QD UC-NRLF $D 3b 700 THE LIBRARY OF THE UNIVERSITY OF CALIFORNIA DAVIS GIFT OF ROBERT I. TEKMEy Indostrial Cbemical Inslitule of Digitized by the Internet Archive in 2007 with funding from Microsoft Corporation http://www.archiye.org/details/examiwaterOOIeffrich Examination of Water ^ SANITARY AND TECHNIC PURPOSES BY CHEMICAL AND BACTERIOLOGIC METHODS. BY HENRY LEFFMANN, A.M., M.D., Ph.D., PROFESSOR OF CHEMISTRY IN THE WOMAn's MEDICAL COLLEGE OF PENNSYL- VANIA AND IN THE WAGNER FREE INSTITUTE OF SCIENCE; PRESI- DENT (19O1) OF THE engineers' CLUB OF PHILADELPHIA; AND VICE-PRESIDENT (190I-C2) BRITISH SOCIETY OF PUBLIC ANALYSTS. SIXTH EDITION, REVISED AND ENLARGED, WITH ILL USTRA TIONS. .<" PHILADELPHIA BLAKISTON'S SONS & CO. IOI2 WALNUT STREET 1909 UttlVliJbC::iITY of CAUFftSaia PAvia Copyright, 1909, by P. Blakiston's Son & Co. . F. FELL COMPANY PRINTERS PHILADELPHIA Wl^T^ I DEDICATE THIS BOOK TO THE MEMORY OF /iDi? /iDotber, TO WHOSE WISE PRECEPT AND EXAMPLE IN MY BABYHOOD I OWE WHATEVER MERIT MY MANHOOD YEARS MAY SHOW. PREFACE In the present edition, numerous revisions have been made, but the plan of the book has not been disturbed. Among the additions to this edition are: the precipitation (nitron) method for ni- trates ; the bile-lactose method for the colon bacil- lus ; a description of the method devised by Jack- son and Melia for detecting the typhoid bacillus ; notice of a special reagent for detecting excess of aluminum compounds. In the five editions which have been issued, the book has seen notable changes in the attitude of experts toward certain methods. When the first edition was issued (jointly with Dr. William Beam) , bacteriologists were enthusiastically claim- ing to be able to determine easily the presence or absence of disease-producing microbes in water, and asserting that sanitary analysis of water by chemical methods was about to become a mat- ter of history. At present, the value of routine chemical analysis is generally recognized. Upon this point, and also on the question of the inherent danger of unfiltered surface water even when not PREFACE. receiving sewage directly, the book took decided stand, and the development of the views of ex- perts has fully justified it. Long experience has convinced me that for determining the potability of water, the determinations of chlorin, nitrates, and nitrites afford the most satisfactory indica- tions, and that the figures •for nitrogen or ammo- nium (so-called ''free ammonia") and nitrogen by permanganate (so-called ''albuminoid ammonia") are of much less value than is generally supposed. The suggestion of Woodman, that phosphates afford a useful datum, is worthy of special atten- tion. 1839 N. 17th Street, Philadelphia, January, 1909. CONTENTS. PAGES Natural History and Classification of Natural Waters. Rain Water — Surface Water — Subsoil Water — Deep Water, 1-7 Analytic Operations. Sanitary Examinations: Collection and Preliminary Examination — Total Solids — Chlorin — ^Nitrogen in Ammonium Com- pounds and Organic Matter — ^Nitrogen as Ni- trates — Nitrogen as Nitrites — Oxygen-consuming Power — Phosphates — Dissolved Oxygen — Pois- onous Metals — Biologic Examinations 8-90 Technic Examinations: General Quantitative Analysis — Spectroscopic An- alysis — Specific Gravity, 91-113 Interpretation of Results. Statement of Analysis — Sanitary Applications — Action of Water on Lead — Technic Applica- tions — Boiler Waters — Sewage Effluents — Puri- fication of Water — Identification of Source of Water, 1 14-141 Data for Calculation 142, 143 Index. NATURAL HISTORY AND CLASSIFICA- TION OF WATER. Pure water is an artificial product. Natural waters always contain foreign matters in solution and suspension, varying from mere traces to very large proportions. The properties, effects, and uses of water are considerably modified by these ingredients, and the object of analysis is to as- certain their character and amount. Since these are largely dependent on the history of the water, a classification based on this will be convenient. We may distinguish four classes of natural waters : Rain Water. — Water precipitated from the atmosphere under any conditions, and therefore including dew, frost, snow, and hail. Surface Water. — All collections of water in free contact with the atmosphere, as in streams, seas, lakes, or ponds. Subsoil or Ground Water. — Water not in free contact with the atmosphere, percolating or flowing thru soil or rock at moderate distance be- low the surface, and derived in large part from the rain or surface water of the district. Deep or Artesian Water. — Water accumulated 2 HISTORY AND CLASSIFICATION. at considerable depth below the surface, from which the subsoil water of the district has been excluded by difficultly permeable strata. Rain water, when gathered in the open country and in the later period of a prolonged rain or snow, is the purest form of natural water. When collected directly, it contains but little solid matter, this consisting principally of ammonium compounds and particles of organic matter, living and dead, gathered from the atmosphere. In districts near the sea an appreciable amount of chlorids will be present. It is obvious that a pro- longed rain will wash out the air, but since storms are usually attended by wind, fresh portions of air are continually flowing in, and thus the water never becomes perfectly pure. Rain water col- lected in inhabited districts is usually quite impure. Surface Water. — Rain water in part flows off on the surface, and gains in the proportion of sus- pended and dissolved matters, the former being found in large amount when the rainfall is profuse. The wearing action of water is dependent on the amount and character of these suspended mater- ials. From the higher levels of a watershed, the streams, more or less in the form of torrents, gather into larger currents, and, reaching lower levels, become slower in movement, and deposit much of the suspended matter. By admixture of the waters from widely separated districts the SUBSOIL WATER. 3 character and amount of the dissolved matters are much modified. It is obviously impossible to establish close standards of composition for surface waters. In the case of rain water, falling on the surface of un- disturbed, unpopulated territory, the amount of solids dissolved will be small, and will consist prin- cipally of carbonates and sulfates. The water of lakes and rivers is, however, in part derived from springs, which may proceed from great depths, and thus introduce substances not easily soluble in surface water, nor derivable from the soil of the district. The exposure to light and air which surface water undergoes results in the absorption of oxy- gen and loss of carbonic acid, together with the oxidation of the organic matter. The diminution of the rapidity of the current permits the deposi- tion of the suspended matters, and this occurs especially as the river approaches the sea, not only from the retarding influence of the tidal wave, but from the precipitating action of the salt water. Subsoil Water. — Water that penetrates the soil passes to different depths, according to the porosity and arrangement of the strata. As a rule, it descends until it reaches formations but slightly pervious, upon the level of which it accumulates. In the upper layers of soil it dissolves mineral and organic ingredients, and becomes impregnated 4 HISTORY AND CLASSIFICATION. with minute organisms, thru the agency of which the organic matter undergoes important trans- formations. The water, constantly accumulating, gradually flows along the incline of the impervious stratum, or thru its fissures, and may either pass downward or emerge in the form of a spring. Much difference is observed in the composition of subsoil waters, but as a general rule they con- tain small amounts of mineral substances and organic matter. In populated districts, however, a marked change is produced through admixture with water containing animal and vegetable prod- ucts in various stages of decomposition. It is probably the organic matters containing nitrogen that are of importance. These are mostly un- stable, and decompose, partly by oxidation, partly by splitting up into simpler forms — changes in most cases brought about by microbes. The nitrogen is in part converted into ammonium compounds, but a considerable portion suffers further oxidation, and in association with the mineral substances present forms nitrites and nitrates, especially the latter. This is called *' nitrification.*' Nitrification takes place under the influence of microbes, the habitat of which does not extend very far below the surface of the soil. Several forms with active nitrifying powers have been isolated and described. The nitrifying action is DEEP WATER. 5 often exerted upon the ammonium compounds formed from the organic matter. The presence of some substance capable of neutraHzing acids is usually necessary to continuous action. Calcium and magnesium carbonates fulfil this function. Nitrates are the final result of this action ; nitrites are present at any given time only in small quan- tity. Denitrification — that is, the reduction of nitrates and nitrites to ammonium compounds — takes place also under the influence of microbes, and is especially apt to occur when considerable quantities of decomposing organic matter are in- troduced. Several species of denitrifying bacilli have been described. A partial reduction some- times occurs, and a notable proportion of nitrites is found, but in the presence of actively decompos- ing organic matter, such as that in sewage, a com- plete reduction, even to the liberation of nitrogen, may occur. Deep Water. — Water which penetrates the fissures of the fundamental rock-formations may pass to great depths, and by following the lines of the lowest and least permeable strata, may be transported to points far removed from those at which it was originally collected. The chemical changes thus induced include most of those which take place at higher points, but the increase of pressure and temperature confers increased solvent power. Carbonic acid will accumulate and pro- 6 HISTORY AND CLASSIFICATION. duce conditions favorable to the solution of cal- cium, magnesium, and iron carbonates; iron and manganese oxids may be converted into car- bonates and then dissolved. Sulfates are reduced to sulfids, and these subsequently, by the action of carbonic acid, yield hydrogen sulfid. Organic matter, living and dead, plays an important part, determining the reduction of ferric compounds to ferrous, and of the sulfates to sulfids, and is itself converted ultimately into ammonium compounds, notable quantities of which are often found in deep waters. Further, it is found that nitrates and nitrites are present only in small amount, except from certain strata rich in organic matter. In some cases the water acquires very high tempera- ture, and dissociation of rocks occurs with solution of considerable amounts of silicic acid, which is ordinarily but sparingly soluble in water. Masses of water thus accumulated under heat and pressure may find their way to the surface either thru natural fissures or be reached by borings. While X J absolute line can be drawn between deep and subsoil waters, yet it will, in most cases, be found that the deep water of a given district, whether obtained thru natural or artificial chan- nels, will be decidedly different in composition from the subsoil or surface water of the same, and that the rocks overlying the veins of water will DEEP WATER. 7 contain one or more strata, difficultly permeable to water, and therefore preventing direct communi- cation. The characteristic differences between surface, subsoil, and deep waters are given in a table under the section on interpretation of re- sults. ANALYTIC OPERATIONS. SANITARY EXAMINATIONS. COLLECTION AND PRELIMINARY EX- AMINATION OF SAMPLES. Great care must be taken in collecting water samples, in order to secure a fair representation of the supply and to avoid introduction of foreign matters. The five-pint green glass-stoppered bottles used for holding acids are suitable for con- taining the samples. The contents of one such bottle will suffice for most sanitary or technic examinations. Boxed bottles which are furnished by dealers are convenient for transportation. They are usually provided with a hinged lid which can be fastened, if deemed necessary, by a pad- lock. The green glass-stoppered bottles may be fitted in such an arrangement. Crated demijohns are also made for forwarding water. The larger sizes are well adapted for samples which are to be subjected to elaborate analysis. Stone jugs, casks, or metal vessels should not be employed. All bottles must be well rinsed several times with the water to be examined, filled, and the stopper tied 8 SANITARY EXAMINATIONS. 9 down or fastened by stretching a rubber finger-cot over the stopper and Hp. If corks are used, they should be new and well rinsed. Wax, putty, plaster, or similar material should not be used. It is not necessary to sterilize the bottles when only chemical tests are to be made. In taking samples from lakes, slow streams, or reservoirs, it is necessary to submerge the bottle so as to avoid collecting any water that has been in immediate contact with the air. In the ex- amination of public water-supplies, the sample should be drawn from a hydrant in direct con- nection with the main, and not from a cistern, storage -tank, or dead end of a pipe. In the case of pump- wells, a few gallons of water should be pumped out before taking the sample, in order to remove that which has been standing in the pipe. In all cases, care should be taken to fill the vessel with as little agitation with air as possible. It is important that with each sample a record be made of those surroundings and conditions which might influence the character of the water, particularly in reference to sources of pollution, such as proximity to cesspools, sewers, or manu- facturing establishments. The character and con- dition of the different strata of the locality should be noted if possible. Determinations of nitrogen existing as ammon- ium compounds and as organic matter, and of lO ANALYTIC OPERATIONS. oxygen-consuming power, should be made upon the sample in the original condition, whether tur- bid or clear, but all other estimations should be made upon the clear liquid. Tur- bid waters may be clarified by standing or by filtration; for the latter purpose Schleicher & Schull's extra heavy 598 paper is the best. In many cases the suspended matter can not be entirely removed by filtration, and subsidence must be resorted to. The use of a small quantity of alum, or aluminum sul- fate, as now applied in the purifica- tion of drinking water, will some- times be satisfactory as a means of clarifying samples. For the quan- titative determination, the sedi- ment from a known volume of the water is collected on a tared filter, dried, and weighed. The water from newly dug wells is generally turbid, and the deter- minations are best made after filtration; but the results will be unsatisfactory, showing a higher proportion of organic matter than will be found when the supply becomes clear. For taking samples at considerable depths the lI'liiiKNTZSiSONS Fig. I. I SANITARY EXAMINATIONS. II bottle shown in Fig. i will answer, but samples so- collected will not serve for determination of dis- solved gases. Collection of Samples for Bacteriologic Exam- ination. — Bacteriologic examinations are of little value unless made promptly on samples that have been collected with precautions against contam- ination. The inoculation of the culture-medium is best done at the source. If this is not possible, glass-stoppered bottles holding about 200 c.c, which have been thoroughly sterilized, with stop- pers in place, in a hot-air oven at 150° C, must be used for collection. They should be rinsed on the outside with the water, dipped below the surface, the stopper withdrawn, and again inserted when the bottle is full. If these are to be transported any distance, they should be packed in ice. For the collection of samples below the surface of the water, the bottle shown in Fig. i is recommended by Abbott. The bottle having been previously thoroughly sterilized, is sunk to the proper depth, and the stopper is then lifted by a special cord and held until the bottle is full, when, the cord being released, the stopper falls. Before taking out portions for test the lip and stopper must be well sterilized by strong alcohol and by care- ful heating, and, after cooling, washing with sterilized water. Color. — A colorless glass tube, two feet long and 12 ANALYTIC OPERATIONS. two inches in diameter, is closed at each end with a disc of colorless glass. An opening for filling and emptying the tube should be made at one end, either by cutting a small segment off the glass disc, or cutting out a small segmental section of the tube itself before the disc is cemented on. A good cement for such purposes is the following : Caoutchouc, 2 parts. Mastic, 6 Chloroform, 100 '' The ingredients are mixed and allowed to stand for a few days. The cement should be used as soon as solution is effected, as it becomes viscid on standing. The tube must be about half filled with the water to be examined, brought into a horizontal position, level with the eye, and directed toward a brightly illuminated white surface. The com- parison of tint has to be made between the lower half of the tube containing the water under ex- amination and the upper half containing air only. A more convenient form of tube is made by attaching brass screw-nipples to each end of the tube, and closing these by screw-caps carrying plate-glass discs. Such tubes can be obtained from dealers in chemical apparatus. It is obvious that various methods of comparing color and turbidity may be devised, but data so obtained are I SANITARY EXAMINATIONS. 13 of little analytic value, and even that little is limited to samples closely analogous in character. Hazen has devised a standard for color com- parison which he claims as capable of most satis- factory use on all ordinary waters. It is based upon the modification of a solution of platinum chlorid by a solution of cobalt chlorid, as follows : 1.246 grams of potassium platinum chlorid (corresponding to 0.5 gram of platinum) and i gram of cobalt chlorid (corresponding to 0.25 gram of cobalt) are dissolved in water, 100 c.c. of strong hydrochloric acid added, and the solution made up to 1000 c.c. It keeps well, even when exposed to the light. For comparison, 1,2,3, ^"^c., of the stock solution are diluted to 50 c.c. in Nessler tubes. These correspond to o.i, 0.2, 0.3, etc., degrees of the color standard. These also keep for a long time if protected from dust. Direct comparison in 200 mm. tubes is generally sufficient. If the shade of color is not exactly that of the water, more cobalt may be added, the platinum being constant. Hazen expresses the result in any case in terms of ''the amount of platinum in parts per 10,000, which in acid solu- tion with so much cobalt as will match the hue, produces an equal color in distilled water." Lovibond's tintometer is probably the best means of making color comparisons. Odon — Put about 150 c.c. of the water into a 14 ANALYTIC OPERATIONS. clean, wide-mouthed 250 c.c. stoppered bottle, which has been previously rinsed with the same water ; insert the stopper and warm the water in a water-bath to 100° F. Remove the bottle from the water-bath and shake it rapidly for a few seconds; remove the stopper, and immediately note if the water has any smell. Insert the stopper and repeat the test. In a polluted water the odor will sometimes give a clue to the origin of the pollution. Turbidity. — Several methods for expressing de- gree of turbidity have been used. Whipple and Jackson, after comparing these, find that finely powdered diatomaceous earth is satisfactory. The material is ignited, ground to a powder that will pass thru a 200-mesh sieve, dried at 100° C, cooled in a desiccator, and kept in a well-stoppered bottle. A strong standard is prepared by adding I gram of this powder to 1000 c.c. of water, and 12 dilute standards by mixing quantities of the strong standard in amount from i to 10 c.c, increasing by 0.5 c.c, to quantities of distilled water sufficient to make 100 c.c in each case. The dilute stand- ards are kept in tightly corked tubes and shaken several times when used for comparison. If new corks are used, they should be well boiled in water to extract coloring-matter. If the sample is of high turbidity, it must be diluted by a known volume of water. The record is made by noting SANITARY EXAMINATIONS. IS the strength of the standard tube that is nearest in turbidity to the sample, the tubes being held together and viewed from the side. The tube con- taining the sample must be uniform in size and quality with those containing the standard. Reaction. — Many indicators have been proposed for testing reaction. As water always contains carbonic acid or acid carbonates, tests may be misleading, unless made both on the cold water and on a sample immediately after thoro boiling. I have had good results with sodium alizarin- monosulfonate. The solution may be made by dissolving one gram in loo c.c. of water. It keeps well. For use a few drops are added to a con- venient volume of the water, and decinormal acid (or alkali as required) added until the color change is noted. The indicator is yellow when acid, red when alkaline, brownish when neutral. A special method of determining reaction has been recommended by the laboratory committee of the American Public Health Association. The indicator is made by dissolving o.i gram of sodium erythrosin in looo c.c. of water, loo c.c. of the sample is put into a 250 c.c. fiask, and 2.5 c.c. of the indicator solution added with 5 c.c. of chloroform. Decinormal acid is added drop by drop, shaking after each addition. The disappearance of the color of the reagent indicates that the alkalinity has been overcome. i6 ANALYTIC OPERATIONS. Other indicators are used in special cases, which will be noted in the proper connection. Most indicators are freely soluble in water, the solu- tion being made by dissolving about i gram in loo c.c. Phenolphthalein must be dissolved in alcohol in the above proportion. TOTAL SOLIDS. A platinum basin holding loo c.c. will be found convenient for this determination. This will weigh about 45 grams. It should be kept clean and smooth by frequent burnishing with sand, a little of which should be placed in the palm of the hand, moistened, and the dish gently rubbed against it. Very fine sea-sand with round, smooth grains is the only kind suitable for this purpose . Coarse river sand, tripoli, or other rough scouring powders must not be employed. If proper care is taken, the luster of the metal will be retained, and the loss in weight will be trifling. The inner surface can generally be cleaned by treatment with hydrochloric acid, rinsing, and, if necessary, burnishing with sand. Neglect of these precau- FlG. 2. SANITARY EXAMINATIONS. 