UNIVERSITY OF CALIFORNIA. Class QDIO/P7 v. i WORKS OF ALFRED I. CORN PUBLISHED BY JOHN WILEY & SONS. Indicators and Test-papers. Their Source, Preparation, Application, and Tests for Sensitiveness. With Tabular Summary of the Ap- plication of Indicators. Second Edition, Revised and Enlarged. i2mo, ix + 267 pages. Cloth, $2.00. Tests and Reagents. Chemical and Microscopical, known by their Authors' Names; together with an Index of Subjects. 8vo, iii + 383 pages. Cloth, $3.00. TRANS LA TIONS. Fresenius's Quantitative Chemical Analysis. New Authorized Translation of the latest German Edition. In two volumes. By Alfred I. Conn, Phar.D. Recalculated on the basis of the latest atomic weights, and also greatly amplified by the translator. 8vo, 2 vols., upwards of 2000 pages, 280 figures. Cloth, $12.50. Techno-Chemical Analysis. By Dr. G. LUNGE, Professor at the Eidgenossische Polytechnische Schule, Zurich. Authorized Transla- tion by Alfred I. Cohn, Phar.D. i2ino, vii + i36 pages, 16 figures. Cloth, $1.00. Toxins and Venoms and Their Antibodies. By EM. Pozzi-EscoT. Authorized Translation by Alfred I. Cohn, Phar.D. iamo, vii+ 101 pages. Cloth, $1.00, net. QUANTITATIVE CHEMICAL ANALYSIS BY THE LATE DR. 0. EEMIGIUS FRESENIUS PRIVY AULIC COUNSELLOR; DIRECTOR OF THK CHEMICAL LABORATORY AT WIESBADEN AUTHORIZED TRANSLATION OF THE SIXTH GERMAN EDITION, GREATLY AMPLIFIED AND REVISED BY ALFRED I. COHN AUTHOR OF "INDICATORS AND TEST-PAPERS," AND "TESTS AND REAGENTS.** MEMBER OF THE AMERICAN CHEMICAL SOCIETY; SOCIETY OF CHEMICAL INDUSTRY; VEREIN DEUTSCHER CHEMJKER; ETC. VOL. I NEW YORK ' JOHN WTLEY & SONS 43-45 EAST NINETEENTH STREET 1907. />< f-c Copyright, 1903, BY ALFRED I. COHN, ROBERT DRUMMOND, PRINTER, NEW VOHK. PREFACE. THE great advances made in analytical chemistry since the publication of the previous edition of this work, and the intro- duction' of numerous new methods of analysis and improvements upon older ones, have necessitated a new translation of the most recent German edition. This translation is here presented. While the German text, however, has been adhered to as closely as possible in the translation, the requirements of the skilled analyst, as well as of the student, have been borne in mind, and hence all the values in the book excepting those of Appendices I and II have been recalculated on the basis of the table of atomic weights published by Prof. F. W. CLARKE in the Journal of the American Chemical Society, March, 1902; furthermore, all the old-style formulas and equations have been made to conform to the chemical notation and nomenclature now in use. It had been the translator's intention to greatly amplify the work by inserting very many of the more recent and approved methods of chemical analysis, but as the work in hand progressed it was found inexpedient to carry out the original intention with- out too greatly enlarging the size of the volumes; and the more so as the appearance in the interim of the works by CLASSEN, SMITH, and others, in part rendered unnecessary any very extended amplification. Nevertheless, quite a number of new methods have been added by the translator, and have been appropriately designated. The matter inserted by Professors ALLEN and JOHNSON in the previous edition has been retained in this, unless superseded by more recent and improved methods. In addition, and be- cause of their undoubted value, it has been deemed useful to add in the form of appendices the official methods of analysis adopted by the Association of Official Agricultural Chemists, and con- iii 1 93937 IV PREFACE. stituting Bulletin No. 46, revised edition, of the U. S. Dept. of Agriculture, 1899, because of the legal recognition they receive; and also the excellent treatise, on "Some Principles and Methods of Rock Analysis," constituting Bulletin No. 176 of the U. S. Geological Survey, 1900. The table of factors and their multiples has also been entirely recalculated, and the translator has moreover added the loga- rithmic values of the factors. Two new tables of weights of gases per litre have been added to Table IX by the translator, both calculated on the atomic values used in the book; the object being to have all the values in each given table agree among themselves, whereby more consistent and uniform results may be obtained and fewer discrepancies are likely to occur. Particular care has been bestowed upon the index so as to render it as complete and comprehensive as possible, and in order that the work may prove of maximum service to the user. ALFRED I. COHN. NEW YORK, November, 1903. CONTENTS. PAOB INTRODUCTION. 1 PART I. GENKRAL FART. SECTION I. Operations, 1 11 L Determination of quantity, 2 11 1. Weighing, 3 11 a. The balance 12 Accuracy, 4 12 Sensibility, 5 14 Testing, 6 15 7 17 6. The weights, 8 19 C. The process of weighing, 9 21 Rules, 10 23 2. Measuring, 11 26 a. The measuring of gases, 12 27 Correct reading-off, 13 30 Influence of temperature, 14 33 Influence of pressure, 15 33 Influence of moisture, 16 34 & The Measuring of fluids 17 36 a. Measuring vessels graduated to hold certain volumes of fluid. aa. Vessels serving to measure out one definite volume of fluid. 1. Measuring flasks, 18 36 H>. Vessels serving to measure out different volumes of fluid. 2. The graduated cylinder, 19 30 v CONTENTS. ft. Measuring vessels graduated to deliver certain vol- umes of fluid. aa. Vessels serving to measure out one definite volume of fluid. 3. The graduated pipette, 20 ................ 39 bb. Vessels serving to measure out different volumes of fluid. 4. The Burette. I. Mohr's burette, 21 .................. 42 II. Gay-Lussac's burette, 22 ........ ..... 48 III. Geissler's burette, 23 ................ 49 H. Preliminary operations. Preparation of substances for the pro- cesses of quantitative analysis. 1. Selection of the sample, 24 ................. . ......... 50 2. Mechanical division, 25 ............................... 51 3. Drying, 26 .......................................... 54 Desiccators, 27 .................................... 56 Water-baths, 28 ......... .......................... 58 Air-baths, 29 ...................................... 62 Oil-baths, 30 ...................................... 66 Drying-disk, 31 ................................... 67 III. General procedure in quantitative analysis, 32 ............... 69 1. Weighing the substance, 33 ........................... 70" 2. Estimation of water, 34 .............................. 72 a. Estimation of water by loss of weight, 35 .......... 72 b. Estimation of water by direct weighing, 36 ......... 75 3. Solution of substances, 37 ............................ 79 a. Direct solution, 38 .......................... .... 79 b. Decomposition by fluxing, 39 .............. ....... 80 4. Conversion of the dissolved substance into a weighable form, 40 ...................................... 81 a. Evaporation, 41 ................................ 81 Weighing of residues, 42 ....................... 89 b. Precipitation, 43 ............................... 91 a. Separation of precipitates by decantation, 44.. 93 0. Separation of precipitates by filtration, 45. ... 94 aa. Ordinary filtration, 45 ................ 94 aa. Filtering apparatus ............... 94 jlfi. Rules to be observed in the process of filtration ................... . 96 7-7-. Washing of precipitates, 46 ...... 98 bb. Filtration by suction, 47. . ............. 100 f. Separation of precipitates by decantation and fil- tration combined, l"8 ...................... 108 CONTENTS. Vil FACE Further treatment of precipitates preparatory to weighing, 49 109 aa. Drying of precipitates, 50 110 bb. Ignition of precipitates, 51. . ..- 112 First method, 52 116 Second method, 53 118 Asbestos filters with Bunsen's apparatus, 53, a ... 120 5. Volumetric analysis, 54 122 SECTION II. Reagents, 55 127 A. Reagents for gravimetric analysis in the wet way. I. Simple solvents, 56 127 II. Acids and halogens. a. Oxygen acids, 57 128 6. Hydrogen acids and halogens, 58 129 c. Sulpho-acids 131 III. Bases and metals. a. Oxygen bases and metals. a. Alkalies, and . Alkaline earths, 59 131 f. Heavy metals and oxides of heavy metals, 60 132 6. Sulpho-bases 134 IV. Salts. a. Salts of the alkalies, 61 135 6. Salts of the alkali-earth metals, 62 137 c. Salts of the heavy metals, 63 139 B. Reagents for gravimetric analysis in the dry way, 64 140 C. Reagents for volumetric analysis, 65 144 D. Reagents for organic analysis, 66 151 SECTION III. Forms and combinations in which substances are separated from each other, or weighed, 67 / 158 A. BASIC RADICALS. FIRST GROUP. 1. Potassium, 68 161 2. Sodium, 69 164 3. Ammonium, 70 . 167 SECOND GROUP. 1. Barium, 71 '. 168 2. Strontium, 72 171 3. Calcium, 73 173 4. Magnesium, 74 176 Vlll CONTENTS. THIRD GROUP. PAGB 1. Aluminium, 75 179 2. Chromium, 76 181 FOURTH GROUP. 1. Zinc, 77 182 2. Manganese, 78 185 3. Nickel, 79 189 4. Cobalt, 80 191 5. Ferrous iron; and 6. Ferric iron, 81 194 FIFTH GROUP. 1. Silver, 82 198 2. Lead, 83 201 3. Mercury in mercurous; and 4. in mercuric compounds, 84 205 5. Copper, 85 208 6. Bismuth, 86 211 7. Cadmium, 87 213 SIXTH GROUP. 1. Gold, 88 215 2. Platinum, 89 215 3. Antimony, 90 ' 216 4. Tin in stannous; and 5. in stannic compounds, 91 219 6. Arsenous acid ; and 7. Arsenic acid, 92 221 B. ACIDS. FIRST GROUP, 93. 1. Arsenous and arsenic acids. 2. Chromic acid 225 3. Sulphuric acid 225 4. Phosphoric acid 226 5. Boric acid 232 6. Oxalic acid 232 7. Hydrofluoric acid 232 8. Carbonic acid 233 9. Silicic acid 233 SECOND GROUP, 94. 1. Hydrochloric acid 235 2. Hydrobromic acid 235 3. Hydriodic acid 236 4. Hydrocyanic acid 237 5. Hydrosulphuric acid 237 THIRD GROUP, 95. 1. Nitric acid 238 2. Chloric acid 238 CONTENTS. SECTION IV. PAGE Determination of radicals, 96 ................................... 239 I. Determination of basic radicals ................................. 242 FIRST GROUP. 1. Potassium. 97 .......................................... 242 2. Sodium, 98 ............................................. 248 3. Ammonium, 99 ......................................... 251 Supplement to first group, 100. 4. Lithium ................................................. 258 SECOND GROUP. 1. Barium , 101 ............................................ 262 2. Strontium, 102 ......................................... 265 3. Calcium, 103 ........................................... 268 4. Magnesium, 104 ........................................ 274 THIRD GROUP. 1. Aluminium, 105 ........................................ 277 2. Chromium, 106 ......................................... 280 Supplement to third group, 107. 3. Titanium ................................................ 284 FOURTH GROUP. 1. Zinc, 108 .............................................. 286 2. Manganese, 109 ................ , ........................ 291 3. Nickel, 110 ................................ . ............ 301 4. Cobalt, 111 ............................................. 305 5. Ferrous iron, 112 ....................................... 310 6. Ferric iron, 113 .................... .' .................... 321 Supplement to fourth group, 114. 7. Uranium ................................................ 335 FIFTH GROUP. 1. Silver, 115 ............................................. 337 2. Lead, 116 .............................................. 351 3. Mercury in mercurous compounds, 117 ..................... 361 4. Mercury in mercuric compounds, 118 ...................... 363 5. Copper, 119 ............................................ 370 6. Bismuth, 120 ........................................... 382 7. Cadmium, 121 .......................................... 387 Supplement to fifth group, 122. 8. Palladium. . 389 CONTENTS. SIXTH GROUP. PAGE 1. Gold, 123 , 391 2. Platinum, 124 393 3. Antimony, 125 395 4. Tin in stannous; and 5. in stannic compounds, 126 403 6. Arsenous acid ; and 7. Arsenic acid, 127 409 Supplement to sixth group, 128. 8. Molybdic acid 1 1 , 1 1 . t 1 1 . t . 1 1 , 420 JL Estimation of the acids. FIRST GROUP. First Division. 1. Arsenous and arsenic acids, 129 422 2. Chromic acid, 130 422 Supplement, 131. 1. Selenous acid 429 2. Sulphurous acid 431 3. Thiosulphuric acid 432 4. lodic acid 432 5. Nitrous acid 433 Second Division. Sulphuric acid, 132 434 Supplement, 133. Hydrofluosilicic acid 442 Third Division. 1. Phosphoric acid. I. Determination, 134 444 II. Separation from the bases, 135 457 2. Boric acid, 136 465 3. Oxalic acid, 137 470 4. Hydrofluoric acid, 138 472 Fourth Division. . 1. Carbonic acid, 139 479 2. Silicic acid, 140 505 SECOND GROUP. 1. Chlorine (Hydrochloric acid), 141 521 Supplement: free chlorine, 142 529 2. Bromine (Hydrobromic acid), 143 532 Supplement: free bromine, 144 536 3. Iodine (Hydriodic acid), 145 536 Supplement: free iodine, 146 542 4. Cyanogen (Hydrocyanic acid), 147 548 5. Sulphur (Hydrosulphuric acid), 148 558 CONTENTS. XI THIRD GEOUP. PAGE 1. Nitric acid, 149. 571 2. Cholricacid, 150. 393 SECTION V. Separation of bodies, 151 596 I. SEPARATION OF BASIC RADICALS FROM EACH OTHER. FIRST GROUP. Separation of the alkalies from each other, 152 599 SECOND GROUP. L Separation of the basic radicals of the second group from those of first, 153 607 II. Separation of the basic radicals of the second group from each other, 154 615 THIRD GROUP. I. Separation of metals of the third group from the alkalies, 155. ... 622 II. Separation of metals of the third group from the alkali-earth metals, 156 623 III. Reparation of metals of the third group from each other, 157. . . . 630 FOURTH GROUP. I. Separation of the metals of the fourth group from the alkalies, 158 631 II. Separation of the metals of the fourth group from those of the second, 159 633 III. Separation of the motals of the fourth group from those of the third and from each other, 160 639 IV. Separation of iron, aluminium, manganese, calcium, magnesium, potassium, and sodium, 161. -. 666 Separation of uranium from the metals of groups I. -IV. 672 FIFTH GROUP. I. Separation of the metals of the fifth group from those of the preced- ing four groups, 162 676 IL Separation of the metals of the fifth group from each other, 163 . 685 SIXTH GROUP. I. Separation of the metals of the sixth group from those of the first five groups, 164 699 II. Separation of the metals of the sixth group from each other, 165 715 Xll CONTENTS. H. SEPARATION OF ACIDS FROM EACH OTHER. FIRST GROUP. PAGE Separation of the acids of the first group from each other, 166. . 730 SECOND GROUP. I. Separation of the acids of the second group from those of the first, 167 739 Supplement. Analysis of compounds containing sulphides of the alkali metals, carbonates, sulphates, and thiosulphates, 168. . 742 II. Separation of the acids of the second group from each other, 169. . 744 THIRD GROUP. I. Separation of the acids of the third group from those of the two first groups, 170 757 IL Separation of the acids of the third group from each other 759 INTRODUCTION. Chemical analysis comprises two branches, viz., qualitative analysis and quantitative analysis, the object of the former being to ascertain the nature, that of the latter to determine the amount, of the several component parts of any compound. By QUALITATIVE ANALYSIS we convert the unknown constituents of a body into certain "known forms and combinations, whereby we are enabled to draw correct inferences respecting the nature of these unknown constituents. In QUANTITATIVE ANALYSIS the object is attained, according to circumstances, often by very different ways ; the two methods most widely differing from each other are analysis by weight, or gravimetric analysis, and analysis by measure, or volumetric analysis. GRAVIMETRIC ANALYSIS has for its object the conversion of the "known constituents of a substance into forms or combinations which will admit, of the most exact determination of their weight, and of which, moreover, the composition is accurately known. These new forms or combinations may be either educts from the analyzed substance, i.e., bodies that were present as such in the analyzed substance, as water in crystallized sodium sulphate or charcoal in gunpowder; or they may be products, i.e., substances that have formed from the constituents of the analyzed substance by the addition of other elements, e.g., carbonic acid and water by the combustion of paraffin, or barium sulphate on bringing together barium-chloride solution and sulphuric acid. In the former case the ascertained weight of the eliminated substance is the direct" ex- pression of the amount in which it existed in the compound under cxnmination; in the latter case, that is, when we have to deal with 2)roducts, the quantity in which the eliminated constituent was originally present in the analyzed compound has to be deduced by 2 INTRODUCTION. calculation from the quantity in which, it exists in its new combi- nation. The following example will serve to illustrate these points: Suppose we wish to determine the quantity of mercury contained in mercuric chloride. We may do this, either by precipitating the metallic mercury from the solution of the chloride, say by means of stannous chloride, or we may attain our object by precipitating the solution by hydrogen sulphide and weighing the precipi- tated mercuric sulphide. 100 parts of mercuric chloride consists of 73*83 of mercury and 26-17 of chlorine; consequently if the process is conducted with absolute accuracy, the precipitation by stannous chloride of the mercury in 100 parts of mercuric chloride will yield 73 '83 parts of metallic mercury. With equally exact manipulation the other method yields 85 '67 parts of mercuric sulphide. Now, in the former case we find the numbers 73*83 directly ;' in the latter case we have to deduce it by calculation (100 parts, of mercuric sulphide contain 86*18 parts of mercury; how much mercury do 85*67 parts contain?): 100 : 85*67 : : 86*18 : x . . x = 73*83. As already hinted, it is absolutely indispensable that the forms into which bodies are converted for the purpose of estimation by weight phould fulfil two conditions. First, they must be capable of being weighed exactly ; secondly, they must be of known composi- tion, for it is quite obvious, on the one hand, that accurate quan- titative analysis must be altogether impossible if the substance the quantity of which it is intended to ascertain does not admit of correct weighing ; and on the other hand, it is equally evident that if we do not know the exact composition of a new product, we lack the necessary basis of our calculation. YOLUMETKIC ANALYSIS is based upon a principle very different from that of gravimetric analysis ; viz., it effects the quantitative determination of a body by converting it from a certain, definite state to another equally definite state by means of a fluid of accu- rately known power of action, and under circumstances which per- mit the analyst to mark with rigorous precision the exact point when the conversion is accomplished. The following example will serve to illustrate the principle of this method: Potassium per- INTRODUCTION. 3 manganate added to a solution of ferrous sulphate acidulated with sulphuric acid immediately converts the ferrous sulphate into fer- ric sulphate, the permanganic acid, characterized by its intense color, yielding up oxygen and forming with the free sulphuric acid present colorless manganous sulphate. If, therefore, to an acidu- lated fluid containing a ferrous salt we add, drop by drop, a solu- tion of potassium permanganate, its red color continues for some time to disappear upon stirring; but at last a point is reached when the coloration imparted to the fluid by the last drop added remains. This point marks the termination of the conversion of the ferrous salt into a ferric salt. If we now accurately determine the effective value of the per- manganate solution, which may be done by noting its action on a known quantity of dissolved ferrous sulphate, we are in a position to determine, by means of this solution, the* quantity of ferrous salt in any solution of unknown strength. For instance, suppose there were required just 100 parts of the permanganate solution to com- pletely oxidize 2 parts of ferrous salt in solution. If, therefore, we used only 50 parts of the permanganate solution, only 1 part of ferrous salt would be indicated as being present, etc. Accordingly, by measuring the quantity of permanganate solution, the propor- tional quantity of ferrous salt is at once determined. Since the quantity of active fluid used is determined by measur- ing, and not by weighing, this method of analysis is termed volu- metric analysis. The object in view is usually much more rapidly attained by its means than by gravimetric analysis. To this brief intimation of the general purport and object of quantitative analysis, and the general mode of proceeding in ana- lytical researches, there must be added that certain qualifications are essential to those who would devote themselves successfully to the pursuit of this branch. These qualifications are, 1, theoretical knowledge ; 2, skill in manipulation ; and 3, strict conscientious- ness. The preliminary knowledge required consists in an acquaintance with qualitative analysis, the stoichiometric laws, and simple arith- metic. Thus prepared, we shall understand the method by which bodies are separated and determined, and v/e shall be in a position to perform our calculations, by which, on the one hand, the formu- las of compounds are deduced from the analytical results, and, on 4 INTRODUCTION. the other hand, the correctness of the adopted methods is tested, and the results obtained are controlled. To this knowledge must be joined the ability to perform the necessary practical operations. This axiom generally holds good for all applied sciences, but if it is true of one more than another, quantitative analysis is that one. 'The most extensive and solid theoretical acquirements will not enable us, for instance, to determine the amount of common salt present in a solution if we are without the requisite dexterity to transfer a fluid from one vessel to another without the smallest loss by spirting, running down the side, etc. The various opera- tions of quantitative analysis demand great aptitude and manual skill, which can be acquired only by practice ; but even the pos- session of the greatest practical skill in manipulation, joined to a thorough theoretical knowledge, will still prove insufficient to in- sure a successful pursuit of quantitative researches unless also combined with a sincere love of truth and a firm determination to accept none but thoroughly confirmed results. Every one who has been engaged in quantitative analysis knows that cases will sometimes occur, especially when commencing the study, in which doubts may be entertained as to whether the result will turn out correct, or in which even the operator is positively convinced that it cannot be quite correct. Thus, for instance, a small portion of the substance under investigation may be spilled or some of it lost by decrepitation ; or the analyst may have rea- son to doubt the accuracy of his weighing ; or it may happen that two analyses of the same substance do not exactly agree. In all such cases it is indispensable that the operator should be conscien- tious enough to repeat the whole process over again. He who is not possessed of this self-command who shirks trouble where truth is at stake who would be satisfied with mere assumptions and guess-work where the attainment of positive certainty is the object must be pronounced just as deficient in the necessary quali- fications for quantitative analytical researches as he who is wanting in knowledge or skill. He, therefore, who cannot fully trust his work, who cannot swear to the correctness of his results, may in- deed occupy himself with quantitative analysis by way of practice, but lie ought on no account to publish or use his results as if they were positive, since such proceeding could not conduce to his own INTRODUCTION. 5 advantage, and would certainly be mischievous as regards the science. The domain of quantitative analysis may be said to extend over all matter that is, in other words, anything corporeal may become the object of quantitative investigation. The present work, how- ever, is intended to embrace only the substances used in pharmacy, arts, trades, and agriculture. Quantitative analysis may be subdivided into two branches, viz., analysis of mixtures, and analysis of chemical compounds. This division may appear at first sight of very small moment, yet it is necessary that we should establish and maintain it if we would form a clear conception of the value and utility of quantitative research. The quantitative analysis of mixtures, too, has not the same aim as that of chemical compounds ; and the- method applied to secure the correctness of the results in the former case is differ- ent from that adopted in the latter. The quantitative analysis of chemical compounds also rather subserves the purposes of the sci- ence, while that of mixtures belongs to the practical purposes of life. If, for instance, we analyze the salt of an acid, the result of the analysis will give the constitution of that acid, its combining proportion, saturating capacity, etc. ; or, in other words, the results obtained will enable us to answer a series of questions of which the solution is important for the theory of chemical science. But if, on the other hand, we analyze gunpowder, alloys, medicinal mixtures, ashes of plants, etc., etc., we have a very different object in view. It is not intended in such cases to apply the results obtained to the solution of any theoretical question in chemistry, but we want to render a practical service either to the arts and industries or to some other science. If in the. analysis of a chemical compound we wish to control the results obtained, we may do this in most cases by means of calculations based on stoichiometric data, but in the case of a mixture, a second analysis is necessary to confirm the correctness of the results afforded by the first. The preceding remarks clearly show the immense importance of quantitative analysis. It may, indeed, be averred that chem- istry owes to this branch its elevation to the rank of a science, since quantitative researches have led us to discover and determine the laws which govern the combinations and transpositions of the ele- 6 INTRODUCTION. ments. Stoichiometry is entirely based upon the results of quanti- tative investigations; all rational views respecting the constitution of compounds rest upon them as the only safe and solid basis. Quantitative analysis, therefore, forms the strongest and most powerful lever for chemistry as a science, and not less so for chemistry in its applications to the practical purposes of life, to trades, arts, manufactures, and likewise in its application to other sciences. It teaches the mineralogist the true nature of minerals, and suggests to him principles and rules for their recognition and classi- fication. It is an indispensable auxiliary to the physiologist, and agriculture has already derived much benefit from it; but far greater benefits may be predicted. We need not expatiate here upon the advantages which medicine, pharmacy, and every branch of industry derive, either directly or indirectly, from the practical application of its results. On the other hand, the benefit thus be- stowed by quantitative analysis upon the various sciences, arts, etc., has been in a measure reciprocated by some of them. Thus, while Stoichiometry owes its establishment to quantitative analysis, the stoichiometric laws afford us the means of controlling the results of our analyses so accurately as to justify the reliance which we now generally place on them. Again, while quantitative analysis has advanced the progress of arts and industry, our manufacturers in return supply us with the most perfect platinum, glass, and por- celain vessels, and with articles of india-rubber, without which it would be next to impossible to conduct our analytical operations with the minuteness and accuracy which we have now attained. Although the aid which quantitative analysis thus derives from Stoichiometry and the arts and manufactures greatly facilitates its practice, 1 and although many determinations are considerably abbre- viated by volumetric analysis, it must be admitted, notwithstanding, that the pursuit of this branch of chemistry requires considerable expenditure of time. This remark applies especially to those who are commencing the study, for they must not allow their attention to be divided upon many things at one time, otherwise the accuracy of their results will be more or less injured. I would therefore advise every one desirous of becoming an analytical chemist to arm himself with a considerable share of patience, reminding him that it is not at one bound, but gradually, and step by step, that the. student may hope to attain the necessary certainty in his work, the INTRODUCTION. 7 indispensable self-reliance which can alone be founded on one's own results. However mechanical, protracted, and tedious the operations of quantitative analysis may appear to be, the attain- ment of accuracy will amply compensate for the time and labor bestowed upon them; while, on the other hand, nothing can be more disagreeable than to find, after a long and laborious process, that our results are incorrect or uncertain. Let him, therefore, who would render the study of quantitative analysis agreeable to himself, from the very outset endeavor, by strict, nay, scrupulous adherence to the conditions laid down, to attain correct results at any sacrifice of time. I scarcely know a better and more imme- diate reward of labor than that which springs from the attainment of accurate results and perfectly corresponding analyses. The satis- faction enjoyed at the success of our efforts is surely in itself a sufficient motive for the necessary expenditure of time and labor, ven without looking to the practical benefits which we may derive from our operations. The following are the substances treated of in this work : I. METALLOIDS. Oxygen, Hydrogen, Sulphur, [Selenium~\, Phosphorus, Chlo- rine, Iodine, Bromine, Fluorine, Nitrogen, Boron, Silicon, Carbon. II. METALS. Potassium, Sodium, [Lithium, ~\ Barium, Strontium, Calcium, Magnesium, Aluminium, Chromium, [Titanium^] Zinc, Manga- nese, Nickel, Cobalt, Iron, [Uranium,'} Silver, Mercury, Lead y Copper, Bismuth, Cadmium, {Palladium,} Gold, Platinum, Tin, Antimony, Arsenic, {Molybdenum. ,] (The elements enclosed within brackets are considered in sup- plementary paragraphs, and more briefly than the rest.) The subject has been divided into three parts. In the first quantitative analysis generally is treated of, describing the execu- tion of analysis. In the second is given a detailed description of several special analytical processes. In the third a number of 8 INTRODUCTION. carefully selected examples, which may serve as exercises for the groundwork of the study of quantitative analysis. The following table will afford the reader a clear and definite notion of the contents of the whole work : I. GENEEAL PAET. A. Execution of analysis. 1. Operations. 2. Reagents. 3. Forms and combinations in which bodies are separated from others, or in which their weight is determined. 4. Determination of bodies in simple compounds. 5. Separation of bodies. 6. Organic elementary analysis. B. Calculation of the results. II. SPECIAL PAET. 1. Analysis of waters, more especially mineral waters. 2. Analysis of such minerals and technical products as are most frequently the object of chemical investigation, including methods for ascertaining their commercial value. 3. Analysis of plant ashes. 4. Analysis of soils. 5. Analysis of manures. 6. Analysis of atmospheric air. III. EXEECISES FOE PEACTICE. APPENDIX. 1. Analytical experiments. 2. Tables for calculating analytical results. PART I. GENERAL PART. THE EXECUTION OF ANALYSIS. SECTION" I. OPEKATIONS. * -4 JL MOST of the operations performed in quantitative research are the same as in qualitative analysis, and have been accordingly described in my work on that branch of analytical science. "With respect to such operations I shall, therefore, confine myself here to pointing out any modifications they may require to adapt them for applica- tion in the quantitative branch ; but there will, of course, be giren a full description of such as are resorted to exclusively in quanti- tative investigations. Operations forming merely part of certain specific processes will be found described in the proper place, under the head of such processes. I. DETERMINATION OF QUANTITY. 2. The quantity of solids is usually determined by weight; -the quantity of gases and fluids, in many cases by measure; and upon the care and accuracy with which these operations are per- formed, depend the value of all our results. We shall therefore dwell minutely upon them. 3. 1. WEIGHING. To enable us to determine with precision the correct weight of a substance, it is indispensable that we should possess, 1st, a BALANCE, and 2d, accurate WEIGHTS. OPERATIONS. [4. a. THE BALANCE. Fig. 1 represents a form of balance well adapted for analytical purposes. There are several points respecting the construction and properties of a good balance, which it is absolutely necessary for every chemist to understand. The usefulness of this instrument depends upon two points : 1st, its accuracy, and 2d, its sensibility or delicacy. M- The ACCURACY of a balance depends upon the following condi- tions : a. The fulcrum or the point on which the beam rests must lie above the centre of gravity of the balance. Fig. 1. This is in fact a condition essential to every balance. If the fulcrum were placed in the centre of gravity of the balance, the beam would not oscillate, but remain in any position in which it is placed, assuming the scales to be equally loaded. If the fulcrum be placed below the centre of gravity, the balance will be overset by the slightest impulse. When the fulcrum is above the centre of gravity the balance represents a pendulum, the length of which is equal to that of the line uniting the fulcrum with the centre of gravity, and this line forms right angles with the beam, in whatever position the latter may be placed. Now, if we impart an impetus to a ball suspended by a thread, the ball, after having terminated its vibrations, will 4] WEIGHING. invariably rest in its original perpendicular position under the point of suspension. It is the same with a properly adjusted bal- ance impart an impetus to it, and it will oscillate for some time, but it will invariably return to its original position ; in other words, its centre of gravity will finally fall back into its perpen- dicular position under the fulcrum, and the beam must consequently ivussuine the horizontal position. But to judge correctly of the force with which this is accom- plished, it must be borne in mind that a balance is not a simple pendulum, but a compound one, i. e., a pendulum in which not one, but many material points move round the turning point. The inert mass to be moved is accordingly equal to the sum of these points, and the moving force is equal to the excess of the material points below, over those above the fulcrum. p. The points of suspension of the scales must be on an exact level with the fulcrum. If the fulcrum be placed below the line joining the points of suspension, increased loading of the scales will continually tend to raise the centre of gravity of the whole system, so as to bring it nearer and nearer the fulcrum ; the weight which presses upon the scales combining in the relatively high- placed points of suspension ; at last, when the scales have been loaded to a certain degree, the centre of gravity will shift alto- gether to the fulcrum, and the balance will consequently cease to vibrate any further addition of weight will finally overset the beam by placing- the centre of gravity above the fulcrum. If, on the other hand, jthe fulcrum be placed above the line joining the points of suspension, the centre of gravity will become more and more depressed in proportion as the loading of the scales is in- creased ; the line of the pendulum will consequently be length- ened, and a greater force will be required to produce an equal turn ; in other words, the balance will grow less sensitive the greater the load. But when the three edges are in one plane, in- creased loading of the scales will, indeed, continually tend to raise the centre of gravity towards the fulcrum, but the former can in this case never entirely reach the latter, and consequently the bal- ance will never altogether cease to vibrate upon the further addi- tion of weight, nor will its sensibility be lessened ; on the contra ry speaking theoretically a greater degree of sensibility is im- parted to it. This increase of sensibility is, however, compensated for by other circumstancas. (See 5.) 14 OPERATIONS. [ 5. y. The beam must be sufficiently rigid to bear without bend- ing the greatest weight that the construction of the balance admits of. The bending of the beam would of course depress the points of suspension so as to place them below the fulcrum, and this would, as we have just seen, tend to diminish the sensibility of the balance in proportion to the increase of the load. It is, therefore, necessary to avoid this fault by a proper construction of the beam. The form best adapted for beams is that of an isosceles obtuse- 'angled triangle, or of a rhombus. d. The arms of the balance must Ue of equal length, i. e., the points of suspension must be equidistant from the fulcrum, for if the arms are of unequal length the balance will not be in equili- brium, supposing the scales to be loaded with equal weights, but there, will be preponderance on the side of the longer arm. 5. The SENSIBILITY of a balance depends principally upon the three following conditions : a. The friction of the edges upon their supports must be as slight as possible. The greater or less friction of the edges upon their supports depends upon both the form and material of those parts of the balance. The edges must be made of good steel, the supports may be made of the same material ; it is better, however, that the centre edge at least should rest upon an agate plane. To form a clear conception of how necessary it is that even the end edges should have as little friction as possible, we need simply reflect upon what would happen were we to fix the scales immov- bly to the beam by means of rigid rods. Such a contrivance would at once altogether annihilate the sensibility of a balance, for if a -weight were placed upon one scale, this certainly would have a tendency to sink ; but at the same time the connecting rods be- ing compelled to form constantly a right angle with the beam, the weighted scale would incline inwards, whilst the other scale would turn outwards, and thus the arms would become unequal, the shorter arm being on the side of the weighted scale, whereby the tendency of the latter to sink would be immediately compensated for. The more considerable the friction becomes at the end edges of a balance, the more the latter approaches the state just now described, and consequently the more is its sensibility impaired. yd The centre of gravity must be as near as possible to theful- .^ ;j ) 6.] WEIGHING. 15 The nearer the centre of gravity approaches the fulcrum, the shorter becomes the pendulum. If we take two balls, the one suspended by a short and the other by a long thread, and impart the same impetus to both, the former will naturally swing at a far greater angle from its perpendicular position than the latter. The same must of course happen with a balance ; the same weight will cause the scale upon which it is placed to turn the more rapidly mid completely, the shorter the distance between the centre of gravity and the fulcrum. We have seen above, that in* a balance where the three edges are on a level with each other, increased loading of the scales will continually tend to raise the centre of gravity towards the fulcrum. A good balance will therefore be- come more delicate in proportion to the increase of weights placed upon its scales; but, on the other hand, its sensibility will be di- minished in about the same proportion by the increment of the muss to be moved, and by the increased friction attendant upon the increase of load ; in other words, the delicacy of a good balance will remain the same, whatever may be the load placed upon it. The nearer the centre of gravity lies to the fulcrum, the slower are the oscillations of the balance. Hence in regulating the position of the centre of gravity we must not go too far, for if it ap- proaches the fulcrum too nearly, the operation of weighing will take too much time. y. The learn must ~be as light as possible. The remarks which we have just now made will likewise show how far the weight of the beam may influence the sensibility of a balance. We have seen that if a balance is not actually to become less delicate on increased loading, it must on the one hand have a tendency to become more delicate by the continual approach of the centre of gravity to the fulcrum. ~Now it is evident, that the more considerable the weight of the beam is, the less will an equal load placed upon both scales alter the centre of gravity of the whole system, the more slowly will the centre of gravity approach the fulcrum, the less will the increased friction be neutralized, and consequently the less sensi- bility will the balance possess. Another point to be taken into account here is, that the moving forces being equal, a less mass or weight is more readily moved than a greater. ( 4 a.) 6. We will now proceed, first, to give the student a few general 16 OPERATIONS. [ 6. rules to guide him in the purchase of a balance intended for the purposes of quantitative analysis ; and, secondly, to point out the best method of testing the accuracy and sensibility of a balance. 1. A balance able to bear TO or 80 grammes in each scale, suf- fices for most purposes. 2. The balance must be enclosed in a glass case to protect it from dust. This case ought to be sufficiently large, and, more especially, its sides should not approach too near the scales. It must be constructed in a manner to admit of its being opened and closed with facility, and thus to allow the operation of weighing to be effected without any disturbing influence from currents of air. Therefore, either the front part of the case should consist of three parts, viz., a fixed centre part and two lateral parts, open- ing like doors ; or, if the front part happens to be made of one piece, and arranged as a sliding-door, the two sides of the case must be provided each with a door. 3. The balance must be provided with a proper contrivance to render it immovable whilst the weights are being placed upon the scale. This is most commonly effected by an arrangement which enables the operator to lift up the beam and thus to remove the middle edge from its support, whilst the scales remain suspended. Older contrivances check the scales without lifting the middle edge from its support. It is very convenient to have a stop both for the scale-pans and the beam. The newer balances are almost always so provided. The usual device for checking the scale- pans consists of two supports immediately below them, which slide up and down, and are provided with crossed silk ribbons or camel's- hair brushes. The supports must move with such perfect steadi- ness that, when carefully removed from the scale-pans, the latter do not shake in the least. This arrangement is of advantage in facilitating the loading of the scale-pans, besides enabling an im- mediate stop to be put to trembling or shaking of the scales, and is convenient also because in cases where the same body has to be weighed repeatedly, the weights may be left on the scale-pan without risk to the balance. Stops that check the beam and scale- pans by one action (a turn) appear to me less practical, because the checking of the scale-pans after every addition of a small weight is purposeless, while it impairs the rapidity of weighing. It is highly advisable to have the checking contrivances arranged to be manipulated from without, while the glass case remains closed. 7.] WEIGHING. 17 4. It is necessary that the balance be provided with an index or pointer to mark its oscillations on a graduated arc ; and it is more advantageous to have the index beneath the axis rather than at the side of the balance. 5. The balance must be provided with a pendulum or spirit level in order that the three edges may be placed on an exactly horizontal level; it is hence practical to have the case rest on three screws. 6. It is very convenient and time-saving for the beam to be graduated decimally, so that by means of a centi- gramme u rider," Fig. 2, milligrammes and their fractions may be weighed. Modern balances are so constructed that the rider may be shifted to any posi- tion on the beam, without opening the case, by means Fig. 2. of a movable arm passing through the side of the case.* 7. The balance must be provided with a screw to regulate the centre of gravity, and likewise with two screws to regulate the equality of the arms, and finally with screws to restore the equi- librium of the scales, should this have been disturbed. 7. The following experiments serve to test the accuracy and sensi- bility of a balance : 1. The balance is, in the first place, accurately adjusted, if necessary, either by the regulating screws, or by means of tinfoil, and a milligramme weight is then placed in one of the scales. A good and practically useful balance must turn very distinctly with this weight; a delicate chemical balance should indicate 0.1 milli- gramme distinctly. It should be noted here that the mere point- ing of the index to zero is not sufficient evidence of equilibrium. It is much better to observe the oscillations of the pointer, which may be effected, if necessary, by a slight move of the hand near one of the scale-pans so as to produce a slight wind. The vibra- tion must be nearly equal on both sides, growing less with each vibration, until the pointer finally comes to rest at zero. * HEMPEL, of Paris, puts a very complete arrangement for placing small weights and shifting the rider, on his balances (see Zeitschr. f. analyt. Chern., iv, 83). I have had no personal experience with it, however. 18 OPERATIONS. [ 7. 2. Both scales are loaded with the maximum weight the con- struction of the balance will admit of; the balance is then accu rately adjusted, and a milligramme added to the weight in the one scale. This ought to cause the balance to turn to the same extent as in 1. In most balances, however, it shows somewhat less on the index. It follows from 5 ft that the. balance will oscillate more slowly in this than in the first experiment. 3. The balance is accurately adjusted (should it be necessary to establish a perfect equilibrium between the scales by loading the one with a minute portion of tinfoil, this tinfoil must be left re- maining upon the scale during the experiment) ; both scales are then equally loaded, say, with fifty grammes each, and, if neces- sary, the balance is again adjusted (by the addition of small weights). The load of the two scales is then interchanged, so as to transfer that of the right scale to the left, and vice versa. A balance with perfectly equal arms must maintain its absolute equi- librium upon this interchange of the weights of the two scales. 4. The balance is accurately adjusted ; it is then arrested and again set in motion ; the same process should be repeated several times. A good balance must invariably reassume its original equi- librium. A balance the end edges of which afford too much play to the hook resting upon them, so as to allow the latter slightly to alter its position, will show perceptible differences in different trials. This fault, however, is possible only with balances of defec- tive construction.* A balance to be practically useful for the purposes of quantita- tive analysis must stand the first, second, and last of these tests. A slight inequality of the arms is of no great consequence, as the error that it would occasion may be completely prevented by the manner of weighing. As the sensibility of a balance will speedily decrease if the steel edges are allowed to get rusty, delicate balances should never be kept in the laboratory, but always in a separate room.' It is also advisable to place within the case of the balance a vessel half filled with calcined potassium carbonate, to keep the air dry. I need hardly add that this salt must be recalcined as soon as it gets moist. * G. WESTPHAL, of Celle, has described a mode of construction which abso- lutely excludes the possibility of this fault (Zeitschr.f. analyt. Chem , vii, 294). 8.] WEIGHTING. 19 5. THE WEIGHTS. Intrinsically, it is quite immaterial what unit of weight is adopted. Most chemists, however, use the gramme weights because of convenience in recording as well as in calculating. With regard to the set of weights, it is generally a matter of indif- ference for scientific purposes whether the gramme, its multiples, and fractions are really and perfectly equal to the accurately adjusted normal weights of the corresponding denominations;* but it is absolutely necessary that they should agree perfectly among themselves, i.e., the centigramme weight must be exactly the one- hundredth part of the gramme weight of the set, etc., etc. Before describing the testing of the weights as to their accuracy, attention must be called to the following points: 1. A set of weights ranging from one milligramme to fifty grammes fully suffices for most purposes. 2. The set of weights should be kept in a suitable, well- closing box ; and it is desirable likewise that a distinct compartment be appropriated to every one, even of the smaller weights. 3. As to the shape best adapted for weights, I think that of short frusta of cones inverted, with a handle at the top, the most convenient and practical form for the large weights; square pieces of foil, turned up at one corner, are best adapted for the small weights. The foil used for this purpose should not be too thin, and the compartments adapted for the reception of the several smaller weights in the box should be large enough to admit of their contents being taken out of them with facility, or else the smaller weights will soon become crumpled and defaced. Every one of the weights (with the exception of the milligramme) should be dis- tinctly marked. 4. So far as the material most suitable for weights is con- cerned, rock crystal, though best adapted for normal weights, is unsuitable for ordinary weights because of its high cost and the inconvenient form the weights would have. Platinum, were it not so costly, would be surely adopted generally, on account of its unchangeability. As a rule, however, platinum is used for weights * It were desirable that makers of analytical weights endeavor to procure normal weights. It is very annoying, in many cases, to find notable differences between weights of the same denomination, but coming from different makers, as I have frequently observed. 20 OPERATIONS. [ 8. smaller than 1 gramme or 0*5 gramme, while brass is used for all the higher denominations. Brass weights must be carefully shielded from the action of acid or other vapors, or their correct- ness will be impaired ; nor should they ever be touched with the fingers, but always with small pincers. It is an erroneous notion to suppose that weights slightly tarnished are unfit for use. In fact, it is scarcely possible to keep weights for any very great length of time from becoming slightly tarnished. I have carefully examined many weights of this description, and have found them to correspond as exactly with one another in their relative propor- tions as they did when first used. The tarnishing coat is so extremely thin that even a very delicate balance will generally fail to point out any perceptible difference in the weight. It will nevertheless be found advantageous to gild the brass weights pre- vious to their final adjustment. The following is the proper way of testing the weights : One scale of a delicate balance is loaded with a one-gramme weight, and the balance is then completely equipoised by taring with small pieces of brass, and finally tinfoil (not paper, since this absorbs moisture). The weight is then removed, and replaced suc- cessively by the other gramme weights, and afterwards by the same amount of weight in pieces of lower denominations. The balance is carefully scrutinized each time, and any devia- tion from the exact equilibrium marked. In the same way it is seen whether the two-gramme piece weighs as much as two single grammes, the five-gramme piece as much as three single grammes and the two-gramme piece, etc. In the comparison of the smaller weights thus among themselves, they must not show the least dif- ference on a balance turning with 0*1 milligramme. In comparing the larger weights with all the small ones, differences of 0*1 to 0"2 milligramme may be passed over. If you wish them to be more accurate, you must adjust them yourself. In the purchase of weights chemists ought always to bear in mind that an accurate weight is truly valuable, whilst an inaccurate one is absolutely worthless." 55 ' It is the safest way for the chemist to test every weight he purchases, no matter how high the reputation of the maker. * Compare W. CROOKES, on adjusting chemical weights (Zeitschr. f. Analyt Chem., vi, 431), and K. L. BAUER (ibid., vin, 390). 9.J WEIGHING. 21 9. c. THE PKOCESS OF WEIGHING. There are two different methods of determining the weight of substances ; the one might be termed direct weighing, the other is called weighing by substitution. In direct weighing, the substance is placed upon one scale, and the weight upon the other. If we possess a balance, the arms of which are of equal length, and the scales in a perfect state of equilibrium, it is indifferent upon which scale the substance is placed in the several weighings required during an analytical pro- cess ; i.e., we may weigh upon the right or upon the left side, and change sides at pleasure, without endangering the accuracy of our results. But if, on the contrary, the arms of our balance are not perfectly equal, or if the scales are not in a state of perfect equili- brium, we are compelled to weigh invariably upon the same scale, otherwise the correctness of our results will be more or less materi- ally impaired. Suppose we want to weigh one gramme of a substance, and to divide this subsequently into two equal parts. Let us assume our balance to be in a state of perfect equilibrium, but with unequal arms, the left being 99 millimeters, the right 100 millimeters, long; we place a gramme weight upon the left scale, and against this, on the right scale, as much of the substance to be weighed as will restore the equilibrium of the balance. According to the axiom, " masses are in equilibrium upon a lever, if the products of their weights into their distances from the fulcrum are equal," we have consequently upon the right scale 0'99 grm. of substance, since 99 X 1*00=: 100 X 0'99. If we now, for the purpose of weighing one half the quantity, remove the whole weight from the left scale, substituting a 0*5 grm. weight for it, and then take off part of the substance from the right scale, until the balance recovers its equilibrium, there will remain 0*495 grm. ; and this is exactly the amount we have removed from the scale : we have consequently accomplished our object with respect to the relative weight ; and as we have already remarked, the absolute weight is not generally of so much importance in scientific work. But if we attempted to halve the substance which we have on the right scale, by first removing both the weight and the substance 22 OPERATIONS. [ 9. from the scales, and placing subsequently a 0'5 grm. weight upon the right scale, and part of the substance upon the left, until the balance recovers its equilibrium, we should have 0*505 grm. of substance upon the left scale, since 100 X 0*5 = 99 X 0'505; and consequently, instead of exact halves, we should have one part of the substance amounting to 0'505, the other only to 0-485, grm. If the balance is equal-armed, but the scale-pans are not in a state of absolute equilibrium, we are obliged to weigh our sub- stances in vessels to insure accurate results (although the arms of the balance be perfectly equal). It is self-evident that the weights in this case must likewise be invariably placed upon one and the same pan, and that the difference between the two scale-pans must not vary during the course of a series of experiments. From these remarks result the two following rules : 1. It is, under all circumstances, advisable to place the sub-, stance invariably upon one and the same pan most conveniently upon the left. 2. If the operator happens to possess a balance for his own private and exclusive use, there is no need that he should adjust it at the commencement of every analysis ; but if the balance be used in common by several persons, it is absolutely necessary to ascer- tain, before every operation, whether the state of absolute equili- brium may not have been disturbed. Weighing by substitution yields not only relatively, but also absolutely accurate results ; no matter whether the arms of the balance be of exactly equal lengths or not, or whether the scales be in perfect equipoise or not. The process is conducted as follows : The material to be weighed say a platinum crucible is placed upon one scale, and the other scale is accurately counterpoised against it. The plati- num crucible is then removed, and the equilibrium of the balance restored by substituting weights for the removed crucible. It is perfectly obvious that the substituted weights will invariably express the real weight of the crucible with absolute accuracy. We weigh by substitution whenever we require the greatest pos' sible accuracy; as, for instance, in the determination of atomic weights. The process may be materially shortened by first placing a tare (which must of course be heavier than the substance to be weighed) upon one scale, say the left, and loading the other scale 10.] WEIGHING. 23 with weights until equilibrium is produced. This tare is always retained on the left scale. The weights after being noted are re- moved. The substance is placed on the right scale, together with the smaller weights requisite to restore the equilibrium of the balance. The sum of the weights added is then subtracted from the noted weight of the counterpoise : the remainder will at once indicate the absolute weight of the substance. Let us suppose, for instance, we have on the left scale a tare requiring a weight of fifty grammes to counterpoise it. We place a platinum crucible on the right scale, and find that it requires an additional weight of 10 grammes to counterpoise the tare on the left. Accordingly, the crucible weighs 50 minus 10, i.e. , 40 grammes. 10. The following rules will be found useful in performing the process of weighing : 1. The balance should be kept in a dry place, protected from acid vapors, etc., and, if possible, not exposed to direct sunlight. It should be placed on a firm support and in a level position ; nor must it be too near the source of heat, should the room be heated, otherwise it may be unequally heated, 2. The safest and most expeditious method of ascertaining the exact weight of a substance, is to avoid trying weights at random ; instead of this, a strictly systematic course ought to be pursued in counterpoising substances on the balance. Suppose, for instance, we want to weigh a crucible, the weight of which subsequently turns out to be 6 '627 grammes; we first place 10 grammes on the other scale against it, and we find this is too much ; we place the weight next in succession, i.e., 5 grammes, and find this too little; next Y, too much ; 6, too little ; 6*5, too little ; 6 -7, too much ; 6 -6, too little; 6*65, too much; 6'62, too little; 6'63, too much; 6-625, too little; 6 '627, right. For the sake of illustration, a most complicated case has been selected ; this systematic way of laying on the weights will, how- ever, in most instances lead to the desired end, in half the time required when weights are tried at random. After a little practice a few minutes will suffice to ascertain the weight of a substance to within O'l milligramme, provided the balance does not oscillate too slowly. 24 OPERATIONS. L10. 3. The milligrammes and fractions of milligrammes are deter- mined by a centigramme rider (to be placed on or between the divisions on the beam) far more expeditiously and conveniently than by the use of the weights themselves, and at the same time with equal accuracy. 4. Particular care and attention should be bestowed on enter- ing the weights in the book. The best way is to write down the weights first by inference from the blanks, or gaps in the weight box, and to control the entry subsequently by removing the weights from the scale, and replacing them in their respective compartments in the box. The student should from the commencement make it a rule to enter the number to be deducted in the lower line; thus, in the upper line, the weight of the crucible + the substance ; in the lower line, the weight of the empty crucible. 5. The balance ought to be arrested every time any change is contemplated, such as removing weights, substituting one weight for another, etc. etc. , or it will soon be spoiled. 6. Substances (except, perhaps, pieces of metal, or some other bodies of the kind) must never be placed directly upon the scales, but ought to be weighed in appropriate vessels of platinum, silver, glass, porcelain, etc., never on paper or card, since these, being liable to attract moisture, are apt to alter in weight. The most common method is to weigh in the first instance the vessel by itself, and to introduce subsequently the substance into it, to weigh again, and subtract the former weight from the latter. In many instances, and more especially where several portions of the same substance are to be weighed, the united weight of the vessel and of its con- tents is first ascertained ; a portion of the contents is then shaken out, and the vessel weighed again ; the loss of weight expresses the amount of the portion taken out of the vessel. 7. Substances prone to attract moisture from the air must be weighed invariably in closed vessels (in covered crucibles, for instance, or between two watch-glasses, or in a closed glass tube); fluids are to be weighed in small bottles closed with glass stoppers. 8. A vessel ought never to be weighed while warm, since it 5 O will in that case invariably weigh lighter than it really is. This is owing to two circumstances. In the first place, every substance condenses upon its surface a certain quantity of air and moisture,, the amount of which depends upon the temperature and hygro- 10.] WEIGHING. 25 scopic state of the air, and likewise on its own temperature. Now suppose a crucible has been weighed cold at the commencement of the operation, and is subsequently weighed again while hot, together with the substance it contains, and the weight of which we wish to determine. If we subtract for this purpose the weight of the cold crucible, ascertained in the former instance, from the weight found in the latter, we shall subtract too much, and consequently we shall set down less than the real weight for the substance. In the second place, bodies at a high temperature are constantly communicating heat to the air immediately around them ; the heated air expands and ascends, and the denser and colder air, flowing towards the space which the former leaves, pro- duces a current which tends to raise the scale, making it thus appear lighter than it really is. 9. If we suspend from the end edges of a correct balance respectively 10 grammes of platinum and 10 grammes of glass, by- wires of equal weight, the balance will assume a state of equili- brium ; but if we subsequently immerse the platinum and glass completely in water, this equilibrium will at once cease, owing to the different specific gravity of the two substances ; since, as i& well known, substances immersed in water lose of their weight a portion equal to the weight of their own bulk of water. If this be borne in mind, it must be obvious to every one that weighing in the air is likewise defective, inasmuch as the bulk of the substance weighed is not equal to that of the weight used. This defect, however, is so very insignificant, owing to the trifling specific gravity of the air in proportion to that of solid substances, that we may generally disregard it alto- gether in analytical experiments. In cases, however, where altfiolutcly accurate results are required, the bulk both of the substance examined, and of the weight, must be taken into account, and the weight of the corresponding volume of air added respectively to that of the substance and of the weight, making thus the process equivalent to weighing in vacuo. 26 OPERATIONS. [ 11. 11. 2. MEASURING. The process of measuring is confined in analytical researches mostly to gases and liquids. The method of measuring gases has been brought to such perfection by BUNSEN, REGNAULT and REISET, FRANKLAND and WARD, WILLIAMSON and RUSSELL, and by others, that it may be said to equal in accuracy the method of weighing. However, such accurate measurements demand an expenditure of time and care which can be bestowed only on the nicest and most delicate scientific investigations.* The measuring of liquids in analytical investigations was resorted to first by DESCROIZILLES (Alkalimeter, 1806). GAY-LUSSAG materially improved the process, and indeed brought it to the highest degree of perfection (measuring of the solution of sodium chloride in the assay of silver in the wet way). More recently F. MOHR f has bestowed much care and ingenuity upon the pro- duction of appropriate and convenient measuring apparatus, and has added to our store the eminently practical pinch-cock burette. The process is now resorted to even in most accurate scientific investigations, since it requires much less time than the process of weighing. The accuracy of all measurings depends upon the proper con- * A detailed description of BUNSEN'S method is to be found in the Hand- worterbuch der Chemie, by LIEBIG, POGGENDORFF, and WOHLER, n, 1053 (KOLBE'S Eudiometer), and i, 2, 2d edit., 930 (Volumetric Analysis of Gases, by KOLBE and FRANKLAND). BUNSEN, further, wrote a valuable monograph on this subject under the title Gasometric Methods, and published by FR. VIEWEG & SON, Brunswick, 1857, and which was translated by ROSCOE. The methods of gas measurement employed by KEGNAULT and REISET, as well as by FRANK- LAND and WARD, differ from the improved BUNSEN method in that in the former the measuring tubes stand in cylinders filled with water, whereby the temperature of the gas is brought in a few minutes to that of the water, thus materially shortening the time required in gas analysis. In the FRANKLAND- WARD method the gasometric determination is also independent of the atmos- pheric pressure. Both methods, as a matter of course, require complicated and expensive apparatus. These are minutely described and figured in the above- mentioned article by FRANKLAND. The WILLIAMSON-RUSSELL gasometric apparatus is described in the Jour. Chem. Soc., xvn, 238; and RUSSELL'S modi- fication, ibid. (2), vi, 128, also in Zeitschr.f. analyt. Chem., vn, 454. f Lehrbuch der Titrirmethode, Dr. Fr. MOHR. 12.] MEASURING OF GASES. 27 strnction of the measuring vessels, and also upon the manner in which the process is conducted. 12. a. THE MEASURING OF GASES. We use for the measuring of gases graduated tubes of greater or less capacity, made of strong glass, and sealed at one end, which should be rounded. The following tubes will be found sufficient for all the processes of gas measuring required in organic elementary analyses : 1. A bell-glass capable of holding from 150 to 250 c. c., and about 4 centimetres in diameter, and divided into cubic centimetres. 2. Five or six glass tubes of about 12 to 15 millimetres bore diameter, and capable of holding from 30 to 40 c. c. each, divided into 0-2 c. c. The sides of these tubes should be fairly thick, otherwise they will be liable to break, especially when used to measure over mer- cury. The sides of the bell-glass should be about 3, of the tubes about 2, millimetres thick. The most important point, however, in connection with meas- uring instruments is that they be correctly graduated, since upon this of course depends the accuracy of the results. For the method of graduating consult BERZELIUS' " Lehrbuch der Chemie^ 4th ed., x, under Messenj also GREVILLE WILLIAMS' " Chemical Manipulation. ' ' In testing the measuring tubes we have to consider three questions : 1. Do the divisions of a tube correspond with each other? 3. Do the divisions of each tube correspond with those of the other tubes ? 3. Do the volumes expressed by the graduation lines corre- spond with the weights used by the analyst ? These three questions are answered by the following experi- ments : a. The tube which it is intended to examine is placed in a per- pendicular position, and filled gradually with accurately measured small quantities of mercury, care being taken to ascertain with the utmost precision whether the graduation of the tube is proportion- ate to the equal volumes of mercury poured in. The measuring- OF THE UNIVERSITY 28 OPERATIONS. [ 12, off of the mercury is effected by means of a small glass tube, sealed at one end, and ground perfectly even and smooth at the other. This tube is filled to overflowing by immersion under mercury y care being taken to allow no air bubbles to remain in it ; the excess of mercury is then removed by pressing a small glass plate down on the smooth edge of the tube.* ~b. Different quantities of mercury are successively measured off in one of the smaller tubes, and then transferred into the other tubes. The tubes may be considered in perfect accordance with each other, if the mercury reaches invariably the same divisional point in every one of them. Such tubes as are intended simply to determine the relative volume of different gases, need only pass these two experiments ; but in cases where we want to calculate the weight of a gas from its volume, it is necessary also to obtain an answer to the third question. For this purpose c. One of the tubes is accurately weighed and then filled with distilled water of a temperature of 17 '5 to the last mark of the graduated scale ; the weight of the water is then accurately deter- mined. If the tube agrees with the weight, every 100 c. c. of water at 17*5 must weigh 99 '78 grammes. f Should it not agree, no matter whether the error is due to faulty weights or incorrect graduations, we must apply a correction to the volume observed before calculating the weight therefrom. For instance, if 100 c. c. had been found to weigh 100 grammes assuming our weight to be perfect then the c. c. divisions would be too large, and to convert 100 c. c. into normal c. c. the following calculation would have to be made : 99-78 : 100 :: 100 : x. In gas analysis proper by BUNSEN'S methods (the simplest and most accurate) a suitable eudiometer is indispensable. BUNSEN'S eudiometer, Fig. 3, is a glass tube 500 to 600 mm. long, with a bore of 20 mm., as uniform as possible throughout, and the thick- ness of the glass not exceeding 2 mm. At the upper, closed end * As warming the metal is to be carefully avoided in this process, it is advis- able not to hold the tube with the hand in immersing it in the mercury, but to fasten it in a small wooden holder. f A gramme is the weight of 1 c. c. of water in vacuo at 4. 12.] MEASURING OF GASES. 29 of the tube there are sealed in at opposite sides two fine platinum wires. These wires are ^ bent internally to lie close to the walls of Fig. 3. Fig. 4. the tube, and approach each other at the apex of the tube until separated by a distance of 1 to 2 mm. The tube is graduated in millimetre divisions by means of an ingenious divider. The volumes corresponding to the several divi- sions are then determined by measurement with equal volumes of mercury, and noted down in a table. This method of dividing and adjusting is unquestionably the most accurate. Besides this large eudiometer there is required also a short one, Fig. 4, similarly graduated in millimetres, and slightly curved at the lower end. Its length is 250 mm., and its bore 20 mm. in diameter ; the glass should be 2 mm. thick. BUNSEN'S method of gas analysis requires a laboratory with a northern exposure and uniform temperature, and consumes much time because of the slow cooling of the gases. In order to adapt the method for the use of those who do not possess a suitable labo- ratory, and to shorten the time, O. KERSTEN* recommends that the BUNSEN eudiometer be provided with a screw stopper like that in BUNSEN'S absorptiometer tube,f and that the readings should be taken after immersing the eudiometer in water. The same result is obtained in another manner in the eudiometer recommended by J. P. COOKE.J In measuring gases attention must be given to the following points: 1. Correct reading of the results; 2, the temperature of the gas ; 3, the pressure under which the gas is confined ; and 4, the circumstance whether the gas is dry or moist. ' The three last points will be readily understood when it is remembered that a *Zeitschr.f. analyt. Chem., I, 281. fBuNSEN, Gasometr. Meth,, p. 147. \Zeitschr.f. analyt. Chem., vn, 86. 30 OPERATIONS. [ 13, given weight of gas undergoes considerable alteration in volume by changes in temperature or pressure, as well as from greater or less tension of the admixed aqueous vapor. 13. 1 . CORRECT READING OF RESULTS. "When mercury is introduced into a tube, it exhibits a convex surface, because of its cohesion, the phenomenon being particularly striking in a narrow glass tube. On the other hand water, under similar circumstances, exhibits a concave surface, owing to the attraction between the tube- wall and the water. These circumstances render accurate reading-oil' diffi- cult. The tube must invariably be placed perpendicularly, with the eye on a level with the surface of the fluid. This is effected by the aid of two plummets suspended a short distance from the tube and at a proper distance from each other, or by the aid of any convenient perpendicular door- or window-edge. A small mirror is then pressed against the back of the tube, and the eye lixed on it across the surface of the liquid. When the eye has thus been placed in the proper position, the mirror is removed, and the reading taken. Instead of a mirror, BUNSEN generally employs a horizontal telescope, movable vertically, and placed at a distance of four to six paces from the eudiometer. This arrangement, while very greatly facilitating the reading, is of especial advantage in the measurement of gases, in that the observer is placed some distance from the eudiometer, thus avoiding any expansion of the gases such as is to be apprehended by the close proximity required in using the mirror. In taking the reading over water, the middle of the dark zone formed by the water drawn up the walls of the tube, is to be taken as the actual surface. When operating with mercury, however, we have to place the real surface in a plane exactly in the middle between the highest point of the surface of the mercury, and the points at which the latter is in actual contact with the walls of the tube. However, the results obtained in this way are only approx- imate. Absolutely accurate results cannot be arrived at, in measuring over water or any other fluid that adheres to glass. But over mer- 13.] MEASURING OF GASES. 31 cury they may be arrived at if the error of the meniscus be deter- mined and the mercury be read off at the highest -point. The determination of the error of the meniscus is performed for each tube, once for all, in the following manner : Some mercury is poured into the tube, and its height read -off right on a level with the top of the convex surface exhibited by it ; a few drops of solu- tion of mercuric chloride are then poured on the top of the metal ; this causes the convexity to disappear. The height of the mercury in the tube is now read-off again and the difference noted. In the process of graduation, the tube stands upright, in that of measur- ing gases, it is placed upside down ; the difference observed must accordingly be doubled, and the sum added to each volume of gas read-off. The mercury used in measuring gases must be as pure as pos- sible, and must more particularly be free from lead and tin, as these impart to it the property of adhering to glass. Should they be present, they may be most readily removed by subjecting the mercury for a day to the action of diluted nitric acid, in a shallow porcelain dish, with frequent stirring. Dust, etc., maybe removed by filtration through a cloth. As a pneumatic trough, that devised by BUNSEN, Fig. 5, will be found convenient. A is a piece of pear-wood 310 to 350 mm. long, 80 to 86 mm. wide, with a cavity chiseled in it 240 to 250 mm. long, 50 mm. wide, and 50 mm. deep. The bottom of the cavity is round, except at one end, where a flat surface 32 mm. wide and 50 mm. long is prepared. On this surface is ce- mented a sheet of vulcanized caoutchouc 3 mm. thick. To A are joined two end pieces, BB, each 19 mm. thick, 100 tol 10 mm. wide, and 150 to 155 mm. Fig. 5. high. They serve as supports for A, as well as ends for a further 32 OPERATIONS. [13. trough, the side walls of which are formed of stout pieces of glass cemented in' grooves in A and BE. The glass plates are each 310 to 320 mm. long, and 55 mm. high. They are placed not quite parallel, the lower edges being separated by a distance of 67 to 70 mm., the upper 85 mm. The trough stands on a board, DD, to which it is fastened by strips of wood, e e. A vertical pillar, F, screwed into D, carries the inclined channel, G, lined with felt; it serves to support the measuring tube, h is a round, slanting cut in _Z?, for the reception of the tube ; i is an incision for the reception of the lower end of the tube ; it prevents the latter from falling into the lower part of the trough. In use the trough is filled to within an inch of the upper edges of the glass plates with mercury, of which 30 [to 35 Ib. will be necessary. In order that the mercury may adhere to the wooden walls, the latter are first- rubbed moist then dry with mercury and mercuric-chloride solution. In order to transfer gases which have been collected in large bottles, a similar but larger trough is employed. Lastly, in order to accurately determine the volume of a gas collected over mercury, it is above all necessary that the tube be first com- pletely filled with mercury, and with the entire exclusion of air bubbles, before introducing the gas. To this end the tube is first washed with water and dried with filter paper by aid of a wooden rod, Fig. 6, the upper end of which is provided with 10 to 20 little spikes. Care must be taken that no filaments of paper be left behind. The filling with mercury is accomplished by means of a funnel, Fig. 7, kept filled with mercury, and having a long stem, with narrow exit, reaching to the bottom of the tube to . be filled. The metal thus introduced from below presents a mirror-like surface on the sides of the tube. If such a funnel as described is not at hand, a glass tube drawn out at the lower end may be fused to a small funnel. Tig. 6. Fig. 7. 14, 15.] MEASURING OF GASES. 33 14. 2. INFLUENCE OF TEMPERATURE. The temperature of gases to be measured is determined either by reducing it to the same temperature as the confining fluid, or, using a closed eudiometer, to that of the water in the cylinder provided for that purpose, and then measuring the temperature of the liquid; or by suspending a sensitive thermometer by the side of the gas, and noting its varia- tion. If the construction of the pneumatic apparatus permits the total immersion of the cylinder in the confining fluid, uniformity of temperature between the latter and the gas which it is intended to measure is most readily and speedily obtained ; but in the reverse case, the operator must always, after every manipulation, allow half an hour, or, in operations combined with much heating, even an entire hour, to elapse, before proceeding to observe the state of the mercury in the cylinder and in the thermometer. Proper care must also be taken, after the temperature of the gas has been duly adjusted, to prevent re-expansion during the reading- off; all injurious influences in this respect must accordingly be care- fully guarded against, and the operator should, more especially, avoid laying hold of the tube with his hand (in pressing it down, for instance, into the confining fluid) ; making use, instead, of a wooden holder. On account of the necessity of bringing the gas and surround- ing air to the same temperature, every sudden change in the latter is prejudicial, hence it is advisable to select for gas analysis a room having a northern exposure, and sheltered so far as possible from the direct influence of solar heat. * 15. 3. INFLUENCE OF PRESSURE. When a gas is confined by a fluid, and the level of the latter is the same inside the tube as outside, then the gas is under the prevailing pressure only, that of the atmos- phere ; and this may be directly found by a barometric reading. On the other hand, when the confining fluid within the tube is higher than that without, then the gas is under less pressure ; if lower, it is under greater pressure than that of the atmosphere. In 34 OPERATIONS. [16. the latter case, the level may be equalized by raising the tube ; in the former, by sinking it, if the trough be deep enough. When operating over water, the level may be usually secured without difficulty; when operating over mercury, however, this is very often impossible, particularly with wide tubes, as in Fig. 8. In this case the gas is under the pressure of the atmosphere minus the pressure of a column of mercury equal in height to the line ab. The pressure is accordingly ascertained by accu- rately measuring the line J, and subtracting its length from the observed height of the barometer. For instance, if the latter is 758 mm., and the line ab measures 100 mm., the actual pressure upon the gas will be 758 100 = 658 mm. mercury. If there is water or some other fluid, e.g., potassa solution, con- fined over the mercury, we as a rule proceed just as if there were no fluid present but mercury, but either bring the mercury inside and outside the tube to the same level, or else measure the differ- ence between the two surfaces of mercury. The pressure of the additional column of water, etc., is usually so insignificant that it may be disregarded. Properly speaking, the fluid ought to be measured, and, according to its specific gravity, reduced to mercu- rial pressure, and this subtracted from the barometric reading. This correction may, however, be omitted as a rule, since, as already stated, an absolutely accurate measurement is impossible under such circumstances. Fig. 8. 16. 4. INFLUENCE OF MOISTURE. When a gas saturated with aqueous vapor is measured, its actual volume is not ascertained, since the aqueous vapor, because of its tension, exerts a pressure on the confining fluid. As, however, the tension of aqueous vapor is 16.] MEASURING OF GASES. 35 known for various temperatures, the necessary corrections may be readily made. This is, however, only then possible when the gas is ftrlly saturated with vapor. Care must be exercised, hence, when measuring gases, that the gas be either fully saturated with aqueous vapor, or else is absolutely dry. The drying of gases confined over mercury is effected by means of a ball of fused calcium chloride fixed on a platinum wire. This may be prepared by inserting the end of a platinum wire, bent in the form of a hook, into a bullet mold of about 6 mm. diameter, and then filling the latter with fused calcium chloride free from caustic lime. After cooling, the neck of calcium chloride adhering to the wire is removed by means of a knife. In order to dry a gas, the ball is pushed, by means of the wire, up through the mercury into the gas, in which it is allowed to remain for about an hour, when it is removed from the now fully dried gas. While the ball is in contact with the gas, care must be taken that the end of the platinum wire is entirely submerged in the mercury in the trough, otherwise there will inevitably occur a diffusion of the confined gas and outside air through the exposed wire. It is more convenient to measure the gas in the moist condition, when this may be done. BUNSEN effects the saturation with moisture by introducing a drop of water the size of a lentil into the empty tube, by means of a glass rod, taking care to deposit it at the top of the tube, and without touching the sides of the tube. This quantity of water is amply sufficient to fully saturate with aqueous vapor at the ordinary temperature the gas subsequently introduced. From the preceding remarks, it will be quite obvious that the volumes of gases can be compared only when reduced to the same temperature, pressure, and degree of moisture. As a rule the reductions are made to 0, 760 mm. barometric pressure, and absolute dryness. The methods by which this is effected, as well as the manner of ascertaining the weights of gases from their vol- umes, will be found under the calculation of analyses. 36 OPERATIONS. [ 17, lt>. IT. 5. THE MEASURING OF FLUIDS. In consequence of the vast development which volumetric analysis has of late undergone, the measuring of fluids has become an operation of very frequent occurrence. According to the dif- ferent objects in view, various kinds of measuring vessels are employed. The number proposed has gradually increased to such an extent that all the forms and arrangements recommended can not be discussed here, but only those will be described that have been found most practical, and that have given the best results in my work. The operator must, in the case of every measuring vessel, care- fully distinguish whether it is graduated for holding or for deliver- ing the exact number of c. c. marked on it. If you have made use of a vessel of the former description in measuring off 100 c. c. of a fluid, and wish to transfer the latter completely to another vessel, you must, after emptying your measuring vessel, rinse it, and add the rinsings to the fluid transferred; whereas, if you have made use of a measuring vessel of the latter description, there must be no rinsing. a. MEASURING VESSELS GRADUATED FOR HOLDING THE EXACT MEAS- URE OF FLUID MARKED ON THEM. aa. Measuring vessels which serve to measure out one definite quantity of fluid. We use for this purpose 18. 1. Measuring Flasks. Fig. 9 represents a measuring flask of the most practical and convenient form. Measuring flasks of various sizes are obtainable, holding re- spectively 200, 250, 500, 1000, 2000, etc., c. c. As a general rule, they have no ground-glass stoppers; it is, however, very desirable, in certain cases, to have measuring flasks with ground stoppers. The flasks must be made of well-annealed glass of uni- 18.] MEASURING OF FLUIDS. 37 form thickness, so that fluids may be heated in them. The line- mark should be placed within the lower third, or at least within the lower half, of the neck. Measuring flasks, before they can properly be employed in analytical operations, must first be carefully tested. The best and simplest method of effecting this is to proceed thus : Put the flask, perfectly dry inside and outside, on the one scale of a sufficiently delicate balance, together with a weight of 1000 grm. in the case of a litre flask, 500 grin, in the case of a half -litre flask, etc., restore the equilibrium by placing the requisite quantity of shot and tinfoil on the other scale, then remove the flask and the weight from the balance, place the flask on a perfectly level surface, and pour in distilled water at IT* 5 until the lower border of the dark zone formed by the top of the water around the inner walls corre- sponds with the line-mark. After having thoroughly dried the neck of the flask above the mark, replace it upon the scale; if this restores the perfect equilibrium of the balance, the water in the flask weighs, in the case of a litre measure, exactly 1000 grm. If the scale bearing the flask sinks, the water in it weighs as much above 1000 grm. as the additional weights amount to which you have to put in the other scale to restore the equilibrium ; if it rises, on the other hand, the water weighs as much less as the weights amount to which you have to put in the scale with the flask to effect the same end. If the water in the litre measure weighs 1000 grm., in the half -litre measure 500 grm., etc., the measuring flasks are correct. Differences up to O'l grm. in the liter measure, up to 0'07 grm. in the half-litre measure, and up to 0'05 grm. in the quarter-litre measure, are not taken into account, as one and the same measuring flask will be found to offer variation to the extent indicated, in repeated consecutive weighings" though filled each time exactly up to the mark with water of the same temperature. Though a flask should, upon examination, turn out not to hold the exact quantity of water which it is stated to contain, it may yet possibly agree with the other measuring vessels, and may accord- 38 OPERATIONS. [ 18. ingly still be perfectly fit for use for most purposes. Two meas- uring vessels agree among themselves if the marked numbers of c. c. bear the same proportion to each other as the weights found; thus, for instance, supposing your litre measure to hold 998 grm. water at 17*5, and your 50 c. c. pipette to deliver 49 -9 grm. water of the same temperature, the two measures agree, since 1000:50 = 998:49-9. To prepare or correct a measuring flask, tare the dry litre, half- litre, or quarter-litre flask, and then weigh into it, by substitution ( 9), 1000 grin., or, as the case may be, the half or quarter of that quantity of distilled water at 17'5. Put the flask on a per- fectly horizontal support, place your eye on an exact level with the surface of the water, and mark the lower border of the dark zone by two little dots made on the glass with a point dipped into thick asphaltum varnish, or some other substance of the kind. Now pour out the water, place the flask in a convenient position, and cut with a diamond a fine distinct line into the glass from one dot to the other. Measuring vessels are sometimes graduated also for delivery. These, however, can only be used where very accurate measuring is unnecessary, because the quantity of water remaining in the flask varies not inconsiderably, and hence in repeated measurings with the same flask notable differences may arise. The graduation or testing of such flasks is effected by filling the flask with water, then emptying it and allowing it to drain for a minute, and then weigh- ing into it the weight of distilled water at 17 '5 corresponding to the number of c. c. In none of these weighings, as will be seen, have the operations been conducted at 4, or reduced to vacuum, hence the measuring vessels will not accurately conform to the standard. If, however, this system is carried out with all the vessels used for measuring fluids, as proposed by FR. MOHR, the measures will correspond per- fectly among themselves, and this will suffice for all the purposes of volumetric analysis. In the exceptional case where such a measuring vessel is used in measuring a gas, the proper correction may be made by multiplying the c. c. by 1-0022, as pointed out by FR. MOHR (Zeitschr. f. Analyt. Chem., vn, 287). 19, 20.] MEASURING OF FLUIDS. 39 55. Measuring vessels which serve to measure out any quan- tities of fluid at will. 19. 2. The Graduated Cylinder. This instrument, represented in Fig. 10, should be from 2 to 3 cm. wide, of a capacity of 100 to 300 c. c., and divided into single c. c. It must be ground at the top, so that it may be covered closely with a ground-glass plate. The measuring with such cylinders is not quite so ac- curate as with, measuring flasks, as in the latter the volume is read off in a narrower part. The accuracy of measuring cylinders may be tested in the same way as in the case of measuring flasks, viz., by weighing into them water at 17*5; or, also, very well, by letting definite quantities of fluid flow into the cylinder from a correct pipette or burette graduated for delivering, and observ- ing whether or not they are correctly indicated by the scale of the cylinder. 100 90 80 J70 60 1 50 /3. MEASURING VESSELS GRADUATED FOR DELIVER- ING THE EXACT MEASURE OF FLUID MARKED ON THEM. aa. Measuring vessels which serve to measure out one definite quantity of fluid. 20. 3. The Graduated Pipette. 40 ( 30 [20 10 Fig. 10. This instrument serves to remove a definite volume of a fluid from one vessel and to transfer it to another ; it must accordingly be of a suitable shape to admit of its being freely inserted into flasks and bottles. We use pipettes of 1, 5, 10, 20, 50, 100, 150, and 200 c. c. capacity. The proper shape for pipettes up to 20 c. c. capacity 40 OPERATIONS. [20. is represented in Fig. 11 ; Fig. 12 shows the most practical form for larger ones. To fill a pipette suction is applied to the upper aperture, either directly with the lips or through a caoutchouc tube, until the fluid stands above the mark ; the upper orifice which is somewhat narrowed and ground) is then closed with the first finger of the right hand (the point of which should be a little moist) ; the outside is then wiped dry, if required, and, the pipette being held in a perfectly vertical direction, the fluid is allowed to drop out, by lifting the finger a little, till it has fallen to the required level ; the loose drop is care- fully wiped off, and the contents of the tube are then finally transferred to the other vessel. In this pro- cess it is found that the fluid does not run out completely, but that a small portion of it remains adhering to the glass in the point of the pipette ; after a time, as this becomes increased by other minute particles of fluid trickling down from the upper part of the tube, there gathers at the lower orifice a drop which may be allowed to fall off from its own weight, or may be made to drop off by a slight shake. If, after this, the point of the pipette be laid against a moist portion of the inner side of the- vessel, another minute portion of fluid will trickle out, and, lastly, another trifling droplet or so may be got out by blowing into the pipette. Now, supposing the operator follows no fixed rule in this- respect, letting the fluid, for instance, in one operation simply run out, whilst in another operation he lets it drain afterwards, and in a third blows out the last particles of it from the pipette ; it is evident that the respective quantities of fluid delivered in the several operations cannot be quite equal. I prefer in all cases the second method, viz., to lay the point of the pipette,, whilst draining, finally against a moist portion of the side of the vessel, which I have always found to give the most accurately cor- responding measurements. The correctness of a pipette is tested by filling it up to the mark with distilled water at 17' 5, letting the water run out, in Fig. 11. Fig. 12. 20.] MEASURING OF FLUIDS. 41 the manner just stated, into a tared vessel, and weighing; the pipette may be pronounced correct if 100 c. c. of water at 17 '5 weigh 100 grm. Testing in like manner the accuracy of the measurements made with a simple hand pipette, we find that one and the same pipette will, in repeated consecutive weighings of the contents, though filled and emptied each time with the minutest care, show differ- ences up to O'Ol grin, for 10 c. c. capacity, up to 0*04 grm. for 50 c. c. capacity. The accuracy of the measurements made with a pipette may be heightened by giving the instrument the form and construction shown in Fig. 13 and fixing it in a holder. It will be seen from the drawing that these pipettes are emptied only to a certain mark in the lower tube, and that they are provided with a pinch-cock, a contrivance which we shall have occasion to describe in detail when on the subject of burettes. This contrivance reduces the dif- ferences of measurements with one and the same 50 c. c. pipette to 0'005 grin. Pipettes are used more especially in cases where it is in- tended to estimate different constituents in separate por- tions of a substance : for instance, 10 grm. of the sub- stance under examination are dissolved in a 250 c. c. flask, the solution is diluted up to the mark, shaken, and 2, 3, or 4 several portions are then taken out with -a 50 c. c. pipette. Each portion consists of I part of the whole, and accordingly contains 2 grin, of the substance. Of course the pipette and the flask must be in perfect harmony. Whether they are may be ascertained by, for instance, emptying the 50 c. c. pipette 5 times into the 250 c. c. flask, and observing if the lower edge of the dark zone of fluid coincides with the mark. If it does not, you may make a fresh mark, which, no matter whether it is really correct or not, will bring the two instruments in question into conformity with each other. Cylindrical pipettes, graduated throughout their entire length, may be used also to measure out any given quanti- ties of liquid; however, these instruments can properly be employed only in processes where minute accuracy is not indispensable, as the limits of error in reading off the divisions in. V OPERATIONS. C the wider part of the tube are not inconsiderable. For smaller quantities of liquid this inaccuracy may be avoided by making the pipettes of tubes of uniform width, having a small diameter only. and narrowed at both ends. (Fu. MOHK'S measuring pipettes.) "When a fluid runs out of a pipette, drops sometimes remain here and there adhering to the tube ; this arises from a film of fat on the inside ; it may be removed by filling the instrument with a concentrated solution of potassium bichro- mate mixed with sul- phuric acid, or with potassa solution, and, after allowing sufficient time for action, thor- oughly washing. bb. Me asur -i n g vessels which serve to measure out quantities of fluid at will. 4. The Burette. Of the various forms and dispositions of this instrument, the following appear to me the most convenient : * 21. I. Molir^s Burette. This excellent measuring apparatus is represented in Fig. 14. It consists of a cylin- drical tube, narrower towards the lower end for about an inch, *In regard to other forms see F. Moira, Lehrbuch der Titrirmethode, 3d edit,, 2 ; G. C. WITTSTEIN, Vierteljaliresschr. f. prakt. Pharm., xvi, 567, and Zeitschr. /. analyt. Chem., vn, 84; A. GAWALOWSKY, Zeitschr. /. Chem. [N. F.], vi, 129, and Zeitschr. f. analyt. Chem., ix, 369 ; GONDOLO, Rev. hebdomad. deChim., Nov 1869, and Zeitschr. f. analyt. Chem., ix, 370. Fig. 14. 21.] MEASURING OF FLUIDS. " 43 with a slight widening, however, at the extreme point, in order that the caoutchouc connector may take a firm hold. I only use burettes of two sizes, viz.., of 30 c. c. , divided into O'l c. c. ; and of 50 c. c., divided into 0*5 c. c. The former are employed principally in scientific, the latter chiefly in technical, investiga- tions. The usual length of my 30 c. c. burette is about 50 cm. ; the graduated portion occupies about 49 cm. The diameter of the tube is accordingly about 10 mm. in the clear; the upper orifice is, for the convenience of filling, widened in form of a funnel, measuring 20 mm. in diameter; the width of the lower orifice is 5 mm. For very delicate processes, the length of the graduated portion may be extended to 50 or 52 cm., leaving thus intervals of nearly 2 mm. between the small divisional lines. In my 50 c. c. burettes the graduated portion of the tube is generally 40 cm. long. To make the instrument ready for use, the narrowed lower end of the tube is warmed a little, and greased with tallow; a caout- chouc tube, about 30 mm. long, and having a diameter of 3 mm. in the clear, is then drawn over it ; into the other end of this is inserted a tube of pretty thick glass, about 40 mm. long, and drawn out to a tolerably fine point; it is advisable to slightly widen the upper end of this tube also, and to covet- it with a thin coat of tallow ; and also to tie linen thread, or twine, round both ends of the con- nector, to insure perfect tight- ness. The space between the lower orifice of the burette and the upper orifice of the small delivery tube should be about 15 mm. The India-rubber tube is now pressed together between the ends of the tubes by the pinch- cock (or clip). This latter instrument is usually made of brass wire; the form represented in Fig. 15 was devised by MOHR. A good clip must pinch so tightly that not a particle of fluid can make its way through the connector when compressed by it ; it must be so constructed that the analyst may work it with perfect facility and exactness, so as to regulate the outflow of the liquid with the most rigorous accuracy, by bringing a greater or less degree of pressure to bear upon it. 44 OPERATIONS. [21. MOHR* lias also devised very practical clips of glass (or horn) and rubber, which are to be highly recommended. Figs. 16 and 17 show the construction of these clips, which are so simple that any one can readily make them by following MOHR'S instructions, as follows: Bend two pieces of flat ther- mometer-tubing, 80 to 90 cm. long, into a very obtuse angle, and place between them, in the middle, a thin piece of cork, about 1 to 2 mm. thick ; pass a rubber ring, Fig. 16. Fig. 17. cut from a somewhat wide rubber tube, over the part inclosing the cork. After placing the rubber tube of the burette between the two glass tubes, press these together and slip another rubber ring over the ends of the glass tubes. These rings serve to tightly com- press and close the rubber tube OH the burette. On pressing together the divergent ends of the glass tubes, however, the pressure on the rubber tube is relieved, and the liquid flows through the delivery tube. On releasing the pressure, the elastic bands again completely close the connecting tube. For supporting MOHR'S burettes, use is made of the holder shown in Fig. 14; this instrument, whilst securely confining the tube, permits its being moved up and down with perfect freedom, and also its being taken out, without interfering with the pinch- cock. The position of the burette must be strictly perpendicular, to insure which, care must be taken to have the grooves of the cork lining, which are intended to receive the tube, perfectly vertical, with the lower board of the stand in a horizontal position. The arm bearing the burette I now have made movable around the upright, so that first one, then another burette may readily be * MOHR, Lehrbuch der Titrirmethode, 3d edit., p. 7. 21.] MEASURING OF FLUIDS. 45 used. If it is desired to fix the arm, a screw (wanting in the illus- tration) serves the purpose. A similar holder, with a brass clamp, is shown in Fig. 18. To best charge the burette for a volumetric operation, the point of the instrument is immersed in the liquid, the pinch-cock opened, and a little liquid, suf- ficient at least to reach into the burette tube, drawn up by ap- plying suction to the upper end ; the cock is then closed, and the liquid poured into the burette until it reaches up to a little above the top mark. The burette having, if required, been duly adjusted in the proper vertical position, the liquid is allowed to drop out to the exact level of the top mark. The instrument is now ready for use. When as much liquid has flowed out as is required to attain the desired object, the analyst, before pro- ceeding to read off the volume used, should wait a few minutes, to give the particles of fluid ad- hering to the sides of the emp- tied portion of the tube proper time to run down. This is an indispensable part of the opera- tion in accurate measurements, since, if neglected, an experiment in which the standard liquid in the burette is added slowly to the fluid under examination (in which, accordingly, the minute particles of fluid adhering to the glass have proper time afforded them during the operation itself to run down), will, of course, give slightly dif- ferent results from those arrived at in another experiment, where the larger portion of the standard fluid is applied rapidly, and the last few drops alone are added slowly. The manner in which the rcading-off is effected, is a matter of Fig. 18. 46 OPERATIONS. [ great importance in volumetric analysis; the first requisite is to bring the eye on a level with the top of the fluid. "We must consequently settle the question What is to be considered the top? If you hold a burette, partly filled with water, between the eye and a strongly illumined wall, the surface of the fluid presents the appearance shown in Fig. 19 ; if you hold close behind the tube a sheet of white paper, with a strong light falling on it, the surface of the fluid presents the appearance shown in Fig. 20. Fig. 19. Fig. 20. Fig. 21, In the one as well as in the other case, you have to read off at the lower border of the dark zone, this being the most distinctly marked line. FR. MOHR recommends the following device for reading off : Paste on a sheet of white paper a broad strip of black paper, and, when reading-off, hold this close behind the burette, in a position to place the border line between white and black from 2 to 3 mm. below the lower border of the dark zone, as shown in Fig. 21 ; then read-off at the lower border of the dark zone. Great care must be taken to hold the paper invariably in the same position, since, if it be held lower down, the lower border of the black zone will move higher up. I prefer to read-off in a light which causes the appearance rep- resented in Fig. 19. By the use of ERDMANN'S float* all uncertainties in reading-off * Journ. f. prakt. Chem., LXXI, 194. 21.] MEASURING OF FLUIDS. 47 may be avoided. Fig. 22 represents a burette thus provided. In tins case we always read-off the mark on the burette which coin- cides with the circle in the middle of the float. The float must be so fitted to the width of the burette that when placed in the filled burette, it will, on allowing the fluid to run out gradually, sink down without wavering ; and when it has been pressed down into the fluid of the closed burette, it will slowly rise again. The weight of the float must, if necessary, be so regulated by mercury that when placed in the filled tube, the surface of the liquid will coin- cide uniformly all around with the upper shoulder of the float. A further important condition of the float is that its axis should coincide as nearly as possible with that of the burette tube, so that the division-mark on the burette may be always parallel with the circular line on the float. The correctness of the graduation of a burette is tested in the most simple way, as follows : Fill tlie instrument up to the highest division with water at 17 '5, then let 10 c. c. of the liquid flow out into an accurately weighed flask, and weigh; then let another quantity of 10 c. c. flow out, and weigh again, and repeat the operation until the contents of the burette are exhausted. If the instrument is correctly graduated, every 10 c. c. of water at IT '5 must weigh 10 grin. Differences up to O'Ol grm. may be disregarded, since even with the greatest care bestowed on the process of reading-off, deviations to that extent will occur in repeated measurements of the upper- most 10 c. c. of one and the same burette. With the float-burettes the weighings agree much more accurately, and the differences for 10 c. c. do not exceed 0*002 grm. MOHR'S burette is unquestionably the best and most convenient instrument of the kind, and ought to be employed in the measurement of all liquids which are not injuriously affected by contact with caoutchouc. Of the standard solutions used at present in volumetric analysis, that of potassium Fig. 22. permanganate alone cannot bear contact with caoutchouc. Excel- 48 OPERATIONS. [22. lent directions for calibrating MOHR burettes have been given by SCHEIBLER. * 22. II. Gay-Lussatf 8 Burette. Fig. 23 represents this instrument in, as I believe, its most practical form. I make use of two sizes, one of 50 c'. c. graduated in 0'5 c. c., the other of 30 c. c. graduated in O'l c. c. The former is about 33 cm. long; the graduated portion occupies about 25 cm. ; the internal diameter of the wide tube meas- ures 15 mm. ; that of the narrow tube 4 mm., which in the upper bent end gradually decreases to 2 mm. The graduated portion of the smaller burette is about 28 cm. long, and'has accordingly an internal diameter of about 11 mm. In use this burette is held in the left hand, with the lower end resting lightly against the chest. Held in this manner, and with an occasional turn of the spout sideways, the outflow of liquid may be regulated at will. The fluid is not allowed, as a rule, to run back into the narrow tube during the course of an opera- tion, as it is frequently difficult to renew the flow of the fluid because of the formation of an air-bubble be- tween the fluid and the drop remaining in the mouth of the spout. To provide a substantial stand for the burette, a solid disk of wood 10 to 12 cm. in diameter and from Fig. 23. 5 to 6 cm. thick has a suitable hole bored in its centre, in which the burette may be inserted. This seems to me more convenient than to cement the burette in a wooden foot. To overcome the difficulty of renewing the flow of liquid when an air-bubble has become enclosed between the fluid and the drop remaining in the tip of the spout, MOHR has proposed closing the wider tube with a perforated cork bearing a short glass tube bent at right angles. On slipping a piece of rubber tubing over this 40 50 *Jour.f. prakt. Chemie, LXXVI, 177. 23.] MEASURING OF FLUIDS. 49 sliort tube, and blowing into it more or less strongly, the outflow of liquid may be regulated at will. Instead of blowing with the" mouth, a rubber bulb may be used, but the latter must be provided with a small hole through which air may enter after the bulb has been compressed, and which is closed by the finger during compres- sion (HERYE-MANGON).* The readings with this burette are taken just as with the MOHR burette. The instrument is preferably held firmly against a per- pendicular wall, a strongly illuminated white door, or a window pane, to insure its being held perfectly vertical. When operating with concentrated, and hence opaque, solutions of potassium per- manganate, the method of reading requires modification, the upper border of the liquid being then observed, and the readings best taken by reflected light against a white background. The GAY-LUSSAG burette is tested as to its correctness just as is the MOHR burette. 23. III. Geissler's Burette. In this instrument, figured in Fig. 24, the narrow tube, which is outside in the GAY-LUSSAC burette, is placed within the wide tube. The glass of that part of the inner tube which projects from the wide tube is very thick, while the part within the wide tube is very thin. This burette is very convenient to use, and is but little liable to fracture. I am very partial to it. What has been stated above regarding reading-oil and testing applies to this burette as well. * Sep. chim. appliquee, I, 68. 50 OPERATIONS. [24 U PRELIMINARY OPERATIONS. PREPARATION OF SUBSTANCES FOB THE PROCESSES OF QUANTITATIVE ANALYSIS. 24. 1. THE SELECTION OF THE SAMPLE. Before the analyst proceeds to make the quantitative analysis of a body, he cannot too carefully consider whether the desired result is fully attained if he simply knows the respective quantity of every individual constituent of that body. This pri- mary point is but too frequently disregarded, and thus false impressions are made, even by the most careful analysis. This remark applies both to scientific and to technical investigations. Therefore, if the constitution of a mineral is to be determined, take the greatest pos- sible care to remove in the first place every particle of gangue, and disseminated imp-un- ties ; remove any adherent matter by wiping or washing, then wrap the substance up in a sheet of thick paper, crush it to pieces on a steel anvil, and pick out with a pair of small pincers the cleanest pieces. Crystalline substances, prepared artificially, ought to be purified by re- crystallization ; precipitates by thorough wash- ing, &c., &c. In technical investigations, when called * upon, for instance, to determine the amount of peroxide present in a manganese ore, or the amount of iron present in an iron ore, the first point for consideration ought to be whether the samples selected correspond as much as possible to the average quality of the ore. What would it serve, indeed, to the purchaser of a manganese mine to know the amount of peroxide present in a select, possibly particularly rich, sample ? These few observations will suffice to show that no universally applicable and valid rules to guide the analyst in the selection of the sample can be laid down ; he must in every individual case, Fig. 24. 25.] MECHANICAL DIVISION. 51 on the one hand, examine the substance carefully, and more par- ticularly also under the microscope, or through a lens ; and, on the other hand, keep clearly in view the object of the investigation, and then take his measures accordingly. 25. 2. MECHANICAL DIVISION. In order to prepare a substance for analysis, i.e., to render it accessible to the action of solvents or fluxes, it is generally indis- pensable, in the first place, to divide it into minute parts, since this will create numerous points of contact for the solvent, and will counteract, and, so far as practicable, remove the adverse influences of the power of cohesion, thus fulfilling all the condi- tions necessary to effect a complete and speedy solution. The means employed to attain this object vary according to the nature of the different bodies we have to operate upon. In many cases, simple crushing or pounding is sufficient ; in other cases it is necessary to reduce the powder to the very highest degree of fineness, by sifting or by elutriation. The operation of powdering is conducted in mortars. The first and absolutely indispensable condition is, that the material of the mortar be considerably harder than the substance to be pulver- ized, so as to prevent, so far as practicable, the latter from being contaminated with any particles of the former. Thus, for crush- ing salts and other substances possessing no very considerable degree of hardness, porcelain mortars may be used, whilst the trituration of harder substances (of most minerals, for instance,) requires vessels of agate, chalcedony, or flint. In such cases, the larger pieces are first reduced to a coarse powder, best effected by wrapping them up in several sheets of writing-paper, and striking them with a hammer upon a steel or iron plate ; the coarse powder thus obtained is then pulverized, in small portions at a time, in an agate mortar, until reduced to the state of an impalpable powder. If we have but a small portion of a mineral to operate upon, and indeed in all cases where we are desirous of avoiding loss, it is advisable to use a steel mortar (Fig. 25) for the preparatory reduc- tion of the mineral to coarse powder. ab and cd represent the two parts of the mortar ; these may be readily taken asunder. The substance to be crushed (having, if OPERATIONS. [25. Fig. 25. practicable, first been broken into small pieces), is placed in the cylindrical chamber ef $ the steel cylinder, which fits somewhat loosely into the chamber, serves as pestle. The mortar is placed upon a solid support, and perpendicular blows are repeatedly struck upon the pestle with a hammer until the object in view is attained. Minerals which are very difficult to pulverize may be strongly ignited, and then suddenly plunged into cold water, and subsequently again ignited. This process is of course applicable only to minerals which lose no essential con- stituent on ignition, and are perfectly insoluble in water. In the purchase of agate mortars, especial care ought to be taken that they have no palpable cracks or indentations ; very slight cracks, however, that cannot be felt, do not render the mortar useless, although they impair its durability. Minerals insoluble in acids, and which consequently require fusing, must especially be finely divided, otherwise we cannot calcu- late upon complete decomposition. This object may be obtained either by triturating the crushed mineral with water, or by elutri- ation, or by sifting ; the two former processes, however, can be resorted to only in the case of substances which are not attacked by water. It is quite clear that analysts must in future be much more cautious on this point than has hitherto been the case, since we know now that many substances which are usually held to be insoluble in water are, when in a state of minute division, strongly affected by that solvent; thus, for instance, water, acting upon some sorts of finely pulverized glass, is found to rapidly dissolve from 2 to 3 per cent, of glass even in the cold. (PELOUZE.*) Thus, again, finely divided feldspar, granite, trachyte and porphyry give up to water both alkali and silica. (H. LuDwia.f) Levigation (trituration with water). Add a little water to the crushed mineral in the mortar, and triturate the paste until all crepitation ceases, or, which is a more expeditious process, transfer * Compt. Rend., XLIII, 117-123. f Archiv d#r Pharm., xci, 147. 25.] MEASURING OF FLUIDS. 53 tlie paste from the mortar to an agate or flint slab, and triturate it thereon with a nmller. Rinse the paste off, with the washing- bottle, into a smooth porcelain basin of hemispherical form, evaporate the water 011 the water-bath, and mix the residue most carefully with the pestle. (The paste may be dried also in the agate mortar, but at a very gentle heat, since otherwise the mortar might crack.) To perform the process of elutriation, the pasty mass, having first been very finely triturated with water, is washed off into a beaker, and stirred with distilled water; the mixture is then allowed to stand a minute or so, after which the supernatant turbid fluid is poured off into another beaker. The sediment, which contains the coarser parts, is then again subjected to the process of trituration, etc., and the same operation repeated until the whole quantity is elutriated. The turbid fluid is allowed to stand at rest until the minute particles of the substance held in suspension have subsided, which generally takes many hours. The water is then finally decanted, and the powder dried in the beaker. The process of sifting is conducted as follows : A piece of fine, well-washed, and thoroughly dry linen is placed over the mouth of a bottle about 10 cm. high, and pressed down a little into the mouth, so as to form a kind of bag ; a portion of the finely triturated sub- stance is put into the bag, and a piece of soft leather stretched tightly over the top by way of cover. By drumming with the finger on the leather cover, a shaking motion is imparted to the bag, which makes the finer particles of the powder gradually pass through the linen. The portion remaining in the bag is subjected again to trituration in an agate mortar, and, together with a fresh portion of the powder, sifted again; the process is repeated until the entire mass has passed through the bag into the glass. When operating on substances consisting of different com- pounds it would be a grave error indeed to use for analysis the powder resulting from the first process of elutriation or sifting, since this will contain the more readily pulverizable constituents in a greater proportion to the more resisting ones than is the ca>e with the original substance. Great care must, therefore, also be taken to avoid a loss of substance in the process of elutriation or sifting, as this loss is likely to be distributed unequally among the several component parts. It is safer in such cases to effect the subdivision by patiently triturating the dry substance, and to avoid elutriation or sifting. 54 OPERATIONS. [ 26. In cases where it is intended to ascertain the average composi- tion of a heterogeneous substance, of an iron ore for instance, a large average sample is selected, and reduced to a coarse powder ; the latter is thoroughly intermixed, a portion of it powdered more finely, and mixed uniformly, and finally the quantity required for analysis is reduced to the finest powder. The most convenient instrument for the crushing and coarse powdering of large samples of ore, &c., is a steel anvil and hammer. The anvil in my own laboratory consists of a wood pillar, 85 cm. high and 26 cm. in diameter, into which a steel plate, 3 cm. thick and 20 cm. in diameter, is let to the depth of one-half of its thickness. A brass ring, 5 cm. high, fits round the upper projecting part of the steel plate. The hammer, which is well steeled, has a striking surface of 5 cm. diameter. An anvil and hammer of this kind afford, among others, this advantage, that their steel surfaces admit most readily of cleaning. To convert the coarse powder into a finer, a smooth-turned steel mortar of about 130 mm. upper diameter and 74 mm. deep is used the final trituration is conducted in an agate mortar. 26. 3. DEYING. Bodies which it is intended to analyze quantitatively must be, when weighed, in a, definite state in a condition in which they can be always obtained again. Now, the essential constituents of a substance are usually accom- panied by an unessential one, viz., a greater or less quantity of water, enclosed either within its lamellae, or adhering to it from the mode of its preparation, or absorbed by it from the atmosphere. It is perfectly obvious that to estimate correctly the quantity of a substance, we must, in the first place, remove this variable quantity of water. Most solid bodies, therefore, require to ~be dried before they can ~be quantitatively analyzed. The operation of drying is of the very highest importance for the correctness of the results ; indeed it may safely be averred that many of the differences observed in analytical researches proceed entirely from the fact that substances are analyzed in different states of moisture. Many bodies contain, as is well known, water which is proper 26.] DESICCATION. 55 to them either as inherent in their constitution, or as so-called water of crystallization. In contradistinction to this, we will employ the ' term moisture to designate that variable adherent or mechanically enclosed water, with the removal of which the operation of drying in the sense here in view is alone concerned. In the drying of substances for quantitative analysis, our object is to remove all moisture, without interfering in the slightest degree with combined water or any other constituent of the body. To accomplish this object, it is absolutely requisite that we should know the properties which the substance under examination mani- fests in the dry state, and whether it loses water or other constitu- ents at a red heat, or at 100, or in dried air, or even simply in contact with the atmosphere. These data will serve to guide us in the selection of the process of desiccation best suited to each sub- 1 stance. The dried substance should always at once be transferred to a well-closed vessel ; glass tubes, sealed at one end, and of suf- ficiently thick glass to bear the firm insertion of tight-fitting smooth corks weighing-tubes are usually employed for this purpose. It is also advisable to cover the corks with tinfoil. The following classification may accordingly be adopted : a. Substances which lose water even on simple contact with the atmosphere such as sodium sulphate, crystallized sodium carbon- ate, etc. Substances of this kind turn dull and opaque when exposed to the air, and finally crumble wholly or partially to a white powder. They are more difficult to dry than many other bodies. The process best adapted for the purpose, is to press the pulverized salts with some degree of force betw r een thick layers of fine white blotting-paper, repeating the operation with fresh paper until the last sheets remain absolutely dry. It is generally advisable in the course of this operation to repow- der the salt. 1. Substances which do not yield water to the atmosphere (unless it is perfectly dry), but effloresce in artificially dried air y such as magnesium sulphate, sodium-potassium tartrate (Rochelle salt), &c. Salts of this kind are reduced to powder, which, if it be very moist, is pressed between sheets of blotting-paper, as in a ; after this operation, it must be allowed to remain for some time spread in a thin layer upon a sheet of blotting-paper, effectually protected against dust, and shielded from the direct rays of the sun. 56 OPERATIONS. [27, 27. c. Substances which undergo no alteration in dried air, but lose water at 100 ; calcium tartrate, for instance. These are finely pulverized ; the powder is put in a thin layer into a watch-glass or shallow dish, and the latter placed inside a chamber in which the air is kept dry by means of concentrated sulphuric acid or calcium chloride. This process is usually conducted in one of the follow- ing apparatus, which are ternied desiccators, and which subserve still another purpose besides that of drying, viz., that of allowing hot crucibles, dishes, etc., to cool in dry air. Fig. 26. Fig. 27. In Fig. 26, a represents a glass plate (ground-glass plates answer the purpose best), #, a bell jar with ground rim, which is greased with tallow; c is a glass basin with concentrated sulphuric acid; d, a round iron plate, sup- ported on three feet, with circular holes of various sizes, for the recep- tion of the watch-glasses, crucibles, etc., containing the substance. In Fig. 27, a represents a beaker with ground and greased rim, and filled to one-fourth or one-third with concentrated sulphuric acid ; 1) is a ground-glass plate; c is a bent wire of lead, which serves to support the watch-glass contain- ing the substance. Fig. 28 is a similarly constructed calcium-chloride desiccator. Fig. 28. .27.] DESICCATION. 57 Fig. 29 represents a readily portable desiccator, used more par- ticularly to receive crucibles in course of cooling, and carry them to the bal- ance. The apparatus consists of a strong glass jar ; the lid must be ground to shut air-tight ; the place on which it joins is greased with tallow. The outer diameter of my jars is 105 mm. ; the sides are 6 mm. thick. The aperture has a diameter of 80 mm. ; the box up to the small part is 65 mm. high; the lid has the same height ; the small part itself is .15 mm. high, and ground to a slightly conical shape. A brass ring, with rim, fits exactly into the aperture ; the rim must not project be- yond the glass. The ring bears a triangle of iron or, better, platinum wire, intended for the reception of crucibles, &c. The vessel is one- third filled with calcium chloride. Fig. 30 represents an exsiccator devised by A. SCHKOTTER; it affords free egress to the air, which expands when a hot crucible is placed within the exsiccator and passes through the small tube, &, escaping through two small holes placed at the base of &, whence it rises through sulphuric acid contained in c, and finally escapes through the bulb d filled with calcium chlo- ride. When the appa- Fig. 30. ratus is cooling, per- fectly dry air re-enters by the same way. The operation may be .58 OPERATIONS. [ %&. considered at an end when no more air-bubbles pass through the sulphuric acid. The small tube, e, serves to catch any sulphuric acid that might be carried down through a ; it must not close air- tight, the lower orifice of the apparatus serving as a stopper for the bell- jar, hence the cork carrying it must be channelled. f serves as a stand for the bell-jar. This desiccator possesses the advantage that the substances placed in it are cooled in dry air at the ordinary atmospheric pressure, and hence, when ^amoved from the apparatus, have no tendency to absorb air and, with this, moisture, which can not be said of substances cooled in air slightly rarefied by heat. The substance to be dried is exposed to the action of the dry air, until it ceases to lose weight. Substances which are acted on by atmospheric oxygen are in a similar manner dried under the bell- jar of an air-pump. Substances which, although they lose no water, yet lose ammonia, in dry air, are dried over caustic lime mixed with a little powdered ammonium chloride, i.e., in an anhydrous ammoniacal atmosphere. 28. d. Substances which 'at 100 completely lose their moisture, without suffering any other alteration, such as hydrogen potassium tartrate, sugar, etc. These are dried in the water-bath ; in the case of slow-drying substances, or where it is wished to expedite the operation, with the aid of a current of dry air. Fig. 31 represents the water-bath most commonly used. It is made of sheet copper. The engraving renders a detailed description unnecessary. The inner chamber, with a double-limbed tube, the outer longer limb of which dips into a cylinder filled with water ; a is in that case closed with a perforated cork bearing a sufficiently tall funnel tube, which fits air-tight in the cork. The lower end of this tube reaches down to one inch from the bottom. In large analytical laboratories water is usually kept boiling all day long, for the production of distilled water. The boilers used in my own laboratory have the shape of somewhat oblong square boxes, about 120 cm. long, 60 cm. broad, and 24 cm. high ; the front of the boiler has soldered into it, one above the other, two rows of drying chambers, of the kind shown in Fig. 31. This gives so many ovens that almost every student may have one for his special use. Most of these ovens are from 11 to 12 cm. deep and broad, and 8 cm. high ; some of them, however, are 16 cm. deep and broad, to enable them to receive large-sized dishes. The substances to be dried are usually put on double watch-glasses, laid one within the other, which are placed in the oven, and the door is then closed. In the subsequent process of weighing, the upper glass, which contains the substance, is covered with the lower one. The glasses must be quite cold before they are placed on the scale. In cases where we have to deal with hygroscopic substances, the reabsorption of water upon cooling is prevented by the selection of close-fitting glasses, which are held Fig. 32. tight together by a clasp (Fig. 32), and allowed to cool with their contents under a bell- glass over sulphuric acid (see Fig. 26). These latter instructions apply equally to the process of drying conducted in other appa- ratus. The clasp used for pressing the watch-glasses together and 60 OPERATIONS. [ 28. which in all cases where it is intended to ascertain the loss of weight which a substance suffers on desiccation, is to be looked upon as belonging to the glasses, and must accordingly be weighed with them is constructed of two strips of thin brass plate, 'about 10 cm. long and 1 cin. wide, which are laid the one over the other, and soldered together at the ends, to the extent of 5 to 6 mm. ; or, they may be made of one piece, as our illustration shows. The following apparatus serves for drying substances in a cur- rent of air : In Fig. 33 the air-current is caused by simply warming the air, hence the apparatus is very convenient to use. a b is a copper or Fig. 33. tinned-iron box into which the canal c d is soldered, and communi- cates with the chimney ef. The latter is surrounded on three sides by the case g 7^, which also communicates with a 1>, but has no- opening at the top ; i is a round hole leading into - the canal, and which may be closed with a cork ; I k is provided with a well-fitting sliding door running in grooves. In use, the aperture n, which serves as an outlet for the water, is closed with a cork, when the outer case is half filled, through the opening m, with water, which is then heated to boiling. The watch-glasses containing the substances to be dried are then placed on the holes in the sliding shelf B (Fig. 33), which is then intro- duced into the canal at I &, and the latter closed. The steam sur- rounding the chimney soon causes an upward current of the warmed air within it, and this causes cold air to be drawn in at the opening ^, and to pass over the substance to be dried, carrying -with it the evaporating moisture. 28.] DESICCATION. 61 The disadvantage that the substances are always kept at a point below 100 C. by the current of cold air, is easily remedied by sol- dering a tube under the bottom of the canal along its entire length and back again, and conducting the air through it into the canal. The air is thus heated to 100 C. before it comes into contact with the substances. This tube is not shown in the illustration, in order to avoid confusion. It is very practical, also, to omit the opening m, and instead to cut in the top of the case round holes of different sizes (provided with suitable covers) for receiving small evaporating dishes. According to requirements, .the apparatus may be made 20 to 30 cm. long, 15 cm. wide, and about 10 cm. high. The chimney should be 6 cm. wide arid 3 cm. high. Should a stronger current of air be desired than that afforded by the small chimney, a current of air previously passed through sulphuric acid or over calcium chloride may be blown through the opening i by means of a gasometer, rubber bulb, or other suitable contrivance. Or, air dried by passing through sulphuric acid may be drawn through the apparatus by means of an air-pump ( 47), or an aspirator (d in Fig. 34) connected with the small chimney by a tube-bearing cork inserted into a short tube with which the chimney is in this case provided. If a higher temperature than that of boiling water is desired, the (copper) apparatus is filled with oil, and the tempera- ture taken with a thermometer inserted in a perforated cork fixed in the opening m. In Fig. 34 the air-current is produced by a stream of water. Fig. 34. Fig. 35. a represents a flask one-third filled with concentrated sulphuric acid; o is a glass vessel (commonly called a LIEBIG'S drying-tube), 62 OPERATIONS. [ 29. and d a tin vessel (the aspirator) provided with a stop-cock at e, and arranged in other respects as the cut shows. Fig. 35 repre- sents a small tin vessel, containing water and covered with a lid ; two apertures are cut into the border of the latter, to receive the ascending limbs of c. The tube c is first weighed with the substance, then placed in the water-bath (Fig. 35), which is placed over an alcohol- or gas- lamp; the aspirator d is then filled with water, and c connected with the flask a by the perforated cork g, and with d by means of a caoutchouc tube f. If the stop-cock e be now opened so as to allow the water to drop from d, the air will pass through the tube , and after being dehydrated by the sulphuric acid, will pass over the heated substance in c. After the operation has been continued for some time, it is interrupted for the purpose of weighing the tube c and its contents, and then resumed again, and continued until the weight of c (and its contents) remains stationary. The current of cold air exercising its constant cooling action upon the substance, the latter never really reaches 100. It is, therefore, sometimes advisable to substitute for the water in the bath a satu- rated solution of common salt. "With this substitution, the apparatus will be found to effect its purpose most expeditiously. It is not adapted, however, for dry- ing such substances as have a tendency to fuse or agglutinate at 100. It is, moreover, less adapted for determining the moisture in substances than for simply drying them, because the glass is some- what attacked by the prolonged action of the boiling water, hence causing a slight loss in weight of the drying-tube in the course of the operation. This loss, too, varies with different kinds of glass. 29. e. Substances which persistently retain moisture at 100, or become completely dry only after a very long time, but which are decomposed by a red heat. The desiccation of such substances is affected by means of air-, oil-, paraffin-, or mercury-baths, or on drying-disks (Fig. 42), at a temperature of 100-120 or even still higher; and, according to circumstances, with or without the aid of a current of air, some- times in a partial vacuum, and sometimes in diluted carbon dioxide. 29.] DESICCATION. Figs. 36 and 37 represent two air-baths of simple construction. The latter is adapted for the desiccation of a single substance ; the former is suited for the simultaneous drying of several substances. In Fig. 36, a 5 % is a case of stout sheet copper soldered with brass, 15 to 20 cm. wide and deep, and of suitable height. In the Fig. 36. Fig. 37. aperture c there is fixed a cork carrying a thermometer, d, which extends into the interior of the case, e is a wire stand on which the watch-glasses containing the substances to be dried are placed. The case is heated by means of a gas-, alcohol-, or oil-lamp. When the temperature has reached the point desired, it is easily main- tained at this point by regulating the flame.* In order to limit as much as possible the cooling from without, it is advisable to cover the apparatus with a pasteboard hood having a movable front. In Fig 37, A is a box of strong sheet copper, about 11 cm. * When using gas, BUNSEN'S improved KEMP regulator (made by DESAGA, of Heidelberg) may be advantageously employed in order to obtain constant tem- peratures. A modification has been recommended by TH. SCHORER (Zeitschr.f. analyt. Chem., ix, 213). SCHEIBLER'S regulator (ibid., vn, 88) is more certain in action, even also under sudden changes in gas pressure, but, as its action depends* on an electromagnet, its construction is more complicated. 64 OPEKATIONS. [29. high and 9 cm. in diameter. Tlie box is closed with the loosely- fitting cover B, which is provided with a narrow rim, and has two apertures, and E ; C is intended to receive the thermometer Z>, which is fitted into it by a perforated cork ; .# affords an exit to the aqueous vapors, and is, according to circumstances, either left open or loosely closed. Within the box, about half-way up, are fixed three pins, for the support of a triangle of moderately stout wire, upon which the crucible with the substance is placed. The ther- mometer bulb should be as near the crucible as possible, but must not touch the triangle. Heating is effected by means of a gas- or alcohol-lamp. When the appa- ratus has cooled to the extent that it may be conveniently grasped, the cover is removed, and the still warm crucible taken out, covered, and allowed to become cold in an exsiccator, when it is weighed. The air-bath shown in Fig. 38 serves for drying substances in a bulb-tube with the simultaneous employment of a current of dry air. The apparatus consists of a sheet-iron box having the follow- ing dimensions in cm. : a ~b = 20 ; ac = 13; ad = 12; ef= 11; e g 6. The diameter of the aperture on each side is 16 mm. The thermometer is thrust so far down until its bulb is on a level with and touches the side of the bulb-tube. To this end the opening h is not placed exactly in the middle line, but 1 cm. behind it. In this apparatus a temperature of 200 to 260 may be easily attained. To produce the current of dry air, one of the projecting ends of the bulb-tube is connected 29.] DESICCATION. with a hydraulic air-pump ( 47) or an aspirator, as in Fig. 34 ; the other end is connected with a calcium -chloride tube or a flask containing concentrated sulphuric acid (Fig. 34, a) ; the current should be somewhat rapid at first, slower afterwards. If the tube with the dried substance is to be weighed, it must be allowed to cool, with a current of dry air still passing through it. In the air-bath shown in Fig. 39, the drying is promoted by alternate exhaustion and readmission of air. a is a vessel of stout Fig. 39. sheet copper, provided with two apertures, and soldered with brass ; b is a glass tube in which the substance is dried ; c is a thermom- eter; d is a calcium-chloride tube; e is a small hand air-pump, which may, of course, be replaced by a hydraulic or mercurial air-pump. In use a is heated to the desired degree ; then 5 and d are ex- hausted. After a few minutes the stop- cock f is opened and air is allowed to re-enter, first becoming perfectly dry by passing over the calcium chloride. The exhaustion and readmission of air are then repeated until not the slightest trace of moisture is visible in the tube g when the latter is cooled by surrounding it with cotton saturated with ether. 66 OPERATIONS. [30. 30. As an oil-bath, the copper drying closet figured in Fig. 31 is employed as a rule, being then filled two-thirds with refined rape- oil. The temperature is ascertained by means of a thermometer borne by a perforated cork inserted in the aperture a. The ther- mometer bulb must reach nearly to the bottom, or must at least be entirely immersed in the oil. As the oil emits a disagreeable and most annoying odor when heated, it is preferable to use paraffin instead. The air-bath shown in Fig. 39 may also serve as an oil- bath. If it is intended to weigh the substance, after drying, in the tube, a shorter tube should be selected which may be readily inserted into the tube standing in the oil. Some organic substances, when dried at high temperatures, suffer alteration from the action on them of atmospheric oxygen (see FR. ROCHLEDER, Jour, fur prakt. Chemie, LXVI, 20'8). When drying such substances, hence, contact with oxygen must be avoided. Figs. 40 and 41 show apparatus devised by ROCHLEDER for this purpose. The former may also be advantageously employed for drying with an air- current ; in the latter, the drying is effected in a rarefied gas. J?, Fig. 40, is a sheet-copper "* cylinder 18 cm. high and 9 cm. in diameter, containing a suitable quantity of oil or paraffin in which is suspended and suitably fixed an iron or glass vessel, A, containing mercury. In the mercury there dip a thermometer, and the glass tube, (7, containing the substance to be dried. The dried gas (hydrogen, car- bonic acid, air, etc.) enters at 5, and escapes at a if necessary, through a weighed cal- cium-chloride tube. To prevent any possible air-currents acting on the substance, the end of & is bent upwards. The advantage of hav- ing mercury in a is that, on removing the tube (7, it is perfectly clean. In Fig. 41 the cock If is screwed on to the air-pump at a ; b is connected by means of rubber tubing with a rubber bag or bladder filled with carbon dioxide. B is an oil-bath, the tern- Fig. 40. 31.] DESICCATION. 67 perature of which is ascertained by a thermometer. In the oil- bath is suspended a wide-mouthed, stout glass vessel, $, in which Fig. 41. is placed the substance to be dried, and contained in a glass tube, as wide as practicable, and sealed at one end. On pumping, while the cock II is open and the cock H r is closed, the air in S is rare- fied; on now closing JTand opening II' , the vessel becomes filled with carbon dioxide, previously dried by passing through the calcium-chloride tube C'. By repeating this procedure the appa- ratus is entirely filled with dried carbon dioxide. H' is then closed, and the pump operated. The oil-bath is then heated to the desired temperature, carbon dioxide being admitted from time to time by opening II' . On closing H' and pumping again, the moisture taken up by the carbon dioxide is removed with the lat- ter, and is retained in the calcium- chloride tube C. Within half an hour the drying is complete. 31. In technical and agricultural chemical investigations, in which a number of specimens are to be simultaneously dried at a high temperature, the drying-disk illustrated in Fig. 42 and devised by me is recommended. The apparatus consists of a turned cast-iron plate 21 cm. in diameter and 37 mm. thick, supported by a tripod, and weighing about 8 kilos. This weight enables the plate to be uniformly heated, and permits the desired temperature to be readily main- tained. At equal distances from the centre of the plate six smooth, 68 OPERATIONS. cylindrical cavities are turned. Each cavity is fitted with a turned brass pan 55 mm. in diameter and 18 mm. deep, and fitting rather loosely, so as to be readily removable after heating. Every pan is provided with a small handle pointing toward the periphery of the plate, and resting in appropriate grooves made in the latter. Each handle, moreover, bears stamped upon it a number, from 1 to 6, corresponding to a similar num- ber stamped in the plate behind the cavities so that each pan has its own proper cavity. The centre of each pan is 6.5 cm. distant from the centre of the plate ; and the rims of the pans are level with the surface of the plate. Five of the pans are intended for the samples (ores, parts of plants, etc.), and the sixth for the thermom- eter. The sixth cavity bears fitted into it a brass rim extending 3 cm- above the surface of the plate; the pan so heightened is filled with brass or copper filings, and the thermom- eter bulb is embedded in these down to the bottom. Heat is applied under the centre of the plate. f. Substances which suffer no alteration at a red heat, such as barium sulphate, pearlash, etc., are very readily freed from mois- ture. They need simply be heated in a platinum or porcelain crucible over a gas- or spirit-lamp until the desired end is attained. The crucible, having first been allowed to cool a little, is put, still hot, under a desiccator, and finally weighed when cold. 32.] DESICCATION. 69 III. GENERAL PROCEDURE IN QUANTITATIVE ANALYSES. 32. It is important, in the first place, to observe that we embrace in the following general analytical method only the separation and determination of the metals and their combinations with the metalloids, and of the inorganic acids and salts. With respect to the quantitative analysis of other compounds, it is not easy to lay down a universally applicable method, except that their constitu- ents usually require to be converted first into acids or bases, before their separation and estimation can be attempted ; this is the case, for instance, with phosphorus sulphide, sulphur chloride, iodine chloride, nitrogen sulphide, &c. The quantitative analysis of a substance presupposes an accurate knowledge of its properties, and of the nature of its several constituents. These data will enable the operator at once to decide whether the direct estimation of each individual constitu- ent is necessary ; whether he need operate only on one portion of the substance, or whether it would be advantageous to deter- mine each constituent in different portions. - Let us suppose, for instance, we have a mixture of sodium chloride and anhydrous sodium sulphate, and wish to ascertain the proportion in which these two substances are mixed. Here it would be superfluous to determine each constituent directly, since the determination either of the quantity of the chlorine, or of the sulphuric acid, is quite sufficient to answer the purpose; still the estimation of both the chlorine and the sulphur trioxide will afford us an infallible con- trol for the correctness of our analysis ; since the united weights of these two substances, added to the sodium and soda respectively equivalent to them, must be equal to the weight of the substance taken. These estimations may be made, either in one and the same portion of the mixture, by first precipitating the sulphuric acid with barium nitrate, and subsequently the hydrochloric acid from the filtrate with solution of silver nitrate ; or a separate portion of the mixture may be appropriated to each of these two operations. Unless there is some objection to its use (e.g., deficiency or hetero- geneousness of substance), the latter method is more convenient, 70 OPERATIONS. [ 33. and generally yields more accurate results; since, in the former method, the unavoidable washing of the first precipitate swells the amount of liquid so considerably that the analysis is thereby delayed, and, moreover, loss of substance less easily guarded against. Before beginning all analyses, at least those of a more complex nature, the student should write out an exact plan, and accurately note on paper, during the entire process, everything that lie does. It is in the highest degree unwise to rely on the memory in a com- plicated analysis. When students, w T ho imagine they can do so, come, a week or a fortnight after they have begun their analysis, to work out the results, they find generally too late that they have forgotten much, which now appears to them of importance to, know. The intelligent pursuit of chemical analysis consists in the projecting and accurate testing of the plan ; acuteness and the power of passing in review all the influencing chemical relations must here support each other. He who works without a thor- oughly thought-out plan, has no right to say he is practising chem- istry ; for a mere unthinking stringing together of a series of filtra- tions, evaporations, ignitions, and weighings, howsoever well these several operations may be performed, is not chemistry. We will now proceed to describe the various operations consti- tuting the process of quantitative analysis. 33. 1. WEIGHING THE SUBSTANCE. The amount of matter required for the quantitative analysis of a substance depends upon the nature of its constituents ; it is, there- fore, impossible to lay down rules for guidance on this point. Half a gramme of sodium chloride, and even less, is sufficient to effect the estimation of the chlorine. For the quantitative analy- sis of a mixture of common salt and anhydrous sodium sulphate, 1 gramme will suffice ; whereas, in the case of ashes of plants, com- plex minerals, &c., 3 or 4 grammes, and even more, are required. 1 to 3 grm. can therefore be indicated as the average quantity suitable in most cases. For the estimation of constituents present in very minute proportions only, as, for instance, sodium and potassium in limestones, phosphorus or sulphur in cast-iron, &c., much greater quantities are often required 10, 20, or 50 grammes. 33.] WEIGHING THE SUBSTANCE. 71 The greater the amount of substance taken the more accurate will be the analysis ; the smaller the quantity, the sooner, as a rule, will the analysis be finished. We would advise the student to endeavor to combine accuracy with economy of time. The less substance lie takes to operate upon, the more carefully he ought to weigh ; the larger the amount of substance, the less harm can result from slight inaccuracies in weighing. Somewhat large quantities of substance are generally weighed to 1 milligramme ; minute quantities, to Ol milligramme. If one portion of a substance is to be weighed off, w^e first weigh tw T o watch-glasses which fit on each other, or else an empty platinum crucible with lid, then we put some substance in, and weigh again ; the difference between the two weighings gives the weight of the substance taken. This mode of weighing off, however, is advisable only when the substance is to be further treated in the watch-glasses or crucible, or when the substance is not of an adherent nature, or when the adherent particles may be washed away with water. If the substance is to be transferred to a flask or beaker, and treated with a concentrated solvent, the weighing is best done in a small tube sealed at one end. In this case the approximate weight of the tube should be known. After receiving the substance, the tube and its contents are carefully weighed ; then nearly the whole, or a suitable quantity, of the substance is shaken out into the flask or beaker, the weight again taken, and the difference in weight, showing the quantity of sub- stance taken, noted. When the substances handled are hygroscopic, the tube must be closed ; this is easily ac- ^ 1 ^* ***' complished by inserting the tube into one slightly larger, as shown in Fig. 43. If several quantities of a substance are to be operated upon, the best way is to weigh off the several portions successively ; which may be accomplished most readily by weighing in a glass tube, or other appropriate vessel, the whole amount of substance, and then shaking out of the tube the quantities required one after another into appropriate vessels, weighing the tube after each time. The w r ork may often also be materially lightened, by weighing off a larger portion of the substance, dissolving this to , i or 1 72 OPERATIONS. [ 34, 35, litre, and taking out for the several estimations aliquot parts, with the 50- or 100-c. c. pipette. The first and most essential condition of this proceeding, of course, is that the pipettes must accurately correspond with the measuring flasks (18 and 20). 34. 2. ESTIMATION OF THE WATER. If the substance to be examined after having been freed from moisture by a suitable drying process ( 26-32) contains water, it is usual to begin by determining the amount of this Water. Thi& operation is generally simple ; in some instances, however, it has its difficulties. This depends upon various circumstances, viz., whether the compounds intended for analysis yield their water readily or not ; whether they can bear a red heat without suffering decomposition ; or whether, on the contrary, they give off other volatile substances, besides water, even at a lower temperature. The correct knowledge of the constitution of a compound depends frequently upon the accurate estimation of the water con- tained in it ; in many cases for instance, in the analysis of the salts of known acids the estimation of the water contained in the analyzed compound suffices to enable us to deduce the formula. The estimation of the water contained in a substance is, therefore, one of the most important, as well as most frequently occurring operations of quantitative analysis. The proportion of water con- tained in a substance may be determined in two ways, viz., , from the diminution of weight consequent upon the expulsion of the water ; J, by weighing the amount of water expelled. 35. a. ESTIMATION OF THE WATER FROM THE Loss OF WEIGHT. This method, on account of its simplicity, is most frequently employed. The modus operandi depends upon the nature of the substance under examination. a. The substance hears ignition without losing other Constituents hesides Water, and without absorhing Oxygen. The substance is weighed in a platinum or porcelain crucible, and placed over the gas- or spirit-lamp ; the heat should be very $ 35.] ESTIMATION OF WATER. 73 gentle at first, and gradually increased. When the crucible has been maintained some time at a red heat, it is allowed to cool a little, put still warm under the desiccator, and finally weighed when cold. The ignition is then repeated, and the weight again ascer- tained. If no further diminution of weight has taken place, the process is at at end, the desired object being fully attained. But if the weight is less than after the first heating, the operation must be repeated until the weight remains constant. In the case of silicates, the heat must be raised to a very high degree, since many of them (e.g., talc, steatite, nephrite) only begin at a red heat to give off water, and require a yellow heat for the complete expulsion of that constituent. (Tn. SCHEERER.*) Such bodies are therefore ignited over a blast-lamp. The flame should be observed ; if it be colored, it indicates some volatilization of alkali. hi the case of substances that have a tendency to puff off, or to spirt, a small flask or retort may sometimes be advantageously sub- stituted for the crucible. Care must be taken to remove the last traces of aqueous vapor from the vessel, by suction through a glass tube. Decrepitating salts (sodium chloride, for instance) are put finely pulverized, if possible in a small covered platinum crucible,. whi;'h is then placed in a large one, also covered ; the whole is weighed, then heated, gently at first for some time, then more strongly ; finally, after cooling, weighed again. ft. The substance loses on ignition other Constituents besides Water (Boric Acid, Sulphuric Acid, Silicon Fluoride, tv?.). Here the analyst has to consider, in the first place, whether the water may not be expelled at a lower degree of heat, which does not involve the loss of other constituents. If this may be done, the substance is heated either in the water-bath, or where a higher temperature is required, in the air-bath or oil-bath, the tempera- ture being regulated by the thermometer. The expulsion of the water may be promoted by the co-operation of a current of air (compare 29 and J30) ; or by the addition of pure dry sand to the substance, to keep it porous, f The process must be continued under these circumstances also, until the weight remains constant. * Jahresber. von LIEBIG u. KOPP, 1851, 610. f Ann. d. Chem. u. Pkarm., LIII, 233. 