ~ s Is. f! i Is. 11 1! !i || i ~ i 2 | |I DO I s - ! " j^ I s 3 ~^ lulv I I2.OO P.M. 1 32 D-.A 0.078 2lS 2 900 214 IT r 2 2OO 116 J u 1 V 1 2 6.OO A.M. T 33 C-A 0.037 I 59 2 2OO 172 1 900 167 71 I OOO I 2 . (X) M . i ^ C-A 0.028 151 2 2IO 148 2 1 OO no " 2 8.30 P.M. i 3 6 A-B 5 000 140 6 ioo 128 4 600 62 2 7OO " 3 5.30 A.M. 137 A-C 2.06 35 4 Soo 143 4300 139 2 500 170 I 400 6 4.30 P.M. 140 A-B 1.67 121 3 900 7 1 I 3IO 6 9.00 " 140 A-B 1.67 127 3 800 57 I 7OO " 7 1 1 . OO A.M. 141 A-C i. 59 552 12 IOO 191 3 700 * 7 2.30 P.M. 142 A-C 2. 02 753 12 6oO 324 6 600 " 7 4 . oo 142 A-C 2-O2 343 13 ioo 226 3 7oo " 7 10.00 " 144 A-C 2.6o 455 400 3 45 8 4.OO A. M. 145 A-B 1 .92 438 14 ioo i 355 8 3-OO P.M. 146 A-B 1 .99 633 1 7 900 M3 3 300 " 9 5 OO A.M. 147 D-B .089 488 7300 Not in use Bas in No. 2 181 3400 9 10.00 " MS D-B 099 419 5 900 was not used 138 2 IOO 9 5.OO P.M. 149 A-B 2.41 455 6 300 " " again after Run lift i 400 9 9.00 " 149 A-B 2.4! 478 7400 No. i ,6. 148 i 890 9 12.00 " 150 D-B .09! 445 S 500 " " 57 2 300 " IO 1.30 A.M. 151 D-B . .096 430 9 ooo 84 2 150 " 10 3.00 " 152 D-B .104 312 9 600 123 I 700 IO 5.OO " 1 53 A-B 2.69 32O " IO 9.00 " 153 A-B 2.6g 384 7 100 124 850 " IO 2.30 P.M. 155 B-B 2.66 379 9400 " " 85 2 900 " 10 5.40 " 156 A-B 2.69 305 9400 142 3 ioo " 14 9.30 " 157 A-D 1.85 9 5 600 186 3 900 92 i 380 " 14 11.30 " i 5 s A-E 1-81 95 5 600 182 4 600 77 4400 " 15 5 00 A.M. 159 A-D 2.22 : 99 5900 183 4 700 107 2 2OO " 15 9 oo " 159 A-D 2.22 :1 71 4600 IOO 2 7OO 71 Soo " 15 10.30 " 1 60 A-B 2.IS 4 71 4600 127 4 600 (>s I IOO " 15 3.00 P.M. 161 A-E 2.ig 5 So 6 ioo 54 3 900 124 2 4OO " 15 8.00 " 162 A-D 2. 16* 85 6 ioo 135 i 930 7 J 210 " 15 11.30 " 163 A-E 2.19 s 89 6000 142 2 5OO 139 I 700 1 16 4.00A.M. 164 A-B I. Si 89 5 800 Not in use 1 20 980 " 16 II.OO " 1 66 A-B 2.14 89 5 300 " 9i i 500 " 16 10.05 P.M. 167 A-B 2.82 29 59oo " " " 79 760 " 17 4.30A.M. 168 A-D 2.32 52 6400 23 940 o 840 " 17 2.OO P.M. 1 68 A-D 2.32 31 6 ioo 71 i 400 3 I 190 " =7 5.30 " 169 A-E 2.30" 25 6 ioo 55 i 250 25 I 180 " 17 12. OO " 169 A-E 2.30 19 5 600 39 i 150 17 680 " 18 6.OO A.M. 170 A-D 2 2O 9 13 4 5o 26 20 810 " 18 II.OO " 170 A-D 2.29 9 09 3900 26 gSo 22 770 " 18 5.30 P.M. 171 A-E 2.32 oS 4800 42 780 33 770 " 18 I2.OO " 172 A-D 2. IS" 09 3900 9 Sio 6 730 " 9 10.00 A.M. 172 A-D 2.18" 20 3 800 37 i 080 17 810 " 9 9.30 P.M. 173 A-F I.5I 78 34oo 16 590 20 57" " 9 12.00 M. 173 A-T I.5I 78 3000 M 860 18 700 " 20 5.00 A.M. 173 A-F 1.51 75 2 7OO 21 SSo If) 495 " 20 4.00 P.M. 73 A-F 51 66 3250 SI l 020 31 960 " 20 10.00 " 1/4 A-B 1. 60 70 2 550 Not in use 49 I 1 20 " 21 9-00 A.M. 174 A-B 1. 60 76 3350 43 2 I So " 21 3.30 P.M. 75 A-C 1. 60 130 4 850 " 86 2 780 " 21 II.OO " 176 A-B 1.02 130 4 250 70 I 670 " 22 5-00 " 178 D-B .040 128 4780 15 3400 " 22 II.3O " 79 A-C 1.04 127 4 800 37 I 900 " 23 1 . 3O A . M . 1 80 A-D 0.7S 140 4600 " " " 1 300 " 23 9.00 P.M. 182 A-G 9-77 7 6 200 84 3400 82 3 600 ; 23 II.OO " 183 A-B 0.77 173 6 200 Not in use 35 3800 3.30 A.M. 184 A C I. 10 174 5 300 163 24 " 24 5.30 " 185 A-B 1.03 1 *t 173 5300 .. .. 108 37 WATER PURIFICATION AT LOUISVILLE. Final Summary, showing the Leading Results of Operation. In the following table will be found a sum mary of all of the leading results of opera tion of the Water Company s devices. The devices, methods of operation, and general periods of operation, have already been pre sented, as well as complete tables of analy tical results, it is only necessary, therefore, to explain the various headings in the table. Settling Basins in Service. Under this heading the basins used for preliminary treatment of the water previous to its en trance to the Jewell settling chamber are given. It is understood that the latter was in use at all times. Treatment. For economy of space letters have been used in the two columns un der this head. The letters refer to the co agulant used, or place of application, as follows: A. Hydrate of alumina from sulphate of alumina. B. Hydrate of iron from persulphate of iron. C. Hydrate of alumina prepared electro- lytically from aluminum. D. Hydrate of iron prepared electrolyti- cally from iron. E. Hydrate of iron from copperas. F. K. with caustic soda. The caustic soda was applied to basin No. 2, and the copperas at the inlet to the Jewell settling chamber. A. Equal amounts at basin No. 2 and Jew ell settling chamber. 1). Jewell settling chamber. C. Top of filter. (Outlet of Jewell set tling chamber. D. Basin No. T and Jewell settling cham ber. E. Basin No. i and top of niter. E. Basin No. i. G. Basin No. 2. In all cases coagulant was applied at the inlets of the settling basins or chamber, un less otherwise recorded. Grains of Chemical per (iallon. The total amount of chemical used per gallon of ap plied water is given under this heading. \Vhere it was applied at more than one point the separate amounts are given as foot-notes in the order they were applied. lilectric Current. The electric horse-power per million gallons of treated water per 24 hours is given from calculations based on the amperage and voltage of the electric current. Ampere hours of electric current per gallon of treated water is used to express the amount of electrolytic treatment. For the amounts of metal used see discussion in the last portion of this chapter on the rate of electrolytic decomposition of the metal. Average Suspended Solids. In so far as was feasible, the suspended solids in the river water, and in the water above the sand layer in the filter, were determined for each run. The results are given in these two columns. The headings in the balance of the table have already been described in Chapter VIII, and do not need further explanation. On runs Nos. 154 and 155 the free acid in the persulphate of iron was neutralized by caustic soda. Several runs were not continued to their normal length, owing to either the comple tion of the special study for which the run was made, the necessity of taking up other work at a certain time, or. in two cases, by closing operations for the week. Where the run ended on a good water for these reasons, the period of service is marked with a star (*); and it is to be noted that in most of these cases the bacterial efficiency is probably lower than it would have been had the run been continued to its normal end, as the samples collected at the beginning of the run gener ally represented the poorest water of the run. INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 1.S97. 371 12 ?: O O OOOO OGC t ; O N O o o q coo 88883 83888888 88 8 jodoj,iv E E E E E E E E E J E E E E E E E E | E E E | E E E E E E | E O O M -000000" . S. SS. S . . S. S S S S . X . . ~ *. i- <~ ~ -~ <. -~ < i. <. " WATER PURIFICATION AT LOUISVILLE. INVESTIGATIONS Ol- THE WATER COM PA NY FRO At APRIL TO JULY, 18W. 373 .30 i i .H = III O O O O O Of S 8 2 a ^ jnm i-iL uuv o" o o 1 o" -r f ~- c> r^ m -f- o O O O O O O O 374 WATER PURIFICATION AT LOUISVILLE. z : : : : : z : : ss .: i : . js O vO O O CO HI O > O O O O O v. -Oi-O : s R : : O O> i-i O t 1 O O O i- INVESTIGATIONS OF THE WATER COMPANY FROM APRIL TO JULY, 1807. 375 at a C i ! Ul 2i&3 s r s ! D en O /i O l^ u-i O c*l O "- r -7 . < 3" " ,-?. I J3)IT A \ J3A1JJ ft: O O O O O O&- U XX OXXXXU a- o w m -r i OUTLINE OF THE METHOD FOLLOWED IN THE DISCUSSION OF THE RESULTS OF THESE INVESTIGATIONS. In the following pages are presented the full discussions of the results of the investi gations made during 1897 U1 connection with the devices arranged by the Water Company, or in connection with laboratory experiments on a small scale. These discussions are pre sented in fifteen main sections as follows: Section No. i. Purification of the Ohio River water by plain sedimentation. Section No. 2. Account of the commercial chemicals available as coagulants for the ( )hio River water, and of the manner of their behavior when applied to this water. Section No. 3. Status at the beginning of this portion of the investigation, with a general description, of the formation of coagulating chemicals by the electro lytic decomposition of metal plates. Section No. 4. Detailed account of the elec trolytic formation of iron hydrate in the Ohio River water. Section No. 5. Detailed account of the elec trolytic formation of aluminum hydrate in the Ohio River water. Section No. 6. Relative efficiency of available coagulants based on equal weights of metal used, and also on the amount of electric current in the case of electro- lytically formed coagulants. Section No. 7. Economical application of coagulants, in terms of sulphate of alu mina, to aid in the removal of suspended matter by sedimentation. Section No. 8. Effect of the period of co agulation of the Ohio River water be fore filtration. Section No. 9. Degree of coagulation of the water before filtration and the minimum amount of coagulant required for that purpose. Section No. 10. On the conditions of suc cessful filtration. Section No. u. Quality of the effluent, after proper sedimentation, coagulation, and filtration independent of the nature of the coagulant. 376 WATER PURIFICATION AT LOUISVILLE. Section Xo. 12. Manner in which the nature of the coagulant affected the quality of the effluent. Section No. 13. Amounts of the different available coagulants which would be re quired with optimum conditions of sub sidence and filtration to purify satisfac torily the Ohio River water. Section No. 14. Degree to which the several coagulants in their respective amounts would affect the quality of the effluent, with its practical significance, and a con sideration of the advisability and cost of the removal of the added constituents. Section No. 15. Relative costs of equivalent amounts of the different available co agulants, together with an estimate of the yearly cost for coagulants for the pu rification of the Ohio River water. SECTION No. i. PURIFICATION OF THE OHIO RIVER WATER BY PLAIN SEDIMENTATION. Plain sedimentation means the removal of suspended matters from the water by gravity in the absence of any coagulating treatment. In many cases in this report subsidence is used synonymously with sedimentation; and in some places other expressions, such as set tling and settlement, are also used in refer ring to this same action. In the early summer of 1896 a series of sedimentation experiments upon a ^ small scale were undertaken to show the relation of coagulation and period of subsidence. The results of these experiments have been re corded in Chapter IV, where it will be seen that a number of them throw light upon the present question. As a matter of conven ience, the results of those experiments in which no coagulants were used are repeated, as follows: Percentage Removal. Parts per Million. -24 Hours. 48 Hours. 590 936 320 220 52 32 41 . 73 64 62 60 26l 44 74 Averages 469 41 67 From a practical point of view the condi tions under which the above experiments were carried on were abnormal in two re spects. In the first place, the diameter of the tank (2 feet) was such that the friction of the water upon the sides caused the vortex mo tion of the suspended particles to decrease more rapidly than in large basins or reser voirs; or, in other words, it caused the water to reach a state of rest more quickly. Owing to the fact that subsidence is very closely as sociated with the vortex motion of the par ticles, the above results, so far as this point is concerned, show higher percentages of re moval than would occur in practice. Sec ondly, the high and varying temperature of the boiler-house in which the tank was placed caused the presence of currents which re tarded the subsidence by increasing the vor tex motion. The significance -of these points is shown clearly by the results of plain sedimentation experiments made in one-gallon bottles, which were kept at approximately the same temperature. It will be noted that the average amount of suspended matter in these experiments was substantially the same as in the case of those recorded in the last table; and the general character of the water was fairlv similar. R 1 r w ,, cr Parts per Million. 24 Huurs. 48 Hours. 521 81 84 5 ,6 472 428 33S So 86 76 80 71 86 77 86 A verages 455 77 S 4 In addition to showing that by 24 hours of quiescent subsidence about 75 per cent, of the suspended matter in fairly normal muddy water may be removed, the above results demonstrate that under these conditions eco nomical subsidence cannot be carried beyond this period (24 hours). Comparing the last results with those in the first table, it is clear that the conditions of subsidence, as already noted, were important factors, and that 24 hours subsidence in gallon bottles was more efficient than 48 hours in the settling tank SUMMARY AND DISCUSSION OF DATA OF 1S!>7. 377 placed in the boiler house. In this connec tion it may be stated that some analyses made in June, 1896, of the water leaving the Cres cent Hill reservoir, which holds about 6 days supply, indicated a removal by subsidence of about 60 per cent, of the suspended matters in the general class of water under considera tion. Concerning the efficiency of basin No. i in the removal of suspended matters by plain subsidence, the summary of results on pages 371 to 375 show that, during the early spring, when the suspended matter was rather coarse, the removal ranged from 20 to 50 per cent., when the basin held about 3 hours supply. \Yith the water later in the spring, and in the summer, when the suspended particles were much finer, the removal of suspended matters in this basin ranged from o to 15 per cent., but for the most part the water as it left this basin showed no substantial purification. With regard to the removal of bacteria by plain subsidence, the influence of the tem perature is considerable. In general terms it appears that the percentage removal of bac teria follows quite closely that of the sus pended lifeless particles. This is to be ex pected, because the bacteria to a considerable degree appear to be attached to the grosser suspended particles. The removal of bacteria by plain subsidence is not a very important factor, however, because with little or no ex tra expense to the general process they may be removed subsequently by the necessary coagulation and nitration. The character of the suspended matter in the river water is a point of great importance in purification by subsidence; and, farther, the amount of suspended matter influences materially the percentage removal. The lat ter point in a measure follows from the firs^ because when the particles arc large the stage of the river, etc.. is such that the total weight of the suspended matter is bound to be great, comparatively speaking. This is shown by the following results, obtained by plain sedi mentation in one-gallon bottles. Farts per Mill! pended Matter Percentage Removal in 3 h " 24 "48 of Original Sus- </>5 85 96 130 17 32 41 The above experiments show that, with the first water (Xo. I), containing coarse and heavy particles, the suspended matters de creased from 965 to 29 parts; while in the third water containing fine clay particles, the corresponding decrease was from 130 to 77 parts. This brings out the important fact that with the water containing clay particles, such as is found here for two or three months in the year, the removal of sus pended matter is not only much less rapid than in the water containing the heavier mud and silt, but the amount (weight) of sus pended matter at the end of practicable lim its of subsidence is greater in the case of the clay-bearing water. In fact, with the third water it would require weeks, if not months, to remove from it substantially all of the clay. With the second water, resembling a mixture of the other two. the removal of suspended matters was intermediate in its nature. From the above statements it will be per fectly clear that the efficiency of plain sub sidence depends very largely, so far as any given period is concerned, upon the condi tions under which subsidence takes place. The main thing is to bring the water into a state of comparative rest, in order to reduce the vortex motion of the particles due to eddies and similar movements of the water. Ex perience shows that the water is brought to a state of rest much more quickly in small receptacles than in large reser voirs such as would be required in practice. The results obtained from small experimental devices, accordingly, can be taken only as indications in general terms of what may be accomplished practically in this manner, and as a guide for the construction of large subsiding basins might be quite mis leading. Furthermore, it was not considered feasible during these investigations to secure conditions on a sufficiently large scale to en able the efficiency of plain subsidence on a practical basis to be studied in a thorough manner. The following conclusions upon the purification of the Ohio River water at Louis ville, therefore, are in part presented in gen eral rather than in specific terms. Conclusions. i. It is possible to remove, economically, 378 WATER PURIFICATION AT LOUISVILLE. about 75 per cent, of the suspended matter in normal muddy water by plain sedimenta tion (subsidence). At times of freshets dur ing the winter and early spring the percent age might exceed 90; while in the late spring and summer it might fall to 50 or less. 2. The removal of bacteria by plain sedi mentation follows the removal of suspended matter in a general way. But the removal of bacteria by this portion of an efficient sys tem of purification is not important, because they can be effectively disposed of subse quently by filtration without extra cost in the operation of the complete system, and their removal by the niters does not affect the quantitative efficiency of nitration, as is the case with mud and other suspended matters. 3. After treatment to its economical limits by plain sedimentation, the Ohio River water would ordinarily be discolored by suspended matters, which it would contain in sufficient amounts to preclude the probability of growth in the open subsidence basins of algae and other organisms, giving rise to objection able tastes and odors. 4. At times of freshets during the spring and summer all the evidence goes to indicate that plain subsidence cannot economically re move a sufficient amount of the fine clay car ried in the Ohio River water at Louisville to prepare the water satisfactorily for filtration; and, regardless of whether the English or American type of filter is used, economy de mands the use of coagulating treatment to aid subsidence at such times. 5. The period of economical plain subsi dence of the Ohio River water does not ex ceed the equivalent of 24 hours quiescent subsidence, such as is secured in one-gallon bottles. With the heavy coarse mud of the winter freshets this period is doubtless shorter than 24 hours; and in the case of the clay- bearing waters, for which the use of coagu lants is imperative, the period could to ad vantage be somewhat shorter than this. But with the intermediate class of water, resem bling a mixture of these extremes and illus- tatecl by No. IT water in the last table, we are led to believe that plain quiescent subsidence could be economically carried to 24 hours, or to very nearly that period. 6. With regard to the period of subsidence, under the conditions of practice, equiva lent to 24 hours quiescent subsidence in one- gallon bottles, the available conditions of these investigations were not such as to make the solution of this problem feasible. 7. Concerning the arrangement of the sub sidence basins, with reference to size, depth, and location of division walls; and their op eration, with regard to the question of con stant How, intermittent flow, or successive fillings and drawings, there are no data available from these investigations, and it will be necessary to rely upon information from other sources. SECTION No. 2. ACCOUNT OF THE COMMERCIAL CHEMICALS AVAILABLE AS COAGULANTS FOR THE OHIO RIVER WATER AND OF THE MANNER OF THEIR BEHAVIOR WHEN APPLIED TO THIS WATER. In this section it is the purpose to take the entire list of metals and show which of them form commercial compounds capable of co agulating the Ohio River water in a safe manner. The way in which the leading available compounds of the suitable metals are decomposed when applied to this water is described, together with the relative advan tages and disadvantages of each. A fall de scription of the formation of coagulating chemicals by the electrolytic decomposition of metal plates is presented in sections Nos. 3, 4, and 5. In Section No. 6 the relative efficiency and economy of the several avail able coagulants is shown, and in section No. i 5 a comparison is made of the costs of the chemicals most adaptable for this use. Chap ter III contains a description of the action of sulphate of alumina when applied to the Ohio River water, as far as it was understood in 1896; and this portion of the present chapter is, in a measure, an elaboration of Chapter III, and contains all of our additional information upon this subject at the close of these investigations. Classification of Afctals in their Applicability to the Purification of the Ohio River Water. The next table contains a list of all rnetals, excepting the rare and precious ones, sub- SUMMARY AND DISCUSSION OF DATA OF 1807. 379 divided into groups according to their gen eral adaptability for the purpose in question. In the first column are given those metals which are either well-known poisons, or which in small quantities are regarded in the absence of precise data as suspicious from a hygienic point of view. The second column contains those metals the normal compounds of which form soluble salts when added to this water; and in the third column are found those metals capable of forming in the Ohio River water compounds of a solid granular nature, wholly or partially insoluble under the existing conditions. In this connection fre quent use will be made of "precipitate," which is the chemical name for a solid compound. Finally, there are presented in the fourth col umn those metals which form insoluble and gelatinous precipitates when applied in a suit able manner to this water. CLASSIFICATION OF METALS. (Division No. i. Division No., Permissible Me als from a Sani ary Standpoint. Group No i. * a Sanitary Group No. i . Forming Granular Group No. 3. Forming EB Standpoint. Foi min^ Soluble Compounds. Precipitates Partly or Wholly Geh tinous Insoluble Precipitates. Insoluble. Lead Sodium Calcium Aluminum Silver Potassium Magnesium Iron Mercury Manganese Tin Antimony Arsenic Copper Bismuth Cadmium Nickel Cobalt Zinc Barium. Strontium The metals of division No. I, in the light of our present knowledge, cannot be considered as applicable for this work. Taking up the metals of division No. 2, the use of the several groups may be briefly outlined as follows: Group No. T. These metals may be use:l in the form of hydrate (caustic soda) to remove carbonic acid from water. This treatment pro duces sodium carbonate (if caustic soda were added), which will decompose incrusting con stituents (permanent hardness). Sodium or potassium may also be added directly to water in the form of carbonates or tribasic phosphates for the purpose of removing in- crusting constituents. Group No. 2. These metals may be applied to the water in the form of hydrates (e. g. lime water) in order to remove carbonic acid. From such an application there is formed calcium carbonate, which is soluble in water free of carbonic acid, to the extent of 2 to 3 grains per gallon. Quantities in excess of this amount settle out, upon standing, in the form of a fine white powder, which has little or no power as a coagulant. Group No. _?. These metals may be added to this water in proper quantities and in a suitable form with the result that ultimately an insoluble gelatinous precipitate is formed, capable of coagulating the suspended matter. The metals of groups Xos. i and 2 refer solely to metals for the reduction of corrod ing and incrusting constituents, and will be taken up subsequently in connection with these matters; at present we will consider the metals of group No. 3, which are the o:ily ones available for the coagulation of the water, preparatory to subsidence of fine clay and the rapid filtration of the water through sand. Most Suitable Compounds of the Metals (Group No. 3) capable of producing Coagulating Precipitates, and a General Description of their Behavior upon Application to the Ohio River Water. A comparative outline of the leading com mercial compounds (salts) of these metals is as follows: Compounds of Aluminum. In addition to sulphate of alumina and pot ash alum there are several other commercial compounds which have been investigated in the laboratory. It was explained in Chap ters II and IX that the sulphate was the bet ter of the two former: and accordingly this compound will be briefly described and the others" referred to it in comparative terms. Sulphate of Alumina. The behavior of this 3 Ho WATER PURIFICATION AT LOUISVILLE. chemical when added to the Ohio River water has been fully described in general terms in Chapter 111. As a matter of con venience it may be repeated that, briefly, it is decomposed for the most part by the alkaline compounds (lime and magnesia) in the river water; and that the increase in carbonic acid and incrusting constituents in the water is proportional to the decrease in alka inity. The rate of decrease in alkalinity (Ci to <) par.s per million for j grain per gallon of the or dinary chemical) depends upon the amount of sulphuric acid in the chemical and the amount of suspended matters in the water capable of absorbing this compound. The alumina in the commercial product is pre cipitated and removed by sedimentation and filtration, while the increased carbonic add and incrusting constituents (principally sul phate of lime) remain in the water. It has already been made plain that the two latter additions to the water are not desirable from an industrial standpoint, although they do not injure the sanitary quality of the water, when the process is carried on under suitable conditions. From an economical point of view the amount of sulphate of alumina wasted by absorption by the surfaces of the suspended particles of mud and silt, and by the organic matter, is a matter of much im portance. For the sake of explicitness this topic for all the chemicals is discussed by it- se f in this section just after this more general account. Potash Alum. The crystals of this com mercial chemical are a mixture of sulphate of alumina and sulphate of potash. The latter portion is of no practical influence in water purification, while the sulphate of alumina in it behaves in a manner similar to commercial su phate of alumina as described in the last paragraph. Potash alum contains only about two-thirds as much sulphate of alumina as the commercial form of this last chemical; costs substantially the same; possesses no advan tages in current methods of use; and, there fore, is eliminated from the problem on the ground of cost. Chloride of Alumina. This compound be haves in a precisely similar manner to sul- nhate of alumina in forming a precipitate of hydrate of aluminum and reducing in the same ratio the alkalinity, with the formation of car bonic acid and incrusting constituents. The only difference is that the increased amounts of incrusting constituents would be composed of chlorides of lime and magnesia in place of the sulphates of these metals. This change would produce only a very slight and nomi nal difference in the character of the water, because when heated in a steam-boiler under pressure of 50 pounds the added chloride of lime reacts with the sulphate of magnesia originally in the water, and the effect is simi lar to the conditions when commercial sul phates are applied. There probably would not, be enough magnesium sulphate in the water to complete this change at all times; but even in this event it is to be stated that the magnesium chloride formed from this chemical by the above reaction is the com pound which is most injurious to boilers, as it is decomposed by heat in boilers with the formation of free hydrochloric acid. This chemical is more expensive than sul phate of alumina, because the hydrochloric acid used in its preparation is more costly than sulphuric acid; and as there are no sub stantial advantages to offset the increased cost its use is not practicable. Acetate of Alumina. Tn trade this chemical is known as " red-liquor " and is used in dye ing. Tt is decomposed by the alkaline con stituents of the river water the same as sul phate of alumina, with the formation of alu minum hydrate and the same rate of reduc tion in alkalinity and increase in carbonic acid. The other resultant compounds, ace tates of lime and magnesia, in place of sul phates, arc soluble and would appear in the filtered water. They are not injurious to health, and do not act as incrusting constitu ents. The absence of increased amounts of the latter compounds would be desirable, but, as the acetate costs about four times as much as the sulphate, the evidence in section No. 14 of this chapter shows that the use of this chemical would not be advisable. Sodium Alnminatc. This compound of alu minum differs essentially from sulphate of alumina in that in this case the aluminum ads as an acid instead of a base. When car bonic acid is applied to sodium aluminate so lutions in certain industrial chemical proc- SUMMARY AND DISCUSSION OF DATA OF 1897. esses aluminum hydrate is formed and so dium carbonate appears as a by-product. In water coagulation such an action would be very desirable if the conditions allowed it to be have like this, as the same gelatinous hydrate would be obtained, with no increase in cor roding or incrusting constituents; in fact the latter would be reduced because the sodium of the applied chemical would unite with car bonic acid to form sodium carbonate, which in turn would decompose an equivalent amount of incrusting constituents without forming any objectionable compounds. In other words, the single compound would give the combined effect of the metals of groups Xos. i and 3 of division No. 2 of the table. Experience, however, showed that its use was impracticable in the case of the Ohio River water because it would not decompose in the manner stated. The reason of this ap peared to be that the solution in the river water of this chemical and of carbonic acid was too weak. Compounds of Iron. Owing to the fact that iron is a cheap metal, and that its hydrate in the oxidized or ferric state is an excellent coagulant, these compounds are entitled to careful considera tion. At the outset it is to be recalled that there are two series of iron compounds, the ferrous (incompletely oxidized) and the ferric (completely oxidized). \Ye shall first con sider the ferrous compounds. I crrous Sulphate. This is also known as the protosulphate of iron and as green vitriol, and is the cheapest form in which iron com pounds are on the market. When added to the Ohio River water it acts similarly to sul phate of alumina except that ferrous hydrate is formed in place of aluminum hydrate. With equal weights of metal the reduction in alkalinity and increase of carbonic acid and incrusting constituents by ferrous sulphate and sulphate of alumina are in the ratio of i.o to 1.3. Ferrous hydrate is not a suitable coagulant because it dissolves in the water to the extent of about / parts per million: and to make the iron compounds available it is necessary to have the iron ultimately in the form of the ferric (oxidized) hydrate. In the case of ferrous sulphate there is enough atmospheric oxygen dissolved in the water to accomplish this under favorable conditions. Experience, however, shows that this is im practicable in this water, owing to complica tions in the oxidation caused by carbonic acid. When ferrous sulphate is added to this water, white ferrous hydrate, mostly insoluble, is formed. Very quickly this precipitate passes into solution, due to the action of carbonic acid and resulting probably in the formation of a soluble basic carbonate. When the iron is in this form the atmospheric oxygen, al though present in excess, oxidizes it very slowly and with great difficulty. Further more, the iron when it does reach the oxi dized state does not form the normal gelat inous ferric hydrate, but a partially granular compound which is some lower hyclration of ferric oxide as nearly as could be learned. In short, the carbonic acid in this water renders the use of ferrous sulphate (and all other ferrous compounds) inadmissible for coagulation, owing to the passage of dis solved iron through the filters. To remove the carbonic acid before applying the ferrous compounds would be too costly to be prac ticable. Ferric Sulphate. Of the commercial forms of iron in the oxidized or ferric condition, ferric sulphate or persulphate of iron is the best one for this line of work when economy is considered, for the same reason that the sulphate is the best compound of aluminum. Ferric sulphate is decomposed by the alka line constituents of the Ohio River water in a manner precisely similar, so far as could be learned, to sulphate of alumina. The result ing precipitate of ferric hydrate is very gelat inous and is insoluble: therefore it makes an excellent coagulant. With equal weights of iron and aluminum, in the form of sulphates, the ratio of the decrease in alkalinity and in crease in carbonic acid and incrusting con stituents is i.o to 2.1. The waste of ferric sulphate by absorption on the surfaces of silt and mud is similar to that in the case of sul phate of alumina. Commercial ferric sulphate is a little cheaper than sulphate of alumina, free of water of crystallization, and contains about three times as high a percentage of metal. It was these features of the compound that WATER PURIFICATION AT LOUISVILLE. originally attracted our attention. It is diffi cult to dissolve, and the sample with which our tests were made contained some free sul phuric acid and insoluble residue. Neutra! ferric sulphate can be procured without ^ diffi culty, however, and the suspended particles could be removed from the solution by ready means. Metallic Iron b\ flic Anderson Process. This process consists in obtaining in the water by contact with metallic iron a carbon ate (ferrcr.s) of iron by the solvent action of the carbonic acid in the water, and the oxi dation of this compound to insoluble ferric hydrate. It is referred to in Chapter IX, page 244. Bottle experiments indicated that its use with the Ohio River water was not satisfactory, owing to the retarding action of large amounts of carbonic acid such as are present for months at a time in this water. The nature of this retarding action is similar to that in the case of the protosulphate of iron. Aeration was tried on a small scale to sup plement this action, but it did not work well. The oxide was granular in form, showing the absence of normal hydration, and the value of the iron as a coagulant was lost for the most part. Compounds of Manganese. Manganese forms manganous, manganic, and permanganate compounds. The man ganous compounds cannot be safely used to advantage with this water, owing to compli cations with carbonic acid in the manner ex plained in the case of ferrous sulphate. Manganic compounds in a suitable form arc not on the market. Permanganates of lime and potash were used in the laboratory; they are manufactured in considerable quantities, but they cost, according to the best quota tions. $12.70 and $0.40 per pound, respect ively. Their expense renders their use in admissible for purification of municipal sup plies. A study of them, however, has made plainer our understanding of the coagulation of the muddy Ohio River water, and for the sake of completeness they will receive brief consideration. Permanganate of Potash. When added to the Ohio River water in proper amounts the organic matter slowly withdraws oxygen from this compound, and the carbon and hy drogen of the organic matter are oxidized to carbonic acid and water, respectively. The result is that after a time the manganese is converted to mangano-manganic hydrate, which is a gelatinous, insoluble precipitate. Experience shows that the action is very slow, at least 3 hours ordinarily being re quired for its completion; but the time varies with the amount and character of the organic matter. \Yhen the reaction is completed the manganese does not pass through the filter, but it will do so until it is converted into the insoluble hydrate. As fast as the carbon di oxide is formed it unites with the potash of the applied chemical to form carbonate of potash, which is an alkaline but not a corrod ing or incrusting constituent. The nominal increase in alkalinity is the only change in the composition of the filtered water, as the re moval of organic matter would be effected by subsidence and filtration independent of this oxidizing action. in addition to the slowness with which this action takes place this process developed an important fact that the manganese com pounds are not at all or very little absorbed by the surface of the mud or silt. Permanganate of Lime. The behavior of this chemical when applied to the Ohio River water is precisely similar to that of perman ganate of potash, except that the resulting carbonate in this case is that of lime instead of potash. Experience shows that the sulphates of alu mina and ferric iron are the most suitable commercial chemicals for the coagulation of the Ohio River water. In order to make more explicit the next topic, on absorption of coagulants by silt and clay, the permanga nates will be briefly reviewed in comparison with the sulphates, although the former are too expensive and are incapable of being ap plied in sufficient amounts to be practicab e. For convenience we will refer to the sul phates as type A, and to the permanganates as type B. SUMMARY AND DISCUSSION Of DATA OF 1897. 33 /. Nature of Reaction. This has been care fully explained above, but in brief a type A chemical is partly and as a rule mostly de composed by alkaline constituents, while the remainder is absorbed by the surface of the matters in suspension. The latter action ap pears- to be largely if not wholly a chemical one. With type B the dissolved organic matter decomposes the chemical, and a re sultant gelatinous precipitate is formed. So far as we could learn type B appears to be affected not at all, practically speaking, by absorption by silt and clay, and its reaction progresses with suspended organic matter only so fast as the latter becomes disinte grated and passes into solution. 2. Gcrmicidal Action. The chemicals of each type if applied in sufficiently large quan tities will destroy bacteria. But when ap plied to the water in such amounts as are practicable for the purification of a municipal water supply they do not kill bacteria, prac tically speaking, although they cause many of them to die either by direct effect or by en veloping them in masses of coagula. In any case under practicable conditions in this con nection the destruction of bacteria would not be complete from a hygienic point of view. ?. Speed of Reaction. With type A the reaction is completed almost instantaneously, although there are indications that at times there is a selective action in respect to the alkaline constituents and the suspended par ticles which absorb the chemicals. Concern ing the time which elapses before the coagula appear in suitable size for efficient subsidence and nitration, this period deals wholly with the period of coagulation following the initial reaction, which occurs immediately. In the case of type B the initial reaction takes place very slowly; in fact it would probably never be complete in less than 3 hours, and in many instances it would con tinue for more than 24 hours. This is due to the nature of the reaction as explained above, as the chemical has first to disinte grate and make soluble a large part of the organic matter which is oxidized. 4. Safe Maximum Limit of Application. The maximum limit of safe application of type A depends upon the absorptive capac ity of suspended matters, and the alkalinity of the river water, and the amount of sul phuric acid in the applied chemicals. Ex pressed in grains per gallon the range of maximum application of sulphate of alumina would be from 4 to 15. For persulphate of iron these figures would range approximately from 3 to 10 grains per gallon. In the case of permanganate of potash, type B, the safe maximum application would range from o. i to 0.2 grain per gallon, with a period of reaction of not less than 3 hours. 