LIBRARY 
 
 OF THE 
 
 UNIVERSITY OF CALIFORNIA. 
 
 94 J 98 
 
gale ^Bicentennial publication? 
 
 RESEARCH PAPERS 
 
 FROM THE 
 
 KENT CHEMICAL LABORATORY OF YALE 
 UNIVERSITY 
 
gale ^Bicentennial publications 
 
 With the approval of the President and Fellows 
 of TTale University, a series of volumes has been 
 prepared by a number of the Professors and In- 
 structors, to be issued in connection with the 
 Bicentennial Anniversary, as a partial indica- 
 tion of the character of the studies in which the 
 University teachers are engaged. 
 
 This series of volumes is respectfully dedicated to 
 
 of tt) 
 
RESEARCH PAPERS 
 
 FROM THE 
 
 KENT CHEMICAL LABORATORY 
 
 OF 
 
 YALE UNIVERSITY 
 
 EDITED BY 
 
 FRANK AUSTIN GOOCH 
 
 Professor of Chemistry in Yale University 
 
 VOLUME II. 
 
 NEW YORK: CHARLES SCRIBNER'S SONS 
 LONDON: EDWARD ARNOLD 
 
 1901 
 
Copyright, 1901, 
 BY YALE UNIVERSITY 
 
 Published, June^ /go/ 
 
 UNIVERSITY PRESS JOHN WILSON 
 AND SON CAMBRIDGE, U. S. A. 
 
CONTENTS 
 
 VOLUME II. 
 
 PAGE 
 I. The Determination of Tellurium by Precipitation as the 
 
 Iodide. By F. A. GOOCH and W. C. MORGAN .... 1 
 
 II. On the Application of Certain Organic Acids to the Estimation 
 of Vanadium. By PHILIP E. BROWNING and RICHARD 
 J. GOODMAN 4 
 
 III. The Determination of Oxygen in Air and in Aqueous Solu- 
 
 tion. By D. ALBERT KREIDER 11 
 
 IV. A Method for the Separation of Aluminum from Iron. By 
 
 F. A. GOOCH and F. S. HAVENS 20 
 
 V. The Estimation of Molybdenum lodometrically. By F. A. 
 
 GOOCH 27 
 
 VI. The Application of lodic Acid to the Analysis of Iodides. By 
 
 F. A. GOOCH and C. F. WALKER 33 
 
 VII. The Action of Urea and Primary Amines on Maleic Anhy- 
 dride. By FREDERICK L. DUNLAP and I. K. PHELPS . 42 
 
 VIII. The Separation of Aluminum and Beryllium by the Action of 
 
 Hydrochloric Acid. By FRANKS S. HAVENS .... 47 
 
 IX. The Titration of Sodium Thiosulphate by lodic Acid. By 
 
 CLAUDE F. WALKER 52 
 
 X. The Combustion of Organic Substances in the Wet Way. By 
 
 I. K. PHELPS 62 
 
 XI. The Estimation of Manganese as the Sulphate and as the 
 
 Oxides. By F. A. GOOCH and MARTHA AUSTIN ... 77 
 XII. On the Condition of Oxidation of Manganese precipitated 
 by the Chlorate Process. By F. A. GOOCH and MARTHA 
 AUSTIN 85 
 
 XIH. On the Estimation of Manganese separated as the Carbonate. 
 
 By MARTHA AUSTIN 96 
 
 XIV. The Action of Carbon Dioxide on Soluble Borates. By Louis 
 
 CLEVELAND JONES . 100 
 
 94198 
 
viii CONTENTS 
 
 PAGE 
 XV. Further Separations of Aluminum by Hydrochloric Acid. 
 
 By FRANKE STUART HAVENS 106 
 
 XVI. The lodometric Determination of Molybdenum. By F. A. 
 
 GOOCH and JOHN T. NORTON, Jr Ill 
 
 XVII. On the Determination of Manganese as the Pyrophos- 
 
 phate. By F A. GOOCH and MARTHA AUSTIN . . 121 
 
 XVIII. On the Detection of Sulphides, Sulphates, Sulphites, and 
 Thiosulphates in the presence of each other. By 
 PHILIP E. BROWNING and ERNEST HOWE . . . . 134 
 
 XIX. On the Separation of Nickel and Cobalt by Hydrochloric 
 
 Acid. By FRANKE STUART HAVENS 141 
 
 XX. The Ethers of Toluquinoneoxime and their bearing on the 
 Space Isomerism of Nitrogen. By JOHN L. BRIDGE 
 and WILLIAM CONGER MORGAN 145 
 
 XXI. The Application of Iodine in the Analysis of Alkalies and 
 Acids. By CLAUDE F. WALKER and DAVID H. M. 
 GILLESPIE 162 
 
 XXII. The Estimation of Boric Acid. By F. A. GOOCH and 
 
 Louis CLEVELAND JONES 172 
 
 XXIH. A Volumetric Method for the Estimation of Boric Acid. 
 
 By Louis CLEVELAND JONES 182 
 
 XXIV. The Constitution of the Ammonium Magnesium Phosphate 
 
 of Analysis. By F. A. GOOCH and MARTHA AUSTIN 190 
 
 XXV. The Influence of Hydrochloric Acid in Titrations by So- 
 dium Thiosulphate, with special reference to the Esti- 
 mation of Selenious Acid. By JOHN T. NORTON, Jr. 206 
 XXVI. -The Volatilization of the Iron Chlorides in Analysis, and 
 the Separation of the Oxides of Iron and Aluminum. 
 By F. A. GOOCH and FRANKE STUART HAVENS . 215 
 
 XXVII. The Titration of Oxalic Acid by Potassium Permanganate 
 in presence of Hydrochloric Acid. By F. A. GOOCH 
 and C. A. PETERS 222 
 
 XXVIII. The Estimation of Iron in the Ferric State by Reduction 
 with Sodium Thiosulphate and Titration with Iodine. 
 By JOHN T. NORTON, Jr 230 
 
 XXIX. The Determination of Tellurous Acid in presence of 
 
 Haloid Salts. By F. A. GOOCH and C. A. PETERS . 238 
 
 XXX. An lodometric Method for the Estimation of Boric Acid. 
 
 By Louis CLEVELAND JONES 244 
 
CONTENTS 
 
 XXXI. The Double Ammonium Phosphates of Beryllium, Zinc, 
 and Cadmium in Analysis. By MARTHA AUSTIN 
 
 IX 
 PAGE 
 
 252 
 
 XXXII. Separation of Iron from Chromium, Zirconium, and 
 Beryllium, by the Action of Gaseous Hydrochloric 
 Acid on the Oxides. By FRANKE STUART HAVENS 
 and ARTHUR FITCH WAY 266 
 
 XXXin. The lodometric Determination of Gold. By F. A. 
 
 GOOCH and FREDERICK H. MORLEY 269 
 
 XXXIV. The Action of Acetylene on the Oxides of Copper. By 
 
 F. A. GOOCH and DEFOREST BALDWIN . . . . 276 
 
 XXXV. Notes on the Space Isomerism of the Toluquinoneoxime 
 
 Ethers. By WILLIAM CONGER MORGAN .... 283 
 
 XXXVI. On the Volumetric Estimation of Cerium. By PHILIP 
 
 E. BROWNING 289 
 
 XXXVII. On the Estimation of Thallium as the Chromate. By 
 
 PHILIP E. BROWNING and GEORGE P. HUTCHINS 300 
 
 XXXVIII. The Ethics of Isonitrosoguaiacol in their relation to the 
 Space Isomerism of Nitrogen. By JOHN L. BRIDGE 
 and WILLIAM CONGER MORGAN 804 
 
 XXXIX. The Constitution of the Ammonium Magnesium Arseni- 
 
 ate of Analysis. By MARTHA AUSTIN .... 309 
 
 XL. On the Estimation of Thallium as the Acid and Neutral 
 
 Sulphates. By PHILIP E. BROWNING 317 
 
 XLI. The Separation and Determination of Mercury as Mer- 
 
 curous Oxalate. By C. A. PETERS 320 
 
 XLII. The Titration of Mercury by Sodium Thiosulphate. By 
 
 JOHN T. NORTON, Jr 328 
 
 XLIII. The lodometric Estimation of Arsenic Acid. By F. A. 
 
 GOOCH and JULIA C. MORRIS 336 
 
 XLIV. On the Qualitative Separation of Nickel from Cobalt by 
 the Action of Ammonium Hydroxide on the Ferri- 
 cyanides. By PHILIP E. BROWNING and JOHN B. 
 HARTWELL 344 
 
 XLV. The Volumetric Estimation of Copper as the Oxalate, 
 with Separation from Cadmium, Arsenic, Tin, Iron, 
 and Zinc. By CHARLES A. PETERS 347 
 
 XL VI. The Sulphocyanides of Copper and Silver in Gravi- 
 metric Analysis. By R. G. VAN NAME . . . . 359 
 
x CONTENTS 
 
 PAGE 
 
 XL VII. On the Estimation of Caesium and Rubidium as the Acid 
 Sulphates, and of Potassium and Sodium as the Pyro- 
 sulphates. By PHILIP E. BROWNING 368 
 
 XL VIII. The Estimation of Calcium, Strontium, and Barium as the 
 
 Oxalates. By CHARLES A. PETERS 373 
 
 XLIX. The Action of Sodium Thiosulphate on Solutions of Me- 
 tallic Salts at High Temperatures and Pressures. By 
 JOHN T. NORTON, Jr. . 384 
 
 INDEX 395 
 
RESEARCH PAPERS 
 
 PROM THE 
 
 KENT CHEMICAL LABORATORY OF YALE 
 UNIVERSITY 
 

OF" THE 
 
 UNIVERSITY 
 
 OF 
 
 RESEARCH PAPERS 
 
 THE DETERMINATION OF TELLURIUM BY 
 PRECIPITATION AS THE IODIDE. 
 
 BY F. A. GOOCH AND W. C. MORGAN.* 
 
 IT was known to Berzelius that hydriodic acid and tellurous 
 acid interact with the formation of tellurium tetraiodide, which 
 is converted by water into an oxyiodide and by excess of an 
 alkaline iodide into a soluble double salt. Wheeler f has 
 shown that the double salt which is formed when tellurous 
 iodide is boiled in a strong solution of potassium iodide in 
 dilute hydriodic acid is definite and has the constitution 
 represented by the formula 2KI . TeI 4 . 2H 2 O. We have 
 observed, however, that when potassium iodide is added to a 
 cold solution of tellurous acid containing at least one-fourth 
 of its volume of strong sulphuric acid, no tendency toward 
 the formation of a double salt becomes apparent until the 
 potassium iodide amounts to more than enough to convert all 
 the tellurous acid present into the tetraiodide according to the 
 equation, 
 
 H 2 Te0 8 + 4H 2 S0 4 + 4KI = TeI 4 + 4KHS0 4 + 3H 2 0. 
 
 The tellurium tetraiodide which is thus formed is extremely 
 insoluble in sulphuric acid of the strength mentioned, though 
 soluble in excess of potassium iodide, and acted upon by water 
 with the formation of tellurium oxyiodide and hydriodic acid. 
 It is produced at first in the condition of a finely divided 
 dark brown precipitate which upon agitation of the liquid 
 containing it gathers in curdy masses and settles, leaving the 
 
 * From Am. Jour. Sci., ii, 271. 
 t Am. Jour. Sci., xlv, 267. 
 VOL. n. 1 
 
2 DETERMINATION OF TELLURIUM BY 
 
 supernatant liquid clear. By taking advantage of this tendency 
 to curd it is possible to determine without great difficulty the 
 exact point during the gradual addition of potassium iodide 
 when the precipitation of the tellurium iodide is complete, and 
 we have been able to found upon this property a very simple 
 titrimetric method for the direct determination of small 
 amounts of tellurium. 
 
 In our test experiments we used tellurium dioxide prepared 
 by oxidizing presumably pure tellurium with nitric acid and 
 igniting the residue at a low red heat. Weighed amounts of 
 the oxide thus prepared were dissolved in Erlenmeyer beakers 
 in a very little of a strong solution of potassium hydroxide, 
 and dilute sulphuric acid was added carefully until the tellu- 
 rous acid which was precipitated upon the neutralization of 
 the alkaline hydroxide was just redissolved. To this solution 
 sulphuric acid of half-strength was added in such amount that 
 the solution finally obtained, after adding the aqueous solution 
 of potassium iodide subsequently, should still contain at least 
 one-fourth of its volume of strong sulphuric acid. The 
 Erlenmeyer beaker was placed upon a pane of window glass 
 supported upon strips of wood about 1 cm. above the level of 
 the work table, which was covered with white paper. A 
 solution of approximately decinormal potassium iodide free 
 from iodate and carefully standardized in terms of iodine by a 
 method described in a former paper from this laboratory * was 
 introduced gradually from a burette into the middle of the 
 Erlenmeyer beaker. As the drops of the potassium iodide 
 touched the liquid the precipitation formed at the centre and 
 travelled in rings toward the outer walls of the beaker. When 
 the liquid became so opaque that the effect of the potassium 
 iodide was distinguished with difficulty, the beaker was rotated 
 and the curded precipitate permitted to settle, and then the 
 process of titration was continued as before. We experimented 
 with amounts of tellurium dioxide varying from approximately 
 0.025 grm. to 0.1 grm., the latter quantity being as large as can 
 be handled with accuracy without intermediate removal of the 
 
 * Am. Jour. Sci., rxxix, 188. Volume I, p. 1. 
 
PRECIPITATION AS THE IODIDE. 
 
 3 
 
 precipitate by filtration. With an Erlenmeyer beaker 10 cm. 
 in diameter across the bottom and a final volume of liquid 
 amounting to not more than 100 cm 3 , we were able to follow 
 the precipitation most easily. 
 
 The results of a series of determinations made according to 
 the method described and recorded in the following table are 
 closely accordant, and, in close agreement with the theory of 
 the process if the atomic weight of the tellurium which we 
 used is taken as 127. We feel justified in taking this number 
 as the atomic weight of our tellurium, because the mean result 
 of twelve oxidations by standard potassium permanganate of 
 tellurium dioxide, prepared similarly to that which we used 
 and from the same lot of material, and the mean result of 
 twelve reductions by hydrobromic acid of the telluric acid 
 thus produced,* point to this figure. 
 
 Final 
 
 volume. 
 
 Strongest 
 H 2 S0 4 
 
 Iodine value 
 of KI used. 
 
 TeOj taken. 
 
 TeO, found. 
 
 Error. 
 
 
 present. 
 
 
 
 
 
 cm 8 
 
 cm* 
 
 grm. 
 
 gTm. 
 
 grm. 
 
 grm. 
 
 50 
 
 17 
 
 0.0706 
 
 0.0223 
 
 0.0221 
 
 0.0002- 
 
 50 
 
 17 
 
 0.0764 
 
 0.0244 
 
 0.0239 
 
 0.0005- 
 
 50 
 
 17 
 
 0.1591 
 
 0.0496 
 
 0.0499 
 
 0.0003+ 
 
 60 
 
 17 
 
 0.1655 
 
 0.0517 
 
 0.0519 
 
 0.0002+ 
 
 60 
 
 17 
 
 0.1578 
 
 0.0498 
 
 0.0494 
 
 0.0004- 
 
 80 
 
 30 
 
 0.1591 
 
 0.0498 
 
 0.0499 
 
 0.0001+ 
 
 100 
 
 30 
 
 0.3179 
 
 0.1001 
 
 0.0997 
 
 0.0004- 
 
 100 
 
 30 
 
 0.3186 
 
 0.1008 
 
 0.0999 
 
 0.0009- 
 
 100 
 
 30 
 
 0.3208 
 
 0.1011 
 
 0.1005 
 
 0.0006- 
 
 100 
 
 30 
 
 0.3208 
 
 0.1010 
 
 0.1005 
 
 0.0005- 
 
 From these results it is obvious that the method, which is 
 very rapid, is accurate. 
 
 * Am. Jour. Sci., xlviii, 377, 378. Volume I, pp. 279, 281. 
 
n 
 
 ON THE APPLICATION OF CERTAIN ORGANIC 
 ACIDS TO THE ESTIMATION OF VANADIUM. 
 
 BY PHILIP E. BROWNING AND RICHARD J. GOODMAN * 
 
 IN a former paperf by one of us a method for the determi- 
 nation of vanadium was described in which tartaric acid was 
 used to reduce vanadic acid to the condition of the tetroxide. 
 The method may be briefly outlined as follows : 
 
 Measured and weighed portions of a solution of ammonium 
 vanadate, the standard of which had been determined by the 
 evaporation and ignition of definite portions, were treated with 
 tartaric acid in excess and boiled, when the appearance of the 
 deep blue color indicated the reduction to the condition of the 
 tetroxide. After cooling, the solution was neutralized with 
 potassium bicarbonate and a moderate excess of that reagent 
 added. To the alkaline solution an excess of a standard 
 solution of iodine was added and the whole allowed to stand 
 about one hour, when no further bleaching of the iodine was 
 noticed. The excess of iodine was then destroyed with a 
 standard solution of arsenious oxide, starch was added, and the 
 blue color obtained with a few drops of the iodine solution. 
 The total amount of iodine used, less the amount equivalent to 
 the arsenious oxide solution used, is the amount necessary to 
 oxidize the vanadium from the condition of the tetroxide to 
 that of the pentoxide, from which, according to the following 
 equation, can be calculated the amount of vanadium present : 
 V 2 4 + I-I + H 2 = V 2 6 + 2HI. 
 
 * From Am. Jour. ScL, ii, 355. 
 t Zeitschr. anorg. Chem., vii, 158. 
 
 t These determinations are best made in small Erlenmeyer beakers, closed 
 with paraffin-coated corks while standing with iodine. 
 
ESTIMATION OF VANADIUM. 
 
 The work to be described in this paper is in part an 
 application of the work described in the paper above mentioned 
 to a series of determinations of vanadium in the presence of 
 molybdenum and tungsten. The solution of vanadium used was 
 one of ammonium vanadate, and the standard was determined 
 by evaporating and igniting, in the presence of a drop of 
 nitric acid, measured and weighed portions, the mean of 
 closely agreeing results being taken as the standard. Our 
 first series of determinations was by the method previously 
 described, that being the natural starting-point for the work 
 contemplated. The results follow in the table : 
 
 Bxp. 
 
 V,0 6 taken. 
 
 V 2 O 8 found. 
 
 Error. 
 
 Tartaric 
 acid. 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.1621 
 
 0.1618 
 
 0.0003- 
 
 2 
 
 2 
 
 0.1620 
 
 0.1624 
 
 0.0004+ 
 
 2 
 
 (3) 
 
 0.1614 
 
 0.1622 
 
 0.0008+ 
 
 2 
 
 (4) 
 
 0.1619 
 
 0.1606 
 
 0.0013- 
 
 1 
 
 (5) 
 
 0.1604 
 
 0.1597 
 
 0.0007- 
 
 2 
 
 (6) 
 
 0.1618 
 
 0.0615 
 
 00003- 
 
 3 
 
 (7) 
 
 0.1298 
 
 0.1305 
 
 0.0007+ 
 
 1 
 
 (8) 
 
 0.1294 
 
 0.1297 
 
 0.0003+ 
 
 1 
 
 (9) 
 
 0.1618 
 
 0.1618 
 
 O.OOOOi 
 
 2 
 
 (10) 
 
 0.2588 
 
 0.2575 
 
 0.0013 
 
 3 
 
 (11) 
 
 0.2722 
 
 0.2726 
 
 0.0004+ 
 
 2 
 
 (12) 
 
 0.3273 
 
 0.3269 
 
 0.0004- 
 
 2 
 
 We next tried the effect of treating a solution of sodium 
 tungstate in the same manner. We found that after the 
 boiling with tartaric acid, neutralizing, adding iodine and 
 allowing to stand as before, the amount of free iodine present, 
 as shown by the amount of arsenious oxide solution necessary 
 to destroy it was the same as that originally added, showing 
 that no reduction had taken place. Accordingly a series of 
 determinations of vanadium in the presence of tungsten was 
 made which is recorded in the next table. 
 
 The results show that vanadium may be easily determined 
 by this method in the presence of tungsten without any evi- 
 dent interfering action on the part of the latter element. 
 
 When the same method of treatment was applied in the 
 presence of molybdenum in the form of ammonium molybdate, 
 
APPLICATION OF CERTAIN ORGANIC ACIDS 
 
 Exp. 
 
 V 2 O 6 taken. 
 
 V 2 O 5 found. 
 
 Error. 
 
 Sodium 
 
 tuiigstate. 
 
 Tartaric 
 acid. 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.1618 
 
 0.1615 
 
 0.0003- 
 
 1 
 
 3 
 
 (2) 
 
 0.1615 
 
 0.1606 
 
 0.0009- 
 
 1 
 
 3 
 
 (3) 
 
 0.1618 
 
 0.1624 
 
 0.0006+ 
 
 1 
 
 3 
 
 w 
 
 0.1619 
 
 0.1624 
 
 0.0005+ 
 
 1 
 
 3 
 
 w 
 
 0.1627 
 
 0.1623 
 
 0.0004- 
 
 1 
 
 3 
 
 (6) 
 
 0.1621 
 
 0.1624 
 
 0.0003+ 
 
 1 
 
 4 
 
 (7) 
 
 0.2587 
 
 0.2574 
 
 0.0013- 
 
 1 
 
 4 
 
 (8 
 
 0.2587 
 
 0.2689 
 
 0.0002+ 
 
 1 
 
 4 
 
 the majority of the determinations gave large plus errors, and 
 a few experiments made with the molybdate alone seemed to 
 show a noticeable reduction of the molybdic acid. In the 
 following table the results are tabulated. In experiments (3), 
 (4) and (5) the mixtures were not boiled with tartaric acid, 
 but warmed on a steam bath, with, however, no very apparent 
 prevention of the reducing action. 
 
 E 
 
 cp. 
 
 V 2 O 6 taken. 
 
 V,O 6 found. 
 
 Error. 
 
 Ammonium 
 molybdate. 
 
 Tartaric 
 acid. 
 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 
 1) 
 
 0.1620 
 
 0.1790 
 
 0.0170+ 
 
 
 2 
 
 
 2) 
 
 0.1624 
 
 0.1619 
 
 0.0005- 
 
 
 2 
 
 
 3 
 
 0.1294 
 
 0.1416 
 
 0.0122+ 
 
 
 2 
 
 
 4) 
 
 0.1296 
 
 0.1361 
 
 0.0065+ 
 
 
 2 
 
 
 5) 
 
 0.1291 
 
 0.1312 
 
 0.0021+ 
 
 
 2 
 
 
 6) 
 
 0.1293 
 
 0.1324 
 
 0.0031+ 
 
 
 2 
 
 
 7 
 
 0.1636 
 
 0.1760 
 
 0.0124+ 
 
 
 2 
 
 
 8) 
 
 0.1640 
 
 0.1724 
 
 0.0084+ 
 
 
 2 
 
 
 [9) 
 
 0.1622 
 
 0.1624 
 
 0.0002+ 
 
 
 3 
 
 (] 
 
 0) 
 
 0.1622 
 
 0.1632 
 
 0.0010+ 
 
 
 3 
 
 (1 
 
 1) 
 
 0.1619 
 
 0.1879 
 
 0.0260+ 
 
 
 3 
 
 (1 
 
 2) 
 
 0.1292 
 
 0.1360 
 
 0.0068+ 
 
 1 
 
 3 
 
 (1 
 
 3) 
 
 0.1860 
 
 0.1917 
 
 0.0057+ 
 
 1 
 
 3 
 
 1 
 
 4) 
 
 0.3274 
 
 0.3733 
 
 0.0459+ 
 
 1 
 
 4 
 
 (1 
 
 6) 
 
 0.2324 
 
 0.2383 
 
 0.0059+ 
 
 1 
 
 4 
 
 We next tried the action of tartaric acid upon the vanadium 
 solution in the cold and found that the reduction could be 
 carried on to completion under these conditions if the tartaric 
 acid was in large excess, the time sufficient, and the volume of 
 the solution small. The following series was made to deter- 
 mine these points : 
 
TO THE ESTIMATION OF VANADIUM. 
 
 Exp. 
 
 V 2 O 5 taken. 
 
 V 2 O S found. 
 
 Error. 
 
 Time. 
 
 Tartaric 
 acid. 
 
 Total 
 vol. 
 
 
 grm. 
 
 grm. 
 
 gnu. 
 
 days. 
 
 grm. 
 
 cms 
 
 (1) 
 
 0.1646 
 
 0.1649 
 
 0.0003+ 
 
 1 
 
 4 
 
 25 
 
 (2) 
 
 0.1640 
 
 0.1606 
 
 0.0034- 
 
 1 
 
 4 
 
 65 
 
 
 3) 
 
 0.1293 
 
 0.1264 
 
 0.0029 
 
 2 
 
 3 
 
 55 
 
 
 4) 
 
 0.1633 
 
 0.1628 
 
 0.0005- 
 
 2 
 
 4 
 
 65 
 
 
 5) 
 
 0.1293 
 
 0.1288 
 
 0.0005- 
 
 3 
 
 2.5 
 
 60 
 
 
 6) 
 
 0.1298 
 
 0.1299 
 
 0.0001+ 
 
 3 
 
 2.5 
 
 50 
 
 
 7) 
 
 0.1295 
 
 0.1279 
 
 0.0016- 
 
 3 
 
 3 
 
 55 
 
 
 8) 
 
 0.1617 
 
 0.1597 
 
 0.0020- 
 
 4 
 
 2 
 
 70 
 
 
 9) 
 
 0.1623 
 
 0.1622 
 
 0.0001- 
 
 4 
 
 3 
 
 80 
 
 Solutions containing sodium tungstate and ammonium 
 molybdate were allowed to stand from one to four days with 
 varying amounts of tartaric acid without giving any evidence 
 of reduction. 
 
 In the series which follows may be seen the results of a 
 
 Exp. 
 
 V 2 5 
 taken. 
 
 V 2 5 
 found. 
 
 Error. 
 
 Ammonium 
 molybdate. 
 
 Sodium 
 tung- 
 
 state. 
 
 Time 
 in 
 days. 
 
 Total 
 volume. 
 
 Tartaric 
 acid. 
 
 
 grm. 
 
 grin. 
 
 grin. 
 
 grm. 
 
 grm. 
 
 
 cms 
 
 grm. 
 
 (1) 
 
 0.1552 
 
 0.1558 
 
 0.0006+ 
 
 
 
 1 
 
 25 
 
 5 
 
 (2) 
 
 0.1289 
 
 0.1301 
 
 0.0012-f 
 
 
 < 
 
 1 
 
 25 
 
 5 
 
 (3) 
 
 0.2583 
 
 0.2587 
 
 0.0004+ 
 
 m 
 
 
 
 1 
 
 50 
 
 5 
 
 (4) 
 
 0.1293 
 
 0.1299 
 
 0.0006+ 
 
 1 
 
 
 1 
 
 25 
 
 6 
 
 (5) 
 
 0.2582 
 
 0.2591 
 
 0.0009+ 
 
 1 
 
 t 
 
 1 
 
 50 
 
 6 
 
 (6) 
 
 0.2582 
 
 0.2588 
 
 0.0006+ 
 
 1 
 
 
 1 
 
 50 
 
 5 
 
 (7) 
 
 0.1297 
 
 0.1308 
 
 0.0011+ 
 
 1 
 
 
 1 
 
 25 
 
 5 
 
 (8) 
 
 0.1291 
 
 0.1289 
 
 0.0002- 
 
 
 1 
 
 1 
 
 25 
 
 6 
 
 (9) 
 
 0.2582 
 
 0.2568 
 
 0.0014- 
 
 e 
 
 1 
 
 1 
 
 50 
 
 5 
 
 (10) 
 
 0.1293 
 
 0.1299 
 
 0.0006+ 
 
 i' 
 
 1 
 
 1 
 
 25 
 
 8 
 
 (11) 
 
 0.2582 
 
 0.2579 
 
 0.0003- 
 
 i 
 
 1 
 
 1 
 
 50 
 
 5 
 
 (12) 
 
 0.1550 
 
 0.1538 
 
 0.0012- 
 
 . 
 
 t 
 
 2 
 
 25 
 
 5 
 
 (13) 
 
 0.1556 
 
 0.1545 
 
 0.0011- 
 
 e 
 
 f 
 
 2 
 
 25 
 
 5 
 
 (14) 
 
 0.1289 
 
 0.1296 
 
 0.0007+ 
 
 t 
 
 f 
 
 2 
 
 25 
 
 5 
 
 (15) 
 
 0.1549 
 
 0.1527 
 
 0.0022- 
 
 0.5 
 
 
 2 
 
 25 
 
 5 
 
 (16) 
 
 0.1553 
 
 0.1548 
 
 0.0005- 
 
 1 
 
 m 
 
 2 
 
 25 
 
 5 
 
 (17) 
 
 0.1556 
 
 0.1554 
 
 0.0002- 
 
 1 
 
 
 2 
 
 25 
 
 5 
 
 (18) 
 
 0.1293 
 
 0.1310 
 
 0.0017+ 
 
 1 
 
 > 
 
 2 
 
 25 
 
 6 
 
 (19) 
 
 0.1295 
 
 0.1299 
 
 0.0004+ 
 
 . 
 
 y 
 
 2 
 
 25 
 
 6 
 
 (20) 
 
 0.1293 
 
 0.1289 
 
 0.0004- 
 
 r 
 
 i 
 
 2 
 
 25 
 
 7 
 
 (21) 
 
 0.1293 
 
 0.1301 
 
 0.0008+ 
 
 ( 
 
 
 3 
 
 25 
 
 5 
 
 (22) 
 
 0.1289 
 
 0.1299 
 
 0.0010+ 
 
 0.5 
 
 t 
 
 3 
 
 25 
 
 5 
 
 (23) 
 
 0.1293 
 
 0.1292 
 
 0.0001- 
 
 1 
 
 f 
 
 3 
 
 25 
 
 7 
 
 (24) 
 
 0.1556 
 
 0.1567 
 
 0.0011+ 
 
 1 
 
 t 
 
 3 
 
 30 
 
 5 
 
 (25) 
 
 0.1291 
 
 0.1289 
 
 0.0002- 
 
 1 
 
 i 
 
 3 
 
 25 
 
 7 
 
 (26) 
 
 0.1550 
 
 0.1557 
 
 0.0007+ 
 
 . 
 
 ^ 
 
 4 
 
 25 
 
 5 
 
 (27) 
 
 0.1554 
 
 0.1557 
 
 0.0003+ 
 
 1 
 
 t 
 
 4 
 
 25 
 
 5 
 
 (28) 
 
 0.1556 
 
 0.1557 
 
 0.0001+ 
 
 0.5 
 
 
 
 4 
 
 25 
 
 6 
 
8 APPLICATION OF CERTAIN ORGANIC ACIDS 
 
 number of determinations of vanadium in the presence of 
 molybdenum and tungsten made in the cold and allowed to 
 stand from one* to four days. It will be noticed that the 
 results on standing one day with five grams of tartaric acid 
 are for the most part satisfactory, and an increase in the 
 length of tune does not cause any apparent reduction of 
 the molybdenum. 
 
 Friedheim f has shown that vanadium is reduced from the 
 condition of the pentoxide to that of the tetroxide by boiling 
 with oxalic acid. The reduction is so complete that he has 
 developed a method for the estimation of vanadium upon this 
 reaction and shows that it may be applied in the presence of 
 molybdenum and tungsten, the acids of these elements not 
 being reduced by the oxalic acid. When the vanadic acid is 
 reduced the oxalic acid is oxidized and a definite amount of 
 carbon dioxide evolved according to the reaction. 
 
 V 2 6 + H 2 C 2 4 = V 2 4 + H 2 + 2C0 2 . 
 
 This carbon dioxide Friedheim conducts by an appropriate 
 form of apparatus into potassium hydroxide and weighs. 
 From this weight the amount of vanadic acid originally 
 present may be readily calculated. 
 
 We have applied the method of oxidation with standard 
 iodine described in the tartaric acid process to the residue 
 after boiling with oxalic acid. The method of treatment was 
 identical with that outlined at the beginning of this paper. 
 The results which follow in the table are for the most part 
 satisfactory and the method is certainly more easily applied 
 than Friedheim's process, the potassium hydroxide absorption 
 apparatus being unnecessary. 
 
 Having applied successfully both tartaric acid and oxalic 
 acid in the manner described, the action of citric acid applied 
 in the same manner suggested itself as a fitting conclusion to 
 the study of the action of this class of organic acids. In this 
 
 * Some of the determinations designated in the table as having stood one 
 day in reality stood only about fifteen hours, from 6 p. M. to 9 A. M. 
 t Zeitschr. anorg. Chem., i, 312. 
 
TO THE ESTIMATION OF VANADIUM. 
 
 Eip. 
 
 V,O 6 taken. 
 
 V,O 6 found. 
 
 Error. 
 
 Oxalic 
 acid. 
 
 Ammonium 
 molybdate. 
 
 Sodium 
 tungstate. 
 
 
 gnu. 
 
 gnu. 
 
 gnu. 
 
 gnu. 
 
 gnu. 
 
 gnu. 
 
 (1) 
 
 0.1806 
 
 0.1803 
 
 0.0003- 
 
 1 
 
 
 
 2 
 
 0.1950 
 
 0.1955 
 
 0.0005+ 
 
 1 
 
 
 
 3) 
 
 0.1959 
 
 0.1955 
 
 0.0004- 
 
 1 
 
 , 
 
 
 (4) 
 
 0.1950 
 
 0.1959 
 
 0.0009+ 
 
 1 
 
 t 
 
 
 (5) 
 
 0.1954 
 
 0.1977 
 
 0.0023+ 
 
 1 
 
 t 
 
 
 (6) 
 
 0.1956 
 
 0.1960 
 
 0.0004+ 
 
 1 
 
 m 
 
 
 (7) 
 
 0.1956 
 
 0.1964 
 
 0.0008+ 
 
 1 
 
 
 
 (8) 
 
 0.1956 
 
 0.1957 
 
 0.0001+ 
 
 1 
 
 
 
 (9) 
 
 0.3900 
 
 0.3899 
 
 0.0001- 
 
 2 
 
 
 
 (10) 
 
 0.3897 
 
 0.3917 
 
 0.0020+ 
 
 2 
 
 t 
 
 
 (11) 
 
 0.3903 
 
 0.3905 
 
 0.0002+ 
 
 2 
 
 
 
 (12) 
 
 0.1954 
 
 0.1959 
 
 0.0005+ 
 
 2 
 
 Y 
 
 
 (13) 
 
 0.1957 
 
 0.1960 
 
 0.0003+ 
 
 2 
 
 i 
 
 
 (14) 
 
 0.1954 
 
 0.1961 
 
 0.0007+ 
 
 2 
 
 i 
 
 
 (15) 
 
 0.1806 
 
 0.1818 
 
 0.0012+ 
 
 3 
 
 e 
 
 
 (16) 
 
 0.1807 
 
 0.1827 
 
 0.0020+ 
 
 3 
 
 
 
 (17) 
 
 0.1809 
 
 0.1803 
 
 0.0006- 
 
 3 
 
 Y 
 
 
 (18) 
 
 0.1956 
 
 0.1961 
 
 0.0005+ 
 
 3 
 
 t 
 
 1 
 
 (19) 
 
 0.3611 
 
 0.3617 
 
 0.0006+ 
 
 5 
 
 , 
 
 
 (20) 
 
 0.3616 
 
 0.3626 
 
 0.0010+ 
 
 5 
 
 i 
 
 
 case, as in the others, the reduction of the vanadic acid is 
 easily and quickly effected, but the oxidation with the iodine 
 is slower than in the presence of alkaline oxalates and tar- 
 trates. In the case of oxalic acid on standing about fifteen 
 minutes with the excess of iodine, when tartaric acid has 
 been used, the bleaching of the iodine continues from thirty to 
 forty-five minutes, but in the presence of the alkaline citrate 
 the time required is fully an hour. A large excess of tartaric 
 or oxalic acids does not seem to materially affect the results, 
 but hi the use of citric acid it is advisable to avoid a large 
 excess, which tends to give high results. Accordingly in the 
 following series of experiments it will be noticed that the 
 amounts of citric acid do not exceed two grams except where 
 ammonium molybdate or sodium tungstate are present, when 
 the ammonium or sodium base combines with part of the free 
 acid. The results follow, on p. 10. 
 
 The mode of proceeding in the estimation of vanadium by 
 the use of either tartaric, oxalic, or citric acid may be briefly 
 summarized as follows : To a solution of a vanadate with or 
 without a tungstate or molybdate, add approximately one 
 
10 
 
 ESTIMATION OF VANADIUM. 
 
 E 
 
 *I 
 
 
 V,O 4 taken. 
 
 V 2 O 6 found. 
 
 Error. 
 
 Citric 
 acid. 
 
 Ammonium 
 molybdate. 
 
 Sodium 
 tungstate. 
 
 
 
 
 gnn. 
 
 grin. 
 
 gmi. 
 
 gnn. 
 
 grm. 
 
 gnn. 
 
 
 (1 
 
 i 
 
 0.1956 
 
 0.1956 
 
 0.0000 
 
 1 
 
 
 
 
 2 
 
 
 0.3905 
 
 0.3921 
 
 0.0016+ 
 
 2 
 
 
 
 
 3 
 
 
 0.1960 
 
 0.1960 
 
 0.0000 
 
 1 
 
 
 
 
 4 
 
 
 0.1953 
 
 0.1960 
 
 0.0007+ 
 
 1 
 
 
 
 
 5 
 
 
 0.2088 
 
 0.2082 
 
 0.0006- 
 
 2 
 
 
 
 
 6 
 
 
 0.2100 
 
 0.2098 
 
 0.0002- 
 
 2 
 
 
 
 
 
 
 0.2092 
 
 0.2107 
 
 0.0015+ 
 
 1 
 
 
 
 
 8 
 
 i 
 
 0.2092 
 
 0.2107 
 
 0.0015+ 
 
 2 
 
 
 
 
 9 
 
 i 
 
 0.2096 
 
 0.2082 
 
 0.0014- 
 
 2 
 
 0.5 
 
 
 (1 
 
 
 
 i 
 
 0.2099 
 
 0.2116 
 
 0.0017+ 
 
 3 
 
 0.5 
 
 
 1 
 
 
 
 0.2095 
 
 0.2101 
 
 0.0006+ 
 
 2 
 
 
 6.5 
 
 (1 
 
 2 
 
 
 0.2099 
 
 0.2095 
 
 0.0004- 
 
 3 
 
 
 .0.6 
 
 gram of the acid for every tenth of a gram of substance to be 
 determined. Heat the solution to boiling, except in case tar- 
 taric acid be present with molybdic acid, when digestion for 
 from fifteen to twenty-four hours in the cold should be substi- 
 tuted. To the cold liquid add about five grams of potassium 
 bicarbonate for every gram of acid used. Add iodine in slight 
 excess and set aside until no further bleaching is noticeable. 
 Destroy the excess of iodine with arsenious oxide solution, add 
 starch, and titrate back with standard iodine. The total 
 amount of iodine used less the equivalent of the arsenious 
 oxide is the measure of the oxidation. 
 
 We have found it of advantage, when starting with a new 
 solution of the vanadate, to make one determination roughly 
 and to get from this rough determination the proportions of 
 acid and iodine to be used in the determinations to follow. 
 Large amounts of the acid and a large excess of the iodine 
 have been employed in many determinations without any 
 apparent unfavorable effect upon the results. The tendency, 
 however, under these circumstances is toward plus errors, 
 which may be avoided by following the above directions. 
 
Ill 
 
 THE DETERMINATION OF OXYGEN IN AIR AND 
 IN AQUEOUS SOLUTION. 
 
 BY D. ALBERT KREIDER.* 
 
 WHILE there is little to be hoped for by way of improve- 
 ment in the accuracy of present known methods for the deter- 
 mination of oxygen in the air, some choice as to manipulation 
 may nevertheless be desirable, and a process which is not lim- 
 ited wholly to the methods and apparatus of ordinary gas 
 analysis will doubtless often be found serviceable. 
 
 The very satisfactory results which I have obtained in the 
 determination of perchlorates by the action of the liberated 
 oxygen upon hydriodic acid through the medium of nitric 
 oxidef has led me to test this action upon the oxygen of the 
 air, where only the smaller amount of oxygen and its greater 
 dilution with nitrogen might be expected to be unfavorable. 
 However, with the apparatus and manipulation herein de- 
 scribed it will be seen that the method affords a means for the 
 determination of the oxygen of the air or of dissolved oxygen 
 with ease and rapidity and with sufficient accuracy for all 
 practical purposes. 
 
 The method in brief is simply the conducting of a known 
 volume of air through a strong solution of hydriodic acid in 
 the presence of nitric oxide ; subsequently neutralizing the 
 acid with potassium bicarbonate and titrating the liberated 
 iodine with standard decinormal arsenic solution from which 
 the equivalent volume of oxygen is readily calculated. By 
 several simple devices, to be described, all calculations may be 
 done away with and the percentage of oxygen seen imme- 
 
 * From Am. Jour. Sci., ii, 361. 
 
 t Am. Jour. Sci., 1, 287. Volume I, p. 316. 
 
12 THE DETERMINATION OF OXYGEN 
 
 diately by the volume of arsenic solution required for the 
 titration. 
 