74 OPERATIONS. [ 35. In cases where, for some reason or other, such gentle heating is insufficient, the analyst has to consider whether the desired end may not be attained at a red heat, by adding some substance that will retain the volatile constituent whose loss is apprehended. Thus, for instance, the crystallized aluminium sulphate loses at a red heat, besides water, also sulphuric acid ; now, the loss of the latter constituent may be guarded against by adding to the sul- phate an excess (about six times the quantity) of finely pulverized, recently ignited, pure lead oxide. But the addition of this sub- stance will not prevent the escape of silicon fluoride from silicates when exposed to a red heat (LIST *). Thus, again, the amount of water in commercial iodine may be determined by triturating the iodine together with eight times the quantity of mercury, and drying the mixture at 100 (BoL- LEY, DINGLER'S Polytech. Journ., cxxvi, 39). For the determination of water in silicofluorides magnesia is added to the substance. For this purpose about twice as much magnesia as is required for decomposing the silicofluoride is ig- nited in a platinum crucible, weighed, stirred with warm water to a thick paste, the weighed silicofluoride added, the whole stirred with a platinum wire of known weight, more water added if necessary to effect solution, and the mixture then carefully dried and ignited. The loss of weight represents the water con- tained in the silicofluoride, since the decomposition products magnesium fluoride, silicic acid, and metallic oxide weigh as much as the anhydrous silicofluoride plus the magnesia. A cor- rection in this case would be necessary only when the separated metallic oxide, e.g., ferrous oxide, takes up atmospheric oxygen on ignition (F. STOLBA f ). y. The substance contains several differently combined quantities of Water which require different Degrees of Temperature for Expulsion. Substances of this nature are heated first in the water-bath, until their weight remains constant; they are then exposed in the oil- or air-bath to 150, 200, or 250, &c., and finally, when * Ann. d. Chem. u. Pharm., LXXXI, 189. \ Zeitschr.f. analyt. Chem., vn, 93. 36.] ESTIMATION OF WATER. 75 practicable, ignited over a gas- or spirit-lamp. In such opera- tions I prefer to use the apparatus illustrated in Fig. 38. The bulb-tube may also be replaced, if desired, by a tube of uniform width in which is slid a small porcelain boat containing the substance. In order to prevent a desiccated substance from attracting water during the weighing, the boat is weighed in a cork-stoppered glass tube. In this manner differently combined quantities of water may be distinguished, and their respective amounts correctly esti- mated. Thus, for instance, crystallized copper sulphate contains 28 "87 per cent, of water, which escapes at a temperature below 140, and 7*22 per cent., which escapes only at a temperature between 220 and 260. It is frequently advisable to assist the action of heat by the aid of a partial vacuum. Thus magnesium sulphate in vacua over sulphuric acid at 100 loses 5 equivalents of water; dried in the air at 132 it loses 6, and at alow red heat 7, equivalents. tf. When the substance has a tendency to absorb oxygen (from the presence of ferrous compounds, for instance) the water is bet- ter determined in the direct way than by the loss. ( 36.) 36. 5. ESTIMATION .OF WATER BY DIRECT WEIGHING. This method is resorted to by way of control, or in the case of substances which, upon ignition, lose, besides water, other con- stituents, which cannot be retained even by the addition of some other substance (e.g., carbon dioxide, oxygen), or in the case of substances containing bodies inclined to oxidation (e.g., ferrous compounds). The principle of the method is to expel the water by the application of a red heat, so as to admit of the condensa- tion of the aqueous vapor, and the collection of the condensed water in an appropriate apparatus, partly physically, partly by the agency of some hygroscopic substance. The increase in the weight of this apparatus represents the quantity of the water ex- pelled. The operation may be conducted in various ways ; the follow- ing apparatus, Fig. 44-, is one of the most appropriate : 76 OPERATIONS. [36, .Z?, Fig. 44, represents a gasometer filled with air ; b a flask half-filled with concentrated sulphuric acid ; c and a o are calcium- chloride tubes ( 66, 7); d is a bulb-tube of highly infusible glass. Fig. 44. The substance intended for examination is weighed in the perfectly dry tube d, which is then connected with c and the weighed calcium chloride tube ao, by means of sound and well- dried perforated corks. The operation is commenced by opening the stop-cock of the gasometer a little, to allow the air, which loses all its moisture in A and is a common combustion furnace ; cf a tube filled as fol- lows : from c to d with lead carbonate, * from d to e the substance intimately mixed with lead carbonate, and from e to f pure lead car- bonate. The calcium-chloride tube r aqueous alkalies usually require decomposition by fluxing, to prepare them for analysis. Substances of this kind are often met with in the min- eral kingdom ; most silicates, the sulphates of the alkali-earth metals, chrome ironstone, &c., belong to this class. The object and general features of the process of fluxing have already been treated of in the qualitative part of the present work. 40,41.] SOLUTION. EVAPORATION. 81 The special methods of conducting this important operation will be described hereafter under "The analysis of silicates," and in the proper places; as a satisfactory description of the process, .with its various modifications, cannot well be given without entering 'more minutely into the particular circumstances of the several special cases. Decomposition by fluxing often requires a higher temperature than is attainable with a spirit-lamp with double draught, or with a common gas-lamp. In such cases, the glass-blower's lamp, fed with gas, is used with advantage.* 40. 4. CONVERSION OF DISSOLVED SUBSTANCES INTO WEIGHABLE FORMS. The conversion of a substance in a state of solution into a form adapted for weighing may be effected either by evaporation or by precipitation. The former of these operations is applicable only in cases where the substance, the weight of which we are desirous to ascertain, either exists already in the solution in the form suit- able for the determination of its weight, or may be converted into such form by evaporation in conjunction with some reagent. The solution must, moreover, contain the substance unmixed, or, at least, mixed only with such bodies as are expelled by evaporation or at a red heat. Thus, for instance, the amount of sodium sulphate present in an aqueous solution of that substance may be ascertained by simple evaporation ; whilst the potassium carbonate contained in a solution would better be converted into potassium chloride, by evaporating with solution of ammonium chloride. Precipitation may always be resorted to, whenever the substance in solution admits of being converted into a combination which is insoluble in the menstruum present, provided that the precipitate is fit for determination, which can never be the case unless it can be washed and is of constant composition. 41. a. EVAPORATION. In processes of evaporation for pharmaceutical or technico- chemical purposes the principal object to be considered is saving * Excellent lamps of this kind are made by DESAGA, of Heidelberg. 82 OPERATIONS. $ 41. of time and fuel; but in evaporating processes in quantitative analytical researches this is merely a subordinate point, and the analyst has to direct his principal care and attention to the means of guarding against loss or contamination of the substance operated upon. The simplest case of evaporation is when we have to concentrate a clear fluid, without carrying the process to dryness. To effect this object, the fluid is poured into a basin, which should not be filled to more than two-thirds. Heat is then applied by placing the basin either on a water-bath, sand-bath, common stove, or heated iron plate, or over the flame of a gas- or spirit-lamp, care being taken always to guard against actual ebullition, as- this in- variably and unavoidably leads to loss from small drops of fluid spirting out. Evaporation over a gas- or spirit-lamp, when con- ducted with proper care, is an expeditious and cleanly process, BUNSEN'S gas-lamp, Fig. 46, may be used most advantageously in operations of this kind ; a little wire-gauze cap, loosely fitted upon the tube of the lamp, is a material improvement. By means Fig. 46. Fig. 47. of this simple arrangement it is easy to produce even the smallest flame, without the least apprehension of the flame striking back. The lamp recently introduced by the MASTE Brothers, of Iserlohn, and illustrated in Fig. 47, affords excellent service both for evaporation and ignition. In it the burner is very similar to 41.] EVAPORATION. 83 Fig. 48. that of the BERZELIUS alcohol lamp, and affords a very small as well as a very powerful large flame, and it has given me excellent results during long-continued use. Five different sizes are made. The gas furnace shown in Fig. 48 is also excellently adapt- ed for evaporations carried en in evapo- rating-dishes. In this furnace the mixture of gas and air issues from many small ori- fices, and the construc- tion enables the flames to be made so small that the contents of the dish can be quietly evaporated without ebullition.* If the evaporation is to be effected on the water-bath, and the operator happens to possess a BEINDOKF, or other similarly con- structed, steam apparatus, the evaporating- dish may be placed simply into an opening corresponding in size. Otherwise recourse must be had to the water-bath illustrated by Fig. 49. It is made of strong sheet copper, and when used is half filled with water, which is kept boiling over a gas-, spirit-, or oil-lamp. The breadth from a to b should be from 12 to 18 cm. Various flat rings of the same outside diameter as' the top of the bath, and adapted to receive dishes and crucibles of difierent sizes, are essential adjuncts to the bath. These rings when required are simply laid on the bath. It is very inconvenient to have the water in the bath com- pletely evaporate unnoticed, because frequently residues become heated to a higher degree than is desirable, or concentrated solu- tions spirt, etc. To avoid such inconveniences, I make use of a water-bath with constant level, as shown in Fig. 50. This ap- paratus consists of a zinc vessel, abed, 10 cm. high and 12 cm. * My furnaces are made by KILIAN, of Wiesbaden. - 49> 84 OPERATIONS. [ 41. in diameter, and connected with the water-bath, g, by means of the short rubber tube. 0, and the copper tube, f. A sheet-zinc Fig. 50. bottle, h i k Z, the cylindrical part of which is 17 cm. high and the neck 3 cm. in diameter, is filled with water, and, inverted, placed in the vessel, abed. The orifice of the bottle at the neck is 15 mm. wide, and, in the inverted position, is closed by a valve, m. On inserting the bottle into the vessel, the wire carrying the valve strikes the bottom of the vessel and opens the valve. The level of the water in g is readily regulated by raising or low- ering the pillar support, 0, and remains constant so long as any water remains in the bottle. The tube/" is bent downward in the water-bath and reaches nearly to the bottom. A simple arrangement for extinguishing the flame when all the water in the water-bath has evaporated has been described by K. REUSS*; the construction of BUNSEN'S constant water-bath has been detailed by W. H. WAHL.f If the operator can conduct his processes of evaporation in a room set apart for the purpose, where he may easily guard against any occurrence tending to suspend dust in the air, he will find it no very difficult task to keep the evaporating fluid clean ; in this * Zeitschr.f. analyt. Chem., ix, 336. f Ibid., x, 88. 41.] EVAPORATION. 85 case it is best to leave the dishes uncovered.* But in a large laboratory, frequented by many people, or in a room exposed to draughts of air, or in which coal fires are burning, the greatest caution is required to protect the evaporating fluid from contami- nation by dust or ashes. For this purpose the evaporating dish is either covered with a sheet of filtering-paper turned down over the edges, or a glass rod twisted into a triangular shape, Fig. 51, is laid' upon it, and a sheet of filtering-paper spread over it, which is kept in position by a glass rod laid across, the latter again b.eing kept from rolling down by the slightly turned up ends, a and >, of the triangle. The best way, however, is the following : Take two small thin wooden hoops, Fig. 52, one of which fits loosely in the other ; spread a sheet of blotting-paper over the smaller ^^ one, and push the other over it. This forms a cover admirably adapted to the purpose ; and whilst in no way interfering with the operation, it completely protects the evaporating fluid from dust, and may be readily taken off ; the paper cannot dip into the fluid ; the cover lasts a long time, and may, moreover, at any time be easily renewed. It must be borne in mind, however, that the common filtering- paper contains always certain substances soluble in acids, such as lime, ferric oxide, &c., which, were covers of the kind just described used over evaporating dishes containing a fluid evolving acid vapors, would infallibly dissolve in these vapors, and the solu- tion dripping down into the evaporating fluid, would speedily con- taminate it. Care must be taken, therefore, in such cases, to use only such filtering-paper as has been freed by washing from sub- stances soluble in acids. Evaporation for the purpose of concentration may be effected also in flasks ; these are only half filled, and placed in a slanting * In my own laboratory separate closets are set apart for evaporations in quantitative analyses. It is best to' have the floor and roof of sandstone, and the walls of brick lined with glazed tiles or finished with plaster of Paris. At the top of the back wall is a horizontal channel of suitable width, and leading into a Russian chimney. No fire must be made under this chimney, but it is very desirable to place this chimney close to another chimney kept constantly warm (by the fire used for the steam apparatus, for instance). The front wall of the evaporating chamber may be of sandstone pillars 18 decimeters high, be- tween which are fitted wooden frames wherein balanced windows may slide up and down. OPERATIONS. [41. position. The process may be conducted on the sand-bath, or over a gas- or spirit-lamp, or even, and with equal propriety, over a char- coal fire. In cases where the operation is conducted over a lamp or a charcoal fire, it is the safest way to place the flasks on wire gauze. Gentle ebullition of the fluid can do no harm, here, since the slanting position of the flask guards effectively against risk of loss from the spirting of the liquid. Still better than in flasks, the object may be attained by evaporating in tubulated retorts with open tubulure and neck directed obliquely upwards. The latter acts as a chimney, and the constant change of air thus effected is extremely favorable to evaporation. The evaporation of fluids containing a precipitate is best con- ducted on the water-bath ; since on the sand-bath, or over the lamp, it is next to impossible to guard against loss from bumping. This Fig. 53. bumping is occasioned by slight explosions of steam, arising from the sediment impeding the uniform diffusion of the heat. Still there remains another, though less safe way, viz., to conduct the evaporation in a crucible placed in a slanting position, as illus- trated in fig. 53. In this process, the flame is made to play upon the crucible above the level of the fluid. Where a fluid has to be evaporated to dry ness, as is so often the case, the operation should always, if possible, be terminated on the water-bath. In cases where the nature of the dissolved sub- stance precludes the application of the water-bath, the object in view mav often be most readily attained by heating the contents 41.] EVAPORATION. 87 of the dish from the top, which is effected by placing the dish in a proper position in a drying closet, the upper plate of which is heated by a flame (that of the water- or sand-bath) passing over it. If the substance is in a covered platinum dish or crucible, place the gas- lamp in such a position that the flame may act on the cover from above. In cases where the heat has to be applied from the bottom, a method must be chosen which admits of an equal diffusion and ready regulation of the heat. An air-bath is well adapted for this purpose, i.e.. a dish of iron plate, in which the porcelain or platinum dish is to be placed on a wire triangle, so that the two vessels may be at all points J to J inch distant from each other. The copper apparatus, fig. 49 may also serve as an air-bath, although I must not omit to mention that this mode of application will in the end seriously injure it. If the operation has to be conducted over a lamp, the dish should be placed high above the flame ; best on wire gauze, since this will greatly contribute to an equal diffusion of the heat. The use of the sand-bath is objectionable here, because with that apparatus we cannot reduce the heat so speedily as may be desirable. An iron plate heated by gas may perhaps be used with advantage. But no matter which method be employed, this rule applies equally to all of them ; that the operator must watch the process, from the moment that the residue begins to thicken, in order to prevent spirting, by reducing the heat, and breaking the pellicles which form on the surface, with a glass rod, or a platinum wire or spatula, Saline solutions that have a tendency, upon their evaporation, to creep up the sides of the vessel, and may thus finally pass over the brim of the latter, thereby involving the risk of a loss of substance, should be heated from the top, in the way just indicated ; since by that means the sides of the vessel will get heated sufficiently to cause the instantaneous evaporation of the ascending liquid, pre- venting thus its overflowing the brim. The inconvenience just alluded to may, however, be obviated also, in most cases, by cover- ing the brim, and the uppermost part of the inner side of the ves- sel, with a very thin coat of tallow, thus diminishing the adhesion between the fluid and the vessel. In the case of liquids evolving gas-bubbles upon evaporating^ particular caution is required to guard against loss from spirting. The safest way is to heat such liquids in an obliquely-placed flask, or in a beaker covered with a large watch-glass ; the latter i? 88 OPERATIONS. [ 41. removed as soon as the evolution of gas-bubbles has ceased, and the fluid that may have spirted up against it is carefully rinsed into the glass, by means of a washing-bottle. If the evaporation has to be conducted in a dish, a rather capacious one should be selected, and a very moderate degree of heat applied at first, and until the evolution of gas has nearly ceased. If a fluid has to be evaporated with exclusion of air, the best way is to place the dish under the bell of an air-pump, over a ves- sel with sulphuric acid, and to exhaust; or a tubulated retort may be used through whose tubulure hydrogen or carbon dioxide is passed by the acid of a tube not quite reaching to the surface of the fluid. The material of the evaporating vessels may exercise a much greater influence on the results of an analysis than is generally believed. Many rather startling phenomena that are observed in analytical processes may arise simply from a contamination of the evaporated liquid by the material of the vessel ; great errors may also spring from the same source.* The importance of this point has induced me to subject it to a searching investigation (see Appendix, Analytical Experiments, 1-4) ; more recently A. EMMERLING has also exhaustively inves- tigated the subject, and fully confirms the results, which I will here briefly intimate. Distilled water kept boiling for some length of time in glass (flasks of Bohemian glass) dissolves very appreciable traces of that material. This is owing to the formation of soluble silicates ; the particles dissolved consist chiefly of potassa, or soda and lime, in combination with silicic acid. A much larger proportion of the glass is dissolved by water containing caustic or carbonated alkali ; boiling solution of ammonium chloride also strongly attacks glass vessels. Boiling dilute acids, with the exception, of course, of hydrofluoric and hydrofluosilicilic acids, exercise a less powerful solvent action on glass than pure water. Porcelain (Berlin dishes) is much less affected by water than glass ; alkaline liquids also exercise a less powerful solvent action on porcelain than on glass ; the quantity dissolved is, however, still notable. Solution of ammonium chloride acts on porcelain as strongly as on glass ; dilute acids, though exercising no very powerful solvent action on porcelain, yet attack that material more strongly than glass. It results from these data, that in analyses pretending to a high * Compare A. SOUCHAY, Zeitschr.f. analyt. Chem., iv, 66. 42.] EVAPORATION. 89 degree of accuracy, platinum or platinum- iridium or silver dishes should always be preferred. The former may be used in all casei where no free chlorine, bromine, or iodine is present in the fluid, or can be formed during evaporation. Fluids containing caustic alkalies may safely be evaporated in platinum, but not to the point of fusion of the residue. Silver vessels should never be used to evaporate acid fluids nor liquids containing alkaline sulphides ; but they are admirably suited for solutions of alkali hydroxides and carbonates, as well as of most normal salts. If the use of porcelain or glass vessels for evaporating large volumes of fluid cannot be avoided, then porcelain dishes are to be preferred ; with alkaline fluids glass vessels are totally inadmissible, at least in accurate analyses. 42. We come now to weighing the residues remaining upon the evaporation of fluids. We allude here simply to such as are soluble in water; those which are separated by filtration will be treated of afterwards. Residues are generally weighed in the same vessel in which the evaporation has been completed, for which purpose platinum dishes, from 4 to 8 cm. in diameter, pro- vided with light covers, or large platinum cruci- bles, are best adapted, since they are lighter than porcelain vessels of the same capacity. However, in most cases, the amount of liquid to be evaporated is too large for so small a vessel, and its evaporation in portions would occupy too much time. The best way, in cases of this kind, is to concentrate the liquid first in a larger vessel, and to terminate the operation afterwards in the smaller weighing vessel. In transferring the fluid from the larger to the smaller vessel, the lip of the former is slightly greased, and the liquid made to run down a glass rod (Fig. 54). Finally the large vessel is carefully rinsed with a washing- bottle, until a drop of the last rinsing leaves no longer a residue upon evaporation on a platinum knife. When the fluid has thus been transferred to the weighing-vessel, the evaporation is com- pleted on the water-bath and the residuary substance finally ignited, provided, of course, it will admit of this process. For this pur- 90 OPERATIONS. [ 42. pose the dish is covered with a lid of thin platinum (or a thin glass plate), and then placed high over the flame of a lamp, and heated very gently until all the water which may still adhere to the sub- stance is expelled ; the dish is now exposed to a stronger, and finally to a red heat. (Where a glass plate is used, this must, of course, be removed before igniting.) In this case it is also well to make the flame play obliquely on the cover from above, so as to run as little risk as possible of loss by spirting. After cooling in a desic- cator, the covered dish is weighed with its contents. When oper- ating upon substances which decrepitate, such as sodium chloride, for instance, it is advisable to expose them after their removal from the water-bath, and previously to the application of a naked flame to a temperature somewhat above 100, either in the air- bath, or on a sand-bath, or on a common stove. If the residue does not admit of ignition, as is the case, for instance, with organic substances, ammonium salts, &c., it is dried .at a temperature suited to its nature. In many cases, the tempera- ture of the water-bath is sufficiently high for this purpose, for the drying of ammonium chloride, for instance ; in others, the air or oil-bath must be resorted to. (See 29 and 30.) Under any cir- cumstances, the desiccation must be continued until the substance ceases to suffer the slightest diminution in weight, after renewed exposure to heat for half an hour. The dish should invariably be covered during the process of weighing. Since saline residues obtained on evaporation are frequently prone to attract moisture after drying or ignition, the first weigh- ing, which always requires some time, may give results which are too high. To avoid this, the dish is reheated after the first weighing, then allowed to cool in the exsiccator ; the weight ob- tained in the first weighing is then placed on one scale-pan and the dish placed on the other, when the second weighing is ac- complished with as little loss of time as possible. If, as will frequently happen, we have to deal with a fluid con- taining a small quantity of a potassium or sodium salt, the weight of which we want to ascertain, in presence of a comparatively large amount of an ammonium salt, which has been mixed with it in the course of the analytical process, I prefer the following method : The saline mass is thoroughly dried, in a large dish, on the water- bath, or, towards the end of the process, at a temperature some- what exceeding 100. The dry mass is then, with the aid of a 43.] EVAPORATION. 91 platinum spatula, transferred to a small glass dish, which is put aside for a time in a desiccator. The last traces of the salt left adhering to the sides and bottom of the large dish are rinsed off with a little water into the small dish, or the large crucible, in which it is intended to weigh the salt ; the water is then evaporated, and the dry contents of the glass dish are added to the residue : the ammonium salts are now expelled by ignition, and the residu- ary fixed salts finally weighed. Should some traces of the saline mass adhere to the smaller glass dish, they ought to be removed and transferred to the weighing vessel, with the aid of a little pounded ammonium chloride, or some other ammonium salt, as the moistening again with water would involve an almost certain loss of substance. 43 1). PRECIPITATION. Precipitation is resorted to in quantitative analysis far more frequently than evaporation, since it serves not merely to convert substances into forms adapted for weighing, but also, and more especially, to separate them from one another. The principal in- tention in precipitation, for the purpose of quantitative estimations, is to convert the substance in solution into a form in which it is insoluble in the menstruum present. The result will, therefore, cceteris paribus, be the more accurate, the more the precipitated body deserves the epithet insoluble, and in cases where precipi- tates are of the same degree of solubility, that one will suffer the least loss which comes in contact with the smallest amount of solvent. Hence it follows, first, that in all cases where other circum- stances do not interfere, it is preferable to precipitate su-bstances in their most insoluble form ; thus, for instance, barium had better be precipitated as sulphate than as carbonate ; secondly, that when we have to deal with precipitates that are not quite insoluble in the menstruum present, we must endeavor to remove that men- struum, as far as practicable, by evaporation ; thus a dilute solution of strontium should be concentrated, before proceeding to precipi- tate the strontium with sulphuric acid ; and, thirdly, that when we have to deal with precipitates slightly soluble in the liquid present, but insoluble in another menstruum, into which the former may 92 OPEEATIONS. [ 43. be converted by the addition of some substance or other, we ought to endeavor to bring about this modification of the menstruum. Thus, for instance, alcohol may be added to water, to induce com- plete precipitation of ammonium platinic chloride, lead chloride, calcium sulphate, &c.; thus again, ammonium magnesium phosphate may be rendered insoluble in an aqueous menstruum by adding ammonia to the latter, &c. Precipitation is generally effected in beakers. In cases, how- ever, where we have to precipitate from fluids in a state of ebulli- tion, or where the precipitate requires to be kept boiling for some time with the fluid, flasks or dishes are substituted for beakers, with due regard always to the material of which they are made (see Evaporation, 41, at the end). The separation of precipitates from the fluid in which they are suspended, is effected either by decantation or filtration, or by both these processes jointly. But, before proceeding to the sepa- ration of the precipitate by any of these methods, the operator must know whether the* precipitant has been added in sufficient quantity, and whether the precipitate is completely formed. To determine the latter point, an accurate knowledge of the properties of the various precipitates must be attained, which we shall en- deavor to supply in the third section. To decide the former ques- tion, it is usually sufficient to add to the fluid (after the precipitate has settled) cautiously a fresh portion of the precipitant, arid to note if a further turbidity ensues. This test, however, is not infallible, when the precipitate has not the property of forming immediately ; as, for instance, is the case with ammonium phos- pho-molybdate. When this is apprehended, pour out (or transfer with a pipette) a small quantity of the clear supernatant fluid into another vessel, add some of the precipitant, warm if necessary ; and after some time look and see whether a fresh precipitate has formed. As a general rule, the precipitated liquid should be allowed to stand at rest for several hours, before proceeding to the separation of the precipitate. This rule applies more particularly to crystalline, pulverulent, and gelatinous precipitates, whilst curdy and flocculent precipitates, more particularly when the precipitation was effected at a boiling temperature, may often be filtered off im- mediately. However, we must observe here, that all general rules, in this respect, are of limited application. 44.] DECANTATION. 93 44. a. SEPARATION OF PRECIPITATES BY DECANTATION. When a precipitate subsides so completely and speedily in a fluid that the latter may be decanted off perfectly clear, or drawn off with a syphon, or removed by means of a pipette, and that the washing of the precipitate does not require a very long time, decantation is often resorted to for its separation and washing ; this is the case, for instance, with silver chloride, metallic mer- cury, &c. Decantation will always be found a very expeditious and accu- rate method of separation, if the process be conducted with due care ; it is necessary, however, in most cases, to promote the speedy and complete subsidence of the precipitate ; and it may be laid down as a general rule, that heating the precipitate with the fluid will produce the desired effect. Nevertheless, there are instances in which the simple application of heat will not suffice ; in some cases, as with silver chloride, for instance, agitation must be resorted to ; in other cases, some reagent or other is to be added hydrochloric acid, for instance, in the precipitation of mercury, &c. We shall have occasion, subsequently, in the fourth section, to discuss this point more fully, when we shall also mention the vessels best adapted for the application of this process to the various precipitates. After having washed the precipitate repeatedly with fresh quantities of the proper fluid, until there is no trace of a dissolved substance to be detected in the last rinsings, it is placed in a crucible or dish, if not already in a vessel of that description ; the fluid still adhering to it is poured off as far as practicable, and the precipitate is then, according to its nature, either simply dried, or heated to redness. A far larger amount of water being required for washing pre- cipitates by decantation than on filters, the former process can be 3xpected to yield accurate results only where the precipitates are absolutely insoluble. For the same reason, decantation is not ordi- narily resorted to in cases where we have to determine other con^ stituents in the decanted fluid. The decanted fluid must be allowed to stand at rest from twelve to twenty-four hours, to make quite sure that it contains no particles of the precipitate ; if, after the lapse of this time, no 94 OPERATIONS. [ 45. precipitate is visible, the fluid may be thrown away ; but if a pre- cipitate has subsided, this had better be estimated by itself, and the weight added to the main amount ; the precipitate may, in such cases, be separated from the supernatant fluid by decantation, or by filtration. 45. /3. SEPARATION OF PRECIPITATES BY FILTRATION. This operation is resorted to whenever decantation is imprac- ticable, and, consequently, in the great majority of cases ; provided always the precipitate is of a nature to admit of its being com- pletely freed, by mere washing on the filter, from all foreign substances. Where this is not the case, more particularly, there- fore, with gelatinous precipitates, aluminium hydroxide for in- stance, a combination of decantation and filtration is resorted to ( 48). Filtration is effected either with or without exhaustion of the liquid ; in the latter case, however, it is greatly accelerated. 4:5. aa. ORDINARY FILTRATION. aa. FILTERING APPARATUS. Filtration, as a process of quantitative analysis, is almost exclusively effected by means of paper. Plain circular filters are most generally employed ; plaited fil- ters are only occasionally used. Much depends upon the quality of the paper. Good filtering paper must possess the three follow- ing properties: 1. It must completely retain the finest precipi- tates ; 2. It must filter rapidly ; and 3. It must be as free as possible from any admixture of inorganic bodies, but more espe- cially from such as are soluble in acid or alkaline fluids. It is a matter of some difficulty, however, to procure paper fully answering these conditions. The Swedish filtering paper, with the water-mark J. H. MUNKTELL, is considered the best, and, consequently, fetches the highest price ; but even this answers only the first two conditions, being by no means sufficiently pure for very accurate analyses, since, it leaves upon incineration about O3> 45.] FILTRATION. 95 per cent, of ash,* and yields to acids perceptible traces of lime, mag- nesia, and ferric oxide. For exact experiments it is, consequently, necessary first to extract the paper with dilute hydrochloric acid, then to wash the acid completely out with water, and finally to dry the paper. In the case of very fine filtering paper, the best way to perform this operation is to place the ready-cut filters, several together, in a funnel, exactly the same way as if intended for immediate filtration ; they are then moistened with a mixture of one part of ordinary pure hydrochloric acid with two parts of water, which is allowed to act on them for about ten minutes ; after this all traces of the acid are carefully removed by washing the filters in the funnel repeatedly with warm water. The funnel being then covered with a piece of paper, turned over the edges, is put in a warm place until the filters are dry. Compare the instruction given in the "Qual. Anal.," Am. Ed., p. 8, on the preparation of washed filters. Filter paper containing lead, and which is consequently blackened by sulphuretted hydrogen, should be rejected, f Iteady-cut filters of various sizes should always be kept on hand. Filters are either cut by circular patterns of pasteboard or tin, or, still better, by Moire's filter- patterns, fig. 55. This little apparatus" is made of tin-plate, and consists of two parts. B is a quadrant fitting in A, whose straight edges are turned up, and which is slightly smaller than B. The sheets of filter- paper are first cut up into squares, which are folded in quarters, and placed in A, then B is placed on the top, and the free edge of the paper is cut off with scissors. Filters cut in this way are per- fectly circular, and of equal size. Several pairs of these patterns of various sizes (3, 4, 5, 6, 6*5,, and 8 cm. radius) should be procured. In taking a filter for a given operation, you should always choose one which, after the fluid has run through, will not be more than half filled with the precipitate, * Plantamour found the ash of Swedish filtering paper to consist of 63*23 silicic acid, 12 '83 lime, 6*21 magnesia, 2*94 alumina, and 13*92 ferric oxide, ill 100 parts. | WICKE, Annal. d. Chem. u. Fharm., cxn, 127. Fig. 55. 96 OPERATIONS. [45. As to the funnels, they should be inclined at the angle of 60, and not bulge at the sides. Glass is the most suitable material for them. Fig. 56. Fig. 57. The filter should never protrude beyond the funnel. It should come up to one or two lines from the edge of the latter. The filter is firmly pressed into the funnel, to make the paper fit closely to the side of the latter ; it is then moistened with water ; any extra water is not poured out, but allowed to drop through. The stands shown in figs. 56 and 57 complete the apparatus for filtering. The stands are made of hard wood. The arm holding the funnel or funnels must slide easily up and down, and be fixable by the screw. The holes for the funnels must be cut conically, to keep the funnels steadily in their place. These stands are very convenient, and may be readily moved about without interfering with the operation. fifi. RULES TO BE OBSERVED IN THE PROCESS OF FlLTRATION. In the case of curdy, flocculent, gelatinous, or crystalline pre- cipitates there is no danger of the fluid passing turbid through the filter. But with fine pulverulent precipitates it is generally neces- sary, and always advisable, to let the precipitate subside, and then filter the supernatant liquid, before proceeding to place the precipi- 45.] FILTRATION. 97 tate upon the filter. We generally proceed in this way also with other kinds of precipitates, especially with those that require to stand long before they completely separate. Precipitates which have been thrown down hot, are most properly filtered off before cooling (provided always there be no objections to this course), since hot fluids run through the filter more speedily than cold ones. Some precipitates have a tendency to be carried through the filter along with the fluid ; this may be prevented in some instances by modifying the latter. Thus barium sulphate, when filtered from an aqueous solution, passes rather easily through the filter the addition of hydrochloric acid or ammonium chloride prevents this in a great measure. If the operator finds, during a filtration, that the filter would be much more than half filled by the precipitate, he would better use an additional filter, and thus distribute the precipitate over the two ; for, if the first were too full, the precipitate could not be properly washed. The fluid ought never to be poured directly upon the filter, but always down a glass rod (see Fig. 54), and the lip or rim of the vessel from which the fluid is poured should always be slightly greased with tallow.* The stream ought invariably to be directed towards the sides of the filter, never to the centre, since this might occasion loss by splashing. In cases where the fluid has to be filtered off, with the least possible disturbance of the precipitate, the glass rod must not be placed, during the intervals, in the vessel containing the precipitate; but it may conveniently be put into a clean glass, which is finally rinsed with the wash- water. The filtrate is received either in flasks, beakers, or dishes, according to the various purposes for which it may be intended. Strict care should be taken that the drops of fluid filtering through glide down the side of the receiving vessel ; they should never be allowed to fall into the centre of the filtrate, since this again might occasion loss by splashing. The best rnethocl is that shown in Fig. 56, viz., to rest the point -of the funnel against the upper part of the inside of the receiving vessel. If the process of filtration is conducted in a place perfectly free from dust, there is no necessity to cover the funnel, nor the * The tallow may be kept under the edge of the work-table at a conv3nient point, where it will adhere by a little pressure. The best way of applying the tallow to the lip of a vessel i> with the greased ringer.' 98 OPERATIONS. [ 46. vessel receiving the filtrate ; however, as this is but rarely the case, it is generally indispensable to cover both. This is best effected with round plates of sheet-glass. The plate used for covering the receiving vessel should have a small U-shaped piece cut out of its edge, large enough for the funnel-tube to go through. The effect desired may be produced by cautiously chipping out the glass bit by bit with the aid of a key. Plates perforated in the centre are worthless as regards the object in view. After the fluid and precipitate have been transferred to the filter, and the vessel which originally contained them has been rinsed repeatedly with water, it happens generally that small par- ticles of the precipitate remain adhering to the vessel, which can- not be removed with the glass rod. From beakers or dishes these particles may be readily removed by means of a feather prepared for the purpose by tearing off nearly the whole of the plumules, leaving only a small piece at the end which should be cut per- fectly straight. From flasks, minute portions of heavy precipitates which are not adherent, are readily removed by blowing a jet of water into the flask, held inverted over the funnel ; this is effected by means of the washing-bottle shown in Fig. 60, after the tube 5 has been properly directed. If the minute adhering particles of a precipitate cannot be removed by mechanical means, solution in an appropriate menstruum must be resorted to, followed by re-pre- cipitation. Bodies for which we possess no solvent, such as barium sulphate, for instance, must not be precipitated in flasks. 46. yy. WASHING OF PRECIPITATES. After having transferred the precipitate completely to the filter, we have next to perform the operation of washing ; this is effected by means of owe. of the well-known washing-bottles, Figs. 58, 59, and 60.* The doubly perforated stoppers are of vulcanized rubber. By the arrangement shown in Fig. 60, in which a short piece of wide glass tubing a is connected by means of pieces of rubber tubing with the tip J, the jet of water may be turned in any direc- * A wash-bottle for odorous liquids has been devised by JACOB, Zeitschr.f. analyt. Chem., v, 168. 46.] FILTRATION. tion, and even upwards, by simply turning b. Care must be taken that no loss is occasioned by too violent a stream of water. Fig. 58. Fig. 59. Fig. 60. Where great caution is required in washing a precipitate, the arrangement shown in Fig. 61 can be used with good results; its construction requires 110 explanation. The point of a is drawn out and broken off. On inverting the flask it delivers a fine, continuous stream of water. Precipitates requiring washing with water are washed most expeditiously with hot water, provided always there be no special reason against its use. The wash- ing-bottle shown in Fig. 59 is particularly well adapted for boiling water. The wooden handle which is fastened to the neck of the flask with wire serves to facil- itate holding it. If this is not desired, the neck of the bottle may be wound with cord of suitable thickness. It is a rule in washing precipitates not to add fresh wash- water to the filter till the old has quite run through. In applying the jet of water you have to take care on the one hand that the upper edge of the filter is properly washed, and on the other hand that Fig. 01. ICO OPERATIONS. [ 47. no canals are formed in the precipitate, through whicli the fluid runs off, without coming in contact with the whole of the precipi- tate. If such canals have formed and cannot be broken up by the jet, the precipitate must be stirred cautiously with a small platinum knife or glass rod. The washing may be considered completed when all soluble matter that is to be removed has been got rid of. The beginner w r ho devotes proper attention to the completion of this operation shuns one of the rocks which he is most likely to encounter. Whether the precipitate has been completely washed may generally be ascertained by slowly evaporating a drop of the last washings upon a platinum knife, and observing if a residue is left. But in cases where the precipitate is not altogether insoluble in water (strontium sulphate, for instance), recourse must be had to more special tests, which we shall have occasion to point out in the course of the work. The student should never discontinue the washing of a precipitate because he simply imagines it is finished he must be certain. Formerly continuous wash-bottles were employed for pro- tracted washings. They have, however, fallen into disuse because in their employment canals readily form in the precipitates, a large quantity of water is required, and the use of hot water is excluded. Hence it is now customary to treat precipitates as described in 48. Those interested in the construction of con- tinuous wash-bottles will find them described and illustrated in the Handwdrterbuch der Chemie, 2d edit., n, 584586. 47. J5. Filtration by Suction. Filtration being a frequent and tedious operation, many at- tempts to facilitate the operation by employing suction have been made for a long time. BUNSEN * has more recently studied the subject exhaustively. In order to avoid the danger of breaking the filter, the fear of which has prevented chemists from generally employing this method, care must be taken that the filter lies close to the funnel down to the point. Hence funnels should be chosen * Ann. d. Chem. u. Pharm., CXLVIII, 269; also Zeitschr. f. analyt. Ckem., viii, 174 47.] FILTRATION. 101 the sides of which are inclined at an angle of 60, arid free from inequalities of surface. In the funnel should be placed a vary thin, exactly fitting platinum cone, and in the latter is placed the filter so that, after being moistened, it will be in contact at all points, and without any intervening air-bubbles. The preparation of the platinum cone is thus given by BUNSEN : A sheet of writing-paper is formed into a filter, and accurately adjusted to the sides of a carefully selected funnel, when it is fixed in place in the funnel by a few drops of sealing-wax applied to the upper margin. The filter is then impregnated with oil, and filled with plaster of Paris, in which a handle is inserted before the plaster has set. After a few hours the plaster cast, with its paper again oiled, is imbedded in a crucible 4 or 5 cm. high and filled with plaster of Paris. After this has set the plaster cone is re- moved and freed from its oiled-paper covering. There are thus obtained a solid cone and a conical hollow which fit each other perfectly and exactly correspond to the funnel. The platinum cone is now prepared by cutting a piece of sheet platinum (weigh- ing about 0*1-54 grm. per square cm.), of suitable size, to the shape shown in Fig. 62. With a scissors a slit, 5, is cut from the centre of the piece to a point midway on the line cd. The foil is next softened by igni- tion, after which it is laid against the solid cone with the point of the latter at #, and abd pressed against the cone arid the foil wrapped around as closely as possible. The platinum cone is now again ignited, after which it is molded to the plaster cone by hand, then inserted into the hollow cone in which it is tightly pressed. The platinum cone, when finished, should let no light pass through its apex, and even without being soldered is sufficiently firm for all uses. The glass funnel, carrying its platinum cone and filter, is now inserted air-tight into one hole of a doubly perforated rubber stopper, so that the stern projects from 5 to 8 cm. from the stopper; the other perforation carries a short tube bent at right angles, the lower end of which should not project below the stopper. On now inserting the stopper in the neck of a bottle, arid applying suction to the tube, the filtration of any liquid in the filter is effected the 102 OPERATIONS. [47. Fig. 63. 47.] FILTRATION. 103 more rapidly the greater the difference in pressure between the external air and that within the flask. If it is intended to filter under a great difference of pressure, an ordinary flask will not suffice, as it may be shattered by the pressure of the external air. For such a purpose, therefore, a stout glass flask must be used, placed conveniently within a tin vessel, f the examined fluid upon a particular reagent, as is the case when a solution of sodium arsenite is added, drop by drop, to a solution of chlorinated lime, until the mixture no longer imparts a blue tint to paper moistened with potassium iodide and starch- paste, &c. The more sensitive a final reaction is, and the more readily, posi- tively, and rapidly it manifests itself, the better is it calculated to serve as the basis of a volumetric method. In cases where it is an object of great importance to ascertain with the greatest practica- ble precision the exact moment when the reaction is completed, the analyst may sometimes prepare, besides the actual standard solu- tion, another, ten times more dilute, and use the latter to finish the process, carried nearly to completion with the former. But a good final reaction is not of itself sufficient to afford a safe basis for a good volumetric method ; this requires, as the first and most indispensable condition, that the particular decomposition which constitutes the leading point of the analytical process should at least under certain known circumstances remain unalterably the same. Wherever this is not the case where the action varies with the greater or less degree of concentration of the fluid, or according as there may be a little more or less free acid present ; or according to the greater or less rapidity of action of the standard solution ; or where a precipitate formed in the course of the process has not the same composition throughout the operation the basis of the volumetric method is fallacious, and the method itself, therefore, of no value. When volumetric analysis first began to find favor, many chemists based new volumetric methods upon final end-reactions, without carefully studying the decompositions involved ; the re- 126 OPERATIONS. [ 54. suit was a great number of volumetric methods, many of which were useless. These have, however, been subjected to a sifting process, particularly by FK. MOHB ; * and in the special part of the present work I have separated the really good methods from the unserviceable. * Lehrbuch d&r Titrirmeihode, 3d edit. SECTION II. REAGENTS. FOR general information respecting reagents, I refer the stu- dent to my volume on ".Qualitative Analysis." The instructions given here will be confined to the preparation, testing, and most important uses of those chemical substances which subserve principally and more exclusively the purposes of quanti- tative analysis. Those reagents which are employed in qualitative investigations, having been treated of already in the volume on the qualitative branch of the analytical science, will therefore be simply mentioned here by name. The reagents used in quantitative analysis are properly arranged under the following heads : A. Reagents for gravimetric analysis in the wet way. B. Reagents for gravimetric analysis in the dry way. C. Reagents for volumetric analysis. 2). Reagents used in organic analysis. The mode of preparing the fluids used in volumetric analysis, will be found where we shall have occasion to speak of their appli- cation. A. REAGENTS FOR GRAVIMETRIC ANALYSIS IN THE WET WAT. I. SIMPLE SOLVENTS. 56. 1. DISTILLED WATER (see " Qual. Anal."). Water intended for quantitative investigations must be perfectly pure. Water distilled from glass vessels leaves a residue upon evaporation in a platinum vessel (see experiment 'No. 5), and is therefore inapplicable for many purposes ; as, for instance, for the determination of the exact degree of solubility of sparingly soluble 127 128 BEAGENTS. [ 57 substances. For certain uses it is necessary to free the water by ebullition from atmospheric air and carbonic acid. 2. ALCOHOL (see " Qual. Anal."). a. Absolute alcohol. . Common alcohol of various degrees of o strength. 3. ETHER. The application of ether as a solvent is very limited. It is more frequently used mixed with alcohol, in order to diminish the solvent power of the latter for certain substances, e.g., ammonium platinic chloride. The ordinary ether of the shops will answer the purpose. 4. CARBON BISULPHIDE (see " Qual. Anal."). This should be purified, if necessary, by shaking with metallic mercury (whereby the disagreeable odor of the commercial article is removed), and then rectifying over the water-bath. In con- ducting this latter operation the use of all rubber tubing must be avoided. Carbon disulphide serves for removing free iodine from aqueous solutions, and for freeing sulphides of metals from ad- mixed sulphur. II. ACIDS AND HALOGENS. a. Oxygen Acids. 57. 1. SULPHURIC ACID. a. Concentrated sulphuric acid (commercial). 1). Concentrated pure sulphuric acid. c. Dilute sulphuric acid. See "Qual. Anal." 2. NITRIC ACID. a. Pure nitric acid of 1*2 sp. gr. (see " Qual. Anal."). 5. Red fuming nitric acid (concentrated nitric acid containing some hyponitric acid). Preparation. Mix 1000 grm. of pure potassium nitrate with 15 grm. starch in lumps, place the mixture in a capacious tubu- lated retort, and add 500 grm. sulphuric acid and 500 grm. fuming sulphuric acid. The retort is then placed on a wire 58.] TCEAGENTS. 129 gauze over a gas-oven, or in a sand-bath. The distillation will begin without the application of heat. If the potassium nitrate is not perfectly free from metallic chlorides, the first portion of the distillate should be collected separately and set aside. When the distillation slackens gentle heat is applied, taking care not to push the distillation too rapidly. The process is complete, when, by the application of moderate heat, no more acid distills over. As it is impossible to prevent a portion of the hyponitric acid from escaping, the process should be conducted in the open air, or under a good vapor hood. Tests. Red fuming nitric acid must be in a state of the greatest possible concentration, and perfectly free from sulphuric acid. In order to detect minute traces of the latter, evaporate a few c. c. of the specimen in a porcelain dish nearly to dryness, dilute the resi- due with water, add some barium chloride, and observe whether a precipitate forms on standing. Uses. A powerful oxidizing agent and solvent ; it serves more especially to convert sulphur and metallic sulphides into sulphuric .acid and sulphates respectively. 3. ACETIC ACID (see " Qual. Anal."). 4. TARTARIC ACID (see " Qual. Anal."). b. Hydrogen Acids and Hologens. 58. 1. HYDROCHLORIC ACID. a. Pure hydrochloric acid of 1-12 sp. gr. (see "Qual. Anal.").* b. Pure fuming hydrochloric acid of about 1*18 sp. gr. Preparation. As in " Qual. Anal." 26, with this modifica- tion, however, that only 3 or 4 parts of water, instead of 6, are put into the receiver, to 4 parts of sodium chloride in the retort. The greatest care must be taken to keep the receiver cool, and to change it as soon as the tube through which the gas is conducted into it begins to get hot, since it is now no longer hydrochloric acid gas which passes over, but an aqueous solution of the gas, in form of * For BETTENDOKFF'S process for the preparation of arsenic-free hydro- chloric acid, and based upon the precipitation of arsenic by stannous chloride, see Zeitschr.f. analyt. Chem., ix, 107. 130 REAGENTS. [ 58. vapor, which would simply weaken the fuming acid, if it were allowed to mix with it. Tests. The fuming acid must, for many purposes, be perfectly free from chlorine and sulphurous acid. For the mode of testing for these impurities, see u Qual Anal." loc. cit. Test for sulphuric acid as under Nitric Acid, above. Uses. Fuming hydrochloric acid has a much more energetic action than the dilute acid; it is, therefore, used instead of the latter in cases where a more rapid and energetic action is desirable. 2. HYDROFLUORIC ACID. This is employed for the decomposition of silicates and borates, sometimes in the gaseous form, sometimes in the condition of aqueous solution. In the first case, the substance to be decomposed is introduced into the leaden box, in which the hydrofluoric gas is being generated ; in the latter case, we must first prepare the aque- ous acid. The raw material employed is fluor spar or kryolite (LuBOLDT*). Both are first finely powdered, and then treated with concentrated sulphuric acid. To 1 part kryolite, 2 parts sulphuric acid are used ; to 1 part fluor spar, 2 parts sulphuric acid are used. If the latter is employed, allow the mixture to stand in a dry place for several days, stirring every now and then, so that the silicic acid (which is generally contained in fluor spar) may first escape in the form of fluosilicic gas. Convenient distil- latory apparatus have been described by LUBOLDT (loc. cit.) and by H. BEiEGLEB.f The latter commends itself especially on account of its relatively small cost. It consists of a leaden retort, with a movable leaden top, which can be luted on. The receiver belong- ing to it is a box of lead, with a tubulure at the side, into which the neck of the retort just enters. The cover of the receiver is raised conical, and is provided at the top with an exit tube of le?.d. In the receiver a platinum dish containing water is placed, all joints are luted, and the retort is carefully heated in a sand-bath. The aqueous hydrofluoric acid found" at the end of the operation in the platinum dish is perfectly pure. The small quantity of impure hydrofluoric acid which collects on the bottom of the receiver is thrown away. The hydrofluoric acid must entirely volatilize when heated in a platinum dish on a water-bath. The pure acid gives no * Jour, fur prakt. Chem., LXXVI, 330. f Annal. d. Chem. u. Pharm , cxi, 380. 59.] REAGENTS. 131 precipitate when neutralized with potash, while potassium silico- fluoride separates if the acid contains hydrofluosilicic acid. The acid is best preserved in gutta-percha bottles, as recommended by STADELER. The acid is now obtainable in the market in gutta- percha bottles. It should at once be tested ; this must never be neglected, as I have often found the acid to be impure. The greatest caution must be observed in preparing this acid, since, whether in the fluid or gaseous condition, it is one of the most injurious substances. 3. CHLORINE AND CHLORINE-WATER (see u Qual. Anal."). 4. NITRO- HYDROCHLORIC ACID (see " Qual. Anal."). 5. HTDROFLUOSILICIC ACID (see " Qual. Anal."). This should be kept in gutta-percha bottles, as when long kept in glass it attacks the latter and takes up some of its constitu- ents. c. Sulphur Acids. 1. HYDROSULPHURIC ACID (see " Qual. Anal."). III. BASES AND METALS. a. Oxygen Bases and Metals. 59. a. Alkali Eases. 1. POTASSIUM HYDROXIDE OR POTASSA, AND SODIUM HYDROXIDE OR SODA (see " Qual. Anal."). All the four sorts of the caustic alkalies mentioned in the quali- tative part are required in quantitative analysis, viz., common solu- tion of soda, potassa purified with alcohol, solution of potassa pre- pared with baryta, and absolutely pure soda. Pure solution of potassa may be obtained also by heating to redness for half an hour in a copper crucible, a mixture of 1 part of potassium nitrate, and 2 or 3 parts of thin sheet copper cut into small pieces, treating the mass with water, allowing the oxide of copper to subside in a tall vessel, and removing the supernatant clear fluid by means of a syphon (WOHLER). 2. AMMONIA (see " Qual. Anal."). 132 KEAGENTS. [ 60. fi. Alkali-earth Bases. 1. BARIUM HYDROXIDE, OR BARYTA (see " Qual. Anal."). 2. CALCIUM HYDROXIDE, OR LIME. Finely divided calcium hydroxide mixed with water (milk of lime), is used more particularly to effect the separation of magne- sium, etc., from the alkali metals. Milk of lime intended to be used for that purpose must, of course, be perfectly free from alka- lies. To insure this the purest lime (calcined white marble) should be used, and slaked lime should be thoroughly washed by repeated boiling with fresh quantities of distilled water. This operation is best conducted in a silver dish. When cold, the inilk of lime so prepared is kept in a well-stoppered bottle. y. Heavy Metals, and their Oxides. 60. 1. ZINC. Zinc has of late been much used as a reagent in quantitative analysis. It serves more especially to effect the reduction of ferric to ferrous salts, and also the precipitation of copper from solutions of its salts. Zinc intended to be used for the former purpose must be free from iron, for the latter free from lead, copper, and other metals which remain undissolved upon treating the zinc with dilute acids. As it is not easy to prepare in quantity zinc that will answer both purposes, it is advisable to keep on hand, besides the ordinary zinc used for preparing hydrogen, the two following kinds also : a. Zinc, free from Iron. The distillation of zinc in the labo- ratory is a tedious and costly operation, hence as a rule the raw product obtained by distillation from the ore is used in the prepara- tion of the iron-free zinc. This product contains, at least in many cases, only such slight traces of iron, that it may be safely used for the reduction of ferrous salts in solution. Ordinary commercial zinc contains much more iron, from having been fused in iron vessels. h. Zinc, free from Lead, Copper, etc. To procure zinc which leaves no residue upon solution in dilute sulphuric acid, there is commonly no other resource but to re-distil the commercial article. This is effected in a retort made of the material of Hessian or 60.] 1JEAGENTS. 133 black-lead crucibles. The operation is conducted in a wind-furnace with good draught. The neck of the retort must hang down as perpendicularly as possible. Over this is placed a small clay drain- pipe, the lower end of which dips into water contained in a tub or large stone- ware dish. The joints are all stopped with clay. Under the neck is placed a basin or small tub, filled with water. The distillation begins as soon as the retort is at a bright red heat. As the neck of the retort is very liable to become choked up with zinc or oxide of zinc, it is necessary to keep it constantly free by means of a pipe-stern. The zinc obtained by this re-distillation is nearly or quite free from lead, but still contains notable traces of iron (from the wire). If the presence of iron is to be totally avoided, a clay pipe-stem or stick of wood must be used instead of the iron wire. Tests. The following is the simplest way of testing the purity of zinc : Dissolve the metal in dilute sulphuric acid in a small flask provided with a gas-evolution tube, place the outer limb of the tube under water, and when the solution is completed, let the water entirely or partly recede into the flask; after cooling, add to the fluid, drop by drop, a sufficiently dilute solution of potassium per- manganate. If a drop of that solution imparts the same red tint to the zinc solution as to an equal volume of water, the zinc may be considered free from iron. I prefer this way of testing the purity of zinc to other methods, as it affords, at the same time, an approx- imate, or, if the zinc has been weighed and the permanganate solu- tion (which, in that case, must be considerably diluted) measured, an accurate and precise knowledge of the quantity of iron present. If lead or copper is present, the metal remains undissolved upon solution of the zinc. 2. COPPEE. Preparation. Commercial copper, with the exception of the Japanese, which is not always obtainable, is seldom sufficiently pure for analytical purposes. Hence the pure metal should be prepared by the chemist, either by the galvanoplastic process, or by the method of FCCHS, in which copper-sulphate solution is precipitated by well-cleaned iron, the precipitate of copper boiled with hydro- chloric acid to remove iron, then washed, dried, and fused, and the regulus so obtained rolled out into thin sheets. Tests. Pure copper must be completely soluble in nitric acid, 134 KEAGENTS. [ 60. and the solution must afford no precipitate with excess of ammonia, even on long standing (iron, lead, etc.) ; nor should it be rendered turbid by hydrochloric acid (silver). After precipitation with hydrogen sulphide, the filtrate should leave no residue on evapo- ration. Uses. Copper serves occasionally in indirect analysis ; for in- stance, in estimating the copper content of a liquid, for estimating iron according to FUCHS, etc. Since the development and use of volumetric analysis, however, it is but seldom used in quantitative analysis. 2. LEAD OXIDE. Precipitate pure lead nitrate or acetate with ammonium car- bonate, wash the precipitate, dry, and ignite gently to complete decomposition . Lead oxide is often used to fix an acid, so that it is not expelled even by a read heat. 3. MEKCUKIC OXIDE. Preparation. Add a solution of mercuric chloride to a hot, moderately dilute caustic-soda solution, taking care that the soda be always in excess. The yellow precipitate is thoroughly washed by decantation, then mixed with water, and preserved in this con- dition in a bottle. Test.- Mercuric oxide must leave no residue on ignition in a platinum crucible. Uses. This reagent serves in quantitative analysis for decom- posing magnesium chloride in the process of separating magnesia from alkalies. J. Sulphur Bases. 1. AMMONIUM SULPHIDE (see " Qual. Anal."). "We require both the colorless monosulphide, and the yellow polysulphide. 2. SODIUM SULPHIDE (see "Qual. Anal."). 61.] KEAGENTS. 135 IV. SALTS. a. Salts of the Alkalies. 61. 1. POTASSIUM SULPHATE (see "Qual. Anal."). 2. AMMONIUM PHOSPHATE. Preparation. Dilute phosphoric acid (sp. gr. 1*13), prepared from phosphorus, is mixed with an equal quantity of water, and pure ammonia water added until the liquid has a strongly alkaline reaction, when it is set aside for some time, then filtered if neces- sary, and kept for use. Tests. Ammonium phosphate must he free from arsenic, nitric, and sulphuric acids, but more particularly from potassa or soda. To test it for these two last, add lead-acetate solution so long as a precipitate still forms, then filter, precipitate the lead excess with hydrogen sulphide, filter again, evaporate the filtrate to dryness, and ignite the residue. If an alkaline residue is left, potassa or soda was present. In most cases sodium phosphate (see " Qual. Anal.") may be used instead of ammonium phosphate, 3. AMMONIUM OXALATE (see u Qual Anal."). 4. SODIUM ACETATE (see ".Qual. Anal."). 5. AMMONIUM SUCCINATE. Preparation. Saturate succinic acid, which has been purified by dissolving in nitric acid and recrystallizing, with dilute ammo- nia. The reaction of the new compound should be rather slightly alkaline than acid. Uses. This reagent serves occasionally to separate ferric iron from other metals. 6. SODIUM CARBONATE (see "Qual. Anal."). This reagent is required both in solution, and in pure crystals ; in the latter form to neutralize an excess of acid in a fluid which it is desirable not to dilute too much. 7. AMMONIUM CARBONATE (see " Qual. Anal."). 8. SODIUM HYDROGEN SULPHITE (see " Qual. Anal."). 9. SODIUM THIOSULPHATK (HYPOSULPHITE) N a S 2 O 3 . This salt occurs in commerce. It should be dry, clear, well 136 REAGENTS. [ 61. crystallized, and completely and easily soluble in water. The solu- tion must give with silver nitrate at first a white precipitate, must not effervesce with acetic acid, and when acidified must give no precipitate with barium chloride, or, at most, only a slight turbidity. The acidified solution must, after a short time, become milky from separation of sulphur. Uses. Sodium thiosulphate is used for the precipitation of several metals, as sulphides, particularly in separations, for instance, of copper from zinc ; it also serves as solvent for several salts (sil- ver chloride, calcium sulphate, etc.) ; lastly, it is employed in volu- metric analysis, its use here depending on the reaction 2(Na,S,O t ) + 21 = 2NaI + Na s S 4 O a . 10. POTASSIUM NITRITE (see " Qnal. Anal."). 11. POTASSIUM BICHROMATE (see "Qual. Anal."). 12. AMMONIUM MOLYBDATE (see "Qual. Anal."). When using the solution of ammonium molybdate in nitric acid for the estimation of phosphoric acid, the filtrates from the ammo- nium phospho-molybdate and magnesium-ammonium phosphate will contain all the molybdic acid. If the filtrates are preserved, therefore, there will be no loss, and the acid may be recovered as follows : Evaporate the residue to dryness in the open air or under a good draught, and heat finally until most of the ammonium nitrate has been decomposed. Digest the residue with ammonia, which dissolves the molybdic acid, and filter. To the filtrate add a ]ittle magnesia mixture ( 62, 6) in order to precipitate any phos- phoric acid present. If a precipitate occurs, add sufficient mag- nesia mixture to assure complete precipitation of all phosphoric acid. After allowing to stand for some time, filter, acidulate the filtrate with nitric acid, and then filter off the precipitated molybdic acid, using suction, and wash it with the smallest quantity of water. The acid is then available for use again in solution. The filtrate and washings from the acid will contain but little acid ; they may be worked up .with the next residues treated. 13. AMMONIUM CHLORIDE (see "Qual. Anal."). 14. POTASSIUM CYANIDE (see "Qual. Anal."). Preparation. Besides the potassium cyanide prepared accord- ing to LIEBIG, and which contains potassium cyanate and carbonate, there is required also a pure cyanide for use in certain separations, e.g., as in WOHLER'S method of separating nickel from zinc. 62.] REAGENTS. 137 The pure cyanide is prepared as follows: 2 parts of crystallized potassium ferrocyanide are powdered and then transferred to a retort, wherein it is heated together with 1J- parts concentrated sulphuric acid and 4 parts water, until the residue begins to bump. The vapors of hydrocyanic acid are conducted into a cooled receiver containing a freshly prepared and filtered solution of 1 part caustic potassa (not fused, but evaporated until it just solidi- fies on cooling) in 3 to 4 parts of not less than 92-per-cent alcohol. The caustic potassa should be present in slight excess at the close of the operation. The crystalline mass is filtered by the aid of suction, then washed with a little alcohol, then dried in a porcelain dish by the aid of heat, and finally preserved for use in a well- closed bottle. b. Salts of the Alkali-earth Metals. 1. BARIUM CHLORIDE (see " Qual. Anal."). The following process gives a very pure barium chloride, free from calcium and strontium : Transmit through a concentrated solution of impure barium chloride hydrochloric gas, as long as a precipitate continues to form. Nearly the whole of the barium chloride present is by this means separated from the solution, in form of a crystalline powder. Collect this on a filter, let the adhering liquid drain off, wash the powder repeatedly with small quantities of pure hydrochloric acid, until a sample of the wash- ings, diluted with water, and precipitated with sulphuric acid, gives a filtrate which, upon evaporation in a platinum dish, leaves no residue. The hydrochloric mother-liquor serves to dissolve fresh portions of witherite. I make use of the barium chloride so obtained, principally for the preparation of perfectly pure barium carbonate, which is often required in quantitative analyses. 2. BARIUM ACETATE. Preparation. Dissolve pure barium carbonate in moderately dilute acetic acid, filter, and evaporate to crystallization. Tests. Dilute solution of barium acetate must not be rendered turbid by solution of silver nitrate. See also " Qual. Anal.," Barium chloride, the same tests being also used to ascertain the purity of the acetate. Uses. Barium acetate is used instead of barium chloride, to effect the precipitation of sulphuric acid, in cases where it is desir- 138 KEAGENTS. [ 62. able to avoid the introduction of a chloride into the solution, or to convert the base into an acetate. As the reagent is seldom required, it is best kept in crystals. 3. BARIUM CARBONATE (see "Qual. Anal."). 4. STRONTIUM CHLORIDE. Preparation. Strontium chloride is prepared from strontian- ite or celestine, by the same processes as barium chloride. The pure crystals obtained are dissolved in alcohol of 96 per cent., the solution is filtered, and kept for use. Uses. The alcoholic solution of strontium chloride is used to effect the conversion of alkali sulphates into chlorides, in cases where it is desirable to avoid the introduction into the fluid of a salt insoluble in alcohol. 5. CALCIUM CHLORIDE (see " Qual. Anal."). 6. MAGNESIUM CHLORIDE, MAGNESIUM SULPHATE, OR MAGNESIA MIXTURE. Dissolve 1 1 parts crystallized magnesium chloride (MgCl 2 -f- 6 H 2 O) and 28 parts ammonium chloride in 130 parts water, add 70 parts dilute ammonia solution (sp. gr. 0*96). Allow the mix- ture to stand one or two days and filter. Tins solution, commonly called " magnesia mixture," is used to precipitate phosphoric acid, and also arsenic acid, from aqueous solutions. An excess is required to effect complete precipitation. Prepared as here de- scribed, about 10 c. c. should be used in ordinary cases for every O'l gramme P 3 O 6 . A solution containing the same per cent, (approximately) of magnesium chloride and other constituents may also be prepared from common calcined magnesia (MgO), provided it is free from the other alkali-earth metals, as follows: Add to 11 parts magnesia sufficient hydrochloric acid to effect solution, next add a slight ex- cess of magnesia and boil to separate traces of iron ; filter, and add 140 parts ammonium chloride and 350 parts dilute ammonia. Dilute with water until volume equals 1000 c. c. for every 11 grammes of MgO used. Allow the mixture to stand two or three days, and filter if necessary. Magnesia mixture may be also made as follows: Dissolve 1 part of crystallized magnesium sulphate and 2 parts pure ammonium chloride in 8 parts water and add 4 parts ammonia water. Let stand several days, then filter. 63.] REAGENTS. 139 c. Salts of the Heavy Metals. 63. 1. FERROUS SULPHATE (see "Qual. Anal."). 2. FERRIC CHLORIDE (see " Qual. Anal."). 3. URANIO ACETATE. Heat finely powdered pitchblende with dilute nitric acid, filter the fluid from the undissolved portion, and treat the filtrate with hydrosulphuric acid to remove the lead, copper, and arsenic ; filter again, evaporate to dryness, extract the residue with water, and fil- ter the solution from the oxides of iron, cobalt, and manganese. Uranic nitrate crystallizes from the filtrate ; purify this by recrys- tallization, and then heat the crystals until a small portion of uranic oxide is reduced. Warm the yellowish-red mass thus obtained with acetic acid, filter arid let the filtrate crystallize. The crystals are uranic acetate, and the mother-liquor contains the undecom- posed nitrate (WERTIIEIM). The salt may be more conveniently made from the commer- cially obtainable sodium uranate (manufactured by the K. K. Bergoberamt, Joachim sthal). Digest 1 part of this salt in 2 parts acetic acid (sp. gr. 1*038), then add 25 parts water, heat, filter, evaporate, and allow to crystallize. The uranic oxide in the last mother-liquors (containing also sodium acetate) is precipitated by ammonia. Uranium being a costly metal, all the residues should be saved, and worked up as follows : The liquid is poured off fr^gn. any sediment of uranium phosphate. All the uranium in it is then precipitated by adding sodium phosphate. The precipitate is washed by decantation, mixed with the uranium phosphate reserved, the whole dissolved in hydrochloric acid, and ferric chloride added until a sample gives a brownish precipitate with ammonium carbonate. The mixture is then diluted, and to the solution, which must contain a sufficient excess of hydrochloric acid, add a solution of crystallized sodium carbonate in excess. All the phosphoric acid is thus precipitated as basic ferric phos- phate; the uranium oxide, however, remains dissolved in the solution of sodium bicarbonate formed. Filter the mixture, wash, acidulate the filtrate with hydrochloric acid, warm until the car- bon dioxide is completely expelled, warm and precipitate the 140 REAGENTS. [ 64. uranium oxide with ammonia. After washing, dissolve in acetic acid (E. REICHARD).* Tests. Solution of uranic acetate after acidification with hydrochloric acid must not be altered by hydrosulphuric acid ; ammonium carbonate must produce in it a precipitate, soluble in an excess of the precipitant. A sample of the dilute solution should acquire a red tint on adding a little sulphuric acid and a drop of potassium permanganate solution (absence of uranous salt). Uses. Uranic acetate may serve, in many cases, to effect the separation and determination of phosphoric acid. 4. SILVER NITRATE (see " Qual. Anal."). 5. LEAD ACETATE (see " Qual. Anal."). 6. MERCURIC CHLORIDE (see " Qual. Anal."). 7. STANNOUS CHLORIDE (see " Qua!. Anal."). 8. PLATINIC CHLORIDE (see " Qual. Anal."). It is convenient to know approximately the strength of this solution. I usually use a solution 10 or 20 c. c. of which contain 1 grm. platinum. 9. SODIUM PALLADIO-CHLORIDE (see " Qual. Anal."). R RE A GENTS FOR GRA VIMETRIC ANAL T8I8 IN THE DR T WAY. 64. 1. SODIUM CARBONATE, pure anhydrous (see " Qual. Anal."). 2. MIXED SODIUM AND POTASSIUM CARBONATES (see " Qual. Anal."). 3. BARIUM HYDROXIDE OR BARYTA (see " Qual. Anal." and 59). 4. POTASSIUM NITRATE (see " Qual. Anal."). 5. SODIUM NITRATE (see " Qual. Anal."). 6. BORAX (fused). Preparation. Heat crystallized borax (see " Qual. Anal.) in a platinum or porcelain dish, until there is no further intumescence ; reduce the porous mass to powder, and heat this in a platinum cru- cible until it is fused to a transparent mass. Pour the semi-fluid, * Zeitschr.f. analyt. Chem., vin, 116. 64.] REAGENTS. 141 viscid mass upon a fragment of porcelain. A better way is to fuse the borax in a net- of platinum gauze, by making the gas blowpipe- flame act upon it. The drops are collected in a platinum dish. The vitrified borax obtained is kept in a well -stoppered bottle. But as it is always necessary to heat the vitrified borax previous to use, to make quite sure that it is perfectly anhydrous, the best way is to prepare it only when required. Uses. Vitrified borax is used to effect the expulsion of car- bonic acid and other volatile acids, at a red heat. 7. POTASSIUM DISULPHATE. Preparation. Mix 87 parts of normal potassium sulphate (see " Qual. Anal."), in a platinum crucible, with 49 parts of concen- trated pure sulphuric acid, and heat to gentle redness until the mass is in a state of uniform and limpid fusion. Pour the fused salt on a fragment of porcelain, or into a platinum dish standing in cold water. After cooling, break the mass into pieces, and keep for use. Uses. This reagent serves as a flux for certain native com- pounds of alumina and chromic oxide. Potassium disulphate is used also, as we have already had occasion to state, for the cleansing of platinum crucibles ; for this latter purpose, however, the salt which is obtained in the preparation of nitric acid will be found sufficiently pure. 8. SODIUM DISULPHATE. Preparation. This is prepared like the potassium salt, usiu^ 71 parts of pure, normal sodium sulphate and 49 parts of concen- trated pure sulphuric acid. Uses. Sodium disi^phate is used like the potassium disul- phate, but is to be substituted for the latter when, as in fusing corundum, the analysis is hampered by alum crystallizing out {L. SMITH, Zeitsohr. f. analyt. Chem., iv, 412). 9. HYDROGEN-POTASSIUM FLUORIDE. Neutralize a definite quantity of hydrofluoric acid in a plati- num dish with pure potassium carbonate or potassium hydroxide, applying heat toward the last; then add a quantity of hydrofluoric acid equal to that first taken, and evaporate the whole to dryness. The preparation is usually made just before required ; if it is to be preserved, gutta-percha vessels must be used. 142 KEAGENTS. [ 64. Tests. Solution of hydrogen -potassium fluoride must remain unaffected on adding hydrogen sulphide, ammonia, ammonia and ammonium sulphide, or ammonium carbonate and sodium phosphate with addition of ammonia. 'Uses. This preparation is an excellent flux for many minerals which are usually very refractory, e.g., tinstone, chrome iron (GIBBS, Zeitschr.f. analyt. Chem., m, 399). 10. HYDROGEN- AMMONIUM FLUORIDE. Preparation. Add ammonia in considerable excess to hydro- fluoric acid or hydrosilicofluoric acid in a platinum dish, gently heat for some time, filter if necessary, and evaporate the filtrate to dryness in a platinum dish. Half the ammonia escapes, while hy- drogen-ammonium fluoride remains behind. If this is to be pre- served, a gutta-percha vessel must be used. Tests. Like those of hydrogen-potassium fluoride. In addi- tion, a sample heated in a platinum dish (in the open or under a good vapor hood) must leave no fixed residue. Uses. The preparation may be advantageously used instead of hydrofluoric acid in the analysis of silicates. 11. AMMONIUM CARBONATE (solid). , Preparation. See " Qual. Anal." This reagent serves to convert the acid alkali sulphates into normal salts. It must com- pletely -volatilize when heated in a platinum dish. 12. AMMONIUM TITRATE. Preparation. Neutralize pure ammonium carbonate with pure nitric acid, warm, and add ammonia to slightly alkaline reaction; filter, if necessary, and let the filtrate crystallize. Fuse the crys- tals in a platinum dish, and pour the fused mass upon a piece of porcelain ; break into pieces whilst still warm, and keep in a well- stoppered bottle. Tests. Ammonium nitrate must leave no residue when heated in a platinum dish. Uses. Ammonium nitrate serves as an oxidizing agent; for instance, to convert lead into lead oxide, or to effect the com- bustion of carbon, in cases where it is desired to avoid the use of fixed salts. 84. J REAGENTS. 143 13. AMMONIUM CHLORIDE. Preparation and Tests. See " Qual. Anal." Uses. Ammonium chloride is often used to convert metallic oxides and adds, e.g., lead oxide, zinc oxide, stannic oxide, arsenic acid, antimonic acid, &c., into chlorides (ammonia and water escape in the process). Many metallic chlorides being volatile, and others volatilizing in presence of ammonium chloride fumes, they may be completely removed by igniting them with ammonium chloride in- excess, and thus many compounds, e.g., alkali antimonates, may be easily arid expeditiously analyzed. Ammonium chloride is also used to convert various salts of other acids into chlorides, e.g., small quantities of alkali sulphates. 14. HYDROGEN GAS. Preparation. Hydrogen gas is evolved when dilute sulphuric acid is added to granulated zinc. It may be purified from traces of foreign gases either by passing first through mercuric-chloride solution, then through potassa solution, or as recommended by STENHOUSE, by passing through a tube filled with pieces of char- coal. If the gas is desired dry, pass through sulphuric acid or a calcium-chloride tube. If the zinc used is new, the evolution of gas may be facilitated by adding a drop of platinic-chloride solution. Tests. Pure hydrogen gas is inodorous. It ought to burn with a colorless flame, which, when cooled by depressing a porcelain dish upon it, must deposit nothing on the surface of the dish except pure water (free from acid reaction). Uses. Hydrogen gas is frequently used, in quantitative analy- sis, to reduce oxides, chlorides, sulphides, &c., to the metallic state ; also to protect certain substances, like metallic sulphides, from the action of atmospheric oxygen during ignition. 12. CHLORINE. Preparation. See "Qua!. Anal." Chlorine gas is purified and dried by transmitting it through a U-tube containing frag- ments of manganese dioxide, then concentrated sulphuric acid, or a calcium-chloride tube. Uses. Chlorine gas serves principally to produce chlorides, and to separate the volatile from the non-volatile chlorides; it is also used to displace and indirectly determine bromine and iodine, as well as to convert lower chlorine compounds into higher. 144 REAGENTS. [ 65. C. REAGENTS USED IN VOLUMETRIC ANALYSIS. 65. Under this head are arranged the most important of those substances which serve for the preparation and testing of the fluids required in volumetric analysis and have not been given under A and B. 1 . PURE CRYSTALLIZED OXALIC ACID, H a C a 4 -f- 2H 2 0. The introduction of crystallized oxalic acid as a basis for alkal- imetry and acidimetry is due to FR. MOHR. It is also employed to determine the strength of, or to standardize, a solution of potas- sium permanganate, 1 molecule of potassium permanganate being required, in the presence of free sulphuric acid, to convert 5 mole- cules of oxalic acid into carbon dioxide and water (K 2 Mn 2 O 8 -f- 5H.CA + 3H,S0 4 = K 2 S0 4 + 2MnSO 4 + 8H.O + lOCO',). "We use in most cases the pure crystallized acid which has the formula H a C a O 4 + 2H a O, and of which the molecular weight is accordingly 126*048. Preparation. The pure acid is prepared by shaking powdered commercial oxalic acid with a small quantity of warm water (so that considerable of the acid remains undissolved), then filtering, and crystallizing by rapidly cooling (MOHR). Spread the crystals, after draining, on blotting paper, and set aside to dry in a dust- free place at the ordinary temperature (not too high), or press them gently between renewed layers of blotting paper until the latter take up no more moisture. Tests. The crystals of oxalic acid must not show the least sign of efflorescence (to which they are liable even at 20 in a dry atmosphere) ; they must dissolve in water to a perfectly clear fluid ; when heated in a platinum dish, they must leave no fixed and incombustible residue (calcium carbonate, potassium carbonate, &c.). If the acid obtained by a first crystallization fails to satisfy these requirements, it must be recrystallized. In this case the strength of the solution must be such that only 10 or 20 per cent of the dissolved oxalic acid crystallizes out, and this, containing the impurities, is then removed, after which the mother-liquor is con- centrated by evaporation. The crop of crystals next obtained is purer. 65.] REAGENTS. 145 2. TINCTURE OF LITMUS. Preparation. Digest 1 part of litmus of commerce witli 6 parts of water on the water-bath for some time, filter, divide the blue fluid into 2 portions, and saturate in one half the free alkali, by stirring repeatedly with a glass rod dipped in very dilute nitric acid, until the color just appears red ; add the remaining blue half, together with 1 part of strong spirit of wine, and keep the tincture which is now ready for use, in a small open bottle, not quite full, in a place protected from dust. In a stoppered bottle the tincture would speedily lose color. Tests. Litmus tincture is tested by coloring with it about 100 cubic centimetres of water distinctly blue, dividing the fluid into two portions, and adding to the one the least quantity of a dilute acid, to the other a trace of solution of soda. If the one portion acquires a distinct red, the other a distinct blue tint, the litmus tincture is fit for use, as neither acid nor alkali predominates. 3. POTASSIUM PERMANGANATE. Preparation. Mix 8 parts of very finely powdered pure pyro- lusite, or manganese dioxide, with 7 parts of potassium chlorate, put the mixture into a shallow cast-iron pot, and add 37 parts of a solution of potassa of 1/27 specific gravity (the same solution as is used in organic analysis *) ; evaporate to dry ness, stirring the mixture during the operation ; put the residue before it has ab- sorbed moisture, into an iron or Hessian crucible, and expose to a dull-red heat, with frequent stirring with an iron rod or iron spa- tula, until no more aqueous vapors escape and the mass is in a faint glow. Remove the crucible now from the fire, and transfer the friable mass to an iron pot. Reduce to coarse powder, and transfer this, in small portions at a time, to an iron vessel containing 100 parts of boiling water; keep boiling, replacing the evaporating w r ater, and passing a stream of carbon dioxide through the- fluid ( MULDER f). The originally dark-green solution of potassium manganate soon changes, with separation of hydrated manganese dioxide, to the deep violet-red of the permanganate. When it is considered that the conversion is complete, allow to settle, take out a small quantity of the clear liquid, boil and pass carbon dioxide through it. If a precipitate forms, the conversion is not yet complete. * Or, instead of the solution, use 10 parts of the hydroxide KOH. In this case fuse the potash and the chlorate together first, and then project the man- ganese into the crucible. \ Jahresbericht von KOPP uud WILL. 1858. 581. 146 REAGENTS. [ 65. The following method, recommended by STAEDELER,* is more rapid : Powder the fused mass and macerate it with an equal weight of cold water in a flask ; then add an equal quantity of water, and conduct chlorine into the mixture until the latter has lost its green color and become a pure red. Then dilute with four times its vol- ume of water and allow to settle. By this method the yield of potassium permanganate is increased one-half, because no manga- nese dioxide is precipitated. On the other hand, if the fused mass contains potassium hydroxide in excess, potassium chlorate may form, and which will render somewhat difficult the purifica- tion of the permanganate crystals. The clear, red solution, obtained by any one of the above methods, is decanted from the precipitate, the latter washed by decantation, and the united liquids evaporated over an open fire to the crystallizing point, and then allowed to cool. The mother- liquor, on evaporation, yields a new crop of crystals. The last mother-liquor contains much potassium chloride, hence it can only be employed for the preparation of manganese dioxide. If the crystals obtained are not sufficiently pure, they may be readily purified by recrystallization. They are freed from adhering in other-liquor by exposure on a block of plaster of Paris. The solution of potassium permanganate may be filtered, if necessary, through gun-cotton, asbestos or previously ignited sand. 4. AMMONIUM FERROUS SULPHATE. FR. MOHR has proposed to employ this double salt, which is not liable to efflorescence and oxidation, as an agent to determine the strength of the permanganate solution. Preparation. Take two equal portions of dilute sulphuric acid, ,and warm the one with a moderate excess of small iron nails free from rust, until the evolution of hydrogen gas has alto- gether or very nearly ceased ; neutralize the other portion exactly with ammonium carbonate, and then add to it a few drops of dihue sulphuric acid. Filter the solution of the ferrous sulphate into that of the ammonium sulphate, evaporate the mixture a little, if neces- sary, and then allow the salt to crystallize. Let the crystals, which are hard and of a pale-green color, drain in a funnel, then wash * Journ. f. prakt. Chem., cur, 107. 65.] REAGENTS. 147 them in a little water, dry thoroughly on blotting-paper in the air, and keep for use. The molecular weight of the salt (892 '376) is almost exactly 7 times the atomic weight of iron (55 -9). The solution of the salt in water which has been just acidified with sulphuric acid must not become red on the addition of potassium sulphocyanate. 5. AMMONIA- IRON-ALUM. Preparation. Bring into a large porcelain dish 58 grins, of pure crystallized ferrous sulphate (see Fresenius' "Qual. Anal." Am. ed., p. 73), together with a quantity of sulphuric acid equiva- lent to 8'3 grms. of sulphuric anhydride (SO,), (see Table, 191). Heat upon a sand-bath, adding nitric acid from time to time, in small portions, until the iron has all passed into ferric sulphate, or until a drop of the solution gives no blue coloration with potassium ferricyamde. Heat further, and evaporate until the excess of nitric acid is expelled, then add 14 grms. of ammonium sulphate,* and, if need be, hot water sufficient to bring the salt into solution; iilter into a porcelain capsule and set aside, under cover, to crys- tallize. The iron-alum separates in cubo-octahedrons, wh\ch may be yel- lowish, lilac, or colorless. If dark in color, dissolve in w r arm water, add a few drops of oil of vitriol, and crystallize again. Rinse the pale or colorless crystals, after separation from the mother-liquor, with cold water, wrap up closely in iilter paper, and allow them to dry at the ordinary temperature.f * If not on hand, this salt may be prepared by saturating sulphuric acid with ammonium carbonate and evaporating to dryness. 30 grammes of sul- phuric acid give somewhat more than is required above. t Examinations of iron-alum thus prepared show that the variations in the color of the salt, from colorless to rose, are not connected with appreciable differences of composition. J. H. Grove, of the Sheffield Laboratory, obtained the following results in the examination of ammonia-iron-alum crystals, the ferric oxide being estimated by ignition : Fe 2 O 3 ' ( 16-59 1st j < 16'55 \ 16-59 2d 16-53 3d 16-57 4th 16-57 5th 16-58 6th j 16-50 16-56 7th 1655 Calculated 16-60 148 KEAGENTS. [ 65. The yield should be about 80 grms. The dry salt should be pulverized, pressed between folds of paper until freed from mechanically adhering water, and preserved in a w r ell-stoppered bottle. Uses. Ammonia-iron-alum furnishes the best means of obtain- ing a definite quantity of iron in a ferric salt for making standard solutions, being easily obtained pure and inalterable if kept away from acid vapors. Its purity may be readily controlled by ascertaining the loss on careful ignition, which should leave a resi- due of 16 '6 per cent, of ferric oxide, corresponding to 11 '62 per cent, of metallic iron. 6. PURE IODINE. Preparation. Triturate iodine of commerce with -J- part of its weight of potassium iodide, dry the mass in a large watch-glass with ground rim, warm this gently on a sand-bath, or on an iron plate, and as soon as violet fumes begin to escape, cover it with another watch-glass of the same size. Continue the application of heat until all the iodine is sublimed, and keep in a well-closed glass bottle. The chlorine or bromine, which is often found in iodine of commerce, combines, in this process, with the potassium, and remains in the lower watch-glass, together with the excess of potassium iodide. Tests. Iodine purified by the process just now described, must leave no fixed residue when heated on a watch-glass. But, even supposing it should leave a trace on the glass, it would be of no great consequence, as the small portion intended for use has to be resublimed immediately before weighing. Uses. Pure iodine is used to determine the amount of iodine contained in the solution of iodine in potassium iodide, employed in many volumetric processes. 7. POTASSIUM IODIDE. Small quantities of this article may be procured cheaper in commerce than prepared in the laboratory. For the preparation of potassium iodide intended for analytical purposes I recommend BAUP'S method, improved by FKEDEKKING, because the product obtained by tjiis process is free from iodic acid. Tests. Put a sample of the salt in dilute sulphuric acid. If the iodide is pure, it will dissolve without coloring the fluid ; but if it contain potassium iodate, the fluid will acquire a brown tint, 65.] REAGENTS. 149 from the presence of free iodine, the sulphuric acid setting free iodic and hydriodic acids which react on each other (HIO, -f- 5HI = 3H,O + 61) with liberation of iodine which remains in solution. Mix the solution of another sample with silver nitrate, as long as a precipitate continues to form ; add solution of ammonia in excess, shake the mixture, filter, and supersaturate the filtrate with nitric acid. The formation of a white, curdy precipitate indicates the presence of chloride in the potassium iodide. Presence of potassium sulphate is detected by means of solution of barium chloride, with addition of some hydrochloric acid. Uses. Potassium iodide is used as a solvent for iodine in the preparation of standard solutions of iodine ; it is employed also to absorb free chlorine. In the latter case every atom of chlorine lib- erates an atom of iodine, which is retained in solution by the agency of the excess of potassium iodide. The potassium iodide intended for these uses must be free from potassium iodate and carbonate; the presence of trifling traces of potassium chloride or potassium sulphate is of no consequence. If a potassium- iodide solution of accurately known strength is to be made, the salt must be dried before weighing. This may be accomplished by exposing it in powdered form to a tempera- ture of 180 until its weight is constant. Exposure to a tempera- ture much above 200 is to be avoided, as then potassium iodate is likely to be formed, and this would render the iodide impure. (PETTERSSON, Zeitschr. f. analyt. Chem., ix, 362). 8. SULPHUROUS ACID. Preparation. ^-The sulphurous acid gas evolved by the action of sulphuric acid on copper turnings (see "Qual. Anal.") is washed and then passed into water until the latter is saturated. The solution is best preserved in small, well- stoppered bottles. This concentrated solution serves for preparing the diluted solution of sulphurous acid used in BUNSEN'S method of estimat- ing iodine. 9. ARSENOUS OXIDE (As 2 O 3 ). The arsenous oxide sold in the shops in large pieces, externally opaque, but often still vitreous within, is generally quite pure. The purity of the article is tested by moderately heating it in a glass tube, open at both ends, through which a feeble current of air is transmitted. Pure arsenous oxide must completely volatilize in this process; no residue must be left in the tube upon the expulsion of the sublimate from it. If a non-volatile residue is left 150 REAGENTS. [ 65. which, when heated in a current of hydrogen gas, turns black, the arsenous oxide contains antimony trioxide, and is unfit for use in analytical processes. Dissolve about 10 grms. of the arsenous oxide to be tested in soda, and add 1 to 2 drops lead-acetate solu- tion. If a brownish color develops, the arsenous oxide contains arsenous sulphide and cannot be used. Arsenous oxide dissolves in a solution of sodium carbonate, forming sodium arsenite, which is used to determine hypochlorous acid, free chlorine, iodine, &c. 10. SODIUM CHLORIDE. Perfectly pure rock-salt is best suited for analytical purposes. It must dissolve in water to a clear fluid ; ammonium oxalate, sodium phosphate, and barium chloride must not render the solu- tion cloudy. Pure sodium chloride may be prepared also by MARGUERITTE'S process, viz., conduct into a concentrated solution of common salt hydrochloric gas to saturation, collect the small crystals of sodium chloride which separate on a funnel, let them' thoroughly drain, wash with hydrochloric acid, and dry the sodium chloride finally in a porcelain dish, until the hydrochloric acid adhering to it has completely evaporated. The mother-liquor which contains the small quantities of calcium sulphate, magne- sium chloride, &c., originally present in the salt, is used instead of a corresponding quantity of water, when next preparing hydro- chloric acid. Uses. Sodium chloride serves as a volumetric precipitating agent in the determination of silver, and also to determine the strength of solutions of silver intended for the estimation of chlo- rine. We usually fuse it before weighing. The operation must be conducted with caution, and must not be continued longer than necessary ; for if the gas-flame acts on the salt, hydrochloric acid escapes, while sodium carbonate is formed. 11. METALLIC SILVER. The silver obtained by the proper reduction of the pure chlo- ride of the metal alone can be called chemically pure. The silver precipitated by copper is never absolutely pure, but contains gener- ally about TfiVff of copper. Chemically pure silver is only used in small quantity for stand- ardizing the NaCl solution employed for the determination of silver. The solution of silver required for the estimation of chlorine need not be made with absolutely pure silver, as the strength of this solution had always best be determined after the preparation, by means of pure sodium chloride. 66.] REAGENTS. 151 D. REAGENTS USED IN ORGANIC ANALYSIS. 1. CUPRIC OXIDE. Preparation. Stir pure* copper scales (which should first be ignited in a muffle) with pure nitric acid in a porcelain dish to a thick paste ; after the effervescence has ceased, heat gently on the sand-bath until the mass is perfectly dry. Transfer the green basic salt produced to a Hessian crucible, and heat to a moderate redness, until no more fames of hyponitric acid escape ; this may be known by the smell, or by introducing a small portion of the mass into a test tube, closing the latter with the finger, heating to redness, and then looking through the tube lengthways. The uniform decom- position of the salt in the crucible may be promoted by stirring the mass from time to time with a hot glass rod. When the cruci- ble has cooled a little, reduce the mass, which now consists of pure cupric oxide, to a tolerably fine powder, by triturating it in a brass or porcelain mortar ; pass through a metal sieve, and keep in a well-stoppered bottle for use. It is always advisable to leave a small portion of the oxide in the crucible, and to expose this again to an intense red heat. This agglutinated portion is not pounded, but simply broken into small fragments. Another method is to dissolve pure copper in pure nitric acid, evaporate to dryness in a porcelain dish, ignite the copper nitrate thus obtained in a Hessian crucible until no fumes arise on stirring the top of the mass with a rod. A portion in the bottom of the crucible will be sintered if a proper heat has been applied, while the upper part will be pulverulent. Treat the sintered portion as above, and reserve each separately. This method gives a reliable product. Tests. Pure cupric oxide is a compact, heavy, deep-black pow- der, gritty to the touch ; upon exposure to a red heat it must evolve no hyponitric acid fumes, nor carbon dioxide ; the latter would indicate presence of fragments of charcoal, or particles of dust. It must contain nothing soluble in water. That portion of the oxide which has been exposed to an intense red heat should be hard, and of a grayish-black color. * If the scales contain lime, digest them with water, containing a little nitric acid, for a long time, wash, and then proceed as above. l.)2 REAGENTS. [ 66. Uses. Cupric oxide serves to oxidize the carbon and hydrogen of organic substances, yielding up its oxygen wholly or in part, according to circumstances. That portion of the oxide which has been heated to the most intense redness is particularly useful in the analysis of volatile fluids. N.B. The cupric oxide, after use, may be regenerated by oxi- dation with nitric acid, and subsequent ignition. Should it have become mixed with alkali salts in the course of the analytical pro- cess, it is first digested with very dilute cold nitric acid, and washed afterwards with water. To purify cupric oxide containing chlo- ride, E. ERLENMEYER recommends to ignite it in a tube, first in a stream of moist air, and finally, when the escaping gas ceases to redden litmus paper, in dry air. By these operations any oxides of nitrogen that may have remained are also removed. 2. LEAD CHROMATE. Preparation. Precipitate a clear filtered solution of lead ace- tate, slightly acidulated with acetic acid, with a small excess oi potassium dichromate ; wash the precipitate by decantation, and at last on a linen strainer ; dry, put in a Hessian crucible, and heat to bright redness until the mass is fairly in fusion. Pour out upon a stone slab or iron plate, break, pulverize, pass through a fine metallic sieve, and keep the tolerably fine powder for use. Tests. Lead chromate is a heavy powder, of a dirty yellowish- brown color. It must evolve no carbon dioxide upon the applica- tion of a red heat ; the evolution of carbon dioxide would indicate contamination with organic matter, dust, &c. It must contain nothing soluble in water. Uses. Lead chromate serves, the same as cupric oxide, for the combustion of organic substances. It is converted, in the pro- cess of combustion, into chromic oxide and basic lead chromate. It suffers the same decomposition, with evolution of oxygen, when heated by itself above its point of fusion. The property of lead chromate to fuse at a red heat renders it preferable to cupric oxide as an oxidizing agent, in cases where we have to act upon difficultly combustible substances. N.B. Lead chromate may be used a second time. For this purpose it is fused again (being first roasted, if necessary), and then powdered. After having been twice used it is powdered, moistened with nitric acidj dried, and fused. In this way the 66.] REAGENTS. 153 lead chromate may be used over and over again indefinitely (VOHL*). 3. OXYGEN GAS. Preparation. Triturate 100 grammes of potassium chlorate with 5 grammes of finely pulverized manganese binoxide, and introduce the mixture into a plain retort, which must not be more than half full ; expose the retort over a charcoal fire or a gas-lamp, at first to a gentle, and then to a gradually increased heat. As soon as the salt begins to fuse, shake the retort a little, that the contents may be uniformly heated. The evolution of oxygen speedily commences, and proceeds rapidly at a relatively low tem- perature, provided the above proportions be adhered to. As soon as the air is expelled from the retort, connect the glass tube fixed in the neck of the retort by means of a tight-fitting cork, with an india-rubber tube inserted in the lower orifice of the gasometer ; the glass tube must be sufficiently wide, and there must be sufficient space left around the india-rubber to permit the free efflux of dis- placed water. Continue the application of heat to the retort till the evolution of gas has ceased. 100 grammes of potassium chlorate give about 27 litres of oxygen. The oxygen produced by this process is moist, and may con- tain traces of carbon dioxide, and also of chlorine. These impuri- ties must be removed and the oxygen thoroughly dried, before it can be used in organic analysis. The gas is therefore passed from the gasometer first through a solution of potassa of 1*27 sp. gr., then through U tubes containing granulated soda lime, and finally > according to circumstances, through II tubes containing calcium chloride or pumice-stone moistened with sulphuric acid. Tests. A chip of wood which has been kindled and blown out so as to leave a spark at the extremity must immediately burst into flame in oxygen gas. The gas must not render lime-water or a solution of silver nitrate turbid when transmitted through these fluids. 4. SODA-LIME. Preparation. T^ke solution of soda (]S"aOH), ascertain its specific gravity, weigh out a certain quantity, calculate the weight of sodium hydroxide present, add twice this latter weight of the best quick-lime, allow the lime to slake, and then evaporate to drym- * Annalen d. CJiem. u. Pharm., cvi, 127. 154 EEAGENTS. [ 66. in an iron vessel. Heat the residue in an iron or Hessian crucible ; keep for some time at a low red heat. Break up while still warm in an iron mortar, and pass the whole through a sieve with meshes about 3 mm. wide. Reject the finest portion (removing it with a fine sieve having 2 mm. meshes) and keep the granulated prod- uct in a well-closed bottle. Tests. Soda-lime must not effervesce much when treated with dilute hydrochloric acid ; nor should it, more particularly, evolve ammonia when mixed with pure sugar and ignited. Use. Granulated soda-lime prepared as above described forms an excellent absorbent for carbon dioxide. It was formerly also used for nitrogen determination instead of the following : 5. SODA-LIME FOR NITROGEN DETERMINATIONS.* Preparation. Equal weights of sal-soda in clean (washed) large crystals and of good white promptly slaking quick-lime are separately so far pulverized as to pass through holes of -^ w inch, then well mixed together, placed in an iron pot which should not be more than half filled, and gently heated, at first without stir- ring. The lime soon begins to combine with the crystal water of the sodium carbonate, the whole mass heats strongly, swells up, and in a short time yields a fine powder, which may then be stirred to effect intimate mixture and to drive off the excess of water so that the mass is not perceptibly moist and yet short of the point at which it rises in dust on handling. When cold it is secured in well-closed bottles or fruit jars, and is ready for use. 6. METALLIC COPPER. Metallic copper serves, in the analysis of nitrogenous substances, to effect the reduction of the nitric oxide gas that may form in the course of the analytical process. It is used either in the form of turnings, or copper scales reduced by hydrogen ; or of small rolls made of fine copper wire gauze. A length of from 7 to 10 centimetres is given to the spirals or rolls, and just sufficient thickness to admit of their being inserted into the combustion tube. To have it perfectly free from dust, oxide, &c., it is first heated to redness in the open air, in a crucible, until the surface is oxidized ; it is then put into a glass or porcelain tube, through which an uninterrupted current of dry hydrogen gas is transmitted ; and when all atmospheric air has been expelled from the evolution apparatus and the tube, the * S. W. Johnson. Report of the Conn. Agr. Expr. Station, 1878, p. 111. 66.] REAGENTS. 155 latter is in its whole length heated to redness. The operator should make sure that the atmospheric air has been thoroughly expelled, before he proceeds to apply heat to the tube ; neglect of this pre- caution may lead to an explosion. 7. POTASSIUM HYDROXIDE OR POTASSA. a. Solution of Potassa. Solution of potassa is prepared from the carbonate, with the aid of milk of lime, in the way described in the " Qualitative Analysis," for the preparation of solution of soda. The propor- tions are 1 part of potassium carbonate to 12 parts of water, and f part of lime, slaked to paste with three times the quantity of warm water. The decanted clear solutio"h is evaporated, in an iron vessel, over a strong fire, until it has a specific gravity of 1-27 ; it is then, whilst still warm, poured into a bottle, which is well closed, and allowed to stand at rest until all solid particles have subsided. The clear solution is finally drawn off from the deposit, and kept for use. J. Fused Potassa (common). The commercial potassa in sticks (impure KOH usually com- bined with more or less H 2 O) will answer the purpose. If you wish to prepare it, evaporate solution of potassa (a) in a silver ves- sel, over a strong fire, until the residuary hydroxide flows like oil, and white fumes begin to rise from the surface. Pour the fused mass out on a clean iron plate, and break it up into small pieces. Keep in a well-stoppered bottle for use. c. Potassa (purified with alcohol), see " Qual. Anal.," p. 43. Uses. Solution of potassa serves for the absorption, and at the same time for the estimation of carbon dioxide. In many cases, a tube filled with fragments of fused potassa is used, in addition to the apparatus filled with solution of potassa. Potassa purified with alcohol^ which is perfectly free from potassium sul- phate, is employed for the determination of sulphur in organic substances. 8. CALCIUM CHLORIDE. a. Pure Calcium Chloride. Preparation. Dissolve marble in commercial hydrochloric acid diluted with four or five times its volume of water. (The waste solution resulting from the preparation of carbon dioxide 156 REAGENTS. [ 66- may be used.) Add to this solution with stirring lime, flaked with sufficient water to give it the consistency of thin paste until it gives an alkaline reaction and a pellicle of calcium carbonate forms on the surface on standing exposed to the air. Iron, man- ganese, and especially magnesium are usually present in such a solution, and are precipitated by the calcium hydroxide the iron, however, not completely. After a few hours, filter and pass hydro- gen sulphide through the alkaline solution until a filtered portion is no longer blackened by this reagent. Let the solution stand for twelve hours, then filter from the iron sulphide. Add next hydro- chloric acid to strongly acid reaction to convert the calcium sul- phide and calcium oxychloride which may be present into chloride. Concentrate in a porcelain dish. If sulphur separates, after a short time filter again, and continue the evaporation to dry ness with addition of a little more hydrochloric acid toward the end of the process. Finally expose the residue to a tolerably strong heat about (200) on the sand-bath, until it is changed throughout to a white porous perfectly opaque mass, which point can be ascertained by breaking up a piece detached from the top. The product is CaCl 2 -f- 2H a O. Reduce while still hot to granules of the proper size ( to ^ of an inch) by means of suitable sieves and a mortar previously warmed, and keep in well-closed bottles. b. Crude fused Calcium Chloride. Preparation. Neutralize the alkaline solution obtained in a (without separating the little iron present with H 2 S) exactly with hydrochloric acid, and evaporate to dryness in an iron pan ; fuse the residue in an iron or Hessian crucible, pour out the fused mass> and break into pieces. Preserve it in well-stoppered bottles. Uses. The crude fused calcium chloride serves to dry moist gases ; the pure chloride is used in elementary organic analysis for the absorption and estimation of water formed by the hydrogen contained in the analyzed substance. A solution of the pure cal- cium chloride should not show an alkaline -reaction. A calcium chloride tube filled with it should not gain weight when a very slow current of perfectly dry carbon dioxide is passed through it an hour. 9. POTASSIUM BICHROMATE. Potassium dichromate of commerce is purified by repeated recrystallization, until barium chloride produces, in the solution of 66.] REAGENTS. 157 a sample of it in water, a precipitate which completely dissolves in hydrochloric acid. Potassium dichromate thus perfectly free from sulphuric acid is required more particularly for the oxidation of organic substances with a view to the estimation of the sulphur contained in them. Where the salt is intended for other purposes, e.g., to determine the carbon of organic bodies, by heating them with potassium dichromate and sulphuric acid, one recrystallizatioa is sufficient. SECTION III. FOKMS AND COMBINATIONS IN WHICH SUB- STANCES ARE SEPARATED FROM EACH OTHER, OR IN WHICH THEIR WEIGHT IS DETERMINED. 67. THE quantitative analysis of a compound substance requires, as the first and most indispensable condition, a correct and accurate knowledge of the composition and properties of the new combina- tions into which it is intended to convert its several individual constituents, for the purpose of separating them from one another, and determining their several weights. Regarding the properties of the new compounds, we have to inquire more particularly, in the first place, how they behave with solvents ; secondly, what is their deportment in the air ; and, thirdly, what is their behavior on ignition ? It may be laid down as a general rule, that compounds are the better adapted for quantitative determination the more insoluble they are, and the less alteration they undergo upon exposure to air or to a high temperature. The composition of a substance is expressed either in per cents or in stoichiometrical formulas ; the latter enable the composition of the more frequently recurring compounds to be readily re- membered. In this section the chemical formula is stated in the first column ; the second gives the equivalents (O = 1 6), while the third gives the percentage composition. A compound is the better adapted for quantitative determina- tion, with respect to its composition, the less it contains relatively of the substance to be determined, since every operative error, loss, or inaccuracy is distributed over a larger mass on weighing, and the error will, hence, be so much the less for the substance to be determined. Thus platinum-ammonium chloride is better adapted for the estimation of nitrogen than is ammonium chloride because the former contains only 6*295 per cent, of nitrogen, whereas the latter contains 26*24 per cent. 1158 67.] FORMULAE. 159 Suppose we analyze a nitrogenous compound, and, with abso- lutely accurate manipulation, obtain 1 grai. of platinum-ammo- nium chloride from 0*3 grm. of the compound. 100 parts of this platinum salt contain 6-295 parts of nitrogen, hence 1 grm. will contain '06295 parts. Since this is afforded by 0*3 grm. of the substance, it follows that 100 parts of the latter will contain 20-983 parts of nitrogen. Let us now make a second analysis, and convert the nitrogen into ammonium chloride. "Working with equal accuracy we ob- tain from 0-3 grm. of the substance 0*2399 grm. ammonium, chloride, corresponding to 0*06295 grm. nitrogen, or 20*983 per cent. Assuming, now, that in both analyses a loss of 10 milli- grammes had occurred, we would have obtained in the first opera- tion only 0'99 parts 'of platinum-ammonium chloride instead of 100, corresponding to 0-06232 nitrogen, or 20*77 per cent. The loss would, hence, have been 0*213 per cent. In the case of ammonium chloride, however, we would have obtained 0*2299 parts instead of 0*2399, corresponding to 0-0603 nitrogen, or 20'1 per cent a loss of 0-873 per cent. We thus see that a similar error would occasion in one case a loss in nitrogen of 0*213 per cent, while in the other the loss would be 0-873 per cent. Having thus touched generally upon the requirements a com- pound must possess in order to be adapted for quantitative ex- amination, we will proceed to enumerate those compounds best adapted and which are as a rule employed. Of course the de- scription of the external form and appearance relates more par- ticularly to the condition in which they are obtained in our analy- sis. In enumerating tho properties of substances reference will be had exclusively to those which bear directly upon the object immediately in view. [The percentage compositions of these compounds are stated in connection with their description. For this purpose the symbols of the constituent elements of the compounds in many cases (viz. : when they are oxygen salts) are grouped in a manner different from that used to express their chemical constitution. This grouping constitutes a kind of formulae differing from either the empirical or rational in ordinary use in modern text-books of chemistry, but identical with that formerly in general use (the old system). These formulae are based upon the fact that in all oxygen salts, whether normal, acid, basic, ortho-, meta-, or pyro~~ 160 FORMS. [ 67. salts, there is just enough oxygen to form with the radicals present, both basic and acid, their corresponding oxides or anhydrides, and with hydrogen, if present, water. They represent oxides (and water) jointly equivalent in weight to the radicals, hydrogen, and remaining oxygen, which rational formulae represent as existing in oxygen salts EXAMPLES. Potassium sulphate, SO 2 < QK = K A SO - Hydrogen potassium sulphate, Potassium disulphate, < 18 i OK = K ' 2SO " Ammonium magnesium phosphate, / , \ \ Magnesium pyrophosphate, / PO < g > Mg O < =2MffO,P,O,. \ PO < g > Mg Most analytical chemists prefer to present the results of analyses of oxygen salts in percentages of oxides (or anhydrides) and water on account of the simplicity of computations required. Accord- ingly, in the following section, the percentage composition of oxygen salts is given in this manner, accompanied by correspond- ing formulae and molecular weights. These formulae are in every case preceded by rational formulae constructed in accordance with the theory of the constitution of oxygen salts which is now generally accepted.] 68.] BASES OF GROUP I. 161 A. FORMS IN WHICH THE BASIC RADICALS ARE WEIGHED OR PRECIPITA TED. BASIC RADICALS OF THE FIRST GROUP. . 68. 1. POTASSIUM. The combinations best suited for the weighing of potassium are POTASSIUM SULPHATE, POTASSIUM CHLORIDE, and POTASSIUM PLATINIC CHLORIDE. a. Potassium sulphate crystallizes usually in small, hard, straight, four-sided prisms, or in double six-sided pyramids ; in the analytical process it is obtained as a white crystalline mass. It dissolves with some difficulty in water (1 part requiring 10 parts of water of 12), it is almost absolutely insoluble in pure alcohol, but slightly more soluble in alcohol containing sulphuric acid (Expt. No. 6). It does not affect vegetable colors ; it is unalter- able in the air. The crystals decrepitate strongly when heated, yielding at the same time a little water, which they hold mechani- cally confined. The decrepitation of crystals that have been kept long drying is less marked. At a good red heat the salt fuses without volatilizing or decomposing. At a white heat a little of the salt volatilizes and also some sulphuric acid, so that the residue possesses an alkaline reaction (AL. MITSCHEKLICH,* BoussiNGAuurf-). When exposed to a red heat, in conjunction with ammonium chloride, potassium sulphate is partly, and, upon repeated applica- tion of the process, wholly converted, with effervescence, into potassium chloride (H. ROSE). COMPOSITION. qo OK _ K a O . . . 94-22 54-06 ^ 3< OK = -SO 8 . . . 80-07 45-94 174-29 100-00 The acid potassium sulphate (KHS0 4 ), which is produced when the normal salt is evaporated to dryness with free sulphuric acid, is readily soluble in water, and fusible even at a moderate heat, * Journ.f. prakt. Chem., IAXXIII, 486. f Zeitschr.f. anal. Chem., vn, 244. 162 FORMS. [ 68. At a red heat it loses sulphuric acid, and is converted into normal potassium sulphate, but not readily the complete conversion of the acid into the normal salt requiring the long-continued applica- tion of an intense red heat. However, when heated in an atmos- phere of ammonium carbonate which may be readily procured by repeatedly throwing into the faint red-hot crucible containing the acid sulphate small lumps of pure ammonium carbonate, and putting on the lid the acid salt changes readily and quickly to the normal sulphate. The transformation may be considered complete as soon as the salt, which was so readily fusible before, is perfectly solid at a faint red heat. b. Potassium nitrate crystallizes ordinarily in the form of long, striated prisms. In analysis it is obtained as a white saline mass. It is readily soluble in water, almost insoluble in absolute alcohol, and but sparingly soluble in alcohol. It does not affect vegetable colors, and is unchangeable in air. On being heated it melts far below a red heat, without decomposition or loss of weight. Strongly heated it evolves oxygen, and becomes con- verted into potassium nitrite ; intensely heated (to redness) oxy- gen arid nitrogen are evolved, the residue being then caustic potassa. On being ignited with ammonium chloride, or in a current of dry hydrochloric-acid gas, it is readily and completely converted into potassium chloride. Repeatedly evaporated with oxalic acid in excess (4 to 6 times), it is completely converted into potassium chloride. COMPOSITION. . . . 46-04 45-52 101-15 100-00 D. Potassium chloride crystallizes usually in cubes, often lengthened to columns ; rarely in octahedra. In analysis we obtain it either in the former shape, or as a crystalline mass. It is readily soluble in water, bat much less so in dilute hydrochloric acid ; in absolute alcohol it is nearly insoluble, and but slightly soluble in common alcohol. It does not affect vegetable colors, and is unalterable in the air. "When heated, it decrepitates, unless it has been kept long drying, with expulsion of a little water 68.] BASES OF GROUP I. 163 mechanically confined in it. At a moderate red heat, it fuses unaltered and without diminution of weight ; when exposed to a higher temperature, it volatilizes in white fumes ; this volatilization proceeds the more slowly the more effectually the access of air is prevented (Expt. No. T). When repeatedly evaporated with solution of oxalic acid in excess, it is converted into potassium oxalate. "When evaporated with excess of nitric acid, it is con- verted readily and completely into nitrate. On ignition with ammonium oxalate, potassium carbonate and potassium cyanide are formed in noticeable quantities. COMPOSITION. K .... 39-11 52-455 Cl 35-45 47-545 74-56 100-000 d. Potassium platinic chloride presents either small reddish- yellow octahedra, or a lemon-colored -powder. It is difficultly soluble in cold, more readily in hot water; nearly insoluble in absolute alcohol, and but sparingly soluble in common alcohol one part requiring for its solution, respectively, 12083 parts of absolute alcohol, 3775 parts of alcohol of 76 per cent, and 1053 parts of alcohol of 55 per cent. (Expt. No. 8, a.} Presence of free hydrochloric acid sensibly increases the solubility (Expt. No. 8, I). In caustic potassa it dissolves completely to a yellow fluid. It is unalterable in the air, and at 100. On exposure to an intense red heat, four atoms of chlorine escape, metallic plati- num and potassium chloride being left ; but even after long-con- tinued fusion, there remains always a little potassium platinic chloride which resists decomposition. Complete decomposition is easily effected, by igniting the double salt in a current of hydrogen gas, or with some oxalic acid. According to ANDREWS, potassium platinic chloride, even though dried at a temperature considerably exceeding 100, retains still -0055 of its weight of water. 164 FORMS. [ 69. COMPOSITION. (KC1) 3 . . . 149-12 30-69 K, . . . 78-22 16-10 PtCL . . . 336-70 69-31 Ft ... 194-90 40-12 01.. . . 212-70 43 78 485-82 100-00 485-82 100-00 e. Potassium silicofluoride is obtained on mixing a solution of a potassium salt with hydrofluosilicic acid in the form of a trans- lucent iridescent precipitate, which increases and completely separates, when an equal volume of strong alcohol is added to the fluid. After being filtered off, washed with weak alcohol and dried, it is a soft white powder. It is difficultly soluble in cold water, far more readily in boiling water, not at all or in merest traces soluble in a mixture of water and strong alcohol in equal parts, but it is decidedly more soluble in the presence of any considerable quan-, tity of free acid, especially hydrochloric or sulphuric acid. When potassa is added to the boiling aqueous solution of the salt the following change takes place : (KF) 2 SiF 4 + 4KOH : : 6KF + Si(OH) 4 , the solution turning from acid to neutral (principle of STOLBA'S volumetric method of estimating potassium). As soon as it is ignited the salt fuses, gives off silicon fluoride and leaves potassium fluoride. 69. 2. SODIUM. Sodium is usually weighed as SODIUM SULPHATE, SODIUM CHLO- RIDE, or SODIUM CARBONATE. It is separated from potassium in the form of SODIUM PLATINIC CHLORIDE, from other bodies occasionally in the form of sodium silicofluoride. a. Anhydrous normal sodium sulphate is a white powder or a white very friable mass. It dissolves readily in water ; but is sparingly soluble in absolute alcohol ; presence of free sulphuric acid slightly increases its solubility in that menstruum ; it is some- what more readily soluble in common alcohol (Expt. No. 9). It does not a-ffect vegetable colors ; upon exposure to moist air, it slowly absorbs water (Expt. No. 10). At a gentle heat it is un- altered, at a strong red heat it fuses without decomposition or loss of weight. At a white heat it loses weight by volatilization of sodium sulphate and also of sulphuric acid (AL. MITSCIIERLICII, 69.] BASKS OF GROUP I. 165 BOUSSINGAULT). When ignited with ammonium chloride, it be* haves like potassium sulphate. COMPOSITION. j ONa Na,O ..... 62 10 43 '68 u * < ONa ' SO 3 .... 80-07 56-32 142-17 100-00 The acid sodium sulphate (sodium hydrogen sulphate, NaHSO 4 ) which is always produced upon the evaporation of a solution of the normal salt with sulphuric acid in excess, fuses even at a gentle heat ; it may be readily converted into the normal salt in the same manner as the acid potassium sulphate (see 68, a). 1). Sodium nitrate crystallizes as obtuse rhombohedra. In analysis it is usually obtained as an amorphous saline mass. It is readily soluble in water, is practically insoluble in absolute alcohol, and is but very slightly soluble in alcohol. It is indifferent towards vegetable colors. Under ordinary circumstances it is unalterable in the air, but attracts moisture from very moist air. It fuses far below a red heat, and without decomposition (comp. Expt. No. 11); at a higher temperature it is decomposed like potassium nitrate ( 68). Ignited with ammonium chloride, or in hydrochloric-acid gas, and evaporated with oxalic acid or aqueous hydrochloric acid, it behaves like the corresponding potassium salt. The decomposition with aqueous hydrochloric acid is more readily effected, i.e., with fewer evaporations, than is the case with potassium nitrate (BAUMHATTER *). COMPOSITION. NT n -vrXO_]SraO . . . 39.05 45-89 ~ . . . 46-04 54-11 85-09 100-00 c. Sodium chloride crystallizes in cubes, octahedra, and hollow * Jour. f. prakt. Chem., LXXVIII, 213. 166 FORMS. [69,. four-sided pyramids. In analysis it is frequently obtained as an amorphous mass. It dissolves readily in water, but is much less soluble in hydrochloric acid ; it is nearly insoluble in absolute alcohol, and but sparingly soluble in common alcohol ; 100 parts of alcohol of 75 per cent, dissolve, at a temperature of 15, 0*7 part (WAGNEK). It is neutral to vegetable colors. Exposed to a somewhat moist atmosphere, it slowly absorbs water (Expt. No. 12). Crystals of this salt that have not been kept drying a considerable time decrepitate when heated, yielding a little water, which they hold mechanically confined. The salt fuses at a red heat without decomposition ; at a white heat, and in open vessels even at a bright red heat, it volatilizes in white fumes (Expt. "No. 13). If a carburetted hydrogen flame acts on fusing sodium chloride, hydro- chloric acid escapes, and some sodium carbonate is formed. On evaporation with oxalic or nitric acid as well as by ignition with ammonium oxalate, it behaves like the corresponding potassium salt. COMPOSITION. !Na . . . . 23-05 39-40 Cl 35-45 60 60 58-50 100-00 d. Anhydrous sodium carbonate is a white powder or a white very friable mass. It dissolves readily in water, but much less so in solution of ammonia (MARGUERITTE) ; it is insoluble in alcohol. Its reaction is strongly alkaline. Exposed to the air, it absorbs water slowly. On moderate ignition to incipient fusion it scarcely loses weight ; on long fusion, however, it volatilizes to a consider- able extent (Comp. Exp fc. 14). COMPOSITION. .ONa_ ]STa a O . . . . 62-10 58-53 < ONa ~ CO, .... 44-00 41-47 106-10 100-00 e. Sodium platinic chloride crystallizes with 6 mol. water, (NaCl) a .PtCl 4 -f- 6H 5 O, in light yellow, transparent, prismatic crystals which dissolve readily both in water and in common alcohol. 70.] BASES OF GROUP I. 167 f. Sodium silicqfluoride is similar in properties to the corre- sponding potassium salt. It has an analogous composition, and is decomposed in the same way by alkalies. It is, however, con- siderably more soluble in water and in diluted alcohol. TO. 3. AMMONIUM. Ammonium is most appropriately weighed as AMMONIUM OHLORIDE, or as AMMONIUM pLATiNic CHLORIDE, or it may be esti- mated from the weight of the PLATINUM in the latter compound. Under certain circumstances ammonium may also be estimated from the volume of the NITROGEN GAS eliminated from it ; and it is frequently estimated by alkalimetry. a. Ammonium chloride crystallizes in cubes and octahedra, but more frequently in feathery crystals. In analysis we obtain it uniformly as a white mass. It dissolves readily in water, but is difficultly soluble in common alcohol. It does not alter vegetable colors, and remains unaltered in the air. Solution of ammonium chloride, when evaporated on the water-bath, loses a small quantity of ammonia, and becomes slightly acid. The diminution of weight occasioned by this loss of ammonia is very trifling (Expt. No. 15). At 100 ammonium chloride loses nothing, or very little of its weight (comp. same Expt.). At a higher temperature it volatilizes readily, and without undergoing decomposition. COMPOSITION. NH 4 . . 18-072 33-77 NH, . . 17-064 31-88 Cl. 35-450 06-23 HC1 36-458 68-12 53-522 100-00 53'522 100-00 100 parts of ammonium chloride correspond to 48*72 parts of ammonium oxide. b. Ammonium platinic chloride occurs either as a heavy, lemon-colored powder, or in small, hard octahedral crystals of a bright yellow color. It is difficultly soluble in cold, but more readily in hot water. It is very sparingly soluble in absolute alcohol, but more readily in common alcohol 1 part requiring of absolute alcohol, 26535 parts ; of. alcohol of 76 per cent., 1406 168 FORMS. [ 71. parts; of alcohol of 55 per cent., 665 parts. The presence of free acid sensibly increases its solubility (Expt. No. 16). It remains unaltered in the air, and at 100. It loses a little water between 100 and 125. Upon ignition chlorine and ammonium chloride escape, leaving the metallic platinum as a porous masa (spongy platinum). However, if due care be not taken in this process to apply the heat gradually, the escaping fumes will carry off particles of platinum, which will coat the lid of the crucible. For properties of metallic platinum, see 89, a. COMPOSITION. (NI-I 4 C1) S . .107-044 24-12 (NH 4 ) 3 . . 36-144 8'15 FtCl 4 . . . 336-700 75-88 Ft ... 194-900 43-92 Cl e . . . 212-700 47-93 443-744 100-00 443-744 100 00 N, . . . . 28-080 6-33 (NH 3 ) a . . 34-128 7-691 H e ... 8-064 1-82 Pt ... 194-900 43-92 (HCl) f . . 72-916 16-432 01. ... 212-700 47.93 FtCl 4 . . 336-700 76-010 443-744 100-00 443-744 100-000 100 parts of ammonium platinic chloride correspond to 11 '76 parts of ammonium oxide. c. Nitrogen gas is colorless, tasteless, and inodorous ; it mixes with air, without producing the slightest coloration ; it does not affect vegetable colors. Its specific gravity is 0-996971.* One litre weighs at 0, and 0-76 metre bar., 1*254035 grm. It is difficultly soluble in water, 1 volume of water absorbing, at 0, and 0-76 pressure, 0-02035 vol. ; at 10, 0-01607 vol. ; at 15% 0-01478 vol. of nitrogen gas (BUNSEN). BASIC RADICALS OF THE SECOND GROUP. 71. 1. BARIUM. Barium is weighed as BARIUM SULPHATE, BARIUM CARBONATE, and BARIUM SILICOFLUORIDE. a. f Artificially prepared barium sulphate presents the appear- ance of a fine white powder. When recently precipitated, it is * According to REGNAULT, 0-97137. 71.] BASES OF GROUP II. difficult to obtain a clear filtrate, especially if the precipitation wa& effected in the cold, and the solution contains neither hydrochloric acid nor ammonium chloride. It is as good as insoluble in cold and in hot water. (1 part of the salt requires more than 400,000 | parts of water for solution.) It has a great tendency, upon pre- cipitation, to carry down with it other substances contained in the solution from which it separates, more particularly barium nitrate, nitrates and chlorates of the alkali metals, ferric oxide, &c. Several of the impurities, such, for instance, as potassium or sodium chlo- rates, may be removed by igniting the barium sulphate, moistening with hydrochloric acid, evaporating the latter off and exhausting the residue with water ; other impurities again, such as potassium or sodium nitrates, cannot be removed by this treatment. Even the precipitate obtained from a solution of barium chloride by means of sulphuric acid in excess contains traces of barium chloride, which it is impossible to remove, even by washing with boiling water, but which are dissolved by nitric acid (SIEGLE). Cold dilute acids dissolve trifling, yet appreciable traces of barium sulphate ; for instance, 1000 parts of nitric acid of 1 -032 sp. gr. dissolve 0-062 parts (CALVERT), 1000 parts of hydrochloric acid containing 3 per cent, dissolve 0*06 parts.* Cold concentrated acids dissolve con- siderably more ; thus, 1000 parts of nitric acid of 1 '167 sp. gr. dis- solve 2 parts (CALVERT). Boiling hydrochloric acid also dissolves appreciable traces; thus 230 c. c. hydrochloric acid of 1*02 sp. gr., were found, after a quarter of an hour's boiling with 0'679 grm. barium sulphate, to have dissolved of it O'O-IS grm. Acetic acid dissolves less barium sulphate than the other acids; thus, 80 c. c. acetic acid of 1*02 sp. gr. were found, after a quarter of an hour's boiling with 0'4 grm., to have dissolved only 0'002 grm. (SIEGLE). Free chlorine considerably increases its solubility (O. L. ERDMANN). Several salts more particularly interfere with the precipitation of barium by sulphuric acid. I observed this some time ago with magnesium chloride, but ammonium nitrate (MITTENTZWEY), alkali nitrates generally,* and more particularly alkali citrates (SPILLER), possess this property in a high degree. In the last case the pre- cipitate appears on the addition of hydrochloric acid. If a fluid contains metaphosphoric acid, barium cannot be completely pre- cipitated out of it by means of sulphuric acid ; the resulting pre- cipitate too contains phosphoric acid (SCHEERER, RUBE). Barium * Zeitschr.f. anal. Chem., ix, 62. 170 FOKMS. [ 71. sulphate dissolves in considerable quantity in concentrated sulphuric acid, but separates again on dilution. It is as good as insoluble in a boiling solution of ammonium sulphate (1 in 4). Barium sulphate remains quite unaltered in the air, at 100, and even at a red heat. At a strong white heat it loses sulphuric acid (Bous- SINGAULT).* On ignition with charcoal, or under the influence of reducing gases, it is converted comparatively easily, but as a rule only partially, into barium .sulphide. On ignition with ammonium chloride, barium sulphate undergoes partial decomposition. It is not affected, or affected but very slightly, by cold solutions of the hydrogen carbonates of the alkali metals or of ammonium carbo- nate ; solutions of normal sodium and potassium carbonates when cold have only a slight decomposing action upon it ; but when boiling, and upon repeated application, they effect at last the complete decomposition of the salt (H. KOSE). By fusion with sodium or potassium carbonate, barium sulphate is readily decom- posed. COMPOSITION. BaO . . . . 153-40 65-70 SO S .... 80-07 34-30 233-47 100 00 I. Artificially prepared barium carbonate presents the appear- ance of a white powder. It dissolves in 14137 parts of cold, and in 15421 parts of boiling water (Expt. No. 17). It dissolves far more readily in solutions of ammonium chloride or ammonium nitrate; from these solutions it is, however, precipitated again, though not completely, by caustic ammonia. In water containing free carbonic acid, barium carbonate dissolves to an acid carbonate. In water con- taining ammonia and ammonium carbonate, it is nearly insoluble, one part requiring about 141000 parts (Expt. No. 18). Its solution in water has a very faint alkaline reaction. Alkali citrates and motaphosphates impede the precipitation of barium by ammonium carbonate. It is unalterable in the air, and at a red heat. When exposed to the strongest heat of a blast-furnace, it slowly yields up the whole of its carbonic acid ; this expulsion of the carbonic acid is promoted by the simultaneous action of aqueous vapor. Upon heating it to redness with charcoal, caustic baryta is formed, with evolution of carbon monoxide. * Z&itschr.f. analyt. Chem., vn, 244. 72.] BASES OF GROUP II. 171 COMPOSITION. _BaO .... 153-4 77-71 - CO a .... 44 22-29 197-4 100-00 c. Bcvrium silicofluoride forms small, hard, and colorless crys- tals, or (more generally) a crystalline powder. It dissolves in 3800 parts of cold water ; in hot water it is more readily soluble (Expt. No. 19). The presence of free hydrochloric acid increases its solu- bility considerably (Expt. No. 20). Ammonium chloride acts also in the same way (1 part siliconuoride of barium dissolves in 428 parts of saturated, and 589 parts of dilute solution of ammonium chloride. J. W. MALLET). In common alcohol it is almost insoluble. It is unalterable in the air, and at 100 ; when ignited, it is decom- posed into silicon fluoride, which escapes, and barium fluoride, which remains. COMPOSITION. BaF 3 . . . 175-5 62-66 Ba . . . 137-4 49-05 SiF 4 ... 104-6 37-34 Si ... 28-4 10-14 F. . . . 114.3 40-81 280-1 100-00 280-1 100-00 72. 2. STRONTIUM. Strontium is weighed either as STRONTIUM SULPHATE, or as STRONTIUM CARBONATE. a. Strontium sulphate, artificially prepared, is a white powder, sometimes dense and crystalline, sometimes loose and bulky. It dissolves in 6895 parts of cold, and 9638 parts of boiling water (Expt. No. 21). In water containing sulphuric acid, it is still more difficultly soluble, requiring from 11000 to 12000 parts (Expt. No. 22). Of cold hydrochloric acid of 8'5 per cent., it requires 474 parts ; of cold nitric acid of 4'8 per cent., 432 parts ; of cold acetic acid of 15-6 per cent, of HC 2 H 3 O a , as much as 7843 parts (Expt. No. 23). It dissolves in solutions of potassium chloride and magnesium chlo- ride, in quantity which increases with the concentration, also in solu- tions of sodium chloride and calcium chloride in greatest quantity 172 FOKMS. L 7%- when the solutions are of medium concentration (A. VIRCK*) ; it it is precipitated from these solutions by sulphuric acid. Meta- phosphoric acid (SCHEERER, RTJBE), and also alkali citrates, but not free citric acid (SPILLER), impede the precipitation of strontium by sulphuric acid. It is as good as insoluble in absolute alcohol, in common alcohol, and in a boiling solution of ammonium sulphate (1 in 4). It does not alter vegetable colors ; and remains unaltered in the air, and at a red heat. When exposed to a most intense red heat, it fuses with loss of a small quantity of sulphuric acid (M. DARMSTADT f) ; all the sulphuric acid will escape on very strong ignition continued for a length of time (BOUSSINGAULT $). When ignited with charcoal, or under the influence of 1 reducing gases, it is converted into strontium sulphide. Solutions of acid and nor- mal carbonates of potassium, sodium, and ammonium decompose strontium sulphate completely at the common temperature, even when considerable quantities of alkali sulphates are present (EL ROSE). Boiling promotes the decomposition. COMPOSITION. _SrO . . . 103-00 56-41 SO . . . 80-0.7 43-59 183-67 100-00 J. Strontium carbonate, artificially prepared, is a white, soft, loose powder. It dissolves, at the common temperature, in 18045 parts of water (Expt. No. 24) : presence of ammonia diminishes its solubility (Expt. No. 25). It dissolves pretty readily in solu- tions of ammonium chloride and ammonium nitrate, but is precipi- tated again from these solutions by ammonia and ammonium car- bonate, and more completely than barium carbonate under similar circumstances. Water impregnated with carbonic acid dissolves it as an acid carbonate. Its reaction is very feebly alkaline. Alkali citrates and metaphosphates impede the precipitation of strontium by -alkali carbonates. Ignited with access of air it is infusible, but when exposed to a most intense heat, it fuses and gradually loses its carbonic acid. On ignition with charcoal, strontium oxide is formed, with evolution of carbon monoxide gas. * Zeitschr.f. analyt. Chem., i, 473. ~\ lb., vi, 370. \lb., vn, 244. 73.] BASES OF GROUP II. 173 COMPOSITION. ^O. _ SrO . . . 103-6 70*19 -O' "CO, ... 44-0 29-81 '147-6 100-00 73. 3. CALCIUM. Calcium is weighed either as CALCIUM SULPHATE, CALCIUM CAR- BONATE, or CALCIUM OXIDE ; to convert it into the latter forms, it is first usually precipitated as calcium oxalate. a. Artificially prepared anhydrous calcium sulphate is a loose, white powder. It dissolves, at the common temperature, in 430 parts, at 100, in 460 parts of water (POGGIALE). Presence of hydrochloric acid, nitric acid, ammonium chloride, sodium sulphate, or sodium chloride, increases its solubility. It dissolves with com- parative ease, especially on gently warming, in aqueous solution of sodium thiosulphate (DIEHL), and also in a boiling solution of ammonium sulphate (1 in 4). The aqueous solution of calcium sulphate does not alter vegetable colors. In alcohol of 90 per cent or stronger it is almost absolutely insoluble. Exposed to the air, it slowly absorbs water. It remains unaltered at a dull-red heat. Heated to intense bright redness, it fuses, losing weight consider- ably from loss of sulphuric acid (AL. MITSCHEELICH *). On long ignition at a white heat all the sulphuric acid escapes (BoussiN- OAULTf). On ignition with charcoal, or under the influence of reducing gases, it is converted into calcium sulphide. Solutions of normal and acid carbonates of the alkali metals decompose cal- cium sulphate more readily still than strontium sulphate. COMPOSITION. ^O^ p CaO . . . 56-10 41-20 U ' < > = SO, ... 80-07 58-80 136-17 100-00 b. Calcium carbonate artificially produced by the precipitation of a calcium salt with ammonium carbonate is at first loose and *Jour.f. prakt. Chem., LXXXIII, 485. f Zeitschr.f. analyt. Chem., vn, 224. 174 FORMS. [ 73. amorphous, but after some time becomes a white, fine, crystalline powder, which under the microscope has sometimes the form of calcite, sometimes that of aragonite. It is very slightly soluble in water. By protracted boiling 1 litre of water dissolves 0*034: grm. according to A. "W. HOFMANN, or 0-036 grm. according to C. WELTZIEN; so one part requires 28500 parts of water for solu- tion. The solution has a barely-perceptible alkaline reaction. In water containing ammonia and ammonium carbonate the crystal- lized salt dissolves much more sparingly (Expt. No. 26), one part requiring about 65000 parts ; this solution is not precipitated by ammonium oxalate. Amorphous calcium carbonate is also much more insoluble in water containing ammonia than in pure water (DIVEKS*). Presence of ammonium chloride and of ammo- nium nitrate increases the solubility of calcium carbonate ; but the salt is precipitated again from these solutions by ammonia and ammonium carbonate, and more completely than barium carbonate under similar circumstances. Normal salts of potassium and sodium, and also normal calcium and magnesium salts (HUNT), likewise increase its solubility, the precipitation of calcium by the alkali carbonates is completely prevented or considerably interfered with by the presence of alkali citrates (SPILLER) or metaphosphates (RUBE). Water impregnated with carbonic acid dissolves calcium carbonate as acid carbonate. Calcium carbonate remains unaltered in the air at 100, and even at a low red heat ; but upon the appli- cation of a stronger heat, more particularly with free access of air, it gradually loses its carbonic acid. By means of a gas blowpipe- lamp, calcium carbonate (about 0*5 grm.), in an open platinum crucible, is without difficulty reduced to calcium oxide ; attempts to effect complete reduction over a spirit lamp with double draught have, however, failed (Expt. No. 27). It is decomposed far more readily when ignited with charcoal, giving off its carbonic acid in, the form of carbon monoxide. COMPOSITION. 0. pa _CaO . . . . 56-1- 56-04 O ' ~ CO, . . . . 44-0 43-96 100-1 100*00 * Jour. Chem. Soc. 1870, 362. 73.] BASES OF GROUP II. 175 c. Calcium oxalate, precipitated from hot or concentrated solu* tions, is a fine white powder consisting of infinitely minute indis- tinct crystals, and almost absolutely insoluble in water. The salt has the formula, CaC 2 O 4 -|- H 9 O. When precipitated from cold, extremely -dilute -solutions, the salt presents a more distinctly crys- talline appearance, and consists of a mixture of CaC 2 O 4 -|- H 2 O and CaC 2 O 4 + 3II 3 O (SOUCHAY and LENSSEN). Presence of free oxalic acid and acetic acid slightly increases the solubility of calcium oxalate. The stronger acids (hydrochloric acid, nitric acid) dissolve it readily ; from these solutions it is precipitated again unaltered, by alkalies, and also (provided the excess of acid be not too great) by alkali oxalates or acetates added in excess. Calcium oxalate does not dissolve in solutions of potassium chloride, sodium chlo- ride, ammonium chloride, barium chloride, calcium chloride, and strontium chloride, even though these solutions be hot and concen- trated ; but, on the other hand, it dissolves readily and in appreci- able quantities, in hot solutions of the salts belonging to the mag- nesium group. From these solutions it is reprecipitated by an excess of alkali oxalate (SOUCHAY and LENSSEN). Alkali citrates (SPILLEK) and metaphosphates (RUBE) impede the precipitation of lime by alkali oxalates. When treated with solutions of many of the heavy metals, e.g., with solution of cupric chloride, silver nitrate, &c., calcium oxalate suffers decomposition, a soluble cal- cium salt being formed, and an oxalate of the heavy metal, which separates immediately, or after some time (REYNOSO). Calcium oxalate is unalterable in the air, and at 100. Dried at the latter temperature, it has invariably the following composition (Expt. No. 28, also SOUCHAY and LENSSEN*). CO O\ CaO . . . 56-100 38-39 | ;> Ca 4- H f O = C 2 3 . . . 72-000 . 49-28 CO-0/ H 2 . . , 18-016 12-33 146-116 100-00 At 205 calcium oxalate loses its water, without undergoing decomposition ; at a somewhat higher temperature, still scarcely reaching dull redness, the anhydrous salt is decomposed, without actual separation of carbon, into carbon monoxide and calcium carbonate. The powder, which was previously of snowy whiteness, * Annal. d. Chem. und Pharm., c, 322. 176 FORMS. L 74. transiently assumes a gray color in the course of this process, even though the oxalate be perfectly pure. Upon continued applica- tion of heat this gray color disappears again. If the calcium oxalate is heated in small, coherent fragments, such as are obtained upon drying the precipitated salt on a filter, the commencement and progress of the decomposition can be readily traced by this transient appearance of gray. If the process of heating be con- ducted properly, the residue will not contain a trace of calcium oxide. Hydrated calcium oxalate exposed suddenly to a dull-red heat, is decomposed with considerable separation of carbon. By ignition over the gas blowpipe calcium oxalate is converted into calcium oxide. d. Calcium oxide obtained by continued strong ignition of the oxalate or carbonate appears as a white, infusible powder, unalter- able by ignition. By standing in the air it attracts water and car- bonic acid, but not rapidly enough to interfere with accurate weighing (Expt. No. 29). By treatment with a little water calcium hydroxide is formed with evolution of much heat ; on igniting again the water of hydration is readily and completely removed. Pure calcium oxide dissolves in dilute hydrochloric acid with evolution of heat, but without effervescence. 4. MAGNESIUM. Magnesium is weighed as MAGNESIUM SULPHATE, MAGNESIUM PYROPHOSPHATE, or MAGNESIUM OXIDE. To convert it into the pyro- phosphate, it is precipitated as NOKMAL AMMONIUM MAGNESIUM PHOS- PHATE. a. Anhydrous magnesium sulphate presents the appearance of a white, opaque mass. It dissolves readily in water. It is nearly altogether insoluble in absolute alcohol, but it is somewhat soluble in common alcohol. It does not alter vegetable colors. Exposed to the air it absorbs water rapidly. At a moderate red heat, it remains unaltered ; but when heated to intense redness, it undergoes partial decomposition, losing part of its acid, after which it is no longer perfectly soluble in water. By means of a gas blowpipe it as tolerably easy to expel 74.] BASES OF GROUP II. 177 the whole of the sulphuric acid from small quantities of magne- sium sulphate (Expt. No. 30). Ignited with ammonium chloride magnesium sulphate is not decomposed. COMPOSITION. so /O Mo ._MgO .... 40-30 33-48 *<0 >M S-S0 3 .... 80-07 66-52 120-37 100-00 b. Ammonium magnesium phosphate is a white crystalline powder. It dissolves, at the common temperature, in 15293 parts of cold water (Expt. No. 31). In water containing ammonia, it is much more insoluble. 1000 grin, of a mixture of 3 parts water and 1 part ammonia solution, dissolved only a quantity correspond- ing to 0-004 grm. pyrophosphate (KISSEL *) ; the salt was consid- erably more soluble when ammonium chloride was also present ; thus, in one of KISSEL'S experiments a quantity corresponding to 0-011 grm. pyrophosphate was dissolved by 1000 grm. fluid con- taining 18 grm. ammonium chloride. Presence of excess of mag- nesium sulphate diminishes the solubility in dilute ammonia, even in the presence of ammonium chloride, to such an extent that the quantity dissolved by 1000 grm. fluid cannot be estimated (KISSEL); the precipitate, under these circumstances, is liable, especially in the absence of much ammonium chloride, and when a large excess of magnesium sulphate is present, to contain some magnesium hydroxide or basic magnesium sulphate (KuBEL,f KISSEL). Sodium phosphate also diminishes (to about the same extent as magnesium sulphate) the solubility of the salt in water containing ammonium chloride and ammonia (W. HEINTZ J). It dissolves readily in acids, even in acetic acid. Its composition is expressed by the formula NH 4 MgPO 4 + 6H 2 O. 5 mol. of water escape at 100, the remain- ing water together with ammonia are expelled, at a red heat, leav- ing Mg Q P 2 O 7 . On the application of a stronger heat the mass passes through a state of incandescence, if the salt were pure ; the weight of the residue is not affected. The incandescence may not take place at all in the presence of small quantities of calcium salts, * Zeitschr. j. analyt. Chem., viir, 173. f lb., vm, 125. $lb. t ix, 16. 178 FOKMS. [ 74. of other magnesium salts, or of silicic acid. It is occasioned not by the passage of the orthophosphate into the pyrophosphate, but by the passage from the crystalline to the amorphous condi- tion (O. POPP *). If ammonium magnesium phosphate is dissolved in dilute hydrochloric or nitric acid and ammonia be then added to the solution, the salt is reprecipitated completely, or more cor- rectly, only so much remains in solution as corresponds to its ordinary solubility in water containing ammonia and ammonium salt. c. Magnesium pyrophosphate presents the appearance of a white mass, often slightly inclining to gray. It is barely soluble in water, but readily so in hydrochloric acid, and in nitric acid. It remains unaltered in the air, and at a red heat ; at a very intense heat it fuses unaltered. Exposed at a white heat to the action of hydrogen, Mg 3 (PO 4 ) 2 is formed, while PH 3 , P and P 2 O 3 escape. 3Mg 2 P 2 Q 7 = 2Mg,(PO 4 ) 1 + P a O B (STRUVEf). It leaves the color of moist turmeric-, and of reddened litmus-paper unchanged. If we dissolve it in hydrochloric or nitric acid, add water to the solu- tion, boil for some time, and then precipitate with ammonia in excess, we obtain a precipitate of ammonium magnesium phosphate which, after ignition, affords less Mg 2 P 2 O 7 , than was originally employed. WEBER gives the loss as from 1-3 to 2-3 per cent. My experiments (Expt. No. 32) confirm this, and show under what conditions the loss is smallest. By long-continued fusion with mixed potassium and sodium carbonates, magnesium pyrophosphate is completely decomposed, the pyrophosphoric acid being re- converted into orthophosphoric. If, therefore, we treat the fused mass with hydrochloric acid, and then, add water and ammonia, we re-obtain on igniting the precipitate the whole quantity of the salt used. If the solution of magnesium pyrophosphate in nitric acid is evaporated to dry ness a white resi- due is left ; if this is heated more strongly hyponitric acid is liber- ated, and the residue turns the color of cinnamon ; on cooling it is yellowish-white. By heating still more strongly to incipient red- ness, rapid decomposition sets in, more hyponitric acid is evolved, and pure-white magnesium pyrophosphate is left. Unless the heat is applied with care the evolution of gas may be so rapid as to- carry away particles of the substance (E. LUCK). * Zeitschr. f. analyt. Chem., xiu, 305. ' f Jour. f. prakt. Chem., LXXIX, 349s, \Pogg. Ann., LXXIIT, 146. 75.] BASES OF GROUP III. 179 COMPOSITION. X PO< () >M g_2MgO . . . 