5. Applicability in the Purification of the Ohio River Water. The permanganates, type B. are not applicable to this problem because of their cost, the slowness of their action, and the low limits in the amount of safe application. A study of them, however, was very fruitful in showing inherent weak ness of type A chemicals, and indicating how those weaknesses might be remedied in part in practice. They relate to absorption and are discussed as the next topic. ABSORPTION OF COMMERCIAL SULPHATES OF ALUMINA AND OF FERRIC IRON BY THE SILT AND CLAY IN THE OHIO RIVER WATER, WITH SPECIAL REFER ENCE TO THE WASTE OF CHEMICALS AND THE NECESSITY FOR THE REMOVAL OF COARSE SILT BY PLAIN SEDIMENTA TION. In the course of these investigations a number of observations were made which co operated to bring out the marked significance from a practical point of view, of a phenome non which we shall call absorption. To illus trate this by an action which is familiar to everyone, we may compare it to the some what similar observation of iron stains as they appear upon linen. This action is not the same, but it is believed that its nature is parallel. At this point it may be stated that of all the chemical actions seen in daily life, there is probably none which is more obscure than the action of liquids upon solids, as illus trated by the ones in question.. To explain these observations and facts in a comprehen sive manner is impossible in the present state of applied chemical science. Accordingly we shall present the evidence in a series of ob servations characteristic of the nature of this WATER PURIFICATION AT LOUISVILLE. phenomenon, and at the close point out its practical significance. This is shown by the following experiment, in which a series of one-gallon bottles were filled with river water containing 424 parts per million of mixed coarse and fine silt and clay. Beginning with none, the samples were treated with sulphate of alumina, each bottle being given 0.25 grain per gallon more than the preceding one. The bottles were then well shaken, and samples of the supernatant liquid removed by a siphon after 24 hours subsidence. The results were as follows: After Settli g 24 Hours. Additional Grainper Gallon. Suspended Solids. Parts per Million. Percentage Removal. Portions of o., 5 Grain. None o. 25 47 74 4 o. 50 44 76 2 0-75 35 Si 5 I.OO 3 97 1 6 1.25 i 99 2 1.50 o IOO I This experiment is not an extreme case, but it serves to illustrate the fact that with successive equal amounts of applied sulphate of alumina the work accomplished is not regularly progressive, and that for some rea son in the ordinary river water the specific efficiency of the first portion of chemicals is very low, and less than that of subsequent ones. II. Necessity of Applying Different Amounts of Coagulants to secure complete Coagulation of Equal Weights of Suspended Matters of Different Character. This was repeatedly noted in the operation of the filter, but is illustrated in a very charac teristic manner by the following experiment: Waters A and B each contained 66 parts per million of suspended matter. A repre sents unusually fine particles, while in B the particles were abnormally coarse. As in the foregoing experiment, successive portions of sulphate of alumina were added to a series of bottles containing the two waters, respect ively, and samples of the supernatant liquid were collected for analysis after the coagu lant and water had been shaken and then al lowed to subside for 18 hours. After Settlir g !8 Hours. Applied Sulphate- Wat -T A. Wat er B. Grains per Gallon. Percentage Removal of Suspended Solids. Additional Removal for Removal** Suspended Solids. Additional Removal for 0.25 Grain. 0.25 10 O 57 7 0.50 10 O 72 15 0.75 15 5 92 20 I.OO 30 ic 96 4 1.50 82 52* 99 3* 2.UO 92 10* TOO i* * For increases of 0.50 grain. These results show that 0.75 grain effected as much purification by coagulation and sub sidence with the water B, containing the coarse matters, as did 2.00 grains in the case of the water A, with very fine clay. With water A the first point in this evidence is brought out very forcibly, as 0.5 grain was applied with no purification in addition to that accomplished by plain subsidence. III. retrying Departures from the Theoretical Rate of Reduction in Alkalinity when Coagu lants are applied to Water containing Equal Amounts of Different Kinds of Suspended Matter. Tn Chapter III it was shown that theoretic ally the reduction in alkalinity would be strictly proportional to the amount of chem ical added and to the percentage of sulphuric acid contained in the applied sulphate, pro vided there were no organic or suspended SUMMARY AND DISCUSSION OF DATA Of 1897. 3*5 matters present. Data were presented a* that time showing that the departure from the theoretical reduction was dependent upon the amount of suspended matter; and here it is the purpose to show that the reduc tion is also affected by the character of the suspended matter. The following experiment illustrates this point. The five samples of river water con tained approximately equal amounts of sus pended matter, while the actual reduction in alkalinity by adequate amounts of the same lot of sulphate of alumina showed wide varia tions in departure from the theoretical reduc tion. These data show clearly that different kinds of suspended matter dispose of vary ing amounts of coagulant by an action which for the want of a better name we call ab sorption. This is most marked in the case of clay, and appears to be largely, if not wholly, a chemical action. The reason of this belief is based on the fact that there is no diminution in the conductivity of a solution in which absorption takes place, and the as sumption that this indicates the absence of a physical change which would withdraw and not interchange ions and thus increase the resistance of the solution containing these particles. The absorption of chemical solutions by various materials containing alumina has been known for some time to agricultural chemists, and at the Lawrence Experiment Station this action was found to be a factor in connection with the efficiency of filters of the English type. In the case of clay this absorption, it is important to note, produces some coagulation. \Yith regard to the co agulation of clay by other salts, such as com mon table salt, or by acids, so far as our ob servations go, there would be nothing prac ticable in their use. " Suspended Solids. Parts per Million. Percentage whicli the Actual Reduction in Alkalinity was of the Theoretical. I 500 57 2 534 74 3 534 77 4 516 80 5 558 84 Averages 550 74 IV. Conclusive Indications of the Necessity of hav ing to Saturate some Capacity of the Suspended Particles before complete Co agulation is possible. The foregoing data bring out very forcibly the fact that with ordinary conditions of the river water the first portions of the coagulant have a very low specific efficiency in purifica tion; and after a certain amount has been applied a very small additional amount causes complete coagulation, provided sufficient time is allowed to elapse after the applica tion of the coagulant. This capacity is the absorption previously referred to. and varies materially with different waters. V. Comparison of flic Relative Efficiencies as Sub siding Coagulants, of Type A (sulphates) and Type B (permanganates), and with Reference to Discrepancies between the above Relation and tim Percentages which the Actual were of the Theoretical Reduc tion in Alkalinity. At the outset it may be stated that in gen eral terms experience indicates that to secure equal efficiencies for coagulation, it is neces sary to provide substantially equal volumes of gelatinous hydrate. This is demonstrated by the- data given in section Xo. 6, but here it may be noted that ordinary commercial sul phate of alumina and persulphate of iron yield about the same volume of hydrate, and as coagulants their efficiency is substantially the same in all conditions which we have studied. On the basis of equal volumes of hydrate, sulphate of alumina and permanganate of potash should have relative efficiencies of i.oo to 1.14. In practice with the unsubsided Ohio River water the relative efficiency of sulphate of alumina was far less than this, as is indicated by the following representative results obtained from a series of experiments, in which in all cases coagulants were added to give a fairly complete and corresponding de gree of efficiency, as shown by the removal of WATER PURIFICATION AT LOUISVILLE. suspended matters by subsidence for 24 hours. The comparative efficiencies of types A and B are expressed with reference to the above ratio. Type A. I. 00 2.OO I. 00 I. 00 0.10 O.2O o 14 0.14 I. 00* PercentHge .Pcrcen ta^e which the which the . spended Actual Efficiency of olids. Reduction of Tvpe A was of rts per Alkalinity that indicated illinn. was of the by the Theoretical Type A. Ratio. 542 58 15 542 64 14 129 89 13 S3 91 -4 200 95 127"- * In this case an excess of dissolved organic matter was applied to the water so as to increase very largely the speed of reaction. Bearing in mind the fact that with type A the reaction is practically instantaneous, while in type P it is exceedingly slow, with river water, it will be understood that the coagu lating hydrate in type A is formed very quickly and in type B very slowly. It is also to be remembered that type A, but not type I!, is absorbed by suspended matter. If the absorption were the only point of difference in the behavior of the two types then the percentage efficiency which type A gave of that indicated by the theoretical ratio would correspond to the percentage which the actual was of the theoretical rate of re duction of alkalinity in a general way. If the absorption produced no coagulation of clay particles this last statement would be true in absolute terms. But it is shown in the last table that in the case of the first four (normal) waters the percentage efficiency of type A fell far below the percentage which the actual reduction in alkalinity was of the theoretical. In other words, the amount of aluminum hy drate, which was proven to be formed by the actual decomposition of alkaline constituents, failed to accomplish as much work as an equal volume of hydrate of manganese as pro duced by type B. The explanation of these results, and of others of a similar nature, was that the hy drate with type A was formed instantane ously, or nearly so, and became attached by some means to the coarse particles, which sub sided quickly and carried to the bottom much of the hydrate before it had an opportunity to coagulate those fine particles which needed it most. This explanation was proved con clusively to be correct by adding enough soluble organic matter to the last water in the case of type B to form this hydrate almost immediately, as was normally the case with type A. Under these circumstances the relative efficiency of type A as compared with type B exceeded the theoretical ratio stated above, which was based on the rela tive volumes of hydrate. It is now seen that in addition to the ab sorptive action which with coarse matters means a waste of chemicals, there is also an attachment of hydrate, in the case of sul phate of alumina and persulphate of iron, and the coarse particles, on which this attachment occurs, subside quickly and thus cause a waste of coagulants. Whether or not this at tachment is entirely physical or mechanical in its nature is not known. Conclusions. 1. The suspended matters in the Ohio River water have a certain, but varying, power of absorbing sulphate of alumina and persulphate of iron. With fine clay particles this absorption produces coagulation in a measure, but with the coarsest particles it ap pears to be a total loss of chemicals. This absorption of the coagulant causes the actual rate of reduction of alkalinity to become va riable, and the departure from the theoretical rate measures the absorption of the applied chemicals by the suspended (and soluble or ganic) matters. 2. In order to secure complete coagulation for any given water containing suspended matter it is necessary to apply a certain defi nite amount of the coagulant, which varies with different kinds and amounts of sus pended matter, in order to saturate their ab sorptive power before substantial coagula ion takes place. When applied in amounts less than this the chemicals are largely wasted. 3. Owing to the fact that with the com mercial sulphates the respective hydrates are formed almost instantaneously, the presence of coarse particles which subside quickly cause a waste of chemicals in amounts equal to the quantities of original chemical ab- SUMMAA Y AND DISCUSSION OF DATA OF 18<>7. sorbed, plus a certain amount of hydrate which becomes attached to their surfaces. The attached hydrate is thus removed before it coagulates to its full power the fine par ticles in the water. 4. The above facts are decisive proof that it is impracticable to apply coagulants to a water which contains suspended matter which may be economically removed by plain subsidence. For the sake of completeness it may be stated that in the electrolytical formation of coagulating hydrates the salts of the metals are formed initially, and the general effect is similar to that recorded for the sulphates in this section. So far as our observations ex tend at present, plain subsidence is the only practical step to take to obviate these actions in part. SECTION No. 3. STATUS AT THE BEGINNING OF THIS POR TION OF THE INVESTIGATION, WITH A GENERAL DESCRIPTION, OF THE FORMA TION OF COAGULATING CHEMICALS BY THE ELECTROLYTICAL DECOMPOSITION OE METAL PLATES. At the outset of this portion of the investi gation the evidence upon this point may be briefly outlined as follows: 1. Copper, lead, tin, and zinc are inadmis sible for electrolytic decomposition for this purpose, because the resultant chemicals are partially soluble in water, and would there fore be liable to injure the health of persons drinking the water after such treatment. 2. Aluminum and iron are the only metals of commerce which can be electrolytically de composed into chemicals adapted to the co agulation of water. The available informa tion concerning them at that time was as follows: 3. One pound of metallic aluminum, elec trolytically decomposed into aluminum hy drate, is substantially equivalent to one pound of aluminum in the form of aluminum sul phate, when the latter is applied to a water containing lime in solution. One pound of [ metallic aluminum in sheet form costs 27 cents, and one pound of aluminum in the form of sulphate of alumina costs 16 cents. The alumina in the form of the commercial chemical, therefore, costs only 60 per cent, as much as in the form of metal plates, dis- I regarding the expensive items of power, elec trolytic cells, and waste of metal in the latter case. 4. One pound of metallic iron electrolytic- ally decomposed into iron hydrate is substan- i tially equivalent to one pound of iron in the form of persulphate of iron, when the latter : is applied to water containing lime in solu tion. One pound of metallic iron, in the form of plates suitably arranged in an elec trolytical cell, costs about 2 cents; and one pound of iron in the form of persulphate of iron costs 5 cents. There was a difference, therefore, of 3 cents per pound, to cover the cost of electric power and waste of metal. This was a substantial margin on the right side, and made the electrolytic production of iron hydrate a factor in the problem. 5. Electrolytically produced hydrates fei- ther aluminum or iron) do not, as in the case of commercial chemicals like the sul phates, add to the water a strong acid, to combine with lime and increase the incrust- ing power of the water when used in steam- boilers; nor is there a practically equivalent amount of carbonic acid gas liberated, to in crease the corrosive action of the filtered water on iron vessels. In short, it will be seen that the electro lytic production of iron hydrate was a promis ing factor, while the electrolytic production of aluminum hydrate gave no indications of being practicable for regular use, owing to j excessive cost. It was decided, however, to investigate the electrolytic production of the hydrates of both of these metals. In the case of aluminum this was done, not only for the purpose of securing comparable data on the same scale as was used in 1895-96, but also with the possibility in mind that the use of aluminum during periods of maximum treat ment of the water might reduce the size of power plant, because it appeared that alu minum is decomposed with less power than iron, relatively speaking. I A description of the formation of iron by- \VATER PURIFICATION AT LOUISVILLE. drate and of aluminum hydrate, by the elec trolytic decomposition of the respective met als, is presented in considerable detail, from a practical point of view, in the next two sec tions. Before this is clone, however, it will be well to consider some of the general fea tures of electrolvsis. Electrolysis is the name of the process by which a liquid is decomposed by means of an electric current. As a rule, such liquids are aqueous solutions of various chemical salts and compounds which are capable of splitting (dissociating) into two component parts. Liquids which can be electrolyzed are called electrolytes. Absolutely pure water cannot be electrolyzed, practically speaking, and liquids possess this capacity by virtue of the chemical compounds dissolved in them. These compounds serve as conductors of the electrical current, and electrolytes are called conductors of the second class, in distinction from the metals, which are known as con ductors of the first class. A receptacle in which electrolysis takes place is called an electrolytic cell. The plates attached to the ends of the wires running from the electric generator to the cell and return are spoken of as the electrodes. To distinguish the two plates, or two sets of plates, the electrode by which the electric current enters is termed the positive pole, or anode, and that by which it leaves, the nega tive pole, or cathode. The dissolved chem icals in the water are dissociated into two component parts, which are called ions. When an electric current is passed through an electrolytic cell the ions move to the electrodes. The metallic (including hydro gen) constituents or ions of the substances dissolved in the water pass to the cathode or negative pole, while the acid ions move to the anode. The former ions are called cathions, and the latter anions. This movement to ward the respective electrodes, of the metallic and acid portions of the compounds dissolved in the liquid, explains the manner in which an electric current is conducted through ordi- narv water. Having made this point clear, we will now proceed to consider the most impor tant point in question, viz.: the action of the ions when they reach electrodes of different composition. Electrodes may be divided into two classes, according to their ability or non-ability to be dissolved by the ions which reach the positive pole, with the formation of new chemical compounds. Some electrodes, such as car bon and platinum, are not dissolved by the anions, which find it easier to attack water and decompose it. Such electrodes are called passive or insoluble. Other electrodes, such as aluminum and iron, form new chem ical compounds by the solvent action of the anions, which find it easier, wholly or in part, to unite with the metal electrodes than to at tack and decompose water. Such electrodes are called active or soluble. Of the two ex pressions, passivity and solubility of elec trodes, the former is preferred, and hereafter we shall use it exclusively. As implied above, all negative poles, regardless of their compo sition, are considered to be passive. I assii c Electrodes. When carbon or other passive electrodes are employed in the elec trolysis of a liquid there are no new chemical compounds permanently formed, but the water is gradually decomposed into its constituent parts, hydrogen and oxygen gases. To illustrate this we will consider the electrolysis with carbon electrodes of a solu tion of common salt, sodium chloride, in pure water. The chemical symbol of salt is Na Cl, in which Xa refers to sodium and Cl to chlorine. When an electric current is applied to an electrolytic cell in which the electrolyte is a salt solution the united action is as follows: Before Application of the Current. Salt Solution. NaCl NaCl NaCl NaCl NaCl NaCl NaCl NaCl NaCl SUMMARY AND DISCUSSION OF DATA OF 18!>7. 39 After Application of the Current. Cl Cl Cl Cl Cl Cl Cl Na Na Na Na Na Na Na That is, the electric current is conducted through the liquid by the passage of the so dium and chlorine ions to the negative and positive poles, respectively. When the ions reach the electrodes their electric charges are neutralized, and they find in each case that the carbon poles are passive and do not offer any opportunity for chemical combination. Under these circumstances the second step in the process consists of the ions at each elec trode attacking water. At the positive pole the chlorine ions unite with water and form hydrochloric acid (HC1), which remains dis solved in the water, and oxygen (O), which escapes as a gas. The sodium ions at the negative pole also unite with water and form sodium hydrate (NaOH), commonly called caustic soda, which remains dissolved in the water, and hydrogen (H), which escapes as a gas. This second step in the process may be illustrated as follows: e . (gas) (gas) H . _ HC1 NaOH Cl Cl Cl Na Na Na If a porous (parchment) partition were placed in the cell between the electrodes, it would be found that the water in the vicinity of the positive electrode becomes more and more acid as the passage of the electric cur rent continues, and the water in the vicinity of the negative electrode becomes corre spondingly alkaline. Hydrochloric acid and sodium hydrate have a strong affinity for each other, and in the absence of a partition unite and form salt, the substance which was started with, and water. This combination of two of the intermediate products to form the original product constitutes the third and last step of the process. O (gas) (gas) H NaCl 11C1 NuOII 1 1 .0 Cl Cl Cl Na Na Na Na It will thus be seen that with passive elec trodes, electrolysis of salt solution effects in directly the separation of water into its com ponent elements, and that by a recombination of other secondary products the original sub stance is produced, and the process is there fore continuous. Active Electrodes. In order to make this parallel with the preceding account of pas sive electrodes we will consider the e ec- trolysis of a salt solution when the electrodes are of iron. Here the first step in the process, the conduction of the electric current by the movement of the chlorine and sodium ions to the positive and negative poles, respectively, is precisely the same as in the foregoing de scription. With regard to the second step in the proc ess, the action of the sodium ions at the nega tive pole is also the same (because all negative poles are theoretically passive), at tacking water with the formation of sodium hydrate and hydrogen gas. The action of the chlorine ions at the positive pole shows the difference between carbon and iron elec trodes. In this latter case it is easier for the chlorine to dissolve the iron electrodes than to attack water. Under the most favorab e conditions iron chloride is formed without any oxygen, and under ordinary circum stances the amount of oxygen formed ap pears to be very small, and perhaps nil. The third step, the combination of iron chloride and sodium hydrate to form sodium 39 WATER PURIFICATION AT LOUISVILLE. chloride (the initial compound) and iron hy drate, is precisely similar to the correspond ing; step in the case of passive electrodes. The only difference in this particular is that iron hydrate instead of water (which may be regarded as hydrogen hydrate) is formed. From the above description it will be seen that the activity of iron and aluminum elec trodes u:akes their use possible as a means of producing hydrates of these metals. The de gree of passivity, even of the same metal, with different salts dissolved in the water va ries widc v under the conditions of practice. A consideration of this, and several other im portant factors, in the electrolytic production of iron hydrate and aluminum hydrate, is taken up in the next two sections, in which the matter is described in detail from a prac tical point of view. Fundamental Lat^s and Principles of Electrol ysis, as Applied to the Electrolytic forma tion of Hydrates of Iron and Aluminum in the Ohio River U atcr. The leading laws and principles dealt with in this work are as follows: /. Faraday s Quantitative Law. This law may be expressed in a number of different ways, among which is the following: The amount of an ion liberated at an electrode in a given length of time is equal to the strength (amperage) of the electric current, multiplied by the electro-chemical equivalent of the ion. The electro-chemical equivalent of hydrogen for one ampere of current for one hour is equal to 0.375 gram (5.78 grains). On this basis the electro-chemical equivalent of any ion may be obtained by multiplying the above figures by the chem ical equivalent weight of the ion. In the case of elementary ions, this chemical equivalent weight is the atomic weight divided by the valency, and in the case of compound ions, it is the molecular weight divided by the va lency. From Faraday s law it follows that, other conditions being equal, the amount of hy drate of iron or aluminum formed is propor tional to the amperage of the current; and the amount of coagulating chemicals is there fore controlled by regulating the amperage of the current. 2. Ohm s Law. Ohm s law, that the num ber of amperes of current flowing through a circuit is equal to the number of volts of elec tro-motive force, divided by the number of ohms of resistance in the entire circuit, holds good for electrolytic construction. ,\ Resistance of Electrolytic Cell. In view of the fact that it is the amperage of the cur rent and not its potential which determines I the rate of formation of hydrates, it is obvious that the resistance of the cell should be kept as nearly as possible at a certain minimum for economical reasons. The minimum poten tial is determined bv the polarization of the cell, as stated more fully in a following para graph. The resistance of the cell is due to several factors, among which are: the area of electrodes: the distance between elec trodes; the amount of dissolved salts in the river water (electrolyte): and the formation of non-conducting coatings on the electrodes. From ( )hm s law it follows that the re sistance of an electrolytic cell increases di rectly with the water space between the elec trodes, and inversely with the cross-section of the electrolyte (or area of the electrodes). ./. Resistance to the Passage of an Electric Current of Ohio River U atcr. This subject has been referred to in Chapter XIII, where it was seen that during the period of flood in February and March, 1897, the resistance of the river water increased nearly threefold, due to the decrease in amount of dissolved chemical compounds. That, practically speak ing, the suspended matters in the water, m- cluding those partially dissolved constituents, exerted no influence on the conductivity was also presented at that time. Estimating the conductivity or resistance (which is the reciprocal of the conductivity) of the river water from the observations on different combinations of various solutions of the salts normally present in it, the resistance in ohms per centimeter cube should be theo retically 6100, 930, and 2080 ohms for maxi mum, minimum, and average, respectively, corresponding to 72, 2<K>, and 122 parts per million of dissolved chemical compounds, not including carbonic acid- gas. As will be shown in connection with the study of pas sivity of iron electrodes, it is not possible to draw specific mathematical conclusions in re- SUMMARY AND DISCUSSION OF DATA OF 1891 39 gard to the behavior of combinations of ions, based on the results of observations in indi vidual ions. It will be also shown in this connection that dissolved carbonic acid gas is only very slightly ionized, and from a prac tical point of view need not be considered as a conductor at all. It is therefore necessary to rely upon ob servations on the river water itself, though as will be seen, the theoretical and observed re sistances follow closely the same curve. On Feb. 22, 1897, the observed resistance of the electrolyte in the Brownell cell was 7600 ohms per centimeter cube, and on Feb. 27 this figure became 16.750. The amounts of dissolved chemical compounds on these days were 146 and 67 parts per mil lion, respectively. On Feb. 17. with 120 parts per million of dissolved compounds, the resistance w r as observed to be 9200 ohms. In connection with the devices of the Water Company the average resistance was observed to be about 7000 ohms per centi meter cube, when the river water contained about 130 parts of dissolved solids. From these results combined with numer ous special observations, including those in Chapter XIII, it is estimated that the maxi mum, minimum, and average resistance of the Ohio River water as an electrolyte would be 17000. 2000, and 7000 ohms per centi meter cube, corresponding to 67, 324, and 130 parts per million of dissolved chemical com pounds, exclusive of carbonic acid, which, as was shown, is only very slightly ionized. In times of great freshets the dilution of the compounds in the river water might be so great that it would be advisable to add com- mont salt to the water to increase its con ductivity to the normal. Investigations during the heavy freshet of February and March, 1897, showed that this could be ac complished by the addition of common salt in amounts not exceeding 5 grains per gal lon. Such additions would not cause the total amount of salt to be greater than was normally present in the river water during the low stages of the river in the fall of 1895. 5. Polarization of Electrodes. When a cur rent of electricity flows through an electro lytic cell, and causes changes in the electrolyte, or on the electrode, the electromotive force of the current is thereby reduced. This action is known as polarization. In explana tion of this point, which determines the mini mum potential of current that can be safely employed, it is to be stated that all ions pos sess a certain force or intensity of fixation wherewith they attempt to retain their elec tric charges when they reach the electrodes. Accordingly, a certain potential, slightly above that corresponding to the intensity of fixation, is necessary in order to overcome this force, and free the ions at the electrodes of their electric charges. The existence of this intensity of fixation, with an opportunity to measure it, is shown by the reverse cur rent which takes place for a short time when the primary current is shut off. With active electrodes, polarization becomes less marked. The potential of polarization varies in the line of work in question. So far as we know, there would be no case where the polarization would require over 2.35 volts to overcome it. Ordinarily it would be much less than this. Records show that in the investigation of electrolysis of iron pipes lying in the ground near electric lines of street-cars (a subject" similar in a measure to the present one), de composition of the iron has taken place at a potential of only o.ooi volt. In all cases a difference in potential of 2.5 volts, or less be tween adjoining electrodes, would suffice to overcome the intensity of fixation of all ions, while much less than this would probably be adequate for a majority of the ions. Practi cal investigations along this line are recorded in section Xo. 4 of this chapter, where it will be seen that potential differences as low as r.o volt can be safely employed with iron electrodes. 6. Passivity of Electrodes. As already stated, all negative electrodes, so far as is known, are passive to the ions, and certain positive electrodes such as iron and alumi num, are active. From Ampere s law it fol lows that the same quantity of electric cur rent always causes in electrolysis the same equivalent amount of acid ions (anions) to go to the positive electrodes, and have their elec tric charges neutralized. They then pass into the atomic state. With passive elec trodes, they attack water, and equal currents produce equal amounts of oxygen gas. Pro- 39 a WATER PURIFICATION AT LOUISVILLE. vided that active electrodes were completely active (not at all passive or insoluble under the action of these liberated acid atoms), the amount of metal decomposed from the posi tive electrode would also be proportional to the amperage of the current, and to the amount of liberated acid ions, in accordance with Ampere s law. In the case of iron and aluminum elec trodes, however, experience .shows that the metal of the positive electrodes is not dis solved in quantities proportional and equiva lent to the total quantity of acid ions liberated at the positive electrode. From a practical point of view this fact is a matter of vital im portance, because it relates to the amount of hydrate formed, and consequently to the com mercial merits of the process. Stating this in another way, we may say that only a por tion of the current forms the hydrate of the metal used as the positive electrode; and therefore such metals as iron and aluminum are only partially active, as a portion of the current causes the formation of oxygen, just as in the case of completely passive electrodes, such as carbon or platinum. It follows from the above statements of facts that, under practical conditions, iron and aluminum electrodes are only partially active, and when employed in this process utilize efficiently only a portion of the cur rent. Hence, in the following sections, in which this process is described in detail, use must be made of degree of activity or degree of passivity of electrode. In view of the fact that the latter expression is in use by chem ists, we shall adopt it in future cases where reference is made to this phase of the process. The degree of passivity of iron and alu minum electrodes is due to the two following factors: 1. The initial passivity of the metal to the various acid ions naturally present in the river water. Thus it is well known that hy drochloric acid has a higher solvent action on these metals than carbonic acid. 2. The acquired passivity of the metal to the various acid ions, due to the formation of thin films of metallic oxide, caused by the oxygen formed by the weaker ions, upon the metal. In practice the varying composition of the river water caused a wide range in the rela tive amounts of the different acid ions, and the consequent total dissolving action upon the electrodes. Experience shows that an other important factor, especially in the case of aluminum, is the fact that the film of metal lic oxide on the positive pole causes a ma terial increase in the resistance which the electric current meets in its passage through the cell. In the case of iron electrodes, how ever, this is relatively slight, owing to the fact that the film cracks and falls off in scales at frequent but irregular intervals. 7. Secondary Reactions. Normally there are formed with active electrodes a hydrate of the metal, hydrogen, and, varying with the de gree of passivity, a certain amount of oxygen. These reactions have been explained in the foregoing account of electrolysis, and may be called primary reactions, or perhaps the pri mary group of reactions. Other reactions (in dependent of coagulation), called secondary reactions, will now be referred to. When iron electrodes are employed the iron is dissolved from the plates in the form of ferrous or unoxidized salts, which would be converted into the partially soluble ferrous hydrate by the alkaline hydrates coming from the negative electrode. Accordingly, oxy gen is necessary, in order to convert the fer rous into the ferric forms. As practically all of the electrolytically formed oxygen attacks the positive iron electrodes, it is necessary that the atmospheric oxygen, naturally dissolved in the water, serves to effect this oxidation. This oxidation is necessary in order to con vert the iron into a completely insoluble form, so that it can coagulate the suspended matter in the water without any dissolved iron passing through the filter into the puri fied water. The secondary reactions in the case of alu minum electrodes are less clearly understood, but are referred to beyond. Concerning hydrogen at the negative elec trode, small portions of it in the nascent con dition combine with atmospheric oxygen dissolved in the water, and reduce iron com pounds and nitrates; while the bulk of it, after saturating the pores of the metal, escapes as a gas in a molecular condition. SUMMARY AND DJSCUSS1ON OF DATA Of- 18! >f. 393 SECTION No. 4. DETAILED ACCOUNT OF THE ELECTROLYTIC FORMATION OF IRON HYDRATE IN THE OHIO RIVER WATER. A full record and discussion of this process is presented under several leading topics as follows: Degree of Passivity of Iron Electrodes due to the Different Acid Ions of flic Ohio Rh cr Water. The dissolved acids which are present in appreciable amounts in the Ohio River water are hydrochloric, sulphuric, nitric, silicic, and carbonic. As is well known to chemists, these acids differ in their ability to dissolve iron. This capacity of being dissolved by acids under the conditions of electrolysis measures the pas- sivitv of iron in electrolvtes containing the acid in question. Several experiments were made in which solutions containing the re spective acids alone and in various combina tions (in the form of salts of the alkalies) were placed in sets of cells containing electrodes of bright wrought iron and of rusty iron from the same sheet. The cells were arranged in series, and an electric current of from 2 to 47 amperes per square foot was passed through them for periods of from 2 to 30 minutes. Aliquot portions of the contents of each cell, after rinsing- the electrodes, were col lected for determinations of the amounts of iron. The average results of these determina tions, together with the average specific re sistances of the several electrolytes, are pre sented in the following table. Percentages which the metal decomposed was of the theoretical rate (1.05 grams per ampere hour) are also given. The potential difference between the plates varied in these experiments from 12 to 220 volts. SUMMARY OF RESULTS SHOWING AVERAGE RATES OF ELECTROLYTIC DECOMPOSITION OF NEW AND OLD IRON IN ELECTROLYTES OF VARIOUS ACID IONS. Avc ras;e Kate of Decomposi A Venice Acid Ions. Parts per Million.* Grams pc Ho r Ampere- ur.t Per Cen Theoreti t. of the Ciil Rate. of Klectrnlytc. Ohms per New Iron. Old Iron. New Iron. Old Iron. Cube. o 85 88 8l 6880 69 60 86 82 76 (>-> D 75 6960 L 1 .00 0.63 95 60 3 3<> IO , e f I . 12 0.90 107 86 4170 IO , 60 f 0.97 0.64 93 61 I 690 10 | 0.90 0-73 86 70 iol 5 5 1.04 0.44 99 42 I 240 60 1 5 . \\ 1 .03 0.48 98 46 1310 60 | 1 Carbonic acid (dissolved gas) 50 J * These quantities are equivalent to the estimated average amounts of the several acids contained in the Ohio River water. f To change grams per ampere-hour into grains per ampere-hour multiply by 15.4. 394 WATER PURIFICATION AT LOUISVILLE. The above data indicate quite clearly one point which is, that the result of combinations of two or more ions under these conditions is not the mathematical result of their sum except by chance. If the observations of the resistances of the combined electrolytes (chlorine with the other ions) be compared it will be seen that there is a marked differ ence from the theoretical resistance of the combinations. The only apparent explanation to account for this is that when two or more ions are combined in these amounts in the same elec trolyte their degrees of ionization are differ ent than, in the case when any one of them is presented alone. The relation between the calculated and observed results becomes then a somewhat variable one, dependent upon the relative de grees of ionization of the several acids. The conclusions which may be drawn from these experiments are that the relative amounts of the different ions present in the river water would affect the rate of decom position, but the influence of the several fac tors under practical conditions is so variable that the present data are insufficient to deter mine the exact laws for the estimation of these amounts. Cause of Passivity Initial and Acquired. In the last section a table of results was presented of some experiments with new bright wrought-iron and old (rusty) iron elec trodes. Both sets were cut from the same plate, and one set cleaned to bright metal while the other remained rusty. It will be seen that the results of observation on the behavior of the new iron gave rates of de composition varying from 79 per cent, of the theoretical with silicic acid to 1 1 1 per cent, with dissolved chlorides, averaging 90.5 per cent. That new bright iron behaves differ ently toward the several ions is clearly shown, and while it is difficult to account for the re sults over 100 per cent., it seems clear that in the case of all but hydrochloric acid a certain percentage of the electric current is not di rectly utilized in dissolving the metal. The explanation of this lies in the relative affinities of these ions for the metal and for water. When the electric current is transferred by means of the ions and the ions are discharged or neutralized at the positive pole, they attack the water of the electrolyte and the metal in the proportion of their affinities for the liquid and the metal. The acid ions which attack the water can be said to represent the passivity of the iron, because they would attack the iron were it not passive. The results on new iron may be taken to indicate the relative passivities which bright iron has to the sev eral acid ions. The data are not sufficient, however, to warrant the use of these figures except in a comparative manner. Those ions to which the metal is passive attack and decompose water, setting free oxy gen gas in a nascent condition. Between this gas and the metal there is at all times great affinity, and therefore a considerable amount of the oxygen attacks the metal and forms the oxide. As this continues the plate becomes covered with a coating of oxide scale which grows thicker and thicker until it begins to crack off. Practically speaking, the rate at which the scale is removed from the plate by cracking and peeling becomes eventually as great as the rate of its forma tion, and equilibrium is established with reference to the respective attacking of the metal and of water. The presence of this oxide scale changes the relation of the acid ions to the metal, as they must either attack and dissolve the oxide or pass through the scale to attack the pure bright metal beneath. As either of these processes requires more energy than the sim ple solution of the metal, an increased per centage of the ions does not attack the metal, but decomposes water, and by this action the iron has an acquired passtivity. During the process of formation of the scale and before equilibrium has been estab lished between the formation of the oxide and its scaling off, the acquired degree of passiv ity increases rapidly as the scale forms. As the formation of the oxide is dependent on the passivity of the metal and this in turn increases with the oxide present on the face of the plate, the process is a reciprocal one, one action increasing the other. For this reason the length of time elapsing between the beginning and end of the action (estab lishment of equilibrium) is comparatively SUMMARY AND DISCUSSION OF DATA OF 1897. 