1 7 tions will soon lead to serious damage to the dish. A small, smooth slab of iron or marble is con- venient to set it on while cooling. When being heated over the naked flame, the dish should rest on a triangle of iron wire, covered with pipe- stems. Dishes of pure nickel are not satisfactory substitutes for those of platinum. Platinum-pointed forceps should be used in handling the dish. The platinum terminals may be kept bright and clean by the use of sand. The low-temperature burner, used as shown in figure 2, will be found a very convenient substitute for the water-bath and hot-air oven. The inlet pipe is very short and soon becomes so hot as to injure the rubber tube. To avoid this it may be lengthened by means of a piece of gas-pipe about 12 centimeters long, or the junction may be wrapped with a rag, the ends of which dip into water. By capillary attraction the rag is kept moist and cool. The determination of total solids is made by evaporating loo c.c. of the water in the platinum basin, which has been previously heated almost to redness, allowed to cool for ten minutes, and weighed. The operation is conducted at a moder- ate heat. When the residue appears dry, the heat may be increased slightly for some minutes. The above method will answer in most cases. In waters of l8 ANALYTIC OPERATIONS. exceptional purity it may be advisable to use larger quantities, such as 250 c.c. When the residue contains deliquescent bodies, the deter- mination will not be accurate, and when appre- ciable amounts of magnesium and chlorin are present, a decomposition will occur by which magnesium oxid will be formed and hydrogen chlorid escape. Deliquescence of residues and decomposition during evaporation may be largely prevented by adding 0.005 gram of sodium carbonate to each 100 c.c. of the sample taken. This converts magnesium and calcium salts into carbonates. The sodium carbonate is conveniently kept in the form of solution of such strength that i c.c. con- tains 0.00 1 gram. The weight of the carbonate is, of course, to be deducted from the weight of the residue. Drown and Hazen have carefully in- vestigated this method and have found it available for a more satisfactory determination of the loss on ignition. For this process a thin-walled platinum basin is placed within another similar basin of such size that an air-space of about one-half of an inch is left all around the inner dish, which is supported upon a spiral of platinum that rests on the bottom of the outer dish. Over the inner dish is sus- pended a disc of platinum foil to reflect the heat. The outer dish is heated to bright redness. After the weight of the residue is obtained, SANITARY EXAMINATIONS. I9 many chemists heat the dish cautiously to low redness, and note the effect. Nitrates and ni- trites, calcium and magnesium carbonates, are de- composed; ammonium salts are driven off; potassium and sodium chlorids are also driven off if the temperature is high. Organic matter is at first charred, and by continued heating burned off. When the quantity of nitrates is considerable, slight deflagration may be observed, or the pro- duction of red fumes of nitrogen dioxid. The organic matter, in decomposing, not infrequently develops odors which indicate its character or source. These are more satisfactorily observed when a rather large quantity, say 250 c.c, is evap- orated at a low heat, preferably on a water-bath. In water of high organic purity, the residue on heating will give no appreciable blackening nor odor, while in forest streams charged with vege- table matter derived from falling leaves, very decided blackening without unpleasant odor will be noticed. The loss on heating can not be taken as a measure of the organic matter, except when present in relatively large amount. The difference between the weight before and after heating is often reported as ''organic and volatile matter" but it has rarely any practical value. In many cases the determination of total solids, being merely for control, may be made in glass dishes. I have found especially convenient a 20 ANALYTIC OPERATIONS. shallow dish made by cutting off a beaker about 8 centimeters in diameter, so as leave a depth of about 3 centimeters. The cutting can be easily done by the use of a hot iron, and the edge can be ground with carborundum or emery on an iron or glass plate. A dish of the above dimen- sions will hold ICG c.c. It is most safely heated by placing it on a piece of wire gauze which rests upon the flanges of the low temperature burner. The gas should be turned so as to give a ring of flames not over i centimeter high. The dish should be warmed, allowed to cool, and weighed, and ICG c.c. of the water put in. After the residue has been weighed, it may be removed by addition of a little strong hydrochloric acid with strong rubbing with a rubber-tipped rod. Of course, the determination of volatile matter by heating to redness cannot be made, but this is of no value in the ordinary sanitary and technic examinations. I have found that a dish made from a good quality of glass weighs about 25 grams and loses little by use. In connection with the determination of the total solids, two qualitative tests should be made. Separate portions of the sample should be tested with ammonium oxalate and barium chlorid. If a precipitate is produced by the latter reagent, a few drops of hydrochloric should be added. A permanent precipitate shows sulfates. If the SANITARY EXAMINATIONS. 21 precipitate disappears, it was probably carbonate. A precipitate with ammonium oxalate shows calcium compounds. As a rule, if a notable amount of either calcium or magnesium car- bonate is present, a crystalline scum forms during evaporation. The residue effervesces actively with hydrochloric acid. CHLORIN. Solutions Required : Standard Silver Nitrate. — Five grams of pure recrystallized silver nitrate are dissolved in dis- tilled water, and the solution made up to looo c.c. The amount of chlorin to which this is equivalent may be determined as follows: Several grams of pure sodium chlorid are finely powdered and heated for five minutes, not quite to redness. When cold, 0.824 gram are dissolved in water and the solution made up to 500 c.c. Twenty-five c.c. of this should be treated as below, and the amount of silver solution required noted. Each c.c. of the sodium chlorid solution is equivalent to 0.00 1 gram chlorin. Potassium Chromate. — Five grams of potassium chromate are dissolved in 100 c.c. of distilled water. Absolution of silver nitrate is added until a permanent red precipitate is produced, which is separated by filtration. 2 2 ANALYTIC OPERATIONS. Analytic Process : If a preliminary test shows the chlorin to be present in considerable amount, the determination may be made on loo c.c. of the water without con- centration. If, however, there is but little present, 250 c.c. should be evaporated to about one-fifth, best with the addition of a little sodium carbonate, and the determination made on the concentrated liquid after cooling. The water is placed in a porcelain dish or in a beaker standing on a white surface, a few drops of potassium chromate solution added, and standard silver nitrate solution run in from a buret until a faint red color of silver chromate remains per- manent on stirring. The proportion of chlorin is then calculated from the number of c.c. of silver solution added. Greater accuracy is secured by operating in yellow light. A second determination may be made, using as a comparison the liquid first titrated, the red color having been previously discharged by a few drops of sodium chlorid solu- tion. The water should always be as nearly neutral as possible before titration. If acid, it may be neutralized by the addition of sodium carbonate. The residue obtained by evaporating the water with sodium carbonate, as described in connection with the determination of the total solids, will often serve conveniently for estimating the I SANITARY EXAMINATIONS. 2^ chlorin. It is best to use 200 c.c. of the sample and redissolve the residue in about 50 c.c. of dis- tilled water, rubbing the sides of the dish well with a rubber-tipped rod, and then titrating as indicated above. NITROGEN IN AMMONIUM COMPOUNDS AND IN ORGANIC MATTER. Determinations of these data are usually carried out as one operation, known as the ** ammonia method." The originators, Wanklyn, Chapman and Smith, applied the term ''free ammonia" to the ammonium compound obtained by the distilla- tion of the sample with sodium carbonate, and the term ''albuminoid ammonia" to that obtained by distilling the sample with a strongly alkaline solu- tion of potassium permanganate (alkaline per- manganate). Both terms are somewhat objec- tionable, and in this book the former will be sub- stituted by "nitrogen as ammonium" and the latter by "nitrogen by permanganate." Apparatus Required : Distilling Apparatus. — That shown in figure 3 is convenient. The still consists of a glass retort of about 1000 c.c. capacity. The beak of the retort should incline slightly upward, to prevent contamination by splashing. At about seven centimeters from the end it should be bent at a 24 ANALYTIC OPERATIONS. right angle, and drawn out so as to enter the con- densing worm for such a distance as to terminate beneath the level of the water. The condenser shown in figure 3 is a copper tank, 33 cm. high, 15 cm. wide, and of length pro- FlG. 3- portioned to the number of distilling vessels operated. The condensing tube is not shown in its course thru the tank. Glass worms are apt to break, and it is more satisfactory to use block tin. A piece of rubber SANITARY EXAMINATIONS. 25 tubing is drawn over the junction between the retort neck and worm. A rapid current of cold water should be maintained through the con- denser. The heat is applied by means of the low- temperature burner, the iron ring of which is re- moved so that the retort rests directly on the gauze. To prevent overheating of the upper part of the retort, a sheet of thick asbestos board, about 20 cm. in diameter, with a central opening about 5 cm. in diameter, may be placed on the gauze. With this arrangement the heat is under control, and the danger of breaking the retort is slight. It is advisable to protect the retort from drafts of cold air, which may be done with a cone made of thin sheet asbestos. Many liquids, boiling percussively, are liable to break glass vessels. Among the most success- ful methods for preventing this is the introduction of a few fragments of pumice. Common pumice is broken into masses about as large as an olive- stone, which are placed in a flask or bottle, filled with water, and corked. In a few days the frag- ments will have become waterlogged and sink. If it is necessary to bring some fragments to this condition at short notice, it may be done by heating them to redness and quenching in cold water. Two or three waterlogged fragments should be put into each distilling flask. They 26 ANALYTIC OPERATIONS. will serve for several operations, but must be occasionally renewed. The stock of fragments should be kept in a bottle filled with water. Fig. 4 shows a distilling arrangement, which Fig. 4. is convenient for many laboratory purposes. The condenser shown is relatively too short. It should be not less than 50 centimeters long. It is ad- visable to use Jena glass for the inner tube. The side-tube should be at such an angle as will permit SANITARY EXAMINATIONS. 2^ of the flask being upright when the condenser has the proper angle. The thermometer shown in the cut is, of course, unnecessary in the process under consideration. The flask should be closed by a rubber stopper. All materials should be in- troduced by the aid of a long-necked funnel, so that no splashing occurs. This funnel should be well rinsed before use. The side tube should be pushed into the condensing tube so that its end will be well within the cooled portion of the condensing tube. The junction can be made tight by slipping a short piece of rubber over the condenser tube and passing the side-tube of the flask thru. Another convenient form of apparatus is shown in figure 5. It is employed in the laboratory of the Massachusetts State Board of Health. The joint between the flask and condenser is made by means of a sound cork, into which the condensing tube fits closely ; the tube from the flask is made slightly smaller than the condensing tube, and passes into it for about four centimeters. A form of condenser applicable to distillations of this character has been devised by Cribb and is shown in figure 6. The vapor passes into a narrow annulus by the tube A; the cooling water enters the central portion and overflows, running down the outside wall, being collected by the projecting rim and carried off by the tube G. For water analysis a retort with the neck bent, about the 28 ANALYTIC OPERATIONS. middle, at an obtuse angle, may be used, or a flask with side-tube. In the latter case the tube Scale m \n.= i foot. Fig. s. must leave the flask at a slight angle upward, and about midway be bent at a slight obtuse angle downward. This prevents contamination of the I SANITARY EXAMINATIONS. 29 D D distillate by spurting. The drawing shows the form given by Cribb, but experience has shown that more space should be allowed between the inner and outer wall at the lowest point, and that the catch-basin should be large. The tube, G, should be at least three times the caliber of F. It will often be advantageous to wrap a piece of muslin around the body of the ap- paratus. A modified form of this apparatus is now sold as ''Hopkins, Condenser." Cylinders for Comparison- color Tests, usually called *'Nessler glasses." They should be made of colorless glass, be about 2.5 centimeters in diameter, and marked at SO c.c. Solutions Required : Ammonium-free Water. — If the distilled water of the laboratory gives a reaction with Nessler reagent, it should be treated with sodium carbonate, about one grain to the liter, and boiled until about one-fourth has been evaporated. Ammonium-free water may be ob- tained by distilling, in a retort, water made slightly acid with sulfuric acid. Fig. 6. 30 ANALYTIC OPERATION^. Messrs. J. B. Weems, C. E. Gray, and E. C. Myers recommend the following method : Sodium dioxid is added to ordinary water in the proportion of one gram to a liter, and the liquid boiled for thirty minutes, or longer if the amount of am- monium compounds is high. It is then cooled and the flask kept closed. Flasks holding several liters are most convenient. If the water be dis- tilled, the distillate may also be used for the pre- paration of standard nitrate and nitrite solutions. Standard Ammonium Chlorid. — Dis- solve 0.382 gram of pure dry ammonium chlorid in 100 c.c. of ammonium-free water. For use, dilute i c.c. of this solution with pure .water to 100 c.c. One c.c. of this dilute solution contains 0.0000 1 gram of nitrogen. Fig. 7. Nessler Reagent. — Dissolve 17.5 grams of potassium iodid in 50 c.c. of water. Dissolve 8 grams of mercuric chlorid in 200 c.c. of water. The liquids may be heated to aid solution, but must be cooled before use. Add the mercuric chlorid solution to that of the potassium iodid, until a permanent precipitate is produced. Then dilute with a 20 per cent, solution of sodium hydroxid to 500 c.c, add mercuric chlorid solution until a permanent precipitate again forms, and allow to stand until clear. Nessler and other reagents are best kept in glass-capped bottles SANITARY EXAMINATIONS. 3 1 (Fig. 7) for water analysis, in which the pipet may remain when not in use. A Ikaline Potassium Permanganate. — Dissolve 200 grams of potassium hydroxid, in sticks, and 8 grams of potassium permanganate, in looo c.c. of distilled water. It was originally recommended to boil this solution for some time to remove am- monium, but the procedure described in the next paragraph renders this preliminary treatment unnecessary. The Chemical Section of the American Associa- tion for the Advancement of Science recommended the following method, which is very satisfactory: 200 c.c. of distilled water, and about i gram of pure sodium carbonate, are distilled down to about 100 c.c. in the retort in which the analysis is to be conducted, and the last portion of 50 c.c. nesslerized to assure freedom from ammonium. Then 500 c.c. of the water to be examined are added, and the distillation is carried on at such a rate that about 50 c.c. are collected in each suc- ceeding ten minutes, and until a 50 c.c. measure of distillate is obtained containing only an in- appreciable quantity of ammonium. In nessler- izing, five minutes are to be allowed for the full development of color ; after this, no change takes place for many hours. The distilling vessel is emptied and rinsed thor- oughly, 200 c.c. of distilled water and 50 c.c. of ^2 ANALYTIC OPERATIONS. alkaline permanganate solution put in, and the liquid distilled down to about loo c.c., the last portions of the distillate being tested to ascertain freedom from ammonium compounds, another portion of 500 c.c. of the water to be tested is added, and the distillation made as before. The difference between the ''nitrogen as ammonium" C'free ammonia") of the first operation and the total ammonia of the second is to be taken as the ''nitrogen by permanganate" ("albuminoid am- monia"). It is convenient to operate the distilling flasks in pairs, using one of each pair for the perman- ganate process. Delay and trouble of rinsing are thus avoided. Before beginning an analysis, the greater part of the residues from a previous operation may be drawn off from the flasks with a siphon, 200 c.c. of distilled water added to each, and the liquids distilled until the reagent shows freedom from ammonium compounds. Suitable portions of the sample are then put in each flask. "Nesslerizing," the term applied to the use of Nessler's reagent for determining the ammonium in the distillates, is performed as follows: To each of the distillates of 50 c.c. collected, as directed, in the comparison cylinders, or other suitable vessels, 2 c.c. of Nessler's reagent are added. A yellowish-brown solution is formed, the intensity of which is proportional to the amount SANITARY EXAMINATIONS. ;^^ of ammonium present. The full color is de- veloped in five minutes. This color is exactly matched by introducing into another cylinder 50 c.c. of ammonium-free water, some of the standard ammonium chlorid solution, and 2 c.c. Nessler reagent as before. According as the color so produced is deeper or lighter than that obtained from the water, other comparison liquids are pre- pared containing smaller or larger proportions of the ammonium chlorid, until the proper color is produced. If the quantity of ammonium is sufficient to cause a precipitate, the color comparison can not be accurately made. In most cases this will not be of serious moment, as the quantity will be "beyond the allowable limit. If accurate deter- mination be desired, it may be made by dividing the first distillate into two equal parts, nessleriz- ing one of these, and then, if necessary, diluting the second part with ammonium-free water and nesslerizing this. Since small quantities of ammonium compounds and nitrogenous matters are everywhere present, the greatest care should be exercised in order to avoid their introduction in any way during the course of the analysis. All measuring vessels, cylinders, pipets, and flasks, etc., should be thor- oughly rinsed before using. The temperatures of the distillates and standards should be ap- 3 34 ANALYTIC OPERATIONS. proximately the same when the colors are com- pared. For nesslerizing and other color comparisons, many forms of apparatus have been proposed. One, devised by Hehner, is shown in figure 8. It consists of a graduated cylinder with a stopcock near the base, by which the liquid can be drawn down at will. Two such cylinders may be used — one for the nesslerized dis- tillate, the other for the comparison liquid. The /N 5ftcc. Uco Sl6c 3Y^ <|Xi> Fig. 8. Fig. 9. darker liquid is drawn out until the tints are equal, when the relative volumes remaining will give the data for calculation. H. J. Watson has modified the Hehner tube as SA*nTARY EXAMINATIONS. 35 shown in figures 9 and 10. The cuts were loaned by the Amer. Jour, of Pharmacy. The jar is 30 Fig. 10. cm. long, 1.8 cm. in diameter, and is graduated into cubic centimeters for about 20 cm. from the base. At the side of the base a small tube pro- 36 ANALYTIC OPERATIONS. jects, which may be provided with a stop-cock, but it will be seen from one of the figures that this is not necessary. A number of tubes similar in size and quality of glass to the graduated tube, but marked only at 50 c.c, should be provided. The distillates from the water are placed in the un- graduated tubes, and compared with the tints of the standard ammonium solution, by making the volume of the latter in the graduated tube increase or decrease by means of the stop-cock on the buret and changing the height of the latter. TOTAL ORGANIC NITROGEN. The ease and certainty with which the nitrogen of most organic bodies may be converted into ammonium sulfate . by boiling with sulfuric acid offers a means of determination free from the' ob- jections of former methods. The method in- troduced by Kjeldahl for general organic analysis was first successfully applied to water analysis by Drown and Martin. In their original process, 500 c.c. was concen- trated to about 300 c.c, and the distillate nessler- ized for determining the nitrogen existing as ammonium compounds. The organic nitrogen is then determined in the residual water. Owing to the fact that organic matter may be decom- posed by moderate heat, there is liability to under- ■fIT SANITARY EXAMINATIONS. 37 estimation of the nitrogen. It is best, therefore, to determine at once the total unoxidized nitrogen, and estimate, without distillation, on a separate portion of the sample, the nitrogen that exists in ammonium compounds. The procedure is as follows : Reagents Required : Concentrated Sulfuric Acid. — This should be as free as possible from nitrogen. , It can now be obtained of high purity. Sodium Hydroxid Solution. — The white granu- lated caustic soda sold for household use will answer; 350 grams are dissolved in water and made up to 1000 c.c. Sodium Carbonate and Hydroxid Solution. — Twenty-five grams of each are dissolved in 250 c.c. of distilled water, and the solution boiled down to 200 c.c. to free it from ammonium. Analytic Process : Determination of Nitrogen Existing as Am- monium. — Two hundred c.c. of the water are placed in a stoppered bottle, two c.c. each of the solutions of sodium carbonate and sodium hy- droxid added, the stopper inserted, the solutions mixed, and allowed to stand for an hour or two. A filter is prepared by inserting a rather large plug of absorbent cotton in a funnel. This should be washed with ammonium-free water until the filtrate gives no color with Nessler reagent. The YliAlllilJ 38 ANALYTIC OPERATIONS. clear portion of the sample is drawn off with a pipet and run thru the filter, the first portions being rejected, since it is diluted by the water retained in the cotton. The filtration is rapid, and when 100 c.c. of the liquid have passed thru it, is nesslerized. If but little ammonium is present, a narrow tube about 60 centimeters long should be used for observing the color. Estimation of the Total Organic and Ammoniac al Nitrogen. — Five hundred c.c. of the water are placed in a round-bottomed Jena glass flask, 10 c.c. of concentrated sulfuric acid added, and a piece of pumice-stone is heated to bright redness and dropped in while hot. The liquid is boiled for an hour after it is colorless, or, at least, very pale yellow. The flask is allowed to cool, and about 250 c.c. of ammonium-free water added. Fifty c.c. of the sodium hydroxid solution should be placed in the distilling apparatus, about 250 c.c. of water added, a piece of red-hot pumice-stone dropped in, and the liquid distilled until the dis- tillate is free from ammonium. It is best to distil until the retort contains not more than 100 c.c. The sulfuric acid solution is then poured in slowly by means of a funnel, the stem of which touches the side of the retort, so that the two liquids do not mingle. The stopper of the retort is inserted, the liquids mixed by gentle agitation, and distilled. If much ammonium is present, it is advisable to I SANITARY EXAMINATIONS. 