80-6 36-21 ^ 222-6 100-00 d. Magnesium oxide is a white, light, loose powder. It dis- solves in 55,368 parts of cold, and in the same proportion of boil- ing water (Expt. No. 33). Its aqueous solution has a very slightly alkaline reaction. It dissolves in hydrochloric and in other acids, without evolution of gas. Magnesium oxide dissolves readily and in quantity, in solutions of normal ammonium salts, and also in solutions of potassium chloride and sodium chloride (Expt. No. 34) and potassium sulphate and sodium sulphate (R. WARINGTON, Jr.) it is more soluble than in water. Exposed to the air, it slowly absorbs carbonic acid and water. Magnesium oxide is highly infusi- .ble, remaining unaltered at a strong red heat, and fusing super- ficially only at the very highest temperature. COMPOSITION. Mg ........ 24-3 60-30 O 16-0 39-70 40-3 10000 BASIC RADICALS OF THE THIRD GROUP. T5. 1. ALUMINIUM. Aluminium is usually precipitated as HYDROXIDE, occasionally as BASIC ACETATE or BASIC FORMATE, and always weighed as ALUMINIUM OXIDE. a. Aluminium hydroxide, recently precipitated from a solu- tion of an aluminium salt by an alkali is translucent, and when dried at 100 has the formula, A1 3 (OH) 6 . The precipitate inva- riably retains a minute proportion of the acid with which the aluminium was previously combined, as well as of the alkali which has served as the precipitant ; it is freed with difficulty from these admixtures by repeated washing. It is insoluble in pure water ; 180 FORMS. [ 75. but it readily dissolves in soda, potassa, and ethylamine (SONNEN- SCHEIN) ; it is sparingly soluble in ammonia, and insoluble in am- monium carbonate; presence of ammonium salts greatly diminishes its solubility in ammonia (Expt. No. 35). The correctness of this statement of mine in the first edition of the present work, has been amply confirmed since by MALAGUTI and DUEOCHEK ;* and also by experiments made by my former assistant, Mr. J. FUCHS. The former chemists state also that, when a solution of aluminium is precipitated with ammonium sulphide, the fluid may be filtered off five minutes after, without a trace of aluminium in it. FUCHS did not find this to be the case (Expt. No. 36). Aluminium hydroxide, recently precipitated, dissolves readily in hydrochloric or nitric acid ; but after filtration, or after having remained for some time in the fluid from which it has been precipitated, it does not dissolve in these acids without considerable difficulty, and long digestion. Aluminium hydroxide shrinks considerably on drying, and then presents the appearance of a hard, translucent, yellowish, or of a white, earthy mass. "When ignited, it loses water, and this loss is frequently attended with slight decrepitation, and invariably with considerable diminution of bulk. Aluminium hydroxide precipitated from a solution of aluminium in potassa or soda by ammonium chloride is milk-white, denser, easier to wash, and much less soluble in ammonia than the variety above de- scribed. When dried at 100, it has the formula A1 2 O 3 -f (II.O), (J. LowEf). b. Aluminium oxide or alumina, prepared by heating the hydroxide to a moderate degree of redness, is a loose and soft mass ; but upon the application of a very intense degree of heat, it con- cretes into small, hard lumps. At the most intense white heat, it fuses to a clear glass. Ignited alumina is dissolved by dilute acids with very great difficulty ; in fuming hydrochloric acid, it dis- solves upon long-continued digestion in a warm place, slowly, but completely. It dissolves tolerably easily and quickly by first heat- ing with a mixture of 8 parts of concentrated sulphuric acid and 3 parts of water, and then adding water (A. MITSCHERLICH^:). Ignition in a current of hydrogen gas leaves it unaltered. By fusion with potassium disulphate, it is rendered soluble in water. Upon igniting alumina with ammonium chloride, aluminium *4nn. de GMm. et de 'Phys. t 3 Ser. 17, 421. \Zeitochr.f. analyt. Chem., iv, 350. Jour. f. prakt. Chem., LXXXI, 110. 76.] BASES OF GROUP III. 181 chloride escapes; but the process fails to effect complete volatili- zation of the alumina (II. ROSE). When alumina is fused at a very high temperature, with ten times its quantity of sodium car- bonate, sodium aluminate is formed, which is soluble in water (R. RICHTER). Placed upon moist red litmus-paper, pure alumina does not change the color to blue. COMPOSITION. A1 2 54-2 53-03 O 3 48-0 46-97 102-2 100-00 c. If to the solution of a salt of aluminium, sodium carbonate or ammonium carbonate be added, till the resulting precipitate only just redissolves on stirring, and then sodium acetate or ammonium acetate poured in in abundance and the mixture boiled some time, the aluminium is precipitated almost completely as basic acetate in the form of translucent flocks, so that if the filtrate be boiled with ammonium chloride and ammonia only unweighable traces of aluminium hydroxide separate. If the quantity of sodium acetate employed be too small, the precipitate appears more granular, the filtrate would then contain a larger amount of aluminium. The precipitate cannot be very conveniently filtered and washed. In washing it is best to use boiling water, containing a little sodium acetate or ammonium acetate. The precipitate is readily soluble in hydrochloric acid. Mn~ P > _^ 284 100 79. 3. NICKEL. Xickel is precipitated as HYDROXIDE, and as SULPHIDE. It is weighed in the form of NICKELOUS OXIDE, of METALLIC NICKEL, or of anhydrous NICKELOUS SULPHATE. a. Nickelous hydroxide forms an apple-green precipitate, almost absolutely insoluble in water. "When precipitated from a solution of the chloride or sulphate, it retains some of the acid even after long washing (TEICHMANN*). It is also very difficult to remove the last traces of alkali. It dissolves with some difficulty in ammonia and ammonium carbonate, far more readily in the presence of an ammonium salt. From these solutions it is com- pletely precipitated by excess of potassa or soda ; application of heat promotes the precipitation. It is unalterable in the air ; on ignition, it passes into nickelous oxide. b. Nickelous oxide is a dirty grayish-green powder. When obtained by heating the nitrate to redness, it always contains some nickelic oxide, and requires very strong and protracted ignition for conversion into the pure green nickelous oxide (W. J. RUSSELL). It is insoluble in water, but readily soluble in hydrochloric acid. It does not affect vegetable colors. It suffers no variation of weight upon ignition with free access of air. Mixed with am- monium chloride and ignited, it is reduced to metallic nickel (II. ROSE); it is also easily reduced by ignition in hydrogen or carbon monoxide. COMPOSITION. Ni . . . 58-7 78-58 O 16-0 21-4:2 74-7 100-00 * Annal. d. Chem. u. Pharm., CLVI, 17. 190 FORMS. [ 79. c. Metallic nickel obtained by the reduction of nickelous oxide with hydrogen lias the form of a gray powder, or if the heat has been very strong, and it lias melted, it is lustrous and white like silver. It is unaltered in weight by ignition in hydrogen ; when ignited in the air it is superficially oxidized. It is attracted by the magnet. It is dissolved slowly by hydrochloric acid and dilute sulphuric acid, and readily by moderately strong nitric acid. d. Anhydrous nickelous sulphate obtained by evaporating a solution of the chloride, nitrate, &c., with sulphuric acid is yellow, soluble in water to a green fluid. The hydrous salt may be rendered anhydrous without loss of acid by cautious heating in a platinum dish, but at low redness it begins to blacken at the edges and loses acid (F. GAUHE*).' e. Ilydrated nickelous sulphide, prepared in the wet way, forms a black precipitate, insoluble in water. I must make some observations on its precipitation.! In order to precitate the nickel from a pure solution completely and with ease, ammonium chloride must be present ; it is not enough to add ammonium sulphide alone. A large quantity even of ammonium chloride produces no injurious effect. In the presence of free ammonia, on the con- trary, some nickel remains in solution. In this case, the super- natant fluid appears brown. As precipitant, colorless or light- yellow ammonium sulphide containing no free ammonia should be used, a large excess must be avoided. If the directions given are adhered to allowing to stand 48 hours the nickel may be pre- cipitated by means of ammonium sulphide, from solutions con- taining only ^nroTrro ^ ^ ie ox ^ e * ^ s ^ ne precipitate is liable to take up oxygen from the air, being transformed into sulphate, a little ammonium sulphide is mixed with the wash-water, to which also it is advisable to add ammonium chloride (less and less at last none); the filter should be kept full (Expt. Xo. 41). Brown filtrates, containing nickel sulphide in solution, may be freed from the latter by acidulation with acetic acid, and boiling some time. The sulphide falls down, and may now be filtered off. It is very sparingly soluble in concentrated acetic acid, somewhat more soluble in hydrochloric acid. It is more readily soluble still in nitric acid, but its best solvent is nitro-hydrochloric acid. It loses its water upon the application of a red heat ; when ignited in the air, it is transformed into a basic compound of nickelous oxide with sulphuric acid. Mixed with sulphur and ignited in a stream * Zeitschr. f. analyt. Chem., iv, 190. \Journ. f. prakt. Chem., LXXXII, 257. 80.] BASES OF GROUP IV. 191 of hydrogen, a fused mass remains, of pale yellow color and me- tallic lustre. This consists of Ni a S, but its composition is not perfectly constant (F. GATJHE *). On heating a solution of a nickel- ous salt with an excess of sodium thiosulphate in a sealed glass tube at 120, all the nickel will be precipitated in the course of half an hour as a sulphide (2N1C1, + 21Sa 2 S 2 O, = ]STi 2 S + 2NaCl + Na 2 S 3 O 6 ). The sulphide so obtained is black, and unchange- able in air; it may be readily washed, is almost unaffected by hydrochloric or dilute sulphuric acid, and it may be converted into nickelous sulphate by dissolving it in nitric acid and evaporating the solution with sulphuric acid (W. GIBBS f). [Xiekel may be precipitated as a sulphide, dense in form, easy to wash, and not readily oxidizing by contact with air, by proceed- ing as follows : To the solution, which should be concentrated and contain a liberal quantity of ammonium salts, add ammonia (if necessary) to alkaline reaction, then acetic acid to slight acid reac- tion, also ammonium or sodium acetate, and heat to boiling. Transmit II 2 S gas through the boiling solution. Since much free acetic acid prevents complete precipitation, it is necessary some- times when much nickel is present to partially neutralize once or twice the acid set free during the process.] 80. 4. COBALT. Cobalt is weighed in the PUKE METALLIC state, or as COBALTOTJS SULPHATE. Besides the properties of these substances, we have to study also those of COBALTOUS HYDROXIDE, of the SULPHIDE, and of the TRIPOTASSIUM COBALTIC NITRITE. a. Cobaltous hydroxide. Upon precipitating a solution of a cobaltous salt with potassa, a blue precipitate (a basic salt) is formed at first, which, upon boiling with potassa in excess, exclud- ed from contact of air, changes to light red cobaltous hydroxide ; if, on the contrary, this process is conducted with free access of air, the precipitate becomes discolored, and finally black, part of the cobaltous hydroxide being converted into cobaltic hydroxide. But the hydroxide prepared in this way, retains always a certain quantity of the acid, and, even after the most thorough washing * Zeitschr. f. analyt. Chem., iv, 191. t/6., in, 389. 192 roRMS. [ 80. with hot water, also a small amount of the alkaline precipitant. The latter, however, is not enough to spoil the accuracy of the results (H. ROSE, F. GAUIIE*). Cobaltous hydroxide is insoluble in water, and also in dilute potassa ; it is somewhat soluble in very concentrated potassa, and readily in ammonium salts. When dried in the air, it absorbs oxygen, and acquires a brownish color. By strong ignition it is converted into cobaltous oxide (even if some higher oxide had formed from boiling or drying in the air) ; if cooled with exclusion of air, as in a current of carbon dioxide, pure light.brown cobaltous oxide will be left ; if cooled, on the contrary, with access of air, it is more or less changed to black protosesquioxide (cobaltoso-cobaltic oxide) (W. J. RUSSELL-)-). By ignition in a current of hydrogen, metallic cobalt is left, from which any traces of alkali may now be almost completely removed by boiling water. b. The metallic cobalt obtained according to a, or by igniting the chloride or the protosesquioxide (produced by igniting the nitrate) in hydrogen is a grayish-black powder, which is attracted by the magnet, and is more difficultly fusible than gold. If the reduction has been effected at a faint heat, the finely divided metal burns in the air to protosesquioxide of cobalt, which is not the case if the reduction has been effected at an intense heat. Cobalt does not decompose water, either at the common temperature, or upon ebullition ^-except sulphuric acid be present, in which case decomposition will ensue. Heated with concentrated sul- phuric acid, it forms cobaltous sulphate, with evolution of sulphur dioxide. In nitric acid it dissolves readily to cobaltous nitrate. c. Cobalt sulphide, produced in the wet way, forms a black precipitate, insoluble in water, alkalies, and alkali sulphides. With regard to its precipitation,^: this is effected but slowly and im- perfectly by ammonium sulphide alone ; in the presence of am- monium chloride however, it takes place quickly and completely. Free ammonia is not injurious ; it is all one, whether colorless or yellow ammonium sulphide is employed. If the directions given are observed, cobalt may be precipitated from a solution contain- ing no more than Tr ooVo"o f ^ ie protoxide. In the moist con- dition, exposed to the air, it oxidizes to sulphate. In washing it, therefore, water containing ammonium sulphide is employed, and the niter is kept full. It is advisable also to mix a little ammo- * Zeitschr.f. analyt. Chem., iv, 54. f H>., n, 471. \ Journ. f. prakt. Chem., LXXXII, 262. 80.] BASES OF GROUP IY. 193 ilium chloride with the wash-water, but its quantity should be gradually decreased, and the last water used must contain none. It is but sparingly soluble in acetic acid and in dilute mineral acids, more readily in concentrated mineral acids, and most readily in warm nitro-hydrochloric acid. Mixed with sulphur and ignited in a stream of hydrogen, we obtain a product which varies in composition according to the temperature employed. The residue is therefore not suited for the determination of cobalt (H. ROSE). Cobalt can be precipitated as sulphide completely in the presence of a very small amount of free acetic acid by hydrogen sulphide, or by heating with an excess of sodium thiosulphate in a sealed tube (\Y. GIBBS, Zeitschr.f. analyt. Chem., in, 390), in the same manner as nickel (see 79, e). Cobalt sulphide may be converted into cobaltous sulphate by heating in the air, moistening with nitric acid, evaporating with sulphuric acid and igniting. d. Cobaltous sulphate crystallizes, in combination with 7 aq., slowly in oblique rhombic prisms of a fine red color. The crystals yield the whole of the water, at a moderate heat, and are con- verted into a rose-colored anhydrous salt, which bears the applica- tion of a low red heat without losing acid. At a stronger heat the edges become black and some sulphuric acid escapes (F. GAUHE*). It dissolves rather difficultly in cold, but more readily in hot water. COMPOSITION. an' > O i _ CoO . 75-00 48-37 ' < 0- ~S0 3 .... 80-07 51-63 155-07 100-00 e. Tripotassium cobaltic nitrite. If a solution of a cobalt salt (not too dilute) is mixed with excess of potassa and then with acetic acid till the precipitate is redissolved, and a concentrated solution of potassium nitrite previously acidified with acetic acid is added, first a dirty, brownish precipitate forms which gradually turns yellow and crystalline, especially on the application of a gentle heat (X. "W". FISCHER-)-). The composition of this precipi- tate corresponds to the formula (KXO,) 6 Co a (XO,) 6 + aq = CoK 8 (XO 2 ) 6 + aq.(SADTLER).- Dried at 100 its composition is somewhat variable (STROMEYER, ERDMANN^:). It is decidedly * Zeitschr.f. analyt. Cliem., iv, 55. \Pogg. Ann., LXXII, 477. \Journ.f. prakt. Chem., cxvu, 385. 194 FORMS. [ 81. soluble in water, less in potassium acetate whether neutral or acidified with acetic acid, not in potassium acetate to which some potassium nitrite has been added, not in potassium nitrite, nor in 80-per cent, alcohol. On washing with water or solution of potas- sium acetate, unless potassium nitrite is added, nitric oxide is con- stantly evolved in small quantities. It is decomposed with separa- tion of brown cobaltic hydroxide, with difficulty by solution of potassa, with ease by soda or baryta. On being moistened with sulphuric acid and ignited (finally with addition of ammonium carbonate) it leaves 2(CoSO 4 ) + 3(K 2 SO 4 ), but there is a diffi- culty in driving off all the excess of acid without decomposing the cobaltous sulphate. The yellow salt is soluble in hydrochloric acid ; potassa precipitates the whole of the cobalt from this solu- tion as hydroxide. 5. FERROUS IRON ; and 6.' FERRIC IRON. Iron is usually weighed in the form of FERRIC OXIDE, occasion- ally as SULPHIDE. We have to study also the FERRIC HYDROXIDE, the FERRIC SUCCINATE, the FERRIC ACETATE, and the FERRIC FORMATE. a. Ferric hydroxide, recently prepared, is a reddish-brown precipitate, insoluble in water, in dilute alkalies, and in ammonium salts, but readily soluble in acids ; it shrinks very greatly on drying. When dry, it presents a brown, hard mass, with shining conchoidal fracture. If the precipitant alkali is not used in excess, the precipitate contains basic salt ; on the other hand, if the alkali has been used in excess, a portion of it is invariably carried down in combination with the ferric hydroxide, on which account ammonia alone can properly be used in analysis for this purpose. Under certain circumstances, for instance, by protracted heating of a solution of ferric acetate on the water-bath (which turns the solution from blood-red to brick-red, and makes it appear turbid by reflected light), and subsequent addition of some sulphuric acid or salt of an alkali, a reddish-brown hydrated ferric oxide is pro- duced, which is insoluble in cold acids, even though concentrated, and is not attacked even by boiling nitric acid (L. PEAN DE ST. GILLES*). Closely allied to ferric hydroxide are the highly basic salts obtained by mixing dilute cold solutions of ferric salts, best ferric chloride, with much ammonium chloride, cautiously adding am- * Journ.f prakt. C'/iem , i.xvi, 137. 81.] BASES OF GROUP IV. 195 nionium carbonate till the fluid on standing in the cold instead of becoming clear turns more turbid if anything, and then boiling. The precipitates, thus produced in the fluid which still retains its acid reaction, contain the whole of the iron present and play an important part in analytical separations. They should be washed with boiling water containing ammonium chloride, being soluble to a slight extent in pure water. They are not suitable for ignition, as ferric chloride might occasionally escape from them. b. Ferric hydroxide is, upon ignition, converted into ferric oxide. If the hydroxide has been superficially dried only, the violent escape of steam from the lumps is likely to occasion loss ; but if it has been dried as much as possible by suction and still remains moist, it may be ignited without fear of loss. Pure ferric oxide, when placed upon moist reddened litmus-paper, does not change the color to blue. It dissolves slowly in dilute, but more rapidly in concentrated hydrochloric acid ; the application of a moderate degree of heat effects this solution more readily than boiling. With a mixture of 8 parts concentrated sulphuric acid and 3 parts water, it behaves in the same manner as alumina. The weight of ferric oxide does not vary upon ignition in the air; when ignited with ammonium chloride, ferric chloride escapes. Ignition with charcoal, in a closed vessel, reduces it more or less. Strongly ignited with sulphur in a stream of hydrogen, it is trans- formed into ferrous sulphide. COMPOSITION. Fe a 111-8 69-96 O 3 48-0 30-04: 159-8 100-00 c. Ferrous -sulphide, produced in the wet way, forms a black precipitate. The following facts are to be noticed with regard to its precipitation.* Ammonium sulphide used alone, whether colorless or yellow, precipitates pure neutral solutions of ferrous salts, but slowly and imperfectly. Ammonium chloride acts very favorably ; a large excess even is not attended with inconvenience. Ammonia has no injurious action. It is all the same whether the ammonium sulphide be colorless or light yellow. If the direc- *Jou/n.f. praki. them., LXXXII, 268. 196 FORMS. [ 81. tions given are observed, iron may be precipitated by means of ammonium sulphide, from solutions containing only y^ffnnj- of ferrous, oxide. In such a case, howeve'r, it is necessary to allow to stand forty-eight hours. Since the precipitate rapidly oxidizes in contact with air, ammonium sulphide is to be added to the wash- water, and the filter kept full. It is well also to mix a little ammonium chloride with the wash-water, but the quantity should be continually reduced, and the last water used should contain none. In mineral acids, even when very dilute, the hydrated sulphide dissolves readily. Mixed with sulphur, and strongly ignited in a stream of hydrogen, anhydrous ferrous sulphide re- mains (H. ROSE). COMPOSITION. x Fe 55-90 63-54 S 32-07 36-46 87-97 100-00 d. When a neutral solution of a ferric salt is mixed with a neutral solution of an alkali succinate, a cinnamon-colored pre- cipitate of a brighter or darker tint of a basic ferric succinate is formed, Fe(OH)C 4 H 4 O 4 , succinic acid being set free. From the nature of this precipitate it must follow that, in its formation, one equivalent of succinic acid (using an excess of ammonium succinate) must be liberated, as follows : Fe,(S0 4 ), + 3(NH.),C,H 4 0, + m,O = 2Fe(OH)C.H,0, + 3(NH,),SO, + H 3 -0,H 4 O.. The free succinic acid does not exercise any perceptible sol- vent action upon the precipitate in a cold and highly dilute solu- tion, but it redissolves the precipitate a little more readily in a warm solution. The precipitate must therefore be filtered cold, if we want to guard against re- solution. Formerly the precipi- tate was erroneously supposed to consist of a normal salt, de- composable by hot water into an insoluble basic and a soluble acid compound. Basic ferric succinate is insoluble in cold, and but sparingly soluble in hot, water, containing for every equivalent of 81.] BASES OF GROUP IV. 197 succinic acid, II a 'C 4 H 4 O 4 , from 18 to 30 equivalents of Fe a O 3 . It dissolves readily in mineral acids. Ammonia, especially if warm, deprives it of the greater portion of its acid, leaving compounds which are highly basic ferric succinates e. If to a solution of a ferric salt, sodium carbonate be added in the cold, till the fluid contains no more free acid, and in con- sequence of the formation of basic salt has become deep red, but remains still perfectly clear, and then sodium acetate be poured in and the mixture boiled, the whole of the iron will be precipi- tated as basic ferric acetate. The precipitation is successfully effected in this operation by having the ferric salt solution sufficiently dilute, that the free acid be properly neutralized, and that sufficient sodium acetate be added. The duration of boiling is of little consequence ; when proper proportions have been used, one boiling-up suffices. Of course it is understood that all the iron must be ferric. Instead of sodium carbonate or acetate the corresponding ammonium salts will answer also. The precipitates may, as a rule, be filtered off and well washed without any ferric oxide passing into the nitrate ; at times, however, the reverse may happen. I should advise not to boil longer than is necessary to precipitate, then filter while hot, and to add to the boiling wash-water some sodium- or ammo- nium acetate ; this can cause no inconvenience, because the pre- cipitate is usually redissolved in hydrochloric acid, and reprecipi- tated with ammonia water. f. Instead of the sodium- or ammonium acetate used in , the corresponding formates may be used. The basic ferric formate obtained in this case is more easily washed than the basic acetate (F. SCHULZE).* *Chem. Centralblatt, 1861, 3. 198 FORMS. [ 82. BASIC RADICALS OF THE FIFTH GROUP. 82. 1. SILVER. Silver may be weighed in the METALLIC state, as CHLORIDE, SUL- PHIDE, or CYANIDE. a. Metallic silver, obtained by the ignition of salts of silver with organic acids, &c., is a loose, white, glittering mass of metallic lustre ; but, when obtained by reducing silver chloride, &c., in the wet way, by zinc, it is a dull-gray powder. It fuses at about 1000. Its weight is not altered by moderate ignition. It may, however, be distilled by the heat of the oxy hydrogen name (CHRISTOMANOS*). It dissolves readily and completely in dilute nitric acid. 1). Silver chloride, recently precipitated, is white and curdy. On shaking, the large spongy flocks combine with the smaller particles, so that the fluid becomes perfectly clear. This result is, however, only satisfactorily effected when the flocks have been recently precipitated in presence of excess of silver solution (com- pare Gr. J. MULDER t). Silver chloride is in a very high degree insoluble in water, and in dilute nitric acid ; strong nitric acid, on the contrary, does dissolve a trace. Hydrochloric acid, especially if concentrated and boiling, dissolves it very perceptibly. Accord- ing to PIERRE, 1 part of silver chloride requires for solution 200 parts of strong hydrochloric acid and 600 parts of a dilute acid, composed of 1 part strong acid and 2 parts water. On sufficiently diluting such a solution with cold water the silver chloride is pre- cipitated so completely that the filtrate is not colored by hydrogen sulphide. Silver chloride is insoluble, or very nearly so, in con- centrated sulphuric acid ; in the dilute acid it is as insoluble as in water. In a solution of tartaric acid silver chloride dissolves per- ceptibly on warming ; on cooling, however, the solution deposits the whole, or, at all events, the greater part of it. Aqueous solu- tions of chlorides (of sodium, potassium, ammonium, calcium, zinc, * Zeitschr.f. analyt. Chem., vn, 299. f Die Silberprobirmethode, translated into German by D. CHR. GRIMM, pp. 19 and 311. Leipzig : J. J. Weber. 1859 . 82.] BASES OF GROUP V. 199 "SO, 80-07 26-43 302-99 10000 e. Lead chloride obtained by precipitation is a white crystalline powder. It separates in needles from a hot solution containing a certain quantity of hydrochloric acid ; occasionally it presents wedge-shaped crystals, or when separated from a strong hydro- chloric solution, hexagonal tables. At 17-7 water dissolves 0*946 * Journ. /. prakt. Ghem., LXXIV, 348. \Pogg. AnnaL, xcv, 426. \ Journ. f. prakt. Chem., LXII, 381. Zeitschr.f. analyt. Chem., vn, 244. 204 FORMS. [ 83. per cent. ; a fluid containing 15 per cent, of hydrochloric acid of 1-162 sp. gr. dissolves 0*09; a fluid containing 20 per cent, acid dissolves 0*111 per cent. ; a fluid containing 80 per cent, acid dis- solves 1 '498 per cent. Pure hydrochloric acid of the above strength dissolves 2-9 per cent. (J. CARTER BELL*). Lead chloride is less soluble in water containing nitric acid than in water (1 part requires 1636 parts, BISCHOF). It is extremely sparingly soluble in alcohol of 70 to 80 per cent., and altogether insoluble in absolute alcohol. It is unalterable in the air. It fuses at a temperature below red heat, without loss of weight. When exposed to a higher temperature, with access of air, it volatilizes slowly, being partially decomposed : chlorine gas escapes, and a mixture of lead oxide and chloride remains. COMPOSITION. Pb 206-92 74-47 CL 70-90 ' 25-53 277-82 100-00 f. Lead sulphide, prepared in the wet way, is a black precipi- tate, insoluble in water, dilute acids, alkalies, and alkali sulphides. In precipitating it from a solution containing free hydrochloric acid, it is necessary to dilute plentifully, otherwise the precipitation, will be incomplete. Even if a fluid only contain 2*5 per cent. HC1, the whole of the lead will not be precipitated (M. MARTIN f). It is unalterable in the air ; it cannot be dried at 100 without decomposition. According to H. ROSE it increases perceptibly in weight by oxidation ; in the case of long-protracted drying even becoming a few per-cents heavier.^ I have confirmed his state- ment (see Expt. No. 48). If lead sulphide mixed with sulphur is heated gently in a current of hydrogen, so that the lower quarter of the crucible is red hot, lead sulphide is left without loss of weight. By continuing a gentle heat the weight gradually dimin- ishes; by strong ignition the loss is rapid. This loss is partly owing to volatilization of lead sulphide, but mainly to escape of sulphur in the form of hydrogen sulphide and formation of Pb 2 S, or even of lead (A. SOUCIIAY ). It dissolves in concentrated hot * Journ.f. prakt. Chem , cv, 188; Jour. Chem. Soc. (2), vi, 355. \Journ.f. prakt. Chem., LXVII, 374. \ Pogg. Annal., xci, 110, and ex, 134,. . /. analyt. Chem., rv, 63. 84.] .BASES OF GROUP V. 205 hydrochloric acid, with evolution of hydrogen sulphide. In mod- erately strong nitric acid lead sulphide dissolves, upon the applica- tion of heat, with separation of sulphur ;- if the acid is rather con- centrated, a small portion of lead sulphate is also formed. Fuming nitric acid acts energetically upon lead sulphide, and converts it into sulphate without separation of sulphur. COMPOSITION. Pb 206-92 86-58 S 32-07 13-42 238-99 100-00 g. For the composition and properties of lead chromate, see Chromic acid, 93. 84. 3. MERCURY IN MERCUROUS COMPOUNDS ; and 4. MERCURY IN MERCURIC COMPOUNDS. Mercury is weighed either in the METALLIC STATE, as MERCUROUS CHLORIDE, or as SULPHIDE, or occasionally as MERCURIC OXIDE. a. Metallic mercury is liquid at .the common temperature ; it has a tin- white color. When pure, it presents a perfectly bright surface. It is quite unalterable in the air at the common tempera- ture. It boils at 360. It evaporates, but very slowly, at the ordinary temperature of summer. Upon long-continued boiling with water, a small portion of mercury volatilizes, and traces escape along with the aqueous vapor, whilst a very minute proportion remains suspended (not dissolved) in the water (comp. Expt. No. 49). This suspended portion of mercury subsides completely after long standing. When mercury is precipitated from a fluid, in a very minutely divided state, the small globules will readily unite to a large one if the mercury be perfectly pure ; but even the slightest trace of extraneous matter, such as fat, etc., adhering to the mercury will prevent the union of the globules. Mercury does not dissolve in hydrochloric acid, even in concentrated ; it is barely soluble in dilute cold sulphuric acid, but dissolves readily in nitric acid. J. Mercurous chloride, prepared in the wet way, is a heavy 206 FORMS. [ 84. white powder. It is almost absolutely insoluble in cold water ; in boiling water it is gradually decomposed, the water taking up chlorine and mercury ; upon continued boiling, the residue acquires a gray color. Highly dilute hydrochloric acid fails to dissolve it at the common temperature, but dissolves it slowly at a higher temperature ; upon ebullition, with access of air, the whole of the mercurous chloride is gradually dissolved ; the solution contains mer- curic chloride (Hg 2 Cl 2 + 2HC1 + O = 2HgCl 2 + H 2 O). When acted upon by boiling concentrated hydrochloric acid, it is rather speedily decomposed into mercury, which remains undissolved, and mer- curic chloride, which dissolves. Boiling nitric acid dissolves it to mercuric chloride and nitrate. Chlorine water and nitrohydrochlo- ric acid dissolve it to mercuric chloride, even in the cold. Solutions of ammonium chloride, sodium chloride, and potassium chloride, decompose it into metallic mercury and mercuric chloride, which latter dissolves ; in the cold, this decomposition is but slight ; heat promotes the action. It is soluble in hot solution of mercurous nitrate, and still more in that of mercuric nitrate; on cooling it crystallizes out almost completely (DERRAY*). It does not affect vegetable colors ; it is unalterable in the air, and may be dried at 100, without loss of weight ; when exposed to a higher degree of heat, though still below redness, it volatilizes completely, without previous fusion. COMPOSITION. Hg, 400-0 84-94 C1 3 70-9 15-06 470- 9 100-00 c. Mercuric sulphide, prepared in the wet way, is a black pow- der, insoluble in water. Dilute hydrochloric acid and dilute nitric acid fail to dissolve it, hot concentrated nitric acid scarcely attacks it, boiling hydrochloric acid has no action on it. By prolonged heating with red fuming nitric acid it is finally converted into a white compound, 2HgS + Hg(NO 3 ) 2 , which is insoluble, or barely soluble, in nitric acid. It dissolves readily in nitrohydrochloric acid. From a solution of mercuric chloride containing much free hydrochloric acid, the whole of the metal cannot be precipitated aa * Compt. Rend , LXX, 995. 84.] BASES OF GROUP V. 207 sulphide by means of hydrogen sulphide, until the solution is prop- erly diluted. Should such a solution be very concentrated, mer- curous chloride and sulphur are precipitated (M. MARTIN*). Solu- tion of potassa, even boiling, fails to dissolve it. It dissolves in potassium sulphide, but readily only in presence of free alkali. It is insoluble in potassium hydrosulphide and in the corresponding sodium compound, and is therefore precipitated from its solution in potassium or sodium sulphide by hydrogen sulphide or by ammonium hydrosulphide (C. BAKFOEDf). Small but distinctly perceptible traces dissolve on cold digestion with yellowish or yel- low ammonium sulphide, but after hot digestion it is scarcely possi- ble to detect any traces in solution.;): Potassium cyanide and sodium sulphite do not dissolve it. On account of the solubility of mer- curic sulphide in potassium sulphide, it is impossible to precipitate mercury by means of ammonium sulphide completely from solutions containing potassium or sodium hydroxides or carbonates. Such solutions may occur, for instance, when a solution of mercuric chloride contains much potassium chloride, or sodium chloride, for, in this case, no mercuric oxide would be precipitated on the addi- tion of potassa or soda (H. ROSE). In the air it is unalterable, even in the moist state, and at 100. When exposed to a higher temperature, it sublimes completely and unaltered. COMPOSITION. Hg 200-00 8618 S 32-07 13-82 232-07 100-00 d. Mercuric oxide, prepared in the dry way, is a crystalline brick-colored powder, which, when exposed to the action of heat, changes to the color of cinnabar, and subsequently to a violet-black tint. It bears a tolerably strong heat without decomposition ; but when heated to incipient redness, it is decomposed into mercury and oxygen ; perfectly pure mercuric oxide leaves no residue upon ignition. Its escaping fumes also should not redden litmus-paper. "Water takes up a trace of mercuric oxide, acquiring thereby a very weak alkaline reaction. Hydrochloric or nitric acid dissolves it readily. *Joun.f. prakt. Chem., LXVII, 376. f Zeitschr. f. analyt. Chem., iv, 436. % lb., m, 140. $P0ffff- Annal., ex, 141. 208 FORMS. [ 85. COMPOSITION. Hg 200 92-59 O 16 7-4:1 216 100-00 5. COPPER. Copper is usually weighed in the METALLIC STATE, or in the form of CUPRIC OXIDE, or of CUPROUS SULPHIDE. Besides these forms, we have to examine CUPRIC SULPHIDE, CUPROUS OXIDE, and CUPROUS SULPHOCYANATE. a. Copper, in the pure state, is a metal of a peculiar well- known color. It fuses only at a white heat. Exposure to dry air,, or to moist air, free from carbon dioxide, leaves the fused metal unaltered; but upon exposure to moist air impregnated with carbon dioxide, it becomes gradually tarnished and coated with a film, first of a blackish-gray, finally of a bluish-green color. Pre- cipitated finely divided copper, in contact with water and air, oxidizes far more quickly, especially at an elevated temperature. On igniting copper in the air, it oxidizes superficially to a varying mixture of cuprous and cupric oxide. In hydrochloric acid, in the cold, it does not dissolve if air be excluded ; in the heat it dissolves but slightly if the metal is in a compact state. Finely divided copper on the contrary dissolves slowly when heated with strong hydrochloric acid, hydrogen being evolved and cuprous chloride being formed (WELTZIEN*). Copper dissolves readily in nitric acid. In ammonia it dissolves slowly if free access is given to the air ; but it remains insoluble if the air is excluded. Metallic copper brought into contact in a closed vessel with solution of cupric chloride in hydrochloric acid, reduces the cupric to cuprous chloride, an atom of metal being dissolved for every molecule of chloride. b. Cupric oxide. If a dilute, cold, aqueous solution of a cupric salt is mixed with solution of potassa or soda in excess, a light blue precipitate of cupric hydroxide, Cu(OH) 2 , is formed, which it is found difficult to wash. If the precipitate be left in the fluid * Ann. d. Chem. u. Pkarm., cxxxvi, 109. 85.] BASES OF GROUP V. 209 from wliicli it lias been precipitated, it will, even at a summer lieat, gradually change to brownish-black, passing, with separation of water, into 6CuO -f- H a O (SOUCHAY). This transformation is immediate upon heating the fluid nearly to boiling. The fluid iiltered off from the black precipitate is free from copper. It follows from this that the black precipitate is insoluble in dilute potassa. Concentrated potassa or soda on the contrary dissolves the hydroxide, and on long warming even the black oxide (O. Low*). The resulting blue solutions remain clear on boiling, even if mixed with some water ; but if boiled after being much diluted the whole of the copper will separate as. black oxide. If a solution of a cnpric salt contains non-volatile organic substances, the addition of alkali in excess will, even upon boiling, fall to precipitate the whole of the copper. The hyd rated cupric oxide, 6CuO -f- H 2 O, precipitated with potassa or soda from hot dilute solutions obsti- nately retains a portion of the precipitant ; it may, however, be completely freed from this by washing with boiling water. The precipitated oxide after ignition, or the oxide prepared by decom- posing cupric carbonate or nitrate by heat, is a brownish-black, or black powder, the weight of which remains unaltered even upon strong ignition over the gas- or spirit-lamp, provided all reducing gases be excluded (Expt. No. 50). If cupric oxide is exposed to a heat approaching the fusing point of metallic copper, it fuses, yields oxygen, and becomes Cu s O, (FAVKE and MAUMENE). It is very readily reduced by ignition with charcoal, or under the in- fluence of reducing gases ; heated in the air for a long time, the reduced metallic copper re-oxidizes. Mixed with sulphur and ignited in a current of hydrogen, towards the end strongly, cupric oxide passes into cuprous sulphide (Cu a S H. ROSE). Cupric oxide, in contact with the atmosphere, absorbs water ; less rapidly after being strongly ignited (Expt. No. 51). It is nearly insoluble in water; but it dissolves readily in hydrochloric acid, nitric acid, &Q. ; less readily in ammonia. It does not affect vegetable colors. COMPOSITION. Cu . . . . 63-6 80-00 O . . . . 16-0 20-00 79-6 100-00 * Zeitschr. /. analyt. Chem. , ix, 463. 210 FOKMS. [ 85. c. Cupric sulphide, prepared in the wet way, is a brownish- black, or black precipitate, almost absolutely insoluble in water.* When exposed to the air in a moist state, it acquires a greenish tint and the property of reddening litmus paper, cupric sulphate being formed. Hence the sulphide must be washed with water containing hydrogen sulphide. It dissolves readily in boiling nitric acid, with separation of sulphur. Hydrochloric acid dis- solves it with difficulty. This is the reason why hydrogen sulphide 'precipitates copper entirely from solutions which contain even a very large amount of free hydrochloric acid (GuuNDMANNf). Only when we dissolve a copper salt directly in pure hydrochloric acid of 1*1 sp. gr. does any copper remain unprecipitated (M. MARTINA). It does not dissolve in solutions of potassa and of potassium sulphide, particularly if these solutions be boiling; it dissolves perceptibly in colorless, and much more readily in hot yellow ammonium sulphide. Potassium cyanide dissolves the freshly pre-' cipitated sulphide readily and completely. Upon intense ignition in a current of hydrogen it is converted into pure Cu 2 S. d. If the blue solution which is obtained upon adding to solu- tion of copper tartaric acid and then soda in excess, is mixed with solution of grape sugar or sugar of milk, and heat applied, an orange-yellow precipitate of cuprous hydroxide is formed, which contains the whole of the copper originally present in the solu- tion, and after a short time, more particularly upon the applica- tion of a stronger heat, turns red, owing to the conversion of the hydroxide into anhydrous cuprous oxide (Cu 2 O). The precipitate, which is insoluble in water, retains a portion of alkali with con- siderable tenacity. When treated with dilute sulphuric acid, it gives cupric sulphate which dissolves, and metallic copper which separates. e. Cuprous sulphocyanate, Cu 2 (CNS) 2 , which is always formed when potassium sulphocyanate is added to a solution of copper, mixed with sulphurous or hypophosphorous acid, is a white precipitate in- soluble in water, as well as in dilute hydrochloric or sulphuric acid. Dried at 115, the salt retains from 1 to 3 per cent, of water, which is driven off only by heating to incipient decomposition ; it is, therefore, not well adapted for direct weighing. When * In some experiments that I made when examining the Weilbach water, I found that about 9.50000 P_art_s of. water are required to dissolve 1 part of CuS. \Journ.f. prakt. Chem., LXXIII, 241. . | Ib., LXVIT, 375. 86.] BASES OF GROUP V. 211 ignited with sulphur, with exclusion of air, it changes to Cu a S (RivoT*). When heated with hydrochloric acid and potassium chlorate, or with sulphuric acid and nitric acid, it is dissolved and suffers decomposition.^ Solutions of potassa and soda separate hydrated cuprous oxide, with formation of sulphocyanate of the alkali metal. f. Cuprous sulphide, produced by heating CuS in a current of hydrogen or Cu a (CNS) ? with sulphur, is a grayish-black crystalline mass, which may be ignited and fused without decomposition if the air is excluded. COMPOSITION. Cu a .... 127-20 80-00 S 32-07 20-00 159-27 100-00 86. 6. BISMUTH. Bismuth is weighed as OXIDE, as METAL, or as CHROMATE (Bi 2 O,2CrO 4 ). Besides these compounds, we have to study here the BASIC CARBONATE, the BASIC NITRATE, the BASIC CHLORIDE, aild the SULPHIDE. a. Bismuth trioxide, prepared by igniting the carbonate or nitrate, is a pale lemon-yellow powder which, under the influence of heat, assumes transiently a dark yellow or reddish-brown color. When heated to intense redness, it fuses, without alteration of weight. Ignition with charcoal, or in a current of carbon mon- oxide, reduces it to the metallic state. Fusion with potassium cyanide also effects its complete reduction (H. RosEf). It is in- soluble in water, and does not affect vegetable colors. It dissolves readily in those acids which form soluble salts with it. When ignited with ammonium chloride it gives metallic bismuth, the reduction being attended with deflagration. COMPOSITION. Bi 3 . . . 0, ... 416-2 48-0 89-66 10-34 464-2 100-00 *Journ.f. prakt. Ghem ., LXII, 252. jib., LXI, 188. 212 FORMS. [ 86. b. Metallic bismuth is white, with a reddish tinge, moderately hard, brittle, with a tendency to crystallize. It fuses at 264, and at a low white heat volatilizes. It does not oxidize in the air at the ordinary temperature, but with the co-operation of water it oxidizes slowly, more speedily on fusion. It dissolves in dilute nitric acid. c. Bismuth carbonate. Upon adding ammonium carbonate in excess to a solution of bismuth, free from hydrochloric acid, a white precipitate of basic bismuth carbonate (Bi 3 O z CO 3 ) is imme- diately formed; part of this precipitate, however, redissolves in the excess of the precipitant. But if the fluid with the precipitate be heated before filtration, the filtrate will be free from bismuth. (Potassium carbonate likewise precipitates solutions of bismuth completely ; but the precipitate in this case invariably contains traces of potassium, which it is very difficult to remove by wash- ing. Sodium carbonate precipitates solutions of bismuth less completely.) The precipitate is easily washed ; it is practically insoluble in water, but dissolves readily, with effervescence, in hydrochloric and nitric acids. Upon ignition it leaves the oxide. d. The basic bismuth nitrate, which is obtained by mixing with water a solution of the nitrate containing little or no free acid, presents a white, crystalline powder. It cannot be washed with pure cold water without suffering a decided alteration. It becomes more basic, while the washings show an acid reaction, and contain bismuth. If the basic salt, however, be washed with cold water containing -g-i^- of ammonium nitrate, no bismuth passes through the filter. The solution of ammonium nitrate must not be warm. These remarks only apply in the absence of free nitric acid (J. LOWE*). On ignition the basic nitrate passes into the oxide. e. Basic bismuth chloride, formed by adding much water to solution of bismuth containing hydrochloric acid or sodium chloride, is a brilliant white powder (BiOCl after drying at 100). It is insoluble in water, but dissolves in concentrated hydrochloric or nitric acid. Fused with potassium cyanide it gives metallic bismuth. * f. Bismuth chromate (Bi 2 O 3 ,2CrO 3 ), which is produced by adding potassium dichromate, slightly in excess, to a solution of *Journ.f. prakt. (JJiem., LXXJV, 341. 87.] BASES OF GROUP V. 213 bismuth nitrate as neutral as possible, is an orange-yellow, dense, readily-subsiding precipitate, insoluble in water, even in presence of some free chromic acid, but soluble in hydrochloric acid and nitric acid. It may be dried at 100-112 without decomposition (LowE*). COMPOSITION. Q Q / ' = Bi.0, . . 464-2 69-87 x Bi < > Cr0 2 ~ 2Cr 3 ' 2QO ' 2 664-4 100-00 g. Bismuth trisulphide, prepared in the wet way, is a brownish black, or black precipitate, insoluble in water, dilute acids, alkalies, alkali sulphides, sodium sulphite, and potassium cyanide. In moderately concentrated nitric acid it dissolves, especially on warming, to nitrate ? with separation of sulphur. Hence in pre- cipitating bismuth from a nitric acid solution, care should be taken to dilute sufficiently. Hydrochloric acid impedes the pre- cipitation by hydrogen sulphide only when a very large excess is present, and the fluid is quite concentrated. The sulphide does not change in the air. Dried at 100, it. continually takes up oxygen and increases slightly in weight ; if the drying is protracted this increase may be considerable (Expt. No. 52). Fused with potassium cyanide, it is completely reduced (H. ROSE). Reduction takes place more slowly by ignition in a current of hydrogen. COMPOSITION. Bi 2 416-20 81-22 S 3 96-21 18-78 512-41 100-00 87. 7. CADMIUM. Cadmium is weighed either as OXIDE or as SULPHIDE. Besides these substances, we have to examine CADMIUM CARBONATE. a. Cadmium oxide, produced by igniting the carbonate or nitrate, is a yellowish-brown or reddish-brown powder. The appli- *Journ.f. prakt. Chem., LXVII, 291 214 FOEMS. [ 87. cation of a white heat fails to fuse, volatilize, or decompose it ; it is insoluble in water, but dissolves readily in acids ; it does not alter vegetable colors. Ignition with charcoal, or in a current of hydrogen, carbon monoxide, or carburetted hydrogen, reduces it readily, the metallic cadmium escaping in the form of vapor. COMPOSITION. Cd 1124 87-54 O 16.0 12-46 128-4 100-00 b. Cadmium carbonate is a white precipitate, insoluble in water and the fixed alkali carbonates, and extremely sparingly soluble in ammonium carbonate. It loses its water completely upon drying. Ignition converts it into oxide. c. Cadmium sulphide, produced in the wet way, is a lemon- yellow to orange-yellow precipitate, insoluble in water, dilute acids, alkalies, alkali sulphides, sodium sulphite, and potassium cyanide (Expt. JTo. 53). It dissolves readily in concentrated hydrochloric acid, with evolution of hydrogen sulphide. In precipitating, there- fore, with hydrogen sulphide, a cadmium solution should not contain too much hydrochloric acid, and should be sufficiently diluted. The sulphide dissolves readily in dilute sulphuric acid on heating. It dissolves in moderately concentrated nitric acid, with separation of sulphur. It may be washed, and dried at 100 or 105, without decomposition. Even on gentle ignition in a current of hydrogen, it volatilizes in appreciable amount (II. ROSE*), partially unchanged, partially as metallic vapor. COMPOSITION. Cd . . . . 112-40 77-80 S .... 32-07 22-20 144-47 100-00 *Pogg. Annal, ex, 134. 88, 89.] METALS OF GROUP VI. 215 METALS OF THE SIXTH GROUP. 88. 1. GOLD. Gold is always weighed in the metallic state. Besides METALLIC GOLD, we have to consider the TRISULPHIDE or AURIC SULPHIDE. a. Metallic gold, obtained by precipitation, presents a blackish- brown powder, destitute of metallic lustre, which it assumes, how- ever, upon pressure or friction ; when coherent in a compact mass, it exhibits the well-known bright yellow color peculiar to it. It fuses only at a white heat, and resists, accordingly, all attempts at fusion over a spirit-lamp. It remains wholly unaltered in the air and at a red heat, and is not in the slightest degree affected by water, nor by any simple acid. Nitrohydrochloric acid dissolves it to trichloride. Hot concentrated sulphuric acid containing a little nitric acid dissolves gold, especially if in a linely divided condition, to a yellow fluid, from which it is thrown down again by water (J. SPILLEE!). b. Auric sulphide. When hydrogen sulphide is transmitted through a cold dilute solution of auric chloride, the whole of the gold separates as auric sulphide. Au 2 S 3 , in form of a brownish- black precipitate. If this precipitate is left in the fluid, it is gradually transformed into metallic gold and free sulphuric acid. Upon transmitting hydrogen sulphide through a warm solution of auric chloride, aurous sulphide Au a S precipitates, with formation of sulphuric and hydrochloric acids, thus : 4AuCl a + 3H a S + 4H 3 O = 2Au,S + 12HC1 + H a SO 4 . Auric sulphide is insoluble in water, hydrochloric acid, and nitric acid, but dissolves in nitrohydrochloric acid. Colorless am- monium sulphide fails to dissolve it ; but it dissolves almost entirely in yellow ammonium sulphide, and completely upon addition of potassa. It dissolves in potassa, with separation of gold. Yellow potassium sulphide dissolves it completely. It dis- solves in potassium cyanide. Exposure to a moderate heat reduces it to the metallic state. 2. PLATINUM. Platinum is invariably weighed in the METALLIC STATE ; it is f Cliem. News, xiv, 256; Zeitschr.f. analyt. Chem.,vi, 228. 216 FORMS. [ 90. generally precipitated as AMMONIUM PLATINIC CHLORIDE, or as POTASSIUM PLATINIC CHLORIDE, rarely aS PLATLNIC SULPHIDE. a. Metallic platinum, produced by igniting ammonium platinic chloride, or potassium platinic chloride, presents the appearance of a gray, lustreless, porous mass (spongy platinum). The fusion of platinum can be effected only at the very highest degrees of heat. It remains wholly unaltered in the air, and in the most powerful furnaces. It is not attacked by water, or simple acids, and scarcely by aqueous solutions of the alkalies. Nitrohydrochloric acid dis- solves it to platinic chloride. I). The properties of potassium platinic chloride, and those of ammonium platinic chloride, have been given already in 68 and TO respectively. c. Platinic sulphide. When a concentrated solution of pla- tinic chloride is mixed with hydrogen sulphide water, or when hydrogen sulphide gas is transmitted through a rather dilute solution of the chloride, no precipitate forms at first ; after stand- ing some time, however, the solution turns brown, and finally a precipitate subsides. But if the mixture of solution of platinic chloride, with hydrogen sulphide in excess, is gradually heated (finally to ebullition), the whole of the platinum separates as platinic sulphide (free from any admixture of platinic chloride). Platinic sulphide is insoluble in water and in simple acids ; but it dissolves in nitrohydrochloric acid. It dissolves partly in caustic alkalies, with separation of platinum, and completely in alkali sulphides, especially the polysulphides if used in sufficient excess. When hydrogen sulphide is transmitted through water holding minutely divided platinic sulphide in suspension, the sulphide, absorbing hydrogen sulphide, acquires a light grayish-brown color ;. the hydrogen sulphide thus absorbed, separates again upon exposure to t^ie air. When moist platinic sulphide is exposed to the air, it is gradually decomposed, being converted into metallic platinum and sulphuric acid. Ignition in the air reduces platinic sulphide to metallic platinum. 90. 3. ANTIMONY. Antimony is weighed as ANTIMONOUS SULPHIDE, as ANTIMONY TETROXIDE (or ANTIMONOUS ANTiMONATE), or more rarely in the METALLIC state. 90.] METALS OF GROUP VI. 217 a. Upon transmitting hydrogen sulphide through a solution of antimonous chloride mixed with tartaric acid, an orange precipi- tate of amorphous antimonous sulphide is obtained, mixed at first with a small portion of basic antimony chloride. However, if the fluid is thoroughly saturated with hydrogen sulphide, and a gentle heat applied, the chloride mixed with the precipitate is decom- posed, and pure antimonous sulphide obtained. Antimonous sulphide is insoluble in water and dilute acids ; it dissolves in con- centrated hydrochloric acid, with evolution of hydrogen sulphide. In precipitating with hydrogen sulphide, therefore, antimony solutions should not contain too much free hydrochloric acid, and should be sufficiently diluted. The amorphous antimonous sul- phide dissolves readily in dilute potassa, ammonium sulphide, and potassium sulphide, sparingly in ammonia, very slightly in ammo- nium carbonate, and not at all in hydrogen potassium sulphite. The amorphous sulphide, dried in the desiccator at the ordinary temperature, loses very little weight at 100; if kept for some time at this latter temperature its weight remains constant. But it still retains a little water, which does not perfectly escape even at 190, but at 200 the sulphide becomes anhydrous, turning black and crystalline (H. KOBE* and Expt. No. 54). Ignited gently in a stream of carbon dioxide, the weight of this anhydrous sulphide remains constant ; at a stronger heat a small amount volatilizes. The amorphous sulphide, if long exposed to the action of air, in presence of water, slowly takes up oxygen, so that on treatment with tartaric acid it yields a filtrate containing anti- mony. Antimonic sulphide is insoluble in water, also in w r ater con- taining hydrogen sulphide. It dissolves completely in ammonia, especially on warming; traces only dissolve in ammonium car- bonate. On heating dried antimonic sulphide in a current of carbon dioxide 2 atoms of sulphur escape, black crystalline anti- monous sulphide remaining. On treating antimonous or antimonic sulphide with fuming nitric acid violent oxidation sets in. We obtain first antimonic acid and pulverulent sulphur ; on evaporating to dryness antimonic acid and sulphuric acid ; and lastly on igniting antimony tetroxide. The same antimony tetroxide is obtained by igniting the sulphide *Journ.f,prakt. Chem., LIX. 381. 218 FORMS. [ 90. with 30 to 50 times its amount of mercuric oxide (BUNSEN*). [According to later investigations of BuNSEN,f the temperature necessary to reduce Sb 2 O 6 to Sb 2 O 4 lies so near that which reduces Sb 2 O 4 to Sb 2 O 3 that it is not easy to bring antimony into Sb 2 O 4 for weighing. It is possible only by using a large covered platinum or rather large open porcelain crucible (by suitable choice of size of crucible and intensity of flame) and heating with a gas blast lamp so that the bottom only of the crucible reaches a strong red heat, to drive off exactly one atom of oxygen from Sb 2 O 5 .] Ignition in, .a current of hydrogen converts the sulphides of antimony into the metallic state. COMPOSITION. Sb a . . . . 240-80 71-45 S 8 .... 96-21 28-55 337-01 100-00 b. Antimony tetroxide is a white powder, which, when heated, acquires transiently a yellow tint ; it is infusible ; it loses weight when ignited intensely in a small platinum crucible with a gas blast flame (BuNSENf). It is almost insoluble in water, and dis- solves in hydrochloric acid with very great difficulty. It undergoes no alteration on treatment with ammonium sulphide. It manifests :.an acid reaction when placed upon moist litmus-paper. COMPOSITION. Sb 2 ...... 240-8 79.00 O 4 64-0 21-00 304-8 100-00 c. Metallic antimony, produced in the wet way, by precipita- tion, presents a lustreless black powder. It may be dried at 100 without alteration. It fuses at a moderate red heat. Upon ignition in a current of gas, e.g., hydrogen, it volatilizes, without formation of antimonetted hydrogen. Hydrochloric acid has very little action on it, even when concentrated and boiling. Nitric acid converts it into antimorious oxide, mixed with more or less * Annal. de Chem. u. Pharm., cvi, 3. \Zeitschr.f.analyt. Chem., 1879, 268 91.] METALS OF GROUP VI. 219 antimony tetroxide, according to the concentration of the nitric acid. 91. 4. TIN IN STANNOUS COMPOUNDS ; and 5. TIN IN STANNIC COMPOUNDS. Tin is generally weighed in the form of STANNIC OXIDE ; be- sides stannic oxide, we have to examine STANNOUS SULPHIDE and STANNIC SULPHIDE. a. Stannic oxide. If a solution of an alkali, sodium sulphate or ammonium nitrate is added to a solution of stannic chloride, stannic acid (H a SnO 3 ) is precipitated. This precipitate is soluble in excess of soda, and does not separate again even on the addition of a large quantity of soda (C. F. BAKFOED*). It is also readily soluble in hydrochloric acid. By the action of nitric acid on metallic tin, or by evaporating a solution of tin with an excess of nitric acid, a white residue is obtained which is metastannic acid (Sn B H ]0 O 1B ?). This residue is insoluble in water, but very slightly soluble in nitric acid, or sulphuric acid. By heating with hydrochloric acid it does not dissolve, but is changed to metastannic chloride, which is soluble in water after removal of the excess of hydrochloric acid. Soda added to a solution of metastannic chloride precipitates sodium inetastannate, which is insoluble in excess of soda and in weak alcohol, whereas when added to ordinary stannic-chloride solution, it affords a precipitate which is soluble in excess, and is not reprecipitated by even a very large excess (C. F. BARFOED*). Upon intense ignition, both stannic and metastannic acids are converted into stannic oxide. Mere heating to redness is not sufficient to expel all the water (DuMAsf). Stannic oxide is a straw-colored powder, which under the influence of heat, transiently assumes a different tint, varying from bright yellow to brown. It is insoluble in water and acids, and does not alter the color of litmus-paper. Mixed with ammonium chloride in excess, and ignited, it volatilizes completely as stannic chloride. If stannic oxide is fused with potassium cyanide, all the tin is obtained in form of metallic globules, which may be com- pletely, and without the least loss of metal, freed from the adhering slag, by extracting with dilute alcohol, and rapidly decanting the fluid from the tin globules (IT. ROSE;}:). * Zeitschr.f. analyt. Chem., vu, 260. f Annal. d. Chem. u. Pharm., cv, 104. \Journ.f. prakt. Chem., LXI, 189. 220 FORMS. [91. COMPOSITION. Sn 119 78-81 O a 32 21-19 151 100-00 b. Hydrated stannous sulphide forms a brown precipitate, insoluble in water, hydrogen sulphide water, and dilute acids. In precipitating tin from stannous solutions by means of hydrogen sulphide, free hydrochloric acid must not be present in too large amount, and the solution must be diluted sufficiently. Ammonia fails to dissolve it ; but it dissolves pretty readily in yellow ammo- nium sulphide, and in yellow potassium sulphide ; it dissolves readily in hot concentrated hydrochloric acid. Heated, with exclu- sion of air, it loses its water, and is rendered anhydrous ; when ex- posed to the continued action of a gentle heat, with free access of air, it is converted into sulphur dioxide, which escapes, and stannic oxide, which remains. c. Hydrated stannic sulphide, precipitated by acids from the solution of its alkali sulphur salts, is a light-yellow precipitate. In washing with pure water, it is inclined to yield a turbid filtrate and to stop up the pores of the filter ; this annoyance is got over by washing with water containing sodium chloride, ammonium acetate, or the like (BUNSEN). On drying, the precipitate assumes a darker tint. It is insoluble in water; it dissolves with difficulty in ammonia, but readily in potassa, alkali sulphides, and hot con- centrated hydrochloric acid. It is insoluble in hydrogen potassium sulphite. In precipitating tin from stannic solutions by hydrogen sulphide, the solution should not contain too much free hydro- chloric acid, and should be sufficiently diluted. According to C. F. BARFOED* the precipitates thus produced are not pure hydrated stannic sulphide, but a mixture of this with stannic or metastannic acid, as the case may be. The precipitate thrown down from ordinary stannic chloride keeps its yellow color even after long standing in the fluid, and dissolves completely in excess of soda ; that thrown down from the metastannic chloride is first white and becomes gradually yellow, it turns brown on standing in the fluid and dissolves in excess of soda, leaving, however, a considerable residue of sodium metastannate. When heated, with * Zeitschr. f. analyt. Cliem. , vii, 261. 92.] METALS OF GROUP VI. 221 exclusion of air, stannic sulphide loses its water of hydration, and, at the same time, according to the degree of heat, one-half or one- fourth of its sulphur, becoming converted either into stannous sulphide or the sesquisulphide of tin ; when heated very slowly, with free access of air, it is converted into stannic oxide, with dis- engagement of sulphur dioxide. 92. 6. ARSENIC OF ARSENOUS COMPOUNDS; and 7. ARSENIC OF ARSENIC COMPOUNDS. ARSENIC is weighed either as LEAD ARSENATE, as ARSENOUS SULPHIDE, as AMMONIUM MAGNESIUM ARSENATE, as MAGNESIUM PYRO- ARSENATE, or as URANYL PYROARSENATE ; besides these forms, we have here to examine also ARSENIO-MOLYBDATE OF AMMONIUM. a. Lead arsenate, in the pure state, is a white powder, which agglutinates when exposed to a gentle red heat, at the same time transitorily acquiring a yellow tint ; it fuses w r hen exposed to a higher degree of heat. When strongly ignited, it suffers a slight diminution of weight, losing a small proportion of arsenic acid, which escapes as arsenous oxide and oxygen. In analysis we have never occasion to operate upon the pure lead arsenate, but upon a mixture of it with lead oxide. b. Arsenous sulphide forms a precipitate of a rich yellow color; it is insoluble in water,* and also in hydrogen-sulphide water. When boiled with water, or left for several days in con- tact with that fluid, it undergoes a very trifling decomposition : a trace of arsenous acid dissolves in the water, and a minute pro- portion of hydrogen sulphide is disengaged. This does not in the least interfere, however, with the washing of the precipitate. The precipitate may be dried at 100, without decomposition ; the whole of the water which it contains is expelled at that tempera- ture. When exposed to a stronger heat, it transitorily assumes a brownish-red color, fuses, and finally rises in vapor, without decomposition. It dissolves readily in alkalies, alkali carbonates, * In some experiments which I had occasion to make, in the course of an analysis of the springs of Weilbach (Chemische Untersuchung der wichtigsten Nassauischen Mineralwasser von Dr. Fresenius, V. Schwefelquelle zu Weil bach. Weisbaden, Kreidel und Niedner. 1856), I found that one part of As a S 3 dis- solves in about one million parts of water. 222 FORMS. [ 92. alkali sulphides, potassium-hydrogen sulphite, and nitrohydro- chloric acid ; but it is scarcely soluble in boiling concentrated hydrochloric acid. Red fuming nitric acid converts it into arsenic ' acid and sulphuric acid. It is insoluble in carbon disulphide. COMPOSITION. As, 150-00 60-92 S 3 96-21 39-08 246-21 100-00 c. Ammonium magnesium ar senate forms a white, somewhat transparent, finely crystalline precipitate, which when dried in a desiccator has the formula NH 4 MgAsO 4 + 6H Q O. After drying at 100, its composition is (NH 4 MgAsO 4 ) 2 + H 2 O. At a higher temperature, say 105 110, more water escapes, and at 130 this loss is considerable (PULLER*). Upon ignition it loses water and ammonia, and changes to magnesium pyroarsenate, Mg 2 As 2 O 7 . On rapid ignition the escaping ammonia has a reducing action on the arsenic acid, and a notable loss is occasioned (H. ROSE) ; by raising the heat very gradually reduction may be avoided (H. ROSE, WiTTSTEiN,f PULLER), or by passing a current of dry oxygen during the ignition. Ammonium magnesium arsenate dissolves very sparingly in water, one part of the salt dried at 100, requir- ing 2656, one part of the anhydrous salt, 2788 parts of water of 15. It is far less soluble in amrnoniated water, one part of the salt dried at 100 requiring 15038, one part of the anhydrous salt, 15786 parts of a mixture of one part of solution of ammonia (0'96 sp. gr.), and 3 parts of water at 15. In water containing ammonium chloride, it is much more readily soluble, one part of the anhydrous salt requiring 886 parts of a solution of one part of ammonium chloride in 7 parts of water. Presence of ammonia diminishes the solvent capacity of the ammonium chloride ; one part of the anhydrous salt requires 3014 parts of a mixture of 60 parts of water, 10 of solution of ammonia (0*96 sp. gr.) and one of ammonium chloride.:): A solution of ammonium chloride, ammonia and magnesium sulphate dissolves much less of the salt than ammoniated water ; thus, PULLER (loc. cit.) found that one *Zettschr.f.analyL Chem., x, 62. \lb., n, 19. | lb., in, 206. PULLER obtained almost the same numbers (Ib. , x, 53). 92.] METALS OF GROUP VI. part of the anhydrous salt dissolved in 32827 parts of a fluid con- taining ^ of magnesia mixture (p. 113). Excess of alkali arsenate still more diminishes the solubility of the salt in water containing, ammonia and ammonium chloride (PULLER). COMPOSITION OF AMMONIUM MAGNESIUM ARSENATE DEIED AT 100. 2MgO. 80-600 21-17 4 NI > - 52-144 13-69 K^^t ' As 2 O 5 . . . 230-000 60-41 H,0 . . . 18-016 4-7S +H.O 380-760 100-00 d. Magnesium pyroarsenate, obtained by careful ignition of the preceding salt, is white, infusible by ignition in a porcelain crucible even over the blowpipe, but agglutinating at a still higher temperature, and finally fusing. After ignition in a porcelain crucible it dissolves readily in hydrochloric acid : ammonia pre- cipitates ammonium magnesium arsenate from the solution in a crystalline form. COMPOSITION. O , 2MgO . . . 80-6 25-95 V Q = \AsO < o > Mg As 3 O 6 . . . 230-0 74-05 310-6 100-00 e. Uranyl pyroarsenate. If a solution of arsenic acid is mixed with potash in slight excess, then w r ith acetic acid to strongly acid reaction, and finally with uranyl acetate, the whole of the arsenic is thrown down as UO 2 HAsO 4 -|- 4H 2 O. In the presence of salts of ammonia the precipitate also contains the whole of the arsenic, and consists of UO 2 KH 4 AsO 4 -f- water. Both precipitates are pale yellowish-green, slimy, insoluble in water, acetic acid and saline solutions, such as ammonium chloride ; soluble in mineral acids. Boiling favors the separation of the precipitate, addition of a few drops of chloroform will help it to settle, the washing is to be effected by boiling up and decanting. Both precipitates give (UO Q ) 2 As 2 O 7 . on ignition. The latter is a light yellovy residue ; if it has turned greenish from the action of reducing gases, it maybe restored to its proper color by moistening w r ith nitric acid and 224 FORMS. [ 92. re-igniting. On igniting the ammonium uranyl arsenate, the ammonia must first be expelled by cautious heating, or a current of oxygen must be passed during the ignition, otherwise the arsenic acid will be partially reduced, and arsenic will be lost (PULLER*). COMPOSITION. / AsO < > U0 2 2U0 2 . . 575-2 71-44 0< N AsO < Q > UO, As 2 O 5 . . 230-0 28-56 805-2 100-00 f. Ferric Arsenate. The white, slimy precipitate obtained when ferric chloride is treated with sodium arsenate has the follow- ing composition : 2Fe 2 .As 6 O a , -f- aq., and it is formed as in the fol- lowing reaction : 2Fe 2 Cl 6 + 6Na a HAsO 4 = 2Fe 2 .As 6 O 21 + 12NaCl -f- 3H a O. It is soluble in mineral acids, but is soluble in arsenic- acid solution only when this is highly concentrated and cold. On either heating or diluting such a solution a precipitate of ferric arsenate occurs; on cooling the solution the precipitate does not again dissolve (LUNGE *). Ferric arsenate is soluble in ammonia with yellow color. Besides this neutral compound there are others with higher iron content, e.g.) FeAsO 4 -j- 5H 2 O, precipi- tated on adding ferric acetate to arsenic acid (KOTSCHOUBET) ; 2Fe 2 .As 2 O n -f- 12H 2 O, obtained when basic ferrous arsenate is oxidized with nitric acid and ammonia added; 16Fe a .As 2 O B3 + 24H 2 O, obtained on boiling less basic compounds with excess of potassium-hydroxide solution (BERZELIUS). The last two com- pounds are insoluble in ammonia; the last is quite like ferric hydroxide. In BERTHIER'S method of estimating arsenic acid a mixture of these various salts is obtained. The more basic they are, the better r adapted they are for estimation, on account of their insolubility in ammonia. They are also then more easily washed. On being very gradually heated to redness, water alone is expelled; if the salt is strongly heated suddenly, however (before the adhering ammonia has escaped), a part of the arsenic acid is reduced to arsenous acid (H. ROSE), g. Arsenio-molybdate of ammonium. If a fluid containing * ZeitscJir. /. analyt. Chem., x, 72. 93.] ACIDS OF GROUP I. 225 arsenic acid is mixed with excess of the nitric acid solution of ammonium molybdate, the fluid remains clear in the cold, but on heating a yellow precipitate of arsenio molybdate of ammonium separates. This precipitate comports itself with solvents like the analogous compound of phosphoric acid ; it is, like the latter, insoluble in water, nitric acid, dilute sulphuric acid and salts, pro- vided an excess of solution of ammonium molybdate, mixed with acid in moderate excess, be present. Hydrochloric acid or metallic chlorides, when present in large quantity, interfere with the thoroughness of the precipitation. SELIGSOHN * found it to be composed of 87*666 per cent, of molybdic acid, 6*308 arsenic acid, 4*258 ammonia, and 1*768 water. B. FORMS IN WHICH THE ACID RADICALS ARE WEIGHED OR PRECIPITATED. ACIDS OF THE FIRST GROUP. 93. 1. AKSENOCS ACID and ARSENIC ACID. See 92. 2. CHROMIC ACID. Chromic acid is weighed either as CHROMIC OXIDE, or as LEAD OHROMATE, or BARIUM CHROMATE. We have also to consider MER- CUROUS CHROMATE. a. Chromic oxide. See 76. J. Lead chromate obtained by precipitation forms a bright-yel- low precipitate, insoluble in water and acetic acid, barely soluble in dilute nitric acid, readily in solution of potassa. When lead chro- mate is boiled with concentrated hydrochloric acid, it is readily decomposed, lead chloride and chromic chloride being formed. Addition of alcohol tends to promote this decomposition. Lead chromate is unalterable in the air. It dries thoroughly at 100. Under the influence of heat it transitorily acquires a reddish-brown tint ; it fuses at a red heat ; when heated beyond its point of fusion, it loses oxygen, and is transformed into a mixture of chro- mic oxide and basic lead chromate. Heated in contact with organic substances, it readily yields oxygen to the latter. *Journ.f. prakt. Chem., LXXVII, 481. 226 FORMS. [ 93. COMPOSITION. \Tn, PbO . . . 222-92 69-01 > "CrO, . . . 100-10 30-99 323-02 100-00 c. Barium chromate is obtained as a light-yellow precipitate on mixing a solution of an alkali chromate with barium chloride. It dissolves in hydrochloric and in nitric acid, but not in acetic acid. On washing with pure water, the latter begins to dissolve it slightly, as soon as all soluble salts are removed, to such an extent that the washings run off yellow. The precipitate is insoluble in saline solutions. Hence it is best to use a solution of ammonium acetate for washing (PEARSON and RICHARDS*). It is not decom- posed by moderate ignition. COMPOSITION, ' ' ' 153 ' 40 60 ' 51 253-50 100-00 d. Mercurous chromate obtained by adding mercurous nitrate to an alkali chromate is a brilliant-red precipitate, which turns black by the action of light. It dissolves very slightly in cold water, more in boiling water, being partially converted into a mer- curic salt ; it dissolves slightly in dilute nitric acid. For washing, it is best to use a dilute solution of mercurous nitrate containing but little free acid ; in this solution it is insoluble (H. ROSE f). 3. SULPHURIC ACID. Sulphuric acid is determined best in the form of BARIUM SUL- PHATE, for the properties of which see 71. 4. PHOSPHORIC ACID. The principal forms into which phosphoric acid is converted are as follows : LEAD PHOSPHATE, MAGNESIUM PYROPHOSPHATE, MAGNE- SIUM PHOSPHATE Mg s (PO 4 ) a , FERRIC PHOSPHATE, URANYL PYROPHOS* * Zeitschr.f. analyt. Chem., ix, 108. \Pogg. Ann., LIU, 124. 93.] ACIDS OF GROUP I. 227 PHATE, STANNIC PHOSPHATE, 1111(1 SILVER PHOSPHATE. Besides these compounds, we have to examine MERCUROUS PHOSPHATE and PHOSPHO-MOLYBDATE OF AMMONIUM. a. The lead phosphate obtained in the course of analysis is rarely pure, but is generally mixed with free lead oxide. In this mixture we have accordingly the normal lead phosphate Pb 3 (PO 4 ) 2 ; in the pure state, this presents the appearance of a white powder ; it is insoluble in water, acetic acid, and ammonia. It dissolves readily in nitric acid. When heated it fuses without decomposi- tion. b. Magnesium pyrophosphate. See 74. c. Magnesium phosfjhate (Mg 3 (PO 4 ) 2 ). A mixture of this com- pound with excess of magnesia is produced by mixing a solution of an alkali phosphate, containing ammonium chloride, with magnesia, evaporating, heating until the ammonium chloride is expelled, and finally treating with water. It is practically insoluble in water and in solutions of salts of the alkalies (FR. SCHULZE *). d. Ferric phosphate. If a solution of phosphoric acid or of calcium phosphate in acetic acid is carefully precipitated with a solution of ferric acetate, or with a mixture of iron-alum and sodium acetate, so that the iron salt may just predominate, the pre- cipitate always contains 1 mol. P 2 O 5 to 1 mol. Fe 2 O 3 corresponding to the formula of normal ferric phosphate, Fe 2 (PO 4 ) 2 (RAWSKY, WITTSTEIN, E. DAVY-f); if, on the other hand, the ferric acetate is in larger excess, the precipitate is more basic. WITTSTEIN obtained, by using a considerable excess of ferric acetate, a precipitate con- taining 3P 3 O B to 4Fe a O 3 . Precipitates obtained with a small excess of the precipitant possess a composition varying between the above- mentioned limits. RAMMELSBERG obtained Fe 2 (PO 4 ) 2 -f- 4H 2 O, and WITTSTEIN subsequently the same compound (with 8H 2 O instead of 4) upon mixing ferric sulphate with sodium phosphate in excess ; with an insufficient quantity of sodium phosphate the latter chem- ist obtained a more yellowish precipitate which had a composition corresponding to the formula 3Fe 2 (PO 4 ) 2 + Fe 2 (OH) 6 + 8H 2 O. If an acid fluid containing a considerable excess of phosphoric acid is mixed with a small quantity of a ferric solution, and an alkali * Journ.f. prakt. Chem. t LXIII, 440. \Phil. Mag., xix, 181. 228 FORMS. [ 93. acetate is added, a precipitate of the formula, Fe 2 (PO 4 ) 2 -f- water, is invariably obtained, which accordingly leaves upon ignition Fe, (PO 4 ) 2 = Fe 2 O 3 -|- P 2 O 6 (WITTSTEIN). Fresh experiments which I have made upon this subject have convinced me of the perfect correctness of this statement. MOHR obtained the same results.* The precipitate is insoluble in a fluid containing salts, but when washing, as soon as the soluble salts are nearly remoA r ed, the pre- cipitate begins to dissolve. The filtrate has an acid reaction, and contains iron and phosphoric acid. The precipitate, under these circumstances, alters in composition, and this explains why different results were obtained in the analysis of precipitates which had been. washing for different lengths of time (Fn. MOHB). COMPOSITION. 142-0 47-05 159-8 52-95 301-8 100-00 If we dissolve ferric phosphate in hydrochloric acid, supersatu- rate the solution with ammonia, and apply heat, we obtain more basic salts, viz., 3Fe 2 O 3 ,2P a O 5 (RAMMELSBEKG) ; 2Fe 2 O 3 ,P 2 O B (WITT- STEIN after long washing). In WITTSTEIN'S experiment, the wash- water contained phosphoric acid. The white ferric phosphate does not dissolve in acetic acid, but it dissolves in a solution of ferric acetate. Upon boiling the latter solution (of ferric phosphate in ferric acetate), the whole of the phosphoric acid precipitates, with basic ferric acetate, as hyperbasic ferric phosphate. Similar extremely basic combinations are invariably obtained (often mixed with ferric hydroxide), upon precipitating with ammonia or barium carbonate a solution containing phosphoric acid and an excess of a ferric salt. The precipitate obtained by barium carbonate can be conveniently filtered off and washed, the filtrate is perfectly free from either iron or phosphoric acid ; on the contrary, the precipi- tate obtained by ammonia, especially if the latter were much in excess, is slimy, and therefore difficult to wash, and the filtrate always contains small traces of both iron and phosphoric acid. * ZeitscJir.f. analyt. Chem., n, 250. 93.] ACIDS OF GROUP I. 229 e. Uranyl pyrophosphate. If the hot aqueous solution of a phosphate soluble in water or acetic acid is mixed, in presence of free acetic acid, with uranyl acetate, a precipitate of uranyl hydro- gen phosphate is immediately formed. If the fluid contains much ammonium salt, the precipitate contains also uranyl ammonium phosphate. The same precipitate forms also if aluminium or ferric salts are present ; but in that case it is always mixed with more or less aluminium or ferric phosphate. Presence of potassium or sodium salts, on the contrary, or of salts of the alkali-earth metals, has no influence on the composition of the precipitate. Aimnonium- uranyl phosphate (UO 2 NII 4 P0 4 -f- a?H,O) is a somewhat gelatinous, whitish-yellow precipitate, with a tinge of green. The best way of washing it, at least so far as the principal part of the operation is concerned, is by boiling with water and decanting. If, after having allowed the fluid in which the precipitate is suspended to cool a little, a few drops of chloroform are added, and the mixture is shaken or boiled up, the precipitate subsides much more readily than without this addition. The precipitate is insoluble in water and in acetic acid ; but it dissolves in mineral acids ; ammonium acetate, added in sufficient excess, completely re-precipitates it from this solution, upon appli- cation of heat. Upon igniting the precipitate, no matter whether containing ammonium or not, uranyl pyrophosphate of the for- mula (UO 2 ) 2 P 2 O 7 is produced. This has the color of the yolk of an egg. If the precipitate is ignited in presence of charcoal or of some reducing gas, partial reduction to uranous phosphate ensues, owing to which the ignited mass acquires a greenish tint ; however, upon warming the greenish residue with some nitric acid, the green ura- nous salt is readily reconverted into the yellow uranyl salt. Uranyl pyrophosphate is not hygroscopic, and may therefore be ignited and weighed in an open platinum dish (A. ARENDT and W. KNOP*). 2UO 3 O . . 575-2 80-20 O ^ w " \_ () _ P a O 5 . . . 142-0 19-80 717-2 100-00 * Chemisches Centralblatt, 1856, 769, 803; and 1857, 177. 230 FORMS. [ 93. The one-fifth part of the precipitate may accordingly be cal- culated as phosphoric anhydride in ordinary analyses.* f. Stannic phosphate is never obtained in the pure state in the analytical process, but contains always an admixture of hydrated metastannic acid in excess, which, upon ignition, changes to meta- stannic acid. It has, generally speaking, the same properties as hydrated metastannic acid, and is more particularly, like the latter, insoluble in nitric acid. Upon heating with concentrated solution of potassa, potassium phosphate and metastannate are formed. g. Normal silver phosphate is a yellow powder ; it is insoluble in water, but readily soluble in nitric acid, and also in ammonia. In ammonium salts, it is difficultly soluble. It is unalterable in the air. Upon ignition, it acquires transiently a reddish-brown color ; at an intense red heat, it fuses without decomposition. 695-52 83-04: 16-96 837-52 10000 h. Mercurous phosphate. This compound is employed for the purpose of effecting the separation of phosphoric acid from many bases, after H. ROSE'S method. Mercurous phosphate presents the appearance of a white crys- talline mass, or of a white powder. It is insoluble in water, but dissolves in nitric acid. The action of a red heat converts it into fused mercuric phosphate, with evolution of vapor of mercury. Upon fusion with alkali carbonates, alkali phosphates are pro- duced, and mercury, oxygen, and carbon dioxide escape. i. Phospho-molybdate of ammonium. This compound also serves to effect the separation of phosphoric acid from other bodies ; it is of the utmost importance in this respect. The composition of ammonium phospho-molybdate is variable ; * The atomic weight of uranium is here taken as 239 '6, according to Clark (0 = 16). If we take it according to Peligot, as 240, the ignited phosphate would contain 80-22 UO 3 and 19 '78 P,O 6 . W. Knop and Arendt found in four experiments 2013, 20'06, 20*04, and 20'04 respectively (in another 20*77). It will be seen that these numbers agree better with the composition as reck- oned from Ebelmen's atomic weight for uranium, 237 - 6, than from Peligot's atomic weight. 93.] ACIDS OF GEOUP I. 231 it is usually given as 2(KH 4 ),P0 4 .22MoO, + 12H a O. It forms a bright yellow, readily subsiding precipitate. Dried at 100, it lias, according to SELIGSOHN, the following (average) com- position : MoO 3 90-744: P 2 O 6 3-142 (XH 4 ) 2 3-570 H n O ' 2-544 100-000* Iii the pure state, it dissolves but sparingly in cold water (1 in 10000 EGGERTZ) ; but it is soluble in hot water. It is readily soluble even in the cold, in caustic alkalies, alkali carbonates and phosphates, ammonium chloride, and ammonium oxalate. It dis- solves sparingly in ammonium sulphate, potassium nitrate, and potassium chloride ; and very sparingly in ammonium nitrate. It is soluble in potassium sulphate and sodium sulphate, sodium chloride and magnesium chloride, and sulphuric, hydrochloric and nitric acids (concentrated and dilute). Water, containing 1 per cent, of common nitric acid, dissolves -^g^ir (EGGERTZ). Appli- cation of heat does not check the solvent action of these substances. Presence of ammonium molybdate totally changes its deportment with acid fluids. Dilute nitric or sulphuric acid containing ammonium molybdate does not dissolve it ; but much hydro- chloric acid, even in the presence of ammonium molybdate, has a solvent action, and this acid consequently interferes with the complete precipitation of phosphoric acid by nitric acid solution of ammonium molybdate. The solution of the phospho-molybdate of ammonium in acids is probably attended, in all cases, with decomposition and separation of the molybdic acid, which cannot take place in the presence of ammonium molybdate (J. CRAW)-)-. Tartaric acid and similar organic substances entirely prevent the * From the varying results of different analysts it is plain that the precipiv tate, prepared under apparently the same circumstances, has not always exactly the same composition. SONNENSCIIEIN (Journ. f. prakt. Client., LIII, 342) found in the precipitate dried at 120, 2 -93 3 -12 P;jO 6 ; LIPOWITZ (Pogg. Annal, cix, 135),' in the precipitate dried at from 20 to 30, 3'607# P a O 6 ; EGGERTZ (Journ. f. prakt. Chem., LXXIX, 496), 3 -7 to 3 -8& t Chem. Gaz. 1852, 216. 232 FORMS. [| 93. precipitation of the phospho-molybdate of ammonium (EGGERTZ). In the presence of an iodide instead of a yellow precipitate, a green precipitate or a green fluid is formed, resulting from the reducing action of the hydriodic acid on the molybdic acid (J. "W. BILL*). Other substances which reduce molybdic acid have of course a similar action. 5. BORIC ACID. POTASSIUM BOROFLUORIDE is the best form to convert boric acid into for the purpose of the direct estimation of the acid. This compound is produced by mixing the solution of an alkali borate, in presence of a sufficient quantity of potassa, with hydrofluoric acid in excess, in a silver or platinum dish, and evaporating to dry- ness. The gelatinous precipitate which forms in the cold, dissolves upon application of heat, and separates from the solution subse- quently, upon evaporation, in small, hard, transparent crystals. The compound has the formula KF,BF 3 . It is soluble in water and also in dilute alcohol ; but strong alcohol fails to dissolve it ; it is insoluble also in concentrated solution of potassium acetate. It may be dried at 100, without decomposition (AUG. STRO- MEYERf). COMPOSITION. K 39-11 30-96 B 11-00 8-71 F 4 76-20 60-33 126-31 100-00 6. OXALIC ACID. When oxalic acid is to be directly determined it is usually pre- cipitated in the form of CALCIUM OXALATE ; and its weight is inferred from the CALCIUM CARBONATE or CALCIUM OXIDE produced from the oxal^e by ignition. For the properties of these bodies see 73. 7. HYDROFLUORIC ACID. The direct estimation of hydrofluoric acid is usually effected by weighing the acid in the form of CALCIUM FLUORIDE. Calcium fluoride forms a gelatinous precipitate, which it is found difficult to wash. If digested with ammonia, previous to * Sillim. Journ., July, 1858. f Annal. d. CJiem. u. Pharm., c, 82. 93.] ACIDS OF GROUP I. filtration, it is rendered denser and less gelatinous. It is not alto- gether insoluble in water ; aqueous solutions of the alkalies fail to decompose it. It is very slightly soluble in dilute, but more readily in concentrated hydrochloric acid. When acted upon by sulphuric acid, it is decomposed, and calcium sulphate and hydro- fluoric acid are formed. Calcium fluoride is unalterable in the air, and at a red heat. Exposed to a very intense heat, it fuses. Upon intense ignition in moist air, it is slowly and partially decomposed into calcium oxide and hydrofluoric acid. Mixed with ammonium chloride,, and exposed to a red heat, calcium fluoride suffers a con- tinual loss of weight ; but the decomposition is incomplete. COMPOSITION. Ca 40.1 51-28 F a . 38-1 48.72 78 -2 100-00 We often determine fluorine, more particularly in presence of silicic acid, by converting it into silicon fluoride (SiF 4 ). This is a colorless gas, fuming in the air, with suffocating odor, of sp. gr. 3'574, which decomposes when mixed with water forming silica and hydrofluosilicic acid thus : 3SiF 4 + 2H a O = 2H 2 SiF.+ SiO a . 8. CARBONIC ACID. The direct estimation of carbonic ucid which, however, i& only rarely resorted to is usually effected by weighing the acid in the form of CALCIUM CARBONATE. For the properties of the latter substance, see 73. 9. SILICIC ACID.* When silicic acid is separated by acids from aqueous solutions- of alkali silicates, it is at first perfectly soluble in water. It be- comes insoluble or rather difficultly soluble when it coagulates. Coagulation is a permanent change and is furthered by concentra- tion and by elevation of temperature. Silicic acid solution con- taining 10 or 12 per cent, of SiO 2 coagulates at the ordinary tem- perature in a few hours, and immediately if heated. A solution of * Five silicic acid in solution is assumed to have the composition expressed by the formula Si(OH) 4 . Silicic anhydride (SiO 2 ) is usually called "silica." Compounds of SiO a with less water than corresponds to the formula Si(OH) 4 = SiO 2 (H 2 O) 3 are here called " hydrates of silica." 234 FORMS. [ 93. 5 per cent, may be preserved without coagulating for five or six: days, one of 2 per cent, for two or three months, and one of 1 per cent, for several years, and solutions containing ^ per cent, or less are not appreciably altered by time. Solid matter in powder such as graphite, hastens coagulation, alkali salts induce it rapidly. Aque- ous solutions of silicic acid may, on the contrary, be mixed with hydrochloric acid, nitric acid, acetic acid, tartaric acid and alcohol without coagulating. The gelatinous silicic acid produced by coagulation may contain more or less water, and it appears to be the more difficultly soluble in water, the less water it contains ; thus a jelly of silicic acid containing 1 per cent, of silica (SiO a ) gives a solution with cold water containing 1 part of silica in about 5000 parts, a jelly of 5 per cent, gives a solution containing 1 part of silica in about 10000 parts of water. A jelly containing less water is still less soluble, and when the jelly is dried up to a gummy mass it is> barely soluble at all ; this is also the case with the pulverulent hydrate of silica obtained in the analysis of silicates by drying a jelly containing much salts at 100 (GRAHAM*). The hydrated silica dried at 100 dissolves but very slightly in acids (with the exception of hydrofluoric acid) ; it dissolves, however, in solutions of fixed alkalies and alkali carbonates, especially on heating. Aque- ous ammonia dissolves the jelly in tolerable quantity and the dry hydrate in very notable quantity (PRiBRAM)f. Regarding the amount of water in the hydrate dried at given temperatures chem- ists do not agree. J On ignition all the hydrates pass into anhydrous silica. As the vapor escapes small particles of the extremely fine powder are liable to whirl up. This may be avoided by moistening the hydrate in the crucible with water, evaporating to dryness on a water bath, and then applying at first a slight and then a gradu- ally increased heat. The silica obtained by igniting the hydrate appears in the amorphous condition, with a sp. gr. of 2'2 to 2*3. It forms a * Fogg. AnnaL, cxui, 529. \Zeitschr.f.analyt. Chem., vi, 119. \ DOVERI (AnnaL de Chim. et de Phys., xxi, 40; AnnaL d. Chem. u. Pharm., LXIV, 256) found in the air-dried hydrate 16 '9 to 17 "8# water; J. FUCHS (AnnaL d. Chem. u. PJiarm., LXXXII, 119 to 123), 9'1 to 9-6; G. LIPPERT, 9-38 to 9-95. DOVERI found in the hydrate dried at 100, 8'3 to 9 '4; J. FUCHS, 6'63 to 6 96; G. LIPPERT, 4-97 to 5'52. H. ROSE (Pogg. AnnaL, cvm, 1; Journ. fur prakt. Chem., LXXXI, 227) found in the hydrate obtained by digesting stilbite with concentrated hydrochloric acid, and dried at 150, 4'85$ water. 94.] ACIDS OF GROUP II. 235 white powder insoluble in water, and acids (hydrofluoric excepted), soluble in solutions of the fixed alkalies and their carbonates, especially in the heat. Hydrofluoric acid readily dissolves amor- phous silica; the solution leaves no residue on evaporation in platinum, if the silica was pure. The amorphous silica, when heated with ammonium fluoride in a platinum crucible, readily volatilizes. The ignited amorphous silica, exposed to the air, eagerly absorbs water, which it will not give up at from 100 to 150 (H. ROSE). The lower the heat during ignition the more hygroscopic is the residue (SOUCHAY*). Silica fuses at the strong- est heat ; the mass obtained being vitreous and amorphous. Amor- phous silica ignited with ammonium chloride, at first loses weight, and then, when the ignition has rendered it denser, the weight remains constant. The amorphous silica must be distinguished from the crystallized or crystalline variety, which occurs as rock crystal, quartz, sand, &c. This has a sp. gr. of 2'6 (SCHAFFGOTSCH), and is far more difficultly, arid in far less amount, dissolved by potash solution or solution of fixed alkali carbonates ; it is also more slowly attacked by hydro- fluoric acid, or ammonium fluoride. Crystallized silica is not hygro- scopic, whether strongly or gently ignited (SOUCHAY). Vegetable colors are not changed either by silica or its hydrates. COMPOSITION. Si 28-4 47-02 O 2 ..... 32-0 52-98 60-4 100-00 ACID RADICALS OF THE SECOND GROUP. 94. 1. HYDROCHLORIC ACID. Hydrochloric acid is almost invariably weighed in the form of SILVER CHLORIDE for the properties of which see 82. 2. HYDROBROMIC ACID. Hydrouromic acid is always weighed in the form of SILVER BROMIDE. * Zeitschr. f. analyt. Chem., vin, 423. 236 FORMS. [ 94. Silver bromide, prepared in the wet way, forms a yellowish- white precipitate. It is wholly insoluble in water and in nitric acid, tolerably soluble in ammonia, readily soluble in sodium thio- sulphate and potassium cyanide. Concentrated solutions of potas- sium, sodium, and ammonium chlorides arid bromides dissolve it to a very perceptible amount, while in very dilute solutions of these salts it is entirely insoluble. Traces only dissolve in the alkali nitrates. It dissolves abundantly in a concentrated warm solution of mercuric nitrate. On digestion with excess of potassium iodide solution it is completely converted into silver iodide (FIELD). On ignition in a current of chlorine silver bromide is transformed into chloride ; on ignition in a current of hydrogen it is converted into metallic silver. Exposed to the light it gradually turns gray, and finally black. Under the influence of heat, it fuses to a reddish liquid, which, upon cooling, solidifies to a yellow, horn-like mass: Brought into contact with zinc and water, it is decomposed ; a spongy mass of metallic silver forms, and the solution contains zinc bromide. COMPOSITION. Ag . . . . 107-92 57-44 Br 79-95 42-56 187-87 100-00 3. HYDRIODIC ACID. Hydriodic acid is usually determined in the form of SILVER IODIDE, and occasionally also in that of PALLADIOUS IODIDE. a. Silver iodide, produced in the wet way, forms a light-yellow precipitate, insoluble in water, and in dilute nitric acid, and very slightly soluble in ammonia. One part dissolves, according to- WALLACE and LAMONT* in 2493 parts of aqueous ammonia sp. gr. 0*89 ; according to MARTINI, in 2510 parts of 0*96 sp. gr. It is copi- ously taken up by concentrated solution of potassium iodide, but it is insoluble in very. dilute; it dissolves readily in sodium .thiosul- phate and in potassium cyanide; traces only are dissolved by alkali nitrates. In concentrated warm solution of mercuric nitrate it is copiously soluble. Hot concentrated nitric and sulphuric acids- convert it, but with some difficulty, into silver nitrate and sulphate respectively, with expulsion of the iodine. Silver iodide acquires a * Chem. Gaz., lS5Q.Jahresbericht, KOPP and WILL, 1859, 670. 94.] ACIDS OF GROUP II. 237 black color when exposed to the light. When heated, it fusea without decomposition to a reddish fluid, which, upon cooling, solidifies to a yellow mass, that may be cut with a knife. Under the influence of excess of chlorine in the heat it is completely con- verted into silver chloride ; ignition in hydrogen reduces it but incompletely to the metallic state. When brought into contact with zinc and water, it is decomposed but incompletley ; zinc iodide is formed, and metallic silver separates. COMPOSITION. Ag . . . . 107-92 45-97 I 126-85 54-03 234-77 100-00 J. Palladious iodide, produced by mixing an alkali iodide with palladious chloride, is a deep brownish-black, flocculent pre- cipitate, insoluble in water and in dilute hydrochloric acid, but slightly soluble in saline solutions (sodium chloride, magnesium chloride, calcium chloride, &c.). It is unalterable in the air. Dried simply in the air it retains one molecule of water=5'05 per cent. Dried long in vacuo, or at a rather high temperature (70 to 80), it yields the whole of this water, without the least loss of iodine. Dried at 100, it loses a trace of iodine ; at from 300 to 400, the whole of the iodine is expelled. It may be washed with hot water, without loss of iodine. COMPOSITION. Pd 107-00 29-66 I 3 253-70 70-34 360-70 100-00 4. HYDROCYANIC ACID. Hydrocyanic acid, if determined gravimetrically and directly, is always converted into SILVER CYANIDE for the properties of which compound see 82. 5. HYDROSULPHURIC ACID. The forms into which the sulphur in hydrogen sulphide or metallic sulphides, is converted for the purpose of being weighed, 238 FOKMS. [ 95. are ARSENOUS SULPHIDE, SILVER SULPHIDE, COPPER SULPHIDE, and BARIUM SULPHATE. For the properties of the sulphides named, see 82, 85, 92 ; for those of barium sulphate, see 71. ACID RADICALS OF THE THIRD GROUP. 95. 1. NITRIC ACID ; and 2. CHLORIC ACID. These two acids are never determined directly that is to say, in compounds containing them, but always in an indirect way ; generally volumetrically. SECTION IV. THE DETERMINATION (OR ESTIMATION) OF RADICALS. 96. IN the preceding Section we have examined the composition and properties of the various forms and combinations in which radicals are separated from each other, or in which they are weighed. "We have now to consider the special means and methods of con- verting them into such forms and combinations. For the sake of greater clearness and simplicity, we shall, in the present Section, confine our attention to the various methods applied to effect the determination of single radicals, deferring to the next Section the consideration of the means adopted for sepa- rating them from each other. We shall here deal with the estimations of substances in the free state, or compounds consisting of one base and one acid, or containing one metal and one metalloid. As in the c ' Qualitative Analysis, ' ' the acids of arsenic will be treated of among the bases, on account of their behavior to hydro- gen sulphide ; and those elements that form acids with hydrogen will be treated of under their respective hydrogen acids. In the quantitative analysis of a compound we have to study first, the most appropriate method of dissolving it ; and, secondly, the modes of determining the quantity of one or more of its con- stituents. "With regard to the latter point, we have to turn our attention, first, to the performance and secondly, to the accuracy of the methods. It happens very rarely in quantitative analyses that the amount of a substance, as determined by the analytical process, corresponds exactly with the amount theoretically calculated or actually pres- ent ; and if it does happen, it is merely by chance. It is of importance to inquire what is the reason of this fact, and what are the limits of inaccuracy in the several methods. The cartse of this almost invariably occurring discrepancy between the quantity present and that actually found, is to be ascribed either exclusively to the execution, or it lies partly in the method itself. 239 240 DETERMINATION. [ 96. The execution of tlie analytical processes and operations can never be absolutely accurate, even though the greatest care and attention be bestowed on the most trifling minutiae. To account for this, we need only bear in mind that our weights and measures are never absolutely correct, nor our balances absolutely accurate, nor our reagents absolutely pure ; and, moreover, that we do not weigh in vacuo ; and that, even if we deduce the weight in vacua from the weight we actually obtain by weighing in the air, the very volumes on which the calculation is based are but approxi- mately known ; that the hygroscopic state of the air is liable to vary between the weighing of the empty crucible and of the cru- cible -f- the substance ; that we know the weight of a filter ash only approximately / that we can never succeed in completely keeping off dust, &c. With regard to the methods, many of them are not entirely free from certain unavoidable sources of error ; precipitates are not absolutely insoluble; compounds which require ignition are not absolutely fixed ; others, which require drying, have a slight tendency to volatilize ; the final reaction in volumetric analyses is usually produced only by a small excess of the standard fluid, which is occasionally liable to vary with the degree of dilution, the temperature, &c. Strictly speaking, no method can be pronounced quite free from defect ; it should be borne in mind, for example, that even barium sulphate is not absolutely insoluble in water. Whenever we describe any method as free from sources of error, we mean, that no causes of considerable inaccuracy are inherent in it. We have, therefore, in our analytical processes, invariably to contend against certain sources of inaccuracy which it is impossi- ble to overcome entirely, even though our operations be conducted with the most scrupulous care and with the utmost attention to established rules. It will be readily understood that several defects and sources of error may, in some cases, combine to vitiate the results ; whereas, in other cases, they may compensate one another, and thus enable us to attain a higher degree of accuracy. The comparative accuracy of the results attainable by an analytical method oscillates between two points these points are called the limits of error. In the case of methods free from sources of error, these limits will closely approach each other ; thus, for instance, in 90.] DETERMINATION. 241 the determination of chlorine, with great care one will always be able to obtain between 99'9and lOO'l for the 100 parts of chlorine actually present. Less perfect methods will, of course, exhibit far greater dis- crepancies ; thus, in the estimation of strontium by means of sul- phuric acid, the most attentive and skilful operator may not be able to obtain more than 99 (and even less) for the 100 parts of strontium actually present. I may here incidentally state that the numbers occasionally given in this manner, in the course of the present work, to denote the degree of accuracy of certain methods, refer invariably to the substance estimated (chlorine, nitrogen, baryta, for instance), and not to the combination in which that substance may be weighed (silver chloride, ammonium platinic chloride, barium sulphate, for instance) ; otherwise the accuracy of various methods would not be comparable. The occasional attainment of results exactly corresponding witli the numbers calculated does not always justify the assumption, on the part of the student, that his operations, to have led to such a result, must have been conducted with the utmost precision and accuracy. It may sometimes happen, in the course of the analyti- cal process, that one error serves to compensate another ; thus, for instance, the analyst may, at the commencement of his operations, spill a minute portion of the substance to be analyzed ; whilst, at a later stage of the process, he may recover the loss by an imperfect washing of the precipitate. As a general rule, results showing a trifling deficiency of substance may be looked upon as better proof of accurate performance of the analytical process than results exhibiting an excess of substance. As not the least effective means of guarding against error and inaccuracies in gravimetric analyses, I would most strongly recom- mend the analyst, after weighing a precipitate, <&c., to compare its properties (color, solubility, reaction, dec.) with those which it should possess, and which have been amply described in .the pre- ceding Section. In my own laboratory, I insist upon all substances that are weighed in the course of an analysis being kept between watch- glasses, until the whole affair is concluded. This affords always a chance of* testing them once more for some impurity, the presence of which may become suspected in the after-course of the process. 242 DETERMINATION. [ 97. I. DETERMINATION OF BASIC RADICALS IN SIMPLE SALTS. First Group. POTASSIUM SODIUM AMMONIUM (LITHIUM). 97. 1. POTASSIUM. a. Solution. Potassa and potassium salts of those inorganic acids which we have to consider here, are dissolved in water, in which menstruum they dissolve readily, or at all events, pretty readily. Potassium salts of organic acids it is most convenient to convert into potassium carbonate by long-continued, gentle ignition in a covered crucible. Heated to fusion, the carbon separated acts on the potassium carbonate ; carbonic oxide escapes, and some potas- sium hydroxide is formed. On simple carbonization a slight loss is caused ; on fusing, which must be avoided, a further loss occurs. b. Determination. Potassium is weighed either as potassium sulphate, ^potas- sium chloride, or as potassium -platinic chloride (see 68). It may also be determined volumetrically. For the alkalimetric estimation of potassa or potassium carbonate, see 219 and 220. For estimating potassium as potassium hydrogen tartrate, and which only gives approximate results, a chapter will be given in the Special Part. "We may convert into 1. POTASSIUM SULPHATE. Potassium salts of strong volatile acids ; e. g. , potassium chlo- ride, potassium bromide, potassium nitrate, etc., and salts of organic acids. 2. POTASSIUM NITRATE. Potassium hydroxide and compounds of potassium witli weak, volatile acids not decomposable by nitric acid, e.g., potassium carbonate (potassium salts with organic acids). 3. POTASSIUM CHLORIDE. In general, caustic potassa and potassium salts of weak volatile acids; also, and more particularly, such as are decomposed by 97.] POTASSIUM. nitric acid, e.g., potassium sulphide, potassium sulphate, chromate, chlorate, and silicate. 4. POTASSIUM PLATINIC CHLORIDE. Potassium Baits of non-volatile acids soluble in alcohol. This method is particularly important for salts of the non-volatile acids ; e.g., potassium phosphate, potassium borate; also for separating potassium from sodium. The potassium in potassium borate may be determined also as sulphate ( 136) ; and the potassium in the phosphate, as potas- sium chloride ( 135). The form of potassium platinic chloride may also be resorted to in general, for the estimation of potassium in all potassium salts of those acids which are soluble in alcohol. This form is, more- over, of especial importance, as that in which the separation of potassium from sodium, etc., is effected. 5. POTASSIUM SILICOFLUOKIDE. Potassium salts of those acids which are soluble in weak alcohol, except borate. 1. Determination as Potassium Sulphate. Evaporate the aqueous solution of the potassium sulphate to dryness, ignite the residue in a platinum crucible or dish, and weigh ( 42). The residue must be thoroughly dried before you proceed to ignite it ; the heat applied for the latter purpose must be moderate at first, and very gradually increased to the requisite degree ; the crucible or dish must be kept well covered neglect of those precautionary rules involves always a loss of substance from decrepitation. If free sulphuric acid is present, we obtain, upon evaporation, acid potassium sulphate ; in such cases the acid salt is to be converted into the normal by igniting first alone (here it is best to place the lamp so that the flame may strike the dish-cover obliquely from above), then with ammonium carbonate. See 68. For properties of the residue, see 68. Observe more particu- larly that the residue must dissolve to a clear fluid, and that the solution must be neutral. Should traces of platinum remain behind (the dish rfot having been previously weighed), these must be care- fully determined, and their weight subtracted from that of the ignited residue. 244 DETERMINATION. [ 97. With proper care and attention, this method gives accurate results. To convert the above-mentioned salts (potassium chloride, &c.) into potassium sulphate, add to their aqueous solution a quantity of pure sulphuric acid more than sufficient to form normal sulphate with the whole of the potassium, evaporate the solution to dry- ness, ignite the residue, and convert the resulting acid potassium sulphate into the normal, by treating with ammonium carbonate ( 68). As the expulsion of a large quantity of sulphuric acid is a very disagreeable process, avoid adding too great an excess. Should too little of the acid have been used, which you may infer from the non-evolution of sulphuric acid fumes on ignition, moisten the residue with dilute sulphuric acid, evaporate, and again ignite. If you have to deal with a small quantity only of potassium chloride, &c., proceed at once to treat the dry salt, cautiously, with dilute sulphuric acid in the platinum crucible ; provided the latter be capacious enough. In the case of potassium bromide and iodide, the use of platinum vessels must be avoided. Potassium salts of organic . acids are directly converted into potassium sulphate by first carbonizing them at the lowest possible temperature, and after cooling adding some crystals of pure ammo- nium sulphate and a little water to the mass. The crucible being covered, the water is evaporated by heating the crucible cover, and the whole is afterwards heated to dull redness, until the excess of ammonium sulphate is destroyed. If the carbon is not fully con- sumed by this operation, add a little ammonium nitrate and repeat the ignition. The potassium sulphate is then weighed. (KAM- MEREK. *) It is usually advisable to ignite finally in an atmosphere of ammonium carbonate. The results are accurate. 2. Determination as Potassium Nitrate. The general method is the same as in 1 . The potassium nitrate must be heated gently to the melting-point, otherwise loss will arise from evolution of oxygen. For properties of the residue see 68. The process is easily carried out, and the results are accurate. In converting potassium carbonate into the nitrate, consult 38. [*Fres. Zeit., vn, 222.] 97.] POTASSIUM. 245 3. Determination as Potassium Chloride. General method the same as described in 1. The residue of potassium chloride must, previously to ignition, be treated in the same way as potassium sulphate, and for the same reason. The salt must be heated in a well-covered crucible or dish, and only to dull redness, as the application of a higher degree of heat is likely to cause some loss by volatilization. No particular regard need be had to the presence of free acid. For properties of the residue, see 68. This method, if properly and carefully executed, gives very accurate results. The potassium chloride may, instead of being weighed, be determined volumetrically by 141, b. This method, however, has no advantage in the case of single estima- tions, but saves time when a series of estimations has to be made. In determining potassium in the carbonate it is sometimes desirable to avoid the effervescence occasioned by treatment with hydrochloric acid, as, for instance, in the case of the ignited resi- due of a potassium salt of an organic acid, which is contained in the crucible. This may be effected by treating the carbonate with solution of ammonium chloride in excess, evaporating and igniting, when ammonium carbonate and the excess of ammonium chloride will escape, leaving potassium chloride behind. . The methods of converting the potassium compounds specified above into potassium chloride, will he found in Part II. of this Section, under the respective hea ,i of the acids which they con- tain. 4. Determination as Potassium- Platinie Chloride, a. Potassium salts of volatile acids (nitric acid, acetic acid, &c.). Mix the solution with hydrochloric acid, evaporate to dryness, dissolve the residue in a little water, add a concentrated solution of platinic chloride, as neutral as possible, in excess, and evaporate in a porcelain dish, on the water-bath, nearly to dryness, taking care not to heat the water-bath quite to boiling. If the platinum con- tent of the platinum- chloride solution is known, the proper quan- tity of the latter to add is more readily used (63, 8). Add alcohol of about 80 per cent, by volume to the residue and let it stand for some time, pour the alcoholic solution through a small iilter, and treat the residue if necessary a few times with small quantities of alcohol of the same strength, until it appears to be pure potassium-platinic chloride. Bring this upon the filter and 246 DETERMINATION. [ 97. wash completely by applying repeatedly small quantities of the same alcohol. Dry next the filter and its contents in the funnel, for it is necessary that the alcohol should be completely volatilized. Transfer the contents of the filter carefully to a watch-glass, and place the filter back into the funnel and dissolve and wash out the small quantity of adhering potassium-platinic chloride with hot water. Evaporate the yellow solution thus obtained to dryness in a weighed platinum vessel. Then bring the chief quantity of the precipitate into the platinum dish and dry the whole to a constant weight at 130 C. If the quantity of potassium-platinic chloride obtained is very small, the whole may be dissolved from the filter, evaporated and dried in the same manner.* The asbestos filtering- tube described on page 108, Fig. 68, is generally to be recommended for filtering. The tube to be dried is freed from water so far as possible by suction, and then inserted into another, wider tube about 4 cm. shorter, which is fixed in the air-bath shown on page 64, Fig. 38. Air is then slowly drawn through the tube, while the air-bath is heated, towards the end at 130, for a long time. The air-current should enter the tube at the wide end, and should be dried by concentrated sulphuric acid. After the drying is effected and the tube weighed, the results may be readily controlled by converting the potassium-platinic chloride into plati- num. For this* purpose dry hydrogen is passed through the tube while it is moderately heated. After the decomposition is com- plete, the potassium chloride is leached out with water, all the water is removed by suction while the tube is being heated, and the residual platinum weighed, 1 equivalent being equal to 2 equivalents of potassium. If a paper filter is used, this must first be dried at 100, weighed, and then the- loss of weight of an aliquot portion dried at 130 determined, from which the loss of weight of the entire filter at 130 may be calculated. ft. Potassium salts of non-volatile acids (phosphoric acid, boracic acid, &c.). * When many successive determinations are to be made, especially in technical analyses, much time can be saved by using Goocn's apparatus (see pp. 120, 121) for washing and weighing the K 2 PtCl 6 . 97.] POTASSIUM. 247 Make a concentrated solution of the salt in water, add some hydrochloric acid, and platinic chloride in excess, mix with a toler- able quantity of the strongest alcohol, let the mixture stand 24 hours, after which filter, and proceed as directed in a. For properties of the precipitate see 68. This method, prop- erly executed, gives satisfactory results. Still there is generally a trifling loss of substance, potassium-platinic chloride not being absolutely insoluble even in strong alcohol. In accurate analyses, therefore, the alcoholic washings must be evaporated, with addition of a little pure sodium chloride, at a temperature not exceeding 75, nearly to dryness, and the residue treated once more with 80-per cent, alcohol. A trifling additional amount of potassium-platinic chloride is thus obtained, which is either added to the principal precipitate or collected on a separate small filter, and weighed by dissolving from the filter arid evaporating to dryness as above de- scribed. The object of the addition of a little sodium chloride to the platinic chloride is to obviate the decomposition to which pure platinic chloride is more liable upon evaporation in alcoholic solu- tion alone, than it is when mixed with sodium-platinic chloride. The atmosphere of a laboratory often contains ammonia, which might give rise to the formation of some ammonium-platinic chlo- ride, and to a consequent increase of weight in the potassium salt. As the collection of a precipitate on a weighed filter-paper is very tedious, and, where small quantities are operated upon, inaccurate as well, it is better, where the filter-tube is not used, to collect small quantities of potassium-platinic chloride (up to 0*03 grm.) in a very small, un weighed filter, dry, transfer the filter, wrapped up around the precipitate, to a small covered porcelain crucible, and slowly carbonize. The cover is then removed, the carbon burned, and the crucible allowed to cool. A very small quantity of pure oxalic acid is next added, the cover placed on, and heat applied, at first gently, finally strongly. By the addition of the oxalic acid the complete decomposition of the potassium-platinic chloride is greatly facilitated, and which is not so readily accomplished by simple ignition. Of course the oxalic acid may be replaced by a current of hydrogen. The cooled contents of the crucible are treated with water, the' residual platinum washed out until the washings give no cloudiness with silver-nitrate solution, then dried and weighed. As a rule the washing may be accomplished by simple decantation. 248 DETERMINATION. [ 98. 5. Volumetric determination as Potassium Silicofluoride. To the moderately concentrated solution of the potassium salt in a beaker add a sufficiency of hydrofluosilicic acid,* and then an equal volume of pure strong alcohol. If the potassium salt was difficultly soluble (such as potassium platinic chloride), warm it with the hydrofluosilicic acid before adding the spirit. The potas- sium silicofluoride will separate as a translucent precipitate ; when it has settled, filter, wash out the beaker with a mixture of equal parts strong alcohol and water, and wash the precipitate with the same mixture till the washings are no longer acid to litmus paper. Put the filter and precipitate into the beaker previously used, treat with water, add some tincture of litmus, heat to boiling, and add standard, or, in the case of very small quantities, decinornial, potassa or soda solution ( 215) till the fluid is just blue, and remains so after continued boiling. The reaction is as follows : (KF),SiF 4 + 4KOH = 6KF + Si(OH) 4 , consequently 2 atoms potassium in the standard solution correspond to 1 at. potassium originally present and precipitated as potassium silicofluoride (FR. STOLBA|). If the solution of the potassium salt contains much free acid, particularly sulphuric acid, this is to be removed by heat before adding the hydrofluosilicic acid. Small quantities of ammonium salts are of no influence, but large quantities should be removed. It need hardly be mentioned that other metals precipitable by hydrofluosilicic acid must be absent. The results are satisfactory. STOLBA obtained 99 -2 to 100 per cent. Potassium-platinic chloride may be easily converted into potassium silicofluoride; hence, in technical analyses, the potassium may be separated in the first form, and then titrated as the latter (STOLBA, loc. cit.). 98. 2. SODIUM. a. Solution. See 97, a, all the directions given in that place applying equally to the solution of NaOH and sodium salts. * W. KNOP and W. WOLF use hydrofluosilicate of aniline instead. Zeitschr. f. analyt. Chem., i, 471. f Zeitschr. f. analyt. Chem., in, 298. 98.] SODIUM. . 249 5 . Determination . Sodium is determined either as sodium sulphate, as sodium chloride, or as sodium carbonate ( 69). For the alkalirnetric esti- mation of caustic soda and sodium carbonate, see 219 and 220. We may convert into 1. SODIUM SULPHATE; 2. SODIUM NITRATE; 3. SODIUM CHLORIDE. In general the sodium salts corresponding to the potassium salts specified under the analogous potassium compounds, 97. 4. SODIUM CARBONATE. Caustic soda, sodium hydrogen carbonate, and sodium salts of organic acids, also sodium nitrate and sodium chlorate. 5. SODIUM SILICOFLUORIDE. Sodium salts of acids soluble in dilute alcohol, excepting sodium borate. In sodium borate the sodium is estimated best as sodium sul- phate ( 136) ; in the phosphate, as sodium chloride, or sodium carbonate ( 135). Sodium salts of organic acids are determined either, like the corresponding potassium compounds, as chloride, or by preference as carbonate. (This latter method is not so well adapted for potassium salts.) The analyst must here bear in mind, that when carbon acts on fusing sodium carbonate, carbon monoxide escapes, and caustic soda in not inconsiderable quantity is formed. 1. Determination as Sodium, Sulphate. If alone and in aqueous solution, evaporate to dryness, ignite and weigh the residue in a covered platinum crucible ( 42). The process does not involve any risk of loss by decrepitation, as in the case of potassium sulphate. If free sulphuric acid happens to be present, this is removed in the same way as in the case of potas- sium sulphate. With regard to the conversion of sodium chloride, &c. , into sodium sulphate, see 97, , 1. For properties of the residue, see 69. The method is easy, and gives accurate results. 2. Determination as Sodium Nitrate. Same method as described in 1 . The rules and observations are the same as those given under the estimation of potassium nitrate ( 97). For properties of the residue, see 69. 250 DETERMINATION. [ 98. 3. Determination as Sodium Chloride. Same method as described in 1. The rules given and the observations made in 97, &, 2, apply equally here. The fact that sodium chloride is more difficultly volatilizable than potassium chloride favors greater accuracy of results. For properties of the residue, see 69. The methods of converting sodium sulphate, chromate, chlorate, and silicate into sodium chloride will be found in Part II. of this Section, under the respective heads of the acids which these salts contain. 4. Determination as Sodium Carbonate. Evaporate the aqueous solution, ignite moderately, and weigh. TJie results are perfectly accurate. For properties of the residue, see 69. Caustic soda is converted into the carbonate by adding to its aqueous solution ammonium carbonate in excess, evaporating at a gentle heat, and igniting the residue. Sodium hydrogen carbonate, if in the dry state, is converted into the normal carbonate by ignition. The heat must be very gradually increased, and the crucible kept well covered. If iu aqueous solution, it is evaporated to dryness, in a capacious silver or platinum dish, and the residue ignited. Sodium salts of organic acids are converted into the carbonate by ignition in a covered platinum crucible, from which the lid is removed after a time. The heat must be increased very gradually. When the mass has ceased to swell, the crucible is placed obliquely, with the lid leaning against it (see 52, fig. 42), and a dull red heat applied until the carbon is consumed as far as practicable. The contents of the crucible are then warmed with water, and the fluid is filtered off from the residuary carbon, which is carefully washed. The filtrate and rinsings are evaporated to dryness with the addition of a little ammonium carbonate, and the residue is ignited and weighed. The ammonium carbonate is added, to con- vert any caustic soda that may have been formed into carbonate. The method, if carefully conducted, gives accurate results ; how- ever, a small loss of soda on carbonization is not to be avoided. If any residue soluble in water remains, dissolve it and add it to the principal solution. 99.] AMMONIUM. 251 Sodium nitrate, or sodium chloride, may be converted into car- bonate, by adding to its aqueous solution perfectly pure oxalic iicid in moderate excess, and evaporating several times to dryne.-^, with repeated renewal of the water. All the nitric acid of the sodium nitrate escapes in this process (partly decomposed, partly undecomposed) ; and equally so all the hydrochloric acid in the case of sodium chloride. If the residue is now ignited until the excess of oxalic acid is removed, sodium carbonate is left. 99. 3. AMMONIUM. a. Solution. Ammonia is soluble in water, as are all ammonium salts of those acids which claim our attention here. It is not always necessary, however, to dissolve ammonium salts for the purpose of determin- ing the amount of ammonium contained in them. b. Determination. Ammonium is weighed, as stated TO, either in the form of ammonium chloride, or in that of ammonium platinic chloride, Into these forms it may be converted either directly or indirectly (i.e., after expulsion as ammonia, and re-combination with an acid). Ammonium is also frequently determined by volumetric analysis, and its quantity is sometimes inferred, from the volume of nitrogen. We convert directly into 1. AMMONIUM CHLORIDE. Ammonia gas and its aqueous solution, and also ammonium salts of weak volatile acids (ammonium carbonate, ammonium sulphide, 2. AMMONIUM PLATINIC CHLORIDE. Ammonium salts of acids soluble in alcohol, such as ammonium sulphate, ammonium phosphate, &c. 3. The methods based on the EXPULSION OF AMMONIA from ammonium compounds, and also that of inferring the amount of ammonium from the volume of nitrogen eliminated in the dry way, are equally applicable to all ammonium salts. The expulsion of ammonia in the dry way (by ignition with soda-lime), and its estimation from the volume of nitrogen elimi- nated in the dry way, being effected in the same manner as the DETERMINATION. [ 99. estimation of the nitrogen in organic compounds, I refer the stu- dent to the Section on Organic Analysis. The process of estimating ammonia by decomposing with a bromized solution of sodium hypo- chlorite will be given under the Analysis of Soils, in the Special Part. For the alkali metric estimation of free ammonia, see 219 and 220 ; and for the calorimetric method based on the use of KESSLEK'S solution, see under the Analysis of Waters, 205. 1. Determination as Ammonium Chloride. Evaporate the aqueous solution of the ammonium chloride on the water-bath, and dry the residue at 100 until the weight re- mains constant ( 42). The results are accurate. The volatiliza- tion of the chloride is very trifling. A direct experiment gave 99-94 instead of 100. (See Expt. 15.) The presence of free hydrochloric acid makes no difference ; the conversion of caustic ammonia into ammonium chloride may accordingly be effected by supersaturating with hydrochloric acid.* The same applies to the conversion of the carbonate, with this addition only, that the process of supersaturation mu^t bo conducted in an obliquely-placed flask, and the mixture heated in ihe samo, till the carbonic acid is driven off. In the analysis of ammonium gulphide we proceed in the same way, taking care simply, after the expulsion of the hydrogen sulphide, and before proceeding to evaporate, to filter off the sul- phur which may have separated. Instead of weighing the ammo- nium chloride, its quantity may be inferred by the determination of its chlorine according to 141, ~b (Comp. potassium chloride, W, 5, 3.) 2. Determination as Ammonium- Platinic Chloride. a. Ammoniacal salts with volatile acids. Same method as described in 97, 4, a. * (potassium -platinic chloride). ft. Ammonium salts of non-volatile acids. Same method as described 97, 4, fi (potassium-platinic chlo- ride). The results obtained by these methods are accurate. If you wish to control the results, f ignite the double chloride, -GUNNING (Zeitschr. /. analyt. Chem., YII, 480) has pointed out that fluids during evaporation may take up ammonia because of the presence of this in illu- minating gas. flf the ammonium-platinic chloride is pure, which maybe known by its color and general appearance, this control may be dispensed with. 99.] AMMONIUM. 253 wrapped up in the filter, in a covered crucible, and calculate the amount of ammonium from that of the residuary platinum. The results must agree. If the double salt is in a filtering-tube, slowly pass a current of air through it, while heating very carefully. If, however, the salt is in a paper filter, it is best to transfer the pre- cipitate, wrapped up in the filter, to the crucible, and continue the application of a moderate heat for a long time, then to remove the lid, place the crucible obliquely, with the lid leaning against it, and burn the charred filter at a gradually-increased heat (H. ROSE). When the salt is pure, which may be known from its color and general appearance, this control may be omitted. Want of due caution in respect to heating is apt to lead to loss, from particles of the double salt being carried away with the ammonium chloride. Very small quantities of ammonium-platinic chloride are collected on an unweighed filter, dried, and at once reduced to platinum by ignition . * 3. Estimation by Expulsion of Ammonia in the Wet Way. This method, which is applicable in all cases, may be effected in different ways, viz. : a. EXPULSION OF THE AMMONIA BY DISTILLATION WITH SOLUTION OF POTASSA, SODA, MILK OF LIME, OR MAGNESIA. Applicable in all cases where no nitrogenous organic matters from which ammonia might be evolved upon boiling with solution of potassa, etc., are present with the ammonium salts. Magnesia is used in cases where nitrogenous substances capable of yielding ammonia on boiling, are present. Weigh the substance under examination in a small glass tube, three centimetres long and one wide, and put the tube, with the substance in it, into a small, tubulated retort, a, Fig. 81, contain- ing a suitable quantity of moderately concentrated solution of potassa or soda, milk of lime, or magnesia mixed with water, from which every trace of ammonia has been removed by protracted ebullition, but which has been allowed to get thoroughly cold again. The further arrangement of the apparatus is shown by the cut. As will be seen, the ammoniacal distillate does not come into contact with *In a series of experiments to get the platinum from pure and perfectly anhydrous ammonium-platinic chloride, by very cautious ignition, Mr. Lucius, one of my pupils, obtained from 44 -1 to 44 '3 per cent, of the metal, instead of 44-3. 254 DETERMINATION. [99. cork or rubber and this is quite important, since otherwise thsee might easily retain some of the ammoniacal fluid. Fig. 81. If you wish to determine volumetrically the quantity of ammo- nia expelled, introduce the larger portion of a measured quantity of standard solution of acid (sulphuric, hydrochloric, or oxalic, 215) into the receiver, the remainder into the U-tube; add to the portion of fluid in the latter a little water, and color the liquids in the receiver and U-tube red with 1 or 2 c. c. of tincture of lit- mus. The cooling-tube must not dip into the fluid in the receiver ; the fluid in the U-tube must completely fill the lower part, but it must not rise high, as otherwise the passage of air-bubbles might easily occasion loss by spirting. The quantity of acid used must of course be more than sufficient to fix the whole of the ammonia expelled. When the apparatus is fully arranged, and yoii have ascertained that all the joints are perfectly tight, heat the contents of the retort to gentle ebullition, and continue the application of the same degree of heat until the drops, as they fall into the receiver, have for some time altogether ceased to impart the least tint of blue to the portion of the fluid with which they first come in contact. Before removing the heat, a strip of turmeric paper is fixed in the tubulure of the retort; it must not turn brown. Then loosen the stopper of the retort, allow to stand half an hour, pour the con- tents of the receiver and U-tube into a beaker, rinse out with small quantities of water, and determine finally with a standard solution 99.] AMMONIUM. 255 of alkali the quantity of acid still free, which, by simple subtrac- tion, will give the amount of acid which has combined with the ammonia; and from this you may now calculate the amount of the latter ( 220). Results accurate.* (Expt. ISTo. 55.) If you wish to determine by the gravimetric method the qua/n- tity of ammonia expelled, receive the ammonia evolved in a quan- tity of hydrochloric acid more than sufficient to fix the whole of it, and determine the ammonium chloride formed, either by simple evaporation, after the directions of 1, or, far preferably, as ammo- nium-platinic chloride, after the directions of 2. 1). EXPULSION OF THE AMMONIA BY MILK OF LIME, WITHOUT APPLICATION OF HEAT. This method, recommended byScHLosiNG, is based upon the fact that an aqueous solution containing free ammonia gives off the latter completely, and in a comparatively short time, when exposed in a shallow vessel to the air, at the com- mon temperature. It finds application in cases where the presence of organic nitrogenous substances, decomposable by boiling alkalies, forbids the use of the method described in 3, a\ thus, for instance, in the estimation of the ammonia in urine, manures, etc. The fluid containing the ammonia, the volume of which must not exceed 35 c. c., is introduced into a shallow flat-bottomed ves- sel from 10 to 12 centimetres in diameter; this vessel is put on a plate filled with mercury. A tripod, made of a massive glass rod, is placed in the vessel which contains the solution of the ammonium salt, and a saucer or shallow dish with 10 c. c. of the normal solu- tion of oxalic or sulphuric acid ( 215) put on it. A beaker is now inverted over the whole. The beaker is lifted up on one side as far as is required, and a sufficient quantity of milk of lime added by means of a pipette (which should not be drawn out at the lower end). The beaker is then rapidly pressed down, and weighted with a stone slab. After forty-eight hours the glass is lifted up, and a slip of moist reddened litmus paper placed in it ; if no change of color is observable, this is a sign that the expulsion of the ammo- nia is complete ; in the contrary case, the glass must be replaced. Instead of the beaker and plate with mercury, a bell- jar, with a ground and greased rim, placed air-tight on a level glass plate, may be used. A bell-jar, having at the top a tubular opening furnished with a close-fitting glass stopper, answers the purpose best, as it * [In thus estimating minute quantities of ammonia, the condensing tube must "be of tin, since glass yields a sensible amount of alkali to hot- water vapor.]) 256 DETERMINATION. [ 99. permits the introduction of a slip of red litmus paper suspended from a thread, thus enabling the operator to see whether the com- bination of the ammonia with the acid is completed, without the necessity of removing the bell- jar. According to SCHLOSING, forty- eight hours are always sufficient to expel O'l to 1 gramme of ammo- nia from 25 to 35 c. c. of solution. However, I can admit this statement only as regards quantities up to 0'3 grin. ; quantities above this often require a longer time. I therefore always prefer operating with quantities of substance containing no more than 0*3 grin, ammonia at the most. When all the ammonia has been expelled, and has entered into combination with the acid, the quantity of acid left free is deter- mined by means of standard solution of alkali, and the amount of the ammonia calculated from the result ( 220). c. INDIRECT METHOD ACCORDING TO FR. MOHR.* In this method a known quantity of alkali in excess, e.g., sodium carbonate, is heated in aqueous solution with the ammonium salt until all the ammonia has been expelled ; the residual alkali is then volumetri- cally estimated, and from the difference the equivalent quantity of ammonia estimated. This method is of limited application, because it can only be used with neutral ammoniacal salts in the absence of organic matter, f It is, however, convenient and exact, and may be conducted in an obliquely supported flask. As alkali, a normal solution of soda or sodium carbonate (53 '05 grm. anhy- drous salt per litre) may be used. The boiling is stopped when the escaping vapors cease to act on red litmus paper or turmeric paper. 4. Estimation by Expulsion of the Nitrogen in the Wet Way. A process for determining ammonium by means of the azo- tometer has been given by "W. KNOP. $ It depends on the sepa- ration of the nitrogen by a bromized and strongly alkaline solution of sodium hypochlorite. The simplest azotometer is that described by KUMPF.|| It * Lehrbuch der Titrirmethode. \ Even organic matter free from nitrogen has a disturbing action, as when boiled with alkali it forms humus-like acid decomposition products, which neu- tralize the alkali. J Hiem. CentralbL, 1860, 244. This is prepared as follows : Dissolve 1 part of sodium carbonate in 15 parts of water, cooi the fluid with ice, saturate perfectly with chlorine, keeping cold all the while, and add strong soda solution (of 25 per cent.) till the mixture on rubbing between the fingers makes the skin slippery. Before using, add to the quantity required for the series of experiments bromine in the proportion of 2-3 grm. to the litre, and shake. || Fres. Zeit., vi, 398. 99.] AMMONIUM. 257 consists of a burette of 50 or 100 c. c. stationed in a glass cylinder nearly filled with mercury, and connected by a stout caoutchouc tube with a small bottle, a, Fig. 82, to which is fitted a soft, thrice- perforated caoutchouc stopper. The stopper carries a thermometer and two short glass tubes, one of which joins it to the burette, and the other has attached a short bit of caoutchouc tubing arid a pinch- cock, e. The weighed ammonium salt (not more than 0'4 grm.) is placed in the tube, /', with 10 C, c. of water, and 50 c. c. of the bromized hypochlorite solution are brought into the bottle, a. The cock, 0, being open, the stopper is firmly fixed in its place, and the burette is depressed in the mercury un- til its uppermost degree exactly coincides with the surface of the metal. The cock is then closed, and the bottle is inclined to bring the two substances in contact. The ammonium salt is speedily decomposed. When no further evolution of gas takes place the burette is so adjusted that the level of the mercury without and within it shall nearly coincide, and the operator waits 10-20 minutes, or until the thermometer in a indicates the same temperature as the sur- rounding air. Then the adjustment of the burette to exact coincidence of the mercury level, within and without, is effected, and the volume of the gas is read off. The stand of the thermometer and barometer are also noted, and the recorded volume of nitrogen is corrected by use of the tables on pp. 259- 261, arranged by DIETRICH. The first table gives a correction for the nitrogen which is absorbed by the 60 c. c. of liquid in the bottle a. The amount varies with the relative volumes of air and nitrogen, and is determined empirically by decomposing known quantities of ammonia and noting the difference between the obtained and the theoretical^ volume of nitrogen. The correction holds strictly, of rourse, only for a solution of such strength as that employed by DIETRICH and at the mean temperatures. The second table serves to spare the labor of calculation. The weight of 1 c. c. of nitrogen, measured e. g. at 751 mm. of barome- Fig. 82. 258 DETERMINATION. [ 10(X ter and 15, is found at the intersection of the vertical column 754 with the horizontal column 15, is, viz., 1-16187. To the observed volume of nitrogen add the amount absorbed as per Table I., and correct the total by Table II. It scarcely requires to be mentioned that good results can only be obtained in an apartment where the temperature is uniform, and when care is exercised to avoid warming the apparatus in handling. See DIET- RICH'S papers.* 100. Supplement to the First Group. LITHIUM. In the absence of other bases, lithium may, like potassium and sodium, be converted into anhydrous SULPHATE, and weighed' in that form (Li 2 SO 4 ). As lithium forms no acid sulphate, the excess of sulphuric acid may be readily removed by simple igni- tion. LITHIUM CARBONATE also, which is difficultly soluble in water, and fuses at a red heat without suffering decomposition, is well suited for weighing ; whilst lithium chloride, which deliquesces in the air, and is by ignition in moist air converted into hydro- chloric acid and lithium oxide, is unfit for the estimation of lithium. In presence of other alkali metals, lithium is best converted into LITHIUM PHOSPHATE (Li 3 PO 4 ), and weighed in that form. This is effected by the following process : add to the solution a sufficient quantity of sodium phosphate (which must be perfectly free from phosphates of the alkali-earth metals), and enough soda to keep the reaction alkaline, and evaporate the mixture to dryness ; pour water over the residue, in sufficient quantity to dissolve the soluble salts with the aid of a gentle heat, add an equal volume of solution of ammonia, digest at a gentle heat, filter after twelve hours, and wash the precipitate with a mixture of equal volumes of water and solution of ammonia. Evaporate the filtrate and first washings to dryness, and treat the residue in the same way as before. If some more lithium phosphate is thereby obtained, add this to the prin- cipal quantity. The process gives, on an average, 99'61 for 100 parts of lithium oxide ( * Fres. Zeit., in, 162; iv, 141, and v, 36. \ Ann. d. Chem. u. Pharm., xcvni, 212. 99.] TABLE OF ABSORPTION OF NITROGEN GAS. 259 8 I s s 8 8 S 8 S? TH (^ s s 3 S 1-H S TH s 9 (^ W <^ 260 TABLE OF WEIGHTS. [99. II. TABLE OF THE WEIGHT OF A In Milligrammes for Pressures from 720 to 770 mm. MILLIMETRES. 720 722 724 726 728 730 732 734 736 738 740 742 744 10 1.13380 1.13699 1.14018 1.14337 1.14656 1.14975 1.15294 1.15613 1.15932 1.16251 1.16570 1.16889 1.17208 !! 1.12881 1.131991.13517 1.13835 1.14153 1.14471 1.14789 1.15107 1.15424 1.15742 1.16060 1.16378 1.16696 lr> 1.12376 1.12693 1.13010 1.13326 1.13643 1.13960 1.142771.145931.14910 1.152271.15543 1.15860 1.16177 13 1.11875 1.12191 1.12506 1.12822 1.13138 1.13454 1.137691.140851.14401 1.14716 1.15032 1.15348 1.15663 14 1.11369 1.11684 1.11999 1.12313 1.12628 1.12942 1.13257 1.13572 1.13886 1.14201 1.14515 1.14830 1.15145 15 16 1.10859 1.10346 1.11172 1.11486 1.106581-10971 1.11799 1.11283 1.12113 1.11596 1.12426 1 11908 1.12739 1 12220 1.13053 1 12533 1.13366 1 12845 1.13680 1 13158 1.13993 '1.14306 1 13470 1 13782 1.14620 1 14095 1.09828 1.10139 1.10450 1.10761 1.11073 1.11384 1.11695 1.12006 1.12317 1.126291.12940 1.13251 1.13562 18 19 20 1.09304 1.08744 1.09614 1.09083 1.09924 1.09392 1.10234 1 09702 1.10544 1 10011 1.10854 1 10320 1.11165 1.10629 1.114751.11785 1 10938 1 1124ft 1.120951.12405 1 11557 1 11866 1.12715 1 12175 1.11035 1.13025 1.12484 1.11943 1.08246 1.08554 1.08862 1.09170 1.09478 1.097861.100941.104021.10710 1.110181.11327 21 1.07708 1.08015 1.08322 1.08629 1.08936 1.092431.09550 1.09857 1.10165 1.104721.107791.11086 1.11393 22 1.07166 I 23 1.06616 1.07472 1.06921 1.07778 1.07226 1.08084 1.07531 1.083901.08696 1.078361.08141 1.090021.093081.09614 1.08446 1.08751 1.09056 1.09921 1.09361 1.102271.10533 1.09666 1.09971 1.10839 1.10276 24 1.06061 1.06365 1.06669 1.06973 1.072771.07581 1.07885 1.081891.08493 1.08796 1.091001.09404 1.09708 25 1.05499 - 1.05801 1.06104 1.06407 1.06710 1.07013 1.07316 1.07619 1.07922 1.08225 1.08538 1.08881 1.09134 720 722 724 726 728 730 732 734 736 738 740 742 744 MILLIMETRES. TABLE OF WEIGHTS. CUBIC CENTIMETRE OF NITROGEN. of Mercury, and for Temperatures from 10 to 25 C. MILLIMETRES. 746 748 750 752 754 756 758 760 762 764 766 768 770 1.17527 1.1784G j 1.18165 1.18484 1 18803 1.19122 1 19441 1 19760 1 20079 1.20398 1.20717 1.21036 1 21355 1.20829 10 11* 1.17014 1.17332 1.17650 1.17168 1.18286,1.18603 1.18921 1.192391.19557 1.19875 1.20193 1.20511 1.16493 1.16810 1.17127 1.17444 ! 1. 17760 1.18077 1.18394 1.18710 1.19027 1.19344 1.19660 1.19977 1.20294 12 1.15979 1.16295 1.16611 1.16926 1.17242 1.17558 1.17873 1.18189 1.18505 1.18820 1.19136 1.19452 1.19768 13 1.15459 1.15774 1.160881.16403 1.16718 1.17032 1.17347 1.17661 1.17976 1.18291 1.18605 1.18920 1.19234 14 1.14933 1.15247 1.155601.15873 1.16187 1.16500 1.16814 1.17127 1.17440 1.17754 1.18067 1.18381 1.18694 15 1.14407 1.14720 1.15032 1.15344 1.15657 1.15969 1.16282 1.16594 1.16906 1.17219 1.17531 1.17844 1.18156 16 1.13873 1.14185 1.14496 1.14807 1.15118 1.15429 1.15741 1.16052 1.16363 1.16674 1.16985 1.17297 1.17608 17 1.13335 1.13645 1.13955 1.14266 1.14576 1.14886 1.15196 1.15506 1.15816 1.16126 1.16436 1.16746 1.17056 18 1.12794 1.13103 1.13412 1.13721 1.14030 1.14340 1.14649 1.14958 1.15267 1.15576 1.15886 1.16195 1.16504 19 1.12251 112559 1.12867 1.13175 1.13483 1.13791 1.14099 1.14408 1.14716 1.15024 1.15332 1.15640 1.15948 20 1.11700 1.12007 1.12314 1.12621 1.12928 1.13236 1.13543 1.13850 1.14157 1.14464 1.14771 1.15078 1.15385 21 1.11145 1.10581 1.11451 1.10886 1.11757 1.12063 1.11191:1.11496 1.12369 1.11801 1.12675 1.12106 1.12982 1,12411 1.13288 1.12716 1.13594 1.13021 1.13900 1.13326 1.14206 1.13631 1.14512 1.13936 1.14818 1.14241 22 23* 1.10012 1.10316 1.106201.10924 1.11228 1.11532 1.11835 1.12139 1.12443 1.12747 1.13051 1.13355 1.13659 24 1.09437 1.09740 1.10043 1.10346 1.10649 1.10952 1.11255 1.11558 1.11861 1.12164 1.12467 1.12770 1.13073 25 746 748 750 752 754 756 758 760 762 764 766 768 770 MILLIMETRES. 262 DETERMINATION. [101. If the quantity of lithium present is relatively very small, the larger portion of the potassa or soda compounds should first be removed by addition of absolute alcohol to the most highly con- centrated solution of the salts (chlorides, bromides, iodides, or nitrates, but not sulphates) ; since this, by lessening the amount of water required to effect the separation of the lithium phosphate from the soluble salts, will prevent loss of lithium (W. MAYER). * The precipitated normal lithium phosphate has the formula 2Li 3 PO 4 + H 2 O. It dissolves in 2539 parts of pure, and 3920 parts of ammoniated water ; at 100, it completely loses its water ; if pure, it does not cake at a moderate red heat (MAYER). The objections raised by RAMMXLSBEBaf to MAYER'S method of estimating lithia I find to be ungrounded. According to my own experience, it appears that the filtrate and wa&'ji- water must be evaporated in a platinum dish not only once, but at least twice in fact, till a residue is obtained which is completely soluble in dilute ammonia. Lithium phosphate may be dried at 100, or ignited according to 53, before being weighed. In the latter case, care must be taken to free the filter as much as possible from the precipitate before proceeding to incinerate it. I have thus obtained, instead of 100 parts lithium carbonate, by drying at 100, 99-84, 99-89, 100-41, by igniting 99*66 and 100-05. The lithium phosphate obtained was free from sodium. Second Group. BARIUM STRONTIUM CALCIUM MAGNESIUM. 101. 1. BARIUM. a. Solution. Caustic baryta is soluble in water, as are many barium salts. Barium salts which are insoluble in water are, with almost the single exception of the sulphate, readily dissolved by dilute hydro- chloric acid. The solution of the sulphate is effected by fusion with sodium carbonate, &c. (See 132.) Barium silico-fluoride * Ann. der Chem. u. Pharm., xcvin, 193, where Mayer has also demonstrated the non-existence of a sodium lithium phosphate of fixed composition (Berzelius), or of varying composition (Rammelsberg). \Pogg. Annal., en, 443. \Zeitschr.f. analyt. Chem., 1, 42. 101.] BARIUM. 263 may be readily converted into barium sulphate by heating and evaporating with moderately dilute sulphuric acid in a platinum vessel. It may also be readily decomposed by fusing it with po- tassium-sodium carbonate. ~b. Determination, Barium is weighed either as sulphate or as carbonate, rarely (in the separation from strontia) as barium silico-fluoride or barium chromote ( 71). Barium oxide or hydroxide, also barium car- bonate, may also be determined by the volumetric (alkalimetric) method. Comp. 223. We may convert into 1. BARIUM STJLPHATE. a. By Precipitation. All barium compounds without exception. I. By Evaporation. All barium salts of volatile acids, if no other non-volatile body is present. 2. BARIUM CARBONATE. a. All barium salts soluble in water. 5. Barium salts of organic acids. Barium is both precipitated and weighed, by far the most fre- quently as sulphate, the more so as this is the form in which it is most conveniently separated from other bases. The determination by means of evaporation (1, 5) is, in cases where it can be applied, and where we are not obliged to evaporate large quantities of fluid, very exact and convenient. Barium is determined as carbonate in the wet way, when from any reason it is not possible or not desir- able to precipitate it as sulphate. If a fluid or dry substance con- tains bodies which impede the precipitation of barium as sulphate or carbonate (alkali citrates, metaphosphoric acid, see 71, a and Z>), such bodies must of course be got rid of, before proceeding to precipitation. The precipitation of barium as a silico-fluoride or chromate will be treated of under the separation of barium from strontium, 154. 1 . Determination as Barium Sulphate. a. By Precipitation. Heat the moderately dilute solution of barium, which must not contain too much free acid (and must, therefore, if necessary, first 264 DETERMINATION. [ 102. be freed therefrom by evaporation or addition of sodium carbon- ate), in a platinum or porcelain dish, or in a glass vessel, to incipi- ent ebullition, add dilute sulphuric acid so long as a precipitate forms, keep the mixture for some time at a temperature very hear the boiling point, stirring if not on a water- bath, and allow the precipitate to subside ; decant the almost clear supernatant fluid on a filter, boil the precipitate once with water and a little dilute sul- phuric acid, then three or four times with water, then transfer it to the filter, and wash with boiling water, until the filtrate is no longer rendered turbid by barium chloride. Dry the precipitate, and treat it as directed in 53, employing only a moderate heat. If the precipitate has been properly washed in the manner here directed, it is perfectly pure. In the presence of alkali salts, however, the precipitate will still contain small quantities of alkali sulphate. Comp. 153. J. By Evaporation. Add to the solution, in a weighed platinum dish, pure sul- phuric acid very slightly in excess, and evaporate on the water- bath ; expel the excess of sulphuric acid by cautious application of heat, and ignite the residue moderately. For the properties of barium sulphate, see 71. Both methods, if properly and carefully executed, give almost absolutely accurate results. 2. Determination as Barium Carbonate, a. In Solutions. Mix the moderately dilute solution of the barium salt in a beaker with ammonia, add ammonium carbonate in slight excess, and let the mixture stand several hours in a warm place. Filter, wash the precipitate with water mixed with a little ammonia, dry ? and ignite moderately ( 53). For the properties of the precipitate, see 71. This method involves a trifling loss of substance, as barium carbonate is not absolutely insoluble in water. The direct experiment, No. 56, gave 99*79 instead of 100. If the solution contains a notable quantity of ammonium salts, the loss incurred is much more considerable, since the presence of such salts greatly increases the solubility of barium carbonate. 102.] STRONTIUM. 265 b. In Barium Salts of Organic Acids. Heat the salt slowly in a covered platinum crucible, until no more fumes are evolved ; place the crucible obliquely, with the lid leaning against it, and ignite, until the whole of the carbon is consumed, and the residue presents a perfectly white appear- ance ; moisten the residue with a concentrated solution of ammo- nium carbonate, evaporate, ignite gently, and weigh. The results obtained by this method are quite satisfactory. A direct experi- ment, No. 57, gave 99.61 instead of 100. The loss of substance which almost invariably attends this method is owing to particles 'of the salt being carried away with the fumes evolved upon igni- tion, and is accordingly the less considerable, the more slowly and* gradually the heat is increased. Omission of the moistening of the residue with ammonium carbonate would involve a further loss- of substance, as the ignition of barium carbonate in contact with carbon is attended with formation of some caustic baryta, carbon, monoxide gas being evolved. 102. 2. STRONTIUM. a. Solution. See the preceding paragraph ( 101, a. Solution of baryta and! barium salts) ; the directions there given apply equally here. Strontium silico-fluoride is readily and completely soluble in water- acidulated with hydrochloric acid. J. Determination. Strontium is weighed either as strontium sulphate or as stron- tium carbonate ( 72). Strontium in the form of oxide, hydrox- ide, or carbonate, may be determined also by the volumetric-- ^alkalimetric) method. Comp. 223. We may convert into 1. STRONTIUM SULPHATE. a. By Precipitation. All compounds of strontium without exception. b. By Evaporation. All strontium salts of volatile acids, if no other non-volatile body is present. 266 DETERMINATION. [ 102. 2. STRONTIUM CARBONATE. a. All strontium compounds soluble in water. /?. Strontium salts of organic acids. The method based on the precipitation of strontium with sul- phuric acid yields accurate results only in cases where the fluid from which the strontium is to be precipitated may be mixed, without injury, with alcohol. Where this cannot be done, and where the method based on the evaporation of the solution of strontium with sulphuric acid is equally inapplicable, the conver- sion into the carbonate ought to be resorted to in preference, if admissible. As in the case of barium, so here, we have to be on- our guard against the presence of substances which would impede precipitation (citrates of the alkalies, metaphosphoric acid, etc.), and if necessary, these must first be removed. 1. Determination as Strontium Sulphate. a. By Precipitation. Mix the solution of the strontium salt (which must not be too dilute, nor contain much free hydrochloric or nitjic acid) with dilute sulphuric acid in excess, in a beaker, and add at least an equal volume of alcohol; let the mixture stand twelve hours, and filter ; wash the precipitate with dilute alcohol, dry and ignite ( 53). If the circumstances of the case prevent the use of alcohol, the fluid must be precipitated in a tolerably concentrated state, and a fairly large excess of sulphuric acid added (this is particularly neces- sary if large quantities of potassium chloride, sodium chloride, or magnesium chloride are present). The mixture is then allowed to stand in the cold for at least twenty-four hours, filtered, and the precipitate washed with cold water, until the last rinsings manifest no longer an acid reaction, and leave no perceptible residue upon evaporation. If traces of free sulphuric, acid remain adhering to the filter, the latter turns black on drying, and crumbles to pieces; too protracted washing of the precipitate, on the other hand, tends to increase the loss of substance. Care must be taken that the precipitate be thoroughly dry, before proceeding to ignite it ; otherwise it will be apt to throw off fine particles during the latter process. The filter, which is to be burnt apart from the precipitate, must be as clean as possible, or some loss of substance will be incurred ; as may be clearly seen 102.] STRONTIUM. 267 from the depth of the carmine tint of the flame with which the filter burns if the precipitate has not been properly removed. For the properties of the precipitate, see 72. When alcohol is used and the directions given are properly adhered to, the results are very accurate ; when the sulphate of strontium is precipitated from an aqueous solution, on the contrary, a certain amount of loss is unavoidable, as strontium sulphate is not absolutely insoluble in water. The direct experiments, No. 58, gave only 98'12 and 98*02 instead of 100. However, the error may be rectified, by calculat- ing the amount of strontium sulphate dissolved in the filtrate and the wash-water, basing the calculation upon the known degree of solubility of strontium sulphate in pure and acidified water. See Expt. No. 59, which, with this correction, gave 99'77 instead of 100. The necessity for making the correction may be obviated by washing with 1 part sulphuric acid mixed with 20 parts water till all substances precipitable by alcohol are removed, then with alco- hol till all the sulphuric acid is removed. Strontium sulphate also carries down sulphates of other strong bases in small quantities, and this point must not be overlooked in making accurate analy- ses. See 153. 1}. By Evaporation. The same method as described for barium, 101, 1, 5. 