 The volume of oxygen found by means of the arsenic 
 solution is, of course, under the standard conditions of tem- 
 perature and pressure (0 and 760 mm.), and it is therefore 
 essential either to calculate this volume into that which it would 
 occupy under the conditions of the experiment or to reduce 
 to the standard conditions of temperature and pressure the 
 volume of air taken. The latter plan is the more satisfactory 
 since by Lunge's ingenious device* the reduction can be 
 readily effected without any calculation and independently of 
 changing temperature and pressure. For my purpose the 
 following arrangement of two burettes answered admirably. 
 One burette graduated to 120 cm 3 contained over mercury 
 the same volume of moist air which 100 cm 3 of air at and 
 760 mm. would occupy under the given conditions, this stand- 
 ard being very carefully determined. By means of a T-tube 
 this standard burette was placed between and in connection 
 with the burette in which the volume of air to be analyzed 
 was measured, and a movable reservoir of mercury. Both 
 burettes were firmly fastened to a movable iron rod and the 
 zero marks accurately adjusted to the same level. By draw- 
 ing into the measuring burette a volume of air greater than 
 that required for which purpose a small bulb was attached 
 to the lower end of the burette, and then by raising the reser- 
 voir of mercury, compressing the air in the standard tube to 
 the 100 cm 3 mark, at the same time allowing the excess of 
 air to escape from the measuring burette, exactly 100 cm 3 of 
 air under the standard conditions of temperature and pressure 
 was obtained. To facilitate the adjustment, two strips 
 of wood were fastened to the rubber connection by means of 
 screw pinch-cocks in such a way that by closing one pinch- 
 cock the flow of mercury from the reservoir could be shut off, 
 and then by gradually tightening the other pinch-cock the 
 mercury would be forced out of the rubber, and thus an easy 
 and accurate adjustment to the 100 cm 3 mark be secured. 
 * Zeitschr. angew. Chem., 1890, 139. 
 
IN AIR AND IN AQUEOUS SOLUTION. 13 
 
 The apparatus in which the action of the oxygen upon 
 hydriodic acid was effected consisted of a 300 cm 3 bulb pipette, 
 both ends of which were cut off short and sealed to glass 
 stop-cocks. The tube from one of the stop-cocks was cut off 
 short after being tapered and constricted so as to hold a rubber 
 connector tightly, while the tube from the other stop-cock 
 was left sufficiently long to reach to the bottom of a 500 cm 8 
 Erlenmeyer beaker. These tubes are preferably of about 
 3 mm. bore, since for the several connections all air may be 
 expelled from tubes of this size by displacement with water. 
 In order to expel all air from the flask, instead of passing a 
 current of carbon dioxide as was done in the determination 
 of perchlorates, tune was saved by first filling the flask with 
 water, which was then displaced by pure carbon dioxide 
 (prepared as described below) and the flask subsequently 
 exhausted, which was accomplished instantaneously by the 
 device described in the article on perchlorates. The required 
 amounts of potassium iodide solution, hydrochloric acid and 
 nitric oxide were drawn in in the order named, after which 
 the measured volume of air was gradually admitted Awhile 
 the bulb was constantly agitated so as to keep the hydriodic 
 acid continually renewed along the surface of the bulb. 
 The shaking was continued for a minute or two until the 
 action was completed, when a dilute solution of potassium 
 bicarbonate was admitted. The carbon dioxide liberated 
 forces the liquid from the bulb into a beaker which contains 
 bicarbonate in amount sufficient, as previously determined, to 
 neutralize all the acid taken. When the exit is too slow more 
 bicarbonate may be admitted through the other stop-cock, and 
 after neutralization has been completed the bulb may be 
 washed out without any danger from the admission of air. 
 
 All the water employed, both for the solution of potassium 
 iodide and for the various connections, was free of oxygen. 
 It was prepared by filling a three-liter flask with distilled 
 water and boiling until the volume of the liquid was reduced 
 about one-third, when the flask was closed by a doubly 
 perforated rubber stopper and fitted as a wash bottle. By 
 
14 THE DETERMINATION OF OXYGEN 
 
 means of the tube which reached below the surface of the 
 water, pure carbon dioxide was passed through while the 
 water was still boiling, which together with the escaping 
 steam was sure to expel all oxygen. Then the heat was 
 removed and the current of carbon dioxide continued until 
 the boiling ceased, when the escape tube was closed by a 
 piece of rubber tubing and screw pinch-cock. As the water 
 cooled it was well shaken while still in connection with 
 the carbon dioxide generator, and thus became saturated 
 with the gas, which was then pumped in under considerable 
 pressure by the little hand pump described in a previous 
 paper from this laboratory. By this means the water could 
 be drawn as needed without the introduction of any air. 
 The escape tube was provided with a rubber tube and screw 
 pinch-cock, and a long, slender nozzle which could be in- 
 serted into the tubes of the absorption apparatus. A bottle 
 thus charged sufficed for all the determinations and required 
 only an occasional supply of carbon dioxide when large 
 draughts of water were required for making the potassium 
 iodide solution. 
 
 The potassium iodide solution was made up to contain one 
 gram of the salt in thirty cubic centimeters of water, and 
 was contained in convenient form in an ordinary wide- 
 mouthed bottle fitted as a wash bottle, and graduated ap- 
 proximately for each thirty cubic centimeters' volume the 
 amount usually taken. The potassium iodide was weighed 
 into the bottle, which was then closed and all air expelled 
 by a current of carbon dioxide, when the desired amount of 
 water, free of oxygen, was drawn in, and attachment again 
 made with the carbon dioxide generator. After allowing the 
 gas to pass for several minutes the exit was closed and the 
 gas pumped in by the little hand pump. Inasmuch as this 
 solution, when it was used, was drawn into an exhausted 
 bulb, the bottle could be emptied without ever exposing its 
 contents to the air. 
 
 Nitric oxide was generated very satisfactorily according to 
 Professor Gooch's method by the action of nitric acid 
 
IN AIR AND IN AQUEOUS SOLUTION. 15 
 
 upon globules of copper in a Kipp generator. When the 
 nitric acid is diluted with an equal volume of water the 
 evolution of the gas is sufficiently rapid without the applica- 
 tion of heat, but contamination by the higher oxide is more 
 likely. However, since it is necessary, in order to be certain 
 of purity, to pass the gas through an acidified solution of 
 potassium iodide before applying it to the determination 
 of oxygen, whatever higher oxide may be present will be 
 reduced. By passing the gas, as it issued from the generator, 
 through a set of Geisler bulbs containing an acidified solution 
 of potassium iodide and washing with potassium iodide solu- 
 tion, the perfectly purified gas was obtained. Theoretically, 
 only a small amount of the nitric oxide is required for the 
 transference of the oxygen to the hydriodic acid, but when 
 too little is taken the action is very slow. On the other 
 hand, too large an amount relieves the vacuum to such an 
 extent as to interfere with the introduction of the air. A 
 little device to measure the volume of gas taken was there- 
 fore attached to the generator. It consisted of a tube filled 
 with water and roughly graduated for every five cubic centi- 
 meters, so attached to the generator that the gas would enter 
 by displacement of the water, which would descend to a 
 lower bulb, and as the gas was withdrawn the water would 
 again take its place. Fifteen cubic centimeters of the gas 
 was found a convenient and satisfactory amount for the 
 analysis. 
 
 Carbon dioxide was generated in a Kipp generator, the 
 acid and marble of which had been previously boiled and 
 contained a little cuprous chloride. To remove a trace of 
 reducing matter which the gas was found to contain, it was 
 first passed through a solution of iodine and washed with 
 potassium iodide. 
 
 For the titration a decinormal solution of arsenious oxide 
 (4.95 grms. to the liter) was employed : 1 cubic centimeter 
 being equal to 0.559846 cm 3 of oxygen at and 760 mm. 
 when the weight of a liter of oxygen at and 760 mm. is 
 taken as 1.42895 grin. When the volume of air taken 
 
16 
 
 THE DETERMINATION OF OXYGEN 
 
 is 100 cm 3 under standard conditions of temperature and 
 pressure, as obtained by Lunge's device, the following table, 
 calculated for the volume of oxygen equivalent to the volume 
 of arsenic solution, shows directly the percentage of oxygen 
 corresponding to the reading of the burette. The correction 
 necessary for the fraction of a tenth of a cubic centimeter 
 of the arsenic solution is obtained with sufficient accuracy 
 by simply multiplying by 0.005. 
 
 RELATION OF ARSENIC TO OXYGEN. 
 
 
 
 Correction for 
 
 SKA. 
 
 Oxygen equivalent 
 at and 760 mm. 
 
 0.01 cms AsjjO,, 
 
 cms 
 
 37.0 
 
 cm 3 
 20.714 
 
 0.005 
 
 37.1 
 
 20.770 
 
 
 37.2 
 
 20.826 
 
 
 37.3 
 
 20.882 
 
 
 37.4 
 
 20.938 
 
 
 37.5 
 
 20.994 
 
 
 37.6 
 
 21.050 
 
 
 37.7 
 
 21.106 
 
 
 37.8 
 
 21.162 
 
 
 37.9 
 
 21.218 
 
 
 38.0 
 
 21.274 
 
 
 Table I shows the results obtained in a series of determina- 
 tions. Experiments (1) to (11) inclusive were made upon por- 
 tions of air collected over water on March 28, measured in an 
 ordinary gas burette and reduced to the standard conditions 
 of temperature and pressure. The remainder of the deter- 
 minations were made upon air collected on April 8, each 
 portion having been measured in the apparatus described, for 
 the reduction to standard conditions. 
 
 No correction was found necessary for the blank determina- 
 tions, since when boiled water was used the solution was only 
 faintly colored with iodine, which requires only a drop of 
 arsenic solution to bleach it. As is evident from the table, the 
 determinations according to this method are not reliable beyond 
 0.05 per cent, but for practical purposes this is sufficiently 
 accurate. For the sake of comparison two determinations by 
 
IN AIR AND IN AQUEOUS SOLUTION. 
 TABLE I. 
 
 17 
 
 Kxp. 
 
 Volume of Air 
 reduced to 
 and 760 mm. 
 
 l>" 
 required. 
 
 Volume of 
 Oxygen found at 
 and 760 mm. 
 
 Per cent of 
 Oxygen in Air. 
 
 
 cm 3 
 
 cms 
 
 cnjS 
 
 
 (1) 
 
 91.18 
 
 34.06 
 
 19.07 
 
 20.91 
 
 (2) 
 
 91.73 
 
 34.47 
 
 19.30 
 
 21.04 
 
 (3) 
 
 90.84 
 
 34.25 
 
 19.17 
 
 21.11 
 
 (4) 
 
 90.60 
 
 34.20 
 
 19.16 
 
 21.13 
 
 (5) 
 
 86.06 
 
 32.55 
 
 18.22 
 
 21.17 
 
 (6) 
 
 85.96 
 
 32.40 
 
 18.14 
 
 21.10 
 
 (7) 
 
 86.49 
 
 32.53 
 
 18.21 
 
 21.06 
 
 (8) 
 
 87.85 
 
 33.00 
 
 18.47 
 
 21.03 
 
 (9) 
 
 44.17 
 
 16.60 
 
 9.29 
 
 21.04 
 
 (10) 
 
 44.11 
 
 16.70 
 
 9.35 
 
 21.19 
 
 (11) 
 
 44.54 
 
 16.80 
 
 9.41 
 
 21.12 
 
 (12) 
 
 100.00 
 
 37.44 
 
 20.96 
 
 20.96 
 
 (13) 
 
 100.00 
 
 37.54 
 
 21.01 
 
 21.01 
 
 (14) 
 
 100.00 
 
 37.50 
 
 20.99 
 
 20.99 
 
 (15) 
 
 100.00 
 
 37.57 
 
 21.03 
 
 21.03 
 
 (16) 
 
 100.00 
 
 37.47 
 
 20.97 
 
 20.97 
 
 (17) 
 
 100.00 
 
 37.50 
 
 20.99 
 
 20.99 
 
 the pyrogallic acid method were made upon a portion of the 
 same air used in the last experiments, the results being 20.93 
 and 20.88 per cent respectively. While the pyrogallic acid 
 method is capable of much greater accuracy when applied in 
 Hempel's improved apparatus, in ordinary burettes it will 
 probably not yield more closely agreeing results than the 
 above method. 
 
 Determination of dissolved Oxygen. A deter- 
 mination of oxygen dissolved in water can be 
 completed by the above method in about ten 
 minutes by means of the apparatus illustrated by 
 the accompanying figure. 
 
 The apparatus consisted of a flask of about 300 
 cm 3 capacity, into the bottom of which was sealed 
 a stop-cock with a long exit tube. Upon the 
 neck was cut the fiducial circle c and immediately 
 above this stop-cock e was sealed as shown. The 
 neck of the flask was drawn out and sealed to 
 stop-cock d and the bulb, a, of about 30 cm 3 
 capacity blown in it. The capacity of the apparatus be- 
 
 VOL. II. 2 
 
 FIG. 17. 
 
18 THE DETERMINATION OF OXYGEN 
 
 tween stop-cock, &, and the fiducial mark, c, was carefully 
 determined. 
 
 The manipulation for the determination of dissolved oxygen 
 was as follows : The flask was held in the position shown by 
 a clamp fastened to a movable support. Stop-cock b being 
 closed, the water was admitted through e and the air allowed 
 to escape through d until the level of water was that indicated 
 by the line /. (When the water to be examined is not 
 saturated with air, the flask must first be filled with carbon 
 dioxide and the water entered by replacement of that gas.) 
 With d closed, sufficient water was allowed to escape through 
 b to bring the surface to e, which was then closed. The nitric 
 oxide generator was then attached to d, and by opening b the 
 gas was allowed to replace the water until the meniscus 
 coincided with c, when d was closed and the generator 
 disconnected. Two cubic centimeters of strong hydrochloric 
 were introduced through e by expelling nitric oxide through 
 d, in which a drop of water formed an effective trap to prevent 
 the entrance of air. Then the potassium iodide was admitted 
 in the same way. The solution of iodide for this purpose 
 was free of oxygen and contained one gram in three cubic 
 centimeters. It was kept under pressure of carbon dioxide in 
 the bottle previously described, and by means of a long nozzle 
 could be conducted to the bottom of eh and thus be admitted 
 with but momentary and slight contact with the air. The 
 tube eh contained approximately three cubic centimeters. 
 With all the stop-cocks closed, the flask was inverted several 
 times and thoroughly shaken, at the same tune washing out 
 the ends of the stop-cocks with distilled water. After again 
 placing the apparatus in its position, enough potassium bicar- 
 bonate solution was admitted through e to expel all the 
 nitric oxide through d\ the bulb, a, holding sufficient of the 
 bicarbonate to neutralize all the acid taken. The bicarbonate 
 being heavier quickly diffuses through the contents of the 
 flask and neutralizes the acid; d and e are kept closed for a 
 minute with b open so as to allow sufficient of the liquid to 
 escape into a beaker containing some bicarbonate to provide 
 
IN AIR AND IN AQUEOUS SOLUTION. 
 
 19 
 
 space for the carbon dioxide evolved. Then the flask is washed 
 out and its contents titrated with arsenic. 
 
 The bleaching, by the aid of starch for the final reaction, 
 can be accurately read to a single drop Usually the reading 
 was verified by adding a drop of $ iodine solution, which 
 produced the characteristic color. 
 
 Table II gives the results of a series of determinations. 
 
 TABLE IL 
 
 Volume of Water 
 taken. 
 
 Temperature. 
 
 As 2 3 required. 
 
 Volume of Oxygen 
 dissolved in lOOOcmS 
 of water at 760 mm. 
 
 cm* 
 
 C. 
 
 cm 8 
 
 cm 3 
 
 314.63 
 
 20 
 
 3.42 
 
 6.04 
 
 314.63 
 
 20 
 
 3.45 
 
 6.09 
 
 314.63 
 
 20 
 
 3.40 
 
 6.00 
 
 314.63 
 
 20 
 
 3.41 
 
 6.02 
 
 314.63 
 
 20 
 
 3.43 
 
 6.05 
 
 314.63 
 
 20 
 
 3.40 
 
 6.00 
 
 314.63 
 
 20 
 
 3.36 
 
 5.93 
 
 314.63 
 
 20 
 
 3.40 
 
 6.00 
 
 314.63 
 
 20 
 
 3.40 
 
 6.00 
 
 314.63 
 
 20 
 
 3.50 
 
 6.18 
 
 314.63 
 
 20 
 
 3.38 
 
 5.96 
 
 314.63 
 
 20 
 
 3.40 
 
 6.00 
 
 The mean of these determinations gives 6.022 cm 3 of oxygen 
 as the amount dissolved in distilled water at 20 C. and 760 
 mm., and while some of the determinations vary considerably 
 from this mean, as a whole they are fairly accordant. This 
 method, moreover, is applicable to carbonated water. 
 
IV 
 
 A METHOD FOR THE SEPARATION OF 
 ALUMINUM FROM IRON. 
 
 BY F. A. GOOCH AND F. S. HAVENS.* 
 
 OF the well-known methods for the separation of aluminum 
 from iron by the action, for example, of an alkaline hydroxide 
 in aqueous solution or by fusion of the mixed oxides in 
 potassium or sodium hydroxide ; by reduction of the iron 
 oxide to the metal by heating in hydrogen, with the 
 subsequent solution of the metallic iron in hydrochloric 
 acid; by boiling the nearly neutral solution of the salts of 
 aluminum and iron with sodium thiosulphate either with or 
 without sodium phosphate ; by acting with hydrogen sulphide 
 or ammonium sulphide upon solutions of the salts containing 
 also an ammoniacal citrate or tartrate no single process can 
 be said to be ideal as regards directness, rapidity and accuracy 
 of working. We have deemed it not superfluous, therefore, 
 to attempt the utilization of a reaction which should apparently 
 be capable of effecting directly and quickly the separation of 
 aluminum from iron under conditions easily attainable. 
 
 It is known t that the hydrous aluminum chloride A1C1 3 .6H 2 O 
 is very slightly soluble in strong hydrochloric acid, while ferric 
 chloride, on the other hand, is extremely soluble in that 
 medium. It is this difference of relation of which we wished 
 to take advantage. 
 
 It appeared at the outset that crude aluminum chloride 
 could be freed from every trace of a ferric salt by dissolving it 
 in the least possible amount of water, saturating the cooled 
 solution with gaseous hydrochloric acid, filtering upon asbestos 
 
 * From Am. Jour. Sci., ii, 416. 
 
 t Gladysz, Ber. Dtsch. chem. Ges., xvi, 447. 
 
SEPARATION OF ALUMINUM FROM IRON. 21 
 
 in a filtering crucible or cone, and washing the crystalline 
 precipitate with the strongest hydrochloric acid. Aluminum 
 chloride prepared in this way gave no trace of color when 
 dissolved in water and tested with potassium sulphocyanide. 
 The correlative question as to how much aluminum chloride 
 goes into solution under the conditions was settled by taking 
 a portion of the pure aluminum chloride, dissolving it in a 
 very little water, diluting the solution with strong hydrochloric 
 acid, saturating the cooled liquid with the gaseous acid, filtering 
 on asbestos, precipitating by ammonia the aluminum salt in the 
 nitrate and weighing the ignited oxide. 
 
 From 10 cm 3 of such a filtrate we obtained in two deter- 
 minations 0.0022 grm. and 0.0024 grm. of the oxide, the mean 
 of which corresponds to 23 parts of the oxide or 109 parts of 
 the hydrous chloride in 100,000 parts of the strong hydrochloric 
 acid. This degree of solubility, though inconsiderable when 
 the objective point is the preparation of the pure salt of 
 aluminum, is obviously incompatible with the attainment of 
 quantitative accuracy in the retention of the aluminum. We 
 have found, however, that various mixtures of anhydrous ether 
 and the strongest hydrochloric acid can be used satisfactorily 
 as solvents for the iron chloride, while the aluminum chloride 
 is insoluble to a very high degree in a mixture of hydrochloric 
 acid and ether taken in equal parts and thoroughly saturated 
 with gaseous hydrochloric acid at the atmospheric temperature. 
 We found that 50 cm 3 of the solution of aluminum chloride, 
 obtained by mixing about 0.1 grm. of the hydrous chloride 
 (dissolved in 2 cm 3 of water) with the mixture of pure, 
 specially prepared aqueous hydrochloric acid and ether in equal 
 parts and again saturating the liquid at 15 C. with gaseous 
 hydrochloric acid, left upon evaporation and ignition 0.0004 
 grm. in each of two experiments results which indicate a 
 maximum solubility corresponding to 1 part of the oxide or 
 approximately 5 parts of the chloride in 125,000 parts of the 
 equal mixture of ether and aqueous hydrochloric acid of full 
 strength. 
 
 Pure aqueous hydrochloric acid of full strength mixes 
 
22 A METHOD FOR THE SEPARATION 
 
 perfectly with its own volume of anhydrous ether, but it is 
 a curious fact that the addition to this mixture of any very 
 considerable amounts of a solution of ferric chloride in strong 
 hydrochloric acid determines the separation of a greenish oily 
 ethereal solution of the ferric salt upon the surface of the 
 acid. The addition of more aqueous acid does not change the 
 conditions essentially, but more ether renders the acid and 
 the oily solution completely miscible. The ferric chloride 
 seems to abstract ether from the ether-acid mixture and, then 
 dissolved in the ether, remains to some extent immiscible with 
 the aqueous acid thus left until the addition of more ether 
 restores to the mixture that which was taken from it by the 
 ferric chloride. Our experiments show that, while for the 
 separation of insoluble aluminum chloride from certain small 
 amounts of soluble ferric chloride the mixture of the strongest 
 aqueous hydrochloric acid and ether in equal parts serves a 
 most excellent purpose, when larger amounts of ferric chloride 
 are to be dissolved ether must be added proportionately in 
 order to prevent the separation of the ethereal solution of 
 ferric chloride from the rest of the liquid. 
 
 Great care was taken to insure the purity of the aluminum 
 chloride used in the test experiments. The so-called pure 
 chloride of commerce was dissolved in the least possible 
 amount of water and this solution was treated with a large 
 volume of strong hydrochloric acid. The chloride thus 
 obtained, free from iron, but possibly contaminated (as we 
 found by experience) with some alkaline chloride, was dis- 
 solved in water and converted by ammonia to the form of the 
 hydroxide, which was thoroughly washed and dissolved in hot 
 hydrochloric acid of half-strength. From this solution, after 
 cooling, gaseous hydrochloric acid precipitated the hydrous 
 chloride in pure condition. The chloride thus prepared was 
 dissolved in water and the strength of the solution was 
 determined by precipitating the hydroxide from definite 
 portions, and weighing the ignited oxide in the usual manner. 
 
 In the experiments recorded in Table I, measured portions 
 of the standardized solution were submitted to the treatment 
 
OF ALUMINUM FROM IRON. 
 TABLE I. 
 
 23 
 
 Bxp. 
 
 A1 2 3 taken 
 in solution as 
 the chloride. 
 
 found. 
 
 Final 
 volume. 
 
 Error. 
 
 
 grin. 
 
 gnn. 
 
 cm 
 
 grm. 
 
 
 1) 
 
 0.0761 
 
 0.0746 
 
 50 
 
 0.0015- 
 
 
 2 
 
 0.0761 
 
 0.0745 
 
 50 
 
 0.0016- 
 
 
 3) 
 
 0.0761 
 
 0.0741 
 
 50 
 
 0.0020- 
 
 
 4) 
 
 0.0761 
 
 0.0734 
 
 50 
 
 0.0027- 
 
 
 5) 
 
 0.0761 
 
 0.0756 
 
 50 
 
 0.0005- 
 
 
 6) 
 
 0.0157 
 
 0.0149 
 
 45 
 
 0.0008- 
 
 
 7 
 
 0.0157 
 
 0.0147 
 
 40 
 
 0.0010- 
 
 8) 
 
 0.0157 
 
 0.0144 
 
 45 
 
 0.0013- 
 
 (9) 
 
 0.0480 
 
 0.0481 
 
 30 
 
 0.00014- 
 
 (10) 
 
 0.0960 
 
 0.0957 
 
 30 
 
 0.0003- 
 
 with hydrochloric acid and ether. The essential thing in the 
 process is to have at the end a mixture of the strongest aqueous 
 hydrochloric acid with an equal volume of anhydrous ether 
 saturated at a temperature of about 15 C. The most con- 
 venient way to secure these conditions seems to be to mix the 
 aqueous solution of the aluminum salt with a suitable volume 
 of the strongest aqueous hydrochloric acid enough to make 
 the entire volume something between 15 and 25 cm 3 to 
 saturate this mixture with gaseous hydrochloric acid while the 
 liquid is kept cool by immersing the receptacle containing it 
 in a current of running water, to intermix a volume of ether 
 equal to the volume of the liquid, and finally, to treat the 
 ethereal mixture once more with the gaseous acid to insure 
 saturation. The precipitated crystalline chloride was collected 
 upon asbestos in a perforated crucible, washed with a previously 
 prepared mixture of hydrochloric acid and ether carefully 
 saturated with the gaseous acid at 15 C., and either ignited 
 after careful drying at 150 or redissolved in water, converted 
 to the hydroxide by ammonia in the usual way and weighed 
 as the oxide after filtration, washing, and ignition. In experi- 
 ments (1) to (4) the precipitated chloride was ignited directly ; 
 in experiment (5) the ignition was made with great care in an 
 atmosphere of superheated steam ; and in experiments (6) to 
 (10) the chloride was dissolved, precipitated as the hydroxide, 
 and weighed as the oxide. 
 
24 A METHOD FOR THE SEPARATION 
 
 The experiments in which the chloride was converted to the 
 hydroxide before ignition show upon the average an absolute 
 loss of about 0.0006 grm. : the single experiment in which the 
 ignition took place in steam shows about the same loss 0.0005 
 grm. ; while in those experiments in which the chloride was 
 dried and then ignited directly, the average loss amounts to 
 about 0.0020 grm. The error of the process which involves 
 the precipitation of the aluminum as the hydroxide, falls within 
 reasonable limits, but it is plain that the direct ignition of the 
 chloride is liable to error, which may possibly be explicable as 
 a mechanical loss occasioned by the too rapid evolution of the 
 hydrochloric acid and water of crystallization, or, possibly, as 
 the result of a very slight volatilization of the aluminum still 
 holding chlorine in spite of the decomposing action of the 
 water upon the chloride. In either case, it would seem to be 
 reasonable to suppose that a layer of some easily volatilizable 
 oxidizer placed upon the aluminum chloride might serve to 
 obviate the difficulty in the one case, by serving as a screen 
 to diminish mechanical transportation of the non-volatile 
 material ; and in the other, by acting as an agent to promote 
 the exchange of chlorine for oxygen on the part of the 
 aluminum chloride. 
 
 We have tried, therefore, the expedient of covering the 
 aluminum chloride before ignition with a layer of mercuric 
 oxide, which of itself left no appreciable residue when it 
 volatilized. The hydrous chloride was collected as usual upon 
 the asbestos in a perforated crucible, dried for a half-hour at 
 150 C., covered with about 1 grm. of the pure mercuric oxide, 
 gently heated with great care under a suitable ventilating flue, 
 and finally ignited over the blast. The results are given 
 below (see Table II). 
 
 It is obvious, therefore, that the precipitation of the crystal- 
 line hydrous aluminum chloride from solutions of the pure salt 
 is perfectly feasible and very complete, when effected by 
 aqueous hydrochloric acid and ether thoroughly saturated with 
 the gaseous acid and kept cool ; and that the conversion of the 
 chloride into the weighable form of the oxide is best effected 
 
OF ALUMINUM FROM IRON. 
 
 25 
 
 TABLE II. 
 
 Exp. 
 
 Al,0,, taken 
 in solution as 
 the chloride. 
 
 Al a O 3 found 
 by ignition 
 with HgO. 
 
 Final 
 volume. 
 
 Error. 
 
 
 grm. 
 
 grm. 
 
 cm 3 
 
 grm. 
 
 (1) 
 
 0.0761 
 
 0.0758 
 
 25 
 
 0.0003- 
 
 
 0.0761 
 
 0.0754 
 
 25 
 
 0.0007- 
 
 (3) 
 
 0.0761 
 
 0.0761 
 
 25 
 
 0.0010- 
 
 by ignition under a layer of mercuric oxide, or by dissolving 
 it in water and precipitating it as the hydroxide to be 
 afterward washed, dried, and ignited. Of the two methods the 
 former is by far the more convenient. 
 
 The precipitation of the aluminum chloride in pure condition 
 from solutions containing ferric chloride ought not, it would 
 seem, to present any difficulty, providing only that the 
 precaution is taken to have present a sufficient excess of ether. 
 The question was put to the test of experiment with the 
 results recorded in Table III. 
 
 TABLE III. 
 
 Exp. 
 
 A1,O 3 taken 
 in solution as 
 the chloride. 
 
 A1 5 O 3 found 
 by ignition 
 with HgO. 
 
 Fe 2 3 
 present as 
 chloride. 
 
 Final 
 volume. 
 
 Error. 
 
 
 grin. 
 
 grm. 
 
 grm. 
 
 cm 3 
 
 grm. 
 
 (1) 
 
 0.0761 
 
 0.0757 
 
 0.15 
 
 25-30 
 
 0.0004- 
 
 2 
 
 0.0761 
 
 0.0756 
 
 0.15 
 
 25-30 
 
 0.0005- 
 
 3 
 
 0.0761 
 
 0.0755 
 
 0.15 
 
 25-30 
 
 0.0006- 
 
 4) 
 
 0.0761 
 
 0.0755 
 
 0.15 
 
 25-30 
 
 0.0006- 
 
 Measured portions of the standardized solution of aluminum 
 chloride were evaporated nearly to dryness in a platinum dish, 
 an amount of pure ferric chloride equivalent to about 0.15 
 grm. of the oxide was added in a very little water, 15 cm 3 of 
 the mixture of strong hydrochloric acid and ether in equal 
 parts were introduced, the liquid was saturated at 15 C. with 
 gaseous hydrochloric acid (the dish being held in a convenient 
 device for cooling it by running water), 5 cm 3 more of ether 
 were added to secure complete miscibility of the solutions, and 
 
26 SEPARATION OF ALUMINUM FROM IRON. 
 
 more gas passed to perfect saturation. The aluminum chloride 
 was collected upon asbestos in a perforated crucible, washed 
 with a mixture of ether and aqueous hydrochloric acid 
 thoroughly saturated with the gaseous acid, dried at 150 C. 
 for a half -hour, covered with 1 grm. of pure mercuric oxide, 
 and ignited at first gently and finally over the blast. 
 
 The results show plainly a very satisfactory limit of error. 
 
THE ESTIMATION OF MOLYBDENUM 
 IODOMETRICALLY. 
 
 BY F. A. GOOCH * 
 
 IN a former paper from this laboratory f several modes of 
 applying hydriodic acid to the reduction of molybdic acid 
 were studied. It was found, first, that the digestion process 
 of Mauro and Danesi f is of very limited applicability, 
 owing to the fact that the reaction of reduction is reversible. 
 Secondly, it appeared that the use of the same reaction by 
 Friedheim and Euler in a distillation process, so arranged 
 that the iodine set free in the reduction might be caught in 
 the distillate and titrated to serve as the measure of the 
 reducing action, was not sufficiently regular because of in- 
 attention to minor details. It was shown that by taking 
 care to adjust the conditions constant results might be 
 obtained. Thirdly, the fact was developed that by simply 
 boiling the solution under well defined conditions in an 
 ordinary Erlenmeyer flask, partly closed by a simple trap, 
 the reduction of the molybdic acid proceeded regularly, and 
 that the addition of standard iodine to the solution made 
 alkaline with sodium bicarbonate served to restore the original 
 condition of oxidation of the molybdic acid. The results 
 of this treatment were shown to be accurate. 
 
 In a recent paper || Friedheim has seen fit to make our 
 modifications of the distillation process the subject of attack. 
 Friedheim's comments upon the third method discussed (as 
 
 * From Am. Jour. Sci., iii, 237. 
 
 t Gooch and Fairbanks, Am. Jour. Sci., ii, 157. Volume I, p. 375. 
 
 J Zeitschr. anal. Chem.. xx, 507. 
 
 Ber. Dtsch. chem. Ges., xxviii, 2066. 
 
 || Ber. Dtsch. chem. Ges., xxix, 2981. 
 
28 THE ESTIMATION OF MOLYBDENUM. 
 
 well as upon a subsequent application of the process) * are 
 evidently prompted wholly by personal opinion and demand 
 no further attention. With reference to Friedheim's denial 
 of the necessity of modification in the Friedheim and Euler 
 treatment the case is different. 
 
 The process of Friedheim and Euler consists, it will be 
 remembered, in treating the soluble molybdate, or the solu- 
 tion of molybdic acid in sodium hydroxide, with potassium 
 iodide and hydrochloric acid in a Bunsen apparatus, boiling 
 until the solution is of a clear green color, collecting the 
 iodine distilled in potassium iodide, and titrating it with 
 sodium thiosulphate. We found that the development of 
 the green color was not a sufficient criterion of the exact 
 reduction of the molybdic acid to the condition of the 
 pentoxide and of the removal of the iodine which should be 
 theoretically set free. To accomplish that end we found it 
 safer and more convenient to start the distillation with a 
 definite volume (40 cm 3 ) of liquid and boil until a definite 
 volume (25 cm 3 ) was reached, care being taken with regard 
 to the strength of acid and the excess of potassium iodide 
 employed. Experience showed unmistakably that in order 
 to avoid the decomposing action of the air upon the hot 
 vaporous hydriodic acid in the retort, it was necessary to 
 go beyond the measures advised by Friedheim and Euler 
 (namely, to warm the retort and its contents slowly, heating 
 to boiling only when the connecting tube was well filled 
 with iodine vapor and the tendency toward back-suction of 
 the liquid in the receiver began to appear) and to conduct 
 the operation in a simple little apparatus (the retort holding 
 about 100 cm 3 ) put together entirely with sealed and ground 
 joints, as shown in the figure of the former paper, so ar- 
 ranged that a current of purified carbon dioxide could be 
 passed through retort and receiver during the distillation. 
 With this apparatus we were able to determine with accuracy 
 the point of concentration at which the free iodine left the 
 liquid, the molybdic acid having been converted to the con- 
 * Am. Jour. jSci. ii, 181. Volume I, p. 391. 
 
THE ESTIMATION OF MOLYBDENUM. 29 
 
 dition of the pentoxide. It was found that if dependence is 
 placed upon the occurrence of the so-called clear green color 
 of the liquid to determine the end of the distillation, it may 
 frequently happen that free iodine remains in the residue. 
 This takes place, it will be observed, in the atmosphere of 
 carbon dioxide, so that the presence of the free iodine can 
 by no possibility be attributed to the action of atmospheric 
 air upon the hydriodic acid remaining after the distillation 
 is complete. On the other hand, it appeared that, if the 
 distillation is pushed too far, the molybdenum pentoxide may 
 be still further reduced with consequent evolution of more 
 than the expected amount of iodine. The attainment of an 
 exact degree of reduction with the expulsion of the corre- 
 sponding amount of iodine becomes, therefore, a matter of 
 chance unless further precautions are taken. We found in 
 our experiments that, if amounts less than 0.3 grm. of the 
 molybdic acid are introduced in soluble form into the 100 cm 3 
 retort with a not too great excess of potassium iodide, and 
 the 40 cm 3 of liquid so constituted that 20 cm 3 of it shall be 
 water and 20 cm 3 the strongest hydrochloric acid, the reduc- 
 tion proceeds with a fair degree of regularity in the manner 
 expected. We found it important to restrict the excess of 
 potassium iodide so that it shall never exceed the theoretical 
 requirement by more than 0.5 grm. 
 
 Our determinations with the pure molybdenum trioxide 
 showed errors varying from 0.0010 grm. -|- to 0.0007 grm. ; 
 the variations from theory in the experiments with ammonium 
 molybdate ranged from 0.0011 grm. -f- to 0.0011 grm. . If 
 these results are compared with those given by Friedheim 
 and Euler, the advantage is a little in favor of the latter; 
 but a scrutiny of the figures given by Friedheim and Euler 
 develops the fact that the apparent accuracy of their work 
 is founded upon miscalculations. This fact was known to us 
 at the tune of our former writing, but we did not consider it 
 essential then to make the matter public. The recent attack 
 of Friedheim makes that course now necessary. 
 
 Herewith is reproduced a table of results obtained by Fried- 
 
30 
 
 THE ESTIMATION OF MOLYBDENUM. 
 
 heim and Euler in the test of their method upon ammonium 
 molybdate, shown by analysis to contain 81.49 per cent of 
 molybdenum trioxide. The figures which are incorrect are 
 enclosed in brackets: 
 
 OEIGINAL FIGUKES OF FRIEDHEIM AND EULEB. 
 
 
 
 
 Per cent of 
 
 Molybdate 
 taken. 
 
 Na 2 S 2 3 
 used. 
 
 Mo0 8 
 found. 
 
 MoO s referred 
 to molybdate 
 
 
 
 
 taken. 
 
 grm. 
 
 cm 8 
 
 grm. 
 
 
 0.2674 
 
 30.8 ) 1 cm 2 = 
 
 0.2184 
 
 [81.71] 
 
 0.4418 
 
 50.8 > 0.00709 
 
 0.3601 
 
 81.51 
 
 0.4075 
 
 [40.71*1 Mo0 8 . 
 
 0.3317 
 
 81.40 
 
 0.3281 
 0.4340 
 0.4098 
 0.4305 
 
 37.33 1 T o 
 
 ACk AQ I 1 Cm = 
 
 I?'!? [ 0.007086 
 46.63 I ii/r^r* 
 49.08 J Mo - 
 
 0.2644 
 0.3502 
 0.3304 
 0.3478 
 
 t 81.85-1 
 81.69 
 81.67 
 81.78J 
 
 Appended is a recalculation of the percentage of the trioxide 
 found, with columns showing the percentage error and the 
 error stated in fractions of a gram. Changes from the figures 
 of Friedheim and Euler are in heavy-faced type. 
 
 RECALCULATION OF THE RESULTS OF FRIEDHEIM AND EULEB. 
 
 Corrected 
 per cent of MoO, 
 found, referred to 
 the molybdate. 
 
 Error in 
 per cent of 
 Mo0 3 found 
 compared with 
 Mo0 3 taken. 
 
 Error 
 of MoO a . 
 
 
 
 grm. 
 
 81.68 
 
 0.23+ 
 
 0.0005+ 
 
 81.51 
 
 0.03+ 
 
 0.0001+ 
 
 81.40 
 
 0.12- 
 
 0.0004- 
 
 80.58 
 
 1.12- 
 
 0.0030- 
 
 80.69 
 
 0.99- 
 
 0.0035- 
 
 80.62 
 
 1.05- 
 
 0.0035- 
 
 80.79 
 
 0.86- 
 
 0.0030- 
 
 These figures of their own (properly calculated) are suffi- 
 cient to show the inadequacy of the method of Friedheim and 
 Euler. We ourselves were occasionally able to get results 
 from the method of Friedheim and Euler quite as good as 
 
 Probably 46.7. 
 
THE ESTIMATION OF MOLYBDENUM. 
 
 31 
 
 these; it must be said, however, that most of our results 
 obtained by their unmodified method have been even worse 
 than their own. 
 
 In another series of six determinations, in which molybde- 
 num trioxide was the starting-point, Friedheim and Euler 
 were more successful, the errors varying from 0.0006 grm. + 
 to 0.0006 grm.. Thus Friedheim and Euler establish by 
 their own results the fact that the hitting of the right point at 
 which to stop their process of boiling is a matter of chance. 
 In spite of the probability that some of the iodine which they 
 found in the receiver was liberated by atmospheric action, the 
 fact remains that their results are in many cases very low. 
 That is, they did not boil long enough. 
 
 The difficulty appears again in the modification of their 
 method which Friedheim and Euler apply to the determination 
 of molybdenum trioxide associated with vanadium pentoxide,* 
 namely, the distillation with phosphoric acid and potassium 
 iodide of the residue left after reducing the vanadium pentox- 
 ide by hydrochloric acid and potassium bromide, according to 
 the method of Holverscheit. We reproduce the part of their 
 table which refers to the determination of the molybdenum, 
 adding, however, columns containing the errors and corrected 
 percentages. 
 
 MoO a 
 taken. 
 
 MoO ? 
 
 Per cent 
 Mo0 3 
 
 Error. 
 
 Per cent 
 Mo0 3 . 
 
 
 
 P. and E. 
 
 
 Recalculated. 
 
 grin. 
 
 gnu. 
 
 
 grm. 
 
 
 0.15037 
 
 0.15005 
 
 99.79 
 
 0.00032- 
 
 99.79 
 
 0.16895 
 
 0.16879 
 
 99.90 
 
 0.00016- 
 
 99.90 
 
 0.17758 
 
 0.17729 
 
 99.84 
 
 0.00029 
 
 99.84 
 
 0.24975 
 
 0.24962 
 
 99.95 
 
 0.00013- 
 
 99.95 
 
 0.33151 
 
 0.33607 
 
 [99.87] 
 
 0.00456+ 
 
 101.38 
 
 Four of the five determinations are accurate, but the fact 
 that all figures are carried out to the fifth decimal place does 
 not keep three good-sized figures out of the error column for 
 the fifth determination. 
 
 * Ber. Dtsch. chem. Ges., xxviii. 2072. 
 
32 THE ESTIMATION OF MOLYBDENUM. 
 
 It is hardly necessary, in the light of a comparison of the 
 results of Friedheim and Euler with ours, to discuss further 
 the unreliability of the unmodified process. The necessity of 
 a proper control of the volume, strength of acid, and excess 
 of potassium iodide, as well as proper protection from atmos- 
 pheric oxidation, is real. 
 
VI 
 
 THE APPLICATION OF IODIC ACID TO THE 
 ANALYSIS OF IODIDES. 
 
 BY F. A. GOOCH AND C. F. WALKER.* 
 
 IT has long been understood that iodic acid is easily and 
 completely reduced by an excess of hydriodic acid with the 
 liberation of iodine according to the equation: 
 
 HI0 8 + 5HI = 61 + 3H 2 O. 
 
 To apply this reaction to the quantitative estimation of iodic 
 acid, it is only necessary to add to the free iodic acid or solu- 
 ble iodate an excess of a soluble iodide, to acidify best with 
 dilute sulphuric acid and to titrate the iodine thus set free 
 with sodium thiosulphate, one-sixth of the iodine found being 
 credited to the iodic acid. 
 
 It has been shown recently by Rieglerf that this reaction 
 may be also applied to the quantitative estimation of iodides, 
 the iodine set free upon the addition of a known excess of 
 iodic acid to the iodide solution being removed by petroleum 
 ether, and the residual iodic acid titrated directly with sodium 
 thiosulphate. 
 