39 distil the first portion into 10 c.c. of very dilute (i : 1000) sulfuric acid, a piece of glass tube being connected to the condensing worm, so that the lower end dips below the surface of the liquid. The distillates are collected and nesslerized in the usual way. A blank experiment should be made to deter- mine the amount of ammonium in the sulfuric acid. NITROGEN AS NITRATES. A. H. Gill has subjected the various indirect methods of estimating nitrates to comparative examination, and finds the following method satisfactory : Solutions Required : Phenoldisulfonic Acid: Sulfuric acid iii grams. Phenol 9 " The mixture is heated for six hours in, not upon, the water-bath. The resulting compound often solidifies to a white mass on standing, but can be easily liquefied on the water-bath during the evaporation of the samples to be tested. Standard Potassium Nitrate. — 0.722 gram of potassium nitrate, previously heated to a tem- perature just sufficient to fuse it, dissolved in water, and the solution made up to looo c.c. One 40 ANALYTIC OPERATIONS. c.c. of this solution will contain o.oooi gram of nitrogen. Analytic Process : A measured volume of the water is evaporated just to dryness in a porcelain basin about six centimeters in diameter. One c.c. of the -phenol- disulfonic acid is added and thoroughly mixed with the residue by means of a glass rod. The liquid is then diluted with about 25 c.c. of water, ammonium hydroxid added in excess, and the solution made up to 50 c.c. The nitrate converts the phenoldisulfonic acid into picric acid, which, by the action of the am- monium hydroxid, forms ammonium pier ate ; this imparts to the solution a yellow color, the intensity of which is proportional to the amount present. One c.c. of the standard solution of potassium nitrate is now similarly evaporated in a platinum basin, treated as above, and made up to 50 c.c. The color produced is compared to that given by the water, and one or the other of the solutions is diluted until the tints of the two agree. The comparative volumes of the liquids furnish the necessary data for determining the amount of nitrate. The results obtained by this method are satis- factory. Care should be taken that the same quantities of phenoldisulfonic acid are used for the water and for the comparison liquid. SANITARY EXAMINATIONS. 41 With subsoil and other waters probably con- taining much nitrates, lo c.c. will be sufficient; but with river and spring waters, 25 c.c. may be used. When the organic matter is sufficient to color the residue, it will be well to purify the water by addition of aluminum hydroxid and filtration, before evaporating. Chlorin interferes with the accuracy of the test, but Gill finds that when not amounting to more than 20 parts per million, it does not impair the practical value of the results. When greater than this, it is best to evaporate in vacuo over sulfuric acid. If the chlorin is more than 70 parts per million, it should be considerably reduced by the addition of silver sulfate which has been ascer- tained to be free from nitrates. Nitrites do not influence the reaction. Nitron Method. — A complex synthetic product, the cumbersome systematic name of which has been replaced by the commercial name ** nitron," forms with nitric acid a definite compound spar- ingly soluble in cold water. The application of this substance is more likely to be made in the general assay of nitrates, but as it may have oc- casional application in water analysis, the analytic method is briefly described. Nitron is a yellowish powder. For use it is dissolved in the proportion of i gram to loo c.c. of 5 % acetic acid. The solution does not keep 42 ANALYTIC OPERATIONS. long, hence large amounts should not be made up at once. The determination is made by adding to a convenient amount of the solution of the nitrate, sufficient dilute sulfuric acid to make the liquid distinctly acid, heating to boiling, adding a few c.c. of the nitron solution, and then allowing the liquid to cool to room temperature, after which it is immersed in melting ice for several hours. Max Busch, to whom the discovery of the use of nitron is due, states that the substances usually occurring in water do not interfere, and that the reagent will give a precipitate with i part of nitric acid in 60,000 of water. This cor- responds to about 4 parts of nitrogen per 1,000,000. The precipitate must be collected on a filter, preferably by means of the gooch crucible, well sucked out by the filter pump, washed with a few c.c. of water as near 0° as possible, again sucked out, dried and weighed. The weight multiplied by 0.037 gives nitrogen. Concentration of water samples will, of course, increase the delicacy of the test. This concen- tration should be carried on at a very low tem- perature, and the water should be neutral or alkaline, or nitrates will be decomposed. NITROGEN AS NITRITES. The following is Ilosvay's modification of Griess's test. It has the advantage over the SANITARY EXAMINATIONS. 43 original method, that the color is developed more rapidly, and the solutions are less liable to change. Solutions Required : . 1-4-Amidobenzenesulfonic Acid Solution (Sulf- anilic Acid). — 0.5 gram dissolved in 150 c.c. of diluted acetic acid, sp. gr. 1.04. a-Amidonaphthalene Acetate Solution. — Boil o.i gram of solid a-amidonaphthalene (<3t-naphthyl- amin) in 20 c.c. of water, filter the solution thru a plug of washed absorbent cotton, and mix the filtrate with 180 c.c. of diluted acetic acid. All water used must be free from nitrites, and all vessels must be rinsed out with such water before tests are applied, since appreciable quantities of nitrites may be taken up from the air. Much of the commercial «-naphthylamin has a very offensive odor, but a pure article less ob- jectionable can now be obtained. Standard Sodium Nitrite. — 0.275 gram pure silver nitrite dissolved in pure water, and mixed with a dilute solution of pure sodium chlorid until a precipitate ceases to form. The solution is diluted with pure water to 250 c.c, and allowed to stand until clear. For use 10 c.c. of this solution are diluted to 100 c.c. It is to be kept in the dark. One c.c. of the dilute solution is equivalent to 0.00001 gram nitrogen. The silver nitrite is prepared thus: A hot con- centrated solution of silver nitrate is added to a 1% 44 ANALYTIC OPERATIONS. concentrated solution of the purest sodium or potassium nitrite available, filtered while hot, and allowed to cool. The silver nitrite will separate in fine, needle-like crystals, which are freed from the mother liquor by filtration by the aid of a filter pump. The crystals are dissolved in the smallest possible quantity of hot water, allowed to cool, and again separated by means of the pump. They are then thoroly dried in the water-bath, and preserved in a tightly stoppered bottle away from the light. The purity may be tested by heating a weighed quantity to redness in a tared porcelain crucible. The residue is silver and should be 70.1% of the weight of nitrite taken. Analytic Process : Twenty-five c.c. of the water are placed in one of the color-comparison cylinders, and two c.c. each of the test solutions are dropped in. It is con- venient to have a pipet for each solution, and to use it for no other purpose. One c.c. of the standard nitrite solution is placed in another clean cylinder, made up with nitrite-free water to 25 c.c, and treated with the reagents as above. In the presence of nitrites the liquid becomes pink. At the end of five minutes the two solutions are compared, the colors equalized by diluting the darker, and the calculation made as explained under the estimation of nitrates. SANITARY EXAMINATIONS. 45 The reactions consist in the conversion of the sulfaniHc acid into diazobenzenesulfonic anhydrid, by the nitrite present; this compound is then in turn converted by the amidonaphthalene into azo-a-amidonaphthalene-i-4-benzenesulfonic acid. The last-named body gives the color to the liquid. OXYGEN-CONSUMING POWER. All organic materials being more or less easily oxidized, several methods have been suggested for determining the oxygen-consuming powers of waters by treatment with active oxidizing agents. These methods are, however, limited in value. The organic matters in water differ much in character and condition, and their oxidability is subject to much variation, according to the cir- cumstances under which the test is made. Never- theless, as a high oxygen-consuming power cer- tainly indicates departure from purity, some ad- ditional evidence may be obtained. Potassium permanganate is especially suitable. The test is usually made by introducing a known amount of the permanganate into the water, which has been rendered acid, and measuring after a definite period the proportion which has been decomposed. It must not be overlooked that if a water con- tains nitrites, ferrous compounds, or sulfur com- pounds other than sulfates, the proportion of 46 ANALYTIC OPERATIONS. oxygen consumed will be greater than that re- quired for the organic matter. It has been pro- posed, in order to remove the nitrites before apply- ing the permanganate, to take 500 c.c. of the water, add 10 c.c. of the dilute sulfuric acid, boil for twenty minutes, allow to cool, and then treat with permanganate. Since, however, the amount of nitrites, if appreciable, can be directly deter- mined, it is more satisfactory to deduct from the oxygen consumed the amount required to con- vert the nitrites present into nitrates, and the remainder will be that required for the other oxidizable ingredients. Fourteen parts of nitrogen existing as nitrite require 16 parts of oxygen for conversion into nitrate. Similarly, 112 parts of iron in a ferrous compound will require 16 parts of oxygen for conversion to the ferric con- dition. Of the following methods, the first, due, in the main, to Tidy, was improved by Dupre, and ap- proved by the British Society of Public Analysts. Solutions Required : Standard Permanganate. — 0.395 gram pure po- tassium permanganate dissolved in distilled water, and the solution made up to 1000 c.c. One c.c. is equal to o.oooi gram oxygen. Diluted Sulfuric Acid. — Pure sulfuric acid is diluted with twice its bulk of water, and then a solution of potassium permanganate added suffi- SANITARY EXAMINATIONS. 47 cient to give a faint pink, permanent when the liquid is heated to 80° F. for four hours. Potassium lodid. — Ten grams of the pure sub- stance dissolved in 100 c.c. of distilled water. Sodium Thiosulfate. — One gram of the pure crystallized substance dissolved in 2000 c.c. of dis- tilled water. Starch Indicator. — One gram of clean starch is mixed smoothly with cold water into a thin paste, then poured gradually into about 200 c.c. of boiling water, the boiling continued for one minute, the liquid allowed to settle, and the clear portion used. It is best freshly prepared. Analytic Process : Two determinations are made — one, of the oxygen consumed in fifteen minutes, which is con- sidered to represent the nitrites, sulfids, or ferrous compounds, and the other, of the oxygen con- sumed by four hours' action. Both determina- tions are made at a temperature of 80° F. Three glass-stoppered bottles, of about 350 c.c. capacity, are rinsed with strong sulfuric acid and then with water. In one is placed 250 c.c. of pure distilled water as a control experiment, and in each of the others 250 c.c. of the water to be tested. The bottles are stoppered and brought to a temperature of 80° F. ; 10 c.c. of the dilute sulfuric acid and 10 c.c. of the standard permanganate are added to each, and the stoppers again replaced. At the I 48 ANALYTIC OPERATIONS. I end of fifteen minutes one sample of water is removed from the bath, and two or three drops of the potassium iodid solution added to remove the pink. After thoro admixture, the thiosulfate solution is run in from a buret until the yellow is nearly destroyed, a few drops of the starch solutiorij added, and the addition of the thiosulfate con-' tinned until the blue is quite discharged. If the addition of the thiosulfate solution has been prop4| erly conducted, one drop of the permanganat^ solution will restore the blue. » The other bottles are maintained at 80° F. for four hours. Should the pink disappear rapidly in the bottle containing the water under examination, 10 c.c. of the permanganate solution must be added to each bottle, in order to maintain a distinct color. At the end of four hours each bottle is removed from the bath, two or three drops of potassium iodid added, and the titration with thiosulfate solution conducted as just described. The calculation is most conveniently made as follows : a = number of c.c. required for the control experiment. b = number of c.c. required for the water under examination. c = available oxygen in permanganate (o.ooi for 10 c.c). X — oxygen consumed by water. Then, a : a-b : :c :x. SANITARY EXAMINATIONS. 49 The following method was recommended by the Chemical Section of the American Association for the Advancement of Science : '' Prepare a solution of potassium permanganate containing 0.2 milligram of available oxygen to I c.c. and a solution of oxalic acid of such strength as to decompose the permanganate solution, volume for volume, the strength being rede- termined from time to time. The water used for making these solutions should be purified by distillation from alkaline permanganate. ''To 200 c.c. of water to be examined, in a 400 c.c. flask, add 10 c.c. of dilute sulfuric acid (i : 3) and such measured quantity of the permanganate as will give a persistent color; boil ten minutes, add, if necessary, more permanganate in measured quantities, so as to maintain the red color; re- move the flask from the lamp, add 10 c.c. of oxalic acid solution to destroy the color, or more if required by the excess of permanganate, and then add permanganate, drop by drop, until a faint pink appears. From the total quantity of permanganate used deduct the equivalent of the oxalic acid used, and from the remainder calculate the milligrams of oxygen consumed by the ox- idizable organic matter in the water/' The oxygen-consuming power may also be indirectly estimated by the action of the organic matter upon silver compounds. H. Fleck's i 50 ANALYTIC OPERATIONS. method depends upon the reduction produced b boiling the water with alkaHne solution of silve: thiosulfate and estimation of the unreduced silver. A. R. Leeds has proposed a method by treating the water with decinormal silver nitrate, exposing to light until it settles perfectly clear, and estimating the reduced silver. These methods are open to practically the same objections as in the use of permanganate, and do not seem to possess any decided advantage. Qualitative results of some interest may occa- sionally be obtained by the following method: Two c.c. of a one per cent, solution of silver nitrate, rendered decidedly alkaline by ammonium hydroxid, are added to 100 c.c. of the water in a stoppered bottle, which is then placed in full sun- light for two hours. Waters containing but little organic matter will not show at the end of this period any appreciable tint. PHOSPHATES. The following method is recommended by A. G. Woodman : Solutions Required : Ammonium Molybdate Solution. — 50 grams in I liter of distilled water. Nitric acid, sp. gr. 1.07. Sodium Phosphate Solution. — 0.5324 gram crys- tallized disodium hydrogen phosphate, 100 c.c. of SANITARY EXAMINATIONS. 5 1 the above nitric acid, distilled water sufficient to make looo c.c. This is equivalent to coooi phosphoric anhydrid in i c.c. Analytic Process : Fifty c.c. of the sample and 3 c.c. of nitric acid are evaporated to dryness in a porcelain dish on the water-bath, the residue heated for two hours at 100"^ C, and treated with 50 c.c. of cold distilled water added in several portions which are mixed in a comparison tube. It is usually unnecessary to filter. Four c.c. of the ammonium molybdate solution and 2 c.c. of nitric acid are added, the contents mixed, and after three minutes the color compared with that given by different quantities of the standard phosphate solution which have been made up to 50 c.c. and the reagents added in the same amount as above. Blank tests must be made to determine the purity of the materials used. Distilled water kept for some time in glass vessels may contain appreciable amounts of substances giving color with the reagents. Ammonium molybdate solution suitable for the gravimetric determination of phosphates may be prepared as follows : Weigh into a beaker 10 grams of pure molyb- denum teroxid, mix well with 40 c.c. cold distilled water, and add 8 c.c. strong ammonium hydroxid, 52 ANALYTIC OPERATIONS. sp. gr. 0.900. When completely dissolved, filter and pour slowly, with constant stirring, into a mixture of 40 c.c. of nitric acid, sp. gr. 1.42, and 60 c.c. of water. Add 0.005 gram sodium am- monium hydrogen phosphate, dissolved in a little water, agitate well, allow precipitate to settle twenty-four hours, and filter before using. DISSOLVED OXYGEN. The method here given, a modification of Mohr's, was proposed by Blarez. It is rapid and satis- factory. Solutions Required : Sodium Hydroxid. — Forty grams of pure sodium hydroxid to 1000 c.c. Ferrous-ammonium Sulfate, — Forty grams dis- solved in about 1000 c.c. of water, and acidified with a few drops of concentrated sulfuric acid. Decinormal Potassium Permanganate. — 3 .156 grams in 1000 c.c. of distilled water. The ac- curacy of this solution should be determined by titration with a known weight of ferrous-am- monium sulfate. One c.c. should be equivalent to 0.0392 gram ferrous ammonium sulfate (0.0008 gram of oxygen). The apparatus employed (shown in Fig. 11) is a globular separator, of about 250 .c.c. capacity. Above the bulb is a caoutchouc stopper carrying SANITARY EXAMINATIONS. 53 a cylindric funnel, of about 12 c.c. capacity, terminating in a tube, 8 mm. caliber, sharply con- tracted at the outlet to a capillary opening. The tube should project about 6 mm. below the stopper. The exact capacity of the apparatus is measured as follows: The bulb is completely filled with water and the stopper inserted; the level of the water will rise slightly in the funnel tube, and should be brought down to its outlet by drawing a little off at the stop-cock, after which the water is run into a graduated measure and its volume noted. Analytic Process : Thirty-five c.c. of mercury and ten c.c. of sodium hydroxid solution are put into the bulb, and then sufficient of the water to be tested to fill it. The funnel stopper is inserted, and the water which rises into the funnel brought into the bulb by cautiously running out at the stop-cock, mer- cury, the volume of which should be noted. The exact volume of water used is thus known. Five c.c. of the ferrous ammonium sulfate solution are poured into the funnel, brought into the bulb by running out mercury, and the liquid thoroughly mixed by giving the apparatus a gyratory move- ment. After standing five or six minutes the Fig. II. 54 ANALYTIC OPERATIONS. oxygen will be completely absorbed; lo c.c. of the diluted sulfuric acid are now added by the same method. On agitating the bulb, the con- tents become clear. The watery liquid is then transferred to a beaker and titrated with deci- normal permanganate. A volume of water equal to that used in the test is poured into another beaker, lo c.c. each of the sodium hydroxid and diluted sulfuric acid added, and then 5 c.c. of ferrous ammonium sulfate solution. The result- ing liquid is titrated with permanganate. The weight of oxygen corresponding to the difference between the two titrations gives the weight of dissolved oxygen in the liquid employed. Nitrates do not appear to impair the accuracy of this method, and the interfering action of nitrites and other reducing compounds is avoided by the control experiment. It is perhaps hardly necessary to add that the exact temperature of the water is to be noted at the time of collection of the sample. In transferring to the bulb, the water should be agitated as little as possible in contact with the air, in order to avoid the absorption of oxygen. A siphon should be used for this purpose, the lower end being allowed to reach to the bottom of the bulb. SANITARY EXAMINATIONS. 55 POISO/NOUS METALS. Under this conventional title are included barium, chromium, zinc, arsenic, copper, and lead. Manganese, iron and aluminum are objectionable when present in notable amounts. Barium is rarely present, and only in water containing no sulfates. It can be detected and determined by slightly acidifying the water with hydrochloric acid, filtering if necessary, and adding solution of calcium sulfate. The precipi- tated barium sulfate is collected and weighed in the usual way. Chromium is rarely present, but may be looked for in the waste waters of dye-works and similar sources. To detect it, a considerable volume of the water is evaporated to dryness with addition of a small amount of potassium chlorate and nitrate, transferred to a porcelain crucible, and brought to quiet fusion ; any chromium present will be found in the residue in the form of chromate. The fused mass, after cooling, is boiled with a little water, filtered, the filtrate rendered slightly acid with hydrochloric acid, and a solution of hydrogen dioxid or sodium dioxid added. In the presence of chromium a transient blue will appear; by adding a little ether and shaking the mixture, the color will pass into the ether, and on standing a blue layer will form on the surface of the water. 56 ANALYTIC OPERATIONS. Zinc is best detected by the test described by Allen. The water is rendered slightly alkaline by addition of ammonium hydroxid, heated to boil- ing, filtered, and the clear liquid treated with a few drops of potassium ferrocyanid; in the presence even of the merest trace of zinc a white precipitate will be produced. Arsenic is most readily detected by Reinsch's test. A considerable volume of the water is rendered slightly alkaline by pure sodium car- bonate, and evaporated nearly to dryness in a porcelain basin. Two or three c.c. of distilled water strongly acidulated with hydrochloric acid are placed in a small test-tube, about ^ of a square centimeter of bright copper foil is added, and the liquid boiled gently for a few minutes. If the copper remains bright, showing that the re- agents contain no arsenic, the water-residue is acidified with hydrochloric acid, added to the contents of the test-tube, and the liquid again boiled for several minutes. If arsenic is present, a steel-gray stain will appear on the copper. The slip is removed, washed with distilled water, thoroly dried by pressure between filter-paper, inserted into a narrow glass tube closed at one end, which has been previously dried by heating nearly to redness. The tube is gently heated at the point at which the copper rests; the deposit will sublime and collect on the cooler portion of the SANITARY EXAMINATIONS. 