2. Determination as Strontium Carbonate. a. In Solutions. The same method as described 101, 2, a. For the proper- ties of the precipitate, see 72. The method gives very accurate results, as strontium carbonate is nearly absolutely insoluble in water containing ammonia and ammonium carbonate. A direct experiment, No. 60, gave 99*82 instead of 100. Presence of ammonium salts exercises here a less adverse influence than the precipitation of barium carbonate. b. In Salts of Organic Acids. The same method as described 101, 2, b. The remarks made? there, respecting the accuracy of the results, apply equally here. 268 DETERMINATION. [ 103v 103. 3. CALCIUM. a. Solution. See 101, a. Solution of barium. Calcium fluoride is, by means of sulphuric acid, converted into calcium sulphate, and the latter again, if necessary, decomposed by boiling or fusing with an alkali carbonate ( 132). [Calcium sulphate dissolves readily in moderately dilute hydrochloric acid. It is much less soluble in strong hydrochloric acid.] b. Determination. Calcium is weighed either as calcium sulphate, as calcium carbonate, or calcium oxide ( 73). It may be converted into sulphate by evaporation, as well as by precipitation ; and into carbonate or oxide by precipitation as oxalate or carbonate, or by ignition. Calcium, in the form of oxide, hydroxide, or carbon- ate, may be determined also by the volumetric (alkalimetric) method. Comp. 223. The volumetric estimation may be also effected by precipitating the calcium as oxalate, either by a direct or an indirect method. We may convert into 1. CALCIUM SULPHATE. a. By Precipitation. All calcium salts of acids soluble in alcohol, provided no other substance insoluble in alcohol be present. b. By Evaporation. All calcium salts of volatile acids, provided no non-volatile body be present. 2. CALCIUM CARBON ATP:, OR CALCIUM OXIDE. a. By Precipitation with Ammonium Carbonate. All calcium salts soluble in water. b. By Precipitation with Ammonium Oxalate. All calcium salts soluble in water or in hydrochloric acid with- out exception. c. By Ignition. Calcium salts of organic acids. Of these several methods, 2, b (precipitation with ammonium 103.] CALCIUM. 269 oxalate) is the one most frequently resorted to. This, and the method 1, >, give the most accurate results. The method, 1, #, is usually resorted to only to effect the separation of calcium from other basic radicals ; 2, a, generally only to effect the separation of calcium together with the other alkali-earth metals from the alkalies. As many bodies (alkali citrates, and metaphosphates) interfere with the precipitation of calcium by the precipitants given, these, if present, must be first removed. 3. The volumetric methods of estimation, which are particularly to be recommended when a large number of deter- minations are to be made, will be described after the gravimetric methods. 1. Determination of Calcium Sulphate. a. 13 y Precipitation. Mix the solution of the calcium salt in a beaker, with dilute sul- phuric acid in excess, and add three or four times the volume of alcohol ; let the mixture stand twelve hours, filter, and thoroughly wash the precipitate with alcohol, dry, and ignite moderately ( 53). For the properties of the precipitate, see 73. The re- sults are very accurate. A direct experiment, No. 61, gave 99-64 instead of 100. J). By Evaporation. The same method as described under barium, 101, 1, 5. 2. Determination as Calcium Carbonate or Calcium Oxide. a. By Precipitation with Ammonium Carbonate. The same method as described 101, 2, a. The precipitate can be most conveniently weighed as calcium carbonate. It must be exposed only to a very gentle red heat, but this must be con- tinued for some time. For the properties of the precipitate, see 73. This method gives very accurate results, the loss of substance incurred being hardly worth mentioning. If the solution contains ammonium chloride or similar ammo- nium salts in considerable proportion, the loss of substance in- curred is far greater. The same is the case if the precipitate is washed with pure instead of ammoniacal water. A direct experi- 270 DETERMINATION. [ 103. ment, No. 62, in which pure water was used, gave 99'17 instead of 100 parts of lime. If it is feared that some calcium oxide has formed, from the heat being too high, the residue is moistened with a little water, a small piece of ammonium carbonate added, the whole slowly evaporated, and heated again to gentle redness, i.e., till the bot- tom of the crucible is just dull-red. If a gas blowpipe is at hand, the calcium carbonate may be converted into oxide by pro- longed, strong ignition, and then weighed as such. Comp. I *. b. B-y Precipitation with Ammonium Oxalate. a. The Calcium Salt is soluble in Water. To the hot solution in a beaker add ammonium oxalate in moderate excess, and then ammonia sufficient to impart an ammo- niacal smell to the fluid ; cover the glass, and let it stand in a warm place until the precipitate has completely subsided, which will require twelve hours, at least. Pour the clear fluid gently and cautiously, so as to leave the precipitate undisturbed, on a filter ; wash the precipitate two or three times by decantation with hot water; lastly, transfer the precipitate also to the filter, by rinsing with hot water, taking care, before the addition of a fresh portion, to wait until the fluid has completely passed through the filter, f Small particles of the precipitate, adhering firmly to the glass, are removed with a feather. If this fails to effect their complete removal, they should be dissolved in a few drops of highly dilute hydrochloric acid, ammonia added to the solution, and the oxalate obtained added to the first precipitate. Devia- tions from the rules laid down here will generally give rise to the passing of a turbid fluid through the filter. After having washed the precipitate, dry it on the filter in the funnel, and transfer the dry precipitate to a platinum crucible, taking care to remove it as * In converting precipitated calcium carbonate into calcium oxide, FRITZ- SCHE (Zeitschr.f. analyl. Chem., in, 179), and A. COSSA (ib. t vm, 141) obtained somewhat too little (99 4 7 instead of 100). This, I think, may be because the calcium' carbonate employed by FRITZSCIIE (which he dried at 160), was not anhydrous, and which he himself hints. t In order to make the calcium oxalate settle more rapidly and filter it off clearly, MUCK recommends adding 1 c. c. of ammonia-alum solution containing O'OOl grm. alumina. An excess of ammonia is to be avoided, and 0*001 grm. must be deducted from the weight of the calcium oxide. (Zeitschr. /. analyt. Chem., ix, 451.) 103.] CALCIUM. 271 completely as possible from the filter ; burn the filter on a piece of platinum wire, letting the ash drop into the hollow of the lid ( 53) ; put the latter, now inverted, on the crucible, so that the filter ash may not mix with the precipitate ; heat at first very gently, then more strongly, until the bottom of the crucible is heated to very faint redness. Keep it at that temperature from ten to fifteen minutes, removing the lid from time to time. I am accustomed during this operation to move the lamp back- wards and forwards under the crucible with the hand, since, if you allow it to stand, the heat may very easily get too high. Finally allow to cool in the desiccator and weigh. After weigh- ing, moisten the contents of the crucible, which must be perfectly white, or barely show the least tinge of gray, with a little water, and test this after a time with a minute slip of turmeric paper. Should the paper turn brown a sign that the heat applied was too strong rinse oft' the fluid adhering to the paper. with a little water into the crucible, throw in a small lump of pure ammo- nium carbonate, evaporate to dryness (best in the water-bath), heat to very faint redness, and weigh the residue. If the weight has increased, repeat the same operation until the weight remains constant. I wish to particularly point out that by closely adher- ing to the rules laid down above regarding the method of ignition, the tedious evaporation with ammonium carbonate may be al- ways avoided. For the properties of the precipitate and residue see 73. This method gives nearly absolutely accurate results. A direct experiment, No. 63, gave 99 -99 instead of 100. Equally accurate results were recently obtained also by A. Sou- CHAY in my laboratory.* If a gas blowpipe is at hand, or any other arrangement by means of which a platinum crucible may be raised to a white heat, the calcium oxalate may be converted into CAUSTIC LIME with results almost equally accurate; and I believe that this method, which requires less patience than the other, is more certain to yield good results in the hands of many persons. The calcium oxalate and the filter ash are transferred to a moderate-sized platinum crucible, which is ignited first over the BUNSEN flame, and then over the blowpipe. The crucible is then weighed, and *Zeitschr.f. analyt. Chern., x, 323. 272 DETERMINATION. [ 103. ignited again over the blowpipe. The second ignition over the blowpipe should not reduce the weight. The duration of the ignition necessary varies from 5 to 15 or more minutes, accord- ing to intensity of heat and quantity of the precipitate. It is well to weigh the empty crucible again at the end of the operation, as platinum sometimes loses weight after violent and prolonged igni- tion. The results obtained by FKITZSCHE, COSSA,"* and SOTJCHAY scarcely differ from the calculated numbers. For properties of calcium oxide, see 73. The calcium oxalate may also be converted into sulphate. SCHKOTTER ignites in a covered platinum crucible with pure ammonium sulphate. Or you may ignite in a covered platinum dish till the precipitate is for the most part converted into oxide, add a little water, then hydrochloric acid to effect solution, then pure sulphuric acid in excess, evaporate and ignite moderately. This process is also quite accurate. Some chemists collect the calcium oxalate on a weighed filter, dry at 100, and weigh. The composition of the precipitate so obtained is CaC 2 O< -f- H a O. This method, however, is more troublesome, as well as less accurate, than the first. /?. The Salt is insoluble in Water. Dissolve the salt in dilute hydrochloric acid. If the acid of the calcium salt is of a nature to escape in this operation (e.g., carbonic acid), or to admit of its separation by evaporation (e.g., silicic acid), proceed, after the removal of the acid, as directed in ex. But if the acid cannot thus be readily got rid of (e.g., phosphoric acid), proceed as follows : Add ammonia until a pre- cipitate begins to form, re-dissolve this with a drop of hydro- chloric acid, add ammonium oxalate in excess, and finally sodium acetate ; allow the precipitate to subside, and proceed for the re- mainder of the operation as directed in ct. In this process the free hydrochloric acid present reacts on the sodium acetate and ammonium oxalate, fprming sodium and ammonium chlorides, with liberation of a corresponding amount of oxalic and acetic acids in which calcium oxalate is nearly insoluble. In this method the loss is very slight. The method yields accurate re- sults. A direct experiment, No. 64, gave 99*78 instead of 100. *FRITZSCHE (Zeitschr. f. analyt. Chem., in, 179) and A. COSSA (lb., vm, 141). 103.] CALCIUM. 273 c. By l(jii'dlon. The same method as described 101, 2, 5 (barium). The residue remaining upon evaporation with ammonium carbonate (which operation it is advisable to perform twice) must be ignited very gently. The remarks made in 101, 2, J, in reference to the accuracy of the results, apply equally here. By way of con- trol, the calcium carbonate may be converted into oxide or into calcium sulphate (see &, ar), or it may be determined alkalimetri- cally ( 223). 3. Volumetric Methods. a. Regarding the alkalimetric determination of calcium oxide or carbonate, see 223. By properly carrying out the process, a mixture of calcium oxide and carbonate, obtained by moderately igniting calcium oxalate in the air, affords very good results (Expt. No. 65). b. Precipitation as calcium oxalate, and direct estimation of the oxalic acid in the precipitate. The oxalic acid in the well- washed, but not dried, calcium oxalate, is determined by means of potassium permanganate ( 137). Results are very good (Expt. No. 65). c. Precipitation of calcium oxalate and indirect estimation of the oxalic acid in the precipitate. In this method (KRAUT*) the calcium salt must be soluble in water. Add to the solutio'n of the calcium salt, contained in a measuring-flask, an exactly measured quantity of decinormal oxalic-acid solution ( 215), more than sufficient to precipitate the calcium, add ammonia until the liquid is alkaline, heat to boiling, cool, fill the flask up to the mark, and shake. Then filter through a dry filter, meas- ure off an aliquot part of the filtrate (at least one-half), determine in it the oxalic acid by potassium permanganate, as in 137, and calculate the quantity for the whole of the filtrate ; the quantity of oxalic acid which lias been used up to combine with the cal- cium, gives the quantity of the latter present. 1 c. c. of deci- normal oxalic-acid solution = 0*0028 grm. lime. The method is rapid, and gives accurate results. If the quantity of calcium is small in comparison with the volume of the liquid, no correc- tion will be necessary for the space the calcium oxalate occupies in the measuring flask. *Chem Centrattl. 1856, 316. ~ 274 DETERMINATION. [ 104. 104. 4. MAGNESIUM. a. Solution. Many magnesium salts are soluble in water ; those which are insoluble in that menstruum dissolve in hydrochloric acid, with the exception of some silicates and aluminates (see 105 and 140). b. Determination. Magnesium is weighed ( 74) either as sulphate or as pyro- phosphate, or as magnesium oxide. In the form of oxide or car- bonate, it may be determined also by the alkalimetric method described in 223. We may convert into V 1. MAGNESIUM SULPHATE. a. Directly. b. Indirectly. All magnesium salts of vola- All magnesium salts soluble tile acids, provided no other in water, and also those which, non-volatile substance be pres- insoluble in that menstruum, ent. dissolve in hydrochloric acid, with separation of their acid (provided no ammonium salts be present). 2. MAGNESIUM PYROPHOSPHATE. All magnesium compounds without exception. 3. MAGNESIUM OXIDE. a. Magnesium salts of organic acids, or of readily volatile inor- ganic oxygen acids. b. Magnesium chloride, and magnesium compounds convertible into that salt. The direct determination as magnesium sulphate is highly recommended in all cases where it is applicable. The indirect con- version into the sulphate serves only in the case of certain separa- tions, and is hardly ever had recourse to where it can possibly be avoided. The determination as pyrophosphate is most generally resorted to ; especially also in the separation of magnesium from other bases. The method based on the conversion of magnesium chloride into oxide is usually resorted to only to effect the separa- 104.] MAGNESIUM. 275 tion of magnesium from the alkali metals. Magnesium phosphates are analyzed as 135 directs. 1. Determination as Magnesium Sulphate. Add to the solution excess of pure dilute sulphuric acid, evapo- rate to dryness, in a weighed platinum dish, on the water-bath;, then heat at first cautiously, afterwards, with the cover on more strongly here it is advisable to place the lamp so that the flame may play obliquely on the cover from above until the excess of sulphuric acid is completely expelled ; lastly, ignite gently over the lamp for some time; allow to cool, and weigh. Should no fumes of hydrated sulphuric acid escape upon the application of a strongish. heat, this may be looked upon as a sure sign that the sulphuric acid has not been added in sufficient quantity, in which case, after allowing to cool, a fresh portion of sulphuric acid is added. The method yields very accurate results. Care must be taken not to use a very large excess of sulphuric acid. The resi- due must be exposed to a moderate red heat only, and weighed rapidly. For the properties of the residue, see 74. 2. Determination as Magnesium Pyrophosphate. The solution of the magnesium salt is mixed, in a beaker, with ammonium chloride, and ammonia added in slight excess. Should a precipitate form upon the addition of ammonia, this may be con- sidered a sign that a sufficient amount of ammonium chloride has not been used ; a fresh amount of that salt must consequently be added, sufficient to effect the re-solution of the precipitate formed. The clear fluid is then mixed with a solution of sodium phosphate or sodium ammonium phosphate* in excess, and the mixture stirred,, taking care to avoid touching the sides of the beaker with the stir- ring-rod ; otherwise particles of the precipitate are apt to adhere so firmly to the rubbed parts of the beaker, that it will be found difficult to remove them ; the beaker is then covered, and allowed to stand at rest for twelve hours, without warming ; after that time the fluid is filtered, and the precipitate collected on the filter, the last particles of it being rinsed out of the glass with a portion of the filtrate, with the aid of a feather; when the fluid has completely passed through, the precipitate is washed with a mixture of 3 parts of water, and 1 part of solution of ammonia of O90 sp. gr., the * According to MOHR (NaNH 4 H)PO 4 is preferable to (Na a H)PO 4 as a pre- cipitant. (See Zeitschr.f. analyt. Chem., xn, 36.) 276 DETERMINATION. [ 104. operation being continued until a few drops of the fluid passing through the filter mixed with nitric acid and a drop of silver nitrate show only a very slight opalescence. The precipitate is now thoroughly dried, and then transferred to a platinum crucible ( 53) ; the latter, with the lid on, is exposed for some time to a very gentle heat, which is finally increased to intense redness. The filter, as clean as practicable, is incinerated in a spiral of platinum wire, and the ash transferred to the cru- cible, which is then once more exposed to a red heat, allowed to cool, and weighed. If the magnesium pyrophosphate is not per- fectly white, moisten with a few drops of nitric acid, and ignite again, applying the heat at first carefully. For the properties of the precipitate and residue, see 74. This method, if properly executed, yields most accurate results. The precipitate must be washed completely, but not over-washed, and the washing water must always contain the requisite quantity of ammonia. 3. Determination as Magnesium Oxide. a. hi Magnesium Salts of Organic or Volatile Inorganic Acids. The magnesium salt is gently heated in a covered platinum crucible, increasing the temperature gradually, until no more fumes escape ; the lid is then removed, and the crucible placed in an oblique position, with the lid leaning against it. A red heat is now applied, until the residue is perfectly white. For the prop- erties of the residue, see 74. The method gives the more accu- rate results the more slowly the salt is heated from the beginning. Some loss of substance is usually sustained, owing to traces of the salt being carried oif with the empyreumatic products. Mag- nesium salts of readily volatile oxygen acids (carbonic acid, nitric acid), may be transformed into magnesium oxide in a similar way, by simple ignition. Even magnesium sulphate loses the whole of its sulphuric acid when exposed, in a platinum crucible, to the heat of the gas blowpipe-flame (SONNENSCHEIN). As regards small quan- tities of magnesium sulphate, I can fully confirm this statement. b. Conversion of Magnesium Chloride into Magnesium Oxide. To the concentrated solution, in a porcelain crucible, add a mixture of water and pure mercuric oxide in more than sufficient quantity to enable the oxygen of the oxide to convert all the magnesium chloride present into magnesium oxide, evaporate the 105.] ALUMINIUM. 277 mixture on a water-bath, dry thoroughly, cover the crucible, and heat to redness until all the mercuric chloride formed and also the excess of mercuric oxide have been expelled. (The operator should carefully guard against inhaling any of the vapors evolved.) The residue, magnesium oxide, may be weighed in the crucible, or, if its separation from alkalies is intended, it is col- lected on a filter, washed with hot water, dried and ignited (53). Regarding other methods whereby the object intended may be attained, and which are frequently more convenient for effecting separations, see 153, B, 4 (17 to 21). THIRD GROUP OF BASIC RADICALS. ALUMINIUM GHROMIUM (TITANIUM). 105. 1. ALUMINIUM. a. Solution. Aluminium compounds which are insoluble in water dissolve, for the most part, in hydrochloric acid. Native crystallized alu- minium oxide (sapphire, ruby, corundum, &c.), and many native aluminium compounds, and also artificially produced aluminium oxide after intense ignition, require fusing with sodium carbon- ate, caustic potassa, or barium hydroxide, as a preliminary step to their solution in hydrochloric acid. Many aluminium com- pounds (e.g., common clay) which resist the action of concen- trated hydrochloric acid, may be decomposed by protracted heat- ing with moderately concentrated sulphuric acid, or by fusion with sodium disulphate ; potassium disulphate also effects the de- composition, but it gives rise to the formation of a double salt of potassium and aluminium difficultly soluble in water or acids, and which renders further analysis more difficult (L. SMITH *). b. Determination. Aluminium is almost invariably weighed as aluminium oxide; occasionally also as phosphate (compare, for instance, 209, 7, n). In the former case the several aluminium salts are converted into aluminium oxide, either by precipitation as aluminium hydroxide, and subsequent ignition, or by simple ignition. Precipitation as basic acetate or basic formate is resorted to only in cases of sepa- * Amer. Jour, of Sc. and Arts, XL, 248; Zeitschr.f. analyt. Chem., iv, 412. 278 DETERMINATION. [ 105. ration. For the indirect (acidimetric) estimation of aluminium in alum, etc., see 215. We may convert into ALUMINIUM OXIDE. a. By Precipitation. b. By Heating or Ignition. All aluminium salts soluble a. All aluminium salts of in water, and those which, in- readily volatile oxygen acids soluble in that menstruum, dis- (e.g.-, aluminium nitrate), solve in hydrochloric acid, with ft. All aluminium salts of separation of their acid. organic acids. With regard to the method #, it must be remembered that the solution must contain no organic substances which would interfere with the precipitation e.g., tartaric acid, sugar, &c. Should such be present, the solution must be mixed with sodiuni carbonate and potassium nitrate, evaporated to dryness in a platinum dish, the residue fused, then softened with water, trans- ferred to a beaker, digested with hydrochloric acid, and the solu- tion filtered, and then, but not before, precipitated. The methods 5, a and fi are applicable only in cases where no other fixed substances or ammonium chloride are present (on igniting the latter with aluminium oxide, aluminium chloride volatilizes). The methods of determining aluminium in its com- binations with phosphoric, boric, silicic, and chromic acids will be found in Part II. of this Section, under the heads of these several acids. Determination as Aluminium Oxide. a. By Precipitation. Mix the moderately dilute hot solution of the aluminium salt, in a beaker or dish, witli a tolerable quantity of ammonium chlo- ride, if that salt is not already present ; add ammonia slightly in excess, boil gently till the fluid gives a neutral or barely alkaline reaction (the fluid adhering to the test paper must be washed back). The fluid must not be heated too long, or it may become acid through decomposition of ammonium chloride, and some of the precipitate may redissolve; and this must, of course, be avoided. Precipitation is best effected in a large platinum dish ; in default of this a porcelain one will answer, but glass is not to 105.] ALUMINIUM. 279 be recommended because it is markedly attacked by hot ammo- niacal liquids (see p. 88). Allow the precipitate to settle; then decant the clear supernatant fluid on to a filter, taking care not to disturb the precipitate; pour boiling water on the latter in the beaker, stir, let the precipitate subside, decant again, and repeat this operation of washing by decantation a second and a third time ; transfer the precipitate now to the filter, and finish the washing with boiling water. Suction is particularly useful in filtering off aluminium hydroxide ( 47), as the precipitate may , without further treatment, be at once ignited as detailed on p. 117. If suction is not employed, the ignition of the moist precipitate is a rather critical operation. If the precipitate is to be dried before igni- tion, however, the drying must first be very thorough, after 'which ignite ( 52), and weigh. The heat applied should be very gentle at first, and the crucible kept well 'covered, to guard against the risk of loss of substance from spirting, which is always to be apprehended if the precipitate is not thoroughly dry. In whichever way the precipitate is ignited, it is always advisable to expose it for some time to an incipient white heat by means of the gas blowpipe, before weighing, in order to be sure that every trace of moisture has been expelled. A. MITSCHERLICH.* In the case of aluminium sulphate the foregoing process is apt to leave some sulphuric acid in the precipitate, which, of course, vitiates the result. To insure the removal of this sulphuric acid, the precipitate should be exposed for 5-10 min. to the heat of the gas blowpipe flame. If there are difficulties in the way, preventing this proceeding, the precipitate, either simply washed or mod- erately ignited, must be re-dissolved in hydrochloric acid (which requires protracted warming with strong acid), and then precipi- tated again with ammonia ; or the sulphate must first be con- verted into nitrate by decomposing it with lead nitrate, added in very slight excess, the excess of lead removed by means of hydro- gen sulphide, and the further process conducted according to the directions of a or I. The precipitation may also be effected by means of ammonium carbonate or ammonium sulphate, in- stead of ammonia ; the accuracy of the results is not, however, thereby increased. *Zeitschr.f. analyt. Chem., i, 67. 280 DETERMINATION. [ 106. For the properties of aluminium hydroxide and ignited alu- minium oxide, see 75. The operator should never neglect to test the aluminium hydroxide for silicic acid (which is often present). This is readily done by heating with dilute sulphuric acid, or fusing with potassium- or sodium disulphate ( 75). The method, if properly executed, gives very accurate results. But if a considerable excess of ammonia is used, more particularly in the absence of ammonium salts, and the liquid is filtered with- out boiling or long standing in a warm place to remove the ammonia, no trifling loss may be incurred. This loss is the greater, the more dilute the solution, and the larger the excess of ammonia. The precipitate cannot well be sufficiently washed on the filter on account of its gelatinous nature ; on the other hand, if it be entirely washed by decantation, a very large quantity of wash- water must be used, hence it is advisable to combine the two methods, as directed.* b. By Ignition. a. Aluminium Salts of Volatile Oxygen Acids. Ignite the salt (or the residue of the evaporated solution) in a platinum crucible, gently at first, then gradually to the very high- est degree of intensity, until the weight remains constant. For the properties of the residue, see 75. Its purity must be care- fully tested. There are no sources of error. ft. Aluminium Salts of Organic Acids. The same method as described 104, 3, a (Magnesium). 106. 2. CHROMIUM. a. Solution. Many chromic salts are soluble in water. Chromic hydroxide, and most of the salts insoluble in water, dissolve in hydrochloric acid. Ignition renders chromic oxide and many chromium salts insoluble in acids ; this insoluble modification must be prepared for * [When a solution of aluminium hydroxide in potassium or sodium hydrox- ide is boiled with excess of ammonium chloride, the aluminium separates com- pletely as a hydrated oxide with two mol. of water, which may be washed with- comparative ease. In certain cases, as where aluminium is separated from ferric iron by boiling their hydroxides with soda, this fact may be taken advantage of. LOWE, Fres. Zeitsckrift, iv, 355.] 106.] CHROMIUM. 281 solution in hydrochloric acid, by fusing with 3 or 4 parts of po- tassa in a silver crucible. In the process of fusing a small quan- tity of potassium chromate is formed by the action of air ; this, however, can be decomposed by heating with hydrochloric acid witli formation of chromic chloride. Addition of alcohol greatly promotes the reduction to chromic chloride. Instead of this fus- ing with potassa, we frequently prefer to adopt a treatment whereby the chromium is at once oxidized and converted into an alkali chromate (see 2). For the solution of chromic iron, see 160. b. Determination. Chromium is always, when directly determined, weighed as chromic oxide. It is brought into this form either by precipitation as hydroxide and ignition, or by simple ignition. It may, how- ever, also be estimated by conversion into chromic acid, and deter- mination as such. We may convert into 1. CHROMIC OXIDE. a. By Precipitation. I. By Ignition. All chromic salts soluble in a. All chromic salts of vola- water, and also those which, in- tile oxygen acids, provided no soluble in that menstruum, dis- non-volatile substances be pres- solve in hydrochloric acid, with ent. separation of their acid. Pro- ft. Chromic salts of organic vided always that no organic acids, substances (such as tartaric acid, oxalic acid, &c.) which interfere with the precipitation be present. 2. CHROMIC ACID, or. more correctly speaking, ALKALI CHROMATE. Chromic oxide and all chromic salts. The methods of analyzing chromic phosphates, borates, silicates, and chromic chromate, will be found in Part II. of this Section, under the heads of the several acids of these compounds. 1. Determination as Chromic Oxide. a. By Precipitation. The solution, which must not be too highly concentrated, is best heated to 100 in a platinum dish. One of porcelain may also answer, but is not so good, but glass should be avoided, 282 DETERMINATION. [ otherwise considerable error is caused by contamination of the precipitate with silica, and the results, therefore, will be too high (A. SOUCHAY*). If porcelain is used, this, error is slight. Am- monia is then added slightly in excess, and the mixture exposed to a temperature approaching boiling, until the fluid over the pre- cipitate is perfectly colorless, presenting no longer the least shade of red ; let the solid particles subside, wash three times by decan- tation, and lastly on the filter, with hot water, dry thoroughly, and ignite ( 52). The heat in the latter process must be in- creased gradually, and the crucible kept covered, otherwise some loss of substance is likely to arise from spirting upon the incan- descence of the chromic oxide, which marks the passing of the sol- uble into the insoluble modification. A suction filter ( 47) is very convenient for washing the precipitate, which may then be transferred, still moist, to the crucible in which it is ignited and weighed. (See p. 117.) For the properties of the precipitate and residue, see 76. This method, if properly executed, gives accurate results. Precipitation may also be effected w r ith ammo- nium sulphide, instead of ammonia. In this case precipitation is complete in the cold ; and it may be carried out in glass vessels. J. By Ignition. a. Chfomic Salts of Volatile Oxygen Acids, The same method as described, 105, Z>, a (Aluminium). /?. Chromic Salts of Organic Acids. The same method as described, 104, 3, a (Magnesium). 2. CONVERSION OF CHROMIUM IN CHROMIC COMPOUNDS INTO ALKALI CHROMATE. (For the estimation of chromic acid, see 130.) The following methods have been proposed with this view : a. The solution of the chromic salt is mixed with solution of potassa or soda in excess, until the chromic hydroxide, which forms at first, is redissolved. Chlorine gas is then conducted into the cold fluid until it acquires a yellowish-red tint ; it is then mixed with potassa or soda in excess, and the mixture evaporated to dry- ness ; the residue is ignited in a platinum crucible. The whole of * Zeitschr. f. analyt. Chem., TV, 66. 106.] CHROMIUM. 283 the potassium (or sodium) chlorate formed is decomposed by this process, and the residue consists, therefore, now of an alkali chro- mate and potassium (or sodium) chloride. (YoHL.) b. Potassium hydroxide is heated in a silver crucible to calm fusion ; the heat is then somewhat moderated, and the perfectly dry chromic compound projected into the crucible. When the substance is thoroughly moistened with the potassa, small lumps of fused potassium chlorate are added. A lively effervescence ensues, from the escape of oxygen ; at the same time the mass acquires a more and more yellow color, and finally becomes clear and trans- parent. Loss of substance must be carefully guarded against (H. SCHWABZ). c. Dissolve chromic hydroxide in solution of potassa or soda, add lead dioxide in sufficient excess, and warm. The yellow fluid produced contains all the chromium as lead chromate in alkaline solution. Filter from the excess of lead dioxide, add to the filtrate acetic acid to acid reaction, and determine the weight of the pre- cipitated lead chromate (G. CHANCEL*). d. Mix finely comminuted chromic hydroxide with some potassium chlorate in a porcelain dish, add nitric acid (sp. gr. 1*367), cover the dish with a funnel of somewhat smaller diam- eter, heat on the w T ater-bath, and add from time to time a frag- ment of potassium chlorate until all the chromic hydroxide is dis- solved and converted into potassium chromate. Even with a hydroxide which has been strongly ignited, the operation does not last longer than 30 to 60 minutes. The chromic acid is most con- O veniently determined in the solution by precipitation as barium chromate (STORES f ; PEARSON ;). [e. Render the solution of chromic salt nearly neutral by a solution of sodium carbonate, add sodium acetate in excess, heat and add chlorine water, or pass in chlorine gas, keeping the solu- tion nearly neutral by occasional addition of sodium carbonate. The oxidation proceeds readily. Boil off excess of chlorine, when the chromic acid may be precipitated as lead chromate or barium chromate (W. GIBBS ).] * Comp. rend., XLIII, 927. -\Zeit8chr.f. analyt. Chem., ix, 71. \ Ibid., ix, 108. [Am. Journ. 8ci. 2 Ser., xxxix, 58.] 284 DETERMINATION. [ 107. 107. Supplement to the Third Group. TITANIUM. Titanium is always weighed as titanic oxide (TiO 2 ), i.e., the oxide or anhydride corresponding to titanic acid (Ti(OH) 4 ). Titanic acid is precipitated with an alkali or by boiling its dilute acid solution. In precipitating acid solutions of titanic acid ammo- nia is employed; take care to add the precipitating agent only in slight excess, let the precipitate formed, which resembles alu- minium hydroxide, deposit, wash, first by decantation, then com- pletely on the filter, dry, and ignite ( 52). If the solution con- tained sulphuric acid, put some ammonium carbonate into the crucible, after the first ignition, to secure the removal of every remaining trace of that acid. Lose no time in weighing the ignited titanic oxide, as it is slightly hygroscopic. Occasionally it is more convenient to precipitate titanic acid from its acid solutions by nearly neutralizing with ammonia, adding sodium acetate and boil- ing. The precipitate thus obtained is easily filtered and washed. If we have titanic acid dissolved in sulphuric acid, as for instance occurs when we fuse it with potassium disulphate and treat the mass with cold water, we may, by largely diluting, and long boil- ing, with renewal of the evaporating water, fully precipitate the titanic acid. If much free acid is present it must be nearly neu- tralized with ammonia before boiling. Boiling is best effected in a platinum dish. After filtration, the free acid in the filtrate is still further neutralized, and the liquid boiled again for sometime, to see that no titanic acid is precipitated. Testing the last filtrate with ammonia affords the certainty that precipitation is complete. In the process of igniting the dried precipitate, some ammonium carbonate is added. From dilute hydrochloric-acid solutions of titanic acid, the latter separates completely only upon evaporating- the fluid to dryness; and if the precipitate in that case were washed Avith pure water, the filtrate would be milky ; acid must, therefore, be added to the water. Titanic acid precipitated in the cold, washed with cold water, and dried without elevation of temperature, is completely soluble in hydrochloric acid ; otherwise it dissolves only incompletely in that acid. The metaiitanic acid thrown down from dilute acid 107.] TITANIUM. 285 solutions by boiling, is not soluble in dilute acids. Titanic oxide resulting from ignition of titanic or metatitanic acid does not dis- solve even in concentrated hydrochloric acid, but it does dissolve by long heating with tolerably concentrated sulphuric acid. The easiest way of effecting its solution is to fuse it for some time with potassium disulphate, and treat the fused mass with a large quantity of cold water. Upon fusing with sodium carbonate, sodium titanate is formed, which, w r hen treated with water, leaves acid sodium titanate, which is soluble in hydrochloric acid. Ti- tanic oxide (TiO 2 ) consists of 60 -07 per cent, of titanium and 39*93 per cent, of oxygen. By fusing titanic oxide with three times its quantity of potassium hydrogen fluoride, potassium titan ium fluoride is formed, which readily dissolves in very dilute hydrochloric acid (of sp. gr. 1-015) in the heat. On fusing, a very low heat must be applied at first, till the excess of hydro- fluoric acid has escaped, then the heat is quickly raised till the mass melts and the titanic oxide is just dissolved (MARIGNAC *) . Or heating with hydrofluoric and sulphuric acids practically no titanium fluoride escapes, but by heating with hydrofluoric acid some loss does occur (RiLEY f). Titanium may be estimated volumetrically by first converting it into titanous oxide, Ti 2 O 3 , and then oxidizing this to titanic oxide by means of potassium permanganate (compare 112, 2) (PisANi ;{;). Solutions in sulphuric acid are to be avoided ; but the ordinary solution in hydrochloric acid, or the solution of titanium-potassium fluoride in dilute hydrochloric acid, is used. The reduction is effected with zinc under exclusion of air, and with or without the application of heat. In case of the hydro- chloric-acid solutions it is accompanied by a violet color ; in solu- tions of titanium-potassium fluoride, with a greenish color. After the reduction is effected the zinc is removed, and solution of potassium permanganate added until the liquid begins to remain red. The weak point in this method lies in the difficulty of ac- curately determining the moment when the reduction is complete. MARIGNAC has fully described the conditions by the observance of which he almost invariably obtained good results. * Zeitschr.f. analyt. Chem., vn, 112. jib., n, 71. J lb. t iv, 419. $lb., vn, 113. 286 DETERMINATION. [ 108, FOURTH GROUP OF BASIC RADICALS. ZINC MANGANESE NICKEL COBALT FERROUS IRON FERRIC IRON (URANIUM AND URANYL). 108. 1. ZINC. a. Solution. Many of the zinc salts are soluble in water. Metallic zinc, zinc oxide, and the salts which are insoluble in water, dissolve in hydrochloric acid. For effecting the solution of precipitated zinc sulphide, hydrochloric acid is also best. To dissolve zinc blende, however, it is best to first subject the finely powdered mineral to- the action of hot, concentrated hydrochloric acid, and then effect complete solution by adding some nitric acid, potassium chlorate, or a little of some solution of bromine in hydrochloric acid. b. Determination. Zinc is weighed either as oxide or as sulphide ( 77). The conversion of zinc salts into the oxide is effected either by precipi- tation as basic zinc carbonate or sulphide, or by direct ignition. Besides these gravimetric methods, several volumetric methods are in use. We may convert into 1. ZINC OXIDE. a. By Precipitation as Zinc b. By Precipitation as Zinc Carbonate. Sulphide. All zinc salts which are solu- All compounds of zinc with- ble in water, and all zinc salts of out exception, organic volatile acids ; also those salts of zinc which, insoluble in water, dissolve in hydrochloric acid, with separation of their acid. c. By direct Ignition. Zinc salts of volatile inorganic oxygen acids. 108.] ZINC. 287 2. ZINC SULPHIDE. All compounds of zinc without exception. The method 1, c, is to be recommended only, as regards the more frequently occurring compounds of zinc, for the carbonate and the nitrate. The methods 1, &, or 2, are usually only resorted to in cases where 1, #, is inadmissible. They serve more especially to separate zinc from other basic radicals. Zinc salts of organic acids cannot be converted into the oxide by ignition, since this, process would cause the reduction and volatilization of a small por- tion 'of the metal. If the acids are volatile, the zinc may be deter- mined at once, according to method 1, a: if, on the contrary, the acids are non- volatile, the zinc is best precipitated as sulphide. For the analysis of zinc chromate, phosphate, borate, and silicate, look to the several acids. The volumetric methods are chiefly employed for technical purposes ; see Special Part. 1. Determination as Zinc Oxide. a. By Precipitation as Zinc Carbonate, Heat the moderately dilute solution nearly to boiling in a capo- clous vessel, a glass vessel is poorly adapted for this purpose, porcelain is better, and platinum best; add, drop by drop, sodium carbonate till the fluid shows a strong alkaline reaction; boil a few minutes; allow to subside, decant through a filter, and boil the precipitate three times with water, decanting each time; then transfer the precipitate to the filter, wash completely with hot water, dry, and ignite as directed 53, taking care to have the filter as clean as practicable, before proceeding to incinerate it. In. order to prevent the reduction of the zinc oxide and volatilization of zinc, it is advisable to carefully saturate with ammonium nitrate the filter after removing as much of the precipitate from the latter as possible, and then to incinerate it. Should the solution contain ammonium salts, the ebullition must be con- tinued until, upon a fresh addition of sodium carbonate, the escap- ing vapor no longer imparts a brown tint to turmeric paper. If the quantity of ammonium salts present is considerable, the fluid must be evaporated boiling to dryness. It is, therefore, in such cases more convenient to precipitate the zinc as sulphide (see b). The presence of a great excess of acid in the solution of zinc must be as much as possible guarded against, that the effervescence from the escaping carbonic acid gas may not be too impetuous. The 288 " DETERMINATION. [ 108. filtrate must always be tested with ammonium sulphide (with addi- tion of ammonium chloride) to ascertain whether the whole of the zinc has been precipitated. This should be done in a flask filled to the neck and then closed. A slight precipitate will indeed invari- ably form upon the application of this test; but, if the process has been properly conducted, this is so insignificant that it may be al- together disregarded, being limited to some exceedingly slight and imponderable flakes, which moreover make their appearance only after many hours' standing. If the precipitate is more considerable, however, it must be treated as directed in &, and the weight of the zinc oxide obtained added to that resulting from the first process. For the properties of the precipitate and residue, see 77. This method yields pretty accurate results, though they are in most cases a little too low, as the precipitation is never absolutely com- plete, and as particles of the precipitate will always and unavoid- ably adhere to the filter, which exposes them to the chance of reduction and volatilization during the process of ignition. On the other hand, the results are sometimes too high ; this is owing to defective washing, as may be seen from the alkaline reaction which the residue manifests in such cases. It is advisable also to ascertain whether the residue will dissolve in hydrochloric acid without leaving silica; this latter precaution is indispensable in cases where the precipitation has been effected in a glass vessel. [It is often better, especially in presence of ammonium salts, to heat the dry zinc salt with excess of sodium carbonate in a plati- num dish cautiously to near redness, then treat with hot water arid wash as directed.] b. By Precipitation as Zinc Sulphide. Mix the solution, contained in a not too large flask and suffi- ciently diluted, with ammonium chloride, then add ammonia, till the reaction is just alkaline, and then colorless or slightly yellow ammonium sulphide in moderate excess. If the flask is not now quite full up to the neck, make it so with water, cork, allow to stand 12 to 24 hours in a warm place, wash the precipitate, if con- siderable, first by decantation, then on the filter with water con- taining ammonium sulphide and also less and less ammonium chlo- ride (finally none). If the zinc sulphide is to be weighed as such, it is best to replace the ammonium chloride by ammonium nitrate. In decanting do not pour the fluid through the filter, 109.] ZINC. 289 but at once into a flask. After thrice decanting, filter the fluid that was poured off, and then transfer the precipitate to the filter, finishing the washing as directed. The funnel is kept cov- ered with a glass plate. If the zinc is not to be determined accord- ing to 2, then put the moist filter with the precipitate in a beaker, and pour over it moderately dilute hydrochloric acid slightly in excess. Put the glass now in a warm place, until the solution smells no longer of hydrogen sulphide ; dilute the fluid with a little water, filter, wash the original filter with hot water, and proceed with the solution of zinc chloride obtained as directed in a. The following method also effects a practically complete pre- cipitation of zinc from acid solution. Add sodium carbonate, at last drop by drop till a lasting precipitate forms, dissolve the latter by a drop of hydrochloric acid, pass .hydrogen sulphide till the precipitate ceases to increase perceptibly, add sodium acetate, and again pass the gas. After washing with water containing hydro- gen sulphide (which when the zinc sulphide had been thrown down by hydrogen sulphide from acetic acid solution, is easily done), treat as above directed. From a solution of zinc acetate the metal may be precipitated completely, or nearly so, with hydrogen sulphide gas, even in pres- ence of an excess of acetic acid, provided always no other free acid be present (Expt. No. 66). The precipitated zinc sulphide is washed with water impregnated with hydrogen sulphide, and, for the rest, treated exactly like the zinc sulphide obtained by precipi- tation with ammonium sulphide. Small quantities of zinc sulphide may also be converted directly into the oxide, by heating in an open platinum crucible, to gentle redness at first, then, after some time, to most intense redness. For the properties of zinc sulphide see 77, c. c. By direct Ignition. The salt is exposed, in a covered platinum crucible, first to a gentle heat, finally to a most intense heat, until the weight of the residue remains constant. The action of reducing gases is to be avoided. 2. Determination as Zinc Sulphide. The precipitated zinc sulphide, obtained as in 1, J, may be ignited in hydrogen and weighed. H. ROSE,* who recom- * Pogg. Anal., ex, 128. 290 DETERMINATION. [ 108. mended the process, employs the apparatus represented by Fig. 83. a contains concentrated sulphuric acid, >, calcium chloride. The porcelain crucible has a perforated porcelain or platinum cover, into the opening of which fits the porcelain or platinum tube, d. The latter is provided with an annular projection which rests on the cover, the tube itself extends some distance into the crucible. When the zinc sulphide has dried in the filter, it is transferred to the weighed porcelain crucible, the filter ashes added, powdered sulphur is sprinkled over the contents of the crucible, the cover is placed on, and hydrogen is passed in a moderate stream, a gentle heat is applied at first, which is afterwards raised for five minutes to intense redness ; finally the crucible is allow r ed to cool with continued transmission of the gas, and the zinc sul- phide is weighed. Fig. 83. Instead of the hydrogen apparatus shown, which may not be at the disposal of the operator, any apparatus that allows the cur- rent of gas to be regulated may be used. An evolution-appara- tus in which the current is not under control, is not suitable. Instead of the porcelain tube and perforated cover, a com- mon tobacco-pipe may be employed, the bowl of the latter being 109.] MANGANESE. 291 inverted over and fitting exactly within a porcelain crucible. [Hydrogen sulphide may be advantageously substituted for hydrogen.] OESTEN'S experiments, which were adduced by ROSE in sup- port of the accuracy of this method, were highly satisfactory. Zinc sulphate, carbonate, and oxide may be converted into sulphide in the manner just described. They must, however, be mixed with an excess of powdered sulphur, otherwise you will lose some zinc from the reducing action of the hydrogen 011 the zinc oxide. Zinc sulphate is best ignited first with excess of air and before mixing with sulphur, and then igniting in a current of hydrogen. (II, ROSE.) The properties of the hydrated and anhydrous zinc sulphide are given 77 ; the results are accurate. Loss occurs only when the ignition is performed over the gas blowpipe (which is quite unnec- essary), and continued longer than five minutes. Compare 77, c. 109. 2. MANGANESE. a. Solution. Many manganous salts are soluble in water. The inanganous salts which are insoluble in that menstruum, dissolve in hydrochloric acid, which dissolves also all oxides of manganese. The solution of the higher oxides is attended with evolution of chlorine equiva- lent to the amount of oxygen which the oxide under examination contains, more than mauganous oxide (MnO) and the fluid, after application of heat, is found to contain inanganous chloride. b. Determination. Manganese is weighed either as protosesquioxide, as sulphide, or as ptjt'<>j)ltoxj)hate ( 78). Into the form of protosesquioxide it is converted either by precipitation as manganous carbonate, or as manganous hydroxide, sometimes preceded by precipitation as- manganous sulphide, or as manganese dioxide ; or, finally, by direct ignition. [When determined as pyrophosphate it is precipitated as ammonium manganous phosphate.] Manganese may be determined volumetrically in three differ- ent ways, one being applicable to any manganous solution, pro- vided it be free from any other substance which exerts a reducing action on alkaline solution of potassium f erricy anide ; the second is 292 DETERMINATION. [ 109* applicable only when iron is absent ; the third is admissible only when we have manganese in the condition of a perfectly definite higher oxide, and free from other bodies, which evolve chlorine on boiling with hydrochloric acid. We may convert into 1. MANGANESE PROTOSESQUIOXIDE. a. By Precipitation as Man- b. By Precipitation as Man- ganous Carbonate. ganese Hydroxide. All soluble manganous salts All the compounds of manga- of inorganic acids, and all man- nese, with the exception of its ganous salts of volatile organic salts of non-volatile organic acids, acids ; also those manganous salts which, insoluble in water, dis- solve in hydrochloric acid with separation of their acid. c. By Precipitation as Man- d. By Separation as Manga' ganese Sulphide. nese Dioxide. All compounds of manganese All compounds of manganese without exception. in a slightly acid solution, espe- cially manganous acetate and ni- trate. e. By direct Ignition. All manganese oxides ; man- ganous salts of readily volatile acids, and organic acids. 2. MANGANESE SULPHIDE. All compounds of manganese without exception. 3. MANGANESE SULPHATE. All manganese oxides, as well as salts of volatile acids, pro- vided no non-volatile substance be present. 4. MANGANESE PYROPHOSPHATE. All the oxides of manganese soluble in water, and those salts, insoluble in water, the acids of which may be removed by solution. The method 1, . There are no sources of error in the pre- cipitation with ammonium sulphide. For the properties of cobalt sulphide, see 80. It cannot be brought into a weighable form by ignition in hydrogen, as the residue is a variable mixture of different sulphides (II. ROSE). Cobalt may also be thrown down as sulphide by the other methods given under Nickel. The thorough precipitation of cobalt is much easier than that of nickel. d. By precipitation as tripotassium cobaltic nitrate. To the moderately concentrated solution of the cobalt salt add potassa in excess, then acetic acid till the precipitate is just redis- solved, then a concentrated solution of potassium nitrite previously just acidified with acetic acid, and allow to stand 24 hours at a gentle heat. Filter, wash with solution of potassium acetate (1 in 10) containing some potassium nitrite, till all foreign substances are removed, dry, dissolve with the filter ash in hydrochloric acid, filter and determine the cobalt according to 1, l>. This method 308 DETERMINATION. [ 111. was introduced by A. STKOMEYEK ; * the present modification, first suggested by H. ROSE, and improved by FR. GATJHE, is the surest to yield good results (GAUHEf). For the properties of the pre- cipitates, see 80, e. 2. Determination as sulphate. a. By direct conversion. To the solution of cobaltous sulphate add a little more sul- phuric acid than will suffice to form cobaltous sulphate with all the cobalt present if a volatile acid is present. Evaporate, using a platinum dish or platinum crucible, at all events, to finish the operation. Heat the residue cautiously over the lamp, gradually increasing the temperature to dull redness, and maintain at this point for 15 minutes. Should the edges blacken, moisten with dilute sulphuric acid, dry, and ignite again with greater caution, Properties of the precipitate, 80. Results quite satisfactory. J b. With previous precipitation as sulphide. Precipitate the cobalt as sulphide according to 1, kt. Chem., LXXXI, 423) employed the method in estimating iron in ferric chloride obtained by oxidizing ferrous chloride with nitric acid, but he improved the process by employing an excess of potassium iodide, facilitating the reaction by warming, and estimating the liberated iodine on cooling. In 1863 FR. MOHR (Zeitschr. f. analyt. Chem., n, 243) described the method, adopting BRAUN'S improvements, but standardizing the thiosulphate solution with iodine liberated by potassium bichromate. In 1864 BRAUN (Zeitschr. analyt. Chem., in, 452) again described his method in detail. 332 DETERMINATION. [ 113, cient hydrochloric acid (about 0'5 to 1 c. c.) of I'l specific gravity is added to each to render the solutions again clear. By this method undue acidity of the liquids is avoided, and the fluids will no longer be brownish- red, but will have a dark-yellow color. Three grm. of potassium iodide are now introduced into each of the flasks, the stoppers tightly inserted and tied down with moistened parchment paper or with wire or cord, and the flasks warmed to 50 or 60( best accomplished by suspending the flasks in the ascending steam of a water-bath). The reduction of the ferric chloride is completed in from 15 to 20 minutes, when the solutions will have a brownish-red color. After allowing to cool, run in from the burette the sodium-thiosulphate solution until the liquid has a wine-yellow color, then add from O5 to 1 c. c. of thin starch paste, and continue again to add the thiosulphate solu-, tion until the blue color of the starch iodide just disappears. The volume used up corresponds to the iodine liberated by Ol grm. iron, and hence also to O'l grm. iron present as ferric chloride. The Estimation of the Iron in a solution of unknown strength is accomplished in the same manner as in standardizing the sodium-thiosulphate solution. Care must be taken that all the iron must be present as ferric oxide or chloride, and that the liquid contains no other substance that will decompose potassium iodide, e.g. , free chlorine or nitric acid. It is also advisable to employ solutions containing as nearly as possible about O'l grm. iron, and that not too little nor too much thiosulphate be used. The free acid must also be reduced in quantity, as detailed above. If it was found that 18 -4 c. c. of the thiosulphate solution corre- sponded to 0*1 grm. iron, and if there had been required 24*5 c. c. of the thiosulphate solution to combine with the iodine liberated by the unknown quantity of iron present, this last would then amount to 0*13315 grm., since 18-4: O'l : : 24-5:0-13315. The method gives good results, and is much to be recommended for determining small quantities of iron. y. Reduction by Sodium TTiiosulpKate in the presence of a Cupric Salt, after OUDEMANS.* If an acid solution of ferric chloride is mixed witli a little cupric * Sodium tliiosalpbate was first employed by SCIIERER (Gel. Ans. der K. Bayerischen Akademie, vom 31 Aug. 1859), afterwards by KREMER and LAN- DOLT (Zeitschr.f. analyt. Chem., I, 214). The method of OUDEMANS is to be found 113.] FERRIC IRON. 333 sulphate and some potassium sulphocyanate and then sodium thiosulphate is added, the red color of the ferric sulphocyanate gets paler and paler, and finally when the ferric salts are reduced to ferrous, disappears altogether. Warming is unnecessary. To hit the point is not easy, so we add a slight excess of sodium thio- sulphate and then titrate back with standard iodine. The reaction is as follows: Fe 2 Cl 6 + 21Sa 2 !S 2 O 3 _ 2FeCl a + 2NaCl -f !S T a 3 S 4 O 6 ; it is promoted by the addition of a small quantity of cupric sul- phate, which is alternately reduced by the thiosulphate and oxi- dized by the ferric chloride. If a small quantity of cuprous salt is produced by the excess of thiosulphate this does not matter, as its action on the iodine solution is the same in extent as the action of the thiosulphate which produced it. The method is not accurate unless the fluid remains clear; neither cuprous sul- phocyanate nor cuprous iodide nor sulphur must be thrown down. Hence care must be taken to maintain the proper amounts of the reagents and to dilute the fluid sufficiently. This method is much like that detailed under /?, in so far as it is most convenient to standardize the thiosulphate against a ferric- chloride solution of known strength, and then to use it on solutions of unknown strength to determine their iron content. "We require 1. A solution of sodium thiosulphate containing about 12 grm. (of the crystallized salt) per litre. 2. A solution of ferric chloride of known strength, prepared by dissolving 10 -04 grm. of clean, fine, and soft iron wire (=10 grm. pure iron) in hydrochloric acid in a slanting long-necked flask, oxidizing the solution with potassium chlorate, completely removing the excess of chlorine by protracted gentle boiling, and finally diluting the solution to 1 litre. 3. A solution of cupric sulphate, 1 in 100. 4. A solution of potassium sulphocyanate, 1 in 100. 5. A solution of iodine in potassium iodide, containing 5 or 6 grm. iodine in the litre (compare 146, 3). 6. Thin starch paste. Measure off some of the sodium thiosulphate, add starch paste ( 146, 3), and then titrate with iodine solution, in order to de- in Zeitschr. f. analyt. Chem., vi, 129; it was criticised and rejected in MOHR'S Lehrb. d. Titrirmethode, 3 Aufl. 291. OUDEMANS replied toMomi in Zeitschr. f. analyt. Chem., ix, 342, and an examination of the method by C. BALLING ap- peared in the same journal, ix, 99. 334 DETERMINATION. [ 113. terraine the relation between the two solutions. Now transfer 10 or 20 c. c. of the ferric chloride to a beaker, add 2 c. c. concen- trated hydrochloric acid, 100 or 150 c. c. water, 3 c. c. copper solution, and 1 c. c. potassium-sulphocyanate solution, titrate with sodium thiosulphate till the fluid just loses its color, add at once some starch paste, and titrate back with iodine solution till the blue color appears. Deduct the thiosulphate equivalent to the iodine solution from the total quantity of thiosulphate used ; the re- mainder will represent the amount required to reduce the iron present. In the analysis the conditions should be similar to those in the standardizing of the thiosulphate. This method is very rapid, and the results, though not so accurate as those by methods a and /?, are quite good enough for many technical purposes. Supplement to 112 and 113. Besides the methods given in 112 and 113, there have been many others, particularly indirect ones, advocated. Since these, however, possess no advantages over those described above, or are capable of only limited application, I will confine myself to a description of only the most important. 1. FITCH'S method* : Add hydrochloric acid to the solution, which must contain the iron as a ferric salt, and be free from nitric acid, and boil in contact with a few strips of metallic copper until the solution acquires a light-green color. Then estimate the iron from the loss in weight of the copper (Fe a Cl 6 -|- Cu = 2FeCl 2 -]- 2CuCl). The method yields good results only when the most careful attention is paid in excluding the air. The conditions most favorable to success have been studied by J. LOWE and KONIG, and are detailed, under the "Analysis of Iron Ores," in the Special Part. 2. The solution containing the iron as a ferric salt, and free from the metals of the fifth and sixth groups, as well as other sub- stances decomposable by hydrogen sulphide, is precipitated by an excess of a clear solution of hydrogen sulphide, avoiding all heat. The precipitated sulphur is determined after a few days, and the quantity of iron calculated therefrom according to the equation *Jour.f. prakt. Ghem., xvn, 160. 5 114.] URANIUM AND URANYL. 335 Fe a O 3 + H a S = 2FeO + H a O + S (H. EOSE). Eesults accurate. Compare also DELFFS.* 3. Add an excess of gold and sodium chloride to the solution containing the iron as a ferrous salt, close the bottle, and deter- mine the precipitated gold : 6FeCl a + 2 AuCl, = 3Fe a Cl. + 2Au (H. KOSE). Supplement to the Fourth Group. H4. 7. URANIUM AND URANYL. If the compound in which the uranium is to be determined contains no other fixed substances, it may often be converted into uranous uranate U(UO 4 ) 2 (called also uranoso-uranic oxide UO,- 2UO 3 ) by simple ignition. If sulphuric acid is present, small por- tions of ammonium carbonate must be thrown into the crucible towards the end of the operation. In cases where the application of this method is inadmissible, the solution of uranium (which, if it contains uranous salts, must ti rat be warmed with nitric acid, until they are converted into uranyl suits) is nearly boiled in a platinum or porcelain dish, and pre- cipitated with ammonia in slight excess. The yellow precipitate formed, which consists of hydrated ammonium uranate, is filtered off hot and washed with a dilute solution of ammonium chloride, to prevent the fluid passing milky through the filter. The precipitate is dried and ignited ( 53). To make quite sure of obtaining the uranous uranate in the pure state, the crucible is ignited for some time in a slanting position and uncovered ; the lid is then put on, while the ignition is still continuing ; the crucible is allowed to cool under the desiccator, and weighed (RAMMELSBERG). If the solution from which the uranyl is to be precipitated con- tains other basic radicals (alkali-earth metals, or even alkali metals), portions of these will precipitate along with the ammonium uranate. For the measures to be resorted to in such cases, I refer to Sec- tion Y. The reduction of the uranous uranate to the state of uranous *Chem. CentralbL, 1856, 839. 336 DETERMINATION. > oxide (UO 2 ) is an excellent means of ascertaining its purity for the purpose of control. This reduction should never be omitted, since PELIGOT has found the uranous uranate to be variable in composi- tion. It is effected by ignition in a current of hydrogen gas, in the way described 111, 1 (Cobalt). In the case of large quantities the ignition must be several times repeated, and the residue must be occasionally stirred with a platinum wire. While cooling increase the current of gas to prevent reabsorption of oxygen. By intense heating the property of spontaneous ignition in the air is destroyed. If after evaporating a solution of uranyl chloride, the residue is to be ignited in hydrogen, heat gently at first in the gas to avoid loss by volatilization. The separation of uranyl from phosphoric acid is effected by fusing the compound with potassium cyanide and sodium carbonate. Upon extracting the fused mass with water, the phosphoric acid is obtained in solution, whilst ura- nium is left as uranous oxide. KNOP and ARENDT* have employed this method. Taking 239 '6 as the atomic weight of uranium, uranous uran- ate, U(UO 4 ) a , contains 84-88 percent, of uranium and 15-12 per cent, of oxygen. UO a , uranous oxide, contains 88*22 per cent, uranium and 11 '78 per cent, of oxygen. According to BELOHOFBECK,t uranium may be also determined volumetrically by reducing the solution of uranyl acetate or sul- phate to uranous salts with zinc, as in the case of iron ( 113, 3, a). As the color of the solution is no safe criterion of the end of the reduction, you must allow the action of the zinc to continue for a considerable time. BELOHOUBECK says, a quarter of an hour is sufficient for small quantities, half an hour for large quantities. The solution of the uranous salt is diluted, mixed with dilute sul- phuric acid, and then titrated with permanganate to incipient red- dening. The permanganate is standardized by 112, 2; 1 at. uranium = 2 at. iron. BELOHOUBECK obtained good results also in hydrochloric solu- tions, but experiments made in this laboratory have shown that these are liable to the error pointed out in the case of iron (Comp. p. 319,7), at least in the presence of considerable quantities of hydrochloric acid. * Ctiem. Centralblatt, 1856, 773. f Zeitschr.f. analyt. Chem., vi, 120. 115.] SILVER. 337 Fifth Group. SILVER LEAD MERCURY IN MERCUROTJS COMPOUNDS MERCURY IN MERCURIC COMPOUNDS COPPER BISMUTH CADMIUM (PALLA- DIUM). 115. 1. SILVER. a. Solution. Metallic silver, and those of its compounds which are insoluble in water, are best dissolved in nitric acid (if soluble in that acid). Dilute nitric acid suffices for most compounds ; silver sulphide, however, requires concentrated acid. The solution is effected best in a flask, which should be heated if necessary, and placed in a slanting position if gas is evolved. In the case of metallic silver, or silver sulphide, the solution is heated finally to gentle boiling to drive off nitrous acid. Silver chloride, bromide, and iodide are insoluble in w r ater and in nitric acid. To get the silver contained in chloride and bromide in solution, proceed as follows : Fuse the salt in a porcelain crucible (this operation, though not absolutely indispensable, had better not be omitted), pour water over it, put u piece of clean cadmium, zinc, or iron upon it, and add some dilute sulphuric acid. Wash the reduced spongy silver, first with dilute sulphuric acid, then with water, and finally dissolve it in nitric acid. However, as we shall see below, the quantitative analysis of these salts does not necessarily involve their solution. 1>. Determination. Silver may be weighed as chloride, sulphide, or cyanide, or in the metallic state ( 82). It is also frequently determined by volu- metric analysis. We may convert into 1. SILVER CHLORIDE : All compounds of silver without excep- tion. 2. SILVER SULPHIDE : 3. SILVER CYANIDE : All compounds so^u ble in water or nitric acid. 4. METALLIC SILVER : Silver oxide and some silver salts of readily volatile acids; silver salts of organic acids; silver chloride, bro- mide, iodide, sulphide, and sulphate. The method -1 is the most convenient, especially when con- ducted in the dry way, and is preferred to the others in all cases 338 DETERMINATION. [ 115. where its application is admissible. The method 1 is that most generally resorted to. 2 and 3 serve mostly only to effect the separation of silver from other metals. In assays for the Mint, silver is usually determined volumetric- ally by GAY-LUSSAC'S method. PISANI'S volumetric method is especially suited to the determination of very small quantities of silver. H. YOGEL'S method is specially useful to photographers. The estimation of silver by cupellation will be detailed under " Analysis of Galena," in the Special Part. 1. Determination of Silver as Chloride. a. In the Wet Way. Mix the moderately dilute solution in a beaker with nitric acid,, heat to about 70, and add hydrochloric acid with constant stirring till it ceases to produce a precipitate. A large excess of hydro- chloric acid must be avoided, as the precipitate is not absolutely insoluble therein. While protecting the contents of the beaker from the action of direct sunlight continue the heat till the precipi- tate has fully settled, pour off the clear fluid through a small filter,, rinse the precipitate on to the latter by means of hot water mixed with some nitric acid, wash with hot water containing nitric acid,, then with pure hot water, dry thoroughly, transfer the precipitate to a watch-glass as nearly as possible, incinerate the filter in a weighed porcelain crucible, treat the ash (which always contains some metallic silver) with a few drops of nitric acid in the heat ; add two or three drops of hydrochloric acid, evaporate cautiously to dryness, add the main bulk of the precipitate, using a camel's- hair brush to transfer the last portions, heat cautiously till it begins to fuse at the edge, allow to cool, and weigh. To remove the fused mass without breaking the crucible, lay a small piece of iron or zinc upon it, and then add very dilute hydrochloric or sulphuric acid. The chloride will be reduced, and the silver may now be detached from the crucible with the greatest ease. For the properties of the precipitate see 82. The method gives very exact results, at all events in the absence of any con- siderable quantities of those salts in which silver chloride is some^ what soluble ; compare 82. To avoid error in this respect, it is well to test the clear filtrate with hydrogen sulphide. 115.] SILVER. 339 Ij. In the Dry Way. This method serves more exclusively for the analysis of silver bromide and iodide, although it can be applied in the case of other compounds. Fig. 87. The process is conducted in the apparatus illustrated by Fig. 87. a is a flask for disengaging chlorine ; it is completely filled with pieces of pyrolusite (native manganese dioxide) of the size of hazel- nuts, and half filled with strong hydrochloric acid ; contains concentrated sulphuric acid ; c contains calcium chloride; d is a bulb containing the silver iodide or bromide; e conducts the chlorine by means of a rubber tube into the open air or into a flask containing calcium hydroxide. The operation is commenced by introducing the compound to be analyzed into the bulb, and applying heat to the latter until its contents are fused ; when cold, the tube is weighed and connected with the apparatus. Chlorine gas is then evolved from a ; when the evolution of the gas has proceeded for some time, the contents of the bulb are heated to fusion, and kept in this state for about fifteen minutes, agitating now and then the fused mass. The bulb-tube is then removed from the apparatus, allowed to cool, and held in a slanting position to replace the chlorine by atmospheric air; it is subsequently 340 DETERMINATION. [ 115. weighed, then again connected with the apparatus, and the former process repeated, keeping the contents of d in a state of fusion for a few minutes. The operation may, in ordinary cases, be con- sidered concluded if the weight of the tube suffers no variation by the repetition of the process. If the highest degree of accuracy is to be attained, heat the silver chloride again to fusion, passing ;at the same time a slow stream of pure, dry carbon dioxide through the tube, in order to drive ont the traces of chlorine absorbed by the fused chloride. Allow to cool, hold obliquely for a short time r so as to replace the carbon dioxide by air, and finally weigh. 2. Determination as Silver Sulphide. Hydrogen sulphide precipitates silver completely from acid, neutral, and alkaline solutions ; ammonium sulphide precipitates it from neutral and alkaline solutions. The precipitate does not settle clearly and rapidly except a free acid or salt be present (such as nitric acid or an alkali nitrate). Recently prepared perfectly clear solution of hydrogen sulphide may be employed to precipitate small quantities of silver ; to precipitate larger quantities, the solu- tion of the salt of silver (which must riot be too acid) is moderately diluted, and washed hydrogen sulphide gas conducted into it. After complete precipitation has been effected, and the silver sul- phide has perfectly subsided (with exclusion of air), it is collected on a weighed filter, washed, dried at 100, and weighed. For the properties of the precipitate, see 82. This method, if properly executed, gives accurate results. The operator must take care to filter quickly, and to prevent the access of air as much as possible during the filtration, since, if this precaution be neglected, sulphur is likely to separate from the hydrogen sulphide water, which, of course, would add falsely to the weight of the silver sulphide. If the presence of a minute quantity of sulphur in the precipitate is suspected, treat it after drying with pure carbon disulphide on the filter repeatedly, till the fluid running through gives no residue on evaporation in a watch-glass ; dry and weigh. The sulphide must, however, never be weighed as just described, imless the analyst is satisfied that no considerable amount of sul- phur has fallen down with it, as would occur if the fluid contained liyponitric acid, a ferric salt, or any other substance which decom- poses hydrogen sulphide. In case the precipitate does contain much admixed sulphur, the simplest process is to convert it into 115.] SILVER. 341 metallic silver (H. ROSE*). For this purpose it is transferred to a weighed porcelain crucible, the filter ash is added, and the whole is heated to redness in a stream of hydrogen, the apparatus described in 108 being employed. Results accurate. Should the apparatus in question not be at the operator's dis- posal, he may, after complete washing of the precipitate, carefully rinse it into a porcelain dish (without injuring the weighed filter), heat it once or twice with a moderately strong solution of pure sodium sulphite, retransfer the precipitate (now freed from admixed sulphur) to the old filter, wash well, dry and weigh (J. LowEf) ; or he may treat the dried precipitate, together with the filter-ash, with moderately dilute chlorine-free nitric acid at a gentle heat, till complete decomposition has been effected (till the undissolved sulphur has a clean yellow appearance), filter, wash well, and pro- ceed according to 1, a. 3. Determination as Silver Cyanide. Mix the neutral solution of silver with potassium cyanide, until the precipitate of silver cyanide which forms at first is redissolved ; add nitric acid in slight excess, and apply a gentle heat. If the solution contains free acid, this must be first neutralized with pot- ash or sodium carbonate. After some time, collect the precipitated silver cyanide on a weighed filter, wash, dry at 100, and weigh. For the properties of the precipitate, see 82. The results are accurate. 4. Determination as Metallic Silver. a. In the Dry Way. Silver oxide, silver carbonate, &c., are easily reduced by simple ignition in a porcelain crucible. In the reduction of salts of organic acids, the crucible is kept covered at first, and a moder- ate heat applied ; after a time the lid is removed, and the heat increased, until the whole of the carbon is consumed. For the properties .of the residue, see 82. The results are absolutely accurate, except as regards silver salts of organic acids ; in the analysis of the latter, it not unfrequently happens that the reduced silver contains a minute portion of carbon, which increases the weight of the residue to a trifling extent. If it is desired to transform silver chloride, bromide, or sulphide * Pogg. AnnaL, ex, 139. \Journ.f. prakt. Chem., LXXVII, 73. 342 DETERMINATION. [ 115. into metallic silver, for the purpose of analysis, they are heated in a current of pure hydrogen to redness, till the weight remains constant. The process may be conducted in a porcelain crucible or a bulb-tube. In the former case, the apparatus described in 108 is used ; in the latter the apparatus represented in Fig. 87, with the substitution, of course, of hydrogen for chlorine ( 64, 14). If the bulb-tube is used, it must, after cooling and before being weighed, be held in an inclined position, so that the hydrogen may be replaced by air. The results are perfectly accurate. Silver iodide cannot be reduced in this way. I. In the Wet Way. If the silver solution is that of a nitrate, as is usual, add a little sulphuric acid and evaporate till all the nitric acid is expelled, dis- solve the silver sulphate in hot water, transfer it to a weighed porcelain crucible, arid immerse in the solution a rod of cadmium. The silver is rapidly reduced, and the precipitated metal may be easily removed from the cadmium and collected into a coherent mass. Warm the latter with the acid liquid until no more hydrogen is evolved, wash with hot water by decantation, dry, and ignite. Results accurate (A. CLASSEN*). Cadmium is preferable to zinc, because the latter usually leaves behind a little lead on solution in sulphuric acid. 5. Volumetric Methods. I. GAY-LUSSAC'S. This, the most exact of all known volumetric processes, was introduced by GAY-LUSSAC as a substitute for the assay of silver by cupellation, was thoroughly investigated by him, and will be found fully described in his work on the subject. This method has been rendered still more precise by the researches of G. J. MULDER, to whose exhaustive monograph f I refer the special student of this branch. I shall here confine myself to giving the process so far as to suit the requirements of the chemical laboratory, taking only for granted that the analyst has the ordinary measuring apparatus, . II. SCHWAKZ'S method.* To the nitric acid solution add ammonia or sodium carbonate, as long as the precipitate redissolves on shaking, mix with sodium acetate in not too small quantity, and then run in from a burette a solution of potassium dichromate (containing 14 '721 grm. in the litre) till the precipitate begins to settle rapidly. Now place on a porcelain plate a number of drops of a neutral solution of silver nitrate, and proceed with the addition of the chromate, two or three drops at a time, stirring carefully after each addition. When the precipitate has settled tolerably clear, which takes only a few seconds, remove a drop of the super- natant liquid and mix it with one of the drops of silver solution on the plate. A small excess of chromate gives at once a distinct red coloration ; the precipitated lead chromate does not act on the silver solution, but remains suspended in the drop. The number of c. c. of solution of chromate used (minus 0*1, which SCHWARZ deducts for the excess) multiplied by 0-0207 = the quantity of lead. If the fluid appear yellow before the reaction with the silver salt occurs, sodium acetate is wanting. In such a case first add more sodium acetate, then 1 c. c. of a solution containing 0'0207 lead in 1 c. c., complete the process in the usual way, and deduct 1 c. c. from the quantity of chromate used on account of the extra lead added. Any iron present must be in the form of a ferric salt ; metals the chromates of which are insoluble must be removed before the method can be employed. c. The lead is precipitated according to 1 , a, the carbonate (its composition is a matter of indifference in the present case) is washed, dissolved in a measured quantity of standard nitric acid ( 215), and a neutral solution of sodium sulphate added, whereby lead sulphate is precipitated and an equivalent quantity of sodium nitrate formed. If the nitric acid still free is now determined with standard alkali, we shall find the quantity of acid that has been neutralized by means of the lead, from which the amount of lead may be calculated, each c. c. of standard nitric acid being the equivalent of 0'1034() lead. You may also determine the free nitric acid by adding standard sodium carbonate till, the vessel being on * Dingl. polyt. Journ., CLXIX, 284; Zeitschr.f. analyt. Chem., n, 378. 117.] MERCURY IN MERCUROUS COMPOUNDS. 361 a black surface, a permanent turbidity is visible. Besults good (F. MOHE*). 117. 3. MERCUKY IN MEKCTJKOUS COMPOUNDS. a. Solution. Mercurous oxide and mercurous salts may generally be dissolved by means of dilute nitric acid, but without application of heat if conversion into mercuric compounds is to be avoided. If all that is required is to dissolve the mercury, the easiest way is to warm the substance for some time with nitric acid, then add hydrochloric acid, drop by drop, and continue the application of a moderate heat until a perfectly clear solution is produced, which now contains all the mercury in form of mercuric salts. Heating the solution to boiling, or evaporating, must be carefully avoided, as otherwise mercuric chloride may escape with the steam. b. Determination. If it is impracticable to produce a solution of the mercurous compound without formation of mercuric salts, it becomes neces- sary to convert the mercury completely into mercuric salts, when it may be determined as directed 118. But if a solution of a mercurous compound has been obtained, quite free from mercuric salts, the determination of the mercury may be based upon the insolubility of mercurous chloride, and effected either gravimetri- cally or volumetrically. The process of determining mercury, described 118, 1, a, may, of course, be applied equally well in the case of mercurous compounds. 1. Determination- as Mercurous Chloride. Mix the cold highly dilute solution with solution of sodium chloride, as long as a precipitate forms ; let the precipitate subside, collect on a weighed filter, dry at 100, and weigh. For the properties of the precipitate, see 84. Results accurate. If the jnercurons solution contains much free nitric acid, the greater part of this should be neutralized with sodium carbonate before adding the sodium chloride. * Lehrluch der Titrirmethode, 3. Aufl. 115. 362 DETERMINATION. [ 117. 2. Volumetric Methods. Several methods have been proposed under this head : the following are those which are most worthy of recommendation : a. Mix the cold solution with decinormal solution of sodium chloride ( 141, &, or), until this no longer 'produces a precipitate, and is accordingly present in excess ; filter and wash thoroughly, taking care, however, to limit the quantity of water used ; add a few drops of solution of potassium chromate, then pure sodium carbonate, sufficient to impart a light yellow tint to the fluid, and determine by means of solution of silver nitrate ( 141, &, a) the quantity of sodium chloride in solution, consequently the quantity which has been added in excess ; this shows, of course, also the amount of sodium chloride consumed in effecting the precipitation. One mol. of Hg 2 O is reckoned for 2 mols. of NaCl, consequently for every c.c. of the decinormal solution of sodium chloride, '0208 grm. of mercurous oxide. As filtering and washing form indis- pensable parts of the process, this method affords no great ad van tage over the gravimetric ; however, the results are accurate (FR. MOHK*). The two methods, 1 and 2, &, may also be advaii tageously combined. J. Precipitate the mercurous solution,f according to 1, with sodium chloride in a stoppered bottle, allow to subside, filter, wash, push a hole through the bottom of the filter, and rinse the precipi- tate into the bottle, which usually has some of the washed mercu- rous chloride adhering to its inside. Add a sufficient quantity of solution of potassium iodide, together with standard iodine solution {to 1 grin, Hg 2 01 2 about 2'5 grm. KI and 100 c.c. decinormal iodine .solution;}:), insert the stopper, and shake till the precipitate has entirely dissolved (Hg 2 Cl 2 + 6fel + 21 = 2[IIgI 2 (KI)J + 2KC1). As iodine is in excess, the solution appears brown. If any mercu- ric iodide separates, add potassium iodide to redissolve it. Now add from a burette solution of sodium thiosulphate correspond- ing to decinormal iodine solution till the fluid is decolorized and appears like water, transfer to a measuring flask, rinse and fill up to the mark, shake, take out an aliquot part, add starch paste to it, and determine the excess of sodium thiosulphate with decinormai iodine solution. After multiplying by the proper number, add the c.c. originally employed, subtract the c.c. of thiosulphate used, and * Lehrbuch der Titrirmethode, 3. Aufl. 395. \ If mercuric oxide is also present, see 118, 2. \ See 146, 2. 118.] MERCURY IN MERCUKIC COMPOUNDS. 363 calculate tlie quantity of mercury from the remainder. 2 at. iodine = 1 moL Hg a Cl a . Results good (!!EMPEL *). US. 4. MERCURY IN MERCUKIC COMPOUNDS. a. Solution. Mercuric oxide, and those mercuric compounds which are insoluble in water, are dissolved, according to circumstances, in hydrochloric acid or in nitric acid. Mercuric sulphide is heated with hydrochloric acid, and nitric acid or potassium chlorate added until complete solution ensues ; it is, however, most readily dis- solved by suspending it in dilute potassa and transmitting chlorine, at the same time gently warming (H. ROSE). When a solution of mercuric chloride is evaporated on the water-bath, mercuric chloride escapes with an aqueous vapor. This fact must not be lost sight of in effecting solutions of mercuric compounds. The methods proposed by YouLf give on this account inaccurate results. FR. MOHR \ and R. RIETH also have not given this source of error proper attention. 5. Determination. Mercury may be weighed in the metallic state, or as mercu- rous chloride, mercuric sulphide, or mercuric oxide ( 84) ; in separations it is sometimes determined as loss on ignition. It may also be estimated volumetrically. The first three methods may be used in almost all cases ; the determination as mercuric oxide, on the contrary, is possible only in mercurous or mecuric nitrates. The methods by which the mercury is determined as mercurous chloride or mercuric sul- phide are to be preferred before those in which it is separated in the metallic form. The volumetric method 5 is of very limited application. The mercurous chloride obtained by method 2, instead of being weighed, may be determined volumetrically as in 117, 2, I. *Annal d. Chem. u. Pharm.. ex, 176. \2bid.,xciv, 230. J Lehrbuch der Titrirmelhode, 3. Aufl., 208. RIETH'S Volumetric, 225. 364 DETERMINATION. [ 118. 1. Determination as Metallic Mercury. a. In the Dry Way. The process is conducted in the apparatus illustrated by Fig. 88. Take a tube 45 cm. long and about 12 mm. wide, made of difficultly fusible glass and sealed at one end. First put into the tube a mixture of sodium bicarbonate and powdered chalk 6 cm. long, then a layer of quicklime ; these two will occupy the space from a to 5. Then add the intimate mixture of the substance with an excess of quicklime (&- ),+ 2JSTa 2 S0 4 + 4HN O 3 ; or, 2H 2 O + 3HgCl f + 2tfa,S a O. = (HgS), . Hg Cl a + 2N~a 2 SO 4 -f- 4.HC1. The process is conducted as follows in the case of mercuric nitrate : Mix the highly dilute solution with a little free nitric acid in a tall glass, and add drop by drop solution of sodium thiosulphate 12*4 grm. in a litre. Each drop produces an intense yellow cloud, which on shaking quickly subsides in the form of a heavy flocculent precipitate (HgS) 2 Iig(NO 3 ) 2 . In order- to distinguish clearly the exact end of the reaction, SCHEBEB recommends to transfer the fluid towards the end to a measuring flask, to take out -J or % of the clear fluid and to finish with this. The portion of thiosulphate last used. is multiplied by 3 or 2, as the case may be, and added to the quantity first used. 1 c.c. of the solution corresponds to 0'015 mercury, or 0'0162 mercuric oxide. The relation is not changed even when the fluid contains another acid (sulphuric, phosphoric). In the case of mercuric chloride, the highly dilute solution is mixed with a little hydrochloric acid and warmed, nearly to boil- ing, before beginning to add the sodium thiosulphate. At first a white turbidity is formed, then the precipitate separates in thick flocks. When the solution begins to appear transparent, the pre- cipitant is added more slowly. In order to hit the end of the reaction exactly, small portions must be filtered off towards the close. The precipitate must be completely white ; if too much thiosulphate has been added, it is gray or blackish, and the experi- ment must be repeated. SCHEEEB obtained very accurate results. Of course no other metals must be present that exert a decompos- ing action on sodium thiosulphate. * Journ. dePharm. et de Chem., XLIII, 477. f Lehrbuch der Ghemie, i, 513. 370 DETERMINATION. [ 119. 5. COPPER. a. Solution Many cupric salts dissolve in water. Metallic copper is best dissolved in nitric acid. Cupric oxide, and those cupric salts which are insoluble in water, may be dissolved in nitric, hydrochloric, cr sulphuric acid. Cupric sulphide is treated with fuming nitric acid, or it is heated with moderately dilute nitric acid, until the separated sulphur exhibits a pure yellow tint ; addition of a little hydro- chloric acid or potassium chlorate greatly promotes the action of the dilute acid. l>. Determination. Copper may be weighed in the form of cupric oxide, or in the 'metallic state, or as cuprous sulphide ( 85). Into the form of cupric oxide it is converted by precipitation, or ignition, sometimes with previous precipitation as sulphide. The determination as cuprous sulphide is preceded usually by precipitation either a& cupric sulphide or as cuprous sulphocyanate. Copper may be deter- mined also by various volumetric and indirect methods. We may convert into 1. CUPRIC OXIDE : a. By Precipitation as hydrated cupric oxide and subsequent ignition : All cupric salts soluble in water, and also those insoluble salts, the acids of which may be removed upon solution in nitric acid, provided no non-volatile organic substances be present. 1}. By Precipitation, preceded by Ignition of the compound : Such of the salts enumerated under a as contain a n on- volatile organic substance, thus more particularly cupric salts of non-vola- tile organic acids. c. By Ignition : Cupric salts of oxygen acids that are readily volatile or decomposable at a high temperature (cupric carbonate, cupric nitrate). 2. METALLIC COPPER : Copper in all solutions free from other metals precipitable by zinc or the galvanic current, also the oxides of copper. 3. CUPROUS SULPHIDE : Copper in all cases in which no other metals are present that are precipitable by hydrogen sulphide or potassium sulphocyanate. 119.] COPPER. 371 Of the several methods of effecting the estimation of copper, ~No. 3 is particularly to be recommended for use in laboratories ; method 2 is also very convenient, and well adapted for assaying. Of the volumetric methods, one is suited for technical purposes, the other for the estimation of small quantities of copper. For technical purposes there are, besides, also several col ori metric methods, proposed by HEINE, VON HUBERT, JACQUELAIN, A. MUL- LER, and others, which are, all of them, based upon the comparison of an ammoniacal solution of copper, of unknown strength, with others of known strength.* LEVOL'S indirect method of estimating copper, which is based upon the diminution of weight suffered by a strip of copper when digested in a close-stoppered flask with ammoniacal solution of copper till decolorization is effected, takes too much time, and is apt to give false results (PmLLiPs,t ERDMANN^:). The latter remark applies also to the indirect method proposed by KUNGE, which con- sists in boiling the solution of copper, free from nitric acid and ferric salts, in presence of some free hydrochloric acid, in a flask, with a weighed strip of copper, and, after decolorization of the fluid, determining the loss of weight suffered by the copper. 1. Determination as Cupric Oxide. a. By direct Precipitation as Oxide. Heat the rather dilute neutral or add solution in a platinum or porcelain dish, to incipient ebullition, add a somewhat dilute solu- tion of pure soda o* potassa until the formation of a precipitate ceases, and keep the mixture a few minutes longer at a tempera- ture near boiling. Allow to subside, filter, wash by decantation twice or thrice, boiling up each time, then collect it on the filter, wash thoroughly with hot water, dry, and ignite in a porcelain or platinum crucible, as directed 53. Do not use the blow-pipe. After ignition, and having added the ash of the filter, let the crucible cool in the desiccator, and weigh. The action of reducing gases must be carefully guarded against in the process of ignition. It will sometimes happen, though mostly from want of proper attention to the directions here given, that particles of the precipi- *This subject hardly comes within the scope of the present work. I there- fore refer to AL. Mr LLKH. das Complementiircolorimeter, Chemnitz, 1854; Bo- DEMANN'S Probirkunst von KERL, 222; also to DKHMS, Zeitschr.f.analyt. Chem. t in, 218, and GUSTAV BISCHOF, jun., jf> , vr, 459. \Annal. d. Chem. u. Pharm., LXXXI, 208. \Journ.f prakt. Chem., LXXV, 211. 372 DETERMINATION. . [ 110. tate adhere so tenaciously to the dish as to be mechanically irremov- able. In a case of this kind, after washing the dish thoroughly, dissolve the adhering particles with a few drops of nitric acid, and evaporate the solution over the principal mass of the precipitated oxide, before you proceed to ignite the latter. Should the solution be rather copious, it must first be concentrated by evaporation, until only very little of it is left. For the properties of the pre- cipitate, see 85. With proper attention to the directions here given, the results obtained by this method are quite accurate, otherwise they may be either too high or too low. Thus, if the solution be not sufficiently dilute, the precipitant will fail to throw down the whole of the copper ; or if the precipitate be not thoroughly washed with hot water, it will retain a portion of the alkali ; or if the ignited pre- cipitate be allowed to stand exposed to the air before it is weighed, an increase of weight will be the result ; and so, on the other hand, a diminution of weight, if the oxide be ignited with the filter or under the influence of reducing gases, as thereby cuprous oxide would be formed. Should a portion of the oxide have suffered reduction, it must be reoxidized by moistening with nitric acid, evaporating cautiously to dryness, and exposing the residue to a gentle heat, increasing this gradually to a high degree of intensity. Let it be an invariable rule to test the filtrate for copper with hydrogen sulphide water. If, notwithstanding the strictest compli- ance with the directions here given, the addition of this reagent produces a precipitate, or imparts a brown tint to the fluid, this is to be attributed to the presence of organic matter ; in that case, concentrate the filtrate and wash-water by evaporation, acidify, precipitate with hydrogen sulphide water, filter, incinerate the filter, heat with nitric acid, dilute, filter, concentrate, precipitate with soda, and add the oxide obtained to the main quantity. Never neglect to test the cupric oxide after weighing for alkali or alkali salt by boiling it with water. If either is present, the oxide must be exhausted with hot water, and then reignited and reweighed. Finally, dissolve the oxide in hydrochloric acid to detect and if necessary to estimate any silicic acid it may contain. In default of sufficiently pure potash or soda, the carbonate may be used, but the solution must not contain more than 1 grm. copper in the litre ; the alkali carbonate must only be added slightly in excess, and the mixture must be boiled for half an hour. 119.] COPPER. 373 The bluish-green precipitate will then turn dark brown and gran- ular, and may b easily washed (GIBBS*). From ammwiiacal solutions, also, copper may be precipitated by soda or potassa. In the main, the process is conducted as above. After precipitation the mixture is heated, until the supernatant fluid has become perfectly colorless ; the fluid is then filtered off with the greatest possible expedition. If allowed to cool with the precipitate in it, a small portion of the latter would redissolve. b. By Precipitation as Oxide, preceded by Ignition of the Substance. Heat the substance in a porcelain crucible, until the organic matter present is totally destroyed ; dissolve the residue in dilute nitric acid, filter if necessary, and treat the clear solution as directed in a. c. By Ignition. The salt is put into a platinum or porcelain crucible, and exposed to a very gentle heat, which is gradually increased to intense redness ; the residue is then weighed. As cupric nitrate spirts strongly when ignited, it is always advisable to put it into a small covered platinum crucible, and to place the latter in a large one, also covered. With proper care, the results are accurate. Cupric salts of organic acids may also be converted into cupric oxide by simple ignition. To this end, the residue flrst obtained, which contains cuprous oxide, is completely oxidized by ignition with mercuric oxide (which leaves no residue on ignition), or, with less advantage, by repeated moistening with nitric acid, and ignition. A loss of substance is generally incurred by the use of nitric acid from the difficulty of avoiding spirting. 2. Determination as Metallic Copper, a. By Precipitation with Zinc or Cadmium.^ Introduce the solution of copper, after having, if required, first freed it from nitric acid, by evaporation with hydrochloric acid or * Zeitschr. f. analyt. Cliem., vir, 258. f The method of precipitating copper by iron or zinc and weighing it in Uie metallic form was proposed long ago; see PFAFP'S Handbuch der analytischen Chemie, Altona, 1822, u, 269; where the reasons are given for preferring zinc as a precipitant, and hydrogen sulphide is recommended as a test for ascertaining whether the precipitation is complete. I mention this with reference to F. MOHR'S paper in the Annal. d. Chem. u. Pharm., xcvi, 215, and BODE- MANN'S Probirkunst von KERL, 220. 374 DETERMINATION. [ 119. sulphuric acid, into a weighed platinum dish , dilute, if necessary with some water, throw in a piece of zinc (soluble in hydrochloric acid without residue), and add, if necessary, hydrochloric acid in sufficient quantity to produce a moderate evolution of hydrogen. If, on the other hand, this evolution should be too brisk, owing to too large excess of acid, add a little water. Cover the dish with a watch-glass, which is afterwards rinsed into the dish with the aid of a washing-bottle. The separation of the copper begins imme- diately ; a large proportion of it is deposited on the platinum in form of a solid coating; another portion separates, more particu- larly from concentrated solutions, in the form of red spongy masses. Application of heat, though "it promotes the reaction, is not abso- lutely necessary ; but there must always be sufficient free acid present to keep up the evolution of hydrogen. After the lapse of about an hour or two, the whole of the copper has separated. To make sure of this, test a small portion of the supernatant fluid with hydrogen sulphide water ; if this fails to impart a brown tint to it, you may safely assume that the precipitation of the copper is complete. Ascertain now, also, whether the zinc is entirely dis- solved, by feeling about for any hard lumps with a glass rod, and observing whether renewed evolution of hydrogen will take place upon addition of some hydrochloric acid. If the results are satis- factory in this respect also, press the copper together with the glass rod, decant the clear fluid, which is an easy operation, pour, with- out loss of time, boiling water into the dish, decant again, and repeat this operation until the washings are quite free from hydro- chloric acid. Decant the water now as far as practicable, rinse the dish with strong alcohol, dry at 100, let it cool, .and weigh. If you have no platinum dish, the precipitation may be effected also in a porcelain crucible or glass dish ; but it will, in that case, take a longer time, because of the lack of the galvanic action between the platinum and zinc; and the whole of the copper will be obtained in loose masses, and not firmly adhering to the sides of the crucible or dish, as in the case of precipitation in platinum vessels. , The results are very accurate. The direct experiment, No. 69, gave 100 and 100-06, instead of 100. FK. Moim (loo. cit.) obtained equally satisfactory results by precipitating in a porce- lain crucible.* * STOKER (On the alloys of copper and zinc, Cambridge, 1860, p. 47) says that the precipitated copper retains water, but I have not found this to be the case. 119 ] COPPER. 375 Zinc being sometimes difficult to obtain of sufficient purity, cadmium may be used instead; it dissolves with less violence in strongly acid copper solutions. It may be used in the form of rod in which it usually occurs in commerce (CLASSEN*). h. By Precipitation with the Galvanic Current. This method makes us independent of pure zinc or cadmium, and yields the copper in a compact form, readily washed and deter- mined. It is now largely used in copper works, constant batteries have been employed for it, and the whole process has been organ- ized for use on a large scale by LTJCKOW, and adopted by the Mans- feld Ober-Berg-und Hiitten-Direction in Eisleben.f A small elec- trolytic apparatus without separate battery, for single precipitations, has been described by ULLGKEN.^: c. By Ignition in Hydrogen. The oxides of copper when ignited in a current of pure hydro gen are converted into metallic copper, and may thus be convex ientiy analyzed. Occasionally the cupric oxide obtained by 1, a <,, J, is reduced either at once, or after weighing ; in the latter casr, the reduction serves as a control. 3. Determination as Cuprous Sulphide. a. By Precipitation as Cupric Sulphide. Precipitate the solution which is best moderately acid, bu, should not contain a great excess of nitric acid according to th? quantity of copper present, either by the addition of strong hydro- gen sulphide water, or by passing the gas. In the absence of nitric acid it is well to heat nearly to boiling while the gas is passing, as this makes the precipitate denser, and it is more easily washed. When the precipitate has fully subsided, and you have made sura that the supernatant fluid is no longer colored or precipitated by strong hydrogen sulphide water, filter quickly, wash the precipi- tate without intermission with water containing hydrogen sulphide, and dry on the filter with some expedition. Transfer to a weighed porcelain crucible, add the filter-ash and some pure powdered sul-* phur, and ignite strongly in a stream of hydrogen ( 108, Fig. 83). It is advisable to use a glass blow-pipe. The results are very accurate (II. KOSE). * Journ. f. prakt. Chem., xcvi, 259. \Zeiischr. f. analyt. Chem., vm, 23 and xi, 1. Compare also G^BBS, ib. t in, 334, and LECOQ DE BOISBAUDAN, ib., vn, 253. $lb. , vii, 442. Pogg. Annal. ex, 138. 370 DETERMINATION. [119-. This method, which was recommended by BERZELIUS, and afterwards by BRUNNER, has only lately received a very practical form from the apparatus introduced by II. ROSE. I feel great pleasure in recommending it. In my own laboratory it is in frequent use. If the precipitated cupric sulphide is ignited instead in a current of hydrogen in a covered porcelain crucible, from which the heat as well as the cover are removed occasionally for a few seconds, the contents will be converted into a variable mixture of Cu a S and CuO, which may contain, according to circumstances, cupric oxide or cuprous sulphide. Since, however, the percent- age content of cupric oxide and cuprous sulphide in copper is the same, the copper content in the residue may also be deter- mined (ULRICI *). This method is simpler than the one detailed above, but is not quite as accurate. 1). By Precipitation as Cuprous /Sulphocyanate, after Rivox.f The solution should be as free as possible from nitric acid and free chlorine, and should contain little or no free acid. Add sul- phurous or hypophosphorous acid in sufficient quantity, and then solution of potassium sulphocyanate in the least possible excess. The copper precipitates as white cuprous sulphocyanate. It is filtered after standing some time, washed and dried, mixed with sulphur, ignited in hydrogen in the apparatus mentioned in , and this ignition with sulphur is repeated till the weight is constant. The precipitate may also be collected on a weighed filter, dried at 100, and then weighed. The experiment, No. 71, conducted in the latter way, gave 99*66 instead of 100. The process yields satisfactory results, but they are always inclined to be a little too low, as the cuprous sulphocyanate is not absolutely insoluble. The loss is larger in the presence of much free acid. c. Cuprous and cupric oxide, cupric sulphate, and many other salts of copper (but not chloride, bromide, or iodide) may be directly converted into cuprous sulphide, by mixing with sulphur and igniting in hydrogen as in a (II. ROSE, loc. cit.). The results are thoroughly satisfactory. *Journ.f. prakt. Chem., cvn, 110. f Compt. Rend., xxxvni, 868; Journ. f. prakt. Chem., LXII, 252. 119.] COPPER. 377 4. Volumetric Methods. a. DE HAEN'S METHOD.* I recommend this method, which was devised in my own laboratory,! as more especially applicable in cases where small quantities of copper are to be estimated in an expeditious way. The method is based upon the fact that, when a cupric salt in solution is mixed with potassium iodide in excess, cuprous iodide and free iodine are formed, the latter remaining dissolved in the solution of potassium iodide : CuSO 4 + 2KI = Cul + K 2 SO 4 + I. Now, by estimating the iodine by BUNSEN'S method, or with sodium thiosnlphate ( 146), we learn the quantity of copper, as 1 a L . iodine (126'85) corresponds to 1 at. copper (63 -6). The following is the most convenient way of proceeding : Dissolve the compound of copper in sulphuric acid, best to a neutral solution ; a moderate- excess of free sulphuric acid, however, does not injuriously affect the process. Dilate the solution, in a measuring flask, to a defi- nite volume ; 100 c.c. should contain from 1 to 2 grm. of copper. Introduce now about 10 c.c. of potassium iodide solution (1 in 10) into a stoppered bottle, add 10 c.c. of the copper solution, mix, allow to stand 10 minutes, and then determine the separated iodine, either with sulphurous acid and iodine ( 146, 1), or with sodium thiosnlphate ( 146, 2). The copper solution must be free from ferric salts and other bodies which, decompose potassium iodide, also free nitric acid, and free hydrochloric acid; -and the* solution must not be allowed to stand too long before titration.. With strict attention to these rules, the results are quite accurate. DE HAEN obtained, for instance 0*3567 instead of O3566 of cupric sulphate, 99 -89 and 100 -1 instead of 100 of metallic copper- Further experiments (No. 72) have convinced me, however, that, though the results attainable by this method are satisfactory, they are not always quite so accurate as would be supposed from the above figures given by DE HAEN. Acting upon FR. MOHR'S suggestion I tried to counteract the injurious influence of the presence of * Annal. d. Chem. u. Pharm., xci. 237. f BROWN (Quart. Journ. of the Chem. Soc., x, 65), who published this as a new method iu 1857, appears to have been ignorant of its publication in 1854. Even the slight variation of determining the iodine with sodium hyposulphite (SCHWAKZ) instead of with sulphurous acid (BUNSEN) was given by MOHR (Lehrbuch der Titrirmethode, i, 387) in 1855. The same may be said of RUMP- LER, who in 1868 (Journ. f prakt. Chem., cv, 193) published the method, with a slight modification, as new. 378 DETERMINATION. [ 119. nitric acid, by adding to the solution containing nitric acid, first, ammonia in excess, then hydrochloric acid to slight excess ; the result was by no means satisfactory. The reason of this is that a solution of ammonium nitrate, mixed with some hydrochloric acid, will, even after a short time, begin to liberate iodine from solution of potassium iodide. 5. PARKES' * METHOD; AND H. FLECK' sf MODIFICATION. PARKES' expeditious method is based on the action of potassium cyanide on ammoniacal copper solution. On adding potassium cyanide to the azure-blue fluid, the color disappears, CuCy, OTI 4 Cy, and KOIi being formed, while one equivalent of cyanogen is liberated, and acting on the free ammonia present, yields urea, urea oxalate, ammonium cyanide, and ammonium formate (LIE- BIG- :{;). The decomposition is not always uniform, however, tlrj quantity and strength of the ammonia having considerable in- fluence ; see LIEBIG (loo. cit.'), as also my experiments (No. 73, &), from which it appears that the neutral ammonium salts present modify the results. See also FLECK (loo. cit.\ v. WOLFSKRON, STEINBECK,| and KIRPITSCHOW.^ FLECK proposed the following modification : Instead of am- monia, a 1 : 10 solution of ammonium sesquicarbonate is used at a temperature of 60, the end of the reaction being rendered more readily determined by adding 2 drops of a 1 : 20 potassium-fer- rocyanide solution, neither the blue color nor its transparency be- ing affected by this addition. The potassium-cyanide solution is standardized against a copper solution of known strength, before being employed for solutions of unknown strength. On adding the potassium-cyanide solution by drops to the blue solution warmed to 60, the odor of cyanogen becomes quite distinct, while the color of the solution becomes gradually paler. As soon as the copper double salt is decomposed, the red color of copper ferrocyanide becomes visible without any precipitate forming, and on adding the last drop of the potassium-cyanide solution this color fades away also and leaves a perfectly colorless liquid. * Mining Journal, 1851. \Polytechn. Centralbl., 1859, 1313. \ Annal. d- CTiem. u. Pharm,, xcv. 118. $Zeitschr.f. analyL Chem., v. 403. || Ibid., vin, 16. 1 Zeilschr. f. Chem. (II), vn, 207. 119.] COPPER. 379 This modification yields results which while concordant are yet only approximate.* Where such suffice, the method may be used, as it is quite convenient. I have found that in this method also, ammonium salts, if present, have an influence on the results (see Exp. ]STo. 73, 5), hence the method appears to he useful only when the standard- ization of the potassium-cyanide solution and the analytical proc- esses are performed under similar circumstances. On this principle is based STEINBECK' sf method, which was devised for estimating the copper in the Mansfeld shales, and which received a premium from the Mansfeld Ober-Berg-und Hiitten-Direction. It depends upon the precipitation of metallic copper from a hydrochloric-acid solution by zinc in contact with platinum. After being washed, the metallic copper is dissolved in a definite quantity of nitric acid, a definite quantity of am- monia added, and the standard solution of potassium cyanide then added until decolorization is effected. Since, in this method, only definite quantities of ammonia and ammonium nitrate are present, the results obtained are very concordant, and also very nearly correct if the cyanide solution is standardized against a copper solution the strength of which is approximately like that of solution to be operated upon. The cyanide solution should be made of such strength that 1 c. c. of it is the equiva- lent of 0*005 grin, of copper. c. METHODS DEPENDING UPON THE PRECIPITATION OF COPPER BY SODIUM SULPHIDE. PELOUZE supersaturates the neutral or acid copper solution with ammonia, heats the solution to between 60 and 80, and adds so- dium sulphide until the blue color just disappears. The precipitate that forms at this temperature has the composition 5CuS + CuO. As the temperature is not without influence on the composition of the precipitate, and as the moment of disappearance of the blue color is not very marked, FR. MOHR J and KUNZEL have * FLECK used in 6 tests, in which varying quantities of ammonium carbonate were purposely taken for 100 c. c. of copper solution, a minimum of 15'2 c. c. and a maximum of 15 75 c. c., an average of 15*46 c. c. of potassium-cyanide solution. \ Zeitsckr. f. analyt. CJiem., vui, 8. i Lehrbuch der Titrirmethode , 3. Ann., 429. Jour.f. prakt. Chem. t LXXXVIII, 486; Zeitschr. f. analyt. C7iem., n, 373. 380 DETERMINATION. [ 119. modified the method. The former precipitates in the cold (whereby cupric sulphide is formed), and ascertains the incipient excess of sodium sulphide by using alkaline-lead solution. The latter precipitates at the boiling temperature (the oxysulphide formed in this case rapidly settles), and ascertains when the pre- cipitation of copper is complete by bringing a drop of the fluid into contact with freshly precipitated hydrated zinc sulphide (it should not be colored brown). The sodium-sulphide solution should be diluted so that 1 c. c. will precipitate about O'Ol grm. copper. It may be standardized by using a solution containing 10 grm. of copper per litre. 20 c. c. are taken, representing 0'2 grin, of copper, supersaturated with ammonia, then diluted with water, heated to boiling, and sodium- sulphide solution then added until the reaction is complete. The zinc sulphide required is pre-< pared by dissolving ordinary zinc in hydrochloric acid, adding an excess of ammonia, and boiling with a small quantity of sodium- sulphide solution, whereby any lead present is precipitated. Suf- ficient sodium-sulphide solution is now added to precipitate nearly all of the zinc (leaving a small quantity unprecipitated) ; the magma obtained is uniformly spread out over several layers of blotting-paper. According to KUNZEL the method, if carefully carried out, gives errors not exceeding 0*25 per cent.; hence it is perfectly suitable for technical purposes. d. METHODS DEPENDING UPON THE REDUCTION OF CUPKIO CHLORIDE BY STANNOUS CHLORIDE. E. MULDER * was the first to base upon this reaction a method of estimating copper, using indigo-carmine as an indicator. FR. WEIL f found that if sufficient hydrochloric acid is present, the end of the reaction is indicated by the decolorization of the hot liquid. He prepared the stannous- chloride solution by dissolving 6 grm. of tinfoil in 200 c. c. of hot hydrochloric acid and diluting the solution with boiled water to make 1 litre. The copper solu- tion, against which the stannous-chloride solution must be stand- ardized before every fresh series of estimations, is prepared by *Jakresber. von KOPP u. WILL, 1860, 613. \Zeitschr.f. analyt Chem., ix, 297. 119.] COPPER. 381 dissolving 7'854 grm. of copper sulphate (= 2 grm. Cu.), pow- dered and dried by pressure between blotting-paper, in water to make 500 c, c. 25 c. c. of this solution (containing O'l grm. cop- per) are then introduced into a 100-c. c. flask, 5 c. c. of pure, concentrated hydrochloric acid added, the whole heated to gentle boiling, and stannous-chioride solution added to the boiling liquid, rapidly at first, but towards the last by drops, until the fluid is perfectly colorless. 5 c. c. of hydrochloric acid are again added ; if a slight color develops, it is discharged by adding a few drops of , stannous-chioride solution. The further certainty that the reaction is complete is afforded on adding a few drops of mercuric-chloride solution to a small quantity of the cooled solu- tion ; if no turbidity is noticeable there is no excess of stannous chloride present, hence a little of the latter may be added until a faint precipitate of mercurous chloride is developed. In this case, however, there must be deducted 0-05 c. c. from the quantity of stannous-chioride solution used. In titrating a copper solu- tion proceed similarly. Any nitric acid present must be evap- orated off after adding an excess of sulphuric acid. If any fer- ric salt is present, it will be reduced with the cupric chloride. In such a case precipitate the copper in a second portion of the solution with zinc and platinum wire in the heat, determine the ferrous salt with potassium permanganate or chrornate ( 112), and calculate how much of the stannous-chioride solution had been used to reduce the ferric salt : the remainder will be that used up for the cupric chloride ; or wash the precipitated cop- per, dissolve it in sulphuric acid, and then reduce it with stan- nous chloride. The test analyses cited by WEIL show very satis- factory results. e. SCHWARZ * precipitates cuprous oxide from the solution of potassio-cupric tartrate by heating with grape sugar, filters off the precipitate, washes it, warms it with ferric chloride and hy- drochloric acid, and, according to the equation Cu 2 O -f- Fe 3 Cl 6 + .2IICl = 2CuCl 1 + 2FeCl f + H 1 O, estimates the ferrous chloride formed by means of potassium permanganate.f * A/Dial d. Chem. n. Pharm . LXXXIV, 84. f Potassium Hiromate is not eligible for use because the cupric chloride im- pairs the distinctness of the end reaction. 382 DETERMINATION. [ 120. f. E. FLEISCHER * precipitates the copper as cuprous sulplio- cyanate ( 119, 3, 5), boils the washed precipitate with potassa lye, arid thus obtains cuprous oxide ; or he adds stannous chloride and potassium iodide and obtains a precipitate of cuprous iodide. In either case the precipitate is brought into contact with ferric- sulphate solution, the ferrous salt formed estimated, and from this the copper calculated. g. F. FLEITMANN f precipitates the copper with zinc, brings the washed precipitate into contact with ferric chloride and hy- drocholoric acid, and estimates the ferrous chloride formed (Cu + Fe.01. = CuCl 2 + 2FeCl a ). h. H. SCHWARZ adds potassium xanthogenate to the acetic- acid solution of copper until no further precipitate forms. Since the other heavy metals, excepting zinc, are also precipitated by the reagent from acetic-acid solutions, the copper must be separated from the precipitate. As noted, the methods e to li require the previous precipita- tion or isolation of the copper in one way or another ; they can- not therefore be preferred to gravimetric methods, excepting in very special cases. 120. 6. BISMUTH. a. Solution. Metallic bismuth, bismuth trioxide, and all other compounds of that metal, are dissolved best in nitric acid more or less diluted. It must be borne in mind that hydrochloric-acid solutions of bis- muth, if concentrated, cannot be evaporated without loss of bismuth chloride. 1). Determination. Bismuth is weighed in the form of trioxide, chromate, sul- pJiide, or arsenate, or in the metallic state. The compounds of bis* muth are converted into trioxide by ignition, by precipitation as basic carbonate, or by repeated evaporation of the nitric-acid solu- * ZeitscJir. /. analyt. Chem., ix, 255. f Annal. d. Chem. u. Pharm., xcvm, 141. \Dingl. polyt. Journ., cxc, 220 and 295; also Zeitschr. f. analyt. Chem.,. vin, 462. 120.] BISMUTH. 383 tion, These are sometimes preceded by separation as sulphide. The determination as metallic bismuth is frequently preceded by precipitation as sulphide or as basic chloride. We may convert into 1. BISMUTH TRIOXIDE: a. By Precipitation as basic Bismuth Carbonate. All com- pounds of bismuth which dissolve in nitric acid to nitrate, no other acid remaining in the solution. b. By Ignition. a. Bismuth salts of readily volatile oxygen acids, /?. Bismuth salts of organic acids. c. By Evaporation. Bismuth in nitric-acid solution. d. By Precipitation as Bismuth Trisulphide. All compounds of bismuth without exception. 2. BISMUTH CHROMATE. All compounds named in 1, a. 3. BISMUTH TRISULPHIDE. The compounds of bismuth without exception. 4. METALLIC BISMUTH : The trioxide and oxygen salts, the sulphide, the basic chloride, in which latter form the bismuth may be precipitated out of allots solutions. 1. Determination of Bismuth as Trioxide. a. By Precipitation as Bismuth Carbonate. If the solution is concentrated add water, taking no notice of any precipitate of basic nitrate that may be formed. Mix with ammonium carbonate in very slight excess, and heat for some time nearly to boiling ; filter, dry the precipitate, and ignite in the man- ner directed 116, 1 (Ignition of lead carbonate) ; the process of ignition serves to convert the carbonate into bismuth trioxide. For the properties of the precipitate and residue, see 86. The method gives accurate results, though generally a trifle too low, owing to the circumstance that bismuth carbonate is not absolutely insoluble in arumonium carbonate. Were you to attempt to precipitate bismuth, by means of ammonium carbonate, from solutions con- taining sulphuric acid or hydrochloric acid, you would obtain incorrect results, since with the basic carbonate, basic sulphate or basic chloride would be precipitated, which are not decomposed by excess of ammonium carbonate. Were you to filter off the precipi- tate without warming, a considerable loss would be sustained, as 384 DETERMINATION. [ 120. the whole of the basic carbonate would not have been separated (Expt. No. 74). l>. By Ignition. a. Compounds like bismuth carbonate or nitrate are ignited in a porcelain crucible until their weight remains constant. ft. Salts of organic acids are treated like the corresponding compounds of copper ( 119, 1, c). c. By Evaporation. The solution of the nitrate is evaporated, in a porcelain dish on the water-bath, till the neutral salt remains in syrupy solution ; add water, loosen the white crust that is formed with a glass rod from the sides, evaporate again on a water-bath, reprecipitate with water, and repeat the whole operation three or four times. After the dry mass on the water-bath has ceased to smell of nitric acid, it is allowed to cool thoroughly, and then treated with cold water containing a little ammonium nitrate (1 in 500) ; after the residue and fluid have been a short time together, filter, wash with the weak solution of ammonium nitrate, dry and ignite ( 53). Results very satisfactory (J. LOWE*). d. By Precipitation as Bismuth Trfeulphide. Dilute the solution with water slightly acidulated with acetic acid (to prevent the precipitation of a basic salt), and precipitate with hydrogen sulphide water or gas ; allow the precipitate to subside, and test a portion of the supernatant fluid with hydrogen sulphide water: if it remains clear, which is a sign that the bismuth is completely precipitated, filter (the filtrate should smell strongly of H a S), and wash the precipitate with water containing hydrogen sulphide. Or mix with ammonia until the free acid is neutralized, then add ammonium sulphide in excess, and allow to digest for some time. The washed precipitate may now be weighed in three different forms, viz., as trisulphide, as metal, or as trioxide. The treatment in the two former cases will be described in 3 and 4 : in the* latter case proceed as follows : Spread the filter out on a glass plate and remove the precipitate to a vessel by means of a jet of water from the wash-bottle or, if this is not practicable, put the precipitate and filter together into the vessel and heat gently with moderately strong nitric acid * Journ.f. prakt. Chem., LXXIV, 344. 120.] BISMUTH. 385 until complete decomposition is effected ; the solution is then diluted with water slightly acidulated with acetic or nitric acid, and filtered, the filter being washed with the acidulated water; the titrate is then finally precipitated as directed in a. 2. Determination of Bismuth as Chromate (J. LOWE*). Pour the solution of bismuth, which must be as neutral as possible, and must, if necessary, be first freed from the excess of nitric acid by evaporation on the water-bath, into a warm solution of pure potassium dichromate in a porcelain dish, with stirring, and take care to leave the alkali chromate slightly in excess. Rinse the vessel which contained the solution of bismuth with water containing nitric acid into the porcelain dish. The precipi- tate formed must be orange-yellow, and dense throughout ; if it is fiocculent, and has the color of the yolk of an egg, this is a sign that there is a deficiency of potassium dichromate ; in which case add a fresh quantity of this salt, taking care, however, to guard against too great an excess, and boil until the precipitate presents the proper appearance. Boil the contents of the dish for ten minutes, with stirring ; then wash the precipitate, first by repeated boiling with water and decantation on to a weighed filter, at last thoroughly on the latter with boiling water ; dry at about 120, and weigh. For the properties and composition of the precipitate, .see 86. Results very satisfactory. 3. Determination of Bismuth as Trisulphide. Precipitate the bismuth as trisulphide according to 1, d. If the precipitate contains free sulphur, extract the latter by boiling with solution of sodium sulphite, or by treatment with carbon disulphide (compare the determination of mercury as sulphide, 118, 3), collect on a weighed filter, dry at 100, and weigh. The drying must be conducted with caution. At first the precipitate loses weight, by the evaporation of w r ater, then it gains weight, from the absorption of oxygen. Hence you should weigh every half hour, and take the lowest weight as the correct one. Compare Expt. No. 52. Properties and composition, 86, y. The bismuth sulphide cannot be conveniently converted into the metallic state by ignition in hydrogen, as its complete decom- position is a work of considerable time. As regards reduction with potassium cyanide, see 4. * Journ.f. prakt. CJiem., LXVII, 464 386 DETERMINATION. [ 120. 4. Determination of Bismuth as Metal. The oxide, sulphide, or basic chloride to be reduced is fused in a porcelain crucible with five times its quantity of ordi- nary potassium cyanide. The crucible must be large enough In the case of oxide and basic chloride, the reduction is completed in a short time at a gentle heat ; sulphide, on the other hand, requires longer fusion and a higher temperature. The operation has been successful if on treatment with water metallic grains are obtained. These grains are first washed completely and rapidly with water, then with weak and lastly with strong alcohol, dried and weighed. If you have been reducing the sulphide, and on treating the fused mass with water a black powder (a mixture of bismuth with bismuth sulphide) is visible, besides the metallic grains,, it is necessary to fuse the former again with potassium cyanide. It sometimes happens that the crucible is attacked, and particles of porcelain are found mixed with the metallic bismuth ; to prevent this from spoiling the analysis, weigh the crucible together with a small dried filter before the experiment, collect the metal on the filter, dry and weigh the crucible with the filter and bismuth again. Results good (H. HOSE*). The precipitation of bismuth as basic chloride, and the reduc- tion of the latter with potassium cyanide, has been recommended by H. RosE.f The process is conducted as follows : Nearly neu- tralize any large excess of acid that may be present with potassa r soda, or ammonia, add ammonium chloride in sufficient quantity (if hydrochloric acid is not already present), and then a rather large quantity of water. After allowing to stand some time, test whether a portion of the clear supernatant fluid is rendered turbid by a. further addition of water ; and then, if required, add water to the whole till the precipitation is complete. Finally filter, wash com- pletely with cold water, dry and fuse according to the directions just given with potassium cyanide. It is less advisable to dry the precipitate at 100, weigh and calculate the metal present from the formula BiOCl, as washing causes a slight alteration in its com- position (unless a little hydrochloric acid is added to the wash- water, which is inconvenient when the precipitate is collected on a weighed filter), and if precipitated in the presence of sulphuric, phosphoric acids, &c., it is liable to contain small quantities of these acids. Results accurate. *Pogg. Annal., xci, 104, and ex, 136. \ lb ex, 425. 121.] CADMIUM. 387 5. Determination of Bismuth as Arsenate. II. SALKOWSKI * recommends the determination of bismuth as a arsenate (BiAsO 4 .H,O) which is dried at 100 to 120, the method being based on the observation by SCHEELE that bismuth arsenate is perfectly insoluble in nitric acid. The solution of bismuthic nitrate is acidulated with nitric acid(but must be free from other acids), precipitated by a moderate excess of arsenic acid, and stirred, avoiding touching the sides of the beaker with the rod (otherwise the crystalline precipitate will adhere fast to the parts touched). The whole is then allowed to stand, without wanning, for a few hours, then the precipitate is collected on a filter previously dried at 120, and washed until the washings begin to pass slightly turbid. Then dry at 120 and weigh. Ignition of the precipitate is not advisable, as the carbon of the filter exercises a reducing action even when ammonium nitrate is used. The test analyses made by SALKOWSKI gave 99*88 to 100-02 instead of 100. 121. 7. CADMIUM. a. Solution. Cadmium, its oxide, and all the other compounds insoluble in water, are dissolved in hydrochloric acid or in nitric acid. 1). Determination. Cadmium is weighed either in the form of oxide, or in that of sulphide ( 87). It may also be weighed as sulphate, and in the absence of other bases precipitable by oxalic acid, it may be esti- mated volumetrically. We may convert into 1. CADMIUM OXIDE: a. By Precipitation. The compounds of cadmium which are soluble in water ; the insoluble compounds, the acid of which is removed upon solution in hydrochloric acid ; cadmium salts of organic acids. 1). By Ignition. Cadmium salts of readily volatile or easily decomposable inorganic oxygen acids. 2. CADMIUM SULPHIDE : All compounds of cadmium without exception. * Journ. f. prakt. Chem., civ, 170; Zeitechr. f. analyt. Chem., vin, 205. 388 DETERMINATION. [ 121. 3. CADMIUM SULPHATE : All compounds of cadmium, in the absence of other non-volatile substances. 1. Determination as Cadmium Oxide. a. By Precipitation. Precipitate with potassium carbonate, wash the precipitated cadmium carbonate, and convert it, by ignition, into oxide. The precipitation is conducted as in the case of zinc, 108, 1, a. The cadmium oxide which adheres to the filter may easily be reduced and volatilized ; it is therefore necessary to be cautious. In- the first place choose a thin filter, transfer the dried precipitate as com- pletely as possible to the crucible, replace the filter in the funnel, and moisten it with ammonium nitrate solution, allow to dry, and then burn carefully in a coil of platinum wire. Let the ash fall into the crucible containing the mass of the precipitate, ignite carefully, avoiding the action of reducing gases, and finally weigh. It is difficult to remove the last portions of carbonic acid ; you must therefore repeat the ignition till the weight remains constant. Properties of precipitate and residue, 87. Results generally a little too low. b. By Ignition. Same process as for zinc, 108, 1, c. 2. Determination as Cadmium Sulphide. It is best to precipitate the moderately acid solution with hydro- gen sulphide water or gas, which must be used in sufficient excess. The presence of a considerable quantity of free hydrochloric or nitric acid may especially if the solution is not enough diluted prevent complete precipitation, hence such an excess should be avoided, and the clear supernatant fluid should in all cases be tested, by the addition of a relatively large amount of hydrogen sulphide water to a portion, before being filtered. Alkaline solutions of cadmium may be precipitated with ammonium sulphide. If the cadmium sulphide is free from admixed sulphur, it may be at once collected on a weighed filter, washed first with diluted hydrogen sulphide water mixed with a little hydrochloric acid, then with pure water, dried at 100, and weighed ; if, on the contrary, it con- tains free sulphur, it may be purified by boiling with a solution of sodium sulphite, or by treatment with carbon disulphide (see Mer- curic Sulphide, 118, 3). Results accurate. The precipitation of sulphur may occasionally be obviated by adding to the cadmium 122.] PALLADIUM. 389 solution potassium cyanide till the precipitate first formed is redis- solved, and then precipitating this solution with hydrogen sulphide. If the cadmium sulphide is not to be weighed as such, warm it, together with the filter, with moderately strong hydrochloric acid, till the precipitate has dissolved and the odor of hydrogen sulphide is no longer perceptible, filter and precipitate the solution as in 1, a, after having removed the excess of free acid for the most part by evaporation. 3. Determination as Cadmium Sulphate. Same process as for magnesium ( 104, 1). The CdSO 4 may be rather strongly ignited without decomposition. 4. W. GIBBS* determines cadmium volumetrically by mixing the concentrated solution of the sulphate, nitrate, or chloride with excess of oxalic acid and a quantity of strong alcohol, filtering, washing with alcohol, dissolving in hot hydrochloric acid and determining the oxalic acid with permanganate ( 13 Y). W. Gr. LEisoNf obtained satisfactory results by this process. Supplement to the Fifth Group. 122. 8. PALLADIUM. Palladium is converted, for the purpose of estimation, into the metallic state or in many separations into potassium palladia chloride. 1. Determination as Palladium. a. Neutralize the solution of palladious chloride almost com- pletely with sodium carbonate, mix with solution of mercuric cyanide ; and heat gently for some time, until the odor of hydro- cyanic acid has gone off. A yellowish-white precipitate of palladi- ous cyanide will subside ; from dilute solutions, only after the lapse of some time. Wash first by decantation, then on the filter, dry thoroughly, ignite cautiously, finally over the gas blowpipe till the palladium paracyanide first formed is decomposed, then ignite in hydrogen, since the palladium has been slightly oxidized As soon as the lamp is removed, stop the hydrogen to prevent absorption, and weigh the metal. If the solution contains palladious nitrate, evaporate it first with hydrochloric acid to dryness; as otherwise * Zeitschr.f. analyt. Chem., vn, 259. f lb., x, 343. 390 DETERMINATION. [ 122. the precipitate obtained deflagrates upon ignition ("WOLLASTOIST). Results exact. 5. Mix the solution of palladious chloride or nitrate with sodium or potassium formate, and warm until no more carbonic acid escapes. The palladium precipitates in brilliant scales (DoBE- KEINEK). . Precipitate the acid solution of palladium with hydrogen sulphide, filter, wash with boiling water, roast, dissolve in hydro' chloric acid and nitric acid, and precipitate as in a. Exposed to a moderate red heat metallic palladium becomes covered with a film varying from violet to blue, but at a higher temperature it recovers its lustre, which it keeps after being sud- denly cooled, for instance, with cold water. This tarnishing and recovery of the metallic lustre is not attended with any percepti- ble difference of weight. Palladium which has taken up oxygen is immediately reduced in hydrogen ; when cooled in the current of gas, it retains some absorbed hydrogen. Palladium requires the very highest degree of heat for its fusion. It dissolves readily in nitrohydrochloric acid, with difficulty in pure nitric acid, more easily in nitric acid containing nitrous acid, with difficulty in boil- ing concentrated sulphuric acid. 2. Determination as Potassium Palladia Chloride. Evaporate the solution of palladic chloride with potassium chloride and nitric acid to dryness, and treat the mass when cold with alcohol of 0'833 sp. gr., in which the double salt is insoluble. Collect on a weighed filter, dry at 100, and weigh. Results a little too low, as traces of the double salt pass away with the alcohol washings (BEKZELIUS). Instead of weighing the double salt you may ignite in hydrogen, remove the potassium chloride with water and weigh the metal obtained. This method is indeed to be pre- ferred, as it prevents any potassium chloride in the precipitate from affecting the result. POTASSIUM PALLADIC CHLORIDE consists of microscopic octa- hedra; it presents the appearance of a vermilion or, if the crystals are somewhat large, of a brown powder. It is very slightly solu- ble in cold water ; it is almost insoluble in cold alcohol of the above strength. It contains 26*89 per cent, palladium. 123.] GOLD. 391 Sixth Group. OOLD PLATESTUM ANTIMONY TIN IN STANNIC COMPOUNDS TIN IN STANNOUS COMPOUNDS ARSENOTJS AND AKSENIC ACIDS ( MO- LYBDIC ACID). 123. 1. GOLD. a. Solution. Metallic gold, and all compounds of gold insoluble in water, are warmed with hydrochloric acid, and nitric acid is gradually added until complete solution is effected ; or they are repeatedly digested with strong chlorine water. The latter method is resorted to more especially in cases where the quantity of gold to be dis- solved is small, and mixed with foreign oxides which it is wished to leave undissolved. According to "W". SKEY* tincture of iodine, or, for larger quantities of gold, bromine water, is better than chlo- rine water. They give solutions freer from other metals than the chlorine water gives. b. Determination. Gold is always weighed in the metallic state. The compounds are brought into this form, either by ignition or by precipitation, as gold, or auric sulphide. We convert into METALLIC GOLD : a. By Ignition. All compounds of gold which contain no fixed acid, or other body. 1). By Precipitation as metallic gold. All compounds of gold without exception in cases where a is inapplicable. c. By Precipitation as auric sulphide. This method serves to effect the separation of gold from certain other metals which may be mixed with it in a solution. Determination as Metallic Gold, a. By Ignition. Heat the compound, in a covered porcelain crucible, very gently at first, but finally to redness, and weigh the residuary pure gold. For properties of the residue, see 88. The results are most accurate. *Z*itschr.f. analyt. Chem., x, 221. 392 DETERMINATION. [ I. By Precipitation as Metallic Gold. a. The solution is free from Nitric Acid. Mix the solution with a little hydrochloric acid, if it does not already contain some of that acid in the free state, and add a clear solution of ferrous sulphate in excess; heat gently for a few hours until the precipi- tated fine gold powder has completely subsided ; filter, wash, dry, and ignite according to 52. A porcelain dish is a more appro- priate vessel to effect the precipitation in than a beaker, as the heavy fine gold powder is more readily rinsed out of the former than out of the latter. There are no sources of error inherent in the method. fi. The solution of Gold contains Nitric Acid. Evaporate the solution, on a water-bath, to the consistence of syrup, adding from time to time hydrochloric acid ; dissolve the residue in water con- taining hydrochloric acid, and treat the solution as directed in a* It will sometimes happen that the residue does not dissolve to a. clear fluid, in consequence of a partial decomposition of auric chlo- ride into aurous chloride and metallic gold ; however, this is a mat- ter of perfect indifference. y. In cases where it is wished to avoid the presence of iron in the filtrate, the gold may be reduced by means of oxalic acid. To this end, the dilute solution freed previously, if necessary, from nitric acid, in the manner directed in /? is mixed, in a beaker, with oxalic acid, or with ammonium oxalate in excess, some sul- phuric acid added (if that acid is not already present in the free state), and the vessel, covered with a glass plate, is kept standing for two days in a moderately warm place. At the end of that time, the whole of the gold will be found to have separated in small yellow scales, which are collected on a filter, washed first with dilute hydrochloric acid, then with water, dried, and ignited. If the gold solution contains a large excess of hydrochloric acid, the latter should be for the most part evaporated, before the solu- tion is diluted and the oxalic acid added. If the gold solution con- tains chlorides of alkali metals, it is necessary to dilute largely, and allow to stand for a long time, in order to effect complete precipi- tation (H. ROSE). d. The gold may also be thrown down in the metallic form by chloral hydrate * in the presence of potassa. Warm the solution, * HAGEK'S Pharmac. Centralhalle, xi, 393. 124.] PLATINUM. 393 add the chloral, then pure potassa in excess, and boil for a minute or so. The gold is precipitated with evolution of chloroform. . Finally, gold may be thrown down by many metals, such as zinc, cadmium, magnesium, &c. The latter has been recommended by SCHEIBLEB* for the analysis of the gold salts of organic bases. The precipitate is first washed with hydrochloric acid, then with water. c. By Precipitation as Auric Sulphide. . Hydrogen sulphide gas is transmitted in excess through the dilute solution containing some free acid ; the precipitate formed is speedily filtered off, without heating, washed, dried, and ignited in a porcelain crucible. For the properties of the precipitate, see 88. 'No sources of error. 2. PLATINUM. a. Solution. Metallic platinum, and the compounds of platinum which are insoluble in water, are dissolved by digestion, at a gentle heat, with nitrohydrochloric acid. b. Determination. Platinum is invariably weighed in the metallic state, to which condition its compounds are brought, either by precipitation as ammonium platinic chloride, potassium platinic chloride, or pla- tiiric sulphide, or by ignition, or by precipitation with reducing agents. All compounds of platinum, without exception, may, in. most cases, be converted into platinum by either of these methods. "Which is the most advantageous process to be pursued in special instances, depends entirely upon the circumstances. The reduc- tion to the metallic state, by simple ignition is preferable to the other methods, in all cases where admissible. The precipitation as platinic sulphide is resorted to exclusively to effect the separation of platinum from other metals. Determination as Metallic Platinum. a. By Precipitation as Ammonium Platinic Chloride. The solution must be concentrated if necessary by evaporation *Ber. der dcutsch. chem. Gesellsch., 1869, 295. 394 DETERMINATION. [ 124. on a water-bath. Mix in a beaker with ammonia until the excess of acid (that is, supposing an excess of acid to be present) is nearly saturated ; add ammonium chloride in excess and mix the fluid with a pretty large quantity of strong alcohol. Cover the beaker now with a glass plate and let it stand for twenty-four hours, after which collect in a weighed asbestos filter, or in an unweighed paper filter, wash the precipitate with alcohol of about 80 per cent., till the substances to be separated are Removed, dry carefully, ignite according to 99, 2, and weigh. In the case of large quan- tities the final ignition is advantageously conducted in a stream of hydrogen ( 108, Fig. 83), in order to be quite sure of effecting complete decomposition. For the properties of the precipitate and residue, see 89. The results are satisfactory, though generally a little too low, as the ammonium platinic chloride is not altogether insoluble in alcohol of the above strength (Expt. No. 16), and as the fumes of ammonium chloride are liable to carry away traces of the yet uiidecomposed double chloride, if the application of heat is not conducted with the greatest care. If the precipitated ammonium platinic chloride were weighed in that form, the results would be inaccurate, since, as I have con- vinced myself by direct experiments, it is impossible to completely free the double chloride, by washing with alcohol, from all traces of the ammonium chloride thrown down with it, without dissolving at the same time a notable portion of the double chloride. As a general rule, the results obtained by weighing the ammonium pla- tinic chloride in that form are one or two per cent, too high. ~b. By Precipitation as Potassium Platinic Chloride. Mix the solution, in a beaker, with potassa, until the greater part of the excess of acid (if there be any) is neutralized; add potassium chloride slightly in excess, and finally a pretty large quantity of strong alcohol ; should your solution of platinum be very dilute, you must concentrate it previously to the addition of the alcohol. After twenty- four hours collect the precipitate weighed asbestos filter, wash with alcohol of 80 per cent. , dry thoroughly at 100, convert into platinum according to 97, 4, a, and weigh. For the properties of the precipitate and residue, see 89. The results are more accurate than those obtained by method #, since, on the one hand, the potassium platinic chloride is more 125.] ANTIMONY. 395 insoluble in alcohol than the corresponding ammonium salt ; and, on the other hand, loss of substance is less likely to occur during ignition. To weigh the potassium platinic chloride in that form would not be practicable, as it is impossible to remove, by washing with alcohol, all traces of the potassium chloride thrown down with it, without, at the same time, dissolving a portion of the double chloride. c. By Precipitation as Plcutinic Sulphide. Precipitate the solution with hydrogen- sulphide water or gas, according to circumstances, heat the mixture to incipient ebulli- tion, filter, wash the precipitate, dry, and ignite according to 52. For the properties of the precipitate and residue, see 89. The results are accurate. d. By Ignition. Same process as for gold, 123. For the properties of the residue, see 89. The results are most accurate. e. By Precipitation with' Reducing Agents. Various reducing agents may be employed to precipitate plati- num from its solutions in the metallic state. The reduction is very promptly effected by ferrous sulphate and potassa or soda (the protosesquioxide of iron being removed by subsequent addi- tion of hydrochloric acid, HEMPEL), or by pure zinc or magnesium (the excess of which is removed by hydrochloric acid) ; somewhat more slowly, and only with application of heat, by alkali formates. Mercurous nitrate also precipitates the whole of the platinum from solution of platinic chloride ; upon igniting the brown precipitate obtained, fumes of inercurous chloride escape, and metallic plati- num remains. 125. 3. ANTIMONY. a. Solution. Antirnonous oxide, and the compounds of antimony which are insoluble in water, or are decomposed by that agent, are dissolved in more or less concentrated hydrochloric acid. Metallic antimony is dissolved best in nitrohydrochloric acid. The ebullition of a hydrochloric acid solution of antimonous chloride is attended with volatilization of traces of the latter ; the concentration of a solution 396 DETEEMINATIONo [ 125. of the kind by evaporation involves accordingly loss of substance. Solutions so highly dilute as to necessitate a recourse to evapora- tion must therefore previously be supersaturated with potassa. Solutions of antimonous chloride, which it is intended to dilute with water, nrast previously be mixed with tartaric acid, to prevent the separation of basic salt. In diluting an acid solution of anti- monic acid in hydrochloric acid, the water must not be added gradually and in small quantities at a time, which would make the fluid turbid, but in sufficient quantity at once, which will leave the fluid clear. 1). Determination. Antimony may be weighed as antimonous- sulphide or anti- mony tetroxide: in separations it is sometimes weighed as metallic antimony j or it is estimated volumetrically. Antimony oxides and the salts, with readily volatile or decomposable oxygen acids, may be converted into antimony tetroxide by simple ignition. Antimony in solution is almost invariably first precipitated as sulphide, which is then, with the view of estimation, converted into anhydrous sulphide or determined volumetrically. Of the volumetric methods the first two are applicable only when the antimony is present as a pure tetroxide or as anti- monous chloride. 1. Precipitation as Antimonous Sulphide. Add to the antimony solution hydrochloric acid, if not already present, then tartaric acid, and dilute with water, if necessary. Introduce the clear fluid into a flask, closed with a doubly perfo- rated cork; through one of the perforations passes a tube, bent outside at a right angle, which nearly extends to the bottom of the flask ; through the other perforation passes another tube, bent out- side twice at right angles, which reaches only a short way into the flask ; the outer end of this tube dips slightly under water. Con- duct through the first tube hydrogen sulphide gas, until it pre- dominates strongly ; put the flask in a moderately warm place, and after some time conduct carbon dioxide into the fluid, until the excess of the other gas is almost completely removed. If there is no reason against it, from the presence of a large quantity of hydrochloric acid, or from the presence of nitric acid, it is well to heat the solution during the passing of the gas, finally even boiling. 125.] ANTIMONY. 397 The precipitate is then denser, and may be very easily washed. (SllARPLES*). If the amount of the precipitate is at all considerable, filter without intermission through a weighed filter, wash rapidly and thoroughly with water mixed with a few drops of hydrogen sul- phide water, dry at 100, and weigh. The precipitate so weighed .always retains some water, and may, besides, contain free sulphur; in fact, it always contains the latter in cases where the antimony solution, besides antimonous salts, contains antimonic acid or antimony pentachloride, since the precipitation under these cir- cumstances is preceded by a reduction of antimonic to antimo- nous compounds, accompanied by separation of sulphur (H. ROSE). A further examination of the precipitate is accordingly indis- pensable. To this end treat a sample of the weighed precipitate with strong hydrochloric acid. If a. The sample dissolves to a clear fluid, this is a proof that the precipitate only contains Sb,S 3 ; but if I). Sulphur separates, this shows that free sulphur is present. In case a (in order to remove the water retained at 100) the greater portion of the dried precipitate is weighed in a porcelain boat, which is then inserted into a glass tube, about 2 decimetres long ; a slow current of dry carbon dioxide is transmitted through the latter, and the boat cautiously heated by means of a lamp, moved to and fro under it, until the orange precipitate becomes black. The precipitate is then allowed to cool in the current of carbon dioxide, and weighed ; from the amount found, the total quantity of anhydrous antimonous sulphide contained in the entire precipitate is ascertained by a simple calculation. The results are accurate. Expt. No. 75 gave 99 -24 instead of 100. But if thu precipitate is simply dried at 100, the results are about 2 per cent, too high see the same experiment. For the properties of the precipitate, see 90. In case ~b, the precipitate is subjected to the same treatment as in a, with this difference only, that the contents of the boat are heated much more intensely, and the process is continued until no more sulphur is expelled. This removes the whole of the admixed sulphur ; the residue consists of pure antimonous sulphide. It must be completely soluble in fuming hydrochloric acid on heating. * Zeilschr. /. analyt. Chem., x, 343. 398 DETERMINATION. [ 125, If the amount of the precipitate is small, collect it in a weighed asbestos filtering tube, dry in a slow current of carbon dioxide at a gentle heat, heat finally rather more strongly till the sulphide has turned black and any free sulphur present has volatilized, allow to cool, replace the gas in the tube by air, and weigh. Results quite satisfactory. 45 ' BUNSEN recommends converting the antimonous sulphide into antimony tetroxide. For the method of estimating the antimony in the sulphide volumetrically and indirectly, see 3, c. 2. Determination as Antimony Tetroxide. a. In the case of antimonous oxide or a compound of the same with an easy volatile or decomposable oxygen acid, evaporate carefully with nitric acid and ignite finally for some time till the weight is constant. The experiment may be safely made in a platinum crucible. With antimonic acid, the evaporation with nitric acid is unnecessary. 5. If antimonous sulphide is to be converted into anntimony tetroxide, one of the two following methods given by BUNSEN f is employed : a. Moisten the dry antimony sulphide with a few drops of nitric acid of 1*42 sp. gr., then treat, in a weighed porcelain crucible with concave lid, with 8-10 times the quantity of fuming nitric acid,J and let the acid gradually evaporate on the water-bath. The sulphur separates at first as a fine powder, which, however, is readily and completely oxidized during the process of evaporation. The white residual mass in the crucible consists of antimonic acid and sulphuric acid, and may by ignition be con- verted without loss -into antimony tetroxide. If the antimony sulphide contains a large excess of free sulphur, this must be removed by washing with carbon disulphide. /?. Mix the antimony sulphide with 30-50 times its quantity *Zeitsckr.f. analyt. Chem., vm, 155. f Annal. d. Chem. u. Pharm., cvi, 3. \ Nitric acid of l'42sp. gr. is not suitable for this purpose, as its boiling-point is almost 10 above the fusing-point of sulphur, whereas fuming nitric acid boils at 86, consequently below the fusing-point of sulphur. With nitric acid of 1 -42 sp. gr., therefore, the separated sulphur fuses and forms drops, which obstinately resist oxidation. 125.] ANTIMONY. 399 of pure mercuric oxide,* and heat the mixture gradually in an open porcelain crucible. As soon as oxidation begins, which may be known by the sudden evolution of gray mercurial fumes, moderate the heat. When the evolution of mercurial fumes diminishes raise the temperature again, always taking care, how- ever, that no reducing gases come in contact with the contents of the crucible. Remove the last traces of mercuric oxide over the blast gas-lamp, then weigh the residual fine white powder of anti- mony tetroxide. As mercuric oxide generally leaves a trifling fixed residue upon ignition, the amount of this should be determined once for all, the mercuric oxide added approximately weighed, and the corresponding amount of fixed residue deducted from the antimony tetroxide. The volatilization of the oxide of mercury proceeds much more rapidly when effected in a platinum crucible instead of a porcelain one. But, if a platinum crucible is employed, it must be effectively protected from the action of antimony upon it, by a good lining of mercuric oxide, f If the antimony sulphide contains free sulphur, this must first be removed by washing with carbon disulphide before the oxidation can be proceeded with, since otherwise a slight deflagration is avoidable. According to later experiments made by B ONSEN, \ it is some- what difficult to obtain good results by this method, because a temperature a little above that required to reduce Sb,O 6 to Sb,O 4 will reduce the latter Sb a O 8 . Ignition over a blast-lamp in a very large covered platinum or rather large open porcelain crucible, keeping only the bottom at a full red heat, is recommended as a * Prepared by precipitation from mercuric chloride by excess of soda solution and thorough washing. f This is effected best, according to BUNSEN, in the following way: Soften the sealed end of a common test-tube before the glass-blower's lamp ; place the softened end in the centre of the platinum crucible, and blow into it, which will cause it to expand and assume the exact form of the interior of the crucible. Crack off the bottom of the little flask so formed, and smooth the sharp edges cautiously by fusion. A glass is thus obtained, open at both ends, which exactly fits the crucible. To effect the lining by means of this instrument, fill the crucible loosely with mercuric oxide up to the brim, then force the glass gradually and slowly down to the bottom of the crucible, occasionally shaking out the oxide of mercury from the interior of the glass. The inside of the crucible is thus covered with a layer of oxide of mercury \ 1 line thick, which, after the removal of the glass, adheres with sufficient firmness, even upon ignition. \ Zeitschr. f. anal. Chem., xvm, 268. 