395 short and is dependent upon the density of the current used. Furthermore, it is proba ble that the passivity of the metal may be come greater, before the scale begins to come off, than it is after equilibrium has been estab lished. The results with old iron, presented in the last section, may be taken as fairly representa tive of the total passivity of iron as it would be used in practice. As will be presented beyond, there are no indications to warrant the belief that the po tential difference between the plates has any effect upon the passivity of the metal. Within the limits employed in these inves tigations the difference in initial passivity due to the composition of the metal was not apparent, cast iron, wrought iron of different grades, and mild steel all giving apparently parallel results, or very nearly so. Form in which Iron leaves the Plates. The iron leaves the positive electrodes only, and in order to make the full set of plates serviceable it is necessary to reverse the electric current from time to time. At the positive electrodes the iron leaves the plates in two ways, namely: 1. Those acid ions, which are neutralized electrically at the pole by dissolving some of the metal, form iron salts of the various acids, such as iron chloride, iron sulphate, and iron carbonate. These compounds, furthermore, are in the ferrous (unoxidized) condition, as explained in the next section. 2. Those acid ions which, by virtue of the degree of passivity of the iron anode, find it easier to react with water upon neutralization than to dissolve the equivalent amounts of iron, form oxygen. This oxygen unites with the iron to form iron oxide, which appears as films. In the first case the solution of iron is regu lar and proportional to the amperage of the current when the degree of passivity is con stant. With those ions to which the iron electrode is passive, the formation of films of iron oxide is regular, but the films crack and peel off from the electrode in an irregular manner. Influence on the Process of O.v\gcn. With the Ohio River water the oxygen in an electrolytic cell comes from two sources, the atmospheric oxygen naturally dissolved in the water, and the oxygen which is formed eleectrolytically at the anode. We shall con sider them separately. Atmospheric O.rygcn. The atmospheric oxygen performs a very important part in this process, by virtue of the fact that it unites witli the ferrous compounds as they are dis solved from the electrode, and changes them to ferric or oxidized salts of iron. Appar ently this action takes place partly before the iron salts are acted upon by the alkaline hy drates coming from the negative pole, and partly after this reaction. The result is that after the completion of the secondary group of reactions, the dis solved iron is converted into the form of in soluble ferric hydrate, which is an excellent coagulant. The importance of this oxida tion from a practical point of view is great, because without it ferrous hydrate alone would be formed; and, owing to its partial solubility, there would be difficulties arising from its passage through the filter. A small amount of atmospheric oxygen also unites with hydrogen, which is given off as a gas at the negative electrodes. As this combination cannot take place except when the hydrogen is in the nascent state, the ac tion is confined to the oxygen in the immedi ate vicinity of the cathode. Electrolytic O.rvgcn. So far as could be learned, practically all of the oxygen which is formed at the anode by the decomposition of water attacks the metal electrodes and forms films of iron oxide. As stated above, these films crack off and leave the cell in an ir regular manner. Of course the scales or films of iron oxide are of no assistance in the purification of water. Comparing the influence of the oxygen from the two sources we see that the atmos pheric oxygen performs a very important part, and without it the process could not be put in practice with satisfactory results. The influence of the electrolytic oxygen and the factors which produce it, on the other hand. 396 WAT lilt PURIF1CA1ION AT LOUISVILLE. is a very serious drawback to the process, be cause it means a large waste of electric power and of metallic iron. The amount of power and of metal wasted is shown bevond. Viewed in the present connection, as a fac tor in a series of secondary reactions which occur in this process after the iron is dis solved from the anode, carbonic acid did not give indications of retarding the oxidation of the resulting ferrous salts, as was the cas e when commercial protosulphate of iron (cop peras) was applied to the water. So far as could be learned the practical effect of car bonic acid upon any secondary reaction was ml. With regard to the relation of free carbonic acid to the passivity of iron electrodes and the degree of its ionization, the evidence has al ready been presented. Briefly, it showed that the degree of ionization of free carbonic acid in the water is very low, and in this con nection its practical significance is very slight. Under all conditions hydrogen gas is pro duced at the negative electrode in amounts proportional to the formation of alkaline hy drates. Owing to the fact that iron pos sesses the capacity to occlude large quanti ties of hydrogen gas within its pores, a portion of the hydrogen is disposed of in this manner, and there are reasons for believing that the negative electrode after a time is practically composed of hydrogen, from an electrical point of view. A small portion of the hydrogen when in a nascent condition unites with atmospheric oxygen to form water. The bulk of it, however, after the saturation of the pores of the iron cathode and of the water, escapes as a gas. The only practical influence of hydrogen, other than the slight consumption of the at mospheric oxygen, is in connection with sub sequent sedimentation and filtration, as men tioned beyond, together with the relation of other factors of the process upon the follow ing steps in the purification of water. It has been stated that when the iron is first dissolved from the positive electrode the compounds are in a form of soluble ferrous salts. Experience shows that some of these iron salts, before they are converted into in soluble ferric hydrate, become conductors of the electric current just like lime and other salts dissolved in the water. The result of this is that a portion of the iron is conducted to the negative electrode; and, in a. manner similar to that in electroplating, is deposited there in what appears to be a metallic form. From a practical point of view the influ ence of this state of affairs is to cause a waste of electric current, as no good is accom plished by transferring the metal from one pole to the other. So far as could be learned this metal is not wasted, but is available for electrolytic decomposition when the direction of the electric current is reversed. Form in which the Iron leaves the Cell. A portion of the iron which leaves the cell is in a form available for coagulation, while a portion is not available. We shall consider the two separately in this connection. Available Iron. Analyses of the treated water as it leaves the cells show that the greater proportion of the available iron com pounds is in the form of ferric hydrate. Some of the iron which may be actually available for coagulation a little farther along in the flow of the water is at times in the form of ferrous compounds as the water leaves the cell. At no time in regular practice was the amount of iron in the form of ferrous com pounds found to exceed 3.8 parts per million. Lender these conditions it is probable that all the ferrous iron was in the form of hydrate. When the rate of electrolytic treatment is so great that the atmospheric oxygen is all util ized, then additional treatment causes the iron to leave the cell in the form of ferrous hydrate. Under these conditions there would be about seven parts per million of iron which would be soluble. Non-Available Iron. As a consequence of the passivity of iron and the formation of SUMMARY AND DISCUSSION OF DATA OF 1X J7. 397 films or scales of oxide of the metal, a portion of the iron leaves the cell at irregular inter vals in the form of flakes or scales. The evidence showing the relative amounts of iron in these t\vo forms is presented just below. Natural Limitations of the Electrolytic Treat ment n ith Iron Electrodes. From the foregoing account of this process it is clear that this treatment cannot be safely applied beyond the point where atmospheric oxygen in the water is entirely utilized to convert the iron compounds into the form of insoluble ferric hydrate. The limit of safe ap plication of this process depends therefore on two factors, namely, the rate of formation of iron compounds and the amount of oxygen in the water; both of these vary, and their mutual relation is referred to beyond, after the presentation of further evidence. Rate of Decomposition of Iron at the Positive Electrode. Owing to faulty insulation of the large electrodes the data prior to May 29, when hard rubber fittings were put in service, are disregarded. In the following tables the weights of iron decomposed electrolytically from the positive electrodes of each of the two sets are recorded for the several periods between weighings, together with the corre sponding electric current (expressed in am pere-hours) for each period. During the first week of June the sets of electrodes were weighed nearly every day, after washing with a stream of water from a hose. It was found, however, that these weighings were of no account, owing to the metal deposited on the negative electrodes and to remaining accumulations of mud and silt. In a number of cases such weighings showed an increase in the weight of the elec trodes. Beginning June 9, the electrodes were dismantled from time to time and the positive and negative plates weighed sepa rately, after cleaning each one carefully with a broom and a stream of water. After June 16 the direction of the electric current was re- reversed from time to time, and the positive and negative plates were also weighed sepa rately from time to time. The periods of service between weighings were considerably longer than formerly, and an effort was made to choose such times for weighing when the amount of metal deposited on the negative plates formed a comparatively small per cent, of the total loss of metal. Afrer June 29 the plates were not weighed until Sept. 25. when the entire electric plant was finally dismantled. From the close of the regular tests until Sept. 25 the electrodes were used for several special tests and exhibitions, the last of which was on Sept. 23. When not in service the water was drawn out of the cells in which the electrodes were placed. In studying the following table it is to be recalled that the rate of decomposition of iron at the positive electrode would be 16.2 grains (1.05 grams) per ampere-hour, pro vided that the iron plates were completely active, or. in other words, soluble in the acid ions discharged at that point. For conve nience in comparison this is called the theo retical rate of decomposition. SUMMARY OF RESULTS. I er ocl. Electric p ofitive Average Rate of Decomposition. Per Tent. of Theo. Grains Grams retical 1 1 Electrodes No. 1 , May )une *l\ 46 760 50.0 7-5 0.49 46 ,u 34 00 86.5 17.8 i . 1 6 no 2 1 23 660 78.75 23.3 1.52 144 2g ( - 3 520 72.75 16.2 1.05 loo Sept *9 | 109500, 288.25* 18.4 1-19 3 Totals 245540 576.25 16.42 1.07 102 June E n t" 5 75 ectrodes 30.5 No. 3. 13.6 0.88 84 II / 18 |" 46 750 104.0 15.6 i .02 96 2Q f 23 600 46.0 13 6 0.88 84 Sept 2J? j- 72 420 200.25 19.4 1.36 12O Totals 158 500 380.75* 16.82 1.09 104 Total for both Electrodes. 404040 957-00 16.57 1.075 102 * Plates scraped with metal scrapers. WATER PURIFICATION AT LOUISVILLE. The slight excess of the total average rate of decomposition over the theoretical was probably due largely to the fact that at the close of the test the plates were freer from rust than at the beginning of these tests, as at the close the plates were cleaned with me .al scrapers. Similarly the irregularities in some of the periods were due largely, if not wholly. to different degrees of thoroughness with which th:? plates were cleaned with a broom and a stream of water. The rate at which the iron, removed from the positive electrodes in a soluble form, may become a conductor of the electric current and may be deposited in the metallic state, is indicated by the following summary: SUMMARY OF RESULTS. Owing to the fact that it was found advis able to reverse the direction of the electric current at the end of the above period, no further data upon this point were obtained. It is possible that slight amounts of silt from the river water were mixed with the deposited metal, and caused the above results to be a trifle high. This was not a serious factor, however, because the river water at this time was very clear, comparatively speaking. During the above periods the current den sity was varied from time to time. On elec trodes No. i, from May 29 to June 7, the current ranged from 100 to 400 amperes and averaged 253 amperes, corresponding to 0.26, 1.03, and 0.65 amperes per square foot of the active surface of the cathodes (corresponding approximately to the cross-section of the electrolyte). With electrodes No. 3 for the above period the maximum, minimum, and average currents were 234, 171, and 196 am peres, respectively, corresponding to 0.60, 0.44, and 0.50 amperes per square foot of cathode surface. From May 29 to June 2, inclusive, it required about O.8 minute for the water to pass through the portion of cell No. i occupied by the electrodes. From June 2 to 9 the rate of (low in both cells Nos. i and 3 was such that the water remained in the por tion of these cells occupied by the electrodes for about i minute. With the same rate of flow of water through a cell, it is probable that the rate of deposition of metal would increase di rectly with the current density and with the specific resistance of the electrolyte. For the composition of the river water at this time sec Chapter J. oi Average Rate of -3 Gain in Deposition. Period. Electric Current Wei K lH of Negative Per Ampere- PerCent. W Ampere- trodes. Ilour. of Theo- f 1 tours. Pounds. Decom Jj Grain*. Grams. position. I May 29 | June 9 f 46 760 14.0 2.1 0.14 S.6 3 June 2 I June 7 )" 15 730 6.0 2.7 o . I S II . I \Ye have considered the rate of decompo sition of iron at the positive electrode, some of which is non-available for coagulation, as it is in the form of scales of iron oxide, and the deposition of some of the soluble iron, capable of forming ferric hydrate, at the nega tive electrode. It remains to show the rate of formation of iron hydrate which is the only form capable of coagulating the river water. From time to time during the regular opera tion the amount of iron leaving the cell was determined by tests with ferricyanide and ferrocyanide. By these well-known tests it was learned, in addition to the total amount. how much iron was in the two states of oxi dation (ferrous and ferric iron). But these tests do not show the difference between the hydrate and the scales of oxide which are given off at an irregular rate. To note the rate of hydrate formation it is necessary to compare a series of observations, and study the low and uniform rates, which are affected but little, comparatively speaking, by scales of oxide. In order to make this evidence as complete as practicable, numerous special tests with electrodes Nos. i and 3 were begun on Jn y 23 and continued until July 28. The results of these tests, with the percentage which the total iron was of the amount indicated by the theoretically normal rate of formation, are presented in the set of tables on pages 400 SUMMARY AND DISCUSSION OF DATA OF 1897. 399 and 401. In these tables sufficient notes will he found to make plain the conditions of operation. It will be seen that the results of these tests, which confirm earlier and more fragmentary data, indicate a normal amount of iron in the water leaving the cells, or something less than 50 per cent, of the quantity corresponding to the theoretical rate of decomposition. In some instances the amount of iron exceeded 100 per cent, of the theoretical rate, showing an abnormal removal of scale of oxide. So far as could be learned the amount of iron corresponding to more than about 50 per cent, of the theoretical rate was ordinarily present for the most part in the form of scales. At times it will be seen that the rate was con siderably less than 40 per cent. Among the principal factors which aided in producing the low results are the reversal of the electric cur rent, exposure of the plates to air when not in service, and coatings of mud mixed with the scale and films. These factors are re ferred to beyond the tables. At the beginning of these experiments electrodes Xo. 3 were dry, not having been used since 8.30 P.M., July 16. Electrodes No. i had been in use on July 22 and had only partially dried off during the night. Through out this series a uniform rate of flow of water of 23.5 cubic feet per minute through each cell was maintained while in operation. At the start a current of 210 amperes was used, which gave an effective current of 0.020 am pere-hour per gallon in cell Xo. 3, and 0.040 ampere-hour per gallon in cell Xo. i, which contained the modified electrodes Xo. i. The potential difference between the plates at the start (at 210 amperes) was 6 volts on Xo. i and 4 volts on Xo. 3. When the current was increased on July 23 to 400 amperes the potential differences were increased to 1 1 and 6 respectively. The current density at 210 amperes was 0.51 ampere per square foot of anode in electrodes Xo. 3 and i.oo ampere per square foot of anode in modified elec trodes Xo. i. The corresponding figures for 400 amperes were 0.97 and 1.91, respectively. Influence on the I onnatioii of Hydrate of Po tential Difference bctiwcn the Plufes. Potential difference can be denned as the electric pressure necessary to cause the pas sage of the desired strength of electric cur rent through the electrolyte from plate to plate of the electrodes. This pressure is necessary to cause the ions to carry their respective charges to the poles, and is therefore dependent on the density of the current and length of the electrolyte. There is, furthermore, a certain potential difference necessary to overcome the resist ance which the ions have against neutrali zation. This is the polarization resistance. In the case of passive electrodes the available data indicated that as a maximum this resist ance may reach as high as 2.35 volts; and in practice it was assumed early in these tests that 3 volts was the lowest safe minimum dif ference in potential to secure a utilization of all ions uniformly. So far as could be learned the practical in fluence upon the formation of hydrate, of the potential difference above this limit, is ;///. During the use of the large electrodes poten tial differences between the plates of from 3 to 12 volts were employed. In laboratory ex periments the potential difference was carried at times as high as 220 volts, although the usual range was from 20 to 50 volts. The most serious factor against the econ omy of this process of formation of coagu lants is the passivity of the iron to the various dissolving ions in the river water. As has been shown in previous portions of this dis cussion, the passivity of iron is represented by the percentage which the actual amount de composed and converted into hydrate is of the total theoretical rate of decomposition, and ranged during the observations on this point, including laboratory tests, from about 30 to 50. As the ions which do not dissolve the metal attack water, it appeared, as the report on these matters neared completion, that by low ering the potential difference to an amount below that which is commonly accepted as necessary to maintain a steady decomposition of water, with passive electrodes, this partial passivity of active electrodes might be over come in a measure. For this purpose two electrodes of steel were prepared and operated from 1.20 P.M, Sept. 28, to 9.00 A.M., Sept. 30. The area 400 WATER PURIFICATION AT LOUISVILLE. TABLE SHOWING THE CONDITION OF OXIDATION AND AMOUNTS OF ELECTROLY TICALLY DECOMPOSED IRON LEAVING THE CELLS, EXPRESSED IN PARTS PER MILLION AND IN PERCENTAGES OF THE THEORETICAL RATE. Cell No. i. Cell No. 3. Dale. Hour. Parts per Mi lion. J jii Parts per M lion. S(S g . Remarks. 1897. 5^3- U v I -. Ferrous Ferric Total 2 oS. Ferrous Ferric Total t - co: Iron. Iron. Iron. a. Iron. Iron. Iron. SH Conditions at start described in text. JU 1 V "\ nee A M . 5 S.o 8. 5 77 : . . uiy 43 -J-JJ ....... " 23 O.O5 " I . O 9 - o 10. o " 23 0.2O " 03 (). ^O o.5 o 3 n. o 11.5 11 O 111 104 o.o I 20 IO.O IO.O 7 " 23 , 1. 06 " O. I 1 J ^ A J J IO.O IO.I 91 0.4 10.0 10.4 73 23 1 .47 o. 3 S.o 8.3 75 "23 .57 " o. 3 R . o S.i 75 " 23 2.25 P.M. o.o 5.5 5.5 50 o.o 9.0 9 . o 63 " 23 T.OO " 0.3 10. o 10.3 93 0.3 10.0 10.3 72 " 23 .20 " 0.2 | 9.0 9.2 83 O.O 9-5 9-5 67 " 23 I.5<) " O. I 9.0 9. 1 82 : O.I 9-5 <)(> 67 " 23 2.05 " 0.4 9.0 9.4 85 0-3 9.0 9.3 6; " 23 2.20 " 0.4 7-5 7-9 71 0.4 7-5 7-9 56 " 23 2.35 " 0-3 7.0 7-3 66 o-3 7.0 7-3 51 " 23 2.50 0.3 7-0 7-3 66 0.2 7-5 7-7 54 " 23 3.05 " 0.2 7.0 7-2 65 0.3 7-2 7-5 53 " 23 3.20 " 0. I 5-5 5-6 50 0. I 4.8 4-9 83 " 23 3.30 " O. I 5-3 5-4 49 0.2 4.0 4.2 75 " 23 3.40 " 0.3 7.0 7-3 66 O. I 4-5 4-6 82 " 23 3.50 " 0. I IO.O IO.I 90 0.2 5.0 5.2 93 Direction of How of electric current " 23 4.10 " 0.6 4.0 4.6 4i O.O I.O I.O 18 reversed in both cells at 4.00 I .M. " 23 4.20 " 0.5 4.0 4-5 40 Trace Trace O.I 2 " 23 4.30 " 3 4-5 4.8 43 0. I Trace 0. I 2 " 23 4.40 " 0.3 4-3 4.6 41 O.I 0.2 2. I 5 .. 23 4-5" " o-3 5-2 5-5 49 0.0 | I.O I.O IS . 23 " 23 5.OO 5.io " 0.2 5-2 IO.O 5-4 10.2 49 48 O.I 3-5 3-f> 34 and 0.038 ampere-hour per gallon " 23 5.20 " O.I 9.8 9-9 47 0.3 4.0 4-3 41 in cells Nos. i and 3 respectively. " 23 5.30 " O.2 it .0 II. 2 53 O. I 3-8 3-9 37 Current shut off and cells drained at 5-45 I -M. " 24 .15 A.M. O.O 4-5 4-5 21 0.3 1-5 1.8 17 Began operations at 11.10 A.M. with " 2 4 .25 " ! o.o 6.5 6-5 31 o.i 3.5 3-6 34 conditions unchanged. " 24 35 " 0.0 7-0 7.0 33 o.i 4.0 4-1 39 " 24 45 0.0 7-5 7-5 35 o.i 4.1 4-2 4 " 24 55 0.3 9-5 9.8 46 0.3 5-2 5-5 52 " 24 .12 P.M. 0.0 o.o o.o 47 O. I 6.0 6.1 57 " 24 .25 " 0.2 2.O ! 2. 2 58 O. 2 6.0 6.2 58 " 24 .40 " 0-3 2.0 2.3 58 O.2 6.0 6.2 | 5 8 !. 24 55 0.0 I.O 1 .0 52 O.O 6.0 (>.o ; 57 24 . IO 0.0 2.O 2.0 57 O.O 5.0 5.0 47 " 24 .25 " o. 2-5 2.6 59 0.2 4.8 5-0 47 .! 24 45 o. 2.2 2-4 58 O.I 4.8 4-9 4(> 24 55 o. 2.2 2.2 58 0. I 4.8 4.9 46 " 24 .10 " 0. 2.2 2-4 58 O. I 5 -O 5.1 48 " 24 25 " o. 2.2 2.4 58 O.I 5-5 5-6 , 53 " 24 .40 " o. 3-o 3-2 52 0.2 5-0 5-2 ; 40 " 24 55 o. 2.O 2.2 58 O.2 5-o 5-2 49 " 24 3.10 " o. 1-5 1.6 55 O. I 5-2 5-3 50 " 24 3-25 o. I.O I.I 52 O. I 5-0 5-i 48 " 24 3-40 " o. 1 .0 I . I 52 O.2 4.8 5-0 47 " 24 3-55 o. I.O I . I 52 O. I 5-0 5-1 48 " 24 4.10 " o. o-5 0.6 50 O.I 5-0 5-i 48 " 24 4-25 o. 0.5 0.6 50 O. I 5-0 5-i 48 " 24 4.40 " o. 0.5 0.6 50 O. I 5-0 5- 1 48 " 24 4-55 o. 0.5 0.6 50 O.I 4.8 4-9 46 " 24 5.10 " 0. o.o O.I 48 O. I 5-o 5-i 48 " 24 5-25 o.o 0.0 0.0 47 o.o 5.0 5-0 47 " 24 5.40 " O. I 0.0 O.I 48 0.0 6.0 6.0 56 Current shut off and cells drained at 5.45 I .M. SUMMARY AND DISCUSSION OF DATA OF 1807. CONDITIONS OF OXIDATION AND AMOUNTS OF ELECTROLY TICALLY DECOMPOSED IRON LEAVING THE CELLS. Concluded. Cell Cell No. 3 Due. 1897. Hour. P;ir Is per M lion. - ? Par s per M Ilion. ^H " Remarks. T u y ~ TV.nl U ,. t : Iron Iron. - Iron. 0. Began operations at 7.00 A.M. with July 26 g.IO A.M. 1 O.O 9.0 9.0 42 O.O 4.0 4-0 38 conditions unchanged. " 26 g.25 " i O.O 8.5 8-5 40 0.0 3-5 3-5 33 " 26 9.40 " 0.6 9.0 9.6 45 0.7 4.0 4-7 I 44 26 9-55 0.0 9-5 9-5 45 O.O 4-0 4-0 38 26 IO.IO " 0.2 9-5 9-7 46 0.2 4-5 4-7 44 26 10.25 " O.I IO.O 10. 48 O. I 4-5 4.6 43 26 o. 40 O.I 10. I 10. 48 O.I 5-0 5-1 48 26 0-55 " O. I IO.O IO. 48 0.0 6.0 6.0 57 26 I. 10 " 0. I IO.O 10. 48 O.O 5-0 5-0 47 26 1.25 " O.I 9-5 9- * 45 0.0 5-0 5-0 47 26 2.55 P.M. 0. I IO.O IO. 48 O. I 5.0 5-1 48 " 26 1. 10 " O. I IO.O 10. 48 O.O 5-0 5-o 47 26 1.25 " 0. I .9-5 9- 45 O.O 5.0 5-0 47 26 1.40 " 0. I IO.O IO. 48 O. I 5-1 5-2 48 26 i-55 O. I IO.O 10. 48 O. I 5-1 5.2 48 26 2.25 " O.O IO.O IO.O 47 O. 1 5-i 5-2 48 26 2-55 " 0.0 IO.O IO.O 47 O. I 5.1 5-2 48 26 3-25 O.O IO.O IO.O 47 O.O 5.0 5-0 47 " 2f) 3-55 0.0 IO.O IO.O 47 0.0 5-o 5-0 47 " 26 4.25 0.0 IO.O IO.O 47 O.O 5-0 5-0 47 " 26 4-55 O.O 9-5 9-5 45 0.0 4-5 4-5 42 " 26 5.25 0.0 IO.O IO.O 47 0.0 5.0 5.o 47 Ran continuously over night. " 27 9.05 A.M. O.O IO.O IO.O 47 0.0 7-5 7-5 7i " 27 9.15 " O.O IO.O IO.O 47 O.O 7.0 7. o 66 " 27 1. 15 0.0 IO.O IO.O 47 2.0 5-o 7- 66 " 27 1.30 " O.O 9.0 9.0 42 2.0 5-5 7-5 71 " 27 1-45 " O.O IO.O IO.O 47 2.O 5 o 7-o 66 " 27 2.OO M. 0.3 II. 11.3 53 4.0 6.0 IO.O 94 " 27 2.45 P.M. O.O IO.O IO.O 47 4-o 5-o 9.0 85 " 27 1. 00 " o.5 IO.O 10.5 50 6.0 8.0 14.0 132 " 27 I.I5 0.0 6.0 6.0 28 1.0 7-0 8.0 75 " 27 I.JO " O.O 5-o 5.0 24 1.0 6-5 7-5 71 Ian continuously over night. " 28 9.15 A.M. 0.0 6.0 6.0 28 O.O 5-0 5-0 47 " 28 9.25 " O.O 7.0 7-0 33 0.0 5-o 5-0 47 " 28 9-35 " 0.0 6.0 6.0 28 0.0 5.0 5-o 47 " 28 9-45 0.0 6.0 6.0 28 O.O 4.0 4.0 38 " 28 9-55 " O.O 5-0 5-0 24 0.0 4o 4-5 42 hut down from 10.00 A.M. to 3.00 P.M. " 28 3.10 I .M. 0.0 9.0 9.0 42 O.2 7-0 7-2 68 and cleaned scale and mud from " 28 3-40 " O.O 9-5 9-5 45 0.0 6.0 6.0 57 electrodes No. I. " 28 4.10 " 0.0 9.0 9.0 42 0.0 6.0 6.0 57 direction of flow of electric current " 28 4-30 " O.O 9.0 9.0 42 O. I 3-O 3-1 29 reversed in both cells at 4.15 P.M. " 28 4 40 " 0.3 9.0 9-3 44 0. I 2.5 2.6 25 " 28 5.10 " 0.5 9.0 9-5 45 0.2 2-5 2-7 25 of the positive plates (equal to one side of all plates) of set No. i was 1510 square inches, and the area of the electrolyte was 1300 square inches. In set No. 2 the correspond ing figures were 3320 and 3170, respectively. A potential difference of 1.5 volts was main tained on set No. i, and of i.o volt on set No. 2. From 10.00 A.M. to 2.00 P.M., Sept. 30, set No. i was operated at a potential differ ence of 3.5 volts. The rate of flow of water was maintained uniformly throughout the first run at 0.5 cubic foot per minute through each cell, and determinations were made every two hours of the amounts of iron in the water as it left the cells. During this run the elec tric current was held uniformly at 6.0 am peres. On the second run the strength of electric current averaged 15 amperes. The rate of How of water was i.o cubic foot per minute. 402 WATER PURIFICATION AT LOUISVILLE. The results of the determinations of the amounts of iron in the water as it left the cells, and the percentages which these amounts were of the theoretical, are given in the next table. During the first run the gas evolved in cell No. i appeared to be almost constant in amount. In cell No. 2 almost no gas was observed up to 6.30 A.M., Sept. 29. From this time the formation of gas steadily in creased in amount till at the close of the run there was nearly as much gas being formed in No. 2 as in No. i. On examination the positive plates of set No. 2 were found to be covered with a heavy coating of what appeared to be an irregularly hydrated form of red oxide, somewhat granu lar in form. In the bottom of this ce l was found a heavy accumulation of red oxide, green hydrate, and scale. TABLE SHOWING THE AMOUNTS OF IRON LEAVING THE TWO CELLS CONTAINING STEEL ELECTRODES ON WHICH POTENTIAL DIFFERENCES OF 1.5 AND 1.0 VOLTS, RESPEC TIVELY, WERE MAINTAINED. Date. 1897. Hour. Cell No. 11.5 Volts. Cell No. 21.0 Volt. Water. Parts per Million. Per Cent of the Theoreti cal. Water. Parts per Million. Per Cent, of the Theoreti cal. Sept. 28 2.00 I .M S.O no 6.4 88 28 3.00 " 7-5 tor 5-2 7i " 28 5.OO " 9.3 134 7-2 98 " 23 y.OO " 8.0 1 08 5-3 72 " 28 9.00 " 5-3 79 6.5 89 " 28 11.00 " 6.0 82 7.0 95 " 29 I.OO A.M. 6.3 86 7-4 IOO " 29 3.OO " 8.0 108 6.5 38 ; 29 5 . OO " 6.0 81 6.0 Si " 29 7-OO " 5-8 79 7-3 99 29 9.OO " 9.0 122 5-8 78 29 1 1. 00 " 8.0 I 08 6.1 S3 " 29 i.oo r.M. 6.1 83 5-8 78- " 29 3.00 " 5-9 80 6.3 85 ; 29 5-oo " 5-2 70 5-7 77 " 29 7.00 " 5.9 So 6.4 86 29 9.00 " 6.2 84 5.8 73 29 II. OO " 7-9 107 6.1 83 " 30 1. 00 A.M. 7-8 105 5 6 77 30 , 3.00 " 6.7 91 5-7 77 30 5.00 " 6.6 89 i 5-8 73 30 7.00 " 6.7 91 8-3 112 " 30 9.00 " 6.7 91 S.o 108 Averages Increased voltage on 6.8 No. i, t 92 6.2 84 -> 3.5 am stopped No. 2. Sept. 30 1 30 1 l.uu . i.m. I.OO I .M. II-5 [0.8 118 30 30 1.30 2.OO " s s 93 It will be seen that the amounts of iron leaving cell No. i were somewhat greater than in the case of cell No. 2, the averages being 92 and 84 per cent, of the theoretical rate, respectively. A slight increase in the amount of iron leaving the cell was noticed after increasing the voltage and current in No. i, but as this soon returned to the normal it was con cluded that the first high results were due to the effect of the increased volume of water and electric current removing the metal pre viously decomposed, from the plates and por tions of the cell upon which they had lodged. It was evident from these experiments that iron could be decomposed at a potential dif ference as low as i.o volt, and that gas was formed in the process. There was some indication that the higher potential was slightly more efficient than the lower, but the differences were not great enough to make certain that differences in arrangement of supporting framework under the electrodes, or slight variations in the dis placement of the water in the cells, were not the explanation of these results. To make this point clear new electrodes were prepared of the same metal, and river water treated at 10, 5, 2.5, 1.75, and 1.5 volts difference in potentials, respectively. Each experiment was continued to the equivalent of 0.05 am pere-hour per gallon of treated water and the amounts of iron were determined. Within the limits of accuracy no difference could be found in the rate of hydrate formation at the several potential differences. It was therefore concluded that potential differences between the limits of i.o and 220 volts exerted no apparent influence on the rate of formation of hydrate; that scale and gas formed with apparently the same rapidity at all potential differences; and that the prac tical limits of construction with references to the area of electrode surface and length of electrolyte would be the controlling factors in determination of the potential difference to employ in practice. Influence on tltc Formation of Hydrate of Cur rent Density. The current density ranged from 0.30 to 2.08 amperes per square foot of active elec trode surface during these tests. For the SUMMARY AND DISCUSSION OF DATA OF 1801 43 most part it was about 1.04 amperes with the large devices. In the case of some labora tory experiments a current density as high as 50.4 amperes per square foot was employed at times, but the usual density was about 15 amperes per square foot. In connection with the low potential experi ment described above, current densities as low as 0.26 ampere per square foot were em ployed. Within these limits no marked influence of current density upon the formation of hydrate was noticed, but it is probable, as noted above, that the deposition of iron at the nega tive pole was somewhat greater with the high than with the low densities, though the in creased rapidity of replacement of electrolyte due to increased flow of water through the cells would probably compensate for this in a measure. Theoretically, as has been ex plained above, the lower the current density the lower the rate of deposition of metal on the negative pole. Current density as low as admissible with economic construction of cells should therefore be employed. Influence on the Formation of Hydrate of the Composition of the Iron. The rolled wrought-iron plates used in this work were of the following percentage com position: Carbon 0.06 Silica 0.28 Sulphur 1.58 Phosphorus o. 14 Manganese o.oo Iron 97-94 So far as was learned the only way in which the composition of the iron might affect the formation of hydrate was indirectly through its effect upon the passivity of the metal to the acid ions of the water. This subject has been dealt with above in connection with pas sivity, when it was stated that no difference in hydrate formation was apparent in the va rious grades of metal used here, which were clearly due to differences in the composition of the several irons and steel. From a practical point of view the differ ences in passivity of electrodes due to varia tions in the composition of the metal would, under the conditions of these tests, have been almost completely disguised by the acquired passivity caused by coatings of iron oxide. Directly, the composition of the river water influences the formation of hydrate princi pally through the action of the atmospheric oxygen dissolved in the water in converting the iron into the insoluble ferric hydrate. This action is a very important one. It is also probable that the suspended matter affects the character of the surface coatings especially on the negative electrode. Indirectly, the relative amounts of the dif ferent acid ions in the river water influence the formation of hydrate through their dif ferent relations to the passivity of the iron. In connection with the current density, also, the composition of the river water is a factor in the consideration of deposition of metallic iron on the cathode. The last two points are referred to in detail at the beginning of this section. Influence on the Formation of Hydrate of Re versing the Direction of the Electric Current. When the direction of flow of the electric current passing through the electrodes and electrolyte is reversed, the positive electrode (which was previously the negative) is at the outset saturated with hydrogen gas and the surface is coated with metallic iron, probably mixed somewhat with suspended matters from the water. Comparison in the last set of tables of the amount of iron in the water as it left the cells before and after reversals of current on July 23 at 4.00 P.M., and on July 28 at 4.15 P.M., shows that in three of the four instances the rate of decomposition c:f iron suffered a marked diminution for more than an hour. The cause of this is not clear , y understood, but it appears to be associated with the occluded hydrogen in the pores of the metal and with the surface coatings, which will vary of course with the frequency of re versal and the character of the river water. In practice these marked diminutions in the WATER PURIFICATION AT LOUISVILLE. formation of hydrate would be a serious mat ter and for a time would require the opera tion of a reserve portion of the plant, both with regard to the cells and the generating appliances. When the electric current was shut off from the electrodes and the cells kept full of water it was repeatedly noted that the decom position of iron and formation of gas continued for a long time. This was due to a galvanic action, the metal being electro-positive to the surface coating of oxide. During the month of April and early part of May, when the elec trical devices were out of service on numer ous occasions while tests with chemicals were being made, it is estimated that the total weight of the electrodes decreased 65 pounds, due to this factor alone. On the grounds of economy and of comparable conditions for reliable data, it became necessary to drain the water from the cells as soon as the electric current was stopped. So far as is known the subsequent formation of hydrate, when the electric current was applied following a pe riod of rest in which the electrodes were cov ered with water, was not seriously influenced by this procedure, which, however, for the reasons stated above, was found to be imprac ticable. After the first of June the water was drained out of the cells as soon as the current was turned off. In consequence of the action of the air it was found that the rusting of the electrodes thus produced, increased the ac quired passivity of- the iron, and, when the electrodes were again put in service, the rate of formation of hydrate was abnormally low for a time. This is shown in the tab e on page 400 by the results on the morning of July 24, when, following a rest after draining the cells, of about 41 hours, the electrodes did not yield the normal amount of iron for half or three- quarters of an hour. In other cases, where the period of rest was longer the evidence shows a more prolonged diminution in the rate. Per Cent, of Metal Wasted in this Process. When the potential difference of the cur rent between adjoining plates was 3 volts or more the evidence showed a rate of decompo sition of iron on the positive plate equivalent to about 100 per cent, of the theoretical rate of 1.05 grams, or 16.2 grains per ampere-hour. Of this iron an amount equivalent to about 10 per cent, of the theoretical rate was found deposited on the negative plate. The amount of iron leaving the cell in the form of avail able hydrate seemed to vary considerably, but averaged about 40 per cent, of the theoretical rate. Taking into consideration the fact that eventually the plates become too thin for use and have to be discarded, it seems fair to con clude that in this process of producing iron hydrate substantially one-half of the metal is wasted by passing into the water in the form of non-hydrated and non-available scales of iron oxide. The experiments of Sept. 28 to 30 indicate that no substantial advantage in this respect would be obtained with potential differences as low as i volt. Resistance to the Passage of Electric Current of Films of Iron Oxide. The results of analyses and of observations of scales in the water leaving the cells showed that the films of iron oxide attached to the positive electrodes remained there only tem porarily, and came off at an irregular rate from time to time. In consequence thereof it is not probable that the entire surface was covered at any one time, in the course of regular operations, with a film which very materially increased the resistance of the elec trodes. Compared with new metal the plates doubtless offered a certain resistance, but within the ranges of service to which these electrodes were subjected the increase in re sistance was within the limits of observation, or less than I volt at 400 amperes or .0025 ohm. Percentage of Electric Power Wasted in this Process. Under the above-described conditions of operation, with potential differences between SUMMARY AND D1SCUSSJON OF DATA OF 18 J7. 