 The present investigation was undertaken to define more 
 particularly the limit of applicability of the reaction and to 
 establish, if possible, a direct method for the quantitative esti- 
 mation of iodides, dependent upon the action of iodic acid or 
 an iodate in the presence of free sulphuric acid, neutralization 
 of the solution by means of an acid carbonate, and titration of 
 the free iodine by arsenious acid five-sixths of the iodine 
 thus found being credited to the iodide to be estimated. It 
 
 * From Am. Jour. Sci. iii, 293. t Zeitschr. anal. Chem., xxxv, 305. 
 
 VOL. II. 3 
 
34 THE APPLICATION OF 10DIC ACID 
 
 has been found that by fulfilling certain necessary conditions, 
 the proposed method is entirely successful, so far as concerns 
 the estimation of iodine in iodide solutions free from large 
 amounts of chlorides or bromides. 
 
 In a system containing a considerable quantity of free 
 iodine with variable amounts of the other reagents mentioned, 
 as well as possible impurities, it is conceivable that secondary 
 reactions may occur, depending largely on conditions of mass, 
 tune, and temperature, and of a sort likely to alter the amount 
 of recoverable iodine, or to exert an excessive oxidizing influ- 
 ence on the arsenious acid finally titrated. It has been estab- 
 lished by Schonbein, Lunge and Schloch, and others, that 
 iodine forms compounds with the alkalies of the type R-O-I, 
 and Phelps* has recently found that the formation of some 
 such compound, accompanying the iodate naturally expected, 
 is distinctly recognizable when iodine and barium hydroxide 
 interact at ordinary temperatures. It has been shown, also, 
 in a former paper from this laboratory! that free iodine or an 
 iodide interacts very easily with iodic acid in the presence of 
 dilute hydrochloric acid with the formation of iodine mono- 
 chloride, according to the equations: 
 
 HI0 3 + 21, + 5HC1 = 3H 2 + 6IC1. 
 
 HI0 3 + 2KI + 5HC1 = 3H 2 + 2KC1 + 3IC1. 
 
 Moreover, organic compounds containing the groups 1 = 
 and I ~ Q, in which iodine seems to be analogous to nitrogen, 
 
 result in great variety from the oxidation of halogen sub- 
 stitution products. It would seem, therefore, that the 
 formation of inorganic reduction products of iodic acid under 
 the conditions likely to obtain in this analytical process might 
 be by no means beyond the bounds of possibility. 
 
 A few simple qualitative tests to determine the possibility 
 of interaction between small quantities of iodine and iodic 
 acid alone met with negative results. Thus, a single drop 
 of a decinormal solution of iodine, made as usual in potassium 
 
 * Am. Jour. Sci., ii, 70. Volume I, p. 370. 
 
 t Roberts, Am. Jour. Sci., xlviii, 157. Volume I, p. 257. 
 
TO THE ANALYSIS OF IODIDES. 35 
 
 iodide, gave when added to 10 cm 3 of decinormal iodic acid a 
 distinctive color to chloroform. Similar results were obtained 
 when the iodine was employed in aqueous solution in which 
 there was no alkaline iodide. A few drops of an aqueous 
 solution of iodine treated (in either order) with 10 cm 3 of a 
 saturated solution of potassium bicarbonate and 10 cm 3 of 
 decinormal iodic acid gave the same distinctive color to 
 chloroform as came from the same amount of iodine in the 
 absence of the iodic acid. So it appears that if in the system 
 under consideration reactions do occur between iodic acid and 
 iodine to alter the amount of iodine recoverable, such action 
 is not appreciable between small amounts of these materials. 
 This, however, does not preclude the possibility of perceptible 
 changes under the mass-action of a large amount of iodine. 
 
 The reactions of hydrochloric acid, and probably of hydro- 
 bromic acid, hi the presence of varying amounts of iodic 
 acid, iodine, and iodide, as well as the reaction of the alkaline 
 carbonate upon such mixtures are doubtless complex, more or 
 less reversible, and dependent upon proportions and dilution. 
 The tendency of the former reactions is toward the reduction 
 of the molecule of iodic acid, and the formation of the chloride 
 or bromide of iodine. Thus, Miss Roberts * demonstrated 
 that a solution of hydrochloric acid, so dilute that by itself it 
 is without effect on iodic acid, acts upon a mixture of iodic 
 acid with either free iodine or an iodide to form iodine 
 monochloride. The action of the acid carbonate upon the 
 iodine chloride or bromide may produce a salt of the oxy-acids 
 and free iodine. 
 
 The practical effects, under the conditions of analysis, of the 
 reaction between iodine, iodic acid and the halogen acids in 
 presence of sulphuric acid, and of reactions which may occur 
 upon neutralization by an acid carbonate, were studied in detail 
 in a number of experiments. 
 
 The preliminary experiments of Table I were made to bring 
 out the effect of neutralizing with the acid carbonate and 
 subsequently titrating with an alkaline arsenite a solution 
 
 * Loc. cit. 
 
36 
 
 THE APPLICATION OF IODIC ACID 
 
 [5 cm3 
 
 TABLE I. 
 
 EFFECT OF THE CARBONATE. 
 (1:3). Total volume of liquid, 250 cm 8 .] 
 
 Ezp. 
 
 I (in KI) taken. 
 
 KHC0 3 in excess. 
 
 I found. 
 
 Error. 
 
 
 grin. 
 
 cm 8 
 
 grin. 
 
 gnu. 
 
 (1) 
 
 0.0713 
 
 Very small. 
 
 0.0707 
 
 0.0006- 
 
 (2) 
 
 0.0715 
 
 Very small. 
 
 0.0710 
 
 0.0005- 
 
 (3) 
 
 0.0713 
 
 10 
 
 0.0710 
 
 0.0003- 
 
 (4) 
 
 0.0710 
 
 10 
 
 0.0706 
 
 0.0004- 
 
 (5) 
 
 0.0723 
 
 10 
 
 0.0717 
 
 0.0006- 
 
 (6) 
 
 0.0713 
 
 20 
 
 0.0709 
 
 0.0004- 
 
 (7) 
 
 0.0713 
 
 20 
 
 0.0709 
 
 0.0004- 
 
 (8) 
 
 0.3565 
 
 Very small. 
 
 0.3560 
 
 0.0005- 
 
 (9) 
 
 0.3568 
 
 Very small. 
 
 0.3561 
 
 0.0007- 
 
 (10) 
 
 0.3667 
 
 10 
 
 0.3563 
 
 0.0004- 
 
 (11) 
 
 0.3596 
 
 10 
 
 0.3588 
 
 0.0008- 
 
 (12) 
 
 0.3565 
 
 10 
 
 0.3565 
 
 0.0000 
 
 (13) 
 
 0.3572 
 
 20 
 
 0.3560 
 
 0.0012- 
 
 (14) 
 
 0.3567 
 
 20 
 
 0.3569 
 
 0.0002-f- 
 
 containing sulphuric acid and a considerable amount of free 
 iodine. The danger of mechanical loss of iodine during the 
 effervescence accompanying neutralization, as well as by 
 spontaneous volatilization from the surface during the process 
 of titration, was minimized by effecting the neutralization in 
 the trapped Drexel washing-bottle to be described later, and 
 making the titration in the same tall washing cylinder without 
 transfer. To varying amounts of a recently standardized 
 decinormal solution of iodine were added successively 5 cm 3 
 of dilute sulphuric acid and varying amounts of potassium 
 bicarbonate in excess of that necessary to neutralize the free 
 acid, decinormal arsenious acid in slight excess of the iodine, 
 5 cm 3 of starch emulsion, and decinormal iodine to coloration, 
 the total volume of the liquid being not greater than 250 cm 8 . 
 The results show plainly that while the loss, mechanical or 
 otherwise, in the treatment of reasonably large amounts of 
 fairly concentrated iodine is perceptible, it is still well within 
 permissible limits (amounting to a little less than 0.0005 grm. 
 in the mean), and obviously independent of the excess of the 
 carbonate in the solution, and of the amount of free iodine 
 present. 
 
TO THE ANALYSIS OF IODIDES. 
 
 37 
 
 TABLE IL 
 EFFECT OF DILUTION. 
 
 Bxp. 
 
 El taken. 
 
 KI found. 
 
 Error. 
 
 Approximate 
 volume upon 
 addition of H 2 S0 4 . 
 
 Volume 
 H 2 S0 4 (1:3) 
 used. 
 
 (1) 
 
 (2) 
 (3) 
 (4) 
 (5) 
 (6) 
 (7) 
 (8) 
 
 grin. 
 
 0.0772 
 0.0772 
 0.1544 
 0.1544 
 
 0.3087 
 0.3087 
 0.3859 
 0.3859 
 
 grlu. 
 
 0.0768 
 0.0765 
 0.1546 
 0.1541 
 0.3090 
 0.3088 
 0.3864 
 0.3860 
 
 grm. 
 0.0004- 
 0.0007- 
 0.0002+ 
 0.0003- 
 0.0003+ 
 0.0001+ 
 0.0005+ 
 0.0001+ 
 
 1SSSSSSSS 
 
 cm** 
 
 6 
 5 
 5 
 5 
 5 
 6 
 5 
 5 
 
 (9) 
 (10) 
 
 (11) 
 (12) 
 
 0.0772 
 0.0772 
 0.1543 
 0.1544 
 
 0.0754 
 0.0757 
 0.1532 
 0.1524 
 
 0.0018- 
 0.0015- 
 0.0011- 
 0.0020- 
 
 300 
 300 
 300 
 300 
 
 5 
 5 
 5 
 5 
 
 (13) 
 (14) 
 (16) 
 (16) 
 (17) 
 (18) 
 
 0.0772 
 0.0772 
 0.1544 
 0.1544 
 0.3859 
 0.3859 
 
 0.0744 
 0.0737 
 0.1521 
 0.1512 
 
 0.3827 
 0.3831 
 
 0.0028- 
 0.0035- 
 0.0023- 
 0.0032- 
 0.0032- 
 0.0028- 
 
 600 
 500 
 500 
 500 
 500 
 500 
 
 5 
 6 
 5 
 5 
 5 
 5 
 
 (19) 
 (20) 
 (21) 
 (22) 
 
 0.0772 
 0.0772 
 0.3859 
 0.3859 
 
 0.0744 
 0.0757 
 
 0.3828 
 0.3827 
 
 0.0028- 
 0.0015- 
 0.0031- 
 0.0032- 
 
 500 
 500 
 500 
 500 
 
 10 
 10 
 10 
 
 10 
 
 In the experiments of Table II the proposed process of 
 analysis was tested upon potassium iodide taken by itself in 
 varying amounts of a J^ normal solution and 
 carefully standardized by the method formerly 
 elaborated in this laboratory.* The apparatus 
 employed was a Drexel washing-bottle of 500 
 cm 3 or 1000 cm 3 capacity, according to require- 
 ments, with stop-cock and thistle-tube fused to 
 the inlet tube and a Will and Varrentrapp 
 absorption trap sealed to the outlet, as shown 
 in the accompanying figure. The iodide for the 
 test was drawn from a burette into the bottle 
 and carefully washed down, and potassium 
 iodate in excess of the amount theoretically necessary (namely, 
 
 FIG. 18. 
 
 * Gooch and Browning, Am. Jour. Sci., xxxix, 188. Volume I, p. 1. 
 
38 THE APPLICATION OF IODIC ACID 
 
 5 cm 3 of a 0.5 per cent solution for every portion of 20 cm 3 
 of the iodide solution), was added and the volume of the 
 liquid was adjusted to the volume at which it was desired 
 that the iodic and hydriodic acids should react. The stop- 
 per with the thistle-tube and trap was now placed on the 
 bottle and the trap was half filled by means of a pipette 
 with a 5 per cent solution of potassium iodide. Five cen- 
 timeters of dilute (1 : 3) sulphuric acid were added through 
 the thistle-tube and washed down ; the stop-cock was closed, 
 .and the solution gently agitated, if necessary, to insure a 
 ; complete separation of iodine. Potassium bicarbonate in 
 saturated solution to an amount about 10 cm 8 in excess of 
 that required to neutralize 5 cm 8 of dilute (1 : 3) sulphuric 
 acid, was poured into the thistle-tube, and allowed to flow 
 into the bottle slowly enough to avoid a too violent evolution 
 of gas. The stop-cock was closed and the solution agitated 
 by giving to the bottle a rotary motion, at the same time 
 keeping the bottom pressed down upon the work-table, to 
 prevent a possible splashing of the iodide out of the trap into 
 the yet acid solution. When the neutralization of the solution 
 had been completed, the bottle was shaken until the last trace 
 of violet vapor was absorbed in the liquid. The greater part 
 of the solution in the trap was then run back into the bottle, 
 the stopper removed, and the tube and trap carefully washed, 
 the washings being added to the bulk of the solution. Deci- 
 normal arsenious acid was introduced from a burette to the 
 bleaching point, 5 cm 3 of starch emulsion were added, and the 
 solution was titrated back with decinormal iodine (usually 
 only a few drops) to coloration. 
 
 Blank tests made upon a solution obtained by mixing the 
 maximum amount of the iodate with 5 cm 3 of dilute sulphuric 
 acid (1 : 3), neutralizing as usual with potassium bicarbonate, 
 adding the iodide from the trap and 5 cm 3 of starch emulsion, 
 showed that a single drop of iodine was invariably sufficient 
 to bring out the starch blue. Occasionally it was found that 
 the mixture, particularly when chlorides or bromides were 
 present, of itself developed a trace of color, but by no means 
 
TO THE ANALYSIS OF IODIDES. 
 
 39 
 
 a reading tint. A correction of the one drop of iodine 
 necessary to bring out the color reaction in the blanks, was 
 applied uniformly in the analytical process. 
 
 The number of centimeters of decinormal arsenious acid 
 required to bleach the free iodine, multiplied by 0.01383 
 (log. 2.140822) gives the number of grams of potassium 
 iodide taken for analysis, being equivalent to five-sixths of 
 the iodine liberated in the solution. 
 
 From these results it appears that the degree of dilution 
 of the solution at the time when the mixed iodide and iodate 
 are acidified has an important influence on the completeness 
 of the reaction. Thus, the mean error of the determinations 
 in which the volume at the time of the reaction did not 
 exceed 150 cm 3 was practically nothing, while the errors at 
 volumes of 300 cm 3 and 500 cm 3 amounted to 0.0016 grm. and 
 0.0028 grm. respectively. It is obvious that the doubling 
 of the amount of sulphuric acid used in acidifying does not 
 increase the amount of iodine liberated at the highest dilution. 
 The plain inference is that the interaction between the iodide 
 and iodate should be brought about in a volume of liquid 
 not much exceeding 150 cm 3 . 
 
 In the following series of experiments, recorded in Table 
 III, the effect of the introduction of a chloride or bromide 
 
 TABLE III. 
 EFFECT OF CHLORIDE AND BROMIDE. 
 
 Exp 
 
 
 KI taken. 
 
 KI found. 
 
 Error. 
 
 NaCl taken. 
 
 KBr taken. 
 
 
 
 grin. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 
 0.0772 
 
 0.0795 
 
 0.0023+ 
 
 0.2 
 
 , 
 
 (2 
 
 
 0.0772 
 
 0.0784 
 
 0.0012+ 
 
 0.2 
 
 
 (3 
 (4 
 
 
 0.0771 
 0.0773 
 
 0.0823 
 0.0819 
 
 0.0052+ 
 0.0046+ 
 
 0.5 
 0.6 
 
 
 (6 
 
 
 0.1544 
 
 0.1588 
 
 0.0044+ 
 
 0.5 
 
 ( 
 
 6 
 
 
 0.1544 
 
 0.1590 
 
 0.0046+ 
 
 0.5 
 
 
 o 
 
 
 0.0772 
 
 0.0802 
 
 0.0030+ 
 
 , , 
 
 6.2 
 
 (8 
 
 
 0.0773 
 
 0.0853 
 
 0.0080+ 
 
 . , 
 
 0.2 
 
 (9) 
 
 
 0.0772 
 
 0.0873 
 
 0.0101+ 
 
 t 
 
 0.5 
 
 (10) 
 
 
 0.0772 
 
 0.0861 
 
 0.0089+ 
 
 
 0.5 
 
 11) 
 
 
 0.1544 
 
 0.1646 
 
 0.0102+ 
 
 t 
 
 0.5 
 
 (12) 
 
 
 0.1543 
 
 0.1626 
 
 0.0083+ 
 
 
 
 0.6 
 
40 
 
 THE APPLICATION OF IODIC ACID 
 
 into the iodide (before the iodate is added) was studied. The 
 volume of the liquid at the time of acidifying was fixed at 
 150 cm 3 approximately, and 5 cm 3 of the dilute sulphuric acid 
 (1 : 3) were used. The mode of procedure was otherwise 
 similar to that of the foregoing series. 
 
 The influence of sodium chloride and potassium bromide 
 in increasing the amount of iodine liberated is plain. The 
 increase comes without doubt from the iodate, and is doubt- 
 less due to the formation of iodine chloride or bromide, during 
 the acidifying, by the interaction of the free iodine, the iodic 
 acid, and the hydrochloric or hydrobromic acid, according to 
 the reactions previously discussed. It is plain, therefore, 
 that the value of the process in the determination of iodine 
 in an iodide is restricted of necessity to those cases in which 
 it is known that chlorides or bromides are not present to 
 
 TABLE IV. 
 ANALYSIS OF PURE POTASSIUM IODIDE. 
 
 Exp. 
 
 KI taken. 
 
 KI found. 
 
 Error. 
 
 
 grm. 
 
 gnn. 
 
 grm. 
 
 (1) 
 
 0.0814 
 
 0.0816 
 
 0.0002+ 
 
 (2) 
 
 0.0814 
 
 0.0813 
 
 0.0001- 
 
 (3) 
 
 0.0814 
 
 0.0805 
 
 0.0009+ 
 
 (4) 
 
 0.0815 
 
 0.0809 
 
 0.0006- 
 
 (5) 
 
 0.0814 
 
 0.0808 
 
 0.0006- 
 
 (6) 
 
 0.0814 
 
 0.0806 
 
 0.0008- 
 
 (7) 
 
 0.0814 
 
 0.0812 
 
 0.0002- 
 
 (8) 
 
 0.1628 
 
 0.1624 
 
 0.0004- 
 
 (9) 
 
 0.1628 
 
 0.1617 
 
 0.0011- 
 
 (10) 
 
 0.1628 
 
 0.1621 
 
 0.0007- 
 
 (11) 
 
 0.1628 
 
 0.1619 
 
 0.0009- 
 
 (12) 
 
 0.1628 
 
 0.1624 
 
 0.0004- 
 
 (13) 
 
 0.1628 
 
 0.1621 
 
 0.0007- 
 
 (14) 
 
 0.1628 
 
 0.1626 
 
 0.0002- 
 
 (15) 
 
 0.2442 
 
 0.2451 
 
 0.0009+ 
 
 (16) 
 
 0.2442 
 
 0.2442 
 
 0.0000 
 
 (17) 
 
 0.2442 
 
 0.2439 
 
 0.0003- 
 
 (18) 
 
 0.3256 
 
 0.3258 
 
 0.0002+ 
 
 (19) 
 
 0.3256 
 
 0.3256 
 
 0.0000 
 
 (20) 
 
 0.3256 
 
 0.3258 
 
 0.0002+ 
 
 (21) 
 
 0.3256 
 
 0.3272 
 
 0.0016+ 
 
 (22) 
 
 0.3256 
 
 0.3256 
 
 0.0000 
 
 (23) 
 
 0.4071 
 
 0.4076 
 
 0.0005+ 
 
 (24) 
 
 0.4071 
 
 0.4080 
 
 0.0009+ 
 
 (25) 
 
 0.4071 
 
 0.4073 
 
 0.0002+ 
 
TO THE ANALYSIS OF IODIDES. 41 
 
 any considerable extent. For determining the standard of a 
 solution of nearly pure potassium iodide, employed in so many 
 laboratory processes, it should find useful application. 
 
 In Table IV are comprised a number of experiments made 
 exactly like those which seemed to give the best results in 
 the series of Table II. The iodide and an excess of iodate 
 (5 cm 3 of the 0.5 per cent solution to every 20 cm 3 of ^ iodide) 
 were made to interact in a volume of about 150 cm 3 , 5 cm 3 of 
 sulphuric acid (1 : 3) were used to bring about the reaction, 
 10 cm 3 of potassium bicarbonate were added after the neutrali- 
 zation of the sulphuric acid was complete, and the free 
 iodine was estimated by titrating decinormal arsenious acid, 
 the manipulation being like that previously described in 
 detail. 
 
 The average result of a series of several determinations in 
 which a great excess (0.1 grm.) of potassium iodate was 
 used, proved to be practically identical with that of a similar 
 series in which only a small excess of the iodate was employed, 
 so that it appears to be unnecessary in any practical work to 
 restrict the amount of iodate below the amount necessary to 
 decompose the maximum quantity of potassium iodide which 
 we have handled, namely, 0.4 grm. 
 
 It appears that for the estimation of iodine in a soluble 
 iodide free from notable amounts of chlorides or bromides, 
 this method, depending as it does upon a single standard 
 solution, is simple, fairly accurate, and rapid. 
 
vn 
 
 THE ACTION OF UREA AND PRIMARY AMINES 
 ON MALEIC ANHYDRIDE.* 
 
 BY FREDERICK L. DUNLAP AND ISAAC K. PHELPS. 
 
 IN a former article, f a method was described by one of us 
 for obtaining imides by the action of urea on the anhydrides 
 of dibasic acids. It was shown that the formation of imides 
 by the interaction of urea and anhydrides was to be explained 
 by the addition of the urea and the anhydrides to form an 
 acid, which when heated, decomposed with the formation of 
 the imide, carbon dioxide and ammonia. These reactions can 
 be shown by the following equations : 
 
 COOH 
 
 . ^O.^T 
 
 <COOH : < CO > 
 
 In some cases this intermediate product, formed by the 
 addition of the reacting substances, was readily isolated, and 
 upon heating to certain temperatures, it was found that this 
 addition-product decomposed with the formation of the imide. 
 
 It was found, in particular, that maleic anhydride formed 
 an addition-product with urea, and we hoped to obtain the 
 unknown imide of maleic acid by the decomposition of this 
 addition-product. 
 
 Male-uric Acid. When equal molecules of maleic anhydride 
 and urea are heated to 100 -105 C. in an oil-bath, the mixture 
 melts, and if this temperature is maintained for a short time, 
 
 * From Am. Chem. Jour., xix, 492. 
 
 t Am. Chem. Jour., xviii, 332. Volume I, 355. 
 
ACTION OF UREA AND PRIMARY AMINES. 43 
 
 the liquid solidifies. This product was purified by crystalli- 
 zation from alcohol, and when pure was perfectly white and 
 melted at 167.5-168 with decomposition. Upon analysis 
 the following results were obtained: 
 
 0.2148 gram substance gave 0.2968 gram C0 2 and 0.0751 
 gram H 2 0. 
 
 Calculated for 
 
 C 37.97 37.69 
 
 H 3.80 3.88 
 
 Maleiiric acid is soluble in hot water and alcohol, fairly 
 insoluble in cold and insoluble in chloroform, benzene, ligrom, 
 carbon disulphide, acetone, and ether. 
 
 Its formation and structure are shown in the following 
 equation : 
 
 H-C-CO V H-C-CONHCONH a 
 
 II ^0 + NH 3 CONH 2 = || 
 
 H-C-CCT H-C-COOH 
 
 No difficulty was expected in the formation of the imide of 
 maleic acid, either from the maleiiric acid, or by the direct 
 heating of maleic anhydride and urea; for, in the cases 
 already studied,* the reaction ran with great smoothness and 
 ease. But, for some unknown reason, the imide of maleic 
 acid could not be obtained at least in quantities for complete 
 identification. A great many experiments were carried out 
 under varying conditions, but only under one set of conditions 
 could any product be isolated, and then unfortunately the 
 yield was so distressingly small that the study of the compound 
 had to be abandoned. When equal molecules of maleic 
 anhydride and urea are heated in a boiling-flask, at first slowly, 
 and finally at the full heat of a Bunsen burner, a vigorous 
 evolution of carbon dioxide and ammonia took place, and a 
 small amount of a dark-colored distillate was obtained, which 
 on cooling solidified and gave a melting-point of 130.5 after 
 crystallization from anhydrous acetone. This same crystalline 
 
 * Loc. cit. 
 
44 THE ACTION OF UREA AND PRIMARY 
 
 product was also obtained on distilling the maleiiric acid. In 
 both cases a large amount of carbonaceous residue remained 
 in the boiling-flask. 
 
 Although the method of formation of this compound is 
 exactly parallel to that employed in the preparation of 
 succinimide,* yet the identity of this product cannot be 
 regarded as established until sufficient quantities of it have 
 been prepared to subject it to elementary analysis. 
 
 The Action of Primary Amines on Maleic Anhydride. 
 
 Anschiitzf in 1889 published a method by which anilic acids 
 could be very easily prepared by dissolving molecular quantities 
 of aniline and the anhydride of dibasic acids in dry chloroform, 
 when after standing for a short tune the anilic acid crystallized 
 out. Following out this line and method of formation, 
 suggested by Anschiitz's work, some derivatives of maleic 
 acid have been prepared and studied. 
 
 p-Tolylmaleamic Acid. Molecular proportions of maleic 
 anhydride and ^>-toluidine were dissolved in dry chloroform, 
 and upon mixing these two solutions, there immediately 
 separated out a light yellow precipitate. After standing for a 
 time to insure the complete precipitation of the product, it 
 was filtered off and purified by crystallization from alcohol, 
 from which it separated in lemon-yellow needles. The pure 
 product melted at 201, with evolution of gas. It is readily 
 soluble in ether, acetone, and hot alcohol, but insoluble in 
 benzene, ligrom, carbon disulphide, chloroform and water. 
 Upon analysis the f ollowing results were obtained : 
 
 0.2072 gram substance gave 0.4881 gram C0 2 , and 0.1068 
 gram H 2 0. 
 
 Calculated for v ^ 
 
 C U H U S N. Found ' 
 
 C 64.39 64.24 
 
 H 5.37 5.73 
 
 * Loc. cit. t Ber. Dtsch. chem. Ges., xx, 3214 
 
AMINES ON MALEIC ANHYDRIDE. 45 
 
 The structure and mode of formation of this ^>-tolylmaleamic 
 acid may be seen from the following equation : 
 
 H-C-CO V X CH 8 H~C-CONHC 6 H 4 CH 8 (p) 
 
 C 6 H 4 
 
 N 
 
 H-C-C(r NH 2 (p)H-C-COOH 
 
 o-Totylmaledmic Acid. This acid was prepared similarly to 
 the ^-tolylmaleamic acid, using molecular proportions of 
 maleic anhydride and 0-toluidine. The product was purified 
 by crystallization from alcohol, from which it separated in 
 bunches of thick light yellow prisms. When pure it melted 
 at 117.5-118. It is readily soluble hi acetone, and in 
 alcohol; sparingly soluble in chloroform, and insoluble in 
 ligroin, carbon disulphide, benzene, water, and ether. Analy- 
 sis gave the following results : 
 
 0.2124 gram substance gave 0.4993 gram C0 2 , and 0.1146 
 gram H 2 0. 
 
 C 64.39 64.12 
 
 H 5.37 5.99 
 
 It has the structure represented by the formula 
 
 H - C - CONHC 6 H 4 CH 8 (o) 
 
 II 
 H-C-COOH 
 
 p-Naphthylmaleamic Acid. When molecular quantities of 
 ^-naphthylamine and maleic anhydride were mixed in dry 
 chloroform solution they reacted and united as readily as did 
 the o- and ^>-toluidines and the anhydride. Almost immediately 
 a yellow crystalline precipitate separated out. This, after 
 standing a short time, was filtered off and crystallized from 
 alcohol. It separated from alcohol in small bright yellow 
 needles, which, when pure, melted at 200 with evolution of 
 gas. It is soluble in acetone and alcohol, but insoluble in 
 ether, carbon disulphide, chloroform, benzene, ligroin, and 
 water. Analysis gave the following results : 
 
46 ACTION OF UREA AND PRIMARY AMINES. 
 
 0.2033 gram substance gave 0.5226 gram C0 2 and 0.0948 
 gram H 2 0. 
 
 Calculated for v , 
 
 C 14 H U 3 N. FoTmd ' 
 
 C 69.71 70.11 
 
 H 4.56 5.18 
 
 This addition-product is obviously formed as follows : 
 H _ C - CO V H-C- CONHC 10 H 7 
 
 + C 10 H 7 NH 2 = 
 
 H-C-CCT H-C-COOH 
 
 a-Naphthylamine appears also to be added readily to 
 maleic anhydride, but the product has not been submitted 
 to analysis. 
 
VIII 
 
 THE SEPARATION OF ALUMINUM AND BERYL- 
 LIUM BY THE ACTION OF HYDROCHLORIC 
 ACID. 
 
 BY FRANKE S. HAVENS* 
 
 IN a former paper f a method was described for the determi- 
 nation of aluminum in the presence of iron, based upon 
 the fact that the hydrous aluminum chloride A1C1 3 . 6H 2 O 
 is practically insoluble in a mixture of strong hydrochloric 
 acid and anhydrous ether saturated with hydrochloric acid 
 gas, while the ferric chloride is entirely soluble in that 
 medium. 
 
 The work to be described in this paper is an extension 
 of this process to cover the separation of aluminum from 
 beryllium, with the subsequent determination of the beryllium 
 by weighing as the oxide after conversion to the nitrate and 
 ignition. 
 
 The aluminum chloride solution was prepared by dissolving 
 the so-called pure aluminum chloride of commerce in as little 
 water as possible, precipitating and washing free from iron 
 with strong hydrochloric acid, dissolving the chloride thus 
 obtained in water, precipitating the hydroxide by ammonia, 
 washing the precipitate free from all alkalies, and redissolving 
 it in hot hydrochloric acid. From this solution, after cool- 
 ing, gaseous hydrochloric acid precipitated the pure hydrous 
 chloride. This prepared chloride was dissolved in water 
 and the solution standardized by precipitating with ammonia 
 the hydroxide from weighed portions and weighing as the 
 oxide. The solution of beryllium used was made by dis- 
 
 * From Am. Jour. Sci., iv, 111. 
 
 t Gooch and Havens, Am. Jour. Sci., ii, 416. This volume, p. 20. 
 
48 SEPARATION OF ALUMINUM AND BERYLLIUM 
 
 solving in water beryllium chloride found to be free from 
 iron by the sulphocyanate test, and giving no precipitate 
 when tested by the gaseous hydrochloric acid process to be 
 described later on. This was standardized by precipitating 
 with ammonia the hydroxide from weighed portions and 
 weighing the ignited oxide in the usual manner. 
 
 In the experiments of Table I, weighed portions of the 
 aluminum solution were mixed with portions of the beryllium 
 chloride solution representing from 0.01 gram to 0.10 gram 
 of the oxide, an equal volume of a mixture of strong hydro- 
 chloric acid and ether (taken in equal parts) was added to the 
 solution of the mixed chlorides, and the whole was completely 
 saturated with gaseous hydrochloric acid while kept at a 
 temperature of about 15 C. by immersing the receptacle 
 
 TABLE I. 
 
 Bxp. 
 
 A1 2 O 8 taken in 
 solution as the 
 chloride. 
 
 found. 
 
 Final 
 volume. 
 
 Error. 
 
 (1) 
 
 grm. 
 
 0.1046 
 
 gnn. 
 0.1044 
 
 cms 
 12 
 
 grm. 
 0.0002- 
 
 (2) 
 
 0.1046 
 
 0.1038 
 
 12 
 
 0.0008- 
 
 (3) 
 
 0.1067 
 
 0.1066 
 
 12 
 
 0.0001- 
 
 (4) 
 
 0.1071 
 
 0.1063 
 
 12 
 
 0.0008- 
 
 (5) 
 
 0.1059 
 
 0.1064 
 
 30 
 
 0.0005- 
 
 in running water. Ether was added, equal in volume to 
 the aqueous aluminum and beryllium solutions originally 
 taken, and the current of gas again turned on until satura- 
 tion was complete. By this treatment there is present at the 
 end of the saturation a volume of ether equal to that of 
 the aqueous hydrochloric acid introduced and generated. 
 The finely crystalline precipitate of aluminum chloride was 
 caught on asbestos in a filter crucible washed with a pre- 
 viously prepared mixture of hydrochloric acid and ether in 
 equal parts saturated at 15 C. with hydrochloric acid gas, 
 and dried for half an hour at a temperature of 150 C. It 
 was next covered with a layer of pure mercuric oxide, which 
 had been tested and found to leave no residue on volatilizing, 
 
BY THE ACTION OF HYDROCHLORIC ACID. 
 
 49 
 
 and the crucible was gently heated over a low flame under 
 a ventilating hood and finally ignited over the blast. 
 
 From these results it is obvious that the aluminum chloride 
 may be determined in the presence of beryllium chloride with 
 reasonable accuracy. 
 
 The beryllium may be recovered in the filtrate from the 
 aluminum chloride by precipitation with ammonia after nearly 
 complete evaporation of the acid. It was found, however, 
 upon trial that the conversion of the chloride to the oxide 
 without precipitation and filtration may be easily accomplished 
 by treatment with nitric acid and ignition. The results of 
 Table II indicate this clearly. In these experiments weighed 
 
 TABLE II. 
 
 
 BeO taken in 
 
 
 
 Exp. 
 
 solution as the 
 
 BeO found. 
 
 Error. 
 
 
 chloride. 
 
 
 
 
 grin. 
 
 gnu. 
 
 grm. 
 
 0) 
 
 0.0483 
 
 0.0481 
 
 0.0002- 
 
 2) 
 
 0.0483 
 
 0.0483 
 
 0.0000 
 
 (3) 
 
 0.1076 
 
 0.1085 
 
 0.0009+ 
 
 portions of the beryllium solution were evaporated just to 
 dryness on a radiator, care being taken not to heat to the 
 volatilizing point of the beryllium chloride, a few drops of 
 strong nitric acid were added, the liquid was evaporated, 
 and the residue heated at first gently, to break up the 
 nitrate safely and finally on the blast. It was found that 
 this conversion of the beryllium to the nitrate can be carried 
 on in platinum without attacking that metal appreciably, 
 providing care be taken to remove thoroughly all excess of 
 hydrochloric acid before the nitric acid is added to the dry 
 residue. 
 
 In Table III, (l)-(9), are given the results of experiments 
 in which both the aluminum and the beryllium were deter- 
 mined the former by precipitation as the hydrous chloride 
 and weighing as the oxide after igniting with mercuric oxide : 
 the latter by the conversion of the chloride, through the 
 
 VOL. II. 4 
 
50 SEPARATION OF ALUMINUM AND BERYLLIUM 
 
 nitrate, into the oxide. In experiment (10) (made to get a 
 comparison of the methods) the beryllium was recovered 
 by precipitating the hydroxide with ammonia from the par- 
 tially evaporated solution of the chloride after removing the 
 aluminum. 
 
 In experiments (1) to (5), inclusive, the aluminum was 
 determined exactly as previously described ; in (6) and (7) 
 the solutions (being originally larger) were concentrated 
 by evaporation previous to the addition of the ether and 
 hydrochloric acid mixture. In experiments (8), (9) and 
 (10), the treatment was varied advantageously by saturating 
 the aqueous solution directly with hydrochloric acid gas 
 before adding an equal volume of ether, and completing the 
 saturation. 
 
 TABLE III. 
 
 
 A1 2 3 taken 
 
 
 
 
 BeO taken 
 
 
 
 Exp. 
 
 in solution 
 as the 
 
 AM>, 
 
 Error. 
 
 Final 
 volume. 
 
 in solution 
 as the 
 
 BeO 
 
 found. 
 
 Error. 
 
 
 chloride. 
 
 
 
 
 chloride. 
 
 
 
 
 grin. 
 
 grm. 
 
 grm* 
 
 cm 3 
 
 gnu* 
 
 grin. 
 
 grm. 
 
 (1) 
 
 0.1059 
 
 0.1058 
 
 0.0001- 
 
 12 
 
 0.0198 
 
 0.0204 
 
 0.0006+ 
 
 (2) 
 
 0.1053 
 
 0.1044 
 
 0.0009- 
 
 12 
 
 0.0194 
 
 0.0196 
 
 0.0002+ 
 
 (3) 
 
 0.1065 
 
 0.1059 
 
 0.0006- 
 
 12 
 
 0.0197 
 
 0.0205 
 
 0.0008+ 
 
 (4 
 
 0.1068 
 
 0.1060 
 
 0.0008- 
 
 12 
 
 0.0199 
 
 0.0207 
 
 0.0008+ 
 
 (5) 
 
 0.1049 
 
 0.1047 
 
 0.0002- 
 
 12 
 
 0.0198 
 
 0.0208 
 
 0.0010+ 
 
 (6) 
 
 0.1060 
 
 0.1057 
 
 0.0003- 
 
 12 
 
 0.0977 
 
 0.0969 
 
 0.0008- 
 
 (7) 
 
 0.1064 
 
 0.1063 
 
 0.0001- 
 
 12 
 
 0.1085 
 
 0.1084 
 
 0.0001- 
 
 (8) 
 
 0.1046 
 
 0.1038 
 
 0.0008- 
 
 30 
 
 0.1083 
 
 0.1087 
 
 0.0004+ 
 
 (9) 
 
 0.1051 
 
 0.1048 
 
 0.0003- 
 
 30 
 
 0.1071 
 
 0.1078 
 
 0.0007+ 
 
 (10) 
 
 0.1076 
 
 0.1075 
 
 0.0001- 
 
 30 
 
 0.1086 
 
 0.1094 
 
 0.0008+ 
 
 These results are plainly very good. 
 
 The manipulation of the process is not difficult. The 
 gaseous hydrochloric acid is most conveniently produced by 
 the well known method of treating with strong sulphuric 
 acid in regulated current a mixture of strong aqueous 
 hydrochloric acid and common salt. A platinum dish hung 
 in an inverted bell-jar, provided with inlet and outlet tubes 
 through which the current of water for cooling is passed, 
 makes the best container for the solution to be saturated 
 with the gas. It is advantageous to arrange the filtration 
 
BY THE ACTION OF HYDROCHLORIC ACID. 51 
 
 upon asbestos so that the filtrate and washings may be 
 caught directly in the crucible (placed under the bell-jar of 
 the filter pump) in which the subsequent evaporation is to 
 be effected. The heating of the strong acid solution must 
 be gradual and conducted with care to prevent mechanical 
 loss by a too violent evolution of the gaseous acid. 
 
IX 
 
 THE TITRATION OF SODIUM THIOSULPHATE 
 BY IODIC ACID. 
 
 BY CLAUDE F. WALKER .* 
 
 THIS investigation was undertaken to determine the nature 
 and limitations of the reaction between iodic acid and thiosul- 
 phuric acid, and to show the expediency of employing iodic 
 acid in standard solution for the direct titration of sodium 
 thiosulphate. Rieglert states that iodic acid is readily ob- 
 tained in the pure state, that it may be accurately weighed 
 out, and that a solution of it may be exactly made up to a 
 desired strength and kept for a long tune unaltered. He 
 further states that when a solution of sodium thiosulphate is 
 titrated with iodic acid the reaction takes place according to 
 the equation, 
 
 6Na 2 S 2 3 + 6HI0 3 = 3Na 2 S<0 6 + 5NaI0 3 + Nal + 3H 2 0, 
 
 under which circumstances no free iodine will be evolved 
 until all the sodium thiosulphate has been oxidized to tetra- 
 thionate ; the first drop of iodic acid in excess, however, will 
 react with the sodium iodide that has been formed, and sepa- 
 rate iodine, as shown by the equation, 
 
 5NaI + 6HI0 3 = 5NaI0 8 + 3H 2 + 3I 2 , 
 
 thus furnishing an accurate means for determining the end 
 point. 
 
 A careful repetition of the work of Riegler has shown that 
 his conclusions are in a large measure erroneous. Thus, it 
 has been found that the ordinary " chemically pure " iodic 
 acid, purchased from reliable manufacturers, is likely to con- 
 
 * From the Amer. Jour. Sci., iv, 235. 
 
 t Riegler, Zeitschr. anal. Chem., xxxr, 308. 
 
TITRATION OF SODIUM THIOSULPHATE, ETC. 53 
 
 tain more than the theoretical amount of iodine, due probably 
 to the presence of the anhydride, although iodic acid can be 
 safely employed for standardizing when it is made in the 
 laboratory by dissolving the purified anhydride, crystallizing 
 out the acid, and drying over sulphuric acid. Such a care- 
 fully prepared product, if used immediately, will be found to 
 contain the theoretical amount of iodine. Riegler's proposed 
 method of titration depends on two different reactions, and to 
 insure the accuracy of the process these must be definite, com- 
 plete and non-reversible under the conditions of analysis. 
 Thus one molecule out of every six of iodic acid should be 
 reduced by six molecules of thiosulphate, with the formation 
 of a neutral mixture of iodide and iodate, free from other 
 oxidizing or reducing substances. Under these circumstances 
 it might be expected that iodine will be liberated by the first 
 trace of iodic acid hi excess. It has been found by investiga- 
 tion, however, that although the main reaction between iodic 
 acid and sodium thiosulphate may result in the formation of 
 sodium tetrathionate in the proportions given, there is never- 
 theless striking evidence of some other obscure action of the 
 thiosulphate, which influences the reduction of the iodic acid 
 in such a way as to make it impossible to calculate the analy- 
 ses according to Riegler's reaction. Moreover, a peculiar 
 " after-coloration " which invariably follows the first formation 
 of the starch blue during the titration of one solution against 
 the other, seems to point to the possibility that the reaction 
 between the iodide and iodic acid is dependent, under these 
 circumstances, on conditions of time and mass for its com- 
 pleteness. It is not impossible that some third compound of 
 iodine unstable in its nature, may be formed as an interme- 
 diate product and thus delay the liberation of iodine. In con- 
 sideration of the results that have been obtained it appears 
 that Riegler's proposed process for standardizing sodium thio- 
 sulphate, as well as his related method for the analysis of 
 iodides,* must remain impracticable unless they can be modi- 
 fied so as to do away with a number of sources of error. 
 