400 DETERMINATION. [ 125. method bj which it is possible to drive off just one atom O from 6b.O.. 3. Volumetric Methods. a. Oxidation of Antimonous Oxide to Antimonic Oxide by Iodine (Fn. MOHR *). The oxidation is effected in alkaline solution and proceeds as follows : Sb 2 O 3 + 41 + 4NaOH = Sb 9 O B + 4NaI + 2H 3 O. The method gives results which are serviceable only under very defi- nite conditions, because the antimonous oxide has not always an equal tendency to change to antimonic oxide in alkaline solution, but this tendency is greater in the presence of much alkali car- bonate than when little is present, and becomes constant only with a certain excess of carbonate. According to my investigations it is best to proceed thus : Dissolve a quantity of the compound containing about O'l grm. of antimonous oxide in about 10 c. c. of an aqueous solution of tartaric acid, then add sufficient sodium-carbonate solution to about neutralize the liquid. Add now 20 c. c. of a cold, saturated solution of sodium bicarbonate and (the liquid remaining clear) some starch paste, then titrate with iodine ( 146) until the fluid just remains blue on being stirred. The fact that the blue color soon disappears, however, must not induce the operator to add more iodine. 4 eq. of iodine corresponds to 1 eq. of antimonous oxide. The results so obtained are entirely satisfactory (Expt. No. 76). I cannot recommend the use of sodium carbonate as employed by FR. MOHR in his experiments, because it itself has the prop- erty of fixing a considerable quantity of iodine, varying, moreover, according to the quantity of water used (Expt. No. 77), whereas this is not the case with sodium bicarbonate (Expt. No. 78). Compare also 127, 5, #, 1, and Expt. No. 79. a. Conversion of Antimonous Chloride to Antimonic Chlo- ride by Hydrochloric Acid and Potassium Chromate or Perman- ganate. F. KES&LER'sf first description of this method was so wanting in precision, that it could not be depended upon. However, he has sincej determined most accurately the conditions under which antimony in acid solution may be satisfactorily titrated either with potassium chromate (the excess of the standard solution being determined with ferrous sulphate) or with potassium permanganate. * Lehrbuch der Titrir method e, 3. Aufl., 276. \Pogg AnnaL, xcv, 204. %lb., cxvin, 17, and Zeitschr f. analyt. Chem., n 3*3. 124.] ANTIMONY. 401 I. Titration with Potassium Dichr ornate. 1. REQUISITES. a. Standard Solution of Arsenous Acid. Dissolve exactly 5 grm. pure arsenous oxide by the aid of some soda solution, add hydrochloric acid till slightly acid, then 100 c.c. more of hydro- chloric acid of 1-12 sp. gr., and dilute to 1000 c.c. Each c.c. con- tains 0-005 grm. arsenous oxide and corresponds to 0*007293 antimonous oxide. ft. Solution of Potassium Dichromate. Dissolve about 2*5 grm. to 1 litre. y. Solution of Ferrous Sulphate. Dissolve about 1*1 grm. iron wire in 20 c.c. dilute sulphuric acid (1 to 4), filter, and dilute to 1 litre. #. Solution of Potassium Ferricyanide. Should be tolerably dilute and freshly prepared. 2. DETERMINATION OF THE SOLUTIONS. a. .Relation between the Solution of Chr ornate and the Solution of Ferrous Sulphate. Run into a beaker 10 c.c. of the chromate solution from the burette, add 5 c.c. of hydrochloric acid and 50 c.c. water, and then add iron solution from a burette till the fluid is green. Continue adding the iron solution, a c.c. at a time, test- ting after each addition whether a drop of the fluid, when brought in contact with a drop of the potassium f erricyanide, on a porcelain plate, manifests a distinct reaction for ferrous iron. As soon as this point is attained, add '5 c.c. of chromate solution and then iron solution two drops at a time, till the blue reaction just occurs.- Now read off both burettes, and calculate how much chromate solution corresponds to 10 c.c. of iron solution. This experiment is to be repeated before every fresh series of analyses, as the iron solution gradually oxidizes. ft. Delation between the Chromate Solution and the Solution of Arsenous Acid. Transfer 10 c.c. of the arsenous solution to a beaker, add 20 c.c. hydrochloric acid of 1/2 sp. gr., and 80 100 c.c.* water, run in chromate solution till the yellow color of the tin id shows an excess, wait a few minutes, add excess of iron solu- tion, then again 0-5 chromate solution, and finally again iron solu- tion till the end-reaction appears (see above). Deduct from the * The water must be measured, for the action of chromic acid on arsenous acid (and also on antimonous chloride) is normal only if the fluid contains at least one sixth of its volume of hydrochloric acid of 1'12 sp. gr. 402 DETERMINATION. [ 125. total quantity of chromate solution employed, the amount corre- sponding to the iron used, and from the datum thus afforded calcu- late how much antimony corresponds to 100 c.c. of chromate solu- tion ; in other words, how much antimony is converted by the quantity of chromate mentioned from SbCl 3 into SbCl 5 . 3. THE ACTUAL ANALYSIS. In the absence of organic matter, heavy metallic oxides, and other bodies which are detrimental to the reaction, dissolve the antimo- nous compound at once in hydrochloric acid. The solution should contain not less than -J- of its volume of hydrochloric acid of T12< sp. gr. It is not advisable, on the other hand, that it should con- tain more than ^, otherwise the end-reaction with potassium fern- cyanide is slower in making its appearance and loses its nicety. Tartaric acid cannot be employed as a solvent, since it interferes with the action of chromic acid on ferrous salts. Now proceed as directed in 2. If the direct determination of antimony in the hydrochloric acid solution is not practicable, precipitate it with hydrogen sulphide. Wash the precipitate, transfer it, together with the filter, to a small flask ; treat it with a sufficiency of hydro- chloric acid, dissolve by digestion on the water-bath, add a suffi- cient quantity of a nearly saturated solution of mercuric chloride- in hydrochloric acid of 1*12 sp. gr. to remove the hydrogen sul- phide, and then proceed as directed. II. Titration with Potassium Permanganate. Here also the fluid must contain at least -J- of its volume of hydrochloric acid of 1.12 sp. gr. The permanganate solution,., which may contain about 1*5 gnu. of the crystallized salt in a litre, is added to permanent reddening. The end-reaction is exact, and the conversion of antiinonous to antimonic chloride goes on uni- formly, although the degree of dilution may vary, provided the above relation between hydrochloric acid and water is kept up. It is not well that the hydrochloric acid should exceed -J of the volume of the fluid, as in that case the end-reaction would be too transitory. Tartaric acid, at least in the proportion to antimony in which it exists in tartar emetic, does not interfere with the reac- tion. Hence the permanganate may be standardized by the aid of solution of tartar emetic of known strength. If you have to analyze antimonous sulphide, proceed as directed I. 3 ; make the fluid mixed with mercuric chloride up to a certain 126.] TIN IN STANNOUS AND STANNIC COMPOUNDS. 403 volume, allow to settle, and use a measured portion of the perfectly c!< ,ir solution for the experiment. My own experiments* have shown that KESSLER'S methods are also suitable for the estimation of very small quantities of anti- mony. o. Volumetric Estimation ~by determining the Hydrogen Sul- phide yielded by the Sulphide (R. SCHNEIDER f). Both antimonous and antimonic sulphides yield under the fcctiou of boiling hydrochloric acid 3 mol. hydrogen sulphide for every 2 atoms of antimony. Hence, if the amount of the gas evolved under such circumstances is estimated, the amount of anti- mony is known. For decomposing the sulphide and absorbing the gas, the same apparatus serves as BUNSEN employs for his iodimetric analyses ( 130). The size of the boiling-flask should depend on the quan- tity of sulphide ; for quantities up to 0*4 grm. Sb a S 3 , a flask of 100 c.c. is large enough; for 0'4 1 grm., use a 200 c.c. flask. The body of the flask should be spherical, the neck rather narrow, long, and cylindrical. If the sulphide of antimony is on a filter, put both together into the flask. The hydrochloric acid should not be too concentrated. The determination of the hydrogen sulphide is best conducted according to the method given in 148, h. The- results obtained by SCHNEIDER are satisfactory. If the precipitate contains anti- monious chloride, the results are of course false, and this would actually be the case if on precipitation with hydrogen sulphide the addition of the tartaric acid were omitted. 126. 4. TIN IN STANNOUS COMPOUNDS, and 5. TIN IN STANNIC COMPOUNDS. a. Solution. In dissolving compounds of tin soluble in water, a little hydro- chloric acid is added to insure a clear solution. Nearly all the compounds of tin insoluble in water dissolve in hydrochloric acid, or in aqua regia. The hydrate of metastannic acid may be dissolved by boiling with hydrochloric acid, decanting the fluid, and treating the residue with a large proportion of water. Ignited stannic oxide, * ZeitscJir. f. analyt. Chem., vm, 155. f Pogg. Annal., ex, 634. 404 DETERMINATION. [ 12 and stannic compounds insoluble in acids, are prepared for solution in hydrochloric acid, by reducing them to the state of a fine pow- der, and fusing in a silver crucible with potassium or sodium hydroxide, in excess. Metallic tin is dissolved best in aqua regia ; the solution frequently contains metastannic chloride mixed with the stannic chloride (Tn. SCHEERER*). It is generally determined, liowever, by converting it into stannic oxide, without previous solution. Acid solutions of stannic salts, which contain hydrochlo- ric acid, or a chloride, cannot be concentrated by evaporation, not even after addition of nitric acid or sulphuric acid, without volatili- zation of stannic chloride taking place. b. Determination. Tin is weighed in the form of stannic oxide, into which it is converted, either by the agency of nitric acid, or by precipitation as stannic (or metastannic) acid, or by precipitation as sulphide. A great many volumetric methods of estimating tin have been pro- posed. They all depend on obtaining the tin in solution in the condition of stannous chloride, and converting this into stannic chloride either in alkaline or acid solution. A few only yield satis- factory results. We may convert into STANNIC OXIDE: a. By Treatment with Nitric Acid. Metallic tin, and those compounds of tin which contain no fixed acid, provided no com- pounds of chlorine be present. b. By Precipitation as Stannic (or Metastannic] Acid. All tin salts of volatile acids, provided no non- volatile organic substances nor ferric salts be present. c. By Precipitation as Sulphide. All compounds of tin with- out exception. In methods a and c, it is quite indifferent whether the tin is present as a stannous or a stannic compound. The method b requires the tin to be present as a stannic salt. The volumetric methods may be employed in all cases ; but the estimation is simple and direct only where the tin is in solution as stannous chloride and free from other oxidizable bodies, or can readily be brought Into this state. For the methods of determining stannous and stannic tin in presence of each other, I refer to Section Y. *Joum.f. prakt. Chem. N. F., in, 472. 126.] TIN IN STANNOUS AND STANNIC COMPOUNDS. 405 1. Determination of Tin as Stannic Oxide. a. By Treatment with Nitric Acid. This method is resorted to principally to convert the metallic tin into stannic oxide. For this purpose the finely-divided metal is put into a capacious flask, and moderately concentrated pure nitric acid (about 1'3 sp. gr.) gradually poured over it ; the flask is overed with a watch glass. When the first tumultuous action of the acid has somewhat abated, a gentle heat is applied until the metastannic acid formed appears of a pure white color, and further action of the acid is no longer perceptible. The contents of the flask are then transferred to a porcelain dish and evaporated on a water-bath nearly to dry ness, water is then added, and the precipi- tate is collected on a filter, washed, till the washings scarcely red- den litmus paper, dried, ignited, and weighed. The ignition is effected best in a small porcelain crucible, according to 53 ; still a platinum crucible may also be used. A simple red heat is not sufficient to drive off all the water ; the ignition must therefore be finished over a gas blowpipe. Compounds of tin which contain no fixed substances may be converted into stannic oxide by treating them in a porcelain crucible with nitric acid, evaporating to dry- ness, and igniting the residue. If sulphuric acid be present, the expulsion of that acid may be promoted, in the last stages of the process, by ammonium carbonate, as in the case of acid potassium sulphate ( 97) ; here also the heat must be increased as much as possible at the end. For the properties of the residue, see 91. There are no inherent sources of error. l>. By Precipitation as Stannic (or Metastannic) Acid. The application of this method presupposes the whole of the tin to be present in the state of stannic salts. Therefore, if a solu- tion contains stannous salts, either mix with chlorine water, or con- duct chlorine gas into it, or heat gently with chlorate of potassa, until the conversion of the stannous into stannic salts is effected. When this has been done, add ammonia until a permanent precipitate just begins to form, and then hydrochloric acid, drop by drop, until this precipitate is completely redissolved ; by this means a large excess of hydrochloric acid in the solution will be avoided. Add to the fluid so prepared a concentrated solution of ammonium nitrate (or sodium sulphate), and apply heat for some time, where- upon the whole of the tin will precipitate as stannic acid. Decant three times on to a filter, then collect the precipitate on the latter, 406 DETERMINATION. [ 126. wash thoroughly, dry, and ignite. To make quite sure that the whole of the tin has separated, you need simply, before proceeding to filter, add a few drops of the clear supernatant fluid to a hot solution of ammonium nitrate, or sodium sulphate, when the for- mation or non-formation of a precipitate will at once decide the question. The tin is also precipitated from metastannic chloride by the above reagents. This method, which we owe to J. LOWENTHAL, has been repeat- edly tested by him in iny own laboratory,* is easy and convenient, and gives very accurate results. The decomposition is expressed by the equation, SnCl 4 + 4Na 2 SO 4 + 3H 2 O = H 2 SnO 3 + 4NaCl -f- 4NaHSO 4 , or in precipitating with ammonium nitrate: SnCl 4 + 4NH 4 NO 3 + 3H 2 O = H 2 SnO 3 + 4KB 4 C1 + 4HNO 3 . Tin may also, according to H. RosE,f be completely precipi- tated from stannic solutions by sulphuric acid. If the solution contains metastannic acid or metastannic chloride, the precipitation is effected without extraordinary dilution ; the other stannic com- pounds, however, require very considerable dilution. If free hydrochloric acid is absent, the precipitation is rapid ; in other cases 12 or 24 hours at least are required for perfect precipitation. Allow to settle thoroughly, before filtering, wash well (if hydro- chloric acid was present, till the washings give no turbidity with silver nitrate), dry and ignite, at last intensely with addition of some ammonium carbonate. The results obtained by OESTEN, and communicated by H. ROSE, are exact. c. By Precipitation as Stannous or Stannic Sulphide. Precipitate the dilute moderately acid solution with hydrogen sulphide water or gas. If the tin was present in the solution as a stannous salt, and the precipitate consists accordingly of the brown stannous sulphide, keep the solution, supersaturated with hydrogen sulphide, standing for half an hour in a moderately warm place, and then filter. If, on the other hand, the solution contain a stan- nic salt, or metastannic acid, and the precipitate is yellow and consists of stannic sulphide mixed with stannic oxide, or yellowish brown and consists of hydrated metastannic sulphide mixed with meta- stannic acid (BARFOED, TH. SCHEERER :{:), put the fluid, loosely covered, in a warm place, until the odor of hydrogen sulphide * Journ. f. prakt. Chem. t LVI, 366. \Pogg. Annal., cxn, 164. J Journ. f. prakt. Chem. N. F., in. 472. 126.] TIN IN STANNOUS AND STANNIC COMPOUNDS. 407 Las nearly gone off, and then filter. The washing of the stan- nic-sulphide precipitate, which has a great inclination to pass through the filter, is best effected with a concentrated solution of sodium chloride, the remains of the latter being got rid of by a solution of ammonium acetate containing a small excess of acetic acid. If there is no objection to having the latter salt in the fil- trate, the washing maybe entirely effected by its means (BUNSEN*). Transfer the dry precipitate as completely as possible to a watch glass, burn the filter carefully in a weighed porcelain crucible, moisten the ash with nitric acid, ignite, allow to cool, add the pre- cipitate, cover the crucible, heat gently for some time (slight decrep- itation often occurs), remove the lid and heat gently with access of air, till sulphur dioxide has almost ceased to be formed. (If too much heat is applied at first, stannic sulphide volatilizes, the fumes of which give stannic oxide.) Now heat strongly, allow to cool, and heat repeatedly with pieces of ammonium carbonate to a high degree, to drive out the last portions of sulphuric acid. When the weight remains constant the experiment is ended (H. HOSE). For the properties of the precipitates, see 91. The results are accu- rate. 2. Volumetric Methods. The determination of tin by the conversion of stannous into stannic chloride with the aid of oxidizing agents (potassium dichro- mate, iodine, potassium permanganate, etc. )offers peculiar difficulties, inasmuch as on the one hand the stannous chloride takes up oxygen from the air and from the water used for dilution, with more or less rapidity, according to circumstances ; and on the other hand, the energy of the oxidizing agent is not always the same, being influenced by the state of dilution and the presence of a larger or .smaller excess of acid. In the following methods, these sources of error are avoided or limited in such a manner as to render the results satisfactory. * Annal. d. CJiem. u. Pharm., cvi, 13. 408 DETERMINATION. [ 126. 1. Determination of Stannous Chloride ly Iodine in Alkaline Solution (after LENSSEN *). Dissolve the stannous salt or the metallic tin f in hydrochloric acid (preferably in a stream of carbon dioxide), add sodium-potas- sium tartrate, then sodium bicarbonate in excess. To the clear slightly alkaline solution thus formed add some starch-solution, and afterwards the iodine solution of 146, till a permanent blue coloration appears. 2 at. free iodine used corresponds to 1 at. tin. LENSSEN'S results are entirely satisfactory. .2. Determination of Stannous Chloride after addition of Ferric Chloride. The fact that stannous chloride in acid solution can be far more accurately converted into stannic by oxidizing agents after being mixed with ferric chloride (or even with cupric chloride) than without this addition, was first settled by LOWENTHAL. \ Sub- sequently STROMEYER published some experiments leading to the same results, together with practical remarks on the best way of carrying out the method in different cases. The processes thus originated, and which have been well tested, are as follows : a. The given substance is a stannous salt. Dissolve in pure ferric chloride (free from ferrous chloride) with addition of hydro- chloric acid, dilute and add standard permanganate from the burette. Now make another experiment with, the same quantity of water similarly colored with ferric chloride to ascertain how much permanganate is required to tinge the liquid, and subtract the quantity so used from the amount employed in the actual analysis, and from the remainder calculate the tin. The reaction between the tin salt and the iron solution is SnCl, -(-Fe 2 Cl 6 =SnCl 4 +2FeCl 2 . The solution thus contains ferrous, chloride in the place of stannous salt, the former being, as is well known, far less susceptible of alteration from the action of free oxygen than the latter. at. iron found correspond to 1 at. tin. * Journ. f.prakt. Chem., LXXVIII, 200; Annal. d. Chem. u. Pharm., cxiv, 113. f The solution of metallic tin is much assisted by the presence of platinum foil, which is accordingly added. LENSSEN found this addition of platimirn ta be objectionable ; but no other experimenter has observed that it interferes, with the accuracy of the results. \ Journ. f. prakt- Chem., LXXVI, 484. Annal d. Chem. u. Pharm., cxvii, 261. 127.] ARSENOUS AND ARSENIC ACIDS. 409 It must not be forgotten that the titration takes place in presence of hydrochloric acid, and that hence the inconveniences mentioned under 112, 2, y y may arise and impair the accuracy of the method. The results cannot be considered accurate unless the standardizing of the permanganate and the analysis take place under similar conditions as regards dilution and amount of hydro- chloric acid. I. The given substance is metallic tin. Either dissolve in hydrochloric acid preferably with addition of platinum and in an atmosphere of carbon dioxide and treat the solution according to a, or place the substance at once in a concentrated solution of ferric chloride mixed with a little hydrochloric acid ; under these cir- cumstances it will, if finely divided, dissolve quickly even in the cold and without evolution of hydrogen. Gentle warming is unobjectionable. Now add the permanganate. The reaction is Sn + 2Fe 2 Cl 6 =SnCl 4 + 4FeCl 2 , therefore every 4 at. iron found reduced correspond to 1 at. tin. The results are of course only correct when iron is not present. Where this is the case, proceed with the impure tin solution according to c. c. The given substance is stannic chloride or stannic oxide, or a compound of tin containing iron. Dissolve in water with addition of hydrochloric acid, place a plate of zinc in the solution and allow to stand twelve hours, then remove the precipitated tin with a brush, wash it, dissolve in ferric chloride, and proceed as in b. d. The given substance is pure stannic sulphide, precipitated out of an acid stannic solution containing no stannous salt. Mix with ferric chloride, heat gently, filter off the sulphur, and then add the permanganate. 4 at. iron correspond to 1 at. tin, for SnS, + 2Fe 3 Cl 8 = SnCl 4 + 4FeCl, + 2S. The results obtained by STROMEYER are quite satisfactory. As regards the precipitated stannic sulphide, see BARFOED, 91, c. 12T. 6. ARSENOUS ACID, and 7. ARSENIC ACID. a. Solution. The compounds of arsenous and arsenic acids which are not soluble in water are dissolved in hydrochloric acid or in nitrohydro- chloric acid. Some native arsenates require fusing with sodium carbonate. Metallic arsenic, arsenous sulphide, and metallic arsen- 410 DETERMINATION. [ 127. ides are dissolved in fuming nitric acid or nitrohydrochloric acid, or a solution of bromine in hydrochloric acid ; those metallic arsenides which are insoluble in these menstrua are fused with sodium carbonate and potassium nitrate, by which means they are converted into soluble alkali arsenates and insoluble metallic oxides ; or they may be suspended in potassa solution and treated with chlorine ( 164, B, 7). In this last manner, too, arsenous sul- phide dissolved in concentrated potassa may be very easily ren- dered soluble. All solutions of compounds of arsenic which have been effected by long heating with fuming nitric acid, or by warm- ing with excess of nitrohydrochloric acid, or chlorine, contain arsenic acid. A solution of arsenous acid in hydrochloric acid cannot be concentrated by evaporation, since arsenous chloride would escape with the hydrochloric-acid fumes. This, however, less readily takes place if the solution contains arsenic acid ; in fact, it only occurs in the presence of a large proportion of hydro- chloric acid (for instance, half the volume of hydrochloric acid of 1*12 sp. gr.*). It is therefore advisable in most cases where a hydrochloric-acid solution containing arsenic is to be concentrated to previously render the solution alkaline. &. Determination. Arsenic is weighed as lead arsenate, as ammonium magnesium arsenate, as magnesium py roar senate, as uranyl py roar senate, or as arsenous sulphide. The determination as ammonium mag- nesium arsenate is sometimes preceded by precipitation as am- monium arseno-molybdate. The method recommended by BER- THIER and modified by v. KOBELL, of separating the arsenic as basic-ferric arsenate, is only used in separations. Arsenic may be estimated also in an indirect way, and by volumetric methods. We may convert into 1. LEAD ARSENATE: Arsenous and arsenic acids in aqueous or nitric-acid solution. (Acids or halogens forming fixed salts with lead, and also ammonium salts, must not be present.) 2. AMMONIUM MAGNESIUM ARSENATE, or MAGNESIUM PYRO- ARSENATE : a. By direct Precipitation. Arsenic acid in all solutions free from bases or acids precipitable by magnesia or ammonia. * Zeitschr. f. Chem., i, 448. 127.] ARSENOUS AND ARSENIC ACIDS. 411 1). Preceded by Precipitation as Ammonium Arseno-molyb- date. Arsenic acids in all cases where no phosphoric acid is present, little or no hydrochloric acid, nor any substance which decomposes molybdic acid. 3. URANYL PYKO ARSENATE : Arsenic acid in all combinations soluble in water and acetic acid. 4. ARSENOUS SULPHIDE : All compounds of arsenic without exception. Arsenic may be determined volumetrically in a simple and exact manner, whether present in the form of arsenous acid or an alkali arsenite, or as arsenic acid or an alkali arsenate. The volu- metric methods have now almost entirely superseded the indirect gravimetric methods formerly employed to effect the determina- tion of arsenous acid. 1. Determination as Lead Arsenate. a. Arsenic Acid in Aqueous Solution. A weighed portion of the solution is put into a platinum or porcelain dish, and a weighed amount of recently ignited pure lead oxide added (about five or six times the supposed quantity of arse- nic acid present) ; the mixture is cautiously evaporated to dryness, and the residue heated to gentle redness, and maintained some time at this temperature. The residue is lead arsenate + lead oxide. The quantity of arsenic acid is now readily found by sub- tracting from the weight of the residue that of the oxide of lead added. For the properties of lead arsenate, see 92. The results are accurate, provided the residue be not heated beyond gentle red- ness. 5. Arsenous Acid in Solution. Mix the solution with nitric acid, evaporate to a small bulk, add a weighed quantity of lead oxide in excess, evaporate to dry- ness, and ignite the residue most cautiously in a covered crucible, until the whole of the lead nitrate is decomposed. The residue consists here also of arsenic acid + ^ ea( i oxide. This method requires considerable care to guard against loss by decrepitation upon ignition of the lead nitrate. 412 DETERMINATION. [ 127. 2. Estimation as Ammonium Magnesium Arsenate, or Magnesium Pyroar senate. a. By direct Precipitation. This method, which was first recommended by LEVOL, presup- poses the whole of the arsenic in the form of arsenic acid. Where this is not the case, the solution is gently heated, in a capacious flask, with hydrochloric acid, and potassium chlorate added in small portions, until the fluid emits a strong smell of chlorous acid ; it is then allowed to stand at a gentle heat until the odor of this gas has nearly disappeared. The arsenic-acid solution is now mixed with ammonia in ex- cess, which must not produce turbidity, even after standing some time ; magnesia mixture is then added ( 62, 6). The fluid, which smells strongly of ammonia, is allowed to stand 24 or 48 hours in < the cold, well covered, and then filtered through a weighed filter. The precipitate is then transferred to the filter, with the aid of portions of the filtrate, so as to use no more washing water than necessary, and washed with small quantities of a mixture of three parts water and one part ammonia, till the washings, on being mixed with nitric acid and silver nitrate, show no opalescence. The precipitate is dried at 102 to 103, and weighed. It has the for- mula (MgNH 4 AsO 4 ) a -|-H 2 O.'* As the drying of ammonium mag- nesium arsenate till its weight is constant, requires much time and repeated weighings, it is a great advantage that we can now con- vert it without loss of arsenic into magnesium pyroarsenate (Mg a As 2 O 7 ), thanks to the researches of H. RosE,f WITTSTEINJ: and PULLER. For this purpose first transfer the dried precipitate as completely as possible to a watch-glass, saturate the filter with a solution of ammonium nitrate, dry and burn it cautiously in a porcelain crucible. After cooling, transfer the precipitate to the crucible, heat in an air-bath to about 130, continue heating for 2 hours on a sand-bath, then heat for an hour or two on an iron plate a little more strongly, and when the ammonia has been thus entirely * If it is dried in a water-bath, the drying must be extremely prolonged, or otherwise more than 1 eq. will be left. After brief drying in the water-bath the compound contains between 1 and 3 eq. water. If it is dried between 105 and 110 part of the 1 eq. water is lost. \ Handbuch der anal. Chem. 6. Aufl., n, 390. $Zeitschr,f, anal, Chem., n, 19. 15., x, 63. 127.] ARSENOUS AND ARSENIC ACIDS. 413 expelled ignite strongly for some time over the lamp. The pro- cess may be shortened by conducting the heating in a ROSE'S cruci- ble in a slow current of oxygen. The ammonia may then be driven off in 10 minutes, and after the precipitate has been at last strongly heated it will be ready to weigh. For the properties of the ammonium magnesium arsenate and magnesium pyroarsenate, see 92. The method yields satisfactory results, since the small loss of precipitate dissolved in the filtrate and washings is coun- terbalanced by the presence of a trace of basic magnesium sulphate (PULLER). PULLER with a quantity of 0*37 grm. ammonium mag- nesium arsenate lost only a fraction of a milligramme ; on the ad- dition of a large proportion of ammonium chloride the loss rose to about 0-002 grm. The correction for the solubility of the pre- cipitate in the ammoniacal filtrate containing excess of magnesia mixture is 0-001 grm. of (Mg]STH 4 AsO 4 ) a + H a O for 30 c.c. b. Preceded ly Precipitation as Ammonium Arseno-molyl- date. Mix the acid solution, which must be free from phosphoric and silicic acids, with an excess of solution of ammonium molybdate. The ammonium molybdate solution should have been previously' mixed with nitric acid in excess, and the whole process is con- ducted exactly as in the case of phosphoric acid see 134, J, ft. After dissolving the ammonium arseno-molybdate in ammonia, neutralize the latter partially with hydrochloric acid. Treat th^ ammonium magnesium arsenate as in a. Results satisfactory. 3. Estimation as Uranyl Pyroarsenate. This method was first proposed by WERTHER.* It lias been carefully studied by PuLLERf in my laboratory, and gives thor- oughly satisfactory results. Mix the arsenic acid solution with potash or ammonia in excess, and then a good excess of acetic acid. (If a precipitate of ferric or aluminium arsenate here remains insoluble, the method would be inapplicable.) Add uranyl acetate in excess, and boil. Wash the slimy precipitate of uranyl arsenat^ or of ammonium uranyl arsenate by decantation with boiling water, and then transfer to a filter. The addition of a few drops of chta roform to the partly cool fluid will hasten the deposition of the pre^ cipitate. Dry, transfer the precipitate to a watch-glass, cleaning *Journ.f. prakt. Chcm., XLIIT, 346. \Zeitschr.f. analyt. Chem., x, 72. 414 DETERMINATION. [ 127. the filter as much as possible ; saturate the latter with ammonium nitrate, dry it, incinerate in a porcelain crucible, and add the pre- cipitate. If the precipitate contains ammonium, heat very cau- tiously, finally adding nitric acid, or ignite in oxygen. (See 2, a.) If the precipitate is free from ammonium, ignite in the ordinary way. Ammonium salts do not interfere. Properties of the pre- cipitate and residue, 92, e. 4. Estimation as Arsenous Sulphide. a. In solutions of Arsenous Acid or Arsenites free from Arsenic Acid. The solution should be strongly acid with hydrochloric acid. Precipitate with hydrogen sulphide and expel the excess with car- bon dioxide. Pass the latter through the solution for an hour, a longer time is useless. (See 125, 1.) Wash the precipitate thor- oughly and dry at 100 till the weight is constant. Particles of the precipitate which adhere so firmly to the glass that they can- iiot be removed mechanically are dissolved in ammonia and repre- cipitated with hydrochloric acid. Properties of the precipitate, 92. Do not omit to test a weighed portion to see whether it completely volatilizes on heating. If a residue remains it is to be weighed and the proportional quantity deducted from the total weight of the precipitate. Results accurate. If the solution contains any substance which decomposes hydro- gen sulphide, such as ferric chloride, chromic acid, etc., the precip- itate produced in the cold contains an admixture of finely divided sulphur. It should be collected in the same manner on a filter dried at 100, arid weighed, washed and dried. Extract the admixed sulphur with purified carbon disulphide (which should leave no residue on evaporation), continuing till the fluid which runs through leaves no residue. Dry at 100 till the weight is constant. From experiments made in my laboratory it appears that the results thus obtained are quite accurate, even when the amount of admixed sulphur is large ; but the precipitation must have been effected in the cold. If, on the contrary, heat is used, the sulphur is in the form of small agglutinated grains and cannot be completely extracted by cold carbon disulphide on the filter. However, it may be extracted by removing the precipitate from the filter and repeatedly digesting it with the disulphide on a water- bath (PULLER*). * ZeitscJir. f. analyt. Chem., x, 46 et seq. 127.] ARSENOUS AND ARSENIC ACIDS. 415 Instead of purifying the arsenous sulphide you may estimate the arsenic in the mixture of the sulphide with sulphur as follows: Dissolve the precipitate in strong potassa, and pass chlorine into the solution ( 148, II. 2, 5). The arsenic and the sulphur are con- verted into arsenic and sulphuric acids respectively ; the former may be estimated according to 2, 0, or the latter according to 132. In the latter case, deduct the sulphur found from the weight of the arsenical precipitate. There is no loss of arsenic in this process from volatilization of the chloride, as the solution remains alkaline. The object may also be conveniently attained by the use of nitric acid. A very strong fuming acid, of 86 boiling point, is employed ; an acid of 1-42 sp. gr. which boils at a higher tempera- ture does not answer the purpose, as the separated sulphur would fuse, and its oxidation would be much retarded. The well dried precipitate is shaken into a small porcelain dish, treated with a tol- erably large excess of the fuming nitric acid, the dish immediately covered with a watch-glass, and as soon as the turbulence of the first action has somewhat abated, heated on a water-bath till all the sulphur has disappeared, and the nitric acid has evaporated to a small volume. The filter to which the unremovable traces of arsenous sulphide adhere is treated separately in the same manner, the complete destruction of the organic matter being finally effected by gently warming the somewhat dilute solution with potassium chlorate (BUNSEN*). Or the filter may instead be extracted with ammonia, the solution evaporated in a separate dish, and the resid- ual sulphide treated as above. In the mixed solution the arsenic acid is finally precipitated as ammonium magnesium arsenate. ( 127, 2, a). Treatment of the impure precipitate with ammonia, whereby the sulphide is dissolved, and the sulphur is supposed to remain behind, only gives approximate results, as the ammoniacal solution of arsenous sulphide takes up a little sulphur. 1. In solutions of Arsenic Acid, or of a mixture of the two Oxides of Arsenic. Heat the solution in a flask (preferably on an iron plate) to about 70, and conduct hydrogen sulphide at the same time into the fluid, so long as precipitation takes place. The precipitate formed is always a mixture of sulphur and arsenous sulphide, since the arsenic acid is first reduced to arsenous acid with separa- * Annal d. Chem. u. Pharm , cvi, 10. 416 DETERMINATION. [ 127. tion of sulphur, and then the latter is decomposed (H. KOSE *). Only in the case when a sulpho-salt containing pentasulphide of arsenic is decomposed with an acid, is the precipitate actually pentasulphide, and not merely a mixture of sulphur with arse- nous sulphide (A. FTJCHS f). To convert this mixture of arsenous sulphide and granular sulphur into pure arsenous sulphide suitable for. weighing, treat it as follows : Extract the washed and still moist precipitate on the filter with ammonia, wash the resid- ual sulphur, precipitate the solution with hydrochloric acid with- out heat, filter, dry, extract with carbon disulphide, dry at 100, and weigh. Results accurate. The mixture of arsenous sulphide and sulphur obtained by hot precipitation may, of course, also be estimated directly or indirectly after one of the other methods in 4, a. 5. Volumetric Methods, a. Method which presupposes the presence of A.rsenous Acid. 1. FR. MOHE'S method.;); This method depends upon the principle already stated under Antimonous Oxide, 125, 3, a\ i.e., arsenous acid in alkaline solution is oxidized by iodine to arsenic acid (As 3 O 8 + 4NaOH + 41 + = As 2 O 6 + 4NaI+ 2H a O). If, therefore, you have an aqueous solution of arsenous acid or an alkali arsenite, weigh or measure off a quantity that will contain about O'l grm. As a O 3 ; add to it 20 c. c. of a saturated solution of sodium bicarbonate (previously purified by washing with water), then add a little thin starch paste, and finally titrate with standard iodine solution ( 146) until the starch-iodide reac- tion appears ; 4 eq. of 'iodine correspond to 1 eq. of arsenous acid. If the arsenous-acid solution is acid, neutralize it with pure sodium carbonate; but, on the other hand, if alkaline, neutralize with hydrochloric acid before adding the sodium bicarbonate. It is of course understood that the solution contains no substances (sul- phides or thiosulphates) that will act upon iodine. The results are accurate. Compare Expt. No. 79 ; also WAITZ. * Pogg. Annal., cvn, 186. \ Zntschr.f. analyt. Chem., I, 189. } Lehrbuch der Titrirmethode, 3. Aufl., 275. %Zeitsc7ir. f. analyt. Chem., x, 1(52. The attempts made by WAITZ to convert the arsenic in arsenous sulphide into alkali arsenite were unsuccessful. [ 127. ARSENOUS AND ARSENIC ACIDS. 417 2. KESSLER'S method.* This method depends upon the principle stated under 125, 3, 5; i.e., the oxidation of arsenous to arsenic acid is effected in hydrochloric-acid solution by the use of potassium dichromate, f and is carried out in exactly a similar manner. The results are reliable only when care is taken that the hydrochloric acid (sp. gr. 1*12) constitutes at least one sixth of the volume of the liquid ; it should not, however, exceed one half the volume, otherwise the end reaction, due to the formation of iron ferricyanide, sets in more slowly and loses in sharpness. If for any reason the direct titration of the hydrochloric- .acid solution is impracticable, precipitate with hydrogen sulphide (if arsenic acid is present, at TO ), wash the precipitate, transfer it together with the filter to a stoppered flask, and treat it with an almost saturated solution of mercuric chloride in hydrochloric acid (sp. gr. 1*12), stopper tightly, digest at a gentle warmth until the precipitate has become white, dilute with a measured quantity of water (so that the proportion of hydrochloric acid of sp. gr. 1*12 does not fall below one sixth), add solution of potassium dichromate, then iron solution, etc., as detailed under 125, 3, b. Results good. Compare also WAITZ.^ 3. BUNSEN'S method. This method is based upon the follow- ing facts : aa. If potassium dichromate is boiled with concentrated hydro- chloric acid, 6 at. chlorine are disengaged for every 2 mol. chromic acid; 2CrO 8 + 12HC1 = CrCl, + 6H,O + 6C1. l>b. But if arsenous acid is present (not in excess) there is not the quantity of chlorine disengaged corresponding to the chromic acid, but so much less of that element as is required to convert the arsenous into arsenic acid (H 3 AsO, + 2C1 -f- H a O = H,AsO 4 -f- 2HC1). Consequently for every 2 at. chlorine wanting there is to be reckoned 1 mol. arsenous acid. * Pogg. AnnaL, xcv, 204; cxui, 184; cxvm, 17; Zeitschr.f. analyt. Chem., IT, 383. f Oxidation may also be effected by potassium permanganate, an excess being added, and this latter then determined with iron. The estimation is inac- curate in hydrochloric-acid solution, hence the permanganate can only be used in sulphuric-acid solution. Compare WAITZ, ZeitscJir.f. analyt. Chem., x, 174. J Zeitschr.f. analyt. Chem., x, 169. AnnaL d. Chem. u. Pharm., LXXXVI, 290. 418 DETERMINATION. [ 127. CG. The quantity of chlorine is estimated by determining the quantity of iodine liberated by it from potassium iodide. These are the principles of BUNSEN'S method. For the man- ner of execution I refer to the Estimation of Chromic Acid. ~b. Method which presupposes the presence of Arsenic Acid. This method depends on the precipitation of the arsenic acid by uranium solution and the recognition of the end of the reaction by means of potassium ferrocyanide. It is therefore the same as was suggested for phosphoric acid by LECOMTE, and brought into use by NEUBAUEK,* and afterwards by PiNcus.f BODEKER,^; who first employed the process for arsenic acid, recommends the employment of a solution of uranyl nitrate, as- this is more permanent than the hitherto used acetate, which is gradually decomposed by the action of light. The uranium solution has the correct degree of dilution, if it contains about 20 grm. of uranium in 1 litre. It should contain as little free acid as possible. The determination of its value may be effected with the aid of pure sodium arsenate or by means of arsenous acid the latter is converted into arsenic acid by boiling with fuming nitric acid. The solution is rendered strongly alka- line with ammonia, and then distinctly acid with acetic acid. The uranium solution is now run in from the burette slowly, the liquid being well stirred all the while, till a drop of the mixture spread out on a porcelain plate, gives with a drop of potassium ferrocya- nide placed in its centre, a distinct reddish-brown line where the two ilnids meet. The height of the fluid in the burette is now read off, the level of the mixture in the beaker is marked with a, strip of gummed paper, and the beaker is emptied and washed, filled with water with addition of about as much ammonia and acetic acid as was before employed, and the uranium solution is cautiously dropped in from the burette, till a drop taken out of the beaker and tested as above, gives an equally distinct reaction. The quantity of uranium solution used in this last experiment is the excess, which must be added to make the end-reaction plai^ for the dilution adopted. This amount is subtracted from that used in the first experiment, and we then know the exact value of the uranium solution with reference to arsenic acid. * Arckivfur wissenchaftHche Heilkunde, iv, 228. \Journ.f. prakt. Chem., LXXVI, 104. $ Annal. de Chem. u. Pharm., cxvn, 195 127.] AKSENOUS AND ARSENIC AOIDS. 419 In an actual analysis, the arsenic is first brought into the form of arsenic acid, a clear solution is obtained containing ammonium acetate and some free acetic acid,* and the process is conducted exactly as in determining the value of the standard solution. The experiment to ascertain the correction must not be omitted here, otherwise errors are sure to arise from the different degrees of dilu- tion of the arsenic acid solutions used in the determination of the value of the standard solution and in the actual analyses. The results of two determinations of arsenic given by BODEKEB are satisfactory. To execute the method well requires practice. The results are not exact enough unless the conditions as regards amount and quality of alkali salts are nearly similar in the standardizing of the uranium solution and in its use. Compare WAixz.f 6. Estimation of Arsenous Acid l}y Indirect Gravimet- ric Analysis. a. ROSE'S method. Add to the hydrochloric acid solution, in the preparation of which care must be taken to exclude oxidizing substances, a solution of sodium- or ammonium-auric chloride in excess, and digest the mixture for several days, in the cold, or, in the case of dilute solutions, at a gentle warmth ; then weigh the separated gold as directed in 123. Keep the filtrate to make quite sure that no more gold will separate. 2 at. gold correspond to 3 rnoL arsenous acid. I). YOHI/S^: method. Mix the solution with a weighed quan- tity of potassium dichromate, and free sulphuric acid ; estimate the chromic acid still present by the method given in 130, c, and deduce from the quantity of that acid consumed in the process, i.e., reduced by the arsenous acid, the quantity of the latter, after the formula 3H, AsO 3 + 2CrO 3 = 3H 3 AsO 4 + O 2 O 3 . * Alkalies, alkali earths, and zinc oxide may be present, but not such metals as yield colored precipitates with potassium ferrocyanide, as, for instance, copper. \Zeit8chr.f, anal. Chem., x, 182. %Annal. de CJiem. u. Pharm., xciv, 219. 420 DETERMINATION. [ 128. Supplement to the Sixth Group. 128. 8. MOLYBDIC ACID. Molybdic acid is converted, for the purpose of its determina- tion, either into molybdenum dioxide, or into lead molybdate, or into molybdenum disulphide. a. Molybdic anhydride (MoO 3 ), and also ammonium molybdate, may be reduced to dioxide by heating in a current of hydrogen gas. This may be done either in a porcelain boat, placed in a wide glass tube, or in a platinum or porcelain crucible with perforated cover (108, Fig. 83). The operation is continued till the weight remains constant. The temperature must not exceed a gentle redness, otherwise the dioxide itself might lose oxygen and become partially converted into metal. In the case of ammonium molybdate the heat must be very low at first on account of the frothing. If you have a platinum tube it is safer to ignite the molybdic acid in this for 2 or 3 hours in a slow current of hydrogen, thus reducing it to the metallic state. When reducing to dioxide the contents of the crucible are frequently gray below, and brown above (RAMMELS- b. The following is the best method of precipitating molybdic acid from an alkaline solution : Dilute the solution, if necessary, neutralize the free alkali with nitric acid, and allow the carbonic acid, which may be liberated in the process, to escape, then add neutral mercurous nitrate. The yellow precipitate formed appears at first bulky, but after several hours' standing it shrinks ; it is insoluble in the fluid, which contains an excess of mercurous nitrate. Collect on a filter, and wash with a dilute solution of mer- curous nitrate, as it is slightly soluble in pure water. Dry, remove the precipitate as completely as practicable from the filter, and deter- mine the molybdenum in it as directed in a (H. ROSE) ; or mix the precipitate, together with the filter-ash, with a weighed quantity of ignited lead oxide, and ignite until all the mercury is expelled ; then add some ammonium nitrate, ignite' again and weigh. The excess obtained, over and above the weight of the lead oxide used, is molybdenum trioxide * Pogg. Anndl., cxxvii, 281; Zeitschr.f. analyt. CJicm., v, 203. \Journ.f. prakt. Chem., LXVII, 472. 128.] MOLYBDIC ACID. 421 c. CHATARP* recommends estimating rnolybdic acid in the solu- tion of its alkali salts by adding lead acetate in slight excess to the boiling solution and boiling for a few minutes. The precipitate which is at first milky becomes granular, deposits well, and may be easily washed with hot water. It is dried, removed from the filter as much as possible, ignited and weighed as PbMoO 4 . The method is only applicable for solutions of pure alkali rnolybdates. d. The precipitation of molybdenum as sulphide is always a difficult operation. If the acid solution is supersaturated with hydrogen sulphide, warmed, and filtered, the filtrate and washings are generally still colored. They must, accordingly, be warmed, and hydrogen sulphide again added, and the operation must after- wards, if necessary, be repeated until the washings appear almost colorless. The precipitation succeeds better when the molybdenum sulphide is dissolved in a relatively large excess of ammonium sul- phide, and, after the fluid has acquired a reddish-yellow tint, precipi- tated with hydrochloric acid. ZENKERf advises then to boil, until the hydrogen sulphide is expelled, and to wash with hot water, at first slightly acidified. To make quite sure that all the molyb- denum is precipitated, treat the filtrate and washings again with hydrogen sulphide and allow to stand for some time. The brown molybdenum sulphide is collected on a weighed filter, and the molybdenum determined in an aliquot part of it, by gentle ignition in a current of hydrogen gas, as in a. The brown molybdenum sulphide changes in this process to the gray disulphide (H. ROSE). .e. F. PISANI;}; gives the following method for estimating molyb- dic acid volumetrically : Digest the molybdic acid with hydro- chloric acid and zinc, dissolving any precipitate which may form from want of acid and also the excess of zinc. The molybdic acid is thus reduced to a molybdenum salt corresponding to molybdenum sesquioxide. Convert the molybdenum in this solution again into molybdic acid by standard potassium permanganate. The brown color of the solution turns first green, and then disappears. RAM- MELSBERG confirms the statements of PISAOT. * Sill Amer. Journ. (3), i, 416. f Journ.f. prakt. Chem., LVIII, 259. \ Compt. Rend., LIX, 301. Pogg. Annal., cxxvn, 281; Zeitschr.f. analyt. Chem., v f 203. 422 DETERMINATION. [ 129, 130. II. DETERMINATION OF ACIDS IN COMPOUNDS CONTAINING ONLY ONE ACID, FREE OR COMBINED; AND SEPARATION OF ACID FROM BASIC RADICALS. First Group, FIRST DIVISION. ARSENOUS ACID ARSENIC ACID CHROMIC ACID (Selenous Acid, Sulphurous and Hyposulphurous Acids, lodic Acid). 129. 1. ARSENOUS AND ARSENIC ACIDS. These have been already treated of among. the bases ( 127) on account of their behavior with hydrogen sulphide ; they are merely mentioned here to indicate the place to which they properly be- long. The methods of separating them from the bases will be found in Section Y. 130. 2. CHROMIC Aero. I. DETERMINATION. Chromic acid is determined either as chromic oxide or lead chr ornate. But it may be estimated also from the quantity of car- bon dioxide disengaged by its action upon oxalic acid in excess, and also by volumetric analysis. In employing the first method it must be borne in mind that 1 mol. chromic oxide corresponds to 2 mol. chromic acid. a. Determination as Chromic Oxide. a. The chromic acid is reduced to the state of a chromic salt and the amount of chromium in the latter determined ( 106). The reduction is effected either by heating the solution with hydro- chloric acid and alcohol ; or by mixing hydrochloric acid with the solution, and conducting hydrogen sulphide into the mixture ; or by adding a strong solution of sulphurous acid, and applying a gen- tle heat. With concentrated solutions the first method is gener- ally resorted to, with dilute solutions one of the two latter. With respect to the first method, I have to remark that the alcohol must be expelled before the chromium can be precipitated as hydroxide 130.] CHROMIC ACID. 423 by ammonia ; and with respect to the second, that the solution supersaturated with hydrogen sulphide must be allowed to stand in a moderately warm place, until the separated sulphur has com- pletely subsided. The results are accurate, unless the weighed pre- cipitate contains silica and lime, which is always the case if the pre- cipitation is effected in glass vessels. ft. The neutral or slightly acid (nitric acid) solution is precipi- tated with mercurous nitrate, after long standing the red precipitate of mercurous chromate is filtered off, washed with a dilute solution of mercurous nitrate, dried, ignited, and the residuary chromic oxide weighed (II. ROSE). Results accurate. b. Determination as Lead Chromate. The solution is mixed with sodium acetate in excess, and acetic acid added until the reaction is strongly acid ; the solution is then precipitated with neutral lead acetate. The washed precipitate is either collected on a weighed filter, dried in the water-bath, and weighed; or it is gently ignited as directed 53, and then weighed. For the properties of the precipitate, see 93, 2. Results accurate. c. Determination as Barium Chromate. Moderately acidulate the alkali-chromate solution with acetic acid, add a slight excess of barium chloride, allow the fine pre- cipitate to stand 12 hours, wash it with a solution of ammonium acetate so far as possible by decantation, displacing the last por- tion of the solution with ammonium nitrate (or the chromate may be partially reduced on ignition), dry the precipitate, and ignite it after removal, so far as is possible, from the filter. Properties and composition of the barium chromate are given tinder 93, 2, c (H. ROSE; PEARSON *). The test analyses given by PEARSON are satisfactory. d. Determination ~by means of Oxalic Acid (after YOHL). When chromic acid and oxalic acid are brought together in the presence of water and excess of sulphuric acid, chromic sulphate and carbon dioxide are formed, 3H 2 C 2 O 4 -f- 2H 2 CrO 4 -f- 3H 2 SO 4 = 6CO, + Cr 2 (SO 4 ), + 8H,O. Accordingly the amount of chromic acid can be calculated from the weight of carbon dioxide evolved. The process is the same as in the analysis of manganese ores ( 230). 1 part of chromic acid requires 2J * Amer. Journ. of Science [2], XLV, 298; Zeitschr.f. analyt. Chlm., ix, 108. 424 DETERMINATION. [ 130. parts of sodium oxalate. If it is intended to determine potas- sium or sodium in the residue, ammonium oxalate is used. e. Determination by Volumetric Analysis, a. SCHWARZ'S method. The principle of this very accurate method is identical with that upon which PENNY'S method of determining iron is based ( 112, 2, Z>). The execution is simple: Acidify the not too dilute solution of the chromate with sulphuric acid, add in excess a measured quantity of solution of a ferrous salt, the strength of which you have previously ascertained according to the direc- tions of 112, 2, #, or &, or the solution of a weighed quantity of ammonium-ferrous sulphate, free from ferric salt, and then determine in the manner directed in 112, 2 #, or 5, the quan- tity of ferrous iron remaining. The difference shows the amount of iron- that has been converted by the chromic acid from a ferrous to a ferric salt. 1 grm. of iron corresponds to 0*5969 of chromic anhydride (CrO 8 ). To determine the chromic acid in lead chromate, the latter is, after addition of the ammonium ferrous sulphate, most thoroughly triturated with hydrochloric acid, water added, and the analysis then proceeded with. /?. BUNSEN'S method.* If a chromate is boiled with an excess of fuming hydrochloric acid, there are disengaged for every atom of chromium 3 at. chlorine; for instance, K a Cr 2 O 7 + 14HC1 = 2KC1 + 2CrCl s + 6C1 -f- 7H a O. If the escaping gas is conducted into a solution of potassium iodide in excess, the 3 at. chlorine set free 3 at. iodine. The liberated iodine may next be determined as de- scribed in 146. 380 '55 of iodine corresponds to lOO'l of chromic anhydride (CrO 3 ). The analytical process is conducted as follows : Put the weighed sample of the chromate (say 0*3 to 0*4 grm.) into the little flask d, Fig. 89 (blown before the lamp, and holding only from 36 to 40 c. c.), and fill the flask two-thirds with pure fuming hydrochloric acid free from Cl and SO 2 ), and add a compact lump of magnesite to keep up a constant current of gas and prevent the fluid from receding. Connect the bulbed evolution tube a with the neck * Annal. d. Ghem. u. Pharm., LXXXVI, 279. 130.] CHROMIC ACID. 425 of the flask by means of a stout india-rubber tube c. As shown in the engraving, a is a bent pipette, drawn out at the lower end into an upturned point. A loss of chlorine need not be apprehended on adding the hydrochloric acid, as the disengage- Fig. 89. ment of that gas begins only upon the application of heat. Insert the evolution tube into the neck of the retort, which is one-third filled with solution of potassium iodide.* This retort holds about 160 c.c. The neck presents two small expansions, blown before the lamp, and intended, the lower one, to receive the liquid which is forced up during the operation, the upper one to serve as an additional guard against spirting. Apply heat now, cautiously, to the little flask. After two or three minutes ebullition the whole of the chlorine has passed over, and liberated its equivalent quan- tity of iodine in the potassium iodide solution. "When the ebulli- tion is at an end, take hold of the caoutchouc tube c with the left hand, and, whilst steadily holding the lamp under the flask with the right, lift a so far out of the retort that the curved point is in the bulb &. Now remove first the lamp, then the flask, dip the retort in cold water to cool it, and shake the fluid in it about to effect the complete solution of the separated iodine in the excess of potas- sium iodide solution. "When the fluid is quite cold, transfer it to a beaker, rinsing the retort into the beaker, and proceed as directed 146. The method gives very satisfactory results. The apparatus here recommended differs slightly from that used by BUNSEN, the retort of the latter having only one bulb in the neck, and the evo- lution tube 110 bulb, being closed instead, at the lower end, by a glass or caoutchouc valve, which permits the exit of the gas from * 1 part of pure potassium iodide, free from iodic acid, dissolved in 10 parts of water. The fluid must show no brown tint immediately after addition of dilute sulphuric acid. 426 DETERMINATION. [ 130. the tube, but opposes the entrance of the fluid into it. I think the modifications which I have made in BUNSEN'S apparatus are calcu- lated to facilitate the success of the operation. Instead of this ap- paratus, that described in 112 may also be very conveniently used. y. There need be only mention made here regarding the method by RUBE,* which is based on the equation 2Cr() 3 -\ 6K 4 Fe(Ctf).+ 12HCl = 6KC1 + 2CrCl i +3K.Fe,(CN) J ,-f 6H 3 O> and also regarding the method devised by ZuLKOwsKY,f which is* based on the direct (i.e., without distillation) estimation of the iodine separated by chromic acid, and which is carried out ex- actly as detailed under 113, /?, in estimating iron. II. SEPARATION OF CHROMIC ACID FROM THE BASIC RADICALS. a. Of the First Group. a. Reduce the chromic acid to a chromic salt, as directed in I., and separate the chromium from the alkalies as directed in 155. /?. Mix the potassium or sodium chrornate with about 5 parts of dry pulverized ammonium chloride, and heat the mixture cau- tiously. The residue contains the chlorides of the alkali metals and chromic oxide, which may be separated by means of water. y. Precipitate the chromic acid according to L, a, ft, and sep- arate the mercury and alkali metals in the filtrate by 162. b. Of the Second Group. a. Fuse the compound with 4 parts of sodium and potassium carbonates, and treat the fused mass with hot water, which dis- solves the chromic acid in the form of an alkali chromate. The residue contains the alkali earth metals in the form of carbonates ; but as they contain alkali, they cannot be weighed directly. The chromic acid in the solution is determined as in I. Strontium and calcium chromates may be decomposed by boiling with potassium or sodium carbonate. Barium chromate may also be decomposed in the same way, but the boiling must be repeated a second time with fresh solution of alkali carbonate (H. ROSE). * Journ. /. prakt. Chem., xcv, 58; Zeitschr.f. analyt. Chem., iv, 444. \Journ.f. prakt. Chem., cm, 351; Zeitschr.f. analyt. Chem., vm, 74. 130.] CHROMIC ACID. 427 J3. Dissolve in hydrochloric acid, reduce the chromic acid according to L, 0, and separate the chromium from the alkali earth metals according to 156. y. Magnesium chromate, as well as other chromates of the alkali earth metals soluble in water, may be easily decomposed also, by determining the chromic acid according to I., a, /?, or L, 5, and separating the magnesium, etc., in the filtrate from the excess of the salt of mercury or lead as directed 162. d. Barium strontium and calcium chromates may also be decomposed by the method described II. , , ft. Compare BAHR, Analysis of barium and calcium dichromates, etc.* H. ROSE recommends using 5 parts of ammonium chloride to 1 part of the very finely powdered substance. One single igni- tion of the mixture usually suffices for complete decomposition, but it is safer to repeat the ignition with ammonium chloride, to make sure that the weight remains constant, before washing out the barium chloride from the residue. c. Of the Third Group. a. From Aluminium. If you have chromic acid to separate from aluminium in acid solution, precipitate the aluminium with ammonia or ammonium carbonate ( 105, #), and determine the chromic acid in the filtrate according to I. If the washed aluminium hydroxide has a yellow color, treat on the filter with ammonia, and wash with boiling water ; this will remove the last traces of chromic acid. However, a little aluminium hydroxide dissolves in the ammonia, therefore heat the ammoniacal fluid in a platinum dish till it has almost lost its alkaline reaction, and collect on a filter the flocks of aluminium hydroxide which separate, and add them to the principal precip- itate. ft. From Chromium. aa. Determine in one portion the quantity of the chromic acid according to L, d, or I., e, a, or ft, and in another portion the total amount of the chromium, by converting it into sesquioxide by cautious ignition with ammonium chloride, or by I., #, or by converting it entirely into chromic acid by 106, 2. 55. In many cases the chromic acid may be precipitated accord- * Journ.f. prakt. Chem., LX, 60. 428 DETERMINATION. [ 130. ing to I., a, y#, or L, J. The chromium and mercury, or lead, in the filtrate, are separated as directed 162. cc. The hydrated compounds of sesquioxide of chromium with chromium trioxide, or chromic chromates, such as are obtained by precipitating a solution of chromic salt with potassium chromate,. etc., may also be analyzed by ignition in a stream of dry air, in a bulb tube, to which a calcium chloride tube is attached (Fig- 44, 36). The loss of weight represents the joint amount of oxygen and water that have escaped. If the increment of the CaCl a tube is deducted, we shall have the oxygen. Now every 3 at. oxygen correspond to 2 mol. CrO 3 . The amount of the latter being thus calculated, we have only to subtract its equivalent quantity of ses- quioxide from the weight of residue after the ignition, and the remainder is the quantity of sesquioxide originally present. VOGEL* and also STOEEE and ELLIOT^ have employed this method. d. Of the Fourth Group. a. Proceed as directed in &, a. Upon treating the fused mass with hot water, oxides of the basic metals are left. In the case of manganese the fusion must be effected in an atmosphere of carbon dioxide. Apparatus, Fig. 83 in 108. /?. Reduce the chromic acid as directed in I., a, and separate the chromium from the metals in question, as directed in 160. e. Of the Fifth and Sixth Groups. a. Acidify the solution, and precipitate, either at once or after reduction of the chromic acid by sulphurous acid, with hydrogen sulphide. The metals of the fifth and sixth groups precipitate in conjunction with free sulphur ( 115 to 127), the chromic acid is reduced. Filter and determine the chromium in the filtrate, as directed in I., a. /?. Lead chromate may be conveniently decomposed by heating with hydrochloric acid and some alcohol ; the lead chloride and chromic chloride formed are subsequently separated by means of alcohol (compare 162). The alcoholic solution ought always to be tested with sulphuric acid ; should a precipitate of lead sulphate form, this must be filtered off, weighed, and taken into account. Compare also 130, 1, d. * Journ.f. prakt. Chem., LXXVII, 484. f Proceedings of the American Academy, V, 198. 131.] SELENOUS ACID. 429 Supplement to the First Division. 131. 1. SELENOUS ACID. From aqueous or hydrochloric- acid solutions of selenous acid, the selenium is precipitated by sulphurous acid gas, or, in presence of an excess of acid, by so'dium sulphite, or ammonium sulphite. The liquid containing the precipitate is heated to boiling for J hour, which changes the precipitate from its original red color to black, and makes it dense and heavy. The liquid is tested by a further addition of the reagent to see whether any more selenium will sep- arate ; the precipitate is finally collected on a weighed filter, dried at a temperature somewhat below 100, and weighed. Since H. ROSE* has shown that the presence of hydrochloric acid is an essen- tial condition to the complete reduction of selenous acid, the for- mer acid must be added, if not already present. To make quite sure that all the selenium has been removed, the filtrate is evapo- rated to a small volume, with addition of potassium or sodium chlo- ride, boiled with strong hydrochloric acid, so as to reduce any sele- nic acid to selenous acid, and tested once more with sulphurous acid. If the solution contains nitric acid it must be evaporated repeatedly with hydrochloric acid, with addition of sodium or potassium chloride. If the latter were omitted there would be considerable loss of selenous acid (RATHKE f). As regards the separation of selenous acid from basic radicals, the following brief directions will suffice : a. If the basic radicals are not liable to be altered by the action of sulphurous acid and hydrochloric acid, the selenium may be at once precipitated in the way just given ; the filtrate, when evap- orated with sulphuric acid, yields the base as sulphate. 1). From basic metals which are not thrown down from acid solu- tion by hydrogen sulphide, the selenous acid may be separated by precipitation with that reagent. The precipitate (according to RATHKE, if a mixture of SeS, , Se a S and S) contains 2 at. sulphur to 1 at. selenium. If it is dried at or a little below 100, the weight * Zeitschr. f. analyt. Chem., i, 73. \Journ. f. prakt. CJiem., cviu, 249; ZeitscJir.f. analyt. Chem., ix, 484. ^Journ.f. prakt. Cliem., cvni, 252. 430 DETERMINATION. [ 131, of the selenium may be accurately ascertained. Should, however, extra sulphur be mixed with the precipitate, the latter is oxidized while still moist with hydrochloric acid and potassium chlorate, or by treatment with potassa solution with simultaneous heating and transmission of chlorine. It is necessary here to oxidize the sul- phur completely, as it may enclose selenium, The solution now containing selenic acid is heated till it smells no longer of chlorine, hydrochloric acid is added, and the mixture is reheated. The sele- nic acid is hereby reduced to selenous acid, and when the solution has again ceased to smell of chlorine, the selenium is precip- itated with sulphurous acid. Instead of this process you may digest the precipitate of sulphur and selenium for some hours with con- centrated potassium cyanide, which will completely dissolve it, and then throw down the selenium from the dilute solution with hydro- chloric acid as in c (RATHKE, loc. cit.). c. In many selenites or selenates the selenium may also be determined by converting first into potassium selenocyanate, and precipitating the aqueous solution of the latter with hydrochloric acid (QppENHEiM*). To this end the substance is mixed with 1 or 8 times its quantity of ordinary potassium cyanide (containing cyanic acid), the mixture is put into a long-necked flask, or a porce- lain crucible, covered with a layer of potassium cyanide, and fused in a stream of hydrogen. The temperature is kept so low that the glass or porcelain is not attacked, and while cooling care must be taken to exclude atmospheric air. When cold, the brown mass is treated with water, and the colorless solution filtered, if neces- sary. The liquid should be somewhat but not immoderately diluted. Now boil some time (in order to convert the small quan- tity of potassium seienide that may be present into potassium sele- nocyanate, by the excess of potassium cyanide, allow to cool, super- saturate with hydrochloric acid, and heat again for some time. At the end of 12 or 24 hours all selenium will have separated, filter, dry at 100, and weigh. The results obtained by this process are accurate (H. Rossf ). If the selenium agglomerates together on heating, it may enclose salts. In such cases, by way of control, it should be redissolved in nitric acid, and, after addition of hydro- chloric acid, precipitated with sulphurous acid. The fluid filtered from the selenium precipitate is, as a rule, free from selenium ; it * Journ.f. prakt. Chem., LXXI, 280. f Zeitschr. f. anatyt. Chem., I, 73. 131.] SULPHUROUS ACID. 431 is, however, always well to satisfy one's self on this point by the addition of sulphurous acid. d. From many basic radicals selenous acid (and also selenic acid) may be separated by fusing the compound with 2 parts of sodium carbonate and one part of potassium nitrate, extracting the- fused mass thoroughly by boiling with water, saturating the filtrate, if necessary, with carbonic acid, to free it from lead which it might contain, then boiling down w T ith hydrochloric acid in excess (to reduce the selenic acid and drive off the nitric acid), and precipi- tating finally with sulphurous acid. Selenium, if pure, must volatilize without residue when heated in a tube. 2. SULPHUROUS ACID. To estimate free sulphurous acid in a fluid which may contain also other acids (sulphuric acid, hydrochloric acid, acetic acid), a weighed quantity of the fluid is diluted with water, absolutely free from air,* until the diluted liquid contains not more than 0-05 per cent, by weight of sulphurous acid, the solution is poured with stirring into an excess of standard solution of iodine, the free iodine remaining is titrated with sodium thiosulphate, and the iodine used for the conversion of sulphurous into sulphuric acid is thus found. The reaction is expressed by the equation, SO a + 2l -f 2H 2 O = H 2 SO 4 + 2III. According to FINKENER, f if the iodine is added to the sulphurous acid the reaction is not quite normal. Anyhow this method of operating prevents any loss of sulphurous acid. For the details, see 146. In case of sulphites soluble in water or acids, water perfectly free from air is poured over the substance, in sufficient quantity to attain the degree of dilution stated above, sulphuric or hydrochloric acid is added in excess, and then the titration is effected as above. The greatest care must be taken in this method, to use, for the purpose of dilution, water absolutely free from air. Sulphurous acid may also be determined in the gravimetric way, by conversion into sulphuric acid, and precipitation of the latter with barium chloride, according to 132. This method is espe- cially applicable in the case of sulphites quite free from sulphuric acid. The conversion of the sulphurous into sulphuric acid is * Prepared by long-continued boiling and subsequent cooling with exclusion of air. f Handb. der analyt. Chem. von H. ROSE, 6. Aufl. von FINKENER, n, 937. 432 DETERMINATION. [ 131. effected in the wet way, best by pouring the dilute solution with stirring into excess of chlorine or bromine water. Sulphites insolu- ble in water are decomposed by boiling with sodium carbonate, and the solution of sodium sulphite is treated as directed. After driv- ing off the excess of chlorine or bromine by heating, the moderately acid solution is precipitated with barium chloride. Sulphites may be oxidized in the dry way by heating in a platinum crucible, with 4 parts of a mixture of equal parts sodium carbonate and potassium nitrate. 3. THIOSULPHTJKIC ACID. Thiosulphuric acid, in form of soluble thiosulphates, may be determined by means of iodine, in a similar way to sulphurous acid. The reaction is represented by the equation, 2Na a S a O, + 21 = 2NaI -f- Na 2 S 4 O 6 . The salt under examination is dissolved in a large amount of water, starch-paste added, and then the neutral solution is titrated with iodine. That this method can give correct results only in cases where no other substances acting upon iodine are present, need hardly be mentioned. Thiosulphuric acid may, like sulphurous acid, be converted into sulphuric acid by means of chlorine or bromine water, and then determined. 4. IODIC ACID. lodic acid may be determined by the following easy method : Distil the free acid or iodate with an excess of pure fuming hydro- chloric acid, in the apparatus described in 130, 0, ft (chromic acid), receive the disengaged chlorine in solution of potassium iodide, and determine the separated iodine as directed in 130, I, 0, /?. The decomposition of iodic acid by hydrochloric acid is represented by the equation, H1O 3 + 5HC1 = Id + 4C1 + 3H 2 O. Since the 4 at. Cl set free 4 at. I, the amount of iodic acid or iodic anhydride can be calculated from the weight of the latter ; 1014-8 iodine cor- respond to 333-7 iodic anhydride (I 2 O 5 ) (BTJNSEN*). The following method also yields good results. Mix the solution with dilute sul- phuric acid, add potassium iodide in excess, and determine the amount of liberated iodine, after 146. One sixth of the iodine thus formed is derived from the iodic acid (HIO 9 + 5HI = 3H 2 O + I e ). See R,AMMELSBEKG.t * Annul, d. Chem. u. Pharm., LXXXVI, 285. \Pogg. Annal., cxxxv, 493; Zeitschr. f. analyt. Chem., vm, 456. 131.] NITROUS ACID. 433 5. NITKOUS ACID. The nitrous acid in nitrites which are free from nitrates may be estimated by converting the nitrogen into ammonia and deter- mining the latter, or by determining the oxidizing action on. ferrous salt. This method is conducted exactly as described under nitric acid ( 149). When nitric acid is also present, nitrous acid may be determined very satisfactorily with a solution of pure potassium permanganate, provided the fluid be sufficiently diluted to prevent the nitrous acid, which is liberated by the addition of a stronger acid, being decomposed by water with formation of nitric acid and nitric oxide. For 1 part of nitrous anhydride at least 5000 parts of water should be present. The decomposition is repre- sented by the following equation : 5HNO 2 + K 9 Mn,O 8 + 3II 2 SO 4 = 5HNO 3 +K 2 SO 4 + 2MnSO 4 + 3H 2 O. If the permanganate be standardized with iron, 4 at. iron correspond to 1 mol. 2s" a O,, since both of these require 2 at. oxygen. Nitrites are dissolved in very slightly acidulated water, the permanganate is added till the oxidation of the nitrous acid is nearly completed, the solution is then made strongly acid, and finally permanganate is added to light- red coloration. To determine nitrogen tetroxide N,O 4 in red fuming nitric acid, transfer a few c.c. to about 500 c.c. cold pure distilled water with stirring, and determine the nitrous acid produced. 1 mol. nitrous anhydride found corresponds to 2 mol. nitrogen tetroxide, for the latter when mixed with such a large quantity of water as is indi- cated above is decomposed in accordance with the following equa- tion : jSr 2 O 4 + H 2 O = HNO 3 + HM) 2 (Sio. FELDHAUS*). Nitrous acid and nitrogen tetroxide in presence of nitric acid may also be estimated by the reduction of chromic acid. An excess of standard potassium dichromate is added, and the unde- composed residue of chromic acid is estimated with standard solu- tion of ferrous salt (Fit. MoHBf). As regards the estimation of nitrous acid with lead dioxide, comp. FELDHAUS, loo. cit. p. 431, also LANG^: and J. LOWENTHAL. Regarding the estimation of nitrous acid in water, see 205. * Zeitsclir. f. analyt. Chem., i, 426. f Lehrbuch der Titrirmethode, 3. Aufl., 236. \Zeitschr.f. analyt. Chem., I, 485. Ib., in, 176. 434 DETEEMnSTATION. [ 132. Second Division of the First Group of the Acids. SULPHURIC ACID ; (Hydrofluosilicic Acid). 132. SULPHURIC ACID. I. DETERMINATION. Sulphuric acid is usually determined in the gravimetric way as barium sulphate. The acid may, however, be estimated also by the acidimetric method (2 15), and by certain volumetric methods, based upon the insolubility of the barium sulphate (and lead sul- phate). . 1. Gravimetric Method. The exact estimation of sulphuric acid as barium sulphate is by no means so simple and easy as it was formerly supposed to be, but requires, on the contrary, great care and attention. This arises from three causes ( 71, a): First, the barium sulphate is found to be far more soluble than was imagined in solutions of free acid& and of many salts ; secondly, it is extremely liable to carry down with it foreign salts, which are of themselves soluble in water' thirdly, when the precipitate has once separated in an impure state, it is often very difficult to purify it completely. The solution should contain but little free hydrochloric acid, and no nitric or chloric acid. If either of the two last are present, evaporate repeatedly, on the water-bath with pure hydrochloric acid. Dilute considerably, heat nearly to boiling, add barium chlo- ride in moderate excess, and allow to settle for a long time at a gentle heat. Decant the clear fluid through a filter, treat the pre- cipitate with boiling water, allow to settle, decant again, and so on, till the washings are free from chlorine. Finally transfer the pre- cipitate to the filter, dry and treat according to 53, using only a moderate red heat. After the precipitate has been weighed it is well to warm it for some time with dilute hydrochloric acid on the water-bath. Then pour off the hydrochloric acid through a small filter, wash the pre- cipitate by decantation with boiling water without removing it to the filter, evaporate the filtrate and washings nearly to dryness in a platinum or porcelain dish, add water, collect the minute amount 132.] SULPHURIC ACID. 435 of barium sulphate here left undissolved upon the small filter, wash, dry, incinerate, add the ash to the bulk of the precipitate, ignite again, and weigh. If the precipitate has lost weight, this shows that it at first contained foreign salts. This method of purification sometimes fails when the precipi- tate contains ferric oxide or platinum (GLAUS*), and it invariably fails when the solution contained any notable quantity of nitric acid.f In such cases there is only one resource, namely, to fuse with about four parts of sodium carbonate, warm with water, filter, wash with boiling water, acidify the filtrate slightly with hydro- chloric acid, and determine the sulphuric acid again. The results are thoroughly satisfactory if these directions are attended to ; if not, the result may be two or three per cent, too high or too low. 2. Volumetric Methods. a. After CARL MOHB.^ We require a normal solution of barium chloride, containing 122-166 grm. of the pure crystallized salt in 1 litre, and also normal nitric or hydrochloric acid and normal soda (215). Add to the fluid to be examined for sulphuric acid which, should it contain much free acid, is previ- ously to be nearly neutralized with pure sodium carbonate a meas- ured quantity of barium chloride solution, best a round number of cubic centimetres, in more than sufficient proportion to precipi- tate the sulphuric acid, but not in too great excess. Digest the mixture for some time in a warm place, then precipitate, without previous filtration, the excess of barium chloride with ammonium carbonate and a little ammonia, filter off the barium sulphate and carbonate, wash until the water running off acts no longer upon red litmus paper, and then determine the barium carbonate by the alkalimetric method given in 223. Deduct the c.c. of normal acid used from the c.c. of barium chloride, and the remainder will be the c.c. of barium chloride corresponding to the sulphuric acid present. The results of this method are quite satisfactory, if the solution does not contain too much free acid ; but in presence of a large excess of free acid, the action of the salt of ammonia will retain barium carbonate in solution, which, of course, will make * JaJiresber. von KOPP und WILL, 1861, 323, note. f Compare my paper in Zeitschr. f. analyt. Chem., ix, 52. \ Ann. der Chem. u. Pharm., xc, 165. 436 DETERMINATION. [ 132 the amount of sulphuric acid appear higher than is really the case. It need hardly be mentioned that this method is altogether inapplicable in presence of phosphoric acid, oxalic acid, or any other acid precipitating barium salt from neutral solutions, and that no basic radicals except the alkalies may be present. J. After AD. CLEMM.* In order to render C. MOHK'S method more expeditious, and hence better adapted for the use of manu- facturers, CLEMM has modified it. In it, also, the absence is required of all other acids which yield insoluble barium salts ; all bases except the alkalies must also be absent. In addition to the standard solutions mentioned under , there is also required a normal solution of pure sodium carbonate (53 '45 grm. anhydrous salt contained in 1 litre). Add a little litmus tincture to the solution contained in a measuring flask, and if necessary exactly neutralize with carbonate-free KaOH solution or hydrochloric acid. Add now a measured excess of barium-chloride solution to pre- cipitate all the sulphuric acid present, and then add a volume of normal sodium -carbonate solution equal to that of the barium- chloride solution used, nil with water to the mark, shake, filter, and in an aliquot portion of the filtrate (about one half) determine the sodium carbonate according to 220. The acid required to neutralize the residual sodium carbonate is, of course, the equiva- lent of the sulphuric acid present : K 2 SO 4 + 2BaCl 2 = BaSO 4 + 2KC1 + BaCl 2 ; and BaSO 4 + KC1 + BaCl a + 2]STa 2 CO 3 = BaSO 4 + KC1 + BaCO s + 2NaCl + Na 2 CO 3 . In dilute solution the slight excess of sodium carbonate has no action on. the barium sulphate, hence no error will arise on this score. The results are suffi- ciently accurate for technical purposes. c. After E. BoHLio.f This method, which is also adapted for technical purposes, depends upon the fact that the sulphates of alkalies are completely decomposed by precipitated barium car- bonate in the presence of an excess of carbonic acid and at 100, barium sulphate and an alkali bicarbonate being formed ; the heating prevents the solution of any notable quantity of barium carbonate because of the presence of free carbonic acid. The alkali which has combined with the carbonic acid corresponds * Zeitschr f. analyt. Chem. ix, 122. -ft., ix, 310. 132.] SULPIIUKIC ACID. 437 to the sulphuric acid originally present as sulphate. Regarding the details see the original paper. d. After R. WILDENSTEIN (first method *). The principle of this method depends upon precipitating the sulphuric with barium chloride and estimating the excess of barium chloride, using potassium chromate. If the solution is neutral, the chromate is added directly ; if acid, after previous addition of ammonia free from carbonate in slight excess. There are required: 1. Barium- chloride solution, 1 c. c. of which should correspond to 0*02 grm. sulphuric anhydride, SO,, and prepared by dissolving 61-03 grm. of pure crystallized barium chloride, BaCl a -(- 2H,O, to make one litre. 2. Potassium-chromate solution, of which 2 c. c. should precipitate 1 c. c. of the barium-chloride solu- tion. It is prepared by dissolving 18 '3853 grm. potassium bi- chromate in some water, adding ammonia until the reddish- yellow color has given place to a pale-yellow, and then diluting to measure 1 litre. The solution should be neutral. The two solutions must first be tested, to see whether they correspond properly. For this purpose dilute 10 c. c. of the barium-chloride solution with about 50 c. c. of water, boil, and add 20 '4 c. c. of the potassium-chromate solution. The precipi- tate rapidly subsides, and the supernatant liquid must be yellow- ish. On now adding barium-chloride solution by drops, exactly 0'2 c. c. of the solution must be required to effect complete decomposition in all, therefore, 10 '2 c. c. To carry out the sulphuric-acid determination, dissolve the substance in about 50 c. c. of water, heat to boiling in a 200 c. c. flask, and run in barium-chloride solution until perfectly certain that all the sul- phuric acid has been precipitated, yet avoiding too great an excess. Boil then for one-half to one minute, and if acid neutralize with ammonia free from carbonate, then add to the hot liquid, whether turbid or not, potassium-chromate solution in quantities of 0*5 c. c. The liquid rapidly clears up on gently agitating, so that the appearance of the yellow color, when the chromate begins to be present in excess, may be readily observed. When this point arrives, add barium -chloride solution slowly drop by drop until * Zeitschr.f. analyL Chem., i, 323. 438 DETERMINATION. [ 132. complete discolorization is just effected, for which purpose a few drops, and at most 0*4 c. c., are required. Half the number of c. c. of potassium-chromate solution used is deducted from the entire number of c. c. of barium -chloride solution used, and from the difference calculate the sulphuric acid. Results good. In applying this method to the sulphates of magnesium, zinc, or cadmium, dissolve the sulphate in ammonia with the addition of ammonium chloride, heat with a little calcium chloride in order to remove any carbonate that may be present, then add barium chloride, and finally the potassium chromate (FLEISCHER *). e. After R, WILDENSTEIN (second method f). Of all the methods for the volumetric estimation of sulphuric acid, the simplest, and that which is capable of the most general application, is to drop in to the solution containing excess of hydro- chloric acid, standard barium-chloride solution, till the exact point is reached when no more precipita- tion takes place. This point is difficult to hit, and ~ h hence the method has only found a very limited use. WILDENSTEIN has given this method a practical form, which renders it possible to complete an analysis in about half an hour, and at the same time to obtain satisfactory results. He employs the apparatus, Fig. 90. A is a bottle of white glass, the bottom of which has been removed; it holds 900 to 950 c. c. is a strong funnel-tube with bell-shaped funnel, and bent as shown, provided below with a piece of india-rubber tubing, a screw compression-cock, and a small piece of tubing not drawn out. The length from c to d is about 7J 8, from d to e about 12, cm. The opening of the funnel-tube /, which should have a diameter of 2 -5 to 3 cm., is covered as fol- lows: Take a piece of fine new calico or muslin free from sul- phuric acid and about 6 cm. square, lay on it two. pieces of Swedish paper of the same size and then another piece of stuff like the first, now bind these all together over the opening/ 1 , care- fully and without injuring the paper, by means of a strong linen * Journ. f. prakt. Chem., N. F. v, 318. Here also a modification is given by which the excess of ammonium chromate maybe detected in colored liquids, but unfortunately the process is far less simple. \Zeitschr.f. analyt. Chem., i, 432. 132.] SULPHURIC ACID. 439 thread which has been drawn a few times over wax, and cut it off even all round. We have now a small siphon -filter, which enables us to filter off a portion of fluid contained in A, and turbid from barium sulphate, clear and with comparative rapidity. On gradually adding barium chloride to the dilute acid solution of a sulphate a point occurs which may be compared with the neutral point in precipitating silver with sodium chloride (see 115, 5, b) ; i.e., there is a certain moment when a portion filtered off will give a turbidity both with sulphuric acid and barium chloride after the lapse of a few minutes. On this account we must either proceed on the principle recommended for the estimation of silver, i.e., dis- regarding the quantity of barium chloride in the solution, to stand- ardize it by adding it to a known amount of sulphate, till a pre- cipitate ceases to be formed ; or else we must and WILDENSTEIN recommends this latter course consider as the end-point of the reaction the point at which barium chloride ceases to produce a distinctly visible precipitation in the clear filtrate after a lapse of two minutes. The barium chloride solution is prepared so that 1 c.c. corre- sponds to 0-02 sulphuric anhydride, by making a solution contain- ing the requisite calculated and carefully weighed amount of the pure salt per litre. A solution of sulphuric acid containing 0'02 grm . SO 3 per c. c. may also be required. The process is as follows : First prepare the solution of the sulphate to be analyzed (using ;about 3 or 4 grin.), then fill A. with hot water, open the cock with the screw or by the aid of a glass rod, and wait till the syphon B is quite full of water. If the water runs down the tube c e with- out filling it entirely, close and open the cock a few times, and this inconvenience will be removed. (It is not allowable to suck at 0, or to fill the syphon with the wash-bottle at t the following is the only direct method which can be relied on : Evaporate on the water-bath to dryness and exhaust the residue with absolute alcohol ; determine the combined acid in the residue, and the free acid in the alcoholic extract, after mixing with water and evaporating off the alcohol. It has been said that the object may be obtained by the use of barium carbonate, which is supposed to throw down the free acid only, but this is erroneous, since alkali sulphates in aqueous solution are partially decomposed at the ordinary temperature by barium carbonate. In some cases the amount of free sulphuric acid present may be calculated after having determined the total amount of basic and acid radicals present. When no other free acid is present, free sulphuric acid may be determined by the acidimetric process. Supplement to the Second Division. 133. HYDROFLUOSILICTC ACID. If you have hydrofluosilicic acid in solution, add solution of potassium chloride, then a volume of strong alcohol equal to the fluid present, collect the precipitated potassium silicofluoride on a * This is best done in a porcelain crucible. \Gompt. Rend., LXXXVIII, 515; Zeitschr.f. analyt. Chem., iv, 219. 133.] HYDROFLUOSILICIC ACID. 443 weighed filter, and wash with a mixture of equal volumes of alco- hol and water. Dry the washed precipitate at 100, and weigh. Mix the alcoholic filtrate with hydrochloric acid, evaporate to dry- ness, and treat the residue with hydrochloric acid and water. If this leaves an undissolved residue of silicic acid, this is a sign that die examined acid contained an excess of silicic acid ; the weight of the residue shows the amount of excess. Potassium silicofluor- ido dried at 100 has the formula (KF) 2 Si F 4 ; for its properties, scr 68. Instead of weighing it, it may be estimated volumetric- ally according to 97, 5. The analysis of metallic silicofluorides is best effected by heating in platinum vessels, with concentrated sul- phuric acid ; silicon fluoride and hydrofluoric acid volatilize, the basic metals are left behind in the form of sulphates, and may, in many cases, after volatilization of the excess of sulphuric acid, be weighed .as such. If the metallic silicofluorides to be analyzed contain water, the latter cannot be estimated by mere ignition, since silicon fluoride would escape with the water. H. ROSE recommends the following method : Mix them most intimately with parts of recently ignited lead oxide, cover the mixture in a small retort, with a layer of pure lead oxide, weigh the retort, heat cautiously until the contents begin to fuse together, remove the aqueous vapor still remaining in the vessel by suction, and weigh the retort again when cold. The diminution of weight shows the quantity of water expelled. Do not neglect testing the drops of th& escaping water with litmus paper; the result is ^^urate only if they have no acid reaction. F. STOLE A* proposes the following process, at least for com- pounds soluble in water : Put into a crucible double as much mag- nesia as is necessary to decompose the silicofluoride to be analyzed, ignite it as strongly as possible, allow to cool, and weigh. Add water to form a thick paste, and then the weighed silicofluoride ; if the amount of water present is not enough to dissolve the compound, add some more, mix with a platinum wire which must afterwards be wiped off clean, dry, ignite, and weigh. The increase in weight shows the amount of anhydrous silicofluoride,*provided no oxide is present which takes up oxygen. * ZeitscJir. f. analyt. Chem., vn, 93. 444 DETEBMINATION. [ 134. Third Division of the First Group of the Adds. PHOSPHORIC ACID BORIC ACID OXALIC ACID HYDROFLUORIC ACID. 134 1. PHOSPHORIC ACID. I. DETERMINATION. Orthophosphoric acid may be determined in a great variety of ways. The forms in which this determination may be effected have been given already in 93, 4. The most appropriate forms for the purpose, however, are magnesium pyrophosphate and ura- nyl pyrophosphate. The determination as magnesium pyrophos- phate is frequently preceded by precipitation in another way, especially as ammonium phospho-molybdate, occasionally as stannic phosphate or mercurous phosphate. The other forms in which phosphoric acid may be determined give also, in part, very good results, but admit only of a more limited application. "With respect to volumetric methods, those which depend upon the use of standard solution of uranium are the best. With regard to meta- and pyrophosphoric acids, I have simply to remark here that these acids cannot be determined by any of the methods given below. The best way to effect their determination is to convert them into orthophosporic acid, as follows : a. In the dry way. By protracted fusion with from 4 to 6 parts of mixed sodium and potassium carbonates. This method is, however, applicable only in the case of alkali meta- and pyrophos- phates, and of those metallic mata- or pyrophosphates which are completely decomposed by fusion with alkali carbonates ; it fails, accordingly, for instance, with the salts of the alkali-earth metals, magnesium excepted. ft. In the wet way. The salt is heated for some time with a strong acid, best with concentrated sulphuric acid (WEBER*). This method leads only to^the attainment of approximate results, in the case of all salts whose basic radicals form soluble salts of the acid added, since in these cases the meta- or pyrophosphoric acid is never completely liberated ; but the desired result may be fully attained by the use of any acid which forms insoluble salts compounds the basic radicals present. Respecting the partial conversion * Pogg. Annal., LXXIII, 137. 134.] PHOSPHORIC ACID. 445 in the former case, I have found that it approaches the nearer to completeness the greater the quantity of free acid added,* and that the ebullition must be long continued. Compare Expt. "No. 32. BUNCE'S statement, f that phosphoric acid volatilizes when a phosphate is evaporated to dryness with hydrochloric or nitric acid and the residue heated a little, is quite erroneous (compare my paper on the subject, in A.nnal. der Chem. und Pharm., LXXXVI, 216). But, on the other hand, it must be borne in mind that orthophosphoric acid under these circumstances changes, not indeed ut 100, but at a temperature still below 150, to pyrophosphoric acid ; thus, for instance, upon evaporating common hydrogen sodium phosphate with hydrochloric acid in excess, and drying the residue at 150, we obtain 2NaCl + Na a H 3 P 2 O, . a. Determination as Lead Phosphate. Proceed as with arsenic acid, 127, 1, a i.e., evaporate with a weighed quantity of oxide of lead, and ignite. This method pre- supposes that no other acid is present in the aqueous or nitric acid solution ; it has this great advantage, that it gives correct results, no matter whether ortho-, meta-, or pyrophosphoric acid is present. 1). Determination as Magnesium Pyrophosphate. OL. Direct determination. Suitable in all cases in which it is quite certain that the acid present is orthophosphoric, either free or combined as an alkali phosphate. The solution should be neutral, or only moderately ammoniacal. Add ammonium chloride, and then the usual magnesia mixture ( 62, 6), or still better, a mixture of magnesium chloride, am- monium chloride, and ammonia,;}: in sufficient but not too ex- cessive quantity (see 62, 6); 10 c.c. of the mixture will pre- cipitate 0*24 grm. P a O 6 . The precipitate being under these conditions somewhat slowly formed, appears distinctly crystalline. * There are, however, other considerations which forbid going too far in this respect. t Sillim. Journ., May, 1851, 405. | This mixture deserves the preference because it can be employed with greater certainty than the sulphate solution (as when the latter is not quite cor- rectly used the precipitate is apt to contain some basic magnesium sulphate) and gives accurate results. It is made as follows : Dissolve 83 grm. crystal- lized magnesium sulphate in boiling water, add 5 c.c. hydrochloric acid, then an aqueous solution of 82 grm. crystallized barium chloride, boil, decant, filter, and test with sulphuric acid (no precipitate should form). Mix the fil- trate and washings, concentrate, cool, transfer to a litre flask, add 165 grm. pure ammonium chloride, 260 c.c. ammonia, and water to 1 litre. Set aside for a few days and filter if necessary. This solution contains the same quan- tity of magnesia as the mixture given under 62, 6, 446 DETERMINATION. [ 134, After some time add ammonia gradually to the amount of one- third of the fluid. Allow to stand 12 hours in a well-covered vessel in the cold, filter, test the filtrate with magnesia mixture and ammonia, and wash the precipitate with ammonia diluted with 3 volumes of water till the washings, when acidified with nitric acid and tested with silver nitrate, are no longer rendered turbid; proceed according to 104, 2. The precipitate is not abso- lutely insoluble in ammoniated water, therefore it is Well to wasli by suction, as this reduces the necessary amount of wash- water to a minimum. The results are accurate (Expt. JSTo. 80 ; also KISSEL*). If there is reason to suspect the purity of the precipitate, dissolve it in hydrochloric acid and throw down again with ammonia, adding some magnesia mixture. If the magnesia mixture is omitted, the solution, being free from magnesia, will dissolve some of the precipitate. Compare KISSEL, loc. cit. For properties of the precipitate and residue, see 74. If the solution contains pyrophosphoric acid, the precipitate is floccnlent and dissolves to a notable degree in ammoniated water (WEBER). /?. Indirect determination, with previous precipitation as ammo- nium pJiospliomolybdate, after SoNNENSCHEHsr.f Applicable in all cases in which the phosphoric acid present is orthophosphoric, even in presence of salts of the alkali-earth metals, aluminium, ferric iron, &c. Tartaric acid, however, and similarly acting organic substances must be absent. No considerable quan- tity of free hydrochloric acid may be present. Large quantities of ammonium chloride, and of metallic chlorides generally, also of certain ammonium salts, especially the oxalate and citrate (KONIG) J, are to be avoided. Ammonium nitrate assists the precipitation and neutralizes the injurious action of very large quantities of nitrates and sulphates (E. RICHTEKS). The molybdenum solution described " Qual. Anal.," 52, is employed as the precipitant. It con- tains 5 per cent, of molybdic acid. The fiuid to be examined for phosphoric acid should be concentrated ; it may contain free nitric or free sulphuric acid. Transfer to a beaker and add a considerable quantity of the molybdenum solu- tion. About 40 parts molybdic acid must be added for every 1 part phosphoric anhydride, therefore 80 c.c. of the molybdic solution for 0*1 grm. Stir without touching the sides, and keep covered 12 hours at about 40; then remove a portion of the clear supernatant fluid with a pipette, mix *Zeitschr.f. analyt. Chem., vni, 170. -\Journ.f. prakt. Chem., LIU, 343. IZeitschr.f, analyt, Chem., x, 305. %lb. x, 469. 134.] PHOSPHORIC ACID. 447 it with an equal volume of molybdenum solution, and allow it to stand some time at 40. If a further precipitation takes place, return the portion to the main quantity, add more molybdenum solution, allow to stand again 12 hours, and test again. "When complete precipitation has been effected pour the fluid off through a small filter and wash the precipitate entirely by decantation, using a mixture of 100 parts molybdate solution, 20 parts nitric acid of 1*2 sp. gr., and 80 parts water.* The washing must be thorough, and the last runnings must not be precipitated by excess of ammonia, even if lime, iron, &c., was present in the solution. Now dissolve the precipitate in the least quantity of ammonia, pour the fluid through the small filter, when the minute amount of precipitate thereon will be dissolved, wash the filter with ammonia diluted with three volumes of water, mix the filtrate and washings, and add hydrochloric acid carefully till the precipitate produced, instead of redissolving instantly, takes a little time to disappear; finally throw down with magnesia mixture (compare a). If the ammonia leaves a small amount of the precipitate undissolved, treat the residue with nitric acid and test the filtrate with molybdic solution in order to save any phosphoric acid. Results accurate.f As this method requires so large a quantity of molybdic acid, it is usually resorted to only in cases where methods >, a, and c are inapplicable ; and the phosphoric acid in the quantity of substance taken is not allowed to exceed 0*3 grm. Arsenic acid and silicic acid,;): if present, must first be removed. Of all the methods for determining phosphoric acid which are admissible in the presence of ferric and aluminium salts, this is the best in my opinion, espe- cially for the estimation of small quantities of the acid in presence of large quantities of these salts. * According to E. RICHTERS (Zeitschr. f. anatyt. Chem., x, 471) you may also wash with a solution of ammonium nitrate containing 15 grm. in 100 c.c. slightly acidified with nitric acid and containing a few per cents of molybdic-acid solution. f Zeitschr. f. analyt. Chem., m, 446, and vi, 403. \ Silicic acid may also be thrown down, in form of a yellow precipitate, by acid solution of ammonium molybdate, especially in presence of much ammo- nium chloride (W. KNOP, Chem. CentralbL, 1857, 691). Mr. GRUNDMANN, who repeated KNOP'S experiments in my laboratory, obtained the same results. The precipitate dissolves in ammonia. If the solution, after addition of some ammo- nium chloride, is allowed to stand for some time, the silicic acid separates, and the phosphoric acid may then be precipitated from the filtrate with magnesia mixture; it is, however, always the safer way to remove silicic acid first. 448 DETERMINATION. [ 134 y. Indirect determination, with previous precipitation as mer- curous phosphate, after II. ROSE.* Applicable for the separation of phosphoric acid (also of pyro- and metaphosphoric acid) from all basic radicals, except aluminium. Comp. 135, k. Dissolve the phosphate in neither too large nor too small a quantity of nitric acid, in a porcelain dish, add pure metallic mer- cury in sufficient quantity to leave a portion, even though only a small one, undissolved by the free acid. Evaporate on the water- bath to dryness. If the warm mass still evolves an odor of nitric acid, moisten it with water, and heat again on the water-bath, until it smells no longer of nitric acid. Add now hot water, pass through a small filter, and wash until the washings leave no longer a fixed residue upon platinum. Dry the filter, which, besides mercurous phosphate, contains also basic mercurous nitrate and free mercury, mix its contents, in a platinum crucible, with mixed sodium and potassium carbonates in excess, roll the filter into the shape of a ball, place it in a hollow made in the mixture, and cover the whole with a layer of the mixed carbonates. Expose the crucible, under a chimney with good draught, for about half an hour to a moderate heat, so that it does not get red-hot. At this temperature, the mercurous nitrate and the metallic mercury volatilize. Heat now over the lamp to bright redness, and treat the residue with hot water, which will dissolve it completely, if no ferric oxide be present, and if no oxide of platinum has been formed. The latter may occur on account of too rapid heating, which might produce sodium nitrate, which would act upon the platinum. Supersatu- rate the clear (if necessary, filtered) solution with hydrochloric acid, add ammonia and magnesia mixture, and proceed as in or.f #. Indirect determination, with previous precipitation as stannic phosphate. aa. After W. KEISSIG.^: Dissolve the substance, which must * Fogg. Annal.,"LXXVi, 218. f If ferric oxide is present to any extent, it may easily retain some phos- phoric acid. (Compare 135, g, a.) ROSE'S modification, to be employed when alumina is present, is given in 135, K, y. %Annal. d. Chem. u. Pharm., xcvm, 339, The method is a serviceable modification of REYNOSO'S process (Journ. f. prakt. Chem., LIV, 261), which, though free from errors in principle, yet presents certain difficulties, particularly the fact that the slight impurities in the tin are a source of considerable error, since at least 8 times as much tin must be taken as the quantity of phosphoric acid present. These statements of Reissig fully coincide with my own investi- gations- 134.] PHOSPHORIC ACID. 449 l>e free from chlorides, in concentrated nitric acid, add at least eight times as much of the purest tin (in the form of foil or grains) as there is phosphoric acid present, and warm for 5 to 6 hours, until the precipitate has completely subsided. (If alumina or ferric oxide is present, some portion of these is carried down in the precipitate. GIKARD.) "Wash by decantation and filtration, rinse into a platinum dish, and digest with a small quantity of very concentrated potassa solution. These form both potassium in etas tan n ate and phosphate, which, on adding hot water, yield a clear solution, dissolving with great readiness if not too much potassa has been used. Any traces of precipitate re- maining adhering to the filter are similarly dissolved. The total alkaline fluid is transferred to a tared litre flask, diluted to weigh 900 grin, and saturated with hydrogen sulphide; some ammonium pentasulphide and acetic acid are then added until the tin sulphide is precipitated and the liquid is weakly acid. Now add water to make the contents weigh 1000 grm. again (or some other round number), shake, let stand 12 to 16 hours, filter off the superna- tant clear liquid into a porcelain dish, and again weigh the flask, which now contains the rest of the fluid and the tin sulphide. The difference between the two weights gives the weight of the filtrate in which the phosphoric acid is to be determined. The proportion which this weight bears to the total quantity of fluid, i.e., 1000 grm. minus the weight of the tin sulphide (which may be calculated with sufficient accuracy from the quantity of tin originally taken, or which may be directly estimated), is readily found. Evaporate to a small volume the filtrate mixed with the wash- ings of the filter, and in this determine the phosphoric acid accord- ing to 5, OL. The above method of separating the phosphoric-acid liquid from the tin sulphide must be resorted to, because in filtering off and washing the sulphide small quantities of tin are dissolved out, no matter whether pure water or hydrogen-sulphide water is used. Results accurate. 55. After GIRARD.* In order to render the method, origi- nally based on the separation of phosphoric acid as stannic plms- * Compt. Rend., LIV, 468; Zeitschr.f. analyt. Chem., i, 366. 450 DETERMINATION. [ 134. phate, applicable also in the presence of alumina and ferric oxide, GIRARD employs the following method : Dissolve the substance in highly concentrated nitric acid, remove all chlorine either by pre- cipitation with silver nitrate or by repeated evaporation with nitric acid, add 8 times as much tinfoil as there is phosphoric acid present, and warm the mixture 5 or 6 hours, until the precipitate has completely subsided, leaving the super- natant fluid clear. Wash with hot water by decantation and finally by filtration. The precipitate consists of meta- stannic acid and stannic phosphate, together with a little ferric and aluminium phosphates. Heat it either at first with a small quantity of aqua regia, and then with ammonia and ammonium sulphide, or immediately with ammonium sulphide in excess. The latter process is recommended by O. BABER,* on the ground that the former leaves a little phosphoric acid in the precipitate. The whole is digested about two hours and then filtered ; the precipi- tate, consisting of ferrous sulphide and aluminium hydroxide, is washed with warm ammonium sulphide, then with water contain- ing a little ammonium sulphide dissolved in nitric acid, and the solution thus formed mixed with the filtrate from the tin precipi- tate which contains the principal quantity of the basic metals. From the ammonium-sulphide filtrate, which contains stannic sul- phide and ammonium phosphate, the phosphoric acid is at once precipitated by magnesia mixture. I may add that GIRARD con- siders 4 to 5 parts tin sufficient for 1 part P 2 O 6 . The results afforded by his test analyses are unexceptionable. According to jANovsKY,f at least 6 parts of tin must be used. The tin should be free from arsenic. If the tin contains arsenic, direct precipi- tation with magnesia mixture will give, besides ammonium -mag- nesium phosphate, some ammonium-magnesium arsenate also, and consequently the results would be too high. In such a case the ammonium-sulphide solution is best treated as in aa. e. Indirect determination after previous precipitation as bis- muthic phosphate, This method was proposed by CHANCEL J and modified by * Zeitechr.f. die gesammten Naturwissensch., 1864, 293. f ZeiUchr, f. analyt. Chem., xi, 157. j Compt. Rend., L, 416; Chem. CentralbL, 1860, 212 ; Compt. Rend., LI, 882 ; Chem. CentralbL, 1861, 221. PHOSPHORIC ACID. 451 BIRNBAUM and CHOJNAKI.* It is not applicable in the pres- ence of sulphuric or hydrochloric acid, nor can it lay claim to rapidity or accuracy. f c. Determination as Uranyl Pyrophosphate. After LECONTE, A. AKENDT, and "W. KNOP.^ (Very suitable in presence of alkali and alkali-earth metals, but not in pres- ence of any notable amount of aluminium ; in presence of ferric iron, the method can be applied only with certain modifications.) Where it is possible, prepare an acetic-acid solution of the com- pound. If you have a nitric- or hydrochloric-acid solution, re- move the greater portion of the free acid by evaporation, add ammonia until red litmus paper dipped into it turns very dis- tinctly blue, and then redissolve the preciptate formed in acetic acid. If mineral acids were present, add also some ammonium acetate ; this addition is beneficial under any circumstances. Mix the fluid now with solution of uranyl acetate and heat the mix- ture to boiling, which will cause the phosphoric acid to separate in form of pale greenish-yellow ammonium uranyl phosphate. Wash the precipitate, first by decantation, boiling up each time, then by filtration ; the operation may be materially facili- tated by adding a few per cents of ammonium nitrate to the water. Dry the precipitate and ignite as directed in 53. It is advisable to evaporate small quantities of nitric acid on the ignited precipitate repeatedly and to reignite. The residue must have the color of the yolk of an egg. For the properties of the pre- cipitate and- residue, see 93, 4, e. Should it be necessary to- dissolve the ignited residue again, for the purpose of reprecipitat- ing it, this can be done only after fusing it with a large excess of * Zeilsckr.f. analyt, Chem., ix, 203. f Compare HOLZBERGER, Archiv. der Pharm. , (2), CXVT, 37 ; BABER, Zeitschr. f. d. ges. Naturwiss., 1864, 293 ; GIRARD, Compt. rend., LIV, 468; FRESENIUS, NEUBAUER, and LUCK, Zeitschr. f. analyt. Chem., x, 135 ; ADRI- AANSZ, ib., x, 473. \ LECONTE was the first to recommend the method of precipitating phos- phoric acid from acetic-acid solutions by means of a salt of uranium (Jahresb. von LIEBIG und KOPP, fUr 1853, 642); A. ARENDT and W. KNOP have subse- quently subjected it to a careful and searching examination (Chem. Centralbl. 1850, 769, 803 ; and 1857, 177). Chem. Centralbl., 1857, 182. DETERMINATION. [ 131 mixed sodium and potassium carbonates, and thereby converting the pyrophosphoric into orthophosphoric aci(L . Results accurate. Compare the test analyses given by the authors, Expt. !N"o. 81, and KISSEL'S experiments.* d. Determination as Basic Ferric Phosphate. a. Mix the acid fluid containing the phosphoric acid with an excess of solution of ferric chloride of known strength, add, if necessary, sufficient ammonia to neutralize the greater portion of the free acid, mix with ammonium acetate in not too large ex- cess, and boil. If the quantity of solution of ferric chloride added was sufficient, the precipitate must be brownish-red. This precipitate consists of basic-ferric phosphate and basic-ferric ace- tate, and contains the whole of the phosphoric acid and of the ferric iron. Filter off boiling, wash with boiling water mixed with some ammonium acetate, and dry carefully. [Detach the greater part of the precipitate from the filter, incinerate the filter, trans- fer to the crucible the main part of the precipitate, moisten with strong nitric acid, dry, moisten again with nitric acid, and dry and ignite in a platinum crucible with access of air ( 53). With- out these precautions reduction of ferric oxide to magnetic oxide is liable to occur.] If the weight of the residue has been in- creased by this operation, which is not the case, however, as a rule, the procedure must be repeated until the weight is constant. Deduct from the weight of the residue that ferric oxide produced from the solution added; the difference is the P 2 O B . [This modification of SCHULZE'S method was first recommended by A. MuLLER;f it has been adopted also by WAY and OGSTON, in their analyses of ashes. ^ MULLER'S improvement consists in the use of a solution of ferric chloride of known strength, whereby the determination of iron in the residue (as given in 113, 3) is dispensed with.] /3. J. WEEREN'S method, suitable for the estimation of the phos- phoric acid in phosphates of the alkali and alkali-earth metals. * Zeitschr f. analyt. Chem., vm, 167. f Journ. f. prakt. Chem., XLVII, 341. \ Journal of the Royal Agricultural Society, viu, part I. % Journ. f. prakt. Chem., LXVII, 8. 3 134.] PHOSPHORIC ACID. 453 Mix the nitric acid solution of the phosphate under examination, which must contain no other strong acid, with a solution of ferric nitrate, of known strength, in sufficient proportion to insure the formation of a basic salt (3 or 4 parts of iron should be present for 1 part P 2 O 6 ) ; evaporate to dryness, heat the residue to 160, until no more nitric acid fumes escape, treat with hot water containing ammonium nitrate until all nitrates of the alkali and alkali-earth metals are removed, collect the yellow-ochreous precipitate on a filter, dry, ignite (see 53), weigh, and deduct from the weight the quantity of iron added reckoned as ferric oxide. LATSCHINOW* recommends heating the residue to 200, warming with water and a few drops of sulphuric acid, adding ammonia and then treating with hot solution of ammonium nitrate. He says that the phos- phoric acid is thus more completely separated, and the precipitate may be more readily filtered off. e. Determination as Normal Magnesium Phosphate Mg 3 (Fu. SCHULZE'S method, suitable more particularly to effect the separation of phosphoric acid from the alkalies, f) Mix the solution of the alkali phosphate, which contains ammo- nium chloride, with a weighed excess of pure magnesium oxide, evaporate to dryness, ignite the residue until the ammonium chlo- ride is expelled, and separate the magnesium, which is still present in form of magnesium chloride, by means of mercuric oxide (104. 3, 5). Treat the ignited residue with water, filter the solution of the chlorides of the alkali metals, wash the precipitate, dry, ignite, and weigh. The excess of weight over that of the magnesium oxide used shows the quantity of the P 2 O 6 . Results satisfactory. f. SCHLOSING'S method J does not appear to offer any advan- tages. The phosphate is mixed with silica and ignited in carbon monoxide, the expelled phosphorus being taken up by copper or by silver nitrate. g. Determination by Volumetric Analysis ( With Uraniutn Solution). This method was recommended originally by LECONTE. It *Zeitschr.f. analyt. Chem., vn, 213. \Journ.f. prakt. Cliem., LXIII, 440. %Zeitschr.f. analyt. Chem., iv, 118 and vn, 473. Jahresber. von LIEBIG u. KOPP, ftir 1853, 642. 454 DETERMINATION. [ Ic54. was improved and described in detail by NEUBAUER,* and was afterwards recommended by PINCTTS, f and subsequently by BODEKER.^: The principle of the method is as follows; Uranyl acetate precipitates from solutions rendered acid by acetic acid, hydrogen uranyl phosphate, or in the presence of considerable quantities of ammonium salts ammonium uranyl phosphate. The proportion between the uranium and the phosphoric acid is the same in both compounds. Both compounds when freshly precipi- (tated and suspended in water are left unchanged by potassium ferrocyanide ; uranyl acetate, on the other hand, is indicated by this reagent with great delicacy by the formation of an insoluble reddish-brown precipitate. According to NEUBAUER the following solutions are employed : a. A solution of phosphoric acid of 'known strength. Pre- pared by dissolving 10-13 grm. pure, crystallized, unehMoresced, powdered, and pressed hydrogen sodium phosphate in water to 1 litre. 50 c.c. contain 0*1 grm. P Q O B . It is well to control this solu- tion by evaporating 50 c.c. in a weighed platinum dish to dry ness, igniting strongly, and weighing. The weight should be 0'1881 grm. I. An acid solution of sodium acetate. Prepared by dissolv- ing 100 grm. sodium acetate in 900 c. c. water and adding acetic acid of 1*04 sp. gr. to 1 litre. c. A solution of uranyl acetate (or nitrate) ( 63, 3). This is standardized against the hydrogen sodium phosphate solution. 1 c. c indicates 0-005 grm. P 2 O 5 . The solution is made at first a little stronger than necessary, so that it may contain in the litre, say, 32-5 grm. UO,(C 2 H 3 O 2 ) 2 +2H 4 O or 34 grm. UO 2 (C 2 H 3 O 2 ),+3H 2 O (corresponding to 22 grm. UO a O); its value is then determined, and it is diluted accordingly. To determine its value proceed as follows : Transfer 50 c.c. of the a solution to a beaker, add 5 c.c. of the 1) solution, and heat in a water-bath to 90 100. Now run in uranium solution, at first a large quantity, at last in i c.c. , testing after each addition whether the precipitation is finished or not. For this purpose spread out one or two drops of the mixture on a ^white porcelain surface and introduce into the middle, by means of a thin glass rod, a small drop of freshly prepared potassium ferrocya- .nide solution or a little of the powdered salt. As soon as a trace of -*Archiv. fur wissemchaftliche Heilkunde, IV, 228. \Journ.f. prakt. Chem., LXXVI, 104. J Anal, de Ghem. etPharm., cxvn, 195. |His Anleitung zur Harnanaly se, 6. Aufl., 171. 134 ] PHOSPHORIC ACID. 455 -excess of uranyl acetate is present, a reddisli-brown spot forms in the drop, which, surrounded as it is by the colorless or almost colorless fluid, may be very distinctly perceived. When the final reaction has just appeared, heat a few minutes in the water-bath and repeat the testing on the porcelain. If now the reaction is .still plain the experiment is concluded. If the uranium solution had been exactly of the required strength, 20 c.c. would have been used ; but it is actually too concentrated, hence less than 20 c.c. must have been used. Suppose it was 18 c.c., then the solution will be right, if for every 18 c.c. we add 2 c.c. of water. If in this first experiment we find that the solution is much too strong, the solution is diluted with somewhat less water than is properly speak- ing required, another experiment is made, and it is then diluted exactly. The actual analysis must be made under as nearly as possible similar circumstances to those under which the standardizing: of the o uranium solution was performed, especially as regards the sodium acetate. This salt retards the precipitation of uranium by potas' sium ferrocyanide, hence the test-drop on the porcelain plate becomes darker and darker. The analyst should accustom himself to observing the first appearance of the slightest brownish colora- tion in the middle of the drop, and should take this as the end- reaction. It need hardly be added that the same person must make the analysis who has standardized the solution (NEUBAUER). The method is applicable to free phosphoric acid, alkali phos- phates, and magnesium phosphate, also in the presence of small quantities of the phosphates of other alkali-earth metals, but can- not be employed in presence of ferric and aluminium salts. Dis- solve the substance in water or the least possible quantity of acetic acid, add 5 c.c. of the b solution, dilute to 50 c.c., and proceed with the addition of uranium as above. The results are very satisfac- tory. Compare KISSEL'S experiments.* If the above process is followed in the presence of much calcium, for instance with a solu- tion of calcium phosphate in dilute acetic acid, the results are almost always too low, as little calcium phosphate is precipitated along with uranyl phosphate. [The best means of obviating that error is, according to ABESSEK, JANI, and MARCKER,f to standardize the uranium solution under the same conditions as nearly as possible * Zcitschr. f. analyt. Chem., vm, 167. f Ib., xn, 262. 456 DETERMINATION. [ 134, as exist when the solution is used for the actual determination of phosphoric acid. It must therefore be standardized with calcium phosphate. Prepare a solution of suitable strength by dissolving- pure Ca 3 (PO 4 ) 2 in the smallest possible quantity of nitric acid and diluting to the desired volume. Determine accurately the amount of Ca 3 (PO 4 ) 2 in this solution by evaporating to dryness in a plati- num vessel 50 c. c., moistening the residue with ammonia and igniting. The residual somewhat hygroscopic calcium phosphate is quickly weighed in the covered platinum vessel.] According to R. FRESENITJS, NEUBAUER, and LUCK,* the dif- ficulty may also be easily avoided by adding the phosphate solu- tion to the uranium solution until the ferrocyanide ceases to give a reaction. The standardization of the uranium solution should in this case be conducted as follows : To 25 c. c. of the uranium solution in a beaker add 5 c. c. of the sodium- acetate solution and 3 c. c. of acetic acid (sp. gr. 1*04), heat on a water-bath and run in from a burette sodium-phosphate solution until a drop brought into contact with a few frag- ments of potassium ferrocyanide on a porcelain plate just ceases to react. After every addition of phosphate the beaker must be replaced in the hot water, and a few minutes must be allowed to elapse before again testing; the sodium-phosphate solution may further be added freely so long as the solution remains yellowish. In carrying out the analysis, take care that the solution of calcium acetate contains no considerable excess of free acetic acid, that its concentration does not differ very greatly from that of the sodium-phosphate solution, and that its total volume is known before any part of it is introduced into the burette. Regarding the volumetric methods proposed by FLEISCHER f (alumina method) and SCHWARZ (lead method), see the sources given. The latter method, although based on correct principles and sufficiently exact for neutral liquids, is nevertheless of very limited application, since the presence of acetic acid seriously impairs its accuracy. See FK. MOHR. * ZeitscJir. f. analyt. Chem., x, 147. }lb., iv, 19, and vi, 28. J Ding. Polyt. Journ., CLXIX, 289 ; Zeitschr. f. analyt. Chern., u, 392. J.b., n, 256. f?135.] PHOSPHORIC AOID. 457 II. SEPARATION OF PHOSPHORIC ACID FROM THE BASIC RADICALS. 135. a. from the Alkalies (see also d, &, and Z). a. Add ammonium chloride or hydrochloric acid, then lead acetate, exactly, till no more precipitate is produced, and lastly some pure lead carbonate (prepared by precipitating lead acetate with ammonium carbonate, BABER *), allow to digest for some time, filter off the precipitate consisting of lead phosphate, chlo- ride, and carbonate, wash, precipitate from the filtrate the slight excess of lead by hydrogen sulphide, filter, and evaporate with hydrochloric acid (in the case of lithium, sulphuric acid). If the phosphoric acid is to be estimated in the same portion, proceed with the first precipitate (after washing, to remove the larger quantity of chloride) according to 135, b. /3. (Only applicable in the case of fixed alkalies.) Separate the phosphoric acid as ferric phosphate, according to one of the methods given in 134, d. Or, if you do not wish to determine the phosphoric acid it is very convenient to acidify with hydrochloric acid, add ferric chloride, dilute rather considerably, add ammonia till the fluid is neutral, and boil ; all the phosphoric acid will then separate with ferric oxychloride as ferric phosphate. This modi- fication is recommended when the phosphoric acid is to be pre- cipitated, but not estimated. The separation of phosphoric acid may also be effected as magnesium phosphate ( 134, e). The alkalies are contained in the filtrate as nitrates or chlorides. 1). From Barium, Strontium, Calcium, and Lead. The compound under examination is dissolved in hydrochloric or nitric acid and the solution precipitated with sulphuric acid in slight excess. In the separation of phosphoric acid from strontium, calcium, and lead, alcohol is added with the sulphuric acid. The phosphoric acid in the filtrate is determined according to 134, J, a, after removal of the alcohol by evaporation. The determination of the phosphoric acid is effected most accurately by saturating the fluid with sodium carbonate, evaporating to dryness, and fusing the residue with sodium and potassium carbonates. The fused mass is then dissolved in water, and the further process conducted as in 134, &, a. *Zeitschr.f. die ges. Naturwiss., 1864, 298; Zeitschr.f. analyt. Chem., iv, 120- 458 DETERMINATION. [ 135. c. From Magnesium (see also d, h, &, Z). Add ferric chloride in sufficient excess, dilute, add excess of barium carbonate, allow to remain for several hours with frequent stirring, filter and separate magnesium and barium in the filtrate after 154. d. From the whole of the Alkali-earth Metals and fixed Alka- lies (comp. A, &, Z). a. Dissolve in the least possible quantity of nitric acid, add a little ammonium chloride, precipitate exactly with lead acetate, add a little lead carbonate (precipitated), digest, filter, precipitate the excess of lead rapidly from the filtrate by hydrogen sulphide, filter and determine the basic metals in the filtrate. Results good. (3. Dissolve in water, and in case of phosphates of the alkali- earth metals the least possible nitric acid, add neutral silver nitrate and then silver carbonate, till the fluid reacts neutral. All phos- phoric acid now separates as Ag 3 PO 4 . Warming is unnecessary. Filter, wash the precipitate, dissolve it in dilute nitric acid, precipi- tate the silver with hydrochloric acid, and determine the phosphoric acid in the filtrate according to 134, 5, a. The filtrate from the silver phosphate is freed from silver by hydrochloric acid, and the basic metals are then determined according to the methods already given (G. CHANCEL*). A good and convenient method unless the proportion of alkali is very large. (If the substance contains alu- minium or ferric iron, they are completely precipitated by the silver carbonate, and are found with the silver phosphate.) y. Separate the phosphoric acid as uranyl phosphate ( 134, c\ and the excess of uranium from the alkali-earth metals, &c., in the filtrate, according to 160 and 161, Supplement. Results good. d. Separate the phosphoric acid according to 134, d, a or ft. The alkali-earth metals are obtained in solution in the first case, as chlorides, together with alkali acetate and chloride; in the second case as nitrates. Results good. e. From Aluminium. The best method of separating phosphoric acid from aluminium is that depending on precipitation by ammonium inolybdate ( 135, Z). The separation of the acid as stannic phosphate (A, a) is also satisfactory. * Compt. rend., XLIX, 997; Journ. /. prakt. Chem., LXXIX, 222. 135.] PHOSPHORIC ACID. 459 The older methods are scarcely ever used now ; hence the two most in use formerly will be but briefly described. a. OTTO'S method. This depends on the precipitation of phosphoric acid with magnesia mixture from the solution to which tartaric acid and ammonia are added. It is difficult to obtain a precipitate free from alumina, even after repeated pre- cipitation ; on the other hand, a certain quantity of phosphoric acid remains in solution. Compare ILYREN,* F. KNAPp,f and K. PRIBRAM.;); ft. BERZELIUS'S method. Mix the finely powdered substance with about IL^ parts of pure silicic acid (the artificially prepared is the best) and 6 parts sodium carbonate, and expose to a strong red heat for half an hour in a platinum crucible. Treat the mass with water, add ammonium bicarbonate in excess, digest for some time, filter, and wash. Aluminium-sodium silicate remains in the filter, while the filtrate contains sodium phosphate, sodium bicarbonate, and ammonium carbonate. (Had the solution been filtered before adding the ammonium bicarbonate, some of the aluminium compound would have gone into solution.) Phosphoric acid is determined in the solution according to 1 34, J, a ; the alumina is separated frpm the residue and determined according to 140. The method is tedious and troublesome, as- the precipi- tate is washed only with difficulty ; the results are accurate, how- ever. Compare SCHWEITZER. [Of several other methods which have been used, the follow- ing (by WACKENRODER and FRESENIUS) is one of the easiest to carry out: Precipitate the not too acid solution with ammonia, taking care not to use a great excess of that reagent, and add barium chloride so long as a precipitate continues to form. Digest for some time and then filter. The precipitate contains the whole of the aluminium and the whole of the phosphoric acid, the latter combined partly with aluminium, partly with barium. Filter it off, wash it a little, and dissolve in the least possible quantity of hydrochloric acid. Warm, saturate the solution with barium carbonate, add pure solution of potassa in excess, apply * Journ. de Pharm., xxi, 28. \Zeitnchr.f. analyt. Chem., iv, 151. \ Vierieljahresschr. f. prakt. Pharm., xv, 184 Zeitschr.f. analyt. Chem., ix, 89. 460 DETERMINATION. [ 135. heat, precipitate the barium which the solution may contain with sodium carbonate, and filter. You have now the whole of the aluminium in the solution, the whole of the phosphoric acid in the precipitate. Acidify the solution with hydrochloric acid, boil with some potassium chlorate, and precipitate as directed in 105. Dissolve the precipitate in hydrochloric acid, precipitate the barium with dilute sulphuric acid, filter, and determine the phos- phoric acid in the filtrate by precipitation with solution of magne- sium in the manner described in 134, J, a. HERMANN has applied a perfectly similar method in his analysis of (impure) gibbsite.] f. From Chromium (see also h, &, I). Fuse with sodium carbonate and nitrate and separate the chromic acid and phosphoric acid in the manner described in 166. g. From the Metals of the Fourth Group (see also A, &, Z). a. The method so often used of fusing with sodium carbonate does not give accurate results on account of the constant presence of some phosphoric acid in the washed residue. Compare W. SCHWEIKERT * and G. SCHWEITZER. f The former has studied the separation of zinc from phosphoric acid by this method ; the latter the separation of iron. ft. Dissolve in hydrochloric acid, add tartaric acid, ammo- nium chloride, and ammonia, and finally, in a flask which is to be closed afterwards, ammonium sulphide, put the flask in a moder- ately warm place, allowing the mixture to deposit until the fluid appears of a yellow color, without the least tint of green ; filter, and determine the metals as directed in 108 to 114. The phosphoric acid is found from the loss cr determined according to 134, 5, a. The magnesia mixture may immediately be added, to the filtrate, which contains ammonium sulphide. The washed precipitate is redissolved in just sufficient hydrochloric acid, and the solution reprecipitated by ammonia with addition of magnesia mixture. This method is not well adapted for nickelous phos- phate. A. From Metals of the Second, Third, and Fourth Groups. a. More especially from the second group, aluminium, manga- * Annal. d. Chem. u. Pharm., CXLV, 57; Zeitschr.f. andlyt. Chem., vn, 246. f Zeitschr.f. anatyt. Chem., ix, 84. 135.] PHOSPHORIC ACID. 461 nose, nickel, cobalt, zinc ; and also from ferric iron, if the quan- tity of the latter is not too considerable. The phosphoric acid is precipitated as stannic phosphate, ac- cording to 134, 5, tf, aa. The filtrate contains the bases free from any foreign body requiring removal, which, of course, greatly facilitates their estimation.* REISSIG obtained very good results by this method. If the precipitation of the phosphoric acid is to be effected in the presence of iron and aluminium by tin, GIRARD'S method should be used ( 134, &, tf, bb). /?. From ferric iron, aluminium, alkali earths, and all other bases not precipitated by barium carbonate (H. ROSE). Evaporate off as much as possible of-the free acid from the hydrochloric-acid solution, neutralize it partially with sodium carbonate, add excess of barium carbonate, digest for several days in the cold, filter, and wash with cold water. The precipitate contains all the phos- phoric acid combined with iron, aluminium, and barium, as well as the excess of barium carbonate. The filtrate contains the re- maining bases. Dissolve the precipitate in the least possible quantity of dilute hydrochloric acid, cautiously precipitate the barium with sulphuric acid, filter, saturate filtrate with sodium carbonate and evaporate it, together with the precipitate in it, to dryness, add to the residue an equal quantity of pure silicic acid and six times its quantity of sodium carbonate, and heat in large platinum crucible, first lightly and gradually very strongly. Con- duct the remaining operations just as detailed under 135, 0, ft. y. From much ferric iron in the presence of alkali earths (FRESENIUs).f While the method detailed under 134, d, is applicable in this case, it is nevertheless exceedingly tedious where a small quantity of phosphoric acid is to be separated from a very large quantity of ferric iron. A better process is as follows : Heat the hydrochloric-acid solution to boiling, remove the heat, and add a solution of sodium sulphite until sodium carbonate causes an almost white precipitate ; then boil the mixture until the odor of sulphurous acid has disappeared, nearly neutralize any free acid witli sodium carbonate, add a few drops chlorine water, and * If the nitric acid is not concentrated, a little stannous nitrate is formed, which dissolves and must afterwards be precipitated from the acid fluid by hydrogen sulphide. BABER, Zeitschr. f. d. ges. Naturwiss. , 1864, 324. \ Journ. /. prakt. Chem. , XLV, 258. 462 DETERMINATION. [ 135. finally add an excess of sodium acetate. The smallest quantities of phosphoric acid are thus immediately precipitated as ferric phos- phate. (Arsenic and silicic acids give a similar precipitate, hence if present they must be previously removed. ) Chlorine water is now added drop by drop until the liquid is reddish, then boil until the precipitate has subsided well, filter while hot, and wash with hot water containing a little ammonium acetate. The pre- cipitate now contains all the phosphoric acid, together with a small quantity of iron, while the filtrate contains the great bulk of the iron and all the alkali earths. The precipitate is treated according to 135 l\ i.e., the phosphoric acid is precipitated from its nitric-acid solution as ammonium phosphomolybdate, and the iron and aluminium from the filtrate by means of ammonium sulphide in excess. If the precipitate is free from aluminium, it may be ignited, weighed, and the iron in it determined volumet- rically ( 113), the difference giving the phosphoric acid. This method may be variously modified ; thus the reduction of the iron solution may be effected by hydrogen sulphide, and the excess of this removed by carbonic-acid gas. Again, the precipitation of the ferric phosphate may be effected by digestion with calcium carbonate (free from phosphate) in moderate excess. The pre- cipitation of ferric phosphate by means of ammonium sesquicar- bonate must be effected below 21, otherwise some phosphoric acid will remain in the filtrate (SPILLER*). i. From the Metals of the Fifth and Sixth Groups. Dissolve in hydrochloric or nitric acid, precipitate with hydro- gen sulphide, filter, determine the bases by the methods given in 115 to 127, and the phosphoric acid in the filtrate by the method described in 134, J, tx. From silver the phosphoric acid is separated in a more simple way still, by adding hydrochloric acid to the nitric-acid solution ; from lead it is separated most readily according to 5. k. From all Basic Metals, except Mercury (li. ROSE). The phosphoric acid is separated as mercurous phosphate by ROSE'S method ( 134, 5, y). a. If the substance is free from iron and aluminium, the filtrate from the mercurous phosphate contains all the metals as nitrates, together with much mercurous nitrate, and occasionally * Journ. Chem. Soc. Ser., 2, TV, 148; Zeitschr.f. analyt. Chem., v, 224. 135.] PHOSPHORIC ACip. 463 also some mercuric salt. The former is removed by the addition of hydrochloric acid. The precipitated mercurous chloride is free from other metals : if large in quantity, it should be separated by filtering ; if slight, filtering may be omitted. Add next ammonia to slight alkaline reaction (with previous addition of ammonium chloride if magnesium is present). Filter rapidly from the mer- cury compound which will be precipitated so as to avoid forma- tion of calcium carbonate by contact with air. The filtrate contains the basic radicals from which phosphoric acid has been separated. The mercury compound which has been separated by ammonia is dried and ignited (under a chimney with good draught). Should a residue remain, this must be examined. If it consists of phos- phates of the alkali-earth metals, the treatment with mercury and nitric acid must be repeated; if, on the contrary, it consists of magnesium oxide or of carbonates of the alkali-earth metals, it is dissolved in hydrochloric acid, and the solution added to the fluid containing the chief portion of the basic metals, which may then be separated and determined in the usual manner. The following method is often advantageously resorted to instead of the one described : The nitrate from the mercurous phosphate is evaporated to dryness, in a platinum dish, and the residue ignited, in a plati^ num crucible, under a chimney with good draught. If alkali nitrates are present, some ammonium carbonate must be added from time to time during the process of ignition, to guard against injury to the crucible from the formation of caustic alkali. The ignited residue is treated, according to circumstances, first with water and then with nitric acid, or at once with nitric acid. ft. If the substance contains iron l)ut not aluminium, the greater part of the iron is left undissolved with the mercurous o Jr phosphate. The dissolved part is separatd from the other bases by the methods given in Section Y. ; the iron in the undissolved part is obtained, after ignition of the residue with sodium carbonate and treating the ignited mass with water, as ferric oxide contain- ing alkali (and generally also some phosphoric acid). This is dis- solved in hydrochloric acid, and precipitated with ammonia. y. If the xnbxtain-<' ronldiii^ aluminium, the process just given cannot be used, as aluminium phosphate is not decomposed by fusion with alkali carbonates, while aluminium nitrate, like ferric nitrate, is decomposed by simple evaporation. In this case proceed as follows ; Dissolve the substance in the least quantity of nitric 464 "DETEKMINATION. [ 135. acid, precipitate hot with mercurous nitrate, add a little mercuric nitrate, and then pure potash or soda, till a permanent red precipi- tate appears. The precipitate contains no aluminium, and is to be treated according to a or /3 (II. ROSE, E. E. MUNKOE*). I. From all Bases without exception. Apply SONNEJSTSCHEIN'S method ( 134, &, /?), and in the filtrate from the ammonium phospho-molybdate separate the bases from the molybdic acid. As molybdic acid comports itself with hydro- gen sulphide and ammonium sulphide like a metal of the sixth group, it is best to precipitate metals of the sixth and also of the fifth group from acid solution with hydrogen sulphide, before pro- ceeding to precipitate the phosphoric acid with molybdic acid ; the latter will then have to be separated only from the metals of the first four groups. This is done in the following manner : Mix the acid fluid, in a flask, with ammonia till it- acquires an alkaline reaction, add ammonium sulphide in sufficient excess, close the mouth of the flask, and digest the mixture. As soon as the solution appears of a reddish-yellow color, without the least tint of green, filter off the fluid, which contains molybdenum and ammonium sulphide, wash the residue with water mixed with some ammonium sulphide, and separate the remaining metallic sulphides and hydrox- ides of the fourth and third groups by the methods which will be found in Section Y. Mix the filtrate cautiously with hydrochloric acid in moderate excess, remove the molybdenum sulphide accord- ing to 128, d, and determine the metals of the first and second groups in the filtrate. This method of separating the phosphoric acid from basic radi- cals is highly to be recommended ; especially in cases where a small quantity of phosphoric acid has to be determined in presence of a very large quantity of ferric and aluminium salts, as, for exam- ple, in iron ores, soils, &c. As arsenic acid and silicic acid give, with molybdic acid and ammonia, similar yellow precipitates, it is necessary, if these acids are present, to remove them first. As the separation of the basic metals from the large excess of molybdic acid used is somewhat tedious, the best way is to arrange matters so that this process may be altogether dispensed with. Supposing, for instance, you have a fluid containing ferric iron, aluminium, and phosphoric acid, estimate, in one portion, by cau- tious precipitation with ammonia, the total amount of the three * Amer. Journ. of Sci. and Arts, May, 1871; Zeitschr. /. analyt. Chem,, x, 467 130.] BORIC ACID AND BORIC ANHYDRIDE. 465 bodies; in another portion the phosphoric acid, by means of molybdic acid; and in a third, the iron, in the volumetric way. The aluminium can then be calculated by difference. Attention has already been called ( 135, A, y) to a method, often very con- venient, which consists in precipitating the phosphoric acid together with a small quantity of the iron and then determining in this precipitate the acid and iron, as well as the aluminium carried down. In this method the molybdic acid need be separated only from the small quantity of iron and aluminium, and not from the other bases, thus greatly simplifying the process. 136. BORIC ACID (H 3 BO 3 ) AND BORIC ANHYDRIDE (B 2 O 3 ). I. Determination. Boric acid is estimated either indirectly or in the form of potas- sium borofiuoride. \. The determination of the boric acid in an aqueous or alco- holic solution cannot be effected by simply evaporating the fluid and weighing the residue, as a notable portion of the acid volatil- izes and is carried off with the aqueous or alcoholic vapor. This is the case also when the solution is evaporated with lead oxide in excess. a. Mix the solution of the boric acid with a weighed quantity of perfectly anhydrous pure sodium carbonate, in amount about 1 \ times the supposed quantity of B 2 3 present. Evaporate the mix- ture to dryness, heat the residue to fusion, and weigh. The residue contains a known amount of Na 2 O, and unknown quantities of CO a and B 2 O 3 combined as sodium borate and carbonate. Determine the CO a by one of the methods given in 139, and find the B a O 3 from the difference (II. ROSE). 1}. In the method a, if between 1 and 2 mol. sodium carbonate (Na a CO 3 ) are used to 1 mol. B 2 O 3 and this can easily be done if one knows approximately the amount of the latter present all the carbonic acid io expelled by the boric acid. Hence we have only to deduct the Na a O from the residue to find the B 2 O 3 . As the tumultuous escape of carbonic acid may lead to loss, it is well, after 1 inving thoroughly dried the residual saline mass, to project it in small portions cautiously into the red-hot crucible. Results good (F. G. SCHAFFGOTSCH).* c. When the amount of acid is quite unknown, and an estima- tion of carbonic acid in the residue is objected to, you may proceed * Pogg. Ann., cvn, 427. 466 DETERMINATION. [ 136. thus : Evaporate the solution of the acid with addition of a weighed quantity of anhydrous neutral borax (sodium metaborate NaBOJ free from carbonic acid to dryness, and heat the residue to redness with great caution (on account of the intumescence) till the weight is constant. The amount of neutral borax must be so adjusted that it may not be entirely converted into common borax (2NaBO 2 B 2 O 3 ) (II. KOSE). d. If a solution contains, besides boric acid, only alkalies or magnesium, the acid may be determined, according to C. MARIG- NAC,* in the following manner : Neutralize the solution with hydrochloric acid, add double magnesium and ammonium chloride in sufficient quantity to give at least 2 parts of MgO to l.part of B 2 O S , then add ammonia and evaporate to drynces. If on adding the ammonia a precipitate is formed which does not redissolve readily on warming, add more ammonium chloride. The evapora- tion is conducted, at least towards the end, in a platinum dish, a few drops of ammonia being added from time to time. Ignite the dry mass, treat with boiling water, collect the insoluble precipitate (consisting of magnesium borate mixed with excess of magnesium oxide) on a filter, and wash with boiling water till the w r ashings remain clear with nitrate of silver. The filtrate and washings are mixed with ammonia, evaporated to dryness, ignited, and washed with boiling water as before. The two insoluble residues are ignited together in the platinum dish before used, as strongly as possible, and for a sufficiently long time, in order to decompose the slight traces of magnesium chlo- ride that might still be present. After weighing determine the magnesium oxide, and find the boric acid from the difference. The determination of the magnesium may be made by dissolving the residue in hydrochloric acid and precipitating as ammonium magnesium phosphate, or more quickly, and almost as accurately, by dissolving in a known quantity of standard sulphuric acid at a boiling temperature and determining the excess of acid with stand- ard soda (comp. Alkalimetry). Should a little platinum remain behind on dissolving the resi- due, it must be weighed and subtracted from the weight of the whole (unless the dish was weighed first). Results satisfactory. MARIGNAC obtained in two experiments 0-276 instead of O28. 2. If boric acid is to be determined as potassium borqftuoride, alkalies only (preferably only potassa) maybe present. The process- * Zeitschr. f. analyt. Chem., I, 405. 136.] BOKIC ACID AND BOEIC ANHYDKIDE. 467 is conducted as follows : Mix the fluid with pure solution of potassa, adding for each mol. boric acid supposed to be present, at least 1 mol. potassa ; add pure hydrofluoric acid (free from silicic acid) in excess, and evaporate, in a platinum dish, on the water-bath, to dry ness. The fumes from the evaporating fluid should redden litmus paper, otherwise there is a deficiency of hydrofluoric acid. The residue consists now of KF,BF 3 and KF,HF. Treat the dry saline mass, at the common temperature, with a solution of 1 part of potassium acetate in 4 parts of water, let it stand a few hours, with stirring, then decant the fluid portion on to a weighed filter, and wash the precipitate repeatedly in the same way, finally on the filter, with solution of potassium acetate, until the last rinsings are no longer precipitated by calcium chloride. By this course of pro- ceeding, the hydrogen potassium fluoride is removed, without a .particle of the potassium borofluoride being dissolved. To remove the potassium acetate, wash the precipitate now with 84 per cent, alcohol, dry at 100, and weigh. As potassium chloride, nitrate, and phosphate, sodium salts, and even, though with some difficulty, potassium sulphate, dissolve in solution of potassium acetate, the presence of these salts does not interfere with the estimation of the boric acid ; however, sodium salts must not be present in consider- able proportion, as sodium fluoride dissolves with very great diffi- culty. The results obtained by this method are satisfactory. STEO- MEYEK'S experiments gave from 97'5 to 10O7 instead of 100. When the amount of alkali salt to be removed is very large, the saline mass left on evaporation should be warmed with the solution of potassium acetate, allowed to stand 12 hours in the cold and then filtered. In this way the quantity of potassium acetate required will be much reduced. For the composition and proper- ties of potassium borofluoride, see 93, 5. As the salt is very likely to contain potassium silicofluoride it is indispensable to test it for that substance ; this is done by placing a small sample of it on moist blue litmus paper, and putting another sample into cold concentrated sulphuric acid. If the blue paper turns red, and effervescence ensues in the sulphuric acid, the salt is impure, and contains potassium silicofluoride. To remove this impurity, dis- solve the remainder of the salt, after weighing it, in boiling water, add ammonia, and evaporate, redissolve in boiling water, add ammonia, 5, when there has been any delay or neglect in completely expelling all adherent ammonia from the precipitate by prolonged boiling. These circumstances, tending to give results too high, are partly compensated for by the fact that the carbonates of the alkaline earths are not absolutely in- soluble in the fluid containing ammonium chloride and in the wash-water. If the mixture of carbonated water and calcium or barium chloride and ammonia is not heated as above detailed, the results may be too low, either from imperfect decomposition of the ammonium carbonate by reason of insufficient heating, or be- cause of loss of ammonium carbonate by too active an ebullition. y. AFTER PETTENKOFER.* The principle of this simple and expeditious process consists in mixing the carbonic-acid water with a measured quantity of stand- *BUCHNEB'S neuesRepert., x, 1; Journ. f. prakt. Chem., LXXXII, 32; Annal. d. Chem. u. Pharm., n, Supplement, i; Zeitschr. f. analyt. Chem., I, 92. 139. J CARBONIC ACID. 485 ard lime water (or, under certain circumstances, baryta water) in excess. After complete separation of the calcium or barium carbo- nate, the excess of calcium or barium in the fluid is determined in an aliquot part by means of standard solution of oxalic acid ; the difference gives the calcium or barium precipitated by the carbonic acid, and consequently the amount of the latter present. If a water contains only free carbonic acid, the analyst has only to bear in mind if lime water is employed that the calcium car- bonate formed is at first, as long as it remains amorphous, very perceptibly soluble in water, to which it communicates an alkaline reaction. Hence the unprecipitated lime in the fluid cannot be estimated till the calcium carbonate has separated in the crystalline form, which takes 8 or 10 hours, unless the mixture is warmed to 70 or 80. On this account it is generally best to use baryta water (see "Analysis of Atmospheric Air"). If, on the contrary, a water contains an alkali carbonate or any other alkali salt whose acid would be precipitated by lime or baryta, a neutral solution of calcium or barium chloride must first be added to decompose the same. This addition, too, prevents any incon- venience arising from the presence of free alkali in the lime or baryta water, or of magnesium carbonate in the carbonic acid water; this inconvenience consists in the fact that oxalate of an alkali or of magnesium enters into double decomposition with cal- cium carbonate (which is seldom entirely absent from the fluid to be analyzed), forming calcium oxalate and carbonate of the alkali or of magnesium, which latter will of course again take up oxalic acid. In the presence of magnesium salts in the carbonic acid water, in order to avoid the precipitation of the magnesium, a little ammonium chloride must also be added, but in this case heat must not be applied to induce the calcium carbonate to become more quickly crystalline, as ammonia would be thereby expelled. In making the determination the first thing to be done is to ascertain the relation between the lime- or baryta water and a standard solution of oxalic acid. PETTENKOFEK makes the latter solution by dissolving 2*8647 grin, pure uneflioresced dry crystal- lized oxalic acid to 1 litre ; 1 c.c. of this is equivalent to 1 mgrm. carbonic acid. The lime water is standardized as follows : Measure 45 c.c. into a little flask which can be closed by the thumb, and then run in from the burette the solution of oxalic acid till the 486 DETERMINATION. [ 139. alkaline reaction has just vanished. During the operation the flask is closed with the thumb and gently shaken. The end is attained as soon as a drop taken out with a glass rod and applied to sensitive turmeric paper * produces no brown ring. The first experiment is a rough one, the second should be exact. The analysis of a carbonic acid water (a spring water, for instance) is performed by transferring 100 c.c. to a dry flask, add- ing 3 c.c. of a neutral and nearly saturated solution of calcium or barium chloride, and 2 c.c. of a saturated solution of ammonium chloride, then 45 c.c. of the standard lime or baryta water ; close the flask with an india-rubber stopper, shake and allow to stand 12 hours. The fluid contents of the flask measure consequently 150 c.c. From the clear fluidf take out by means of a pipette two por- tions of 50 c.c. each, and determine the free lime or baryta by means of oxalic acid, in -the first portion approximately, in the second exactly. Multiply the c.c. used in the last experiment by 3 and deduct the product from the c.c. of oxalic acid which corre- spond to 45 c.c. of lime or baryta water. The difference shows the lime or baryta precipitated by carbonic acid, each c.c. corresponds to 1 mgrm. carbonic acid. The method is convenient and good ; it is especially to be recommended for dilute carbonic acid water. When calcium sul- phate or carbonate is present, as is almost always the case in spring water, you must always before titrating await the conversion of the amorphous calcium carbonate to the crystalline state, even if baryta water is used (K. KNAPP J). Baryta water therefore possesses no advantages over lime water for the analysis of spring waters. *For the preparation of this bibulous paper should be used, the ash of which is free from carbonate of lime. Swedish filtering-paper answers best. J. GOTTLIEB (Journ.f. prakt. Chem., cvn, 488; Zeitschr. f. analyt. Chem., ix, 251) prefers aqueous tincture of litmus, prepared from litmus first exhausted with alcohol, and used in a very dilute state. E. SCHULZE and M. MARCKER (Zeitschr. /. analyt. Chem., ix, 334) employ corallin or rosolic acid, which they say is specially adapted for the purpose. The alcoholic solution is cautiously neutralized with potassa and a drop or two of this tincture is added. F. SCHULZE (Zeitschr. f. analyt. Chem., ix, 292) recommends alcoholic tincture of turmeric. f It is not admissible to use a filter (A. MULLER, Zeitschr. f. analyt. Chem., I, 84). \Annal. d. Chem. u. Pharm., CLVIII, 112; ZeitscUr.f. analyt. Chem., x, 361. 