45 the electrodes of 3 volts or more, the evidence shows that between 50 and 60 per cent, of the electric current was wasted in removing iron in the form of scales, due to the formation of oxygen at the surface of the plates, and in depositing some of the available metal upon the negative electrodes. This does not in clude the effect of scales in offering increased resistance to the electric current, as noted in the last paragraph. Combining the three items, the waste of electric power may be safely placed at 60 per cent. The experiments of Sept. 28 to 30 with po tential differences as low as i volt indicate that there is no reason under these conditions for modifying the above figures. Influence of this Process on Subsidence and Fil tration. Comparing the effect of this process upon the efficiency of subsidence and nitration of water with that of chemical treatment such as persulphate of iron or sulphate of alumina. when the degree of coagulation is the same, it is to be pointed out that under the condi tions tested the hydrogen gas evolved at the cathode has a slight effect in retarding sub sidence, and also at times reduces the length of run between washings of the filter by plug ging up some of the pores of the sand layer. Compared with aluminum the iron elec trodes offer much less difficulty with gas for mation when equal coagulation is obtained. Some laboratory observations in May showed that the ratio of gas formed by equivalent co agulation with aluminum and iron electrodes was about 2 to i. In practice as the elec trodes become covered with scale this ratio is greatly increased. Thus the ratio of gas formation with aluminum and iron electrodes after long service in connection with the filter, and under the same conditions other than the surface coatings, was found on June 8 to be i ^o to 4. It is to be noted that these ratios refer to amounts of evolved gas. As the amount required to saturate the water formed different percentages of the total, the ratios are not absolutely exact. Influence of the Process on the Composition of Filtered Water. Like all electrolytic processes of this gen eral type, the iron process does not add to the filtered water any mineral acid to make the water less desirable when used in boilers, and it does not add to the water any free carbonic acid to affect corrosion. In addition to the ordinary removal of suspended mineral and organic matters and a slight removal of dis solved organic matter, however, it removes the atmospheric oxygen in the water in amounts proportional to the iron obtained as hydrate. Up to a certain point this factor is of little practical significance, but when the process is carried to a degree where the oxy gen is all or nearly all used for this purpose there is danger of some of the iron passing into the filtered water. A filtered water con taining no oxygen would also be undesirable in several ways. Up to a certain point, there fore, this process can be used with satisfaction so far as composition of the filtered water is concerned. Beyond this- point (exhaustion of atmospheric oxygen) the application of this process is inadmissible. From the evidence presented in this section it may be concluded: 1. Under practical conditions this process can be used to produce ferric hydrate, a good coagulant, up to the point where the atmospheric oxygen dissolved in the water is not completely exhausted. 2. The evolution of gas is fairly small, com paratively speaking, but the indications were that at times the gas might exert a retarding influence upon subsidence and a clogging effect upon filters. 3. The rate of production of ferric hydrate was reasonably uniform at its minimum limit, except for periods in the vicinity of one hour following a reversal of the direction of the current and an exposure of the plates to the atmosphere. 4. Owing to galvanic action when the coated plates were allowed to remain in water when out of service, the loss of metal made it imperative to avoid this procedure except for very short intervals. 5. Under conditions of good practice the amount of metal wasted as oxide scale would be substantially 50 per cent. 406 WATER PURIFICATION AT LOUISVILLE. 6. Under conditions of good practice the amount of power wasted would reach about 60 per cent. SECTION No. 5. DETAILED ACCOUNT OF THE ELECTROLYTIC FORMATION OF ALUMINUM HYDRATE IN THE OHIO RIVER WATER. During the latter part of 1896 and first part of 1897, several factors operated together to bring forward again the process of the forma tion of a coagulant by the electrolytic decom position of metallic aluminum. As was explained at the close of Chapter XI 1, inves tigations in July and August, 1896, led to the conclusion that the use of electrolytically pre pared aluminum hydrate was out of the ques tion because of its excessive cost, but that the use of an electrolytically prepared coagulant presented certain distinct advantages over the use of sulphate of alumina. Chief among the factors which led to further consideration of this process were the following: 1. The operations with iron electrodes showed clearly that potential differences very much lower than those employed during July and August, 1896, could be used in this general process with equally satisfactory re sults, thus greatly decreasing the cost of power. 2. The cost of metallic aluminum had de creased to about one-half its cost in August, 1896. 3. It appeared that the electric current gave results nearly twice as effective when new aluminum electrodes were used as when iron plates were employed. This was of much practical significance in consideration of the construction and maintenance of a plant. 4. President Long requested that the pro posed investigations of coagulants be made as full and thorough as consistent with the prac ticability of the results obtained. For these reasons the use of electrolytically prepared aluminum hydrate was again given attention in connection with other coagulants. Early in the course of the investigations it was found that aluminum in the form of elec trolytically prepared hydrate was no more effective than when in the form of hydrate obtained from the decomposition of the sul phate by the lime in the river water. It was also seen that the efficiency of aluminum plates in practice was very much less, both in amount and in regularity, than was indicated to be the case by the results with new plates. It seemed inadvisable, therefore, to make the study of the details of this process as ex haustive as those of the formation of iron hy drate electrolytically, although in the main the investigations of both processes were car ried on simultaneously. The following ac count, while it covers the bulk of the ground fully investigated in connection with iron electrodes, is, accordingly, not as complete as the investigation in the case of iron as re corded in section Xo. 4; and, furthermore, owing to the peculiar and widely varying re sults, it has been necessary in some instances in order to account for certain observations to introduce explanations and theories, as an exhaustive study of these points was not war ranted by the available time and the imprac ticability of the process. The same general plan as was employed in the presentation of the investigations of the iron process is followed here. As has been presented in the general dis cussion of the process of decomposition of metals under the action of an electric current, the percentage of the acid ions which attacks and decomposes water may be stated to repre sent the degree of passivity of the electrodes, on the assumption that when an electrode is perfectly non-passive (completely active), all of the acid ions attack the plate and no water is decomposed at the anode. The principal points learned in regard to the action of the acid ions upon aluminum electrodes, so far as it was considered prac ticable to investigate the subject, were as fol lows : i. The fact that in all cases within the lim its employed a gas appeared to be liberated at the positive pole when aluminum elec trodes of bright metal were used in electro lytes formed of pure solutions of each of the several acid ions normally present in the Ohio SUMMARY AND DISCUSSION OF DATA OF 1897. 407 River water, indicates that aluminum is nor mally passive to all of these ions to a certain but variable degree. 2. Aluminum seems to be most passive to the ions of nitric acid. This would be ex pected from the fact that aluminum is not readily soluble in nitric acid. 3. The acquired passivity of aluminum electrodes, that is, the passivity due to the formation of a coating of oxide on the face of the plates, continually increases in prac tice; and while it fluctuates somewhat, due to the scaling of the plates, there is strong evi dence to indicate that after considerable ser vice the rate of formation of the hydrate might decrease to almost nothing on this ac count. It is to be noted in this connection that the metallic aluminum used during these inves tigations was the very best grade of commer cial rolled plate. How far this may have affected the results is difficult to say, but the indications are that a less pure grade of metal might be more readily soluble, although pos sibly no more efficient so far as the formation of available hydrate is concerned. It is fur ther to be borne in mind that the passivity of the metal is dependent upon the solvent ac tion of the secondary compounds as well as of initial ions; and, as presented beyond, the general instability of aluminum compounds is very important in this connection. Number of Experiment! Averaged. Electrolyte. Rate of per Ampere-Hour. Grams o 51 0.77 O.C)I 0.57 0.67 0.64 0.58 0.76 0.77 0.58 0.57 0.61 1 8 (1896) 2 2 2 2 2 3 3 3 3 3 By relative mentatio Filtered water 7.84 II.QO 9.40 8.80 10.30 9.80 8.90 11.70 11.90 8.90 8.80 9.42 Carbonic acid* Hydrochloric and sulphuric Hydrochloric and nitric acids* Hydrochloric and carbonic Hydrochloric, sulphuric, nitric, and carbonic acids* Hydrochloric, sulphuric, nitric, carbonic acids,* and carbonic efficiency in sedi- n (section No. 6). .River water water. Il was Mined in several of these experiments that dissolved aluminum was present in the treated water, although its fcrm was nut ascertained. Rate and Form in which Aluminum leaves the Positive Pole. A. New Metal. The rate at which aluminum leaves the positive pole under the action of an electric current when the electrodes were composed of bright metal plates was determined at dif ferent times and under varying conditions, as shown in the opposite table. B. In Practice. The results in this connection, obtained from long use of the aluminum electrodes on a large scale, are given in the next table. On account of faulty insulation of the large electrodes no results of positive value were obtained with them previous to May 30. From this date to June 18 one or both of the large aluminum electrodes were kept in prac tically continuous service. They were re moved, rinsed and weighed nearly everv day up to June 8. The deposition of the metal and silt on the negative pole caused the weight of the electrodes to increase regularly, however, so that these first weighings were of no value. On June 8 electrodes No. 2 were dismantled and the separate plates weighed and this set rebuilt. At this time one-half of the plates was left out of the set, the space be tween plates doubled, and the direction of the current reversed. The second run from June 8 to 12 does not therefore represent the ac tual loss of metal of the positive pole, but this loss plus the loss in weight due to the removal of the deposited matter on the surface. The total ampere-hours during this run were 58^0 and the loss of metal and deposit 81 pounds, or at the rate of 6.3 grams per ampere-hour. Electrodes No. 4 were operated from June 3 to / and were dismantled and weighed on June 10. They were operated again from June 10 to 16, after which frequent reversals of the direction of the current prevented any study of the loss in weight of any particular portion of the set. The results of these determinations are given in the following table: 4o8 WATER PURIFICATION AT LOUISVILLE. SUMMARY OF RESULTS. Weight of Electrodes No. 2. Electrodes No. 4-. June 12 ) " 15 ) Total.. . Average 21 470 20.75 6.8 0.44 66 850 37.00 . . . .3.89 0.25 In connection with the above data the fol lowing points are to be noted: 1. The atomic weight of aluminum is 27 and its combining weight is generally ac cepted as 9, or in other words its valency is 3. On this basis the amount of metal decom posed per ampere-hour, assuming that the full current is active in bringing to the anode surface decomposing ions to form normal trivalent salts, is 0.34 gram, or 5.23 grains. 2. As is generally known, certain salts of aluminum, notably those of the mineral acids, are comparatively unstable and have the power of decomposing the metal. 3. This supplementary solvent action is not limited to the formation of a basic salt, and in fact may not occur in this manner at all. The evidence indicates that under certain conditions which appear to occur in electro lytic cells, the initially formed salt becomes decomposed; the metal separates out as a more or less hydrated oxide; while the acid is free to attack new metal and form new salt, thus making the action in a measure con tinuous. 4. In an electrolyte composed of a com bination of several acid ions it is believed that hydrochloric acid attacks the metal to the greatest extent, if not solely, and that the re sultant chlorides are acted upon by the other acids, leaving the hydrochloric acid free for further solvent action. 5. According to Watts, the authority quoted in Chapter XII, the rate of decompo sition of metallic aluminum is 0.51 gram, or 7.84 grains, per ampere-hour. The total average results of our experiments in August, 1896, agree with this very closely. If this rate were specifically correct, and not depend ent upon circumstantial factors, it would im ply that the metal left the electrodes in the form of divalent salts, as in the case of iron. 6. In view of the fact that the ultimate compound of aluminum, the hydrate, gives every indication of being in the trivalent form, it would require some atmospheric oxygen dissolved in the water in order to convert the initial compounds into the trivalent hydrate, supposing that initially they were divalent. The facts show that no appreciable diminu tion in the dissolved oxygen occurred; and this seems to be substantial proof against the divalent form of the initial compound, and the theoretical rate of 0.51 gram per ampere hour, as was indicated to be correct in 1896. 7. With regard to the formation of oxygen at the anode the direct observations were con flicting, but the evidence shows clearly that even with bright metal there is formed some oxygen which attacks the metal, producing a scale of alumina rust (oxide). The rust thus formed, together with a similar accumu lation coming from the secondary reactions of the initial salts, produces a surface coating which not only reduces the supplementary solvent action, but also gives to the bright metal beneath the coating an acquired pas sivity. 8. In Chapter XII it is recorded that alu minum hydrate was obtained when metallic .aluminum was used only for the negative elec trodes, with the comment that it was not in harmony with the present views of electro chemistry. Except in a few instances where small amounts of aluminum were obtained in special laboratory experiments, there were no indications from the work in 1897 that the electrolytic decomposition of the negative electrodes was a factor under practical condi tions. However, it is not unreasonable to suppose that under some circumstances the strong alkaline solution present at the surface of the cathode might dissolve this metal, which has the power of acting as an acid and of forming aluminates. In conclusion it may be stated that the available evidence taken as a whole points clearly to the decomposition of aluminum un der the conditions of practice only at the posi tive pole. Here it is initially removed from SUMMARY AND DJSCUSSION OF DATA OF 1891 409 the plate as a trivalent salt of the strongest acid ions ( perhaps only in the form of chlo ride). With bright metal the plates are not only free from marked passivity, but the ini tially formed salts appear to have a supple mentary solvent action, with an accompany ing deposition of rust upon the surface of the anode. This coating, formed in this manner and by the action of oxygen, which in small quantities is set free at the anode, gives event ually to the electrode a marked acquired passivity and reduces materially the supple mentary solvent action. In consequence of this the rate at which the metal leaves the positive pole diminishes in time to almost nil, although at first a supplementary solvent ac tion causes it to exceed the theoretical by a large, although variable, percentage. I onn and Rate of Deposition of Aluminum on the Negative Pole. During the course of the experiments by the Water Company in August, 1896, it was found, as has been presented in Chapter XII, that aluminum is deposited on the negative pole simultaneously with the formation of alu minum hydrate. The reason of this seems to be the same as in the case of iron electrodes, that a portion of the metal from the positive pole, which is dissolved by the acid ions, be comes itself ionized, and acting as a negative ion or cathion transfers a charge of electricity from the positive to the negative pole, where on neutralization of the charge it is deposited on the face of the plate. It is well known that electroplating with aluminum is ex tremely difficult and in some cases practically impossible, owing to the rough and weakly adherent characteristics of the deposit. This is just the condition which was found to exist in the process of formation of the hydrate electrolytically, and to the character of the de posit one of the most serious factors against this process, aside from the cost of the metal, is due. It has been found in the course of many experiments that very finely divided silt or clay particles suspended in the electrolyte are deposited upon the negative pole to a greater or less extent. This deposition of silt, taking place at the same time as the depo sition of the metal in its porous form, results in a combined layer or deposit of silt and metal of high electric resistance. The effect of this formation on the amount of power re quired is shown in one of the latter portions of this section. The amount of metallic alu minum contained in any of the deposits found in practice could not be readily determined, owing, as stated above, to the presence of silt. Some idea of the significance of this point can be obtained from the following table, in which the weights of deposits for the several runs with electrodes Nos. 2 and 4 are recorded, together with the ampere-hours of service and average rates of deposition. TABLE OF RESULTS. Electric Period. Current in A Hr- Weight of Negative Electrodes. ~^~ Electrodes No. 2. June 3 s| -! 53So I24 25.4 I.flJ I! ,*} 58 3 o ,4 24.0 1.56 Total 51210 33 27.0 1.75 Electrodes No. 4. JU .? C 3 6 \ .5730 to 6.4 0.42 I! \l\ *I470 22 7-2 0.47 Total 37200 32 6.0 0.39 Total for both electrodes.. 88410 170 13.3 0.88 Influence on flic I onnation of H\dratc of the Composition of tlic Rircr Water. The formation of aluminum hydrate elec trolytically is affected chiefly by three factors: 1. Initial passivity. 2. Supplementary solvent action of the salt formed when the metal leaves the plates. 3. Acquired passivity due to coatings formed on the surface of the plates. So far as our observations went, no decisive evidence was obtained that the composition of the river water materially affected the first two of these factors, but if it did do so the re sult was disguised by the third factor. The formation of hydrate was influenced very ma terially by the surface scales or coatings (caus ing acquired passivity), which are next dis cussed with reference to the composition of the river water. 4io WATER PURIFICATION AT LOUISVILLE. So far as the data go the indications are that the variations in the character of the river water with regard to dissolved constituents caused only a very small percentage variation in the amount of scale formation on the posi tive plates. The character and amount of the suspended matter, however, exerted consider able influence on the formation of the deposit on the negative plate. This is quite clearly shown by the difference in the increased weights of electrodes Nos. 2 and 4, the very finely divided suspended matter of the last days of May greatly increasing the amount of deposit on electrodes No. 2 as compared with electrodes Xo. 4, which were operated some what later. That a certain percentage of this fine material was transferred from the positive to the negative pole after reversing the direc tion of the current seems to be the explana tion of the high rate of deposition with elec trodes No. 2 from -June 8 to 12. This sup position was borne out by the change in the character of the matter on the negative pole after the current was reversed, the presence of the silt being quite marked after the run from Alay 29 to June 8, while almost no silt was found in the scale after removal by the reversed current. Some idea of the signifi cance of this action can be obtained from the following table, in which the initial amount of power per ampere-hour per gallon at the rate of treatment of 23.5 cubic feet per minute is given for the start of the first run of each of the two sets, after 15,000 ampere-hours, and for electrodes No. 2 at the end of the first run (see " Form and Kate of Deposition on Negative Pole." page 409). HORSE-POWER REQUIRED PER AMPERE-HOUR PER GALLON. May 30, i i.oo A.M June i, =;.oo I .M " 8, 6.OOA.M " 3, 6. oo r.M " 6, S.OOA.M The phites were only rinsed and i Ampere- Hours Service Since Last Cleaned. El-clric H.P. per Ampere- Hour per Gallon. o* 450 15 ooo 45 ooo I 2OO 3080 o 240 15 ooo 435 :raped at last cleaning. After the discussion of foregoing topics it is obvious that the coatings or scales, com posed of aluminum oxide (rust) and clay par ticles, not only increased the required amount of power very largely (shown above to be sevenfold), but also increased the acquired passivity so that with the increased power used there was actually less aluminum hy drate produced. Influence on the formation of Hydrate of the Presence of Scale. The presence of coatings or scales, princi pally of aluminum rust, upon the surface of the plates exerts a marked influence upon the process, as follows: 1. It reduces the supplementary solvent ac tion of the aluminum salts and thus makes the actual rate of loss of metal at the positive pole fall to the normal rate, approximately. 2. It gives to the metal an acquired pas sivity which makes the actual rate of hydrate formation drop far below the normal; in fact we have seen instances where the formation of hydrate was reduced to almost nothing. 3. When the scale becomes very thick and has sufficient tenacity to remain in place with out breaking in pieces, we have repeatedly noted, especially for a short time after revers ing the direction of the current, that practi cally all of the aluminum hydrate remained between the metal and the layer of alumina rust. 4. From the last table it is plain that the scale makes a large increase in the resistance to the electric current, and the reduced rate of formation is concomitant with an increased amount of electric power. Influence on the Process of Reversing the Direc tion of the Electric Current. After electrodes have been in service for some time the positive plates become covered with a scale of oxide and the negative ones with a more or less heavy coating of deposi ted metal and silt. Under such conditions a reversal of the electric current results in the complete removal of the deposit from the old negative poles and the cessation of removal of oxide scale from the old positive poles. SUMMARY AND DISCUSSION OF DATA OF 1897. 411 It is difficult to determine in how far the reversing of the current affects the rate of formation of hydrate. That hydrate was formed in large amounts after a reversal was evident from inspection, but, owing to the fact that the coating on the old negative plates retained a large amount of the newly formed hydrate between it and the plate, it was not possible to determine the rate of formation of the hydrate. This retention of the hydrate would be a very serious factor in practice, however, necessitating a mechanical removal of the coating when the current was reversed. In regard to the effect of reversing the cur rent on the resistance of the electrodes, the evidence indicates that if a suitable means could be provided to remove the coating as it cracks off the old negative plates, a con siderable increase in conductivity might be gained. In practice, however, it is probable that the combination of the new deposit and old oxide scale which would form on the negative (old positive) plates after reversal would result in increasing the resistance as fast as or faster than the removal of the coat ing from the positive (old negative) plates would reduce the resistance. Furthermore, the almost immediate formation of a layer of hydrate which would be retained on the new positive poles would, unless removed, result in an increase of the resistance. These effects on resistance are shown by the records of potential differences on cell Xo. 4 on July 16, 17, and 1 8, in the next table. The electrodes were rinsed at 6.08 P.M. on July 16 and the direction of the electric cur rent reversed after 7960 ampere-hours service since the plates were last cleaned. From 6.30 P.M. to 12.00 P.M., July 16, the resistance of the electrodes remained constant at 0.071 ohm. At 12. oo P.M. the direction of the elec tric current was reversed. Tt was again re versed at 6.00 A.M. and 12.43 P.M., July 17, and at 12.30 P.M., July 18. The records of the observed resistance are opposite. Percentage of Metal U asfed in tliix Process. On account of the very irregular manner in which the scale came off from the positive pole during practice, no observations were attempted to learn the relative percentages of R esistanc e in ims. Hour. Ju y >7- July ,8. A.M. .. * M. 12.30 O. IO2 o 130* O 10 0.124* I .OO 0. 120 o 104 04 0.088 I .30 O.II6 o 57 03 0.124 2.00 0. Ill o 168 04 0.150 2.30 0.094 o 164 o 08 o . i 60 3-00 0.089 157 o 08 0.162 3.30 0.089 o M4 10 0.162 4.00 0.086 131 12 i O.lSj 4-3" 0.089 o 131 o 3 5.00 0.089 o 122 o 15 5.30 0.089 o 122 o 17 6.00 0.089* o 124 o 17 6.30 0.069 o I 3 o 19 7.00 0.094 o I 3 24 7.30 0.094 o 112 o "3 8.00 0. 102 108 o 17 8.30 0.089 o 104 o 17 9.00 O. I IO o 1 06 o "9 9-3 O. I IO o 106 22 o oo o. no o 106 o 9 0.30 o. no 106 22 1. 00 O. I IO o 106 o 24 1.30 0. I 10 106 o 24 2.OO 0.139 o 104 24 metal removed as hydrate and as oxide. Judging from the results in July and August, 1896 (see Chapters XI and XII), and also from observations of the amounts of scale re moved from the plates in cleaning during 1897, it is estimated that at least 50 per cent, of the metal removed from the positive pole was lost in the form of scale, and that from i o to 20 per cent, more was deposited on the negative plate in a manner which made it very slightly or not at all available. During practice it would seem reasonable to figure on an average loss of about 65 per cent, of the metal removed from the plates. Influence of Scale and Deposit on the Amount of Potvcr Required. This subject was referred to under the dis cussion of the effect of the composition of the river water. As presented there it was found in practice that the formation of scale on the positive poles and the deposit on the nega tive pole increased the amount of power re quired per ampere-hour per gallon from 240 electric II.! . with new plates (electrodes No. 4) to 3080 with old plates of 45,000 am pere-hours service (electrodes No. 2). This increase of 740 per cent, does not probably 412 WATER PURIFICATION AT LOUISVILLE.. represent the limit of the increase, as the re sistances of both electrodes Nos. 2 and 4 were steadily increasing at the time they were taken out of service. Assuming an efficiency of 50 per cent, in the formation of hydrate the above figures represent an increase from 246 to 1825 electric H.P. to treat 25 million gal lons per 24 hours with the equivalent of i grain per gallon of sulphate of alumina. The effect of reversal on scale formation has already been presented. So far as could be learned the only possible way to reduce the resistance of the electrodes was to remove the plates and scrape the sur faces fairly clean. This of course involves a considerable loss of metal, and its cost for la bor alone would make it prohibitive. No practicable means was found of doing away with these scales and deposits with their at tending effects. Percentage of Electric Power Wasted in this Process. In the formation of aluminum hydrate by this process power is wasted in two ways: 1. A certain percentage of the current is transferred by ions which do not attack the pole, but decompose water. This current is entirely wasted and, as power is required to produce it, such power is lost. Owing to ir regularities in removal of scale it was found that determinations of the amount of hydrate formed during practice were not feasible. Basing the conclusions on the result of many laboratory tests and on the amount of scale formed during practice, it is estimated that not more than 50 per cent, of the current is avail able in the formation of hydrate, and there are indications that the loss might at times approach very nearly to TOO per cent. 2. The loss of power due to increased re sistance of the electrodes as presented in pre vious sections of this chapter reached as high as 740 per cent, of the initial power required, and the indications were that the limit had not been reached. Frequent scrapings of the plates, such as would be required to prevent this increase of resistance, are impracticable. Conclusions. With the present knowledge of electro chemical actions, and with the present cost of aluminum in the form of plates as contrasted with equal amounts of metal in the form of commercial sulphate, the use of hydrate of aluminum prepared by the electrolytic decom position of the metal is out of the question on account of cost. This is shown by the fol lowing summary: 1. Aluminum in sheet form costs in car load lots according to the latest quotation 27 cents per pound. In the form of sulphate of alumina one pound of metal costs 16 cents. The ratio of cost of equal amounts of coagu lant prepared by the electrolytical decompo sition of the metal and by the chemical de composition of the commercial sulphate by lime is therefore 17 to 10 for aluminum alone. 2. In operating there would be a constant loss of about 50 per cent, of the metal re quired for the formation of the hydrate due to the passivity of the electrodes to the acid ions, the supplementary solvent action of the salts formed, and the consequent formation of oxide scale. This might at times approach 100 per cent. 3. The amount of power required would constantly increase with the age of the elec trodes, necessitating at frequent intervals the removal and scraping of the plates. This last step would be very expensive. Under nor mal conditions probably 50 per cent, of the normal power required would be wasted in overcoming the resistance of the surface coat ings. The normal amount of power would also be increased from 50 to 100 per cent, to offset the reduced rate of formation of hy drate, due to the acquired passivity of the metal. 4. In short, the process was impracticable, under the conditions of these tests, both with regard to economy and regularity of produc tion of hydrate. While it is probable, if not certain, that prolonged investigation would improve the process by using a different grade of metal and devising mechanical appli ances for the removal of scales, yet in the light of our knowledge, owing to the inherent character of the metal and the narrow range of conditions as applied to this line of work, commercial success of this process seems to be an impossibility. 4 3 SECTION No. 6. RELATIVE EFFICIENCY OF AVAILABLE CO AGULANTS, BASED ON EQUAL WEIGHTS OF METAL USED AND ALSO ON THE AMOUNT OF ELECTRIC CURRENT IN THE CASE OF ELECTROLYTICALLV FORMED COAGULANTS. In order to arrive at the most economical coagulant to apply to the water, it is essential to know the relative efficiencies of those available for the purpose. From foregoing sections it is clear that the available coagu lants are four in number, obtained by the fol lowing treatments: 1. Sulphate of alumina. 2. Persulphate of iron. 3. Electric current on aluminum elec trodes. 4. Electric current on iron electrodes. With the first two (sulphates) the compari sons are expressed in their final form with ref erence to the amount of metal contained in the commercial products used for the tests. This is necessary because the amount of metal which determines the quantity of hydrates varies in different lots of sulphates of the same kind. The reason that this is so important is that the coagulation is associated very closely with the volume of hydrate formed. Further more it may be mentioned in passing that the volume of hydrate formed depends upon the specific gravity of the metal. Thus it has been found that one part by weight of aluminum forms about three times as much hydrate as does one part by weight of iron, while the specific gravity of iron is about 2.8 as great as that of aluminum. These general compari sons will be of much assistance in understand ing the different amounts of work done in this line by equal weights of the two metals. A record will be found at the foot of each table showing the amount of metal contained in the commercial sulphate used. With regard to the electrolytic formation of hydrates of these two metals, what has been said above about volumes of hydrate and specific gravity of metals also holds good in this case. Fn sections Nos. 4 and 5 it was pointed out that much of the metal left the plate in a form non-available for coagulation, and was therefore wasted. Practically it is necessary under these conditions to learn the amount of electric current necessary to decompose and convert enough available metal in form of hy drate to equal the efficiency of a known amount of metal in the form of sulphate. With this information in hand, and knowing the total amount of metal decomposed and removed from the plate, it is possible to esti mate how much of the metal (and also electric power) served directly in producing coagula tion. The evidence obtained upon the relative efficiencies of these coagulants was obtained in two different ways, as follows: 1. As coagulants in aiding subsidence in one-gallon bottles. 2. As coagulants in connection with the Jewell filter and devices operated therewith. In the first method the percentages of re moval of suspended matter were obtained after 24 hours subsidence. The electrolyti- cally formed hydrates were obtained regularly from bright metal electrodes, and a current was applied for such a period that the theo retical rate of decomposition would give an amount of metal equal to that in the cor responding bottle in the set with the sul phates. In all cases a series of tests was made with each of the coagulants under considera tion, and for a direct comparison of effi ciencies there were selected, so far as possible, those results which came midway between the results of plain subsidence and complete clari fication. In this manner the fairest compari sons were made, and it is these results which are given in the tables beyond. Concerning the second portion of the evi dence in connection with the Jewell filter, it is to be stated that, of the 185 runs described and recorded in the first half of this chapter, those runs are selected which enable a direct comparison to be made of the several coagu lants in the purification of similar waters under the same general conditions. It is also to be borne in mind that these conditions of the second portion of the evidence are those of practice, and as the electrodes were more or less rusty the efficiency of the electric cur rent was less than on the first set of data, where the tests were made with bright metal. 414 WATER PURIFICATION AT LOUISVILLE. Supplementary to these comparable data were a number of runs in which the coagula tion was insufficient for satisfactory purifica tion. Such runs cannot be included in these tables of comparable results, yet they were of much value in showing and confirming the range in relative efficiencies of the several co agulants. TABLE SHOWING THE RELATIVE EFFICIENCY OF SULPHATE OF ALUMINA* AND PERSUL PHATE OF IRON.f IN CONNECTION WITH 24 HOURS SUBSIDENCE. Percentage Removal. Suspended Solids in Grains of Each River Water. Chemical per Parts per Million. Gallon. Sulphate of Persulphate of Alumina. Iron. 360 2.25 97 94 562 2.50 93 97 I 560 4.00 96 97 216 0.75 80 So 553 1-75 89 90 416 1-75 88 90 Averages 611 2.17 90.5 9>-3 * Containing 9.87 per cent, of aluminum, f Containing 24. 73 per cent, of iron. The above data show that under these con ditions sulphate of alumina and persulphate of iron, containing 9.87 and 24.73 P er cent, of metal respectively, possess substantially equal efficiency, or that the advantage lies very slightly with the persulphate of iron. TABLE SHOWING THE RELATIVE EFFICIENCY OF PERSULPHATE OF IRON* AND ELECTRIC CURRENT WITH IRON ELECTRODES, IN CON NECTION WITH 2J HOURS SUBSIDENCE. Suspended 1 crsulpha te of Iron. Electric Current. Solids in River Water. Million. Gallon. Removal. Hour per Gallon. Removal 562 2 50 93 0.054 95 I 560 2 00 91 0.050 96 243 I 25 85 0.031 87 216 I OO 90 0.025 85 553 i 75 89 0.045 88 272 2 00 88 0.031 90 364 2 OO 90 0.037 91 Aver. 554 1.79 90 0.039 90 * Containing 24.75 per cent, of iron. These data show that 0.039 ampere-hour of electric current on bright iron electrodes Suspended Solids in River Sulphate of Alumina. Electric Current. Water. C Ampere- j Parts per Million. Gallon" RemSv af Hour per Gallon. Percentage 562 2.50 95 0.023 94 I 560 2.00 9i 0.023 92 243 1-25 85 0.013 88 553 1-75 89 0.015 90 416 1-75 88 0.018 90 322 1-75 88 o.oiS 90 987 1.50 94 0.015 94 184 0.88 89 0.015 88 [Aver. 603 1.67 go 0.0175 91 per gallon was equal under these conditions to 1. 79 grains per gallon of persulphate of iron, containing 24.73 P er cent, of metallic iron. From this it follows that the amount of iron decomposed and converted into the form of available ferric hydrate per ampere- hour was 11.3 grains, which is 70 per cent, of the theoretical rate. TABLE SHOWING THE RELATIVE EFFICIENCY OF SULPHATE OF ALUMINA * AND ELECTRIC- CURRENT WITH ALUMINUM ELECTRODES, IN CONNECTION WITH 24 HOURS SUBSIDENCE. *Containing 9.87 per cent, of aluminum. From the above data it is seen that 0.0175 ampere-hour of electric current on bright aluminum electrodes per gallon was equal under these conditions to 1.67 grains per gal lon of sulphate of aluminum containing 9.87 per cent, of metallic aluminum. This shows that there were decomposed and converted into the form of available aluminum hydrate 9.42 grains per ampere-hour. This is 181 per cent, of the theoretical rate. Comparing the efficiency of an electric cur rent in the formation of coagulants by the de composition of bright metallic iron and bright metallic aluminum, and remembering that in the last two tables 1.67 grains per gallon of sulphate of aluminum is equal to 1.65 grains of persulphate of iron, it is noted that 0.0175 ampere-hour on aluminum electrodes is equal to 0.036 ampere-hour on iron electrodes. According to these data the electric current is twice as efficient when applied to aluminum electrodes as when applied to iron electrodes. From the theoretical rate of decomposition it would be expected that the relative effi ciency of the current on aluminum and iron electrodes would be 0.9 to i. SUMMARY AND DISCUSSION OF DATA OF 1807. TAHLE SHOWING THE RELATIVE EKKICIENCV Electric Current with Iron Electrodes. OK SULPHATE OK AH N IN A * AND PER- SULPHATE OF IRON .f WHEN USED IN CON Suspended Number of Ampere- Solids in River Filleted H-incnal NECTION WITH THE JEWELL KILTER AND .MiintK.r Hour per Water. . Water. "j^ Ku " Gallon. Parts per Cubic Feet Efficiency. ASSOCIATED DEVICES. Million. Suspended 89 0.040 330 15551 98.4 Number of Grains per Gallon. Water. Efficiency. 122 0.097 ! 535 7494 98.4 Million 132 0.078 218 , 10598 97.5 147 0.089 488 4392 98.3 148 0.099 4 9 5082 97-8 Sulphate of Alui nina. 178 0.040 128 7274 98.4 i 1.32 336 5004 95-9 iSl 0.040 160 7426 97.5 8 1-33 586 8540 99-2 19 3-98 452 II 986 95-3 Averages 0.070 374 8260 98.2 347 97.6 28 3.01 347 4356 99.1 * Containing 8.46 per cent, of aluminum. 36 1.05 231 II 296 99-2 5 1.17 136 16 282 99 .4 153 3-33 352 7"79 98.7 For the sake of a more extended compari Averages 2.29 34S 8898 - 98.0 son, sulphate of alumina is presented here in - Persulphate of Iron. stead of persulphate of iron. From the fore 4 1.23 351 5S59 97-9 going table the corresponding amou it of the 10 3-99 586 1 2 446 99 . 3 latter chemical may be substituted if desired. 