57 tube, in crystals which the microscope shows to be octahedral. Since small amounts of arsenic frequently occur in reagents and in glass vessels, care must be taken to avoid such sources of error. Sodium carbonate solution may contain arsenic dissolved from the glass bottle in which it is kept. It is best, therefore, to use the solid carbonate for rendering the water alkaline, and to determine its purity before use. For very delicate testing for arsenic. Marsh's test should be used, according to the methods described in works on toxicology. Iron is detected by the addition of a drop of ammonium sulfid to the water in a tall glass cylinder. Ferrous sulfid is formed having a greenish-black color, instantly discharged by acidifying the water with dilute hydrochloric acid. A still better test is the production of a blood-red color, with potassium thiocyanate, due to the formation of ferric thiocyanate. The water should be first boiled with a few drops of nitric acid, to convert the iron to the ferric condition, cooled, and a drop or two of the solution of potassium thio- cyanate added. The test is very delicate. Either of the above tests may be made quantitative by matching the color produced in loo c.c. of the water with that obtained from a known weight of iron. The method with potassium thiocyanate 58 ANALYTIC OPERATIONS. is preferable, as it is more delicate and there are fewer interfering conditions. The following is the method as elaborated by Thompson and described in Sutton's ''Volumetric Analysis" : Solutions Required : Standard Ferric Sulfate. — 0.7 gram ferrous am- monium sulfate is dissolved in water acidified with sulfuric acid, and potassium permanganate solu- tion added until the solution turns a very faint pink color. The solution is diluted to 1000 c.c. One c.c. contains o.oooi gram iron. Diluted Nitric Acid. — Thirty c.c. concentrated nitric acid diluted with water to about 100 c.c. Potassium Thiocyanate. — Five grams of the salt dissolved in about 100 c.c. water. Analytic Process : About 100 c.c. of the water are evaporated to small bulk, acidified with hydrochloric acid, and just sufficient dilute potassium permanganate solution added to convert all the iron to the ferric condition. The liquid is evaporated nearly to dryness to drive off excess of acid, then diluted to its original volume, 100 c.c. In two tall glasses marked at 100 c.c, 5 c.c. of the nitric acid and 15 c.c. of the thiocyanate solution are placed. To one of these a measured volume of the treated water is added and both vessels filled up to the mark with distilled water. If iron is present, a blood-red color will be produced. Standard iron SANITARY EXAMINATIONS. 59 solution is added to the second vessel until the color agrees. The amount of water which is added to the first glass will depend upon the quantity of iron it contains; not more should be used than will require two or three c.c. of the standard to match it, otherwise the color will be too deep for comparison. Manganese. — The following method is described by Wanklyn in his treatise on water analysis. A considerable volume (i. e., 1000 c.c.) of the water is evaporated to small bulk, nearly neutralized by hydrochloric acid, and treated with a few drops of a solution of hydrogen dioxid. The formation of a brown precipitate indicates the presence of manganese. The test is very delicate. The pre- cipitate may be collected on a filter, the filter ashed, and the residue fused with a mixture of sodium carbonate and potassium nitrate. Green potassium manganate will be produced, which, when boiled with water, will give a bright-red solution of potassium permanganate. The quan- titative determination is given elsewhere. Aluminum. — Traces of this element are to be expected in all waters, and it is not usual to test for it except in elaborate analysis of the mineral ingredients, as described in another section. The use of aluminum sulfate as a coagulant in many rapid-filtration methods makes it necessary to examine effluents for excess of precipitant, and 6o ANALYTIC OPERATIONS. this may be done by the following method devised by Mrs. Richards : To 25 c.c. of the water to be tested (concentrated from a large volume, if necessary) is added a few drops of freshly prepared logwood decoction ; any alkali is neutralized and the color is brightened by the addition of two or three drops of acetic acid. By comparison with standard solutions, the araount of alum present may be determined. One part of alum in 1,000,000 of water can be detected with certainty. In cases of greater dilution, concentration of larger volumes may be necessary to obtain a decisive test. The log- wood chips yield the right color only after having been treated with boiling water two or three times, rejecting the successive decoctions. The first portion gives a yellow color, the third or fourth usually a deep red. The logwood chips must be fresh. I have found the solution of sodium alizarin- monosulfonate (see p. 15) a useful reagent for detecting excess of aluminum compounds in water. The sample should be filtered and a few drops of the indicator solution added. When aluminum compounds are present, especially aluminum sulfate, the reagent assumes a yellow- ish color and becomes much less sensitive, so that a considerable amount of decinormal acid can be added without producing the acid reaction. If SANITARY EXAMINATIONS. 6l another equal portion of the sample is treated with a small amount of sodium carbonate, allowed to stand an hour or so, filtered and tested with the indicator, the difference in reaction will be marked if the sodium carbonate has been added in quan- tity sufficient to overcome the aluminum com- pound present. Lead may be readily detected by adding to the water in a tall glass cylinder a drop of ammonium sulfid; brownish-black lead sulfid is formed, which does not dissolve either by acidulating the water with dilute hydrochloric acid (distinction from iron), nor by the addition of about one c.c. of a strong solution of potassium cyanid (dis- tinction from copper). S. Harvey gives the following method: 250 c.c. are placed in a pre- cipitating jar, about o.i gram of crystallized potassium dichromate is added and dissolved by agitation. The same volume of lead-free water is treated in the same manner, and the two solu- tions placed side by side. Water containing 0.3 part per million will show a turbidity in fifteen minutes which will be rendered more distinct by contrast with the clear water alongside. By allowing the jar to stand for about twelve hours undisturbed, the precipitate will settle and will become still more distinct. No other metal likely to be present in water will give a similar reaction. 62 ANALYTIC OPERATIONS. In the absence of copper the amount of lead present may be determined as follows : A solution is prepared containing 1.6 grams of lead nitrate to 1000 c.c. ; one c.c. of this contains one milligram lead. One hundred c.c. of the water to be tested are placed in a tall glass vessel, made acid by the addition of a few drops of acetic acid, and five c.c. of hydrogen sulfid added. In a similar vessel 100 c.c. of distilled water are placed, together with the same quantities of acetic acid and hy- drogen sulfid, and sufficient of the standard lead solution to match the tint in the first cylinder. The amount of lead in the water under examina- tion is thus known. Copper. As copper sulfate is now used for preventing vegetable growths in water, it is often necessary to test for it. When present in appre- ciable amount, it may be detected by acidifying with acetic acid and adding hydrogen sulfid. The precipitate is distinguished from lead sulfid by the fact that the color is discharged on the ad- dition of about I c.c. of a strong solution of pure potassium cyanid. The presence of copper may be confirmed by the addition of a solution of potassium ferrocyanid to another portion of the water. In the presence of even a very small amount of copper, a mahogany-red color is pro- duced. H. C. Bradley has found that logwood hema- I SANITARY EXAMINATIONS. 63 toxylon is a delicate test for copper. It is best used in weak alcoholic solution to which a trace of sodium hydroxid has been added so as to render it faintly pink. A few drops of this reagent added to a moderate volume of the sample will in the course of a few minutes develop a distinct blue, even if very small amounts of copper are present. Larger amounts of copper produce a deep blue precipitate. The liquid should be allowed to stand for an hour or so before deciding that the test is negative. Bradley has found that free mineral acid interferes with the test. In the absence of lead, copper is determined in the same way as that metal, using, however, a standard solution of copper for the comparison liquid. This is made by dissolving 3.929 grams of crystallized copper sulfate in one liter of water. One c.c. of the solution contains one milligram copper. If both lead and copper are present, a large quantity of the water should be evaporated to small bulk, and the metals separated and deter- mined by any one of the ordinary laboratory methods. BIOLOGIC EXAMINATIONS. In a comprehensive sense the living organisms of water include representatives of all the great 64 ANALYTIC OPERATIONS. groups of animals and plants. The higher orders of organic forms are absent from very foul water. From an analytic point of view, observation is limited to the determinations of those forms which are inappreciable to the unassisted eye. So far as regards some of the moderately com- plex organisms, such as the minute crustaceans, algae, diatoms, and even ameb^, it may be said that while some general inferences as to the character and history of the water may be de- duced from an identification of the specific forms, no sanitary signification can be attached to them, except that they are objectionable. From what is now known of the life-history of parasitic organisms, it is evident that water that is freely accessible to animal forms is liable to be danger- ously polluted. Moreover, the dead bodies of such animals will furnish food to many forms of microbes and thus assist in the multiplication of the latter. The ova of the entozoa might in some cases be detected by careful search, and would indicate recent pollution of a highly dangerous character. The number of the higher forms present in any sample will depend very much upon the point at which it is collected, they being more numerous in the neighborhood of abundant vegetable growths, and at the bottom and sides of streams. Several observers, notably Sedgwick and Rafter, SANITARY EXAMINATIONS. 65 have paid considerable attention to the recog- nition of the animal and vegetable forms in surface waters. Some of these forms cause disagreeable odors and colors ; in the warm season of the year, when such water is stored in reservoirs, consider- able annoyance is felt by the users, and the en- gineer-in-charge is subjected to much criticism. It has been found that even crude filtration methods, such as allowing the water to pass thru a dike of porous soil before storing it in a reservoir, will diminish the tendency to these conditions. Cleansing a reservoir — disinfecting the inner surface, for instance, by whitewashing — has also improved the condition. Observation, especially in Massachusetts, has shown that reservoirs, intended for even moder- ately prolonged storage of water, should be clean — that is, organic matter of any kind should not be allowed to accumulate on the bottom and sides. Drown states that while the water in one basin became foul from stagnation, in another which was carefully prepared by the removal of all soil and vegetable matter, and is supplied by a brown, swampy water from a district almost entirely free from pollution, the water is good at a depth of forty feet. In Philadelphia, where large storage reservoirs are used for water that is often very muddy, but little trouble from the growth of microscopic 5 66 ANALYTIC OPERATIONS. organisms occurs. These reservoirs are artificial basins. Sedgwick's method of collecting organisms other than microbes, with some modifications by Williston, is as follows : In ordinary cases about loo c.c. are employed. Sometimes it will be advantageous to use double this quantity, at other times much less. In rare cases the examination can be made upon unfiltered water. Originally sand was employed for a filter material, but Williston finds that precipitated silica, made by decomposing silicon fiuorid with water, is more satisfactory. This precipitated silica is a commercial article, and its method of preparation is given in all the larger manuals of chemistry. A small glass funnel with an even-calibered stem is selected, and the lower end of the stem plugged with a little absorbent cotton, upon which a layer three or four mm. deep of the filter-mate- rial is placed. The requisite volume of water is then allowed to filter thru. The pledget of cotton is removed, and the filter-material is washed down with filtered or distilled water into a cell intended for microscopic examination. This cell is a glass plate accurately ruled, to which is attached a brass cell 50 mm. long by 10 mm. wide, of depth sufficient to hold about two c.c. of water. After the material has been allowed to distribute itself SANITARY EXAMINATIONS. 67 and settle in the cell, it is examined with a moder- ate power, and the different organisms in a vary- ing number of the squares counted. Each or- ganism may be counted by itself, if occurring in large numbers, the average of a few squares being sufficient for the purpose. Organisms less numer- B Sedgwick'-Rifter B1587. ously represented may be counted by averaging a larger number of squares. Figure 12 (loaned by the Arthur H. Thomas Co.) shows a more elab- orate form of this filter. Filtering in this manner can not be relied upon in all cases. Indeed, in most cases the unfiltered water also should be examined. Some of the 68 ANALYTIC OPERATIONS. minute unicellular organisms pass thru a small extent of sand or precipitated silica, or even filter-paper. It is not unlikely that the high-speed centrif- ugal apparatus now used in laboratories, associ- ated with the employment of some fine precipitant, will aid in these investigations. Dibdin prepares as follows, a ''micro-filter," for the collection of minute suspended matters: A piece of combustion-tubing 20 to 25 cm. long is cleaned and drawn out in the middle to a capillary tube, and broken by a file scratch at a point at which the caliber is not more than two milli- meters. Each of these pieces serves for a filter. A mixture of equal parts of air-dried clay and infusorial earth is made into a smooth, stiff paste with water and spread out on a slab in a layer about two millimeters deep. The capillary end of a tube is pressed down into the mass and moved in a circle until a plug is formed. This is warmed until dry, and heated to redness, forming a close filter. The water to be examined is filtered in con- siderable amount, — 1000 c.c, for example, if there is but little suspended matter, — first thru a hardened paper filter placed in a funnel, pre- cautions being taken to exclude dust. The deposit is washed from the filter-paper into the micro-filter by means of a jet of pure water. The SANITARY EXAMINATIONS. 69 suspended matter collects on the top of the clay plug and is measured by noting its height. If the clay plug is blocked, the application of a filter- pump may be needed. When the column of water in the small tube is only about a centi- meter in height, the main body of the tube is cut away by means of a file scratch and the deposit loosened from the filter plug, if necessary, by the use of a platinum wire. The tube is inverted so as to bring the deposit to the open end, and then cut off close to the plug. By this means the sus- pended matter is collected in a short capillary tube open at both ends. By gentle shaking, the contents may be brought onto a glass slide. Owing to the great differences in the size of microscopic organisms, the mere enumeration of their numbers is not always an index of the araount of living matter in suspension. To obviate this, Whipple has suggested a standard unit of size, estimating by means of it the total volume of the organisms, and not their number. He finds by this method that the analytic and biologic results correspond much more closely than when mere numbers are recorded. The unit is an area of 400 microns — that is, a square of 20 microns on a side. The results are stated in number of standard units per cubic centimeter. Whipple has investigated the conditions in- fluencing the growth of the microscopic organisms 70 ANALYTIC OPERATIONS. in water. He finds that diatoms thrive best with a supply of nitrates and a free circulation of air; temperature alone has no very direct effect. Infusoria will be found in largest numbers when the water contains the greatest amount of finely divided organic matter. When the conditions bring about a circulation of the water, the or- ganisms are not only brought constantly in con- tact with new food materials, but are enabled to reach the upper layers of the water where oxygen is abundant. Bacteriologic examinations may be qualitative or quantitative. The former involves the determina- tion of the species of microbes present, especially those having disease-producing power, or charac- teristic of some form of pollution. The processes are usually laborious, requiring extensive lab- oratory facilities. Quantitative examination — microbe-counting, as it may be called — is the deter- mination of the number of microbes, or microbe- colonies, that can be grown from a given volume of water under specified conditions. As the growth of living organisms is influenced by all external conditions, the results of the culture of microbes are not comparable with one another, unless strict uniformity of methods has been observed. Neg- lect of this fact renders a very large part of the earlier work and some of the present-day work of little statistical value. Among the conditions SANITARY EXAMINATIONS. 7 1 materially affecting the growth of microbes are temperature, reaction of the culture-medium to different indicators, degree of exposure to light and air, and duration of cultivation. The composition of the culture -medium has much influence, and it is difficult to control this exactly, owing to the irregularity of quality of some of the materials used. At the present day microbe-counting for water analysis is done almost entirely with culture- media that are solid at ordinary temperatures, but may be liquefied at or near blood-heat. Gelatin or agar is used for producing the solidity. The former is the most convenient, but its jelly melts at such a low temperature that it is of limited ap- plication, and agar is largely employed. Apparatus for bacteriologic work is now all furnished of good quality by dealers, and will not need special description. For the ordinary meth- ods of microbe-counting the following will be needed : Open-steam Sterilizer. A modification of the Arnold sterilizer j.s now much used. Autoclave, a closed-steam sterilizer, permitting the application of temperatures much above the boiling point of water. Hot-air Oven for special sterilizations. Culture Oven, with thermostat. Double Boiler of agate or other good culinary 72 ANALYTIC OPERATIONS. ware. The inner vessel should have a capacity of a little more than looo c.c. Test-tubes, about 12 cm. long and 1.5 cm. in diameter. Petri dishes, about 10 cm. in diameter and i.o cm. deep. As far as possible, dishes of uniform size should be selected. Each dish and cover should be marked in the center by a diamond with a distinguishing number. Wire baskets for holding several dozen test- tubes. Fermentation-tubes, such as used in the detection of sugar in urine. Ordinary laboratory appliances, such as pipets, burets, funnels, beakers, and cotton-wool. Tin- foil cut in squares 5 cm. on the side. Sterilization in the autoclave is much in favor in laboratories in which much work is done, as it is efficient and rapid, but the high temperature has a greater chemical action upon some of the media than that of the open steam sterilizer. Sufficient sterilization for work in water analysis can be accomplished by the latter method, especially if several short treatments, say of a half hour each, at intervals of twenty-four hours, are employed. The materials for preparing culture-media should be obtained from responsible dealers, who will furnish the grades regularly used. The follow- ing will assist in the selection. SANITARY EXAMINATIONS. 73 Gelatin, A grade made in Germany and dis- tinguished by a monogram of the initials WH is used. Agar. A colorless grade is preferable. Peptone. Witte's dry peptone is used. Dextrose. The grade termed ' ' crystallized pure ' ' (often called '* glucose") is preferred. Meat-extract. That made by the Liebig Meat- Extract Company, limited, of London, is largely used by bacteriologists, but there seems to be no reason for preferring it to the best American extracts. Sodium chlorid. A good quality of table salt will suffice. Glycerol should be as free as possible from acid and mineral matters. Lactose should be of high purity, especially free from milk-proteids. I have found samples of the imported (WH) gelatin so largely used by American bacteriolo- gists, showing a reaction with iodin like that given by sulfites. As sulfurous acid may be used for bleaching, some of this might remain in the finished product. It would be, of course, very objectionable in culture-media. To avoid this impurity, each lot of gelatin should be tested by allowing a weighed amount (say lo grams) to soak overnight in cold water, then completing the solution by gentle warming, allowing the liquid 74 ANALYTIC OPERATIONS. to cool, and titrating with centinormal iodin in presence of sulfuric acid and starch solution, as in the usual process for estimation of total sulfites in food products. A sample of gelatin so tested should require less than 2 c.c. of the iodin solution. It is not unlikely that the preference for imported gelatin involves a useless expense to bacteriolo- gists. I have found a sample of American manu- facture known as '' Marblehead" to show a very slight reaction with iodin. Preparation of Culture-media : Bouillon is the term applied to many forms of liquid media, prepared with meat-juice or meat- extract. The ordinary bouillon is prepared ac- cording to the following formula : Five hundred grams of finely chopped meat, as free as possible from fat and gristle, are soaked overnight in about a liter of cold water, at a temperature between 0° C. and 10^ C. The mass is then strained thru a coarse towel and pressed until as much as possible of the liquid is obtained. To this is added 10 grams of peptone and 5 grams of common salt. It is then heated to boiling, best in the open-steam sterilizer, to coagulate albumin, after which it is filtered. The most difficult point in the work is neutralization. This is often ac- complished by the use of sodium carbonate, which is added in small amounts until the liquid no longer affects red litmus paper. The better I SANITARY EXAMINATIONS. 