139.] CARBONIC ACID. 487 II. Separation of Carbonic Ac'nl from the Basic Radicals, and its Estimation hi Carbona1< *. a. Exit inatioii in Normal Alkali Carbonates and Alkali-earth Carbonates. .If the salts are unquestionably normal carbonates, and there is no other salt with power to neutralize an acid present, we may determine the quantity of the basic radical by the alkalimetric method ( 219, 220, 223), and calculate the amount of CO, necessary to form with it normal carbonate. b. Separation from Basic Metals in Salts which upon ignition readily and completely yield their Carbonic Acid. Such are, for instance, the carbonates of zinc, cadmium, lead, copper, magnesium, &c. a. Anhydrous Carbonates. Ignite the weighed substance, in a platinum crucible (cadmium and lead carbonates in a porcelain crucible), until the weight of the residue remains constant. The results are, of course, very accurate. Substances liable to absorb oxygen upon ignition in the air are ignited in a bulb-tube, through which a stream of dry carbon dioxide gas is conducted. The car- bonic acid is inferred from the loss. /?. Hydrated Carbonates. The substance is ignited in a bulb- tube through which dried air or, in presence of oxidizable sub- stances, carbon dioxide is transmitted, and which is connected with a calcium ^chloride tube, by means of a dry, close-fitting cork. During the ignition, the posterior end of the bulb-tube is, by means of a small lamp, kept sufficiently hot to prevent the con- densation of water in it, care being taken, however, to guard against burning the cork. The loss of weight of the tube gives the amount of the water -(-the carbonic acid; the increase of weight gained by the calcium chloride tube gives the amount of the water, and the difference accordingly that of the carbonic acid. A somewhat wide glass tube may also be put in tiie place of the bulb-tube, and the substance introduced into it in a little boat, which is weighed before and after the operation. c. Srjxi ration front all fixed Basic Radicals, without exception, in Anhydrous Carbonat< *. Fuse vitrified borax in a weighed platinum crucible, allow to cool in the desiccator, weigh, then transfer the well-dried substance to the crucible and weigh again. The weights of both carbonate 488 DETERMINATION. [ 139. and borax are thus ascertained. They should be in about the pro- portion of 1 : 4. Heat is then applied, which is gradually increased to redness, and maintained at this temperature until the contents of the crucible are in a state of calm fusion. The crucible is now allowed to cool, and weighed. The loss of weight is carbonic acid. The results are very accurate (SCHAFFGOTSCH). I must add that borax-glass may be kept in a state of fusion at a red heat for J to an hour without the occurrence of any vola- tilization, but that at a white heat (by igniting over the gas-bel- lows), even in a few minutes, it suffers a decided loss.* A few bubbles of carbonic acid remaining in the fusing mass are without any influence on the result. Instead of vitrified borax fused potassium dichromate may be used, in the proportion of 5 to 1 of the carbonate (H. RosEf). The heat applied in this case must be low, and great caution must be used, or the dichromate will lose weight of itself.;}; The carbonic acid may be expelled from alkali carbonates, by strong ignition with ignited silica (H. KOSE). d. Separation from all bases without exception. (Estimation from the loss of weight.) aa. Carbonates of Bases yielding Soluble Salts with Sulphuric Acid. The process is conducted in the apparatus illustrated by Fig. 93. The size of the flask depends upon the capacity of the balance. B may be smaller than A. The tube a is closed at b with a little wax ball or a small piece of india-rubber tube stopped with half an inch of rod ; the other end of the tube a is open, as are also both ends of c and d. The flask B is nearly half filled with concentrated sulphuric acid, free from oxides of nitrogen and sulphurous acid. The tubes must fit air-tight in the corks, and the latter equally so in the flasks. The weighed substance is put into A ; this flask is then filled about one third with waiter, the cork properly inserted, and the apparatus tared on the balance. A few bubbles of air are now sucked out of d, by means of an india-rubber tube. This serves to rarefy the air in A also, and causes the sulphuric acid in B to ascend in the tube c. The latter is watched for some time, *Zeitschr. f. analyl. Chem., i, 65. f Pogg. An?tal., cxvr, 131. \Zdtschr.f. analyt. C/iem., I, 183. Pogg. Annal., cxvi, 686. 139.] CARBONIC ACID. 489 to ascertain whether the column of sulphuric acid in it remains stationary, which is a proof that the apparatus is air-tight. Air is then again sucked out of ^7, which causes a portion of the sulphuric acid to flow over into A. The carbonate in the latter flask is decomposed by the sulphuric acid, and the liberated carbonic acid, completely dried in its passage through the sulphuric acid in B y escapes through d. When the evolu- tion of the gas slackens a fresh portion of sulphuric acid is made to pass over into A, by renewed suction through d\ the operation being repeated until the whole of the carbonate is decom- posed. A more vigorous suction is now applied, to make a large amount of sulphuric acid pass over into A, whereby the contents of that flask are considerably heated ; when the evolu- tion of gas bubbles has completely ceased, the stopper on a is opened, and suction applied to d, until the air sucked out tastes no longer of carbonic acid.* When the apparatus is quite cold it is replaced upon the balance, and the equilibrium restored by additional weights. The sum of the weights so added indicates the amount of carbonic acid originally present in the substance. If the flasks A and B are selected of small size, the apparatus may be so constructed that, together with the contents, it need not weigli above TO grammes, admitting thus of being weighed on a delicate balance. The results obtained by the use of this apparatus, first suggested by WILL and myself, are very accurate, provided the quantity of the carbonic acid be not too trifling. Various modifications of the apparatus have been proposed, principally in order to make it lighter. See foot-note, p. 492. If sulphites or sulphides are present, together with the carbon- ates, their injurious influence is best obviated by adding to the carbonate solution of normal potassium chroinate in more than sufficient quantity to effect their oxidation. If chlorides are pres- ent, in order to prevent the evolution of hydrochloric acid, add to * In accurate experiments it is advisable to connect the end b of the tube with a calcium-chloride tube during the process of suction, and to use an aspira- tor or hydraulic air-pump instead of the mouth. 490 DETERMINATION. [ 139. the evolution flask a sufficient quantity of silver sulphate in solu- tion, or connect the exit tube d with a small prepared TJ-tnbe, which is, of course, "first tared with the apparatus, and afterwards weighed with it. This U-tube is prepared in accordance with the happy proposal of STOLBA by filling with fragments of pumice which have been boiled with an excess of concentrated solution of cupric sulphate, till the air has been expelled, and then dried and heated to complete dehydration of the copper salt. If the U-tube is only 8 cm. high and has a bore of 1 cm., it answers the purpose very well. The outer end is provided with a perforated cork and short glass tube. We apply suction to this by means of a flexible tube, instead of to d. "bb. Carbonates of Bases yielding Insoluble Salts with Sul- phuric Acid. The method aa is unsuitable for these bases, because the insoluble sulphate formed, e.g., calcium sulphate, partially protects tne still undecompcsed portion of carbonate from decomposition. The apparatus is hence modified as shown in Fig. 94. The modification consists simply in the tube ab, which, as the cut shows, is provided with a glass bulb, and is drawn out to a fine point at the lower end. The process is carried out as follows : The weighed substance is introduced into A, with some water. . The bulb-tube ab con- tains dilute nitric acid (or, if substances are present which decom- pose nitric acid, e.g., ferrous oxide, 10-per cent, hydrochloric acid) in quantity more than sufficient to decompose the carbonate present. The end b is closed by a well-kneaded piece of wax, or with a short section of rubber tubing closed by a small piece of a glass rod, in order to prevent the acid from running out. The tip of the tube a should not, at first, dip into the water in A. Place the apparatus on a balance and ascertain its tare, then care- fully push the tube a down into the liquid until the tip nearly reaches the bottom, then loosen the wax plug or open the rubber tube for a moment, and allow some of the acid to run out, and repeat this now and again until all the carbonate has been decom- posed. Now heat the contents of A to incipient boiling, remove the stopper from 5, and draw the carbonic acid out of the apparatus as detailed under aa\ after cooling, determine the loss of weight. It will be seen at a glance that the apparatus is susceptible of 139.] CARBONIC ACID. 491 a different construction ; for instance, the flask E may be omitted, and the tube C connected instead with a calcium-chloride tube or Fig. 94. Fig. 95. a tube containing pumice-stone or asbestos saturated with sulphuric acid ; or the substance to be decomposed may be placed in a small tube arranged to stand upright at first, or suspended by a thread, and which, after the apparatus is tared, may be, upset or lowered into the dilute acid ; the closure of the tube J may also be effected by a pinch-cock, etc. Such modifications, if made judiciously, affect the results but little, if at all. An apparatus of this kind, modified by FR. MOHR, is shown in Fig. 95. One of the most convenient apparatus is that proposed by GEISSLER * and shown in Fig. 96. This consists of two parts, AB and C. C is ground into the neck at a so as to fit airtight, yet be readily removable, whereby A may be filled. C carries a tube, >* c S C3 s S s S S 1 e s 10 J 1 c= 8 CO 1 5* ^ CO id T t O5 ^Q 10' s id CO GO ^ cci 1 S8 S 10 ^ id 88 10 !> 8 . 09 S J i fe S - S C5 co 05 . CO 05 S 1 - s a 1 J> 1C "^ 05 a 1 * 2 g 10 s s IO 1 1 -<# *^ N - 3 3 S 10 ^ ^ . co . 1O *S CO CO - 1 - 1 (M s 3- CD r3 ^ 10 CO g id aj O o w s 3 s ^ 10' lO a e 10 - s T-I 1C s 3 5! S to i i S 10 S r- g 00 5 Evolved Absorbed Evolved Absorbed Evolved. Absorbed Evolved Absorbed Evolved Absorbed 140.] SILICIC ACID. 509 aqueous or acid solution free from other fixed bodies, simply evaporate the solution in a platinum dish, ignite and weigh the residue. Respecting a volumetric estimation of silicic acid (conversion into and acidimetric determination of potassium silicofluoride, see 97, 5), I must refer to STOLBA.* II. SEPARATION OF SILICIC ACID FROM THE BASIC RAD- ICALS. a. In all compounds which are decomposed by Hydrochloric or Nitric Acid, on digestion in open vessels. To this class belong the silicates soluble in water, as well as many of the insoluble silicates, as, for instance, nearly all zeolites. Several minerals not decomposable of themselves by acids become so by persistent ignition in a state of fine powder (F. MOHR f). If the ignition is too strong, particles of alkali may be lost. The substance is very finely powdered, J dried at 100, and put into a platinum or porcelain dish (in the case of silicates whose solu- tion might be attended with disengagement of chlorine, platinum cannot be used) ; a little water is then added, and the powder mixed to a uniform paste. Moderately concentrated hydrochloric acid, or if the substance contains lead or silver nitric acid, is now added, and the mixture digested at a very gentle heat, with con- stant stirring, until the substance is completely decomposed, in other terms, until the glass rod, which is rounded at the end, encounters no more gritty powder, and the stirring proceeds smoothly without the least grating. The silicates of this class do not all comport themselves in the same manner in this process, but show some differences ; thus most -of them form a bulky gelatinous mass, whilst in the case of others the silicic acid separates as a light pulverulent precipitate ; again, many of them are decomposed readily and rapidly, whilst others require protracted digestion. When the decomposition is effected, the mixture is evaporated to dry ness on the water-bath, and the residue heated, with frequent stirring, until all the small lumps have crumbled to pieces, and the * Zeitschr. f. annlyt. Chcm , iv, 163. f lb. t vn, 293. $ Very hard silicates cannot be powdered in an agate mortar without taking up silica ; these must, therefore, be powdered in a steel mortar, sifted, and freed from particles of steel with the magnet. 510 DETERMINATION. [ 140. whole mass is thoroughly dry, and until no more acid fumes escape. It is always the safest way to conduct the drying on the water-bath. Occasionally it is well to moisten the dry mass with water and evap- orate again. In cases where it appears desirable to accelerate the desiccation by the application of a stronger heat, an air-bath may be had recourse to ; which may be constructed in a simple way, by suspending the dish containing the substance, with the aid of wire, in a somewhat larger dish of silver or iron, in a manner to leave everywhere between the two dishes a small space of uniform width. Direct heating over the lamp is not advisable, as in the most strongly heated parts the silicic acid is liable to unite again with the separated bases to compounds which are not decomposed, or only imperfectly, by hydrochloric acid. When the mass is cold, it is brought to a state of semi-fluidity by thoroughly moistening it with hydrochloric acid ; after w r hich it is allowed to stand for half an hour, then warmed on a water- bath, diluted with hot water, stirred, allowed to deposit, and the fluid decanted on to a filter ; the residiiary silicic acid is again stirred with hydrochloric acid, warmed, diluted, and the fluid once more decanted ; after a third repetition of the same operation, the precipitate also is transferred to the filter, thoroughly washed with hot water, well dried, and ignited at last as strongly as possible, as directed in 52 or 53. For the properties of the residue, see 93, 9. The results are accurate. The basic metals, which are in the filtrate as chlorides, are determined by the methods given above. Devia- tions from the instructions here given are likely to entail loss of substance ; thus, for instance, if the mass is not thoroughly dried, a not inconsiderable portion of the silicic acid passes into the solu- tion, whereas, if the instructions are strictly complied with, only traces of the acid are dissolved ; in accurate analyses, however, even such minute traces must not be neglected, but should be separated from the metals precipitated from the solution. The separation may, as a rule, be readily effected by dissolving them, after ignition and weighing, in hydrochloric or sulphuric acid, by long digestion in the heat, the traces of silicic acid being left undissolved. Some- times it is better to fuse the metallic oxides with potassium disul- phate, or to reduce them to the metallic state by ignition in hydro- gen, and then to treat with hydrochloric acid. Again, if the silicic acid is not thoroughly dried previous to ignition, the aqueous vapor disengaged upon the rapid application of a strong heat may carry 140.] SILICIC ACID. 511 away particles of the light and loose silica. If a suction appara- tus has been used, however, and the precipitate has been quite thoroughly freed from water, the precipitate may be ignited at once, as described in 52, p. 116. In this case, however, the incineration of the filter is frequently imperfect. The silicic acid may be tested as follows : This testing must on no account be omitted if the silica has been separated in a pulveru- lent and not in a gelatinous form. Heat a portion on a water- bath with moderately concentrated solution of sodium carbonate for an hour in a platinum or silver dish ; with less advantage in a porce- lain dish. EGGERTZ* recommends, for 0*1 grm. silicic acid, 6 c.c. of a saturated solution of sodium carbonate and 12 c.c. of water. Pure silica would dissolve. If a residue remains, pour off the clear fluid and heat again with a small quantity of sodium carbonate. If a residue still remains, weigh the rest of the impure silica and treat it according to &, to estimate the amount of impurity. If you have pure hydrofluoric acid, you may also test the silicic acid in a very easy manner, by treating it with this acid and a few drops of sulphuric acid in a platinum dish ; upon the evaporation of the solution, the silicic acid, if pure, will volatilize completely (as silicon fluoride). If a residue remains, moisten this once more with hydrofluoric acid, add a few drops of sulphuric acid, evaporate, and ignite ; the residue consists of the sulphates of the metals retained by the silicic acid, as well as any titanic acid that was present (BERZELIUS). Ammonium fluoride may be used instead of hydrofluoric acid. 1). Compounds which are not decomposed "by Hydrochloric or Nitric Acid on digestion in open vessels. a. Decomposition by fusion with Alkali Carbonate. Reduce the substance to an impalpable powder, by trituration and, if necessary, sifting ( 25) ; transfer to a platinum crucible, and mix with about 4 times the weight of pure anhydrous sodium carbonate or sodium and potassium carbonate, with the aid of a rounded glass rod ; wipe the rod against a small portion of sodium carbonate on a card, and transfer this also from the card to the crucible. Cover the latter well, and heat, according to size, over a gas or spirit-lamp with double draught, or a blast gas-lamp ; or * Zeitschr. f. analyt. Chem., vir, 502. 512 DETERMINATION. [ 140. insert in a Hessian crucible, compactly filled up with calcined magnesia, and heat in a charcoal fire. Apply at first a moderate heat for some time to make the mass simply agglutinate ; the carbonic acid will, in that case, escape from the porous mass with ease and unattended with spirting. Increase the heat afterwards, finally to a very high degree, and terminate the operation only when the mass appears in a state of calm fusion, and gives no more bubbles. The platinum crucible in which the fusion is conducted must not be too small ; in fact, the mixture should only half fill it. The larger the crucible, the less risk of loss of substance. As it is of O " importance to watch the progress of the operation, the lid must be easily removable ; a concave cover, simply lying on the top, is there- fore preferable to an overlapping lid. If the process is conducted over the spirit or simple gas-lamp, the mixed sodium and potas- sium carbonates are preferable to sodium carbonate, as they fuse much more readily than the latter. In heating o\ r er a lamp, the crucible should always be supported on a triangle of platinum wire, with the opening just sufficiently wide to allow the crucible to drop into it fully one third, yet to retain it firmly, even with the wire at an intense red heat. When conducting the process over a spirit-lamp with double draught, or over a simple gas-lamp, it is also advisable, towards the end of the operation, when the heat is to be raised to the highest degree, to put a chimney over the cruci- ble, with the lower border resting on the ends of the iron triangle which supports the platinum triangle ; this chimney should be about 12 or 14 cm. high, and the upper opening measure about 4 cm. in diameter. The little clay chimneys recommended by O. L. ERDMANN are still more serviceable (Fig. 20,]). 22, fcl Qnal. Anal."). When the fusion is ended, the red-hot crucible is removed with tongs, and placed on a cold, thick, clean iron plate, on which it will rapidly cool ; it is then generally easy to detach the fused cake in one piece. The cake (or the crucible with its contents) is put into a beaker, from 10 to 15 times the quantity of water poured over it, and heat Jipplied for half an hour, then hydrochloric acid is gradually added, or, under certain circumstances, nitric acid ; the beaker is kept covered with a glass plate, or, which is macli better, with a large watch-glass or porcelain dish, perfectly clean outside, to prevent the loss of the drops of fluid which the escaping carbonic acid car- 140.] SILICIC ACID. 613 ries along with it ; the drops thus intercepted by the cover are afterwards rinsed into the beaker. The crucible is also rinsed with water mixed with dilute acid, and the solution obtained added to the fluid in the beaker. "The solution is promoted by the application of a gentle heat, which is continued for some time after this is effected to insure the complete expulsion of the carbonic acid ; since otherwise some loss of substance might be incurred, in the subsequent process of evapo- ration, by spirting caused by the escape of that gas. If in the pro- cess of treating the fused mass with hydrochloric acid, a saline powder subsides (sodium or potassium chloride), this is a sign that more water is required. If the decomposition of the mineral has succeeded to the full extent, the hydrochloric acid solution is either perfectly clear, or light flakes of silicic acid only float in it. But if a heavy powder subsides, which feels gritty under the glass rod, this consists of undecomposed mineral. The cause of such imperfect decomposi- tion is generally to be ascribed to imperfect pulverization. In such cases the undecomposed portion may be fused once more with alkali carbonate ; the better way, however, is to repeat the process with a fresh portion of mineral more finely pulverized. The hydrochloric or nitric acid solution obtained is poured, together with the precipitate of silicic acid, which is usually floating in it, into a porcelain or, better, into a platinum dish, and treated as directed in II., a. That the fluid may not be too much diluted, the beaker should be rinsed only once, or not at all, and the few remaining drops of solution dried in it ; the trifling residue thus obtained is treated in the same way as the residue left in the evapo- rating basin. This is the method most commonly em ployed 'to effect the decomposition of silicates that are undecomposable by acids ; that it cannot be used to determine alkalies in silicates is self-evident. ft. Decomposition ~by means of Hydrofluoric A cid. aa. By Aqueous Hydrofluoric Acid. The silicate should be finely pulverized, dried at 100 (in some cases ignition is advisable"-). It is mixed, in a platinum dish, with * Many minerals are much more readily decomposed by hydrofluoric acid also if they are previously ignited in a state of fine division (HERMANN, RAM- MELSBERG, FR. MOHR, Zeitschr. /. analyt. Chem., vn, 291). 514 DETERMINATION. [ 140, rather concentrated, slightly fuming hydrofluoric acid, the acid being added gradually, and the mixture stirred with a thick plati- num wire. The mixture, which has the consistence of a thin paste, is digested some time on a water-bath at a gentle heat, and pure concentrated sulphuric acid, diluted with an equal quantity of water, is then added, drop by drop, in more than sufficent quantity to convert all the basic metals present into sulphates. The mixture is now evaporated on the water-bath, during which operation sili- con fluoride gas and hydrofluoric acid gas are continually volatiliz- ing ; then it is finally exposed to a stronger heat at some height above the lamp, until the excess of sulphuric acid is almost completely expelled. The mass, when cold, is thoroughly moistened with con- centrated hydrochloric acid, and allowed to stand at rest for one hour ; water is then added, and a gentle heat applied. If the decomposition has fully succeeded, the whole must dissolve to a clear fluid. If an undissolved residue is left, the mixture is heated for some time on the water-bath, then allowed to deposit, the clear supernatant fluid decanted as far as practicable, the residue dried, and then treated again with hydrofluoric acid and sulphuric acid r and, lastly, with hydrochloric acid, which will now effect complete solution, provided the analyzed substance was very finely pulver- ized, and free from barium, strontium (and lead). The solution i& added to the first. The basic metals in the solution (which con- tains them as sulphates, and contains also free hydrochloric acid) are determined by the methods which will be found in Section V. This method, which is certainly one of the best to effect the decomposition of silicates, was proposed by BERZELIUS. It has- been but little used hitherto, because we did not know how to pre- pare hydrofluoric acid, except with the aid of a distilling appa- ratus of platinum, or, at least, with a platinum head ; nor to keep it, except in platinum vessels. These difficulties can now be con- sidered as overcome, comp. 58, 2. Xever omit testing the acid before using it. The hydrofluoric acid may also be employed in combination with hydrochloric acid ; thus 1 grm. of finely elutriated felspar,, mixed with 40 c.c. water, 7 c.c. hydrochloric acid of 25$ and SJc.c. hydrofluoric acid, and heated to near the boiling point, dissolves- completely in three minutes. 4 c.c. sulphuric, acid are then added, the barium sulphate which may separate is filtered oft', and the 140.] SILICIC ACID. 515 filtrate evaporated till no more hydrofluoric acid escapes (AL. MlTSCHEKLICH *). The execution of the method requires the greatest possible care, both the liquid and the gaseous hydrofluoric acid being most injurious substances. The treatment of the silicate with the acid and the evaporation must be conducted in the open air, otherwise the windows and all x glass apparatus will be attacked. As the silicic acid is in this method simply inferred from the loss,f a combination with method a is often resorted to. lb. By Gaseous Hydrofluoric Acid. Instead of the aqueous solution of hydrofluoric acid, the gaseous acid may also be used for decomposing silicates. This method, which was formerly much used, was proposed by BRTJN- NER,; and is as follows: Place 1 or 2 grammes of the very finely powdered silicate in as thin a layer as possible in a shallow platinum dish, moisten the powder with diluted sulphuric acid, and place the dish on a leaden tripod or other support within a leaden box about 6 inches in diameter and about 6 inches high, and on the bottom of which there has been just placed a half -inch layer of powdered fluor-spar mixed into a paste with concentrated sul- phuric acid. (Take care to avoid the vapors evolved ; the mixing of the fluor-spar and sulphuric acid should be done with a long glass rod, or, better still, with a leaden rod). As soon as the plat- inum dish has been placed in the box, by the aid of a pair of pin- cers or tongs, put on a tightly-fitting cover, lute the joints airtight with plaster-of-paris, and set the whole for 6 to 8 days in a warm place. If it is desired to facilitate the process do not lute air- tight, but heat the apparatus in the open air over a gas- or alcohol-lamp ; by this method 1 to 2 grammes of the silicate may be decomposed in a few hours, provided the silicate has been spread out in a very thin layer, or else stirred from time to time, an operation which must be cautiously effected. * Journ. f. prakt. Chem., LXXXI, 108. f The silicon escaping in the form of fluoride may sometimes be determined directly by the method of STORY M. \SKEL YNE ( Zeitschr. f. analyt. Chem., ix, 380), which, however, requires a platinum retort of peculiar construction. J Pogg. Annal. , XLIV, 134. An apparatus that may be used in the laboratory has been described by A. MULLER (Journ. f. prakt. CTiem., xcv, 51). 516 DETERMINATION. [ 140. If the decomposition lias succeeded, tile residue in the plati- num dish will consist of silicofluorides of the metals, and sulphates. Place the dish now in a larger platinum dish, add sulphuric acid drop by drop, using a little more than is sufficient to convert the bases into sulphates, evaporate in an air-bath, nearly but not quite expel finally the excess of sulphuric acid over the naked flame, and treat the residue with hydrochloric acid and water as detailed under aa. The decomposition may be considered as complete only when a perfect solution results (apart from the presence of a little barium sulphate). If a platinum tube adapted for the purpose is at hand the de- composition may also be effected by heating the finely powdered mineral placed in a platinum boat inserted into the tube while, passing through the latter a current of dry hydrofluric-acid gas. The platinum tube is bent downwards in front, and the end should dip into water ; the water takes up the volatile fluorides, while the iion- volatile remain ^in the platinum boat. (SAINT-CLAIRE DEVILLE, KUHLMANN.*) cc. By Ammonium Fluoride. Mix the very finely powdered substance in a platinum dish witli four times its weight of ammonium fluoride, moisten well with concentrated sulphuric acid, heat on the water-bath till the evolution of silicon fluoride and hydrofluoric acid slackens, add more sulphuric acid, heat again, finally somewhat more strongly till the greater part of the sulphuric acid has escaped, and treat the residue according to aa (L. v. BABO, J. POTYKA, R. HOFF- MANN f). H. ROSE J first warms the silicate gently with seven times its amount of the fluoride and some water, then heats gradu- ally to redness till no more fumes escape, and finally treats with sulphuric acid. dd. By Hydrogen Potassium Fluoride, etc. In silicates, which more or less resist the action of hydrofluoric acid, such as zircon and beryl, the basic metals with the exception of the alkalies may be determined by fusing with hydrogen potassium fluoride (MARIGNAC, GIBBS ), or by mixing with 3 parts of sodium fluoride, adding 12 parts of potassium disulphate * Compt. Rend., LVITT, 545. f Zeitschr.f. analyt. Chem., vi, 366. \Pogg. Annal., cvin, 20. % Zeitschr.f. analyt. Chem., m r 399. 140.] SILICIC ACID. 517 to the crucible, and then heating at first very gently, afterwards more strongly till the mass fuses calmly. The residue is dissolved in water or hydrochloric acid (CLARKE*). y. Decomposition by Fusion with Barium Hydroxide or Bariwn Carbonate. The fusion of silicates with barium carbonate requires a very high heat, obtainable only with a good blast-lamp, a SEFSTROM fur- nace, a DEVILLE turpentine lamp, etc. ; even the highest temper- ature afforded by a wind furnace is insufficient to effect the melt- ing together of the barium carbonate and silicate, and only when this occurs is decomposition complete. When this does occur, however, it is so energetic that even the most refractory fossils are easily and completely decomposed. From 4 to 6 parts of barium carbonate are taken for 1 part of the very finely powdered mineral. The fusion is effected in a platinum crucible, which, if a SEFSTROM furnace is used, is placed within another crucible of refractory fire-clay filled with magnesia. The crucible is left in the furnace for at least half an hour. The greater the quan- tity of barium carbonate taken the greater is the danger of alka- lies volatilizing. DEVILLE, in fact, recommends taking only 0*8 part barium carbonate for 1 part of felspathic mineral. More readily decomposable minerals may be more easily de- composed by means of barium hydroxide freed from its water of crystallization. To 1 part of the mineral, from 4 to 5 parts of barium hydroxide are taken, the whole intimately mixed and covered with a layer of barium carbonate. The fusion may be effected over an ordinary gas- or BERZELIUS alcohol -lamp ; and it is best to use silver crucibles, as platinum is attacked. The mass fuses either completely or at least melts together into a mass. In order to render platinum crucibles also applicable, v. FELLEN- BERtt-RrviER f recommends melting 4 or 5 parts of calcium chlo- ride in the platinum crucible, shaking the crucible around while cooling, then adding 1 part of barium hydroxide and fusing this in turn. After cooling, 1 part of the finely powdered silicate is introduced and heat is applied, gently at first, but strongly later when no gas appears to be evolved. SMITH $ recommends fusing *Zeil8chr.f. analyt. Chem , vn, 463. \ 2b., ix, 459. \Journ.f. prakt. Chem., LX, 246. 518 DETERMINATION. [ 140. 1 part silicate with 3 to 4 parts barium carbonate and 2 parts barium chloride. When the operation is at an end no matter whether barium carbonate or barium hydroxide has been used allow the crucible to cool, clean its outside, cover it with 10 to 15 parts of water in a beaker, allow to macerate for some time, then add hydrochloric or nitric acid, and proceed as in &, a. Care must be taken to avoid adding too much hydrochloric acid at a time, because the barium chloride is difficultly soluble in it, and hence may retard or check the solution of the still un dissolved portion by forming over this an insoluble protective coating. In the filtrate from silicic acid the bases are estimated according to the methods given under Section Y. The silicic acid is to be tested as to its purity according to the method described in #, before the operation may be regarded as having been successful. These methods, which were formerly much employed in determining the alkalies in sili- cates, have been more or less superseded by decomposition with aqueous hydrofluoric acid and with ammonium fluoride, as both of these are now readily obtainable commercially. d. Decomposition ~by fusion with Calcium Carbonate and Ammonium Chloride. DEVILLE * recommends fusing 1 part of the powdered silicate with 0'3 to 0'8 part calcium carbonate, but I have not found the process to answer with many silicates. L. SMITH f recommends fusing 0*5 to 1 grm. of powered silicate with 1 grm. finely granu- lated ammonium chloride (prepared by interrupted crystallization) and 8 grm. pure calcium carbonate (obtained by precipitation with ammonium carbonate with heat). Should the temperature during fusion rise too high, however, a portion of the alkali chloride may be lost by volatilization. SMITH employs crucibles 9 mm. high, 22 mm. wide at the mouth, and 16 mm. wide at the base. The crucibles are fixed in a metal clamp or in the iron plate of a special gas furnace,^: and in such a manner that about 15 mm. remains outside. Gentle heat is applied first to * Journ. f. prakt. Chem., LX, 246. \ Ib., LX, 246; also Chem. News, xxnt, 222 and 234; Zeitschr. /. analyt. Chem., xi, 85. % Zeitschr. f. analyt. Chem., xi, 87. 140.] SILICIC ACID. 519 the part of the crucible above the mixture, and is then gradually- moved downwards, so that in about 5 minutes all the ammonium chloride is decomposed. The heat is then increased and the crucible kept at a bright-red heat for 40 to 60 minutes. By this method of heating all fear of volatilization of alkali chloride is avoided. After cooling, proceed with the semi-fused mass accord- ing to y. SMITH states, however, that a solution of the total alkalies may also be obtained by heating the ignited mass with water for several hours, filtering, and washing the residue. From this solution of alkalies, calcium chloride, and calcium hydroxide, the calcium is precipitated by ammonium carbonate and a little oxalate. [Prof. J. L. SMITH'S METHOD in detail for separating alkalies : Mix 1 part of the pulverized silicate with 1 part of dry ? mi no- mum chloride, by gentle trituration in a smooth inortar, then add 8 parts of calcium carbonate (" Qual. Anal." p. 87) and mix inti- mately. Bring the mixture into a platinum crucible, rinsing the mortar with a little calcium carbonate. Warm the crucible gradu- ally over a small Bunsen burner until fumes of ammonium salts no longer appear, then heat with the flame of a Bunsen burner until the lower three-fourths only of the crucible are brought to a red heat. Keep this temperature constant from 40 to 60 minutes. The temperature desired is that which suffices to keep in state of fusion the calcium chloride formed by the reaction of ammonium chloride with calcium carbonate. The mass, however, does not become liquid since the fused calcium chloride is absorbed by the large quantity of calcium carbonate present. If the silicate is fused by application of too strong heat, disintegration of the mass at the end of the operation with water cannot be effected. More- over, too high a temperature causes volatilization of alkali chlo- rides. Certain silicates e.g., those which contain much ferrous iron may fuse when heated with the above mixture, even if no higher temperature is employed than is necessary to effect decom- position. If this occurs, it is better to repeat the ignition with a new portion jof the silicate, using 8 to 10 parts of calcium carbo- nate. The mass contracts in volume during the ignition, and is usually easily detached from the crucible. Boil it in a covered porcelain dish, with 50-75 c.c. water, half an hour, replacing water lost by evaporation. Decant the solution from the residue upon a filter, boil the residue a few minutes with water, and decant again. 620 DETERMINATION. [ 140. If the residue is now all in a finely disintegrated state, it may be brought upon the filter and washed. But if, as is often the case, a portion remains coherent or in a coarsely granular state, it must be reduced to a fine state of division by trituration with a porcelain or agate pestle in the dish, and boiling with water again. By a few repetitions of the trituration, boiling and decanting, allowing the fine suspended portion to pass upon the filter each time, the whole can usually be transferred to the filter in properly disinte- grated condition in course of an hour. Next wash until a few drops of the washings acidified with nitric acid give but a slight turbid- ity with silver nitrate. The filtrate now contains the alkalies of the silicate as chlorides together with calcium chloride and hydrox- ide. It is not advisable to concentrate this filtrate in a glass vessel, since it might take an appreciable quantity of sodium from the glass. Precipitate, therefore, the calcium at once with ammonium carbonate ; allow the precipitate, to settle, and concentrate the supernatant solution in a porcelain (or platinum) dish, decanting it into the latter, portionwise if necessary, rinsing finally the precipi- tate into the porcelain dish. When the whole is thus reduced to about 30 c.c., add a little more ammonium carbonate and ammonia, heat and filter into a platinum (or porcelain) dish, evaporate to dry ness on a water-bath, expel ammonium chloride by ignition, dissolve the residual alkali chlorides in 3 to 5 c.c. of water. A little black or dark-brown flocculent matter usually remains undis- solved, while the solution may still contain traces of calcium. Add two or three drops of ammonium carbonate and ammonia, warm gently, and filter through a very small filter into a weighable plati- num vessel. Evaporate to dryness on a water-bath, heat to in- cipient fusion of the alkali chlorides, and after cooling weigh. Prof. SMITH'S method is the most convenient of all methods for extracting alkalies from silicates, and is universally applicable, except perhaps in presence of boric acid. When carried out as here described, the results are sufficiently accurate in most cases. If, however, the silicate is rich in alkalies, a loss amounting to 0-1 or 0-2 per cent of the mineral is possible. If great accuracy is desired in such cases, a repetition of the whole process may be applied to the residue left by treatment of the ignited mass with water. It need hardly be mentioned that unless care be taken to use reagents perfectly free from soda, and to avoid the action of solu- 141.] CHLORINE. 521 tions on glass, an amount of soda may be introduced from these sources equal to O'l or 0*2 per cent of the silicate.] e. Decomposition with Hydrochloric or Sulphuric Acid in Sealed Tubes (under Pressure), according to AL. MlTSCHERLICH.* Many silicates, as well as altiminates, which are scarcely or not at all attacked on being digested with hydrochloric or sulphuric acid in open vessels, are completely decomposed on being heated for two hours in sealed tubes to 200 to 210 with 25-per cent hydrochloric acid or with a mixture of 1 part by weight of water and 3 parts by weight of concentrated sulphuric acid. To carry out the process introduce about 1 grm. of the finely elu- triated or sifted substance into a tube of difficultly fusible Bohe- mian glass sealed at one end and drawn out somewhat at the other ; then add the acid, carefully seal the tube, enclose it in the wrought-iron tube of a metallic bath,f and heat in the manner prescribed. After cooling, carefully open the tube, rinse out its contents into a platinum or porcelain dish, and proceed as in 140, II, a. This method possesses the advantage over others that any ferrous salt present is obtained in solution as such and can be readily determined. Second Group. CHLORINE BROMINE IODINE CYANOGEN SULPHUB. 141. 1. CHLORINE. I. Determination. Chlorine may be determined very accurately in the gravimetric as well as in the volumetric way.;f a. Gravimetric Method. Determination as Silver Chloride. Solution of silver nitrate, mixed with some nitric acid, is added in excess to the solution of the chloride, the precipitated chloride is made to unite by heating and agitating, washed by decantation * Jour ii, f. prakt. Chem., LXXXI, 108, and LXXXIII, 455. fSuch a bath is described and illustrated in the Joitrn. f. prakt. Chem., LXXXIII, 489, and in the ZeitscJir. f. analyt. C/tem., i, 55. \ For the acidimetric estimation of free hydrochloric acid, see 215. DETERMINATION. [ 141. and filtration, dried, and ignited. The details of the process have been given in 115, 1, a. Care must be taken not to heat the solution mixed with nitric acid, before the nitrate of silver has been added in excess. As soon as the latter is present in excess, the silver chloride separates immediately and completely upon shaking or stirring, and the supernatant fluid becomes perfectly clear after standing a short time in a warm place. The determina- tion of chlorine by means of silver is therefore more readily effected than that of silver by means of hydrochloric acid. 5. Volumetric Methods. a. By Solution of Silver Nitrate. In 115, 5, we have seen how the silver in a fluid may be esti- mated by adding a standard solution of sodium chloride until no further precipitation ensues ; in the same way we may determine also, by means of a standard solution of silver, the amount of hydro- chloric acid in a fluid, or of chlorine in combination with a metal. PELOUZE has used this method for the determination of several atomic weights. LEVOL* proposed a modification which serves to indicate more readily the exact point of complete precipitation. To the fluid, which must be neutral, he added one tenth volume of a saturated solution of sodium phosphate. When the whole of the chlorine has been precipitated by the silver, the further addi- tion of the solution of silver produces a yellow precipitate which does not disappear upon shaking the vessel. FE. MOHE has since replaced, with the most complete success, the sodium phosphate by potassium chromate. This convenient and accurate method requires a perfectly neu- tral solution of silver nitrate of known value. The strength most convenient is, 1 litre = 0*1 at. Cl. I recommend the following method of preparation : Dissolve 18-80 to 18-85 grin, pure fused .silver nitrate in 1100 c.c. water, and filter the solution if required ; the solution is purposely made too strong at first. Now weigh off exactly four portions of pure sodium chloride, each of 0' 10 to 0-18 grm., one after another. The salt should be moderately ignited, not fused, powdered roughly while still warm, and introduced into a small dry tube, that can be well closed. The weighing off is per- formed by first weighing the filled tube, then shaking out into a dry" beaker the quantity required, weighing again, dropping a * Jo-urn./, prakt. Chem., LX, 384. 141.] CHLORINE. second portion into beaker No. 2, weighing again, and so on. Each portion is dissolved in 20 to 30 c.c. water, and about 3 drops of a cold saturated solution of pure normal potassium chromate added. Fill a MOHE'S burette (in very accurate analysis an ERDM ANN'S float should be used) with the silver solution, and run it slowly, with constant stirring, into the light yellow solution contained in one of the beakers. Each drop produces, where it falls, a red spot, which on stirring disappears, owing to the instant decomposition of the silver chromate w r ith the sodium chloride. At last, how- ever, the slight red coloration remains. Now all chlorine has com- bined with silver, and a little silver chromate has been permanently formed. Read off the burette and reckon how much silver solu- tion would have been required for O'l mol. sodium chloride, i.e., 5-85 grm. Suppose we have used to 0*11 sodium chloride 18*7 c. c. silver solution : 0-11 : 5-85 :: 18'T : x\ x= 994-5. Now, without throwing away the contents of the first beake* , make a second and third experiment in the same manner, of course always taking notice to regard the same shade of red as the sign of the end. The results of these are reckoned out in the same way as the first. Suppose they gave for 5 ; 85 NaCl 995*0 and 993-0 respectively, we take the mean of the three numbers, which is 994-2, and we now know that we have only to take this number of c.c. of silver solution, and make it up to 1000 c.c. with 5-8 water, in order to obtain a solution of the required strength, i.e., 1000 c.c. = 0-1 mol. NaCl. But if 994-2 requires 5'8 water, 1000 requires 5-83. Hence we fill a litre-flask (previously dried or rinsed with a small portion of the solution) up to the " holding " mark with the solution, add 5*83 c.c. water, insert a caoutchouc stopper, and shake. The solution must now be correct ; however, to make quite sure, we perform another experiment with it. To this end rinse the empty burette with the new solution, fill it with the same and test with the portion of salt in beaker No. 4. The c.c. used of silver solution must now, if multiplied by 0-00585, give exactly the weight of the salt. Being now in possession of a standard silver solution, and bei'iiir practised in exactly hitting the transition from yellow to the sh:idr 524 DETERMINATION. [ 141. of red, we are in the position to determine with precision chlorine in the form of hydrochloric acid or of a metallic chloride soluble in water. The fluid to be tested must be neutral free acids dissolve the silver chromate. The solution of the substance is therefore, if necessary, rendered neutral by addition of nitric acid or sodium carbonate (it should be rather alkaline than acid), about 3 drops of the solution of chromate added, and then silver from the burette, till the reddish coloration is just perceptible. The number of c.c. used has only to be multiplied by the atomic weight of chlorine or the mol. weight of the metallic chloride and divided by 10,000 to give the amount of these respectively present. If the operator fears he has added too much silver solution, i.e., if the red color is too strongly marked, he may add 1 c.c. of a solu- tion of sodium chloride containing 5*85 in a litre (and therefore corresponding to the silver solution), and then add the silver drop by drop again. Of course in this case 1 c.c. must be deducted from the amount of silver solution used. The results are very satisfactory. The fluid to be analyzed should be about the same volume as the solutions employed in standardizing the silver solution, and also about the same strength, otherwise the small quantity of silver which produces the colora- tion w r ill not stand in the same proportion to the chlorine present. This small quantity of silver solution is extremely small, varying between 0'05 and '1 c.c. : the inaccuracy hereby arising even in the case of quantities of chlorine differing widely from that originally used in standardizing the silver solution is therefore almost incon- siderable. If the amount of silver solution necessary to impart the coloration always remained the same, we should have simply to deduct the amount in question in all experiments, in order to avoid this small inaccuracy entirely ; since, however, the greater the quantity of silver chloride the more silver chromate is required for visible coloration, this method of proceeding would not increase the exactness of the results. ft. By Solution of Silver Nitrate and Iodide of Starch (PISANI'S method*). Add to the solution of the chloride, acidified with nitric acid, a slight excess of standard solution of silver nitrate, warm, and filter. Determine the excess of silver in the filtrate by means of solution * Annal. d. Mines, \, 83; LTEBIG and KOPP'S Jahrefbericht, 1856, 751. 141.] CHLORINE. 525 of iodide of starch (see p. 349), and deduct tliis from the amount of silver solution used. The difference shows the quantity of silver which has combined with the chlorine; calculate from this the amount of the latter. Results satisfactory. y. By Mercuric- Nitrate Solution (LiEBio's method,* par- ticularly recommended for estimating chlorine in the chlorides in urine). aa. Principle of the method. Mercuric-nitrate solution causes an immediate, dense, white precipitate in a solution of urea; mercuric chloride, however, does not. On mixing a mercuric- nitrate solution with an alkali chloride, mercuric chloride and alkali nitrate are formed. Hence on adding sodium chloride to a urea solution and then dropping in a dilute mercuric-nitrate solu- tion a white cloudiness forms at the point where the drops fall, but on shaking it disappears immediately so long as the mercu- ric nitrate continues to react with the sodium chloride. The moment the double decomposition is complete, however, an additional drop of mercuric-nitrate solution causes a permanent turbidity. Hence, if the volume and strength of the mercuric- nitrate solution added be known, the chlorine strength of the salt solution is also known, since 1 eq. of mercury in the mercuric solution corresponds to 2 eq. of chlorine. bb. Preparation of the mercuric-nitrate solution. As this solution must be perfectly free from other metals, it is advisable to prepare it from mercury oxide obtained by precipitating mer- curic chloride with soda solution and thoroughly washing the precipitate. 10 '8 grin, of the dried oxide so obtained are dissolved in nitric acid, the solution evaporated to a syrupy consistency, and then diluted to 550 c. c. with water. The solution may also be made by dissolving repeatedly recrystallized mercurous nitrate in water, with the addition of some nitric acid, boiling, adding strong nitric acid until red fumes no longer are evolved, evapo- rating to syrupy consistency, and diluting with enough water to yield a solution of approximately correct strength. cc. Determining the strength of the solution. This is effected by means of a sodium-chloride solution of known strength, pre- pared according to LIEBIG by mixing 20 c. c. of a saturated (at * Annal. d. Chem. u. Pharm., LXXXV, 297. 526 DETERMINATION. [ 141* ordinary temperatures) solution of pure rock salt or chemically pure sodium chloride, with 298 '4 c. c. water. Every c. c. of the solution will contain 20 mg. of sodium chloride. Of this solution measure 10 c. c. into a beaker and add 3 c. c. of a urea solution containing 4 grm. in every 100 c. c. Drop the mercury solution to be standardized from a burette into the mixture, with shaking, until a just perceptible precipitate, which fails to dissolve on shaking, forms.* dd. Having thus ascertained how many c. c. of mercuric- nitrate solution are equivalent to 10 c. c. of sodium-chloride solu- tion (=0*2 grm. NaCl), the mercuric solution is applicable for immediate use, if a little calculation is not objected to. If this is rather avoided, dilute the mercuric solution so that every c. c. may correspond to a given number of milligrammes of sodium chloride or chlorine. LIEBIG dilutes the solution so that 1 c. c. corresponds to 0*01 grm. of sodium chloride. ee. If the test-fluid is to be used for testing solutions which contain much foreign salts or an excess of urea, add to 10 c. c. of the sodium- chloride solution 3 c. c. of the urea solution and also 5 c. c. of a cold saturated sodium-sulphate solution before drop- ping in the mercuric solution f . Results accurate. d. Alkalimetrically (according to BOHLJG ). Add to the solution, if necessary, potassium carbonate in not too great excess, to precipitate the alkali earths, earths, or metallic oxides, dilute to 250 c. c., mix, filter, and determine the alkalinity of 50 c. c. of the filtrate according to 220. To 125 c. c. of the filtrate in a 250- c. c. flask add an excess of pure silver oxide, fill to the mark with water and shake repeatedly, with exclusion of light. After a few minutes filter through a dry folded filter, pipette oft* 100 c. c. of the filtrate (corresponding to 50 c. c. of the original liquid),, and determine its alkalinity also. The difference in the c. c. of * A mere opalescence of the fluid is to be disregarded, as this depends upon a trace of foreign metals, and has no bearing on the reaction, as may be readily seen from the fact that the cloudiness is not increased by a further addition of the mercuric solution. t The reason for this addition is that the mercuric nitrate and urea are more readily soluble in pure water than in saline solution, hence the solvent powers oi the solutions should be as nearly alike as possible when standardizing and wh^n> performing the analysis, if accurate results are desired. \Zeitschr.f. analyt. Chem., ix, 314 141.] CHLOKINE. 527 standard acid used in the two determinations of alkalinity cor- responds to the chlorine content of the solution. The result is naturally correct only when another portion of the filtrate has been tested and found free from chlorine. BOHLIG'S method is particularly well adapted for technical purposes. Of these volumetric methods of estimating chlorine, the first deserves the preference in all ordinary cases. It is not, however, applicable in urinalysis, because compounds of the silver oxide, with coloring matters, etc. , are precipitated with the silver chloride (C. NEUBAUEK*). PISANI'S method (b, ft) is especially suited for the estimation of very minute quantities of chlorine, but is not applicable when, as in nitre analyses, large quantities of alkaline nitrate are present (p. 344). II. Separation of Chlorine from the Metals, a. In Soluble Chlorides. The same method as in I, a. The metals in the filtrate are separated from the excess of the salt of silver by the methods which will be found in Section Y. Chlorides soluble in water may also be completely decomposed by cold digestion with oxide or carbonate of silver. Silver chloride is obtained, while the metal combined with the chlorine is converted into oxide or carbonate and either remains in solution or falls down with the silver chlo- ride. Take care that no traces of oxide or carbonate of silver pa~*$ into the filtrate. Stannous chloride, mercuric chloride, platinic chloride, the chlorides of antimony, and the green chloride of chromium, form exceptions from the rule. a. From stannic chloride, silver nitrate would precipitate, besides silver chloride, a compound of stannic oxide and silver oxide. To precipitate the tin, therefore, the solution is mixed with concentrated solution of ammonium nitrate, boiled, allowed to deposit, decanted, and filtered (compare 126, 1, J), and the chlo- rine in the filtrate is precipitated with solution of silver. LOWEN- THAL, the inventor of this method, has proved its accuracy. f * To apply this method to urine also, R. PRIBRAM (Zeitschr. f. analyt. Chem., ix, 428) heats 10 c. c. of urine with 50 c. c. of a solution of pure potassium permanganate (1 or 2 : 1,000) to- gentle boiling, filters off the brown flocks, washes these, and determines the chlorine in the filtrate according to b, a. \Journ. /. prakt. Chem., LXVJ, 371. 528 DETERMINATION. [ 141. ft. When 'mercuric chloride is precipitated with solution of silver nitrate, the silver chloride thrown down contains an admix- ture of mercury. The mercury is, therefore, first precipitated by hydrogen sulphide, and the chlorine in the filtrate determined as directed in 169. y. The chlorides of antimony are also decomposed in the man- ner described in ft. The separation of basic salt upon the addi- tion of water may be avoided by addition of tartaric acid. The antimonous sulphide should be tested for chlorine. $. Solution of silver fails to precipitate the whole of the chlo- rine from solution of the green chloride of chromium (PELIGOT). The chromium is, therefore, first precipitated with ammonia, the fluid filtered, and the chlorine in the filtrate precipitated as in I., a. e. From platinic chloride silver nitrate throws down a com-, pound of platinons chloride and silver chloride (COMAILLE *). We may either ignite the platinic chloride in a current of hydrogen and pass the hydrochloric acid produced into solution of silver (BONSDORFF), or we may evaporate the solution with sodium car- bonate, fuse the residue in a platinum crucible, and determine the chloride in the aqueous solution of the fusion. Or, thirdly, we may (after TOPSOE f) digest the moderately dilute solution in the cold with zinc clippings till hydrogen ceases to escape, add ammo- nia in excess, heat on a water-bath till the fluid is fully decolorized, all the platinum being precipitated, and finally determine the chlo- rine in the filtrate. &. In Insoluble Chlorides. a. Chlorides soluble in Nitric A.cid. Dissolve the chloride in nitric acid, without applying heat, and proceed as in I., a. ft. Chlorides insoluble in Nitric Acid (lead chloride, silver chloride, mercurous chloride). aa. Lead chloride is decomposed by digestion with alkali hydrogen carbonate and water. The process is exactly the same as for the decomposition of lead sulphate ( 132, II., &, ft). ~bb. Silver chloride is ignited in a porcelain crucible, with 3 parts of sodium and potassium carbonate, until the mass com- mences to agglutinate. Upon treating with water, the metallic * Zeitschr.f. analyt. Chem., vi,,121. * f7&., ix, 30. 142.] CHLORINE. 529 silver is left undissolved ; the solution contains the alkali chloride, which is then treated as in I., a. Silver chloride may also be readily decomposed by long diges- tion with pure iron (reduced by hydrogen) and dilute sulphuric acid. Zinc may be used instead of iron, but it does not answer so well. The separated metallic silver may be washed, heated with dilute sulphuric acid, washed again and weighed ; it must after- wards be ascertained, however, whether it dissolves in nitric acid. The chlorine is determined in the chloride of iron or zinc as in L,a. cc. Mercurous chloride is decomposed by digestion with solu- tion of soda or potassa. The hydrochloric acid in the filtrate is determined as in I., a. The mercurous oxide is dissolved in nitric or nitrohydrochloric acid, and the mercury determined as directed in 117 or 118. c. The soluble chlorides of the metals of the fourth, fifth, and mxth groups may generally be decomposed also by hydrogen sul- phide or ammonium sulphide. The chlorine in the filtrate is determined as in 169. It must not be omitted to test the pre- cipitated sulphides for chlorine. Several chlorides, cadmium chlo- ride for instance, give sulphides free from chlorine with ammonium sulphide, but not with hydrogen sulphide. d. In many metallic chlorides, for instance in those of the first and second groups, the chlorine may be determined also by evapo- rating with sulphuric acid, converting the metal thus into a sul- phate, which is then ignited and weighed as such ; the chlorine being calculated from the loss. This method is not applicable in the case of silver chloride and lead chloride, which are only imper- fectly and with difficulty decomposed by sulphuric acid ; nor in the case of mercuric chloride and stannic chloride, which sulphuric acid fails almost or altogether to decompose. Supplement. 142. DETERMINATION OF CHLORINE IN THE FEEE STATE. Chlorine in the free state may be determined both in the volu- metric and in the gravimetric way. The volumetric methods, however, deserve the preference in most cases. They are very numerous. 530 DETERMINATION. [ I shall only here adduce that one which is undoubtedly the most accurate and at the same time the most convenient.* 1. Volumetric Method. With Potassium Iodide (after BUNSEN). Bring the chlorine, in the gaseous form or in aqueous solution^ into contact with an excess of solution of potassium iodide in water. Each at. chlorine liberates 1 at. iodine, which remains dissolved in the excess of potassium iodide. By determining the liberated iodine by means of sodium thiosulphate as in 146, you will accordingly learn the quantity of chlorine, and, in fact, with the greatest accuracy. If you have to determine the chlorine of chlorine water, measure a portion off with a pipette. So as to prevent any of the gas entering the mouth, connect the upper end of the pipette with a tube containing moist potassium hydroxide laid between cotton. When the pipette has been correctly filled allow its contents to flow, with stirring, into an excess of solution of potassium iodide (1 in 10). There is no difficulty about knowing whether the latter is sufficiently in excess, for if not, a black precipitate is formed. If the chlorine is evolved in the gaseous condition, you may employ either the apparatus given in 130, I, e^ /?, or the following, Fig. 103, Pig. 103. which is especially suitable where the chlorine is not pure, but is mixed with other gases. * Compare " Chlorimctry " in the [Special Part. 142.] CHLORINE. 531 a is a little flask, from which the chlorine is evolved by boiling the substance with hydrochloric acid, a small lump of magnesite being added ; it is connected with the tube, b, by means of a flexible tube. The latter must be free from sulphur should it contain sulphur it is well boiled with dilute potassa and then thoroughly washed. The thinner tube, c, which has been fused to the bulb of b, leads through the caoutchouc stopper (which has been deprived of sul- phur) to the bulbed U-tube, d, which contains solution of potassium iodide, and which for safety is connected with the plain U-tube, e, also containing potassium iodide solution. Both tubes stand in a beaker filled with water. The apparatus offers the advantages that the fluid cannot return, that the potassium iodide remains cold, and that the absorption is complete. After all the chlorine has been expelled by boiling long enough, rinse d and e out into a beaker and titrate with standard sodium thiosulphate ( 146). 2. Gravimetric Method. The fluid under examination, which must be free from sulphu- ric acid, say, for instance, 30 grin, chlorine water, is mixed in a stop- pered bottle, with a slight excess of sodium thiosulphate, say 0*5 grin., the stopper inserted, and the bottle kept for a short time in a warm place ; after which the odor of chlorine is found to have gone off. The mixture is then heated to boiling with some hydro- chloric acid in excess, to destroy the excess of sodium thiosulphate, filtered, and the sulphuric acid in the filtrate determined by barium chloride ( 132). 1 mol. sulphuric acid corresponds to 4 at. chlorine (WlCKE*). In fluids containing, 'besides free chlorine, also hydrochloric acid, or a metallic chloride, the chlorine existing in a state of combination may be determined, in presence of the free chlorine, in the follow- ing way : A weighed portion of the fluid is mixed with solution of sulphur- ous acid in excess, after some time nitric acid is added, and then potas- sium chromate to destroy the excess of sulphurous acid, and the whole of the chlorine is precipitated as silver chloride. The quantity of the free chlorine is then determined in another weighed portion, by means of potassium iodide ; the difference gives the amount of combined chlorine. f * Annal. d. Chem. u. Pharm., xcix, 99. f If chlorine water is mixed at once with silver nitrate, | only of the chlorine 532 DETERMINATION. [ 143. Having thus seen in how simple and accurate a manner the quantity of free chlorine may be determined by BTJNSEN'S method, it will be readily understood that all oxides and peroxides which yield chlorine when heated with hydrochloric acid, may be analyzed by heating them with concentrated hydrochloric acid, with addition of a small lump of magnesite, and determining the amount of chlorine evolved. As regards the procedure, compare 142, 1. 143. 2. BROMINE. I. Determination. a. Gravimetric Methods. Estimation as silver bromide. Free hydrobromic acid in & solution free from hydrochloric acid or chlorides is precipitated by silver solution, and the further process is conducted as in the case of hydrochloric acid ( 141). For the properties of silver bro- mide, see 94, 2. The results are perfectly accurate. b. Volumetric Methods. Like chlorine in hydrochloric acid and alkali chlorides, bromine may be estimated in the analogous compounds by standard silver solution ( 141, 1., b, <*), by solution of silver and iodide of starch ( 141, I., 5, /?), and also alkalimetrically ( 141, I., 6, 6). But these methods are seldom applicable, as they cannot be used in the presence of hydrochloric acid and metallic chlorides. The following methods must therefore be detailed ; they are especially useful for the estimation of small quantities of bromine in solutions containing chlorides, but in point of accuracy they leave much to be desired.* of. With chlorine water and chloroform (after A. REiMANxf). This method depends on the facts that chlorine, when added to bromides first liberates the bromine and then combines with it, and that bromine colors chloroform yellow or orange, while bromine is obtained as silver chloride : 6C1 -f 3Ag 2 O 5AgCl + AgClO 3 (H. ROSE, WELTZIEN, Annal. d. Chem. u. Pharm., xci, 45). If chlorine water is mixed with ammonia in excess, there are formed at first ammonium chloride and am- monium hypochlorite; the latter then gradually decomposes into nitrogen and ammonium chloride. However, a little ammonium chlorate is also formed be- sides (SciioNBEiN, Journ. /. prakt. Cliem., LXXXIV, 386 ; Zeitschr. /. analyt. Chem., u, 59). * Compare 169. f Annal. a. Onem. u. Pharm., cxv, 140. 143.] BROMINE. 533 chloride merely communicates a yellowish tinge to that fluid. The process is as follows : Mix the liquid containing a bromide of an alkali metal in neutral solution, in a stoppered bottle with a drop of pure chloroform about the size of a hazel-nut, then add standard chlorine water from a burette, protected from the light by being surrounded with black paper. On shaking, the chloroform becomes yellow, on further addition of chlorine water, orange, then yellow again, and lastly at the moment when 2 at. chlorine ha've been used for 1 at. bromine yellowish white (KBr -f- 2C1 = KC1 + BrCl). Considerable practice and skill are required before the operator can tell the end-reaction. He will be assisted by placing the bottle on white paper and comparing the color of the chloro- form with that of a dilute solution of yellow potassium chromate of the required color. The strength of the chlorine water should depend on the amount of the bromine to be determined. It should be so adjusted that about 100 c.c. may be used. The chlorine water is standardized with potassium iodide and sodium thiosulphate ( 142, 1). The method is especially suited for the determination of small quantities of bromine in mother liquors, kelp, &c. The results are approximate: e.g., 0*018 instead of 0*0185 0*055 instead of 0*0590*0112 instead of 0*01, &c. If the fluid con- tains organic substances, it is after being rendered alkaline with caustic soda evaporated to dryness, the residue ignited in a silver dish, extracted with water, the solution neutralized exactly with hydrochloric acid, and then tested. /3. With cMorine water and heat (after FIGUIER*). The principles underlying this method are that in a solution of a metallic bromide, 1 eq. of bromine is liberated by 1 eq. of chlorine, and that bromine gives an aqueous yellow solution from which it readily escapes on boiling and leaves behind a colorless liquid . The chlorine is used in the form of a dilute solution. It must be standardized immediately before use, its strength being deter- mined by its action on a sodium-bromide solution of known strength acidulated with a few drops of hydrochloric acid (or more simply by testing with potassium iodide and sodium thiosulphate according to 142, 1). The mother-liquor is heated almost to, * Annal. d. Chem. et de Pkys., xxxm, 303; Journ. f. prakt. Chem., LIV, 293. Proposed for the determination of bromine in mother-liquors. 534 DETERMINATION. [ 143. boiling in a flask, then the chlorine water is allowed to flow in from the burette covered with black paper, and the mixture heated for about three minutes, whereby the liquid becomes again color- less. After allowing to cool for two minutes, drop some more chlorine water into the mixture and continue in this manner until the liquid is no longer colored on adding chlorine water. If the experiments take several hours to carry out, titrate the chlorine w^ater again at the close of the operations and base the calcula- tions on the mean of the two chlorine determinations. Alkaline fluids should be acidulated with a little hydrochloric acid. Fer- rous and manganous salts, iodine, and organic substances must be absent. Mother-liquors colored yellow by organic matters are best evaporated to dry ness, gently ignited, arid the residue treated with water and filtered. On evaporating the solution to dryness, sodium carbonate must be added, because in this process magnesium chloride or bromide evolves hydrochloric or hydrobromic acid. According to my investigations, the process is best carried out by heating the mother-liquor in a flask the stopper of which has three perforations. Through one of these carbonic acid is con- ducted nearly to the bottom of the flask and escapes, together with the liberated bromine, through another, while the chlorine water is introduced through the middle perforation, in which is inserted the somewhat elongated tip of the burette. The process is carried out while the liquid is kept gently boiling. The deter- minations may in this manner be quite rapidly carried out, and then afford satisfactory results. ft. HEINE'S colorimetric method.* The bromine is liberated by means of chlorine, and taken up with ether; the solution is compared, with respect to color, with an ethereal solution of bro- mine of known strength, and the quantity of bromine in it thus ascertained. FEHLING f obtained satisfactory results by this method. It will at once be seen that the amount of bromine contained in the fluid to be analyzed must be known in some measure before this method can be resorted to. As the brine examined by FEHL- ING could contain at the most 0*02 grm. bromine in 60 grm., he prepared ten different test fluids by adding to ten several portions of 60 grm. each of a saturated solution of common salt, increasing *Journ.f. prakt. Chem., xxxvi, 184. Proposed to effect the determination of bromine in mother- liquors. \Journ.f. prakt. Chem., XLV, 269. 143.J BROMINE. 535 quantities of potassium bromide, containing respectively from 0*002 grm. to 0-02 grm. bromine. He added an equal volume of ether to the test fluids, and then chlorine water, until there was no fur- ther darkening observed in the color of the ether. It being of the highest importance to hit this point exactly, since too little as well as too much chlorine makes the color appear lighter, FEHLING pre- pared three samples of each test fluid, and then chose the darkest of them for the comparison. 60 grm. are now taken * of the mother liquor to be examined, the same volume- of ether added as was added to the test fluids, and then chlorine water. Every experiment is repeated several times. Direct sunlight must be avoided, and the operation conducted with proper expedition. In my opinion it is well to replace the ether by chloroform or car- bon disulphide. CAIGNET f substituted sodium hypochlorite for the chlorine water, and removed the colored carbon disulphide from time to time. II. Separation of Bromine from, the Metals. The metallic bromides are analyzed exactly like the correspond- ing chlorides ( 141* II., a to d) 9 the whole of these methods being applicable to bromides as well as chlorides. In the decomposition of bromides by sulphuric acid ( 141, II., $), porcelain crucibles must be used instead of platinum ones, as the latter would be attacked by the liberated bromine. Some bromides, it must be remembered, are not completely decomposed by sulphuric acid ; for instance, mercuric bromide is not. The soluble bromides may be converted into chlorides by evaporation with hydrochloric acid and excess of chlorine water ; but this process cannot be applied where the chloride is liable to be carried away with the steam ; for instance, in the case of mercuric bromide. * The best way is to take them by measure, f ZeiUclir. /. analyt. Chem., ix, 427. 536 DETERMINATION. [ 144, 145. Supplement. DETERMINATION OF FREE BROMINE. Free bromine in aqueous solution, or evolved in the gaseous form, is caused to act on excess of solution of potassium iodide. Each at. bromine liberates 1 at. iodine, which is most conveniently determined by means of sodium thiosulphate ( 146). As regards the best mode of bringing about the action of the bromine on the potassium iodide, compare 142, 1. The determination of free bromine in presence of hydrobromie acid or metallic bromides is effected in the same manner as that of free chlorine in presence of hydrochloric acid (see 142). 145. 3. IODINE. I. Determination* a. Gravimetric Methods. a. Estimation as silver iodide. If you have hydriodic acid in solution, free from hydrochloric and hydrobromie acids, precipitate with silver nitrate, and proceed exactly as with hydrochloric acid ( 141). If the solution is colored with free iodine, first add sulphurous acid cautiously till the color is removed. The particles of silver iodide adhering to the filter are not reduced on incinera- tion, but a little of the iodide is liable to volatilize if the heat is too high. Hence the filter should be got as clean as possible, and the heat during incineration should not be unduly raised. For the properties of silver iodide, see 94, 3. The results are perfectly accurate. ft. Estimation as palladious iodide. The following method, recommended first by LASSAIGNE, is resorted to exclusively to effect the separation of iodine from chlorine and bromine, for which pur- pose it is extremely well adapted. The solution may not contain any alcohol. Acidify it slightly with hydrochloric acid, and add a solution of palladious chloride, as long as a precipitate forms ; let * For the methods to be adopted in the presence of bromine and chlorine,, see 169. 145.] IODINE. 537 the mixture stand from 2-t to 48 hours in a warm place, filter the brownish-black precipitate off on a weighed filter, wash with warm water, and dry at 100, until the weight remains constant. For the properties of the precipitate, see 94, 3. This method gives very accurate results. Instead of simply drying the palladious iodide, and weighing it in that form, you may ignite it in a current of hydrogen in a crucible of porcelain or platinum,* and calculate the iodine from the residuary palladium (H. ROSE). Compare 122, 1. b. Volumetric Methods. a. The methods given for hydrochloric acid by precipitating with silver solution ( 141, I., b, a); by silver solution and iodide of starch (141, I., 6, /?), and also alkalimetrically ( 141,, I., 6, tf), may be used for hydriodic acid and alkali iodides; the absence of chlorine and bromine being of course presupposed. j3. With nitrous acid and carbon disidphide. This excellent method has been in frequent use in my laboratory for a length of time ; it may be used for small or large quantities of iodine. We require : aa. Solution of potassium iodide of known strength. Made by drying the pure salt at 180 (see p. 149) and dissolving an exactly weighed quantity (about 5 grm.) to 1 litre. bb. Solution of sodium thiosulphate containing about 13 or 13'5 grm. of the pure crystallized salt in 1 litre. cc. Solution of nitrous acid in sulphuric acid. Prepared by passing nitrous acid gas into sulphuric acid to saturation. dd. Pure carbon disulphide. ee. Solution of sodium hydrogen carbonate. Made by dissolv- ing 5 grm. in 1000 c.c. cold water and adding 1 c.c. of hydrochloric acid to the solution. Begin by standardizing the thiosulphate as follows: Take a well-stoppered bottle of about 400 c.c. capacity, transfer to it 50 c.c. of the potassium iodide solution, add about 150 c.c. water, 20 c.c. carbon disulphide, some dilute sulphuric acid, and 10 drops of the solution of nitrous acid in sulphuric acid. Insert the stopper and shake the bottle violently for some time, allow to settle, and ascertain by adding a few more drops of the nitrous acid that the whole of the iodine has been liberated. Shake again, allow to- * This substance is not injured by the operation. 538 DETERMINATION. [ 145. settle, and pour the supernatant fluid as completely as possible into a flask, leaving the carbon disulphide in the bottle, add 200 c.c. water to the latter, shake well, pour off the water into the flask and repeat the washing till the last water has no acid reaction. To the contents of the flask add 10 c.c. carbon disulphide, shake well, pour off into a second flask, w r ash the disulphide a little, and finally shake the contents of the second flask again with some fresh disul- phide, which should now be barely tinged. Collect the disulphide from both flasks on a filter moistened with water, wash it till the washings are no longer acid, place the funnel in the bottle and pierce the point of the filter so that the disulphide from all the operations may be mixed. Add 30 c.c. of the sodium hydrogen carbonate and then the thiosulphate from a burette, with continual shaking, till the disulphide has lost its color. The number of c.c. of thiosulphate used will correspond to the iodine in 50 c.c. of potassium iodide solution. The analysis is performed exactly as above. The thiosulphate requires to be standardized before every fresh series of experi- ments, as it is liable to slight alteration. The presence of chlorides has no influence whatever on the results. In determining minute quantities of iodine let the solutions be ten times weaker, and use smaller quantities and smaller vessels. The results are entirely concordant and exact. y. With potassium permanganate, according to REINIGE.* This method, which is accurate and gives good results, depends upon the fact that all alkali iodides decompose potassium-perman- ganate according to the following equation : KI + 2KMnO 4 = KIO 3 + K 2 O + 2MnO a . This reaction was first recommended by PEA-N DE SAINT-GILLES t as a basis for the volumetric determination of iodine. Boiling facilitates the decomposition ; in very dilute solutions a little alkali carbonate is added in order to start the reaction. Metallic chlorides and bromides have no disturbing influence on the reaction, since they are unaffected by the permanganate. In this process there are required a solution of potassium permanganate (standardized according to 112, 2, a, or by a po- * Zeitschr. f. analyt. Chem., ix, 39. f Compt. rend., XLVI, 624. 145.] IODINE. 539 tassium -iodide solution of known strength, as described below) and a dilute solution of sodium thiosulphate, each solution con- taining about 5 grm. per litre. The thiosulphate solution being used to determine the excess of permanganate, it must therefore be standardized against the latter, according to the following reaction : 2KMn0 4 + 6Xa 2 S,0 3 = 2MnO, + 3^a a S 4 O 6 + K S O + 3]STa a O. To effect standardization, measure off 1 c. c. of the permanganate solution, add to it a large volume of water and a few drops of a sodium-carbonate solution, and then run in sodium-thiosulphate solution until the red color just disappears, a point readily observed in dilute solutions notwithstanding the presence of the precipitated hydrated manganese peroxide. After having added a little potassium- or sodium carbonate to the solution to be analyzed, and containing all the iodine as an alkali iodide, heat to gentle boiling and gradually add perman- ganate solution until the liquid containing the suspended precipi- tate of hydrated manganese peroxide acquires a decided red tint, which it retains even after repeated boiling. To better observe the color, remove the heat a few seconds after each 'ebullition in order to allow the precipitate to settle. Now pour the whole into a 500- c. c. flask, allow to cool, fill up to the mark, pipette off 100 c. c. of the liquid, and add to this thiosulphate solution until decoloriza- tion ensues. Multiply the number of c. c. required for this purpose by 5, calculate from this the equivalent of permanganate solution, and deduct this from the permanganate solution used ; the remainder corresponds to the metallic iodide decomposed, accord- ing to the equation given above. The direct titration of the excess of the permanganate in the liquid containing the suspended hydrated manganese peroxide, as recommended by HEINIGE, is less easily accomplished. It need scarcely be mentioned that organic and other reducing substances must be carefully excluded. tf. With silver solution and starch iodide (PISANI *). For this process there are required a titrated decinormal silver solution (p. 522) and standardized starch-iodide solution (p. 34-9). To the solution, containing the iodine as an alkali iodide, and * Compt. rend., XLIV, 352 ; Journ. f prakt. CTiem., LXXII, 266. 540 DETERMINATION. [ 145. which must be neutral or faintly acid, add first a little pure, precipitated calcium carbonate, then 0*5 to 1 c. c. of the starch- iodide solution, and then run in from a burette the silver solution, with constant stirring, until the starch iodide is just decolorized. The volume of silver solution used, after deducting the small quantity required to decolorize the 0*5 or 1 c. c. starch-iodide solution taken, corresponds to the iodine content. The method depends, as seen, upon the fact that silver solution decom- poses metallic iodides first, then starch iodide, and lastly any chloride that may be present. The process is rapid and gives good results in the absence of metallic chlorides and bromides. If but little chloride is present, the results are still approximate, but if considerable is present, the results are altogether unreliable, as the silver chloride precipitated is not decomposed with sufficient rapidity by the metallic iodide and starch iodide present. Metallic bromides interfere to a still greater extent than the chlorides. e. By distillation with ferric chloride (DUFLOS). When hydriodic acid or a metallic iodide is heated in a retort with solu- tion of pure ferric chloride, the whole of the iodine escapes with the aqueous vapor, and ferrous chloride is formed (Fe a Cl 8 -f- 2HI = 2FeCl 9 -f- 2HC1 -|- 21). The iodine passing over is received in so- lution of potassium iodide and determined by sodium thiosulphate, as directed in 146. In employing this method it must be borne in mind that the ferric chloride must be free from chlorine and nitric acid. It is best to prepare it from ferric oxide and hydro- chloric acid. We must not forget too that the separated iodine is liable to act on cork and caoutchouc ; the apparatus should therefore be constructed according to Fig. 81. C. KERSTING'S method,* depending upon precipitation with palladious-chloride solution until no further precipitate forms, gives good results, but is rather inconvenient, and is hence but little used. The same may be said of the method devised by A. and F. DUPKE f, which depends upon the action of chlorine water on an alkali iodide. This method gives good results only when metallic chlorides are absent \. * Annal. d. CJiem. u. Pharm., LXXXVII, 25. f 76. , xciv, 365. j H. ROSE'S Handbuch der analyt. Chem., 6. Aufl. von FINKENEK, n, 628 j also my own experiments. 145.] IODINE. 541 V. II. STRUVE'S* colorimetric method may be used in many cases. In this method the amount of iodine is estimated by the depth of color which the separated iodine gives to a measured quantity of carbon disulphide. II. Separation of Iodine from the Metals. The metallic iodides are in general analyzed like the corre- sponding chlorides. From iodides of the alkali metals containing free alkali the iodine may be precipitated as silver iodide, by first saturating the free alkali almost completely with nitric acid, then adding solution of silver nitrate in excess, and' finally nitric acid to strongly acid reaction. If an excess of acid were added at the beginning, free iodine might separate, which is not converted com- pletely into silver iodide by solution of silver nitrate. In com- pounds soluble in water the iodine may generally be precipitated as palladious iodide ; you may also determine the base in one por- tion (decomposing the compound with concentrated sulphuric acid) and the iodine in another portion according to 145, I., J, 6. Iodine cannot be separated from platinum directly with silver nitrate, as insoluble platinum salts would be thrown down with the silver iodide. For this purpose H. TOPSOE f recommends the fol- lowing process : Dissolve the substance in a good amount of water, add solution of sodium hydrogen sulphite and sulphurous acid, heat on a water-bath till the cplor has entirely disappeared, and the platinum is consequently converted into platinous sulphite. In this operation a white flocculent precipitate of sodium platinous sulphite which is difficultly soluble separates ; it redissolves on addition of sulphurous acid. After heating on the water-bath for some time, allow to cool completely, precipitate with silver solu- tion, which should not be added in large excess, add nitric acid, heat for about an hour to redissolve the silver sulphite first thrown down with the iodide, and then filter off the latter. Occasionally it is to be preferred to add sulphurous acid instead of the sulphite, and then, when the fluid has been heated and the color has gone, to add an excess of ammonia. In this way the platinum compound * Zeitschr. /. analyt. Chem., vm, 230. f lb., ix, 30. 542 DETERMINATION. [ 146. is not thrown down, and the silver sulphite does not separate after the addition of silver solution till nitric acid is added, and is imme- diately redissolved by the excess of the same. For the analysis of insoluble iodides, especially silver and lead iodides, mercurous and cuprous iodides, E. MEUSEL* strongly recommends sodium thiosulphate, in which these salts dissolve. Very little water should be used, and as small a quantity of the thiosulphate as possible. The metal is precipitated from the solu- tion by ammonium sulphide in the form of sulphide. Evaporate the filtrate with soda, and heat the residue in a platinum dish to incipient redness to destroy sodium thiosulphate and tetrathionate. Dissolve the melt in water by the aid of heat, and determine the iodine in it by 145, L, J, e. A large quantity of ferric chlo- ride will be required to decompose the sodium sulphite ; the resi- due in the retort should have a deep reddish-brown color. Silver iodide may be decomposed also by fusing with sodiunj carbonate, but not by igniting in a current of hydrogen, and not completely by zinc or iron. Mercurous iodide may be easily decomposed by distilling with 8 or 10 parts of a mixture of 1 part potassium cyanide and 2 parts quicklime. For the apparatus, see Fig. 88; ab is filled with magnesite (H. RosEf). Palladi- ous iodide may be decomposed by igniting in hydrogen. Cuprous iodide and many other iodides may be decomposed by boiling with potassium or sodium carbonate. Portions of metal, which may pass into the alkaline solution, may be thrown down by ammonium sulphide, or by acidifying with acetic acid, and passing hydrogen sulphide. Supplement. 146. DETERMINATION OF FREE IODINE. The determination of free iodine is an operation of great impor- tance in analytical chemistry, since, as BUNSEN^: first pointed out, it is a means for the estimation of all those substances which, when brought in contact with potassium iodide, separate from the same a definite quantity of iodine (e.g., chlorine, bromine, &c.), or, when * Zeitschr.f. analyt. Chem., ix, 208. . f 2b., n, 1. J Annal. d. Chem. u. Pharm. , LXXXVI, 265. 146.] IODINE. 543 boiled with hydrochloric acid, yield a definite quantity of chlorine (e.g., chromic acid, peroxide of manganese, &c.). By causing the chlorine produced to act on potassium iodide, we obtain the equiva- lent quantity of free iodine. Of the various methods which have been proposed for the esti- mation of free iodine, the oldest is that of SCHWARZ.* It is based upon the following reaction : 2Na 2 S 3 O 3 + 21 = 2NaI + Na a S 4 O 6 . 24-832 grm. pure crystallized sodium thiosulphate are dissolved to 1 litre. 1000 c.c. of the solution correspond to 12 '685, i.e., to 0*1 at. iodine. This solution is added to the solution of the substance in potassium iodide until the fluid appears bright yel- low, 3 or 4 c.c. thin and very clear starch-paste are then added, which must produce blue coloration, and finally again sodium thiosulphate, until the blue fluid is decolorized. This method, though in itself excellent, is open to this objec- tion, that it is difficult to obtain a solution of absolutely exact value by weighing off sodium thiosulphate, as the salt is not readily pro- curable in a perfectly pure and dry condition, and although the solution does not change rapidly or to any great extent, it is still liable to gradual alteration, especially under the influence of light. BUNSEN'S researches on the volumetric estimation of iodine cited above produced a very important and beneficial effect on the whole domain of chemical analysis. His process depends on the fact that when iodine comes in contact with an aqueous solution of sulphurous acid, a decomposition takes place in accordance with the equation H 2 SO 3 + H 2 O + 21 = H 2 SO 4 + 2HI, provided the solution does not contain more than 0*04 to 0*05 per cent, of an- hydrous sulphurous acid. If the solution is more concentrated, another reaction also takes place to a greater or less extent namely, H a SO 4 + 2HI = H 9 SO, + H,O + 21. In this method, a solution of iodine in potassium iodide con- taining a known quantity of free iodine is employed, and we com- mence by determining the relation between it and a sufficiently dilute solution of sulphurous acid. In applying the method, the iodine to be. estimated is dissolved in potassium iodide, the stand- ard sulphurous acid is added to decoloration, then thin starch-paste, and finally standard iodine solution till the blue color of iodide of starch is just visible. *Anleit. zu Maassamil. Nachtrage, 1853, 22. 544 DETERMINATION. [ 146. We calculate now the c.c. of iodine solution which correspond to the sulphurous acid employed, and deduct therefrom the c.c. of iodine added to destroy the excess of sulphurous &cid. The remainder gives the number of c.c. of iodine solution which contain a quantity of iodine equal to that in the substance ana- lyzed. On account of the rapidity with which solution of sulphurous acid changes, this method is somewhat inconvenient, and has given place to the following, which is now universally employed. It retains the basis of BUNSEN'S method, but substitutes sodium thio- sulphate for sulphurous acid, employing the reaction of SCHWARZ'S method. With F. MOHR* I give this " combined method " the preference, because, first, we are not bound to a definite strength of the thiosulphate ; secondly, the solution of thiosulphate is far less .affected by the oxygen of the air than sulphurous acid ; and thirdly, it loses nothing by evaporation. FINKENER")* even says, that the use of thiosulphate makes the method more accurate, his experi- ments having shown that in using BUNSEN'S method the results differ if, on one occasion, we add the sulphurous acid to the iodine, and, on another, the iodine to the sulphurous acid. a. REQUISITES FOR THE COMBINED METHOD. a. Iodine solution of known strength. Dissolve 6*2 to 6*3 :grm. iodine with the aid of about 9 grm. potassium iodide (free from iodic acid) to about 1200 c.c. /?. Solution of sodium thiosulphate. Dissolve 12*2 to 12*3 ,grm. of the pure and dry salt to about 1200 c.c. y. Solution of potassium iodide. Dissolve 1 part of the salt (free from iodic acid) in about 10 parts of water. The solution must be colorless and must remain so immediately after the addi- tion of dilute sulphuric or hydrochloric acid (either must be iron- free). 6. Starch solution. Stir the purest starch powder gradually with about 100 parts cold water and heat to boiling with constant stirring. Allow to cool quietly, and pour off the fluid from any deposit. The solution should be almost clear and free from all lumps. The starch solution is best prepared fresh before each series of experiments. * Lehrbuchti. chem.-unalyt. Titrirmethode, 3. Aufl., 256. f H. ROSE, Handb. d. analyt. Che?>, 6. Aufl. von FINKENER, n, 937. 146.] IODINE. 545 5. PRELIMINARY DETERMINATIONS. a. Determination of the relation between the Iodine Solution and ThiosuipJiate Solution. Fill two burettes with the solutions. Hun 20 c. c. of the thiosulphate into a beaker, add some water and 3 or 4 c. c. starch solution, then add the iodine till a blue coloration is just pro- duced. If you have added a drop too much, run in one or two drops more of the thiosulphate and then more cautiously the iodine solution. After a few minutes read off the height of the fluid in both burettes. Let us suppose we had used 20 c. c. thio- sulphate to 20-2 c. c. iodine. ft. Exact Determination of the Iodine in the Solution This is done immediately before each series of analyses with the aid of an exactly weighed quantity of pure and dry iodine. Experience has convinced me that solution of iodine in potassium iodide, even when kept cool and in the dark, is much more liable to change than is usually supposed.* The determination is best made as follows : The tubes shown in Fig. 104 are heated, al- lowed to cool in an exsiccator, and then weighed. 0*2 grm. of pure, resublimed iodine f is then introduced into the inner tube, the tube placed obliquely in a small sand-bath, heated until the iodine melts, and cooled in a very oblique position, so that it may be held in the hand. Now slip on the outer tube, let the whole cool in an exsiccator, weigh, and thus ascertain the exact quantity of iodine in the tube. Now place the inner tube (the outer tube also, of course, if any iodine adheres to it) in a stoppered bottle containing about 10 c. c. potassium-iodide solution. As soon as all the iodine is dissolved dilute with water, run in sodiurn- thiosulphate solution from a burette until decolorization is just effected, add 3 or 4 c. c. starch paste, and then iodine solution *I filled several small well-stoppered bottles with some solution of iodine in potassium iodide, the standard of which had been accurately determined, and placed them in. a cellar. Even in the course of a few weeks the standard had altered. I now never rely on the strength of a solution of iodine unless I have determined it shortly before. f Regarding the preparation of absolutely pure iodine, compare STAS, ZeitscJif f. analyt. Chem., vi, 419. 546 DETERMINATION. [ 146. (, <*) until a blue tinge appears. Read off botli burettes and calculate the iodine content of the solution 0, a as follows : Suppose we had weighed off 0*15 grm. iodine and used 29 '5 c. c. thiosulphate and 0'3 c. c. iodine solution. From 5, a, we know that 20 c. c. thiosulphate correspond to 20'2 c. c. iodine solution; 29'5 c. c. therefore correspond to 29-8 c. c Now 29-5 c. c. thiosulphate correspond to 0*15 grm. iodine -\- 0*3 c. c. iodine solution. But 29 '5 c. c. thiosulphate also correspond to 29*8 c. c. iodine solution. .*. 0*15 grm. iodine -f- 0-3 c. c. iodine solution = 29*8 c. c. iodine solution. .'. 0'15 grm. iodine = 29 -5 c. c. iodine solution. .. 1 c. c. iodine solution = 0-0050847 grm. iodine. The experiment just described is repeated and the mean of the two results taken, provided they exhibit sufficient uniformity. Where tubes are not at hand the process may be conducted in the following manner : Select three watch-glasses, &, 5, and , which fit each other ; weigh b and c together accurately. Put about 0'5 grm. pure dry iodine into #, place it on an iron plate, and heat gently till dense fumes of iodine escape. Now cover it with 5 and regulate the heat so that the iodine may sublime en- tirely, or almost entirely, into Z>. Next remove J while still hot and give it a gentle swing in the air to remove the still uricon- densed iodine fumes and any traces of aqueous vapor, cover it with and 1 c. c. corresponds to 0*005 grin, iodine, the calcula- tion is in the highest degree simple ; for suppose we had used 21 c. c. Na a S a O, and 1 c. c. iodine, the quantity of iodine present is 0-1 grm. 21 1 = 20, and 20 X 0*005 = 0*1. 548 DETERMINATION. [ 147. Where you are analyzing chromic acid or manganese dioxide by boiling with hydrochloric acid, and passing the chlorine evolved into potassium iodide, you must allow the solution to cool before titrating with thiosulphate ; for at a high temperature a portion of the sodium tetrathionate produced is converted into sodium sul- phate by the iodine (WRIGHT*). Free acid in the iodine solution to be estimated is not injuri- ous ; when such is present, however, the excess of the thiosulphate must be titrated without delay, or the free thiosulphuric acid may be decomposed before the iodine is added. d. KEEPING OF THE SOLUTIONS. The iodine solution and the thiosulphate solution are kept in glass-stoppered bottles in a cool, dark place. ' But the relation between the two solutions must be tested before each new series of experiments, and the iodine in the iodine solution must be rede- termined. If a fluid contains free iodine in presence of iodine in combina- tion, determine the former in one portion by the combined method, and the total quantity in another portion. For this purpose you may either (1) add sulphurous acid to decoloration, precipitate with silver nitrate ( 145, I., a, a\ digest the precipitate with nitric acid to remove any silver sulphite which it may contain, filter, &c. ; or (2) distil with ferric chloride as directed, 145, I. ,5, e. 1*7. 4. CYANOGEN, f I. Determination. a. Gravimetric .Estimation. If you have free hydrocyanic acid in solution run it into an excess of solution of silver nitrate, add a little nitric acid, allow to settle without warming, and determine the precipitated silver cyanide either by collecting on a weighed filter, drying at 100 and weighing ( 115, 3), or by collecting on an unweighed filter and converting into metallic silver. The latter operation is performed by igniting the precipitate in a porcelain * ZeitscliT. f. analyt. Chem., ix, 482. f With regard to HERAPATH'S colorimetric method which is founded on the intensity of the color of a solution of persulphocyanide of iron, compare (faun. Gaz., Aug. 1853, 294. SI 47.] CYANOGEN. 649 crucible for J hour, or till it ceases to lose weight (H. ROSE). If you wish to determine in this way the hydrocyanic acid in bitter almond water or cherry laurel water, add ammonia after the addi- tion of the solution of silver nitrate till the fluid is strongly alka- line (it is not necessary to dissolve all the silver cyanide), and at once acidify with nitric acid. When the precipitate has settled, filter. The whole of the cyanogen in the fluid will have been now converted into silver cyanide. (The cyanogen was originally pres- ent partly as hydrocyanic acid, partly as ammonium cyanide, but principally as hydrocyanate of benzaldehyd S. FELDHAUS.*) FELDHAUS recommends the following proportions : 100 grin, bitter almond water, about 1'2 grm. silver nitrate, dissolved in water and 2 to 3 c.c. ammonia sp. gr. 0*96. A portion of the filtrate should be tested to make sure that it contains silver salt in excess, another portion should be tested by making it strongly alkaline with ammonia, and then acid again with nitric acid. If a precipi- tate is formed in the latter case it shows that the whole of the hydrocyanate of benzaldehyd was not decomposed, and the precipi- tation must be repeated. If you want to measure off a fluid con- taining hydrocyanic acid with a pipette, insert a little tube with soda-lime between the pipette and the flexible tube which you put into your mouth. b. LIEBIG'S Volumetric Method-)-. If hydrocyanic acid is mixed with potassa to strong alkaline reaction, and a dilute solution of silver nitrate is then added, a permanent turbidity of silver cyanide or, if a few drops of solution of sodium chloride have been added, of silver chloride forms only after the whole of the cyanogen is converted into double cyanide of silver and potassium. The first drop of solution of silver nitrate added in excess produces the per- manent precipitate. 1 at. silver consumed in the process corre- sponds, therefore, exactly to 2 mol. hydrocyanic acid (2KCy + Ag NO 3 = AgCy.KCy + KNO 8 ). A decinormal solution of silver nitrate, containing consequently 10-792 grm. silver in the litre, should be used; 1 c.c. of this solution corresponds to 0-0054096 of hydrocyanic acid. In examining medicinal hydrocyanic acid 5 to 10 grm. ought to be used, but of bitter almond water about 50grm. ; if exactly 5-4096 or 54-096 grm. are used, the number of c.c. of the silver solution, divided by 10, or by 100, expresses exactly * Zeitschr.f. analyt. Chem., in, 34. \ Annal. d. Chem. u. Pharm., LXXVII, 102. 50 DETERMINATION. [ 147. the percentage of hydrocyanic acid. Medicinal hydrocyanic acid is suitably diluted first by adding from 5 to 8 volumes of water ; bitter almond water also is slightly diluted ; if the latter is turbid the end-reaction will not be sufficiently distinct, and the gravimetric method is to be preferred. LIEBIG has examined hydrocyanic acid of various degrees of dilu- tion, and has obtained results by this method corresponding exactly with those obtained by a. SOUCHAY,* too, obtained results almost identical ; with pure dilute hydrocyanic acid, the gravimetric results were to the volumetric as 100 to 100*5 101 ; with clear or nearly clear bitter almond water as 100 to 102. FELDHAUS (loo. cit.) obtained very nearly similar results. The slightly higher results of the volumetric process are to be explained from the fact that a small excess of silver solution is necessary to produce the final reaction. The less the amount of the substance taken the greater importance does this error assume. We should also notice that in the bitter almond water, which contains ammonium cyanide, some ammonia is set free which has a solvent action on the silver cyanide. In this method it does not matter whether the hydrocyanic acid contains an admixture of hydrochloric acid or formic acid. A considerable excess of potassa must be avoided. If it is intended to determine potassium cyanide by this method, a solution of that salt must be prepared of known strength, and a measured quantity used containing about O'l grm. of the salt. Should it contain potassium sulphide, a small quantity of freshly precipitated lead carbonate must be first added and the solution filtered before proceeding to the determination. c. Fordos and Gelis^s Volumetric Method.^ This method is founded on the reaction of free iodine on potassium cyanide, described by SERTJLLAS and WOHLER,, and which is as follows: KCN + 21 = KI + KIN". According to this, 2 eq. of iodine correspond to 1 eq. of cyanogen or 1 eq. hydrocyanic acid, or 1 eq. of potassium cyanide. The iodine solution is best pre- pared according to 146. If free hydrocyanic acid is to be deter- mined, add first some soda solution cautiously until an alkaline reaction, then add carbonic-acid water to convert any possible excess of alkali into carbonate (the fluid must not render curcuma paper brown), and finally sufficient iodine solution to just perrna- * Zeitschr.f. analyt. Chem., u, ISO. f Journ. de Gliim. et de Pharm., xxm, 48; Journ. /. prakt. Chem., LIX, 255. 147.] CYANOGEN. 551 nently tinge the colorless solution yellowish. "When analyzing potassium cyanide prepare a solution first of known strength and use a volume containing about 0'05 grm. of potassium cyanide. In this case, too, the addition of carbonic-acid water is necessary. The potassium cyanide must contain no potassium sulphide, as this vitiates the results. The method on the whole gives good results. Compare SOTJCHAY (loo. cit.)\ it is not applicable for bitter- almond water, however. II. Separation of Cyanogen from the Metals. a. In Cyanides of the Alkali Metals. Mix the substance (if solid, without previous solution in water) with excess of silver nitrate solution, then add water, finally nitric acid in slight excess, allow to settle without warming, and deter- mine the silver cyanide as in L, a. The basic metals are deter- mined in the filtrate after separating the excess of silver. b. In Cyanides and double Cyanides, which are completely decomposed by Silver Nitrate and Nitric Acid or Silver Nitrate and Ammonia. Digest for some time with a dilute solution of silver nitrate, stirring frequently,* then add nitric acid in moderate excess, and digest at a gentle heat, till the foreign cyanide is fully dissolved and the silver cyanide has become pure and quite white. Then add water and filter. As a precautionary measure it is well to test the metal obtained by long ignition of the silver cyanide, whether it is free from those metals which were combined with the cyano- gen. The filtrate is used for estimating the basic metals, the silver being first precipitated with hydrochloric acid. This method affords us an exact analysis of the double cyanides of potassium w^ith nickel, copper, and zinc (II. HOSE). W. WEiTiif recommends a solution of silver nitrate in ammo- nia for the decomposition of many cyanogen compounds, such as potassium ferrocyanide, Prussian blue, and even potassium cobalti- cyanide. He digests them in sealed tubes at 100 (in the case of potassium cobalticyanide, 150) for 4 or 5 hours. Warm the con- tents of the tube gently in a dish ; until the crystals of ammonio- cyanide of silver are dissolved, filter off the separated metallic * Double cyanide of nickel and potassium yields by this process a mixture of silver cyanide with nickel cyanide. Like double cyanides are similarly decom- posed. -\Zeitschr.f. analyt. Chem., ix, 379. 552 DETERMINATION. [ 147. oxide, wash it with ammonia, dilute, and precipitate the silver cyanide by acidifying with nitric acid. In the filtrate separate the silver from the alkalies, &c. In respect to the undissolved oxides it should be noted that metallic silver is always mixed with the ferric oxide. c. In Mercuric Cyanide. Precipitate the aqueous solution with hydrogen sulphide ; the mercuric sulphide may be filtered without difficulty if a little ammonia or hydrochloric acid be added ; it is determined accord- ing to 118, 3. If the compound is in the solid condition, the cyanogen may be determined in another portion by ignition with cupric oxide, the nitrogen and carbonic acid being collected and separated (comp. Organic Analysis). H. ROSE and FINKENEK* have, after much trouble, succeeded in finding out a method for determining cyanogen with precision also in solutions of mercuric cyanide. Mix the solution of the mer- curic cyanide with zinc nitrate dissolved in ammonia. To 1 part of mercuric salt you may add about 2 parts of the zinc salt. Add to the clear solution hydrogen sulphide water gradually till it pro- duces a perfectly white precipitate of zinc sulphide. The precipi- tate, w T hich is a mixture of the mercuric and zinc sulphides, settles well. After a quarter of an hour filter it off and wash with very dilute ammonia, The filtrate contains zinc cyanide dissolved in ammonia, together with ammonium nitrate. It does not smell of hydrocyanic acid, and consequently no escape of the latter takes place. Mix it with silver nitrate and then add dilute sulphuric acid in excess. The silver cyanide is next washed a little by decantation, then to free it from any zinc cyanide simultaneously precipitated heated with a solution of silver nitrate, finally filtered off, washed, arid determined after 147, I., a. The precipitated sul- phides may be dissolved in aqua regia, and the mercury precipitated as mercurous chloride according to 118, 2. The test-analyses com- municated by ROSE yielded excellent results. d. In compounds decomposable by Mercuric Oxide in the Wet Way. Many simple cyanides, and also double cyanides both of the character of the double cyanide of nickel and potassium, and of the ferro- or ferricyanides (not, however, cobalticyanides) may, as * Zeitschr. /. analyt. Chem., i, 288- 147] CYANOGEN. 553 is well known, be completely decomposed by boiling with excess of mercuric oxide and water, all cyanogen being obtained as mer- curic cyanide and the metals passing into oxides. H. ROSE (Loc. cit.} has shown that Prussian blue, potassium ferro- and ferricyanide, more particularly, may be readily analyzed in this manner. Boil a few minutes with water and excess of mercuric oxide till complete decomposition is effected, add in order to render the ferric hydroxide and mercuric oxide removable by filtration nitric acid in small portions, till the alkaline reaction has nearly disap- peared, filter, wash with hot water, dry the precipitate, ignite very gradually raising the heat under a hood (with a good draught), and weigh the ferric oxide remaining. In the filtrate the cyanogen is determined according to 0, and any potassium that may be present is determined in the filtrate from the silver cya- nide. e. Determination of Metals contained in Cyanides with decom- position and volatilization of the Cyanogen. Of the various means for completely decomposing compounds of cyanogen, especially also the double cyanides, according to H. HOSE (Loc. cit.} three particularly are worthy of recommendation viz., concentrated sulphuric acid, mercuric sulphate, and ammo- nium chloride. The nitrates seemed decidedly less suitable on account of their too violent action. a. DECOMPOSITION BY SULPHURIC ACID. All cyanogen com- pounds, simple or double, are completely decomposed and con- verted into sulphates or oxides, as the case may be, if treated in a powdered condition in a platinum dish or a capacious platinum crucible with a mixture of about 3 parts concentrated sulphu- ric acid and 1 part water, and heated till almost all the sulphuric acid had been expelled. The residual mass is then free from cyan- ogen. It is dissolved in water, if necessary with addition of hydrochloric acid, and the metals determined by the usual methods. This way is not adapted for mercuric cyanide, as a little of the metal would escape with the fumes of the sulphuric acid. /?. DECOMPOSITION BY MERCURIC SULPHATE. Of the mercuric sulphates, those suitable to our present purpose are the normal and the basic (Turpeth mineral). The substance is mixed with (\ parts of the latter, heated in a platinum crucible gradually, and finally maintained for a long time at a red-heat, till all the mercury has 554 DETEKMINATION. [ 147. volatilized, and the weight of the crucible remains constant. If alkalies are present, a little ammonium carbonate is added from time to time during the final ignition, in order to convert the acid sulphates into normal. The residue may usually be analyzed by simple treatment with water; in the case of potassium ferro- cyanide, for instance, the potassium sulphate dissolves and pure (alkali-free) ferric oxide remains behind. The test-analyses that have been communicated show excellent results. y. DECOMPOSITION BY AMMONIUM CHLORIDE. Mix the sub- stance with twice or thrice the amount of this salt and ignite the mixture moderately in a stream of hydrogen (apparatus, Fig. 83). From the cooled mass water extracts alkali chloride, while the reducible metals remain in the metallic state. The method is peculiarly adapted for the analysis of double cyanide of nickel and potassium "and cobalticyanide of potassium, not so for iron com- pounds, since the iron obtained is not pure, but contains carbon. If one of the methods described in e is employed, the nitrogen and carbon (the cyanogen) must be determined by a combustion, if an estimation by the loss is not sufficient. f. Determination of the Alkalies, especially of Ammonia, in Soluble Ferrocyanides. . Mix the boiling solution with a solution of cupric chloride in moderate excess, filter off the precipitated cupric ferrocyanide, free the filtrate from copper by means of hydrogen sulphide, and then determine the alkalies (REINDEL *). In the case of fixed .alkalies the object may also be obtained by igniting with barium thiosulphate (FROHDE f). g. Volumetric Determination of Ferro- and Ferricyanogen. OL. After E. DE HAEN. This method, devised in my labora- tory, is founded upon the simple fact that a solution of potassium ferrocyanide acidified with sulphuric acid (and which may accord- ingly be assumed to contain free hydroferrocyanic acid) is by ad- dition of potassium permanganate converted into the correspond- ing ferricyanide. If this conversion is effected in a very dilute fluid, containing about 0-2 grin, potassium ferrocyanide in frorq *Journ.f. prakt. Chem., LXV, 452. t Zcitschr. /. analyt. Chem., in, 181. 147.] CYANOGEN. 555 100 to 200 c. c., the termination of the reaction is clearly and unmistakably indicated by the change of the originally pure yel- low color of the fluid to reddish-yellow.* The process requires two test-fluids of known strength, viz. : 1. A solution of pure potassium ferrocyanide. 2. A solution of potassium permanganate. The former is prepared by dissolving 20 grin, perfectly pure and dry crystallized potassium ferrocyanide in water to 1 litre; each c.c. therefore contains 20 mgrm. The latter is diluted so that somewhat less than a buretteful is required for 10 c.c. of the solu- tion of potassium ferrocyanide. To determine the strength of the potassium permanganate solu- tion in its action upon the potassium ferrocyanide, measure off, by means of a pipette, 10 c.c. of the solution of potassium ferrocyanide (containing 0'2 grm.), dilute with 100 to 200 c.c. water, acidify with sulphuric acid, place the glass on a sheet of white paper, and allow the permanganate to drop into the fluid, stirring it at the same time, until the change from yellow to reddish-yellow indicates that the conversion is complete.f Repetitions of the experiment always give very accurately corresponding results. If at any time you have reason to suspect that the permanganate has suffered altera- tion, recourse must be had again to this experiment. If after acidifying the potassium ferrocyanide with sulphuric acid you add a trace of ferric chloride to produce a bluish-green color, the latter will disappear at the end of the reaction, which is thus rendered very distinct (GINTL^;). To determine the amount of real potassium ferrocyanide con- tained in any given sample of the commercial article, dissolve 5 grm. to 250 c.c. ; take 10 c.c. of this solution, and examine as just directed. Suppose, in determining the strength of the permanga- nate, you have used 20 c.c., and you find now that 19 c.c. is suffi- cient, the simple rule-of-three sum, 20: -2:: 19: a; - Instead of the permanganate you may use potassium cbromate. The solu- tion is added till spots of iron sesquichloride on a plate are no longer colored blue or green, but brownish. E. MEYER, Zeitschr. f. analyt. Chem., vin, 508. f If you wish at first for some additional evidence besides the change of color, add to a drop of the mixture on a plate, a drop of solution of ferric chloride; if this fails to produce a blue tint, the conversion is accomplished. % Zeitschr. f. analyt. Chem., vi, 446. 556 DETERMINATION. [ 147. will inform you how much pure potassium ferrocyanide 0*2 grin, of the analyzed salt contains. And even this small calculation may be dispensed with by diluting the permanganate so that exactly 50 c. c. correspond to 0'2 of potassium ferrocyanide, as, in that case, the number of half-c.c. consumed expresses directly the percentage of pure ferrocyanide. Instead of determining the strength of the permanganate by means of pure potassium ferrocyanide, which is unquestionably the best way, one of the methods given in 112, 2, may also be employed ; bearing in mind, in that case, that 2 mol. potassium ferrocyanide = 845*256, 2 at. iron = 111-8, and 1 mol. oxalic acid == 126-048 are equivalent in their action upon solution of potassium permanganate. The analysis of soluble ferricyanides by this method is effected by reducing them to ferrocyanides, acidifying, and then proceeding in the way described. The reduction is effected as follows : Mix the weighed ferricyanide with a solution of soda or potassa in excess, boil and add concentrated solution of ferrous sulphate gradually, and in small portions, until the color of the precipitate appears black, which is a sign that protosesquioxide of iron has precipitated. Dilute now to 300 c.c., mix, filter, and proceed to determine the ferrocyanide in portions of 50 or 100 c.c. of the fluid. As the space occupied by the precipitate is not taken into account in this process, the results are not absolutely accurate ; the difference is so very trifling, however, that it may safely be disre- garded. GINTL (loc. eit.) suggests to put the neutral or alkaline fluid in a tall vessel and add a few lumps of sodium amalgam as big as peas : in ten minutes the reduction will be effected and with- out the aid of heat. Insoluble ferro- or ferricyanides, decomposable by boiling solu- tion of potassa (as are most of these compounds), are analyzed by boiling a weighed sample sufficiently long with an excess of solu- tion of potassa (adding, in the case of ferricyanides, ferrous sul- phate), and then proceeding as directed above. ft. After E. LENSSE^. Ferricyanides may also be analyzed according to the following method, which too was devised in my laboratory: The method is based on the fact that on bringing together potassium ferricyanide, potassium iodide, and concen- 147.] CYANOGEN. 557 trated hydrochloric acid, for every eq. of potassium ferricyanide 1 eq. of iodine are precipitated, thus : K,Fe(CN). + KI = K.Fe(UN).+ I. On estimating the liberated iodine according to 146, the quantity of potassium ferricyanide is then determined. In four experiments LENSSEN obtained 99 -22, 101*7, 102-1, and 100-5, instead of 100. The solution may be diluted only after the hy- drochloric acid has been added. C. MOHR * obtained still more accurate results, as he avoided the formation of hydroferricyanic acid by adding zinc-sulphate solution, which is not at all decomposed by iodine. He directs adding potassium iodide and hydrochloric acid in excess to the diluted ferricyanide solution, then to add an excess of iron-free zinc-sulphate solution, neutralize the free acid with sodium bicarbonate in slight excess, and to then determine the liberated iodine according to 146. y. To determine potassium ferrocyanide in dyers' baths, which contain oxidizable organic substances, and which, hence, cannot be estimated with permanganate, II. RHEINECK f recom- mends a process based on the fact that potassium -ferrocyanide solution, on gradually adding a solution of a ferric salt, whether a mineral acid is added or not, yields a clear, blue solution which becomes turbid and which, when all the ferrocyanogen is thrown down, forms a clear, colorless liquid containing Berlin blue sus- pended in flocculent form. Hence on adding a solution of a fer- ric salt to equal volumes of a potassium-ferrocyanide solution of known strength and of the bath, until in both cases the flocculent precipitate forms, the unknown quantity of ferrocyanide may be readily calculated. d. After E. BOHLIG.^ In the case of a fluid containing potassium ferrocyanide, and also sulphocyanide (for instance, the red liquor of the prussiate works), the method given in a cannot be employed, as the hydro- sulphocyanic acid also reduces permanganic acid. The following method depending on the precipitation of the ferrocyanogen with solution of cupric sulphate may then be used; it is accurate enough for technical purposes: Dissolve 10 grm. pure cupric sul- *Annal. d. Chem. u. Pharm., cv, 62. f Chem. Centralbl., 1871, p. 778. % Polytechn. Notizblatt, xvi, 81. 558 DETERMINATION. [ 148. phate to 1 litre, also 4 grm. pure dry potassium ferrocyanide to 1 litre. Add to 50 c.c. of the latter solution (which contain 0'2 grm. potassium ferrocyanide) copper solution from a burette to complete precipitation of the ferrocyanogen. In order to hit this point exactly, from time to time dip a strip of filter-paper into the brownish-red fluid which will imbibe the clear filtrate, leaving the precipitate of copper ferrocyanide behind. At first the moist strips of paper, when touched with ferric chloride, become dark blue, the reaction gradually gets w r eaker and weaker, and finally vanishes altogether. We now know the value of the copper solution with reference to its action on potassium ferrocyanide, and can, there- fore, by its means test solutions containing unknown amounts of ferrocyanogen. If alkali sulphides are present, they are first removed by boiling with lead carbonate. After filtering off the lead sulphide, acidify with dilute sulphuric acid, and then proceed. 148. 5. SULPHUR. I. Determination. To determine hydrogen sulphide in a mixture of gases confined over mercury* it may be absorbed by a ball made of 2 parts precipi- tated lead phosphate and 3 parts plaster of Paris. The mixture is made into a paste with water, and pressed into a bullet mould in which the platinum wire is inserted. The mould should previously be oiled. The balls are dried at 100, saturated with concentrated phosphoric acid, and are then ready for use (LuDwiaf). To determine sulphuretted hydrogen dissolved in water the following methods are in use : a. The method of determining hydrogen sulphide volumetri- cally by solution of iodine, was employed first by DUPASQUIER ; it is very convenient and accurate. That chemist used alcoholic solu- tion of iodine. But as the action of the iodine upon the alcohol alters the composition of this solution somewhat rapidly, it is bet- ter to use a solution of iodine in potassium iodide. The decom- position is as follows : H,S + 21 = 2HI + S * When this gas remains long in contact with mercury, sulphide of mercury Is liable to be formed. f Annal, d. Chem. u. Pharm., CLXII, 55. 148.] 2 at. I = 253-70 correspond, lience, to 1 iriol. II 3 S = 34*086. However, this exact decomposition can be relied upon with cer- tainty only if the amount of hydrogen sulphide in the fluid does not exceed 0-04 per cent.(BuNSEN). Fluids containing a larger pro- portion of hydrogen sulphide must therefore first be diluted to the required degree with boiled water cooled out of the contact of air. The iodine solution of 146 may be used for the estimation of larger quantities of hydrogen sulphide ; for weak solutions, e.g., sulphuretted mineral water, it is advisable to dilute the iodine solu- tion 5 times, so that 1 c.c. may contain O'OOl grm. iodine. The process is conducted as follows : Measure or weigh a certain quantity of the sulphuretted wateiy dilute, if required, in the manner directed, add some thin starch- paste, and then solution of iodine, with constant shaking or stir- ring, until the permanent blue color begins to appear. The result of this experiment indicates approximately, but not with positive accuracy, the relation between the examined water and the iodine solution. Suppose you have consumed, to 220 c.c. of the sulphu- retted water, 12 c.c. of a solution of iodine containing 0*000918- grm. iodine in the c.c.* Introduce now into a flask nearly the quantity of iodine solution required, add the sulphuretted water in quantity either already determined, or to be determined, by weight or measure ;f then to the colorless fluid add thin starch- paste, and after this iodine solution until the blue color just begins to show. By this course of proceeding, you avoid the loss of hydrogen sulphide which would otherwise be caused by evaporation and oxidation. In my analysis of the Weilbach water, 256 c.c. of the water required, in my second experiment, 16'26 c.c. of iodine solution, which, calculated to the quantity of sulphuretted water used in the first experiment, viz., 220 c.c., makes 13'9 c.c., or 1'9 e.c. more. But even now the experiment cannot yet be considered quite conclusive, when made with a solution of iodine so dilute ; it being still necessary to ascertain how much iodine solution is required to impart the same blue tint to the same quantity of ordinary water mixed with starch-paste, of the same temperature,^: and as nearly as possible in the same condition! as the analyzed sulphuretted *Tlie numbers here stated are those which I obtained in the analysis of the Weilbach water. f Compare Experiment No. 82. \ Atinal d. Chem. u. Pharm., en, 186. In this connection I would recommend, in cases where the sulphuretted 5(30 DETERMINATION. [ 148. water, and to deduct tins from the quantity of iodine solution used in the second experiment. Thus, in the case mentioned, I had to deduct 0*5 c, c. from the 16*26 c. c. used. If the instruc- tions here given are strictly followed, this method gives very accurate results. (See Expt. No. 82.) 5. FK. MOHR'S method slightly modified. Add to the sulphuretted water a slight excess of sodium-arsen- ite solution standardized against iodine solution ( 127, 5, a\ then add hydrochloric acid until the liquid is distinctly acid. Dilute to 300 c. c., pass through a dry filter-paper, make sure that the solution still contains sodium arsenite by testing a sample with hydrogen -sulphide water, and then determine in 100 c. c. , after adding powdered sodium bicarbonate, the remainder of the arsenous acid. Deduct the c. c. of iodine solution last used, mul- tiplied by 3 (because only 100 c. c. of the 300 c. c. have been operated upon), from that corresponding to the entire quantity of arsenous acid used ; the remainder will express the quantity of iodine solution equivalent to the hydrogen sulphide present. In calculating it must be remembered that here 2 eq. of iodine cor- respond to 3 eq. of hydrogen sulphide, since 1 eq. of As a O, decomposes 3 eq. of H a S on the one hand, forming As a S, and 3H a O, and requires on the other hand 2 eq. of iodine for its con- version into arsenic acid. Yery dilute hydrogen-sulphide solutions cannot be estimated by this method, as the arsenic sulphide formed in them takes a long time to deposit, and a very small portion always remains in solution.* c. Mix the sulphuretted fluid with an excess of solution of sodium arsenite, add hydrochloric acid, allow to deposit, and deter- mine the arsenous sulphide as directed in 127, 4. The results are accurate, unless the solution is very dilute, in which case the slight solubility of arsenous sulphide occasions loss. (See Expt. No. 82.) In an analysis of the Weilbach waters, this method hence gave 0-006621 and 0-006604 per 1000, whereas water taken water contains bicarbonate of soda, to add to the ordinary water an equal quan- tity of this salt, as its presence has a slight influence on the appearance of the final reaction. * Hydrogen-sulphide water containing 0'003 grin. H a S in a litre gave with a solution of arsenous acid in hydrochloric acid a precipitate that could be filtered off only after 12 hours. 148.] SULPHUK. 561 from the well at the same time, and titrated with iodine, gave O -007025 II a S per 1000. Instead of arsenous acid the precipi- tation may be effected by means of cupric acetate together with a little acetic acid, or by 'means of a silver solution, and the sul- phur determined in the precipitated copper sulphide as barium sulphate, according to 148, II, or the silver may be determined in the silver sulphide in the metallic state. When copper salts are used in very dilute solutions, the results are also too low ; whether this is also the case with the silver solution I cannot say, from lack of personal experience. The silver solution best adapted for the purposes is recommended by LYTE * to be prepared by dissolv- ing silver chloride in sodium-thiosulphate solution and adding a few drops of ammonia. In an analysis of water containing iron sul- phate, LYTE f threw down the hydrogen sulphide with freshly pre- cipitated lead sulphate, filtered, washed, and extracted the lead sul- phate with hot ammonium-acetate solution, converted the residual lead sulphide into lead sulphate by oxidation with nitric acid, etc., and then weighed the sulphate. In the case of mineral waters the method a is always to be preferred, unless thiosulphates should be present and impair its accuracy. d. If the hydrogen sulphide is evolved in the gaseous state, and large quantities are to be determined, the best way is to conduct it first through several bulbed U-tubes (Fig. 103), containing an alkaline solution of sodium arsenite, then through a tube connected with the exit of the last U-tube, which contains pieces of glass moistened with solution of soda, to mix the fluids afterwards, and proceed as in b. If, on the other hand, we have to determine small quantities of hydrogen sulphide contained in a large amount of air, etc., it is well to pass the gaseous mixture in separate small bubbles through a very dilute solution of iodine in potassium iodide of known volume and strength, which is contained in a long glass tube fixed in an inclined position and protected from sun- light. The free iodine remaining is finally estimated by means of a solution of sodium thiosulphate ( 146); the difference gives us the quantity of iodine which has been converted by hydrogen sulphide into hydriodic acid, and consequently corresponds to the * Compt. rend., XLIII, 765. f Zeilschr. f. analyt. Chem., v, 441. 562 DETERMINATION". [ 148. amount of the hydrogen sulphide present. The volume of the gaseous mixture may be known by measuring the water which has escaped from the aspirator used. The arrangement of the absorption tube is the same as is figured in connection with the Determination of Carbonic Acid in Air. The thin glass tube conducting the gas into the absorption tube, however, must not be provided with an india-rubber elongation. From my own experiments* it appears that hydrogen sul- phide, whether in small or large quantities, may be also estimated by the increase in weight of absorption tubes. We have only to take care that the mixture of gases is first thoroughly dried by passing over calcium chloride. To take up the hydrogen sulphide we use U-tubes, five-sixths filled with copper sulphate on pumice, one-sixth at the exit containing calcium chloride. To prepare the pumice with copper sulphate, proceed as follows : Treat 60 grin, pumice in lumps the size of peas in a small porcelain dish with a hot concentrated solution of 30 or 35 grm. copper sulphate, dry the whole with constant stirring, place the dish in an air or oil bath of the temperature of 150 to 160, and allow to remain therein four hours. A tube containing 14 grm. of this prepared pumice will absorb about 0*2 grm. hydrogen sulphide. It is well always to employ two such tubes. If the prepared pumice is dried at a lower temperature it takes up much less of the gas, if dried at a higher temperature the gas is decomposed and sulphur- ous acid is formed. This method is more completely detailed under the Analysis of Black Ash. Finally, small quantities of hydrogen sulphide mixed with other gases may be estimated by passing through bromine water and con- verting into sulphuric acid. II. Separation and Determination of Sulphur in Sulphides. A. METHODS BASED ON THE CONVERSION OF THE SULPHUR INTO SULPHURIC ACID. 1. 'Methods in the Dry Way. a. Oxidation ~by Alkali Nitrates (applicable to all compounds of sulphur). If the sulphides do not lose any sulphur on heating, mix the pulverized and weighed substance with 6 parts of anhy- * Zeitschr.f. analyt. Chem., x, 75 ^ 148.] SULPHUR. 563 drous sodium carbonate and 4 of potassium nitrate, with the aid of a rounded glass rod, wipe the particles of the mixture which adhere to the rod carefully off against some sodium carbonate, and add this to the mixture. Heat in a platinum or porcelain crucible (which, however, is somewhat affected by the process), at a grad- ually increased temperature to fusion ;* keep the mass in that state for some time, then allow it to cool, heat the residue with water, filter the fluid, boil the residue with a solution of pure sodium car- bonate, filter, wash, remove all nitric acid from the filtrate by repeated evaporation with pure hydrochloric acid, and determine the sulphuric acid as directed in 132. The metal, metallic oxide r or carbonate, which remains undissolved, is determined, according to circumstances, either by direct weighing or in some other suit- able way. In the presence of lead, before filtering, pass carbonic acid tlirough the solution of the fused mass, to precipitate the small quantity of that metal which has passed into the alkaline solution. Should the sulphides, on the contrary, lose sulphur on heat- ing, the finely powdered compound is mixed with 4 parts sodium carbonate, 8 parts nitre, and 24 parts pure and perfectly dry sodium chloride, and the process otherwise conducted as already given. b. Oxidation by Potassium Chlorate. The oxidation of metallic sulphides by a mixture of potassium chlorate and sodium carbonate has been repeatedly recommended. It is advantageous- in so far that the sulphuric acid in the melt may be more readily converted into barium sulphate than when nitrates are present; on the other hand, it is dangerous because, when the mixture is used in the proportions usually recommended 1 part sulphide, 3 parts potassium chlorate, and 3 parts sodium carbonate (or 4 parts sodium-potassium carbonate) many sulphides, e.g. , fahl- erz, antimony sulphide, etc., afford violent explosions. f Also with many sulphides, like iron pyrites and copper pyrites (Fu. MOHR), the decomposition is not complete. Great caution must hence be exercised in using potassium chlorate in this method. * If gas not free from sulphur is used for heating, some sulphur is likely to be absorbed (PRICE, Journ. Chem. Soc. (2), n, 51). If a platinum crucible is used, do not raise the heat more than is necessary or the crucible will be attacked. \ Annal. d. Chem. u. Pharm., cvn, 128. 564 DETERMINATION. [ 148. II. EOSE recommends taking 6 to 8 parts of sodium carbonate and 1 part of potassium chlorate to 1 part of substance. I. Oxidation ly Chlorine Gas (after BERZELIUS and II. EOSE, especially suitable for sulphosalts of complicated composition). The following apparatus (Fig. 105), or one of similar con- struction, is used. Corks should be used, not india-rubber stop- Fig. 105. pers, and wherever there is an india-rubber connection the glass tubes should be close to each other. A is a chlorine-evolution flask,* B contains concentrated sul- phuric acid, and C calcium chloride. The sulphide to be decom- posed is placed in the bulb-tube D, the straight tube of which should be rather narrow and somewhat inclined, to prevent the heavy fumes of the chlorine compound from returning. .S'is the receiver containing water (or, if antimony is present, a solution of tartaric acid in diluted hydrochloric acid), F is a U-tube contain- ing water, and G conducts the escaping chlorine into a carboy containing moist calcium hydroxide. When the apparatus is arranged weigh off the substance in a narrow glass tube sealed at one end, and carefully transfer * 18 parts of salt mixed with 15 parts finely powdered manganese dioxide are treated with a perfectly cold mixture of 45 parts sulphuric acid and 21 parts of water. On shaking, chlorine is evolved, and when the evolution slackens, it may be promoted by a gentle heat. 148.] STLPIIUR. 565 from this to the bulb D in the manner shown in Fig. 106, so that no part of the substance is allowed to get into the ends of the bulb-tube. When the apparatus is filled with chlorine, con- nect D and (7, and allow the chlorine to act on the sulphide, and at first without applying heat. When no further change is observed, and the receiver E is completely filled with chlorine, heat the bulb D gently, and take care to keep the tube warm also, in order to prevent it from being closed up by the sublimate from a volatile chloride. The sulphide is completely decomposed by the chlo- rine, the metals being converted into chlorides, part of which remain in the bulb and part (the volatile ones, like antimony, arsenic, and mercury chlorides) distil over into the receiver. The sulphur unites with the chlorine to form sulphur chloride, which flows into the receiver E, where, coming into contact with water, it decomposes, hydrochloric and thiosulphuric acids being formed and sulphur precipitating. The thiosulphuric acid in turn decom- poses into sulphur and sulphurous acid, and this last is converted by the action of the chlorine in ^into sulphuric acid. The final result of the decomposition is hence sulphuric acid, with more or less precipitated sulphur. As the separation of sulphur renders troublesome the further treatment of the contents of the receiver, the separation is usually prevented by slowly heating, so that only small portions of sulphur chloride reach the fluid (saturated with chlorine) in E. The operation is concluded when no more products- (excepting perhaps a little ferric chloride, the complete expulsion of which need not be waited for) distil over from the bulb. Then heat the bulb- tube from D to in a manner to drive all the sulphur chloride and volatile metallic chlorides into E, or at least to the end of the bulb-tube. Let the apparatus stand undisturbed for a short time longer, then cut off the tube under the bend 0, and close the separated end, containing usually a part of the volatile chlorine compounds, with a smooth cork, or by inverting over it a glass tube sealed at one end and moistened within. Let the whole now stand for 24 hours to allow the volatile metallic chlorides to absorb moisture and thus become soluble in water without generating heat. The chlorides in the cut-off end are dissolved in diluted hydrochloric 566 DETERMINATION. [ 148. acid, the tube-end rinsed out, and the solution added to the con- tents of the tubes E and F\ a very gentle heat is applied until the free chlorine is expelled, and the fluid is then allowed to stand until the sulphur, if any is present, has solidified. The sulphur is filtered off on a weighed filter, washed, dried, and weighed. The filtrate is precipitated with barium chloride ( 132), by which operation the amount of that portion of the sulphur is determined which has been converted into sulphuric acid. The fluid filtered from the barium sulphate contains, besides the excess of barium chloride added, also the volatile metallic chlorides, which latter are finally determined in it by the proper methods, which will be found in Section Y. The chloride remaining in the bulb-tube is either at once wc-ighed as such (silver chloride, lead chloride), or where this is impracticable as in the case of copper, for instance, which remains partly as cuprous, partly as cupric chloride it is dissolved in water, hydrochloric acid, nitrohydrochloric acid, or some other suitable solvent, and the metal or metals in the solution are determined by the methods already described, or which will be found in Section V. To be enabled to ascertain the weight of the bulb-tube con- taining silver chloride, it .is advisable to reduce the chloride by hydrogen gas, and then dissolve the metal in nitric acid. In cases where you have only to estimate the, sulphur, say in substances containing also sulphuric acid, O. LINDT* recommends conducting the chloride of sulphur and the volatile metallic chlorides into pure solution of soda, when decomposition immedi- ately takes place, producing sodium sulphide, sodium thiosulpliate, sodium chloride, and hypochlorite. When the decomposition is over, continue passing the chlorine for two hours through the soda, evaporate then to dryness, ignite the residue cautiously to destroy the sodium chlorate, dissolve in water, and estimate the sulphuric acid according to 13i5. c. Oxidation ~by Oxide of Mercury (after BUNSENJ. This method, which will be found in detail, 188, is particu- larly suited to the estimation of sulphur in volatile compounds, or in substances which when heated lose sulphur. *Zeitschr.f. analyt. Chem., iv, 870. 148.] SULPHUR. 567 2. Methods in the Wet Way. a. Oxidation of the Sulphur lj Acids yielding Oxygen, or ~by Halogens* a. Weigh the finely pulverized sulphide in a small glass tube sealed at one end, and drop the tube into a tolerably capacious strong bottle with glass stopper, which contains red fuming nitric acid (perfectly free from sulphuric acid f ) in more than sufficient quantity to effect the decomposition of the sulphide. Immediately after having dropped in the tube, close the bottle. When the action, which is very impetuous at first, has somewhat abated, shake the bottle a little ; as soon as this operation ceases to cause renewed action, and the fumes in the flask have condensed, take out the stopper, rinse this with a little nitric acid into the bottle, and then heat the latter gently. aa. The whole of the Sulphur has been oxidized, the Fluid is perfectly clear : J Evaporate with some sodium chloride, towards the end adding pure hydrochloric acid repeatedly, cooling the dish each time before adding the acid. Dilute with much water, and determine the sulphuric acid as directed 132. Make sure that the precipitate is pure ; if it is not, purify it according to 132. Separate the bases in the filtrate from the excess of the barium salt by the methods given in Section Y. bb. Undissolved Sulphur floats in the Fluid: Add potassium chlorate in small portions, or strong hydrochloric acid, and digest some time on a water-bath. This process will often succeed in dis- solving the whole of the sulphur. Should this not be the case, and the undissolved sulphur appear of a pure yellow color, dilute with water, collect on a weighed filter, wash carefully, dry, and weigh. After weighing, ignite the whole, or a portion of it, to ascertain whether it is perfectly pure. If a fixed residue remains (consisting * In presence of lead, barium, strontium, calcium, tin, and antimony, method b is preferable to a. f To test for sulphuric acid in nitric or hydrochloric acid, it is necessary to evaporate on a water-bath nearly to dryness and take up with water before add- ing barium chloride. When the acid cannot be got pure, determine the sul- phuric acid and allow for it. % This can of course be the case only in absence of metals forming insoluble salts with sulphuric acid. If such metals are present, proceed as in bb, as it is in that case less easy to judge whether complete oxidation of the sulphur has been attained. 568 DETERMINATION. [ 148.. commonly of quartz, gangue, &c., but possibly also of lead sul- phate, barium sulphate, &e.), deduct its weight from that of the impure sulphur. In the filtered fluid determine the sulphuric acid as in aa, calculate the sulphur in it, and add the amount to that of the undissolved sulphur. If the residue left upon the ignition of the undissolved sulphur contains an insoluble sulphate, decompose this as directed in 132, and add the sulphur found in it to the principal amount. In the presence of bismuth, the addition of potassium chlorate or of hydrochloric acid, is not advisable, as chlorine interferes with the determination of bismuth. /?. Mix the finely pulverized metallic sulphide in a dry flask, by shaking, with powdered potassium chlorate (free from sulphuric acid), and add moderately concentrated hydrochloric acid in small portions. Cover the flask with a watch-glass, or with an inverted small flask. After digestion in the cold for some time, heat gently,, finally on the water-bath, until the fluid smells no longer of chlo- rine. Proceed now as directed in X or THE \ UNIVERSITY } OF . 149.] NITRIC ACID. 575 fi. Iii nitrates from which the basic metals are precipitated by barium or calcium hydroxides or their carbonates (or by barium sulphydrate, recently precipitated and free from barium thiosul- phate GLAUS*), the nitric acid may be estimated with great accuracy if no other acids are present by filtering, after pre- cipitation has been effected, warm or cold, passing carbonic acid through the filtrate, if necessary, till all the barium is precipitated, warming, filtering, and determining the barium in the filtrate by sulphuric acid. 1 at. of the barium corresponds to 1 mol. nitric anhydride (N a O 5 ). In case of bismuth-nitrate, boil after adding the barium hydroxide until the separated oxide is perfectly yellow. (HUGE; LuDDECKEf). y. In many nitrates the bases of which are precipitable by hydrogen sulphide, the nitric acid may be determined accord- ing to GIBBS by adding to the salt in solution about its own weight of some neutral organic salt, e.g., Rochelle salt, and throwing down the metal by H 2 S. The filtrate and washings are brought to a definite bulk, and the free acid is determined in aliquot por- tions alkalimetrically.J d. Methods based upon the decomposition of Nitric Acid ~by Ferrous Chloride. a. PELOUZE If was the first to utilize the action of free nitric acid on ferrous chloride in the determination of nitric acid. The decomposition is as follows : 6FeCl a + 2KX0 3 + 8HC1 = 3Fe,Cl. + 2KC1 + 4H a O + N a O a . PELOTJZE used a known quantity of ferrous chloride in excess, and titrated the excess with potassium permanganate. The method used by him, and given in the foot-note, affords variable results, sometimes good, sometimes not reliable ; on this point all * Zeitsehr. /. analyt. Chem., I, 372. \lb., vi, 233. | Amer. Jour. Sci. t XLIV, 209. If Journ. /. prakt. Chem. XL, 324. Dissolve 2 grm. piano wire in 80 to 100 c. c. pure concentrated hydrochloric acid with the aid of heat, in a 150-c. c. flask the cork of which is fitted with a glass tube. Then add 1' 2 grm. of the potassium nitrate, or an equivalent quan- tity of another nitrate to be analyzed, stopper, and rapidly heat to boiling. After five or six minutes pour the fluid, which has again cleared, into a larger flask, dilute with much water, and estimate the ferrous chloride present with. permanganate. 576 DETERMINATION. [ 149. who used the method agree (compare FK. MOHR,* and ABEL and BLOXAM f). The numerous experiments made in my own labor- atory also confirm this. The reasons for the lack of accuracy of the method are as follows : a. Action of the air on the nitric oxide in the flask in the presence of the aqueous vapor therein, whereby nitric acid is regenerated ; this is the principal cause of inaccuracy. l>. Incomplete expulsion of the nitric oxide from the liquid, in consequence of which more permanganate is reduced than cor- responds to the ferrous chloride present; this is to be appre- hended only in dilute solutions. c. Escape of nitric acid before it has acted upon the ferrous chloride; this is to be apprehended when the liquid is boiled rapidly after adding the nitrate, and when the excess of ferrous chloride is comparatively small. d. Occasionally loss of iron from incautious boiling ; this is to be apprehended when a part of the ferrous chloride deposits in solid form on the sides of the vessel above the fluid. I have succeeded in so modifying the process as to avoid all these sources of error and to obtain results which, so far as relia- bility arid accuracy are concerned, a.re perfectly satisfactory. My process is as follows : Select a tubulated retort of about 200 c. c. capacity, with a long neck, and fix it so that the latter is inclined a little upwards. Introduce into the body of the retort about 1*5 grm. fine piano- forte wire, accurately weighed, and add about 30 or 40 c. c. pure fuming hydrochloric acid. Conduct now through the tubulure, by means of a glass tube reaching only about 2 cm. into the retort, hydrogen gas washed by solution of potassa, or pure carbonic acid, and connect the neck of the retort with a U-tube containing some water. Place the body of the retort on a water-bath, and heat gently until the iron is dissolved. Let the contents of the retort cool in the current of hydrogen gas or carbonic acid ; increase the latter, and drop in, through the neck of the retort, into the body, a small tube containing a weighed portion of the nitrate under examination, which should not contain more than about 0*2 erm. * LehrbucJi der Titrirmeihode, i, 216. f Quart. Journ. of the Chem. Soc., ix, 97 ; also Journ. f. prakt. Chem., I.XIX, 262. 149.] NITRIC ACID. 577 of N a O 5 . After restoring the connection between the neck and. the U-tube, heat the contents of the retort in the water-bath for about a quarter of an hour, then remove the water-bath, heat with the lamp to boiling, until the fluid, to which the nitric oxide had imparted a dark tint, shows the color of ferric chloride, and con- tinue boiling for some minutes longer. Care must be taken to give the fluid an occasional shake, to prevent the deposition of dry salt on the sides of the retort. Before discontinuing boiling, increase the current of hydrogen or carbonic-acid gas, so that no air may enter through the U-tube when the lamp is removed. Let the contents cool in the current of gas, dilute copiously with water, and determine the iron still present as ferrous chloride volumetrically by potassium dichromate or permanganate 335-4 of iron con- verted by the nitric acid from ferrous to ferric chloride correspond to 108*08 (N a O B ). My test-analyses of pure potassium nitrate gave 100-1 100-03 100-03, and 100-05, instead of 100.* [The iron remaining as ferric chloride may also be determined by sodium thiosulphate.] fi. Since we have learned to titrate ferric salts di- rectly with accuracy, it is, as a rule, more convenient and exact not to estimate (as in a) the residual ferrous salt, but to determine the ferric salt produced, as first pointed out by C. D. I would recommend the following method as the best. J Besides the requisites for titrating ferric chloride by means of stannous chloride (p. 327), there is required a solution prepared by dissolving 100 grin, ferrous sulphate as free from ferric salt as possible in 150 to 200 c. c. of hydrochloric acid (sp. gr. 1-1 1-2) by the aid of heat in a 500-c. c. flask, and finally filling the flask to the mark with fuming hydrochloric acid, and shaking. As, however, a solution absolutely free from ferric salt cannot be obtained, esti- mate according to the method given on page 327 how much stan- nous chloride solution is required to reduce the ferric chloride present in 50 c. c. of the solution made as a~bove. It is advisable * Annal. de Chem. u. Pharm., cvi, 217. \Journ.f- prakt. Chem., LXXXI, 421. i It is particularly convenient when several analyses are to be made ; when only one or two estimations are to be made, however, the iron wire may be dis- solved in hydrochloric acid as in a. 578 DETERMINATION. [ 149^ to heat the solution in an atmosphere of carbon dioxide, and to titrate it either immediately before or directly after the analysis. Place the weighed nitrate (the quantity must be such as to con- tain not more than 0*2 grin, nitric acid) in a long-necked flasK, provided with a doubly- perforated stopper carrying two glass tubes. One of these reaches to the body of the flask, while the other but just enters it. Through the former pass in a current of carbon dioxide, and when all the air has been displaced introduce 50 c. c. of the hydrochloric-acid solution of ferrous sulphate ; con- tinue to pass in the carbon dioxide for some time longer, then heat, at first gently for some time, and then gradually to boiling, until the liquid has lost its blackish color and exhibits the pure color of ferric chloride, and until the escaping gas, passed into dilute starch paste containing a little potassium iodide, ceases to exhibit the blue color of starch iodide. Now remove the stopper from the flask, rinse off the longer tube if necessary, dilute the residue with double its volume of water, and estimate the ferric chloride as on p. 327. The cooling for determining the slight excess of stannous chloride with iodine is best conducted in a current of car- bon dioxide. From the stannous-chloride solution used altogether deduct first the small excess ascertained by the iodine solution, and then the small quantity corresponding to the ferric salt pre- sent in 50 c. c. of the acid ferrous-sulphate solution. The re- mainder gives the iron in the ferric salt produced ; and, when multiplied by 0*32224, gives the nitric acid. This factor is ob- tained thus : 6 eq. of iron (335-4) : 1 eq. of N 2 O B (108 -08) : : ferric iron present : x. It will be seen that it is best, once for all, to multiply the- known quantity of iron in the ferric-chloride solution used for standardizing the stannous-chloride solution by the above factor, and to mark the product on the label as the quantity of nitric acid corresponding to 10 c. c. of ferric -chloride solution. If no standardized ferric-chloride solution is at hand, the stannous-chlo- ride solution may be standardized directly against nitric acid by adding a weighed quantity of pure potassium nitrate to 50 c. c. of the acid ferrous-sulphate solution, and then determining the fer- ric chloride formed according to the method given above. The 149.] NITRIC ACID. 679 results are perfectly satisfactory if the process is properly carried out, and the estimations succeed each other immediately.* y. SCHLOSING'S method, f which was employed more espe- cially for estimating nitric acid in tobacco, affords the important advantage that it ^may be employed in the presence of organic matter. Numerous experiments have shown this method to be thoroughly satisfactory. It is conducted in the apparatus shown in Fig. 107. Fig. 107. The dissolved nitrate is introduced into the flask A, the drawn-out neck of which is connected by means of a rubber tube, a, with a narrow glass tube, &; c is another rubber tube, 15 cm. long, and connected with ~b. The solution of the salt, which must be neutral or alkaline, is boiled down to a small volume, the aqueous vapor completely expelling all the air from A and the tubes, c is then immersed in a solution of ferrous chloride in hydrochloric acid contained in a glass vessel, the lamp removed, and the receding regulated by compressing the tube c with the fingers. When the iron solution is almost entirely absorbed a little hydrochloric acid is allowed to recede, in separate portions, three or four times, in order to entirely free the tube from fer- rous chloride, this being absolutely necessary. Before the air can force its way into the tubes c is closed by means of an iron com- pression-cock, and its end immersed into the mercury in the trough and brought up under the bell B. The lamp is again * Zeit&clir. /. analyt. Chem., i, 38. HOLLAND gives a method in which the use of an indifferent gas is unnecessary. Zeitschr. f. analyt. Chem., vin, 452 ; . News, xvn, 219. \Annal. de Chem. 3 ser., XL, 479 ; Journ.f. prakt., Chem., LXII, 142. 