 * Riegler, Zeitschr. anal. Chem., xxxv, 305. 
 
54 TITRATION OF SODIUM THIOSULPHATE 
 
 The analyses of solutions of iodic acid, during the entire 
 course of the work, was invariably performed by adding to 
 the portion of the solution to be analyzed an excess of potas- 
 sium iodide, acidifying with 5 cm 3 of dilute (1 : 3) sulphuric 
 acid, and recovering the liberated iodine by directly titrating 
 the acid solution with sodium thiosulphate, or by neutralizing 
 with potassium bicarbonate in excess, and directly titrating 
 the alkaline solution with arsenious acid. In the latter case 
 the neutralization was performed in a trapped Drexel wash- 
 ing bottle such as has been described in connection with the 
 analysis of iodides.* In either case one-sixth of the iodine 
 recovered was calculated to iodic acid, according to the terms 
 of the equation, 
 
 SHI + HI0 3 = 31, + 3H 2 0. 
 
 It follows from these proportions that to bring the analyses 
 within the range of the decinormal solutions ordinarily 
 employed, the iodic acid taken for analysis must be restricted 
 to comparatively small amounts. In the present work it was 
 found convenient to analyze the iodic acid in quantities not 
 much exceeding one-tenth of a gram, in which case the varia- 
 tion in the results in the same series is found to be almost inap- 
 preciable. In both variations of the process one or two blank 
 analyses were invariably made, by performing the operation as 
 detailed, except that no iodic acid was employed, and the cor- 
 rection of one drop of iodine thereby shown to be necessary to 
 bring out the starch blue was uniformly applied in the ana- 
 lytical work. 
 
 To determine whether or not the purity of the ordinary 
 iodic acid is sufficient to admit of its direct application in 
 standard solutions, a series of experiments was made. Two 
 different samples of " chemically pure " iodic acid were used. 
 The first was in coarse granular crystals, and the second was 
 in the form of fine powder. Quantities of both of these 
 were dried in a desiccator over sulphuric acid to constant 
 weight. Neither sample lost weight appreciably when left 
 
 * Gooch and Walker, Am. Jour. Sci. iii, 293. This volume, p. 33. 
 
BY IODIC ACID. 
 
 55 
 
 for a considerable time on the scale pan. A third sample of 
 iodic acid was prepared by dissolving a quantity of the 
 purest obtainable iodic anhydride in water, and evaporating at 
 ordinary temperature. The resulting crystalline mass was 
 dried over sulphuric acid in a desiccator for one week, until it 
 ceased to lose weight, when it was presumed to consist of the 
 pure normal acid. Two presumably decinormal solutions of 
 each of the first two samples, and one such solution of the 
 third sample of iodic acid were made by weighing out 17.585 
 grms. and dissolving in exactly one liter of water at 15 C. 
 Convenient portions of each of these solutions were analyzed 
 in the manner described, with results shown in the following 
 table, averaged from many determinations. 
 
 TABLE I. 
 ANALYSES OF APPROXIMATELY ^ IODIC ACID. 
 
 Solution 
 analyzed. 
 
 Sample 
 used. 
 
 mO 3 taken. 
 
 HIO S found. 
 
 Error. 
 
 
 
 grin. 
 
 grm. 
 
 grm. 
 
 I 
 
 A 
 
 0.1055 
 
 0.1066 
 
 0.0011+ 
 
 II 
 
 A 
 
 0.1055 
 
 0.1062 
 
 0.0007+ 
 
 III 
 
 B 
 
 0.1055 
 
 0.1065 
 
 0.0010+ 
 
 IV 
 
 B 
 
 0.1055 
 
 0.1073 
 
 0.0018+ 
 
 V 
 
 C 
 
 0.1055 
 
 0.1053 
 
 0.0002- 
 
 These results show that while the deviation from the 
 theoretical strength of the solution in the case of the acid 
 prepared from the anhydride is hardly appreciable, and will 
 not affect the accuracy of any work in which the solution 
 may be applied as a means of standardization, the solutions 
 made from the purchased product, on the other hand, contain 
 a very appreciable amount of iodine in excess of the theoretical. 
 That iodic acid is somewhat unstable at 30-40 C., gradually 
 losing water with the formation of the anhydride,* is well 
 known, and it is quite possible that to some such gradual 
 change as this must be attributed the fact that the ordinary 
 iodic acid cannot be safely employed as a means of standard- 
 ization unless its purity be directly determined by analysis. 
 
 * Dammer, Anorg. Chem., i, 564. 
 
56 
 
 TITRATION OF SODIUM THIOSULPHATE 
 
 To determine whether a solution of iodic acid, once prepared 
 and standardized, will retain its strength for a long period of 
 time, two such solutions were kept for four months (in the 
 dark) and then again analyzed. The results (averages of 
 several determinations), given in Table II, substantiate the 
 observation of Riegler that a solution of iodic acid will remain 
 of constant strength. 
 
 TABLE II. 
 CONSTANCY OF STRENGTH OF IODIC ACID SOLUTIONS. 
 
 Iodic acid 
 solution. 
 
 First analysis. 
 HI0 3 found. 
 
 Second analysis 
 (after four months). 
 HIO 3 found. 
 
 Variation. 
 
 I 
 II 
 
 grm. 
 0.1073 
 0.1049 
 
 grm. 
 
 0.1072 
 0.1046 
 
 grm. 
 0.0001- 
 0.0003- 
 
 An approximately one-twentieth normal solution of " chemi- 
 cally pure" sodium thiosulphate was made and its exact 
 strength ascertained by titrating it with standardized iodine. 
 A series of analyses made by oxidizing the sodium thiosulphate 
 to sulphate, and precipitating and weighing as barium sulphate, 
 gave results identical with those obtained with iodine, proving 
 that all the sulphur present in the solution was in the form 
 of thiosulphate. According to Riegler's equation, sodium 
 thiosulphate and iodic acid react molecule for molecule, and 
 solutions of these substances should therefore require for 
 their mutual saturation volumes inversely proportional to 
 their concentration. It was found, however, that when the 
 one-twentieth normal solution of sodium thiosulphate that has 
 been described was titrated in the presence of starch emulsion 
 with an approximately decinormal solution of iodic acid, 
 prepared from the anhydride, a distinctly blue color was 
 produced long before the theoretical amount of iodic acid had 
 been added. It was further noticed that the end-point of the 
 reaction was far from distinct, a faint tinge of blue at first 
 being visible, then suddenly becoming deeper, and immediately 
 reappearing when bleached with sodium thiosulphate. The 
 
BY IODIC ACID. 57 
 
 deficiency in the amount of iodic acid actually required to 
 produce the blue color was not lessened by introducing only 
 three-fourths of the theoretical amount of iodic acid, and 
 estimating the residual thiosulphate with iodine. It was 
 found, however, that the addition of a considerable quantity 
 of potassium iodide to the solution, either before or during 
 the titration, had the marked effect of making the reaction 
 sharp and distinct, entirely preventing the " after separation " 
 of iodine, at the same time postponing the appearance of the 
 starch blue until a quantity of iodic acid had been added 
 considerably in excess of the theoretical. These experiments 
 were repeated with entirely different reagents, and under 
 varied conditions of concentration, the results in every case 
 exactly confirming those already observed. 
 
 For the purpose of more particularly investigating this 
 subject, there were prepared and standardized an approxi- 
 mately decinormal solution of sodium thiosulphate, and an 
 approximately one-fiftieth normal solution of iodic acid. 
 Measured portions of the sodium thiosulphate solution were 
 titrated with the iodic acid in the presence of starch emulsion 
 under varying conditions of mass, time and dilution. 
 
 To determine the variability of the end-point of the reaction 
 when the titration was conducted as directed by Riegler, a 
 series of experiments was made. Measured amounts of the 
 sodium thiosulphate 'solution were drawn from a burette into 
 an Erlenmeyer beaker of suitable capacity, the sides of the 
 beaker were carefully washed down with a small amount of 
 water, 5 cm 3 of starch emulsion were added, and the iodic acid 
 was slowly dropped into the small bulk of acid and starch, 
 with constant agitation of the mixture, until the first tint of 
 blue coloration appeared. The results obtained are given in 
 Table III. 
 
 These experiments indicate that the constancy of the end 
 reaction in different titrations of equal volumes of the same 
 solution depends to a certain degree on the volume of sodium 
 thiosulphate taken. The results in the case of the maximum 
 amounts vary within a range of 1.04 cm 3 , which corresponds 
 
58 
 
 TITRATION OF SODIUM THIOSULPHATE 
 
 TABLE III. 
 
 VARIATION OF THE END REACTION BETWEEN ~ SODIUM THIOSULPHATE 
 AND lODIC ACID, IN THE ABSENCE OP POTASSIUM IODIDE. 
 
 Eip. 
 
 NaAO, 
 taken. 
 
 HIO, 
 introduced. 
 
 Mean 
 value. 
 
 Variation. 
 
 
 1) 
 
 cm 8 
 6 
 
 cm 8 
 28.131 
 
 
 cm 8 
 
 
 cm 8 
 0.19- 
 
 
 2) 
 
 6 
 
 27.79 
 
 
 
 
 0.53- 
 
 
 3) 
 
 6 
 
 28.03 
 
 
 
 
 0.29- 
 
 
 1 
 
 6 
 6 
 
 28.32 
 28.32 
 
 
 28.32 
 
 
 0.00 
 0.00 
 
 
 6) 
 
 6 
 
 28.71 
 
 
 
 
 0.39+ 
 
 
 7) 
 
 6 
 
 28.83 
 
 
 
 
 0.51+ 
 
 8) 
 
 6 
 
 28.43 
 
 
 
 
 0.11+ 
 
 (9) 
 
 4 
 
 18.94 1 
 
 
 
 0.26+ 
 
 (10) 
 (11) 
 
 4 
 
 4 
 
 18.67 1 
 18.60 f 
 
 18.68 
 
 
 0.01- 
 0.18 
 
 (12) 
 
 4 
 
 18.60 J 
 
 
 
 0.08- 
 
 to 0.0035 grm. of iodic acid, while the average variation is 
 0.25 cm 3 , corresponding to 0.0009 grm. The variation in the 
 analyses of the smaller amounts is less, the range being 0.44 
 cm 3 , corresponding to 0.0015 grm., and the average variation 
 being 0.13 cm 3 , or 0.0005 grm. The probable error which 
 these irregularities would introduce in any series of practical 
 analyses by this method is obviously greater than can ordinarily 
 be permitted in iodometric work. 
 
 TABLE IV. 
 VARIATION OP THE END REACTION BETWEEN ^ SODIUM THIOSULPHATE 
 
 AND lODIC ACID, IN THE PRESENCE OF POTASSIUM IODIDE. 
 
 Eip. 
 
 Na 2 S 2 3 
 taken. 
 
 HIO S 
 introduced. 
 
 Mean 
 value. 
 
 Variation. 
 
 (1) 
 
 cm 8 
 6 
 
 cm 8 
 32.53 ] 
 
 
 cm 8 
 
 
 cm 8 
 \ 0.05+ 
 
 (2) 
 
 6 
 
 32.45 
 
 
 
 
 0.03- 
 
 (3) 
 (4) 
 
 6 
 6 
 
 32.67 
 32.37 
 
 
 32.48 
 
 
 
 0.19+ 
 0.11- 
 
 (5) 
 
 6 
 
 32.36 
 
 
 
 
 0.12- 
 
 (6) 
 
 6 
 
 32.50 
 
 
 
 
 0.02+ 
 
 (7) 
 
 4 
 
 22.30 ' 
 
 
 
 
 0.11+ 
 
 (8) 
 (9) 
 
 4 
 4 
 
 21.98 
 22.17 
 
 
 22.19 
 
 i 
 
 0.21- 
 0.02- 
 
 (10) 
 
 4 
 
 22.30 J 
 
 
 
 
 0.11+ 
 
BY IODIC ACID. 59 
 
 The experiments detailed in Table IV were performed 
 exactly similarly to those of the last series except that two 
 grams of potassium iodide were added to the sodium thiosul- 
 phate before the titration was commenced. 
 
 These experiments indicate plainly that in the presence of 
 potassium iodide the end reaction of different titrations of 
 equal volumes of the same solution is practically independent 
 of the amount taken for analysis. The results in the case of 
 the maximum amounts vary within a range of 0.31 cm 3 , or 
 0.0011 grm. of iodic acid, while the average variation is 0.09 
 cm 3 , corresponding to 0.0003 grm. The variation in the 
 analyses of the smaller amounts is practically the same as that 
 of the larger, the range being 0.32 cm 3 , corresponding to 0.0011 
 grm., and the average variation being 0.11 cm 3 , or 0.0004 grm. 
 It is therefore evident that the presence of potassium iodide in 
 the sodium thiosulphate to be titrated will bring the variation 
 of the formation of the reading tint within permissible limits. 
 
 A series of experiments was made to determine the nature 
 and effect of the " after coloration" observed to take place 
 when a solution of sodium thiosulphate, free from potassium 
 iodide, was titrated with iodic acid to blue coloration, and then 
 bleached with sodium thiosulphate. The titrations were 
 performed in the usual manner except that the volume was 
 adjusted just before the addition of the iodic acid, and the 
 iodine that was set free after the formation of the first reading 
 tint was destroyed at fixed intervals with measured amounts of 
 sodium thiosulphate. The results are given in Table V. 
 
 In the experiments with small volumes the evolution of 
 iodine in any considerable quantity ceased after two or three 
 hours, although the solution would become recolored as often 
 as it was bleached for a number of days. The traces of iodine 
 thus set free, however, were seldom equivalent to more than 
 one or two drops of sodium thiosulphate. The larger volumes, 
 however, continued to separate iodine in abundance for a very- 
 long time. The amount of iodine thus liberated after the first 
 coloration evidently varies with the amount of iodic acid 
 required for the titration, although not strictly proportional to 
 
60 
 
 TITRATION OF SODIUM THIOSULPHATE 
 
 TABLE V. 
 EFFECT OF DILUTION AND LAPSE OF TIME ON THE "AFTER COLORATION." 
 
 
 
 ^03 
 
 HIO 8 
 
 intwi 
 
 Na,S 2 O s introduced. 
 
 
 taken. 
 
 intro- 
 duced. 
 
 15m. 
 
 45m. 
 
 lh.45m. 
 
 2h.45in. 
 
 20 h. 
 
 Total. 
 
 Volume. 
 
 
 
 cm 3 
 
 cm 8 
 
 cm 3 
 
 cm 3 
 
 cm 3 
 
 cm 3 
 
 cm 3 
 
 cm 8 
 
 cm 3 
 
 (1) 
 
 6 
 
 27.68 
 
 0.25 
 
 0.13 
 
 0.08 
 
 0.00 
 
 0.03 
 
 0.49 
 
 50 
 
 
 (2) 
 
 6 
 
 27.70 
 
 0.20 
 
 0.10 
 
 0.03 
 
 0.03 
 
 0.03 
 
 0.39 
 
 50 
 
 
 '3) 
 
 6 
 
 28.17 
 
 0.16 
 
 0.10 
 
 0.03 
 
 0.01 
 
 none. 
 
 0.30 
 
 50 
 
 
 (4) 
 
 6 
 
 27.03 
 
 0.60 
 
 0.26 
 
 0.09 
 
 0.03 
 
 none. 
 
 0.98 
 
 150 
 
 (5) 
 
 6 
 
 27.60 
 
 0.93 
 
 0.28 
 
 0.06 
 
 0.04 
 
 0.04 
 
 1.35 
 
 150 
 
 
 6) 
 
 6 
 
 28.60 
 
 1.34 
 
 0.46 
 
 0.17 
 
 0.03 
 
 0.14 
 
 2.14 
 
 200 
 
 
 7) 
 
 6 
 
 28.85 
 
 1.20 
 
 0.50 
 
 0.28 
 
 0.06 
 
 0.27 
 
 231 
 
 200 
 
 
 8) 
 
 6 
 
 31.63 
 
 1.46 
 
 0.74 
 
 0.10 
 
 0.21 
 
 0.23 
 
 2.74 
 
 250 
 
 
 9) 
 
 6 
 
 29.90 
 
 1.04 
 
 0.60 
 
 0.23 
 
 0.15 
 
 0.46 
 
 2.48 
 
 250 
 
 ( 
 
 0) 
 
 6 
 
 36.09 
 
 1.60 
 
 1.23 
 
 0.63 
 
 0.34 
 
 0.18 
 
 3.98 
 
 300 
 
 11) 
 
 6 
 
 37.59 
 
 1.65 
 
 1.33 
 
 0.72 
 
 0.27 
 
 0.10 
 
 4.07 
 
 300 
 
 (12) 
 
 6 
 
 37.23 
 
 1.92 
 
 1.05 
 
 0.64 
 
 0.33 
 
 * 
 
 
 
 300 
 
 it. Both of these quantities increase at a regular rate with the 
 volume of the solution. 
 
 To show with what accuracy the reaction between sodium 
 thiosulphate and iodic acid may be applied to the direct 
 estimation of one of these substances by the other, the 
 averaged results of a large number of titrations are compared 
 in Table VI. The operations were conducted as directed by 
 Riegler, equal measured volumes of standardized sodium 
 thiosulphate being titrated with iodic acid of known strength, 
 in the presence of starch and under different conditions of 
 time, dilution, and mass, the volume of iodic acid required to 
 produced the blue coloration being in each case compared with 
 the volume theoretically required by the terms of Riegler's 
 equation. 
 
 These results show plainly that the amount of iodic acid 
 required to decompose a given amount of sodium thiosulphate 
 may be considerably above or below that required by the terms 
 of Riegler's equation. Thus, with small volumes, and in the 
 absence of potassium iodide, the thiosulphate is destroyed and 
 the separation of iodine commences when only 93 per cent of 
 the theoretical amount of acid has been titrated. At higher 
 * No observation. 
 
BY IODIC ACID. 
 
 61 
 
 TABLE VI. 
 TITRATION OP SODIUM THIOSULPHATB WITH 
 
 IODIC ACID. 
 
 Exp. 
 
 NaAO. 
 
 taken. 
 
 HIO S used. 
 
 HI0 8 
 
 required 
 by theory. 
 
 Error. 
 
 Error. 
 
 KI 
 
 present. 
 
 Volume. 
 
 
 cm 3 
 
 cm 3 
 
 cm 3 
 
 cm 8 
 
 per cent. 
 
 grin. 
 
 cm 3 
 
 (1) 
 
 4 
 
 
 f 18.68 
 
 20.32 
 
 1.64- 
 
 8.0- 
 
 
 50 
 
 (2) 
 
 6 
 
 
 28.32 
 
 30.48 
 
 2.16- 
 
 7.0- 
 
 
 60 
 
 (3) 
 
 6 
 
 
 27.32 
 
 30.48 
 
 3.16- 
 
 10.0- 
 
 
 150 
 
 (4) 
 
 6 
 
 3fc 
 
 28.73 
 
 30.48 
 
 1.75- 
 
 6.0- 
 
 
 200 
 
 (5) 
 
 6 
 
 
 30.77 
 
 30.48 
 
 0.29+ 
 
 0.01+ 
 
 
 250 
 
 (6) 
 
 6 
 
 
 36.97 
 
 30.48 
 
 6.49+ 
 
 21.0+ 
 
 
 300 
 
 (7) 
 
 6 
 
 
 27.46 
 
 30.48 
 
 3.02- 
 
 10.0- 
 
 
 50 
 
 8 
 
 6 
 
 
 26.15 
 
 30.48 
 
 4.33- 
 
 14.0- 
 
 
 150 
 
 (9) 
 (10) 
 
 6 
 6 
 
 t" 
 
 26.50 
 28.16 
 
 30.48 
 30.48 
 
 3.98- 
 2.32- 
 
 13.0- 
 8.0- 
 
 
 200 
 250 
 
 (11) 
 
 6 
 
 1 32.93 
 
 30.48 
 
 2.45+ 
 
 8.0+ 
 
 
 300 
 
 (12) 
 
 4 
 
 .i 22.19 
 
 20.32 
 
 1.87+ 
 
 9.0+ 
 
 2.6 
 
 50 
 
 (13) 
 
 6 
 
 * I 32.48 
 
 30.48 
 
 2.00+ 
 
 7.0+ 
 
 2.0 
 
 60 
 
 dilutions the action is retarded, so that at 250 cm 3 very nearly 
 the theoretical amount of acid is required to produce the first 
 blue color, and at 300 cm 3 an excess of 21 per cent over the 
 theoretical amount must be added. If the " after separa- 
 tion " of iodine is considered to be a measure of the excess of 
 iodic acid, and if its amount is accordingly applied as a 
 correction, it appears that for all volumes below 300 cm 3 the 
 original thiosulphate is completely destroyed when about 90 
 per cent of the theoretical amount of iodic acid has been added. 
 The presence of potassium iodide in the system retards the 
 action, so that at small volumes an excess of about 8 per cent 
 of iodic acid must be added to completely destroy the thiosul- 
 phate and commence the separation of iodine. It is obvious 
 from the preceding experiments that the reaction between 
 iodic acid and sodium thiosulphate is so indefinite in its nature, 
 and so dependent for completeness on conditions of time, 
 dilution, and mass, that its direct application as a means of 
 standardizing solutions must remain impracticable. 
 
 * HI0 8 added to first blue color. 
 
 t Calculated by subtracting from the amount of iodic acid originally 
 introduced, the volume of thiosulphate of equal strength required to bleach 
 the solution after standing twenty hours. 
 
THE COMBUSTION OF ORGANIC SUBSTANCES 
 IN THE WET WAY. 
 
 BY I. K. PHELPS . 
 
 IN a former paper f I have shown that carbon dioxide may 
 be estimated iodometrically with a fair degree of accuracy. 
 Inasmuch as this method is not dependent upon the rate of 
 flow or rapidity of generation of the carbon dioxide, it seemed 
 possible that some advantage might follow its application to 
 the determination of organic carbon, oxidized by liquid 
 reagents. 
 
 Method of Oxidation by Potassium Permanganate. 
 
 The first experimental test in this direction was made with 
 oxalic acid, which was oxidized according to the well-known 
 reaction of potassium permanganate in the presence of sulphuric 
 acid. The apparatus used was the same as that previously 
 described in the iodometric process, referred to above. It 
 consisted, in the main, of an evolution flask, and an absorption 
 flask, properly connected. As an evolution flask, a wide- 
 mouthed flask of about 75 cm 3 capacity was used. This was 
 closed by a doubly perforated rubber stopper, carrying a 
 separating funnel for the introduction of liquid into the flask 
 and a glass tube of 0.7 cm. internal diameter, which was 
 expanded to a small bulb just above the stopper, to carry off 
 the gas. This exit tube was joined by means of a rubber 
 connector to a tube which passed through the rubber stopper 
 of the absorption flask, which was an ordinary round-bottom 
 flask of 250 cm 3 capacity. This tube ended in a valve of the 
 
 * From Am. Jour. Sci., iv, 372. 
 
 t Am. Jour. Sci., vol. ii, p. 70. Volume I, p. 369. 
 
COMBUSTION IN THE WET WAY. 63 
 
 Kreider pattern,* which was enclosed in a larger tube, reaching 
 nearly to the bottom of the absorption flask. The second hole 
 of the stopper of this absorption flask, was filled by a glass 
 tube closed by a rubber connector and screw pinch-cock. 
 
 The barium hydroxide solution for use in the determination 
 of the carbon dioxide was prepared by filtering a cold saturated 
 solution of the commercial salt into a large bottle, which was 
 connected with a self-feeding burette. The solution was 
 standardized in the manner described in my former paper by 
 boiling with an excess of decinormal iodine solution in an 
 ether wash bottle. The short tube of the glass ground stopper 
 of the bottle was sealed to a Will and Varrentrapp absorption 
 apparatus, which was charged during the operation with a 
 solution of potassium iodide to prevent the loss of elementary 
 iodine in the boiling ; the long tube of the bottle was used as 
 an inlet tube and was closed externally by a rubber cap during 
 the boiling. After cooling, the excess of iodine used was 
 determined by titration with decinormal arsenious acid solution 
 and the iodine lost calculated on barium hydroxide molecule 
 for molecule. 
 
 Potassium permanganate was prepared for use by dissolving 
 the commercial salt in water, and boiling this solution, made 
 acid with sulphuric acid, until free from carbon dioxide. 
 Water was also prepared free from carbon dioxide by boiling 
 distilled water until one-third had been driven off hi steam 
 and was kept until used in full-stoppered flasks. 
 
 For the first determinations of carbon, crystallized ammo- 
 nium oxalate was weighed out and introduced into the boiling 
 flask with 10-15 cm 3 of pure water and the flasks connected 
 as described above with an appropriate amount of barium 
 hydroxide solution (3-5 cm 8 in excess of the amount required 
 to precipitate the carbon dioxide to be determined) in the 
 absorption flask. The whole system was then evacuated 
 with the water pump to a pressure of 200-225 mm. and the 
 oxalate solution in the boiling flask warmed. An excess of 
 potassium permanganate solution was then run in through 
 * Am. Jour. Sci., 1, p. 132. Volume I, p. 307. 
 
64 COMBUSTION OF ORGANIC SUBSTANCES 
 
 the funnel tube and the mixture warmed again, when the 
 oxidation of the oxalate was shown by the carbon dioxide 
 evolved. The carbon dioxide was completely set free by 
 the introduction of 10 cm 3 of sulphuric acid (1 : 4) and was 
 driven completely to the absorption flask by boiling for five 
 minutes. During the passage of the gas into the absorption 
 flask, it was shaken frequently and was kept cool by standing 
 in a dish of water and by pouring cold water over it from 
 time to tune. If, during the boiling, any fears are enter- 
 tained as to the strength of the vacuum in the flasks, they 
 may be easily allayed by opening momentarily the stop-cock 
 of the funnel tube and noting the direction of the flow of 
 water, contained in the funnel. After the boiling was ended, 
 the atmospheric pressure was restored by allowing air, purified 
 from carbon dioxide by passage through potash bulbs, to 
 enter through the funnel tube of the boiling flask. Then 
 the flasks were disconnected and the stopper of the absorp- 
 tion flask with its attachments was removed, the valve and 
 its tube being carefully washed free from barium hydroxide. 
 A second stopper, which was provided with a separating 
 funnel, and a Will and Varrentrapp absorption apparatus, 
 containing water to serve as a trap, was inserted into the 
 mouth of the absorption flask and the emulsion brought to 
 the boiling point. Decinormal iodine solution was then run 
 in through the funnel tube in sufficient quantity to destroy 
 the larger part of the excess of barium hydroxide and the 
 emulsion brought to the boiling point again, after which iodine 
 was again run hi but this time to the permanent red color of 
 the excess of free iodine. After cooling, this excess of iodine 
 was determined by titration with decinormal arsenious acid 
 solution. Thus, the excess of barium hydroxide taken being 
 determined by the iodine lost, the barium hydroxide used, 
 now hi the form of carbonate, was known, from which the 
 carbon dioxide which precipitated this carbonate, may be 
 calculated. 
 
 The following results were obtained by this procedure. 
 
IN THE WET WAY. 
 TABLE I. 
 
 65 
 
 Exp. 
 
 Ammonium 
 
 oxalate 
 taken. 
 
 Ba0 2 H, 
 
 taken. 
 
 BaO 2 H 2 
 found. 
 
 found. 
 
 C0 2 
 calculated. 
 
 Error on C0 2 . 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.2522 
 
 0.7267 
 
 0.1170 
 
 0.1565 
 
 0.1561 
 
 0.0004+ 
 
 2 
 
 0.2542 
 
 0.7267 
 
 0.1113 
 
 0.1579 
 
 0.1574 
 
 0.0005+ 
 
 3) 
 
 0.5020 
 
 1.4535 
 
 0.2417 
 
 0.3110 
 
 0.3108 
 
 0.0002+ 
 
 (4) 
 
 0.5058 
 
 1.3954 
 
 0.1753 
 
 0.3131 
 
 0.3131 
 
 O.OOOOi 
 
 (5) 
 
 1.0033 
 
 2.6163 
 
 0.1955 
 
 0.6213 
 
 0.6211 
 
 0.0002+ 
 
 (6) 
 
 1.0003 
 
 2.6951 
 
 0.1836 
 
 0.6189 
 
 0.6192 
 
 0.0003- 
 
 (7) 
 
 1.0010 
 
 2.6163 
 
 0.2037 
 
 0.6192 
 
 0.6197 
 
 0.0005- 
 
 In experiments (5) and (6), a few drops of ammonia 
 were added to the oxalate solution before running in the 
 permanganate ; in (3) and (7), the permanganate was treated 
 to alkalinity with barium hydroxide ; in the remaining ex- 
 periments, (1), (2), and (4), the permanganate was slightly 
 acid with the sulphuric acid used in its purification from 
 carbon dioxide, as already described. The results obtained 
 are good and it is plain that the oxidation proceeded regularly, 
 whether the first action of the permanganate was in the 
 alkaline or slightly acid solution. 
 
 Jones * has shown that formates may be determined volu- 
 metrically by titration with potassium permanganate in 
 alkaline solution. In an attempt to determine formates by 
 the process outlined above, the pure barium salt was used. 
 This was prepared by treating the aqueous solution of formic 
 acid with pure barium carbonate to neutrality and crystalliz- 
 ing the product. It was proved pure by ignition and weigh- 
 ing in the form of carbonate. 
 
 In making determinations of carbon in this formate, weighed 
 portions were introduced into the boiling flask, together 
 with sodium hydroxide solution, which was taken in such 
 quantity as to more than neutralize the acid in the potassium 
 permanganate. Naturally, the sodium hydroxide must be 
 freed from carbonate and this was effected by treatment 
 with an excess of barium hydroxide and filtering. An excess 
 
 VOL. II. 5 
 
 * Amer. Chem. Jour., xvii, 539. 
 
66 
 
 COMBUSTION OF ORGANIC SUBSTANCES 
 
 of potassium permanganate is then run into the flask and 
 the solution heated to boiling. An excess of dilute sulphuric 
 acid is introduced into the mixture and the carbon dioxide, 
 thus set free, completely driven over to the absorption flask 
 and determined as before. Table II shows results obtained 
 by the process. 
 
 TABLE II. 
 
 Exp. 
 
 Barium 
 formate 
 taken. 
 
 BaO,H 2 
 taken. 
 
 BaO,H, 
 
 found. 
 
 CO. 
 found. 
 
 CO, 
 
 calculated. 
 
 Error on CO S . 
 
 
 grm. 
 
 gnu. 
 
 grin. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 ft 
 
 (3) 
 
 0.5001 
 0.5033 
 1.0002 
 
 0.9302 
 0.9012 
 1.6861 
 
 0.1745 
 0.1402 
 0.1793 
 
 0.1939 
 0.1953 
 
 0.3867 
 
 0.1935 
 0.1947 
 0.3870 
 
 0.0004+ 
 0.0006+ 
 0.0003- 
 
 (4) 
 
 1.0059 
 
 1.6279 
 
 0.1093 
 
 0.3897 
 
 0.3892 
 
 0.0005+ 
 
 (5) 
 
 1.3750 
 
 2.2529 
 
 0.1820 
 
 0.5315 
 
 0.5320 
 
 0.0005- 
 
 (6) 
 
 1.5028 
 
 2.4419 
 
 0.1754 
 
 0.5816 
 
 0.5814 
 
 0.0002+ 
 
 These results show plainly that the carbon of formic acid 
 may be determined accurately by the method outlined. 
 
 It was found incidentally that ammonia cannot take the 
 place of the sodium hydroxide in this process, probably 
 because the ammonia volatilizes to the absorption flask dur- 
 ing the boiling and is acted on by the iodine subsequently 
 used and is thus registered as barium hydroxide. 
 
 It is a well-known fact that tartrates are oxidized by 
 permanganates. I have found, however, that when tartaric 
 acid is treated under the conditions of analysis outlined 
 above in acid solution, the oxidation is incomplete; but that 
 oxidation is complete if the tartrate is heated in a solution 
 alkaline with sodium hydroxide and then acidified with 
 sulphuric acid. 
 
 The tartrate used was a recrystallized tartar emetic, dried 
 at 100 C. The following results were obtained with such 
 a tartrate by this process. 
 
 It seems possible to draw the general conclusion from the 
 results recorded that organic substances which are oxidized 
 completely by the permanganate may be determined by the 
 process outlined above. It will also be seen that the use of 
 
IN THE WET WAY. 
 TABLE IH. 
 
 67 
 
 Exp. 
 
 Tartar 
 emetic, 
 taken. 
 
 BaOjH, 
 
 taken. 
 
 BaO^ 
 
 found. 
 
 CO, 
 found. 
 
 calculated. 
 
 Error on C0 2 . 
 
 
 grlil. 
 
 grm. 
 
 gnu. 
 
 grm. 
 
 grm. 
 
 grm, 
 
 (1) 
 
 0.5051 
 
 1.2450 
 
 0.1709 
 
 0.2756 
 
 0.2751 
 
 0.0005+ 
 
 (2) 
 
 0.5030 
 
 1.2226 
 
 0.1536 
 
 0.2743 
 
 0.2739 
 
 0.0004+ 
 
 (3) 
 
 0.7509 
 
 1.7355 
 
 0.1401 
 
 0.4094 
 
 0.4091 
 
 0.000:3+ 
 
 (4) 
 
 0.7541 
 
 1.7430 
 
 0.1410 
 
 0.4111 
 
 0.4107 
 
 0.0004+ 
 
 (5) 
 
 1.0018 
 
 2.3456 
 
 0.2187 
 
 0.5458 
 
 0.5456 
 
 0.0002+ 
 
 (6) 
 
 1.0005 
 
 2.2435 
 
 0.1196 
 
 0.5451 
 
 0.5450 
 
 0.0001+ 
 
 the rubber stopper in the boiling flask, with due care to prevent 
 its contact with the solution, does not introduce an appreciable 
 error. 
 
 Wanklyn and Cooper * and others have noted the fact that 
 potassium permanganate, whether in acid or alkaline solution, 
 will not oxidize all organic substances (acetates, for example), 
 even at the boiling temperature. It is well known that a 
 mixture of concentrated sulphuric and chromic acids has 
 a much wider field of action in oxidizing organic compounds 
 than the permanganate. With hopes of applying this reagent 
 more widely to the determination of organic carbon, the 
 experiments about to be recorded were tried. 
 
 Method of Oxidation by Chromic Acid. 
 
 A concentrated mixture of chromic and sulphuric acids, 
 although a much more powerful oxidizer than potassium 
 permanganate in aqueous solutions, fails to oxidize completely 
 many organic compounds. Thus Cross and Higgin f have 
 shown that carbohydrates are among the number of such 
 organic substances ; and later Cross and Be van find that car- 
 bohydrates and many other substances are oxidized to a mix- 
 ture of carbon dioxide and monoxide. Messinger J has proved 
 that carbon may be determined in organic compounds by pass- 
 ing the mixed products, resulting from the oxidation with 
 chromic and sulphuric acids, through a short combustion tube, 
 
 * Phil. Mag. (5), vii, 138. t Jour. Chem. Soc., xli, 113. 
 
 J Ber. Dtsch. chem. Ges., xxiii, 2756. 
 
68 
 
 COMBUSTION OF ORGANIC SUBSTANCES 
 
 filled with granular copper oxide and heated in a furnace all 
 of which facts have been confirmed in my own experience. 
 
 Ludwig * has observed that the contact of carbon monoxide 
 with a mixture of chromic and sulphuric acids, especially 
 when hot, results hi the oxidation of that gas to carbon dioxide. 
 This fact suggested the idea of substituting for the apparatus 
 described above a new form, adapted to retain the first products 
 of oxidation in prolonged contact with the oxidizing mixture. 
 
 This apparatus, shown in the ac- 
 companying figure, by means of 
 which, as the sequel shows, it has 
 been found possible to extend the 
 availability of the oxidizing mix- 
 ture, is put together as follows: 
 A thick-walled, round-bottomed 
 flask of a liter's capacity, serving 
 as an oxidizing chamber, is closed 
 by a rubber stopper with two 
 perforations, through one of which 
 passes the tube of a separating 
 funnel of about 100 cm 3 capacity. 
 The tube of this funnel reaches 
 nearly to the bottom of the flask 
 and is drawn out at the lower end. 
 
 A disc of platinum foil is hung in the neck of the flask, nearly 
 closing it, and held in place by a platinum wire passing 
 through the foil and tucked under the rubber stopper where 
 the funnel tube enters. The second hole of the stopper is 
 filled by the exit tube, a glass tube of 0.7 cm. internal 
 diameter. This tube is expanded just above the stopper to 
 a small bulb, which serves to prevent mechanical loss of 
 the solid contents of the flask during the boiling, and is 
 joined by means of a rubber connector (provided with a screw 
 pinch-cock) to the inlet tube of the absorption flask, which 
 is an ordinary 500 cm 3 round-bottomed flask. This flask is 
 also closed by a rubber stopper with two perforations, through 
 
 * Ann. Chem. (Liebig), clxii, 47. 
 
 FIG. 19. 
 
IN THE WET WAY. 69 
 
 one of which passes the inlet tube described above and through 
 the other the exit tube, which is also enlarged to a small bulb 
 just above the stopper and is closed by a rubber connector and 
 screw pinch-cock. The glass ground stopper of the funnel 
 tube is carefully cleaned and lubricated with a thick solution 
 of metaphosphoric acid. 
 
 Instead of getting the vacuum by the water pump, it may 
 be got almost as quickly and certainly more simply by boiling 
 the water in the evolution flask and the barium hydroxide 
 solution in the absorption flask at the same tune both flasks 
 being connected, ready for making a determination. When 
 steam issues in good quantity from the exit tube, the burner 
 is removed from under the evolution flask, the attached pinch- 
 cock closed, the burner under the absorption flask taken away, 
 and the screw pinch-cock upon the exit tube quickly closed. 
 The flasks are then allowed to cool. 
 
 In making a determination, the organic substance is weighed 
 out in a counterbalanced bulb, so thin that it may be easily 
 broken later and made with a wide mouth for convenience 
 in introducing the solid substance. After the substance is 
 weighed, the mouth of the bulb is sealed by heating hi a 
 small blow-pipe flame and the tube introduced into the evolu- 
 tion flask, together with an amount of pure potassium dichro- 
 mate, which is known to be hi excess of that required to 
 oxidize the organic substance. The flasks are connected, as 
 already described, with an appropriate amount of barium 
 hydroxide solution in the absorption flask and 10 cm 3 of pure 
 water in the evolution flask, and the vacuum is obtained (as 
 described above) by boiling both flasks, the boiling being 
 stopped when the water in the evolution flask has decreased to 
 2 or 3 cm 3 . Naturally, this boiling must be so regulated as 
 not to allow loss of the solid material in either flask. The 
 vacuum obtained, the tube containing the organic substance is 
 broken by shaking the flask, and 20 cm 3 of concentrated 
 sulphuric acid, previously purified from organic material by 
 heating to the fuming point with a few crystals of potassium 
 dichromate, are run in through the funnel tube, when reduc- 
 
70 
 
 COMBUSTION OF ORGANIC SUBSTANCES 
 
 tion of the chromic acid soon becomes evident. While still 
 hot, the acid is shaken in the flask violently, the platinum foil 
 hung in the neck serving to protect the rubber stopper. The 
 flask is warmed to approximately 105 C., the highest temper- 
 ature to which, as shown by Cross and Be van,* a mixture of 
 chromic and sulphuric acids may be safely heated without the 
 disengagement of oxygen gas. Water is then run in until 
 the crystals of chromic anhydride have disappeared and the 
 danger of the evolution of oxygen is past. The solution is 
 heated to its boiling point, care being taken that it shall not 
 get under pressure, which can easily be observed by opening 
 momentarily the stop-cock of the funnel tube and noting the 
 direction of the flow of water, contained in the funnel. The 
 flask is shaken and heated alternately for five minutes a 
 period of tune which appears to be sufficient to bring about 
 the oxidation of the small amount of carbon monoxide origi- 
 nally produced. Then more water (60-70 cm 3 ) is introduced 
 through the funnel, and the stop-cock between the boiling and 
 absorption flasks is opened, when the carbon dioxide enters the 
 
 TABLE IV. 
 
 Exp. 
 
 Substance 
 taken. 
 
 Ba0 2 Hj 
 taken. 
 
 Ba0 2 H, 
 
 found. 
 
 C0 2 
 
 found. 
 
 CO, 
 
 calculated. 
 
 Error on C0 2 . 
 
 ANALYSIS OF AMMONIUM OXALATB. 
 
 (1) 
 (2) 
 (3) 
 (4) 
 (5) 
 
 grm. 
 0.5009 
 0.5006 
 0.5005 
 1.0002 
 1.0010 
 
 grm. 
 1.3534 
 1.3400 
 1.3400 
 2.5460 
 2.5192 
 
 grm. 
 0.1469 
 0.1308 
 0.1343 
 0.1347 
 0.1094 
 
 grin. 
 
 0.3097 
 0.3103 
 0.3094 
 0.6188 
 0.6185 
 
 grm. 
 
 0.3101 
 0.3099 
 0.3098 
 0.6192 
 0.6197 
 
 1*111. 
 
 0.0004- 
 0.0004+ 
 0.0004- 
 0.0004- 
 0.0012- 
 
 ANALYSIS OF CANE SUGAR. 
 
 Pj 
 
 % 
 
 0.2001 
 0.2000 
 0.2001 
 0.2014 
 
 1.3926 
 1.3926 
 1.3926 
 1.3400 
 
 0.1905 
 0.1936 
 0.1857 
 0.1279 
 
 0.3085 
 0.3077 
 0.3097 
 0.3111 
 
 0.3088 
 0.3086 
 0.3088 
 0.3108 
 
 0.0003- 
 0.0009- 
 0.0009+ 
 0.0003-f- 
 
 absorption flask, which is kept cool and shaken as before. 
 
 The contents of the evolution flask are then heated to boiling 
 
 * Jour. Chem. Soc., liii, 889. 
 