75 method is to titrate a portion of the bouillon with sodium hydroxid solution, and calculate from this the amount of that solution necessary to neutral- ize the whole of the liquid. Fuller has devised a good method of procedure. The bouillon is made up when cool to a definite volume, say looo c.c. ; 5 c.c. are mixed with 45 c.c. of distilled water in a porcelain dish, boiled for three minutes, i c.c. of solution of phenolphthalein added, and quickly titrated with twentieth normal sodium hydroxid. The neutral point is the slight pink color not dis- appearing on gentle stirring. From the number of cubic centimeters used the amount of alkali needed to neutralize the whole solution is cal- culated, but this alkali should be added in the form of normal solution in order to avoid much dilution of the bouillon. Meat-extract is often used instead of the in- fusion of chopped meat. Five grams of a good commercial extract are used for each 1000 c.c. of bouillon. Bouillon may be modified in many ways, by the addition of different substances, but the inherent or possible acidity or alkalinity of these must be ascertained and corrected if culture results are to be kept standard. Dextrose Bouillon. For special -fermentation work a solution is made by adding dextrose in the proportion of 20 grams to 1000 c.c. of the liquid. 76 ANALYTIC OPERATIONS. Gelatin Media. — The ingredients, other than the gelatin, are dissolved and treated as described in the making of bouillon. After neutralization, the gelatin is dissolved by gentle heating. If this contributes any acidity, it must be neutralized. The liquid should not be heated strongly or for a long time, as the gelatinizing property may be injured. The solution is made up to the proper volume and filtered thru paper. Meat-extract peptone-gelatin: Meat-extract, 5.0 grams Peptone, lo.o '' Gelatin, 1 50.0 '' Dextrose, 2.0 '' Sodium chlorid, 5.0 " Water, 1 000.0 c.c. Agar Media: The preparation of agar solution is more diffi- cult than that of gelatin. Several methods have been suggested. Ravenel uses the following : Preferable Method: (A) Chopped meat, 500 grams Water, 500 c.c. These are mixed and allowed to stand over- night. (B) Agar,.". 12 grams Water, 500 c.c. SANITARY EXAMINATIONS. 77 Solution B is put into the autoclave, the pres- sure run up to 2 atmospheres, the heat withdrawn, and the boiler opened when the temperature has fallen a little below ioo° C. The solution is allowed to cool to about 75° C. (below the co- agulating point of albumin) 10 grams of dried peptone and 5 grams of sodium chlorid are added, A and B mixed, the liquid boiled for about three minutes, neutralized and filtered. The filtration is very quick — from ten to twelve minutes for 1000 c.c. A hot-water funnel is not needed, but the filter must be moistened with boiling water immediately before pouring in the agar. In the process with fresh meat the clarification is effected by the coagulation of the albumin in the meat- water, hence solution B must not be added to A until cool enough to avoid coagulation. Alternative method: (A) Dried peptone, . . .- 10 grams Common salt, S '' Meat-extract, 5 ** Water, 500 c.c. Boil for three minutes and neutralize. (B) Agar-agar, 12 grams Water, 500 c.c. The agar is chopped fine and heated in the auto- clave to two atmospheres. As soon as this pres- sure is reached, the heat is withdrawn and the 78 ANALYTIC OPERATIONS. liquid allowed to cool until below ioo° C. before opening. The two solutions A and B are then mixed, cooled to 60^ C, the whites of two eggs beaten in 50 c.c. of water added, well stirred in, and the whole then boiled and filtered thru paper. Instead of the white of egg, blood-serum may be used, which seems to add also to the nutritive value of the medium. Agar made with meat- extract will often form a precipitate during the sterilization. Abbott gives the following method of preparing agar solution: The bouillon is prepared and neutralized in the usual way, then 15 grams of finely chopped agar are added, and water suffi- cient to make the volume 1250 c.c. The mass is boiled gently over a direct flame, stirring occa- sionally for several hours. If the fluid goes below the liter level, enough water should be added to make up the amount. The boiling should be con- tinued until about 1000 c.c. is left in the vessel. When the solution of the agar is attained, the: vessel is placed in a large dish of cold water untill it has cooled to about 70° C, the white of one egg that has been beaten up with water added, mixedl well, and boiled again for a half -hour, avoiding' the evaporation of the liquid below 1000 c.c. The liquid is filtered thru heavy folded filter-paper at room-temperature. It is necessary that the solution should be not above 70° C. when the SANITARY EXAMINATIONS. 79 white of egg is added or it will become lumpy. Commercial egg-albumin in lo per cent, solution in water may be used instead of white of egg. The solution thus prepared should filter rapidly. Potato Culture, — Cultivation on potatoes has been much used as a method of distinguishing certain microbes. Large, sound potatoes should be selected, thoroly washed, and cut into disks about five centimeters in diameter and one cen- timeter thick. These are placed in glass boxes (pomade boxes) which have lids with ground joint, and heated for about one-half hour in the sterilizer. Another method is to cut out cylinders with the aid of an apple -corer, or largest size cork-borer, slice these obliquely, and place them in test-tubes, which are then closed with cotton plugs and sterilized. The latter method does not give a large surface, but the growth of any in- oculation may be easily watched: Milk Culture. Milk, deprived of most of its fat, is used for the purpose of detecting microbes that produce notable amounts of acid. The fol- lowing is the procedure generally recommended: Milk as pure as can be obtained is allowed to stand over night in a refrigerator, the cream removed, and the skimmed milk siphoned off from any sediment. It should show not more that i% of acidity to normal alkali (that is, loo c.c. of the milk should not require more than i c.c. of normal 8o ANALYTIC OPERATIONS. sodium hydroxid to neutralize, using phenol- phthalein as an indicator) . Whqn the proper con- dition in this respect has been obtained, the milk may be sterilized in tubes as usual. Azolitmin solu- tion in small amount may also be added if desired. The tubes containing the sterilized milk are inoculated with small amounts of the sample, to be tested, kept for twenty-four hours or even longer, at blood heat, and then tested by heating to the boiling-point. If coagulation occurs, acid- producing microbes are present. Many special forms of culture-media are em- ployed for bacteriologic investigations which do not come within the line of water analysis, and need not be described. One form is much em- ployed in the search for the specific germ of ty- phoid fever, namely, Wurtz's litmus-lactose me- dium. This may be with either agar or gelatin, in conjunction with meat-extract. The nutrient medium must be made so as' to possess such a degree of alkalinity that lo c.c. will neutralize 0.5 c.c. of decinormal sulfuric acid. Lactose is added in the proportion of two or three grams to 100 c.c. of medium and the mixture sterilized, after which sufficient sterilized azolitmin solution is added to give the fluid a distinct but not deep-blue tint. Cultivation in gelatin at ordinary temperatures usually yields a larger number of points of mi- crobic life than in agar at blood-heat. This is SANITARY EXAMINATIONS. 8l due to the fact that many common water-bacteria do not grow well at the higher temperatures. Culture-media when ready for use are distri- buted in test-tubes. These must be well cleaned. In laboratories in which regular chemical work is done, the solution of crude chromic and sulfuric acids used for voltaic batteries is a good cleaning agent, the tubes being soaked in this for about a day, and then rinsed thoroughly and sterilized as noted below. In bacteriologic laboratories it is usual to cleanse the tubes with a 3 per cent, solu- tion of sodium hydroxid. The tubes are boiled in the solution, rinsed, swabbed out with a brush, and allowed to dry in the inverted position. A cotton- wool plug is made for each tube, care being taken that it fits neatly, without creases or channels and not too tightly. The projecting part of the plug is clipped moderately close and a tinfoil cap placed on each. The arrangement is sterilized in the hot-air oven at 150° C. When cold, the tinfoil and plug are carefully removed, about 10 c.c. of culture-medium put into each tube, with as little outside contamination as possible, the plug and cap replaced, and the tubes and contents sterilized in the open-steam sterilizer. Wire baskets are used to hold the tubes during the sterilizations. For making cultures definite volumes of the water sample are introduced into the culture- medium, and if this is a solidifying form, it is put 52 ANALYTIC OPERATIONS. into the Petri dish. All the manipulations must be conducted with great care to avoid contamina- tion. When there is no clue to the amount of microbes present, it will be necessary to make cultures with different proportions of water. Some tubes may be inoculated with a few drops, some with three to five drops, and some with i c.c. Some operators dilute the water considerably and take small measured volumes. If this method is used, the diluting water must be distilled and have been well sterilized, by at least five minutes' boil- ing, and cooled out of contact of air. All pipets, dishes, and other apparatus that come in contact with the water must have been first sterilized. The test-tubes containing the culture-medium are warmed gently, just enough to render the culture-medium fluid, the desired volume of the water added, the mixture shaken and promptly poured into the petri dishes, covered and placed in the oven, which should have been already raised to the temperature at which it is desired to conduct the work. Unless specially desired otherwise, cultures should be made in the dark. They may be made at any temperature short of that at which the medium melts, but either ordinary temperature or 37° C. is usually selected. The condition of the plates should be observed at intervals of twenty- four hours, and the points of microbic life counted and recorded. After some days the growth will SANITARY EXAMINATIONS. S^ become ?o luxuriant or the liquefaction of the medium so extensive that accurate observation is not possible. If the microbic points are numerous, it will be necessary to employ a counting scale. For the petri dish, Pake's modification of Lafar's scale is cheap and sufficient. General Character of the Microbes in Natural Waters. — The microorganisms of natural waters are principally included in the genera Bacillus and Spirillum, especially the former. Microbes are differentiated to some extent by their action upon the culture -medium. Some species rapidly or slowly liquefy the jelly with evolution of foul-smelling gases; others produce characteristic colors. Many do not produce any positive modification, and for purposes of dis- tinction it is usual to transfer portions of the colonies to other culture-media. Such special cultures are obtained by taking up a portion of the colony on the end of a wire which has been just sterilized by heating to redness and in- oculating the prepared medium. Indol Reaction. — Indol, more properly indin, is a nitrogenous substance produced by growth of many species of microbes, and the detection of it may, therefore, be utilized as a differentiation test. The following is a method for performing the test : Ten c.c. of a peptone infusion, previously inocu- 84 ANALYTIC OPERATIONS. lated with the microbes to be tested, and kept for twenty-four hours at blood-heat, are treated with I c.c. of solution of pure potassium nitrite (0.02 gram in 100 c.c.) and then with a few drops of concentrated sulfuric acid. In the presence of indol a rose or deep-red color is developed. Spi- rillum cholercB and Bacillus coli communis give the reaction strongly; 5. Finkleri feebly; the so- called B. typhosus ordinarily does not give it. The reagents and the culture-medium to be tested should be quite cold and the mixture should stand for about an hour before deciding upon the result. Bacteriologists frequently fail to isolate the typhoid bacillus from natural waters. In default of a method for such detection, resort is had to methods for the detection of the microbe known as Bacillus coli communis (often called ** colon bacillus''). This being a usual inhabitant of the intestinal canal of the higher animals, and being almost always associated with dangerous pollu- tion, the detection of it in water indicates previous contamination. This organism is not constant in character, but, in common with many other dangerous microbes, grows well at blood-heat, while many common water, air, and soil organisms do not. Theobald Smith proposed a method of cul- tivating the water sample at blood-heat with a bouillon containing dextrose (see page 75) and I SANITARY EXAMINATIONS. 85 noting the amotint of gas evolved. The operation is carried out in a tube similar to that used in the fermentation test for sugar in urine, but slightly larger. The upright part of the tube should have a capacity of about 15 c.c, and the bulb should be nearly 4 centimeters in diameter. In judging of the amount of gas produced, the tube must be allowed to stand for at least half an hour at room-temperature, as the volume is much in- creased by even slight heating. Sufficient bouillon is put in to fill the upright stem and the curved part, but very little of the bulb. The whole is then well sterilized in the steam sterilizer, a cotton plug with tinfoil cover having been previously placed in the opening of the bulb. The apparatus is cooled ; a few drops of the water sample are introduced into the bouillon, taking care not to allow outside contamination, the plug and tinfoil are replaced, and the tube kept at 37° C. for about forty hours. An accumulation of gas at the top of the tube indicates that microbes of the type of the B. coli communis are present. Several tubes and one or two control tubes, that is, tubes which are not inoculated with water, should be tried together. Portions of the gas- producing bouillon may be inoculated into sterile agar and cultivated at the same temperature after pouring into the petri dish. The latest and apparently the most satisfactory 86 ANALYTIC OPERATIONS. medium for the fermentation test is that pro- posed by D. D. Jackson. This is a mixture of ox-bile and lactose. Jackson found that the inspissated preparations commonly sold are not satisfactory. Fresh ox-bile, however, may be filtered and sterilized (Jackson uses the auto- clave at 15 pounds for 30 minutes), after which it can be kept for a long while. Bile also may be filtered, evaporated to dryness, and kept in this form. On an average, 1000 c.c. of bile will yield no grams of residue and, therefore, 11 grams of the residue will suffice to make 100 c.c. of the culture solution. One gram of lactose is added to each 100 c.c. of bile solution. The mixture is placed in the fermentation tubes and sterilized. The gas evolved by the colon bacillus in these media is principally a mixture of carbon dioxid and hydrogen in the proportion of one volume of the former to two of the latter, but other gases may be present in small amount. The ratio is not always the same. Attempts to draw conclusions from a change in ratio have not been satisfactory. The present opinion of bacteriologists is that a water that shows marked gas production with the fermentation method, especially the lactose-bile solution, is contaminated with the colon bacillus or closely allied organisms. It is customary to make tests on different vol- umes of the sample to get an approximation to the number of bacilli present. Thus one or more SANITARY EXAMINATIONS. 87 tub.es are inoculated with i c.c. each of the sample ; other tubes with the same volume more or less diluted. Thus, if lo c.c. of the sample are mixed with 90 c.c. of sterilized water and a culture tube inoculated with i c.c. of this mixture, gas pro- duction will show that the bacillus is present in 0.1 c.c. of the sample. According to the same method a culture of cooi c.c. may be made. If the sample is supposed to contain very few bacilli, cultures may be made with 5 c.c. or even 10 c.c. Much attention has been paid to the distinct- ions between the colon bacillus and the typhoid bacillus. The following synopsis has been given by Abbott : Characteristics. B. typhosus. B. coli communis. Motility, Conspicuous. Not marked. Growth in gelatin, . Slow. Not very slow. " " potato, ... Usually inconspicu- Always rapid and ous. visible. " " milk, .... No coagulation ; Acidity and coagu- no acidity. lations in forty- eight hoiu-s in in- cubator at blood heat Growth in media con- taining dextrose, lactose, or su- crose, No evolution of gas. Marked evolution of gas. Growth in media containing lactose and litmus, Colonies pale blue ; Colonies pink ; sur- no reddening of rounding medium medium. red. Indol reaction in pep- tone solution (for- ty-eight hours at 37° C.) Rarely present. Always present. 88 ANALYTIC OPERATIONS. D. D. Jackson and T. W, Melia have devised a process for isolation of B, typhosus, taking ad- vantage of the fact that when B. coli communis and B. typhosus grow in the bile-lactose medium, the acid produced soon acts as a restraining agent on the former, thus giving opportunity for the latter to increase. A description of this method was presented at the Winnipeg (1908) meeting of the American Public Health Association, and I am indebted to Dr. Jackson for permission to transcribe from the manuscript, in advance of publication, the method employed. It is one of the most important advances in the sanitary ap- plication of culture methods made in recent years. An essential point is upon the use of a special agar suggested by Hesse. The formula is as follows: Dry agar 4.5 grams. Peptone 10 '' Extract of beef 5 '' Sodium chlorid 8.5 '' Distilled water 1000 c.c. The agar must be dried for thirty minutes at 105° C. Jackson and Melia found that commer- cial agar contains considerable moisture, and if weighed in this condition, the medium may not be of the proper consistence. It was also found that Liebig's extract was more satisfactory than others. The medium must be stored in an ice-chest, the SANITARY EXAMINATIONS. 89 atmosphere of which is saturated with moisture, and the culture must be carried out at 37° C, also in an atmosphere so saturated. The following are the details of the preparation of the medium. The requisite amount of the dried agar is dissolved in 500 c.c. of water by heating over a free flame. In another vessel the peptone, salt, and meat extract are dissolved in the remaining water by the aid of heat. The two solutions are mixed, boiled for thirty minutes, cooled somewhat, the loss by the different boilings made up by adding distilled water, and the liquid filtered thru absorbent cotton held' in the funnel by cotton flannel. The liquid should be clear, and may require more than one filtration to se- cure this. The medium should be adjusted in reaction to not more than 1% normal acid (see page 75). The medium should be distributed in tubes containing 10 c.c. each, sterilized at 120° C. (15 pounds) for twenty minutes. When the auto- clave can be opened, the tubes should be cooled as quickly as possible in running water, and stored, as noted above, in a cold atmosphere saturated with moisture. The test is carried out as follows : The sample to be examined is cultivated for at least twenty-four hours in the bile-lactose medium (see page 86). Eight tubes, each containing 9 c.c. of sterilized dis- tilled water, are arranged in a rack in convenient 90 ANALYTIC OPERATIONS. relation with eight petri dishes, each set being num- bered consecutively. In tube i is placed i c.c. of bile-lactose culture. The contents of tube i are well mixed, and i c.c. of the liquid is placed in dish I, and i c.c. in tube 2. The contents of tube 2 are mixed, i c.c. transferred from tube 2 to dish 2 and i c.c. to tube 3. The dilution is thus carried out until the series is completed. Each dish is then charged with 10 c.c. of Hesse agar, which has been melted and cooled to 40° C, the contents of each dish well mixed, all the dishes chilled in the ice-chest until the contents are solid, and then incubated in a moist atmosphere for twenty-four hours at 37° C. By this method, characteristic growths are ob- tained from B. typhosus, but only when the dilu- tion is high enough to give but few colonies to a dish. It is distinguished from B. colt communis by forming colonies of much larger size, often several centimeters in diameter, showing a broad translucent or scarcely turbid zone between the white center and the narrow white edge. The colonies may be taken off for identification by other cultures, or microscopic examination, accord- ing to the data on page 87. TECHNIC EXAMINATIONS. 9I TECHNIC EXAMINATIONS. GENERAL QUANTITATIVE ANALYSIS. Silica, Iron, Aluminmn, Manganese, Calcium, and Magnesium. — looo c.c. of the water slightly acidified with hydrochloric acid are evaporated to complete dryness, best in a platinum dish, the residue treated with hydrochloric acid and water, and the separated silica filtered, washed, dried, ignited in a platinum crucible, and weighed. To the filtrate, previously boiled with a few drops of strong nitric acid, slight excess of am- monium hydroxid is added, the liquid boiled several minutes, the precipitate collected, washed thoroughly with boiling water, dried, ignited, and weighed. It consists chiefly of FCjOgand AI3O3, contains all the phosphates and some manganese if much is present in the water. In such cases the precipitate before drying is redissolved in hydro- chloric acid and neutralized with a dilute solution of ammonium carbonate until the water becomes almost turbid. It is then boiled, and the precipi- tate, now free from manganese, washed, dried, ignited, and weighed. The iron may be deter- mined by dissolving the precipitate in strong hydrochloric acid and employing the colorimetric method described on page 57. If no manganese or only traces are present, the 92 ANALYTIC OPERATIONS. filtrate from the iron is mixed with sufficient am- monium chlorid to prevent the precipitation of the magnesium ammonium hydroxid, and then am- moniura oxalate added in quantity sufficient to precipitate the calcium and to convert all the magnesium into oxalate, and thus hold it in solu- tion. The precipitate contains all the calcium and some of the magnesium. If the magnesium is present only in relatively small quantity, the amount carried down may be disregarded ; other- wise a second precipitation should be made as follows : The solution is allowed to stand until the precipitate has subsided; this will require some hours. The supernatant liquid is poured off thru a filter, the precipitate washed by decantation, then dissolved in hydrochloric acid, water added, then ammonium hydroxid and a small quantity of ammonium oxalate. After the calcium oxal- ate has subsided it is filtered off, washed, and dried. If quite small in amount, it is placed with the filter in a weighed platinum crucible, ignited over the Bunsen burner for a short time, and then over the blast lamp for from five to fifteen minutes. The calcium is thus obtained in the form of oxid, which is allowed to cool in the desiccator and weighed. The weight thus ob- tained multiplied by 0.7147 gives the weight of calcium. When the amount of precipitate is large, it is better to remove it from the filter and TECHNIC EXAMINATIONS. 