25 2.83 322 n 304 99.6 These data indicate that the relation in ques 31 35 3-37 1.81 407 *93 8013 14 556 99.6 99- T tion was somewhat variable but on an average 51 1.23 3i 13507 99.8 0.070 ampere-hour of electric current per gal 155 2 . (>(> 379 2617 99.1 lon upon rustv iron electrodes as found in Averages 2.31 361 10 560 99-2 practice (with a potential difference between * Containing 8.46 per cent. of aluminum. plates of 3 to 5 volts) was equal to 2.25 grains f Containing 24.43 per c :nt. of iron. per gallon of sulphate of alumina containing The above comparisons show that persul 8.46 per cent, of aluminum. This corresponds phate of iron containing 24.43 per cent, of iron to an electrolytic conversion of iron into the is slightly hut distinctly superior under these form of hydrate of 9.33 grains per ampere- conditions of sulphate of alumina contain hour, or 58 per cent, of the theoretical rate. ing 8-4C) per cent of aluminum. This lot of The difference between this and the 70 per sulphate of alumina was not so rich in alu cent, found with bright iron electrodes was minum as that used in the subsidence experi due of course o the increased >assivitv ments, and it is fair to assume that the relative caused bv rusting. efficiencies stated under those conditions At this point it may be noted that in the would hold true in connection with filtration. later comparisons the relation was found to TABLE SHOWING THE RELATIVE EKKICIENCY be in the ratio of about o. 10 ampere-hour, OK SULPHATE OK ALT M INA * AND ELECTRIC equivalent to 2.5 grains of sulphate of alu CURRENT WITH IKON ELECTRODES, WHEN mina, containing 9.87 per cent, of aluminum USED IN CONNECTION WITH THE JEWELL or to the same amount of persulphate of iron KILTER AND ASSOCIATED DEVICES. containing 24.73 P er cent, of iron. From careful inspection at the time of the tests this Suspended relation was considered to be best and was Number of Grains per Ga . n. Solids in River Water. Filtered Water. Bacterial Efficiency. so reported to you. In view of the fact that Million. the lowest rate is the safest one upon which 08 I Sf) 290 98.0 to base computations, we conclude that the 622 last comparison is the safest ratio under prac 125 3.82 454 10381 99.6 tical conditions. On this basis the amount of 131 149 2.70 2.55 4"4 8034 ?6l8 98.9 98.0 iron decomposed and converted into avail 153 1 80 3-3" 0.96 352 140 7679 10322 98.7 97-9 able hydrate would be 6.17 grains per am I8 3 0.94 .73 5457 96.6 pere-hour, or 38 per cent, of the theoretical TAverages 2.25 349 7920 98.4 rate of decomposition. WATER PURIFICATION AT LOUISVILLE. TABLE SHOWING THE RELATIVE EFFICIENCY OF SULPHATE OF ALUMINA* AND ELECTRIC CURRENT WITH ALUMINUM ELECTRODES, WHEN USED IN CONNECTION WITH THE JEWELL FILTER AND ASSOCIATED DEVICES. Sulphate of Alumina. Snspended Number of Run. per Gallon. Solids in River Water. Parts per Filtered Water. Cubic Feet. Bacterial Efficiency. Million. J9 3.93 452 II 986 95-5 36 1.05 231 II 296 99.2 50 1.17 136 16282 99-4 59 2.47 301 6 899 99-7 131 2.70 288 8034 93-9 Averages 2.28 282 10 899 98.5 Electric Current with Aluminum Electrodes. Am ere Suspended Solids in Filtered River Water. Water. Run Gallon. Parts per Cubic Feet. Million. 20 0.056 452 5 74i 95-9 41 0.020 242 5569 99.0 49 o.on 133 9784 98.5 58 0.028 453 3949 99-5 133 0.037 59 6556 95-2 Averages 0.031 288 5920 97.6 * Containing 8.46 per cent, of aluminum. These data, Which were obtained on the whole under favorable conditions, and when the plates were only slightly covered with oxide coating, comparatively speaking, indi cate that 0.031 ampere-hour of electric cur rent on aluminum electrodes per gallon was nearly as efficient as 2.28 grains of sulphate of alumina containing 8.46 per cent, of alu minum. One of the disadvantages of the electric treatment which is indicated by the above data is the diminution in the length of runs between washings and consequently in the quantity of water filtered per run. This was due largely to accumulations of gas, which retarded subsidence under the given conditions and closed some of the interstices of the sand layer. Later experience with the electric current on aluminum electrodes demonstrated con clusively that the above comparison was not representative of what would occur in prac tice. After the plates of the composition used here (a very pure commercial grade) were continued in service, coatings of oxide made the formation of hydrate very irregular and much less than the rate indicated above. Furthermore, as was repeatedly seen in special tests, continuous service caused a con siderable portion of the aluminum hydrate to be non-available, because it became lodged between the bright metal and the surface coating. Comparing these results with those ob tained with bright aluminum electrodes it is seen that t he efficiency dropped to about 75 per cent, of that recorded in the subsidence experiments. In view of the fact that under the conditions of practice the percentage effi ciency steadily decreased, the regular tests with aluminum electrodes were discontinued during the latter portion of the investigations; and the conclusion was drawn that unless some practicable method of removing surface coatings could be found, the electrolytic method of forming aluminum hydrate was unsafe and impracticable, independent of its cost, on account of the very low and irregu lar formation of the hydrate. No practicable means of removing the surface coatings on a large plant could be devised. These comparative tests show that one grain per gallon of sulphate of alumina, con taining 9.87 per cent, of aluminum, is equaled in efficiency as a subsiding and filtration co agulant by one grain per gallon of persul phate of iron, containing 24.73 P er c ent. of iron, and by 0.040 ampere-hour per gallon of electric current on iron electrodes. With re gard to the action of an electric current on aluminum electrodes, the process as tested under the most favorable conditions which could be made applicable to a large plant, was found to be unsafe and impracticable. SECTION No. 7. ECONOMICAL APPLICATION OF COAGULANTS, IN TERMS OF SULPHATE OF ALUMINA, TO AID IN THE REMOVAL OF SUSPENDED MATTER BY SEDIMENTATION. At the outset of the consideration of this portion of a system of purification applicable to the Ohio River water the following facts are to be recalled: In the application of coagulants to this SUMMARY AND DISCUSSION OF DATA OF 1897. water, as it is taken from the river, there is a great waste of chemicals, with an undesirable consequential effect upon the quality of the water from an industrial standpoint; a large percentage of the normal filter plant would have to be duplicated and held in reserve for use at times of muddy water; the suspended matter in the river water varies so rapidly in amount and character that in the absence of adequate subsidence it would be difficult to manage a plant economically and at the same time efficiently; and, to correct these difficul ties, experience shows that it is essential for economical and other reasons to remove as much suspended matter as practicable from the river water by plain sedimentation before the application of coagulants. I Yom what has been said concerning plain sedimentation in section No. T of this discus sion, it will be understood that during the greater portion of the year there would be re quired a fairly low and approximately uni form application of coagulants to the water after the removal of the bulk of the suspended matter by plain subsidence and before filtra tion. Yet at times (hiring the spring and summer the amount of clay is so great and the size of the particles so small, that the ap plication of coagulants can be divided to ad vantage from an economical point of view, and a portion of the coagulants employed solely as an aid to subsidence. Many data obtained in the course of these investigations point in this direction, but the fact is brought out most clearly by a com parison of the results of runs Nos. 167 and 1 68. which may be briefly summarized as fol lows : 167 Grains pe r Gallon ol Sulphate !ji . C ; = 8 Jewell Settling Chamber. Total. a v " o o 1.49 2.82 0.83 2.82 2.32 129 141 6556 15296 99-3 99-5 In comparing these two runs it will be noted that the quality of the river water was approximately the same, although the amount of suspended matter was slightly higher in the case of No. 168. With the latter run the total amount of the divided application of co agulants was 0.5 grain per gallon less than in the single application in No. 167. The in creased provision for subsidence aided by co agulation, about 3 hours as compared with 0.5 hour, caused a further removal of sus pended matter as the water reached the sand layer, as shown by the fact that the suspended matter in the water at the top of the filter in runs Xos. 167 and i(>X was found to be 79 and 15 parts per million, respectively. In consequence of this clearer but well co agulated water at the sand layer on run No. 168, the quantity of water filtered between washings was considerably more than double what it was on run No. 167, and the percent age of wash-water on Nos. 167 and 168 was 8.1 and 3.5, respectively. In brief, with the divided application of chemicals to give greater facilities for subsidence, the amount of coagulants was reduced 0.05 grain per gal lon; the quality of the effluent was fully main tained; and the capacity of the filter was materially increased. With i .06 grains per gallon of sulphate of alumina applied at basin No. I and i.io to 1.13 grains at the Jewell settling chamber, there was a slight loss in economy with the divided application, in purifying a water con taining rather more suspended matter than in the case of runs Nos. 167 and 168, as shown by runs Nos. 163 and 164. The use of 0.65 grain per gallon of sul phate of alumina at basin No. i, in addition to i. 20 grains at the Jewell settling chamber, was found to be inadequate for the purifica tion of this water, as shown by runs 157 and 158. The practical conclusions to be drawn from this experience are that with preliminary co agulation, followed by subsidence for a period of about 3 hours, the application of coagu lants may be divided to advantage, and a con siderable portion of the suspended matter kept off the filter, when the total amount of required coagulant ranges from 2 to 2.5 grains or more of ordinary sulphate of alu mina per gallon. In the case of a water re quiring more than this amount of coagulating treatment, a proper division of the application would increase the saving of coagulants and 4 i8 WATER PURIFICATION AT LOUISVILLE. would diminish the frequency of washing the filter. To place an estimate upon the minimum limit of suspended matter in the water where a division in the application could be profit ably made is difficult, owing to the wide range in the character of the suspended mat ter; and. further, the period of subsidence fol lowing the preliminary application of coagu lants is an important factor and was too short in these devices. Under the existing condi tions it was found necessary to apply about 1.5 grains per gallon of sulphate of alumina in order to make preliminary coagulation and subsidence effective. With a longer period of subsidence this quantity could possibly be lessened somewhat. But before consider ing further the practical significance of these facts we will show why it is necessary to apply a certain considerable amount of coagulating chemicals in order that coagulation and sub sidence may be efficient. Relative Efficiencies in Sedimentation of Differ ent Amounts of Coagulants. in studying the behavior of coagulants in connection with the sedimentation and filtra tion of the Ohio River water, the most notice able feature is that in the case of sulphate of alumina very little appears to be accom plished until the quantity of applied chemical reaches a certain amount, ranging with differ ent waters from 0.75 to 1.50 grains per gal lon. The indications are that, before any practi cal coagulation is effected, a certain amount of coagulant must be applied in order that the absorptive and perhaps other similar capaci ties of the suspended matter in the water be completely satisfied. In connection with the explanation of this, reference is made to the close of section No. 2 of this discussion. Persulphate of iron and electrolytically formed hydrates of iron and aluminum behave in a similar manner to sulphate of alumina, as shown by the results of experiments recorded in the next set of tables. These results are representative of many data obtained in the laboratory, where sub sidence for 24 hours after the application of the coagulant took place in one-gallon bot tles. Attention is especially called to the last col umn, where the increase in the removal of suspended matter for successive portions of the coagulant is shown. Sulphate of Alumina. (River water contained 424 parts per million of suspended matter.) After Settli ,g 24 Hours Applied Additional per Gallon. Solids. Parts per Million. Percenta.ee Removal. for Successive Portions of 0.25 Grain. N 0.25 0.50 0.75 I.OO 1.25 1.50 47 44 35 3 o 74 76 97 99 too 4 2 5 1 6 T Persulphate of Iron. (River water contained 364 parts per million of suspended matter.) After Settli g 24 Hours. Additional A pplied Chemical. Grains per Gallon. Suspended Solids. Percentage Removal Parts per Million. Removal. 65 0.40 o.So 122 123 66 66 I O 1.20 120 67 I 1. 6() 67 82 15 2.OO 38 90 8 2.40 28 92 2 2. SO 7 98 6 3.20 2 99 I Electric Current with Iron Electrodes. (River water contained 424 parts per million of suspended solids.) After Settlin ; -24 Hours. Additional Treatment. Ampere-Hour Suspended Removal for Successive per Gallon. Parts per Million. Removal. Ampere- Hour. None 185 cfi O.OI2 I So 58 2 o.oiS 167 6 1 5 0.025 161 62 i o . 03 1 156 63 I 0.038 44 66 3 0.044 123 71 5 0.062 26 94 23 0.096 5 98 4 o 124 o IOO 2 SUMMARY AND DISCUSSION OF DATA OF 1897. 419 Electric Current ivith Aluminum Electrodes. (River water contained 364 parts per million of suspended solids.) After Settlin J 24 Hours. Additional Removal for Suspended R e emo n vaf Million. Ampere- Kour. None 67 O.OO2 96 74 7 0.005 7 76 2 0.008 82 77 I 0.010 So 78 I 0.013 44 88 IO 0.015 34 91 3 0.018 20 94 3 O.O2O JO 97 3 0.023 3 99 " The above results show that with these waters it was necessary to apply from i.o to 1.6 grains of sulphate of alumina per gallon before subsidence caused a material removal of suspended matter, in addition to that re moved by plain subsidence. It is true that these waters contained more silt than ought ordinarily to be the case in practice after plain subsidence had taken place. With suspended matter of a clayey nature this minimum amount of coagulant for efficient coagulation would probably be reduced. Just how far the conditions of practice would cause the minimum efficient amount of coagulant to depart from that indicated above was impracticable to ascertain ac curately under the conditions of these inves tigations. There are indications that the minimum limit, where the division in the application of coagulant would be economical, would fall below 2 grains of ordinary sulphate of alu mina and perhaps as low as 1.5 grains per gallon. As it requires about 0.75 grain as a minimum application to secure coagulation prior to filtration, this would leave an equal quantity of coagulant to facilitate subsidence. SECTION No. 8. EFFECT OF THE PERIOD OF COAGULATION OF THE OHIO RIVER WATER BEFORE FILTRATION. Recently there have developed in some lo calities differences in opinion as to the most advantageous period of time to intervene be tween the application of the coagulant and the entrance of the water into the sand layer. The. data bearing on this point are presented in the next two tables, and the conditions un der which they were obtained are outlined as follows: Table No. i. In this table a comparison is made of the principal data in 13 pairs of runs showing the efficiency of the Jewell filter when the coagu lant (sulphate of alumina) was applied at the inlet and outlet of the settling chamber, re spectively. It will be recalled that the outlet of the settling chamber was at the top of the filter. When the coagulant was applied at the inlet to the settling chamber the period of coagulation averaged about 30 minutes, and when applied to the mouth of the pipe lead ing from the settling chamber to the upper compartment (containing about 1.6 feet of water) above the sand layer, the period was about 8 minutes. The character of the Ohio River water was such during these runs that the average quantity of coagulant was ap proximately equal to the estimated annual average amount required for the water under favorable conditions for purification. Table No. 2. A comparison of the principal data of runs Xos. 173, 174, and 175 with the Jewell filter is made in this table. At this time the river water contained much less suspended matter than on the runs recorded in Table No. I. The coagulant on these three runs was applied at the inlet to basin Xo. i, the inlet to the Jewell settling chamber, and to the outlet of the latter cham ber, respectively. This made the average periods for coagulation about 199, 30, and <S minutes, respectively. Comparing the average results of Table No. i, it is seen that with the same character of river water the change in the point of ap plication of the coagulant from the inlet to the outlet of the settling chamber (reducing the period of coagulation from 30 to 8 min utes) caused the quantity of water filtered 420 WATER PURIFICATION AT LOUISVILLE. TABLE No. 1. COMPARISON OF THE EFFICIENCY OF THE JEWELL FILTER WHEN THE COAGULANT WAS APPLIED AT THE INLET TO THE SETTLING CHAMBER, AND WHEN IT WAS APPLIED TO THE WATER AT THE TOP OF THE FILTER. Coagulant Applied at the Inlet to the Settling Chamber. Number of Run. Suspended Solids in River Water. Parts per Million. Applied Sulphate of Alumina. Grains per Gallon. Filtered Water. Cubic Feet. Bacterial Efficiency. 94 149 2.H 17 898 97-9 97 300 1.33 2 IO5 95.6 98 296 1 . 46 5942 98.0 IO2 368 1.69 4 609 96.7 104 438 i. 88 8367 98.8 140 124 1.67 26 640 98.2 143 548 1-52 (> 137 97-9 145 438 1.92 II 584 99.2 I6 4 189 I.8l 4 200 98.9 1 66 189 2.14 4 218 99 " 176 130 I. O2 18435 98.0 1 80 140 0.78 TO 322 97-9 183 173 0.77 5457 96 . 6 Averages 268 1-55 9 686 97.8 Coagulant Applied at the Top of the Filter. 92 2IO 2-47 44f>3 98.1 93 140 1.92 5 5f>7 97-9 95 184 2-74 I 013 98.7 96 3OO i-3i I 048 96.7 99 295 1.42 55iS 95-7 101 350 1.87 999 97-5 103 368 i .92 2856 98.1 141 552 1-59 I 300 97-7 142 548 2.02 4 210 98.9 i65 189 2. IO I 187 97 . 8 175 130 I. 60 10 565 99-2 179 127 1.04 I 007 97.0 184 174 I . IO I 071 97-0 Averages 274 1.78 3139 97-7 TABLE No. 2. COMPARISON OF THE EFFICIENCY OF THE JEWELL FILTER WHEN THE COAGULANT WAS APPLIED AT THE INLET TO HASIN NO. 1, THE INLET TO THE JEWELL SET TLING CHAMBER AND THE OUTLET OF THE LATTER (TOP OF FILTER), RESPECT IVELY. Application of Sulpha teof i Alumina. ^.*s Sj . J g Place. ex c ||| o u ! 3 SO 2 c C-. U rt ^ 7, O X fc B3 173 Inlet, Basin No. I.. I-5I 74 32 227 99.0 174 j Inlet Jewell Set- ( tling Chamber. ... I. 60 73 20438 99.0 T TCI between washings to be diminished to about one-third; and this reduction was accom plished when the quantity of coagulant was increased 15 per cent. The character of the filtered water was unchanged, except perhaps it should be noted in this connection that it was a failure in the quality of the effluent which caused the filter to be washed in all runs re corded in Table No. i, excepting No. 140, on which run the rate failed with a satisfactory effluent. This means that with the shorter period of coagulation the quality of the effluent failed more quickly than in the case of the regular period. In this connection it is to be pointed out that, while the amount of the suspended matter in the river water was the same in each case, the suspended matter in the water as col lected from above the sand layer averaged 90 and 1 60 parts per million with the long and short periods, respectively. With regard to the data in Table No. 2 the river water contained less suspended -matter than in the case of the water dealt with in the first table, and the amount of coagulant was relatively greater, with a consequently higher bacterial efficiency. Under these conditions the ratio of the quantities of filtered water per run, when the coagulant was applied at the inlet and the out let of the settling chamber was 2 to i instead of 3 to i, as in the case cited above. In a gen eral way this diminution in the effect of the period of coagulation was repeatedly ob served as the water became clearer. On runs Nos. 173, 174, and 175 the sus pended matter in the water on the top of the filter was found to be 21, 45, and 86 parts per million, respectively. As would naturally be expected, the longest and most satisfactory run was the one on which the amount of sus pended matter going on the filter was the least. All things considered, this run, No. 173, on which the period of coagulation was about 3.3 hours, was the most satisfactory one ob tained during the entire series of experiments. It brought out very clearly the fact that the period of coagulation could with marked ad vantage be made much longer than custom formerly supposed. It is not to be inferred that in all cases a SUMMARY AND DISCUSSION OF DATA OF 1897. period of 3.3 hours for coagulation is desir able or even admissible from a practical standpoint. The facts as illustrated by runs Nos. 182 and 183 show that this is not true. With about 0.75 grain per gallon of sulphate of alu mina applied to river Water containing 172 parts per million of suspended matter, at the inlet to basin No. i and the Jewell settling chamber, respectively, the application at the first point resulted in the run being a failure, while in the second case fair results were ob tained. As stated in section No. i, the decom position of commercial sulphates in moderate amounts by the alkaline constituents of the Ohio River water is practically instantaneous. Butthe amount and character of the suspended matter in the water exert considerable in fluence upon the optimum period of coagula tion for a given water. This is shown by inspection of the data, when it will be noted that different amounts of coagulants were re quired to yield satisfactory results from waters apparently similar, so far as could be told from amounts of suspended matter. In general terms these last observations hold true, practically speaking, for electrolytically formed coagulants. Conclusions. 1. The Ohio River water as it comes from the river, and also after the coarse matters are removed from it by plain subsidence, requires, for its most economical and efficient treat ment, different periods of coagulation at dif ferent times, according to the character and amount of suspended matter. 2. When the water is very clear the in dications are that but little difference would be noted in periods of from i to 30 minutes; and, further, with clear water it is proba ble that if the period were extended be yond this range to a certain but not well- defined point, a loss in efficiency would result. 3. But when the water contains its usual amount and character of suspended matter the period of coagulation to give the best re sults is a variable one, and reaches several hours in length before a division in the appli cation of coagulants (as discussed in section No. 7) becomes advisable. 4. To fix upon any given period to give uniformly, under the conditions of success ful practice, the optimum degree of coagula tion, or very nearly so, does not seem prac ticable in the light of our present knowledge; and it is recommended that for a large plant the devices for .the application of coagulants be made adjustable, so as to vary the period of coagulation as the character of the water demands. Whether or not it would ever be advisable to fix the point of application, so as to give a constant period of coagulation, can only be told by practical experience. 5. With regard to the best period of coagu lation for subsidence, when a division is made in the application of coagulants, it must be borne in mind that under these conditions two objects are sought, the coagulation of the suspended articles and their removal by sedi mentation. While it is probable that sedi mentation might take place more rapidly if the coagulation was completed before the suspended matters began to subside, than when these two actions took place simultane ously, experience indicates quite clearly that a saving is made by allowing sedimentation to take place during coagulation. The period of coagulation, being plainly the shorter, be comes therefore unimportant, as the optimum period of subsidence with coagulation would be the controlling factor. The conditions of these investigations were not such as to allow the study of the optimum period of subsidence with coagulation, of a water which had already been partially puri fied by plain subsidence. It may be stated, however, that as the water after proper pre liminary treatment by plain subsidence would contain only relatively fine suspended parti cles, the optimum period would probably be considerably longer than would be indicated by the results of subsidence with coagulation of a water which had not been properly set tled. Experiments on the direct treatment of river water such as were recorded in Chapter IV, and in previous sections of this chapter, do not therefore apply to the subject in hand except in a very general way. 4 22 WATER PURIFICATION AT LOUISVILLE. SECTION No. 9. DEGREE OF COAGULATION OF THE WATER BEFORE FILTRATION, AND THE MINIMUM AMOUNT OF COAGULANT REQUIRED FOR THAT PURPOSE. In all cases experience showed that for suc cessful filtration the coagulation of the water as it enters the sand layer must be practically complete. To a trained operator of a filter this condition of the water can be noted quite accurately by inspection. It can be described by the statement that the suspended matters in the water must have a " curdled " or " flakey " appearance, which is such that ulti mately the suspended matters would subside and leave the water in a practically clear con dition. The rate at which such subsidence would take place depends largely upon the size and specific gravity of the matters in suspension. It was found that the proper degree of co agulation of this water as it entered the sand layer was the sine qua non of economical and efficient purification by the American type of filters. From what is said in the paragraph above it is not to be inferred that subsidence alone is adequate under practicable condi tions for satisfactory purification. It is essen tial to have filtration in order to make the purification complete. The amount of sus pended matter in the water affects the degree of coagulation only in that the amount of co agulant must be sufficient to yield enough gelatinous hydrate to envelop practically all of the suspended particles, including the bac teria. Coagulation of the water entering the sand need not if necessity be absolutely complete so far as maximum formation of size of flakes is concerned, because the friction in the sand layer will supplement this action if the volume of hydrate is sufficient. This is especially true of the middle and latter portions of runs, when considerable hydrate is accumulated in and upon the sand layer. Concerning a volume of hydrate in excess of that capable of giving the above conditions, it is to be avoided not only because it is a waste of chemicals (and unnecessarily increases the corroding and in- crusting constituents of the effluent in the case of sulphates), but because it increases the frequency of washing the sand layer and consequently reduces the capacity of the fil ter. In view of the fact that during the greater portion of the year the Ohio River \vater con tains clay particles which are smaller than bacteria, the bacterial efficiency is generally satisfactory if the filtered water is clear and free from turbidity. As a matter of fact, in many cases during the tests recorded in this chapter very fair bacterial efficiency was ob tained when the effluent was so turbid that the run was stopped, the filter washed, and the amount of coagulant increased for the next run. However, there are also times dur ing the winter when the suspended matter is so coarse that comparatively little difficulty is experienced in getting a bright or even brilliant effluent, while a satisfactory removal of bacteria was a less easy matter This consideration of the relative difficulty in removing bacteria and finely divided clay brings us to the question of the minimum amount of coagulant which can properly be applied to this water. Experience indicates that under ordinary circumstances this river water would rarely if ever reach a condition where less than 0.75 grain per gallon of sul phate of alumina, containing about 9.87 per cent, of aluminum, could be used with safety in this method of purification. SECTION No. 10. ON THE CONDITIONS OF SUCCESSFUL FILTRATION. In addition to a confirmation of the evi dence in Chapter IX, our knowledge upon the conditions of successful filtration was ad vanced in several particulars. But as it was considered feasible to operate only one filter during the tests described in this chapter, comparative data are scant}- or lacking along- several lines, notably those related to the character of the sand layer, such as thickness and size of grain. The principal information of practical value obtained in this connection during 1897 is as follows: Amount of Suspended Matter in the Water Reaching tJie Sand Layer, and the . Coagulation of the Same. Experience during the last portion of the tests (April to July, 1897) demonstrated con- SUMMARY AND DISCUSSION OF DATA OF ISO", 423 clusively that for the uniform, efficient, and economical filtration of the Ohio River water it is imperative to remove the suspended mat ter so far as practicable from the water before it reaches the sand layer. The data confirm in a decisive manner the conclusion drawn in 1896, that filtration alone is inadequate for the successful purification of the silt and clay- bearing Ohio River water. Filtration is the last step, and a very important and essential one, in a method in which sedimentation, plain and with coagulation, precedes filtra tion. By plain subsidence, supplemented at times by subsidence with coagulation, there arc gained a number of distinct advantages, as were pointed out before, as follows: i. A considerable wastage of coagulants would be prevented an amount much ex ceeding the cost of subsidence. j. The cost of construction and main tenance of a large reserve portion of a plant, to aid in handling muddy water, would be obviated. 3. The efficiency of the plant would be more uniform and satisfactory with regard to the quality of the effluent. 4. The difficulties and cost of operating a plant to purify with uniform satisfaction the very variable Ohio River water would be re duced largely. In this connection reference is especially made to the adjustment of the optimum amount of coagulant and to the cleansing of the sand layer. At this point the very important question arises, What amount of suspended matter can be properly handled by filters of the American type ? Much thought has been directed to an ex pression of this amount by weight in parts per million. \Ve have not succeeded, how ever, in fixing the limit in this manner, ow ing to the wide discrepancies obtained in handling equal weights of suspended matter of different character. Thus at times 100 parts per million of fine clay were more diffi cult to remove than 500 parts of silt. The best way in which we can express this amount is by the statement that by plain sub sidence, aided by coagulation when necessary, the suspended matter in this river water should be reduced to a point where, by filtra tion and the coagulation just preceding it, the remaining suspended matter can be re moved by the final application of coagulants not exceeding 1.5 to 2 grains per gallon of ordinan- sulphate of alumina, or its equiva lent. In section Xo. 9 it was stated that for suc cessful filtration the final application of co agulants must be such that the coagulation of the suspended particles is practically com plete. When the coagulation is complete, it follows, as shown by experience, that the period of coagulation may be extended so as to cause a still further removal of suspended matter by subsidence, and consequently an additional relief to the sand layer. This is possible with an application of chemical or electrolytic treatment sufficient to produce complete coagulation, because there will then remain, as the water reaches the sand layer, sufficient hydrate to accomplish filtration sat isfactorily. Failure to provide proper coagulation is in admissible, as this is the sine qua non of suc cessful filtration bv this method. The Jewell filter was operated normally in 1897. as was the case in 1895-96, at about roc million gallons per acre daily, equivalent to 1.58 gallons per square foot per minute. Tn the last operations (1897) the only point tested in this connection was the possibility of lowering the rate of filtration to advantage, with regard to a reduction in the amount of coagulant and greater uniformity in effi ciency, especially just after washing the sand layer. A comparison of the efficiency of the filter at normal and half-normal rates, respect ively, is shown by the following representa tive averages of leading data on (1) runs Nos. 67, 72, 81, and 8_>, and (II) Nos. 68, 70, 75, and 83: i n Rate of filtration in million gallons per acre daily 90-95 45-50 Parts per million of suspended solids in river water 181 187 Applied sulphate of alumina in grains per gallon 1.25 i-2i Cubic feet of filtered water 12302 i -I 7-19 Bacterial efficiency 97 - 2 9 6 - 6 424 WATER PURIFICATION AT LOUISVILLE. Comparative Summary. These data, with many others which are not directly comparable, show clearly that, when the amount of coagulant is such as to give at the normal rate only a moderately satis factory bacterial efficiency, a reduction of the rate to one-half of the normal does not in crease the bacterial efficiency, although it in creases slightly the quantity of water between washings. It may be safely concluded that a material reduction in a rate of filtration of ioo million gallons per acre daily would not diminish the required amount of coagulant, and there would be no substantial advantages to offset the lessened capacity of the filter. From what has been said concerning the proper degree of coagulation of the water this fact seems to be almost obvious. In passing it may be stated that, when by any chance the coagulation was inadequate, it was repeatedly noticed that with low rates the bacterial efficiency and appearance of the effluent departed less from the normal than in the case of the regular rate of filtration. The evidence presented in Chapter IX in dicated that under suitable conditions the rate of satisfactory filtration could be materially increased above ioo million gallons per acre daily. Our investigations in 1897 strength ened that view, although with only one filter, and electrical appliances of too small capacity, it was not considered advisable to make a study of high rates, especially with the un- subsided river water. In the judgment of the writer it would be advisable to construct a plant on the basis of ioo million gallons per acre daily, equiva lent to 435.6 square feet per million gallons per 24 hours, with the knowledge that in all probability the rate would be safely increased to a considerable degree in meeting the de mand for a greater consumption of filtered water. Available Head or Pressure for Filtration. The total available head on the Jewell filter, as normally operated by the Water Company, was 10 feet, of which 5.5 feet was a positive head as measured from the surface of the water to the bottom of the sand layer. The remaining 4.5 feet was a negative head or suction, and was obtained by means of a siphon. So far as our observation went, no necessity for modifying this head was noted. The claim has been advanced that a nega tive head possesses a specific advantage for economical and efficient filtration. All the conditions being equal, we have seen nothing to lead us to suppose that there would be any difference in practice due to the head being either positive or negative, beyond the fact that with a negative head there is ordinarily a smaller amount of coagulated water drained into the sewer and wasted just before wash ing the sand layer, or similar operations. Cleaning the Sand Layer to Relieve Clogging. There are three methods which can be em ployed to advantage in the removal of accu mulated matters from the sand in order to re lieve clogging and to keep the sand in a clean and efficient condition, as follows: 1. By washing the sand with a reverse stream of filtered water under pressure. 2. By agitation of the surface accumula tions and their removal into the sewer by flushing with water standing upon the sand. 3. By the application of caustic soda, either with or without steam. The following information was obtained upon the subject in 1897: Washing. Experience confirmed the early conclusion that whenever washing of the sand layer is required it should be accom plished thoroughly, so that the water passing from the top of the sand to the sewer should be comparatively clear. In the Jewell filter this was ordinarily accomplished with the aid of mechanical agitation, by about 600 cubic feet of wash-water supplied at the bottom of the sand layer at a pressure of about 7.5 pounds. The rate of delivery of wash-water ranged from 22 to 134, and averaged 71.0, cubic feet per minute. Assuming that the voids in the normal sand layer were 35 per cent., and that the thickness of the layer when floated was in- SUMMARY AND DISCUSSION OF DATA OF 1897. 425 creased from 30 to 33 inches, this would mean an average vertical velocity between the sand grains of 1.37 linear feet per minute. There seems to be no room for doubt but i hat .he use of mechanical agitation in the process of washing was a distinct advantage. In this case the teeth of the agitator extended to within three inches of the bottom of the sand layer. The indications are that it would be better to have them reach as nearly to the strainer system as safety would allow. With regard to the effect of washing the sand layer, upon the quality of the effluent, it will be recalled that in 1896 it was concluded, from the results obtained from the Warren and Jewell niters, that with complete coagula tion of the applied water and thorough wash ing of the sand layer there was practically no diminution in bacterial efficiency following a washing. In a strict sense this conclusion is correct, but the evidence of 1897 causes a cer tain modification of these views. That is to say. when the coagulation was sufficient to give a satisfactory bacterial efficiency during the latter and major portion of a run. it was repeatedly found that the coagulation might not be complete enough to yield a normal effluent, bacterially, for some little time after the sand laver was washed. This is shown by the following average numbers of bacteria obtained during the first portion of runs Nos. 84. S<J. 91. 93, I O2, 1O6, 122, 126, 130, 131, 133. 142, 147. 148, and 1 68. The bacterial efficiency at the times of col lection of the samples and the average num ber of bacteria in the river water and in the effluent, together with the average bacterial efficiency, are given. In preparing this table all normal runs on which the four or more samples were collected were averaged. Quantity of Water Avcraee Filtered between the .Lut Wash and the Collection of the Sample. Cubic Feet. 15" 273 250 225 500 < 206 moo 178 Average, river water, entire run 9 970 " effluent, entire run. ... 186 98.1 At the regular rate of filtration the above quantities of water correspond to 6.4, 10.6, 21.3, and 42.5 minutes of service after filtra tion was resumed following a wash. The difference between these figures and IXiiinhci Bacteria Cubi. 97-3 97-7 97-9 98.2 corresponding ones obtained in 1896 is ex plained by the fact that in 1897 the coagula tion was relatively less, as every effort was made to keep the amount of coagulant as low as was consistent with good purification. The entire evidence taken as a whole shows that when coagulation is absolutely complete there is no appreciable diminution in the quality of the effluent just after a thorough washing of the sand layer. Nevertheless, when coagulation is very slightly incomplete, but sufficient to give sat isfactory average purification, the data show that just after washing there is a slight dimi nution in efficiency. This slight diminution in efficiency, at this time, could be corrected by an extra application of coagulant for a short period: by reducing the rate of filtra tion, or by discharging the effluent at a nor mal rate (or higher) into the sewer so long as it appeared unsatisfactory. This question is one which will have to be settled by experi ence obtained in practice. Surface Agitation. Surface agitation in a manner similar to that employed in 1896 was used in 39 per cent, of all runs on which the wash and waste water was less than 10 per cent, of the filtered water, and on 49 per cent, of the runs on which the wash and waste water did not exceed 10 per cent, of the filtered water, and on which the average bacterial efficiency was97 per cent, or over. When the Jewell filter was operated by the McDougall Company it was their custom to employ surface agitation, and after stirring up the surface accumula tions to flush them off into the sewer so far as practical. This modification is one of merit, although this filter was not constructed in a manner to allow of its performance in an entirely satis factory way. The reason was that the top of the inner tank was about 10.5 inches above the top of the sand, and accordingly this depth of very muddy water remained on the sand after the operation was completed. The indications are that this is the cheapest manner of removing the bulk of solid matters from the surface of the sand layer, and it seems advisable to modify the construction of filters so that this procedure can be employed to the greatest advantage. This would call for means of removing all 426 WATER PURIFICATION AT LOUISVILLE. of the muddy water from the sand during or after the completion of the agitation, and probably some other changes, especially in the character of the sand layer. It must he understood that surface agita tion could not completely do away with wash ing, and while the indications of its dispensing in part wit h washing are promising, how far it could be carried into successful practice cannot be foretold. .