IN THE WET WAY. 71 
 
 and a slow current of air, freed from carbon dioxide by passing 
 through potash bulbs, is allowed to enter through the funnel 
 tube to keep the liquid from undue bumping. The boiling is 
 continued for fifteen minutes, after which the excess of barium 
 hydroxide is determined iodometrically and thus the carbon 
 dioxide present estimated as before. Table IV shows results 
 obtained by the treatment of crystallized ammonium oxalate 
 and cane sugar, recrystallized from dilute alcoholic solution, 
 in this manner. 
 
 The results are evidently very satisfactory. 
 
 The Determination of the Oxygen required to Oxidize an 
 Organic /Substance. 
 
 Several different methods have been proposed for estimating 
 the oxygen present in organic substances, depending, in gen- 
 eral, upon the determination of the oxygen which must be 
 supplied to burn the substance to a known amount of carbon 
 dioxide and water thus discovering by difference the oxygen 
 originally contained in the substance. Lavoisier is said to have 
 measured directly the oxygen used in burning organic sub- 
 stances ; Gay-Lussac and The'nard determined the oxygen used 
 by measuring the amount of potassium chlorate reduced hi 
 burning the organic compound; Baumhauer* determined the 
 oxygen used by measuring the volume of oxygen entering the 
 combustion furnace and subtracting the measure of the gas 
 coming from the combustion tube, which was set up according 
 to the well known method for determining carbon and hydro- 
 gen ; Stromeyerf determined the amount of copper reduced by 
 the ignition of the substance in copper oxide; LadenburgJ 
 oxidized the substance by heating in a sealed tube with a 
 known amount of iodic acid, determining at the end of the 
 operation the amount of iodic acid left; Mitscherlich has 
 estimated the oxygen in organic substances directly by decom- 
 posing the substance by ignition hi a stream of chlorine gas, 
 
 / 
 
 * Ann. Chem. (Liebig), xc, 228. t Ann. Chera. (Liebig), cxvii, 247. 
 \ Ann. Chem. (Liebig), cxxxv, 1. Ann. Phys. ccvi, 536 (1867). 
 
72 COMBUSTION OF ORGANIC SUBSTANCES 
 
 estimating the oxygen content by determining the resulting 
 carbon dioxide and monoxide. 
 
 As it has been shown in the work described that carbon may 
 be determined in organic substances by oxidation with chromic 
 and sulphuric acids without the evolution of oxygen gas, it 
 would seem that the determination of the oxygen in the sub- 
 stance might be effected by determining the amount of chromic 
 acid used in the operation, taking into consideration the 
 products of combustion. This can be readily accomplished by 
 taking a weighed amount of pure potassium dichromate as the 
 oxidizing agent and determining, at the end of the operation, 
 by treatment of the residue with hydrochloric acid, absorp- 
 tion of the chlorine evolved in an alkaline arsenite of known 
 strength, and titration of the excess of that substance with 
 decinormal iodine solution, the amount of chromic acid left. 
 
 To test the accuracy of the determination of chromic acid 
 under these conditions of analysis, weighed portions of pure 
 fused potassium dichromate were introduced into a Voit flask, 
 whose outlet tube was sealed to the inlet tube of a Drexel 
 wash bottle, the outlet of which, in turn, was sealed to a 
 Will and Varrentrapp absorption apparatus. An amount of 
 hydrochloric acid, more than enough to completely reduce 
 the chromate (15-40 cm 3 of the strongest acid), was added 
 with 20 cm 3 of strong sulphuric acid and the total volume 
 made up to 120-140 cm 3 of liquid. The sulphuric acid used 
 here was purified from carbonaceous matter (as in the carbon 
 determination above) by heating with a few crystals of potas- 
 sium dichromate, the excess of which was destroyed by hold- 
 ing the acid at the fuming point for about two hours, when 
 a portion diluted with water gave no color with potassium 
 iodide and starch paste. Pure arsenious oxide, in amount 
 slightly in excess of that required to take up the oxygen to 
 be given up by the chromate, was dissolved by the aid of 
 heat in a solution of pure sodium hydroxide, taken in such 
 quantity as to more than neutralize the arsenious acid and 
 the hydrochloric acid used to reduce the chromate, and this 
 solution was introduced into the Drexel wash bottle. The 
 
IN THE WET WAY. 
 
 73 
 
 flask was then connected with the wash bottle, using a thick 
 solution of metaphosphoric acid to lute the joint between the 
 flask and its stopper. The absorption apparatus was charged 
 with a dilute solution of sodium hydroxide. Carbon dioxide 
 was generated in a Kipp generator by the action of hydro- 
 chloric acid on marble and purified from reducing matter by 
 bubbling through a strong solution of iodine in potassium 
 iodide and finally washed with a solution of potassium iodide 
 alone. A slow stream of this purified carbon dioxide was 
 allowed to enter the inlet tube of the Voit flask, the con- 
 tents of which were then boiled. When concentration to a 
 volume of 30-40 cm 3 was reached, the boiling was discon- 
 tinued and, after cooling and disconnecting the flask, the 
 contents of the receiver were made acid with sulphuric acid 
 and then alkaline with acid potassium carbonate, and the 
 excess of arsenite was determined by titration with deci- 
 normal iodine solution. Sometimes during the reduction of 
 the chromic acid, the red fumes of the chlorochromic an- 
 hydride Volatilized to the receiver; but since the chromic 
 acid thus produced is reduced later by the arsenite,* this 
 transfer is of no account in the working of the process. 
 The following results were thus obtained. 
 
 TABLE V. 
 
 Exp. 
 
 KjCr 2 O 7 
 taken. 
 
 As 2 3 
 taken. 
 
 found. 
 
 K 2 Cr 2 7 
 found. 
 
 Error on 
 K 2 Cr a 7 . 
 
 
 gnu. 
 
 grill. 
 
 grin. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 5.0002 
 
 5.1025 
 
 0.1144 
 
 4.9447 
 
 0.0555- 
 
 (2) 
 
 5.0018 
 
 5.0799 
 
 0.0526 
 
 4.9849 
 
 0.0169- 
 
 (3) 
 
 5.0005 
 
 5.0801 
 
 0.0582 
 
 4.9782 
 
 0.0223- 
 
 U) 
 
 5.0013 
 
 5.0706 
 
 0.0908 
 
 4.9365 
 
 0.0648- 
 
 The cause of the error shown in these experiments was 
 traced finally to too great concentration of the sulphuric acid 
 in the process. When the boiling begins the chromate is 
 reduced gradually and if the evaporation of the water is 
 pushed too rapidly, the sulphuric acid may reach a strength 
 
 * Browning Am. Jour. Sci., i, 35. Volume I, p. 344 . 
 
74 
 
 COMBUSTION OF ORGANIC SUBSTANCES 
 
 at which it begins to cause the reduction of the chromic acid 
 with the evolution of oxygen instead of chlorine. 
 
 The obvious remedy is to conduct the boiling operation 
 more slowly. It was found that, if from 5-6 hours' time was 
 taken for the proper concentration of the contents of the 
 Voit flask, the presence of the sulphuric acid worked no 
 harm, as will be seen from the following results. Experi- 
 ments (1) and (5) were made with 5 cm 3 of sulphuric acid 
 present and the others with 20 cm 3 , as used before. 
 
 TABLE VI. 
 
 Bxp. 
 
 K 2 Cr 2 T 
 taken. 
 
 As 2 s 
 taken. 
 
 As 2 3 
 found. 
 
 K*Cr 7 
 
 found. 
 
 Error on 
 K 2 Cr 2 0,- 
 
 
 grm. 
 
 grm. 
 
 grin. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 1.0004 
 
 1.0500 
 
 0.0398 
 
 1.0014 
 
 0.0010+ 
 
 (2) 
 
 1.0007 
 
 1.0531 
 
 0.0437 
 
 1.0006 
 
 0.0001- 
 
 (3) 
 
 2.0013 
 
 2.0501 
 
 0.0299 
 
 2.0026 
 
 0.0013+ 
 
 4 
 
 2.0037 
 
 2.0727 
 
 0.0502 
 
 2.0049 
 
 0.0012+ 
 
 (6) 
 
 5.0020 
 
 5.1002 
 
 0.0495 
 
 5.0068 
 
 0.0048+ 
 
 (6) 
 
 5.0037 
 
 6.1018 
 
 0.0513 
 
 5.0066 
 
 0.0029+ 
 
 In applying this method to the determination of oxygen 
 used in the oxidation of an organic substance, the carbon 
 determination was made as already described, the amount of 
 water used being such as to leave 60-80 cm 3 of liquid in the 
 boiling flask after the carbon dioxide had been driven to the 
 absorption flask by boiling. This liquid was then washed 
 into the Voit flask and the chromic acid remaining determined 
 by a second distillation (this time with hydrochloric acid) 
 in the manner described above. In each of the experiments 
 recorded below, 20 cm 8 of purified sulphuric acid were used 
 in the carbon determination and 35 cm 3 of hydrochloric acid 
 (sp. gr. 1.2) in the chromic acid determination. The ammo- 
 nium oxalate used was the pure crystallized salt ; the phthalic 
 acid was recrystallized from its water solution and dried for 
 a short time over sulphuric acid ; the cane sugar was selected 
 crystals of rock candy, recrystallized from dilute alcoholic 
 solution and dried for a long time over sulphuric acid; the 
 
IN THE WET WAY. 
 
 75 
 
 paper was ashless filter paper, dried to a constant weight over 
 sulphuric acid; the tartar emetic was recrystallized from 
 water solution and air dried; the barium formate was pre- 
 pared by treating formic acid with an excess of pure bar- 
 ium carbonate, filtering hot and allowing the product to 
 crystallize. 
 
 TABLE VII. 
 
 Exp. 
 
 Sub- 
 
 staiice 
 taken. 
 
 found. 
 
 Error 
 on 
 CO, 
 
 K 2 Cr 2 7 
 taken. 
 
 A-,0, 
 
 taken. 
 
 found. 
 
 Oxygen 
 
 used. 
 
 Oxygen 
 required 
 by 
 theory. 
 
 Error 
 on 
 Oxygen. 
 
 ANALYSIS OF AMMONIUM OXALATE. 
 
 (1) 
 (2) 
 
 grin. 
 
 1.0122 
 1.0019 
 
 grin. 
 
 0.8265 
 0.6212 
 
 grm. 
 0.0001- 
 0.0010+ 
 
 grm. 
 2.0009 
 2.0002 
 
 grin. 
 
 1.3002 
 1.3517 
 
 grm. 
 0.0000 
 0.0440 
 
 grm. 
 
 0.1160 
 0.1147 
 
 grm. 
 0.1139 
 0.1128 
 
 grm. 
 0.0021+ 
 0.0019+ 
 
 ANALYSIS OF PHTHALIC ACID. 
 
 (1) 
 (2) 
 
 0.1002 
 0.1093 
 
 0.2138 
 0.2324 
 
 0.0014+ 
 0.0007+ 
 
 2.0012 
 2.0000 
 
 1.2004 
 1.1031 
 
 0.0814 
 0.0634 
 
 0.1456 
 0.1582 
 
 0.1448 
 0.1580 
 
 0.0008+ 
 0.0002+ 
 
 ANALYSIS OF CANE SUGAR. 
 
 (1) 
 (2) 
 
 0.2025 
 0.4012 
 
 0.3117 
 0.6166 
 
 0.0008- 
 0.0024- 
 
 3.0000 
 5.0000 
 
 1.7002 
 2.3022 
 
 0.0796 
 0.0366 
 
 0.2275 
 0.4495 
 
 0.2273 
 
 0.4502 
 
 0.0002+ 
 0.0007- 
 
 ANALYSIS OF PAPER. 
 
 (1) 
 
 (2) 
 
 0.3034 
 0.4523 
 
 0.4932 
 0.7334 
 
 0.0010- 
 0.0033- 
 
 3.5015 
 5.0035 
 
 1.4017 
 1.8000 
 
 0.0879 
 0.0710 
 
 0.3589 
 0.5368 
 
 0.3598 
 0.5358 
 
 0.0009- 
 0.0010+ 
 
 ANALYSIS OF TARTAR EMETIC. 
 
 (1) 
 (2) 
 
 0.5057 
 1.0099 
 
 0.2671 
 0.5321 
 
 0.0009- 
 0.0030- 
 
 2.5018 
 3.5003 
 
 1.7000 
 1.7520 
 
 0.0766 
 0.0198 
 
 0.1459 
 0.2911 
 
 0.1462 
 0.2919 
 
 0.0003- 
 0.0008- 
 
 ANALYSIS OF BARIUM FORMATE. 
 
 (1) 
 (2) 
 
 1.0079 
 1.5014 
 
 0.3906 
 0.5814 
 
 0.0006+ 
 0.0005+ 
 
 3.0026 
 3.0010 
 
 2.2002 
 1.8080 
 
 0.0496 
 0.0890 
 
 0.1423 
 0.2118 
 
 0.1422 
 0.2118 
 
 0.0001+ 
 0.0000 
 
 From these results, it will be seen that the process works 
 with accuracy upon a great variety of organic substances. It 
 was found impossible, however, to determine the elements 
 in bodies which are at the same time volatile and hard to 
 oxidize ; for instance, ether oxidizes easily to acetic acid 
 but difficultly beyond that stage; although the liquid acid 
 
76 COMBUSTION OF ORGANIC SUBSTANCES. 
 
 is oxidized vigorously by chromic and sulphuric acids, the 
 gaseous acid is hardly attacked at the temperature used; 
 naphthaline was also found to be volatilized, and hence not 
 attacked, to such an extent as to render its determination 
 by this process valueless. 
 
XI 
 
 THE ESTIMATION OF MANGANESE AS THE 
 SULPHATE AND AS THE OXIDES. 
 
 BY F. A. GOOCH AND MARTHA AUSTIN.* 
 
 THE estimation of manganese by the conversion of salts of 
 that element with volatile acids to the form of the anhydrous 
 sulphate by the action of an excess of sulphuric acid, evapora- 
 tion, and gentle heating was formerly a recognized procedure. 
 On the authority of Rose,f however, this method was set aside 
 on account of the supposed difficulty of removing the excess 
 of acid without disturbing the composition of the normal salt. 
 Thus, Oesten, working under Rose's direction, obtained, upon 
 submitting the crystalline hydrous sulphate MnSO 4 . 5H 2 O to 
 heat, results which may be summarized and compared with the 
 results obtained by Rose's sulphide method (the ignition of 
 the residue with sulphur in hydrogen) as follows : 
 
 MnS0 4 . 5H 2 
 taken. 
 
 MnS0 4 
 found. 
 
 Theory. 
 
 Error. 
 
 MnS 
 found. 
 
 Theory. 
 
 Error. 
 
 grm. 
 
 grm. 
 
 gnu. 
 
 grm. 
 
 grm. 
 
 grin. 
 
 grm. 
 
 1.659 
 
 1.043J 
 
 1.037 
 
 0.006+ 
 
 0.597 
 
 0.595 
 
 0.002+ 
 
 
 1.023 
 
 
 0.014- 
 
 
 
 
 1.481 
 
 0.934} 
 
 0.926 
 
 0.008+ 
 
 
 
 
 
 0.905 
 
 
 0.021- 
 
 
 
 
 
 0.725H 
 
 
 0.201- 
 
 
 
 
 1.430 
 
 0.880 
 
 0.893 
 
 0.013- 
 
 0.509 
 
 0.512 
 
 0.003- 
 
 The residues remaining after the gentle ignition of the sulphate 
 weighed apparently several milligrams more than should have 
 been the case if the salt had been reduced to the normal 
 anhydrous sulphate. At higher temperatures the sulphate 
 
 * From Am. Jour. Sci., v, 209. t Ann. Phys., clxxxvi, 125 (1860). 
 
 t Ignited gently. Ignited at low red heat. || Ignited at strong red heat. 
 
78 
 
 ESTIMATION OF MANGANESE AS THE 
 
 turned brown and lost altogether too much weight. A 
 comparison of the errors of the process in which the ignition 
 was at low temperature with those of the sulphide process 
 would seem to justify Rose's rejection of the former method 
 for the latter. Upon recalculating these results, however, 
 using atomic weights in use at present viz. : Mn = 55, S = 
 32.06, O = 16 it becomes plain that the errors of the two 
 processes, as shown in Oesten's work, are not very different 
 numerically, though with opposite signs. 
 
 MnS0 4 . 5H 2 
 taken. 
 
 MnS0 4 
 found. 
 
 Theory. 
 
 Error. 
 
 MnS 
 found. 
 
 Theory. 
 
 Error. 
 
 grin. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grill. 
 
 grm. 
 
 1.659 
 
 1.043 
 
 1.039 
 
 0.004+ 
 
 0.597 
 
 0.599 
 
 0.002- 
 
 1.481 
 
 0.934 
 
 0.928 
 
 0.006+ 
 
 
 
 
 1.430 
 
 
 
 
 0.509 
 
 0.516 
 
 0.007- 
 
 The most uncertain element in these experiments is the 
 difficulty, well-recognized at present, of getting the hydrous 
 manganous sulphate, upon which the experiments were made, 
 in a perfectly definite condition of hydration. 
 
 Volhard* subsequently studied the sulphate process, and 
 showed that manganous sulphate may be dehydrated, separated 
 from an excess of sulphuric acid, and brought into definite 
 condition for weighing as the anhydrous salt by careful and 
 protracted heating over a special device of his own a ring 
 burner enclosed in a sheet-iron casing. Thus, on evaporating 
 and dehydrating a solution of pure neutral manganous sul- 
 phate, Volhard obtained the results recorded in the following 
 statement : 
 
 Residue of MnS0 4 left by evaporation and dehydration . 0.1635 
 " after treatment with 3 drops of H 2 S0 4 and heating 
 
 3 hours 0.1635 
 
 after heating 2 hours 0.1638 
 
 " after treatment with 4 drops of H 2 S0 4 and heating 
 
 2 hours 0.1635 
 
 " after heating 3 hours 0.1635 
 
 * Ann. Chem. (Liebig), cxcviii, 328. 
 
SULPHATE AND AS THE OXIDES. 79 
 
 Similar results were obtained on evaporating with sulphuric 
 acid and igniting in like manner an aqueous solution of 
 manganous chloride. Volhard's recommendation of the method 
 has not secured for it the acceptance which its simplicity and 
 exactness would seem to demand possibly because the 
 periods of ignition appear to be considerable and the manner 
 of heating special. 
 
 In our own experiments with the sulphate process we have 
 found that special apparatus is unnecessary, that the time of 
 treatment may be short, and that the process is in every respect 
 simple as well as very exact. We took for a starting point 
 manganous chloride prepared in the manner to be detailed. 
 An aqueous solution of the so-called pure manganous chloride 
 of commerce was boiled with pure manganous carbonate 
 (to throw out aluminum, iron, and chromium), filtered and 
 precipitated with ammonium sulphide. The precipitate thus 
 obtained was dissolved in a very slight excess of hydrochloric 
 acid (to leave behind possible traces of nickel, cobalt, and 
 copper), the solution was boiled to expel hydrogen sulphide 
 and precipitated with sodium carbonate. The manganous 
 carbonate thus thrown down was boiled repeatedly with 
 successive portions of water, and washed until the washings 
 were free from chloride. The greater part of this purified 
 carbonate was dissolved in the least possible amount of pure 
 hydrochloric acid, the reserved portion of the carbonate was 
 added, the mixture was boiled, and the solution of the purified 
 and neutral manganous chloride was filtered from the excess 
 of undissolved carbonate. Definite portions of this solution 
 were precipitated with silver nitrate, and from the weight of 
 the silver chloride thus obtained the amount of manganous 
 chloride present was calculated. Portions of the solution thus 
 standardized were drawn, for our experiments, from a burette 
 into a weighed platinum crucible, sulphuric acid was added 
 in amount more than equivalent to the manganese, the solution 
 was evaporated on the water-bath until the water was removed, 
 and then, supported by means of a porcelain ring, or triangle, 
 within a larger porcelain crucible used as a radiator so that 
 
80 
 
 ESTIMATION OF MANGANESE AS THE 
 
 the bottom and walls of the one were distant from the bottom 
 and walls of the other by an interval of about 1 cm., the 
 crucible was heated more strongly. The outer porcelain 
 crucible may be heated over a good Bunsen flame to a red 
 heat without risk of overheating the manganese sulphate 
 within the inner crucible, and the ignition may proceed as 
 rapidly as is consistent with the avoidance of mechanical loss 
 by spattering. The results obtained by treatment of equal 
 portions (50 cm 3 ) of the same solution are given, together 
 with the results of standardizing the solution by precipitation 
 with silver nitrate, hi columns A of the following table. In 
 the other columns are given comparative results got in the 
 treatment of equal portions of several other solutions employed 
 subsequently in other work. 
 
 
 MnSO 4 
 
 
 MnSO 4 
 
 
 
 
 calculated 
 
 
 found by 
 
 
 
 Exp. 
 
 from 
 AgCl found 
 in 50 cm 3 of 
 
 Exp. 
 
 treatment 
 of50cm3of 
 solution A 
 
 Exp. 
 
 MnS0 4 found by treatment of 50 cm 8 of 
 various solutions with H 2 SO 4 . 
 
 
 solution A. 
 
 
 withH 2 SO 4 . 
 
 
 
 
 A. 
 
 
 A. 
 
 
 B. 
 
 C. 
 
 D. 
 
 E. 
 
 F. 
 
 G. 
 
 
 grm. 
 
 
 grm. 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.3518 
 
 (1) 
 
 0.3513 
 
 (1) 
 
 0.3100 
 
 0.3256 
 
 0.3534 
 
 0.3524 
 
 0.3355 
 
 0.5475 
 
 (2) 
 
 0.3512 
 
 (2) 
 
 0.3514 
 
 (2) 
 
 0.3104 
 
 0.3254 
 
 0.3543 
 
 0.3520 
 
 0.3357 
 
 0.5476 
 
 
 
 (3) 
 
 0.3518 
 
 (3) 
 
 0.3096 
 
 
 
 
 
 
 These results show plainly that the process of estimating 
 manganese in the form of the anhydrous sulphate is both 
 simple and accurate. 
 
 The estimation of manganese as the manganese-manganic 
 oxide Mn 3 O 4 , has been so frequently criticised unfavorably 
 that the method may be said to have passed from very general 
 use excepting in certain cases in which the directness of the 
 process is a temptation to incur the risk of some uncertainty. 
 The production of the other oxides of manganese in definite 
 condition is thought to be even more uncertain. Manganese 
 dioxide, MnO 2 , begins, as Wright and Menke have shown * to 
 
 * Jour. Chem. Soc., xxx, 775. 
 
SULPHATE AND AS THE OXIDES. 81 
 
 lose oxygen at a temperature (about 210 C.) to which the 
 hydrated oxide must be heated to free it from water, or very 
 nearly that at which the nitrate is converted into the dioxide ; 
 so that the chance of producing an undecomposed dioxide by 
 the ignition of the hydrated dioxide (the form in which the 
 dioxide generally appears in analytical processes), or of the 
 nitrate, is small. Manganic oxide, Mn 2 O 8 , is produced, it is 
 said, from the other oxides by ignition at a low red heat under 
 the ordinary conditions. The manganoso-manganic oxide, 
 Mn 3 O 4 , forms, presumably, when an oxide of manganese is 
 submitted, under ordinary atmospheric conditions, to the high 
 heat of the blast>lamp. If the proportion of oxygen in the 
 surrounding atmosphere is reduced below the normal, the 
 conversion of Mn 2 O 3 to Mn 3 O 4 goes on very easily, as Dittmar 
 has shown,* at a temperature between the melting points of 
 silver and aluminum, while if the proportion of oxygen in the 
 surrounding atmosphere rises much above the normal, the 
 reverse change, from Mn 3 O 4 to Mn 2 O 3 tends to take place at 
 the same temperature. It is not surprising, in view of these 
 phenomena, that the estimation of manganese as the oxide 
 Mn 3 O 4 should have fallen into disrepute; and yet, if the 
 condition most favorable to the production of that oxide a 
 low proportion of oxygen in the surrounding air can be 
 maintained during the ignition, it is not impossible that the 
 indications of the process might prove to be, under the con- 
 ditions, reasonably accurate. Now, this may be exactly the 
 condition of affairs when the ignition takes place ordinarily ; 
 for, if the products of combustion displace the ordinary air 
 about the crucible, the proportion of oxygen about the oxide 
 falls to a low limit. We have made the experiment of 
 enclosing the ignited crucible within an inverted crucible, so 
 that the products of combustion should be held immediately 
 about and above the ignited oxide, but our experience has 
 shown that the object in view is attained, apparently, quite as 
 well when the ignition is so arranged that the crucible simply 
 rests well within the upper part of the flame of a strong Bunsen 
 
 * Jour. Chem. Soc., xvii, 294. 
 VOL. ii. 6 
 
82 ESTIMATION OF MANGANESE AS THE 
 
 burner, or blast-lamp, in such manner that an oxidizing flame 
 covers nearly the entire wall of the crucible. 
 
 In the following experiments we have put to the test this 
 matter of getting definitely the different oxides of manganese. 
 We started with a known amount of pure anhydrous sulphate, 
 prepared from the pure chloride in the manner previously 
 described. This sulphate was converted by ignition into the 
 oxide presumably the oxide Mn 3 O 4 the containing crucible 
 being well within the upper flame of a powerful burner. 
 
 In the next step, this oxide was further oxidized by 
 moistening it with nitric acid and heating the residue gently 
 until the evolution of fumes ceased, the containing crucible 
 being placed well above a porcelain crucible used as a radiator 
 and heated so that only the bottom showed a faint red heat. 
 In this process the attempt was made to arrest the ignition 
 at the point where the anhydrous dioxide was produced. As 
 the table shows, and as would be expected, this attempt was 
 only occasionally and partly successful. 
 
 The residue of the last process was then submitted to a 
 higher heat. The platinum crucible containing the oxide was 
 placed within and touching the bottom of a larger porcelain 
 crucible which was heated to redness. Under these conditions 
 the temperature should not be too hot, and the products of 
 combustion should naturally be thrown so far away from the 
 oxide undergoing ignition that circumstances should be 
 favorable for the formation of the oxide Mn 2 O 3 . The event 
 proved that the attainment of the exact condition corresponding 
 to the symbol Mn 2 O 3 is a matter of some uncertainty. 
 
 Next, the oxide was subjected to the highest heat of a strong 
 Bunsen burner (or in some cases, the broad flame of a blast 
 lamp), the crucible being well surrounded by the products of 
 combustion. The results of this treatment, it will be seen, 
 agree, with a single exception out of ten experiments, reason- 
 ably well with the theory for Mn 8 O 4 . 
 
 By treating the final oxide with nitric acid and repeating 
 the cycle of operations described, the observations of the 
 phenomena were multiplied, until, finally, the oxide formed 
 
SULPHATE AND AS THE OXIDES. 
 
 83 
 
 I 4 
 
 o'oo 
 
 CO CO CO 
 
 ooo 
 
 4 I 
 
 OOO 
 
 II II 
 
 oo oo 
 
 li 
 
 rH t~ <M OS 
 OO 00 GO GO 
 O OOO 
 
 t O T* COT* 
 
 O O O* O" O 
 
 1-1 O OO <N U3 
 
 ,4444- J^^ 
 
 ' C^ O CO O O T ( i I 
 
 'oooo ooo 
 
 oooo ooo 
 
 88 888 
 
 oo ooo 
 
 (M t- 
 
 CO CO CO CO CO 
 
 CO CO CO CO CO 
 
 o' o ooo 
 
 CCOOSO COCDOS 
 
 4 I 4- 
 
 oooooooo 
 
 ^H CO <M O (N 7^ t 
 
 o o o' o' o o o 
 
 I 
 
 M CO CD Oi O CO CD 
 
84 ESTIMATION OF MANGANESE, ETC. 
 
 last was treated with sulphuric acid, ignited in the manner 
 previously detailed, and weighed as the anhydrous sulphate, 
 thus showing that no significant loss of material had taken 
 place in the series of manipulations. The table comprises the 
 results of these experiments. The numbers in parentheses 
 indicate the order of treatment. 
 
 The inference is plain that the estimation of the manganese 
 in the form of the manganoso-manganic oxide, Mn 8 O 4 , is by no 
 means to be considered utterly untrustworthy when the process 
 is conducted in the manner described, though it must be 
 recognized that an irregular result may occur occasionally. 
 The danger of accepting such an irregularity as a correct 
 indication may be eliminated to a very considerable extent if 
 the precaution is taken invariably to moisten the ignited oxide 
 with nitric acid, and ignite again. The indications of 
 harmonious results thus got may be taken with a fair degree 
 of confidence. However, it is, in our judgment, by far the 
 wiser and simpler plan to convert an oxide of manganese 
 obtained in course of analysis into the sulphate and to weigh 
 the manganese in that form. 
 
XII 
 
 ON THE CONDITION OF OXIDATION OF 
 
 MANGANESE PRECIPITATED BY THE 
 
 CHLORATE PROCESS. 
 
 BY F. A. GOOCH AND MARTHA AUSTIN.* 
 
 HANNAY,| wno was the fr" 8 * ^ propose the precipitation of 
 manganese from its solution in nitric acid by the use of 
 potassium chlorate, states that precipitation is complete, but 
 that the oxide produced is not of constant composition. 
 While, therefore, precipitation by this method serves an 
 excellent purpose in separating manganese from other sub- 
 stances, it was Hannay's opinion that reliance cannot be placed 
 upon the determination of the oxygen value of the oxide to 
 estimate the manganese. Beilstein and Jawein,f who proposed 
 the same method, subsequently, regarded the precipitate as 
 the oxide MnO 2 . Hannay's reaction was developed, inde- 
 pendently, by Hampe and Ford|| into the method which is 
 now known as the " chlorate process" for the estimation of 
 manganese. The discussion of the exact condition of the 
 precipitated oxide was very active ten years ago, and occasional 
 echoes of it are heard at the present day ; and yet, in all this 
 discussion we find no account of an adequate test of the 
 process upon an exact amount of a salt of manganese known 
 to be pure. The discussion for the most part has centered 
 about the degree of oxidation of the precipitate, but there is 
 obviously another condition to be taken into account, viz. : 
 the possibility of the mechanical inclusion of the comparatively 
 insoluble chlorate in the precipitated oxide. As to the 
 existence of the latter source of error we have had in the 
 
 * From Am. Jour. Sci., v, 260. f Jour. Chem. Soc., vol. xxxiii, 269. 
 
 t Ber. Dtsch. chem. Ges., xii, 1528. Chem. Central-Blatt, 1885, 714. 
 || Trans, last. Am. Min. Eng., ix, 347. 
 
86 
 
 CONDITION OF OXIDATION OF MANGANESE 
 
 course of our work very strong affirmative evidence, the 
 apparent condition of oxidation of the precipitate being 
 sometimes so high as to be otherwise inexplicable. This 
 difficulty does not occur, however, when a more soluble 
 chlorate is chosen to do the work of oxidation, and we have 
 found quite as convenient and much safer the substitution of 
 sodium chlorate for the comparatively insoluble potassium 
 chlorate. Besides, the rapidity with which the sodium 
 chlorate is decomposed makes its use an advantage. 
 
 With regard to the completeness of the precipitation our 
 experience has shown that with due precaution the method is 
 practically perfect. Thus, after boiling manganous nitrate 
 (free from chlorides and sulphates) with strong nitric acid 
 (85 cm 3 ) and sodium chlorate (5 grm.) for five minutes, 
 adding, subsequently, 15 cm 3 of the nitric acid and a few 
 crystals more of the sodium chlorate, and discontinuing the 
 heating as soon as the liquid again boils, the insolubility of the 
 manganese is so great that no more than insignificant traces 
 may be recovered from the filtrate after cooling, filtering upon 
 asbestos and washing with water. The test for manganese in 
 the filtrate and washings was made after evaporation and 
 solution of the residue in distilled water by treating the hot 
 solution with bromine and ammonia. In the first division of 
 the table below are results obtained by treating the manganese 
 
 TABLE I. 
 
 MnSO 
 
 taken. 
 
 Mn found by KI 
 treatment in 
 filtrate. 
 
 Mn found by 
 As 2 O 3 treatment 
 in filtrate. 
 
 grm. 
 0.3361 
 0.3361 
 0.3361 
 0.3361 
 0.3361 
 0.3361 
 0.3361 
 
 grm. 
 
 None. 
 None. 
 0.00006 
 0.00005 
 0.00002 
 0.00008 
 None. 
 
 grm. 
 
 0.4128 
 0.4128 
 0.4128 
 0.4128 
 
 '..'.'. 
 
 000003 
 000003 
 Trace. 
 Trace. 
 
PRECIPITATED BY THE CHLORATE PROCESS. 87 
 
 precipitated from the filtrate with potassium iodide and 
 sulphuric acid, the iodine set free being determined by sodium 
 thiosulphate. In the second series the precipitated manganese 
 dioxide was reduced by a known amount of decinormal 
 arsenious acid and the amount remaining unoxidized was 
 estimated by titration with iodine in the presence of acid 
 potassium carbonate. 
 
 It will be seen that in no case did the manganese which 
 escaped precipitation that which corresponded to the iodine 
 freed or the arsenious acid oxidized exceed 0.0001 grm. 
 Plainly this modified method of handling the chlorate process 
 may be trusted to precipitate the manganese with gratifying 
 rapidity and approximation to completeness. Our experience 
 has shown plainly that prolonged boiling results in a considera- 
 ble loss of manganese (from 0.0010 0.0030 grm.). This we 
 think is due to the solvent effect of the lower oxides of nitro- 
 gen naturally produced (as is always the case in boiling nitric 
 acid) after the chlorine dioxide has been thoroughly expelled. 
 An excess of the chlorate at the end of the boiling seems to be 
 essential and a slight yellow color in the solution, due to chlo- 
 rine dioxide, is a favorable indication rather than the reverse. 
 We find it best to filter the undiluted nitric acid, under pressure, 
 upon asbestos on a perforated cone with a filtering surface of 
 about 40 cm 2 . The dilution of the nitric acid before filtra- 
 tion tends to produce some solubility of the manganese, and 
 the loss then introduced, though trifling if the filtration is 
 rapid, may be considerable if the process of filtration is pro- 
 longed, as is the case in the method approved by the " Verein 
 der deutschen Eisenhiitteleute." * 
 
 Our experiments upon the chlorate process have been made 
 with manganous chloride prepared as detailed in a former 
 paper, viz.: by boiling manganous chloride with manganous 
 carbonate, precipitating the filtered solution with ammonium 
 sulphide, dissolving the washed manganous sulphide in dilute 
 hydrochloric acid, precipitating the solution thus obtained with 
 sodium carbonate (after boiling out hydrogen sulphide), dis- 
 
 * Von Reis, Zeitschr. angew. Chem., 1891, 376. 
 
88 CONDITION OF OXIDATION OF MANGANESE 
 
 solving the greater part of the manganous carbonate (thor- 
 oughly washed by repeatedly boiling it in successive portions 
 of water) in the least amount of hydrochloric acid, and boiling 
 the solution thus obtained with the remainder of the pure 
 carbonate and filtering. The standard of the solution thus 
 prepared, neutral and probably very pure, was fixed by evapo- 
 rating definite portions with sulphuric acid and weighing 
 the residue as the normal sulphate in accordance with the pro- 
 cedure outlined in a former paper.* 
 
 Any method, by means of which the oxidizing power of the 
 higher oxygen compounds of manganese is discoverable, may, 
 obviously, be employed to determine the condition of the 
 manganese precipitated in this chlorate process. Convenient 
 processes for the determination of the available oxygen in the 
 higher oxides of manganese are the iodometric methods of 
 Bunsen and Pickering. Bunsen's method is applicable to any 
 of the higher oxygen compounds of manganese though some- 
 what inconvenient because it involves the distillation of the 
 chlorine liberated by the action of strong hydrochloric acid 
 upon the substance and its collection in potassium iodide, the 
 iodine thus set free being estimated by standard thiosulphate. 
 According to Pickering's! method the higher oxide is treated 
 immediately with potassium iodide and hydrochloric acid and 
 the iodine liberated is estimated by sodium thiosulphate. 
 Plainly, the latter method is limited to the treatment of the 
 less refractory or more finely comminuted oxides, and it fails 
 in the presence of ferric salts and all other substances capable 
 of liberating iodine from the acidified iodide. 
 
 Still another general iodometric method for determining the 
 oxygen value of the higher oxides of manganese is suggested 
 by Deshayes's titration of permanganic acid in nitric acid by 
 means of standard arsenious acid.J Our experience in follow- 
 ing out this idea shows that the precipitated oxides of manga- 
 nese, as well as the soluble permanganate, may be easily 
 reduced with the aid of gentle heating by arsenious acid in 
 
 * Am. Jour. Sci., v, 209. This volume, p. 77. 
 
 t Jour. Chem. Soc., xxxvii, 128. \ Bull. Soc. Chim., xxix, 541. 
 
PRECIPITATED BY THE CHLORATE PROCESS. 89 
 
 the presence of sulphuric acid, and that the determination of 
 the excess of arsenious acid by titration with iodine after neu- 
 tralization of the free sulphuric acid by an alkaline carbonate 
 gives exact data for estimating the oxidizing power of the 
 manganese compound. We found, however, that if the iodine 
 is allowed to come into contact with the manganous carbonate 
 thrown down by the alkaline carbonate, as is inevitable 
 for at least short intervals during the titration of the 
 arsenious acid in presence of the floating carbonate, the 
 danger arises of more or less reoxidation of the manganous 
 carbonate by the iodine and the consequent introduction of 
 error. Fortunately the difficulty may be obviated by adding 
 to the solution, while still acid, enough tartaric acid or alkaline 
 tartrate to prevent the precipitation of the manganese in the 
 subsequent neutralization by the bicarbonate. 
 
 Confining our attention to the last two simpler iodometric 
 methods the reduction of the higher oxide by an acidified 
 iodide on the one hand and by arsenious acid on the other 
 we made, first, some experiments to determine the accuracy 
 with which manganese may be thus estimated. We used 
 for the manganese compound of known oxidizing power a 
 solution of potassium permanganate filtered carefully through 
 asbestos and standardized against ammonium oxalate which 
 had been found to be the exact equivalent of a specially 
 prepared lead oxalate. For each experiment a definite por- 
 tion of this solution was drawn from a burette and treated 
 with a solution of pure manganous sulphate until the color 
 of the permanganate had vanished, thus precipitating a 
 hydrous oxide approximating quite closely probably to the 
 condition of oxidation of the dioxide, but containing at all 
 events, whatever its actual composition might be, the exact 
 amount of available oxygen originally in the permanganate. 
 In the experiments of the following table this precipitate 
 was treated with a solution of potassium iodide (6 grm.) and 
 tartaric acid (10 grm.), by which the freshly prepared hydrate 
 is dissolved quite as well as by the iodide and hydrochloric 
 acid of Pickering's original method and with less risk of 
 
90 
 
 CONDITION OF OXIDATION OF MANGANESE 
 
 evolution of iodine outside the main reaction. From the 
 iodine found by titration with this sulphate we have cal- 
 culated the weight of manganese dioxide which would liberate 
 it; and a comparison of this value with the amount of the 
 dioxide theoretically precipitated by the interaction of the 
 known permanganate and the sulphate, upon the assumption 
 that two molecules of the former throw down five molecules 
 of the hydrated dioxide, should disclose the error of the 
 analytical process when applied to the estimation of man- 
 ganese dioxide. In all probability the assumption that it 
 is the dioxide that is precipitated, and which afterwards acts 
 upon the iodide, is not quite true under the conditions, since 
 the precipitation takes place in presence of an excess of a 
 manganous salt ; but for our purpose it does not matter, since 
 we are in effect simply dealing with the oxidizing power of 
 a known amount of permanganate. 
 
 TABLE II. 
 
 Mn theoretically 
 precipitated by 
 KMnO 4 as MnO 2 . 
 
 Mn in MnO 2 
 corresponding 
 to iodine found. 
 
 Error of the analytical 
 process (applied to 
 MuO 2 ) in terms of Mn. 
 
 grm. 
 0.1351 
 0.1351 
 0.1351 
 0.1351 
 0.1351 
 0.1351 
 
 gnu. 
 0.1347 
 0.1347 
 0.1350 
 0.1353 
 0.1358 
 0.1353 
 
 grin. 
 
 0.0004- 
 0.0004- 
 0.0001- 
 0.0002+ 
 0.0007+ 
 0.0002+ 
 
 Thus, it is obvious that the mean error of the results 
 is practically inconsiderable, varying between extremes of 
 0.0004 grm. and + 0.0007 on 0.1351 grm. of manganese 
 dioxide. 
 
 In the experiments of Table III the precipitated oxide was 
 treated with an excess of a standard arsenious acid solution 
 and 5 cm 3 of sulphuric acid of half strength, and the whole 
 was heated until the manganese dioxide was dissolved. To 
 this liquid was added tartaric acid (10 grm.) to prevent the 
 precipitation of the manganese and the oxidation by iodine in 
 
PRECIPITATED BY THE CHLORATE PROCESS. 91 
 
 the subsequent titration, the acid was neutralized by acid 
 potassium carbonate, and the arsenic still remaining in the 
 arsenious condition titrated by standard iodine. 
 
 TABLE IIL 
 
 Mn precipitated by 
 action of KMnO 4 
 on MnSO 4 as MnO 2 . 
 
 MninMnO 2 
 corresponding to 
 As-jOg oxidized. 
 
 Error of the process 
 in terms of Mn. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 0.1392 
 
 0.1396 
 
 0.0004- 
 
 0.1109 
 
 0.1117 
 
 0.0008- 
 
 0.1112 
 
 0.1117 
 
 0.0005- 
 
 0.1109 
 
 0.1117 
 
 0.0008- 
 
 0.1109 
 
 0.1117 
 
 0.0008- 
 
 0.1117 
 
 0.1125 
 
 0.0008- 
 
 It is clear that either of these methods of reduction, the 
 action of an acidified iodide or that of arsenious acid, is 
 capable of yielding fairly accurate indications when we have 
 to deal with a pure salt of manganese. When, however, the 
 manganese is associated with a considerable amount of iron, 
 as is frequently the case, it becomes a matter of necessity to 
 separate the manganese before attempting its estimation. 
 For this purpose the " chlorate process " is by far the simplest 
 of those generally used, and though it has been the subject 
 of much discussion, it is at present the method of separation 
 most widely used by practical chemists, whether the final 
 estimation of the manganese is made gravimetrically as in 
 Ford's process, or volumetrically, as in the methods of 
 Volhard, Williams, or Pattinson. 
 