93 heat it just short of redness until it assumes a grayish tint. It then consists of calcium car- bonate. To this is added the ash of the filter. The weight of the calcium carbonate multiplied by 0.4 gives the weight of calcium. The calcium may be determined by titration. The precipitate of calcium oxalate, after thorough washing, is dissolved from the filter by warm, dilute, sulfuric acid, heated to about 65° C, and titrated with decinormal permanganate until a pink tint is obtained. One c.c. of the perman- ganate is equivalent to 0.0020 calcium; 9.0028 calcium oxid; 0.0050 calcium carbonate. The filtrates are mixed, slightly acidified with hydrochloric acid, concentrated and cooled, am- monium hydroxid and sodium phosphate added in excess, stirred briskly, and allowed to stand in the cold for about twelve hours. The precipitated ammonium magnesium phosphate is brought upon a filter, that adhering to the sides of the vessel being dislodged by rubbing with a glass rod tipped with a piece of clean rubber tubing. It is washed with a solution made by mixing one part of the ammonium hydroxid of 0.96 sp. gr. with three parts of water. The precipitate is dried, trans- ferred to a platinum crucible, the filter ashed separately and added to it, and the whole heated at first gently and then to intense redness for several minutes. After cooling it is weighed. 94 ANALYTIC OPERATIONS. It consists of magnesium pyrophosphate; the weight multipHed by 0.2187 gives the weight of magnesium. Manganese, if present in appreciable quantity, is separated before the precipitation of the cal- cium, as follows: The filtrate from the iron pre- cipitate is slightly acidulated with hydrochloric acid, concentrated, and the manganese precipi- tated as sulfid by colorless or slightly yellow solu- tion of ammonium sulfid. The flask, which should be nearly full, is stoppered, allowed to rest in a moderately warm place until the precipitate has thoroughly settled, filtered, washed with dilute ammonium sulfid water, and purified by dissolv- ing in a small quantity of hydrochloric acid and reprecipitating with ammonium sulfid. It is filtered off, washed as before, dried, placed in a weighed porcelain crucible, covered with a little sulfur, and ignited in a current of hydrogen intro- duced into the crucible by a tube passing thru a hole in the cover. The pure manganese sulfid thus obtained is allowed to cool and weighed. The weight multiplied by 0.631 gives manganese. Sulfates. — Five hundred c.c. of the clear water are slightly acidulated with hydrochloric acid, heated to boiling, and barium chlorid solution added in moderate excess. The precipitated barium sulfate is allowed to subside completely, collected upon a filter, washed thoroly, dried, TECHNIC EXAMINATIONS. 95 and incinerated. The weight multipHed by 0.42 gives SO. If the proportion is very low, it will be advisable to concentrate the water to one-fifth or one-tenth its bulk before precipitating. Control. Potassium, Sodium, and Lithium. — From 250 to looo c.c. of the water, according to the amount of solid matters present, are evapo- rated to dryness in a platinum dish, and the residue treated with a small amount of water and suffi- cient dilute sulfuric acid to decompose the salts present. The dish should then be covered and placed upon the water-bath for five or ten minutes, after which any liquid spurted on the cover is washed into the dish, the whole evaporated to dryness and heated to redness. A few drops of ammonium carbonate solution should then be mixed with the residue, and the ignition repeated to insure the removal of the last portions of free acid. In the ifiajority of cases the only basic elements present in considerable quantity are calcium, magnesium, and sodium. The sodium may be determined indirectly, therefore, by cal- culating from the amount of calcium and mag- nesium found, the calcium and magnesium sulfate in the residue, and subtracting this sum, together with the silica, from the total residue. For the determination of potassium and sodium in ordinary well and river waters, not less than 2000 c.c. should be employed. When lithium is 96 ANALYTIC OPERATIONS. to be determined, it is generally necessary to use much more. In any case, as the alkalis are to be weighed as chlorids, it is advisable, if notable amounts of sulfates are present, to precipitate them by addition of barium chlorid. The water is evaporated to about 200 c.c, a slight excess of pure calcium hydroxid added to the hot liquid, — generally a few c.c. of thin milk of lime will be sufficient, — and the heat continued for several minutes. It is then washed into a 250 c.c. flask, disregarding the insoluble portion adhering to the dish, which, however, should be thoroly washed, and the washings added to the flask. After cooling, the flask is filled up to the mark with distilled water, thoroly mixed, the precipitate allowed to settle, and the liquid fil- tered thru a dry filter. Two hundred c.c. of the filtrate are measured into another 250 c.c. flask, ammonium carbonate • and ammonium oxalate added, filled with water up to the mark, mixed, allowed to settle, filtered thru a dry filter, 200 c.c. of the filtrate measured off, and evaporated to complete dryness in a platinum crucible, heating very cautiously at the last stages, to avoid loss by spurting. The low-temperature burner is suited for this purpose. The crucible is now covered and cautiously heated to dull redness, cooled, and weighed. The residue contains the potassium, lithium, and sodium as chlorids. It TECHNIC EXAMINATIONS. 97 contains, sometimes, also, traces of magnesium, which may be removed by treating again with lime and ammonium carbonate and oxalate. It is frequently of advantage, in evaporating these saline solutions, to add, when the solution be- comes Concentrated, several cubic centimeters of strong hydrochloric acid. This precipitates the greater portion of the salts in a finely granular condition, and renders loss by spurting less liable to occur. If potassium and sodium chlorids only are present, the residue is dissolved in a small quan- tity of water, an excess of a concentrated neutral solution of platinum chlorid added, evaporated to small bulk at a low heat on the water-bath, some eighty per cent, alcohol added, allowed to stand, the clear liquid decanted off on a small filter, and the residue washed in this way several times by fresh, small portions of eighty per cent, alcohol. The precipitate is then washed on to the filter with alcohol, washed again with eighty per cent, alcohol, dried well, and transferred as far as possible to a watch-glass. The small portion on the filter is dissolved off and the solution placed in a weighed platinum dish and evaporated to dry- ness. The main portion on the watch-glass is then added, and the whole dried to a constant weight at about 260° F., cooled, and weighed. The weight thus found, multiplied by 0.307, gives 7 98 ANALYTIC OPERATIONS. the weight of potassium chlorid. This, subtracted from the combined weight of the chlorids, gives the weight of sodium chlorid. Lithium, if present, is best separated before the treatment with platinum chlorid. The follow- ing method, devised by Grooch, gives good re- sults : To the concentrated solution of the weighed chlorids, amyl alcohol is added and heat applied, gently at first, to avoid bumping, until the water disappears from the solution and the point of ebullition becomes constant at a temperature which is approximately that at which the alcohol boils (182° C), the potassium and sodium chlorids are deposited, and the lithium chlorid is dehy- drated and taken into solution. The liquid is then cooled, and a drop or two of strong hydro- chloric acid added to reconvert traces of lithium hydroxid in the deposit and the boiling continued until the alcohol is again free from water. If the amount of lithium chlorid be small, it will be found in the solution and the potassium chlorid and sodium chlorid in the residue, excepting traces which can be allowed for. If the lithium chlorid exceed ten or twenty milligrams, the liquid may be decanted, the residue washed with amyl alco- hol, dissolved in a few drops of water, and treated as before. For washing, amyl alcohol, previ- ously dehydrated by boiling, is to be used and the filtrates are to be measured apart from the TECHNIC EXAMINATIONS. 99 washings. In filtering, the Gooch filter with asbestos felt may be used with advantage, apply- ing gentle pressure by the aid of the filter-pump. The crucible and residue are ready for weighing after gentle heating over the low-temperature burner. The weight of the insoluble chlorids is to be corrected by adding 0.00041 for every 10 c.c. of amyl alcohol in the filtrate, exclusive of the washings, if only sodium chlorid be present; 0.00051 for every 10 c.c. if only potassium chlorid, and 0.00092 in the presence of both these chlorids. The filtrate and washings are evaporated to dryness in a platinum crucible heated with sul- furic acid, the excess driven off, and the residue ignited to fusion, cooled, and weighed. From the weight is to be subtracted, for each 10 c.c. of filtrate, 0.0005, 0.00059, ^^ 0.00109, according as only sodium chlorid, potassium chlorid, or both were present in the original mixture. Hydrogen Sulfid. — The following method is taken from Sutton's ''Volumetric Analysis'': Reagents Required : Centinormal lodin. — Dry, commercial iodin is intimately mixed with one-fourth its weight of pure potassium iodid and gently heated between two clock-glasses by resting the lower on a hot plate. The iodin sublimes in a perfectly pure condition. It is allowed to cool under the desic- cator, 1.2797 grams weighed out, together with lOO ANALYTIC OPERATIONS. 1.8 grams of pure potassium iodid, dissolved in about 50 c.c. of water, and the solution made up exactly to 1000 c.c. The liquid must not be heated, and care should be taken that no iodin vapor is lost. One c.c. is equivalent to 0.00017 H2S. The solution is best prepared in stoppered bottles, which should be completely filled and kept in the dark. It will not even then keep very- long, and should be standardized by titration with a weighed amount of pure sodium thiosulfate, which should be powdered previous to weighing, and pressed between filter-paper to absorb any moisture. Fifty c.c. of the iodin solution, when of full strength, will require 0.124 gram of sodium thiosulfate. Starch Indicator. — See page 47. Analytic Process : Ten c.c, or any other necessary volume of the iodin solution, is measured into a 500 c.c. flask, and the water to be examined added until the color disappears. Five c.c. of starch liquor are then added, and the iodin solution run in until the blue copper appears; the flask is then filled "tb the mark with distilled water. The respective vol- umes of iodin and starch solution, together with the added water, deducted from the 500 c.c, will show the volume of water actually titrated by iodin. A correction should be made as follows for the excess of iodin required to produce the blue TECHNIC EXAMINATIONS. lOI color: Five c.c. starch solution are made up with distilled water to 500 c.c, iodin run in until the color matches that in the test, and the volume of iodin solution so used subtracted from the figure obtained in the first titration. Hardness. CO 3 in Normal Carbonates. — Waters containing considerable quantities of calcium and magnesium are said to be hard. Since the solu- tion of calcium and magnesium carbonate in water depends partly upon the presence of carbon dioxid, boiling precipitates the greater portion of the carbonates, the result being to diminish the hardness — i, e., to soften the water. Magnesium and calcium sulfates and chlorids are not precipi- tated in this way. Hardness, therefore, is divided into two classes, temporary and permanent, the former being that which may be removed by boiling. The determination may be made by titration with acid or soap solution. Acid Titration Method. — This method is due to Hehner. His description is followed, but there seems to be no reason why the standard acid can not be made by diluting decinormal acid with three times its bulk of water. It would also be advisable to try the alizarin indicator (see p. 15) in place of methyl orange. Reagents Required : Standard Sodium Carbonate. — 1.061 grams of recently ignited pure sodium carbonate are dis- I02 ANALYTIC OPERATIONS. solved in water and the solution diluted to looo c.c. One c.c. = 0.00106 gram Na2C03, equiv- alent to o.ooi gram CaCOg. Standard Sulfuric Acid. — One c.c. of pure con- centrated sulfuric acid is added to about 1000 c.c. of water. Fifty c.c. of the standard sodium car- bonate are placed in a porcelain dish, heated to boiling, a few drops of a solution of methyl orange added, and the sulfuric acid cautiously run in from a buret until the proper change of color occurs. From the figure thus obtained, the ex- tent to which the acid should be diluted in order to make i c.c. of the sodium carbonate equivalent to I c.c. of the acid may be calculated. The proper amount of water is then added, and the solution verified by again titrating. Analytic Process : Temporary Hardness. — One hundred c.c. to 250 c.c. of the water tinted with the indicator are heated to boiling, and the sulfuric acid cautiously run in until the color-change occurs. Each cubic centimeter required will represent one part of calcium carbonate or its equivalent per 100,000 parts of the water. Permanent Hardness, — To 100 c.c. of the water is added an amount of the sodium carbonate solution more than sufficient to decompose the calcium and magnesium sulfates, chlorids, and nitrates present; usually a bulk equal to the TECHNIC EXAMINATIONS. IO3 water taken will be more than sufficient. The mixture is evaporated to dryness in a nickel or platinum dish, and the residue extracted with distilled water. The solution is filtered thru a very small filter, and the filtrate and washings titrated hot with sulfuric acid as above; or 25 c.c. of distilled water may be poured on the residue, and the solution obtained filtered thru a dry filter, the filtrate measured and titrated. The difference between the number of cubic centimeters of sodium carbonate used and the acid required for the residue will give the permanent hardness. If the water contains sodium or potassium car- bonate, there will be no permanent hardness, and there will be more acid required for the filtrate than the equivalent of the sodium carbonate added. From this excess the quantity of sodium carbonate in the water may be determined. Since any alkali carbonate in the water would be erroneously calculated as temporary hardness by the direct titration, the equivalent, in terms of calcium carbonate, of the alkali carbonate present should be deducted from the figure given by the titration in order to get the true temporary hard- ness. The total CO 3 in normal carbonates is given by the direct titration of the water with dilute sulfuric acid. One c.c. of the acid is equivalent to 0.0006 gram of CO3. I04 ANALYTIC OPERATIONS. Soap Titration Method. — This is done with a solution of soap in dilute alcohol. Commercial soap is irregular in composition, even apart from intentional adulteration. '' Castile" soap is re- commended, but it must be borne in mind that this title has now no positive value. Good soaps can be obtained from responsible dealers. Trans- parent or dark soaps should not be used. The following method has been used largely in the laboratory of the State Board of Health of Massachusetts and will be found satisfactory. Reagents Required : Standard Calcium Chlorid Solution. — 0.2 gram of pure calcium carbonate is dissolved in dilute hydrochloric acid in a porcelain dish, the solution evaporated to dryness, redissolved and reevapor- ated, until a perfectly neutral salt remains. This is dissolved in water and made up to 1000 c.c. One c.c. contains calcium equivalent to 0.0002 gram calcium carbonate. Soap Solution. — Castile soap is cut into thin shavings, dried, 10 grams weighed out and dis- solved in a mixture of fifty c.c. of alcohol (95%) and 50 c.c. of water, and the solution allowed to settle over night. Fifty c.c. of it should be diluted to 1000 c.c. principally with water, but using enough alcohol to prevent soap separating. The liquid may be slightly turbid, in which case it may be filtered. As alcohol and water produce slight TECHNIC EXAMINATIONS. 10$ contraction when mixed, the solution will contain a little over 5 grams to 1000 c.c. Fifty c.c. of the standard solution of calcium chlorid, which, according to the table, should take exactly 14.25 c.c. of standard soap, are used to test the strength. The soap solution thus prepared does not change perceptibly if air fias no access to it, and, if used with a siphon buret attached to the bottle, will keep for five or six weeks or longer. For the standardization of the soap, and for the determination of the hardness of any water, 50 c.c. of the sample or of the standard calcium chlorid solution are placed in a flask or bottle of 200 c.c. capacity, and of a convenient shape, and the soap solution added, 0.2 or 0.3 c.c. at a time, shaking well after each addition, until a lather is obtained which is permanent for five minutes and covers the entire surface of the liquid with the bottle placed on its side. The soap solution must be added in small quantities, especially in the presence of mag- nesium compounds. If much carbonic acid is liberated, it is well to remove it by suction. The table (page 106) does not apply to hardness above 12.5. If the water tested requires more than 10 c.c. of the standard soap solution, a smaller portion of 25 c.c, 10 c.c, or even 2 c.c, as the case may require, is measured out and made up to a volume of 50 c.c with recently distilled water. io6 ANALYTIC OPERATIONS. This will keep the results comparable with each other, altho the dilution introduces some error into the calculation. The following table gives the hardness cor- responding to the number of cubic centimeters of soap solution used in the analyses : C.C. OF C.C. OF C.C. OF Soap Soap Soap Solu- Hard- Solu- Hard- Solu- Har tion. ness. tion. ness. tion. NESS 0.7 0.0 - 3.7 4.1 6.7 8.4 0.8 0.1 3.8 4.2 6.8 8.5 0.9 0.3 3.9 4.4 6.9 8.7 I.O 0.4 4.0 4.5 7.0 8.8 I.I 0.6 4.1 4-7 7.1 9.0 1.2 0.7 4.2 4.8 7.2 9.1 1.3 0.9 4.3 5.0 7.3 9.2 1.4 I.I 4.4 5.1 7.4 9.4 1.5 1.2 4.5 5.2 7.5 9.5 1.6 1.4 4.6 5.4 7.6 9.7 1.7 1.5 4.7 5.5 7.7 9.8 1.8 1.6 4.8 5.7 7.8 lO.O 1.9 1.8 4.9 5.8 7.9 lO.I 2.0 1.9 5.0 6.0 8.0 10.3 2.1 2.0 5.1 6.1 8.1 10.4 2.2 2.2 5.2 6.2 8.2 10.6 2.3 2.3 5.3 6.4 8.3 10.7 2.4 2.4 5.4 6.5 8.4 10.9 2.5 2.6 5.5 6.7 8.5 II. 2.6 2.7 5.6 6.8 8.6 II. 2 2.7 2.8 5.7 7.0 8.7 II-3 2.8 2.9 5.8 7.1 8.8 11.5 2.9 3.1 5.9 7.2 8.9 11.6 3.0 3.2 6.0 7.4 9.0 11.8 3.1 3.3 6.1 7.5 9.1 11.9 3.2 3.5 6.2 7.7 9.2 12.1 3.3 3.6 6.3 7.8 9-3 12.2 3.4 3.7 6.4 8.0 9.4 12.4 3.5 3.9 6.5 8.1 9.5 12.5 3.6 4.0 6.6 8.2 TECHNIC EXAMINATIONS. 107 If the hardness of a water is given as 9.0, it means that in 100,000 pounds of water there is of calcium and magnesium salts a quantity which gives the same hardness to water which would be given by nine pounds of calcium carbonate. In order to soften this water for manufacturing purposes, about nine pounds of soda ash will be required, and for laundry purposes about ninety pounds of soap. Free and Half-bound Carbonic Acid. — Several methods for the determination of these data have been devised. Pettenkofer's is much used, but F. B. Forbes and G. H. Pratt, after careful study of different methods, favor the Lunge-Trillich (also termed the Seyler) method, which they describe as follows : One hundred c.c. of the sample are placed in a tall glass cylinder, by means of a siphon (in order to avoid contact with air), 6 drops of neutral alcoholic solution of phenolphthalein added, and ^/so sodium carbonate run in from a buret with careful stirring, until a faint permanent pink is obtained. If the water contains much free car- bonic acid, it is better to take less than 100 c.c, and in every case care must be taken not to stir the sample so vigorously as to cause loss of the acid, nor to proceed so slowly that it may be ab- sorbed from the air. The solutions must be ca^- fuUy standardized and preserved so that they do Io8 ANALYTIC OPERATIONS. not absorb the acid. The fixed carbonic acid is determined on another portion of the sample by Hehner's method, as given above. In waters acid to phenolphthalein this will be equal to the half- bound. When the water is alkaline to phenolphthalein, the alkalinity with this indicator is first deter- mined, then the total alkalinity by Hehner's method. Twice the phenolphthalein alkalinity subtracted from the total alkalinity gives the half- bound acid, no free acid being present in this case, and the half-bound being less than the fixed. If the water is neutral to phenolphthalein, the half -bound may be equal to the fixed. Forbes and Pratt have modified the Pettenkofer methods as follows : Ground-glass-stoppered bottles, holding ap- proximately 480 c.c, are accurately calibrated by weighing completely filled with water. The bottle is filled with the water to be analyzed by means of a siphon, the glass stopper inserted, leaving no air-bubble, and the neck of the bottle wiped dry. The glass stopper is then carefully removed, and the 57 c.c. of the water withdrawn by means of an accurately calibrated pipet, in order to make room for the reagents. Three c.c. of strong barium chlorid solution (8 grams per liter), 2 c.c. of sat- urated ammonium chlorid solution, and 50 c.c. of standard barium hydroxid are then introduced, TECHNIC EXAMINATIONS. I09 the bottle quickly stoppered, well shaken, and set aside to settle. There is now in the bottle an air space of only 2 C.C., which is left to avoid the possibility of loss of liquid when the stopper is inserted. After the precipitated carbonates have completely settled out, several portions of loo c.c. are siphoned off and titrated, with ^/s© sulfuric acid, which is prepared from decinormal acid, against which the barium hydroxid is standardized, by carefully diluting with water freed from carbonic acid by boiling. The barium hydroxid used is approximately ^/is, and is carefully preserved out of contact with the air, the bottle in which it is kept being fitted with an arrangement whereby the air is drawn through soda-lime before entering either the bottle or the buret. The figure obtained by averaging several results of titration of portions of 100 c.c. is taken as the true value. The use of this large quantity of water and the titration of loo c.c. portions reduce considerably the errors due to the difficulty of obtaining the exact end-point, and those due to inaccuracies of measurement. Boric Acid. — To detect this, add to looo c.c. of the water sufficient sodium carbonate to render it distinctly alkaline. Evaporate to dryness, acidify with hydrochloric acid, moisten a slip of turmeric paper with the liquid, and dry it at a moderate no ANALYTIC OPERATIONS. heat. In the presence of boric acid the paper will assume a distinct brown-red tint. Analysis of Boiler Scale. — If the scale is made up of pieces of decidedly different quality, some being hard and gritty, others soft and friable, separate tests should be made on representative samples of each sort; but if the general character is fairly uniform, it will be sufficient to sample the entire mass and reduce about 5 grams to a fine powder, finishing in an agate mortar. All of the quantity selected as the sample should be equally finely powdered. 