580 DETERMINATION. [ 149. placed under A, in order to let the reaction proceed, and the compression-cock, immediately replaced by the pressure of the fingers, this compression being in turn relieved as soon as a press- ure is felt from within. The reaction is ordinarily at an end in eight minutes, when c is removed, from under B. B is a small bell- jar, made from an adapter, and must have a capacity three to four times the volume of gas to be received. In cases where the evolution of gas is impetuous it is at times necessary to immerse it in the trough in order to better con- 108. dense the vapors. The upper part of B is drawn out as shown in Fig. 108, in order to facilitate its ready insertion into the rubber tube and also the breaking off of its point. The bell is first filled with water in order to expel all the air, and then with mercury ; well-boiled milk-of-lime is then introduced by means of a curved pipette. The nitric oxide entering B is thus freed from the slightest trace of acid vapor. The nitric oxide is now to be transferred into the flask (7, there to be reconverted into nitric acid by means of oxygen. The flask C contains a little water ; it is connected by means of the rubber tube d with the glass tube . The solution of the nitrate in the flask is further concentrated by boiling, and finally the lower end of the tube efg h is brought into the soda solution so that a part of the steam escapes through it. After a few minutes the rubber tube at g is pressed together with the fingers; if the air has been completely displaced from the flask by boiling, the soda solution will rise suddenly in the tube as in a vacuum, and a slight blow against the finger will be perceptible. In this case the rubber tube at g is closed with a clamp and the steam is allowed to escape through a b o d until only 10 c. c. of fluid remain in the flask. The lamp is now removed and the rubber tube at c is closed with a clamp, and the tube c d filled by a jet of water. If an air bubble remains in the rubber tube at c, it must be removed by pressure with the fingers. The graduated measuring tube is now brought over the upcurved end of the evo- lution tube efg h so that the end rises in it 2-3 cm. The flask must next be allowed to stand a few minutes until a partial vacuum is produced in it, which is manifested by a contraction of the rubber tubes at c and g. A nearly saturated solution of ferrous chloride is poured into a small beaker, the upper part of which is marked so as to show the space occupied by 20 c.c. ; two other beakers must also be at hand partly filled with concentrated hydrochloric acid. The tnbecd is now dipped into the ferrous- chloride solution, and the clamp at c is loosened until 15-20 c.c. are drawn into the flask. The ferrous chloride remaining in the tube is next removed by drawing in a small quantity of hydrochloric acid in two suc- cessive portions. Small bubbles may frequently be observed at #, occasioned by evolution of hydrochloric gas caused by dimin- ished pressure in the flask. They disappear almost completely so soon as the pressure rises. Heat is applied, at first very gently, until the rubber tubes at o and g are slightly expanded ; then the rubber tube at g is held com- pressed by the fingers, the clamp being removed, until the pressure becomes stronger, when the gas is allowed to pass over to the grad- uated tube. Toward the end of the operation heat is increased and distillation continued until the volume of gas in the measuring tube no longer increases. The hydrochloric gas, abundantly evolved in the last part of the process, is absorbed with violence by the soda solution with a peculiar clattering sound ; there is nd 584 DETERMINATION. [ 149. danger, however, of breaking the evolution tube if care has been taken to enclose the lower end with a rubber tube as above directed. The measuring tube is brought into a large cylinder containing; cold water, best of 15-18 C., and by means of some suitable fix- ture held wholly submerged in the same. The transfer is effected with the help of a small porcelain dish filled with soda solution. After 15-20 minutes, the temperature of the water in the^ cylinder is ascertained with a sensitive thermometer, and the state of the barometer is also observed. Then the tube is taken hold of at the upper end with a strip of paper or cloth, in order to avoid imparting heat to it by direct contact of the hand, and drawn up perpendicularly so far that the level of the fluids within and without it exactly coincide, and the volume of the gas is read off. From the data thus obtained, the volume which the dry gas would occupy at C. and 760 mm. bar. pressure is to be computed. (See pp. 160, 161, on Calculation of Analyses.) 1 c.c. ]ST a O 3 at C. and 760 mm. bar. pressure corresponds to 0'002415 grm. N 2 O 6 . A condition indispensable for the success of the operation is. the complete expulsion of air from the apparatus in the beginning. When an abundant quantity of nitric acid is present in the sub- stance, enough to produce about 80 c.c. nitrogen dioxide is a suit- able .quantity to use for its determination, and a somewhat larger quantity of ferrous chloride and hydrochloric acid than above indi- cated may be used. An unnecessary amount of these reagents should, however, be avoided, since it is difficult to boil a small quan- tity of nitrogen dioxide out of a large volume of liquid. This method is easy to carry out and gives satisfactory results, e. Methods lased on the conversion of Nitric Acid into Ammonia. On heating a nitrate in an alkaline fluid in which nascent hydrogen is being evolved in sufficient quantity, all the nitric acid of the nitrate is converted into ammonia,* from the volume of which the quantity of nitric acid may be accurately determined. FK. SCHULZE f was the first to base on this principle a method for the estimation -of nitric acid, and he was soon followed by *The conversion takes place in acid solution also, but is then only partial (L. GMELIN; MARTIN). f Chem. CentralbL, 1861, 657 and 833. 149.J . NITEIC ACID. 585 W. WOLF,* HARCOURT^ and SIEWERT. :f Later on BUNSEN, and also HAGER, | modified both the methods and the apparatus. SCHULZE effected reduction with platinized zinc ; W. WOLF, HAR- COURT, and SIEWERT with zinc and iron filings; BU.NSEN with a zinc-iron spiral. Zinc and iron appear to afford the most satis- factory results, therefore I shall first describe HARCOURT' s process, in which an aqueous potassa solution is used, and then describe SIEWERT' s method, in which an alcoholic potassa solution is em- ployed. If organic substances are present, these methods do not afford good results (FRUHLIN&T). The reliability of the results, however, is questioned even when organic substances are absent. While the test-analyses afforded HARCOURT and SIEWERT uniformly good results, WOLF (loo. cit.) states that the three following con- ditions are essential for the success of the method: 1. The con- version of nitric acid into ammonia must take place in the cold (on heating, while the hydrogen is being evolved, some ammonia is lost, most probably from the escape of nitrogen as such). 2. A copious and uniform evolution of hydrogen is necessary, and is best secured by using zinc in conjunction with iron. 3. The potassa or soda must be dissolved in not less than 7 or more than 8 parts of water. It will be observed that these con- ditions are at direct variance with the directions given by HAR- COURT. FINKENER *'* rejects all the methods based on the above principle, because, although all the nitric acid is decomposed, yet all the nitrogen is not converted into* ammonia. I have not studied the methods sufficiently to give a decided opinion, but I must say that in my laboratory the methods of HARCOURT and of SIEWERT have generally given good results. HARCOURT employs the apparatus shown in Fig. 110. Bring the tube e into a vertical position by turning it half round in the tubulure ; then run into it from a burette standard acid (more than is sufficient to fix the ammonia) into d, add a little litmus * Chem. CentralbL, 1862, 379; also Journ. f. prakt. Chem., LXXXIX, 93, and Zeitechr.f. analyt. Chem., n, 401. f Journ. of the Chem. Soc., xv, 385; also Zeitschr.f. analyt. Chem., ir, 14. \Annal. d. Chem. u. Pharm., cxxv. 293. Zeitschr.f. analyt. Chem., x, 414. 1/6., x, 334. \Landwirthschaftl. Versuchsstat., vin, 473. ** H. KOBE, Handb. d. analyt. Chem. , 6 Aufl. von FINKENER, n, 829. 586 DETERMINATION. [ 149. tincture, turn the tube e back to its horizontal position, and let a little more of the standard acid run into the bulbs. Now remove Fig. 110. the flask #, while its stopper carrying the glass tube, and also the small flask J, containing a little water, are allowed to retain their position on the sand-bath unchanged. Into a introduce about 50 grm. of finely granulated zinc and about 25 grm. of iron filings (purified by first sifting and then heating in a current of hydro- gen), add the weighed quantity of nitrate (for instance 0*5 potassium nitrate), 20 c. c. of water, and 20 c. c. of potassa solu- tion of sp. gr. 1*3. That part of the sand-bath c, directly under #, is now heated until the contents of a boil. When the bubbles of air and hydrogen pass quietly through the bulbs , a loss of ammonia is not to be feared. As soon as distillation begins, place the lamp so that also the contents of the flask l> boil gently. In this -manner the fluid is by one operation distilled twice, and the traces of potassa carried over from a are completely retained in J. The end of each exit-tube, as a further precaution, is drawn out and bent upwards in the form of a hook. The distillation re- quires from 1 to 2 hours. It may be stopped when the hydro- gen, which is evolved more freely as the potassa solution becomes more concentrated, has passed through the bulb-tube e for 5 or 10 minutes regularly. As soon as the fluid e has receded to d on the cooling of the apparatus, remove the rubber stopper from the small tubulure/*, and pass a stream of water through the condenser in order to rinse the last traces of ammonia into the receiver. Once 149.] NITRIC ACID. 587 more bring the tube e to a vertical position by a half-turn, rinse it out with water, then remove it, and close the tubulure of the re- ceiver with a cork. Finally remove the receiver, and rinse off the outside of the lower end of the condenser, and proceed to titrate the residual free acid. The metals remaining in a need only be washed with water, diluted acid, and again with water, in order . to render them serviceable for a second determination. Metals which have been once used evolve hydrogen more slowly than do bright zinc and recently ignited iron, but the evolu- tion of ammonia proceeds equally well in both cases. Metallic chlorides and sulphates have no influence on the result. If lead is present it appears advisable to add some potassium sulphate. SIEWEET employed for every gramme of saltpetre 4 grm. iron filings and 8 to 10 grin, zinc filings, and also 16 grin, potas- sium hydroxide and 100 c. c. alcohol of 0-825 sp. gr. By the iise of alcohol the danger of the boiling fiuid receding is avoided. The apparatus used by him consists of a 300- to 350-c. c. flask with an evolution tube connected with the two flasks./? and C, arranged as shown in Fig. 111. These flasks have a capacity of 150 to 200 c. c. each and contain standard acid. The connecting tube b is cut off obliquely at both ends; c serves for the introduction of a strip of litmus paper during the operation, and after the latter is complete, for the transference of the liquid from one flask to the other at will. After put- ting the apparatus together, the disen- gagement of gas may be allowed to pro- ceed first in the cold, or it may be Tj^J _ 1 "1 "I assisted from the beginning by the aid of a small flame. After half an hour the ammonia formed begins to pass over in proportion as the alcohol distils off. As soon as the latter has completely disappeared from the evolution flask, apply heat very cautiously (in order to drive off the last traces of am- monia) until steam appears in the evolution tube ; or 10 to 15 c. c. of alcohol are rapidly introduced once or twice into the evolution flask and distilled off. 588 DETERMINATION. [ 149. f. Method of estimating nitric acid from the loss of hydro- gen, after FK. SCHULZE.* On dissolving aluminium in potassa lye, a potassium-alumin- ium compound is formed and hydrogen is evolved, the quantity Fig. 112. evolved corresponding to the weight of the aluminium dissolved. If a nitrate is added to the mixture evolving the hydrogen, less hydrogen is obtained than were no nitrate present, since part of the nascent hydrogen serves to convert the ]ST a O 5 of the nitrate *Zeitsc7ir.f. anatyt. Chem., n, 300. 149.] NITRIC ACID. 589 into ammonia (K a O 6 + 16H 2OTI 3 + 5H 9 O), and the loss of hydrogen is, of course, proportional to the quantity of N a 6 con- verted into ammonia. Since, according to FR. SCHULZE, this conversion is complete when the process is conducted slowly (FINKENER, * however, contradicts this), and since a small quan- tity of N 2 O 6 is able to effect a relatively large deficit of hydrogen, this method can be applied for the accurate determination of even small quantities of nitric acid. The method cannot, according to E. SCHULZE, f be used if organic matter is present, because then the results are inaccurate. In such a case the substance must first undergo the following preliminary treatment: Heat with dilute potassa lye until all ammonia is expelled, add a concentrated solution of pure potassium permanganate until the fluid retains a red color even on continuous boiling during 10 minutes, then add a little formic acid to decompose the excess of permanganate present, filter, wash, concentrate the filtrate, neutralize accurately with dilute sulphuric acid, and then subject the fluid so obtained, and concentrated by evaporation, if necessary, to the treatment to be described below (FRANZ SCHULZE^:). I shall first describe the apparatus used, resembling KNOP'S azotometer || , and then the process. The flask A (Fig. 112) has a capacity of about 50 c. c. Into its neck is ground airtight the tube ./?, expanded above into a bulb. A glass rod, extract with water, and determine the sodium chloride in the solution. For small quantities of fluid this method will be found convenient, y. A. MITSCHER- LICH recommends to mix the filtrate with sulphuric acid, evapo- rate to dryness, ignite the residue, extract the sodium sul- phate with water, and determine it according to 98, 1. These methods, of course, yield the sodium salt in a pure condition only when the separation of the potassium has been perfect. They present the advantage that the sodium salt is brought under one's eyes and may be tested after weighing. Should the solution contain sulphuric acid, it may be in 2 presence of hydrochloric acid or of some volatile acid, convert the alkalies first into normal sulphates ( 97, 98), and weigh them as such. For the estimation of the potassium, one of the two following methods may be used : a. First convert the sulphates into chlorides and then pro- 602 SEPAEATIOTT. [ 152. ceed as above. For this purpose barium salts were formerly employed, or, better, an alcoholic solution of strontium chloride. The barium sulphate, however, carries down considerable quan- tities of alkali salt, and the strontium sulphate noticeable quantities ; hence the employment of these reagents, more par- ticularly barium, cannot be recommended. H. ROSE advises repeated ignition of the alkali sulphates with ammonium chloride till the weight remains constant; this process is simple and well adapted for small quantities ; no loss of alkali need be feared if the heat is not unnecessarily raised. L. SMITH advises the use of lead salts. Dissolve the alkali sulphate, precipitate with pure neutral lead acetate, avoiding a large excess, add some alcohol, filter, precipitate the excess of lead with sulphuric acid, and evaporate to dryness with addition of sulphuric acid. This method, when carefully conducted, yields excellent results. /?. Precipitate the potash directly out of the solution of the sulphates. R. FINKENER* gives the following process : To the rather dilute solution of the salts in a capacious porcelain dish add platinic chloride in quantity more than sufficient to throw down all the potassium, evaporate on a water-bath down to a few c.c., allow to cool, add, at first in small quantities, 20 times the volume of a mixture of 2 parts absolute alcohol- and 1 part ether, with stirring ; filter after a short time, and wash the precipitate with alcohol and ether till the washings are colorless. If, when the alcohol and ether are first added, a strong aqueous solution of sodium sulphate separates, add some hydrochloric acid till the fluids mix. Dry the precipitate con- sisting of potassium platinic chloride and sodium sulphate, heat with the filter in a porcelain crucible till the filter is car- bonized, then in a current of hydrogen to scarcely visible redness extract the residue with hot water, ignite the platinum in the air, weigh and calculate from the weight the quantity of potassium. The separation of potassium from sodium by platinic chloride gives results which are fully satisfactory, and at all events far more exact than any method depending on another principle ; provided that the platinum solution is pure and the operations have been carefully performed in accordance with the directions. If you have any occasion to doubt the perfect II. ROSE, Handbuch der analyt. Chem., 6. Aufl. von FINKENER, n, 923. 152.] BASES OF GROUP I. 603 purity of the weighed double salt, you may always dissolve it in boiling water, evaporate with addition of a little platinum solution, and re weigh the salt thus purified. Where a series of analyses is being made, the potassium in the potassium-platinic chloride may be volu metrically estimated. For this purpose triturate it with double its quantity of pure sodium oxalate (free from chlorides), heat the mixture in a platinum crucible to fusion, leach the residue with water, neutralize the filtrate nearly with acetic acid, determine the chlorine in the alkali chloride with decinormal silver solution ( 141, I., 5, <*), and calculate 1 eq. of potassium for 3 eq. chlorine. If the quantities of potassium-platinic chloride are very small, moisten with a cencentrated solution of neutral potassium oxalate, dry, ignite in, a covered crucible, and pro- ceed as above. The separated platinum, if weighed, will afford a good control (F. MOHR *). b. AMMONIUM FKOM SODIUM. The process is conducted exactly as in #, when the alka- 3 lies are present as chlorides. See also 99, 2. If potassium also is present, the precipitate produced by platinic chloride is a mixture of ammonium platinic chloride and potassium platinic chloride ; in which case the weighed precipitate is cautiously ignited for a sufficient length of time, but not too strongly, until the ammonium chloride is expelled, the gentle ignition continued in a stream of hydrogen or with addition of oxalic acid, the residue extracted with water, a few drops of hydro- chloric acid added if oxalic acid was employed, and the potas- sium chloride in the solution determined as directed 97, 3. The weight found is calculated into potassium platinic chloride, and the result deducted from the weight of the whole precipi- tate : the difference gives the animmonium platinic chloride. The weighing of the separated platinum affords a good control. The method is seldom employed, as that given in V. yields more exact results. * Zeitschr. f. analyt. Chem., xn, 137. 604 SEPARATION. [ 152. 2. Methods based upon the Volatility of Ammonium Salts and Ammonia. AMMONIUM FKOM POTASSIUM AND SODIUM. a. The salts of the alkalies to he separated contain the same 4 volatile acid, and admit of the total expulsion of their water hy drying at 100, without losing ammonia (e.g., the chlorides). Weigh the total quantity of the salts in a platinum crucible, and heat, with the lid on, gently at .first, but ultimately for some time to faint redness ; let the mass cool, and weigh. The decrease of weight gives the quantity of the ammonium salt. If the acid present is sulphuric acid, you must, in the first place, take care to heat very gradually, as otherwise you will suffer loss from the decrepitation of ammonium sulphate ; and, in the second place, bear in mind that part of the sulphuric acid of the ammonium sulphate remains with the fixed alkali sulphates, and that you must accordingly convert them into normal salts, by ignition in an atmosphere of ammonium car- bonate, before proceeding to determine their weight (compare 97 and 98). Ammonium chloride cannot be separated in this manner from fixed alkali sulphates, as it converts them, upon ignition, partly or totally into chlorides. b. Some one or other of the conditions given in " a " is not fulfilled. If it is impracticable to alter the circumstances by simple 5 means, so as to make the method a applicable, the fixed alkalies and the ammonium must be determined separately in different portions of the substance. The portion in which it is intended to determine the potassium and sodium is gently ignited until ammonium is .completely expelled. The fixed alkalies are con- verted, according to circumstances, into chlorides or sulphates, and treated as directed in 1, 2, or 6. The ammonium is esti- mated in another portion according to 99, 3. 3. Indirect Methods. Of course, a great many of these may be devised ; but the 6 following is the only one in general use. POTASSIUM FKOM SODIUM. Convert both alkalies into neutral sulphates or chlorides 152.] BASES OF GROUP I. 605 ( 97 and 98), and weigh as such ; estimate the sulphuric acid ( 132) or chlorine ( 141); and from the amount of this cal- culate the quantities of the sodium and potassium (see " Calcu- lation of. Analysis," 200*). The indirect method of determining sodium and potassium is applicable only in the analysis of mixtures containing toler- ably large quantities of both bases ; but where this is the case, the process answers very well, affording also, more particularly, the advantage of expedition, if the chlorine in the weighed chlorides is titrated ( 141, I., &). Supplement to the First Group. SEPARATION OF LITHIUM FROM THE OTHER ALKALIES. Lithium may be separated from potassium and sodium in the 7 indirect way, and by two direct methods: a. Treat the nitrates or the chlorides, dried at 120, with a mixture of equal volumes of absolute alcohol and anhydrous ether, digest at least for 24 hours, with occasional shaking (the salts must be completely disintegrated), decant rapidly on to a filter covering the funnel, and treat the residue again several times with smaller portions of the mixture of alcohol and ether. Determine, on the one part, the undissolved potassium and sodium salts ; on the other, the dissolved lithium salt, by dis- tilling the fluid off, and converting the residue into sulphate. This method is apt to give too much lithium, as the potassium and sodium salts, especially the chlorides, are not absolutely insoluble in a mixture of alcohol and ether. The results may be rendered more accurate by treating the impure lithium salt, obtained by distilling off the ether and alcohol, once more with alcohol and ether, with addition of a drop of nitric or hydro- chloric acid, adding the residue left to the principal residue, and then converting the lithium salt into sulphate. If the salts, which it is intended to treat with alcohol and ether, have been ignited, however so gently, caustic lithia is formed in the case of the chloride by the action of water and lithium carbonate by attraction of carbonic acid ; in that case it is neces- sary, therefore, to add a few drops of nitric or, as the case may * Other methods are given by STOLE A (Zeitschr. f. analyt. Chem., n, 397) and MOHR (lb., vn, 173). 606 SEPARATION. [ 152. be, hydrochloric acid, in the process of digestion. The separa- tion of the alkali chlorides by means of ether- alcohol was first proposed by KAMMELSBERG * and later on recommended by jENZscH.f This method, however, can yield only approximate results, as the lithium salt obtained on evaporating the alcoholic extract is found by spectroscopic examination to be always im- pure (DlEHL^). If we have to separate the sulphates, they must be converted into nitrates or chlorides before they can be subjected to the above method. This conversion is best effected by means of lead salts, see 2. Ignition with ammonium chloride does not answer for lithium sulphate, nor can the sulphuric acid be removed by barium, or strontium, as the precipitated sulphates would contain lithium (DIEHL). 1). Weigh the mixed alkalies, best in form of sulphates, and 8 then determine the lithium as phosphate according to 100. If the quantity of lithium is relatively very small, convert the weighed sulphates into chlorides (7), separate, in the first place, the principal amount of the potassa and soda by means of alco- hol ( 100), and then determine the lithium (MAYER |). c. When exact results are required, the indirect method is 9 to be preferred. Proceed first according to $, evaporate the spirituous solution of the lithium chloride containing the remain- der of the other chlorides to dryness, heat moderately, weigh, dissolve in water, estimate the chlorine, and calculate therefrom the lithium and sodium or potassium. BUNSEN 1" also applied the method to the indirect estimation of lithium in presence of potassium and sodium by removing the silver from the filtrate, and separating the potassium with platinum. But I must here point out, that according to JENZSCH** the potassium double salt will contain lithium apparently in the form of the platino- chloride of potassium and lithium. The sulphuric acid in weighed quantities of the sulphates of lithium, and of potassium and sodium, cannot be determined as barium sulphate (see end of 7). * Pogg. AnnaL, LXVI, 79. f Ib., civ, 105. \Annal. d. Chem. u. Pharm., cxxi, 97. Ib., cxxi, 98. I Ib., xcvin, 193. *f\Annal. d. Chem. u. Pharm., cxxn, 348. ** Pogg. Annal., civ, 102. 153.] BASES OF GROUP II. 607 The separation of lithium from ammonium may be effected like that of potassium and sodium from ammonium (4 and 5). Second Group. BARIUM STRONTIUM CALCIUM MAGNESIUM. I. SEPARATION OF THE BASIC RADICALS OF THE SECOND GROUP FROM THOSE OF THE FlRST. 153. INDEX. (The numbers refer to those in the margin.) Barium from potassium and sodium, 10, 12. ammonium, 11. Strontium from potassium and sodium, 10, 13. ammonium, 11. Calcium from potassium and sodium, 10, 14. ammonium, 11. Magnesium from potassium and sodium, 15-26. ammonium, 11. A.. General Method. 1. THE WHOLE OF THE ALKALI-EARTH METALS FROM Po- TASSIUM AND SODIUM. Principle on which the method is based : Ammonium car- 10 bonate precipitates, from a solution containing ammonium chloride, only barium, strontium, and calcium. Mix the solution, in which the metals are assumed to be contained in the form of chlorides, with a sufficient quantity of ammonium chloride to prevent the precipitation of the magne- sium by ammonia ; dilute rather considerably, add some ammo- nia, the^n ammonium carbonate in slight excess, let the mixture stand covered for an hour in a moderately warm place, filter, and wash the precipitate with water to which a few drops of ammonia have been added. The precipitate contains the barium, strontium, and cal- cium ; the filtrate the magnesium and the alkalies. So at least we may assume in cases where the highest degree of accuracy is not required. Strictly speaking, however, the solution still contains exceedingly minute traces of calcium and somewhat more considerable traces of barium, as the car- 608 SEPARATION. [ 153. bonates of these two metals are not absolutely insoluble in a fluid containing ammonium chloride; the precipitate also may contain possibly a little ammonium magnesium carbonate. Treat the precipitate according to 154, and the filtrate in rigorous analyses as follows : Add 3 or 4 drops (but not much more) of dilute sulphuric acid, then ammonium oxalate, and let the fluid stand* again for 12 hours in a warm place. If a precipitate forms, collect this on a small filter, wash, and treat on the filter with some dilute hydrochloric acid, which dis- solves the calcium oxalate, and leaves the barium sulphate undissolved. Since a little magnesium oxalate may have sepa- rated with the former, add some ammonia to the hydrochloric solution, filter after the precipitate has settled, and mix the filtrate with the principal filtrate. Evaporate the fluid containing the magnesium and the alka- lies to dryness, and remove the ammonium salts by gentle igni- tion in a covered crucible, or in a small covered dish of platinum or porcelain.* In the residue, separate the magnesium from the alkalies by one of the methods given 15 24. 2. THE WHOLE OF THE ALKALI-EARTH METALS FROM AM- 11 MONIUM. The same principle and the same process as in the separation of potassium and sodium from ammonium (4 and 5). B. Special Methods. SINGLE ALKALI-EARTH METALS FROM POTASSIUM AND SO- DIUM. 1. BARIUM FROM POTASSIUM AND SODIUM. Precipitate the barium with dilute sulphuric acid ( 101, 1, ^), 12 evaporate the filtrate to dryness, and ignite the residue, with addition towards the end of ammonium carbonate ( 97, 1 and '98, 1). Take care to add a sufficient quantity of sulphuric acid to convert the alkalies also completely into sulphates. In exact analyses, in order to save the alkali salts adhering to the * This operation effects also the removal of the small quantity of sulphuric acid added to precipitate the traces of barium, as sulphates of the alkalies are converted into chlorides upon ignition in presence of a large proportion of ammonium chloride. 153.] BASES OF GROUP II. 609 barium sulphate, remove the dry barium sulphate from the filter, heat it with a sufficient quantity of pure strong sulphu- ric acid to dissolve it completely, allow to cool, dilute largely, collect the barium sulphate (now almost absolutely pure) on the first filter, ignite, and weigh. Evaporate the filtrate in a plati- num dish, drive off the sulphuric acid, and estimate the traces of the alkalies. This method is, on account of its greater accuracy, prefer- able to the one in A, in cases where the barium has to be sepa- rated only from one of the two fixed alkalies; but if both alka- lies are present, the other method is more convenient, since the alkalies are then obtained as chlorides. 2. STRONTIUM FKOM POTASSIUM AND SODIUM. Strontium may be separated from the alkalies like barium, 13 t>y means of sulphuric acid ( 102, 1, ); but this method is not preferable to the one in 10, in cases where the choice is permitted (comp. 102). 3. CALCIUM FROM POTASSIUM AND SODIUM. Precipitate the calcium with ammonium oxalate ( 103, 2, 14 &, or), evaporate the nitrate to dryness, and determine the alka- lies in the ignited residue. In determining the alkalies, dis- solve the residue, freed by ignition from the ammonium salts, in water, filter if necessary, acidify the filtrate, according to cir- cumstances, with hydrochloric acid or sulphuric acid, and then evaporate to dryness ; this treatment of the residue is neces- sary, because ammonium oxalate partially decomposes chlorides of the alkali metals upon ignition with formation of alkali car- bonates, except in presence of a large proportion of ammonium chloride. The results are still more accurate than in A, except where ammonium oxalate has been used, after the precipitation by ammonium carbonate, to remove the minute traces of lime from the filtrate. 610 SEPARATION. [ 135. 4. MAGNESIUM FROM POTASSIUM AND SODIUM.* a. Methods Ixised upon the sparing solubility of Magnesium Hydroxide in Water. a. Make the solution as neutral as possible, and free from 1& ammonium salts (it is a matter of indifference whether the mag- nesium and alkali metals are present as sulphates, chlorides, or nitrates), add baryta-water as long as a precipitate forms, heat to boiling, filter, and wash the precipitate with boiling water. The precipitate contains the magnesium as hydroxide. Dis- solve it in hydrochloric acid, precipitate the barium with sul- phuric acid, and then the magnesium as ammonium-magnesium phosphate ( 104, 2). The alkalies, which are contained in the solution, according to circumstances, as chlorides, nitrates, or caustic alkalies, are separated from the barium as directed in 10 or 12. LIEBIG, who was the first to employ this method, proposes crystallized barium sulphide as precipitant. The method is riot very exact, as magnesium is somewhat more soluble in solutions of alkali salts than in water. On this account the weighed alkali salt must always be tested for magnesium, and the latter determined if required. ft. Precipitate the solution with a little pure milk of lime, 1$ boil, filter, and wash. Separate the calcium and magnesium in the precipitate according to 36; the calcium and the alkalies -in the filtrate according to 10 or 14. This method may be em- ployed when magnesium has to be removed from a fluid con- taining calcium and alkalies, provided the alkalies alone are to be determined. Minute quantities of magnesium also in this case remain with the alkali salt from the cause mentioned in a. y. Evaporate the solution of the chlorides (which must IT contain no other acids) to dryness, and if ammonium chloride is present, ignite ; warm the residue with a little water (this will dissolve it with the exception of some magnesium oxide, which separates). Add mercuric oxide shaken up with water, evaporate to dryness on the water-bath with frequent stirring, dry thoroughly, ignite with increasing temperature till all the resulting mercuric chloride is volatilized, proceeding exactly as detailed in 104, 3, b. (Avoid inhaling the fumes.) There is *The methods a, a, and J3 are suitable for the separation of magnesium from lithium. 153.] BASKS OF GROUP II. 611 no need to continue the ignition until the whole of the mer- curic oxide is expelled ; on the contrary, part of it may be fil- tered off together with the magnesium oxide and subsequently volatilized upon the ignition of the latter. Treat the residue with small quantities of hot water, filter off rapidly, and wash the magnesium oxide with hot water, using small quantities at a time, and not continuing the operation unnecessarily. The solution contains the alkalies in form of chlorides. This method, proposed by BEKZELIUS, gives satisfactory results, and, so far as my experience goes, is the best of those given under a. Take care to add the mercuric oxide only in proper quantity, and always test the alkali chlorides for magnesium, a trace of which will generally be found. d. If the bases are present as chlorides, add as much pure 18 oxalic acid * as may be necessary to unite with all the bases present, viewed as potassium, to form a tetroxalate, add a little water, evaporate to dryness in a platinum dish, and ignite. By this operation the magnesium chloride is completely, the alkali chlorides partially, converted into oxalates, which on ignition yield alkali carbonates and magnesia. Treat the residue repeatedly with small quantities of boiling water ; it is immaterial whether the precipitate is transferred to the filter or remains in the platinum dish. "When all the alkali salt is washed, dry the filter, incinerate it in the dish, ignite strongly, and weigh the magnesia. Should the solution ob- tained be slightly turbid, evaporate it to dryness, take up the residue with water, and filter off the trifling residual magnesia; finally add hydrohcloric acid to the filtrate and esti- mate the alkalies as chlorides. If the bases are present as sulphates, add barium chloride 19 to the boiling solution until a precipitate just ceases to form, evaporate the filtrate with excess of oxalic acid, and proceed as in 18. Separate the slight quantity of barium carbonate remaining with the magnesia as directed in 29. These methods were devised by E. MITSCHERLICH and 20 described by LAScii.f According to my experience the * TH. SCHERER (Zeitschr. f. analyt. Chem., xi, 197) recommends pure ammonium oxalate instead of oxalic acid. \Journ.f. prakt. CJiem., LXIII, 343. 612 SEPARATION. [ 153. results are not particularly good. As a rule too little mag- nesia is obtained ; hence the weighed alkali salt should always be tested with sodium phosphate and ammonia for magnesia. Not infrequently a quite weighable precipitate is obtained, which must not be neglected.* The method described in 18 is also applicable to nitrates, and was recommended for these by DEVILLED During evaporation there are evolved carbon dioxide and nitrous acid. h. Methods based on the Precipitation of Magnesium ~by Ammonium Phosphate (or Ar senate}. Add ammonia in excess to the solution containing mag- 21 nesium, potassium, and sodium, and add some ammonium chloride should this not be already present ; then precipitate the magnesium with only a slight excess of pure ammonium phosphate. Expel the free ammonia from the filtrate by evaporation, and precipitate the phosphoric acid with lead acetate as a compound of lead phosphate and lead chloride. Remove the excess of lead oxide with ammonia and ammonium carbonate, or with hydrogen sulphide, from the still warm fluid, and in the filtrate determine the potassium and sodium according to .97 and 98 (O. L. EKDMANN J ; HEINTZ). The method is rather inconvenient, but quite accurate. If the solution contains much ammonium chloride, the greater part should be first removed by volatilization. The excess of phosphoric acid may also be removed with 22 ferric oxide or silver oxide instead of with lead oxide a. With ferric oxide. Expel any ammonia from the liquid with heat, neutralize if necessary with hydrochloric acid, and add ferric-chloride solution until the liquid is yellowish ; then add ammonium carbonate until the liquid is neutral or only acid from the carbonic acid present, boil, filter off the basic ferric phosphate (which, if sufficient ferric chloride has been used, will have a reddish-brown color), wash, evaporate * SONNENSCHEIN'S method (boiling the chlorides with silver carbonate) I cannot recommend, as the filtrate always contains magnesia, and in fact more than mere traces. \Journf.prakt. Chem., LX, 17. j lb., XXXTX, 278. Pogg. AnnaL, LXXIII, 119. 153.] BASES OF GROUP II. 613 the filtrate to dryness, expel the ammonium salt, and deter- mine the potassium and sodium according to 97, 98. A good and convenient method. (3. With silver oxide. Evaporate to dryness the fluid fil- tered off from the ammonium-magnesium phosphate, ignite cautiously, dissolve in water, and mix with silver nitrate and a slight excess of silver carbonate. After filtering, remove the excess of silver from the filtrate with hydrochloric acid, and evaporate the solution to dryness with an excess of hydro- chloric acid (CHANCEL *). The separation is somewhat shorter, if less precise and convenient, when the magnesium is precipitated with ammo- nium arsenate( 127, 2) instead of ammonium phosphate, and the liquid, with some ammonium chloride added, evaporated to dryness and the residue ignited under a good chimney. By this treatment the excess of arsenate added is volatilized, while the alkalies remain behind as chlorides (always retaining, however, a little magnesium chloride). C. v. HAUERf recommended a similar method. c. Method based on the Precipitation of the Magnesium as Ammonium- Magnesium Carbonate. Mix the solution of sulphates, nitrates, or chlorides (it 23 must be very concentrated) with an excess of a concentrated solution of sesquicarbonate of ammonia in water and ammonia (230 grm. of the salt, 360 c. c. solution of ammonia, sp. gr. 0*96, and water to one litre). After twenty-four hours filter off the precipitate (MgCO,*[NH 4 ] a CO 3 + 4H 2 O), wash it with the solution of ammonia and ammonium carbonate used for the precipitation, dry, ignite strongly and for a sufficient length of time, and weigh the magnesium oxide. Evaporate the filtrate to dryness, keeping the heat at first under 100, expel the ammonium salts, and determine the alkalies as chlorides or sulphates. When sodium alone is present the results are fairly satisfactory. In the presence of potassium the ignited magnesium oxide must be extracted with water before weigh- ing, as it contains an appreciable quantity of potassium carbo- * Compt. rend., L, 94. f Jahrb. der k. k. geolog. JKeicJisanstalt, iv, 863. 614 SEPARATION. [ 153. nate ; the washings are to be added to the principal filtrate. This last measure is unnecessary in the absence of potassium. The magnesium is always a little too low. Mean error 0'009 (F. G. SCHAFFGOTSCH,* PI. WEBEK f). d. Method based on the Precipitation of the Alkalies as jSilicofluorides (STOLE A ;). When a solution contains a mixture of potassium and mag- 24 nesium chlorides or potassium and magnesium nitrates, the potassium in one aliquot portion may be precipitated and de- termined as silicofi uoride ( 97, 5), while in another the magne- sium may be precipitated as ammonium-magnesium phosphate ( 104, 2). If it is desired to make both determinations in the same portion of fluid, remove the excess of silicofluoric acid from the fluid filtered off frorn the potassium silicofluoride, by an alcoholic solution of potassium acetate, wash the precipitate with a mixture of equal volumes of alcohol and water, and determine the magnesium in the filtrate. If sulphates are pres- ent, the method, in my opinion, is rendered so difficult as to make it untrustworthy, because of the difficult solubility of the magnesium sulphate in dilute alcohol. This method is less adapted for the separation of sodium from magnesium than for potassium from magnesium, because sodium silicofluoride is more soluble in alcohol than is the po- tassium salt. In the case of sulphates it is unserviceable, and in the case of chlorides or nitrates, to obtain results of any value, it is necessary to add 2 volumes of strong alcohol after adding the hydrosilicofluoric acid, and to allow the precipi- tate to settle completely before filtering. e. Indirect methods which give simultaneously the quan- tity of Potassium and Sodium, if both are present. a. Weigh the sulphates, dissolve, divide the solution into 25 two parts, and in one determine the magnesium according to 104, 2; in the other determine the potassium as in 2, cal- culate the magnesium sulphate and potassium sulphate, and from the difference find the sodium sulphate. * Pogg. Annal., civ, 482. f Vierteljahresschrift f. prakt. Pharm., vm, 161 IZeitschr.f. analyt. Chem., iv, 160. 154.] BASES OF GROUP II. 615 ft. With proper caution convert the bases into pure neutral 26 sulphates, weigh these, dissolve in water, determine the sul- phuric acid present with barium chloride ( 132), precipitate the excess of barium chloride in the filtrate with sulphuric acid, filter again, and in the filtrate, concentrated by evaporation, determine the magnesium as directed in 104, 2 (K. LIST*). Deduct the weight of the magnesium calculated as mag- nesium sulphate from the weight of the total sulphates; the difference will give the weight of the alkali sulphates. Further, deduct the weight of the sulphuric acid combined with the magnesium from the total sulphuric aci$ ; the difference will give that combined with the alkalies. See 152, 3 (6). It is, of course, evident that the indirect methods can give accurate results only when the most painstaking care is exer- cised. The accuracy of method /? is, besides, impaired by the tendency of barium sulphate to carry down with it readily soluble salts. II. SEPARATION OF THE BASIC RADICALS OF THE SECONP GROUP FROM EACH OTHER. 154. INDEX. (The numbers refer to those in the margin.) Barium from strontium, 28, 31, 40. calcium, 28, 30, 31, 35, 40. " magnesium, 27, 29. Strontium from barium, 28, 31, 40. calcium, 34, 38, 39. " magnesium, 27, 29. Calcium from barium, 28, 30, 31, 35, 40. strontium, 34, 38, 39. magnesium, 27, 32, 33, 36, 37. \ Magnesium from barium, 27, 29. " strontium, 27, 29. calcium, 27, 32, 33, 36, 37. f * Annal. d. Chem. u. Pharm., LXXXI, 117. f Compare also the method of OEFFINGEB, Zeitschr. f. analyt. Chem., vm, 456. 616 SEPARATION. [ 154. A. General Method. THE WHOLE OF THE ALKALI-EARTH METALS FROM EACH OTHER. Proceed as in 10. The magnesium is precipitated from the 27 filtrate as ammonium-magnesium phosphate (see foot-note, p. 620). The precipitated carbonates of barium, strontium, and calcium are dissolved in hydrochloric acid, and the bases sepa- rated as directed in 28. The traces of magnesium, which may be present in the ammonium carbonate precipitate, are obtained by evaporating the nitrate from the strontium or calcium sul- phate to dryness, taking up the residue with water, and precipi- tating the solution with sodium phosphate and ammonia. B. Special Methods. 1. Methods based upon the Insolubility of Barium Silicofluoride. BARIUM FROM STRONTIUM AND FROM CALCIUM. Mix the neutral or slightly acid solution with hyclrofluosi- 28 licic acid * in excess, add one third of the volume of alcohol of 0-81 sp. gr., let the mixture stand twelve hours, collect the pre- cipitate of 'barium silicofluoride on a weighed filter, wash with a mixture of equal parts of water and alcohol until the wash- ings cease to show even the least trace of acid reaction (but no longer), and dry at 100. Precipitate the strontium or calcium from the filtrate by dilute sulphuric acid ( 102, 1, a, and 103, 1). The results are satisfactory. For the properties of barium silicofluoride, see 71. If both strontium and calcium are pres- ent, the sulphates are weighed, and then separated according to 34, or they are converted into carbonates ( 132, II., J), and separated according to 38 or 39. * If not kept in a gutta-percha bottle it should be freshly prepared, 154.] BASES OF GROUP II. 617 2. Methods based upon the Insolubility of Barium Sulphate or Strontium Sulphate, as the case may be, in Water and in Solution of Sodium T hiosulphate. a. BARIUM AND STRONTIUM FROM MAGNESIUM. Precipitate the barium and strontium witli sulphuric acid 29 ( 101, 1, a and 102, 1, a\ and the magnesium from the fil- trate with ammonia and sodium-ammonium phosphate ( 101, 2). b. BARIUM FROM CALCIUM. Mix the solution with hydrochloric acid, then with highly 30 dilute sulphuric acid (1 part acid to 300 water), as long as a pre- cipitate forms ; allow to deposit, and determine the barium sul- phate as directed 101, 1, a. Concentrate the washings by evaporation and add them to the filtrate, neutralize the acid with ammonia, and precipitate the calcium as oxalate ( 103, 2, b, a). The method is principally to be recommended when small quantities of barium have to be separated, from much cal- cium. If we have to separate calcium sulphate from barium sulphate, the salts may (in the absence of free acids) be treated repeatedly with a solution of sodium thiosulphate at a gentle heat. The barium sulphate remains undissolved, the calcium sulphate dissolves. The calcium is precipitated from the fil- trate by ammonium oxalate (DiEHL*). 3. Method based upon the different deportment with Alkali Carbonates of Barium Sulphate on the one hand, and Strontium and Calcium Sulphates on the other. BARIUM FROM STRONTIUM AND CALCIUM. Digest the three precipitated sulphates for twelve hours at 31 the common temperature (15 20), with frequent stirring, with a solution of ammonium carbonate, decant the fluid on to a filter, treat the residue repeatedly in the same way, wash finally with water, and in the still moist precipitate, separate the undecomposed barium sulphate by means of cold dilute hydrochloric acid from the strontium and calcium carbonates formed. To hasten the separation you may boil the sulphates for some time with a solution of potassium (not sodium) car- bonate, to which \ the amount of the carbonate, or more, of *Jouru.f. prakt. Chem., LXXIX, 430. 618 SEPARATION. [ 154. potassium sulphate has been added. By this process, also, the strontium and calcium sulphates are decomposed, the barium sulphate remaining unacted on. If the basic metals are in solu- tion, the above solution of potassium carbonate and sulphate is added in excess at once, and the whole boiled. The precipitate, consisting of barium sulphate and strontium and calcium car- bonates, is to be treated as above with cold hydrochloric acid (H. ROSE*). 4. Methods based on the Insolubility of Calcium Sul- phate in Alcohol. CALCIUM FROM MAGNESIUM. a. Remove water and free hydrochloric from a solution of 32 the chlorides by evaporation, dissolve the residue in strong (but not absolute) alcohol, add a slight excess of pure strong sulphu- ric acid, digest in the cold, allow to stand for some hours, trans- fer the precipitate consisting of calcium sulphate and some magnesium sulphate to a filter, wash away the acid thoroughly with nearly absolute alcohol, and then, but only after the acid has been completely removed, continue the washing with alcohol, sp. gr. 0-96 0-95, till a few drops of the washings give no residue on evaporation. Weigh the calcium sulphate accord- ing to 103, 1. Evaporate the alcohol from the filtrate, and determine the magnesium according to 104, 2. The method is in itself not new, but A. CmzYNSKijf adopting the precautions here given, has obtained excellent results, even in the presence of phosphoric acid. b. SMALL QUANTITIES OF CALCIUM FROM MUCH MAGNESIUM. 33 Convert into neutral sulphates, dissolve the mass in water, and add alcohol, with constant stirring, till a slight permanent tur- bidity is produced, Wait a few hours and then filter, wash the precipitated calcium sulphate with alcohol which has been diluted with an equal volume of water, and determine it after 103, 1, a (in which case the weighed sulphate must be tested for magnesium), or dissolve the precipitate in water containing hydrochloric acid and separate the calcium from the small quan- tity of magnesium possibly coprecipitated according to 36 (SCHEERER^). *Pogg. Ann., xcv, 286, 299, 427. f Zeilschr.f. analyt. Chem., iv, 348. %Annal. d. Chem u. Pharm., ex, 237. 154.] BASES OF GROUP II. 619 5. Methods 'based on the Insolubility of Strontium, and Barium Sulphates in solution of Ammonium Sulphate. STRONTIUM FROM CALCIUM. If the mixture is soluble, dissolve in the smallest quantity 34 of water, add about 50 times the quantity of the substance of ammonium sulphate dissolved in four times its weight of water, and either boil for some time with renewal of the water that evaporates and, addition of. a very little ammonia (as the solu- tion of ammonium sulphate becomes acid on boiling), or allow to stand at the ordinary temperature for twelve hours. Filter and wash the precipitate, which consists of strontium sulphate and a little ammonium strontium sulphate, with a concentrated solu- tion of ammonium sulphate, till the washings remain clear on addition of ammonium oxalate. The precipitate is cautiously ignited, moistened with a little dilute sulphuric acid (to convert the small quantity of strontium sulphide into sulphate), reig- nited and weighed. The highly dilute filtrate is precipitated with ammonium oxalate, and the calcium determined according to 103, 2, &, a. If you have the solid sulphates to analyze, they are very finely powdered and boiled with concentrated solu- tion of ammonium sulphate with renewal of the evaporated water and addition of a little ammonia. Results very close, e.g., 1 -048 Sr(NO 3 ) 2 instead of 1-053, and -497 CaCO 3 , instead of 0-504: (H. ROSE*). . BARIUM may be separated FROM CALCIUM in the same way. 35 6. Methods based upon the Insolubility of Calcium Oxalate in Ammonium Chloride and in Acetic Acid. CALCIUM FROM MAGNESIUM. a. Mix the properly diluted solution with sufficient ammo- 36 ninm chloride to prevent the formation of a precipitate by ammonia, which is added in slight excess ; add ammonium oxa- late as long as a precipitate forms, then a further portion of the same reagent, about sufficient to convert the magnesium also into oxalate (which remains in solution). This excess is abso- lutely indispensable to insure complete precipitation of the cal- cium, as calcium oxalate is slightly soluble in magnesium chlo- ride not mixed with ammonium oxalate (Expt. No. 83). Let * Pogg. Annal., ex, 296. 620 SEPARATION. [ 154 the mixture stand twelve hours, decant the supernatant clear fluid, as far as practicable, from the precipitated calcium oxa- late, mixed with a little magnesium oxalate, on to a filter, wash the precipitate once in the same way by decantation, then dis- solve in hydrochloric acid, add water, then ammonia in slight excess, and a little ammonium oxalate. Let the fluid stand until the precipitate has completely subsided, then pour on to the previous filter, transfer the precipitate finally to the latter, and proceed exactly as directed 103, 2, 5, a. The first filtrate contains by far the larger portion of the magnesium, the second the remainder. Evaporate the second filtrate, acidified with hydrochloric acid, to a small volume, then mix the two fluids, and precipitate the magnesium with sodium ammonium phos- phate (HNaNH 4 )PO 4 ,*as directed 104, 2. If the quantity of ammonium salts present is considerable, the estimation of the magnesium is rendered more accurate by evaporating the fluids in a large platinum or porcelain dish to dryness, and igniting the residuary saline mass, in small portions at a time, in a smaller platinum dish, until the ammonium salts are expelled. The residue is then treated with hydrochloric acid and water, warmed, allowed to cool, and rendered just alkaline w r ith ammo- nia. If enough ammonium chloride is present, no magnesium hydroxide will fall down, but occasionally small flocks of silica or alumina are to be seen. Filter them off and finally precipi- tate with ammonia and (IEN"aNH 4 )PO 4 . If the precipitate pro- duced by ammonia is at all considerable, dissolve it in hydro- chloric acid, evaporate the solution on a water-bath to dryness, treat the residue with hydrochloric acid and water, render alka- line with ammonia, filter, and add the filtrate to the principal solution. Numerous experiments have convinced me that this method, which is so frequently employed, gives accurate results only if the foregoing instructions are strictly complied with. It is only in cases where the quantity of magnesium present is relatively small that a single precipitation with ammonium oxalate may be found sufficient (conip. Expt. No. 84 f). * This is preferable to sodium phosphate as a precipitant. See Zeitschr.f. analyt. Chem., xn, 36 f Further experiments will be found in Zeitschr. f. analyt. Chem., vn, 310. 154.] BASES OF GROUP II. 621 b. In the case of calcium and magnesium phosphates, dis- 37 solve in the least possible quantity of hydrochloric acid, add ammonia until a copious precipitate forms ; redissolve this by addition of acetic acid, and precipitate the calcium with an excess of ammonium oxalate. To determine the magnesium, precipitate the filtrate with ammonia and (H!NaNH 4 )PO 4 . As free acetic acid by no means prevents the precipitation of small quantities of magnesium oxalate, the precipitate contains some magnesium, and as calcium oxalate is not quite insoluble in acetic acid, the filtrate contains some calcium ; these two sources of error compensate each other in some measure. In accurate analysis, however, these trifling admixtures of magnesium and calcium are afterwards separated from the weighed precipi- tates of calcium carbonate or oxide and magnesium pyrophos- phate respectively. Y. Method based upon the Insolubility of Strontium Nitrate in Alcohol and Ether. STRONTIUM FROM CALCIUM (after STROMEYER). Digest the perfectly dry nitrates in a closed flask with abso- 38 lute alcohol, to which an equal volume of ether should be added (H. HOSE). Filter off the undissolved strontium nitrate in a covered funnel, wash with the mixture of alcohol and ether, dis- solve in water, and determine as strontium sulphate ( 102, 1). Precipitate the calcium from the filtrate by sulphuric acid. The results are satisfactory. 8. Indirect Method. STRONTIUM FROM CALCIUM. Determine both bases first as carbonates or oxides, precipi- 39 tating them either with ammonium carbonate or oxalate ( 102, 103) ; then estimate the amount of carbonic acid in them, and calculate the amount of strontium and calcium as directed in " Calculation of Analyses " ( 200). The determination of the carbonic acid may be effected by fusion with vitrified borax ( 139, II., .i carbonate, 61 add sodium acetate in excess, pass chlorine or add bromine and warm. The chromium will readily be converted into chromic acid, especially if sodium carbonate is added every now and then to keep the fluid nearly neutral. As soon as t\ ds is effected proceed according to 130, IL, ased upon, the Solubility of Aluminium Hydroxide in Caustic Alkalies. a. ALUMINIUM FROM FERROUS AND FERRIC IRON, AND SMALL QUANTITIES OF MANGANESE (but not from nickel and cobalt). Mix the hydrochloric solution with sodium carbonate or 7& pure potassa till the greater part of the free acid is neutralized and pour the solution gradually into excess of pure potassa * Tartaric acid often contains aluminium, therefore this is best made from the acid lartrate. f Chromium and zinc cannot be obtained together in alkaline solution (CHANCEL, Compt. rend., XLIII, 927; Journ. f. prakt. Ohem., LXX, 378). 160.] BASES OF GROUP IV. 643 heated nearly to boiling in a platinum or silver dish, stirring all the while. Porcelain does not answer so well, and glass should on no account be used. The iron, if present as ferric chloride, separates as ferric hydroxide, while the aluminium remains in solution as alkali aluminate. Hydrated protosesquioxide of iron is more easy to wash than ferric hydroxide, hence when much iron is present it is better to reduce a part by cautiously adding sodium sulphite and heating, so that when the solution is added to the boiling potash a black granular precipitate may be formed. The iron precipitate is sure to contain alkali, and must be dissolved in hydrochloric acid, the solution boiled with nitric acid if necessary, and reprecipitated with ammonia. To the alkaline filtrate add a few drops of hydrochloric acid. If the potassa was present in sufficient excess the precipi- tate will redissolve readily on stirring. Continue adding hydro- chloric acid till in excess, boil with a little potassium chlorate (to destroy traces of organic matter), concentrate by evapora- tion, and throw down the aluminium according to 105, a. The above is the best method of procedure, but it is always to be feared that small quantities of aluminium will be retained by the iron precipitate. It. ALUMINIUM FROM FERROUS AND FERRIC IRON, COBALT, AND NlCKEL. Fuse the oxides with potassium hydroxide in a silver era- 79 cible, boil the mass with water, and filter the alkaline fluid, which contains the aluminium, from the oxides, which are free from aluminium, but contain potassa (II. ROSE). 2. Methods based on the different behavior of Am- monia or Ammonium Carbonate in the presence of Chlo- r>. MANGANESE FROM NICKEL AND ZINC. The solution should be slightly acid and contain ammonium 81 chloride. Precipitate the manganese as white carbonate with ammonium carbonate, allow to settle in a warm place, filter through a thick paper (double, if necessary), wash with hot water, dry the precipitate, and convert it into protosesquioxide by ignition with access of air. This excellent method was proposed by TAMM,f and has given me good results. J It is not adapted to the separation of cobalt from manganese, as the former is partly precipitated with the latter. 3. Method based upon the different deportment of neutralized Solutions at boiling heat. a. FERRIC IRON FROM MANGANESE, NICKEL, AND COBALT, AND OTHER STRONG BASIC METALS, AFTER IlERSCHEL, SCHWARZENBERG, | AND MY OWN EXPERIMENTS. Mix the dilute solution largely with ammonium chloride 82 (at least 40 of ]S T II 4 C1 to 1 of MnO, NiO, etc.), add ammo- nium carbonate in small quantities, at last drop by drop and in very dilute solution, so long as the precipitated iron redis- Bolves, which takes place promptly at first, but more slowly towards the end. As soon as the fiuid has lost its transpar- ency, without showing, however, the least trace of a distinct precipitate in it, and fails to recover its clearness after stand- * This method, which was recommended in the previous German edition for the separation of small quantities of iron from nickel and cobalt, has been found by BAUMHAUEB, (Zeitschr. /. analyt. Chem., x, 218) to be well adapted for the separation of large quantities as well. f Chem. News, xxvr, 37. ^.Zeitschr. f. analyt. Chem., xi, 425. %Annal. de Chim. et de Phys., XLIX, 306. | Annal. de Chem. u. Pharm., xcvii, 216. 160.] BASES OF GROUP IY. 645 ing some time in the cold, but, on the contrary, becomes rather more turbid than otherwise, the reaction may be considered completed. When this point has been attained, heat slowly to boiling, and keep in ebullition for a short time after the carbonic acid has been entirely expelled. The iron separates as a basic ferric salt, which rapidly settles if the solution was not too concentrated. Pour off the hot fluid through a filter and wash by decantation com- bined with filtration with boiling water containing a little ammonium chloride. It is well to redissolve the precipitate in hydrochloric acid and throw down the iron with ammonia. The first filtrate should be mixed with excess af ammonia. If a small portion of ferric hydroxide is thrown down here, filter it off, dissolve in hydrochloric acid, precipitate with ammonia and thus free the small quantity of iron entirely from the strong basic metals ; if, on the other hand, a large quantity of iron is thrown down, this is a sign that the operation has been con- ducted improperly, and the hydrochloric solution of the pre- cipitate must be reprecipitated as above. The fluid should not contain more than 3 or 4 grm. of iron in the litre, and should be tolerably free from sulphuric acid, as when this is present it is impossible to hit the exact point of saturation. 5. FERROUS IKON FROM FEBEIC IRON. In compounds difficultly soluble in hydrochloric acid, but 83 which are decomposed below 326 by moderately concentrated sulphuric acid,* ScHEERERf separates ferric from ferrous iron by effecting solution in an atmosphere of carbonic acid (which is maintained during the entire operation), diluting the solution by adding pieces of air-free ice, adding ammonium carbonate until the acid has been nearly neutralized, then add- ing finely triturated magnesite (not the artificial magnesium carbonate), and boiling for 10 to 15 minutes. All the ferric iron is precipitated by this process. The washing is carried out as in 76, with water which, after admixture of some ammonium sulphate, has been boiled and allowed to cool with * On boiling ferrous sulphate is oxidized, the sulphuric acid being reduced to sulphurous, v. KOBELL, Annal. d. Chem. u. Pharm., xc, 244. \Pogg. Annal., LXXXVI, 91, and xcm, 448. 646 SEPARATION. 160. exclusion of air. v. KOBELL* uses as a solvent a mixture of 1 volume concentrated sulphuric acid, 2 volumes water, and 1 volume strong hydrochloric acid. Solution may generally be effected by heating with hydrochloric acid, or a mixture of 4: parts concentrated sulphuric acid and 1 part of water, in sealed tubes heated to 210 (A. MITSCHERLICH f). Silicates anay be easily dissolved by hydrofluoric and hydrochloric acids, or by hydrofluoric acid and diluted sulphuric acid. To ex- clude the air, invert a plaster-of-Paris cylinder (which may be readily made in the laboratory) provided with a cover over the silicate with its solvent contained in a platinum dish. Through a hole in the cover conduct a moderate current of carbonic acid, taking care that the cylinder is completely filled with the gas before applying heat. Similar methods and appa- ratus have been described by WERTHER,:): J. P. COOKE, and WILBUR and WHITTLESEY. J Care must be taken to see that the hydrofluoric acid is free from hydrogen sulphide and sul- phurous acid. c. FERRIC IRON FROM ALUMINIUM. To the dilute solution, which may contain aluminium and 84 iron chlorides or even sulphates, add sodium carbonate if necessary, to neutralize any too large quantity of free acid that may be present; then add solution of sodium thiosulphate until all ferric iron has been reduced to a ferrous state ; now add a further quantity of thiosulphate and boil continuously until every trace of sulphurous-acid odor has disappeared. The precipitation of the alumina takes place according to the following equation : Al a (SO 4 ) 3 + 3Na a S a O, + 3H 2 O = Al a (OH) 6 + 3JSTa a SO 4 + 3SO a + 3S. Filter off, wash the precipitate well, and incinerate; this will give the aluminium. Heat the filtrate with hydrochloric acid to decompose the ex- cess of sodium thiosulphate, filter off the precipitated sulphur, and in the filtrate estimate the iron (CHANCEL T). * Annal. d. Chem. u. Pharm., xc, 244. f Zeitschr. f. analyt. Chem., i, 54. J Journ.f. prakt. Chem., xci, 329. % Zeitschr. f. analyt. Chem., vn, 99. | 2b., x, 98. 1 Compt. Tend., XLVI, 987. Compare also WERTHER, Journ.f. prakt. Chem., XCI, 329, and GIBBS, ZeitscJir. f. analyt. Chem., m, 391. I have felt it necea- 160.] BASES OF GROUP IV. Gl7 i 4. Method based on the behavior of the Acetates at a boiling heat. FERRIC IRON AND ALUMINIUM FROM MANGANESE, ZINC, COBALT, NICKEL, AND FERROUS IRON. The metals should be present in the form of chlorides. The 85 solution should be in a flask. If much free acid is present first nearly neutralize with sodium or ammonium carbonate ; the solution should remain clear, but if there is much ferric chloride it should be of a deep red color. Add a concentrated solution of neutral sodium or ammonium acetate, not in large excess, and boil for a short time long-continued boiling would make the precipitate slimy. "When the lamp is removed the precipitate should settle rapidly, leaving'the supernatant fluid clear. Wash the precipitate immediately by decantation and filtration with boiling water containing a little sodium or ammonium acetate. In very particular analyses it would be well after washing the precipitate a little to redissolve it in hydrochloric acid and reprecipitate. In separating ferric from ferrous iron REICHARDT* recom- mends a slight addition of ammonium chloride or of sodium chloride to prevent oxidation of the ferrous salt. The precipitate of basic ferric acetate or basic aluminium acetate is best dissolved in hydrochloric acid, in order to precipi- tate the basic metals from this solution again by ammonia. This method is more suitable to the separation of ferric iron or ferric iron and aluminium from the strong basic metals than to the separation of aluminium alone. It is a good method and is very generally used. Instead of the alkali acetates, the formates may also be used with excellent results ( 81, y). [The results obtained by this method depend greatly on the proper adjustment of free acetic acid to the volume of the solu- tion which is boiled. The solution at this point may contain sary to give this frequently recommended method, but I would add that I have not found it to be perfectly trustworthy. * Zeitschr. /. analyt. Chem., v, 64. 648 SEPAKATION. [ 160 about four per cent, (by volume) of acetic acid sp. gr. 1-044 (JEWETT*). If too little acetic is present, zinc, manganese, nickel, and cobalt are precipitated in notable quantity along with the iron. If too much is present the precipitation of iron is incomplete. The operator may control the amount of acid within narrow limits by proceeding as follows : Add the alkali carbonate to the cold and preferably concentrated solution until a slight precipitate forms which no longer redissolves in four or five minutes with occasional shaking, but imparts a turbidity to the deep red solution ; HC1 is then added without further delay, slowly, drop by drop, until the fluid, though still dark, becomes clear. Next the' amount of acetic acid required to form four per cent, of the final volume is added, then sodium acetate (about ten times as much of the crystallized salt as there is iron present, or more if but little iron is present). Dilute now to the final volume, which should amount to at least 100 c.c, per O'l grm. iron and heat to boiling. After boiling two or three minutes only, allow the iron precipitate to settle. Pour the clear liquid through a filter, then bring the precipitate upon the filter at once and wash as above directed. The iron pre- cipitate contains no zinc and but an inappreciable trace of man- ganese. Small quantities of cobalt and still more nickel will,, however, be precipitated with the iron. When these two metals are present in considerable quantity a repetition of the process, is indispensable when accuracy is required. Coprecipitation of nickel is lessened but not entirely prevented by presence of ammonium chloride.f In carrying out the process according to this plan great care must be taken in the preliminary neutralization with alkali carbonate to leave as little free mineral acid as possible without formation of a permanent precipitate, otherwise this free acid will liberate enough acetic acid from the soaium acetate to prevent (with that intentionally added) the precipitation of iron in a form easy to wash. In separating large quantities of iron from small quantities of manganese the addition of 2 or 3 per cent, of acetic acid will secure a separation satisfactory enough for most purposes (e.g. in iron and iron ores), and the danger that the acetic acid present * Amer. Chem. Journ., i, 251. -j- Loc. cit. 160.] BASES OF GKOUP IV. 649 may accidentally exceed the proper limit will of course be lessened.] 5. Method based on the different 'behavior of the Suo- cinates. FERRIC IRON (AND ALUMINIUM) FROM ZINC, MANGANESE, NICKEL, AND COBAT. The solution should contain no considerable quantity of sul- 86 phuric acid. If acid, as is usually the case, add ammonia till the color is reddish brown, then sodium or ammonium acetate (H. ROSE) till the color is deep red, finally precipitate with neutral alkali succinate at a gentle heat, and when cool filter the ferric succinate from the solution which contains the rest of the metals. Wash the precipitate first with cold water, then with warm ammonia, which removes the greater part of the acid, leaving it darker in color. Dry and ignite, moisten with a little nitric acid, and ignite again. With proper care the sepa- ration is complete, and especially to be recommended when a relatively large quantity of iron is present! The method may also be used in the presence of aluminium. The latter falls down completely with the iron (E. MITSCHEKLICH, PAGELS*). 6. Methods based upon ih# different deportment of the several Sulphides with Acids, or of Acid Solutions with Hydrogen Sulphide. a. ZINC FROM ALUMINIUM AND MANGANESE. The solution of the acetates, which must be free from in- 87 organic acids, and must contain a sufficient excess of acetic acid, is precipitated with hydrogen sulphide, which throws down the zinc only ( .108, I). The metals are usually most readily obtained in the form of acetates, by converting them into sulphates, and adding a sufficient quantity of barium acetate. Hydrogen sulphide is then conducted, without application of heat, into the unfiltered fiuid, to which, if necessary, some more acetic acid has been added. The precipitate, which consists of a mixture of zinc sulphide and barium sulphate, is washed with water containing hydrogen sulphide. It is then heated with dilute hydrochloric acid, the solution filtered, and the zinc * Jahresbericht von KOPF imd WILL, 1858, 617. 650 SEPARATION. [' 160. in the filtrate determined as directed 108, a. The other metals acre determined in the fluid filtered from the zinc sul- phide, after removal of the barium by precipitation. BKUNNER has proposed a modification of this process, especially for the separation of zinc from nickel. b. ZINC FROM NICKEL, COBALT, AND MANGANESE. To the hydrochloric solution add sodium carbonate till a 88 permanent precipitate just forms, and then a drop or two of hydrochloric acid to redissolve the precipitate. Now pass hydrogen sulphide till the precipitate of zinc sulphide ceases to increase. Add a few drops of a very dilute solution of sodium acetate, and continue passing the gas for some time. When all the zinc is precipitated, allow to stand for twelve hours, filter, wash with hydrogen sulphide water, and determine the nickel and cobalt in the filtrate (SMITH and BRUNNER*) A good method ; compare KLAYE and DEUS.')' The method is also adapted for separating zinc from manganese. [Precautions. Bear in mind that Zn can be precipitated from solutions containing free HC1, but only in case the amount of the latter is very small.;): When ZnS is precipi- tated the amount of HC1 set free may be sufficient to prevent complete precipitation of the Zn. Addition of sodium acetate converts this HC1 into NaCl, and allows the formation of ZnS to continue. Care must be taken not to add enough sodium acetate to convert all the HC1 into Nad, for in that case JSHS and CoS will be precipitated.] [Zinc can be precipitated by hydrogen sulphide from a cold solution containing a sufficient amount of free acetic acid to prevent precipitation of nickel and cobalt. To effect separation by this means add sodium or potassium carbonate to the solu- tion till it is slightly alkaline. If a large quantity of any free volatile acid is present it may be previously remove'd by evaporation. Dissolve the precipitate produced by the alkali carbonate (without filtering) in acetic acid, and add a large quantity more of acetic acid. Precipitate the zinc by passing * Dingler's polyt. Journ., CL, 369 ; Ghem. Centralbl., 1859, 26. ^Zeitschr.f. analyt. Chem., x, 200. j STOKER and ELIOT, Mem. Am. Acad. Arts and Sciences, vin, 95. ROSE and HINKENER, Anal. (Jhem. t u, 129 and 143. 160.] BASES OF GKOUP IV. 651 IT 2 S through the cold moderately diluted solution. "Wash the sulphide of zinc with water to which hydrogen sulphide and a little ammonium acetate has heen added. The zinc sulphide should not be dark-colored. This will only be the case when not enough acetic is present to prevent precipitation of nickel or cobalt. Cobalt and nickel may be best separated from the filtrate by evaporating till the greater part of the acetic acid is removed, then adding some ammonium chloride and ammonia to slight alkaline reaction, evaporating further till the reaction becomes acid, heating finally to boiling, and passing hydrogen sulphide through the solution, as directed in 110, 1, b, ft. A good method.] c, COBALT AND NICKEL FROM MANGANESE AND IKON. Add ammonia to the nitric-acid free solution to neutralize 89 any other free acid present, precipitate with ammonium sul- phide, then add very dilute hydrochloric acid, and saturate with hydrogen sulphide while the liquid is frequently stirred. By this treatment the manganese and iron sulphides are dissolved, while the cobalt and nickel sulphides remain undissolved, the latter, it is true, less completely. On filtering, precipita- ting the filtrate with ammonia and ammonium sulphide and treating the precipitated sulphides also as above, the results are quite accurate. Caution requires, however, that the weighed cobalt and nickel compounds be tested for manga- nese, and also more particularly for iron. d. COBALT AND NICKEL FROM MANGANESE. Add first an excess of sodium carbonate to the acid solu- 90 tion, then considerable of an excess of acetic acid, then add to the clear fluid (containing, say, 1 grm. of nickel or cobalt) 30 to 50 c. c. of a sodium-acetate solution (1 : 10), warm the fluid to 70, and saturate it with hydrogen sulphide. After the precipitation is complete, filter off the precipitated nickel and cobalt sulphides, wash, and dry. Concentrate the filtrate by evaporation, and add hydrogen and ammonium sulphide and an excess of acetic acid, thus obtaining a further precipitate of nickel and cobalt sulphides. For the sake of caution test the filtrate once more in a similar manner. In the combined pre- cipitates determine the nickel or cobalt as in 110, 1, J, <*, 652 SEPARATION. [160. and 111, 1, c\ in the filtrate determine the manganese as in 109, 2. T. Methods based upon the deportment of the several oxides with Hydrogen Gas at a red heat. a. FERRIC IRON FROM ALUMINIUM AND CHROMIUM. RIVOT'S method.* Precipitate with ammonia, heat, filter, 91 ignite, and weigh. Triturate, and weigh a portion in a small porcelain boat, which insert in a horizontal porcelain tube, through which is passed from one end a current of hydrogen dried by sulphuric acid and calcium chloride. Close the other end of the tube with a stopper bearing a narrow open glass tube. After all the air has been expelled from the apparatus, gradu- ally heat the porcelain tube to redness and maintain it at this heat so long as water still forms (about one hour) . Let the tube cool while maintaining the current of hydrogen, then re- move the boat and weigh it. The loss of weight gives the oxygen which was combined to form ferric iron. To deter- mine the oxides separately, which may be deemed necessary when but little ferric iron is present, treat the mixture of aluminium, chromium, and metallic iron with a mixture of 1 part nitric acid and 30 to 40 parts of water (or with water to which a very little nitric acid is added from time to time). The iron dissolves, while the aluminium and chromium re- main behind. Weigh these direct ; precipitate the iron with ammonia after boiling the solution. The test analyses given by RIVOT were very satisfactory. The method is particularly to be recommended when much aluminium, etc., and but little iron are present. ~b. FERRIC IRON FROM ALUMINIUM. After reduction by hydrogen (as in &), DEVILLE conducts 92 a current of, first, hydrochloric-acid gas, then again hydrogen, through the tube. The aluminium remains behind, while the iron volatilizes as ferrous chloride, and is determined either from the loss in weight or direct. In the latter case, dissolve all the ferrous chloride in the tubes and tubulated receiver by heating with diluted hydrochloric acid to boiling, and conduct . * Annal. de Ghim. et dePhys., xxx, 188; Journ. f. prakt. Chem., LI, 338. 160.] BASES OF GROUP IV. 653 the vapors into the porcelain tube. The tubulure of the re- ceiver is directed downwards during the operation. By the use of a platinum tube the operation is greatly facilitated (COOKE *). 8. Methods based upon the different capacity of the several Oxides to be converted by Oxidizing Agents into higher Oxides, or by Chlorine into higher Chlo- rides. a. CHROMIUM FROM ALL THE METALS OF THE FOURTH GROUP, AND FROM ALUMINIUM. ex. Fuse the oxides with potassium nitrate and sodium car- 93 bonate (comp. 59), boil the mass with water, add a small quantity of alcohol, and heat gently for several hours. Filter and determine in the filtrate the chromium as directed in 130, and in the residue the metals of the fourth group* The following is the theory of this process : The oxides of zinc, cobalt, nickel, iron, and partly that of manganese, sepa- rate upon the fusion, whilst, on the other hand, potassium manganate (perhaps also some ferrate) and chromate are formed. Upon boiling with water, part of the manganic acid of the potassium manganate is converted into permanganic acid at the expense of the oxygen of another part, which is reduced to the state of bin oxide ; the latter separates, whilst the potas- sium salts are dissolved. The addition of alcohol, with the application of a gentle heat, effects the decomposition of the potassium manganate and permanganate, manganese dioxide being separated. Upon filtering the mixture, we have there- fore now the whole of the chromium in the filtrate as alkali chromate, and all the oxides of. the fourth group on the filter. Aluminium, if present, will be found partly in the residue, partly as alkali aluminate in the filtrate; proceed with the latter according to 59. If you have to deal with the native compound of sesqui- oxide of chromium with ferrous oxide (chromic iron) the above method does not answer. In this case proceed according to one of the methods detailed in the Special Part. The * Zeitschr. /. aiuilyt. Chem., vi, 226. 654 SEPARATION. [ 160. substance may also be decomposed by fusion witli cryolite and potassium disulphate. . The radicals to be separated may be in the form of a 94 solution of their salts; nearly neutralize the solution, add sodium acetate, heat, and convert the chromium into chromic acid by passing chlorine. Compare 61. If ferric iron and aluminium are present, they will separate during boiling by the action of the sodium acetate, while the chromic acid and any zinc will remain in solution. If manganese, nickel, and cobalt are present, the method loses its simplicity ; the manga- nese is precipitated as hydrated peroxide with a portion of the cobalt, almost the whole of the nickel and some zinc, while the chromic acid remains in solution with the principal amount of the zinc and the rest of the cobalt and nickel (W. GIBBS*). 5. COBALT FROM NICKEL. OL. After II. RosE.f Dilute the hydrochloric-acid solu- 95 tion, contained in a capacious flask, with water so that a litre of the solution will contain about 2 grm. of the metal, conduct chlorine gas into the fluid until the latter is saturated and the upper part of the flask is entirely filled with the gas, add an excess of calcium- or barium carbonate shaken up with water, shake frequently during 5 or 6 hours in the cold, and filter off the precipitated cobalt hydroxide from the liquid containing the nickel in solution. Instead of chlorine, HENRY uses bromine. DENHAM SMITH recommends adding a dilute solution of chlorinated lime which has been completely decomposed with sulphuric acid, so as to leave no hypochlorite present. According to FR. GATJHE,;}: HOSE'S method is unsafe, as an insufficiently prolonged action of the carbonates of the alkaline earths precipitates the cobalt incompletely, while a too pro- longed action precipitates nickel as well. The method may be serviceable by observing special conditions, and when applied with great experience ; it is not suitable for accurate analyses. ft. The method of GIBBS, elaborated by H. ROSE (boiling the sulphuric -acid solution with lead peroxide), gives only approximate results also. Compare GAUHE (loo. cit.). * GIBBS and CLARK, Amer. Jour. Sci., 2d ser., XLVIII, 198. \Pogfj. Annul , LXXI, 545, and Handb. d. analyt. Chem., 6. Aufl., u, 143. %Zeitschr.f. analyt. Chem., v, 84. %Pogg. Annal., ex, 413. 160.] BASES OF GROUP IV. 655 9. Method based upon the different deportment vf the Nitrates. COBALT FROM NICKEL, ALSO FROM MANGANESE AND ZINC. The separation of cobalt as tripotassium cobaltic nitrite was 96 recommended first by FISCHER,* afterwards by A. STROMEYER^ GENTH and GIBBS, J IL ROSE, FR. GAUHE,|| and myself (com- pare last edition of this work). The results are quite satisfac- tory both in presence of much cobalt and little nickel, and in the presence of little cobalt and much nickel; but the process is peculiarly good for the latter case. However, it is absolutely necessary that barium, strontium, and calcium should be absent, as in their presence nickel is thrown down as triple nitrite of nickel, potassium, and alkali-earth metal (KUNZEL, O. L. ERD- MANNl[). The best way of proceeding is as follows: The solution (from which any iron must first be separated) is evapo- rated to a small bulk, and then, if much free acid is present, neutralized with potassa. Then add a concentrated solution of potassium nitrite (previously neutralized with acetic acid and filtered from any flocks of silica and alumina that may have separated) in sufficient quantity, and finally acetic acid, till any flocculent precipitate that may have formed from excess of potassa has redissolved and the fluid is decidedly acid. Allow it to stand at least for twenty-four hours in a warm place, take out a portion of the supernatant fluid with a pipette, mix it with more potassium nitrite, and observe whether a further precipita- tion takes place in this after long standing. If no precipitate is formed the whole of the cobalt has fallen down, otherwise the small portion must be returned to the principal solution, some more potassium nitrite added, and after long standing the same test applied. Thus, and thus alone, can the analyst be sure of the complete precipitation of the cobalt. Finally filter and treat the precipitate according to 111, 1, d. Boil the filtrate with excess of hydrochloric acid, precipitate with potash, redissolve the precipitate in hydrochloric acid, throw down the nickel according to 110, 1, , 7, as sulphide, and then * Pogg, Annal , LXXII, 477. \Annal. d. Chem. u. Pharm., xcvi, 218. \ lb., civ, 309. Pogg. Annal. . ex, 412. || Zeitschr. /. analyt. Chem., v, 74. Tf Jb., in. 161: Journ.f. prakt. CTiem., xcvn, 387. 656 SEPARATION. [ 160, convert into metal. In this manner alone can the nickel be obtained pure, as the original filtrate contains so much alkali salt and also generally alumina and silica. [When nickel and cobalt are obtained in the form of sulphides in the process of separation from other metals, the mixed sulphides may be converted into metals without previous separation, by the same process that is described for nickel sulphide 110, 1, &, and 2. Cobalt may then be separated from a nitric acid solution of the two metals and nickel estimated by difference.] 10. Methods Seised upon the different deportment with Potassium Cyanide. a. ALUMINIUM FROM ZINC, COBALT, AND NICKEL. Mix the solution with sodium carbonate, add potassium 97 cyanide in sufficient quantity, and digest in the cold, until the precipitated zinc, cobalt, and nickel carbonates are redissolved. Filter off the undissolved aluminium precipitate, wash, and remove the alkali which it contains, by resolution in hydro- . chloric acid and reprecipitation by ammonia (FRESENIUS and HAIDLEN *). fr. COBALT FROM NICKEL. LIEBIG'S method,f which depends upon the conversion of 98 the cobalt into potassium cobalticyanide, and of the nickel into double nickel and potassium cyanide, has been carefully studied in my laboratory by FR. GAUHE.^: It has been thus found that boiling the solution containing potassium cyanide and hydro- cyanic acid (LIEBIG'S first method) does not completely convert the double cobalt and potassium cyanide first formed into potassium cobalticyanide, but that passing chlorine (LIEBIG'S second method) effects a ready and thorough conversion. The method then gives a very excellent separation, and is more par- ticularly to be recommended where the quantity of nickel is small in proportion to the cobalt. We proceed as follows, taking a hydrochloric solution of the metals : Remove the greater part of the free acid by evaporation or neutralize it by potash, add pure potassium cyanide till the precipitate first * Annal. d. Chem. u. Pharm. , XLIII, 129. f 2b., LXV, 244, and LXXXVII, 128. $ Zeitschr, f. analyt. Chem., v, 75. 160.] BASES OF GROUP IV. 657 formed has redissolved-; then add more cyanide, dilute, boil for some time or not, as you like, pass chlorine through the cold fluid, adding potash or soda occasionally, so that the fluid may remain strongly alkaline to the end. Bromine may be used instead of chlorine, and indeed is far more convenient. In the course of an hour the whole of the nickel will have precipi- tated as black hydrate of the sesquioxide. Having taken out a portion and satisfied yourself of this by addition of a further quantity of chlorine or bromine, filter, and wash with boiling water. The precipitate always retains alkali, and must be redis- solved in hydrochloric acid, and estimated according to 110, 1, a, or 2. As regards the cobalt, it is most convenient to estimate it by difference. But if you wish to make a direct estimation, it will be advisable, in consequence of the large quantity of salts present in solution, first to evaporate to dryness with excess of hydrochloric acid, to take up the residue with a little water, and to heat in a large platinum dish, with the addition of excess of pure concentrated sulphuric acid till the greater part of the sulphuric acid has escaped. The red mass, consisting principally of alkali disulphate, is then treated with water, and the cobalt estimated according to 111, 1, c. Another method of separating nickel and cobalt by means of potassium cyanide has been described by FLECK ; * it does not, however, appear to be in any way better. The method is based on the fact that cobalt monosulphide, as well as nickel sulphide, dissolves readily in potassium -cyanide solution, but that this is not the case with the cobalt sulphide precipitated by ammonium sulphide from a solution of cobalt which has been treated with ammonia in excess and exposed to the air until its color no longer changes. c. COBALT FROM ZINC. Add to the solution of the two metals, which must con- 99 ttim some free hydrochloric acid, common potassium cyanide (prepared by LIEBIG'S method) in sufficient quantity to redis- solve the precipitate of cobalt cyanide and zinc cyanide which * Journ. f. prakt. Chem., xcvn, 303, Zeitsclir.f. analyt. CTiem., v, 399. 658 SEPARATION. [ 16(X forms at first; then add a little more potassium cyanide and boil some time, adding occasionally one or two drops of hydro- chloric acid, but not in sufficient quantity to make the solu- tion acid. After cooling, add some chlorine or bromine, and digest for some time to complete the conversion of the cobalt into potassium cobalticyanide. Mix the solution with hydro- chloric acid in an obliquely placed flask and boil until the zinc cobalticyanide which precipitates at first is redissolved, and the hydrocyanic acid is completely expelled. Add solution of soda or potassa in excess and boil until the fluid is clear ; the solu- tion may now be assumed to contain all the cobalt as potas- sium cobalticyanide, and all the zinc as a compound of zinc oxide and alkali. Precipitate the zinc by hydrogen sulphide ( 108). Filter, and determine the cobalt in the filtrate as in 98. The process is simple and the separation complete (FKE- SENIUS and HAIDLEN). d. NICKEL FROM ZINC. Add to the concentrated solution of the two rnetals an 100 excess of pure concentrated potassa lye, then sufficient aqueous hydrocyanic acid to completely redissolve the precipi- tate, add a solution potassium monosulphide (not ammonium sulphide), let the precipitated zinc sulphide deposit at a gentle heat, filter, wash the sulphide with a dilute potassium-sul- phide solution, treat the precipitate with hydrochloric acid, and from the solution precipitate the zinc with sodium car- bonate, as in 108, 1, a. In the filtrate estimate the nickel by heating for some time with fuming hydrochloric and nitric acids, or instead of the latter, potassium chlorate, evap- orating, and finally precipitating with potassa lye (WOHLER *). KLAYE and DEus,f who tested the process in my laboratory , found that instead of potassa lye and hydrocyanic acid, * potassium cyanide could be used if perfectly pure and recently dissolved. If the solution of the cyanide contains ammonium carbonate or formate, or potassium cyanate (as is the case even on short exposure), the complete precipita- tion of the zinc as sulphide is greatly interfered with. On * AnnaL de Chem. u. Pkarm., LXXXIX, 376. \Zeitschr.f. analyt. Chem., x, 197. 160.] BASES OF GROUP IV. 659 finally washing the precipitated zinc sulphide completely with, water containing hydrogen sulphide, the zinc may be esti- mated according to 108, 2. e. COBALT AND NICKEL FROM MANGANESE AND ZINC (W. GIBBS *). Add sodium acetate to the solution of the chlorides and 101 pass in hydrocyanic-acid gas. Zinc cyanide is immediately more or less completely precipitated as a white powder. Now add sodium sulphide, which converts the zinc and manganese into sulphides, while the cobalt and nickel remain in solution as double cyanides, and may be separated as in 98. The employment of gaseous hydrocyanic acid renders the method very unpleasant. Another method for separating cobalt and nickel from manganese is as follows: To the acid solution add sodium carbonate in excess, then acetic acid in liberal excess, then to the clear fluid, containing say 1 grm. of nickel or cobalt, 30 to 40 c. c. of sodium-acetate solution (1 in 10), and pass hydro- gen sulphide to saturation, keeping at 70. Filter off the precipitated nickel or cobalt sulphide, wash and dry it. Con- centrate the filtrate by evaporation, add ammonium sulphide and then acetic acid, thus obtaining a second precipitate of nickel or cobalt sulphide. Test the filtrate again in the same manner. In the united precipitates determine the nickel or cobalt according to 110, 1, J, or 111, 1, c\ in the fil- trate, the manganese according to 109, 2. 11. Methods Abased on the Volatility of Zinc, a. COBALT AND NICKEL FROM ZINC. BERZELIFS f gives the following method for the absolute 102 separation of cobalt and nickel from zinc : Precipitate the solution with an excess of potassa lye, boil, and filter the fluid containing the greater portion of the zinc oxide dis- solved in the potassa solution from the precipitated nickel and cobalt hydroxides also containing much zinc, wash completely * Zeitschr. f. analyt. Chem., in, 832. f BEIIZELIUS' Jahresber., xxi, 144. 660 SEPARATION. [ 160. with boiling water, and determine the zinc in the filtrate (see 108). Dry the precipitate, ignite, and weigh ; then mix with pure sugar (recrystallized from alcohol) in a porcelain crucible and heat slowly until all the sugar is completely -carbonized. Then place the crucible, covered with its lid, in a bath of magnesia within a covered crucible of larger isize, and heat in a wind furnace for one hour at the highest temperature obtainable. By this treatment the metals are reduced, the nickel and cobalt mixed with carbon remaining behind, while the zinc is volatilized. Treat the residue with nitric acid, and determine the metals by precipitating with potassa solution and weighing the precipitate. The differ- ence between this weight and that obtained before gives the weight of the zinc which had been conjointly precipitated. KLAYE and DEUS,* who tested the method in my laboratory, obtained good results with it. They recommend replacing the sugar by sugar-carbon, as the former causes much intu- mescence. An attempt to use the gas blowpipe, as given by BERZELIUS, gave poor results. lf>. ZINC FROM IRON IN ALLOYS. According to BOBIERRE such alloys may be readily and 103 accurately analyzed by ignition in a current of hydrogen. 12. Methods based upon the Volumetric Determi- nation of one of the Metals, and the finding of the other from the difference. a. FERRIC IRON FROM ALUMINIUM. Precipitate both metals with ammonia ( 105, ' shown in Fig. 115. As soon as the reduction is complete, and which may be recognized by the colorless appearance of the fluid, cool the flask by immersing it in cold water, lift the upper stopper, Fig. 115. 160.] BASES OF GROUP IV. 665 throw a few fragments of sodium carbonate into the acid, draw the zinc ball up the tube 5, wash off the fluid adhering to the ball into the flask, and remove W. Now quickly add the weighed potassium dichromate and proceed as above directed. c. MANGANESE FEOM ALUMINIUM AND IKON (KRIEGER *). Precipitate with sodium carbonate, digest the precipitate 108 for some time with the fluid, wash first by decantation, then on the filter, and as thoroughly as possible, dry, ignite, and deter- mine the manganese in a portion according to 72. Care must be taken that the precipitate contains the manganese as Mn,O 4 , and also that, in the case of highly accurate analyses, the small quantity of manganese passing in the filtrate be not disregarded ( 109, 1, a). The bases may also be precipitated with ammonium carbonate instead of with sodium carbonate (65) ; in fact it deserves the preference. d. MANGANESE FROM ZINC (KEIEGEK). Precipitate boiling with sodium carbonate, wash the pre- 109 cipitate with boiling water, dry, and ignite. If sufficient zinc is present, the precipitate will consist of ZnO -)- a?Mn,O,. Weigh off a portion of the precipitate and in it determine the manganese as in 72. If insufficient zinc is present, proceed as in 72 N. B. Eegarding the small quantity of manganese passing into the filtrate see 109, 1 a. e. COBALT FROM NICKEL. Determine both metals as in 110, 1, &, and 2, and 111, 110 1, &, dissolve the reduced metals in hydrochloric acid with the addition of some nitric acid, evaporate the solution repeat- edly with hydrochloric acid to dryness until all the nitric acid has been expelled, and in the solution of the chlorides then determine the cobalt according to 111, 3, the nickel being found by difference. The method is applicable only in the presence of small quantities of nickel, and gives only fair results. * Annal. de Chem. u. Pharm., LXXXVII, 261. 666 SEPARATION. [161. 13. Indirect Method. FERRIC IRON FROM FERROUS IRON. Of the many indirect methods proposed, but which are 111 now seldom resorted to since the introduction of volumetric methods, I will give only the following : Dissolve in hydro- chloric acid in a current of carbonic acid, add an excess of gold and sodium chloride, stopper the flask, and allow the precipitated gold to subside. Then filter, and determine the gold as in 123. Determine the total quantity of iron in the filtrate or in another portion of the substance. The calcula- tion is readily made if it be remembered that 2 eq. of precipi- tated gold are the equivalent of 6 eq. of ferrous chloride (or oxide) thus : 6FeCl a + 2AuCl 9 = 2Au + 3Fe Q Cl 8 (H. EOSE). IV. SEPARATION OF IRON, ALUMINIUM, MANGANESE, CAL- CIUM, MAGNESIUM, POTASSIUM, AND SODIUM. 161. As these metals are found together in the analysis of most silicates, and also in many other cases, I devote a separate para- graph to the description of the methods which are employed to effect their separation. 1. Method based upon the employment of Barium Car- bonate (particularly applicable in cases where the mixture con- tains only a small proportion of calcium). The solution should contain no free chlorine, 'and the iron 112 should be all in the form of ferric salt. Precipitate the iron and aluminium by barium carbonate * (54 and 76), dissolve the precipitate in hydrochloric acid, throw down the barium with sulphuric acid, filter, and estimate the iron and aluminium according to one of the methods given in 160, by preference 104, at least when the quantity of aluminium is not too small. To the filtrate from the barium -carbonate precipitate add hydrochloric acid, heat, throw down the barium with sulphuric * Before adding the barium carbonate, it is absolutely indispensable to ascer- tain whether a solution of it in hydrochloric acid is completely precipitated by sulphuric acid, so that the filtrate leaves no residue upon evaporation in a platinum dish. $ 161.] BASES OF GROUP IV. 667 acid, added just in excess. Filter off the precipitate, wash till free from soluble sulphate, concentrate if necessary, precipitate, and determine the manganese as sulphide ( 109, 2). To the filtrate add hydrochloric acid, heat, filter oft' the sulphur, pre- cipitate the lime with oxalate of ammonia, and finally separate the magnesia from the alkalies by one of the methods p-iven 153. 2. 3frthod based upon the application of Alkali Acetates or Formcvtes. Eemove by evaporation any very considerable excess of acid 113 which may be present, dilute, add sodium carbonate * until the fiuid is nearly neutral, then sodium acetate (or sodium formate) and precipitate iron and aluminium, observing all directions given in 85. Wash the precipitate well, dissolve in hydrochloric acid, precipitate the solution with ammonia (45), dry, ignite, and weigh. Dissolve in concentrated hydrochloric acid and determine the iron volumetrically with stannous chloride, as in 113, 3, 5, or digest it with 16 times its weight of a mixture of 8 parts sulphuric acid and 3 parts water, or fuse it for a long time with potassium bisulphate, dissolve in water, and determine the iron volumetrically. as in 113, 3, a. The difference gives the quantity of the aluminium. If any silicic acid remains behind on dissolving the pre- cipitate, it is to be collected on a filter, ignited, weighed, and deducted from the alumina. The filtrate contains the manganese, the alkali- earth rnetals, and the alkalies. Pre- cipitate the manganese with ammonium sulphide ( 109, 2), boil with hydrochloric acid and filter off the sulphur, precipi- tate the calcium, after addition of ammonia, with ammonium oxalate, and lastly, after removing the ammonium salts by igni- tion, precipitate the magnesium from the hydrochloric acid solution of the residue with ammonium sodium phosphate. However, if it is intended to estimate the alkalies, the magne- sium must be separated by one of the processes in jj 1 ';], -1-. This * In cases where it is intended to estimate the alkalies in the filtrate, ammo- nium salts must be used instead of the sodium salts. If, however, it is intended to precipitate manganese subsequently with bromine, ammonium salts must not be introduced into the solution. 668 SEPARATION [ 161. method is convenient, mid gives good results, especially in the presence of much iron and little aluminium. Since aluminium is not precipitated by alkali acetates or formates with the same certainty as iron, it is necessary to test the weighed manganese sulphide for aluminium. [This method is to be recommended when manganese is pres- ent with iron, or with iron and a moderate proportion of alumin- ium. If, however, the amount of aluminium is large in propor- tion to the iron, it is difficult to precipitate it completely with sodium acetate. Instead of precipitating manganese with ammonium sulphide it may be separated from calcium and magnesium by precipitation with bromine. Add aqueous solu- tion of bromine to the filtrate from the iron precipitate with- out previous concentration of the filtrate, unless its volume exceeds 600 or 700 c.c., and proceed according to 159, 72, d. 3. Method based upon the application of Ammonium Sul- phide. Mix the fluid in a flask with ammonium chloride, then with 114 ammonia, until a precipitate just begins to form, then with yellow ammonium sulphide, fill the flask nearly up to the top with water, cork it, allow to settle in a warm place, filter, and wash the precipitate consisting of iron and manganese sulphides and aluminium hydroxide without interruption with water containing ammonium sulphide. Separate the calcium, magne- sium, and alkalies in the filtrate as in 113. Dissolve the precipi- tate in hydrochloric acid, and separate the aluminium from the iron and manganese according to 77 or 78, and then the iron from the manganese, say by 82 or 85. The following method is particularly suitable in cases where no manganese is present, or only inappreciable traces: 4. Method based upon the application of Ammonia. a. The solution must contain all the iron in the state of a 115 ferric salt. Add a relatively large quantity of ammonium chloride, and observing the precautions indicated in 45 precipitate with ammonia. The precipitate contains the whole of the iron and aluminium ; at most an inappreciable amount of the latter remains in solution if the free ammonia has been almost but not entirely driven off by heat, if the solution was 161.] BASES OF GROUP IV. 669 diluted sufficiently, and if enough ammonium chloride was present. It may also contain small quantities of calcium and magnesium and a little manganese. It is well, therefore, usually to redissolve the washed precipitate in hydrochloric acid, and reprecipitate with ammonia. In this way the pre- cipitate will be obtained free from alkali-earths and manganese. Wash the precipitate completely, dry, ignite, and treat accord- ing to 113. If silicic acid remains undissolved, it is to be determined and deducted. The solution filtered from the aluminium and ferric hydroxide is concentrated by evaporation, and the manganese is precipitated and determined according to 109, 2, as sulphide; the alkali-earth metals and alkalies in the filtrate are determined according to 113. The weighed sulphide of manganese is digested with dilute hydrochloric acid ; any residue that may remain is fused with potassium bisulphate, dissolved in water, and tested for aluminia. 5. Precipitate the aluminium, iron, and calcium by add- 116 ing ammonia and ammonium carbonate and oxalate, decant, and filter. Dissolve the precipitate in hydrochloric acid, add pure tartaric acid to prevent the aluminium and iron from being precipitated, and then precipitate the calcium with ammonia as an oxalate. In the solution separate the iron and aluminium as in 77 ; and in the first filtrate the magnesium and alkalies according to 18. Should sulphuric acid be pres- ent in the first filtrate, remove it by means of barium chloride, then separate the alkali-earths from the alkalies by evaporat- ing with oxalic acid, igniting, and treating the residue with boiling water, and finally separate the barium from the mag- nesium as in 29 (E. MITSCHERLICH ; LEWTNSTEIN *). As alu- minium in the presence of ammonium oxalate is only grad- ually precipitated on warming (PISANI), the liquid must be digested for some time with heat before the first filtration; and as the precipitate always contains a portion of the magne- sium, I would advise that, after separating the iron from the aluminium, the filtrate from the latter, as well as the alumina itself, be tested for magnesia. If weighable quantities of manganese are present, the method is inapplicable. * Jourji. f. prakt. Chem., LVIII, 99. 670 SEPARATION. [ 161. c. Precipitate with ammonia, digest for some time with 117 heat, and until the greater part of the excess of ammonia has been expelled, filter, carefully and thoroughly wash the pre- cipitate, ignite, and add to the residue, without reducing it to powder, at least ten times its quantity of anhydrous sodium carbonate, cover the crucible and heat the mixture in a blast- lamp or other suitable flame (an alcohol-lamp with double draught is not sufficiently powerful) until no further decom- position of the sodium carbonate is observed, for at least 45 minutes. Now add some caustic potassa to the fused mass, and boil it with water (heat in a silver dish) until thoroughly ex- tracted ; if a green color indicates that sodium manganate is present add a few drops of alcohol, wash the precipitate by decantation and filtration, first with water containing potassa, then with pure water. Dissolve the precipitate in hydro- chloric acid, add a few drops of alcohol, and heat in order to more readily reduce the manganese chloride, and finally add ammonium acetate to separate the iron from the portions of manganese, calcium, and magnesium which were contained in the ammonia precipitate, and which may be estimated sepa- rately or together with the main quantities according to 113. The aluminium is determined in the alkaline solution as in 78. (R. KIOHTEB.*) 5. Method based on the Decomposition of the Ni- trates (DEVILLE). This method assumes that the bases are present as nitrates 118 only. Proceed first as in 46. The nitrous acid evolved dur- ing the heating of the nitrate is no indication of the total de- composition of the ferric or aluminium nitrate, because these vapors may also be due to the conversion of manganous nitrate into manganese dioxide. When all vapors cease to be evolved, and the substance acquires a uniform black color, interrupt the heat. After treatment with ammonium nitrate there remains in solution the nitrates of calcium, magnesium, and the alka- lies, while the residue will contain aluminium, iron, manga- nese dioxide, and if much manganese is present small quan- * Journ. /. prakt. Chem., LXIV, 378. 161.] BASES OF GROUP IV. 671 titles of alkaline earths. (That under certain circumstances some manganese dissolves has already been stated in 71 ; this trace is found with the magnesium, from which it is finally separated.) DEVILLE recommends the following methods to further effect the separation : a. Heat the precipitate with moderately strong nitric acid until the iron and aluminium are dissolved, leaving the man- ganese dioxide as a pure- black residue, which is ignited, and the sesquioxide then weighed. Evaporate the solution in a platinum crucible, ignite the residue, and weigh the mixture of ferric oxide, alumina (and possibly some manganese sesqui- oxide). Now treat a portion according to 91, and thus find the alumina. If manganese was present, the iron cannot be determined by difference. DEVILLE, therefore, evaporates the solution of the chlorides (92) with sulphuric acid, ignites gently, and treats the residual mixture of ferric oxide and manganous sulphate with water to remove the manganese salt. (In case too strong a heat has been applied, in which case the manganous sulphate may also have been decomposed, moisten the residue with a mixture of oxalic and nitric acids, add a little sulphuric acid and repeat the ignition.) b. From the filtrate precipitate first the calcium with ammonium oxalate and then separate the magnesium as in 153, 4. In the presence of manganese this method is not to be recommended. 6. Method which combines 4- and 5. Precipitate with ammonia (45), decant, filter, wash, 119 remove the still moist precipitate so far as possible from the filter, dissolve the remainder in nitric acid and transfer this to the dish to effect solution of the bulk of the precipitate, pro- ceed according to 118, and mix the fluid separated from the ferric oxide and alumina (and which contains small quantities of magnesium, possibly also traces of calcium) with the main filtrate. This method is to be recommended when manganese is absent. The estimation of aluminium is best effected by determining the total weight of the ferric oxide and alumin- ium, and then determining the iron volumetrically (104). If 67:2 SEPARATION. [ 161. on dissolving the precipitate of ferric oxide and alumina there remains any silica, this must be deducted. Supplement to the Fourth Group. To 158, 159, 160. SEPAKATION OF UEANIUM FROM THE OTHER METALS OF GROUPS I. IY. It has already been stated, in 114, that uranium in uranyl 120 compounds cannot be completely separated from the alkalies by means of ammonia, as the precipitated ammonium uranate is likely to contain also fixed alkalies. The precipitate should therefore be dissolved in hydrochloric acid, the solution evapo- rated in the platinum crucible, the residue gently ignited in a current of hydrogen gas (Fig. 83), the chlorides of the alkali metals extracted with water, and the uranous oxide (UO a ) ignited in hydrogen in order to weigh it as UO 2 , or in the air, whereby it is converted into uranous uranate, TJ(UO 4 ) a . Instead of dissolving the precipitate in hydrochloric acid and treating the solution as directed, you may .heat the precipitate cautiously * with ammonium chloride and treat the residue with water (H. HOSE). Uranium may be completely separated from the alkalies also by ammonium sulphide, as H. ROSE found. REMELE f has examined this subject with great care and recommends the following method of pre- cipitation : The solution being neutral or slightly acid, add an excess of yellow ammonium sulphide and keep nearly boiling for an hour to convert the first-formed precipitate of uranium oxysulphide entirely into a mixture of uranous oxide and sul- phur. The fluid, at first dark from presence of dissolved uranium, will now appear yellow and transparent. Filter off the precipitate containing all the uranium and wash it with cold or warm water, first by decantation, finally on the filter. It is well to mix a little ammonium sulphide or chloride with * Strong ignition would occasion the volatilization of uranium chloride. f Zeiischr.f. analyt. Chem., iv, 379. 161.] BASES OF GROUP IV. 673 the water, as when pure water is used the last filtrate is apt to be turbid. The dried precipitate is roasted and then converted into uranons uranate by ignition in the air, or into uranous oxide by ignition in hydrogen ( 114). FR. STOLBA * recommends separating uranyl from alkalies 121 by means of hydrosilicofluoric acid with the addition of alcohol. Treat the substance with a sufficient quantity of 3- to 5-per cent, aqueous silicon 1 uoric acid and warm gently. As soon as the yellow powder has disappeared, allow to cool, add 3 to 4 volumes of 75- to 80-per cent, alcohol, mix, allow to settle in the dark, or at least in a place not exposed to direct sunlight, filter, wash with alcohol until the washings are -absolutely free from acidity, and determine the alkali volumetrically accord- ing to 97. 5. Direcj: sunlight renders the alcoholic solution cloudy, an insoluble uranium silicofluoride precipitating. If the uranyl is to be estimated also, evaporate the alcoholic liquid, heat the residue with an excess of sulphuric acid to expel the hydrosilicofluoric acid, dissolve the residue in water with the addition of some nitric acid, filter, and in the filtrate determine the uranyl according to 114. This method is also applicable for the analysis of uranyl- alkali salts soluble in alcohol. It should be remarked here that moderate quantities of hydrochloric or nitric acid do not noticeably interfere, while sulphuric acid, by causing a pre- cipitation of alkali sulphates, gives too low an alkali value. From barium, uranyl may be separated by sulphuric acid ; 122 from strontium and calcium, by sulphuric acid and alcohol. Ammonia fails to effect complete separation of uranyl from the alkali-earth metals, the precipitate always containing not inconsiderable quantities of the latter. In such precipitates, however, the uranium and the alkali-earth metals may like- wise be separated by gentle ignition with ammonium chloride and treatment of the residue with water. Uranyl may be separated from strontium and calcium also 123 by precipitation with ammonium sulphide by the method given above in the separation from the alkalies. As carbonates of the alkali-earth metals may be coprecipitated, treat the washed pre- *Zcitsc7ir. f. analyt. Chem., in, 71. 074 SEPARATION. [ 161, cipitate of uranous oxide and sulplmr in the cold with dilute hydrochloric acid which will not dissolve uranous oxide. Ammonium sulphide will not answer for the separation of uranium from barium (REMELE*). Magnesium may be separated from uranyl not only by 124 ammonium sulphide in presence of ammonium chloride, but also by ammonia. Add enough ammonium chloride to the solution, heat to boiling, supersaturate with ammonia, continue boiling till the odor of ammonia is but slight, filter the hot fluid, and wash the precipitate, which is free from magnesium, with hot water containing ammonia (H. ROSE). It is always well to test the uranous oxide obtained by ignition in hydrogen for magnesium by treating with dilute hydrochloric acid. Aluminium is best separated from uranyl by mixing the somewhat acid fluid" with ammonium carborfate in excess. The uranyl passes completely into solution, while the aluminium remains absolutely undissolved. Filter, evaporate, add hydro- chloric acid to resolution of the precipitate produced, heat till all the carbonic acid is expelled, and precipitate with ammonia ( 1U). Uranyl is best separated from chromium (W. GiBBsf) by adding to the solution soda in slight excess, heating to' boiling and adding bromine water, when the chromium is rapidly converted into chromic acid. Filter the solution containing sodium chromate from the precipitate which has a deep orange- red color and consists of a compound of soda and uranic oxide mixed with some uranyl chromate. Wash the precipitate with hot water containing a little soda, dissolve it in hot nitric acid, boil the solution a few minutes to drive off any nitrous acid, and precipitate the chromic acid according to 130. L, -, ft with mercurous nitrate (according to GIBBS at a boiling heat). The filtrate now contains the whole of the uranium, of course in presence of mercury. The separation of uranyl from the metals of the fourth 125 group may be based simply on the fact that ammonium carbonate prevents the precipitation of uranyl, but not that of the other metals by ammonium sulphide. Mix the solution with a mixture of ammonium carbonate and ammonium sulphide, allow * Zeitschr.f. analyt. Chem., iv, 383. ^ Ib. t xn, 310. 161.] BASES OF GROUP IY. 675 to subside in a closed flask, and wash the precipitate with water containing ammonium carbonate and ammonium sulphide. Remove the greater part of the excess of ammonium car- bonate from the filtrate by a very gentle heat, acidify with hydrochloric acid, warm, filter off the separated sulphur, and throw down the uranium either by ammonium sulphide (see above, Separation of Uranium from the Alkalies] or by heating with nitric acid and then adding ammonia (H. ROSE,* REMELEf). The method is not so suitable in presence of nickel, as a little of this metal is very liable to pass into the filtrate on precipitation with ammonium carbonate and ammonium sulphide. Ferric iron may be also^eparated from uranyl by means of an excess of ammonium carbonate. The small quantity of- iron which passes with the uranium into solution will fall down on allowing the solution to stand for several hours, or it may be precipitated with ammonium sulphide, before the uranium is thrown down (PiSANiJ). From nickel, cobalt, manganese, zinc, and magnesium the uranyl may also be separated by barium carbonate. The fluid, which should contain a little free acid, is mixed with the precipitant in excess, and allowed to stand in the cold for 24 hours with frequent shaking (76). From cobalt, nickel, and zinc, uranyl may also be separated 126 (GIBBS and PERKINS) by taking the neutral or slightly acid solutions of the chlorides, adding sodium acetate in excess and a few drops of acetic acid, and passing a rapid current of hydro- gen sulphide for half an hour through the boiling fluid. The uranium remains dissolved while the other metals are precipi- tated. I should advise testing the filtrate with a mixture of ammonium carbonate and ammonium sulphide to see if any nickel, cobalt, or zinc remain in solution. * Zeitsctir. f. analyt. Chem., I, 412. \ Compt. rend., LII, 106. f/6., iv, 385. Zeit8chr.f. analyt. Chem., in, 334. 676 SEPARATION. [ 162. Fifth Group. SILVER MERCURY (iN MERCUROUS AND MERCURIC COMPOUNDS) LEAD BISMUTH COPPER CADMIUM. I. SEPARATION OF THE METALS OF THE FIFTH GROUP FROM THOSE OF THE FIRST FOUR GROUPS. 162. INDEX. (The numbers refer to those in the margin.) Silver from the metals of Groups L IY., 127, 128. Mercury (in mercurous and mercuric compounds) from the metais of Groups I. IV., 127, 129. Lead from the metais of Groups L V., 127, 130. " manganese, 142. Binmuth from the metais of Groups I. IV., 127, 140. " manganese, 142. Copper from the metals of Groups I. IV., 127, 131, 132, 133, 134, 135. Copper from zinc, 136, 137. " manganese, 142. iron, 138. " nickel, 139. Cadmium from the metals of Groups I. IV., 127. " zinc, 105. " manganese, 142. A. General Method. ALL THE METALS OF THE FIFTH GROUP FROM THOSE OF THE FIRST FOUR GROUPS. Principle : Hydrogen Sulphide precipitates from Acid Solutions the Metals of the Fifth Group, but not those of the first Four Groups. The following points require especial attention in the execu- tion of the process : OL. To effect the separation of the metals of the fifth group 127 from those of the first three groups, by means of hydrogen sulphide, it is necessary simply that the reaction of the solution should be acid, the nature of the acid to which the reaction is due being of no consequence. But, to effect the separation of the metals of the fifth group from those of the fourth, the presence of a free mineral acid is indispensable ; otherwise zinc and, under certain circumstances, also cobalt and nickel may be coprecipitated. 162.] BASES OF GROUP V. 677 (3. But even the addition of hydrochloric acid to the fluid will not always entirely prevent the coprecipitation of the zinc. RIVOT and BOUQUET * declare a complete separation of copper from zinc by means of hydrogen sulphide altogether imprac- ticable. CALVERT f states that he has arrived at the same con- clusion. On the other hand, SPIRGATIS J concurs with H. ROSE in maintaining that the complete separation of copper from zinc may be effected by means of hydrogen sulphide in presence of a sufficient quantity of free acid. In this conflict of opinions, I thought it necessary to subject this method once more to a searching investigation. I there- fore had R. GRUNDMANN make a series of experiments in the matter in my laboratory with a view to settling the question.f The following process is founded on the results which we obtained : Add to the COPPER and ZINC solution a large amount of hydrochloric acid (e.g., to 0'4 grm. oxide of copper in 250 c.c. of solution, 30 c.c. hydrochloric acid of 1-1 sp. gr.), conduct into the fluid at about 70 hydrogen sulphide largely in excess, filter before the excess of hydrogen sulphide has had time to escape 01- become decomposed, wash with hydrogen sulphide water, dry, roast, redissolve in nitrohydrochloric acid, evaporate nearly to dryness, add water and hydrochloric acid as above, and pre- cipitate again with hydrogen sulphide. This second precipi- tate is free from zinc ; it is treated as directed in 119, 3. If CADMIUM is present, it is well to have less acid present, e.., LXXIII 241. | Ib , i. xvn, 371. 678 SEPARATION. [ 162. sary, be tested by addition of a large quantity of hydrogen sul- phide to see if the precipitation of the fifth group was com- plete. #. If hydrochloric acid produces no precipitate in the solution, it is preferred as acidifying agent, otherwise sulphuric or nitric acid must be used. In the latter case the fluid must be rather largely diluted. ELIOT and STOKER * arrived at the same conclusion as ourselves, and showed that the cause of CALVERT'S unfavorable results was the too large dilution of his solutions. For, to prevent the precipitation of zinc you have not merely to preserve a certain proportion between the zinc and the free acid, but also a certain degree of dilution. Although I agree with the above-named chemists in the opinion that it is possible to produce a condition of the fluid, under which one precipitation will effect complete separation, still it appears to me better, for practical purposes, to precipitate twice, as this is sure to lead to the desired result. e. A somewhat extended experience in the separation of COP- PER from NICKEL (and COBALT) which so frequently occurs has led me to the opinion that a double precipitation is unnecessary. If the solution which is to be treated with hydrogen sulphide contains enough free hydrochloric acid and riot too much water, the copper falls down absolutely free from nickel, while, on the other hand, if the quantity of free acid is not too large, the fil- trate will be quite free from copper. The method of sepa- rating copper from zinc given in j3 is also to be recommended in this case. C- CADMIUM and ZINC may, according to FoLLENius,f also be completely separated by a single precipitation, if the metals are present in a sulphuric acid solution containing 25 or 30 per cent, of dilute acid of 1*19 sp. gr. Precipitate with hydrogen sulphide at 70. Collect the precipitate on a weighed asbestos filter, dry in a current of heated air, ignite gently in a stream of pure hydrogen sulphide (to convert small quantities of cad- mium sulphate into sulphide), remove the small quantity of .separated sulphur by gentle ignition in a current of air, and weigh. * On the Impurities of Commercial Zinc, etc. Memoirs of the American Academy of Arts and Sciences. New Series. Vol. 8. f Zeitschr.f. analyt. Chem., xm, Part 4. 162.] BASES OF GROUP V. 679 B. Special Methods. SINGLE METALS OF THE FIFTH GROUP FROM SINGLE OB MIXED METALS OF THE FIRST FOUR GROUPS. 1. SILVER is most simply and completely separated from the 128 METALS OF THE FIRST FOUR GROUPS by means of hydrochloric acid. The hydrochloric acid must not be used too largely in excess, and the fluid must be sufficiently dilute ; otherwise a portion of the silver will remain in solution. Care must be taken also not to omit the addition of nitric acid, which pro- motes the separation of the silver chloride. The latter should be treated according to 115, 1, a. 2. The separation of MERCURY from the METALS OF THE 129 FIRST FOUR GROUPS may be effected also by ignition, which will cause the volatilization of the mercury or the mercurial com- pound, leaving the non-volatile bodies behind. The method is applicable in many cases to alloys, in others to oxides, chlorides, or sulphides. If the mercury is estimated only from the loss, the operation is conducted in a crucible ; otherwise in a bulb- tube, or a wide glass tube with porcelain boat. In the latter case it is well to use a current of hydrogen (compare 118, 1, a ; also Examination of Mercury Ores in the Special Part). The precipitation of mercury as mercurous chloride with phosphorous acid, according to 118, 2, is also well adapted for its separation from metals of the first four groups. If the mer- cury is already present as a mercurous salt, it may be separated and determined in a simple manner, by precipitation with hydrochloric acid ( 117, 1). 3. FROM THOSE BASIC RADICALS WHICH FORM SOLUBLE SALTS 130 WITH SULPHURIC ACID, LEAD may be readily separated by that acid. The results are very satisfactory, if the rules given in 116, 3 are strictly adhered to. If you have lead in presence of barium, both in form of sulphn'es, digest the precipitate with a solution of ordinary ammonium sesquicarbonate, without application of heat. This decomposes the lead salt, leaving the barium salt unaltered. Wash, first with solution of ammonium carbonate, then with w;itc>r, and separate finally the lead carbonate from the barium sulphate, by acetic acid or dilute nitric acid (II. HOSE*). The *Journ.f. praL-t. Clicm., LXVI, 166. GSO SEPARATION. [ 162. same object may also be attained by suspending the washed insoluble salts in water and digesting with a clear concentrated solution of sodium thiosulphate at 1520 (not higher). The barium sulphate remains undissolved, the lead sulphate dis- solves. Determine the lead in the filtrate (after 116, 2) as lead sulphide (J. LOWE *). The method recommended, by EJVOT, BEUDANT, and DAGUIN f for separating lead by adding sodium acetate to the solution, heating, and passing in chlo- rine gas, requires to be carried out with great caution, ac- cording to H. RosE,J since portions of other metals, even such as are not converted into higher oxides e.g., zinc are very likely to be precipitated with the lead. 4. COPPER FROM ALL METALS OF THE FIRST FoiJR GROUPS. a. Free the solution as far as possible from hydrochloric 131 and nitric acids by evaporation with sulphuric acid. Dilute if necessary, boil, and add sodium thiosulphate as long as a black precipitate continues to form. As soon as this has deposited, and the supernatant fluid contains only suspended sulphur, the whole of the copper is precipitated. The precipitate is cuprous sulphide (Cu 2 S), and may be readily washed without suffering oxidation. Convert it into anhydrous cuprous sul- phide by ignition in hydrogen ( 119, 3). The other metals are in the filtrate and washings. Evaporate with some nitric acid, filter, and determine the metals in the filtrate. || Results good. The method requires practice, as the end of the pre- cipitation of the copper is not so easy to hit as when hydro- gen sulphide is employed. If the solution contained hydrochloric or nitric acid, and this was not first removed before the addition of the thiosul- phate, the precipitant would be required in much larger quan- tity ; in the presence of hydrochloric acid because the cuprous *Journ. f. prakt. Chem., LXXVIT, 75. f76., LXT. 136. \ Pogg. Annal, ox, 417. The commercial salt is often not sufficiently pure, in which case some ^odium carbonate must be added to its solution and the mixture filtered, *|| As far back as 1842. C. HIMLY made the first proposal to employ sodium thiosulphate for the precipitation of many metals as sulphides (Annal. d Chem. u. Pharm., XLIII, 150). The question, after long neglect, was afterwards taken up again by VOHL (Annal. d. Chem. u. Pharm., xcvi, 237), and SLATER (Chem,. Gaz., 1855, 369) FLAJOLOT, however, made the first quantitative experiment (Annal des Mines, 1853, 641; Journ. f. prakt. Chem., LXI, 105). The results obtained by him are perfectly satisfactory, 162.] BASES OF GROUP V. 681 chloride produced is only decomposed by a large excess of tliiosulphate, in the presence of nitric acid because the thio- sulphate does not begin to act on the copper salt till all the nitric acid is decomposed. ~b. Precipitate the copper as cuprous sulphocyanate 132 according to 119, 3, &, or 119, 4, e\ the other metals remain in solution (Kivox). If alkalies were present and it were desired to determine them in the filtrate, ammonium sulphocyanate must be used instead of the potassium salt usually employed. This method is particularly well adapted for the separation of copper from zinc. The zinc can be precipitated at once from the filtrate by sodium carbonate. The method is also suitable for separating copper from iron (H. ROSE *) ; in this case it is unnecessary that ferric salts be completely reduced by the sulphurous acid added ; the sepa- ration may be effected, even if the solution becomes blood- red on the addition of the precipitant. c. The method proposed by FLAJOLOT,f and which has 133 been so frequently recommended, consisting in precipitating the copper by adding a solution of iodine in aqueous sul- phurous acid, after removing the greater part of the free acid present and adding sulphurous acid, gives inaccurate results, according to II. EOSE,^ because a not inconsiderable quantity of copper remains dissolved in the sulphurous liquid. This difficulty may be avoided by adding to the hydrochloric-acid solution, containing a slight excess of acid, an excess of stannous chloride, ammonium chloride, and potassium iodide, until this last just predominates (E. FLEISCHER ). As, how- ever, the excess of stannous chloride in the filtrate and the stannic chloride formed must first be removed before the bases of groups 1 to 4 can be determined, this method offers no advantages. d. If the solution is not too dilute, the bases being pres- 134 ent as sulphates, while hydrochloric and nitric acids are absent, the copper may also be completely precipitated by * Pogrj. Annal., ex, 424 f Annal. des Mines, 1853, 641; Journ.f. prakt. Chem., LXI, 105. \Pogg. Annal., ex, 425. Zeil8chr.f. analyt. Chem., ix, 256. 682 SEPAKATION. [ 162. means of an alkali hypophosphite. At about 70 copper hydride is precipitated, and this, on heating to a still higher temperature, which should not, however, exceed 90, decom- poses into copper and hydrogen. The precipitation is com- plete when a drop of the liquid is no longer colored brown by hydrogen sulphide. "Wash the spongy copper by decan- tatiori, dry, and ignite in a current of hydrogen. The sepa- ration is complete (W. GIBBS and R. CHAUVENET*). The method is adapted particularly for separating copper from the metals of group 4, which may be precipitated from the filtrate by ammonium sulphide. e. The solution should be free from hydrochloric acid, 135 and should contain a certain quantity of free nitric acid (20 c. c. nitric acid of 1-2 sp. gr. to 200 c. c.) and some sul- phuric acid. Throw down the copper by a galvanic current, so that the metal may be firmly deposited on a platinum ves- sel (preferably a platinum cone), which forms the negative pole. Take care that the current is strong enough, and, without interrupting it, remove the cone from the fluid occasionally to see when the copper is all precipitated. With proper execution the separation of copper from all metals of groups 14 is thorough. All metals of groups 1-4 remain dis- solved, except manganese, which separates as dioxide at the positive pole. The method requires practice and strict atten- tion to the conditions which have been determined by a long course of experiments. It is particularly suited for mining assays and manufacturers. The electrolytic method of sepa- rating copper was, I believe, first recommended by GIBBS, f and afterwards improved by LUCKOW. J LECOQ DE BOISBAU- DRAN, ULLGREN, I and MERRICK^[ hate also written on this subject. Finally the method was very accurately and minutely described by the Mansfelder Ober-Berg- und Hut- tend irection at Eisleben,** who, after giving a prize to LUCKOW 's method, afterwards adopted it, and still further * Zeitschr.f. analyt. Chem., vn, 256. f 75., m, 334. Dingier s polyt. Journ., CLXXVII, 296, and (in detail) Zeilschr, f. analyt. ., vin, 25. %Zeitschr. f. analyt. Chem., vn, 253, and ix, 102. | 75., 7, 255. ^American Chemist, n, 136. ** Zeitschr. /. analyt. Chem., xi, 1 162.] BASES OF GROUP V. 683 improved it. I must refer the reader for details to the last mentioned memoir and LUCKOW'S paper. 5. COPPER FROM ZINC. a. BOBIERRE * employed the following method with satis- 136 factory results in the analysis of many alloys of zinc and copper : The alloy is put into a porcelain boat lying in a por- celain tube, and heated to redness for three-quarters of an hour at the most, a rapid stream of hydrogen gas being con- ducted over it during the process. The zinc volatilizes, the copper remains behind. If the alloy contains a little lead (under 2 to 3 per cent.) this goes off entirely with the zinc, and is partly deposited in the porcelain tube in front of the boat ; if more lead is present part only is volatilized, the rest remaining with the copper (M. BURSTYN f). I. A. W. HOFMANN'S method given below (159) for 137 separating copper from cadmium (boiling the precipitated sulphides with dilute sulphuric acid, whereby the cadmium sulphide is dissolved while copper sulphide remains behind), is also adapted for separating copper from zinc (G. C. WITTSTEIN J). 6. COPPER FROM IRON. One of the oldest methods of separating the oxides con- 138 sists in precipitating the solution with ammonia and filtering off the precipitated iron hydroxide from the amrnoniacal copper solution. To obtain accurate results by this method, however, the precipitation must be repeated according to the quantity of copper present, two or even three times, or until the filtrate is no longer blue, otherwise the iron hydroxide will contain copper. 7. COPPER FROM NICKEL. Evaporate the nitric-acid solution, if it be such a one, 139 with hydrochloric acid to dry ness, dissolve the chlorides in water, add about twice as much pure potassium tartrate as there are metals present, warm slightly to facilitate solution, * Compt. rend., xxxvi, 224; Journ.f. pi*akt. Chem., LVIII, 380. "\Zeitschr.f. analyt. Chem., xi, 175. \ VwrteljahresscJir. f. prakt. Pharm., xvn, 461 ; Zeitschr. f analyt. Chem., vni, 202. 684 SEPARATION. [ 162. and then add gradually alcoholic solution of potassa until the hydrated oxides precipitated redissolve. After cooling, add a solution of pure grape sugar, and boil for one or two minutes. The copper is precipitated as cuprous oxide. Make certain that precipitation is complete by adding a drop of grape- sugar solution to the clear liquid, then filter, and determine the copper as oxide, either by ignition (treating with nitric acid and reigniting), or as cuprous sulphide ( 119, 3, c), or volumetrically ( 119, 4, e). Evaporate the liquid contain- ing the nickel to dryness, ignite the residue, remove the potassium carbonate by washing, reignite, dissolve the resi- due in nitrohydrochloric acid, and precipitate the nickel with potassa solution,, as in 110, 1, a (DEWILDE*). The cu- prous oxide must be rapidly filtered off and washed, otherwise a part will redissolve. The method is inconvenient, and is by no means more accurate than the separation by hydrogen sulphide. 8. BISMUTH FKOM THE METALS OF THE FIBST FOUR GROUPS, WITH THE EXCEPTION OF FERRIC IRON. Precipitate the bismuth according to 120, 4, as bismuth 140 oxychloride, and determine it as metal ; all the other basic metals remain completely in solution. Results very satis- factory (H. KosEf). 9. CADMIUM FROM ZINC. Prepare a hydrochloric- or nitric-acid solution of the two 141 oxides as neutral as possible, add. a sufficient quantity of tar- taric acid, then solution of potassa or soda, until the reaction of the clear fluid is distinctly alkaline. Dilute now with a sufficient quantity of water, and boil for 1^-2 hours. All the cadmium precipitates as hydroxide, free from alkali (to be determined as directed in 121), whilst the whole of the zinc remains in solution ; the latter metal is determined as directed in 108, 1, I (AuBEL and KAMDOHK^:). The test-analyses communicated are satisfactory. As the separation only suc- ceeds when the substances are present in correct proportions, * Chem. News, 1863, vn, 49; Zeitschr. f. analyt. Chem., n, 72. f Pogg. AnnaL, ex, 429. \Annal. d. Chem. u. Pharm., cm, 33. 163.] BASES OF GROUP V. 685 I will give the quantities employed by AUBEL and HAMDOHK with especially good effect. About 1 grm. oxide of zinc and 1 grm. oxide of cadmium were dissolved in hydrochloric acid, 30 grm. solution of tartaric acid (containing 0*23 grrn. acid in 1 grm.), 50 grm. soda solution of 1*16 sp. gr., and 120 grm. water added, and the whole boiled 2 hours. (The boiling must on no account be done in glass; a platinum or .silver dish should be used.) 10. LEAD, BISMUTH, CADMIUM, AND COPPER FROM MAN- OANESE, If the solution contains a manganous salt and one of the 142 other bases, precipitate the hot solution with sodium carbonate, wash the precipitate first by decantation, then on the filter, with boiling water, dry, ignite for some time, weigh, and in a portion of the residue estimate the manganese volumetric- ally (72). If sufficient bismuth, lead, cadmium, or copper is present, the residue will have the formula Mn 3 O s -\- a?MO, or MnO, -j- ^Bi 2 O, (KKIEGER*). Never neglect testing the filtrate by adding ammonium sulphide, to ascertain whether the metals have been completely precipitated by the sodium carbonate. When precipitating copper by alkali carbonates, dilute the liquid so that it contains about 1 grm. per litre, add the alkali carbonate in very slight excess, and boil the mixture for about half an hour, whereby the bluish-green basic carbonate becomes dark, granular, and more easily washed ("W. GIBBS and E. R. TAYLOR f). II. SEPARATION or THE METALS OF THE FIFTH GROUP FROM EACH OTHER, if 163. INDEX. (The numbers refer to those in the margin.) Silver from copper, 143, 148, 150, 164, 165. " cadmium, 143, 148, 150. bismuth, 143, 147, 150, 161. " mercuricuni,! 143, 148, 150, 158, 160. lead, 143, 146. 147, 150, 155, 164, 165. * Arwal. de Chem. u. Pharm., i/xxxvir, 264. \Zeit8chr.f. analyt. CJiem., vn, 258. t For the sake of brevity the terms " inercuriciim " and " inercurosum " are used to designate respectively mercury in mercuric and mercurous compounds. 686 SEPARATION. [ 163. Mercuricum* from silver, 143, 148, 150, 158, 160. " mercurosum,* 144. ' lead, 145, 146, 147, 150, 158, 160. bismuth, 145, 147, 150, 151, 158. copper, 145, 149, 150, 158, 160. cadmium, 145, 150, 158. Mercurosum* from inercaricum, 144. 44 copper, 144. <4 cadmium, 144. lead, 144, 146. Compare also uiercuricuin from other metals. Lead from silver, 143, 147, 150, 155, 164, 165. mercuricum, 145, 146, 147, 150, 158, 160. " mercurosum, 144, 146. " copper, 146, 147, 150, 152. " bismuth, 146, 147, 152, 161, 162. " cadmium, 146, 147, 150. Bismuth from silver, 143, 147, 150, 161. lead, 146, 147, 152, 161, 162. copper, 147, 150, 151, 153, 161. cadmium, 147, 150, 151, 152, 157. mercuricum, 145, 147, 150, 151, 158. Copper from silver, 143, 148, 150, 164, 165. lead, 146, 147, 150, 152. bismuth, 147, 150, 151, 153, 161. mercuricum, 145, 149, 150, 158, 160. " mercurosum, 144. " cadmium, 149, 150, 152, 154, 156, 159. Copper in cupric from copper in cuprous compounds, 163, 165. Cadmium from silver, 143, 148, 150. lead, 146, 147, 150. bismuth, 147, 150, 151, 152, 157. copper, 149, 150, 152, 154, 156, 159. " mercuricum, 145, 150, 158. mercurosum, 144. 1. Methods based upon the Insolubility of certain of the Chlorides in Water or Alcohol. a. SILVER FROM COPPER, CADMIUM, BISMUTH, MERCURICUM, AND LEAD. a. To separate silver from copper, cadmium, and bismuth, 143 add to the nitric acid solution containing excess of nitric acid, hydrochloric acid as long as a precipitate forms, and separate the precipitated silver chloride from the solution which con- tains the other metals, as directed 115, 1, a. In the presence '"' For the sake of brevity the terms "mercuricum" and "mercurosum " are used to designate respectively mercury in mercuric and mercurous compounds. 163.] BASES OF GROUP V. 687 of bismuth, after pouring off the supernatant fluid, heat again with nitric acid, and wash with dilute nitric acid before wash- ing with water. ft. If you wish to separate mercuricum from silver by hydrochloric acid, special precautions must be taken, as a solu- tion of mercuric nitrate possesses the property of dissolving silver chloride (WACKENRODER, v. LIEBIG,* H. DEBRAYf). Although the silver chloride in solution for the most part separates on the addition of enough hydrochloric acid to con- vert the mercuric nitrate into chloride, or on addition of sodium acetate, still we cannot depend upon the complete pre- cipitation of the silver. On this account, mix the nitric acid solution which must not contain any mercurous salt, and is to be in a sufficiently dilute condition and acidified with nitric acid w r ith hydrochloric acid, as long as a precipitate forms. Allow to deposit, filter off the clear fluid, heat the precipitate to 'free it from any possibly coprecipitated basic mercuric salts with a little nitric acid, add water, then a few drops of hydrochloric acid, and filter off the silver chloride. In the filtrate determine the mercury as sulphide ( 118, 3), and finally test this for silver, by ignition in a stream of hydrogen any silver that may happen to be present will remain behind in the metallic state. y. In the separation of silver from lead, the precipitation is advantageously preceded by addition of sodium acetate. The fluid must be hot and -the hydrochloric acid rather dilute; no more must be added of the latter than is just necessary. In this manner, the separation may be readily effected, since lead chloride dissolves in sodium acetate (ANTHON). The sil- ver chloride is washed with hot water. The lead is thrown down from the filtrate with hydrogen sulphide. If you desire to prevent the occasionally injurious influence of sodium ace- tate, great care must be given to the washing of the silver chloride. It is also well to reduce the weighed chloride by gentle ignition in a current of hydrogen, and to test the silver obtained for lead. Regarding the estimation of very small quantities of silver in lead, compare " Analysis of Refined Lead ' " in the Special Part. * Annal. d. Chem. u. Pharm., LXXXI, 128. f Compt. rend., LXX, 847 ; Zeitschr.f. analyt. G'hem., xiu, 349. 688 SEPARATION. [ 163. tf . The volumetric method ( 115, 5) is usally resorted to in mints to determine the silver in alloys. In presence of a mer- curic salt, sodium acetate is mixed with the fluid, immediately before the addition of the solution of chloride of sodium. In the East India mint the silver is separated and weighed as chloride.* b. MERCUROSUM FROM MERCURICTJM, COPPER, CADMIUM, AND LEAD. Mix the very dilute cold solution with hydrochloric acid 144 as long as a precipitate (mercurous chloride) forms ; allow this to deposit, filter on a weighed filter, dry at 100, and weigh. The filtrate contains the other metals. If you have to analyze a solid body, insoluble in water, either treat directly, in the cold, with dilute hydrochloric acid, or dissolve in highly dilute nitric acid, and mix the solution with a large quantity of water before proceeding to precipitate. Care must always be taken that the mode of solution is such as not to convert mercurous into mercuric compounds. If lead is present the washing of the mercurous chloride must be executed with special care with water of 60 70, till the filtrate ceases to be colored with hydrogen sulphide. As an additional security, it is well to test at last whether the weighed mercurous chloride leaves no lead sulphide behind on cautious ignition with sulphur in a stream of hydrogen. c. MERCUROSUM AND MERCURICUM FROM COPPER, CADMIUM, AND (but less well) FROM BISMUTH AND LEAD. If mercury is present as a mercuric compound, or partly 145 in a mercuric and partly in a mercurous compound, it is pre- cipitated according to 118, 2, by means of hydrochloric acid and phosphorous acid as mercurous chloride. The precipitate, particularly when bismuth is present, is first washed with water containing hydrochloric acid, then with pure water, till the washings are no longer colored with hydrogen sulphide (H. RosEf). In the presence of lead, the remarks in 144 must be attented to. * Ckem. CentralbL, 1872, 202. \Pogg. Annal., ex, 534. 163.] BASES OF GROUP V. 689 d. The method of separating lead from silver, copper, and bismuth by highly concentrating the nitric-acid solution, adding hydrochloric acid and alcohol, and washing the lead chloride with alcohol, is not to be recommended. It is inferior to 146 in accuracy. 2. Methods based upon the Insolubility of Lead Sulphate. LEAD FROM ALL OTHER METALS OF THE FIFTH GROUP. Mix the nitric acid solution with pure sulphuric acid in not 146 too slight excess, evaporate until the sulphuric acid begins to volatilize, allow the fluid to cool, add water (in which, if there is a sufficient quantity of free sulphuric acid present, the mer- curic and bismuth sulphates dissolve completely), and then filter the solution, which contains the other metals, without delay from the undissolved lead sulphate. If it is feared that the residue no longer contains enough free sulphuric acid, add some dilute acid to it before adding the water. Wash the precipitate with water containing sulphuric acid, displace the latter with alcohol, dry, and weigh ( 116, 3). Precipitate the other metals from the nitrate by hydrogen sulphide. If silver is present in any notable quantity, this method cannot be recommended, as the silver sulphate is not soluble enough. In this case you may follow ELIOT and STORER,* viz., mix the solution with ammonium nitrate, warm, precipitate the greater portion of the silver with ammonium chloride, evaporate the filtrate, remove the ammonium salts by ignition, and in the residue separate the small remainder of the silver from the lead with sulphuric acid as just directed. For the separation of lead from bismuth, on the above principle, H. ROSE t gives the following process as the best. If both oxides are in dilute nitric acid solution, as is usually the case, evaporate to small bulk, and add enough hydrochloric acid to dissolve all the bismuth; the lead separates partially as chloride. Should a portion of the clear fluid poured off become turbid on the addition of a drop of water, you must add some more hydro- * Proceedings of the American Academy of Arts and Sciences, Sept. 11, i860, p. 52 ; Zeitschr.f. analyl. Chem., i, 389. f Pogg. AnnaL, ex, 432. 690 SEPARATION. [ 163. chloric acid, till no permanent turbidity is produced unless several drops of water are added. The turbid fluids should all be returned, and the glasses rinsed with alcohol. Add now dilute sulphuric acid, allow to stand some time with stir- ring, add alcohol of 0'8 sp. gr., stir well, allow to settle for a long time, filter, wash the lead sulphate first with alco- hol mixed with a small quantity of hydrochloric acid, then with pure alcohol. Determine it after 116, 3. Mix the filtrate at once with a large quantity of water, and proceed with the precipitated basic bismuth chloride according to 120, 4. 3. Methods lased upon different deportment with Cyanide of Potassium (FRESENIUS and HAIDLEN*). a. LEAD AND BISMUTH FROM ALL OTHER METALS OF THE FIFTH GROUP. Mix the dilute solution with sodium carbonate in slight 14T excess, add solution of potassium cyanide (free from sulphide), heat gently for some time, filter and wash. On the filter you have lead and bismuth carbonates (containing alkali) ; the fil- trate contains the other metals as cyanides in combination with potassium cyanide. The method of effecting their further separation will be learnt from what follows. In very accurate analyses bear in mind that the filtrate generally contains traces of bismuth, which may be precipitated by ammonium sulphide. 1}. SILVER FROM MERCURICUM, COPPER, AND CADMIUM. Add to the solution, which, if it contains much free acid, 148 must previously be nearly neutralized with soda, potassium cyanide until the precipitate which forms at first is redissolved. The solution contains the cyanides of the metals in combina- tion with potassium cyanide as soluble double salts. Add dilute nitric acid in excess, which effects the decomposition of the double cyanides ; the insoluble silver cyanide precipitates permanently, whilst the mercuric cyanide remains in solution, and the cyanides of copper and cadmium redissolve in the excess of nitric acid. Treat the silver cyanide as directed 115, 3. If the filtrate contains only mercury and cadmium, precipitate at once with hydrogen sulphide, which completely * Annal. d. Chem. u. Pharm., XLIII, 129. 163.] BASES OF GROUP V. 691 throws down the sulphides of the two metals ; but if it con- tains copper, you must first heat with sulphuric acid, until the odor of hydrocyanic acid is no longer perceptible, and then precipitate with hydrogen sulphide ( 119, 3). c. COPPER FROM MERCURICUM AND CADMIUM. Mix the solution, as in J, with potassium cyanide until the 149 precipitate w r hich is first thrown down redissolves ; add some more potassium cyanide, then hydrogen sulphide water or ammonium sulphide, as long as a precipitate forms. The cadmium and mercury sulphides are completely thrown down, wliilst the copper remains in solution, as sulphide dissolved in potassium cyanide. Allow the precipitate to subside, decant repeatedly, treat the precipitate, for security, once more with solution of potassium cyanide, heat gently, filter, and wash the sulphides of the metals. To determine the copper in the filtrate, evaporate the latter, with addition of nitric and sul- puric acids, until there is no longer any odor of hydrocyanic acid, and then precipitate with hydrogen sulphide ( 119, 3). d. ALL THE METALS OF THE FIFTH GROUP FROM EACH OTHER. Mix the dilute solution with sodium carbonate, then with 150 potassium cyanide in excess, digest some time at a gentle heat, and filter,, On the filter you have lead carbonate and bismuth carbonate (containing alkali); separate the two metals by a suitable method. Add to the filtrate dilute nitric acid in excess, warm gently till the cuprous sulphocyanate first pre- cipitated with the silver cyanide has redissolved, and filter off the undissolved silver salt, w T hich is to be determined as directed 115, 3. Neutralize the filtrate with sodium car- bonate, add potassium cyanide, and pass hydrogen sulphide in excess. Add now some more potassium cyanide, to redissolve the copper sulphide which may have fallen down, and filter the fluid, which contains the whole of the copper, from the precipitated sulphides of mercury and cadmium. Determine the copper as directed in c, and separate the mercury and cad- mium as in 145 or 158. 692 SEPARATION. [ 163 e 4. Methods based on tfie Formation and Separation of insoluble Basic Salts. a. BISMUTH FROM COPPER, CADMIUM, AND MERCURICUM (also from the basic radicals of tlie first four groups, with the excep- tion of ferric iron). Precipitate the bismuth as basic chloride according to 120, 151 4, and throw down the copper, &c., in the nitrate by hydro- gen sulphide. Results thoroughly satisfactory (II. ROSE*). b. BISMUTH FROM LEAD AND CADMIUM. Separate the bismuth according to 120, 1, ) ARSEN- OUS ACID AKSENIC ACID. L SEPAKATION OF THE METALS OF THE SIXTH GROUP FROM THOSE OF THE FIRST FlVE GROUPS. 164. INDEX. (The numbers refer to those in the margin.) Gold from the metals of Groups III. I., 166, 171. Group IV., 166, 169, 171. " silver, 169, 188. " mercury, 169, 182. " lead, 169, 194. " copper, 169, 171. bismuth, 169, 171, 194. " cadmium, 169, 171. Platinum from the metals of Groups I. III. , 166, 172. Group IV., 166, 170, 172. silver, 170, 188. " mercury, 170, 172. lead, 170. copper, 170, 172. " bismuth, 170, 172. " cadmium, 170, 172. Tin from the metals of Groups I. and II., 166, 175, 181. " " Group III., 166, 175. " zinc, 166, 168, 173, 175. manganese, 166, 168, 175. nickel and cobalt, 166, 168, 173, 175, 180. " iron, 166, 168. silver, 167, 168, 173, 180. mercury, 167, 168, 173. " lead, 167, 168, 173. 180. " copper, 107, 168, 178, 175, 180. " bismuth, 167, 168. " cadmium, 167, 168, 173, 175. 700 SEPARATION. [ 164. Antimony from the metals of Groups I. and II., 166, 178. " " Group III., 166. zinc, 166, 168, 174. " manganese, 166, 168. nickel and cobalt, 166, 168, 174, 179, 180. iron, 166. 168, 178. silver, 167, 168, 174, 180. mercury, 167, 168, 174, 176, 189. lead, 167, 168, 174, 180, 191. t copper, 167, 168, 174, 178, 180, 192. bismuth, 167, 168. cadmium, 167, 168, 174. Arsenic from the metals of Group I., 166, 178, 184, 186, 187. " " II., 166, 177, 178, 184, 186, 187, 190. III., 166, 185, 186. zinc, 166, 168, 177, 183, 184, 186, 187. manganese, 166, 168, 177, 183, 185, 186, 187. nickel and cobalt, 166, 168, 177, 179, 180, 183, 184, 185, 186, 187. iron, 166, 168, 177, 178, 183, 185, 186. silver, 167, 168,177, 180, 186. mercury, 167, 168, 186, 189. lead, 167, 168, 177, 180, 183, 184, 186, 190. copper, 167, 168, 177, 178, 180, 183, 184, 185, 186, 192, 193. bismuth, 167, 168, 177, 186. cadmium, 167, 168, 177, 184, 185, 186. A. General Methods. 1 . Method based upon the Precipitation of Metals of the Sixth Group from Acid Solutions ~by Hydro- gen Sulphide. ALL METALS or THE SIXTH GROUP FROM THOSE OF THE FIRST FOUR GROUPS. Conduct into the acid* solution hydrogen sulphide in 166 excess, and filter off the precipitated sulphides (correspond- ing to the oxides of the sixth group). The points mentioned 127, a, f3, and y, must also be attended to here. As regards 7, antimony and tin are to be inserted between cadmium and mercury, in the order of metals there given. With respect to the particular conditions- required to secure the proper precipitation of certain metals of the sixth group, I refer to Section IY. I have to remark in addition : * Hydrochloric acid answers best as acidifying agent. 164.] METALS OF GKOUP VI. 701 a. That hydrogen sulphide fails to separate arsenic acid from .zinc, as, even in presence of a large excess of acid, the whole or at least a portion of the zinc precipitates with the arsenic (WOHLER). To secure the separation of the two bodies in a solution, the arsenic acid must first be converted into arsenous acid, by heating with sulphurous acid, before the hydrogen sulphide is conducted into the fluid. ft. That in presence of antimony, tartaric acid should be added, as otherwise the sulphide of antimony will contain chloride ; and that sulphide of antimony, when thrown down from a boiling solution by hydrogen sulphide, becomes black after a time, and so dense that it is deposited like sand, whereby the filtration and washing are much facilitated (S. P. SCHAFELER *). 2. Method based upon the Solubility of the Sulphides of the Metals of the Sixth Group in Sulphides of the Alkali Metals. a. THE METALS OF GROUP VI. (with the exception of Gold and Platinum) FROM THOSE OF GROUP V. Precipitate the acid solution with hydrogen sulphide, pay- 167 ing due attention to the directions given in Section IV. under the heads of the several metals, and also to the remarks in 166. The precipitate consists of the sulphides of the metals of Groups V. and VI. Wash, and treat at once with yellow ammonium sulphide in excess. (It is usually best to spread out the filter in a porcelain dish, add the ammonium sulphide, cover w r ith a large watch-glass, and place on a heated water- bath. Unnecessary exposure to air should be avoided.) Add some water, filter off the clear fluid, treat the residue again with ammonium sulphide, digest a short time, repeat the same operation, if necessary, a third and fourth time, filter, and wash the residuary sulpliides of Group V. with water contain- ing ammonium sulphide. If stannous sulphide is present, some flowers of sulphur must be added to the ammonium sul- phide, unless the latter be very yellow. In presence of copper, * BericJite der deutschen chem. Gesellsch., 1871, 279. I have myself confirmed these observations. 702 SEPARATION. [ 164 the sulphide of which is a little soluble in ammonium sulphide, sodium sulphide should be used instead. However, this sub- stitution can be made only in the absence of mercury, since the sulphide of that metal is soluble in sodium sulphide. Add to the alkaline nitrate, gradually, hydrochloric acid in small portions, until the acid predominates ; allow to subside, and then filter off the sulphides of the metals of the sixth group, which are mixed with sulphur. SCHNEIDER* states that he was unsuccessful in effecting complete separation of bismuth disulphide.and stannic sulphide by digestion with potassium sulphide, but did succeed by con- ducting hydrogen sulphide into the potassium-hydroxide solu- tion of bismuthic tartrate and stannous oxide (which decom- pose into bismuthous oxide and stannic oxide). If a solution contains much arsenic acid in presence of small quantities of copper, bismuth, &c., it is convenient to precipitate these metals (together with a very small amount of arsenous sulphide) by a brief treatment with hydrogen sul- phide. Filter, extract the precipitate with ammonium sulphide (or potassium sulphide), acidify the solution obtained, mix it with the former filtrate containing the principal quantity of the arsenic, and proceed to treat further with hydrogen sul- phide ( 127, 4, J). ~b. THE METALS OF GROUP VI. (with the exception of Gold and Platinum) FROM THOSE OF GROUPS IY. AND Y. a. Neutralize the solution with ammonia, add ammonium 168 chloride, if necessary, and then yellow ammonium sulphide in excess ; digest in a closed flask, for some time at a moderate heat, and then proceed as in 167. Kepeated digestion with fresh quantities of ammonium sulphide is indispensable. On the filter, you have the sulphides of the metals of Groups TV and V. Wash with water containing ammonium sulphide. In presence of nickel, this method offers peculiar difficul- ties ; traces of mercuric sulphide, too, are liable to pass into the filtrate. In presence of copper (and absence of mer- cury), soda and sodium sulphide are substituted for ammonia and ammonium sulphide. f * Annal. d. Chem, u. Pharm., ci. 64. t The accuracy of this method has been called in question by BLOXAM (Quart. Journ. Chem. Soc., v, 119). That chemist found that ammonium sulphide fails 164.] METALS OF GROUP VI. 703 ft. In the analysis of solid compounds (oxides or salts), it is in most cases preferable to fuse the substance with 3 parts of dry sodium carbonate and 3 of sulphur, in a covered porce- lain crucible. When the contents are completely fused, and the excess of sulphur is volatilized, the mass is allowed to cool, and then treated with water, which dissolves the sulphosalts of the metals of the sixth group, leaving the sulphides of Groups IV. and Y. undissolved. By this means, even ignited stannic oxide may be readily tested for iron, &c., and the amount of the admixture determined (H. ROSE). The solu- tion of- the sulphosalts is treated as in 167. In the presence of copper, traces of the sulphide may be dissolved with the sulphides of Group YI. Occasionally a little ferrous sulphide dissolves, coloring the solution green. In that case add some ammonium chloride, and digest till the solution has turned yellow. Instead of the mixture of sodium carbonate and sul- phur you. may also use already prepared hepar sulphuris, or, as FROHDE* says, you may fuse the substance with 4 or 5 parts of sodium thiosulphate. E. Special Methods. 