 Definite portions of the solution of pure manganous chloride 
 were drawn from a burette into an Erlenmeyer flask of 300 cm 3 
 capacity, evaporated to dryness, precipitated by the " chlorate 
 process " with the modifications given in detail above. The 
 oxide, after careful washing, was returned with the asbestos 
 to the flask and treated by one or other of the methods to 
 be described. It was either treated with potassium iodide 
 (5 grm.) and sulphuric acid (10 cm 3 ) of half strength, the 
 iodine set free being estimated by thiosulphate ; or it was 
 
92 
 
 CONDITION OF OXIDATION OF MANGANESE 
 
 heated with an excess of standard arsenious acid and 10 cm 3 
 of sulphuric acid of half strength, and after cooling, adding 
 5 grm. of the Rochelle salt and neutralizing with acid 
 potassium carbonate, the arsenious acid remaining unoxidized 
 was estimated with standard iodine. In Table IV are given 
 the results obtained in this work. 
 
 TABLE IV. 
 
 Mn taken in 
 the form of 
 manganous 
 chloride. 
 
 Mn found upon 
 the hypothesis 
 that MnOi is 
 precipitated. 
 
 Error. 
 
 Mn found in the nitrate 
 after evaporation and 
 treatment with bromine 
 and ammonia. 
 
 BY REDUCTION WITH POTASSIUM IODIDE. 
 
 grin.. 
 
 0.1225 
 0.1225 
 0.1225 
 0.1225 
 
 grm. 
 0.1183 
 0.1177 
 0.1180 
 0.1169 
 
 grill. 
 
 0.0042- 
 0.0048- 
 0.0045- 
 0.0056- 
 
 grm. 
 0.00006 
 Trace. 
 0.00008 
 Trace. 
 
 BY REDUCTION WITH ARSENIOUS ACID. 
 
 0.1222 
 0.1222 
 0.1222 
 0.1222 
 0.1222 
 0.1222 
 0.1222 
 0.1222 
 0.1222 
 
 0.1189 
 0.1191 
 0.1199 
 0.1200 
 0.1186 
 0.1187 
 0.1189 
 0.1194 
 0.1205 
 
 0.0033- 
 0.0031- 
 0.0023- 
 0.0022- 
 0.0036- 
 0.0035- 
 0.0033- 
 0.0028- 
 0.0017- 
 
 Not determined. 
 Not determined. 
 Not determined. 
 Not determined. 
 None. 
 0.0001 
 0.0002 
 Trace. 
 0.0001 
 
 The results show plainly that, while the manganese is so 
 completely precipitated in the chlorate process of oxidation 
 when properly conducted that only insignificant traces may 
 escape, the condition of oxidation cannot be taken to be 
 that of the dioxide. The average error thus put upon the 
 determination of the manganese known to be present is more 
 than 2 per cent. It follows, as a matter of course, that the 
 indications of any process which rests upon the assumption 
 that the oxygen value of the manganese compound preci- 
 pitated in the chlorate process corresponds to that of the 
 dioxide must of necessity be erroneous. If, therefore, the 
 chlorate method is to be employed for the separation of 
 
PRECIPITATED BY THE CHLORATE PROCESS. 93 
 
 the manganese, it is obvious that precautions must be taken 
 to secure a definite condition of oxidation of the manganese 
 before processes which depend upon the oxygen value of 
 the higher oxide may be applied to the estimation of that 
 element. The process which in our hands seems to give 
 the oxide in definite condition is based upon the observations 
 of Wright and Menke* that a dilute solution of potassium 
 permanganate acting in excess, at 80 C., hi the presence of 
 zinc sulphate, and in thorough mixture upon manganous 
 sulphate, yields an oxide which, though combined with alkali, 
 holds the oxygen exactly in the proportion corresponding to 
 the dioxide. Three-fifths of the manganese in such a pre- 
 cipitate represents the amount of that element originally 
 present in the manganous salt. In the following table are 
 given the results of experiments in which manganese was 
 determined iodometrically after the interpolation of the 
 permanganate treatment. 
 
 In these experiments a solution of manganous chloride of 
 known strength was drawn from a burette, evaporated to 
 dryness in a small beaker, heated with nitric acid until there 
 was no evidence of the presence of nitrogen oxides. Strong 
 nitric acid was poured in until the volume was 85 cm 3 , sodium 
 chlorate (5 grm.) was added carefully, the liquid was boiled 
 five minutes, more nitric acid (15 cm 3 ) and a few crystals of 
 the chlorate were introduced, and the solution brought to 
 boiling temperature again. After cooling, the liquid was 
 filtered on asbestos and washed with water, and the oxide 
 upon the asbestos and walls of the beaker was dissolved in 
 2 cm 3 of hydrochloric acid. After diluting a little the 
 solution was evaporated with 5 cm 8 of strong sulphuric acid 
 until no more hydrochloric acid remained. The solution of 
 manganous sulphate (not exceeding 0.5 grm. of the salt), 
 very nearly neutralized by potassium carbonate, was mixed 
 with a solution of zinc sulphate (2 grm.) and a freshly and 
 carefully filtered dilute solution of potassium permanganate 
 (1.5 grm. of the salt) ; the liquid, amounting now to about 
 * Jour. Chem. Soc.,xxxvii, 36. 
 
94 
 
 CONDITION OF OXIDATION OF MANGANESE 
 
 500 cm 3 , was heated to 80 C., and acid potassium carbonate 
 added, in quantity a little more than enough to neutralize the 
 remnant of the acid present. The precipitate was collected 
 upon asbestos, and after careful washing was returned to the 
 flask in which the precipitation had been made. The oxygen 
 value of the oxide was determined by one or other of the 
 methods described. In the one case the flask was fitted with 
 a paraffined stopper having two bores, one holding a Will 
 and Varrentrapp absorption apparatus (in which a solution of 
 potassium iodide dissolved any escaping iodine), the other a 
 small separating funnel. Sulphuric acid and potassium iodide 
 in solution were run in through the funnel, the iodine set 
 free was titrated with thiosulphate the amount of man- 
 ganese being reckoned from the iodine set free. The results 
 of this work follow in the first part of Table V. In the 
 second case the dioxide obtained in the manner described above 
 was reduced by warming gently with a decinormal solution 
 of arsenious acid. After cooling, and neutralizing with acid 
 
 TABLE V. 
 
 Mn taken 
 in the 
 form of chloride. 
 
 Mn found upon 
 the hypothesis that 
 MnO 2 is the oxide 
 finally obtained. 
 
 Error. 
 
 BY REDUCTION WITH POTASSIUM IODIDE. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 0.0643 
 
 0.0637 
 
 0.0006- 
 
 0.0643 
 
 0.0642 
 
 0.0001- 
 
 0.0643 
 
 0.0642 
 
 0.0001- 
 
 0.0651 
 
 0.0651 
 
 0.0000 
 
 0.1125 
 
 0.1121 
 
 0.0004- 
 
 0.1125 
 
 0.1121 
 
 0.0004- 
 
 0.1125 
 
 0.1120 
 
 0.0005- 
 
 0.1214 
 
 0.1206 
 
 0.0008- 
 
 0.1214 
 
 0.1207 
 
 0.0007 
 
 0.1214 
 
 0.1223 
 
 0.0009+ 
 
 0.1214 
 
 0.1214 
 
 0.0000 
 
 BY REDUCTION WITH ARSENIOUS OXIDE. 
 
 0.1213 
 
 0.1212 
 
 0.0001- 
 
 0.1213 
 
 0.1201 
 
 0.0012- 
 
 0.1213 
 
 0.1203 . 
 
 0.0010- 
 
 0.1213 
 
 0.1208 
 
 0.0005- 
 
PRECIPITATED BY THE CHLORATE PROCESS. 95 
 
 potassium carbonate in the presence of Rochelle salt, the 
 excess of the arsenious acid was estimated with iodine in the 
 presence of starch. The estimation by this method gave 
 the results recorded hi the second part of Table V. 
 
 These results show plainly, that if the precautions to which 
 attention has been directed are taken, viz.: dilution of the 
 solution and heating to 80 C., presence of zinc sulphate, 
 and (most essential of all) the almost complete neutralization 
 of free acid before the addition of the potassium permanganate, 
 the manganese dioxide precipitated by the chlorate process 
 from pure manganous nitrate may subsequently, after reduc- 
 tion, be brought by the permanganate treatment so nearly to 
 the full degree of oxidation represented by the symbol MnO 2 , 
 that the amount of manganese originally treated may be 
 calculated with a very fair degree of accuracy from the 
 oxygen value of three-fifths of the oxide found. We do not 
 recommend this procedure as a rapid analytical method ; our 
 purpose is accomplished when the fact is brought plainly to 
 view that the oxide precipitated by the chlorate process is 
 not the dioxide, but that it may be made such by subsequent 
 treatment. 
 
XIII 
 
 ON THE ESTIMATION OF MANGANESE 
 SEPARATED AS THE CARBONATE. 
 
 BY MAKTHA AUSTIN * 
 
 THE estimation of manganese precipitated as the manganous 
 carbonate, when that salt is obtained by the action of sodium 
 or potassium carbonate, has been regarded as very undesirable 
 for the reasons that, even if the conditions of the precipitate 
 is such that it does not run through the filter, the manganous 
 carbonate can never be freed entirely from alkaline salt, and 
 that the conversion of the carbonate to the manganoso-manganic 
 oxide the form in which it is customary to weigh is too 
 uncertain. It had been supposed, also, that the presence of 
 ammoniacal salts (as well as of carbonic acid) causes solution 
 of the manganous carbonate, until the work of Guyard (Hugo 
 Tamm) f showed that when the precipitation is accomplished 
 by ammonium carbonate, even hi the presence of ammonium 
 chloride, complete separation of the manganese is possible. 
 No data are given by Guyard to show the completeness of the 
 separation of the manganese by this process ; but Fresenius J 
 examined the method and speaks favorably of it. In this 
 process the main difficulty of the older method of estimation 
 as the carbonate viz., the inclusion of the alkaline salt is 
 avoided. We know now how to avoid the difficulty in the 
 way of weighing as the oxide by converting that substance to 
 the form of the sulphate, as shown in a former paper. 
 
 For a careful study of the separation of manganous carbonate 
 by Guyard's method a solution of pure manganous chloride 
 was prepared and standardized as the anhydrous sulphate in 
 
 * From Am. Jour. Sci., v, 382. t Chem. News, xxvi, 37. 
 
 J Zeitschr. anal. Chem., xi, 290. 
 
 Am. Jour. Sci., v, 209. This volume, p. 77. 
 
ON THE ESTIMATION OF MANGANESE. 
 
 97 
 
 the manner detailed in the paper to which reference has been 
 made above. A definite volume of the manganous chloride 
 was carefully drawn into a platinum dish and diluted to a 
 volume of 200 cm 3 . To the solution heated to 100 C. ammo- 
 nium chloride (about 10 grm.) was added and ammonium car- 
 bonate in excess. The solution was kept warm until the 
 precipitate subsided, and then was filtered off on asbestos on 
 a perforated crucible under pressure. The presence of ammo- 
 nium chloride is necessary to insure such a condition of the 
 precipitate that it will not run through the felt. 
 
 Inasmuch as the precipitate was collected under conditions 
 which readily permit an attempt to weigh as the carbonate, a 
 trial of that method was made incidentally. The event proved, 
 as Rose * has stated previously, that when the carbonate is 
 gently heated, evolution of carbon dioxide and oxidation of 
 the residue begins before the water is thoroughly removed; 
 for, though nearly all the results are above the theory, the 
 solution of the residue in hydrochloric acid indicated plainly 
 the presence of a small amount of a higher oxide of manganese. 
 In the following table are found the results of a series of 
 experiments in which the attempt was made to weigh first 
 as carbonate and again after strong ignition well within the 
 oxidizing flame of a powerful burner f as the manganoso- 
 manganic oxide. The application of the bromine test to the 
 hot ammoniacal filtrate showed that in every one of these 
 experiments, the precipitation of the manganese in the form 
 of the carbonate had been complete. 
 
 
 
 MnC0 3 
 
 Mn 3 4 
 
 MnCl z . 
 
 MH 4 C1. 
 
 
 
 
 
 
 
 
 
 
 
 Found. 
 
 Theory. 
 
 Error. 
 
 Found. 
 
 Theory. 
 
 Error. 
 
 cms 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 50 
 
 10 
 
 0.2685 
 
 0.2680 
 
 0.0005+ 
 
 0.1770 
 
 0.1776 
 
 0.0006- 
 
 50 
 
 10 
 
 0.2704 
 
 
 0.0024+ 
 
 0.1788 
 
 
 0.0012+ 
 
 50 
 
 10 
 
 0.2710 
 
 
 0.0030+ 
 
 0.1770 
 
 
 0.0006- 
 
 50 
 
 10 
 
 0.2720 
 
 
 0.0040+ 
 
 0.1774 
 
 
 0.0002- 
 
 * Annal Phys. Chem., Ixxxiv, 52. 
 
 t Am. Jour. Sci., v, 209. This volume, p. 82. 
 
 VOL. II. 7 
 
98 
 
 ON THE ESTIMATION OF MANGANESE 
 
 As shown in this table, weighing as the carbonate is out of 
 the question; the errors of the process when the residue is 
 ignited in the manner described to form the manganoso- 
 manganic oxide are much smaller though rather variable. 
 The estimation of manganese as the anhydrous sulphate had 
 given in the work to which reference has been made results 
 agreeing so much more closely than could be obtained by 
 any other procedure, that the attempt was made to estimate 
 the amount of manganese precipitated as the manganese 
 carbonate by converting it first to the oxide, then to the 
 sulphate. A given weight of sulphate was precipitated as 
 the manganous carbonate, after the employment of all the 
 precautions mentioned previously in this paper, and then 
 filtered off on ashless filter paper. After washing thoroughly 
 with hot water, the filter was burned, the residue ignited for 
 the condition of the manganoso-manganic oxide and weighed 
 as such. Then the oxide was converted to the sulphate by 
 heating with three or four drops of concentrated sulphuric 
 acid. The agreement of the results as shown in the following 
 table is considerably better. 
 
 Exp. 
 
 MH 4 C1. 
 
 Mn 3 4 
 
 MnSO 4 
 
 Found. 
 
 Theory. 
 
 Error. 
 
 Found. 
 
 Theory. 
 
 Error. 
 
 Ill 
 
 18 
 
 grm. 
 
 10 
 10 
 10 
 10 
 
 grm. 
 
 0.2463 
 0.1110 
 0.1584 
 0.1672 
 
 grm. 
 0.2478 
 0.1121 
 0.1581 
 0.1699 
 
 grim* 
 0.0015- 
 0.0011- 
 0.0003+ 
 0.0027- 
 
 grm. 
 0.4903 
 0.2225 
 0.3126 
 0.3355 
 
 grm. 
 0.4905 
 0.2219 
 0.3128 
 0.3364 
 
 grm. 
 0.0002- 
 0.0006+ 
 0.0002- 
 0.0009- 
 
 By treatment of the filtrates of (1), (2), and (3) with 
 bromine and ammonia at boiling temperature no manganese 
 was found. In the filtrate from number (4) by the same 
 treatment, a small amount of manganese dioxide was precipi- 
 tated, which when heated with concentrated sulphuric acid 
 gave 0.0006 grm. of manganous sulphate; and, hence, the 
 error in that determination is really 0.0003 grm. on the 
 
SEPARATED AS THE CARBONATE. 99 
 
 sulphate. The slightly larger deficiency recorded in the table 
 was probably due to imperfect filtering. 
 
 It seems to be evident that Guyard's method of separation 
 of manganese as the manganous carbonate, when handled with 
 precautions, gives complete separation of that element. It 
 must be recognized clearly, however, that the precipitation 
 should be made in the presence of a considerable amount of 
 ammonium chloride, and that great care must be used in the 
 filtering and washing of the finely divided precipitate. It is 
 altogether preferable to weigh in the condition of the sulphate. 
 
XIV 
 
 THE ACTION OF CARBON DIOXIDE ON 
 SOLUBLE BORATES. 
 
 BY LOUIS CLEVELAND JONES* 
 
 IN a process for the separation and estimation of boric acid 
 devised by Morse and Burton, f the liberation of carbonic, 
 silicic, and boric acids from a mixture of inorganic salts is 
 effected by the action of sulphuric acid, using tropseolin 
 as an indicator of acidity. In the solution thus prepared, con- 
 taining in free condition only carbonic, silicic and boric acids, 
 the silicic acid is dehydrated and made insoluble by anhydrous 
 copper sulphate. The boric acid is then extracted with abso- 
 lute alcohol. To this alcoholic solution of boric acid, a known 
 amount of barium hydroxide solution is added in excess over 
 that required to form a barium metaborate, BaB 2 O 4 . Carbon 
 dioxide is then passed into the solution in accordance with the 
 hypothesis that the excess only of barium is acted upon. The 
 aqueous mixture of barium metaborate and barium carbonate 
 is evaporated and the residue is heated to a constant weight 
 over a triple burner. From the following proportion the boric 
 acid present may be calculated. The molecular weight of 
 boric acid the molecular weight of carbon dioxide : the 
 molecular weight of boric acid = the total weight found the 
 theoretical weight of barium as carbonate : the weight of boric 
 acid present. It is obvious, inasmuch as the difference be- 
 tween the calculated weight of the barium as carbonate and 
 the actual weight of the residue is multiplied nearly three times 
 to get the boric oxide, that the actual error of the process, 
 whatever it may be, is magnified threefold by the method of 
 computation. 
 
 * From Am. Jour. Sci., v, 442. t Am. Chem. Jour., x, 154. 
 
ACTION OF CARBON DIOXIDE ON BORATES. 101 
 
 I have made a study of this method applied to pure boric 
 acid, but have been unable to obtain results similar to those of 
 Morse and Burton. 
 
 For this investigation boric acid of standard strength was 
 made by dissolving in a given amount of water a known 
 weight of anhydrous boric oxide, prepared by igniting over a 
 blast lamp boric acid several tunes recrystallized and washed. 
 A solution of barium hydroxide was filtered free from carbon- 
 ate and then standardized by precipitation as carbonate and 
 also by the Phelps method with iodine.* To a measured 
 amount of the boric acid solution an excess of barium hydrox- 
 ide was added, carbon dioxide passed, and the whole evapo- 
 rated and by successive ignitions brought to a constant weight. 
 Below are tabulated some of the results : 
 
 Exp. 
 
 Ba(OH) 8 
 
 taken. 
 Calculated 
 as BaCO,. 
 
 B 2 3 
 
 taken. 
 
 Weight of residue 
 after successive 
 ignitions. 
 
 B 2 3 
 found. 
 
 Error. 
 
 
 grin,. 
 
 grin. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.9391 
 
 0.2200 
 
 ( 1st wt., 0.9860 
 } 2d " 0.9786 
 
 0.1263 
 0.1063 
 
 0.0937- 
 0.1137- 
 
 
 
 
 ( 1st " 0.9605 
 
 0.0744 
 
 0.0523- 
 
 (2) 
 
 0.9318 
 
 0.1295 
 
 1 2d " 0.9558 
 
 0.0646 
 
 0.0649- 
 
 
 
 
 ( 3d " 0.9510 
 
 0.0517 
 
 0.0778- 
 
 
 
 
 
 1st " 1.0357 
 
 0.2972 
 
 0.0780+ 
 
 
 
 
 
 2d " 1.0248 
 
 0.2679 
 
 0.0487+ 
 
 (3) 
 
 0.9253 
 
 0.2192 
 
 - 
 
 8d " 1.0149 
 4th " 1.0064 
 
 0.2412 
 0.2183 
 
 0.0220+ 
 0.0009- 
 
 
 
 
 
 5th " 0.9975 
 
 0.1944 
 
 0.0248- 
 
 
 
 
 
 6th " 0.9855 
 
 0.1621 
 
 0.0671- 
 
 
 
 
 
 1st " 0.8017 
 
 0.1927 
 
 0.1117+ 
 
 
 
 
 
 2d " 0.7777 
 
 0.1281 
 
 0.0471+ 
 
 
 
 
 
 3d " 0.7642 
 
 0.0918 
 
 0.0108+ 
 
 
 
 
 
 4th " 0.7582 
 
 0.0757 
 
 0.0053- 
 
 (4) 
 
 0.7301 
 
 0.0810 
 
 - 
 
 5th " 0.7517 
 6th " 0.7482 
 
 0.0582 
 0.0487 
 
 0.0228- 
 0.0323- 
 
 
 
 
 
 7th " 0.7447 
 
 0.0393 
 
 0.0417- 
 
 
 
 
 
 8th " 0.7427 
 
 0.0339 
 
 0.0471- 
 
 
 
 
 
 9th " 0.7422 
 
 0.0326 
 
 0.0484- 
 
 
 
 
 
 10th " 0.7407 
 
 0.0285 
 
 0.0525- 
 
 Plainly the results vary with the degree of the ignition. 
 At the outset the residue may or may not weigh more than 
 the theory requires for the known amounts of barium hydrox- 
 * Am. Jour. Sci., ii, 70. Volume I, p. 369. 
 
102 ACTION OP CARBON DIOXIDE 
 
 ide and boric acid taken upon the assumption that the residue 
 is barium metaborate and barium carbonate. This is obviously 
 natural if the carbon dioxide acts upon the barium borate as 
 well as upon the excess of barium hydroxide ; for, it is to be 
 expected that in the evaporation more or less of the free boric 
 acid will volatilize, and that in the subsequent ignition the 
 boric acid remaining will tend to recombine more or less com- 
 pletely, replacing carbon dioxide. If the boric acid present 
 were to recombine completely with the barium carbonate to 
 form a metaborate, the final result would always be low by 
 just the amount of free boric acid volatilized in the process 
 of evaporation and ignition. The evidence of an experiment, 
 however, in which 0.25 grm. of previously prepared barium 
 metaborate was fused in contact with 0.5 grm. of barium car- 
 bonate, resulting in a loss of 0.0871 grm., goes to show that 
 the metaborate and carbonate of barium interact still further 
 to liberate carbon dioxide. 
 
 These results were so surprising in the light of the experi- 
 ence of Morse and Burton that the question of the possibility 
 of breaking up by carbon dioxide the barium metaborate 
 already formed was put to the test directly. A known amount 
 of barium hydroxide was taken in solution and to it added an 
 amount of boric acid very little in excess of that theoretically 
 necessary to form the barium metaborate. The solution was 
 evaporated to dryness and the residue ignited. The weight 
 obtained proved to be 0.0008 grm. less than the sum of the 
 barium and boric oxides taken, doubtless because the slight 
 excess of boric acid was somewhat volatile in the evaporation. 
 The mass, presumably barium metaborate, was now dissolved 
 as completely as possible in hot water, carbon dioxide was 
 passed through the solution, the whole was evaporated, and 
 the residue ignited and weighed. The increase in the weight 
 showed that carbon dioxide had been absorbed, while a corre- 
 sponding amount of boric acid had not volatilized. 
 
 grm. K rm - P rm - grm. 
 
 (5) 0.8377 0.2990 0.9491 0.9771 
 
OAT SOLUBLE BORATES. 
 
 103 
 
 After passing in carbon dioxide and igniting the increase in 
 weight was 0.0280 grm., representing the gas absorbed less 
 the boric acid volatilized. 
 
 It was plain that barium metaborate is decomposed in solu- 
 tion by carbon dioxide. The possibility remained, however, 
 that the action of carbon dioxide might be so regulated as to 
 leave the metaborate practically unattacked. In experiment 
 (6), therefore, carbon dioxide was passed above the stirred 
 solution until no further precipitate formed upon the surface, 
 the barium present being in excess of that required to form a 
 metaborate. 
 
 Ba(OH) 2 taken. ,, n . . Weight of residue 
 Calculated as BaC0 3 . B *s taken ' after ignition. 
 
 (6) 
 
 grm. 
 
 0.7094 
 
 grin. 
 
 0.2070 
 
 grm. 
 
 0.7710 
 
 BjO 8 found, 
 grm. 
 
 0.1658 
 
 Error. 
 
 grm. 
 0.0412- 
 
 The variation of this result from the theory shows that 
 under these conditions the metaborate is not unaffected by 
 carbon dioxide, the loss being due, of course, to the escape of 
 boric acid. 
 
 An attempt was now made therefore to gauge the amount 
 of the carbon dioxide introduced by means of an indicator. 
 In experiment (7) phenolphthalein was added to the solution 
 of boric acid containing an excess of barium hydroxide and 
 the current of gas was stopped when the color of the indi- 
 cator disappeared. 
 
 Exp. 
 
 
 
 Calculated 
 
 B 2 S 
 taken. 
 
 Weight of residue 
 after ignition. 
 
 found. 
 
 Error. 
 
 
 as BaCO 3 . 
 
 
 
 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grill. 
 
 grm. 
 
 
 
 
 
 1st wt., 0.6820 
 
 0.2011 
 
 0.0438+ 
 
 
 
 
 
 2d " 0.6783 
 
 0.1911 
 
 0.0338+ 
 
 
 
 
 
 3d " 0.6720 
 
 0.1742 
 
 0.0169+ 
 
 (7) 
 
 0.6073 
 
 0.1573 
 
 . 
 
 4th " 0.6700 
 
 0.1688 
 
 0.0115+ 
 
 
 
 
 
 5th " 0.6655 
 
 0.1567 
 
 0.0006- 
 
 
 
 
 
 6th " 0.6630 
 
 0.1499 
 
 0.0074- 
 
 
 
 
 
 7th " 0.6609 
 
 0.1443 
 
 0.0130- 
 
 This result is manifestly an improvement over those obtained 
 without the careful restriction of the supply of carbon dioxide. 
 
104 ACTION OF CARBON DIOXIDE 
 
 A similar experiment, differing only in the single point that 
 the carbon dioxide was made to act upon a boiling solution, 
 resulted in like manner. 
 
 In the light of these observations it is plain that a sufficiently 
 prolonged action of carbon dioxide should result in the 
 displacement of all the boric acid if that acid can be removed 
 from the field of action as fast as it is liberated. Experiments 
 were made which clearly demonstrate the truth of this 
 hypothesis. A small side-necked flask was charged with a 
 solution of boric acid (0.1143 grm.) and barium hydroxide 
 (0.3227 grm.) in proportion to form the metaborate. The 
 mass was brought nearly to dryness by distillation and methyl 
 alcohol (15 cm 3 ) added. Through this flask, in which the 
 alcohol was kept boiling by a Bunsen burner, was passed the 
 vapor of methyl alcohol, while carbon dioxide, purified by a 
 neutral solution of silver nitrate, bubbled continually through 
 the entire system. The methyl alcohol vapor coming from the 
 side-neck flask was kept lighted by contact with the flame of 
 a Bunsen burner and the distillation continued for two hours 
 until the flame showed not the slightest tinge of green. The 
 residue in the flask originally containing barium metaborate 
 was brought to dryness and tested for boric acid. Only a 
 trace was found, and this was thought to be due to inclusion 
 by the insoluble barium carbonate. In a similar experiment 
 in which borax was used instead of barium borate, no trace of 
 boric acid was found in the residue of sodium carbonate, either 
 by turmeric or the flame test, while the distillation was 
 continued only one-half the time of the preceding experiment. 
 This result is quite in harmony with the views of P. 
 George vie* to the effect that the large absorption of carbon 
 dioxide in solutions of borax indicates that the boric acid is 
 displaced from its union with the base. 
 
 Obviously, the division of the base between boric acid and 
 carbonic acid falls under the principle of mass action, and if 
 the boric acid is taken in sufficient excess over the barium 
 hydroxide, the action of the carbonic acid should be inappreci- 
 
 * Jour. prak. Chem. xxxviii, 118. 
 
ON SOLUBLE BORATES. 105 
 
 able. This idea is sustained by an experiment in which about 
 1 grm. of boric acid was dissolved with 0.15 grm. of barium 
 hydroxide. No precipitate could be obtained by passing 
 carbon dioxide and boiling. In fact, in the French process of 
 borax manufacture just this action of an excess of boric acid 
 upon a boiling solution of sodium carbonate is used. On the 
 other hand to prevent the formation of a carbonate this excess 
 of boric acid must be considerable if the action of carbon 
 dioxide is prolonged. Thus in an experiment in which a 
 current of carbon dioxide was passed into a solution containing 
 0.1219 grm. of boric anhydride and one-half the amount of 
 barium needed to form a metaborate, the solution (60 cm 3 ) 
 deposited on boiling 90 per cent of the barium present in the 
 form of the carbonate. 
 
 In view of the facts which I have described, it is difficult to 
 see under what conditions Morse and Burton prevented the 
 excessive action of carbon dioxide and obtained hi their 
 analytical method the excellent results which they record. 
 
XV 
 
 FURTHER SEPARATIONS OF ALUMINUM BY 
 HYDROCHLORIC ACID. 
 
 BY FRAME STUART HAVENS.* 
 
 IN former papersf from this laboratory methods have been 
 described for the separation of aluminum from ferric iron, and 
 from beryllium, based on the insolubility of the hydrous 
 aluminum chloride in a mixture of equal parts of aqueous 
 hydrochloric acid and ether saturated with hydrochloric acid 
 gas, the ferric and beryllium chlorides being exceedingly 
 soluble in this mixture. 
 
 It was the purpose of the investigations herein described to 
 discover how far this process could be extended, with certain 
 modifications, to cover the separation of aluminum from such 
 other metals as might occur with it, either in artificially 
 prepared alloys or in naturally-occurring compounds. 
 
 The aluminum used in all the following experiments was in 
 the form of a solution of the chloride. This chloride was 
 purified, as previously described, J from iron by precipitation 
 with hydrochloric acid, and from the alkalies by precipitation 
 as the hydroxide and continued washing with water until the 
 washings gave no test with silver nitrate. The hydroxide 
 thus obtained was dissolved in hot hydrochloric acid to get it 
 into the form of the chloride. The chloride solution was 
 standardized by precipitating weighed portions with ammonia 
 and weighing as the oxide. 
 
 * From the Am. Jour. Sci., vi, 45. 
 
 t Gooch and Havens, Am. Jour. Sci., vol. ii, p. 416. This volume, p. 20 ; 
 Havens, Am. Jour. Sci., vol., iv, p. 111. This volume, p. 47. 
 J Loc. cit. 
 
FURTHER SEPARATIONS OF ALUMINUM, ETC. 107 
 
 Separation of Aluminum from Zinc. 
 
 A solution of pure zinc chloride made by dissolving metallic 
 zinc, free from impurities, in hydrochloric acid, is not 
 precipitated when treated with an equal volume of ether 
 and saturated with hydrochloric acid gas. 
 
 To prepare a definite zinc salt free from traces of the alkalies 
 which would be precipitated with the aluminum by strong 
 hydrochloric acid, pure metallic zinc was dissolved in hydro- 
 chloric acid, the dilute solution precipitated with ammonium 
 carbonate, and the resulting carbonate ignited to a constant 
 weight as zinc oxide. This oxide dissolved in hydrochloric 
 acid gave a pure chloride. 
 
 The aluminum in all these experiments was determined in 
 the following manner: Portions of the prepared solution of 
 aluminum chloride were weighed in a small beaker, weighed 
 portions of zinc oxide added and sufficient aqueous hydrochloric 
 acid to dissolve it. The beaker was then cooled by immersion 
 in an inverted bell-jar supplied with running water by means 
 of inlet and outlet tubes, and a current of gaseous hydrochloric 
 acid (generated by the gradual addition of sulphuric acid to a 
 mixture of hydrochloric acid and salt) passed through the 
 solution in the beaker to complete saturation. Ether was 
 added in volume equal to that of the original solution, and the 
 whole again saturated with hydrochloric acid gas. The 
 crystalline chloride precipitated was caught on asbestos in a 
 filter crucible, washed with a previously prepared solution of 
 equal parts of ether and hydrochloric acid, saturated with 
 hydrochloric acid gas, dried for half an hour at 150-180 C., 
 covered with a layer of pure mercury oxide, heated gently 
 over a low flame under a ventilating hood, ignited over the 
 blast, and weighed as the oxide. The results show that 
 aluminum can be determined with reasonable accuracy in the 
 presence of a pure zinc salt. 
 
 The zinc can be determined, after the evaporation of the 
 strong acid filtrate, by any of the well known methods. It 
 was found, however, that after thorough conversion to the 
 
108 
 
 FURTHER SEPARATIONS OF ALUMINUM 
 
 nitrate by repeated evaporation with nitric acid the salt could 
 be ignited directly to the oxide with satisfactory results. This 
 is shown clearly in Table I, (3) to (5). In experiments 
 (3) and (4), zinc oxide was dissolved in nitric acid and the 
 nitrate ignited again to the oxide. In experiment (5), the zinc 
 oxide was first dissolved in hydrochloric acid and the chloride 
 thus obtained was converted to the nitrate by evaporating the 
 solution (5 cm 3 ) with nitric acid (2 cm 3 ), treating the residue 
 with nitric acid (2 cm 3 ) and evaporating to dryness. 
 
 TABLE I. 
 
 Bxp. 
 
 A1 2 O 3 taken 
 as the 
 chloride. 
 
 found. 
 
 Error. 
 
 ZnO 
 taken. 
 
 ZnO 
 found. 
 
 Error. 
 
 Error 
 corrected. 
 
 Final, 
 vol. 
 
 
 gnu. 
 
 grin. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 cm 8 
 
 (1) 
 
 0.0562 
 
 0.0562 
 
 0.0000 
 
 0.1110 
 
 . 
 
 > 
 
 . 
 
 
 (2) 
 
 0.0580 
 
 0.0577 
 
 0.0003- 
 
 0.1034 
 
 _ 
 
 
 t 
 
 t 
 
 (3) 
 
 f 
 
 . 
 
 . 
 
 0.1019 
 
 0.1016 
 
 0.0003- 
 
 t 
 
 
 
 (4) 
 
 , . 
 
 . . 
 
 . . 
 
 0.1010 
 
 0.1007 
 
 0.0003- 
 
 . . 
 
 , 
 
 (5) 
 
 
 . . 
 
 . . 
 
 0.1100 
 
 0.1095 
 
 0.0005- 
 
 
 
 (6) 
 
 0.0572 
 
 0.0572 
 
 0.0000 
 
 0.1014 
 
 0.1027 
 
 0.0013+ 
 
 0.0007- 
 
 12 
 
 (7) 
 
 0.0563 
 
 0.0550 
 
 0.0013- 
 
 0.1026 
 
 0.1038 
 
 0.0012+ 
 
 0.0008- 
 
 16 
 
 8) 
 
 0.0577 
 
 0.0576 
 
 0.0001- 
 
 0.1000 
 
 0.1014 
 
 0.0014+ 
 
 0.0006- 
 
 16 
 
 (9) 
 
 0.0559 
 
 0.0558 
 
 0.0001- 
 
 0.1020 
 
 01035 
 
 0.0015+ 
 
 0.0005- 
 
 16 
 
 (10) 
 
 0.0563 
 
 0.0556 
 
 0.0007- 
 
 0.2024 
 
 0.2046 
 
 0.0022+ 
 
 0.0002+ 
 
 20 
 
 (11) 
 
 0.1111 
 
 0.1107 
 
 0.0004- 
 
 0.2092 
 
 0.2116 
 
 0.0024+ 
 
 0.0004+ 
 
 20 
 
 In Table I, (6) to (11), are given the results of experiments 
 in which both the aluminum and zinc were determined, the 
 former, as described, by precipitating as the hydrous chloride 
 and weighing as the oxide, and the latter by carefully evapo- 
 rating the strongly acid filtrate (best with a small current of 
 air playing on the surface of the liquid to avoid spattering due 
 to the too violent evolution of the ether and gaseous acid) and 
 finally converting the chloride through the nitrate into the 
 oxide. It is, of course, absolutely necessary that the treatment 
 with nitric acid shall be thorough, so that no zinc chloride may 
 remain to volatilize when the residue is ignited. On account 
 of the danger to platinum from the aqua regia generated by 
 the action of nitric acid on zinc chloride, the evaporations of 
 the filtrates from the aluminum chloride and the treatment 
 
BY HYDROCHLORIC ACID. 
 
 109 
 
 with nitric acid were carried on in porcelain and the residual 
 nitrate was transferred to a small crucible for ignition. In 
 this process the porcelain was evidently attacked somewhat, 
 so that the residual nitrate was slightly contaminated with 
 material from the large porcelain dish. This fact accounts 
 for the high results given in the first column of errors. 
 However, on introducing a correction (0.0020) found by 
 carrying through the process in blank with the quantities of 
 reagents employed in the regular process, the results on zinc, 
 slightly deficient, agree closely with those obtained in (3)-(5), 
 Table I, where the zinc nitrate was converted directly to 
 the oxide without the previous evaporation in porcelain of 
 a large volume of strongly acid liquid. The errors thus 
 corrected stand in another column of the table. 
 
 These results show clearly that aluminum and zinc may be 
 separated from one another by the action of hydrochloric acid 
 gas in aqueous ethereal solution with a reasonable degree of 
 accuracy. 
 
 Separation of Aluminum from Copper, Mercury, and Bismuth. 
 
 The separation of aluminum from copper, mercury, and 
 bismuth does not differ materially from the separation of 
 
 TABLE II. 
 
 Exp. 
 
 A1 2 O 8 taken 
 as the 
 chloride. 
 
 found. 
 
 Error. 
 
 CuO 
 
 taken. 
 
 CuO 
 found. 
 
 Error. 
 
 HgCl, 
 taken. 
 
 Bi 2 3 
 
 taken. 
 
 gj 
 
 (3) 
 
 18 
 
 (6) 
 
 gnu. 
 
 0.0576 
 0.0561 
 0.0570 
 0.0548 
 0.0565 
 0.0576 
 
 gnn. 
 0.0571 
 0.0557 
 0.0574 
 0.0557 
 0.0571 
 0.0577 
 
 gnn. 
 0.0005- 
 0.0004- 
 0.0004+ 
 0.0009+ 
 0.0006+ 
 0.0001+ 
 
 gnn. 
 0.0500 
 0.0400 
 
 gnu. 
 
 gnn. 
 
 gTIH. 
 
 0.1000 
 0.1000 
 
 grin. 
 
 o.iobo 
 
 0.2000 
 
 J3 
 
 (9) 
 10) 
 
 11) 
 12) 
 13) 
 
 0.0558 
 0.0538 
 0.0566 
 0.0577 
 
 0.0545 
 0.0536 
 0.0562 
 0.0575 
 
 0.0013- 
 0.0002- 
 0.0004- 
 0.0002- 
 
 0.0437 
 0.0359 
 0.0345 
 0.0319 
 0.0343 
 0.0337 
 0.0651 
 
 0.0432 
 0.0359 
 0.0340 
 0.0324 
 0.0356 
 0.0349 
 0.0644 
 
 0.0005- 
 0.0000 
 0.0005- 
 0.0005+ 
 0.0013+ 
 0.0012+ 
 0.0007- 
 
 
 
110 FURTHER SEPARATION OF ALUMINUM, ETC. 
 
 aluminum and zinc. Aluminum chloride is precipitated 
 quantitatively in the presence of pure salts of these elements 
 as shown in experiments of Table II. 
 
 In determining the copper in the acid filtrate it was found 
 advantageous to weigh as the oxide, but to arrive at that 
 condition through the sulphate rather than through the nitrate 
 (which was the transition salt in the case of zinc), as this 
 process can be carried on safely in platinum. 
 
 In Table II, (10)-(13), are given results of experiments in 
 which the aluminum was determined as previously described 
 by precipitation as the hydrous chloride and conversion to the 
 oxide. The acid nitrate was evaporated in platinum and the 
 copper determined by treating the residue with a few drops of 
 strong sulphuric acid, heating gently to drive off the excess of 
 sulphuric acid, and then igniting the sulphate to the oxide at 
 a red heat. That the copper sulphate is converted to the 
 oxide by ignition at a red heat over a Bunsen burner is shown 
 in experiments (7) to (9) of Table II. 
 
XVI 
 
 THE IODOMETRIC DETERMINATION OF 
 MOLYBDENUM. 
 
 BY F. A. GOOCH AND JOHN T. NORTON JR.* 
 
 A PROCESS for the iodometric determination of molybdic acid, 
 which consists in treating a soluble molybdate in a Bunsen 
 distillation-apparatus with potassium iodide and hydrochloric 
 acid, has been advocated by Friedheim and Euler.f Accord- 
 ing to this process the molybdate, containing from 0.2 grm. 
 to 0.3 grm. of molybdenum trioxide, is treated with from 
 0.5 grm. to 0.75 grm. of potassium iodide and enough hydro- 
 chloric acid, of sp. gr. 1.12, to fill two-thirds of the flask of 
 the apparatus. The liquid is warmed until heavy vapors of 
 iodine fill the flask and then boiled until iodine vapor is no 
 longer visible and the color of the liquid residue is a clear 
 green. The iodine liberated is collected in the distillate 
 and titrated with sodium thiosulphate, every atom of iodine 
 found indicating presumably the reduction of a molecule of 
 molybdic acid to the condition of the pentoxide Mo 2 O 5 . 
 
 It was pointed out in a former article from this laboratory, t 
 that greater precaution than was taken by Friedheim and 
 Euler is necessary in order that the reduction may proceed 
 according to theory, and that the iodine collected may serve 
 as a reliable measure of the molybdic acid. It was found 
 that the green color of the liquid comes gradually and that 
 it may develop distinctly before the molybdic acid is fully 
 reduced. It was found, also, that since even a trace of 
 oxygen liberates iodine from the hot mixture of potassium 
 
 * From Am. Jour. Sci., vi, 168. 
 
 t Ber. Dtsch. chem. Ges., xxviii, 2066. 
 
 t Gooch and Fairbanks, Am. Jour. Sci., ii, 156. Volume I, p. 375. 
 