0.5 gram should be heated in a covered beaker with moderately strong hydrochloric acid, until all soluble matter is dissolved; the liquid is then evaporated to dryness on the water-bath, re- dissolved in water containing some hydrochloric acid, and filtered. The precipitate is silica. The filtrate and washings are mixed and divided into convenient parts. One part is used for the de- termination of sulfates, and the other for iron oxid, alumina, calcium, and magnesium, accord- ing to the methods given on pages 91-5. Scale often contains an appreciable amount of oil. This may be determined by extracting a known weight of the finely powdered material with a petroleum spirit that leaves no residue on evapora- tion on the water-bath. TECHNIC EXAMINATIONS. Ill SPECTROSCOPIC EXAMINATION. For the ordinary spectroscopic examination of a water a simple apparatus will suffice. The arrangement shown in the cut (Fig. 13) is a- small direct-vision s p e c t r o- scope, held in a uni- versal stand, with an adjustable burner as the source of heat. For the examination 1000 c.c. or more should be evaporated nearly to dryness, a little hydro- chloric acid being added near the end of the pro- cess, the residue placed in a narrow strip of platinum foil, having] the sides bent so as to retain the liquid, and heated in the flame. While this method will be sufficient in many cases, a far better plan is to separate the substance sought for in a state of approximate purity and then examine with the spectroscope. Very small traces of lithium, for instance, may be detected as follows: Fig. 13. 112 ANALYTIC OPERATIONS. To about I GOO c.c. of the water sufficient sodium carbonate is added to precipitate all the calcium and magnesium, and the liquid boiled down to about one-tenth its bulk; it is then filtered, the filtrate rendered slightly acid with hydrochloric acid, and evaporated to dryness. The residue is boiled with a little alcohol, which will dissolve out the lithium chlorid. The alcoholic solution is evaporated to dryness, the residue taken up with a little water and tested in the flame. In order to identify with certainty any line which may be obtained, it is only necessary to hold in the flame at the same time a wire which has been dipped in a solution of the substance supposed to be present and to note whether the lines produced by it and the material under ex- amination are identical. SPECIFIC GRAVITY. In the great majority of cases the determination of specific gravity is not essential. Ordinary river, spring, and well waters contain such small proportions of solid matter that it is usually the practice to take a measured volume and to assume its weight to be that of an equal bulk of pure water. If the proportion of solids be high, a determination of the specific gravity may be desirable. For this purpose the specific gravity bottle may be used. This consists merely of a TECHNIC EXAMINATIONS. II3 small flask provided with a finely perforated glass stopper. The bottle is weighed first alone, then filled with distilled water at 60"^ F., and finally with the water under examination at the same temperature. In filling the bottle, the liquid is first brought to the proper temperature, the bottle completely filled, the stopper inserted, and the excess of water forced out thru the perforation and around the sides of the stopper carefully re- moved by bibulous paper. The weight of the water examined divided by the weight of the equal bulk of distilled water at the same temper- ature gives the specific gravity. Another method, and one which gives very satisfactory results, is by the use of a plummet. This may conveniently consist of a piece of a thick glass rod of about 10 c.c. in bulk, or of a test-tube weighted with mercury and the open end sealed in the flame. The plummet is suspended to the hook of the balance by means of a fine platinum wire and its weight ascertained. It is then immersed in distilled water at 60° F., and the loss in weight noted. The figure so obtained is the weight of a bulk of water equal to that of the plummet. This having been determined, the specific gravity of any water may be found by immersing in it the plummet and noting the loss in weight. This, divided by the loss suffered in pure water, gives the specific gravity. INTERPRETATION OF RESULTS. STATEMENT OF ANALYSIS. The composition of water is generally stated in terms of a unit of weight in a definite volume of liquid, but much difference exists as to the stand- ard used. The decimal system is very largely employed, the proportions being expressed in milligrams per liter, nominally parts per million; or in centigrams per liter, nominally parts per hundred thousand. The figures are often given in grains per Imperial gallon of 70,000 grains, or the U. S. gallon of 58,328 grains. In this work the composition is always expressed in milligrams per 1000 c.c. This ratio is practically equivalent to parts per million, except in case of water very rich in solids, a 1000 c.c. of which will weigh notably more than one million rnilligrams. Factors for converting the different ratios are given at the end of the book. From the analysis of a water it is rarely possible to ascertain the exact arrangement of the elements determined, but it is the custom to assume ar- rangements based upon the rule of associating in combination elements having the highest affinities, 114 STATEMENT OF ANALYSIS. II5 modifying this system by any inferences derived from the character or reactions of the water itself. It has been demonstrated that, even in the case of mixtures of salts producing no insoluble sub- stances, partial interchange of the basylous and acidulous radicles takes place. In a solution of sodium chlorid and potassium sulfate, sodium sulfate and potassium chlorid will be found, as well as the original salts. When the conditions are rendered more complex by the addition of other substances, it is obviously impossible to determine the exact arrangement. In view of these facts, it is preferable to express the com- position of a water by the proportion of each ele- ment or radicle present. ' In this way a water containing sodium chlorid will be expressed in terms of sodium and chlorid, respectively. In the case of silica which may exist free in the water, the proportion is expressed as such. It frequently occurs that the characteristics of some of the compounds in a water are sufficiently marked to indicate their presence, and there is no objection to suggesting, in connection with the analytic statement, the inferences which may thus be drawn. The organic matters, or derived products, are best stated in terms of the nitrogen which they contain, thus permitting a comparison of the different stages of decomposition. It is inad- Il6 INTERPRETATION OF RESULTS. visable to represent the amount of unchanged organic matter in terms of oxalic acid, as has been suggested, or to express the nitrogen in terms of albumin, or any other supposititious compound. The results of microbe counting should, as a rule, be reported as "points of microbic life" in the given volume of water. Many operators, however, report the number of points as *' colonies" or even ''bacteria." SANITARY APPLICATIONS. Judgment upon the analytic results from a given sample of water depends upon the class to which it belongs, and to the particular influences to which it has been subjected. A proportion of total solids which would be suspicious in a rain or river water, would be without significance in that from an artesian well. On the other hand, a subsoil water of unobjectionable character would contain a proportion of nitrates which would be inadmissible in the case of a river or deep water. Location has also much bearing in the case ; sub- soil waters near the sea will be found to contain, without invoking suspicion, proportions of chlorin which would be ample to condemn the same sample if derived from a point far inland. Hence the importance of recording, at the time of collec- tion, all ascertainable information as to the sur- roundings and probable source of the water. SANITARY APPLICATIONS. II 7 Analyses of surface waters have no value unless supplemented by a careful survey of the watershed to determine sources of pollution. Such survey will often discover conditions sufficient to con- demn the supply, even though the analyses may be satisfactory. Indeed, it may be taken as a fundamental principle that surface water from even a sparsely populated district will be unsafe for use unless efficiently filtered. Color, Odor, and Taste. — Water of the highest purity will be clear, colorless, odorless, and nearly tasteless. While in some cases a decided de- parture from this standard may give rise to sus- picion, analytic observations are necessary to decide the point. Water highly charged with mineral matters will possess decided taste, vege- table matters may communicate distinct color; but, on the other hand, it may be highly con- taminated with dangerous substances and give no indications to the senses. Well-waters oc- casionally become offensive in odor, from pene- tration of tree roots. The odor often recalls that of hydrogen sulfid. Sulfids are, indeed, often formed in such cases by the abstraction of oxygen from sulfates under the influence of microbes. Such waters are often used without apparent in- jury, but it is probable that if direct pollution occurs, the danger would be enhanced by the presence of the vegetable matter. Il8 INTERPRETATION OF RESULTS. Surface waters collected in reservoirs or ponds often become very offensive from the growth of algae, but apart from the disgust created by the water, it is not known that any harmful results occur to those using it. Turbidity may be due to several causes, of different degrees of danger, but is always objec- tionable. Total Solids. — Excessive proportions of mineral solids, especially of marked physiologic action, are known to render water non-potable, but no absolute maximum or minimum can be assigned as the limit of safety. Distilled water and waters very highly charged with mineral matter have been used for long periods without ill effects. The popular notion that the so-called hard waters conduce to the formation of urinary calculi is not borne out by surgical experience or statistical inquiry. Many urinary calculi are composed of uric acid, and are the results of disorders of the general nutritive functions. Sanitary authorities have fixed an arbitrary limit of total solids of about six hundred parts per million, but many artesian waters in constant use exceed this. The odor produced on heating the water residue is often of much use in detecting contamination. Odors similar to those produced by heating glue, hair, rancid fats, urine, or other animal products. SANITARY APPLICATIONS. II9 will give rise to grave suspicion. On the other hand, a more favorable judgment may be given when the odor recalls those given off in the heating of non-nitrogenous vegetable materials, such as wood-fiber. Poisonous Metals. — The proportion of iron in water constantly used for drinking purposes should not much exceed three parts per million. Lead, copper, arsenic, and zinc must be considered dangerous in any amount, tho it appears that zinc and copper, being least cumulative, are rather less objectionable in minute amount than the others. Concerning the limit of safety with manganese and chromium very little is known, but their presence in appreciable quantity must be looked upon with suspicion. Chlorids and Phosphates. — Chlorids — principally sodium chlorid — and phosphates are abundantly distributed in rocks and soils, and find their way into natural waters; but while the former are freely soluble and remain in undiminished amount under all conditions to which the water is sub- jected, all but small amounts of the latter are either precipitated or removed by the action of living organisms. Surface and subsoil waters ordinarily contain but a few parts per million. Both chlorids and phosphates being constant and characteristic ingredients of animal excretions, it is obvious that an excess of them in natural 120 INTERPRETATION OF RESULTS. waters, unless otherwise accounted for, will suggest direct contamination. Proximity to lo- calities in which sodium chlorid is abundant, such as the sea- or salt-deposits, will deprive the figure for the chlorin of diagnostic value, nor can any indication of sewage or other dangerous pollution be inferred from high proportion of chlorin in deep waters. Further, it has been shown that the proportion of chlorin in uncontaminated waters is tolerably constant, while in water subjected to the infiltration of sewage the chlorin undergoes marked variation in amount. In most cases, therefore, a correct judgment can only be attained by comparison with the average character of the waters of the same type in the district, and by examination at intervals of the water in question. As regards phosphates, Hehner, who has pub- lished a series of analyses, states that the presence of more than 0.6 part per million — calculated as PO4 — should be regarded with suspicion. On the other hand, the absence of phosphates affords no positive proof of the freedom from pollution. Woodman, who has carefully investigated this question, regards Hehner 's limit as too strict. He would fix I part per million as the minimum. He regards this datum as valuable in judging of the sanitary quality of the sample. Nitrogen from Ammonium Compounds. — Am- monium compounds are usually the results of the SANITARY APPLICATIONS. 121 putrefactive fermentation of nitrogenous organic matter; they may also be the product of the re- duction of nitrites and nitrates in presence of excess of organic matter. In either case, there- fore, they suggest contamination. Deep waters often contain an excess of ammonium compounds, derived, in large part, from the reduction of nitrates. Their presence here is hardly ground for adverse judgment, since the water, even tho originally contaminated, has undergone extensive filtration and oxidation, its organic matter con- verted into bodies presumably harmless, and microbes have perished. Such waters, indeed, usually show only traces of unchanged organic matter. Rain water often contains large proportions of ammonium compounds; but here, also, the fact can not condemn the water, since it does not in- dicate contamination with dangerous organic matter. Nitrogen by Alkaline Permanganate (Nitrogen of ''Albuminoid Ammonia"). — A large yield of am- monia by boiling with alkaline potassium per- manganate will, of course, point to an excess of nitrogenous organic matter. The inferences to be drawn depend upon the origin and condition of the organic material. If animal, the water may at once be condemned as unsafe. Waters con- taining excessive amounts even of vegetable 122 INTERPRETATION OF RESULTS. matter are not free from objection, since they have frequently caused persistent diarrhea. If the organic matter, whether animal or vegetable, is in a state of active decomposition, it is doubly objectionable. Smart has observed that water containing fermenting vegetable matter is colored yellow by boiling with sodium carbonate. Inferences as to the source of the organic matter can usually be drawn from the amount of chlorin and nitrates present. If the chlorin is high, — i. e., in excess of the average of the district, — it may be inferred that the material is, in great part, of animal origin. In this case the nitrates will either be high or entirely absent, according as the contaminating matter has passed through soil or enters the water directly. A large amount of vegetable matter will, as a rule, show itself by color imparted to the water. Total Nitrogen. — Drown and Martin's results with surface waters indicate that the total nitrogen obtained by their process is about twice that ob- tained by alkaline permanganate. The experi- ments made by Dr. Beam and myself accord with this. Further observation on different waters • and by different observers will be required to determine the value to be assigned to the figures obtained by this method. This method is es- pecially suitable for studying the effects of filtra- SANITARY APPLICATIONS. I23 tion, storage, etc., on the nitrogenous organic matter in water. Nitrogen as Nitrites. — Nitrites are present in water as the result either of incomplete nitrifica- tion of ammonium, or the reduction of already- formed nitrates, under the influence of reducing agents or microbes. Since they are transition products, their presence in water is usually evi- dence of existing fermentative changes, and, further, may be taken as indicating that the water is unable to dispose of the organic contamination. When, however, the conditions are such that oxi- dation can not take place, nitrites may persist for a long time. This sometimes occurs in deep waters in which fermentative changes have long since ceased, but oxygen is not available. These contain not infrequently small amounts of nitrites, to which the same degree of suspicion can not be attached. When nitrites are found in these waters, the possibility of their introduction from polluted subsoil water, thru defective tubing, must not be overlooked. Rain water, also, some- times contains nitrites derived from the air, and therefore not indicative of any putrefactive change. The presence of measurable quantities of nitrites in river or subsoil water is sufficient ground for condemnation. Nitrogen as Nitrates. — Nitrates are the final point in the oxidation of nitrogenous organic 124 INTERPRETATION OF RESULTS. matter, especially animal matters. Rain water and that from mountain streams and deep wells, except from cretaceous strata, generally contain only traces, but river and subsoil waters will always contain appreciable amounts, unless some reducing action, such as recent sewage -pollution, is at work. When, therefore, a water contains enough mineral matter to demonstrate its perco- lation thru soil, and at the same time is free from nitrates or contains only traces, the occurrence of a destructive fermentation may be inferred. These cases are not uncommon among well-waters, and the samples are generally turbid from sus- pended organic matter. Decided departure, either by increase or decrease, from the proportion of nitrates usual in the same class of water in any district may be taken as evidence of contamination. Oxygen-consuming Power. — Sanitary authori- ties differ very much as to the significance of this datum. Attempts have been made to fix maxi- mum limits for the various types of water, and also to gage the character and condition of the organic matter by observing the rate at which the oxidation takes place, but no positive conclusions can be given. In general, it may be said that a sample which has high oxygen-consuming power will be more likely to be unwholesome than one which is low in this respect ; but the interferences are so numerous, and the susceptibility to oxida- SANITARY APPLICATIONS. 12 5 tion of different organic matters, of even the same type, is so different, that the method is at best only of accessory value. It is especially suitable for consecutive determinations on the same supply. The following proportions are given by Frank- land and Tidy as the basis of interpreting the results of this method : Oxygen Absorbed in Three Hours. High organic purity, .0.05 part per million. Medium purity, 0.5 to 1.5 parts '* Doubtful, 1. 5 to 2.1 " " Impure, over 2.1 '' '' ** For the method with acidified permanganate at the boiling heat, the German chemists, who employ it largely, regard an absorption of 2.5 parts of oxygen per million as suspicious, and some sanitary authorities have fixed 3.8 parts of oxygen per million as the highest permissible limit. Dissolved Oxygen. — Full aeration of water is favorable to the destruction of organic matter; a decided diminution in the quantity of dissolved oxygen may show excess of such matter and > of microbic life. These changes are more likely to take place in still waters, and are frequently ac- companied by disagreeable odor and taste. In cases in which stored waters become unpalatable, these facts should be borne in mind. 126 INTERPRETATION OF RESULTS. Hardness. — The degree of hardness, unless very high, has but httle bearing on the sanitary value of water, but is important in reference to its use for general household purposes, in view of the soap-destroying power which hard waters possess. USUAL ANALYTIC RESULTS FROM UNCONTAMI- NATED WATERS. Milligrams per Liter. Rain. Surface. Subsoil. Deep. Total solids, 5 to 20 15 upward 30 upward 45 upward Chlorin, Traces to i i to 10 2 to 12 Traces to large quantity Nitrogen by per- manganate, .... 