-Implication of Caustic Soda. The applica tion of caustic soda to this filter on July 3 demonstrated conclusively that it had a marked effect in cleansing the sand grains of organic matter and other materials attached to them. As niters continue in service the use from time to time of caustic soda would doubtless be an advantage in keeping the sand layer in a satisfactory condition. Character of the Sand Layer. The sand layer of the Jewell filter was 30 inches in thickness and the sand grains had an effective size of 0.43 millimeter. It is the judgment of the writer that the frictional re sistance of the sand could be increased to ad vantage, especially in an effort to reduce the cost of cleansing, by allowing the use of the partial but more frequent cleansings by sur face agitation in place of thorough washing of the whole sand layer. This could be ac complished by increasing the thickness of the layer or by using a sand with a finer grain, or both. The indications are that this could be done best bv maintaining the thickness of the sand layer at 30 inches, and using a finer sand. It is recommended that a sand be employed having an effective size of about 0.35 milli meter. As the resistance of the sand to the How of water varies inversely as the square of the effective size, this would increase the friction in the ratio of 2 to 3. SECTION No. n. QUALITY OF THE EFFLUENT AFTER PROPER SEDIMENTATION, COAGULATION, AND FILTRATION INDEPENDENT OF THE NATURE OF THE COAGULANT. Under suitable conditions for economical and efficient treatment, as noted above, the quality of the Ohio River water after purifi cation is presented in the following para graphs. It will be noted that the statements in this section are independent of the nature of the coagulants. Following this the effect of the several available coagulants is dis cussed. Appearance. The appearance of the Ohio River water after purification under the above conditions was very satisfactory, as it was practically free from turbidity and color. Taste and Odor. As a rule the taste of the effluent was some what different from that of the river water, in that the earthy taste due to suspended earthy matters in the unpurified water was re moved. The odor of the effluent was the same as before purification, and was never found to be sufficient to be in any way objectionable. Organic Matter. The amount- of organic matter remaining in the effluent was found to range, when ex pressed as nitrogen in the form of albuminoid ammonia, from .030 to .110 part per million, and averaged about .070 part, expressed as oxygen consumed the range was from 0.5 to 1.6 parts, and the average was i.o part. In such small amounts the organic matter in the effluent was entirely unobjectionable. Mineral Matter. Upon purification the changes in the min eral contents of the Ohio River water is char acterized chiefly by the complete removal of suspended mud, silt, and clay. This is un questionably an advantage, although from a sanitary standpoint the evidence indicates that such an action does not specifically im prove the healthfulness of the water, except perhaps in the case of some abnormal indi viduals. From an industrial point of view the re moval of suspended mineral matter is dis cussed below. SUMMARY AND DISCUSSION OF DATA OF 1807. 427 With regard to the dissolved mineral con stituents of the effluent, they do not differ materially from those of the river, except as influenced by the nature of the coagulants as shown in the next section. Gases. The principal gases in the river water, car bon dioxide (uniting with water to form free carbonic acid) and atmospheric oxygen, were practically unaffected by purification ex cept through the nature of the coagulant. Alg(c and other Grosser Micro-oganisms. It was found that the effluent was practi cally free from all diatoms, alga?, and other micro-organisms which may be called large when compared with bacteria. Bacteria. Under favorable conditions of coagulation and filtration the bacteria in the effluent were reduced to a point which was satisfactory in the light of modern sanitary science. Undecomposed Coagulants. This topic is entered into in subsequent sections, and the conditions of proper coagu lation as stated above in the title lead to the inference that the presence of Undecomposed coagulants in the effluent is not a factor for consideration. For the sake of explicitness, however, it may be mentioned here that, with suitable conditions for the employment of subsidence to its economical limit, with com mercial sulphates there would be no occasion for the chemicals to be applied in amounts exceeding that capable of complete decom position by the river water. Concerning the use of electrolytically formed iron hydrate, it is not probable that dissolved iron would ever appear in the effluent; although it is possible that large amounts of clay in the river water in midsummer, when the amount of dissolved atmospheric oxygen in the water is least, might press closely and perhaps overreach the safe limit in the amount of iron which could be completely oxidized and rendered insoluble. If such an occasion should arise it would be of very short duration and could be obviated by the use of small amounts of commercial sulphates to supplement the elec trolytic process. In 1897 there was no instance where any difficulty was experienced in keeping the effluent free from Undecomposed coagulants. Storage of the Effluent. The conditions of successful practice de mand that between the purification plant and the distributing pipes there should be provid ed a reservoir in which sufficient filtered water may be stored to compensate for all inequali ties in the rate of consumption at different hours of the day, and also to allow the plant to be stopped when repairs, etc., require it. The question of storage of the effluent is one of much practical importance. While it is true that the effluent as it leaves the filter is free from algae, diatoms, etc., it is a fact that the spores of these micro-organisms are present in the atmosphere; and the fil tered water contains a considerable amount of food (principally in the form of nitrates) for the growth of these organisms, many of which give rise to objectionable tastes and odors. The growth of these organisms re quires the presence of sunlight; and the effect of purification is marked in this respect, be cause in open reservoirs the removal of all suspended matters from the water permits the sun s rays to penetrate the filtered water, while with the river water this is impossible. This was demonstrated conclusively by mi croscopical examinations of the river water with different amounts of suspended matter in it during warm weather (the period of maximum growth of these organisms), as shown in Chapter I. To estimate the period of storage during which the filtered water might be stored in open reservoirs, before the growth of algae would begin in warm weather, is a difficult matter. It would vary widely with the tem perature and the frequency of sunshiny days, the amount and specific character of the par ticles of floating matter coming from the at mosphere, and the amount of dissolved mat ter in the filtered water, adaptable as food for 4 z8 WATER PURIFICATION AT LOUISVILLE. these organisms. Inspection of isolated por tions of the Ohio River where the current was almost nil, lead to the belief that sudh a growth mig ht occur in much less than one week. General information concerning the life history of these organisms indicates that at times the period would be as short as 4 j days, but it is possible that under some cir cumstances it might be no longer than 2 days. It is certain, however, that it would not be safe to expose the filtered water to the rays of the sun for an average period of about 6 days, as would be the case if the uncovered reservoir at Crescent Hill were used under present conditions. Difficulties with alga; growths may be obviated by making the period of exposure to the sun s rays very short, or by using a covered reservoir, or both. The covered res ervoir would be safest, especially as it would preclude trouble from growths of organisms which might become seeded upon the walls of the reservoir. Corrosion by the Effluent of Metal Receptacles. i This subject was discussed in Chapter IX, but in 1897 additional information of value was obtained. In all probability there would be no diffi culty whatever in the action of the effluent upon lead pipes, or iron pipes which were properly coated with a protective paint. With uncoated iron pipes or receptacles, especially those of wrought iron, the effluent would have an increased corroding action. As al ready explained, water normally produces corrosion by the joint action of the carbonic acid and the atmospheric oxygen dissolved in the water. This action produces iron (ferric) hydrate, just as in the Anderson process of se curing coagulation. When first formed, iron hydrate is flocculent and fairly porous, and it will be understood from what has been said concerning the nature of this compound as a coagulant, that it has the power of incorporat ing within itself suspended matters and some dissolved matters. The resulting mass is much less porous than when no suspended matters are embodied in the hydrate. Facts show that this is of practical import ance in the consideration of corrosion by the Ohio River water before and after purifi cation. Partial Protective Influence of Suspended Matter against Corrosion. Comparison of the rc ative corroding action of the Ohio River water before and after the removal of silt and c ay, by filtration through filter-paper, a Pas teur filter or other device in which the chem ical character of the dissolved compound would not be materially changed in quality or amount, showed repeatedly and without exception that the suspended matter acted as a partial protection to the iron. This is illustrated by the following representative experiment, in which 8 pieces of o. 5-inch wrought-iron rods weighing 1 150 grams were placed in 2200 cubic centimeters of river water (chemical sample No. 985) and the effluent after pass ing this water through a Pasteur filter. The experiment was continued for ten days, and in order that the effluent might not be dis similar to practical conditions by an exhaus tion of carbonic acid and oxygen, air (con taining these gases) was constantly passed through each water. At the end of this ex periment it was found that in the case of the river water i.i grams of iron had been lost from the rods by corrosion, while in the efflu ent the corresponding amount was 2. i grams. To prove that this was not due essentially to an action of dissolved chemical compounds (exclusive of course of carbonic acid and oxygen), this experiment was repeated for 14 days, using distilled water with 10 parts of common salt in each bottle, and in one of them some fine clay (kaolin) was suspended. The results show that in the case with sus pended clay there was a loss of iron by cor rosion of 0.90 milligram, while in the clear water the iron lost 1.81 milligrams. Bearing in mind the decisive proof that on a laboratory scale the suspended matter gave a marked although incomplete protection to iron from corrosion by the water, an exam ination was made of the experience of this Company in the corrosion of uncoated pipes. The most notable instance of the slowness with which the Ohio River water corrodes uncoated iron is in the case of the intake at the old pumping station. This wrought-iron pipe. 0.37 inch thick and 50 inches in" diam eter, was put in service in 1860, and remained SUMMARY AND DISCUSSION OF DATA OF 1897. 429 in continuous service without any artificial protective coating until 1894. After 34 years of exposure to the river water it was corroded to a considerable degree, but not enough to warrant its removal, and it was coated with a layer of cement and continued in service. In concluding it may be stated that the rea son that the effluent has a greater corroding action upon iron than has the river water, is because the suspended matters mixing with the hydrate diminish materially the contact of the water with the surface of the bright metal at the point of corrosion. In a measure this effect is similar to that in the case of lead pipes. The Ohio River water dissolves new bright lead rapidly, but with great prompt ness it forms a very thin layer of basic carbon ate of lead which, practically speaking, is absolutely impervious to water and conse quently arrests all further action. Adaptability of Effluent for Boiler Use. Compared with the average boiler waters in this section of the country, and farther West, the Ohio River water in an unpurified condi tion is a fairly satisfactory water for use in steam-boilers. In comparison with the clear and soft waters in the East, however, it has t\vo marked disadvantages. In the first place, during the greater part of the year the amount of suspended matter is so great that it forms large quantities of sludge, which at times cannot be removed by " blowing off," so that it is necessary to enter the boiler and remove it by manual labor. At such times the incrusting constituents are very low in amount and are deposited upon the separate particles of the sludge for the most part, and seem to leave the metal al most free from sulphate of lime, etc. The re moval of these large amounts of suspended matterj would be an advantage, unquestion ably, and the effluent would therefore be su perior to the river water. The second disadvantage of the river water for use in boilers is seen at times of very low water in the river, such as is found during the fall months. At these times the water con tains not only sulphate of lime and other in- crusting constituents in considerable quanti ties (and far in excess of the average amounts), but also some very fine clay. In boilers this clay and the incrusting con stituents unite and form a coating resembling cement, which is very difficult to remove from the surface of the boiler. The removal of the suspended clay from the water before its entrance into boilers would therefore im prove the water for boiler use. With regard to the dissolved chemical com pounds such as incrusting constituents, there is no difference between the river water and the effluent, independent of the nature of the coagulant. Disregarding the influence of the coagulant, which is discussed beyond, the effluent is more suitable for boiler use than is the river water. Uniformity in Quality of Effluent. With proper conditions for sedimentation, coagulation, and filtration, and independent of the nature of the coagulant, there ought not, and need not, be any objectionable variations in the quality of the effluent in consequence of the purification. It was found that the qual ity of the effluent does vary owing to the in herent variations in the river water. From a practical point of view these variations would occur in the dissolved mineral compounds, especially the carbonic acid and the incrust ing constituents. The evidence obtained in 1897 showed that with regard to these con stituents the composition of the river water is more variable than was considered to be the case in 1896. The normal and extreme amounts of carbonic acid and incrusting con stituents in parts per million in the river water are as follows: Carbonic acid 133 Incrusting constituents 51 SECTION No. 12. MANNER IN WHICH THE NATURE OF THE COAGULANT AFFECTED THE QUALITY OF THE EFFLUENT. There are three different coagulants, each of a somewhat different nature, which have been considered as factors in this problem, namely: sulphate of alumina, persulphate of iron, and electrolytically formed iron hydrate. As 43 WATER PURIFICATION AT LOUISVILLE. stated in Chapter III, the passage of unde- composed sulphates into the effluent would not only be inadmissible, but inexcusable. Our experience in 1897 allows of no modifi cation of this view. It may be mentioned in passing, however, that the presence of unde- composed sulphates in the effluent would be exceedingly objectionable in connection with corrosion. With regard to the electrolytic iron method there would be no danger of any iron getting into the effluent during cold weather, but in midsummer if a heavy rise in the river should occur there might not be sufficient oxygen in the water to convert all the iron into insoluble ferric hydrate. Under these circumstances it would be necessary to supplement the safe limit in this treatment with some other coagulant, or to aerate the water. In view of the fact that it is practicable to coagulate this water properly without the passage of undecomposed sulphates or of dis solved hydrates into the effluent, and that the necessity for doing so is imperative, we will consider the nature of the effect of the several coagulants upon the quality of the effluent only when applied in permissible (but ade quate) amounts. In general terms the quality of the effluent is affected in two ways by the nature of the coagulant: first, with regard to the amount of oxygen and carbonic acid; and, second, with reference to the increase in amount of incrusting constituents in consequence of the lime and magnesia passing from the (alkaline) carbonates and bicarbonates into the neutral sulphates. From a sanitary point of view there is no reason to believe that these factors under suit able conditions of practice would be of any practical importance. They would influence, however, the corroding and incrusting power of the effluent. The manner in which the several available coagulants affect the quality of the efflu ent, expressed quantitatively in equivalent amounts of i grain per gallon of sulphate of alumina containing 9.87 per cent, of alumi num, is shown in the table opposite. In order to consider the practical effect of the nature of the coagulant upon the corrod ing and incrusting power of the water, it is Coagulant. Oxygen. Alkalinity. Constituents. Dioxide. Sulphate of None 9.04 9.04 3-97 alumina Persulphate (decrease) u. -8 (increase) 11.78 (increase) 5-16 of iron (decrease) (increase) (increase) Electrolytic iron 0.76 (decrease) None None None COMPARATIVE EFFECT UPON THE QUALITY OF THE EFFLUENT OF ONE GRAIN PER GALLON OF SULPHATE OF ALUMINA,* AND ITS EQUIVALENT OF OTHER COAGULANTS. (Changes in Constituents of Effluent expressed in Parts per Million.) .87 pe necessary to know the amount of the above changes. This depends upon the amounts of coagulants used, and the estimated quantities of coagulants which would be required are presented in the next section. In section No. 14 the degree to which the several co agulants would affect the effluent is shown and its practical significance discussed. SECTION No. 13. AMOUNTS OF THE DIFFERENT AVAILABLE COAGULANTS WHICH WOULD BE RE QUIRED, WITH OPTIMUM CONDITIONS OF SUBSIDENCE AND FILTRATION, TO PURIFY SATISFACTORILY THE OHIO RIVER WATER. Taking into consideration the employment to their economical limits of plain subsidence and an extended but varying period of sub sidence with coagulation, and in some cases a division in the application of coagulants, it is estimated that the annual average amounts of required coagulants for the satisfactory purification of the Ohio River water would be as follows: ESTIMATED REQUIRED AMOUNT OF COAGULANT PER GALLON OF RIVER WATER. Coagulant. Max. Min. Aver. Grains of sulphate of alumina* 4.00 0.75 1.75 Grains of sulphate of iron f 4.00 0.75 1.75 Ampere-hour of electric current on iron electrodes 0.16 0.03 0.07 The average amount of coagulant in the respective equivalent forms would probably range according to the rainfall and other con ditions in amounts from 1.50 to 2.00 grains SUMMARY AND D/SCUSSION OF DATA OF 1897. 43 in the case of the sulphates, and from 0.06 to 0.08 ampere-hour of electric current on iron electrodes. The mean average is given above. In regard to the maximum amount of elec trolytic iron treatment it is to be borne in mind that the amount above given could only be safely applied at times when the amount of oxygen dissolved in the river water was rela tively large. As a rule, the maximum amount of electrolytic treatment which could be safely applied would be about o. 12 ampere-hour per gallon. SECTION No. 14. DEGREE TO WHICH THE SEVERAL COAGU LANTS IN THEIR RESPECTIVE AMOUNTS WOULD AFFECT THE QUALITY OF THE EFFLUENT, WITH ITS PRACTICAL SIG NIFICANCE AND A CONSIDERATION OF THE ADVISABILITY AND COST OF THE REMOVAL OF THE ADDED CONSTITUENTS. Taking the annual average amounts of the required coagulants as estimated in the last section, the amount of changes in the several constituents of the effluent which would ac tually occur may be taken as follows: CHANGES IN CONSTITUENTS OF RIVER WATER. (Parts per Million.) in Oxyeen. in Alkalinity. The first step in the consideration of the practical significance of these data is to note the range of these constituents as they natu rally occur in the Ohio River water, ANNUAL RANGES OF AFFECTED CONSTITU ENTS AS THEY NATURALLY OCCUR IN THE OHIO RIVER WATER. (Parts per Million.) Alkalinity. 4.6 8.6 ! 1080 ! 21.0 I 65.0 Incrusting Constituents Carbon Dioxide.* 2<J.O 70.0 Coagulant. Max. Min. Av. Max. Min. Av. Sulphate of alumina .... None None None 27.0 6 . o 12.5 Persulphate of None 7 8 Electrolytic ! Increase in IncrustinR Increase in Cnrbon Constituents. Dioxide. Coagulant. Max. Min. Av. Max. Min. Av. Sulphate of alumina .... 27.0 6.0 12.5 II. g 2.6 5-5 Persulphate of 7 8 16 2 Electrolytic None None None None None None Atmospheric Oxygen Electrolytic Iron Process. The atmospheric oxygen dissolved in the water is affected only in the electrolytic iron process. It will be seen that while the aver age decrease in oxygen is equal to less than half the average amount in the river water, yet the maximum (calculated) decrease is nearly double the minimum amount in the water. Owing to the influence of temperature the amount of oxygen in the water is least during the summer months. During cold weather the indications are that there would be enough oxygen for the satisfactory use of this process. An exhaustion of the oxygen and the passage of soluble iron into the filtered water is inadmissible. Such a state of affairs would probably never occur except at times when the river water might be heavily charged with clay in midsummer. To guard against this effectively it would be necessary to provide facilities for the use of sulphates to supplement, at times in hot weather, the electrolytic iron process. With regard to alkalinity, incrusting con stituents, and carbonic acid, they remain un affected by the electrolytic iron process. Changes in Alkalinity, Incrusting Constituents, and Carbonic Acid Process with Com mercial Sulphates. The above data show that in this process the atmospheric oxygen is unaffected, while the alkalinity is reduced and the incrusting constituents and carbonic acid increased. In the case of persulphate of iron the changes 43 2 WATER PURIFICATION AT LOUlSVILLh. are 30 per cent, greater than with sulphate of alumina, other conditions being" equal. In view of the fact that the two sulphates are of about equal cost, the sulphate of alumina is therefore the better chemical to employ as a coagulant. As stated and explained in Chap- tor 111, the use of this product in such amounts that it would pass through the filter in an undecomposed form would not only be inadmissible but inexcusable. \Ye shall therefore consider sulphate of alu mina more especially in the balance of this section, but will take up the effect of each sul phate in the following connection: 1. Sanitary character of effluent in this re gard. 2. Use of soap. 3. Incrustations and adaptability for use in steam-boilers. 4. Corrosion of iron receptacles. 5. Corrosion of lead receptacles. Effect upon the Sanitary Character of Effluents due tn Changes Caused by the Use of Sulphates. So far as \ve have been able to learn the sanitary character of the effluent would not be appreciably affected by the reductions in the alkalinity and the corresponding increases iu incrusting constituents and carbonic acid. within the ranges noted above. There is no reason to believe that the carbonic acid has any significance in this respect, and we have only to consider the effect of changing the lime (and some magnesia) from the carbonate to the sulphate. In this connection it is to be noted that during the greater part of the year the amount of sulphate of lime in the chem ically treated effluent would be far less than is naturally present in the river water during the fall months. And at that season of the year the percentage increase of sulphate of lime due to applied chemicals would be very small. Effect upon the Amount of Soap Required by the Filtered Water due to Changes Caused by the Use of Sulphates. The amount of soap required by a water for washing purposes depends upon the total amount of lime and magnesia present in the water. This is indicated by the total hard ness of the water, which is measured by the sum of the alkalinity and incrusting -constitu ents, approximately equal to the temporary and permanent hardness, respectively. As the decrease in alkalinity and increase in in- crusting constituents are proportional, the soap consuming power would be constant under ordinary conditions. After prolonged boiling of the filtered water it would require slightly more soap, be cause upon the expulsion of carbonic acid gas there would be less lime in the form of car bonate to settle out than in the case of the river water. Changes in Adaptability of the Effluent for Use in Steam-boilers due to tlte Employment of Sulphates as Coagulants. The adaptability of a water for boiler use, independent of matters in suspension, is gov erned by the amounts of incrusting constitu ents, and in this respect the two sulphates in question produced relative changes substan tially as follows: PERCENTAGE INCREASE OF INCRUSTING CON STITUENTS OF THE EFFLUENT ABOVE THOSE IN THE RIVER WATER. Ma: Mi With sulphate of alumina. . With persulphate of iron. . 300 400 "7 90 Owing to the fact that the increase in the incrusting constituents is low when the amounts naturally present in the river water are high (fall months) the maximum amount in the filtered water as shown by these data would not be far in excess of that in the river water. COMPARISON OF THE MAXIMUM AND AVER AGE AMOUNTS OF INCRUSTING CONSTITU ENTS OF THE EFFLUENT AND OF THE RIVER WATER. (Parts per Million.) Maximum Avpratrp maximum. rtVLr.i^i. Ohio rievr water 51 18 Effluent with sulphate of alumina 57 30 Effluent with persulphate of iron 59 35 From the above figures it is seen that the average annual amount of incrusting con stituents in the effluents when sulphate of alu mina and persulphate of iron are used are only 60 and 70 per cent., respectively, of the SUMMARY AND DISCUSSION OF DATA OF 1807. 433 maximum amount naturally present in the Ohio River water, and at times when the river water is least suitable for use in boilers the increase in incrusting constituents is only 12 and 16 per cent., respectively. These additions in incrusting constituents cannot be regarded as other than undesirable, hut it is finite possible, if not probable, that the practical effect of the additions when proper subsidence is availed of. would be off set by the freedom of the effluent from sus pended matters. Experience alone can demonstrate this conclusively. There is another way of looking at the ap plicability of the effluents obtained with com mercial sulphates in connection with use in boilers. That is to compare these data with the available results showing the incrusting constituents (permanent hardness) of the water supplies of other cities. So far as they were available these data are as follows: COMPARISON OF INCRUSTING CONSTITUENTS (PERMANENT HARDNESS) OF THE WATER SUPPLIES OF VARIOUS CITIES. (Parts per Million.) Supply. ronstituenis. Unfiltered Ohio River water (average), Louis ville, Ky 18 Filtered Ohio River water with sulphate of alumina, Louisville, Ky 3" Filtered Ohio River water with persulphate of iron, Louisville, Ky 35 Lynn Mass 4 Holyoke, Mass 18 New York. N. Y I? Scranton, Pa 4 Cincinnati, O 20 St. Louis, Mo 48 London, England 5 Liverpool, England 57 Manchester, England 19 Bradford, England 21 Birmingham, England 64 Glasgow, Scotland 9 Paris, France 3 Geneva, Switzerland 53 Vienna, Austria 2O St. Petersburg, Russia 4 With regard to the amounts of incrusting constituents in the water supplies of other cities it is not known how widely they may vary. The data are averages of all available results from reliable sources, but as a rule only one figure was given in a single work. In the case of most of the European results the figures appear to be given as representa tive ones, and it is believed that they suffice for the present purpose. While the above evidence shows that the Ohio River water, after purification in which plain subsidence preceded coagulation with sulphate of alumina or persulphate of iron, would not be an especially soft water as viewed by Eastern standards, yet as com pared with the waters of the Western part of this country it would not be an objectionable one, nor would this factor be of sufficient weight to offset the advantages of filtration. Furthermore, it is possible to remove these incrusting constituents from the water by the application of caustic soda followed by sub sidence or filtration. On the basis of $1.85 per 100 pounds for caustic soda, containing 60 per cent, available sodium oxide, the cost of chemical per million gallons of water to remove i part per million of incrusting con stituents would be 12 cents. In the judgment of the writer this step would not be justifiable so far as the entire supply is concerned, and it is hardly probable that it would be worth while for large manu facturing, establishments to adopt it. Corrosion of Iron Receptacles due to Changes in the Effluent Caused by the Use of Sul phates as Coagulants. It has already been explained in detail that corrosion of uncoated iron is due chiefly to carbonic acid and dissolved oxygen in a water; and in 1897 it was learned that the degree of corrosion by the water would be greater after purification due simply to the re moval of the suspended matter which served in a measure as a protective coating. With adequate facilities for the proper employment of subsidence, the amount of applied sulphate could be reduced much below that employed at times in 1895-96, and the increase in car bonic acid may be considered as follows: PERCENTAGE INCREASE OF CARBONIC ACID (CARBON DIOXIDE) IN THE EFFLUENT ABOVE THAT IN THE RIVER WATER. Maximum. Minimum. Averape. With sulphate of alumina 41 9 With persulphate of iron 55 Expressing these changes in actual quanti ties, the following comparisons are obtained: COMPARISON OF THE MAXIMUM AND AVER AGE AMOUNTS OF CARBON DIOXIDE IN THE EFFLUENT AND IN THE RIVER WATER. (Parts per Million.) Maximum. Averagt. Ohio River water 33 Effluent with sulphate of alumina 145 Effluent with persulphate of iron 149 WATER PURIFICATION AT LOUISVILLE. At this point it is to be stated that the evi dence obtained in 1897 upon the corroding action of the effluent upon iron was very dif- | ferent (independent of the influence of sus pended matter), and much more favorable than was the case with the limited data in 1890. The reasons for this are twofold. 1. In 1890 there were times when the ap- , plied sulphate of alumina for considerable pe riods averaged as high as from 6 to 8 grains per gallon; while in 1897 it was learned that with a proper use of subsidence the maximum limit could be held at about one-half of that stated above. 2. The limited data in 1896 indicated that the amount of free carbonic acid in the Ohio River water ranged from 20 to 30 parts per million, while the more extended series of op erations in 1897 showed that the amount reached as high as J 50 and averaged about 70 parts. Further, the later results showed that it was very seldom that the amount was less than 50 parts, clearly proving that the amounts found when in 1896 were abnor mally low to an extreme degree. The ex planation of this is not entirely known, but it was partly due to the inaccuracies of Patten- kofer s method of determining carbonic acid. From these and other facts it was computed that the average percentage increase in car bonic acid was about 40 per cent, under the conditions and data of 1896. and only about 9 per cent, for 1897. Another very significant condition which obtained from time to time during the latter part of the tests of 1896. and which was absent in 1897, was the presence of undecomposed chemicals in the effluent. This was of great importance in this connec tion, because when corrosion is once started by an effluent containing sulphuric acid, the conditions are much more favorable for a continuance of the action by carbonic acid. As has been stated repeatedly, the presence of undecomposed chemicals in the effluents is inadmissible for many reasons, and with proper subsidence facilities its occurrence would be inexcusable. The carbonic acid liberated by the decom position of carbonate of lime is the same, so far as its corroding nature is concerned, as an equal amount of carbonic acid naturally present in the water. As the results of a large number of experiments, under the con ditions indicated to be most suitable for practical purification, it was found that the small increase in carbonic acid produced only very slightly greater corroding action than possessed by the river water after removing the suspended matter with a Pasteur filter. In fact in a large portion of the tests made in bottles with rods of bright wrought iron the increased corrosion was not appreciable. To remove the small amounts of carbonic acid added to the water by the application of sulphate of alumina or persulphate of iron, and leave the large amounts naturally and normally present in the river water would be impracticable. The cost of chemical per mil lion gallons for removing i part per million of carbonic acid, with lime at $3.75 per ton, would be 2.2 cents. As lime is only sparingly soluble in water it would be necessary to pump daily for a 25-million-gallon plant about 1750 gallons of water in order to pre pare sufficient lime-water to remove t part per million of carbonic acid. In view of the fact that corrosion would af fect onlysuchiron pipes or receptacles as were not properly coated with a protective paint, it would not be justifiable to remove carbonic acid to a point where it would not possess a corroding action. A better way would be to protect the piping system as it is extended as far as possible by protective paints, and to keep the quantities of applied chemicals as low as possible by taking full advantage of subsid ence. With regard to the piping system al ready in service, it is probable that such portions as are not already protected by a coat of paint are protected in a considerable measure by the deposit of suspended matter of the water which has passed through them. Action of tlic Effluent on Lead Pipe. As in the case of unpnrified Ohio River water, the effluent would contain enough car bonic acid and carbonate of lime to form very quickly basic carbonate of lead which is in soluble and makes an impervious protective coatino-. SUMMARY AND DISCUSSION OF DATA OF 1807. 435 SKCTION No. 15. COMPARATIVE COSTS OF EQUIVALENT AMOUNTS OF THE SEVERAL AVAILABLE COAGULANTS, TOGETHER WITH AN ES TIMATE OF THE YEARLY COST OF TREATMENT OF THE OHIO RIVER WATER BY EACH OF THEM. As has been shown in the preceding sec tions of this chapter, the coagulants available for use in the purification of the Ohio River water are hydrate of alumina prepared by the decomposition of sulphate of alumina by the lime in the river water, hydrate of iron pre pared by the similar decomposition of per sulphate of iron, and hydrate of iron prepared by the electrolytical decomposition of metal plates. The relative advantages and disad vantages of each have been presented at con siderable length both in absolute and com parative terms, and it remains to show the exact relative and annual costs of these three coagulants. In regard to the sulphates, this evidence has already been presented in a gen eral way, but, for completeness, they will be taken up again here. Comparative Cost of Equivalent Amounts of the Available Coagulants. In section Xo. 6 of this chapter it was shown that the amounts of treatment with persulphate of iron and with electric current on iron electrodes, equivalent to i grain per gallon of sulphate of alumina (containing 9.87 per cent, of aluminum) were as follows: Per sulphate of iron (containing 24.43 per cent, of iron), i grain per gallon; electric current on iron electrodes. 0.04 ampere-hour per gallon. In regard to the two sulphates the com parison of cost is a. simple one. but in regard to the electrolytically prepared coagulant there are several separate items which must be taken into consideration. In the following comparisons no account is taken of cost of attendance, which would be slightly greater in the case of the electro lytic treatment than in the case of the sul phates. The cost of construction of devices for the application of the sulphates is also not considered as it would be comparatively small for a gravity flow, and, owing to the limitation in the amount of safe electrolytic treatment, these devices would be required in the use of any of the coagulants. Cost of Electrolytic Treatment with Iron Elec trodes. In the cost of preparation of hydrate of iron by the action of an electric current on iron electrodes the following items must be considered: 1. Cost of construction of electrolytic cells. 2. Cost of construction of electrodes. 3. Cost of construction of electric gen erating appliances, together with the neces sary building to cover them. 4. Cost of operation of electric appliances. 5. Cost of metal used in the formation of the hydrate and wasted in the process. In this connection it is considered that the Water Company owns the necessary avail able land on which to construct the buildings and cells. For the several items the following esti mates of cost are used: i. Cost of Construction of Electrolytic Cells. As will be seen the practical size of the nec essary cells would be so great that open chan nels of masonry would be, apparently, most suitable. For ease of handling it is assumed that plates 4 feet wide, 3 feet deep, and 0.5 inch thick would be employed for the elec trodes. It is further assumed that a o.5-inch length of electrolyte (water space between the plates) would be most advantageous. With walls and bottom i foot thick the masonry required on this basis would be 0.090 cubic foot, or 0.00334 cubic yard per square foot of cross-section of electrolyte. At $30.00 per cubic yard the cost for masonry would therefore be $0.100 per square foot of cross-section of electrolyte. At 5 per cent, interest per annum this would represent an expenditure of $0.0000137 per day (or per twenty-five million gallons) for each square foot of electrolyte, or $0.0137 per 1000 square feet of electrolyte. 436 WATER PURIFICATION AT LOUISVILLE. In the following computations the letters C. S. E. will be used to represent the cross- section of electrolyte in thousands of square feet. 2. Cost of Construction of Electrodes. The area of one side of all plates would cor respond practically to the area of cross-sec tion of electrolyte. The weight of metal re quired would therefore be (on the above assumption of size of plates) approximately 20.0 pounds per square foot of cross-section of electrolyte, and would cost at 2 cents per pound (in place) 40 cents. The daily interest on this amount would be $0.0554 x C.S.E. j. Cost of Construction of Electric Generating Appliances, together with the Necessary Building to Cover Them. From preliminary estimates on the re quired size of these appliances it is assumed* that they could be constructed at a cost of $170.00 per indicated H.P., or allowing one- third for a reserve plant, $220.00 per actual average I.H.P. The interest on this would be $0.030 per day, per I.H.P. 4. Cost of Operation of Electric Generating Ap pliances. It is assumed that a combined efficiency of 80 per cent, could be expected, and that the consumption of coal would be 1.33 pounds per I.H.P. per hour. The required amount of current to treat 25 million gallons per 24 hours with 0.04 am pere-hour per gallon would be 41,600 am peres. Using the average resistance of the electro lyte as presented in section No. 3 of this chapter, 7000 ohms per centimeter cube or 9.65 ohms per sqdare foot of cross-section of electrolyte with a o.5-inch length of electro lyte (water space between plates), the re- 17 900 quired amount of power would be C.S.E. The daily cost would be $571.00 C.S.E. 5. Cost of Metal Used in the Formation of the Hydrate and Wasted in the Process. As was shown in section No. 4 of this chapter, the total amount of metal used in this process is 1.05 grams per ampere-hour. To treat 25 million gallons with 0.04 ampere- hour per gallon there would be required therefore 2305 pounds of metal, which at 2 cents per pound would cost $46.10. Summary of Cost. The several items may now be summed up as follows: 1. Daily interest on cost of construction on electrolytic cells, $0.0137 x C.S.E. 2. Daily interest on cost of construction of electrodes, $0.0554 x C.S.E. 3. Daily interest on cost of construction of electric generating appliances, $0.030 per $537- [.II. P., or - . C.S.E. 4. Daily cost of generating electric power, $570.00 $0.0319 per I.H.P. per 24 hours, or . C.S.E. 5. Daily cost of metal used, $46.10. Total ($0.0137 + $0.0554) x C.S.E. $537 + $570 + + $46.10. C.S.E. This is evidently a minimum when the values of the two variables are equal, or when the cross-section of the electrolyte is 126,500 square feet. With this cross-section the potential differ- 8.65x41 600 ence between the plates would be , 126 500 or 3.18 volts. The cost would be as follows: 1. Electrolytic cells, 422 cubic yards masonry, daily interest. $ T -73 2. Electrodes, 1265 tons of iron, daily interest 7.00 3. Construction of generating appli ances, 142 average actual indi cated H.P., 42 I.H.P. reserve, daily interest 3.90 SUMMAKY AND DISCUSSION OF DATA OF 1897. 