1. Methods based upon the Insolubility of some Metals of the Sixth Group in Acids. a. GOLD FROM METALS OF GROUPS IY. AND Y. IN ALLOYS. a. Boil the alloy with pure nitric acid (not too concen- 169 .trated), or, according to circumstances, with hydrochloric acid. The other metals dissolve, the gold is left. The alloy must be reduced to filings, or rolled out into a thin sheet. If the alloy were treated with concentrated nitric acid, and at a tem- perature below boiling, a little gold might dissolve in conse- quence of the co-operation of nitrous acid. In the presence of silver and lead, this method is only applicable when they to separate small quantities of stannic sulphide from large quantities of mercuric sulphide or cadmium sulphide (1 : 100); and thai more especially the separation of copper from tin and antimony (also from arsenic) by this method is a failure, as nearly the whole of the tin remains with the copper. The latter statement I cannot confirm, for Mr. Lucius, in my laboratory, has succeeded in separating copper fFom tin by means of yellowish sodium sulphide completely; but it is indispensable to digest three or four times with sufficiently large quantities of the solvent, as stated in the text. * Zeilschr.f. analyt. C/tem., v, 405. 704 SEPARATION. [ 164 amount to more than 80 -per cent., since otherwise they are not completely dissolved. Alloys of silver and gold contain- ing less than 80 per cent, of silver are therefore fused with 3 parts of lead, before they are treated with nitric acid. The residuary gold is weighed ; but its purity must be ascertained, by dissolving in cold dilute nitrohydrochloric acid, not in con- centrated hot acid, as silver chloride also is soluble in the latter. In the presence of silver, a small quantity of its chloride is usually obtained here. If it can be weighed, it should be reduced and deducted. At the Mint Conference held at Vienna in 1857, the fol- lowing process was agreed upon for the mints in the several states of Germany. Add to 1 part of gold, supposed to be present, 2-J parts of pure silver ; wrap both the alloy and the silver in a paper together, and introduce into a cupel in which the requisite amount of lead is just fusing.* After the lead has been absorbed, f the button is flattened by hammering or rolling, then ignited and rolled. The rolls are treated first with nitric acid of 1*2 sp. gr., afterwards with nitric acid of 1*3 sp. gr., rinsed, ignited, and weighed.^ Even after boiling again with nitric a"cid of 1*3 sp. gr., they retain 0-75 to O'OOl of silver which will remain as chloride if the rolls are treated with cold dilute aqua regia (H. ROSSLEK, loc. cit.). (3. Heat the alloy (previously filed or rolled) in a capacious platinum dish with a mixture of 2 parts pure concentrated sulphuric acid and 1 part water, until the evolution of gas has ceased and the sulphuric acid begins to volatilize ; or fuse the alloy with potassium disulphate (H. ROSE). Separate the gold from the sulphates of the other metals, by treating the mass with water which should finally be boiling. It is advisable to repeat the operation with the separated gold, and ultimately *If the weighed sample, say 0'25 grm., contains 98-92$ gold, 3 grm. of lead are required; if 92-87'5, 4 grm.; if 87*5-75, 5 grm.; if 75-60, 6 grm.; if 60-35, 7 grm. ; if less than 35, 8 grm. f A small quantity of gold from one to three thousandths is always lost in cupellation. The loss increases .with the amount of lead, and is also depend- ent on the proportion of silver to gold. The more silver present the less is the loss of gold. In large buttons the loss is less than in small ones (H. ROSSLEK, Ding, poly 1. Journ , ccvi, 185; Zeitschr. f. analyt. Chem., xin/87). t Kunsl- und Gewerbeblatt f. Baiern, 1857, 151; Chem. CentralbL, 1857, 307; Polyt. CentralbL, 1857, 1151, 1471, 1639. 164.] METALS OF GKOTJP VI. 706 test the purity of the latter. In presence of lead this method is not good. y. The methods given in a. and fi may be united, i.e., the cupelled and thinly-rolled metal may be first warmed with nitric acid of 1*2 sp. gr., then thoroughly washed, the gold boiled 5 minutes with concentrated sulphuric acid, washed again, and ignited (MASCAZZINI, BUGATTI). I). PLATINUM FROM METALS OF GROUPS IY. AND Y. IN ALLOYS. The separation is effected by heating the alloy in filings 170 or foil with pure concentrated sulphuric acid, with addition of a little water, or by fusing with potassium disulphate (169, ft); but not with nitric acid, as platinum in alloys will, under cer- tain circumstances, dissolve in that acid. 2. Method based upon the Separation of Gold in the metallic state. GOLD FROM ALL METALS OF GROUPS I. Y., with the excep- tion of LEAD, MERCURY, AND SILVER. Precipitate the hydrochloric acid solution with oxalic acid 171 as directed 123 J, y, or with ferrous sulphate, 123, , a, .and filter off the gold when it has completely separated. Take care to add a sufficient quantity of hydrochloric acid after the reduction to insure solution of any oxalates. In the presence of copper the addition of hydrochloric acid does not suffice, since the coprecipitated cupric oxalate will dissolve with diffi- culty in this acid. E. PURGOTTI* recommends in this case, after precipitation, adding potash cautiously to the boiling hot fluid till it is neutral, and then if necessary some normal potassium oxalate. Double oxalate of copper and potash will be formed which dissolves with a blue color. The gold after washing will now be pure. 3. Method based upon the Precipitation of Pla- ti until' <(* Potassium Platinic, or Ammonium Platinic Chi <> id'. PLATINUM FROM THE METALS OF GROUPS IY. AND V., with the exception of MERCURY IN MERCUROUS COMPOUNDS, LEAD, AND SILVER. Precipitate the platinum with potassium chloride or 172 * Zeitschr. f. analyL Chem., ix, 128. 706 SEPARATION. [ 164. ammonium chloride as directed 124, and wash the precipi- tate thoroughly with alcohol. The platinum prepared from the precipitated ammonium- or potassium salt is to be tested after being weighed, to see whether it yields any metal (especially iron) to fusing potassium disulphate. 4. Methods based upon the Separation of Oxides insoluble in Nitric Acid. a. TIN FROM METALS OF GROUPS IY. AND Y. (not from Bismuth, Iron, or Manganese*) IN ALLOYS. Treat the finely divided alloy, or the metallic powder 17$ obtained by reducing the oxides in a stream of hydrogen with nitric acid, as directed 126, 1, a. The filtrate contains the other metals as nitrates. As stannic oxide is liable to retain traces of copper arid lead and iron, you must, in an accurate analysis, test an aliquot part of it for these bodies, and determine their amount as directed 168, /?. BRUNNER recommends the following course of proceeding,, by which the presence of copper in the tin may be effectually guarded against. Dissolve the alloy in a mixture of 1 part of nitric acid, 4 parts of hydrochloric acid, and 5 parts of water ; dilute the solution largely with water, and heat gently. Add crystals of sodium carbonate until a distinct precipitate has- formed, and boil. (In presence of copper, the precipitate must, in this operation, change from its original bluish-green to a brown or black tint.) When the fluid has been in ebulli- tion some 10 or 15 minutes, allow it to cool, and then add nitric acid, drop by drop, until the reaction is distinctly acid ; digest ilie precipitate for several hours, when it should have acquired a pure white color. The stannic oxide thus obtained is free from copper ; but it may contain some iron, which can be removed as directed in 168, (3. Before the stannic oxide can be considered pure, it must be tested also for silicic acid, as it frequently contains traces of this substance. To this end, an aliquot part is fused in plati- * If the alloy of tin contains bismuth or manganese, there remains with the stannic oxide, bismuth trioxide or manganese sesquioxide, which cannot be extracted by nitric acid; if it contains iron, on the contrary, some stannic oxide always dissolves with the iron, and cannot be separated even by repeated evapo- ration (H. ROSE, Pogg. Annal., cxn, 1G9, 170, 172) 164.] METALS OF GKOUP VI. 707 num with 3 4 parts of sodium and potassium carbonate, the fused mass boiled with water, and the solution filtered ; hydro- chloric acid is then added to the filtrate, and, should silicic acid separate, the fluid is filtered off from this substance. The tin is then precipitated by hydrogen sulphide, and the silicic acid still remaining in the filtrate is determined in the usual way ( 140). If hydrochloric acid has produced a precipitate of silicic acid, the last filtration is effected on the same filter (KHITTEL*). 1}. ANTIMONY FROM THE METALS OF GROUPS IV. AND Y. IN ALLOYS (not from Bismuth, Iron and Manganese). Proceed as in 173, filter off the precipitate, and convert it 174 by ignition into antimony tetroxide according to 125, 2. Results only approximate, as a little antimony dissolves. Alloys of antimony and lead, containing the former metal in excess, should be previously fused with a weighed quantity of pure lead (YARRENTRAppf). 5. Methods based on the Precipitation of Tin in Stannic Salts ~by Normal Salts (e.g., Sodium Sulphate) or by Sulphuric Acid. TIN FROM THE METALS OF GROUPS I., II., Ill, ; ALSO FROM MANGANESE, ZINC, NICKEL AND COBALT, COPPER, CADMIUM (GOLD). Precipitate the hydrochloric acid solution, which must 175 contain the tin entirely as stannic chloride, according to 126, 1, , by ammonium nitrate or sodium sulphate (LOWENTHAL), or by sulphuric acid, which, H. ROSE says, answers equally well. Alloys are always treated as follows : First, oxidize by digestion with nitric acid ; when no more action takes place, evaporate the greater portion of the nitric acid in a porcelain dish, moisten the mass with strong hydrochloric acid, and after half an hour add water, in which the nietastannic chloride and the other chlorides dissolve. Alloys of tin and gold are dis- solved in aqua regia, the excess of acid evaporated, and the solution diluted with much water, before precipitating with sulphuric acid. It must be remembered that in this process any phosphoric * Chem. CentralbL, 1857, 929. \ Dingier' s polyt. Journ., CLVIII, 316. 708 SEPARATION. [ 164. acid that may be present is precipitated entirely or partially with the tin. After the precipitate has been well washed by decantation, LOWENTHAL recommends to boil with a mixture of 1 part nitric acid (sp. gr. 1-2) and 9 parts water, then to transfer to the - filter, and wash thoroughly. Results very satisfactory. If the fluid contains a ferric salt, a portion of the iron always falls down with the tin. Hence the stannic oxide must be tested for iron according to 168, /?, which, if present, must be determined and deducted. 6. Method based on the Insolubility of Mercuric Sulphide in Hydrochloric Add. MERCURY FROM ANTIMONY. Digest the precipitated sulphides with moderately strong 17(5 hydrochloric acid in a distilling apparatus. The sulphide of antimony dissolves, while the mercuric sulphide remains behind. Expel all the hydrogen sulphide, then add tartaric acid, dilute, filter, mix the filtrate with the distillate which contains a little antimony, and precipitate with hydrogen sulphide. The mercuric sulphide may be weighed as such (Fu. FIELD*). 7. Methods based upon the Conversion of Arsenic and Antimony into Alkali Ar senate and Antimonate. a. ARSENIC FROM THE METALS OF GROUPS II. , IY., AND Y. If you have to do with arsenites or arsenates, fuse with 3 177 parts of sodium and potassium carbonates and 1 part of potas- sium nitrate ; if an alloy has to be analyzed it is fused with 3 parts of sodium carbonate and 3 parts of potassium nitrate. In either case the residue is boiled with water, and the solution, which contains the arsenates of the alkalies, filtered from the undissolved oxides or carbonates. The arsenic acid is deter- mined in the filtrate as directed 127, 2. If the quantity of arsenic is only small, a platinum crucible may be used, other- wise a porcelain crucible must be used, as platinum would be seriously injured. In the latter case, bear in mind that the fused mass is contaminated with silicic acid and alumina, If the alloy contains much arsenic a small quantity may be readily lost by volatilization, even though the operation be cautiously * Quart. Journ. Chem. Soc., xu, 32. 164.] METALS OF GROUP VI. 709 c< ni ducted. In such a case, therefore, it is better first to oxidize with nitric acid, then to evaporate, and to fuse the residue as above directed with sodium carbonate and potassium nitrate. b. ARSENIC AND ANTIMONY FROM COPPER AND IRON, especially in ores containing sulphur. Diffuse the very finely pulverized mineral through pure 178 solution of potassa, and conduct chlorine into the fluid (comp. p. 467). The iron and copper separate as oxides, the solution contains sulphate, arsenate, and antimonate of potassium (RivoT, BEUDANT, and DAGUIN*). c. ARSENIC AND ANTIMONY FROM COBALT AND NICKEL. Dilute the nitric acid solution, add a large excess of potassa, 179 heat gently, and conduct chlorine into the fluid until the pre- cipitate is black. The solution contains the whole of the arsenic and antimony, the precipitate the nickel and cobalt as sesquioxides (RivoT, BEUDANT, and DAGUIN, loc. cit.) 8. Methods based upon the Volatility of certain Chlorides or Metals. a. TIN, ANTIMONY, ARSENIC FROM COPPER, SILVER, LEAD, COBALT, NICKEL. Treat the sulphides with a stream of perfectly dry chlorine, 180 proceeding exactly as directed in 160. In presence of anti- mony, fill E and F (Fig. 116) with a solution of tartaric acid in water, mixed with hydrochloric acid. The metals may be also separated by this method in alloys. The alloy must be very finely divided. Arsenical alloys are only very slowly decomposed in this way. In separating arsenic and copper the temperature must not exceed 200, and chlorine water should be put into the receiver (PARNELLf). If tin and copper are separated in this manner, according to the experience of H. ROSE,;); a small trace of tin remains with the copper chloride. b. STANNIC OXIDE, ANTLMONIOUS OXIDE (AND ALSO ANTTMONIC ACID). ARSENOUS AND ARSENIC ACIDS, FROM ALKALIES AND ALKALINE EARTHS. Mix the solid compound with 5 parts of pure ammonium 181 chloride in powder, in a porcelain crucible, cover this with a * Compt. rend., 1853, 835; Journ. f. prakt. C7iem.,-Lxi, 133. f Chem. News, xxi, 133. j Pogg. Annal, cxn, 169. 710 SEPARATION. [ 164. concave platinum lid, on which some ammonium chloride is sprinkled, and ignite gently until all ammonium chloride is driven off ; mix the contents of the crucible with a fresh por- tion of that salt, and repeat the operation until the weight remains constant. In this process, the chlorides of tin, anti- mony, and arsenic escape, leaving the chlorides of the alkalies and alkali-earth metals. The decomposition proceeds most rapidly with alkali salts. With regard to salts of alkali-earth metals it is to be observed that those which contain antimonic acid or stannic oxide are generally decomposed completely by a double ignition with ammonium chloride (magnesium alone cannot be separated perfectly from antimonic acid by this method). The arsenates of the alkali-earth metals are the most troublesome to decompose ; barium, stroiiium, and cal- cium salts usually require to be subjected 5 times to the opera- tion, before they are free from arsenic, and magnesium arsenate it is impossible thoroughly to decompose in this way (H. ROSE*). According to SALKowsKif barium arsenate may be converted into chloride quite free from arsenic by one ignition with ammonium chloride ; however calcium arsenate was found to leave a residue containing arsenic acid even after six igni- tions with ammonium chloride. c. MERCURY FROM GOLD (SILVER, AND GENERALLY FROM THE NON-VOLATILE METALS). Heat the weighed alloy in a porcelain crucible, ignite till 182 the weight is constant, and determine the mercury from the loss. If it is desired to estimate it directly, the apparatus (Fig. 88) may be used. In cases where the separation of mer- cury from metals that oxidize on ignition in the air is to be effected by this method, the operation must be conducted in an atmosphere of hydrogen (Fig. 83). 50). 9. Methods based on the Volatility of Arsenous Sulphide. ARSENIC ACID FROM THE OXIDES OF MANGANESE, IRON, ZIMC, COPPER, NICKEL, COBALT (NOT so WELL FROM OXIDE oir. LEAD, AND NOT FROM OXIDES OF SILVER, ALUMINIUM, OR MAG- NESIUM). Mix the arsenic-acid compound (no matter whether it has 183 1. Annal., LXXIII 582; LXXIV, 578; cxu, 173. Journ. f. prakl. Chem , civ, 138. 164.] METALS OF GROUP VI. 711 been air-dried or gently ignited) with sulphur, and ignite under a good draught in an atmosphere of hydrogen (Fig. 83) ; the perforated lid must in this case be of porcelain ; platinum would not answer). The whole of the arsenic vola- tilizes, the sulphides of manganese, iron, zinc, lead, and copper remain behind ; they may be weighed directly. After weigh- ing, add a fresh quantity of sulphur to the residue, ignite as before, and weigh again ; repeat this operation until the weight remains constant. Usually, if the compound was intimately mixed with the sulphur, the conversion of the arsenate into sulphide is complete after the first ignition. Results very good. In separating nickel the analyst will remember that the residue cannot be weighed directly, since it does not possess a constant composition ; hence the ignition in hydrogen may be saved ; nickel arsenate loses all its arsenic on being simply mixed with sulphur and heated. The heat should be moderate and continued till no more red sulphide of arsenic is visible on the inside of the porcelain crucible. It is advisable to repeat the operation. The separation of arsenic from cobalt cannot be completely effected in this manner even by repeated treat- ment with sulphur, but it can be effected by oxidizing the resi- due with nitric acid, evaporating to dryness, mixing with sul- phur, and reigniting. Smaltine and cobaltine must be treated in the same manner (II. ROSE*). 1 should not forget to men- tion that Ei5ELMEN,f a long while ago, noticed the separation of arsenic acid from sesquioxide of iron by ignition in a stream of hydrogen sulphide. 10. Method based upon the Separation of Arsenic as Mercurous Arsenate. ARSENIC ACID FROM ALKALIES, ALKALI EARTHS, ZINC, COBALT, NICKEL, LEAD, COPPER, AND CADMIUM. Proceed exactly as in separating phosphoric acid by mer- 184 cury ( 134, 5, y). The arsenic acid can not be determined in the insoluble residue as is done with phosphoric acid. If it is to be estimated directly, it must be separated from the * Zetischr. /. analyt. Chem., i, 413. f Annal. de Chim. et de Phys. (3), xxv, 98. 712 SEPARATION. [ 164 mercurous mercury by one of the methods given in this sec- tion. Treat the filtrate as in 135, k, a (H. KOSE). 1 1 . Method 'based upon the Separation of Arsenic as Ammonium Magnesium Arsenate. ARSENIC ACID FROM COPPER, CADMIUM, FERRIC IRON, MAN- GANESE, NICKEL, COBALT, ALUMINIUM. Mix the hydrochloric acid solution, which must contain 18& the whole of the arsenic in the form of arsenic acid, with enough tartaric acid to prevent precipitation by ammonia, pre- cipitate the arsenic acid according to 127, 2, as ammonium magnesium arsenate, allow to settle, filter, wash once with a mixture of 3 parts water and 1 part ammonia, redissolve in hydrochloric acid, add a very minute quantity of tartaric acid, supersaturate again with ammonia, add some more magnesium chloride and ammonium chloride, allow to deposit, and deter- mine the now pure precipitate according to 127, 2. In the filtrate the "bases of Groups IV. and V. may be precipitated by ammonium sulphide ; if aluminium is present, evaporate the filtrate from the sulphides with addition of sodium carbonate and a little nitre to dryness, fuse, and estimate the aluminium in the residue. The method is more adapted to the separation of rather large than of very small quantities of arsenic from the above-named metals, since in the case of small quantities the minute portions of ammonium magnesium arsenate that remain in solution may exercise a considerable influence on the accu- racy of the result. 12. Method based upon the Separation of Arsenic as Ammonium Arsenio-molybdate. ARSENIC ACID FROM ALL METALS OF GROUPS I. Y. Separate the arsenic acid as directed in 127, 2, b ; long 18$ continued heating at 100 is indispensable. The determination of the basic metals is most conveniently effected in a special portion. 164.] METALS OF GROUP VI. 713 13. Method based upon the Insolubility of Ferric Ar senate. AKSENIC ACID FROM THE METALS OF GROUPS I. AND II., AND FROM ZlNC, MANGANESE, NlCKEL, AND COBALT. Mix the hydrochloric-acid solution with a sufficient quantity 187 of pure ferric chloride, neutralize the greater part of the free acid with sodium carbonate, and precipitate the iron and arse- nic acid together with barium carbonate in the cold or with sodium acetate at a boiling heat. The precipitate should be so basic as to have a brownish-red color. The method is espe- cially suitable for the separation of arsenic acid when its esti- mation is not required. However, the precipitate may be dis- solved in hydrochloric acid and the arsenic determined by precipitation with hydrogen sulphide. 14. Methods based upon the Insolubility of some Chlorides. a. SILVER FROM GOLD. Treat the alloy with cold dilute nitrohydrochloric acid, 188 dilute, and filter the solution of auric chloride from the undis- solved silver chloride. This method is applicable only if the alloy contains less than 15 per cent, of silver ; for if it contains a larger proportion, the silver chloride which forms protects the undecomposed part from the action of the acid. In the same way silver may be separated also from platinum. b. MERCURY FROM THE OXYGEN COMPOUNDS OF ARSENIC AND ANTIMONY. Precipitate the mercury from the hydrochloric solution by 189 means of phosphorous acid as mercurous chloride ( 118, 2). The tartaric acid, which in the presence of antimony must be added, does not interfere with the reaction (H. ROSE*). 15. Methods based upon the Insolubility of certain Sulphates in Water or Alcohol. a. ARSENIC ACID FROM BARIUM, STRONTIUM, CALCIUM, AND LEAD. Proceed as for the separation of phosphoric acid from the 190 same metals ( 135, I). The compounds of these basic radicals with arsenous acid are first converted into arsenates, before * Pogg. Annal., ex, 536. 714 SEPARATION. [ 164. the sulphuric acid is added ; this conversion is effected by heating the hydrochloric acid solution with potassium chlo- rate or by means of bromine. b. ANTIMONY FROM LEAD. Treat the alloy with a mixture of nitric and tartaric acids. 191 The solution of both metals takes place rapidly and with ease. Precipitate the greater part of the lead as sulphate ( 116, 3), filter, precipitate with hydrogen sulphide, and treat the sul- phides according to 168, with ammonium sulphide, in order to separate the antimony from the lead left unprecipitated by the sulphuric acid (A. STRENQ *). 16. Method based upon the Separation of Copper as Cuprous Sulphocyanate. COPPER FROM ARSENIC AND ANTIMONY. From the properly prepared solution precipitate the cop- 192 per by 119, 3, 5, as cuprous sulphocyanate, allow to settle, filter, wash with water containing ammonium nitrate (to pre- Tent the washings being muddy), and determine antimony and arsenic in the filtrate, preciptating first with hydrogen sulphide. Results good. The following method, depending upon the precipitation of the copper as an iodide, is not good. Dissolve in nitric or sulphuric acid, taking care to have a slight excess of acid, dilute with pure water (if antimony is present use water con- taining some tartaric acid), add sulphurous acid, and precipi- tate the copper with potassium iodide as cuprous iodide. Arsenic and antimony remain in solution (FLAJOLOT). Results are approximate, because the liquid retains some cuprous iodide in solution in consequence of the presence of the sulphurous acid. It is impracticable to use stannous chloride for reducing the copper as recommended by FLEISCHER, f because then the separation of the tin from the arsenic and antimony would be too difficult. 17. Method based upon the Precipitation of Cop- per as an Oxalate. COPPEE FBOM AKSENIC. Add ammonia to the nitric-acid solution until the blue 193 precipitate no longer redissolves, and then effect solution by * Dingl. polyt. Journ., CLI, 389. -\Zeitschr.f. analyt. Chem., ix, 256. 165.] METALS OF GROUP VI. 715 an excess of ammonium oxalate. Now add very cautiously hydrochloric or nitric acid to acid reaction, and allow to stand. The copper separates almost completely as oxalate, which is converted into oxide by ignition in the air. Make the filtrate animoniacal and precipitate the trace of copper still in solution by adding a few drops of ammonium-sul- phide solution (F. FIELD *). 18. Method based upon the different Deportment with Potassium Cyanide. GOLD FEOM LEAD AND BISMUTH. These metals may be separated in solution by potassium 194 cyanide in the same way in which the separation of mercury from lead and bismuth is effected (see 147). The solution of the double cyanide of gold and potassium is decomposed by boiling with aqua regia, and, after expulsion of the hydro- cyanic acid, the gold determined by one of the methods given in 123. II. SEPAEATION or THE METALS OF THE SIXTH GEOUP FEOM EACH OTHEE. 165. INDEX. (The numbers refer to those in the margin.) Platinum from gold, 195, 214, 215. tin, antimony, and arsenic, 196. Gold from platinum, 195, 214, 215. tin, 196, 213. " antimony and arsenic, 196. Tin from platinum, 196. gold, 175, 196, 213. arsenic, 199, 206, 207, 208, 211, 212, 216, 217. antimony, 197, 201, 208, 209, 210, 212, 216. Tin in staunous from tin in stannic compounds, 221. Antimony from platinum and gold, 196. arsenic, 200, 201, 202, 203, 204, 206, 207. tin, 197, 201, 208, 209, 210, 212, 216. Antimony of antimonous compounds from antimonic acid, 220. Arsenic from platinum and gold, 196. tin, 199, 206, 207, 208, 211, 212, 216, 217. antimony, 200, 201, 202, 203, 204, 206, 207, 218. Arsenous acid from arsenic acid, 198, 205, 219. * Chem. Gaz., 1857, 313; Journ.f. prakt. Chem., LXXU, 183. 716 SEPAEATION. [ 165. 1. Method based upon the Precipitation of Plati- num as Potassium Platinic Chloride. PLATINUM FROM GOLD. Precipitate from the solution of the chlorides the plati- 195 num as directed 124, , and determine the gold in the filtrate as directed 123, I. 2. Methods based upon the Volatility of the Chlo- rides of the inferior Metals. a. PLATINUM AND GOLD FROM TIN, ANTIMONY, AND ARSENIC. Heat the finely divided alloy or the sulphides in a stream 196 of chlorine gas. Gold and platinum are left, the chlorides of the other metals volatilize (compare 160). 1). ANTIMONY FROM TIN. The tin should be present wholly as a stannous salt. 197 Precipitate with hydrogen sulphide, filter (preferably through an asbestos filtering tube), dry the precipitate, and pass through it a current of dry hydrochloric gas at the ordinary tempera- ture. The sulphides are converted into the corresponding chlorides ; the chloride of antimony alone escapes, and may be received in water. Dissolve the residual stannous chloride in water containing hydrochloric acid, and estimate the tin according to 126 (C. TOOKEY*). The method can only be used in rare cases, as it is difficult to obtain a precipitate quite free from stannic sulphide. c. ARSENOUS ACID FROM ARSENIC ACID. The amount of substance taken should not contain more 198 than 0*2 grm. arsenous acid. Heat with 45 grm. sodium chloride, 135 grm. sulphuric acid (free from arsenic) of 1*61 sp. gr., and 30 grm. water in a tubulated retort containing a spiral of platinum, and provided with a thermometer. The temperature should rise to about 125. In order to condense the arsenous chloride in the products of distillation, a LIEBIG'S condenser is connected with the retort ; a tubulated receiver is connected with the condenser ; a TJ-tube is connected with the receiver, and finally a calcium chloride tube containing fragments of glass moistened with weak soda solution is fixed * Jour ii. Chem. Koc., xv, 462. 165.] METALS OF GROUP VI. 717 upright in the exit end of the U-tube. In the receiver and U-tube water is placed. I can recommend the apparatus shown in Fig. 7" 8. At the end of the operation rinse the calcium chloride tube, and mix with the contents of the receiver. Determine the arsenic in the distillate according to 127, 4, a, in the residue according to 127, 4, I. The sul- phide obtained from the former corresponds to the arsenous acid, from the latter to the arsenic acid. Results satisfactory (RiECKHEU*). If the substance given is a dilute fluid, render slightly alkaline with sodium carbonate, and concentrate to .about 20 c.c., finally in a tubulated retort. 3. Methods based upon the Volatility of Arsenic and Arsenous Sulphide. a. ARSENIC FROM TIN (H. ROSE). Convert into sulphides or oxides, dry at 100, and heat a 199 weighed portion with addition of a little sulphur in a bulb- tube, gently at first, but gradually more strongly, conducting a stream of dry hydrogen sulphide gas through the tube during the operation. Sulphur and arsenous sulphide vola- tilize ; sulphide of tin is left. The arsenous sulphide is received in U-tubes containing dilute ammonia, which are connected with the bulb-tube in the manner described in 160. "When upon continued application of heat no sign of further sublimation is observed in the colder part of the bulb-tube, drive off the sublimate which has collected in the bulb, allow the tube to cool, and then cut it off above the coating. Divide the separated portion of the tube into pieces, and heat these with a little solution of soda until the sublimate is dissolved; unite the solution with the amrnoniacal fluid in the receivers, add hydrochloric acid, then, without filtering, potassium chlorate, and heat gently until the arsenious sulphide is com- pletely dissolved. Filter from the sulphur, and determine the arsenic acid as directed 127, 2. The quantity of tin cannot be calculated at once from the blackish-brown sulphide of tin in the bulb, since this contains more sulphur than SnS. It is therefore weighed, and the tin determined in a weighed por- tion of it, by converting it into stannic oxide, which is effected . by moistening with nitric acid, and roasting ( 12t>, 1, c\ * Plniri/i. ( i ntntllmlle, XI, 1)2. 718 SEPARATION. [ 165. Tin and arsenic in alloys are more conveniently converted into oxides by cautious treatment with nitric acid. If, how- ever, it is wished to convert them into sulphides, this may readily be effected by heating 1 part of the finely divided alloy with 5 parts of sodium carbonate and 5 parts of sulphur, in a covered porcelain crucible until the mass is in a state of calm fusion. It is then dissolved in water, the solution filtered from the ferrous sulphide, &c., which may possibly have formed, and then precipitated with hydrochloric acid. If the tin only in the alloy is to be estimated directly, while the arsenic is to be found from the difference, convert as above directed into sulphides or oxides, mix with sulphur and ignite in a porcelain crucible with perforated cover in a stream of hydrogen sulphide. The residual arsenic-free stannous sulphide is to be converted into stannic oxide and weighed as such. b. ARSENIC FROM ANTIMONY IN ALLOYS. Heat a weighed portion of the substance with two parts 200 of sodium carbonate and two parts of potassium cyanide in a bulb-tube through which dry carbonic acid is being trans- mitted. Heat at first gently, then more and more strongly, and until no more arsenic volatilizes. (Take great care not to inhale the escaping fumes. It is advisable to insert the open end of the tube into a flask in which sublimed arsenic will condense.) After cooling, treat the contents of the bulb first with a mixture of equal volumes of alcohol and water, then with water alone, and finally weigh the residual antimony. The arsenic is found from the loss. The results are only approximate. If it is desired to fuse the alloy, itself not under a slag, in a current of carbonic-acid gas, the heating must be very carefully done, otherwise much antimony will volatilize. H. EOSE recommends the latter process. 4. Methods based upon the Insolubility of Sodium Metantimonate . a. ANTIMONY FROM TIN AND ARSENIC (H. KOSE). If the substance is metallic, oxidize the finely divided 201 weighed sample in a porcelain crucible with nitric acid of 165.] METALS OF GKOUP VI. 719 1-4 sp. gr., adding the acid gradually. Dry the mass on the water-bath, transfer to a silver crucible, rinsing the last par- ticles adhering to the porcelain into the silver crucible with solution of soda, dry again, add eight times the bulk of the mass of solid sodium hydroxide, and fuse for some time. Allow the mass to cool, and then treat with hot water until the undissolved residue presents the appearance of a fine powder ; dilute with some water, and add one-third the volume of alcohol of 0*83 sp. gr. Allow the mixture to stand for 24: hours, with frequent stirring; then filter, transfer the last adhering particles from the crucible to the filter by rinsing with dilute alcohol (1 vol. alcohol to 3 vol. water), and wash the undissolved residue on the filter, first with alcohol diluted with twice its volume of water, then with a mixture of equal volumes of alcohol and water, and finally with a mixture of 3 vol. alcohol and 1 vol. water. Add to each of the alcoholic fluids used for washing a few drops of solution of sodium carbonate. Continue the washing until the color of a portion of the fluid running off remains unaltered upon being acidi- fied with hydrochloric acid and mixed with hydrogen- sul- phide water. Rinse the sodium metantimonate from the filter, wash the latter with a mixture of hydrochloric and tartaric acids, dis- solve the metantimonate in this mixture, precipitate with hydrogen sulphide, and determine the antimony as directed in 125, 1. In presence of much tin it is well to fuse the metantimonate again with caustic soda, etc. To the filtrate, which contains the tin and arsenic, add hydrochloric acid, which produces a precipitate of stannic arsenate; conduct now into the unfiltered fluid hydrogen sulphide for some time, allow the mixture to stand at rest until the odor of that gas has almost completely gone off, and separate the weighed sulphides of the metals which contain free sulphur, as in 199. If the substance contains only antimony and arsenic, the alcoholic filtrate is heated, with repeated addition of water, until it scarcely retains the odor of alcohol ; hydrochloric acid is then added, and the arsenic acid determined as magnesium pyroarsenate ( 127, 2), or as arsenous sulphide ( 127, 4, 5). 720 SEPARATION. [ 165. ~b. Small quantities of the sulphides of arsenic and anti- 202 mony mixed with sulphur are often obtained in mineral analysis. The two metals may in this case be conveniently separated as follows: Exhaust the precipitate with carbon disulphide, oxidize with chlorine-free red fuming nitric acid, evaporate the solution nearly to dryness; mix the residue with a copious excess of sodium carbonate, add some sodium nitrate, and treat the fused mass as given in 201, a. If, on the other hand, you have a mixture of sulphides of tin and antimony to analyze, oxidize it with nitric acid of 1*5 sp. gr., and treat the residue obtained on evaporation as given in 201, a. . c. DETERMINATION OF AKSENIC SULPHIDE IN COMMERCIAL ANTIMONY SULPHIDE (WACKENRODER). Deflagrate 20 grm. of the finely triturated antimony sul- 203 phide with 40 grm. sodium nitrate and 20 grm. sodium car- bonate by projecting the mixture in portions into a red-hot Hessian crucible, then extract the strongly ignited mass by repeated treatment with water, acidulate the filtrate with hydrochloric acid, add some sulphurous acid, and precipitate the arsenic together with a small part of the antimony by means of hydrogen sulphide. Digest the still moist precipi- tate with ammonium carbonate, filter, acidulate the filtrate, conduct in hydrogen sulphide, and determine the arsenic as arsenic sulphide according to 127, 4. 5. Methods "based upon the Precipitation of Arsenic as Ammonium Magnesium Ar senate. a. ARSENIC -FROM ANTIMONY. Oxidize the metals or sulphides with nitrohydrochloric 204 acid, with hydrochloric acid and potassium chlorate, with bromine dissolved in hydrochloric acid, or with chlorine in alkaline solution, page 568, &; add tartaric acid, a large quantity of ammonium chloride, and then ammonia in excess. (Should the addition of the latter reagent produce a precipi- tate, this is a proof that an insufficient quantity of ammo- nium chloride or of tartaric acid has been used, which error must be corrected before proceeding with the analysis.) 165.] METALS OF GROUP VI. 721 Then precipitate the arsenic acid as directed in 127, 2, and determine the antimony in the filtrate as directed in 125, 1. As basic magnesium tartrate might precipitate with the ammonium magnesium arsenate, the precipitate should always, after slight washing, be redissolved in hydrochloric acid, and reprecipitated with ammonia with addition of a little magnesia mixture. An excellent method. b. ARSEXOUS ACID FROM ARSENIC ACID. Mix the sufficiently dilute solution with a large quantity 205 of ammonium chloride, precipitate the arsenic acid as directed in 127, 2, and determine the arsenous acid in the filtrate by precipitation with hydrogen sulphide (127, 4). LUDWIG* has observed that if the solution is too concentrated, magne- sium arsenite falls clown with the ammonium magnesium arsenate, hence it is necessary to dissolve the weighed magne- sium precipitate in hydrochloric acid and test the solution with hydrogen sulphide. The presence of arsenous acid will be betrayed by the immediate formation of a precipitate. c. TIN AND ANTIMONY FROM ARSENIC ACID. LKXSSEN! separated tin from arsenic acid with good 206 results by digesting the oxides obtained by oxidation with nitric acid with ammonia and yellow ammonium sulphide, and precipitating the arsenic acid from the clear solution accord- ing to 127, 2, as ammonium magnesium arsenate. On acidify- ing the filtrate the tin separates as stannic sulphide. The method can only give good results when the whole of the arsenic was present as arsenic acid before the addition of ammonium sulphide, for the arsenic in a solution of arsenous acid in yellow ammonium sulphide is not thrown down by magnesia mixture. The method also answers for separating antimony from arsenic. * Archivfur Pharm., xcvn, 24. \ Annal. d. Chem. u. Pharm., cxiv, 116. 722 SEPARATION. [ 165. 6. Methods based on the different behavior of the freshly Precipitated Sulphides towards Solution of Potassium Hydrogen Sulphite or Oxalic Acid. a. ARSENIC FROM ANTIMONY AND TIN (BUNSEN*). If freshly precipitated arsenous sulphide is digested with 207 sulphurous acid and potassium sulphite, the precipitate is dis- solved ; on boiling, the fluid becomes turbid from separated sulphur, which turbidity for the most part disappears again on long boiling. The fluid contains, after expulsion of the sul- phurous acid, potassium arsenite and thiosulphate. The sul- phides of antimony and tin do not exhibit this reaction. Both therefore maybe separated from arsenous sulphide by diluting the solution of the three sulphides in potassium sulphide to about 500 c.c. and precipitating with a large excess (about a litre) of saturated aqueous sulphurous acid, digesting the whole for some time in a water-bath, and then boiling till one-third of the water and the whole of the sulphurous acid are expelled and the sulphur has disappeared ; this will take about an hour and a half. The residuary sulphide of antimony or tin is arsenic- free, the filtrate contains the whole of the arsenic and maybe immediately precipitated with hydrogen sulphide. BUNSEN determines the arsenic by oxidizing the dried sulphide together with the filter with fuming nitric acid, diluting the solution a little, warming gently with a little potassium chlorate (in order to oxidize more fully the substances formed from the paper), and finally precipitating as ammonium magnesium arsenate. With regard to the separation of stannic sulphide from the solution of potassium arsenite, it is to be observed that the stannic sulphide must be washed with concentrated solution of sodium chloride, as, if water w r ere used, the fluid would run through turbid. As soon as the precipitate is thoroughly washed with the sodium chloride, the latter is displaced by solution of ammonium acetate, containing a slight excess of acetic acid. These last washings must not be added to the first, as the ammonium acetate hinders the complete precipita tion of the arsenous acid by hydrogen sulphide, * Annal. d. Chem. u. Pharm., en, 3. 165.] METALS OF GROUP VI. 723 The test-analyses adduced by BUNSEN show very satisfac- tory results. b. TIN FBOM ABSENIO AND ANTIMONY (F. W. CLAKKE *). Moist freshly precipitated tin disulphide completely dis- solves on boiling for a moderate length of time with excess of oxalic acid, and therefore tin in the form of dichloride is not thrown down by hydrogen sulphide from a hot solution con- taining excess of oxalic acid. The sulphides of arsenic are barely aifected by boiling with oxalic acid, and hydrogen sul- phide immediately reprecipitates the traces dissolved. Anti- mony sulphide dissolves more copiously on boiling with oxalic acid, but hydrogen sulphide reprecipitates the antimony from the solution. CLARKE hence recommends adding to the solution of the three metals (and in which the tin is present in the form of dioxide) oxalic acid in twenty times the weight of the tin. The solution must be so concentrated that on cooling the oxalic acid crystallizes out. Now conduct into the solution, main- tained at a boiling heat, hydrogen-sulphide gas for 20 min- utes, let stand for half an hour in a warm place and filter. According to CLARKE all the arsenic and antimony, free from tin sulphide, or at least nearly so, is thus precipitated, while all the tin remains in solution. The tin is obtained by mak- ing the solution weakly alkaline with ammonia and adding ammonium sulphide until the precipitate first formed redis- solves ; then decompose the sulphosalt with an excess of acetic acid, set aside in a warm* place for the tin disulphide to settle, and determine it according to 126, 1, c. Acids which are stronger than acetic, and which liberate oxalic acid, must not be employed. CLARKE recommends, in order to secure very accurate results, to again dissolve the precipitate of antimony and arsenic sulphides in an alkaline sulphide solution, add an excess of oxalic acid, and boil with hydrogen-sulphide water, thereby obtaining the last portions of tin in solution. According to the investigations made by FR. PHILLIPS in my laboratory, this last operation would appear to be absolutely required in order to obtain any proper results whatever. * CJiem. News^ XXT, 124 ; Zetischr. f. analyt. Chem., ix, 487. 724 SEPARATION. [ 165. The very unfavorable results obtained by G-. C. "WITTSTEIN * .are perhaps referable to the fact that the solution used by him contained too much free hydrochloric acid, whereby the precipitation was .rendered incomplete at the boiling heat, arid he was compelled to complete it in the cold. In the ex- periments made by PHILLIPS the free hydrochloric acid was neutralized as nearly as possible with potassa. [CLARKE'S method, with some important modifications, has been successfully applied to the separation of tin from anti- mony in alloys by F. P. DEWEY,f who proceeds as follows: Dissolve in a mixture of 1 part strong nitric acid, -i parts strong hydrochloric acid, and 5 parts water. Since even small quantities of free mineral acids prevent complete precipitation of antimony, they are removed by evaporating to dryness on a water-bath, with previous addition of enough potassium chloride to form double salts with the tin and antimony chlo- rides present. The presence of the potassium chloride entirely prevents loss of tin and antimony by volatilization as chlorides during the evaporation. Add to the salts thus obtained a large quantity of pure" oxalic acid (at least 20 parts crytallized acid to 1 part tin) and dilute with water to about 125 c. c. per O'l grm. antimony present. The salts dissolve readily. Boil and pass H 3 S through the boiling solution half an hour. Filter immediately while hot, and wash the greater part of the soluble matter out of the precipitate with hot water. The precipi- tated antimonous sulphide will contain a little stannic sulphide. Dissolve in ammonium sulphide, avoiding an unnecessary quantity of the sol vent, and pour the solution into a strong hot solution of oxalic acid. A liberal excess of oxalic acid should be present after decomposition of the sulphur salts. Heat the oxalic solution with the suspended precipitate of antimonous sulphide to boiling and pass H 2 S gas ten minutes. Collect the Sb,S 3 now free from tin on a weighed filter, wash with hot water, and proceed to determine the antimony as directed in 125, 1, J. To recover tin from the filtrate, evaporate nearly to dryness, add strong sulphuric acid, and heat till all * Vierteljahresschr. f. prakt. Pharm, xix, 551. \ Am. Chem. Journ., i, 244. 165.] METALS OF GROUP VI. 726 the oxalic acid present is decomposed and removed. Dilute largely and precipitate the tin \vith hydrogen sulphide according to 126, 1, . The platinum may be precipitated from the filtrate by hydrogen sulphide according to 124, c. 8. Method based on the Extraction of Gold ~by Mercury. DETERMINATION OF GOLD IN PLATINUM ORE. Treat the mineral for several hours with small quantities 215 of pure, boiling mercury, pour off and repeat the operation ; then wash thoroughly with boiling mercury and distil off all the mercury very cautiously. The gold remains behind (DEVILLE and DEBRAV). Prudence requires that the residue should be tested. Pogg. AnnaL, cxn, 172. 728 SEPARATION. [ 165. 9. Method based on the Precipitation of the Indi- vidual Metals as Sulphides ~by Sodium Thiosulphate. AKSENIC AND ANTIMONY FROM TIN. Add an excess of hydrochloric acid to the solution, heat 21$ to boiling, and add sodium thiosulphate until the precipitate is no longer orange or yellow, but white, and the liquid is opalescent from separated sulphur. Arsenic and antimony are completely precipitated, while all the tin remains in solu- tion (YoHL*). Estimate the former, if one alone of the metals is present, according to 125, 1 and 127, 4. If both together are present, separate according to 201 or 204. The tin in the filtrate is best determined according to 126, 1, c. LENSSEN f employed this method with apparently good results. My experience has not, however, been so favor- able. As tin is also precipitated by sodium thiosulphate unless free hydrochloric acid is present, the separation can be successful only when hydrochloric acid present prevents pre- cipitation of tin, while not hindering that of the antimony. 10. Method based upon the Precipitation of Tin as Stannic Ar senate. TIN FROM AKSENIC. ED. HAFFELY ^ has proposed the following method of deter- 217 mining both the tin and the arsenic in commercial sodium stannate, which often contains a large admixture of sodium arsenate. Mix a weighed sample with a known quantity of sodium arsenate in excess, add nitric acid also in excess, boil, filter off the precipitate, which has the composition 2SnO,*As a O, -f- 10H 3 O, and wash; expel the water by igni- tion and weigh the residue, which consists of 2SnO a *A.s 2 O 6 . In the filtrate determine the excess of arsenic acid as directed in 127, 2. The amount of the stannic oxide is found from the weight of the precipitate^ that of the arsenic acid is obtained by adding the quantity in the precipitate to the quantity in the filtrate and deducting the quantity added. * Annal. d. Chem. u. Pharm., xcvi, 240. f 75. , cxiv, 118. \Phil. Mag., x, 220 ; Journ. f. prakt. Chem., LXVII, 209. 165.] METALS OF GROUP VI. 729 1 1 . Method based on the Separation of A rsenic and Antimony from their Hydrogen Compounds. To determine both metals in a mixture of arsenic and 218 antimony hydrides, conduct the gas into a solution of neutral silver nitrate. Antimony hydride yields silver antimonide, whereas arsenic goes into solution as arsenous acid, with reduc- tion of silver. This method was recommended by A. W. HOFMAN * for the qualitative detection of arsenic and anti- mony. Filter off the precipitate, consisting of silver and silver antimonide, and wash it. To the solution add a slight excess of hydrochloric acid, filter off the silver chloride, and pre- cipitate with hydrogen sulphide. The precipitate is arsenous sulphide containing a small quantity of antimonous sulphide, which is to be separated according to 202 or 207. The pre- cipitate of silver and silver antimonide heat with tartaric acid and a very little nitric acid, and determine the antimony according to 125, 1. All methods of determining antimony and arsenic in solu- tions, based on treating the solution with zinc and hydro- chloric acid, passing the gas into silver-nitrate solution, etc., are unreliable, because only a certain part of the arsenic and antimony are evolved as hydrides, while the balance remains * in the flask in the form of metals. 1 2 . Volumetric Methods. a. ARSENOUS FROM ARSENIC ACID. Convert the whole of the arsenic in a portion of the sub- 219 stance into arsenic acid and determine the total amount of this as directed 127, 2 ; determine in another portion the arsen- ous acid as directed in 12T, 5, #, and calculate the arsenic acid from the difference. 1}. ANTIMONY OF ANTIMONOUS COMPOUNDS FROM ANTIMONIO ACID. Determine in a sample of the substance the total amount 220 of the antimony as directed 125, 1, in another portion esti- mate the antimony present as an antimonous compound as * Annal. d. Chem. u. Pharm., cxv, 287. 730 SEPARATION". [ 166. directed 125, 3, and calculate the antimonic acid from the difference. c. TIN OF STANNOUS, FROM TIN OF STANNIC Ccmporxns. In one portion of the substance convert the whole of the 221 stannons into stannic salts by digestion with chlorine water or some other means, and determine the total quantity of tin as directed 126, 1, b ; in another portion, which, if necessary, is to be dissolved in hydrochloric acid in a stream of carbonic acid, determine the stannous tin according to 126, 2. II. SEPARATION OF THE ACIDS FROM EACH OTHER. It must not be forgotten that the following methods of separation proceed generally upon the assumption that the acids exist either in the free state, or as alkali salts ; compare the introductory remarks, (p. 597. Where several acids are to be determined in one and the same substance, we very often use a separate portion for each. The methods here given do not embrace every imaginable case, but only the most important cases, and those of most frequent occurrence. First Group. ARSENOUS ACID ARSENIC ACID CHROMIC ACID SULPHURIC ACID PHOSPHORIC ACID BORIC ACID OXALIC ACID HYDROFLUORIC ACID SILICIC ACID CARBONIC ACID. 166. 1. ARSENOUS ACID AND ARSENIC ACID FROM ALL OTHER ACIDS. Precipitate the arsenic from the solution by hydrogen sul- 222 phide ( 127, 4, a or &), filter, and determine the other acids in the nitrate. It must be remembered, that the arsenous sulphide will be obtained mixed with sulphur if chromic acid, ferric salts, or any other substances which decompose hydro- gen sulphide are present. The estimation of sulphuric acid in the nitrate cannot be accurate unless air is excluded, and oxidizers such as chromic acid are absent ; sulphuric acid is, therefore, best estimated in a separate portion (223). From those acids which form soluble magnesium salts, arsenic acid 166.] ACIDS OF GROUP I. 731 may be separated also by precipitation as ammonium magne- sium arsenate ( 127, 2). 2. SULPHURIC ACID FROM ALL THE OTHER ACIDS.* a. from Arsenous, Arsenic, Phosphoric, f Boric, Oxalic, and Carbonic Acids. Acidify the dilute solution strongly with hydrochloric acid, 223 mix with barium chloride, and filter the barium sulphate from the solution, which contains all the other acids. Determine the barium sulphate as directed 132. If acids are present \vkli which barium forms salts insoluble in water but soluble in acids, the barium sulphate is apt to carry down with it such salts, and this is all the more liable to happen, the longer the precipitate is allowed to settle. This remark applies especially to barium oxalate, and tartrate, and the barium salts of other organic acids (H. ROSE). In such cases I would recom- mend, after washing, to stop up the neck of the funnel, and digest the precipitate with a solution of hydrogen sodium car- bonate, then to wash with water, with dilute hydrochloric acid, and again with water. In every case, however, the purity of the weighed barium sulphate must be tested as directed 132, 1. In the fluids filtered from the barium sulphate the other acids are determined according to the directions of the Fourth Section, after the removal of the excess of barium chloride. Or the other acids may be estimated in separate portions of the substance, which is indeed usually the best way, and for carbonic acid is of course the only way. 1). From Hydrofluoric Acid. of. When sulphuric acid and hydrofluoric acid are present 224 in the free state in aqueous solution, it is best to estimate the acidity in one portion by means of standard soda ( 215), and the sulphuric acid in another ( 132, I., 1), finding the hydro- fluoric acid by difference. The barium sulphate should be purified by fusion with sodium carbonate ( 132, I., 1). * With respect to the separation of sulphuric acid from selenic acid, comp. WOHLWILL (Annal. d. C/iem. u. Fharm., cxiv, 183). f If metaphosphoric acid is present, it must first be converted into ortho- phosphoric by fusion with alkali carbonate. 732 SEPARATION. [ 166, /?. To estimate both acids in minerals or other dry sub- 225 stances, it is safest, provided the fluoride can be decomposed by sulphuric acid, to determine the fluorine in one portion according to 138, 3, #, and to fuse another portion for a long time with four times its amount of sodium carbonate, which will decompose the sulphate thoroughly, the fluoride generally but partially. The fused mass is soaked in water, the solution filtered, acidified with hydrochloric acid and pre- cipitated with barium chloride. The barium sulphate thus obtained generally contains barium fluoride, and must be purified according to 132, I., 1, by fusion with sodium car- bonate, &c. y. An actual separation of both acids may be effected, 226 when both are in the form of alkali salts, by adding sodium ' carbonate if necessary, and then precipitating the fluorine according to 138, I., adding the calcium chloride cautiously in very slight excess. The sulphuric acid is for the most part found in the filtrate from the calcium carbonate and fluoride, a very small part is generally also found in the calcium acetate filtered from the calcium fluoride. Both filtrates are acidified and precipitated with barium chloride ( 132, I., 1. H. EOSE). ft. Insoluble compounds may also be decomposed by fusion 227 with six parts of sodium and potassium carbonates, and two parts of silica. The fused mass, after cooling, is treated with water, the solution is mixed with ammonium carbonate, and heated, more ammonium carbonate is added to replace what evaporates, the silicic acid thrown down is filtered off and washed with water containing ammonium carbonate, a solu- tion of zinc oxide in ammonia is added to precipitate the remaining silica, the fluid is evaporated till all ammonia is driven oif, filtered and the process concluded as in y. The precipitate produced by the zinc should be tested for sulphuric acid. c. From Chromic Acid. Boil the dry compound with strong hydrochloric acid 228 (p. 357, ft) and estimate the chromic acid from the evolved chlorine. Neutralize some of the acid with ammonia, dilute and precipitate the sulphuric acid by long boiling w r ith excess of barium chloride. The barium sulphate thus obtained 166.] ACIDS OF GKOTJP I. 733 retains chromic oxide (H. ROSE) and must always be fused with sodium carbonate, &c. ( 132, I., 1). d. From Hydrqfluosilicic Acid. First throw down the hydrofluosilicic acid according to 229 133, as potassium silicofluoride, then the sulphuric acid in the filtrate with barium chloride. e. From Silicic Acid. Compare 242. 3. PHOSPHORIC ACID FROM THE OTHER ACIDS. * a. From the acids of arsenic, see 222 ; from sulphuric 230 add, see 223 ; from silicic acid, see 242. b. from Chromic Acid. Precipitate the phosphoric acid by adding ammonium nitrate and ammonia, and then magnesium nitrate, and deter- mine the chromic acid in the nitrate as directed 130, L, a, ft or I.,, b. c. From Boric Acid. Precipitate the phosphoric acid with a solution of double 231 chloride of magnesium and ammonium ( 134, 6, or), wash the precipitate partially, redissolve it in hydrochloric acid, repre- cipitate with ammonia, adding a little magnesium and ammo- nium chloride, and estimate the phosphoric acid as magnesium pyrophosphate. In the filtrate estimate the boric acid as magnesium borate ( 136, L, 1, 6). d. From Oxalic Acid. a. If the two acids are to be determined in one portion, 232 the aqueous or hydrochloric solution is mixed with sodium auric chloride in excess, heat applied, and the oxalic acid cal- culated from the reduced gold ( 137, c). The gold added in excess is separated from the nitrate by hydrogen sulphide, and the phosphoric acid then precipitated by double chloride of magnesium and ammonium. /3. If there is enough of the substance, the oxalic acid is 233 determined in one portion according to 137, b, or d, and the phosphoric acid in another portion. If the substance is solu- 734 SEPARATION. [ 166. ble in water, and the quantity of oxalic acid inconsiderable, the phosphoric acid may be precipitated at once with magne- sium chloride, ammonium chloride, and ammonia : if not, the substance is ignited with potassium carbonate and sodium car- bonate, and the oxalic acid being thus destroyed, the phos- phoric acid is determined in the nitric acid solution of the residue according to 134, I., 5, ft. e. From Hydrofluoric Acid. a. Phosphates and fluorides are frequently found together 234 in minerals. In the analysis of phosphorites, for instance, we have to estimate small quantities of fluorine, often, too in the presence of aluminium and iron, which increase the difficulty. According to my own experience,* it is always safest in such cases to estimate in one portion the fluorine as silicon fluoride ( 138, II., 3, #), and in another portion the phosphoric acid. Regarding the first estimation, it must be mentioned that car- bonic acid if present must first be removed. To this end heat the finely powdered weighed substance with water, add acetic acid in slight excess, and also, if the fluoride present is soluble in water, some calcium acetate ; evaporate to dryness on a water bath, treat with water, filter, wash the insoluble matter, dry, separate as far as possible from the filter, add the filter ash, weigh, test a small portion for carbonic acid by heating with hydrochloric acid, and weigh the rest for the fluorine estima- tion. For the estimation of the phosphoric acid, dissolve the finely powdered substance in hydrochloric acid, evaporate to dryness on a water-bath, moisten with a little hydrochloric acid, add nitric acid, warm, dilute, filter, evaporate filtrate and washings to dryness, dissolve in nitric acid, and proceed according to 134, I., &, ft. ft. Where you have an alkali phosphate and an alkali 235 fluoride together in aqueous solution the phosphoric acid may be separated according to 135, II., d, ft, as silver phosphate, or according to 135, II., &, as mercurous phosphate. The fluoride will be all in the filtrate. If the former method is adopted the silver is removed from the filtrate by sodium chloride, and the fluorine estimated as calcium salt ( 138, 1.). * Zeitschr. f. analyt. Chem., v, 190, and vi, 403. 166.] ACIDS OF GROUP I. 735 If the latter method is adopted, as the solution is always acid, the use of glass and porcelain must be avoided. The mercury is remove') f'-om the filtrate by neutralizing with sodium car- bonate {in- without filtering passing hydrogen sulphide. The fluorine is estimated in the filtrate as calcium salt, accord- ing to 138, I. (II. KOBE). stances which are insoluble in water, and cannot 236 be decomposed by acids, are fused with sodium carbonate and silica (227), the fused mass is treated with water, and Jie solution with ammonium carbonate. In this way all the fluorine and all, or nearly all, the phosphoric acid will be brought into solution. The solution is treated as in 235, and any remainder of phosphoric acid in the undissolved residue is estimated according to 234. #. In compounds decomposable by water, fluorine may 237* be occasionally estimated indirectly also. Dissolve in hydro- chloric acid, evaporate with a slight excess of sulphuric acid until all the hydrofluoric acid has escaped (the amount must not be increased to a point where the sulphuric acid will be driven off, otherwise some phosphoric acid will also escape), and determine in the residue the phosphoric acid on the one hand ; on the other the oxides. If now the proportion be- tween the phosphoric acid and the bases in the compound investigated is known, the escaped fluorine may be calculated from the excess of bases. It is assumed, of course, that other acids must not be present, or must be estimated in separate portions. 4. HYDROFLUORIC ACID FROM OTHER ACIDS. a. Fluorides from B orates. Mix the solution containing alkali borate and fluoride with 238 some sodium carbonate, and add calcium acetate in excess. A precipitate is formed, which contains the w r hole of the fluorine MS calcium fluoride, and besides this, calcium carbonate and some calcium borate; the greater portion of the latter having been redissolved by the excess of the calcium salt added. Determine the calcium fluoride in the precipitate as directed 138, I. The small quantity ot boric acid in the precipitate is, in this process, partly volatilized, partly dissolved after evaporating the mass with acetic acid and extracting with 736 SEPARATION. [ 166. water. It is therefore necessary to determine the boric acid in a separate portion of the substance, according to 136, I. 2 (A. STROMEYEK).* ~b. Fluorides from Silicic Acid and Silicates. A great many native silicates contain fluorides : care must, therefore, always be taken, in the analysis of minerals, not to overlook the latter. If the silicates containing .fluoride are decomposable by acids which is only rarely the case and the silicic acid is separated in the usual way by evaporation, the whole of the fluorine may volatilize. a. BERZELIUS'S method. Fuse the elutriated substance 239 with 4 parts of sodium carbonate for some time at a strong red heat, digest the mass in water, boil, filter, and wash, first with boiling water, then with ammonium carbonate. The fil- trate contains all the fluorine as sodium fluoride, and, besides this, sodium carbonate, silicate, and aluminate. Mix the fil- trate with ammonium carbonate and heat the mixture, replac- ing the ammonium carbonate, which evaporates. Filter off the precipitate of hydrate of silicic acid and aluminium hydroxide, and wash with ammonium carbonate. To separate the last portions of silica from the filtrate add a solution of zinc oxide in ammonia, evaporate till no more ammonia escapes, and filter off the precipitate of zinc silicate and oxide. Determine the silica in this precipitate by dissolving in nitric acid, evaporating to dryness, taking up with nitric acid, and filtering off the undissolved silica. In the alkaline filtrate estimate 1 the fluorine as calcium salt ( 138, I.). The residue, insoluble in water, and the precipitate produced by ammonium carbonate are finally treated with hydrochloric acid according to 140, II., a, in order to separate the silica. fi. In substances readily decomposed by sulphuric acid you 240 may also separate and weigh the silica according to 239 in one portion, and determine the fluorine in another portion accord- ing to 138, II., 3, a. c. Fluorides, Silicates and Phosphates together. Compounds of this kind are not rare in nature, and may 241 be decomposed according to 239. We cannot alway rely on * Annal, d. Chem. u. Pharm., c, 91. 166.] ACIDS OF GROUP I. 737 complete decomposition of the phosphate, as, for instance, cal- cium phosphate is but partially decomposed on fusion with sodium carbonate. The solution, obtained after separation of the silica by ammonium carbonate and the zinc solution, is made up to a definite volume, and a portion is tested for phos- phoric acid with molybdic solution. If none is present the fluorine is estimated in the measured remainder of the fluid as fluoride of calcium ( 138, I.). If on the other hand phos- phoric acid is still present, treat the measured remainder of the fluid according to 235. In the original residue and the ammo- nium carbonate precipitate estimate the principal amounts of the silicic and phosphoric acids and the basic metals. In the zinc precipitate estimate the remainder of the silicic acid, and in the filtrate from the latter estimate the portion of the phosphoric acid which was thrown down by zinc oxide. As the phosphoric acid is so divided by this method, it is well to make a direct estimation of it in another portion of the substance, especially when only a small quantity is present. For this purpose decompose the silicate with hydrofluoric and hydrochloric acids, add enough but not too large an excess of sulphuric acid, and evaporate till all the fluorine has escaped as silicon fluoride and hydrofluoric acid. Do not increase the heat to the escape of sulphuric acid, or phosphoric acid may be lost, Take up the residue with nitric acid, dilute, filter, and estimate the phosphoric acid in the filtrate by the molybdic method, page 446. If the substance can be easily decomposed with sulphuric acid, the fluorine may of course also be expelled as silicon fluoride and estimated according to 138, II., 3, a. 5. SILICIC Aero FKOM ALL OTHER ACIDS. a. In 'compounds which are decomposed ~by hydrochloric acid. Decompose the substance by digestion with hydrochloric 242 or nitric acid, evaporate the whole on the water hath to dryness ( 140, II., #), treat with water, hydrochloric acid or nitric acid according to circumstances, filter off the silica, and estimate the other acids in the filtrate. The following points require attention. a. In the presence of borates or fluorides this method cannot be used; employ 243. 738 SEPARATION. [ 166. /3. In the presence of phosphoric acid the silica always retains a small portion, which cannot be extracted by washing with acidified water (H. ROSE, W. SKEY*). After washing the silica with water, treat it repeatedly with ammonia, which will leave only a very minute quantity of the phosphoric acid. Evaporate the ammoniacal fluid, finally adding a little hydro- chloric acid, dissolve in water with addition of a little nitric acid, filter off the small amount of silica which was taken up by the ammonia, and estimate the remainder of the phosphoric acid in the filtrate. 1). In compounds which are not decomposed l>y hydrochlo- ric acid. Fuse with carbonate of potash and soda (p. 511) and treat 243 the residue either at once cautiously with dilute hydrochloric or nitric acid, in order to proceed with the solution according to 242 (not applicable in presence of boric acid or fluorine) ; or taking the fluid obtained by boiling the residue with water, precipitate the dissolved silica by warming with ammonium carbonate, and throw down the last portion of silica from the filtrate of zinc oxide dissolved in ammonia (239). The silicic acid is then found partly in the residue left un- dissolved by water, partly in the precipitate produced by ammonium carbonate, and partly in the precipitate produced by the zinc solution. Separate it according to 140, II., a. Boric acid and fluorine will be found entirely in the last alka- line filtrate (239). Regarding phosphoric acid see 241. Sulphuric acid passes for the most part into the last alkaline filtrate, yet it is well also to examine the acid filtrates from the silica. 6. CARBONIC ACID FROM ALL OTHER ACIDS. When carbonates are heated with stronger acids, the car- 244 bonic acid is expelled ; the presence of carbonates, therefore, does not interfere with the estimation of most other acids. And as, on the other hand, the carbonic acid is determined by the loss of weight or by combination of the expelled gas, the presence of salts of non-volatile acids does not interfere with the determination of the carbonic acid. Accordingly, with compounds containing carbonates, sulphates, phosphates, &c., * Zeitschr. f. analyt. Chem., vm, 70. 167.] ACIDS OF GROUP II. 739 either the carbonic acid is determined in one portion, and the other acids in another, or both estimations are pei formed on one portion. In the latter case the process described on p. 500, , or p. 493, . The known quantity of silver used for precipitating is weighed either directly and is dissolved in nitric acid, or it is added in the form of standard silver solution. This method is more convenient than that given under #, but I do not con- sider it quite so accurate, particularly if only small quantities of bromine are present. It is assumed that from a weighed quantity of silver the absolutely correct quantity of silver chloride equivalent to it is obtained; and this assumption cannot be realized in practice. Errors to the extent of some milligrammes cannot be avoided, hence the difference might be calculated as bromine, even when none is present at all. The method given under a is not nearly so likely to afford errors, or at least to the same extent. Further, one can ascertain without trouble whether, on carefully heating silver chloride in a current of chlorine, any change in weight takes place, and thereby rendering an error of 0*5 milligramme less excusable than one of 2 milligrammes incurred by converting 2 or 3 grm. silver into chloride; and this is scarcely avoid- able, particularly if a filter is used in the process, as is re- quired in a partial precipitation, in which case the precipi- tate always subsides less. d. FisANi'sf method may be regarded as a modification 260 of 0, wherein a known quantity of silver solution is added in slight excess, the precipitate filtered off, and the silver in the fil- trate estimated with starch iodide (page 349) . The precipitate is weighed as in c. This method precludes partial precipitation. e. Determine in a portion of the solution the chlorine 261 -f-bromine (by precipitating with silver), either gravimetrically or volu metrically ; in another portion the bromine, either by the colorimetric method ( 143, L, J, or or /?) or volumetrically ( 143, I., >, y}. Calculate the chlorine from the difference. * Annal. d. Chem. u. Pharm., xcm, 76. f Compt. rend., XLIV. 352 ; Journ.f. prakt. Chem., LXXII, 266. 748 SEPARATION. [ 169. The method is very suitable for an expeditious analysis of mother-liquors. /. Compare also 271 and 272. 2. CHLORINE FROM IODINE. a. Add to the solution palladious nitrate, and determine 262 the precipitated palladious iodide as directed 145, I., a, fi. Conduct hydrogen sulphide into the filtrate to remove excess of the palladium, destroy the excess of hydrogen sulphide by solution of ferric sulphate, and precipitate the chlorine finally with solution of silver. It is generally found more simple and convenient to precipitate from one portion the iodine, by means of palladious chloride, as directed 145, I., , ytf, from another portion the chlorine and iodine jointly with silver nitrate, and to calculate the chlorine from the difference. If you have no solution of palladious nitrate ready, and the chlorine and iodine must be determined in one portion of the solution under examination, add a measured quantity of a solution of palladious chloride, determine the amount of chlo- rine in this and in another exactly equal portion of the same solution, and deduct it. The results are accurate. In the case of fluids containing a large proportion of alkali chlorides to a small quantity of iodide and such cases often occur the iodide is concentrated by adding sodium carbonate to the fluid, evaporating to dryness, extracting the residue with hot alcohol, evaporating the alcoholic solution with addition of a drop of solution of soda, and taking the residue up with water. J. Proceed exactly as for the indirect determination of 263 bromine in presence of chlorine (255). The greatest care must be taken that as little as possible of the mixed silver chlo- ride and iodide adheres to the filter, for silver iodide dissolves only very slightly in ammonia. Any particles of silver iodide remaining attached to the filter may be saved by incinerating the filter and evaporating the ash with a drop of nitric acid and a drop of hydriodic acid. The loss of weight suffered by the silver precipitate on fusion in chlorine multiplied by 2*569 gives the amount of silver iodide present. The methods given under 259 and 260 are also applicable. These methods give still better results than in the separation of bromine from chlorine, inasmuch as the difference between the atomic weights of iodine and chlorine is far greater than the differ- 169.] ACIDS OF GROUP II. 749 ence between those of bromine and chlorine. Regarding the concentration of the iodide, if necessary, see 262. c. Liberate the iodine by nitrous acid, take it up with car- 264 bon disulphide, wash the latter, and then estimate the iodine in it by sodium thiosulphate (p. 537, /?). In this process the chlorine is determined either in the fluid separated from the violet carbon disulphide, or with greater accuracy by precipitating the chlorine -f- iodine in a second portion with silver and deducting the weight of silver iodide corresponding to the iodine already found from the weight of the precipitate. A good and approved method. If the quantity of iodine is small, the following method may also be used with advantage for estimating it: The carbon disulphide should be thoroughly washed, and covered with a layer of water in a stoppered bottle. Add drop by drop, with shaking, dilute chlorine water (of unknown strength) till the coloration has just vanished and all the iodine is consequently converted into IC1 6 . Separate the solu- tion from the disulphide, add potassium-iodide solution in suf- ficient excess, and determine the free iodine after 146. Six parts of the iodine found correspond to 1 part originally pres- ent. If the analyst would avoid the trouble of pouring off the fluid from the disulphide, and of washing the latter, he may transfer the mixture, after the addition of chlorine to decolora- tion, to a somewhat narrow measuring cylinder, note the vol- ume occupied by the iodine-pentachloride solution, take out a portion with a pipette, and proceed as above directed. Instead of carbon disulphide MOEIDE* uses benzene, while RoGERf employs chloroform; and instead of nitrous acid, the latter uses iodic acid for liberating the iodine, as pre- viously recommended by LIEBIG, a dilute solution of the reagent being added to the dilute fluid acidulated with sul- phuric acid. From the equation 5HI + HIO, = 61 + 3H 2 O, it follows that only of the iodine found belonged to the iodide originally present. d. Determine in one portion chlorine and iodine as in 265 141, I., 5, a-, and in another portion the iodine alone as in 145, I., 5, y^ tf, or e. The chlorine is found by difference. *Compt. rend., xxxv, 789; Journ. f. prakt. Cfom., LVIII, 317. \Journ. de P/tarm., xxxvn, 410. 750 SEPAKATIOH. [ 169. The method in 145, I., 5, (PisANi's) is very rapid, and still gives approximately accurate results in the presence of small quantities of chloride; if much chloride is present, however, the results are altogether inaccurate (see page 540). The method in 145, I., Z>, y (REINIGE'S) cannot be em- ployed if the solution contains any organic or other sub- stances capable of reducing potassium permanganate. The method in 145, I., 5, e is inapplicable if the fluid contains chlorates, nitrites, or nitrates. e. For technical purposes the following method is also 266 suitable. It was recommended by WALLACE and LAMONT * for the estimation of iodine in kelp. The kelp-lye is nearly neutralized with nitric acid, evaporated to dryness, and the residue fused in a platinum vessel to oxidation of all the sul- phides. Treat with water, filter, add silver nitrate till the precipitate appears perfectly white, wash, digest with strong ammonia, and weigh the residual silver iodide. Finally add to the weight of the latter the amount which passes into solution in the ammonia; it is -5-^3- of the aqueous am- monia (sp. gr. 0*89) used. See also 268, 271, and 272. 3. CHLORINE, BROMINE, AND IODINE FROM EACH OTHER. a. The three acid radicals are determined jointly in u por- 267 tion of the fluid by precipitating with solution of silver nitrate ( 141, I., a or , a). To determine the iodine, another portion is precipitated with palladious chloride in the least pos- slble.exeess ( 145, L, a, ft). The fluid filtered from the pre- cipitate is freed from palladium by hydrogen sulphide and the excess of the latter removed by means of ferric sulphate ; the chlorine and bromine are then precipitated jointly either com- pletely or partially with silver nitrate, and the bromine deter- mined as directed 255. If the compound contains a large proportion of chlorine to a small proportion of bromine, the iodine may be precipitated also by palladious nitrate, as there is no danger, in that case, of palladious bromide being coprecipitated. The filtrate is treated as above. These methods give accurate results ; but they are appli- *Gliem. Gaz , 1859, 137. 169.] ACIDS OF GROUP II. 751 cable only if the quantity of iodide present is somewhat con- siderable. 1). Mix the neutral dilute and cold solution containing alkali 268 iodide with alkali chloride or alkaki bromide, or both, with a saturated neutral solution of thallium nitrate, stirring well till, on repeated trial, you obtain a transient white precipitate the first and permanent precipitate being yellow. It is best to have the thallium polution in a burette, so that you can easily add it by drops. If the white precipitate of thallium chloride or bromide does not at once disappear on stirring, add more water, but not an unnecessary quantity, or some of the thal- lium iodide will remain in solution. Allow to stand eight or twelve hours in a cold place, pour ofl the clear fluid through a weighed filter dried at 100, wash the filter a little so that no more water than necessary may pass through the precipitate, turn the precipitate on to the filter, wash with as little water as you can, dry at 100, and w r eigh. Precipitate the chlorine and bromine in the filtrate by silver solution. If they are both present, the mixed silver precipitate is to be treated according to 255 . Results quite satis- factory (HtJBNEB and SPEZIA,* and HUBNER and FKERiCHsf). c. Remove the iodine from the solution by carbon disul- 269 phide or chloroform, as in 264. In the fluid separated from the iodized carbon disulphide determine the chlorine and bro- mine as directed in 255, arid in the iodized carbon disulphide, the iodine as directed in 145, I., 5, /?. This method is particularly recommended for the separation of small quanti- ties of iodine, and in this respect is supplementary to 267. d. Determine in a portion of the compound the chlorine, 270 bromine, and iodine jointly by adding a known quantity of standard silver solution in slight excess, filtering and deter- mining the small excess of silver in the filtrate by iodide of starch (p. 349). The precipitate* is weighed. Compare 263. We now know the total of the chloride, bromide, and iodide of silver and also the silver therein contained. Determine the iodine separately as in 269, calculate the quantity of silver iodide and of silver corresponding to the amount found, deduct the calculated amount of silver iodide from the mixed iodide, chloride, and bromide of silver, that * Zeitschr. /. analyt. Chem., xi, 397. f/6., xi, 400. 752 SEPAEATION. [ 169. of the silver from the known quantity of the metal contained in the mixed compound ; the remainders are respectively the joint amount of chloride and bromide of silver and the quan- tity of the metal contained therein ; these are the data for . calculating the chlorine and bromine (258). e. On the fact that freshly precipitated silver chloride is 271 converted into silver bromide by a solution of sodium bromide, and that freshly precipitated silver bromide and chloride are converted into silver iodide by potassium iodide in solution, F. FIELD* has based the following method of estimating all three halogens, if present, and combined with metals : Introduce three weighed portions each into a stop- pered flask, add to each about 30 c. c. water and an excess of silver solution, shake vigorously, and thoroughly wash the precipitates Nos. I, II, and III with water. Dry and weigh No. I ; the weight represents the sum of the silver chloride, bromide, and iodide present. Then digest ~No. II with potassium-bromide solution, and No. Ill with potassium- iodide solution, for 10 hours, taking care that the solutions are dilute and not added in too great excess, and avoiding warming, otherwise notable quantities of silver salts will be dissolved. After II is washed, ignited, and weighed, it gives the quantity of silver bromide and iodide; while III finally gives pure silver iodide. The calculation is as follows : a. The difference between the equivalents of iodine and chlorine (= 91*4) : eq. of silver chloride (= 143*37) : : dif- ference between the weights of I and II : the silver chloride contained in I. P. The difference between the equivalents of iodine and bromine (= 46-9) : eq. of silver bromide (= 187*87) : : dif- ference between II and III : silver-bromide content of II. On deducting the silver bromide found from the weight of the precipitate II, the silver-iodide Content is obtained. y. Finally on subtracting the silver chloride found in a, together with the silver iodide found in /?, from the precipi- tate I, the weight of the silver bromide is obtained. The method is of interest theoretically. FIELD obtained quite satisfactory results. The method was later on thoroughly investigated by * Quart. Journ. Chem. Soc., x, No. 39, 234; Journ. f. prakt. Chem., LXXIII, 404; also Chem. News, n, 325. 169.] ACIDS OF GROUP II. 753 O. HUSCHKE,* and also by M. SiEWERT.f The former used a 1 : 48 potassium-bromide solution and a 1 : 34 potassium- iodide solution, and digested with a moderate excess of solution for 1 Lour. He obtained 5*248 and 5*206 grains of iodine instead of 5-287; 3-313 and 3-349 grains bromine instead of 3-333, and 1-477 and 1-496 grains chlorine instead of 1-503 grains. SIEWERT worked with both cold and hot solutions, but obtained less satisfactory results. According to his investiga- tions, the conversion of silver chloride into bromide is incom- plete, and further, on boiling silver bromide with sodium- chloride solution, silver chloride is found. The conversion of silver chloride and bromide into iodide, however, he found to be perfectly complete. FIELD'S method, hence, can at most be used only when relatively large quantities of all three halogens are present, and when approximate results will suffice. The method is absolutely inapplicable in the analyses of mineral waters J and more particularly when only very small quantities of iodides and bromides are present with comparatively large quantities of chlorides. f. II. HAGER'S method depends upon the fact that freshly 272 precipitated silver chloride is soluble in a boiling solution of ammonium carbonate, while only traces of silver bromide dis- solve in such a solution, and silver iodide is almost absolutely insoluble. Regarding the details of this method, in which the silver bromide and iodide are separated by ammonia, I refer to the original paper. The method can be employed only when approximate results suffice. In SONSTADT'S J method the iodine is precipitated as barium iodate. 4. ANALYSIS OF IODINE CONTAINING CHLORINE. a. Dissolve a weighed quantity of the dried iodine in 273 cold sulphurous acid, precipitate with silver nitrate, digest the precipitate with nitric acid to remove the silver sulphite which may have coprecipitated, and weigh. The calculation * Zeitschr.f. analyt. Chem., vu, 434. t ZeitscJir. f. die gesammt. Naturwiss., 1868, No. 1; Zeitschr.f. analyt. Chem., vii, 469. {J. MITTEREGGER, however, employed it thus, using only 500 grin, of mineral water. See Chem. Analyse des Radeiners Sauerbrun, by Dr. Jos. Mn> TEREGGER, Vienna, 1872, published W. by BRAUMULLER, p. 5. %Pharm. Centralbl., xn, 42 ; Zeitschr.f. analyt. Chem., x, 341. || Chem. News, xxvi, 173; Zeitschr. f. analyt. Chem., xii, 91. 754 SEPARATION. [ 169. of the iodine and chlorine is made by the following equations, in which A represents the quantity of iodine analyzed, x the iodine contained in it, y the chlorine contained in it, and B the amount of silver chloride and iodide obtained : x + y = A and Now as and we have 2-1935 J. If you have free iodine and free chlorine in solution, deter- 274 mine in one portion, after heating with sulphurous acid, the iodine as palladium iodide ( 145, I., #, ft\ and treat another portion as directed 146. Deduct from the apparent amount of iodine found by the latter process, the actual quantity calcu- lated from the palladium iodide ; the difference expresses the amount of iodine equivalent to the chlorine contained in the substance. 5. ANALYSIS OF BROMINE CONTAINING CHLORINE. a. Proceed exactly as in 273, weighing the bromine in a 275 small glass bulb. Taking A to be equal to the analyzed bro- mine, B to the silver bromide and chloride obtained, x to the bromine contained in A, y to the chlorine contained in A, the calculation is made by the following equations : x 4- y = A -i ' *J and ^g -2 -34984 J. 1-69444 5. Mix the weighed anhydrous bromine with solution of 27$ iodide of potassium in excess, and determine the separated iodine as directed 146. 169.] ACIDS OF GROUP II. 755 From these data, the respective quantities of bromine and chlorine are calculated by the following equations. Let A represent the weighed bromine, i the iodine found, y the chlorine contained in A, % the bromine contained in A, then _i 1-5866 A 1-9907 BUNSEN, the originator of methods 4 and 5, has experi- mentally proved their accuracy.* 6. CYANOGEN FROM CHLOKINE, BKOMINE, OR IODINE. a. Precipitate with solution of silver, collect the precipi- 277 tate upon a weighed filter, and dry in the water-bath until the weight remains constant ; then determine the cyanogen by the method of organic analysis ; the quantity of the chlorine, bro- mine, or iodine is found by difference. J. Precipitate with solution of silver as in 277, dry the pre- 278 cipitate at 100 and weigh. Heat the precipitate, or an ali- quot part of it, in a porcelain crucible, with cautious agitation of the contents, to complete fusion ; add dilute sulphuric acid to the fused mass, then reduce by zinc, filter the solution from the metallic silver and silver paracyanide, and determine the chlorine, iodine, or bromine in the filtrate, in the usual way by silver. The silver cyanide is the difference. NEUBATJER and KEENER f obtained very satisfactory results by this method. c. Precipitate with solution of silver as in 277, weigh the pre- 279 cipitate and heat it, or an aliquot part, with nitric acid of 1'2 sp. gr. in a sealed tube at 100 for several hours, or at 150 for one hour. The silver cyanide is completely decomposed, while the chloride, bromide, and iodide are unaffected. Filter the contents of the tube, wash the precipitate and w^eigh it, the loss indicates the amount of silver cyanide (K. KRAUT;):). d. Determine the radicals jointly in a portion by precipi- 280 tating with solution of silver, and the cyanogen in another portion, in the volumetric way ( 147, I., 5 or c). * Annal. d. Cliem. u. P7iarm., LXXXVI, 274, 276. f 76., ci, 344. \ Zeitschr.f. analyt. Chem., n, 243. 756 SEPARATION. [ 169- 7. FERRO- OK FERRICYANOGEN FROM HYDROCHLORIC ACID. To analyze say potassium ferro- or ferricyanide, mixed with 281 an alkali chloride, determine in one portion the ferro- or ferri- cyanogen as directed " 147, II., g ; acidify another portion with nitric acid, precipitate with solution of silver, wash the precipitate, fuse with 4 parts of sodium carbonate and 1 part of potassium nitrate, extract the fused mass with water, and determine the chlorine in the solution as directed in 141. 8. SULPHUR (IN SULPHIDES) FROM CHLORINE. The old method of separating the two radicals by means of a 282 metallic salt is liable to give false results, as part of the chlo- rine may fall down as chloride with the sulphide. We, there- fore, precipitate both as silver compounds, dry the precipitate at 100, weigh it, and determine the sulphur in a weighed portion ; or and this is usually preferred determine in a portion of the solution the sulphur as directed in 148, I. , a, J, or asic (ic) 197 Iron-alum 147 -ammonium sulphate (ous) 146 arsenal? (ic) 224 converting ferrous into ferric 311 772 INDEX. PAGE Iron ferric, determination as oxide or hydroxide 323 as sulphide 323, 325 by Oudeman's method 332 by reduction with hydrogen sulphide 326 by reduction with stannous chloride 327 by reduction with zinc 325 volumetrically 325 with thiosulphate 331 with thiosulphate and copper sulphate ...... 332 Fuch's method of determining 334 separation from aluminium 646, 652, 660 from aluminium and chromium 652 from barium and strontium 633, 634 from calcium and magnesium . 633, 634 from ferrous iron 664, 666 from ferroUo iron, zinc, and nickel 661 from manganese, nickel, cobalt, and zinc .... 644, 649 from manganese, zinc, cobalt, nickel, and ferrous iron 647 from potassium and sodium 632 from radicals of the fourth group 640 from uranium 675 ferrous, determination 311 as metal 313 by Penny's method 319 volumetrically . 312 with ammonium-ferrous sulphate 315 with oxalic acid ; 316 with permanganate 313 separation from ferric iron 645 formate, basic (ic) 197 hydroxide (ic) 194 oxide (ic) 195 phosphate (ic) 227 separation from copper 683 succinate, basic (ic) 196 sulphide (ous) 195 Kersting's method of determining iodine 540 Kessler's method of determining antimony . . . . 400 of determining arsenic 417 Kolbe's method of determining carbonic acid 493 Lamp, Haste's 82 Lassaigne's method of determining iodine 536 Lead arsenate . 221 INDEX. 773 PAGE Lead carbonate, normal , 201 chloride 203 chromate 152, 225 determination as chloride 357 as chromate 356 as metal 358 as oxide 353 as oxide 4- lead 357 as sulphate 355 as sulphide 354 by Schwarz's method 360 volumetrically 359 oxalate 202 oxide 134, 202 phosphate 227 separation from antimony 714 from bismuth 697 from other metals 689, 690 from silver 693 sulphate 202 sulphide 204 Lenssen's method of determining f erricyanides 556 of determining tin 408 Levigation 52 Liebig's method of determining chlorine 525 of determining cyanogen volumetrically 549 Lime 132 Liquids, reading-off 46 Lithium, determination of 258 separation from other alkalies 605 Litmus, tincture 145 Lowe's method of determining bismuth 385 Magnesium-ammonium arsonate 222 ammonium phosphate 177 chloride , 138 determination as oxide 276 as pyrophosphate 275 as sulphate 275 oxide 179 phosphate 227 pyroarsenate 223 pyrophosphate 178 separation from barium and strontium 617 from calcium 619 from potassium and sodium 610 774 INDEX. PAGE Magnesium, separation from uranium 674 sulphate .- 176 Manganese-ammonium phosphate 188 carbonate ^ 185 determination as carbonate 293 as dioxide 294 as hydroxide 294 as protosesquioxide 293 as pyrophosphate 297 as sulphate 297 as sulphide 295 volumetric-ally 298 with potassium f erricyanide 298 with potassium permanganate 300 dioxide 186 hydroxide (ous) 186 protosesquioxide 186 pyrophosphate 189 separation from alkalies 632 from aluminium and iron 665 from barium and strontium 633, 634, 635, 636 from cobalt and nickel 651 from lead, bismuth, cadmium, and copper 685 from nickel and cobalt 633, 638 from nickel and zinc 644 from zinc 665 sulphate, anhydrous (ous) 188 sulphide 187 Marguerite's method of ferrous determination 312 Maste's Lamp 82 Measuring 26 Mechanical division 51 Mercury 205 chloride (ous) 205 chromate (ous) 226 mercuric, determination as choride (ous) 366 determination as metal 364 as oxide 367 as sulphide 366 by Scherer's method 369 volumetrically , 367 separation from mercury (ous) copper, cadmium, and lead 688 mercurous, determination as chloride 361 determination volumetrically 362 INDEX. 775 PAGE Mercury, mercurous, separation from mercury (ic), copper, cadmium, bismuth, and lead 688 oxide (ic) 134, 207 phosphate (ous) 230 separation from antimony 708 from arsenic and antimony oxides 713 from gold and silver 710 from metals 679 from silver, bismuth, copper, cadmium, and lead 694 sulphide (ic) 206 Metals in cyanides, determination of 553 Mitscherlich's method of determining silicic acid 521 Mohr's burette 42 method of determining antimony 400 of determining arsenic 416 of determining hydrogen sulphide- . 560 of determining sulphuric acid 435 Moisture, influence of upon gases, in reading-off 34 Molybdic acid, see acid molybdic. Mortreux's method of determining sulphur in free state 570 Mviller's modified Schulze's method of determining phosphoric acid as ferric phosphate 452 Neubauer's method of determining phosphoric acid 454 Nickel 190 and cobalt, separation from barium and strontium 633, 634, 638 determination as metal 304 as nickel tripotassium nitrate 307 as oxide and hydroxide 302 as sulphate . 304, 308 as sulphide 303, 307 volumetrically 305, 308 hydroxide (ous) 189 oxide (ous) 189 separation from alkalies 632 from copper 683 from zinc 658 sulphide, hydrated (ous) 190 Nitric acid, see acid nitric. Nitrogen 168 table of absorption . . 259 table of weight of 1 c.c. at different temperatures and pressures 260 Nitrous acid, see acid nitrous. 776 INDEX. PAGE Oil-baths 66 Operations 11 Otto's method of separating phosphoric acid from aluminium 459 Oudeman's method of determining ferric iron 332 Oxalic acid, see acid oxalic. Oxygen 153 Palladium, determination as chloride (ic) 390 as metal 390 iodide (ous) : 237 Pelouze's method of determining nitric acid with ferrous chloride. . . 573 Penfield's method of determining silicon fluorides evolved from fluorides 478 Penny's method of ferrous iron determination 319 Pettenkofer's method of determining carbonic acid 484 Phosphates, see acid phosphoric. Phosphoric acid, see acid phosphoric. Pinchcocks 43, 44 Pipettes 39 Pisani's method of determining chlorine 524 of determining iodine 539 of determining molybdic acid 421 of determining silver 349 Vogel's modification of 351 Platinum , 216 and gold, separation from tin, antimony, and arsenic 716 determination as metal 393 as potassium-platinic chloride 394 as sulphide (ic) 395 separation from gold 716, 727 from metals of groups iv and v in alloys 705 sulphide (ic) 216 Potassa 131 fused 155 solution 155 Potassium borofluoride 232 chloride 162 -cobaltic nitrite 193 cyanide 136 determination as chloride 245 as nitrate 244 as potassium-platinic chloride 245 as silicofluoride 248 as sulphate 243 dichromate 156 disulphate 141 INDEX. 777 Potassium hydroxide . 131 -hydrogen fluoride 141 iodide 148 nitrate 162 permanganate 145 -platinic chloride 163 separation from sodium 599, 604 silicofluoride ' 164 sulphate 161 Precipitates, drying 110 igniting 112 washing 98 Precipitation, effecting 91 .Pressure, influence of, upon gases in reading-off 33 Radicals, determination of 239 Reagents 127 Reimann's method of determining bromine 532 Reinige's method of determining iodine 538 Reissig's method of determining phosphoric acid 448 Rheineck's method of determining ferrocyanides 557 Rivot's method of determining copper 376 Rivot-Beudant-Daguin's method of determining sulphur 568 Rose's method of determining arsenous acid 419 carbonic acid 496 oxalic acid 470 phosphoric acid 448 Rose-Finkener's method of determining cyanogen in mercuric cyanide 552 Samples, selection of 50 Scheibler's method of determining carbonic acid 500 Scherer's method of determining mercury (ic) "... . . 369 Schneider's method of determining antimony 403 Schlosing's method of determining nitric acid 579 Schulze's method of determining nitric acid 5S2 nitric acid from loss of hydrogen .... 588 phosphoric acid as magnesium phos- phate 453 Schwarz's method of determining chromic acid 424 copper 3S1 , 3S2 free iodine 542, 543 lead 360 Selenium, determination 429 Selenous acid, see acid selenous. Sifting 53 Silica Csee also acid silicic) 233 778 INDEX. PAGE Silicic acid, see acid silicic. Siewert/s method of determining nitric acid as ammonia . 587 Silver 1 50, 198 bromide 236 chloride 198 cyanide 201 determination as chloride 338 as cyanide 341 as metal 341 as sulphide , 340 by Gay-Lussac's method 342 by Pisani's method 349 volumctrically 342 oxide 236 phosphate, normal 230 separation by cupellation 698* from copper, cadmium, bismuth, mercury, and lead.. 686 from gold 713 from lead 693 from mercury (ic), copper, and cadmium 690 from metals 679 sulphide 200 Smith's method of determining silicic acid 519 Soda 131 -lime 153, 154 Sodium carbonate 135 anhydrous 166 chloride 150, 165 determination as carbonate 250 as chloride 250 as nitrate 249 as sulphate 249 disulphate 141 hydroxide 131 nitrate 165 -platinic chloride 166 separation from ammonium 603 from potassium 599, 604 silicofluoride , 167 sulphate, anhydrous 164 thiosulphate 135 Solution of substances 79 potassa 155 stannous chloride for ferric-iron determination 329 Sonnenschein's method of determining phosphoric acid 446 Starch iodide . 349 INDEX. 779 PAGE Stolba's method of determining hydrofluosilicic acid 443 Strontium carbonate e 172 determination as carbonate 267 as sulphate 266 separation from calcium 619^ @21 from potassium and sodium 607, 609 sulphate 17': Struve's method of determining iodine colorimetrically 541 Substances, converting into weighable forms 81 Sulphides, determining in presence of carbonates 742 in silicates 742 Sulphur, determination as hydrogen sulphide 558, 562, 569 by Berzelius-Rose's method 564 by Bunsen's method 566 by Rivot-Beudant-Daguin's method 568 in free state by Mortreux's method 570 in sulphides 562, 569 in sulphides, separation from chlorine 756 Sulphuric acid, see acid sulphuric. Sulphurous acid, see acid sulphurous. Table of absorption of carbonic acid 508 of weight of 1 c.c. of carbonic acid at various temperatures and pressures 506 Temperature, influence of on gases, in reading-off 33 Tirmann-Schulze's method of determining nitric acid 582 Thiosulphuric acid, see acid thiosulphuric. Tin, determination as oxide (ic) 405 as stannic (or metastannic) acid ...... 405 as sulphide (ous or ic) 406 by alkaline iodine solution 408 by Lenssen's method 408 with ferric chloride 408 volumetrically 407 oxide (ic) 219 phosphate (ic) 230 separation from antimony 716, 725 from antimony and arsenic 723, 726, 72S from arsenic 717, 728 from gold 727 from metals of groups i, n, and in 707 from metals of groups iv and v 706 from stannic tin 730 sulphide, hydrated (ic) and (ous) 220 Titanium, determination of 2s4 Tripotassium cobaltic nitrite 193 Uranium acetate : 139 OF THE UNIVERSITY 780 INDEX. PAGE Uranium determination 335 by Belohoubeck's method 336 separation from aluminium 674 from barium, calcium, and strontium 673 from chromium 674 from cobalt, nickel, and zinc 675 from iron (ic) 675 from magnesium 674 from other metals of groups i-iv 672 Uranyl pyroarsenate 223 pyrophosphate 229 Vogel's modification of Pisani's method 351 Vohl's method of determining arsenous acid 419 of determining chromic acid .-423 Wackenroder-Fresenius' method of separating phosphoric acid from aluminium 459' Water, distilled 127 estimating. 72 Water-bath 58 Weeren's method of determining phosphoric acid by Miiller's modi- fied Schulze's method 452 Weighing, process of 21, 70 Weights, testing, etc .* 19 Weil's method of determining copper 380 Wells' apparatus for determining carbonic acid 499 Werther's method of determining arsenic as uranyl pyroarsenate 413 of examining gunpowder residues 742 Wildenstein's method of determining sulphuric acid 437, 438 Wohler's method of determining silicon fluoride evolved from fluo- rides 478 Zinc 132 carbonate, basic 182 oxide 183 sulphide 184 determination as carbonate 287 as oxide 287 as sulphide . . . . .288, 289 separation from aluminium and manganese 649 from barium and strontium 633, 634 from cadmium 684 from calcium 633, 634 from copper ' 683 from iron in alloys 660 from nickel, cobalt, and manganese 650 from potassium and sodium 632 SHORT-TITLE CATALOGUE OF THE PUBLICATIONS OF JOHN WILEY & SONS NEW YORK LONDON: CHAPMAN & HALL, LIMITED ARRANGED UNDER SUBJECTS Descriptive circulars sent on application. Books marked with an asterisk (*) are Old at net prices only. All books are bound in cloth unless otherwise stated. AGRICULTURE HORTICULTURE FORESTRY. Armsby's Principles of Animal Nutrition 8vo, $4 00 Budd and Hansen's American Horticultural Manual: Part I. Propagation, Culture, and Improvement 12mo, 1 50 Part II. Systematic Pomology 12mo, 1 50 Elliott's Engineering for Land Drainage 12mo, 1 50 Practical Farm Drainage. (Second Edition, Rewritten.) 12mo, 1 50 Graves's Forest Mensuration 8vo, 4 00 Green's Principles of American Forestry 12mo, 1 50 Grotenfelt's Principles of Modern Dairy Practice. (Woll.) 12mo, 2 00 * Herrick's Denatured or Industrial Alcohol 8vo, 4 00 Kemp and Waugh's Landscape Gardening. (New Edition, Rewritten. In Preparation.) * McKay and Larsen's Principles and Practice of Butter-making 8vo, 1 50 Maynard's Landscape Gardening as Applied to Home Decoration 12mo, 1 50 Sanderson's Insects Injurious to Staple Crops 12mo, I 50 Sanderson and Headlee's Insects Injurious to Garden Crops. (In Prepa- ration.) * Schwarz's Longleaf Pine in Virgin Forests 12mo, 1 25 Stockbridge's Rocks and Soils 8vo, 2 50 Winton's Microscopy of Vegetable Foods 8vo, 7 50 Woll's Handbook for Farmers and Dairymen 16mo, 1 50 ARCHITECTURE. Baldwin's Steam Heating for Buildings 12mo, 2 50 Berg's Buildings and Structures of American Railroads 4to, 5 00 Birkmire's Architectural Iron and Steel 8vo, 3 50 Compound Riveted Girders as Applied in Buildings 8vo, 2 00 Planning and Construction of American Theatres 8vo, 3 00 Planning and Construction of High Office Buildings 8vo, 3 50 Skeleton Construction in Buildings 8vo, 3 00 1 Briggs's Modern American School Buildings 8vo, $4 00 Byrne's Inspection of Materials and Wormanship Employed in Construction. 16mo, 3 00 Carpenter's Heating and Ventilating of Buildings 8vo, 4 00 * Corthell's Allowable Pressure on Deep Foundations 12mo, 1 25- Freitag's Architectural Engineering 8vo, 3 50 Fireproofing of Steel Buildings 8vo, 2 50 Gerhard's Guide to Sanitary Inspections. (Fourth Edition, Entirely Re- vised and Enlarged.) 12mo, 1 5O * Modern Baths and Bath Houses 8vo, 3 00 Sanitation of Public Buildings 12mo, 1 50 Theatre Fires and Panics 12mo, 1 5O * The Water Supply, Sewerage and Plumbing of Modern City Buildings. 8vo, 4 OO Johnson's Statics by Algebraic and Graphic Methods 8vo, 2 00 Kellaway 's How to Lay Out Suburban Home Grounds 8vo, 2 00 Kidder's Architects' and Builders' Pocket-book 16mo, mor., 5 OO Merrill's Stones for Building and Decoration 8vo, 5 00 Monckton's Stair-building : 4to, 4 OO Patton's Practical Treatise on Foundations 8vo, 5 00 Peabody's Naval Architecture 8vc, 7 50 Rice's Concrete-block Manufacture 8vo, 2 OO Richey's Handbook for Superintendents of Construction 16mo, mor. 4 00 Building Foreman's Pocket Book and Ready Reference. . 16mo, mor. 5 OO * Building Mechanics' Ready Reference Series: * Carpenters' and Woodworkers' Edition 16mo, mor. * Cement Workers' and Plasterers' Edition 16mo, mor. * Plumbers', Steam-Fitters', and Tinners' Edition. . . 16mo, mor. 50 50 50 50 00 * Stone- and Brick-masons' Edition. 16mo, mor. Sabin's House Painting 12mo, Siebert and Biggin's Modern Stone-cutting and Masonry 8vo, 1 5O Snow's Principal Species of Wood 8vo, 3 5O Towne's Locks and Builders' Hardware 18mo, mor. 3 OO Wait's Engineering and Architectural Jurisprudence 8vo, 6 00 Sheep, 6 5O Law of Contracts 8vo, 3 00 Law of Operations Preliminary to Construction in Engineering and Architecture 8vo, 5 00 Sheep, 5 50 Wilson's Air Conditioning 12mo, 1 50 Worcester and Atkinson's Small Hospitals, Establishment and Maintenance, Suggestions for Hospital Architecture, with Plans fora Small Hospital 12mo, 1 25 ARMY AND NAVY. Bernadou's Smokeless Powder, Nitro-cellulose, and the Theory of the Cellu- lose Molecule 12mo, 2 50 Chase's Art of Pattern Making 12mo, 2 50 Screw Propellers and Marine Propulsion 8vo, 3 00 * Cloke's Enlisted Specialists' Examiner 8vo, 2 00 Gunner's Examiner 8vo, 1 50 Craig's Azimuth 4to, 3 50 Crehore and Squier's Polarizing Photo-chionograph 8vo, 3 50' * Davis's Elements of Law 8vo, 2 5O * Treatise on the Military Law of United States 8vo, 7 OO DeBrack's Cavalry Outpost Duties. (Carr.) 24mo, mor. 2 OO * Dudley's Military Law and the Procedure of Courts-martial. ..Large 12mo, 2 SO Durand's Resistance and Propulsion of Ships 8vo, 5 00 * Dyer's Handbook of Light Artillery 12mo, 3 00 Eissler's Modern High Explosives 8vo, 4 00 * Fiebeger's Text-book on Field Fortification Large 12mo, 2 00 Hamilton and Bond's The Gunner's Catechism 18mo, 1 00 *Hoff's Elementary Naval Tactics 8vo, 1 5O 2 Ingalls's Handbook of Problems in Direct Fire 8vo, $4 00 * Lissak's Ordnance and Gunnery 8vo, 6 00 * Ludlow's Logarithmic and Trigonometric Tables 8vo, 1 00 * Lyons's Treatise on Electromagnetic Phenomena. Vols. I. and II..8vo,each, 6 00 * Mahan's Permanent Fortifications. (Mercur.) 8vo, half mor. 7 50 Manual for Courts-martial IGmo.mor. 1 50 * Mercur's Attack of Fortified Places 12mo, 2 00 * Elements of the Art of War 8vo, 4 00 Nixon's Adjutants' Manual 24mo, 1 00 Peabody 's Naval Architecture 8vo, 7 50 * Phelps's Practical Marine Surveying 8vo, 2 50 Putnam's Nautical Charts 8vo, 2 00 Rust's Ex-meridian Altitude, Azimuth and Star-Finding Tables 8vo 5 00 Sharpe's Art of Subsisting Armies in War 18mo, mor. 1 50 * Tupes and Poole's Manual of Bayonet Exercises and Musketry Fencing. 24mo, leather. 50 * Weaver's Military Explosives 8vo. . 3 00 * Woodhuil's Military Hygiene for Officers of the Line Large 12mo, 1 50 ASSAYING. Betts's Lead Refining by Electrolysis 8vo, 4 00 Fletcher's Practical Instructions in Quantitative Assaying with the Blowpipe. 16mo, mor. 1 50 Furman and Pardoe's Manual of Practical Assaying. (Sixth Edition, Re- vised and Enlarged.) 8vo, 3 00 Lodge's Notes on Assaying and Metallurgical Laboratory Experiments.. 8vo, 3 00 Low's Technical Methods of Ore Analysis 8vo, 3 00 Miller's Cyanide Process 12mo, 1 00 Manual of Assaying 12mo, 1 00 Minet's Production of Aluminum and its Industrial Use. (Waldo.). . .12mo, 2 50 O'Driscoll's Notes on the Treatment of Gold Ores 8vo, 2 00 Ricketts and Miller's Notes on Assaying 8vo, 3 00 Robine and Lenglen's Cyanide Industry. (Le Clerc.) 8vo, 4 00 Seamen's Manual for Assayers and Chemists. (In Press.) Ulke's Modern Electrolytic Copper Refining 8vo, 3 00 Wilson's Chlorination Process 12mo, 1 50 Cyanide Processes , 12mo, 1 50 ASTRONOMY, Comstock's Field Astronomy for Engineers 8vo, 2 50 Craig's Azimuth 4to, 3 50 Crandall 's Text-book on Geodesy and Least Squares 8vo, 3 00 Doolittle's Treatise on Practical Astronomy 8vo, 4 00 Hayford's Text-book of Geodetic Astronomy. 8vo, 3 00 Hosmer's Azimuth 16mD, mor. 1 00 Merriman's Elements of Precise Surveying and Geodesy * . .8vo, 2 50 * Michie and Harlow's Practical Astronomy 8vo, 3 00 Rust's Ex-meridian Altitude, Azimuth and Star-Finding Tables 8vo, 5 00 * White's Elements of Theoretical and Descriptive Astronomy 12mo, 2 00 CHEMISTRY. * Abderhalden's Physiological Chemistry in Thirty Lectures. (Hall and Defren.) 8vo, 5 00 * Abegg's Theory of Electrolytic Dissociation, (von Ende.) 12mo, 1 25 Alexeyeff's General Principles of Organic Syntheses. (Matthews.) 8vo, 3 00 Allen's Tables for Iron Analysis 8vo, 3 00 Armsby's Principles of Animal Nutrition 8vo, 4 00 Arnold's Compendium of Chemistry. (Mandel.) Large 12mo, 3 50 3 Association of State and National Food and Dairy Departments, Hartford Meeting, 1906 8vo, $3 00 Jamestown Meeting, 1907 8vo, 3 00 Austen's Notes for Chemical Students 12mo, 1 50 Baskerville's Chemical Elements. (In Preparation.) Bernadou's Smokeless Powder. Nitre-cellulose, and Theory of the Cellulose Molecule 12mo, 2 50 * Biltz's Introduction to Inorganic Chemistry. (Hall and Phelan.). . . 12mo, 1 25 Laboratory Methods of Inorganic Chemistry. (Hall and Blanchard.) 8vo, 3 00 * Blanchard's Synthetic Inorganic Chemistry 12mo, 1 00 * Browning's Introduction to the Rarer Elements 8vo, 1 50 * Claassen's Beet-sugar Manufacture. (Hall and Rolfe.) 8vo, 3 00 Classen's Quantitative Chemical Analysis by Electrolysis. (Boltwood.).8vo, 3 00 Cohn's Indicators and Test-papers 12mo, 2 00 Tests and Reagents 8vo, 3 00 * Danneel's Electrochemistry. (Merriam.) 12mo, 1 25 Dannerth's Methods of Textile Chemistry 12mo, 2 00 Duhem's Thermodynamics and Chemistry. (Burgess.) 8vo, 4 00 Effront's Enzymes and their Applications. (Prescott.) 8vo, 3 00 Eissler's Modern High Explosives 8vo, 4 00 Erdmann's Introduction to Chemical Preparations. (Dunlap.) 12mo, 1 25 * Fischer's Physiology of Alimentation Large 12mo, 2 00- Fletcher's Practical Instructions in Quantitative Assaying with the Blowpipe. 12mo, mor. 1 50- Fowler's Sewage Works Analyses 12mo, 2 00 Fresenius's Manual of Qualitative Chemical Analysis. (Wells.) 8vo, 5 00 Manual of Qualitative Chemical Analysis. Part I. Descriptive. ( Wells. )8vo, 3 00- Quantitative Chemical Analysis. (Cohn.) 2 vols 8vc, 12 50 When Sold Separately, Vol. I, $6. Vol. II, $8. Fuertes's Water and Public Health 12mo, 1 50' Furman and Pardoe's Manual of Practical Assaying. (Sixth Edition, Revised and En larged.) 8vo, 3 00 * Getnian's Exercises in Physical Chemistry 12mo, 2 00 Gill's Gas and Fuel Analysis for Engineers 12mo, 1 25 * Gooch and Browning's Outlines of Qualitative Chemical Analysis. Large 12mo, 1 25 Grotenfelt's Principles of Modern Dairy Practice. (Woll.) 12mo, 2 00 Groth's Introduction to Chemical Crystallography (Marshall) 12mo, 1 25 Hammarsten's Text-book of Physiological Chemistry. (Mandel.) 8vo, 4 00 Hanausek's Microscopy of Technical Products. (Winton.) 8vo, 5 00 * Haskins and Macleod's Organic Chemistry 12mo, 2 00- Hering's Ready Reference Tables (Conversion Factors) 16mo, mor. 2 50 * Herrick's Denatured or Industrial Alcohol 8vo, 4 00 Hinds's Inorganic Chemistry 8vo, 3 00 * Laboratory Manual for Students 12mo, 1 00 * Holleman's Laboratory Manual of Organic Chemistry for Beginners. (Walker.) 12mo, 1 00- Text-book of Inorganic Chemistry. (Cooper.) 8vo, 2 50 Text-book of Organic Chemistry. (Walker and Mott.) 8vo, 2 50 * Holley's Lead and Zinc Pigments Large 12mo, 3 00 Holley and Ladd's Analysis of Mixed Paints, Color Pigments, and Varnishes. Large 12mo, 2 50' Hopkins's Oil-chemists' Handbook 8vo, 3 00 Jackson's Directions for Laboratory Work in Physiological Chemistry. .8vo, 1 25 Johnson's Rapid Methods for the Chemical Analysis of Special Steels, Steel- making Alloys and Graphite Large 12mo, 3 00' Landauer's Spectrum Analysis. (Tingle.) 8vo, 3 00 * Langworthy and Austen's Occurrence of Aluminum in Vegetable Prod- ucts, Animal Products, and Natural Waters 8vo, 2 00 Lassar-Cohn's Application of Some General Reactions to Investigations in Organic Chemistry. (Tingle.) 12mo, 1 00 Leach's Inspection and Analysis of Food with Special Reference to State Control 8vo, 7 50 Lob's Electrochemistry of Organic Compounds. (Lorenz.) 8vo, 3 00 Lodge's Notes on Assaying and Metallurgical Laboratory Experiments.. 8vo, 3 00 Low's Technical Method of Ore Analysis 8vo, 3 00- 4 Lowe's Paint for Steel Structures 12mo, $1 00 Lunge's Techno-chemical Analysis. (Cohn.) 12mo, 1 00 * McKay and Larsen's Principles and Practice of Butter-making 8vo, 1 50 Maire's Modern Pigments and their Vehicles 12mo, 2 00 Mandel's Handbook for Bio-chemical Laboratory 12mo, 1 50 * Martin's Laboratory Guide to Qualitative Analysis with the Blowpipe 12mo, 6O Mason's Examination of Water. (Chemical and Bacteriological.) 12mo, 1 25 Water-supply. (Considered Principally from a Sanitary Standpoint.) 8vo, 4 00 * Mathewson's First Principles of Chemical Theory 8vo, 1 00 Matthews's Laboratory Manual of Dyeing and Textile Chemistry 8vo, 3 50 Textile Fibres. 2d Edition, Rewritten 8vo, 4 00 * Meyer's Determination of Radicles in Carbon Compounds. (Tingle.) Third Edition 12mo, 1 25 Miller's Cyanide Process 12mo, 1 00 Manual of Assaying 12mo, 1 00 Minet's Production of Aluminum and its Industrial Use. (Waldo.). ..12mo, 2 50 * Mittelstaedt's Technical Calculations for Sugar Works. (Bourbakis.) 12mo, 1 50 Mixter's Elementary Text-book of Chemistry 12mo, 1 50 Morgan's Elements of Physical Chemistry 12mo, 3 00 Outline of the Theory of Solutions and its Results 12mo, 1 00 * Physical Chemistry for Electrical Engineers 12mo, 1 50 Morse's Calculations used in Cane-sugar Factories 16mo, mor. 1 50 * Muir's History of Chemical Theories and Laws 8vo, 4 00 Mulliken's General Method for the Identification of Pure Organic Compounds. Vol. I. Compounds of Carbon with Hydrogen and Oxygen. Large 8vo, 5 00 Vol. II. Nitrogenous Compounds. (In Preparation.) Vol. III. The Commercial Dyestuffs Large 8vo, 5 00 O'Driscoll's Notes on the Treatment of Gold Ores 8vo, 2 00 Ostwald's Conversations on Chemistry. Part One. (Ramsey.) 12mo, 1 50 Part Two. (Turnbull.) 12mo, 2 00 Owen and Standagc's Dyeing and Cleaning of Textile Fabrics 12mo, 2 00 * Palmer's Practical Test Book of Chemistry 12mo, 1 00 * Pauli's Physical Chemistry in the Service of Medicine. (Fischer.) . . 12mo, 1 25 Penfield's Tables of Minerals, Including the Use of Minerals and Statistics of Domestic Production 8vo, 1 00 Pictet's Alkaloids and their Chemical Constitution. (Biddle.) 8vo, 5 00 Poole's Calorific Power of Fuels 8vo, 3 00 Prescott and Winslow's Elements of Water Bacteriology, with Special Refer- ence to Sanitary Water Analysis 12mo, 1 50 * Reisig's Guide to Piece-Dyeing '. . 8vo, 25 00 Richards and Woodman's Air, Water, and Food from a Sanitary Stand- point 8vo. 2 00 Ricketts and Miller's Notes on Assaying 8vo, 3 00 Rideal's Disinfection and the Preservation of Food 8vo, 4 00 Sewage and the Bacterial Purification of Sewage 8vo, 4 00 Rigg's Elementary Manual for the Chemical Laboratory 8vo, 1 25 Robine and Lenglen's Cyanide Industry. (Le Clerc.) 8vo, 4 00 Ruddiman's Incompatibilities in Prescriptions 8vo, 2 00 Whys in Pharmacy 12mo, 1 00 * Ruer's Elements of Metallography. (Mathewson.) 8vo, 3 00 Sabin's Industrial and Artistic Technology of Paint and Varnish 8vo, 3 00 Salkowski's Physiological and Pathological Chemistry. (Orndorff.) 8vo, 2 50 Schimpf's Essentials of Volumetric Analysis 12mo, 1 25 Manual of Volumetric Analysis. (Fifth Edition, Rewritten) 8vo, 5 00 * Qualitative Chemical Analysis 1 ... .8vo, 1 25 Seamen's Manual for Assayers and Chemists. (In Press.) Smith's Lecture Notes on Chemistry for Dental Students 8vo, 2 50 Spencer's Handbook for Cane Sugar Manufacturers 16mo, mor. 3 00 Handbook for Chemists of Beet-sugar Houses 16mo, mor. 3 OO Stockbridge's Rocks and Soils 8vo, 2 5O Stone's Practical Testing of Gas and Gas Meters 8vo, 3 50 * Tillman's Descriptive General Chemistry 8vo, 3 00 * Elementary Lessons in Heat 8vo, 1 5O Treadwell's Qualitative Analysis. (Hall.) 8vo. 3 OO Quantitative Analysis. (Hall.) 8vo, 4 00 Turneaure and Russell's Public Water-supplies 8vo, $5 00 Van Deventer's Physical Chemistry for Beginners. (Boltwood.) 12mo, 1 50 Venable's Methods and Devices for Bacterial Treatment of Sewage 8vo, 3 00 Ward and Whipple's Freshwater Biology. (In Press.) Ware's Beet-sugar Manufacture and Refining. Vol. 1 8vo, 4 00 Vol. II 8vo, 5 00 Washing; o 1's Manual of the Chemical Analysis of Rocks 8vo, 2 00 * Weaver's Military Explosives 8vo, 3 00 Wells's Laboratory Guide in Qualitative Chemical Analysis 8vo, 1 50 Short Course in Inorganic Qualitative Chemical Analysis for Engineering Students 12mo, 1 50 Text-book of Chemical Arithmetic 12mo, 1 25 Whipple's Microscopy of Drinking-water 8vo, 3 50 Wilson's Chlorination Process 12mo, 1 50 Cyanide Processes 12mo, 1 50 Winton's Microscopy of Vegetables Food 8vo, 1 50 Zsigmondy's Colloids and the Ultramicroscope. (Alexander.). .Large 12mo, 3 00 CIVIL ENGINEERING. BRIDGES AND ROOFS. HYDRAULICS. MATERIALS OF ENGINEER- ING. RAILWAY ENGINEERING. Baker's Engineers' Surveying Instruments. 12mo, 3 00 Bixby's Graphical Computing Table Paper 19 X 24J inches.. 25 Breed and Hosmer's Principles and Practice of Surveying. Vol. I. Elemen- tary Surveying 8vo, 3 00 Vol. II. Higher Surveying 8vo, 2 50 * Burr's Ancient and Modern Engineering and the Isthmian Canal 8vo, 3 50 Comstock's Field Astronomy for Engineers 8vo, 2 50 * Corthell's Allowable Pressure on Deep Foundations 12mo, 1 25 Crandall's Text-book on Geodesy and Least Squares 8vo, 3 00 Davis's Elevation and Stadia Tables 8vo, 1 00 Elliott's Engineering for Land Drainage 12mo, 1 50 Practical Farm Drainage. (Second Edition Rewritten.) 12mo. 1 50 * Fiebeger's Treatise on Civil Engineering 8vo, 5 00 Flemer's Photographic Methods and Instruments 8vo, 5 00 Folwell's Sewerage. (Designing and Maintenance.) 8vo, 3 00 Freitag's Architectural Engineering 8vo, 3 50 Goodhue's Municipal Improvements 12mo, 1 50 * Hauch and Rice's Tables of Quantities for Preliminary Estimates. . . 12mo, 1 25 Hayford's Text-book of Geodetic Astronomy 8vo, 3 00 Hering's Ready Reference Tables (Conversion Factors.) 16mo, mor. 2 50 Hosmer's Azimuth 16mo, mor. 1 00 Howe' Retaining Walls for Earth 12mo, 1 25 * Ives's Adjustments of the Engineer's Transit and Level 16mo, bds. 25 Johnson's (J. B.) Theory and Practice of Surveying Large 12mo, 4 00 Johnson's (L. J.) Statics by Algebraic and Graphic Methods 8vo, 2 00 Kinnicutt, Winslow and Pratt's Purification of Sewage. (In Preparation.) * Mahan's Descriptive Geometry 8vo, 1 50 Merriman's Elements of Precise Surveying and Geodesy 8vo, 2 50 Merriman and Brooks's Handbook for Surveyors 16mo, mor. 2 00 Nugent's Plane Surveying 8vo, 3 50 Ogden's Sewer Construction 8vo, 3 00 Sewer Design 12mo, 2 00 Parsons's Disposal of Municipal Refuse 8vo, 2 00 Patton's Treatise on Civil Engineering 8vo, half leather, 7 50 Reed's Topographical Drawing and Sketching 4to, 5 00 Rideal's Sewage and the Bacterial Purification of Sewage 8vo, 4 00 Riemer's Shaft-sinking under Difficult Conditions. (Co/ning and Peele.).8vo, 3 00 Siebert and Biggin's Modern Stone-cutting and Masonry 8vo, 1 50 Smith's Manual of Topographical Drawing. (McMillan.) 8vp, 2 50 6 Soper's Air and Ventilation of Subways 12mo, $2 50 * Tracy's Exercises in Surveying 12mo, mor. 1 00 Tracy's Plane Surveying 16mo, mor. 3 00 * Trautwine's Civil Engineer's Pocket-book 16mo, mor. 5 00 Venable's Garbage Crematories in America 8vo, 2 00 Methods and Devices for Bacterial Treatment of Sewage 8vo, 3 00 Wait's Engineering and Architectural Jurisprudence 8vo, 6 00 Sheep, 6 50 Law of Contracts 8vo, 3 00 Law of Operations Preliminary to Construction in Engineering and Architecture 8vo, 5 OO Sheep, 5 5O Warren's Stereo tomy Problems in Stone-cutting 8vo, 2 50 * Waterbury's Vest-Pocket Hand-book of Mathematics for Engineers. 2JX5f inches, mor. 1 OO * Enlarged Edition, Including Tables mor. 1 50 Webb's Problem's in the Use and Adjustment of Engineering Instruments. 16mo, mor. 1 25 Wilson's Topographic Surveying 8vo, 3 50 BRIDGES AND ROOFS. .-\ Boiler's Practical Treatise on the Construction of Iron Highway Bridges.. 8vo, 2 00 * Thames River Bridge Oblong paper, 5 00 Burr and Falk's Design and Construction of Metallic Bridges 8vo, 5 00 Influence Lines for Bridge and Roof Computations 8vo, 3 00 Du Bois's Mechanics of Engineering. Vol. II Small 4to, 10 00 Foster's Treatise on Wooden Trestle Bridges 4to, 5 00 Fowler's Ordinary Fopndations 8vo, 3 50 Greene's Arches in Wood, Iron, and Stone 8vo, 2 50 Bridge Trusses 8vo, 2 50 Roof Trusses 8vo, 1 25 Grimm's Secondary Stresses in Bridge Trusses 8vo, 2 50 Heller's Stresses in Structures and the Accompanying Deformations. . . . 8vo, 3 00 Howe's Design of Simple Roof-trusses in Wood and Steel 8vo. 2 00 Symmetrical Masonry Arches 8vo, 2 50 Treatise on Arches 8vo, 4 00 * Jacoby's Structural Details, or Elements of Design in Heavy Framing, 8vo, 2 25 Johnson, Bryan and Turneaure's Theory and Practice in the Designing of Modern Framed Structures Small 4to, 10 00 Merriman and Jacoby's Text-book on Roofs and Bridges: Part I. Stresses in Simple Trusses 8vo v 2 50 Part II. Graphic Statics 8vo, 2 50 Part III. Bridge Design 8vo, 2 50 Part IV. Higher Structures 8vo, 2 50 Morison's Memphis Bridge Oblong 4to, 10 00 Sondericker's Graphic Statics, with Applications to Trusses, Beams, and Arches 8vo, 2 00 Waddell's De Pontibus, Pocket-book for Bridge Engineers 16mo, mor. 2 00 * Specifications for Steel Bridges 12mo, 50 Waddell and Harringtoon's Bridge Engineering. (In Preparation.) Wright's Designing of Draw-spans. Two parts in one volume 8vo, 3 50 HYDRAULICS. Barnes's Ice Formation 8vo, 3 00 Bazin's Experiments upon the Contraction of the Liquid Vein Issuing from an Orifice. (Trautwine.) 8vo, 2 00 Bovey's Treatise on Hydraulics 8vo, 5 00 Church's Diagrams of Mean Velocity of Water in Open Channels. Oblong 4to, paper, 1 50 Hydraulic Motors 8vo, 2 00 7 Coffin's Graphical Solution of Hydraulic Problems l6mo, mor. $2 50 Flather's Dynamometers, and the Measurement of Power 12mo, 3 00 Folwell's Water-supply Engineering 8vo, 4 00 Frizell's Water-power 8vo, 5 00 Fuertes's Water and Public Health 12mo, 1 50 Water-filtration Works 12mo, 2 50 Ganguillet and Kutter's General Formula for the Uniform Flow of Water in Rivers and Other Channels. (Hering and Trautwine.) 8vo, 4 00 Hazen's Clean Water and How to Get It Large 12mo, 1 50 Filtration of Public Water-supplies 8vo, 3 00 Hazelhurst's Towers and Tanks for Water-works 8vo, 2 50 Herschel's 115 Experiments on the Carrying Capacity of Large, Riveted, Metal Conduits 8vo, 2 00 Hoyt and Grover's River Discharge 8vo, 2 00 Hubbard and Kiersted's Water-works Management and Maintenance. 8vo, 4 00 * Lyndon's Development and Electrical Distribution of Water Power. 8vo, 3 00 Mason's Water-supply. (Considered Principally from a Sanitary Stand- point.) 8vo, 4 00 Merriman's Treatise on Hydraulics 8vo, 5 00 * Molitor's Hydraulics of Rivers, Weirs and Sluices 8vo, 2 00 * Richards's Laboratory Notes on Industrial Water Analysis 8vo, 50 Schuyler's Reservoirs for Irrigation, Water-power, and Domestic Water- supply. Second Edition, Revised and Enlarged Large 8vo, 6 00 * Thomas and Watt's Improvement of Rivers 4to, 6 00 Turneaure and Russell's Public Water-supplies 8vo, 5 00 Wegmann's Design and Construction of Dams. 5th Ed., enlarged 4to, 6 CO Water-Supply of the City of New York from 1658 to 1895 4 to, 10 00 Whipple's Value of Pure Water Large 12mo, 1 00 Williams and Hazen's Hydraulic Tables 8vo, 1 50 Wilson's Irrigation Engineering 8vo, 4 00 Wood's Turbines 8vo, 2 50 MATERIALS OF ENGINEERING. Baker's Roads and Pavements 8vo, 5 00 Treatise on Masonry Construction 8vo, 5 00 Black's United States Public Works Oblong 4to, 5 00 Blanchard's Bituminous Roads. (In Press.) Bleininger's Manufacture of Hydraulic Cement. (In Preparation.) * Bovey's Strength of Materials and Theory of Structures 8vo, 7 50 Burr's Elasticity and Resistance of the Materials of Engineering, 8vo, 7 50 Byrne's Highway Construction 8vo, 5 00 Inspection of the Materials and Workmanship Employed in Construction. 16mo, 3 00 Church's Mechanics of Engineering 8vo, 6 00 Du Bois's Mechanics of Engineering. Vol. I. Kinematics, Statics, Kinetics Small 4to, 7 50 Vol. II. The Stresses in Framed Structures, Strength of Materials and Theory of Flexures Small 4to, 10 00 * Eckel's Cements, Limes, and Plasters 8vo, 6 00 Stone and Clay Products used in Engineering. (In Preparation.) Fowler's Ordinary Foundations 8vo, 3 50 * Greene's Structural Mechanics 8vo, 2 50 * Holley's Lead and Zinc Pigments Large 12mo, 3 00 Holley and Ladd's Analysis of Mixed Paints, Color Pigments and Varnishes. Large 12mo, 2 50 Hubbard's Dust Preventives and Road Binders. (In Preparation.) Johnson's (C. M.) Rapid Methods for the Chemical Analysis of Special Steels, Steel-making Alloys and Graphite Large 12mo, 3 00 Johnson's (J. B.) Materials of Construction Large 8vo, 6 00 Keep's Cast Iron 8vo, 2 50 Lanza's Applied Mechanics 8vo, 7 50 Lowe's Paint for Steel Structures 12mo, 1 00 8 Ma-re's Modern Pigments and their Vehicles 12mo, $2 00 Martens's Handbook on Testing Materials. (Henning.) 2 vols 8vo, 7 50 Maurer's Technical Mechanics 8vo, 4 00 Merrill's Stones for Building and Decoration 8vo, 5 00 Merriman's Mechanics of Materials 8vo, 5 00 * Strength of Materials 12mo, 1 00 Metcalf 's Steel. A Manual for Steel-users 12mo, 2 00 Morrison's Highway Engineering 8vo, 2 50 Patton's Practical Treatise on Foundations 8vo, 5 00 Rice's Concrete Block Manufacture 8vo, 2 00 Richardson's Modern Asphalt Pavements 8vo, 3 00 Richey's Building Foreman's Pocket Book and Ready Reference. 16mo,mor. 5 00 * Cement Workers' and Plasterers' Edition (Building Mechanics' Ready Reference Series) 16mo, mor. 1 50 Handbook for Superintendents of Construction 16mo, mor. 4 00 * Stone and Brick Masons' Edition (Building Mechanics' Ready Reference Series) 16mo, mor. 1 50 * Ries's Clays : Their Occurrence, Properties, and Uses 8vo, 5 00 * Ries and Leighton's History of the Clay-working Industry of the United States 8vo. 2 50 Sabin's Industrial and Artistic Technology of Paint and Varnish 8vo, 3 00 * Smith's Strength of Material 12mo 1 25 Snow's Principal Species of Wood 8vo, 3 50 Spalding's Hydraulic Cement 12mo, 2 00 Text-book on Roads and Pavements 12mo, 2 00 Taylor and Thompson's Treatise on Concrete, Plain and Reinforced 8vo, 5 00 Thurston's Materials of Engineering. In Three Parts 8vo, 8 00 Part I. Non-metallic Materials of Engineering and Metallurgy.. . .8vo, 2 00 Part II. Iron and Steel 8vo, 3 50 Part III. A Treatise on Brasses, Bronzes, and Other Alloys and their Constituents 8vo, 2 50 Tillson's Street Pavements and Paving Materials 8vo, 4 00 * Trautwine's Concrete, Plain and Reinforced 16mo, 2 00 Turneaure and Maurer's Principles of Reinforced Concrete Construction. Second Edition, Revised and Enlarged 8vo, 3 50 Waterbury's Cement Laboratory Manual 12mo, 1 00 Wood's (De V.) Treatise on the Resistance of Materials, and an Appendix on the Preservation of Timber 8vo, 2 00 Wood's (M. P.) Rustless Coatings: Corrosion and Electrolysis of Iron and Steel 8vo, 4 00 RAILWAY ENGINEERING. Andrews's Handbook for Street Railway Engineers 3X5 inches, mor. 1 25 Berg's Buildings and Structures of American Railroads 4to, 5 00 Brooks's Handbook of Street Railroad Location 16mo, mor. 1 50 Butts's Civil Engineer's Field-book 16mo, mor. 2 50 Crandall's Railway and Other Earthwork Tables 8vo, 1 50 Transition Curve 16mo, mor. 1 50 * Crockett's Methods for Earthwork Computations 8vo, J 50 Dredge's History of the Pennsylvania Railroad. (1879) Papei 5 00 Fisher's Table of Cubic Yards Cardboard, 25 Godwin's Railroad Engineers' Field-book and Explorers' Guide. . 16mo, mor. 2 50 Hudson's Tables for Calculating the Cubic Contents of Excavations and Em- bankments 8vo, 1 00 Ives and Hilts's Problems in Surveying, Railroad Surveying and Geodesy 16mo, mor. 1 50 Molitor and Beard's Manual for Resident Engineers 16mo, 1 00 Nagle's Field Manual for Railroad Engineers 16mo, mor. 3 00 * Orrock's Railroad Structures and Estimates 8vo, 3 00 Philbrick's Field Manual for Engineers 16mo, mor. 3 00 Raymond's Railroad Engineering. 3 volumes. Vol. I. Railroad Field Geometry. (In Preparation.) Vol. II. Elements of Railroad Engineering 8vo, 3 50 Vol. III. Railroad Engineer's Field Book. (In Preparation.) 9 Searles's Field Engineering 16mo, mor. $3 00 Railroad Spiral 16mo, mor. 1 50 Taylor's Prismoidal Formulae and Earthwork 8vo, 1 50 * Trautw ; ne's Field Practice of Laying Out Circular Curves for Railroads. 12mo, mor. 2 50 * Method of Calculating the Cubic Contents of Excavations and Em- bankments by the Aid of Diagrams 8vo, 2 00 Webb's Economics of Railroad Construction Large 12mo, 2 50 Railroad Construction 16mo, mor. 5 00 Wellington's Economic Theory of the Location of Railways Large 12mo, 5 00 Wilson's Elements of Railroad-Track and Construction 12mo, 2 00 DRAWING. Barr's Kinematics of Machinery. , 8vo, 2 50 * Bartlett's Mechanical Drawing 8vo, 3 00 * " AbridgedEd 8vo, 150 Coolidge's Manual of Drawing 8vo, paper, 1 00 Coolidge and Freeman's Elements of General Drafting for Mechanical Engi- neers Oblong 4to, 2 50 Durley's Kinematics of Machines 8vo, 4 00 Emch's Introduction to Projective Geometry and its Application 8vo, 2 50 French and I ves' Stereotomy 8vo, 2 50 Hill's Text-book on Shades and Shadows, and Perspective 8vo, 2 00 Jamison's Advanced Mechanical Drawing 8vo, 2 00 Elements of Mechanical Drawing 8vo, 2 50 Jones's Machine Design: Part I. Kinematics of Machinery 8vo, 1 50 Part II. Form, Strength, and Proportions of Parts 8vo, 3 00 Kimball and Barr's Machine Design 8vo, 3 00 MacCord's Elements of Descritpive Geometry 8vo, 3 00 Kinematics ; or, Practical Mechanism 8vo, 5 00 Mechanical Drawing 4to, 4 00 Velocity Diagrams 8vo, 1 59 McLeod's Descriptive Geometry Large 12mo, 1 50 * Mahan's Descriptive Geometry and Stone-cutting 8vo, 1 50 Industrial Drawing. (Thompson.) 8vo, 3 50 Moyer's Descriptive Geometry 8vo, 2 00 Reed's Topographical Drawing and Sketching 4to, 5 00 Reid's Course in Mechanical Drawing .8vo, 2 00 Text-book of Mechanical Drawing and Elementary Machine Design.. 8vo, 3 00 Robinson's Principles of Mechanism 8vo, 3 00 Schwamb and Merrill's Elements of Mechanism 8vo, 3 00 Smith (A. W.) and Marx's Machine Design , 8vo, 3 00 Smith's (R. S.) Manual of Topographical Drawing. (McMillan.) 8vo, 2 50 * Titsworth's Elements of Mechanical Drawing Oblong 8vo, 1 25 Warren's Drafting Instruments and Operations 12mo, 1 25 Elements of Descriptive Geometry, Shadows, and Perspective 8vo, 3 50 Elements of Machine Construction and Drawing 8vo, 7 50 Elements of Plane and Solid Free-hand Geometrical Drawing 12mo, 1 00 General Problems of Shades and Shadows 8vo, 3 00 Manual of Elementary Problems in the Linear Perspective of Forms and Shadow 12mo, 1 00 Manual of Elementary Projection Drawing 12mo, 1 50 Plane Problems in Elementary Geometry 12mo, 1 25 Problems, Theorems, and Examples in Descriptive Geometry 8vo, 2 50 Weisbach's Kinematics and Power of Transmission. (Hermann and Klein.) 8vo, 5 00 Wilson's (H. M.) Topographic Surveying 8vo, 3 50 * Wilson's (V. T.) Descriptive Geometry 8vo, 1 50 Free-hand Lettering 8vo, 1 00 Free-hand Perspective 8vo, 2 50 Woolf's Elementary Course in Descriptive Geometry Large 8vo, 3 00 10 ELECTRICITY AND PHYSICS. * Abegg's Theory of Electrolytic Dissociation, (von Ende.) 12mo, $1 25 Andrews's Hand-book for Street Railway Engineering 3X5 inches, mor. 1 25 Anthony and Brackett's Text-book of Physics. (Magie.) ... .Large 12mo, 3 00 Anthony and Ball's Lecture-notes on the Theory of Electrical Measure- ments 12mo, 1 00 Benjamin's History of Electricity 8vo, 3 00 Voltaic Cell 8vo, 3 00 Betts's Lead Refining and Electrolysis 8vo, 4 00 Classen's Quantitative Chemical Analysis by Electrolysis. (Boltwood.).Svo, 3 00 * Collins's Manual of Wireless Telegraphy and Telephony 12mo, 1 50 Crehore and Squier's Polarizing Photo-chronograph 8vo, 3 00 * Danneel's Electrochemistry. (Merriam.) 12mo, 1 25 Dawson's "Engineering" and Electric Traction Pocket-book. . . . 16mo, mor. 5 00 Dolezalek's Theory of the Lead Accumulator (Storage Battery), (von Ende.) 12mo, 2 50 Duhem's Thermodynamics and Chemistry. (Burgess.) 8vo, 4 00 Flather's Dynamometers, and the Measurement of Power 12mo, 3 00 Getman's Introduction to Physical Science 12mo, 1 50 Gilbert's De Magnete. (Mottelay ) 8vo, 2 50 * Hanchett's Alternating Currents 12mo, 1 00 Bering's Ready Reference Tables (Conversion Factors) 16mo, mor. 2 50 * Hobart and Ellis's High-speed Dynamo Electric Machinery 8vo, 6 00 Holman's Precision of Measurements 8vo, 2 00 Telescopic Mirror-scale Method, Adjustments, and Tests.. . .Large 8vo, 75 * Karapetoff 's Experimental Electrical Engineering 8vo, 6 00 Kinzbrunner's Testing of Continuous-current Machines 8vo, 2 00 Landauer's Spectrum Analysis. (Tingle.) 8vo, 3 00 Le Chatelier's High-temperature Measurements. (Boudouard Burgess. )12mo, 3 00 Lob's Electrochemistry of Organic Compounds. (Lorenz.) 8vo, 3 00 * Lyndon's Development and Electrical Distribution of Water Power. .8vo, 3 00 * Lyons's Treatise on Electromagnetic Phenomena. Vols, I .and II. 8vo, each, 6 00 * Michie's Elements of Wave Motion Relating to Sound and Light 8vo, 4 OO Morgan's Outline of the Theory of Solution and its Results 12mo, 1 00 * Physical Chemistry for Electrical Engineers I2mo, 1 50 * Norris's Introduction to the Study of Electrical Engineering 8vo, 2 50 Norris and Dennison's Course of Problems on the Electrical Characteristics of Circuits and Machines. (In Press.) * Parshall and Hobart's Electric Machine Design 4to, half mor, 12 50 Reagan's Locomotives: Simple, Compound, and Electric. New Edition. Large 12mo, 3 50 * Rosenberg's Electrical Engineering. (Haldane Gee Kinzbrunner.) . .8vo, 2 00 Ryan, Norris, and Hoxie's Electrical Machinery. Vol. 1 8vo, 2 50 Schapper's Laboratory Guide for Students in Physical Chemistry 12mo, 1 00 * Tillman's Elementary Lessons in Heat 8vo, 1 50 Tory and Pitcher's Manual of Laboratory Physics Large 12mo, 2 00 Ulke's Modern Electrolytic Copper Refining 8vo, 3 00 LAW. * Brennan's Hand-book of Useful Legal Information for Business Men. 16mo, mor. 5 00 * Davis's Elements of Law 8vo, 2 50 * Treatise on the Military Law of United States 8vo, 7 00 * Dudley's Military Law and the Procedure of Courts-martial. . Large 12mo, 2 50 Manual for Courts-martial 16mo, mor. 1 50 Wait's Engineering and Architectural Jurisprudence 8vo, 6 00 Sheep, 6 50 Law of Contracts 8vo, 3 00 Law of Operations Preliminary to Construction in Engineering and Architecture 8vo, 5 00 Sheep, 5 50 11 MATHEMATICS. Baker's Elliptic Functions 8vo, $1 50 Briggs's Elements of Plane Analytic Geometry. (Bocher.) 12mo, 1 00 * Buchanan's Plane and Spherical Trigonometry 8vo, 1 00 Byerley's Harmonic Functions 8vo, 1 00 Chandler's Elements of the Infinitesimal Calculus 12mo, 2 00 * Coffin's Vector Analysis 12mo, 2 50 Compton's Manual of Logarithmic Computations 12mo, 1 50 * Dickson's College Algebra Large 12mo, 1 50 * Introduction to the Theory of Algebraic Equations Large 12mo, 1 25 Emch's Introduction to Projective Geometry and its Application 8vo, 2 50 Fiske's Functions of a Complex Variable 8vo, 1 00 Halsted's Elementary Synthetic Geometry 8vo, 1 50 Elements of Geometry 8vo, 1 75 * Rational Geometry 12mo, 1 50 Synthetic Projective Geometry 8vo, 1 00 Hyde's Grassmann's Space Analysis 8vo, 1 00 * Johnson's (J. B.) Three-place Logarithmic Tables: Vest-pocket size, paper, 15 * 100 copies, 5 00 * Mounted on heavy cardboard, 8 X 10 inches, 25 * 10 copies, 2 00 .Johnson's (W. W.) Abridged Editions of Differential and Integral Calculus. Large 12mo, 1 vol. 2 50 Curve Tracing in Cartesian Co-ordinates 12mo, 1 00 Differential Equations 8vo, 1 00 Elementary Treatise on Differential Calculus Large 12mo, 1 50 Elementary Treatise on the Integral Calculus. Large 12mo, 1 50 * Theoretical Mechanics 12mo, 3 00 Theory of Errors and the Method of Least Squares 12mo, 1 50 Treatise on Differential Calculus Large 12mo, 3 00 Treatise on the Integral Calculus Large 12mo, 3 00 Treatise on Ordinary and Partial Differential Equations. . .Large 12mo, 3 50 Karapetoff's Engineering Applications of Higher Mathematics. (In Preparation.) Laplace's Philosophical Essay on Probabilities. (Truscott and Emory.) . 12mo, 2 00 * Ludlow and Bass's Elements of Trigonometry and Logarithmic and Other Tables 8vo, 3 00 * Trigonometry and Tables published separately Each, 2 00 * Ludlow's Logarithmic and Trigonometric Tables 8vo, 1 00 Macfarlane's Vector Analysis and Quaternions 8vo, 1 00 McMahon's Hyperbolic Functions 8vo, 1 00 Manning's Irrational Numbers and their Representation by Sequences and Series 12mo, 1 25 Mathematical Monographs. Edited by Mansfield Merriman and Robert S. Woodward Octavo, each 1 00 No. 1. History of Modern Mathematics, by David Eugene Smith. No. 2. Synthetic Projective Geometry, by George Bruce Halsted. No. 3. Determinants, by Laenas Gifford Weld. No. 4. Hyper- bolic Functions, by James McMahon. No. 5. Harmonic Func- tions, by William E. Byerly. No. 6. Grassmann's Space Analysis, by Edward W. Hyde. No. 7. Probability and Theory of Errors, by Robert S. Woodward. No. 8. Vector Analysis and Quaternions, by Alexander Macfarlane. No. 9. Differential Equations, by William Woolsey Johnson. No. 10. The Solution of Equations, by Mansfield Merriman. No. 11. Functions of a Complex Variable, by Thomas S. Fiske. Maurer's Technical Mechanics 8vo, 4 00 Merriman's Method of Least Squares 8vo, 2 00 Solution of Equations 8vo, 1 00 Rice and Johnson's Differential and Integral Calculus. 2 vols. in one. Large 12mo, 1 50 Elementary Treatise on the Differential Calculus Large 12mo, 3 00 Smith's History of Modern Mathematics 8vo, 1 00 * Veblen and Lennes's Introduction to the Real Infinitesimal Analysis of One Variable 8vo, 2 00 12 - * Waterbury's Vest Pocket Hand-book of Mathematics for Engineers. 2X5f inches, mor. $1 00 * Enlarged Edition, Including Tables mor. 1 50 Weld's Determinants gvo, 1 00 Wood's Elements of Co-ordinate Geometry 8vo', 2 00 Woodward's Probability and Theory of Errors 8vo, 1 00 MECHANICAL ENGINEERING. MATERIALS OF ENGINEERING, STEAM-ENGINES AND BOILERS. Bacon's Forge Practice 12mo, 1 50 Baldwin's Steam Heating for Buildings 12mo, 2 50 Barr's Kinematics of Machinery 8vo, 2 50 * Bartlett's Mechanical Drawing gvo, 3 00 Abridged Ed 8vo, 1 50 * Burr's Ancient and Modern Engineering and the Isthmian Canal 8vo, 3 50 Carpenter's Experimental Engineering 8vo, 6 00 Heating and Ventilating Buildings 8vo, 4 00 Clerk's Gas and Oil Engine. (New edition in press.) Compton's First Lessons in Metal Working 12mo, 1 50 Compton and De Groodt's Speed Lathe 12mo, 1 50 Coolidge's Manual of Drawing 8vo, paper, 1 00 Coolidge and Freeman's Elements of Geenral Drafting for Mechanical En- gineers Oblong 4to, 2 50 Cromwell's Treatise on Belts and Pulleys 12mo, 1 50 Treatise on Toothed Gearing 12mo, 1 50 Dingey's Machinery Pattern Making 12mo, 2 00 Durley's Kinematics of Machines 8vo, 4 00 Flanders's Gear-cutting Machinery Large 12mo, 3 00 Flather's Dynamometers and the Measurement of Power 12mo, 3 00 Rope Driving 12mo, 2 00 Gill's Gas and Fuel Analysis for Engineers 12mo, 1 25 Goss's Locomotive Sparks 8vo, 2 00 Greene's Pumping Machinery. (In Preparation.) Hering's Ready Reference Tables (Conversion Factors) 16mo, mor. 2 50 * Hobart and Ellis's High Speed Dynamo Electric Machinery 8vo, 6 00 Hutton's Gas Engine 8vo, 5 00 Jamison's Advanced Mechanical Drawing 8vo, 2 00 Elements of Mechanical Drawing 8vo, 2 50 Jones's Gas Engine 8vo. 4 00 Machine Design: Part I. Kinematics of Machinery 8vo, 1 50 Part II. Form, Strength, and Proportions of Parts 8vo, 3 00 Kent's Mechanical Engineer's Pocket-Book 16mo, mor. 5 00 Kerr's Power and Power Transmission 8vo, 2 00 Kimball and Barr's Machine Design 8vo, 3 00 * Levin's Gas Engine 8vo, 4 00 Leonard's Machine Shop Tools and Methods 8vo, 4 00 * Lorenz's Modern Refrigerating Machinery. (Pope, Haven, and Dean). .8vo, 4 00 MacCord's Kinematics; or, Practical Mechanism 8vo, 5 00 Mechanical Drawing 4to, 4 00 Velocity Diagrams 8vo, 1 50 MacFarland's Standard Reduction Factors for Gases 8vo, 1 50 Mahan's Industrial Drawing. (Thompson.) 8vo, 3 50 Mehrtens's Gas Engine Theory and Design Large 12mo, 2 50 Oberg's Handbook of Small Tools Large 12mo. 3 00 * Parshall and Hobart's Electric Machine Design. Small 4to, half leather, 12 50 Peele's Compressed Air Plant for Mines 8vo, 3 00 Poole's Calorific Power of Fuels 8vo, 3 00 * Porter's Engineering Reminiscences, 1855 to 1882 8vo, 3 00 Reid's Course in Mechanical Drawing 8vo. 2 00 Text-book of Mechanical Drawing and Elementary Machine Design. 8vo, 3 00 13 Richards's Compressed Air 12mo, $1 50 Robinson's Principles of Mechanism 8vo, 3 00 Schwamb and Merrill's Elements of Mechanism 8vo, 3 00 Smith (A. W.) and Marx's Machine Design 8vo, 3 00 Smith's (O.) Press-working of Metals 8vo, 3 00 Sorel's Carbureting and Combustion in Alcohol Engines. (Woodward and Preston.) Large 12mo, 3 00 Stone's Practical Testing of Gas and Gas Meters 8vo, 3 50 Thurston's Animal as a Machine and Prime Motor, and the Laws of Energetics. 12mo, 1 00 Treatise on Friction and Lost Work in Machinery and Mill Work. . .8vo, 3 00 * Tillson's Complete Automobile Instructor 16mo, 1 50 * Titsworth's Elements of Mechanical Drawing Oblong 8vo, 1 25 Warren's Elements of Machine Construction and Drawing 8vo, 7 50 * Waterbury's Vest Pocket Hand-book of Mathematics for Engineers. 2IX5| inches, mor. 1 00 * Enlarged Edition, Including Tables mor. 1 50 Weisbach's Kinematics and the Power of Transmission. (Herrmann Klein.) 8vo, 5 00 Machinery of Transmission and Governors. (Hermann Klein.). . 8vo, 5 00 Wood's Turbines 8vo, 2 50' MATERIALS OF ENGINEERING. * Bovey's Strength of Materials and Theory of Structures 8vo, 7 50* Burr's Elasticity and Resistance of the Materials of Engineering 8vo, 7 50 Church's Mechanics of Engineering 8vo, 6 00 * Greene's Structural Mechanics 8vo, 2 50' * Holley's Lead and Zinc Pigments Large 12mo 3 00' Holley and Ladd's Analysis of Mixed Paints, Color Pigments, and Varnishes. Large 12mo, 2 50 Johnson's (C. M.) Rapid Methods for the Chemical Analysis of Special Steels, Steel-Making Alloys and Graphite Large 12mo, 3 Ofr Johnson's (J. B.) Materials of Construction 8vo, 6 00 Keep's Cast Iron 8vo, 2 50 Lanza's Applied Mechanics 8vo, 7 50 Maire's Modern Pigments and their Vehicles 12mo, 2 00 Maurer's Techincal Mechanics 8vo, 4 00 Merriman's Mechanics of Materials 8vo, 5 00' * Strength of Materials 12mo, 1 00 Metcalf 's Steel. A Manual for Steel-users 12mo, 2 00 Sabin's Industrial and Artistic Technology of Paint and Varnish 8vo, 3 00 Smith's ((A. W.) Materials of Machines 12mo, 1 00 * Smith's (H. E.) Strength of Material 12mo, 1 25 Thurston's Materials of Engineering 3 vols., 8vo, 8 00 Part I. Non-metallic Materials of Engineering 8vo, 2 00 Part II. Iron and Steel 8vo, 3 50 Part III. A Treatise on Brasses, Bronzes, and Other Alloys and their Constituents 8vo, 2 50 Wood's (De V.) Elements of Analytical Mechanics 8vo, 3 00- Treatise on the Resistance of Materials and an Appendix on the Preservation of Timber 8vo, 2 00 Wood's (M. P.) Rustless Coatings: Corrosion and Electrolysis of Iron and Steel '. 8vo, 4 00 STEAM-ENGINES AND BOILERS. Berry's Temperature-entropy Diagram 12mo, 2 00 Carnot's Reflections on the Motive Power of Heat. (Thurston.) 12mo, 1 50 Chase's Art of Pattern Making 12mo, 2 50- 14 Creighton's Steam-engine and other Heat Motors 8vo, 85 00 Dawson's "Engineering" and Electric Traction Pocket-book. .. . 16mo, mor. 5 00 Ford's Boiler Making for Boiler Makers 18mo, 1 00 * Gebhardt's Steam Power Plant Engineering 8vo, 6 00 Goss's Locomotive Performance 8vo, 5 00 Hemenway's Indicator Practice and Steam-engine Economy 12mo, J no Button's Heat and Heat-engines 8vo, 5 00 Mechanical Engineering of Power Plants 8vo, 5 00 Kent's Steam boiler Economy 8vo, 4 00 Kneass's Practice and Theory of the Injector 8vo, 1 50 MacCord's Slide-valves 8vo, 2 00 Meyer's Modern Locomotive Construction 4to, 10 00 Moyer's Steam Turbine 8vo, 4 00 Peabody's Manual of the Steam-engine Indicator 12mo, 1 50 Tables of the Properties of Steam and Other Vapors and Temperature- Entropy Table 8vo, 1 00 Thermodynamics of the Steam-engine and Other Heat-engines. . . . 8vo, 5 00 Valve-gears for Steam-engines 8vo, 2 50 Peabody and Miller's Steam-boilers 8vo, 4 00 Pupin's Thermodynamics of Reversible Cycles in Gases and Saturated Vapors. (Osterberg.) 12mo, 1 25 Reagan's Locomotives: Simple, Compound, and Electric. New Edition. Large 12mo. 3 50 Sinclair's Locomotive Engine Running and Management 12mo, 2 00 Smart's Handbook of Engineering Laboratory Practice 12mo, 2 50 Snow's Steam-boiler Practice 8vo, 3 00 Spangler's Notes on Thermodynamics 12mo, 1 00 Valve-gears 8vo, 2 50 Spangler, Greene, and Marshall's Elements of Steam-engineering 8vo, 3 00 Thomas's Steam-turbines 8vo, 4 00 Thurston's Handbook of Engine and Boiler Trials, and the Use of the Indi- cator and the Prony Brake 8vo, 5 00 Handy Tables 8vo, 1 50 Manual of Steam-boilers, their Designs, Construction, and Operation 8vo, 5 00 Manual of the Steam-engine 2vols., 8vo, 10 00 Part I. History, Structure, and Theory 8vo, 6 00 Part II. Design, Construction, and Operation 8vo, 6 00 Wehrenfenning's Analysis and Softening of Boiler Feed-water. (Patterson. ) 8vo, 4 00 Weisbach's Heat, Steam, and Steam-engines. (Du Bois.) 8vo, 5 00 Whitham's Steam-engine Design 8vo, 5 00 Wood's Thermodynamics, Heat Motors, and Refrigerating Machines. . .8vo, 4 00 MECHANICS PURE AND APPLIED. Church's Mechanics of Engineering 8vo, 6 00 Notes and Examples in Mechanics 8vo, 2 00 Dana's Text-book of Elementary Mechanics for Colleges and Schools .12mo, 1 50 Du Bois's Elementary Principles of Mechanics: Vol. I. Kinematics 8vo, 3 50 Vol. II. Statics 8vo. 4 00 Mechanics of Engineering. Vol. I Small 4to, 7 50 Vol. II Small 4to, 10 00 * Greene's Structural Mechanics 8vo, 2 50 James's Kinematics of a Point and the Rational Mechanics of a Particle. Large 12mo, 2 00 * Johnson's (W. W.) Theoretical Mechanics 12mo, 3 00 Lanza's Applied Mechanics 8vo, 7 50 * Martin's Text Book on Mechanics, Vol. I, Statics 12mo, 1 26 * Vol. II, Kinematics and Kinetics. 12tno, 1 50 Maurer's Technical Mechanics 8vo, 4 00 * Merriman's Elements of Mechanics 12mo, 1 00 Mechanics of Materials 8vo, 5 00 15 * Michie's Elements of Analytical Mechanics 8vo, $4 OO Robinson's Principles of Mechanism 8vo, 3 00 Sanborn's Mechanics Problems Large 12mo, 1 50 Schwamb and Merrill's Elements of Mechanism 8vo, 3 00 Wood's Elements of Analytical Mechanics 8vo, 3 00 Principles of Elementary Mechanics 12mo, 1 25 MEDICAL. * Abderhalden's Physiological Chemistry in Thirty Lectures. (Hall and Defren.) 8vo, 5 00 von Behring's Suppression of Tuberculosis. (Bolduan.) 12mo, 1 00 Bolduan's Immune Sera 12mo, 1 50 Bordet's Studies in Immunity. (Gay.) 8vo, 6 00 Davenport's Statistical Methods with Special Reference to Biological Varia- tions 16mo, mor. 1 50 Ehrlich's Collected Studies on Immunity. (Bolduan.) 8vo, 6 00 * Fischer's Physiology of Alimentation Large 12mo, 2 00 de Fursac's Manual of Psychiatry. (Rosanoff and Collins.).. . .Large 12mo, 2 50 Hammarsten's Text-book on Physiological Chemistry. (Mandel.) 8vo, 4 00 Jackson's Directions for Laboratory Work in Physiological Chemistry . .8vo, 1 25 Lassar-Cohn's Practical Urinary Analysis. (Lorenz.) 12mo, 1 00 Mandel's Hand-book for the Bio-Chemical Laboratory 12mo, 1 50 * Pauli's Physical Chemistry in the Service of Medicine. (Fischer.) ..12mo, 1 25 * Pozzi-Escot's Toxins and Venoms and their Antibodies. (Cohn.). . 12mo, 1 00 Rostoski's Serum Diagnosis. (Bolduan.) 12mo, 1 00 Ruddiman's Incompatibilities in Prescriptions 8vo, 2 00' Whys in Pharmacy 12mo, 1 00 Salkowski's Physiological and Pathological Chemistry. (Orndorff.) .. . .8vo, 2 50 * Satterlee's Outlines of Human Embryology 12mo, 1 25- Smith's Lecture Notes on Chemistry for Dental Students 8vo, 2 50 * Whipple's Tyhpoid Fever Large 12mo, 3 00 * Woodhull's Military Hygiene for Officers of the Line Large 12mo, 1 50 * Personal Hygiene 12mo, 1 00' Worcester and Atkinson's Small Hospitals Establishment and Maintenance, and Suggestions for Hospital Architecture, with Plans for a Small Hospital 12mo, 1 25- METALLURGY. Betts's Lead Refining by Electrolysis 8vo, 4 OO Bolland's Encyclopedia of Founding and Dictionary of Foundry Terms used in the Practice of Moulding 12mo, 3 OO Iron Founder 12mo, 2 5O Supplement 12mo, 2 50 Douglas's Untechnical Addresses on Technical Subjects 12mo, 1 OO Goesel's Minerals and Metals: A Reference Book 16mo. mor. 3 OO * Iles's Lead-smelting 12mo, 2 5O Johnson's Rapid Methods for the Chemical Analysis of Special Steels, Steel-making Alloys and Graphite Large 12mo, 3 00> Keep's Cast Iron 8vo, 2 50 Le Chatelier's High-temperature Measurements. (Boudouard Burgess.) 12mo, 3 OO Metcalf 's Steel. A Manual for Steel-users 12mo, 2 00 Minet's Production of Aluminum and its Industrial Use. (Waldo.). - 12mo, 2 5O * Ruer's Elements of Metallography. (Mathewson.) 8vo, 3 00 Smith's Materials of Machines 12rno, 1 OO Tate and Stone's Foundry Practice 12mo, 2 OO Thurston's Materials of Engineering. In Three Parts 8vo, 8 00 Part I. Non-metallic Materials of Engineering, see Civil Engineering, page 9. Part II. Iron and Steel 8vo, 3 50 Part III. A Treatise on Brasses, Bronzes, and Other Alloys and their Constituents 8vo. 2 50 16 Ulke's Modern Electrolytic Copper Refining 8vo, $3 00 West's American Foundry Practice 1 2mo, 2 50 Moulders' Text Book 12mo. 2 50 MINERALOGY. Baskerville's Chemical Elements. (In Preparation.) * Browning's Introduction to the Rarer Elements 8vo, 1 50 Brush's Manual of Determinative Mineralogy. (Penfield.) 8vo, 4 00 Butler's Pocket Hand-book of Minerals 16mo, mor. 3 00 Chester's Catalogue of Minerals 8vo, paper, 1 00 Cloth, 1 25 * Crane's Gold and Silver 8vo, 5 00 Dana's First Appendix to Dana's New "System of Mineralogy". .Large 8vo, 1 00 Dana's Second Appendix to Dana's New " System of Mineralogy." Large 8vo, Manual of Mineralogy and Petrography 12mo, 2 00 Minerals and How to Study Them 12mo, 1 50 System of Mineralogy Large 8vo, half leather, 12 50 Text-book of Mineralogy 8vo, 4 00 Douglas's Untechnical Addresses on Technical Subjects 12mo, 1 00 Eakle's Mineral Tables 8vo, 1 25 Eckel's Stone and Clay Products Used in Engineering. (In Preparation.) Goesel's Minerals and Metals: A Reference Book I6mo, mor. 3 00 Groth's Introduction to Chemical Crystallography (Marshall) 12mo, 1 25 * Hayes's Handbook for Field Geologists J6mo, mor. 1 50 Iddings's Igneous Rocks , 8vo, 5 00 Rock Minerals 8vo, 5 00 Johannsen's Determination of Rock-forming Minerals in Thin Sections. 8vo, With Thumb Index 5 00 * Martin's Laboratory Guide to Qualitative Analysis with the Blow- 1 50 pipe 12mo, 60 Merrill's Non-metallic Minerals: Their Occurrence and Uses 8vo, 4 00 Stones for Building and Decoration 8vo, 5 00 * Penneld's Notes on Determinative Mineralogy and Record of Mineral Tests. 8vo, paper, 50 Tables of Minerals, Including the Use of Minerals and Statistics of Domestic Production 8vo, 1 00 * Pirsson's Rocks and Rock Minerals 12mo, 2 50 * Richards's Synopsis of Mineral Characters 12mo, mor. 1 25 * Ries's Clays: Their Occurrence, Properties and Uses 8vo, 5 00 * Ries and Leighton's History of the Clay-working Industry of the United States 8vo, 2 50 * Tillman's Text-book of Important Minerals and Rocks 8vo, 2 00 Washington's Manual of the Chemical Analysis of Rocks 8vo, 2 00 MINING. * Beard's Mine Gases and Explosions Large 12mo, 3 00- * Crane's Gold and Silver 8vo, 5 00 * Index of Mining Engineering Literature 8vo, 4 00 * 8vo, mor. 5 00 Mining Methods. (In Press.) Douglas's Untechnical Addresses on Technical Subjects 12mo, 1 00 Eissler's Modern High Explosives 8vo, 4 00 Goesel's Minerals and Metals: A Reference Book 16mo, mor. 3 00 Ihlseng's Manual of Mining 8vo, 5 00 * Iles's Lead Smelting 12mo, 2 50 Peele's Compressed Air Plant for Mines 8vo, 3 00 Riemer's Shaft Sinking Under Difficult Conditions. (Corning and Peele. )8vo. 3 00 * Weaver's Military Explosive* 8vo. 3 00 Wilson's Hydraulic and Placer Mining. 2d edition, rewritten 12mo, 2 50 Treatise on Practical and Theoretical Mine Ventilation 12mo, 1 25 17 SANITARY SCIENCE. Association of State and National Food and Dairy Departments, Hartford Meeting, 1906 8vo, $3 00 Jamestown Meeting, 1907 8vo, 3 00 * Bashore's Outlines of Practical Sanitation 12mo, 1 25 Sanitation of a Country House 12mo, 1 00 Sanitation of Recreation Camps and Parks 12mo, 1 00 Folwell's Sewerage. (Designing, Construction, and Maintenance.) 8vo, 3 00 Water-supply Engineering 8vo, 4 00 Fowler's Sewage Works Analyses 12mo, 2 00 Fuertes's Water-filtration Works 12mo, 2 50 Water and Public Health 12mo, 1 50 Gerhard's Guide to Sanitary Inspections 12mo, 1 50 * Modern Baths and Bath Houses 8vo, 3 00 Sanitation of Public Buildings 12mo, 1 50 * The Water Supply, Sewerage, and Plumbing of Modern City Buildings. 8vo, 4 00 Hazen's Clean Water and How to Get It Large 12mo, 1 50 Filtration of Public Water-supplies 8vo, 3 00 Kinnicut, Winslow and Pratt's Purification of Sewage. (In Preparation.) Leach's Inspection and Analysis of Food with Special Reference to State Control 8vo, 7 50 Mason's Examination of Water. (Chemical and Bacteriological) 12mo, 1 25 Water-supply. (Considered principally from a Sanitary Standpoint). 8vo, 4 00 * Merriman's Elements of Sanitary Engineering 8vo, 2 00 Ogden's Sewer Construction 8vo, 3 00 Sewer Design 12mo, 2 00 Parsons's Disposal of Municipal Refuse 8vo, 2 00 Prescott and Winslow's Elements of Water Bacteriology, with Special Refer- ence to Sanitary Water Analysis 12mo, 1 50 * Price's Handbook on Sanitation 12mo, 1 50 Richards's Cost of Cleanness 12mo, Cost of Food. A Study in Dietaries 12mo, Cost of Living as Modified by Sanitary Science 12mo, Cost of Shelter 12mo, 00 * Richards and Williams's Dietary Computer 8vo, 50 Richards and Woodman's Air, Water, and Food from a Sanitary Stand- point 8vo, 2 00 * Richey's Plumbers', Steam-fitters', and Tinners' Edition (Building Mechanics Ready Reference Series) 16mo, mor. 1 50 Rideal's Disinfection and the Preservation of Food 8vo, 4 00 Sewage and Bacterial Purification of Sewage 8vo, 4 00 Soper's Air and Ventilation of Subways 12mo, 2 50 Turneaure and Russell's Public Water-supplies 8vo, 5 00 Venable's Garbage Crematories in America 8vo, 2 00 Method and Devices for Bacterial Treatment of Sewage 8vo, 3 00 Ward and Whipple's Freshwater Biology. (In Press.) Whipple's Microscopy of Drinking-water. . . 8vo, 3 50 * Typhoid Fever Large 12mo, 3 00 Value of Pure Water Large 12mo, 1 00 Winslow's Systematic Relationship of the Coccacea? Large 12mo, 2 50 MISCELLANEOUS. Emmons's Geological Guide-book of the Rocky Mountain Excursion of the International Congress of Geologists Large 8vo. 1 50 Ferrel's Popular Treatise on the Winds ; 8vo, 4 00 Fitzgerald's Boston Machinist 18mo, 1 00 Gannett's Statistical Abstract of the World 24mo, 75 Haines's American Railway Management 12mo, 2 50 Hanausek's The Microscopy of Technical Products. (Winton) 8vo, 5 00 18 Jacobs's Betterment Briefs. A Collection of Published Papers on Or- ganized Industrial Efficiency 8vo, $3 50 Metcalfe's Cost of Manufactures, and the Administration of Workshops. .8vo, 5 OO Putnam's Nautical Charts 8vo, 2 00 Ricketts's History of Rensselaer Polytechnic Institute 1824-1894. Large 12mo, 3 00 Rotherham's Emphasised New Testament Large 8vo, 2 00 Rust's Ex-Meridian Altitude, Azimuth and Star-finding Tables 8vo, 5 00 Standage's Decoration of Wood, Glass, Metal, etc 12mo, 2 OO Thome's Structural and Physiological Botany. (Bennett) 16mo, 2 25 Westermaier's Compendium of General Botany. (Schneider) 8vo, 2 00 Winslow's Elements of Applied Microscopy 12mo, 1 50 HEBREW AND CHALDEE TEXT-BOOOKS. Gesenius's Hebrew and Chaldee Lexicon to the Old Testament Scriptures. (Tregelles.) Small 4to, half mor, 5 OO Green's Elementary Hebrew Grammar 12mo, 1 25 OF THE UNIVERSITY OF 19 g *L. THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW AN INITIAL FINE OF 25 CENTS WILL BE ASSESSED FOR FAILURE TO RETURN THIS BOOK ON THE DATE DUE. THE PENALTY WILL INCREASE TO 50 CENTS ON THE FOURTH DAY AND TO $1.00 ON THE SEVENTH DAY OVERDUE. APR 2 1 1965 JAN 7 REC'O Book Slip-20m-9,'60(B3010s4)458 193957 F7 5 7 3 7