112 THE IODOMETRIC DETERMINATION 
 
 iodide and hydrochloric acid of the strength employed, it is 
 not sufficient to rely upon the volatilization of iodine to 
 expel the air originally in the apparatus, but that it is 
 essential to conduct the distillation in an atmosphere devoid 
 of oxygen. The suggestion was made therefore that the 
 operation should be carried on in a current of carbon dioxide 
 and that the mixture, constituted definitely, should be boiled 
 between stated limits of concentration which were determined 
 by experiment. It was found that when amounts of a soluble 
 molybdate containing less than 0.3 grm. of molybdenum 
 trioxide are treated with potassium iodide, not exceeding the 
 theoretical proportion by more than 0.5 grm., and 40 cm 3 of 
 a mixture of the strongest hydrochloric acid and water in 
 equal parts, the reduction proceeds with a fair degree of 
 regularity and is practically complete when the volume has 
 diminished to 25 cm 3 . If the operation is properly conducted 
 in an atmosphere of carbon dioxide, it was shown that the 
 iodine in the distillate may be trusted to indicate the molybdic 
 acid within reasonable limits of accuracy. It appeared, how- 
 ever, that too great an excess of potassium iodide tends to 
 induce excessive reduction, and that the same tendency 
 shows when the liquid is concentrated to too low a limit. 
 
 To this criticism Friedheim took exception * and contrasted, 
 to their disadvantage, our results by the modified method 
 with those of Friedheim and Euler by the original method. 
 It became necessary, therefore, to point out f the fact that 
 of the results published by Friedheim and Euler, upon which 
 reliance was placed to prove the reliability of their method, 
 five out of seven in one series and one out of five in another 
 series had been calculated incorrectly from data given. 
 Another series of six determinations was, however, apparently 
 faultless in this respect. More recently f Euler has explained 
 that the errors were not really arithmetical. Two of them 
 may be presumed, inferentially, to be due to careless copying 
 
 * Ber. Dtsch. chem. Ges., xxix, 2981. 
 t Gooch, Am. Jour. Sci., iii, 237. 
 J Zeit. anorg. Chem., xv, 454. 
 
OF MOLYBDENUM. 113 
 
 or proof-reading; and four, we are told by Euler, were 
 introduced into the series by mistake, and actually represent 
 (as Prof. Friedheim kindly informs him) the analysis of a 
 sample of ammonium molybdate of undetermined constitution : 
 that is to say, the figures now given by Euler represent the 
 original percentages of molybdenum trioxide which had been 
 changed by some unconscious process from 
 
 80.62 per cent to 81.85 per cent. 
 80.71 " 81.69 
 
 80.63 " " 81.67 " 
 80.78 " " 81.78 
 
 Curiously enough, Euler's corrected figures, as given here, 
 are still affected by trifling arithmetical errors of from one to 
 four units in the second decimal place. The agreement of 
 these results among themselves is no proof of the correctness 
 of the process of analysis. The great variation between the 
 average percentage of molybdenum trioxide in ammonium 
 molybdate as found by Euler in a molybdate of known con- 
 stitution and the percentage of the trioxide as found by 
 Friedheim (if we understand Euler aright) may be due con- 
 ceivably to either or both of two causes, viz. : the change of 
 material analyzed, and the change of operator or conduct of 
 the operation. We shall show in the following account of 
 our work that the exact control of the conditions of treatment, 
 along the lines laid down formerly, is actually essential to the 
 reduction of molybdic acid according to the theory of the 
 process. 
 
 Our experiments were made with ammonium molybdate 
 twice recrystallized from the presumably pure salt. The con- 
 stitution of the preparation was determined by careful ignition 
 per se, and, for greater security, with sodium tungstate free 
 from carbonate. It contained 81.83 per cent of molybdenum 
 trioxide. 
 
 The potassium iodide which we used was prepared by act- 
 ing with re-sublimed iodine upon iron wire, and precipitating 
 by potassium carbonate the proportions of iodine and iron 
 
 VOL. II. 8 
 
114 
 
 THE IODOMETRIC DETERMINATION 
 
 having been adjusted to secure the formation of the hydrous 
 magnetic oxide of iron. The filtrate from the iron hydroxide 
 gave on evaporation and crystallization potassium iodide which 
 was free from iodate. 
 
 The hydrochloric acid was taken of sp. gr. 1.12, because 
 this is the strength used by Friedheim and Euler. 
 
 The sodium thiosulphate employed was taken in nearly 
 decinormal solution, and was standardized by running it into 
 an approximately decinormal solution of iodine which had 
 been determined by comparison with decinormal arsenious 
 acid made from carefully re-sublimed arsenious oxide. We 
 chose this method of standardizing the introduction of the 
 thiosulphate into the iodine rather than the reverse opera- 
 tion, in order that the conditions of the actual analysis might 
 be followed in the standardization. 
 
 The distillation apparatus was constructed with sealed or 
 ground joints of glass wherever contact with iodine was a 
 
 possibility. It was made by 
 sealing together a separating 
 funnel A, a 100 cm 3 Voit 
 flask B, a Drexel wash-bottle 
 C, and a bulbed trap g, as 
 shown in the figure. Upon 
 the side of the distillation- 
 flask B was pasted a gradu- 
 ated scale by means of which 
 the volume of the liquid 
 within the flask might be 
 known at any time. Carbon 
 dioxide, generated in a Kipp 
 apparatus by the action of 
 
 dilute hydrochloric acid (carrying in solution cuprous chloride 
 to take up free oxygen) upon marble previously boiled in 
 water, was passed through the apparatus before and during 
 the operation, so that it was possible to interrupt the process 
 of boiling at any point of concentration, to remove the receiver 
 by easy manipulation, to replace the receiver, and to continue 
 
 FIG. 20. 
 
OF MOLYBDENUM. 115 
 
 the distillation without danger of admitting air to the distilla- 
 tion flask. 
 
 In experiments to be described (1) to (5) of the table 
 the proportions of potassium iodide and molybdic acid, and 
 the strength of the hydrochloric acid recommended by Fried- 
 heim and Euler were retained. The essential change of 
 condition is the removal of atmospheric air from the distillation 
 flask before the acid is admitted to contact with the other 
 reagents. Potassium iodide (3 grm.) and water (200 cm 3 ) 
 were put into the receiver C, and a little of this solution was 
 allowed to flow into the trap g. Ammonium molybdate care- 
 fully weighed (0.3 grm.) and potassium iodide (0.5 grm. to 
 0.75 grm.) were introduced into the distillation flask B, the 
 apparatus was connected as shown in the figure, and carbon 
 dioxide was passed freely through the whole apparatus for 
 some minutes. The stop-cock d, between the bulb of the 
 funnel A and the flask B, was closed, and hydrochloric acid 
 (40 cm 3 , sp. gr. 1.12) was poured into the funnel ; the air 
 above the liquid in the funnel was displaced by carbon dioxide 
 through the space between the neck of the funnel and the 
 loosely adjusted stopper carrying the inlet tube ; the connec- 
 tion between the funnel and inlet tube was tightened, the 
 stop-cock opened, and the acid, under the pressure of carbon 
 dioxide, was permitted to flow into the flask. In this way 
 the acid, iodide, and molybdate were made to interact with 
 little danger of the presence of oxygen. The flask was heated 
 by the Bunsen burner, and the iodine evolved, passing over 
 quietly in the slow current of carbon dioxide, collected in the 
 receiver. The liquid was boiled until fumes of iodine were 
 no longer visible above the liquid in the flask and connecting 
 tubes backed by a ground of white, and then a full minute 
 more. At this stage, the green color of the liquid having 
 developed fully, the apparatus was permitted to cool, the 
 current of carbon dioxide was increased, the cap of the receiver 
 was loosened at /, the contents of the trap were washed back 
 into the receiver, the rest of the apparatus was lifted bodily 
 from the receiver, the liquid adhering to the inlet tube was 
 
116 THE IODOMETRIC DETERMINATION 
 
 washed off into the receiver, and the end of the tube was 
 dipped immediately into a solution of potassium iodide. The 
 constant flow of carbon dioxide prevented reflux of air during 
 the transfer, and as soon as the end of the tube had been 
 submerged in the solution of potassium iodide (which was 
 employed not only as a water-seal, but to catch any iodine 
 still carried in the gas), it was possible to reduce the rapidity 
 of the current. 
 
 After titrating the iodine in the distillate the receiver was 
 again placed in the train and the process of distillation was 
 resumed under the former conditions and continued until the 
 volume of the liquid, as indicated upon the scale, had dimin- 
 ished to 25 cm 3 , when the distillation was interrupted. The 
 apparatus was manipulated as before to prevent access of air, 
 and the iodine evolved in the second treatment determined. A 
 third period of distillation served to show the iodine liberated 
 during the concentration of the liquid from 25 cm 3 to 10 cm 3 . 
 
 During the first period of distillation the liquid assumed the 
 clear green color, which changed but slightly until the begin- 
 ning of the third period, when the tint verged upon olive, and 
 at the end of the operation the color of the liquid was an 
 olive brown which grew browner on cooling. The addition 
 of considerable hydrochloric acid to the residual liquid re- 
 stored the clear green color, while water changed the olive 
 brown to reddish yellow, the tint varying with the dilution. 
 The results of these experiments are recorded in (1) to (5) of 
 the accompanying table. In division A are given the weights 
 of molybdenum trioxide corresponding to the amounts of 
 iodine found in the three stages of distillation ; in division B, 
 the molybdenum trioxide corresponding to the iodine evolved 
 from the beginning of the process to the end of each stage. 
 
 The mean error of the indications taken during the period 
 of distillation advocated by Friedheim and Euler is 0.0045 
 grm. ; * that of the period of concentration from 40 cm 3 to 
 
 * Even this figure does not disclose the full error, which is partly counter- 
 balanced, as will appear later, by the effect of oxygen dissolved in the acid 
 used in the process. 
 
OF MOLYBDENUM. 
 
 11T 
 
 A. 
 
 Exp. 
 
 HCl 
 
 Sp &i 12 
 
 Klin 
 
 retort. 
 
 MoO s taken 
 as ammonium 
 molybdate. 
 
 Mo0 3 corresponding to iodine found. 
 
 First stage 
 40cm 3 to 
 32cm 3 . 
 Green color. 
 
 Second stage 
 32cm3 to 
 25 cm 3 . 
 
 Third stage 
 25cm 3 to 
 10cm 3 . 
 
 (1) 
 (2) 
 (3) 
 (4) 
 (5) 
 
 cm 8 
 40 
 40 
 40 
 40 
 40 
 
 grm. 
 
 0.5 
 0.5 
 0.5 
 0.75 
 0.75 
 
 grm. 
 0.2455 
 0.2455 
 0.2455 
 0.2455 
 0.2455 
 
 grm. 
 0.2399 
 0.2402 
 0.2414 
 0.2404 
 0.2431 
 
 grm. 
 0.0076 
 0.0053 
 0.0040 
 0.0061 
 0.0037 
 
 grm. 
 
 0.0004 
 0.0013 
 0.0004 
 0.0004 
 0.0004 
 
 (6) 
 (7) 
 
 40 
 40 
 
 1 
 
 2 
 
 0.2455 
 0.2455 
 
 0.2404 
 
 0.0085 
 
 0.0019 
 
 B. 
 
 Ezp. 
 
 MoO 3 corre- 
 sponding to 
 iodine found 
 during period 
 of Friedheim 
 and Euler. 
 (1st stage.) 
 
 Error. 
 
 MoO 3 corre- 
 sponding to 
 iodine found in 
 concentrating 
 from 40cm 3 
 to 25cm 3 . 
 
 Error. 
 
 Mo0 3 corre- 
 sponding to 
 iodine found in 
 concentrating 
 from 40cm 3 
 to 10 cm 3 . 
 
 Error. 
 
 (1) 
 (2) 
 (3) 
 (4) 
 (5) 
 
 grm. 
 
 0.2399 
 0.2402 
 0.2414 
 0.2404 
 0.2431 
 
 grin. 
 0.0056- 
 0.0053- 
 0.0041- 
 0.0051- 
 0.0024- 
 
 grin. 
 
 0.2475 
 0.2455 
 0.2454 
 0.2465 
 0.2468 
 
 gnu. 
 0.0020+ 
 0.0000 
 0.0001- 
 0.0010+ 
 0.0013+ 
 
 grm. 
 0.2479 
 0.2468 
 0.2458 
 0.2469 
 0.2472 
 
 grm. 
 0.0024+ 
 0.0013+ 
 0.0003+ 
 0.0014+ 
 0.0017+ 
 
 (6) 
 (7) 
 
 0.2404 
 
 0.0051- 
 
 0.2489 
 
 0.0034+ 
 
 0.2508 
 ( 0.2495 
 \ 0.2529* 
 
 0.0053+ 
 0.0040+ 
 0.0074+ 
 
 25 cm 3 is 0.0008 grm. + ; and that of the full period of distilla- 
 tion is 0.0014 +. It is plain beyond a peradventure that in 
 the process as conducted by Friedheim and Euler, except- 
 ing the protection against atmospheric action the theoretical 
 reduction of the molybdic acid does not take place. The best 
 results are obtained when the distillation is prolonged until 
 the original volume of 40 cm 3 has been diminished to 25 cm 8 . 
 Concentration beyond the limit of 25 cm 3 tends to develop a 
 tendency toward over-reduction, especially when the amount 
 of potassium iodide is increased beyond about 0.5 grm. in 
 
 * On repeating distillation with a fresh charge of acid. 
 
118 THE IODOMETRIC DETERMINATION 
 
 excess of that theoretically required. This is shown in exper- 
 iments (6) and (7), conducted otherwise similarly to those 
 described above, in which the amount of potassium iodide was 
 increased to 1 grm. and 2 grms. The error after distilling 
 from 40 cm 3 to 10 cm 3 , the lowest limit of the preceding ex- 
 periments, was 0.0053 grm. + and 0.0040 grm. -f, and the 
 latter error was increased to 0.0074 grm. -f on repeating 
 the distillation with a fresh portion (30 cm 3 ) of the acid. It 
 is interesting to note incidentally that in the experiment in 
 which the largest amount of iodide (2 grms.) was used the 
 solution did not take the green color at any stage of the dis- 
 tillation, probably because the large excess of iodide held the 
 free iodine and so masked the color until the degree of con- 
 centration was reached at which the olive-brown color dis- 
 places the green. 
 
 The possibility of the interaction of atmospheric oxygen 
 and gaseous hydriodic in the analytical process, even to the 
 extent of producing errors of from one to three per cent 
 reckoned as molybdenum trioxide, was recognized by Fried- 
 heim and Euter ; and it was to obviate this difficulty that the 
 recommendation was made by them to warm very gradually 
 the distillation flask filled two-thirds with the mixture of 
 iodide, molybdate, and acid, and to raise the liquid to actual 
 boiling only when the space above the liquid in the retort and 
 in the connecting tube is filled as completely as possible with 
 iodine vapor, while the liquid in the receiver begins to rise in 
 the tube. 
 
 The action of atmospheric oxygen upon the solution of hydri- 
 odic acid must, however, be also taken into account. It is a 
 familiar fact that when a considerable excess of strong hydro- 
 chloric acid is allowed to act in contact with air upon potassium 
 iodide (free from iodate) dissolved in a little water, the mix- 
 ture is colored by free iodine. The amount of iodine liberated 
 by atmospheric action is insignificant when the acid is very di- 
 lute, but is considerable when the acid is strong, and increases 
 with tune and rise in temperature, as shown in the experiments 
 recorded in the accompanying table. 
 
OF MOLYBDENUM. 
 
 119 
 
 
 
 Per- 
 
 
 
 
 
 KI 
 
 taken. 
 
 Volume. 
 
 centage 
 of HC1 
 in 
 aqueous 
 
 Time 
 in 
 minutes. 
 
 Temperature. 
 Centigrade. 
 
 MoO s 
 equivalent 
 to iodine 
 found. 
 
 Remarks. 
 
 
 
 acid. 
 
 
 
 
 
 gnu. 
 
 cm 8 
 
 
 
 
 gnu. 
 
 
 1 
 
 66 
 
 2 
 
 1 
 
 23 
 
 None. 
 
 
 1 
 
 66 
 
 2 
 
 10 
 
 23 
 
 0.0001 
 
 
 1 
 
 66 
 
 24* 
 
 10 
 
 23 
 
 0.0017 ^ 
 
 Diluted to 500 cm 3 
 
 1 
 
 1 
 
 66 
 66 
 
 24* 
 24* 
 
 4 
 10 
 
 ( From 23 to 
 J the boiling 
 ( point. 
 
 0.0067 [ 
 0.0121 ) 
 
 before titrating 
 with Na 2 S 2 8 . 
 
 Even the precaution to conduct the operation in an atmos- 
 phere of carbon dioxide does not eliminate all chance of error 
 of this sort unless the liquid of the mixture the hydro- 
 chloric acid is free from air. The experiments of the fol- 
 lowing statement, which were conducted in the apparatus and 
 manner previously described, show this point clearly. Thus, 
 
 
 
 Per- 
 
 
 TUff\f\^ 
 
 
 KI 
 
 taken. 
 
 VoL 
 
 centage 
 of HC1 in 
 aqueous 
 acid. 
 
 Concentration 
 by 
 boiling. 
 
 Jjfl.OU o 
 
 equivalent 
 to iodine 
 found. 
 
 Remarks. 
 
 grm. 
 
 cm 8 
 
 
 cm 8 . 
 
 grm. 
 
 
 
 
 
 ( 40-30 
 
 0.0013 
 
 Iodine determined 
 
 1 
 
 40 
 
 24 
 
 } 
 
 
 in distillate. 
 
 
 
 
 (30-20 
 
 0.0002 
 
 1 grm. of KI added 
 
 
 
 
 
 
 to retort at the 
 
 
 
 
 
 
 beginning of the 
 
 
 
 
 
 
 2d stage. 
 
 1 
 
 40 
 
 20 
 
 40-25 
 
 0.0005 
 
 The acid taken, 
 
 
 
 
 
 
 sp. gr. 1.1, was 
 
 
 
 
 
 
 freshly boiled 
 
 
 
 
 
 
 and introduced 
 
 
 
 
 
 
 at once upon KI 
 in retort in CO 2 . 
 
 40 cm 3 of unboiled acid, sp. gr. 1.12, introduced enough air 
 into the apparatus to cause an error of 0.0013 grm. reckoned 
 in terms of molybdenum trioxide, while the iodine set free by 
 the action of the residual acid of this experiment upon another 
 gram of potassium iodide introduced without admission of air 
 corresponded to only 0.0002 grm. in terms of molybdenum 
 
 * This corresponds nearly to sp. gr. 1.12. 
 
120 IODOMETRIC DETERMINATION OF MOLYBDENUM. 
 
 trioxide. The use of acid of sp. gr. 1.1, freshly boiled in the 
 air, obviously reduces the error due to the unboiled acid, but 
 even in this case the effect of included oxygen was not wholly 
 obviated. 
 
 It is obvious that the procedure recommended by Friedheim 
 and Euler can by no possibility eliminate the effect of atmos- 
 pheric action upon the mixture of acid and iodide. The 
 extent of such action must depend upon such conditions as the 
 size of the apparatus, the time of exposure, the body of air 
 above and dissolved in the liquid, and the rate of displacement 
 of the air. How great the error due to atmospheric action 
 actually was in the process as conducted by Friedheim and 
 Euler we, of course, have no means of knowing. It is to be 
 hoped, however, that it was sufficiently great to counterbalance 
 that other inevitable error (of about five milligrams) which 
 exists by reason of the incompleteness with which molybdic 
 acid is reduced under the conditions which these investigators 
 prescribe ; for, the value of Euler's work upon the vanadio- 
 molybdates rests upon the chance that these two very 
 considerable and indisputable tendencies to error may have 
 neutralized one another. 
 
 It has been shown clearly that our former criticism of the 
 procedure of Friedheim and Euler is justified in every par- 
 ticular. We have no change to make in the recommendation 
 made therein as to necessary modifications. 
 
 If the conditions seem difficult, there is an alternative in 
 the method proposed in the former article,* according to 
 which the molybdate is reduced by the acid and iodide in an 
 Erlenmeyer beaker (trapped loosely by means of a short 
 bulbed tube hung in the neck) and the molybdenum pentoxide, 
 freed from iodine by boiling, is reoxidized by standard iodine 
 in alkaline solution. 
 
 * Am. Jour. Sci., ii, 156. Volume I, p. 375. 
 
XVII 
 
 ON THE DETERMINATION OF MANGANESE 
 AS THE PYROPHOSPHATE. 
 
 BY F. A. GOOCH AND MAKTHA AUSTIN.* 
 
 FOR the estimation of manganese in a gravimetric way when 
 accuracy is a consideration, recourse is usually taken to the 
 excellent method of Wolcott Gibbs. f This method consists in 
 the precipitation of a manganous salt by an alkaline phosphate, 
 the conversion of the tribasic phosphate into the ammonium 
 manganese phosphate, and the weighing of the product of 
 ignition as the pyrophosphate. 
 
 By Gibbs' original method the orthophosphate of manganese 
 was precipitated by hydrogen disodium phosphate in large 
 excess above the quantity required to cause the precipitation. 
 The flocky white precipitate was dissolved either in sulphuric 
 or hydrochloric acid, and precipitated again at the boiling 
 temperature by ammonia in excess. This semi-gelatinous 
 precipitate on boiling or long standing even in the cold becomes 
 crystalline, the crystals forming beautiful talcose scales which 
 have a pearly luster and a pale rose color. The precipitate 
 was filtered off, washed with hot water, dried and ignited. The 
 results obtained by Gibbs' students for the pyrophosphate 
 accord closely with the theory. 
 
 Fresenius J showed subsequently that ammonium manganese 
 phosphate dissolves in cold water, in hot water, and in an 
 aqueous solution of ammonium chloride [1 : 70] to the extent 
 of 1 part in 32,000, 1 part in 20,000, and 1 part in 18,000, 
 respectively. It is clear, however, that the solubility of this 
 
 * From Am. Jour. Sci., vi, 233. 
 t Am. Jour. Sci., xliv, 216. 
 J Zeitschr. anal. Chem., vi, 415. 
 
122 DETERMINATION OF MANGANESE 
 
 precipitate is not indicated necessarily by the proportions 
 given so long as an excess of the precipitant is present during 
 the washing, though Fresenius did find in the filtrate traces 
 of manganese which to his mind were sufficient to account 
 for losses indicated by his test analyses, viz., one to three 
 milligrams of oxide, or from two to six milligrams of 
 phosphate. 
 
 Another mode of manipulation has been advocated by Blair * 
 in order that the precipitate may be obtained more easily in 
 crystalline condition. According to this method dilute ammonia 
 is added drop by drop to the hot acid solution until the precipi- 
 tate begins to form, the boiling and stirring are continued 
 until the small amount of flocky precipitate is converted 
 completely to crystalline condition, and the process of adding 
 ammonia drop by drop is repeated until the manganese is all 
 down in crystalline condition. The dilute ammonia is added 
 in excess and the liquid filtered after cooling in ice water. 
 
 In discussing these methods of precipitation, McKenna f 
 points out that both give good and accordant results, and that 
 the process may be carried on in glass as well as in platinum, 
 if the time of crystallization is made short enough. 
 
 When a manganous salt is precipitated in the cold by an 
 excess of an alkaline phosphate, it falls, as Heintz f has shown, 
 in the form of the trimanganous phosphate of the formula 
 Mn 3 P 2 O 8 . This same phosphate constitutes, as we have found, 
 the greater part of the precipitate which forms when a man- 
 ganous salt reacts in the cold in the presence of ammonium 
 chloride with microcosmic salt and ammonia in slight excess. 
 Boiling or even subsequent standing may, as is well known, 
 effect a more or less complete conversion of the manganese 
 phosphate to the ammonium manganese phosphate. Thus, in 
 one experiment hi which an amount of manganous chloride 
 enough to produce 0.2214 gram of the pyrophosphate was 
 precipitated in the cold by 5 cm 3 of a saturated solution of 
 microcosmic salt, with the subsequent addition of ammonia 
 
 * The Chemical Analysis of Iron, 106. t Jour. Anal. Chem., v, 141. 
 i Ann. Phys., cl, 449. 
 
AS THE PYROPHOSPHATE. 123 
 
 in excess, in a volume of 200 cm 3 containing also 5 grams of 
 ammonium chloride, the residue after ignition weighed 0.1904 
 gram. Presuming this residue to consist entirely of the 
 pyrophosphate and the trimanganous orthophosphate, the pro- 
 portion of the former to the latter calculated from the relation 
 of symbols, and the weights taken and found, is nearly one to 
 six. That is to say, about six-sevenths of the precipitate fell 
 in this experiment in the form of the tribasic orthophosphate. 
 In another experiment made exactly similarly, excepting 
 that the .liquid was heated to boiling, the proportion of the 
 manganese pyrophosphate to the trimanganous orthophosphate 
 in the only partially crystallized precipitate proved to be two 
 to one. That is, in this case, two-thirds of the precipitate was 
 in the form of the pyrophosphate. In the former of the 
 experiments a small amount of manganese was found in the 
 filtrate, but not enough to change materially the ratio recorded. 
 The slight solubility appears to be connected with the incomplete 
 conversion of the trimanganous phosphate to the ammonium 
 manganese phosphate, for as will appear later, the manganese 
 found in the nitrate, when the conversion is known to be 
 nearly complete, is inappreciable unless extraordinary amounts 
 of the ammonium salt are present. The success of the analytical 
 process under discussion turns, therefore, upon the change of 
 the trimanganous phosphate Mn 3 P 2 O 8 to the ammonium man- 
 ganese phosphate NH 4 MnPO 4 . In the work to be described 
 the attempt was made to learn the conditions under which this 
 conversion may be best and most completely accomplished. 
 
 The conversion of a molecule of trimanganous phosphate 
 to the ammonium manganese phosphate might be due, con- 
 ceivably, either to the action of free ammonia or to the action 
 of a salt of ammonium. The action of ammonia could only 
 take place at the expense of a partial loss of manganese from 
 the phosphate and its appearance as a hydroxide, two-thirds 
 of the manganese going into two molecules of the ammonium 
 manganese phosphate. In the presence of ammonium salts 
 it is possible that the manganese oxide thus replaced might 
 enter into union with the acid radical of the ammonium salt 
 
124 DETERMINATION OF MANGANESE 
 
 setting free ammonia ; but if the ammonium salt present were 
 the phosphate, or if an alkaline phosphate were present with 
 other suitable ammonium salts, it is conceivable that the 
 replaced manganese might appear as a constituent of a third 
 molecule of ammonium manganese phosphate. In any event, 
 it would be the ammonium salt and not the free ammonia 
 which would determine the formation of the third molecule 
 of the ammonium manganese phosphate. Plainly, too, the 
 ammonium salt by itself, if it were a phosphate, or if a soluble 
 phosphate were also present, might accomplish the conversion 
 without the intermediate action of free ammonia. Unless, 
 therefore, free ammonia favors the insolubility of the ammo- 
 nium manganese phosphate, its presence would be unnecessary 
 and might even be an actual disadvantage if the hydroxide 
 naturally formed by its action upon the manganese phosphate 
 were to fail to reunite fully with a phosphoric acid radical. 
 It is plain, too, that the action of free ammonia might not 
 stop with the replacement of one out of the three of the 
 manganese atoms present in the molecule, but might even 
 proceed under favorable conditions to the formation of phos- 
 phate richer in ammonium and to the separation of more 
 manganese from its union with the acid radical. As a matter 
 of fact Munroe * has shown that the prolonged action of hot 
 ammonia upon the precipitate produced by the interaction of 
 a manganous salt and an alkaline phosphate does actually 
 produce a hydroxide which blackens as it takes oxygen from 
 the air. Our attention has been given, therefore, more 
 especially to a study of the conditions of action under which 
 a salt of ammonium the chloride may bring about the 
 conversion of the precipitate first thrown down by an alkaline 
 phosphate to the form of the ammonium manganese phosphate. 
 Experiments were made upon solutions of pure manganous 
 chloride prepared and standardized by means of the sulphate 
 method, as described in a former paper, f to show the effect 
 of varying amounts of ammonium chloride on the condition 
 
 * Amer. Chemist, 1877. 
 
 t Am. Jour. ScL, v, 209. This volume, p. 77. 
 
AS THE PYROPHOSPHATE. 
 
 125 
 
 of the precipitate and upon the solubility of the precipitate 
 when once formed. The ammonium chloride for this work 
 was prepared pure by boiling the chemically pure salt of 
 commerce with a faint excess of ammonium hydrate and 
 filtering to free it from traces of iron, silica and alumina. 
 In the first series of experiments dilute ammonia was added 
 slowly to the hot faintly acidulated solution containing the 
 manganous chloride and more than enough, theoretically, of 
 a saturated solution of microcosmic salt to precipitate the 
 manganese present. The liquid was heated and stirred until 
 the flocky mass was changed to a crystalline condition. The 
 addition of ammonia drop by drop, with constant stirring and 
 heating, was continued until the manganese was all precipi- 
 tated in crystalline form. A slight excess of ammonia was 
 added and the liquid with the precipitate was allowed to 
 stand for a half hour, cooling gradually or chilled in ice 
 water. The precipitate was filtered off on asbestos under 
 pressure, washed carefully in water made faintly ammoniacal, 
 dried and ignited. The filtrates were tested for manganese 
 by treatment with bromine and heating. The results of these 
 experiments are given in the following table. 
 
 TABLE I. 
 
 Mn 2 P 2 O 7 equivalent 
 to MnCl,. 
 
 Error 
 in terms of 
 Mn 2 P 2 7 . 
 
 Error in 
 terms of 
 Manganese. 
 
 Saturated 
 solution of 
 HNH 4 NaPO 4 . 
 4H 2 0. 
 
 Total 
 volume. 
 
 Manganese 
 in nitrate. 
 
 Taken. 
 
 Found. 
 
 grm. 
 
 grm. 
 
 gnn. 
 
 grm. 
 
 cm 
 
 cm 3 
 
 
 0.4033 
 
 0.3769 
 
 0.0264- 
 
 0.0102- 
 
 5 
 
 60* 
 
 None. 
 
 0.4033 
 
 0.3728 
 
 0.0305- 
 
 0.0118- 
 
 5 
 
 60* 
 
 None. 
 
 0.3770 
 
 0.3530 
 
 0.0240- 
 
 0.0090- 
 
 5 
 
 60 
 
 None. 
 
 0.3770 
 
 0.3620 
 
 0.0150- 
 
 0.0058- 
 
 5 
 
 60 
 
 None. 
 
 0.4033 
 
 0.3751 
 
 0.0282- 
 
 0.0109- 
 
 10 
 
 60 
 
 None. 
 
 0.4033 
 
 0.3774 
 
 0.0259- 
 
 0.0100- 
 
 10 
 
 60 
 
 None. 
 
 0.4033 
 
 0.3871 
 
 0.0162- 
 
 0.0062- 
 
 5 
 
 200 
 
 None. 
 
 0.3226 
 
 03066 
 
 0.0160- 
 
 0.0062- 
 
 5 
 
 200 
 
 None. 
 
 In this method of precipitation of the manganese in a pure 
 solution of a manganous salt the results are all wrong. The 
 
 * Chilled in ice-water. 
 
126 DETERMINATION OF MANGANESE 
 
 proportion of the trimanganous phosphate to the pyrophos- 
 phate in the residue, calculated from the symbols and the 
 weights taken and found, is in the average two to five. 
 That is. to say, five-sevenths of the trimanganous phosphate 
 has been converted to the form of the ammonium manganous 
 phosphate. 
 
 The precipitate obtained in this manner is white and granu- 
 lar but not silky, and after ignition it shows the same dead 
 white color, and is powdery. Evidently the regulation of the 
 volume in which the precipitation is made is not essential, 
 and the chilling of the liquid is of no importance in changing 
 the manganese to the ammonium manganese salt under the 
 given conditions. It is plain, moreover, that the assumption 
 of a crystalline condition cannot serve as an indication that the 
 composition of the salt is ideal. It is to be noted, however, 
 that the conditions obtaining here are essentially different 
 from those hi common practice ; for, ordinarily, when man- 
 ganese is to be determined ammonium salts are abundantly 
 present as the result of previous steps in analysis. 
 
 In the experiments of the next series the conditions are 
 varied simply in this respect, that ammonium salts are in- 
 troduced before the precipitation. The precipitate was less 
 granular and more silky. After ignition the mass was white 
 with a faint rose color. In the experiments of section A of 
 the table the precipitate first thrown down was redissolved, 
 reprecipitated and filtered after cooling; in those of section 
 B, the precipitate was filtered after cooling without re-solution 
 and without reprecipitation ; and hi those of section C, the 
 first precipitate was filtered at once while the solution was 
 still hot. The length of digestion before filtering and the 
 indications of manganese in the filtrate are recorded in the 
 table. 
 
 It was observed in these experiments that when the amount 
 of ammonium chloride is present in considerable quantity a 
 fine crystalline condition is got much more readily than when 
 the amount of that salt is small: with maximum amounts of 
 ammonium chloride the change from the flocky to the crystal- 
 
AS THE PYROPHOSPHATE. 
 TABLE IL 
 
 127 
 
 Mn 2 P,O T equiva- 
 lent to the MnClj. 
 
 Error in 
 terms of 
 Mn,P,O 7 . 
 
 Error in 
 terms of 
 Mangan- 
 ese. 
 
 Saturated 
 solution of 
 HNaNH 4 P0 4 . 
 4H,0. 
 
 NH 4 C1. 
 
 Total 
 volume. 
 
 Time of 
 stand- 
 
 a. 
 
 Mangan- 
 ese in the 
 filtrate. 
 
 Taken. 
 
 Found. 
 
 A. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 cm 3 
 
 grm. 
 
 cm 3 
 
 hrs. 
 
 
 0.1542 
 
 0.1520 
 
 0.0022- 
 
 0.0008- 
 
 5 
 
 5 
 
 200 
 
 15 
 
 None. 
 
 0.1542 
 
 0.1540 
 
 0.0002- 
 
 0.0000 
 
 5 
 
 10 
 
 200 
 
 15 
 
 None. 
 
 0.1542 
 
 0.1536 
 
 0.0006- 
 
 0.0002 
 
 5 
 
 10 
 
 100 
 
 5 
 
 None. 
 
 0.1542 
 
 0.1535 
 
 0.0007- 
 
 0.0002- 
 
 5 
 
 20 
 
 200 
 
 21 
 
 None. 
 
 0.3770 
 
 0.3712 
 
 0.0058- 
 
 0.0022 
 
 5 
 
 20 
 
 200 
 
 i 
 
 None. 
 
 0.3770 
 
 0.3724 
 
 0.0046- 
 
 0.0018- 
 
 5 
 
 20 
 
 200 
 
 1 
 
 None. 
 
 0.3084 
 
 0.3069 
 
 0.0015- 
 
 0.0006- 
 
 5 
 
 40 
 
 200 
 
 1 
 
 None. 
 
 0.3084 
 
 0.3060 
 
 0.0024- 
 
 0.0009- 
 
 5 
 
 40 
 
 200 
 
 1 
 
 None. 
 
 0.3084 
 
 0.3059 
 
 0.0025- 
 
 0.0009- 
 
 5 
 
 40 
 
 200 
 
 15 
 
 Trace. 
 
 0.3084 
 
 0.3057 
 
 0.0027- 
 
 0.0010- 
 
 5 
 
 60 
 
 200 
 
 15 
 
 None. 
 
 B. 
 
 0.1542 
 
 0.1521 
 
 0.0021- 
 
 0.0008- 
 
 5 
 
 10 
 
 100 
 
 40 
 
 None. 
 
 0.1542 
 
 0.1512 
 
 0.0030- 
 
 0.0010- 
 
 5 
 
 10 
 
 200 
 
 40 
 
 None. 
 
 0.1542 
 
 0.1532 
 
 0.0010- 
 
 0.0003- 
 
 6 
 
 20 
 
 200 
 
 15 
 
 None. 
 
 0.1542 
 
 0.1531 
 
 0.0011- 
 
 0.0004- 
 
 5 
 
 20 
 
 100 
 
 15 
 
 None. 
 
 0.3770 
 
 0.3720 
 
 0.0050- 
 
 0.0019- 
 
 5 
 
 20 
 
 200 
 
 i 
 
 None. 
 
 0.3770 
 
 0.3745 
 
 0.0035- 
 
 0.0014- 
 
 5 
 
 20 
 
 200 
 
 
 
 None. 
 
 C. 
 
 0.1542 
 
 0.1519 
 
 0.0023- 
 
 0.0009- 
 
 5 
 
 16 
 
 200 
 
 
 None. 
 
 0.1542 
 
 0.1530 
 
 0.0012- 
 
 0.0004- 
 
 5 
 
 20 
 
 200 
 
 
 None. 
 
 0.1542 
 
 0.1525 
 
 0.0017- 
 
 0.0007- 
 
 5 
 
 30 
 
 200 
 
 
 None. 
 
 0.3084 
 
 0.3020 
 
 0.0064- 
 
 0.0025- 
 
 5 
 
 10 
 
 200 
 
 
 None. 
 
 0.3084 
 
 0.3053 
 
 0.0031- 
 
 0.0012- 
 
 5 
 
 20 
 
 200 
 
 
 None. 
 
 0.3084 
 
 0.3033 
 
 0.0051- 
 
 0.0020- 
 
 5 
 
 20 
 
 200 
 
 
 None. 
 
 0.3084 
 
 0.3039 
 
 0.0045- 
 
 0.0017- 
 
 5 
 
 60 
 
 200 
 
 
 Trace. 
 
 line condition is almost immediate ; even in the cold the 
 change takes place to a marked extent in a few seconds. No 
 manganese was found in the nitrate by boiling with bromine 
 and ammonia a test which is capable of indicating 0.0001 
 grm. of manganous sulphate in 500 cm 3 of water containing 
 60 grm. of ammonium chloride until the ammonium chlo- 
 ride amounted to 20 per cent of the mass, or to 40 grm. 
 in 200 cm 3 of the liquid, and even then but once in three trials : 
 even when the proportion was 30 per cent 60 grm. in 200 
 cm 3 the solvent action of the ammonium chloride upon the 
 
128 
 
 DETERMINATION OF MANGANESE 
 
 manganese salt was trifling. The pyrophosphate residues 
 obtained in these experiments, as well as in all those recorded 
 in this paper, were dissolved in nitric acid and tested for 
 contamination by a chloride ; in no single case did silver nitrate 
 produce more than an inappreciable opalescence in the solution. 
 It is plain, therefore, that the variations of the results from 
 theory are occasioned by variation in the degree of conversion 
 of the trimanganese phosphate to the ammonium manganese 
 phosphate, and that, while the ammonium chloride shows no 
 appreciable solvent action on the precipitate in the presence of 
 the precipitant, its effect in the process of conversion is plainly 
 evident. For the smaller amounts of the manganese salts 
 (equivalent to 0.1542 grm. of the pyrophosphate) the effect 
 of the ammonium chloride reaches a maximum when that salt 
 amounts to 10 per cent of the solution ; for twice that amount 
 of manganese salt, the best results were obtained by doubling 
 the amounts of ammonium chloride. Either line of treatment 
 yields under the most favorable conditions, results which are 
 passably good, but the advantage inclines slightly to the first 
 method in which the first precipitate was dissolved and 
 reprecipitated while the liquid was cooled before filtering. 
 
 TABLE IH. 
 
 Mn 2 P 2 O 7 equivalent 
 
 Error in 
 terms 
 of MN 2 P 2 O 7 . 
 
 Error 
 in terms 
 of 
 Manganese. 
 
 Saturatecl 
 solution of 
 HNaNH 4 
 P0 4 . 4H 2 0. 
 
 NH 4 C1. 
 
 Total 
 volume. 
 
 Manganese 
 in the 
 nitrate. 
 
 Taken. 
 
 Found. 
 
 grin. 
 
 grm. 
 
 grin. 
 
 grin. 
 
 cm 
 
 grm. 
 
 cm 3 
 
 
 0.2214 
 
 0.2202 
 
 0.0012- 
 
 0.0005- 
 
 5 
 
 20 
 
 200 
 
 None. 
 
 0.2214 
 
 0.2202 
 
 0.0012- 
 
 0.0005- 
 
 5 
 
 20 
 
 200 
 
 None. 
 
 0.2214 
 
 0.2191 
 
 0.0023- 
 
 0.0009- 
 
 5 
 
 20 
 
 200 
 
 None. 
 
 0.2214 
 
 0.2191 
 
 0.0023- 
 
 0.0009- 
 
 5 
 
 20 
 
 300 
 
 None. 
 
 0.2214 
 
 0.2191 
 
 0.0023- 
 
 0.0009- 
 
 5 
 
 20 
 
 300 
 
 None. 
 
 0.2214 
 
 0.2185 
 
 0.0029- 
 
 0.0011- 
 
 10 
 
 20 
 
 200 
 
 None. 
 
 0.2214 
 
 0.2186 
 
 0.0028- 
 
 0.0010- 
 
 20 
 
 20 
 
 300 
 
 None. 
 
 0.2214 
 
 0.2192 
 
 0.0022- 
 
 0.0009- 
 
 20 
 
 20 
 
 300 
 
 None. 
 
 In Table III are recorded results obtained by precipitating 
 the cold acid solution of the manganese salt and the microcos- 
 mic salt with a strong excess of ammonia. The mixture was 
 
AS THE PYROPHOSPHATE. 
 
 129 
 
 heated to boiling for from five to ten minutes and filtered hot. 
 In this series of determinations the amount of ammonium 
 chloride present was constant while the volume of the liquid 
 present was varied and the amounts of the microcosmic salt. 
 
 These results are possibly a trifle less satisfactory than those 
 obtained for the smaller amounts of manganese by the method 
 of Table II, it may be because the prolonged boiling tends to 
 form a trifling amount of free oxide ; but the fact is disclosed 
 that an increase of the microcosmic salt is without influence 
 and that a variation in volume from 200 cm 8 to 300 cm 8 is the 
 occasion of little change in the indications of the process. 
 