0.08 to 0.20 o.os to 0.15 0.05 to o.io 0.03 to o.io Nitrogen as am- monium, 0.20 to 0.50 0.00 to 0.03 0.00 to 0.03 Generally high Nitrogen as nitrites, None or None None None or traces traces Nitrogen as ni- trates, Traces 0.75 to 1.25 1.5 to 5 0.00 to 3 Inferences from Culture Methods. — No absolute limit as to the number of ordinary microbes can be fixed. Some bacteriologists have fixed the maximum of loo per cubic centimeter, but this is arbitrary. An appreciable number of microbes of the colon class will be a basis for condemnation of the water. There is, however, one field of inquiry in which even mere microbe-counting has value ; that is, in comparing samples of the same water before and after some treatment for purification or in some other incident. In these studies the method . is sufficiently free from fallacy to make the results trustworthy when they are conducted in a strictly ACTION OF WATER ON LEAD. 1 27 uniform manner ; thus, if a river water supplied to a filter is studied daily by repeated examination of samples before and after filtration, inoculating separate portions of the same culture -medium, and multiplying the results to such an extent as to eliminate accidental differences, a comparison between the water before and after filtration may be safely made as to the proportion of microbes removed. Moreover, special microbes of highly characteristic properties may be introduced in large quantities into the water, and by subsequent culture the extent to which these are removed may be satisfactorily recognized. ACTION OF WATER ON LEAD. The almost universal use of lead pipes for con- veying water, and the facility with which some waters corrode and dissolve the metal, make it a question of moment to determine the cause of this action and to devise means for its prevention. As a rule, it is found that waters free from mineral matter dissolve lead with facility, especially in the presence of oxygen. Some very soft waters are entirely without action. Messrs. Crookes, Odling, and Tidy found that the action was controlled by the amount of siHca contained in the water. They found that those soft waters which, when taken from the service pipes, contained a notable quantity of lead, gave, on the average, three parts 128 INTERPRETATION OF RESULTS. of silica per million ; in those in which there was no lead, the silica present amounted to 7.5 per million, and in those in which the action was intermediate, 5.5. parts per million. That it was really the silica that conditioned the corrosion was con- firmed by laboratory experiments. They also found that the most effective way of silicating a water is by passing it over a mixture of fiint and limestone. The reason for this was pointed out later by Messrs. Camelly and Frew, who showed that while calcium carbonate and silica both exert a protective influence, calcium silicate is more effective than either; and, further, that in almost all cases in which corrosion took place, it was greater in the presence of oxygen. This is partic- ularly the case with potassium and ammonium nitrates and with calcium hydroxid. The reverse is true of calcium sulfate, which is more corrosive when air is excluded. Their experiments also show that the presence of calcium carbonate or calcium silicate, altogether prevents corrosion by potassium and ammonium nitrates. As the result of an elaborate series of experi- ments Miiller concludes that, while chlorids, nitrates, and sulfates all act upon lead pipes, no corrosion takes place in the presence of sodium acid carbonate, and that calcium carbonate, by taking up carbonic acid, acts in the same way. This latter conclusion is at variance with the TECHNIC APPLICATIONS. 129 observations of Carnelly and Frew, who found that calcium carbonate is equally effective when carbonic acid is excluded. MuUer also states that surface waters, contaminated by sewage and containing large amounts of ammoniacal com- pounds, will dissolve lead under all circumstances. Allen has shown that water containing free acid, including sulfuric acid, acts energetically upon lead. This is not surprising in view of the later experiments, which prove that even calcium sulfate is corrosive. Later, W. Carleton- Williams found that even in the presence of free acid, cor- rosion may be prevented by the addition of suffi- cient silica. His experiments also confirm the view generally held, that soluble phosphates pro- tect lead to a marked degree. The following is a summary of the more im- portant observations on this subject : Corrosive : Free acid or alkalis, oxygen, nitrates, particularly potassium and ammonium nitrates, chlorids, and sulfates. Non-corrosive and preventing corrgsion by the above: Calcium carbonate, sodium acid carbon- ate, ammonium carbonate, calcium silicate, silica, and soluble phosphates. TECHNIC APPLICATIONS. Boiler Waters. — The main conditions affecting the value of a water for steam-making purposes 130 INTERPRETATION OF RESULTS. are its tendency to cause corrosion and the forma- tion of scale. Corrosion may be due to the water itself, to the presence of free acids, or to sub- stances which form acids under the influence of the heat to which the water is subjected. Pure water — e. g., distilled water — exhibits a power- fully corrosive action upon iron. The dissolved oxygen which all waters contain also aids in the corrosion, and especially when accompanied, as is usually the case, by carbonic acid. There is always greater rusting at the point at which the water enters the boiler, since there the gases are driven out of solution and immediately attack the metal. This is an evil that obtains with all waters, and it is not customary, in making ex- amination for technic purposes, to determine the amount of these bodies. In water that has had free access to air, the oxygen in solution is a tolerably constant quantity, and it is sufficient to note the temperature and refer to the table of amounts of oxygen dissolved in water. The corrosive action of oxygen and carbonic acid is especially noticeable in waters that are com- paratively pure, such as those derived from moun- tain springs. This was repeatedly observed by Dr. William Beam, in the examination of the waters used for the locomotives of the Baltimore and Ohio Railroad. The waters which caused the most corrosion were mainly those containing TECHNIC APPLICATIONS. I31 small quantities of solid matter, the full amount of oxygen, and considerable carbonic acid, but no other acid or acid-forming body. Waters of this type are often materially improved by adding a small amount of slaked lime. Free acid, other than carbonic acid, is not often found in water, and if, present, renders the water unfit for use, unless it be neutralized. Mine waters are the most likely to contain free acid, sulfuric acid being generally present. Sometimes the acidity is due to organic acids. These act very injuriously on iron. Magnesium chlorid is frequently present in waters, and if in considerable quantity, may be very harmful. At a temperature of 310° F., corresponding to an effective pressure of four at- mospheres, magnesium chlorid reacts with water to form magnesium oxid and hydrochloric acid, the latter attacking the boiler, especially at the water-line. If considerable calcium carbonate is present, the evil may be somewhat lessened, but, as Allen has pointed out, and as we also have noticed, there may still be corrosion, so that the presence of more than a small quantity of the salt, say a grain or two to the gallon, may be considered objectionable. Allen remarks that the presence of a certain amount of sodium chlorid may pre- vent this decomposition, the two chlorids com- bining to form a stable double salt. The addition, 132 INTERPRETATION OF RESULTS. therefore, of common salt to a water containing magnesium chlorid may act to diminish corrosion, a point which will bear further investigation. It has not been determined how far the presence of nitrites, nitrates, and ammonium compounds affects the quality of water for steam-making purposes ; but it is more than probable that they act harmfully, especially the nitrates, which are frequently present in large amount. Scale is composed of matters deposited from the water either by the decompositions induced by the heat or by concentration. When the deposit is loose, it is termed sludge or mud, and usually con- sists of calcium carbonate, magnesium oxid, and a small amount of magnesium carbonate. The magnesium oxid is formed by the decomposition of the magnesium carbonate and chlorid. The formation of sludge is the least objec- tionable effect, since it may readily be removed by ''blowing off,'' provided that care is previously taken to allow the flues to cool down so that when the water is removed the heat of the flues may not bake the deposit to a hard mass. Waters con- taining calcium sulfate form hard incrustations difficult to remove and causing great loss of fuel by interfering with the transmission of the heat to the water. It not only forms a hard incrustation in itself, but becomes incorporated with the mud, and renders it also hard. The hard scale will also TECHNIC APPLICATIONS. I33 contain practically all the silica and the iron and aluminum present in the water, besides any matters originally held in suspension. It follows from the above that a water only temporarily hard will, if care is taken in the management of the boiler, cause the formation merely of a loose deposit of sludge — temporary hardness being due in the main to calcium and magnesium carbonates. A water permanently hard will probably form a hard scale, since such hardness is usually due to calcium sulfate. In accordance with these principles, the analysis of a water for steam-making purposes may include the determinations of free acid, total solid residue, sulfates, chlorin, calcium, magnesium, temporary and permanent hardness. In cases in which the qualitative tests show but small amounts of sulfates and chlorin, the analysis may be limited to the determinations of the temporary and per- ■ manent hardness. In the laboratory of the Pennsylvania Railroad an approximate determination of scale-forming in- gredients has been made in the following manner : The total solids obtained by evaporation are treated with diluted alcohol (fifty per cent.), and the undissolved residue is denominated ' ' scale-forming material. ' ' It has been pointed out in an earlier chapter that it is not possible to deduce from the analytic 134 INTERPRETATION OF RESULTS. result the exact forms in which the various ele- ments are combined, but since it is known that at the high temperature ordinarily reached in boilers definite chemical changes occur, it is safest to exhibit the maximum amount of corrosive and scale -forming ingredients which the water under these circumstances could develop. Thus, since calcium sulfate is practically insoluble in water above 212^ F., the proportion of calcium sulfate may be regarded as such as would be formed by the total quantity of calcium or the total quantity of SO 4, according to which is present in the larger amount. Similarly, as the decomposition of mag- nesium chlorid is induced by the high temperature of the boiler, the analytic statement should in- dicate the maximum proportion of this compound obtainable from the magnesium and chlorin pres- ent. These rules can not apply absolutely to waters rich in alkali carbonates, since these would neutralize any acid formed from the magnesium chlorid, or even prevent its formation, and would prevent, to a large extent, the formation of calcium sulfate. Much remains to be determined con- cerning the effects of the high temperature and concentration to which boiler waters are subjected. General Technic Uses. — In regard to the quality of water for technic other than steam-making purposes, such as brewing, dyeing, tanning, etc., no detailed methods or standards can be laid TECHNIC APPLICATIONS. I35 down. The nearest approach to purity that can be secured in the supply will be of the greatest advantage. The more objectionable qualities will be large proportion of organic matter, es- pecially if it distinctly colors the water, excessive hardness, and notable amounts of iron or free mineral acid. It is said that one part of iron per million will render water unsuitable for bleaching establishments. It has been noted that a large proportion of active microbes is injurious in the manufacture of indigo. In artificial ice making, a very pure water must be used if a clear and colorless product be desired. Any suspended or dissolved coloring-matter will be concentrated by the freezing and appear in the bottom or center of the mass. The examination of sewage-effluents and waste waters from manufacturing establishments is to be conducted upon the same principles as for ordinary supplies, but especial attention must be given to the presence of poisonous metals and free mineral acids. The latter interfere with the nor- mal self -purification of the water. For the nitro- gen determination, the Kjeldahl process will be found more satisfactory than that by alkaline permanganate. 136 INTERPRETATION OF RESULTS. PURIFICATION OF WATER. The purification of drinking water is a question of sanitary engineering that is outside the scope of this work. Purification of Boiler Waters. — The problems present in the treatment of boiler waters are usually the removal of the calcium carbonate and sulfate, and magnesium carbonate and chlorid. Both carbonates are appreciably soluble in pure water. About 1 5 parts of calcium carbonate per million is usually stated to be the proportion dis- solved, but it has been pointed out by Allen that solutions can be obtained containing twice this amount. If the water contains carbonic acid, it will take up a much greater proportion of the car- bonates, but in this case they will be deposited from the solution by boiling. This has been ac- counted for by supposing the existence of soluble bicarbonates, which are decomposed by the boiling. Nearly all of these carbonates can be thrown out of solution by any means that will deprive the water of the carbonic acid. ^Sodium hydroxid is often employed for the purpose, and should be added in quantity just sufficient to form normal sodium carbonate. If there are present in the water calcium and magnesium chlorids and sulfates, these also will be decomposed and precipi- tated by the sodium carbonate so formed. If the PURIFICATION OF WATER. I37 amount of sodium carbonate formed is not suffi- cient to decompose all of these bodies, a sufficient quantity should be added with the sodium hy- droxid to effect the complete decomposition. The precipitate is allowed to settle or filtered off. In cases in which the feed-water is heated be- fore it enters the boiler, it may only be necessary to add to the water sodium carbonate in quantity sufficient to decompose the calcium and mag- nesium chlorids and sulfates, since the heat alone will suffice to throw down the carbonates. Care should be taken in these precipitations that no more sodium hydroxid is added than is required for the precipitation, since any excess would tend to corrode the boiler. Clark's process consists in treating the water with calcium hydroxid (lime-water). This pre- cipitates the calcium and magnesium carbonates by depriving the water of its free carbonic acid. It has, of course, no effect upon the calcium sulfate. It is to be noted that the proportion of calcium hydroxid which is to be added must be calculated from the amount of free carbonic acid existing in the water, and not from the amount of carbonates to be removed. The precipitate will usually re- quire at least twelve hours for complete subsidence, but after three or four hours the water will be sufficiently clear for some purposes. If a filter press is used, as in Porter's process, the time re- 138 INTERPRETATION OF RESULTS. quired for clarification is very much shortened. Another advantage of this process is the use of a so- lution of silver nitrate, in order to determine more conveniently the proportion of calcium hydroxid which is to be employed. The lime is first slaked and dissolved in water, and the water to be soft- ened run in and well mixed with it. From time to time small portions are taken out and a few drops of a solution of silver nitrate added. As long as the lime is in excess, a brownish coloration is pro- duced. When this has become quite faint, and just about to disappear, the addition of the water is discontinued, and, after a short time, the water is filtered by means of the press. Soluble phosphates added to a water precipitate completely in a flocculent condition any calcium, magnesium, iron, or aluminum. This reaction can be best applied by using the trisodium phos- phate, which is now a commercial article. By reason of the facility with which this substance loses a portion of its sodium to acids, it acts not only as a precipitant to the above materials, but will neutralize any free mineral acid present in the water. From evidence submitted by those who have used the process on the large scale, it ap- pears that not only is no hard scale formed, but that scale already existing prior to its use is gradually disintegrated and removed with the IDENTIFICATION OF THE SOURCE OF WATER. 139 sludge. Experiments indicate that no injury results from an excess of the material; but the economical employment of the method, especially with very hard waters, can only be based upon a correct analysis, and an estimation of the phos- phate required for the precipitation. In many cases the composition of the water will be such that a partial precipitation will be sufficient. Considerable success has been obtained by the use of fluorids as precipitants of scale-forming elements. Sodium aluminate has also been re- commended. Waters rich in ferrous compounds may be purified by aeration and filtration, the iron being separated as ferric hydroxid. The corrosive action of very pure waters is partially abated by filtration thru bone charcoal or by addition of small amounts of lime. IDENTIFICATION OF THE SOURCE OF WATER. The determination of the course of underground streams, and of communications between collec- tions of water, is often an important practical problem. In geologic and sanitary surveys, val- uable information may occasionally be gained. The method generally pursued when connection between water at accessible points is to be de- I40 INTERPRETATION OF RESULTS. tected, is to introduce at one point some substance not naturally existing in the water, and capable of recognition in small amount. Lithium com- pounds are among the best for this purpose. They are not frequent ingredients of natural waters, and are easily recognized by the spec- troscope. Lithium chlorid is the most suitable. The quantity to be employed will vary with cir- cumstances. It scarcely needs to be stated that the waters under examination should be carefully tested for lithium before using the method. When the lithium method is inadmissible, re- course must be had to other substances of distinct character, such as strontium chlorid, but this possesses the disadvantage that a considerable amount may be rendered insoluble, and thus lost in the ordinary transit through soil. Use has been made of organic coloring matters of high tinctorial power, one of the most suitable of which is fluorescein. This will communicate a charac- teristic and intense fluorescence to many thousand times its weight of water. The coloration is distinct only in alkaline liquids. Other colors, such as anilin-red, may be employed. A more important feature of the problem from a sanitary point of view is the determination of the source of a given current or collection of water, when such source is inaccessible. Problems of this character are not infrequent in large cities in which the systems of water-supply and drainage IDENTIFICATION OF THE SOURCE OF WATER. I4I are defective, thus giving occasion to accumu- lations of water in cellars and similar places. Often, in these cases, no extended explorations can be made, by reason of the adjacent buildings and conflicting property interests, and the ques- tion raay arise whether the water proceeds from a leaky hydrant, drain, sewer, or subsoil current. It is obvious that in the case of the collection of water in a cellar from causes other than surface washings or entrance of rain, it must have passed thru some distance of soil, and in built-up dis- tricts will almost certainly be charged with or- ganic refuse. To correctly interpret the results, it will be necessary to know the general character of the subsoil water of the district and the com- position of the public supply. As a rule, the transmission of water thru raoderate distances of soil will not raaterially increase the mineral constituents. Hence, if the sample contains an excess of dissolved matters as compared with the water-supply of the district, it may reasonably be inferred that it is derived from a drain, sewer, or subsoil current. In these investigations it will generally be sufficient to determine the total solids, odor on heating, chlorin, nitrates, and nitrites. Occasionally, the analytic results will be am- biguous, and it is advisable to make examinations of more than one sample, since accidental cir- cumstances, rain-fall, etc., may affect the com- position of the* water. 142 INTERPRETATION OF RESULTS. DATA FOR CALCULATION Parts per 100,000 X 0.7 = grains per Imperial a u 1,000,000 X 0.07 n a (( Loo^fOoa X •0.583 u u 100,000 X 0.583 U.S. ,, ,, 1,000,000 X 0.0^83 a n (( Grains per Imp. gal . X 1.429 = parts per 100,000 " " ^," « " X 14.29 ^^ (( n 1,000,000 « .. u- s. .. X 1.724 ^^ u u 100,000 it n n ^^ X 17.24 It (( 1,000,000 Al,03 X 0.53 = Al Agcl X 0.247 = C1 BaSO^ X 0.42 = 804 BaSO^ X 0.343 = S03 B2O3 X 0.314 = B CaO X 0.7147 = Ca CaO X 1.785 = CaC03 CaCOa X 0.4 = Ca CI X 1.65 = NaCl Fe^Os X 0.7 = Fe KCl X 0.525 = K K^PtCle X 0.161 = K K^PtCl^ X 0.307 = KC1 Mg,PA X 0.2187 = Mg Mg^PA X 0.853 = P0, Mg^PA X 0.757 = MgCOs ♦ Mg^PA X 0.6375 = P205 MnS X 0.631 = Mn NaCl X 0.394 = Na N X 4.425 = N03 N X 3.284 = N02 N X 5.857 = Ca(N03)2 N X 1.215 = NH3 NH3 X 0.822 = N ATOMIC WEIGHTS. 143 ATOMIC WEIGHTS (Report International Committee, Jour. Amer. Chem. Soc, Jan^ry, 1908.) Aluminum 27.1 Barium I37'4 Calcium 40.1 Carbon 12.0 Chlorin 35-45 Chromium 52.1 Copper 63.6 Hydrogen 1.008 lodin 126.97 Iron 55.9 Lead 206.9 Lithium 7.03 Magnesium 24.36 Manganese 55.0 Nitrogen 14.01 Oxygen 16.0 Phosphorus 31.0 Platinum 194.8 Potassium 39.15 Silver 107.93 Sodium 23.05 Sulfur 32.06 Zinc 65.4 INDEX Agar, 73. media, 76. , Hesse's, 88. Albuminoid ammonia, 23, 32,121. Alizarin indicator, 15. Alkaline permanganate, 31. Alum, detection, 60. Aluminum, detection, 59. , determination, 91. Aluminum sulfate, detection, 60. Ammonia, albuminoid, 23, 32, 121. , free, 23, 31, 120. method, 23. Ammonium molybdate, 50. Analysis, statement of, 114. Arsenic, detection, 56. Artesian water, i, 5. Atomic weights, 143. Bacillus colt communis, 84. • typhosus, 87. Bacteriologic examinations, 70, 126. Barium, 55. Bile-lactose medium, 86. Biologic examinations, 63. Boiler scale, no, 132. Boiler waters, 109, 129. Boric acid, 109. Bouillon, 71. , dextrose, 75. CALcroM, determination, 92. Calculation table, 142. Carbonic acid, determination, 107. Chlorin, determination, 21. , significance, 119, Chromium, 55. Clark's process, 137. Classification of waters, i. Colon bacillus, 84-9. Color, determination, 11. , significance, 117. Condensers, 23. Copper, detection, 62. , determination, 63. Deep water, i, 5, 126. Denitrification, 5. Dextrose bouillon, 75. Dissolved oxygen, 52. Distilling apparatus, 23. Free ammonia, 23, 32, 37, 121. Ground water, 1, 3. Gelatin, 73. medium, 76. Hardness, ioi. Hesse's agar, 88. Hydrogen sulfid, 99. Indol reaction, 83. Iron, detection, 57. -Ih^ — , determination, 58, 91. Lactose-bile medium, 86. litmus medium, 80. Lead, action of water on, 127. , detection, 61. , determination, 62. Lithium, determination, 98. Litmus-lactose media, 80. Manganese detection, 59. , determination, 92, 94. Milk medium, 79. Nessler glasses, 29. — ; reagent, 30. Nitrates 39, 123. Nitrification, 4. Nitrites, 42, 123. Nitrogen as ammonium, 23, 32, 37 121. total organic, 36, 122. Nitron method, 41. Odor, 13, 117. Organic and volatile matter, 19. Oxygen-consuming power, 42, 124. , dissolved, 52, 125. Permanganate, alkaline, 31. Phenoldisulfonic acid, 39. Phosphates, 50, 119. Poisonous metals, 55, 119. Potassium, determination, 95. Rain water, i, 2, 126. Reaction, 15. Samples, collection, 8, 11. Scale, no, 132. Silica, 91. Sludge, 132. Specific gravity, 112. Spectroscopy, in. Soap test, 104. Subsoil water, i, 3 126. Sulfates, 94. Surface water, i, 2, 126. Taste, 117. Total soUck, 16, 118. Turbidity, 14. Typhoid bacillus, 87. Water, classes of, i. Zmc, detection, 56. 144 9^'M THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW RENEWED BOOKS ARE SUBJECT TO IMMEDIATE RECALL LIBRARY, UNIVERSITY OF CALIFORNIA, DAVIS Book Slip-50m-12,'64(F772s4)458 358532 QDlIi2 Leffmann, H. Lk Examination of water 1?09 for sanitary & technic purposes • LIBRARY UNIVERSITY OF CALIFORNIA DAVIS