437 4. Operation of generating appli ances, 142 average I.H.P., daily cost 4.51 5. Metal used per day 46. 10 Total cost $63.24 Cost of Treatment with Persulphate of Iron. In regard to the sulphates it has already been show that their value is dependent upon the amount of available metal which they contain. In the purchase of these chemicals it is necessary, therefore, to consider their com position. The persulphate of iron which was used in these tests contained 24.43 P er cent, of iron and was of approximately equal effi ciency to sulphate of alumina containing 9.87 per cent, of aluminum. It is stated that this chemical could be purchased in carload lots F.O.B. cars, Louisville, for $1.25 per hun dred pounds. To treat 25 million gallons with i grain per gallon would cost, therefore, $44.62. Cost of Treatment with Sulphate of Alumina. During the investigations of 1897 three dif ferent lots of sulphate of alumina were em ployed. These lots contained different amounts of alumina and were purchased at different prices as follows: Lot. A.. 15 . . C.. 9.87 8.46 10.41 1.50 cents 1.40 " i.f 5 " As the value of a coagulating chemical de pends on the amount of hydrate forming metal which it contains, the costs of these three lots of chemicals must be reduced to the cost per pound of aluminum in order to compare them. These figures are as follows: [ o( Pounds Aluminum per Cost per Pound I ound Sulphate. Aluminum. A 0.0987 15.2 cents B 0.0846 16.5 C 0.1041 15.9 The above comparisons illustrate the ne cessity of purchasing these chemicals by the amount of metal which they contain. In the purchase of large lots it would undoubtedly be best to receive bids based on the amount of available aluminum in the sulphate offered. In all of these estimates lot A is used as a basis. The cost of treating 25 million gallons of water with i grain per gallon of lot A would be $53-55- Summary. The costs of treatment of 25 million gal lons of Ohio River water with the equivalent of i grain per gallon of sulphate of alumina containing 9.87 per cent, of aluminum, by the three available coagulants, would be as fol lows: Electrolytic iron treatment $63.24 Persulphate of iron 44-62 Sulphate of alumina " 53-55 Animal Cost of Treatment of Twenty-five Million Gallons Daily of Ohio River Water. The average amounts of the different meth ods of treatment which would be required to purify the Ohio River water at Louisville have been presented in section No. 13 of this chapter. Combining the averages given there with the relative cost of the different treat ments as given above, the average annual cost of treatment by the three methods of treat ment is obtained as follows: Estimated Average Annual Cost of Treatment. Electrolytic iron $40 400 Persulphate of iron 28 500 Sulphate of alumina 34 300 438 WATER PURIFICATION AT LOUISVILLE. CHAPTER XVI. FINAL SUMMARY AND CONCLUSIONS. FOR the sake of convenience and explicit- ness, the leading points of practical signifi cance are brought together in brief terms in this chapter. The circumstances under which the investigations and tests were conducted caused the evidence upon many of the points to appear in several chapters; but with the aid of the accompanying index detailed in formation upon the important features of the work may be obtained readily. Character of the Unpurified Ohio River Water. The suspended mud, silt, and clay in the Ohio River water make it in many respects an undesirable water for a municipal supply, and the large amounts and wide variations in the size and character of the suspended mat ter make it a difficult and expensive water to purify. With regard to the sanitary character of the river water, the large amounts, during the greater portion of the year, of suspended organic and mineral matter cannot be con sidered other than as objectionable, although there is no evidence to lead to the belief that these matters exert a specifically injurious effect upon persons in normal health. In fact it is in the low stages of the river when the water is comparatively clear that its hygienic character is least satisfactory. During the fall months there were repeatedly noted unmis takable signs of contamination of this water supply by the sewage of the cities located above it; and from time to time throughout the year, the conditions manifested them selves in the presence in the water of bacillus coli communis, which is the most prevalent germ in the feces of man and certain domes tic animals. The result of all tests for specific germs of disease, however, were negative. Practically speaking, the significance of this is that when the river is high and the water muddy the water is not dangerous, al though it is not free of suspicion for drinking purposes. When the river is low and the water clear, however, the healthfulness of the water is always questionable, and the degree of danger which its use involves depends upon the prevalence of disease in the cities higher up in the valley. If an epidemic of cholera or typhoid fever should break out in any of the upper cities, there are at present no reliable means of preventing the specific germs of disease from passing in more or less diminished numbers from the outfall sewers of the upper city by the river, to and through the reservoir and distributing mains, to the service pipes of the water consumers at Louis ville. It is true that several natural agencies such as dilution and sedimentation in the river and reservoir tend to remove these germs in a large measure, but such means cannot be depended upon now. As the population on the watershed increases, with no correspond ing and compensating changes in the natural conditions causing the removal from the water of sewage germs, the healthfulness of the river water will continue to decrease steadily and surely. Under these conditions it is imperative that whatever method of puri fication be adopted, it shall be capable of pro tecting the water consumers from water- borne diseases, because if this were not done the expenditure of the large sums of money necessarily involved in purification would not be justifiable. With regard to the storage and distribu tion of the unpurified Ohio River water, the amount of suspended matter in the water pre vents the penetration of the rays of the sun, when stored in an open reservoir, to such a degree that under ordinary circumstances no FINAL SUMMARY AND CONCLUSIONS. 43 J growths occur of algse, etc., which give rise to the objectionable tastes and odors. The water contains an ample supply, however, of soluble mineral matter suitable as a food for the mi cro-organisms. Carbonic and oxygen gases are dissolved in the water in such amounts that the water has considerable corroding ac tion upon iron pipes which are not coated with a protective paint. This corroding ac tion is much retarded by the suspended mat ters of the water, as they serve in a measure in forming protective coatings. The Ohio River water under the conditions met with in practice has no objectionable action on lead pipes, because the water quickly forms insoluble basic carbonate of lead which gives to the pipe an impervious protective coating. For use in steam-boilers the Ohio River water is not very desirable when compared with the soft and clear waters of the East. Nevertheless, comparing it with the still harder waters met with farther West, it is fairly satisfactory for boiler use. When the river water is muddy it forms in boilers large amounts of sludge which are removed with much difficulty on some occasions, as this sludge cannot be " blown off." At such times the amounts of sulphate of lime and of other incrusting constituents are small and they arc deposited upon the sludge for the most part and not upon the metal of the boil ers. During the fall months, however, when the river water is fairly clear the amounts of incrusting constituents are much larger than usual, and the fine clay in the water at such times unites with the sulphate of lime, etc.. to form a scale resembling cement and which is very difficult to remove. Applicability to the Purification of the Oliio River Water of the Three Methods Investi gated during these Tests. Three methods of purification were tested with general results as follows: i. The general method embodying subsi dence (sedimentation), coagulation, and fi 1 - tration, such as was practiced in part in the Warren. Jewell, and two Western systems, is correct in principle for the practicable purifi cation of this water. Tt had several weak nesses, as practiced in these tests, the most important one being the totally inadequate facilities in all cases for the employment of subsidence to its proper economical limits. This is shown more clearly beyond. _ . The I larris Magneto-Electric System was a complete failure. 3. The MacDougall Polarite System as it was tested by this Company was not ap plicable to the purification of the ( )hio River water. Imperativeness of the Use of Coagulants. Owing to the fact that at times this river water contains large amounts of very minute clay particles (many of which are as small as o.ooooi inch in diameter), it may be stated in unqualified terms that at least for several suc cessive weeks in the spring and early summer, successful and economical purification of this water makes the use of coagulation impera tive in connection with subsidence. Whether or not it is absolutely essential or desirable to employ coagulation in connec tion with filtration of a properly subsided water is a problem which would depend upon the rate of filtration, but upon which no specific data were obtained in these tests. Relative Applicability of American anil English Types of filters. With regard to the filtration of the Ohio River water after partial purification by plain subsidence and subsidence aided at times by coagulation, by the American and English types of filters, no comparisons were made during these investigations. Taking into consideration, however, the general informa tion obtained in these tests as to the character of the water, combined with the results of the tests made with English filters in 1884. the indications point to the superiority of the American filters for this water, owing to their improved facilities for cleansing the sand layer. Here it may be noted that the tests in 1884-85 were made for about eight months in tanks 12 feet in diameter. The water pass ing on to the filter was subsided for about six days in the Crescent Hill reservoir without coagulants. During this period the excessive amounts of fine clay, as found frequently in the spring, were largely absent. The filters were constructed after the English plan, as 440 WATER PURIFICATION AT LOUISVILLE. recommended and described by Kirkwoocl, and the sand had an effective size of 0.36 millimeter. This agrees very closely with the sixe of sand employed in the best filters in Kurope. As a result of the tests of 1884-85 it was learned that the clay could be removed and an effluent free of turbidity secured by Eng lish filters at a net rate of about 1.5 million gallons per acre daily, lint the principal point of practical significance was the marked indication of the clay passing into the sand layer, and the necessity of cleaning and recon structing the sand layer at periods of com paratively short duration. Removal of Coarse Mailers by Plain Subsidence. The entire absence of this very important and essential feature of successful purification of water of this character comprised the greatest weakness of all the systems tested. This subject was given considerable attention on a laboratory scale, and it was found that with ordinary muddy water about 24 hours of quiescent subsidence in one-gallon bottles caused a removal of about 75 per cent, of the suspended matters by weight. As it is well known that on a large scale the quiescent state of the water cannot be obtained in so short a time, the above period is only of pass ing significance in connection with laboratory experiments, and for the most practicable conditions on a large scale it is necessary to rely upon information from other sources. This subject is dealt with in section No. T of Chapter XV, and it may be added that in practice the period should, in all probability, be much longer than 24 hours. Most Suitable Coagulant for the Ohio River Water. All things taken into consideration, the most suitable coagulant at present for the treatment of the Ohio River water is sulphate of alumina. Persulphate of iron in equiva lent amounts is now slightly cheaper, but the difference is not sufficient to offset certain advantages of sulphate of alumina. The electrolytic production of aluminum hydrate from metallic aluminum electrodes is impracticable, both on the grounds of ex cessive cost and of irregularities in efficiency. With regard to the electrolytic production of iron hydrate from iron electrodes this process yields a satisfactory coagulant up to the equivalent of 3 grains per gallon of sulphate of alumina. Beyond this point it could not be safely employed in midsummer, when the amount of dissolved oxygen in the water is inadequate to oxidize larger quantities. Combining the cost, the limitations in the amount which can be safely applied without the presence of iron in the effluent and cer tain irregularities upon reversing the direc tion of the electric current, this process is not considered advisable. Concerning the Anderson process for the preparation of iron hydrate directly from metallic iron, the results of laboratory tests indicate that it is not applicable for the eco nomic and efficient purification of the Ohio River water. Of the various chemicals mentioned in section No. 2 of Chapter XV, no others were found practicable. The method of dissolving known amounts of the chemical in known volumes of water is the best. The passage of a stream of water through a cylinder containing the chemical is not practicable. Concerning the application of the solutions of the chemical the Warren device was fairly automatic, but possessed several faults, the chief one of which was the failure in the op eration of the device when the flow of water fell below a certain quantity. The pumps used in the Jewell and modified Western systems were satisfactory, but required great care and close attention. Taking everything into consideration it is believed that in practice the discharge of a solution by gravity would be the most ad visable. Coagulation and Subsidence. In addition to plain subsidence and to co- FINAL SUMMARY AND CONCLUSIONS. 441 agulation given to water just prior to filtra tion, there are times when coagulation in con junction with subsidence can be employed to advantage in keeping clay and other sus pended matters from passing on to the sand layer. Such times would occur in practice when the water after plain subsidence would require more than from 1.5 to 2.0 grains per gallon of sulphate of alumina for thorough coagulation. In this respect all of the sys tems were lacking, although the practical sig nificance of this point was realized by the operators of the Warren System, as shown by their division of the application of coagulants in July, 1896. With regard to the optimum period of time to provide for the accomplishment of coagu lation and subsidence, it would vary accord ing to the amount and character of the sus pended matters present in the water after plain subsidence had taken place. The indi cations are that it might reach or exceed 6 hours in many instances, but the economic period would be limited by the cost of sub siding facilities. Coagulation and Filtration. After the river water has been properly treated by subsidence for the removal of sus pended matters, it is imperative that the water as it reaches the sand layer be thoroughly coagulated from a practical point of view. In the absence of complete coagulation, or very nearly complete, the efficiency of nitra tion cannot be uniformly depended upon. Concerning the optimum period of coagu lation of the water prior to nitration, it would vary widely from time to time in the purifica tion of this water. When the water contains very little suspended matter the period should probably be not more than half an hour. But as the quantity of suspended matter increases, the period of coagulation (and supplementary subsidence) should increase. In some cases the optimum period would be at least 3 hours, and probably longer. The longest period of course would be found when the amount of coagulant fell just below the point where economy demands a division in the applica tion (1.50 to 2.00 grains per gallon). Point of Application of Coagulant with Refer ence to the Period of Time Elapsing be tween Application and the Entrance of the Water into the Sand Layer. The results of these investigations prove conclusively that at the present time no fixed point of application of coagulant would fulfill the demands of economy. In the light of our present knowledge, the devices for the appli cation of coagulant should be made adjust able with reference to the point of applica tion. Whether or not it would ever be practicable in the treatment of this water to confine the application of coagulants to a range of three or four points can be told only by experience under the conditions of suc cessful practice. Total Annual Average Amounts of Sulphate of Alumina Required for Coagulation. Taking into consideration the fact that two periods of extended droughts occurred dur ing the tests of 1895-96, the data show that with the systems tested at that time the annual average amounts of sulphate of alu mina would range from 2.5 to 3.5, and aver age about 3.0 grains per gallon of filtered water. By taking advantage of subsidence to its economical limit, the investigations of 1897 indicate clearly that this could be held at from 1.5 to 2.0 grains, with an annual average of about 1.75 grains. In these comparisons it is assumed that in each case a good grade of sulphate of alumina would be used. Filtration. In respect to filtration proper, independent of subsidence and coagulation, the Jewell fil ter on the whole was found to be more satis factory than the others examined in these tests. The capacity of niters of this type is con sidered to be 100 million gallons of filtered water per acre daily. This means that to ob tain one million gallons of filtered water in 24 hours it would be necessary to provide 435.6 square feet of filtering surface. To rate these filters at a lower capacity is out of question, 442 WATER PURIFICATION AT LOUISVILLE. and the indications are that when the bulk of the suspended matters is removed by sub sidence, and the operation of a system placed on a practical basis, this capacity could be safely increased to meet the increased con sumption of water. It is probable that this capacity under the stated conditions could be raised 50 per cent, with satisfactory results. The Jewell filter did not contain all 01 the best features of filters of this type, especially when compared with the Warren filter, and it could be improved in a number of ways, both with regard to construction and operation. An outline of the more important features which successful filters in practice should comprise is as follows: Essential Features of American Filters for the Successful Filtration of Twenty-five Million Gallons of Ohio River Water Dail\. Experience obtained during these investi gations shows the practical importance of the following points: Condition of the Water Entering the Sand Layer. The evidence is very decisive that so far as practicable the suspended matters should be removed before reaching the sand layer, and that, at that point, the water should be thoroughly coagulated. Further, it is clear that subsidence should be employed with waters of this character to a degree where the amount of coagulant to be applied at or just before the entrance to the filter should not frequently exceed 2 grains per gallon. Failure to make suitable provisions in this respect caused the Western gravity filter to be voluntarily withdrawn from the tests be cause it was unable to purify enough water, when the river water was very muddy, to wash its own sand layer; and in the two best niters, the Warren and Jewell, it may be con servatively stated that to maintain the full supply at times of heavy freshets it would be necessary to provide reserve filters equa 1 to 75 per cent, of the normal capacity of the plant. Furthermore, the failure in these sys tems to remove the coarser particles by sub sidence would in practice cause a large waste of coagulants, as stated above. Structure of Filters. For a permanent plant the use of metal, as in the case of the Western pressure filter, would be preferable to wood. The foul odors in the bottom of the Warren filter when it was removed at the ciose of the tests, shows an objection to the use of wood for closed compartments. Size of Filters. The several filters repre sent the prevailing size in practice, but tor economy in operation the individual filters should be made much larger, the limit to be determined by the successful operation of me chanical appliances to stir the sand layer effectively while it is being washed by a re verse flow of water. Location of Sand Layers. The location of the sand layer near the top of the filter tank, as in the case of the Jewell filter, is an advan tage, because it guards against the waste of coagulated water above the sand layer just prior to washing and it would also reduce the cost of construction. Character of Sand Layers. Experience in dicates that in all cases the frictional resist ance to the flow of water was too small. This could be remedied by using a sand of finer grain or a layer of greater thickness, or both. In the judgment of the writer it would be ad visable to maintain a thickness of 30 inches and employ a sand having an effective size of 0.35 millimeter. This would increase the frictional resistance of the sand layer in the Jewell filter about 50 per cent., other condi tions being equal. Filtered Water E.rits. The Western filter did not satisfactorily meet this difficult prob lem, as the sand passed into the slotted brass tubes. In the Warren filter there was little chance for lateral and irregular flow of water at the bottom of the sand layer, except as caused by the supports beneath the per forated plate. All things considered, it is be lieved that the Jewell filter was superior in this respect, although it would probably be advisable to double the number of strainer cups. Amount and Nature of Pressure (Head). The indications are that 10 feet for a maxi mum acting head would be advisable under the conditions of practice. So far as could be learned the negative head (suction) in the Jewell filter gave directly no advantages over a positive head with regard to the efficiency FINAL SUMMARY AND CONCLUSIONS. 443 of the filter. In consequence of a negative head, however, there are several advantages as noted above in connection with the waste of coagulated water and the cost of construc tion. There were no indications that the use of a pressure filter, as represented in the Western pressure filter, would be advisable in purify ing this water supply. Rate of Filtration. The evidence and con clusions upon this point are presented above in reference to the capacity of filters. ]V ashing the Sand Layer. Experience showed that when the sand layer requires washing it should be done thoroughly with filtered water, and that accompanying agita tion of the sand layer is an advantage. The agitator of the Jewell filter was of a type su perior to that in the Warren filter, but the teeth of the rake arms should penetrate as nearly to the bottom of the sand layer as safety would allow. Surface Agitation. This process could be profitably employed to a greater degree in practice than was the case in the Jewell filter, the only one of these filters in which advan tage of this was taken at all. In practice the filter tank should be designed so as to allow the water above the sand, together with the surface accumulations, to be flushed off into the sewer during agitation. The use of a finer sand would also be an advantage in this connection. Application of Caustic Soda. From time to time the use of caustic soda, to keep the sand free of matters which are absorbed and at tached to it, is advisable. Attention. This is a very important factor in the efficiency of filters of this type in the purification of the Ohio River water, and economy as well as efficiency demands that they shall receive skilled attention, especially to prevent a waste of coagulants. With suit able provisions for subsidence, the necessary amount of care and skill would be much less than indicated by these tests, after the plant had been placed upon a practical and sys tematic basis of operation. Accessibility of Parts. Improvements in all of these filters should be made with regard to accessibility of parts, in order to facilitate ex amination and repairs whenever necessary. Quality of the Ohio River Water after Puri fication by Coagulation and Filtration, pre ceded by Subsidence so far as Practi cable. With proper attention to the operation of a system as outlined above, and an adequate degree of coagulation (by sulphate of alu mina) of the water as it enters the sand layer, this method could produce a quality of filtered water which would be thoroughly satisfactory under all ordinary conditions with regard to appearance and sanitary character. Owing to the inherent character of the Ohio River water and the local conditions, the filtered water could not be stored in open reservoirs, except for very short periods, with any reasonable assurance that algri 1 , etc., coming from the air would not grow in the presence of sunlight and give to the water objectionable odors and tastes. To guard against this effectively the reservoir in which the filtered water is stored should be covered. The filtered water would not give any trouble in the case of lead pipes, or in iron pipes which are properly coated with pro tective paints. In uncoated iron vessels the corrosive action would be somewhat greater than in the case of river water, owing princi pally to the removal of suspended matters, which in a measure act as a protective coating. With regard to use in steam-boilers there would be more incrusting constituents than in the river water, although the annual aver age amount in the filtered water would be only about 60 per cent, of the quantity nor mally present in the river water during the fall months. The effect of this addition would be largely if not wholly offset by the removal of the suspended matters; and, compared with the waters of other cities, it \\ould be classed as a satisfactory boiler water. Final Conclusion. The general method of subsidence, coagu lation, and filtration is applicable to the satis factory purification of the Ohio River water at Louisville; but, as practiced by the Warren, Jewell, and Western systems during these tests, its practicability is very questionable if not inadmissible. By removing the bulk of 444 WATER PURIFICATION AT LOUISVILLE. the suspended matters from the water, large reductions could be made in the size of filter plant, amount of coagulant, and cost of op eration. On the basis of twenty-five million gallons daily, these reductions when capital ized at 5 per cent, would represent about $700,000. There is no room for doubt but that for a less sum than this satisfactory pro visions for subsidence as outlined herein could be provided, which would not only aid in furnishing a filtered water of better quality, but would also give the water consumers a better service in other regards. Very respectfully submitted, GEORGE W. FULLER, Chief Chemist and Bacteriologist. LOUISVILLE, Kv., Oct. 7, 1897. APPENDIX. THE appendix consists of a brief technical resume of the methods of analysis which were employed during these tests and investiga tions, together with some notes on the collec tion of samples. COLLECTION OF SAMPLES. The only departure from the standard methods and technique of collection of sam ples was in the use of a device termed the " automatic sampler." Automatic Sampler. For the purpose of collecting samples which should be represen tative of the effluent for long periods of time, the device described below was arranged and used during March, April, and early part of May. 1896, for the collection of samples from all the systems, and from May n to July 23, 1896, for the collection of samples for chem ical analysis from the Western Pressure Sys tem. The device consisted of a brass cylinder (A) closed at one end. of about 5 cubic centi meters capacity, and with a single orifice on one side. This cylinder was ground into a covering cylinder (B) in which it was revolved by a feathering paddle-wheel in the main efflu ent pipe. The wheel was operated by the cur rent and was designed to turn at a rate pro portional to the flow of water through the pipe. A small pipe led from the opposite side of B to the collecting bottle. The openings for these pipes and for the orifice in A were in the same plane. An air pipe with protected end was provided from cylinder A, as was also an escape from the collecting bottle. The operation was as follows: As cylinder A re volved its orifice connected with the inlet ori fice through cylinder B and water from trie effluent pipe entered and filled cylinder A and the air tube. As cylinder A turned, it cut off the connection with the inlet pipe, and when one-half around, connected with the outlet orifice through cylinder B and the contents of the cylinder and air tube were dis charged through the outlet tube into the col lecting bottle. In this manner a small sample was taken at each revolution, the gearing- being so proportioned that at the normal rate of 23.5 cubic feet per minute, about six sam ples were taken each minute. A gallon bottle was used as a collecting bottle. The whole device was inclosed in a wooden box. . REGULAR CHEMICAL ANALYSES. Form of Expression. The results of. the determinations of the several chemical constituents arc expressed in this report in parts by weight per million parts of water by volume. This form was adopted mainly for the reason that the re sults of a previous series of analyses for this Company had been expressed in this manner. Turbidity. During the early part of these tests, use was made of the adjectives commonly employed in describing the results of inspection of the samples as they appeared after settling over night in one-gallon bottles. During 1896 a " diaphanometer " was used for a time. This instrument consisted of a brass tube in which the sample of; water was placed and through which light was reflected from a Welsbach gas lamp. The turbidity of the waterwas estimated as the reciprocal of the length of a column of water which would cause the image of a cross of black lines at the bottom of the tube to dis appear. Fairly satisfactory results were ob- 445 446 APPENDIX. tained with waters of a slight but noticeable turbidity; but it appeared to be inadequate for regular use in this connection with satis factory results upon the fairly clear effluents or verv muchly river water. The relative amounts of sediment in the river water were estimated in the early part of the work by inspection of the samples after settling over night in one-gallon bottles. These observations seemed to be of but little value, and later were abandoned, comparison of sediment then being made upon the parts by weight of suspended matter. Odor. The odor of the water at room temperature and after heating in a beaker to 100 C. was observed and recorded, respectively, as the " cold " and " hot odor," in substantially the same manner as described by Drown in the Report upon the Examination of Water Sup plies by the Mass. State Board of Health, 1890, Part 1. Color. The dissolved color of the water was meas ured by the platinum-cobalt standards of Hazen, as described in the American Chem ical Journal, Vol. XIV, page 300. O.v\gcn Consumed. The Kiibel method was used in substan tially the same form as described by Drown in the Report of the Mass. State Board of Health for 1892, page 328. In this determination the period of boiling with potassium permanganate exerts great in fluence on the results obtained; and. further more, this period differs considerably in differ ent laboratories. It was the custom here to boil exactly five minutes after adding the potassium permanganate, the water and sul phuric acid having been previously raised to the boiling temperature. In order that the results obtained in this laboratory may be compared with previous work at other places, Oh Sample. III HCS O u io Rh erw IO 2 5 10 2 5 10 5-4 6.8 8-5 1.9 2-3 2.8 0.7 1-4 1-4 filtered without coagulant with coagulant. . the following average results of a number of experiments are presented. EXPERIMENTS TO SHOW THE EFFECT OF DIFFERENT PERIODS OF BOILING ON THE AMOUNT OF OXYGEN CONSUMED. Nitrogen as Albuminoid Ammonia. The method of \Vanklyn as modified by Drown, Hazen, and Clark was used substan tially as described in the Report of the Mass. State Board of Health for 1890, Tart II, page 710, and also in the American Chemical Jour nal, Vol. XII, page 425. Determinations of the " total " nitrogen as albuminoid ammonia were made on the unfil- tered water; of the " dissolved " after the pas sage of the water through filter-paper or a Pasteur filter. And the " suspended " nitro gen as albuminoid ammonia was obtained by difference. Nitrogen as Free Ammonia. The method of Wanklyn was used as de scribed by Drown. Hazen, and Clark in con nection with the nitrogen as albuminoid am monia, referred to above. Nitrogen as Nitrites. In the determination of nitrogen as nitrites the Griess-Warrington method was used, as described by Drown and Hazen in the Report of the Mass. State Board of Health for 1890, Part II, page 715. Nitrogen as Nitrates. The " aluminum method " was used as modified by Hazen and Clark and described APPENDIX. 447 in the Report of the Mass. State Board of Health for 1890, Part II, page 711. The following method was employed in the preparation of nitrate-free water used in blanks. Eight liters of ordinary distilled water were treated with 100 cubic centimeters of a 50 per cent, sodium hydrate solution, and 5 grams of pure aluminum foil. After some hours the water was placed in a still with 3 grams of potassium permanganate and dis tilled. The middle portion of the distillate was free from nitrates. In preparing the sodium hydrate solution one liter of this water and 250 grams of the purest sodium hydrate obtainable were brought together in a porcelain dish with about 2 grams of pure aluminum foil. When the foil was all dissolved, the solution was boiled down to a volume of 500 cubic centi meters, and, after being allowed to settle, fil tered through asbestos. Two cubic centimeters of this solution with 50 cubic centimeters of water and 0.35 gram of aluminum foil, should indicate the pres ence of only a very slight amount of ammonia when treated in the same manner as samples for analysis. Chlorine. Chlorine was determined according to the method of Mohr, as modified by Hazen and described in the American Chemical Journal, Vol. XI, page 409. Residue on Evaporation. For this determination 100 cubic centi meters of the water were evaporated to dry- ness on a water bath, in a platinum dish of known weight. After drying in a steam oven (temperature 96 to 98 C.) for two hours, the dish with its contents was cooled in a desiccator over sulphuric acid and weighed. This gave the "total" residue on evaporation. A similar determination on the filtered sample gave the " dissolved " and the difference be tween these two gave the " suspended " resi due on evaporation. Fixed Residue after Ignition. The fixed residue after ignition was de termined by igniting the residue on evap oration at a low heat in a radiator, and weighing as usual after cooling in a desicca tor. Alkalinity. The method of Hehner for determining the alkalinity was used as described in Leff- mann s Examination of Water, edition of 1895, page 82. Owing to the fact that the color of the Ohio River water on many oc casions obscures the end point when methyl orange is used as an indicator, lacmoid was employed in the hot sample. Methyl orange as procured at this laboratory was also lacking in sensitiveness. Furthermore, this indicator is incapable of showing the presence of unde- composed sulphate of alumina, a property which is possessed by lacmoid. The determinations of alkalinity with lacmoid as an indicator, in preference to tem porary hardness by the soap method, were necessary in this work for two reasons: 1. Normal carbonates, which are deter mined as permanent hardness or incrusting constituents by the soap method, have the power of decomposing sulphate of alumina in the same manner as the temporary-hardness constituents. 2. The presence of undecomposed chem icals (sulphates) in the effluent is not shown by the ordinary temporary-hardness deter mination, but is readily detected by the alka linity determination when lacmoid is em ployed as an indicator, as the sulphate of alu mina reacts acid to this indicator. The alkalinity determination, therefore, is a measure of the capacity of the water for de composing sulphate of alumina, and an acid reaction to lacmoid indicates the presence of this chemical in an undecomposed form. Incrusting Constituents Permanent Hardness. The salts determined by the Hehner per manent-hardness method have been classed as incrusting constituents, and the normal car bonates have been determined as alkalinity. The reasons for this are presented above. This method is described by Leffmann in Examination of Water, edition of 1895, page 82. 448 APPENDIX. Total Hardness. In regard to the total hardness of the water as indicated by its soap-destroying power, the sum of the determinations of the alkalinity and incrusting constituents is comparable to such results, although not necessarily iden tical. For the determination of dissolved alumina the method of Richards was used as described in Leffmann s Examination of Water, edi tion of 1895, page 58. Iron, For the determination of iron the general method of Thompson was used, as described in Leffmann s Examination of Water, edi tion of 1895, page 57. Dissolved Oxygen. The method for the determination of dis solved oxygen devised by Winkler and modi fied by Drown and Hazen was used as de scribed in the Report of the Mass. State Board of Health for 1890, Part II, page 722. Carbonic Acid (Carbon Dioxide). The determination of carbonic acid (carbon dioxide) in the water was made according to Trillich s modification of Pettenkofer s method, substantially as described in Ohl- miiller s Untersuchung des Wassers, edition of 1896. MINERAL ANALYSES. Silica, Barium, Strontium, Iron, Aluminum, Nickel, and Manganese. The methods employed in the determina tions of silica, barium, strontium, iron, alu minum, nickel, and manganese, were as fol lows: Silica. One or two liters of the water (de pending upon its character) were acidulated by the addition of 20 cubic centimeters of hydrochloric acid, and evaporated to dryness. The dish containing the residue was trans ferred to an air bath, and .heated to 130 C., until the hydrochloric acid was driven off. There were then added 10 cubic centimeters of hydrochloric acid and 150 cubic centi meters of distilled water, and the whole was gently warmed. The silica was filtered off, washed, dried, ignited, and weighed. The residue was then treated with hydrofluoric and sulphuric acids, and the process repeated. The difference in weight of the residue before and after this latter treatment was taken as the amount of silica present. Barium and Strontium. Any residue re maining after the treatment with hydrofluoric and sulphuric acids was fused with sodium carbonate, and the barium and strontium pre cipitated and weighed together as sulphates. Iron and Aluminum. The filtrate after the removal of silica, barium and strontium was treated for iron, manganese, etc., as follows: The iron and aluminum were separated by the basic acetate method, the combined oxides weighed, and the aluminum deter mined by difference, the iron being first determined by the Thompson colorometric method already referred to. Nickel and Manganese. The filtrate from the iron and aluminum determination was rendered slightly acid by the addition of acetic acid, and nickel, if present, was precipi tated with hydrogen sulphide, and the man ganese separated in the filtrate after neutral ization by the addition of ammonium sulphide. Calcium and Magnesium. The filtrate from the manganese determina tion was boiled and filtered if necessary, and the calcium and magnesium determined by precipitation as calcium oxalate and mag nesium pyrophosphate, respectively. Sulphuric Acid. One liter of the water was evaporated to dryness in a porcelain dish with 10 cubic centimeters of hydrochloric acid. The silica was removed (see silica determination) and the sulphur determined by precipitation as barium sulphate. A I l END IX. 449 The phosphoric acid was precipitated from tlie filtrate obtained in the determination of the sulphuric acid by ammonium molybdate. The " yellow precipitate " \vas dissolved in ammonia and the phosphorus determined as magnesium pyrophosphate. Sodium and Potassium. Two liters of the water were evaporated to dryness in a platinum dish with 20 cubic centimeters of hydrochloric acid. The residue was taken up with water, a few crystals of barium hydrate added and the solution boiled. The solution was cooled and the precipitate allowed to settle. The volume was then made up to 250 cubic centimeters and 200 cubic centimeters were filtered off. Ammonium carbonate and ammonium oxalate in the solid form were added to the filtrate and the mixture brought to a boil. After cooling the volume was again made up to 250 cubic centimeters and 200 cubic centi meters filtered off. Manganese or nickel, if present in sufficient amount to interfere, were precipitated as sulphides. The filtrate was evaporated to dryness in a platinum dish, sufficient sulphuric or hydro chloric acid being added to unite with the sodium and potassium. After ignition to ex pel ammonium salts, redissolving in hot water, filtering and evaporating to dryness, the potassium and sodium were ignited and weighed as sulphates or chlorides according to the acid added. The percentages of potassium and sodium in the mixture were calculated after determin ing the common constituent (i.e., the Cl or SO :! ) from the formula given by F. K. Landis, Jour. Anier. Chem. Soc., Feb., 1896. QUANTITATIVE BACTKRIAI. AXALVSKS. The general technique of the quantitative bacterial methods used in these investigations was substantially the same as in the case of those described by Fuller and Copeland in the Report of the Mass. State Hoard of Health for 1895, page 585. A brief summary of some of the principal features is as fol lows: At the outset of these investigations glycer ine agar was used as the culture medium, ow ing to its adaptability under the existing local conditions. The laboratory itself had a widely variable temperature, owing to climatic con ditions and to the presence in the building of a number of steam-pipes used for heating and general laboratory purposes. The fact that, at a number of points on the steam-pipes be tween the laboratory and the boilers, steam was used from time to time in varying amounts by the experimental devices, caused the steam pressure to vary; and in conse quence it was difficult to regulate the labora tory temperature. The maintenance of a uni form temperature in the thermostat was difficult, also, on account of the necessity of using gasoline gas which varied widely not only in quality but in pressure. I nder these circumstances glycerine agar was at first selected as the culture medium with the view to getting satisfactory data upon the efficiency of the filters and at the same time preventing heavy losses of data which seemed imminent with the use of gela tine. The precaution was taken, however, of making comparative studies of the numbers of bacteria obtained by glycerine agar and gelatine, respectively. As a result of these studies it was found that while satisfactory data could be obtained for comparing the efficiency of the several filters, yet the actual numbers of bacteria were normally, but not always, materially lower when glycerine agar was used than in the case of gelatine. Accordingly, when the scope of the investi gations was enlarged on Feb. i, 1896. it was decided to adopt nutrient gelatine as the regular culture medium, and to guard against loss of results through liquefaction of the medium by employing a sufficiently low tem perature for incubation. Reaction of Culture Media. After considerable study it was found that the reaction of media used in the regular quantitative work which gave the most satis factory results under the local conditions was 45 APPENDIX. 1.5 percent, (equal to 15 cubic centimeters of normal hydrochloric acid added to every liter of neutral medium), especially in the case of the effluent, and this reaction was, therefore, maintained throughout the tests. The rela tive effects of reactions ranging from 0.5 to 2.0 per cent, are shown in the following table of average results: BACTERIA PER CUBIC CENTIMETER ON GEL ATINE OF DIFFERENT REACTIONS. Sampl liact ria per Cu )ic Centim ter. Number of Reaction (r er cents.t. 0.5 ,.o i-5 ,.o Effluent 41 > i 1 IlS 2 4 S 298 243 Sterilization. All glassware, such as Petri dishes, pipettes, sample bottles, etc., were sterilized for one and one-half hours in a hot- air sterilizer at a temperature of 150 degrees C, or a little higher. All media and water for dilution purposes were autoclaved for 10 minutes under a steam pressure of 20 pounds. Dilution. In the case of the river water when it contained high numbers of bacteria, one cubic centimeter of the sample was di luted in 100 cubic centimeters of sterile dis tilled water, and in the case of very turbid effluents a dilution of i to 10 was used. The normal period of cultivation em ployed was 4 days in an ice-chest 10 feet long in which the bacterial compartment and ice compartment were in the opposite ends. In some cases, however, when the temperature of the bacterial compartment fell below 16 to 18 degrees C., a longer period was allowed in order that the maximum growth might be more nearly obtained. It was deemed advisa ble at times, owing to the local conditions facilitating melting and liquefaction of the gelatine, to maintain a low temperature in this compartment, and thus guard more effectively against the possibility of loss of plates through liquefaction. Owing to this precautionary procedure the loss of plates due to this cause was trifling, and it is believed that it was more advantageous to prevent the possibility of loss of data upon samples which were collected under conditions which might not occur again, than to strive at all times towards the maximum possible growth. It is to be noted, further, that the conditions were the same for all samples at the same time, and thus strictly comparable results were obtained in the case of the respective filters. The relation between the fourth and succes sive day growths and the influence exerted by temperatures ranging from 10 to 18 de grees C., are indicated by the results and per centages presented in the next table. In 1897 temperatures of less than 14 degrees C. were avoided so far as possible by partially open ing the door of the ice-chest as occasion de manded. So far as our knowledge goes the bacterial results obtained during the investi gations are very appreciably higher than would be obtained by two days cultivation at 20 degrees C., which is the conventional procedure in Europe. SUMMARY SHOWING THE RELATION BETWEEN FOURTH , FIFTH-, AND SIXTH-DAY GROWTHS OF BACTERIA ON GELATINE AT DIFFERENT TEMPERATURES. Number of Range of * per Cublc Ccn line r. Samples Averaged. T D ei > r e e r e s U c!" Source of Sample. Fourth Day. Fiflll Day. Sixth Day. Fourth to FiflhDay. Fourth to Sixth Day. 27 10-12 River water Effluent 20 500 29 600 211 31 700 395 44-4 32.7 54-6 142. i 42 12-14 River water Fffluent . ... 23 ()OO 32 700 47000 36.8 96.7 IO Fffluent 33 s 7 68 16-18 16-18 River water Effluent 12 2OO 155 14 100 167 12 IOO 170 15.6 7-7 o.o 9-7 APPENDIX. 45 In the next table are presented the monthly averages of temperatures of the quantitative bacterial compartment. MONTHLY AVERAGES OF THE TEMPERATURE OF THE QUANTITATIVE BACTERIAL COMPARTMENT. -895. 1896. 1897. Total Averages Nov. Dec. Jan. Feb. Mar. Apr. May. June. ; July. Aug. Feb. Mar. Apr. May. June. July. Maximum .... 22.0 20.7 20.1 12.8 14.6 14.2 16.9 I6.S 19.4 15." 16.1 15-2 16.1 >7-7 19.5 19.0 17.3 Minimum. . . . 19.6 20. 2 IQ.2 10.7 10.3 II.7 15.0 I 5 .8 16.7 14.0 I5.I 14.0 15-5 15-5 18.0 16.7 15.4 Mean 20.8 20.4 19.6 II. 8 12.5 13.0 16.0 16.3 18.0 14.3 15.6 I 14.6 I5.S 1 6. 6 18.7 17.9 16.4 Qualitative Bacterial Analyses. The qualitative bacterial work was divided into examinations for sewage bacteria, no tably B. coli communis, and a comparison of the species of bacteria in the water before and after purification, with an incidental classifica tion of the more common forms of bacteria. For this purpose the methods and tests were employed as found in the best manuals, and they were usually in harmony with modi fications suggested at the Convention of Bac teriologists held at New York in June, 1895. Examinations for B. coli communis. The method of procedure in the search for this organism was substantially the same as that recommended by Smith in the American Journal of the Medical Sciences for Sept., 1895. Dextrose and lactose broths were both used for this work and the reactions of all so lutions used in the fermentation test were ad justed to an alkalinity of 1.5 per cent. A temperature of 37 degrees C. was employed in all the tests covering the examination for this species, and all cultures (with the excep tion of the fermentation cultures, which were allowed to develop for four days) were al lowed to develop for 48 hours before observa tions were made. The cultures were started in flasks containing 100 cubic centimeters of sterile nutrient dextrose broth, 50 cubic centi meters of water being used each time for in oculation. Lactose litmus agar plates were made from these cultures, and from these plates colonies which presented in two days an appearance resembling that of B. coli com munis were transferred to gelatine tube, agar tube, peptone solution, litmus milk, and fer mentation tube. Observations were made after the customary period of two days on the gelatine and agar tube cultures, mi croscopically for si/.e and biologically for a characteristic growth; on the peptone solution for indol production: and on the litmus milk for coagulation both before and after boiling. The fermentation culture was inspected day by day, the quantity of gas recorded; and on the fourth day the total gas, the percentage of carbon dioxide, and the end reaction were determined. COXVERSION TABLES. There are given in the next tables the rela tions which exist between different methods of expression of several quantities which are used in connection with this report. Conver sion can be made from one form of expression to another by the use of the corresponding multiplication factors as given in these tables. Table No. II is copied from Kirkwood s Filtration of River Water. 452 APPENDIX. TABLE No. I. CONVERSION OF STATEMENTS OF CHEMICAL COMPOSITION. Grains per U. S. Gallon. Grains per British Gallon. (277 cubic inches.) Parts per ,00,000. Parts per ,,000,000. i grain per U. S. gallon i grain per British gallon I. 0.830 0.580 0.056 I. 2O I . 0.70 0.07 1.71 17.1 1-43 14-3 I. IO.O 0. 10 I. TABLE No. II. EQUIVALENTS OF VARIOUS MEASURES. Imperial Pounds Gallons. (Water at o C.) U S gallon o 8311 8.3388822 imperial gallon I . 20(J32 I . 4-54346 o. 16046 0.004543 277.274 IO. 08.31529 , . .... TABLE No. III. APPROXIMATE EQUIVALENTS OF VARIOUS MEASURES OF RATE OF FILTRATION. Vertical Million Million ! Tr c Velocity U. S. British i r U j, S British Cubic Feet Vertical Vertical -\j ( ,[.. rs Gallons Gallons iJi" 5 Gallons per Velocity Velocity per ; 4 Hours per Acre per Acre SquareFoot per Square Foot SquarcYard in Inches Millimeters ^ ub c Meter s 24 Hours. 24 Hours. pl per Hour. per Hour. per Hour. per Hour. ^^^7 ; 24 Hours. I million U. S. gallons per acre I . 0.833 o. 06 o.So 1 . 15 j . 53 39.0 0.935 i million British gallons per acre A hours I . 2OO I . I 15 O . 96 T . 1 S 1.84. 46- S I . 122 I U. S. gdlon per square foot per I .(>45 o . S 70 T . O.S3 1 , 2O i. 60 40 . 7 0.978 i British gallon per square foot 1.255 i ( ->45 1 . 20 1 . 44 1.92 48 . 9 I . I~4 i cubic foot per square yard per o. 724 0.83 o . 69 i . i . 33 33 9 0.813 i lineal inch, vertical velocity, per hour o. 652 O- 543 0.62 o. 52 o. 75 i . 25.4 o. 610 i hundred lineal millimeters, ver- tic il vplori v ner hour ^ ~f>fi o T TO 9.16 2 . 05 2 . 95 -\ OJ. IOO. 2 . 400 I lineal meter, vertical velocity - j i j ---,.- J VI per 24 hours = I cubic meter per square meter per 24 hours. . 1 . 069 0.891 i .02 0.85 1.23 i.6 4 41.7 I. INDEX. Accessibility of pans 253, 443 Agitation of surface of sand layer 100, 274, 326, 443 Agitator 75,8o Alg;e, see Microscopical analyses, uuJ Biological character. Alkalinity, determinations of, see Chemical analyses. increase of, by permanganates 382 reduction of, by sulphates 42, 53, 384, 429 Alum, potash, action of, as a coagulant 42, 380 application of, by Western systems.. 44, 47 amounts of, see Alumina, sulphate of. composition of lots of, used 41, 337 cost of, comparative 380 effect of, see Alumina, sulphate of. preparation of solutions of, 47, 251 Alumina, acetate of, action of, as a coagulant 380 sulphate of. effect of application of 42, 53, 57 on cost of purifica tion 259 on quality of efflu ent, 42, 53-56, 384, 427, 434 amount required to purify Ohio river water 270, 430, 441 amounts used (ta8les) 269 with MacDougall system. ... 323 Jewell system, 49, 196- 204, 234, 240, 241 Warren system, 48, 187-196, 232, 240, 241 Water Company s de vices 342, 371-375 Western gravity sys tem, 49, 204-207, 236, 240, 241 Western pressure sys tem, 49, 207-214, 237, 240, 241 composition of lots of 40-42, 337 cost of, annual average estimated, to purify Ohio river water 437 comparative 437 decomposition of, effects of, 53, 244, 251, 384, 428-434 efficiency of, as a coagulant, before and , after decomposition 294 efficiency of, comparative 413-419 presence of undecomposed, in effluents.. 55, 427 Alumina, sulphate of, solutions ot, preparation of. .251, 259, 272, 440 suitability of, as a coagulant for use in purifying Ohio river water 440 see also Chemicals and Coagulants. Aluminate, sodium, action of, as a coagulant 380 Aluminum, cost of, as metal plates 299, 412 sulphate . 259, 299, 437 see also the various compounds, and Elec- trolytical preparation of aluminum hy drate. Aluminum hydrate, efficiency of, as a coagulant, when prepared in different ways, 294, 413-416 see also Electrolytical preparation of aluminum hydrate. Ameiican filters, see Filters, American. Analyses, see specific heading. Analytical work, methods employed in 445-451 plan of 10 Anderson process, use of, with Ohio river water 382 Appurtenances of Warren, Jewell, Western gravity and Western pressure systems, inventory of 89-93 Attention given to the respective systems 107 influence of, on cost of treatment 267 on quality of the effluent 258 required for efficient purification of Ohio river water 271, 443 Averages, grand, for the investigations from Oct. 1895, to Aug. 1896 240 B Bacteria, effect of coagulants on electricity on in city tap water (table) in effluent of (tables), Harris magneto-electric system Harris Company s devices. 282- Jewell system, 148-169, 196- 204, 234, 240, MacDougall polarite system, 328, Warren system, 128-148, 187- 196, 232, 240, Water Company s devices, 350- 366, 368- Western gravity system, 169- 174, 204-207, 236, 240, Western pressure system, 174- 186, 207-214, 238, 240, 453 179 j . 841 330 a M 37 241 141 454 INDEX. Bacteria, in Ohio river water, numbers of (table) 39 species of 37 in sand layers .... 257 passage of, through filters 248 Bacterial analyses, methods of 449-451 Biological character of Ohio river water 16, 36-39 after purifica tion 248, 427 Boilers, u<e of purified Ohio river water in .... 244, 429, 430 Brownell, see Devices, Marl; and Brownell. Carbon dioxide, see Carbonic acid. Carbonaceous mailer, amounts of, see Chemical analyses. daily removal of, by Warren, Jewell, Western gravity, and Western pressure sys tems (tables) 223-227 see also records <{ operation of the several devices and systems. Carbonate of lime, effect of chemical treatment on, see Alkalinity. Carbonic acid, amounts of, naturally present in Ohio river water 33, 34, 433 cost of removal of 434 influence of, on electrolytic preparation of iron hydrate 391 on the use of ferrous com pounds as coagulants. . 381 in purified Ohio river water, 54, 244 428, 433 Caustic soda, see Soda, caustic. Chemical analyses, explanation of 20, 32-35 of effluent of, regular sanitary, (tables) 445-449 Harris system 279 Harris Company s devices.. .282-289 Jewell system ... 117-122 Mark and Brownell devices, 310, 3U, 316 MacDougall system 328 Warren system 1 12-1 17 Water Company s devices 348 Western gravity system 122 Western pressure system. . . 124-126 Ohio river water, regular sani tary 21-31 special 32-35 see also Carbonic acid, Incrusting constituents, Mineral analyses, and Oxygen. Chemicals, coagulating, used, see Alum, potash ; Alumina, sulphate of; Iron, persulphate and protosulphate of; and Electrolytic prepara tion of coagulants. see also Coagulants. Chemicals, commercial, available as coagulants. ...378, 385 Chloride of alumina, use of, as a coagulant 380 Chlorine, application of, with Jewell system 46 Clay exiraclor, see System, MacDougall polarite. Cleaning, set part in question. Clearness, degree of, explanation of 215 of effluent of (tables), Jewell system, 117- 122, 196-204 MacDougall system. 329 Mark and Brownell devices. .310, 314, 316 Warren system, 1 12- 117, 222 Water Company s devices.. .348, 371-375 Western gravity sys tem 122, 222 Western pressure system... 124-126, 222 Coagulants, absorption of, by silt and clay m 383 action of various 381 , 384 effect of, on cost of purification 258 on quality of purified water, 42, 53-56, 384, 405, 429, 431 application of 324 devices for, 42-46, 260, 272, 338-341 uniformity of 251, 260 amount required, effect of character of sus pended matter on 384 to purify Ohio river water 273, 422, 430, 441 cost of available, annual average to purify Ohio River water. . . . 437 comparative 435 economical application of, to aid in sedi mentation 416 efficiency of available, comparative. . . 382, 413 relative, of those absorbed and not absorbed by clay.. 384 different amounts of, in sedi mentation 418 germicidul action of 383 kinds of, used 40-42, 259, 310, 323, 337 loss of, when applied in small quantities... 384 maximum safe amount of (with Ohio river water) electro- lytically pre pared hydrate of iron 397 sulphates 383 metals available as 378 most suitable one for use in the purification of Ohio river water 440 necessity of use of, to purify Ohio river water 40, 439 point of application of 441 reaction of, speed of 383 waste of 384 INDEX. 455 Coagulation, application of, to Ohio river water 57 degree of 253, 262, 422, 440 devices for 58-61, qi, 272, 336 effect of 252, 261, 273 period of 273, 440 effect of 262, 419 process of 57 see also Filtration and Sedimentation. Conditions of these tests 1 1 Contents, table of iii-vii Controller, automatic, of Jewell system 79 Copperas, see Iron, protostilphnte of. Corrosion of iron by purified Ohio river water 244. 428 partial protective influence of sus pended matter against 428 lead by Ohio river water 434 D Decomposition of aluminum, electrolytic, see Electro lytic preparation of alu minum hydrate. chemicals, sec the specific compound, iron, electrolytic, see Electrolytic preparation of iron hydrate. Delays, table of, of Warren, Jewell, Western gravity and Western pressure systems 94 Deposition of metals, in electrolysis, see Electrolytic preparation of aluminum and iron hydrates. Devices arranged by the Water Company in 1897, adaptation of construction to existing con ditions 341 analyses in connection with, bacterial 350-366 chemical 348 see also Water, Ohio river, an alyses of. chemicals, kinds of, used 337 coagulants, quantity of, used 259, 342 comparison of coagulants 342 conditions and methods of operation 341-346 description, general, of 335 discussion of results, plan of 375 electric generating appliances 338 electrodes 339 electrolytic cells. 338 filter 341 filtration, rate of 342 interruptions of tests 341 objects of tests 334 operation, summary of results of (table)... 371-374 periods of operation 342 plans for testing 292 results accomplished, general description of.. . 346 results, plan of presentation of 334 runs, length of 342 special notes on 345 sedimentation, devices for 335 removal of bacteria and sus pended matter by 37-3?8 Devices, mechanical, to aid in the operation of the Warren, Jewell, Western gravity and Western pres sure systems. ... . ... 105-107 Devices of Harris Company : Device No. I, bacterial analyses in connection with. 282 chemical analyses in connection with 282 description of 281 electric appliances 281 electrodes 281 magnets, electro- 281 operation of 282 stand pipe 281 Device No. 2, description of 282 results accomplished 283 Device No. 3, description of 283 filter 284 operation of 284 results accomplished 284 Device No. 4, aluminum used with 286 description of 285 operation of 286 results accomplished 286 Device No. 5, aluminum used with 290 bacterial analyses in connection with. 288 chemical analyses in connection with 28q description of 287 operation of 287 results accomplished 289 glass jar experiments 280 Devices of Mark and Brownell : bacterial analyses in connection with .. 308, 314 Brownell electrodes, results accomplished with 309 Brownell cell, description of 304 passage of untreated water through 306 chemical analyses in connectionwith... 309, 314, 316 comparison of efficiency of filtration of water treated by Brownell electrodes, with that of water without treatment 313 water treated by similar aluminum electrodts 315 conclusion in regard to 313 construction of 304 electrical connections 305 electric generating appliances 304 electrodes 305 electrolytic cells 304 loss of hydrate by construction of cells 311 Mark electrodes, operation of 306 metal decom posed 310 plans for investigation of 303 preliminary experiments leading to 301 unofficial runs of 307 Devices, Palmer and Brownell, see Devices, Maik and Brownell. Electric aluminum process, see Electrolytic preparation of aluminum hydrate. Electric generating appliances, cost of construction of.. 435 operation of. . . . 436 jet also Devices of Mark and Brownell, and System, Harris. 456 INDEX. Electric iron process, see Electrolytic preparation of iron hydrate. Electric power required for the purification of Ohio river water by the use of electrolytically prepared iron hydrate 436 the preparation of alumi num hydrate electrolyti- cally (1896 data). 302 used in the Mark and Brownell devices. 312 Water Company s devices. . 370 waste of, in the electrolytical prepara tion of coagulants 404. 4Og Electric resistance, of coat ings on plates, of aluminum. 411,412 iron 404 Ohio river water 312, 390 solutions of various salts 393 Electricity, effect of, on bacteria and organic matter. . . 292 in purifying Ohio river water 293 Electrodes, active 389 iron, cost of construction of 436, 603 passivity of 391 polarization of 391 see also Devices, Harris Company, Mark and Brownell, Water Company; and Sys tem, Harris. Electrolysis, general description of 388 Electrolytic cells, cost of construction of- 435 see also same as for electrodes. Electrolytic decomposition, see Electrolytic preparation of coagulants. Electrolytic preparation of aluminum hydrate, conclusions in regard to, lSy6 308 1897 412 cost of the hydrate as compared with that prepared by the decomposition of the sulphate (1896 data) 291) decomposition of the positive plates.. 296,406 deposition on the negative plates. . . . 299, 408 description of 406 direction of the electric current, influence of reversing the 410 efficiency of the coagulant 413-416 formation of gas 402 influence of the composition of the river water 406 magnets, electro-, effect of 296 metal wasted 299, 411 passivity of electrodes (06 power wasted 412 regularity of formation of hydrate 299 scale, influence of the composition of the river water on the formation of. 410 presence of 410 Electrolytic preparation of coagulants, advantages of 300, 313, 387 electric laws of 390 polarization of electrodes 391 relative efficiency of iron and aluminum electrodes in 315 secondary reactions 392 see also Electrolysis. Electrolytic preparation of iron hydrate, effect of ill process on subsidence and filtration 405 amount of treatment required to purify Ohio river water 430 carbonic acid, influence of 395 conclusions in regard to, at the close of the i n vestigations of the Mark and Brownell de vices 313 at the close of the entire in vestigations . . . 405 cost of, annual average, to purify Ohio river water 436 cost of, comparative 436 current density, influence of 402 decomposition of the positive plates. . 310, 396 deposition on the negative plnles 398 effect of allowing the electrodes to remain out of service ... 404 efficiency of the coagulant 413-416 electric current, direction of, influence o u reversing the 403 electric resistance of films of oxide 404 form in which the iron leaves the cell 396 plales. . . 395 hydrate, rate and uniformity of formation of available 398 hydrogen, influence of 396 limitations of 397 metal wasted 396, 404 oxygen, influence of 395 passivity of electrodes 31)3 potential, influence of 399 power wasted 404 river water, influence of the composition of 403 solubility of initial iron compounds, influ ence of 396 see also Devices of Mark and Brownell, and of Water Company, and Electrolytic preparation of coagulants. English fillers, see Filters, English. European filters, s,-c Filters, European. Filter, set the various systems and devices. Filtered water exits, see Strainer system. Filters, American type, efficiency and cost of 10-1 1 general description of 10, 70 relative adaptability of, for purification of Ohio river water (39 types tested 70 English type of 5-9 relative adaptability of, for puri fication of Ohio river water. . 439 Filtration, effect of period of coagulation on 419-421 use of electrolytically prepared co agulants on 404 INDEX. 457 Filtration, conditions for successful, by American filters, 422-426 degree of coagulation for 422 minimum amount of coagulant required for, of Ohio river water by American filters. . . 430 principals of sand, without coagulants 7 rate of 254.264, 274 required for purification of Ohio river " ter 441 special devices for regulating the rate of. ... iu6 Freshets in the Ohio river, effect of, on the composition of the water 15 records of (table) 16-17 Gases, in purified Ohio river wate Germicidal action of coagulants. . Listing constituents, its of, in Oh water 33-35 water supplies of various cities... 433 cost of removal of 434 increase of, due to the use of sulphates as coagulants. . 54, 432 in purified Ohio river water. ..54, 432 Investigations, outline of entire 1-2 Ions, see F.lectrolysis. Iron, application of. with Jewell system 46 compounds, influence of the solubility of the unoxidizcd, on the electrolytic prepara tion of iron hydrate 30,6 on their use as coagulants 381 metallic, by the Anderson process, applicability of, in the pnrilication of Ohio river water. . . 382 persulphate of, action of, as a coagulant 381 composition of lot of, used by the Water Company in 1897 337 cost of, annual average estimated, for the purification of Ohio river water 437 effect of, on the quality of the Jewell system, see System, Jewell. MacDougall system, see System, MacDougall polarite. Magnets, Electro-, effect of, on the rate of electrolytic formation of aluminum hydrate 296 s, e also Devices, Harris Company ; and System, Harris mag neto-electric. Manganese, permanganates of lime and potash, action of, as coagulants 382 cost of 382 effect of 382 Microscopical analyses of effluents (tables) 127 Ohio river water (table) .. 36-37 Ohio river water, see Water, Ohio river. Operation, manner of, of Warren, Jewell, Western gravity and Western pressure systems... 96-105 records of, .< the several systems and devices. Organic matter, amounts of, see Chemical analyses. effect of electricity on 292 removal of, by Warren, Jewell, Western gravity and Western pressure systems . . 223-227 Oxygen consumed, see Carbonaceous matter. influence of, on the electrolytic preparation of iron hydrate 392 in Ohio river water 33 after purification by Warren, Jewell, Western gravity and Western pressure systems 243 influence of the use of electrolytically prepared coagulants on the amount of, after purifica- t 43" purified water 381,432 efficiency of, as a coagulant. . 413-416 Passivity, see Electrodes, passivity of. protosulphate of, action of, as a coagulant 381 i Periods, outline of the twenty, into which the in- composition of lot of, used by vestigations from Oct. 1895 to Aug. 1896 were the Water Company, in 1897. 337 divided 216-219 INDEX. Permanganates, see Manganese. Persulphate of iron, see Iron, persulphate of. Polarite, see System, MacDougall polarite. Potash alum, see Alum, potash. Pressure, see Head. Prodigiosus, B., passage of, through Warren, Jewell, Western gravity and Western pressure filters 2.49-250 Protosulphate of iron, see Iron, prolosulphale of. Purification of water, outline of history of 5 see also Coagulation, Filtration, Sedimentation, and the several systems and devices. R Rate of filtration, see Filtration, rate of. Regulators, see the several systems and devices Repairs, see the several systems and devices. Resistance, electric, see Electric resistance. Ruhmkorff coil, see System, Harris magneto-electric. S Samples, manner of collection of 18, 445 Sand layer, area of surface of, of Jewell filter 79 Warren filter 74 Western gravity fil ter 83, 84 Western pressure fil ter 87 character of, for successful filtration of Ohio river water by American filters 426, 442 cleaning, devices for 274, 443 Jewell system 79 Warren system 74 Western gravity sys- tei Western pressure sys tem 88 influence of. on cost of purifica tion 264 quality of effluent 256 manner of operation of, 98, 101, 103, 104, 321, 323, 343 water used, kind of 256 kind of sand used, influence of 253, 262 Jewell filter 79 Warren filter 73 Western gravity filter, 82. 84 Western pressure filter 87 location of 273, 371 thickness of, influence of 253, 262 Jewell filter 79 Warren filter 73 Western gravity filter 82, 84 Western pressure filter 87 Sedimentation, necessity of, with Ohio river water. . . 383 plain, bacteria in city lap water (table). 69 purification of Ohio river water by 252, 261, 440 -Sedimentation, with coagulation, effect of use of electro- lytically prepared hydrate of iron on. . 414 coagulants to aid 416-419 provisions for, 57-64, 9 , 335 influence of, 252, 261 process of 57 purification of Ohio river water by, 58, 65-69, 78, 271, 367, 440 Soap, amount required with purified Ohio river water.. 432 Soda, CHUstic, composition of lot of, used by the Water Company in 1897 33 8 to clean sand layer 426, 443 remove incrusting constituents 433 Spark drum, see System, Harris magneto-electric. Storage of purified Ohio river water 427, 443 Strainer system of 4-I 2 Jewell filter. 78 Warren filter 7 2 Western gravity filter 82, 83 Western pressure filter 86 Sulphate of alumina, see Alumina, sulphate of. Sulphates of lime and magnesia, see Incrusting con stituents. Summaries, of results of investigations from Oct. 1895 to Aug. 1896, explanations of 215-221 tables of 222-241 Suspended matter in water for successful filtration 422 partial protective influence of, against corrosion of iron 428 see also Chemical analyses. System, Harris magneto-electric, bacterial analyses in connection with 279 chemical analyses in connection with 279 description of, general 276 electric circuits 278 electric generating appliances 278 efficiency of. . . . 290 electrodes 277 electrolytic cells 276 magnets, electro 277 operation of 278 spark drum 276 status of, July I, 1896 281 System, Jewell, alumina, sulphate of, amounts of (table), by days 50 by runs 196-204 composition of lots of 41 devices for application of. ... 42-46 solutions of, strength of (table) 47 appurtenances of, inventory of 9-93 attention given to 107 chlorine, application of t . . . 40 efficiency of, bacterial, average by days... .228-231 runs. . ..196-204 removal of organic matter 223-227 INDEX. 459 System, Jewell ConcluJeii. effluent of, analyses of, bacterial 148-169 chemical 117-122 microscopical 127 mineral 127 bacteria in, see bacterial efficiency, clearness of, average daily degree of .... 222 filter of, agitator So controller, automatic 79 description of, general . 77 filter tank 77 prodigiosus B., application of, tn. . . . 249 sand layer, area of surface of 79 bacteria in 257 device for cleaning 79 kinds of sand 79 nitrogen as albuminoid ammonia in 257 thickness of 79 strainer system 78 iron, application of metallic 46 lime, amounts used, average by clays 52 devices for the application of 45 solutions of, strengths of 52 operation of, delays of 94 manner of 99 records of, description of tables., no summaries of, 234, 240, 241 tables of 196-204 regulating devices, special 105 repairs of 93 sedimentation with coagulation, devices for.. . . 59 purification by 61 time occupied by 95 System, MacDougall polarite, alumina, sulphate of, application of. . bacterial analyses in connection with. . bacterial results, summary of, by days, chemical analyses in connection with. . see also Water, Ohio river. chemical results, summary of, by days 328 clay extractor 319 conclusions in regard to 327 description of, general 318 efficiency of, summary of, by days 328 effluent, quality of 323 iron tank with baffle plates 318 operation of 32 1-326 polarite, analysis of 322 polarite filter, description of 320 effect of air pocketing in. . . filtering materials of modifications of System, Warren, alumina, sulphate of, amounts of (tables), by day: by runs.. 187-196 composition of lots of 41 devices for the application of. ... 42 solutions of, strength of (tables). 48 321 322 321 " System, Warren- Concluded. appurtenances of, inventory of 9 ^~93 attention given to 107 efficiency of, bacterial, average by days . . 228-231 runs. . 187-196 removal of organic matter. . 223-227 effluent of, analyses of, bacterial 128-148 chemical 112-1 17 microscopical 127 mineral 127 bacteria in, see bacterial efficiency, clearness of, average daily de gree of 22: filter of, agitator 75 description of, general 71 filter tank. 71 filtered water chamber 73 prodigiosus R., application of. to .... 249 sand layer, area of surface of 74 bacteria in 257 device for cleaning 74 kinds of sand 73 nitrogen as albuminoid ammonia in 257 thickness of 73 strainer system 72 weir 73 operation of, delays of 94 manner of 96 records of, description of tables no summaries of, 232, 240, 241 tables of 187-196 regulating devices, special 105 repairs of 93 sedimentation with coagulation, devices for... 58 purification by 61 time occupied by 95 System, Western (parts common to both), appurtenances, inventory of 90-93 attention given to 107 chemicals used 41 amounts of potash alum used. . 51 sulphate of alumina used (table) 51 solutions of, strength of (table) 49 sedimentation with coagulation, devices for. , . 60 purification by 63 System, Western gravity, appurtenances of, inventory of 90-93 chemicals used, average amounts of, by runs, 204-207 efficiency of, bacterial, average by days.. 228-231 average by runs. . . 204-207 removal of organic matter... 223-227 effluent of, analyses of, bacterial "9-174 chemical 122 microscopical 127 mineral 127 effluent of, bacteria in, sre bacterial efficiency, clearness of, average daily degree of... . . 222 INDEX. System, Western gravity Concluded. filter of, description of filters (A) and (H) Si prodigiosus B., application of, to .... 249 filter (A), description of 82 filter tank . , 82 sand layer, area of surface of 83 bacteria in 257 device for cleaning 83 kinds of sand 83 nitrogen as albuminoid ammonia in 257 thickness of 82 strainer system 82 filter (B), description of 83 filter tank 83 sand layer, area of surface of 84 bacteria in 257 device for cleaning 84 kinds of sand 84 nitrogen as albuminoid ammonia in 257 thickness of 84 strainer system 84 operation of, delays of 94 manner of 102, 103 records of, description of tables., no summaries of, 236, 240, 241 tables of 204-207 regulating devices, special 105 repairs of 93 time occupied by 95 see also System, Western (parts common to both). System, Western pressure, appurtenances, inventory of 9~93 chemicals used, average amounts of, by runs, 207-214 efficiency of, bacterial, average by days.. . 228-231 runs... 207-214 removal of organic matter . . 223-227 effluent of, analyses of, bacterial 174-186 chemical 124-126 microscopical 127 mineral 127 bacteria in, see bacterial efficiency, clearness of, average daily degree of. 222 filter of, description of, general 85 filter chamber 86 prodigiosus B., application of, to,. . 249 sand layer, area of surface of 87 bacteria in 257 device for cleaning 88 kinds of sand 87 nitrogen as albuminoid ammonia in 257 thickness of 87 strainer system 86 operation of, delays of 94 manner of 104 records of, description of tables, no System, Western pressure Concluded. operation of, records of, summaries of, 238,240,241 tables of 207-214 regulating devices, special 105 repairs of 93 time occupied by 95 Systems of purification investigated 1-4 Tables, see specific topic. Tap water, city, bacteria in (table) 69 Tests, from Oct. 1895 to Aug. 1896, conditions of n outline of I Time occupied in various ways by the Warren, Jewell, Western gravity and Western pressure systems... . 93-95 Totals, grand, for the investigations from Oct. 1895 to Aug. 1896 240 U Warren system, see System, Warren. Water, Ohio river, after purification, appearance of 242, 426 adaptability for boiler use, 245, 429,432 biological character of 248, 427 carbonic acid in .... 53, 244, 431-434 color of 243 corrosion of iron by 244, 43I~434 lead by 434 mineral matter in 426 odor of 243, 426 organic matter in 243, 426 oxygen, dissolved, in 243, 427 soap required 432 storage of 427, 443 taste of 243, 426 undecomposed coagulants in 427 uniformity of quality of 429 analyses of, bacterial 39 chemical 21-31 microscopical 36 mineral 34 application of coagulation to 57 character of, general 15 biological, 16, 36-39 chemical, 16, 20-34 factors affect ing the 15 physical 1 8 composition of, bacteria in, numbers of, see bacterial analyses, species of 36 carbonaceous matter 223-227 influence of, on the electrolytic prep aration of coag ulants 403, 410 IhDEX. 461 Water, Ohio river, afier purification Concluded. composition of, nitrogen as albu minoid ammonia 223-227 factors which influenced the quality of, after purification, 250-257, 405, 43 -434 purification of, accessibility of parts of system 443 applicability of meth ods investigated, 271, 439 attention required, 258, 2(^7, 443 coagulant, amount, 273, 430, 441 kind 440 imperative ness of use of. . 439 point of ap plication Water, Ohio river, purification of, cost of, coagulant for, 4S5--437 elements of.. . 268 factors which affected the. 267 filters for 442 filtration 422-425.441 sand layer for. . .426, 442 cleaning of. .424, 443 location of 442 sedimentation, plain, 359. 37<> -378, 44 with co agula tion ... 440 Weir used in ihe Warren system 73 ;., A I II II I II PLATE V. J SECTION OF JEWELL GRAVITY SYSTEM." UNIVERSITY JJLtAiifo*^ "". </ f"3 I C i r^orS^*^ UNIVERSITY Of ( PLATE VJII. PLAN SECTION ON A- B FLAN SECTION ON A-B .... WARREN FILTER JEWELL FILTER :_ PLAN - PLAN SECTION ON A- B SECT/ OH OH A- B WESTERN PRESSURE FILTER WESTERN GRAVITY FILTER "TYPICAL AREAS OF STRAINER FLOORS. * >? ? v&*^ Pi*- ! w . -S. Iw , %vi#. .^^.. . ?.,. %-M, -P* 3 ? , S-M- -^r^^isa, - & t^, sf C5^^tV jV^^ "S- ef J VXcOT jvjjf T*^^^ ^^%&v^^fe- ^% "fi^i^ ^ 9% ^5^T & . 7 * ^"^C ^, ** r/ ^-^ J*- T r ^ ^4^1*^^ ^tPi l^ *fo-i RETURN CIRCULATION DEPARTMENT ^T0^ 202 Main Library v ALL BOOKS MAY BE RECALLED AFTER 7 DAYS 1 LOAN PERIOD 1 HOME USE 2 3 4 1 5 6 u^iiPIW 1 f "^ ^^ 4 ^F- : *ylf*^* &^~*?* . ""^-.-" 1 -month loans may be renewed by calling 642-3405 .. . _ . ite* ^V>1 ^ , toT4W\^ ^fe,< 6-month loans may be recharged by bringing books to Circulation Desk Renewals and recharges may be made 4 days prior to due date ~ apj^*rP ^^v v -fet,^> - r^"^ "*1\ wTjft^ ^"r ^rCJT ^ r * re ) *..]<. DUE AS STAMPED BELOW 4 *" Wr - . .W /e ; /w Jj u^0.1^1^ v^/ j 1 i - A -X . " ^ ffi... ^ M ffrwSSi -^^ . f & FORM NO. DD 6, UNIVERSITY OF CALIFORNIA, BERKELEY BERKELEY, CA 94720 i^.ft ^^jW ?5*S?^fc^ . * &^rO J?*?** , ^ ^--W ^.^"> , *^%1, -^S^ - ^^ i^W L.Tm * 4^ * >\^ " y-7Ty A*. *~ ">r 57 *!*- " *ii t*? ~ - xai/ * t T a rfvp^wt S* j;. * $* tf^pT^ rt& - ~*2\ V, .f^||-. -<\ <Vv&^ ^ ,v j^, ^^A 7 . -^ <x. >54*.,- wl ., -^/Pt" -y ^ftss^ -w ..VH ;t K. _ . . UNIVERSITY OF CALIFORNIA LIBRARY ^ e ^"^C ^ "*> ^-4^ ^- ^ -^-^4 K^^i& ,:V P^l