 In another series of experiments the solution of manganous 
 chloride was added drop by drop to the mixture of microcosmic 
 salt and ammonium chloride made alkaline with ammonia. 
 The precipitate which fell in the cold was crystallized by 
 boiling the mixture a few minutes. The results are given 
 below: 
 
 TABLE IV. 
 
 ^oVct VaIent 
 
 
 Error 
 
 Saturated 
 
 
 
 
 
 Error. 
 
 in terms 
 of 
 manganese. 
 
 solution 
 of 
 HNaNH 4 P0 4 . 
 
 NH 4 C1. 
 
 Total 
 volume. 
 
 in the 
 filtrate. 
 
 Taken. 
 
 Found. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 cms 
 
 grm. 
 
 cm.3 
 
 
 0.1542 
 
 0.1521 
 
 0.0021- 
 
 0.0008- 
 
 5 
 
 5 
 
 200 
 
 None. 
 
 0.2214 
 
 0.2203 
 
 0.0011- 
 
 0.0004- 
 
 5 
 
 10 
 
 275 
 
 None. 
 
 0.2214 
 
 0.2192 
 
 0.0022- 
 
 0.0009- 
 
 5 
 
 15 
 
 275 
 
 None. 
 
 0.2214 
 
 0.2197 
 
 0.0017- 
 
 0.0007- 
 
 5 
 
 20 
 
 275 
 
 None. 
 
 0.2214 
 
 0.2223 
 
 0.0009+ 
 
 0.0003+ 
 
 5 
 
 20 
 
 200 
 
 None. 
 
 0.1542 
 
 0.1528 
 
 0.0014- 
 
 0.0005- 
 
 5 
 
 30 
 
 275 
 
 None. 
 
 The experience of this series of experiments demonstrated 
 again that the ease with which the flocky precipitate is 
 converted to the crystalline ammonium manganese phosphate 
 is proportioned to the ammonium chloride present, and the 
 mean error of the results for the phosphate when the ammonium 
 chloride reached 20 grams (0.0007 grm.) is considerably less 
 than the mean error (0.0018 grm.) when the amount of the 
 ammonium salt was less than 20 grms. 
 
 Experiments were also made according to the modifications 
 
 VOL. II. 9 
 
130 DETERMINATION OF MANGANESE 
 
 suggested by Monroe,* viz., the boiling of the manganous salt 
 with an excess of microcosmic salt until the precipitate becomes 
 crystalline and just neutralizing with dilute ammonia ; but we 
 have been unable to find the conditions of this treatment by 
 which uniform results may be obtained in even moderate 
 agreement with the theory. 
 
 We have tried also the effect of substituting ammonium 
 nitrate for ammonium chloride in the conversion process ; but, 
 so far as our experience goes, the nitrate is not so effective 
 weight for weight in producing the change of the trimanganous 
 phosphate to the ammonium manganese phosphates, while 
 the solubility of the product in the solution of the ammonium 
 nitrate becomes appreciable more rapidly with the increase of 
 the amount present than is the case when the ammonium salt 
 is the chloride. 
 
 In the light of the experiments described it would seem to 
 be reasonable to expect the best results from the phosphate 
 method for determining manganese when the conditions are 
 so arranged that precipitation may take place in the cold 
 solution in the presence of but little free ammonia, and of 
 enough ammonium chloride to bring about the rapid conversion 
 of the precipitate to the crystalline condition. Under such 
 circumstances it should be possible to secure the conversion of 
 the phosphate to the ideal constitution as completely as possible 
 without danger of subsequent decomposition by the prolonged 
 action of the hot free ammonia. In carrying out this idea, 
 the solution of manganese chloride was treated as before with 
 microcosmic salt and a large amount of ammonium chloride, 
 the precipitate first formed was redissolved in hydrochloric 
 acid and precipitation again brought about by the very careful 
 addition of dilute ammonia in slight but distinct excess. The 
 mixture was heated only until the precipitate became silky 
 and crystalline, when it was allowed to stand and cool for a 
 half hour. The precipitate was filtered off upon asbestos in 
 a perforated platinum crucible under pressure, ignited and 
 weighed. Table V comprises the results of experiments made 
 
 * Loc. cit. 
 
AS THE PYROPHOSPHATE. 
 
 131 
 
 in this manner. In those of section A the precipitation was 
 made in platinum vessels ; in those of section B the treatment 
 was in glass. 
 
 TABLE V. 
 
 A. IN PLATINUM. 
 
 Mn 2 P 2 7 equivalent 
 to MnO 2 . 
 
 Error in 
 terms of 
 Mn,P 2 7 . 
 
 Error In 
 terms of 
 
 Mil 11 T<1 11686. 
 
 Saturated 
 solution of 
 HNaNH 4 P0 4 . 
 
 NH 4 C1. 
 
 Total 
 volume. 
 
 Manganese 
 in the 
 filtrate. 
 
 Taken. 
 
 Found. 
 
 grm. 
 0.1885 
 0.1885 
 0.1885 
 0.1885 
 0.3770 
 0.3770 
 0.3770 
 0.3770 
 
 gmi. 
 
 0.1903 
 0.1910 
 0.1913 
 0.1911 
 0.3776 
 0.3773 
 0.3778 
 0.3783 
 
 grin. 
 
 0.0018+ 
 0.0025+ 
 0.0028+ 
 0.0026+ 
 0.0006+ 
 0.0003+ 
 0.0008+ 
 0.0013+ 
 
 grm. 
 
 0.0007+ 
 0.0010+ 
 0.0011+ 
 0.0010+ 
 0.0002+ 
 0.0001+ 
 0.0003+ 
 0.0005+ 
 
 cm 8 
 5 
 5 
 5 
 5 
 5 
 5 
 5 
 5 
 
 grm. 
 20 
 20 
 20 
 20 
 20 
 20 
 20 
 20 
 
 cm 
 200 
 200 
 200 
 200 
 200 
 200 
 200 
 200 
 
 None. 
 None. 
 None. 
 None. 
 None. 
 None. 
 None. 
 None. 
 
 B. IN GLASS. 
 
 0.1885 
 0.1885 
 0.3770 
 0.3770 
 
 0.1904 
 0.1898 
 0.3767 
 0.3784 
 
 0.0019+ 
 0.0013+ 
 0.0003- 
 0.0014+ 
 
 0.0007+ 
 0.0005+ 
 0.0001- 
 0.0005+ 
 
 6 
 6 
 5 
 5 
 
 20 
 20 
 20 
 20 
 
 200 
 200 
 200 
 200 
 
 None. 
 None. 
 None. 
 None. 
 
 In this series of experiments the mean indication is, for the 
 first time, in excess of the theory. Previously the error has 
 been one of deficiency, and that in proportion to the amount 
 of manganese handled, no doubt because the amount of uncon- 
 verted trimanganese phosphate is proportioned to the entire 
 amount of the phosphate. The positive error which is devel- 
 oped in this last series of determinations is probably due to 
 the appearance of the natural error of all precipitation pro- 
 cesses viz., the tendency on the part of the precipitate to 
 include matter in solution. In the previous experiments this 
 effect was doubtless obscured by the incompleteness of the 
 conversion of the trimanganous phosphate to the ammonium 
 manganese phosphate. Indeed it is quite possible that even 
 in the last determinations the conversion is not absolute, and 
 that this is so suggested by the fact that the errors of excess 
 are larger in the case of the smaller amounts of manganese for 
 
132 DETERMINATION OF MANGANESE 
 
 which the conversion throughout the entire work has appeared 
 to be more complete. From the consideration of the results 
 tabulated and described it would seem to be obvious that not 
 only is the presence of ammonium chloride not objectionable 
 in this analytical process, which depends upon obtaining the 
 ammonium manganese phosphate from the trimanganese phos- 
 phate precipitated from a pure solution of manganese, but that 
 its presence in not too small amount, or that of a substitute, is 
 absolutely essential to make this conversion complete. For a 
 given amount of manganese and a given volume of solution 
 it seems essential that the amount of ammonium chloride 
 should reach a certain limit. According to our experience 
 the proportion of ammonium chloride to the pyrophosphate 
 should be at least 50 : 1 ; or, speaking approximately, more 
 than 200 molecules of ammonium chloride must be present in 
 the liquid (100 cm 3 or 200 cm 3 ) to every molecule of the 
 ammonium manganese phosphate to be formed. However, 
 the quantity of the ammonium salt may be increased almost 
 to the point of saturation of the liquid without causing more 
 than a trifling solubility of the ammonium manganese phos- 
 phate in the presence of an excess of the precipitant. The 
 statement of Fresenius and Munroe that ammonium manganese 
 phosphate is soluble in ammonium chloride does not hold if 
 there is an abundance of the soluble precipitating phosphate 
 present. Further, our experience goes to show that the pre- 
 cipitate may be washed with perfect safety with pure water as 
 well as with slightly ammoniacal water, or with ammoniacal 
 water containing ammonium nitrate, if the nitration is per- 
 formed rapidly and the precipitate is gathered in small space, 
 as is the case when the phosphate is collected on asbestos in 
 a perforated crucible. The finely granular precipitate which 
 may be obtained by slow action of dilute ammonia added 
 gradually to the hot solution of the manganese salt apparently 
 includes a portion of unconverted phosphate which resists the 
 replacement of the manganese by ammonium. On the other 
 hand, the precipitate of flocky condition thrown down hi the 
 cold passes easily to the silky and crystalline condition when 
 heated with the proper amount of ammonium salt and pos- 
 
AS THE PYROPHOSPHATE. 133 
 
 sesses a constitution approaching the ideal under such condi- 
 tions. The conversion of the flocky manganous phosphate is 
 so rapid that the precipitation may be carried on safely in 
 glass vessels. If the ammonium chloride in the solution were 
 to be included in the precipitate it would volatilize entirely 
 during the ignition, leaving no trace unless, possibly, a por- 
 tion of its chlorine were to combine with the manganese. 
 Tests for chlorine in the residue of pyrophosphate resulted 
 negatively no more than a mere trace being found in any 
 case, so that the contaminating effect of the ammonium chlor- 
 ide proves to be insignificant and the responsibility for the 
 increase in weight above the theory must apparently rest with 
 the included microcosmic salt. 
 
 In the practical determination of manganese by the phos- 
 phate method of Gibbs, therefore, we advocate strongly the 
 presence of large amounts of ammonium chloride. Good 
 results may be obtained by the method of precipitation origi- 
 nally laid down by Gibbs, or by the modification proposed 
 by Blair, if the ammonium salt is present in sufficient quantity. 
 On the whole trustworthy results are obtained most easily and 
 surely, according to our experience, by following the method 
 of the experiments of Table V. The slightly acid solution, 
 containing in a volume of 200 cm 3 (in platinum or glass) an 
 amount of manganese not more than enough to make 0.4 grm. 
 of the pyrophosphate, 20 grm. of ammonium chloride and 5 
 to 10 cm 3 of a cold saturated solution of microcosmic salt, is 
 precipitated in the cold by the careful addition of dilute 
 ammonia in only slight excess. The mixture is heated until 
 the precipitate becomes silky and crystalline, the whole is 
 allowed to stand and cool a half hour, the precipitate is col- 
 lected upon asbestos in a perforated platinum crucible, washed 
 (best with slightly ammoniacal water), dried at gentle heat 
 and ignited as usual. By this process determinations of the 
 larger amounts of manganese 0.4 grm. of the pyrophosphate 
 approximate rather more closely to the theoretical values 
 than do those of the smaller amounts 0.15 grm. In either 
 case the average error should not exceed 0.0010 grm. in terms 
 of manganese. 
 
XVIII 
 
 ON THE DETECTION OF SULPHIDES, SULPH- 
 ATES, SULPHITES AND THIOSULPHATES IN 
 THE PRESENCE OF EACH OTHER. 
 
 BY PHILIP E. BROWNING AND ERNEST HOWE * 
 
 SOME three years ago R. Greig Smith f published a method 
 for the detection of sulphates, sulphites and thiosulphates in 
 the presence of each other, which promised much toward the 
 solution of this most difficult problem. The method may best 
 be described in the author's own language : To a solution of 
 the salts of the above mentioned acids "barium chloride is 
 added in excess, together with a good quantity of ammonium 
 chloride, which, like many salts of ammonium, potassium and 
 calcium, acts as a flocculent or coagulant, and facilitates the 
 filtration of the barium sulphate. Hydrochloric acid is next 
 added, drop by drop, until it is evident that there is no further 
 solution of barium sulphite and thiosulphate, and that only the 
 sulphate remains undissolved; the solution is then filtered 
 through a moistened double filter paper, which should be free 
 from 'pin holes.' The filtrate will probably be clear, but if 
 not it should be returned to the filter for a second filtration. 
 When too much thiosulphuric acid is present, the clear filtrate 
 will visibly become clouded, or from being whitish will become 
 more opaque; if this occurs the solution should be thrown 
 out, and a fresh portion made more dilute. A solution of 
 iodine is added to half of the filtrate until the color is of a 
 permanent yellow tinge ; a white precipitate indicates the 
 presence of a sulphite which has been oxidized by the iodine 
 
 * From Am. Jour. Sci., vi, 317. t Chem. News, Ixxii, 39. 
 
DETECTION OF SULPHIDES, SULPHATES, ETC. 135 
 
 to sulphate. In the absence of a decided precipitate traces of 
 sulphite may easily be detected by comparing the treated and 
 untreated halves of the filtrate a procedure which very often 
 saves a good deal of time, as it is unnecessary to wait until a 
 clear filtrate is obtained. The two halves are mixed, and if 
 the yellow color disappears more iodine is added, the solution 
 filtered and the filtrate divided into two halves as before. 
 With a slight turbidity filtration may be omitted. Bromine 
 water is added to one of the halves when any thiosulphate in 
 the original solution shows itself as a white precipitate of 
 barium sulphate, readily seen on comparing the two test-tubes. 
 The thiosulphate is by iodine converted to tetrathionate, which 
 is oxidized by bromine water to sulphate." Three objections 
 to this method as described will readily occur to the reader : 
 first, the readiness with which the thiosulphate is decomposed 
 by free hydrochloric acid; second, the comparatively large 
 amount of acid necessary to effect the complete solution of the 
 barium sulphite and thiosulphate when precipitated with the 
 sulphate as compared with the amount required to prevent 
 the precipitation ; third, the lack of delicacy necessitated by a 
 comparison of portions of a colored solution in looking for 
 small precipitates. The work to be described was undertaken 
 to overcome these difficulties and to test the accuracy of a 
 modified method. Solutions of potassium sulphite and sodium 
 thiosulphate were made approximately decinorrnal and stand- 
 ardized in the usual manner against an iodine solution of 
 known value. It was found that by making a solution 
 containing sulphates, sulphites, and thiosulphates very faintly 
 acid, the sulphites and thiosulphates were held completely in 
 solution when the barium sulphate was precipitated. The 
 extreme sensitiveness of a thiosulphate to the decomposing 
 action of free hydrochloric acid suggested the possible substi- 
 tution of acetic acid to hold the sulphites and thiosulphates 
 in solution. This being a weaker acid, we hoped to avoid the 
 decomposition of the thiosulphate into sulphur and sulphurous 
 acid, or at least to delay the decomposing action. The results 
 of these experiments appear in the following table : 
 
136 DETECTION OF SULPHIDES, SULPHATES, ETC., 
 
 TABLE I. 
 
 
 
 TTWI,, , a 
 
 Hydro- 
 
 
 
 
 E 
 
 bcp. 
 
 \ oiume 
 cm 3 of 
 water. 
 
 chloric 
 acid 
 (1:4). 
 
 Acetic 
 acid. 
 
 NaAO, 
 taken. 
 
 Result. 
 
 
 
 
 drops. 
 
 drops. 
 
 grin. 
 
 
 
 
 1) 
 
 10 
 
 2 
 
 . . 
 
 0.01 
 
 No sulphur in 20 minutes. 
 
 
 
 10 
 
 2 
 
 . . 
 
 0.1 
 
 Sulphur in 45 seconds. 
 
 
 3) 
 
 100 
 
 3 
 
 . . 
 
 0.1 
 
 Sulphur in 15 minutes. 
 
 
 4) 
 
 10 
 
 . 
 
 8 
 
 0.01 
 
 No sulphur in 20 minutes. 
 
 
 5) 
 
 10 
 
 . 
 
 8 
 
 0.1 
 
 Sulphur in 90 seconds. 
 
 
 (6) 
 
 100 
 
 , 
 
 10 
 
 0.1 
 
 No sulphur in 20 minutes. 
 
 
 (7) 
 
 100 
 
 , 
 
 10 
 
 0.25 
 
 Sulphur in 15 minutes. 
 
 i 
 
 (8) 
 (9 
 
 100 
 100 
 
 ' 
 
 10 
 10 
 
 0.6 
 1.0 
 
 Sulphur in 60 seconds. 
 Sulphur in 30 seconds. 
 
 From these results it would seem that the decomposition of 
 a thiosulphate is more rapid in presence of hydrochloric acid 
 than hi presence of a much larger amount of acetic acid. 
 
 Our next experiments were directed toward a determination 
 of the effect of adding stannous chloride to bleach the color 
 of the free iodine and bromine used in the oxidation and of 
 acidifying with acetic acid, before treating with barium 
 chloride. That is to say, the process as we used it, consisted 
 in acidifying the solution to be tested with acetic acid, adding 
 barium chloride, filtering to remove precipitated sulphate 
 (always present in the sulphite), adding iodine to the nitrate 
 until the color was permanent, bleaching with stannous 
 chloride, filtering off the sulphate which represents the sul- 
 phite originally present, adding bromine in excess to the 
 filtrate and again bleaching with stannous chloride to increase 
 the visibility of the sulphate which now represents the 
 thiosulphate originally present. The details of experiments 
 in which the sulphite was taken alone and oxidized with 
 iodine are given in Table II. 
 
 A corresponding series of experiments was made in which 
 hydrochloric acid was substituted for acetic acid and essentially 
 the same results were obtained. 
 
 A similar series of experiments was made to test the effect 
 of treating the thiosulphate in an acidified solution, first with 
 
IN THE PRESENCE OF EACH OTHER. 
 
 137 
 
 TABLE H. 
 
 Brp. 
 
 KJO. 
 
 taken. 
 
 Volume 
 of 
 water. 
 
 BaS0 4 precipitated 
 after oxidation 
 with iodine. 
 
 Remarks. 
 
 (1) 
 (2) 
 (3) 
 (4) 
 (5) 
 
 gnu* 
 
 0.1 
 0.01 
 0.001 
 0.0005 
 0.0001 
 
 cm 8 
 10 
 10 
 
 10 
 10 
 10 
 
 Very abundant. 
 Abundant. 
 Distinct. 
 Fair. 
 Faint. 
 
 Plainly visible before adding SnCl 2 . 
 Plainly visible before adding SnCl 2 . 
 More distinct after adding SnCl 2 . 
 Hardly visible before adding SnCl 2 . 
 Invisible before adding SnCLj. 
 
 iodine and then after filtration (if a precipitate had formed) 
 with bromine. In the experiments of division A hydrochloric 
 acid (a few drops) was added before treating with barium 
 chloride, and in those of division B acetic acid was used 
 similarly. Stannous chloride was employed to bleach the 
 excess of iodine and bromine. 
 
 TABLE m. 
 
 
 
 
 BaS0 4 pre- 
 
 BaSO 4 precip- 
 
 
 Exp. 
 
 Na 2 S 2 O s 
 taken. 
 
 Volume 
 of water. 
 
 cipitated 
 by action 
 
 itated by 
 action 
 
 Remarks. 
 
 
 
 
 of iodine. 
 
 of bromine. 
 
 
 A. 
 
 
 grm. 
 
 cm 3 
 
 
 
 
 (1) 
 
 0.1 
 
 10 
 
 Distinct. 
 
 Abundant. 
 
 Sulphur separated in 30 seconds. 
 
 (2) 
 
 0.01 
 
 10 
 
 Faint. 
 
 Abundant. 
 
 No sulphur in 90 seconds. 
 
 (3) 
 
 0.001 
 
 10 
 
 None. 
 
 Distinct. 
 
 No sulphur in several minutes. 
 
 (4) 
 
 0.0005 
 
 10 
 
 None. 
 
 Faint. 
 
 No sulphur ; SnCl 2 necessary. 
 
 (5) 
 
 0.0001 
 
 10 
 
 None. 
 
 Very faint. 
 
 No sulphur ; SnCl 2 necessary. 
 
 B. 
 
 (1) 
 
 0.1 
 
 10 
 
 Faint. 
 
 Abundant. 
 
 No sulphur separated in 1 minute. 
 
 (2) 
 
 0.01 
 
 10 
 
 None. 
 
 Abundant. 
 
 No sulphur separated in several 
 
 
 
 
 
 
 minutes. 
 
 (3) 
 
 0.001 
 
 10 
 
 None. 
 
 Distinct. 
 
 No sulphur. 
 
 (4) 
 
 0.0005 
 
 10 
 
 None. 
 
 Faint. 
 
 No sulphur ; SnCl 2 necessary. 
 
 (5) 
 
 0.0001 
 
 10 
 
 None. 
 
 Very faint. 
 
 No sulphur ; SnCl 2 necessary. 
 
 From these experiments the advantage of the use of acetic 
 acid becomes apparent, as does also the use of stannous 
 chloride in increasing the delicacy of this indication, so that a 
 small fraction of a milligram may easily be detected. 
 
138 DETECTION OF SULPHIDES, SULPHATES, ETC., 
 
 If relatively large amounts of thiosulphate are present 
 with small amounts of sulphite, we have sometimes found it 
 advantageous to manipulate so that even the slow decom- 
 position of the thiosulphate by acetic acid may be avoided 
 by first attempting precipitation with barium chloride in a 
 dilute ammoniacal solution. By this method the barium 
 sulphate and sulphite are separated from the thiosulphate 
 and identified the sulphate by its insolubility in dilute 
 hydrochloric acid, and the sulphite by the action of iodine 
 upon the acid filtrate from the barium sulphate. After 
 filtering, the thiosulphate may be detected in the filtrate by 
 the use of iodine and bromine as described above. Table IV 
 gives some results by this treatment. 
 
 TABLE IV. 
 
 Exp. 
 
 Na 2 S 3 8 
 taken. 
 
 BaS0 4 pre- 
 cipitated by 
 iodine. 
 
 BaS0 4 pre- 
 cipitated by 
 bromine. 
 
 Remarks. 
 
 
 grm. 
 
 
 
 
 (1) 
 
 0.1 
 
 None. 
 
 Abundant. 
 
 
 (2) 
 
 0.01 
 
 None. 
 
 Good. 
 
 
 (3) 
 
 0.001 
 
 None. 
 
 Fair. 
 
 SnCl 2 necessary. 
 
 (4) 
 (5) 
 
 0.0005 
 0.0001 
 
 None. 
 None. 
 
 Faint. 
 None. 
 
 SnCl 2 necessary. 
 
 As will be seen, the test for the thiosulphate by this method 
 of treatment is not so delicate, probably on account of 
 mechanical holding of the barium thiosulphate by the pre- 
 cipitated sulphate and sulphite. 
 
 Having determined the limits of accuracy of the method as 
 applied to the sulphite and thiosulphate taken separately, our 
 next experiments were directed toward an investigation of the 
 working of the method when these two acids are found 
 together in solution. Sulphates, almost invariably present 
 with sulphites, are of course quite easily separated by filtration 
 and treating with the barium salt in acid solution. Sulphides 
 if present in the solution would seriously interfere with the 
 working of this method if not removed, being readily oxidized 
 by the iodine or bromine to sulphite, sulphate, or, should 
 
IN THE PRESENCE OF EACH OTHER. 139 
 
 sulphur also separate, to thiosulphate. We found in course 
 of our work that in attempting to neutralize a mixture of 
 freshly prepared alkaline sulphide together with a sulphite 
 we often obtained a precipitate of sulphur. After the removal 
 of the sulphide and sulphate, we were surprised to find on 
 treating with iodine scarcely a trace of sulphite. On treating 
 with bromine however an abundant indication of thiosulphate 
 was obtained. It is well known of course that thiosulphate 
 may be formed by boiling a sulphite with sulphur, but that 
 this reaction should take place so readily and completely 
 seemed to us rather unusual. 
 
 For the removal of a sulphide before proceeding with the 
 tests for sulphite and thiosulphate Greig Smith recommends 
 the passing of carbon dioxide through the solution until the 
 escaping gas gives no indication of hydrogen sulphide, but 
 Bloxam* calls attention to the tedious and wholly unsatis- 
 factory character of this method of removal and recommends 
 a mixture of zinc chloride, cadmium chloride, ammonium 
 chloride and ammonia. We have found that the addition of 
 zinc acetate to a faintly alkaline solution accomplishes the 
 same purpose in an entirely satisfactory manner. The sul- 
 phide used in our work was freshly made by passing hydrogen 
 sulphide through a dilute solution of sodium hydroxide. 
 When portions of this solution, still alkaline, were treated 
 with zinc acetate in excess, and the zinc hydroxide and 
 sulphide removed by filtration, the filtrate gave no test for 
 either sulphite or thiosulphate by the application of iodine 
 and bromine as described, and the vapor evolved on boiling 
 caused no darkening of lead paper. The following table 
 shows the results of a few experiments in which tests were 
 made for the sulphite and thiosulphate, after removing a con- 
 siderable amount of the sulphide in the manner described, and 
 of the sulphate by acidifying and adding barium chloride. 
 
 The method as we have modified it may be summarized as 
 follows : To about 0.1 grm. of the substance to be analyzed 
 dissolved in 10 cm 8 of water or more, add sodium, potassium or 
 * Chem. News, Ixxii, 63. 
 
140 DETECTION OF SULPHIDES, SULPHATES, ETC. 
 
 TABLE V. 
 
 E 2 SO 3 taken. 
 
 N^SjOj, taken. 
 
 BaS0 4 precipitated 
 after oxidation 
 with iodine. 
 
 BaS0 4 precipitated 
 after oxidation 
 with bromine. 
 
 grin* 
 
 0.1 
 0.1 
 
 0.01 
 0.001 
 0.001 
 
 grin. 
 
 0.01 
 0.001 
 0.1 
 0.1 
 0.001 
 
 Abundant. 
 Abundant. 
 Good. 
 Faint. 
 Fair. 
 
 Good. 
 Distinct. 
 Abundant. 
 Abundant. 
 Fair. 
 
 ammonium hydroxide to distinct but faintly alkaline reaction. 
 The solution should be neutral or alkaline rather than even 
 faintly acid, owing to the readiness with which sulphur 
 separates. To the alkaline solution add zinc acetate in 
 distinct excess and filter. The precipitate may be tested for 
 hydrogen sulphide, on acidifying, in the usual manner. To 
 the filtrate add acetic acid, a few drops in excess of the 
 amount necessary to neutralize, and barium chloride, and filter 
 through a double filter. To the filtrate add iodine until the 
 solution takes on a permanent yellow tinge, and then bleach 
 with stannous chloride, best after adding a few drops of 
 hydrochloric acid to prevent the possible precipitation of a 
 basic salt of tin. A precipitate at this point indicates the 
 sulphite. Filter, add bromine water in faint excess to the 
 filtrate, bleaching again with stannous chloride. A pre- 
 cipitate on adding bromine indicates a thiosulphate originally 
 present. 
 
XIX 
 
 ON THE SEPARATION OF NICKEL AND 
 COBALT BY HYDROCHLORIC ACID. 
 
 BYFRANKE STUART HAVENS. 
 
 A QUANTITATIVE separation of nickel and cobalt by a process 
 analogous to that published from this laboratory for the 
 separation of aluminum and ironf has been put forward in a 
 recent paper by E. Pinerua.J The process may be described 
 briefly as follows: The hydrous chlorides of nickel and 
 cobalt (0.3-0.4 grms.) are dissolved in a little water and to the 
 solution are added 10 to 12 cm 3 of aqueous hydrochloric acid 
 and 10 cm 3 of ether, and the whole, contained in a little 
 beaker surrounded with water, and ice, is saturated with 
 gaseous hydrochloric acid. The cobalt, which remains in 
 solution, is decanted off and the yellow insoluble nickel 
 chloride washed with a previously prepared solution of ether 
 saturated with hydrochloric acid gas at a low temperature. 
 The nickel is determined by known methods, preferably as 
 the sulphate. The author claims very precise results for the 
 process, but gives no experimental proof of his work. Previous 
 to the appearance of this paper my experiments upon the 
 solubility of nickel chloride in an ether-hydrochloric acid 
 solution, such as used in our process for the separation of 
 aluminum and iron, which is practically the same in proportions 
 as that used by Pinerua to effect precipitation, had shown that, 
 while nickel chloride is somewhat insoluble in such a mixture, 
 the degree of insolubility is not sufficient for a quantitative 
 separation. Since the appearance of Pinerua's work I have 
 
 * From Am. Jour. Sci., vi, 396. 
 
 t Gooch and Havens, Am. Jour. Sci., ii, 416. This volume, p. 20. 
 
 t Gaz. chim. ital., xxvii, 56. 
 
142 ON THE SEPARATION OF NICKEL AND 
 
 been over the ground again and have reached the same 
 conclusions as before, as shown in the following experiments. 
 
 When a solution of 0.02 grm. of nickel chloride (free from 
 iron and cobalt) in 7 cm 3 of aqueous hydrochloric acid, was 
 saturated with hydrochloric acid gas at a temperature of 2 
 C. (obtained by immersing the container in a mixture of ice 
 and salt) no precipitation resulted. When, however, an equal 
 volume of ether was added and the whole was again saturated 
 with hydrochloric acid gas a yellow precipitate formed, while 
 the supernatant liquid still remained of a deep green color. 
 The solution was filtered quickly through asbestos in a filter 
 crucible, and the clear filtrate after evaporation with sulphuric 
 acid was electrolyzed. The metallic deposit of 0.0020 grm. 
 proved to be pure nickel ; for when dissolved in nitric acid it 
 gave no test for iron with potassium sulphocyanide or ferro- 
 cyanide, and neither the apple-green hydroxide nor the black 
 sulphide, prepared by the usual methods, showed any trace of 
 cobalt in the borax bead. It is obvious, therefore, that nickel 
 chloride is not fully precipitated under these conditions and 
 that the green color of the solution is due to nickel in solution 
 and not to traces of iron, as Pinerua has supposed.* A second 
 experiment similar to the first showed a solubility of the 
 nickel chloride represented by 0.0018 grm. of metallic nickel. 
 It is evident, then, that the solubility of nickel chloride in this 
 mixture of aqueous hydrochloric acid and ether thoroughly 
 saturated with hydrochloric acid gas is not far from an amount 
 represented by 0.0020 grm. of metallic nickel for every 14 cm 8 
 of solution. 
 
 Still another experiment, in which nickel chloride repre- 
 senting 0.0020 grm. of metallic nickel was treated with 14 cm 3 
 of the ether-hydrochloric acid solution and the whole saturated 
 for one hour at a low temperature with hydrochloric acid gas 
 without precipitation, showed the same thing. 
 
 When the nickel chloride remaining on the asbestos was 
 washed with about 40 cm 8 of a mixture of equal parts ether and 
 aqueous hydrochloric acid saturated with hydrochloric acid gas, 
 
 * Loc. cit. ' 
 
COBALT BY HYDROCHLORIC ACID. 143 
 
 the washings evaporated with sulphuric acid and treated by 
 the battery gave a deposit of metallic nickel weighing 0.0027 
 grm. an amount proportionately less than that found in the 
 nitrate proper. 
 
 Although employing a mixture of aqueous hydrochloric acid 
 and ether saturated with gaseous hydrochloric acid for the 
 precipitation, Pineriia has advised the use of pure ether 
 saturated with gaseous hydrochloric acid for the washing. In 
 my experiments with such a mixture I find that in it the 
 hydrous nickel chloride is practically insoluble and that 30 cm 3 
 of the washings of the precipitated chloride with such a 
 mixture gave no deposit of nickel by the battery. It seemed 
 possible, therefore, that by reducing the water present to the 
 lowest possible amount necessary to dissolve the chlorides to 
 be treated the precipitation of the nickel might be made more 
 complete. The experiments of the following table were made 
 to put this idea to the test. 
 
 Solutions of the pure chlorides of nickel and cobalt, carefully 
 purified and freed from other metals and each other, were, 
 after conversion to the form of the sulphate, standardized 
 by the battery. Weighed portions of these solutions were 
 taken in a small beaker, evaporated to dryness, the dry salts 
 dissolved in as little water as possible (about 1 cm 3 ), 10 to 15 cm 3 
 of ether added, and the whole saturated with hydrochloric acid 
 gas, the beaker being meanwhile immersed in running water 
 and cooled to about 15 C. When saturation was complete 
 the precipitated chloride was caught on asbestos in a filter 
 crucible, washed thoroughly with a previously saturated 
 solution of ether, dissolved in water, evaporated with sulphuric 
 acid and determined as metallic nickel by the battery. The 
 cobalt in the filtrate was recovered by evaporation and 
 electrolysis in like manner. 
 
 Experiments (1), (2), and (3) of the accompanying table 
 show that by this process the nickel is thrown down quantita- 
 tively, and experiments (2) and (3) show that in the presence 
 of a few milligrams of the cobalt salt the separation of a small 
 amount of nickel is sharp. The residue of nickel in these 
 
144 
 
 SEPARATION OF NICKEL AND COBALT. 
 
 experiments gave no test for cobalt with the borax bead. 
 When, however, the cobalt is present to the amount of a few 
 centigrams as in (4), (5), and (6), the precipitated nickel 
 chloride, which forms a hard mass, includes the cobalt salt so 
 that even a large quantity of washing solution (100 cm 3 was 
 used in experiment 6) cannot remove it. 
 
 Exp. 
 
 Nickel taken 
 as the 
 hydrous 
 chloride. 
 
 Nickel 
 found. 
 
 Error. 
 
 Cobalt taken 
 as the 
 hydrous 
 chloride. 
 
 Cobalt 
 found. 
 
 Error. 
 
 (1) 
 
 (2) 
 (3) 
 (4) 
 (5) 
 (6) 
 
 giro. 
 
 0.0068 
 0.0090 
 0.0090 
 0.0469 
 0.0468 
 0.0472 
 
 grlu. 
 
 0.0066 
 0.0090 
 0.0091 
 0.0490 
 0.0503 
 0.0493 
 
 grm. 
 0.0002- 
 
 0.0000 
 0.0001+ 
 0.0021+ 
 0.0035+ 
 0.0021+ 
 
 grm. 
 
 0.0030 
 0.0123 
 0.0700 
 0.0700 
 0.0700 
 
 grin. 
 
 0.0127 
 
 grin. 
 
 0.0004+ 
 
 From the experiments described it is obvious that the pro- 
 cess as proposed by Pinerua will not give a complete precipi- 
 tation of the nickel chloride. Nickel chloride is, however, 
 practically insoluble in pure ether saturated with hydrochloric 
 acid gas and can be separated from small quantities of the solu- 
 ble cobalt salt in that medium. In the presence of even a few 
 centigrams of the cobalt chloride, however, the process is not 
 practicable on account of the inclusion of the cobalt by the 
 massive nickel chloride. It is possible that by repeated solu- 
 tions and reprecipitations the nickel salt might be sufficiently 
 freed from the cobalt, but the process must naturally be long 
 and tedious. 
 
XX 
 
 THE ETHERS OF TOLUQUINONEOXIME AND 
 
 THEIR BEARING ON THE SPACE ISOMER- 
 
 ISM OF NITROGEN. 
 
 BY JOHN L. BRIDGE AND WILLIAM CONGER MORGAN * 
 
 IN an article on the ethers of quinoneoxine (isonitrosophenol) 
 published by one of us,f it was stated, that when boiled with 
 alcohol, the benzoyl ether of quinoneoxime dichloride gave 
 gave two monochlor substitution-products. Naturally it was 
 supposed that the chlorine atom occupied, in the one, an ortho 
 position, and, in the other, a meta position to the radical 
 NOR, the reaction being : 
 
 000 
 
 + + 2HC1. 
 
 Hk /"HC1 HL JH Hi JCl 
 
 N N N 
 
 O 00 
 
 R R R 
 
 It was also found that these same isomers were formed when 
 monochlorquinone was treated with hydroxylamine hydro- 
 chloride, and the sodium salt of the chlorquinoneoxime thus 
 formed, treated with benzoyl chloride. 
 
 The preceding explanation regarding the splitting off of 
 hydrochloric acid from the dichloride is not in accord with re- 
 sults of work done by Kehrmann,J published in his article on 
 the influence of radicals in the action of hydroxylamine on 
 
 * From Am. Chem. Jour., xx, 761. 
 t Ann. Chem. (Liebig), cclxvii, 79. 
 t Ber. Dtsch. chem. Ges., xxi, 3315 ; Jour, prakt. Chem. [2], xl, 268. 
 
 VOL. II. 10 
 
146 THE ETHERS OF TOLUQUINONEOXIME. 
 
 quinones, in which he generalizes the results of his observa- 
 tions in the statement that the presence of a radical attached 
 to the ring so much lessens the replaceability of the quinone 
 oxygen atom neighboring to the radical that the principal 
 part if not the whole of the resulting product, is a meta-sub- 
 stituted quinoneoxime. The work of his former article has 
 been repeated by Kehrmann,* who finds the same isomeric 
 benzoyl ethers of monochlorquinoneoxime ; but believing 
 that observations of their behavior indicate the substantiation 
 of his rule, he states that both these ethers have the chlorine 
 atom in the meta position to the oximido group, and attributes 
 their difference to space isomerism of nitrogen, writing the re- 
 action thus : 
 
 + 2HC1. 
 
 Kehrmann designates the compound represented by symbol I. as 
 " chlorquinonemetaantioxime ether " and II. as " chlorquinone- 
 metasynoxime ether." The question as to whether the chlor- 
 ine atom occupies the same position in both compounds could 
 be definitely settled if ortho- and metachlorphenols could be 
 changed to the corresponding chlorquinoneoximes or so-called 
 isonitrosophenols. This transformation has, however, unfor- 
 tunately not yet been accomplished, and repeated efforts to 
 obtain the corresponding bromquinoneoximes have resulted 
 in failure, orthobromphenol not being attacked by nitrous 
 acid or amyl nitrite. 
 
 The toluquinoneoximes, obtained by the action of nitrous 
 acid on ortho- and metacresol, we have taken up for study as 
 being the most closely analogous compounds in which the 
 position of the side groups is definitely known, believing that 
 
 * Ann. Chem. (Liebig), cclxxix, 27. 
 
THE ETHERS OF TOLUQUINONEOXIME. 147 
 
 by an examination of these bodies, light may be thrown upon 
 the nature of the others. 
 
 When orthocresol is acted upon by nitrous acid, a toluqui- 
 nonemetaoxine is formed according to the following reaction: 
 
 H 
 
 O 
 
 + HONO = 
 H 
 
 H 
 
 The metacresol forms similarly a corresponding orthooxime : 
 
 H 
 
 O O 
 
 H r XXx i H 
 
 + HONO == + H 2 0. 
 
 H LJ CH 8 
 
 H N 
 
 O 
 H 
 
 The benzoyl ethers of the toluquinoneoximes prove to be 
 exceedingly well fitted to characterize these bodies since 
 they are formed from the oximes in quantitative proportions, 
 are easily crystallized, and readily distinguished from each 
 other. 
 
 The benzoyl ether of toluquinoneorthooxime produced 
 from metacresol, crystallizes in light brownish-yellow crystals 
 melting at 177 C. and appearing under the microscope as 
 long rectangular blades, suggesting the orthorhombic system. 
 The benzoyl ether of toluquinonemetaoxime produced from 
 orthocresol is obtained in the form of yellow crystals which 
 begin to soften at about 150 and do not melt entirely 
 until at about 190. Subjected to fractional crystallization, 
 a portion of the metaoxime is readily obtained consisting of 
 branching needle-like crystals, melting at 193 C. These, as 
 
148 THE ETHERS OF TOLUQUINONEOXIME. 
 
 well as lower-melting fractions, can be readily distinguished 
 from the orthooxime by their less regular appearance under 
 the lens, as well as by their crystal habit, tending to produce 
 curved forms. 
 
 The marked difference in the form and habit of the crystals 
 of the benzoyl ethers of the ortho- and metatoluquinone 
 oximes makes it possible to study carefully the product of the 
 action of toluquinone on hydroxylamine. Goldschmidt and 
 Schmidt * have shown that the principal product of this reac- 
 tion is toluquinonemetaoxime. This they demonstrated by 
 oxidizing to a dinitro compound; but as some of the theo- 
 retically possible nitro-derivatives of the cresols are not 
 known, this method can scarcely be considered to prove con- 
 clusively the absence of toluquinoneorthooxime. We have, 
 therefore, studied further the product of the action of hydroxyl- 
 amine on toluquinone by the aid of the benzoyl ethers. When 
 the sodium salt of the oxime thus formed was treated in alco- 
 holic solution with benzoyl chloride, and the benzoyl ether 
 obtained was subjected to fractional crystallization, not a 
 trace of the orthooxime ether, crystallizing in blades and 
 melting at 177, was found. This proves that when tolu- 
 quinone is treated with hydroxylamine the whole of the re- 
 sulting product is toluquinonemetaoxime. The inference is 
 plain that Kehrmann's rule concerning the influence of side- 
 chains, attached to the ring in quinones, upon the entering 
 oximido radical is quantitatively true in this case ; and similar 
 indications furnished by the corresponding methyl ethers, as 
 will appear later, strengthen this conclusion. 
 
 As has been previously shown, the benzoyl ether of the 
 metaoxime does not consist of a single compound, but is a 
 mixture. The body melting at 193 is readily separated in 
 considerable quantities, while the remainder consists of a 
 very intimate mixture of this body with one of much lower 
 melting-point, separated, if it can be separated at all, only 
 with the greatest difficulty. By concentration of the mother- 
 liquor from which the high-melting fractions have been 
 
 * Ber. Dtsch. chem. Ges., xvii, 2063. 
 
THE ETHERS OF TOLUQUINONEOXIME. 149 
 
 obtained, and carefully crystallizing, fractions may be separated 
 melting almost completely at temperatures, varying with the 
 fraction, between 130 and 150. Repeated crystallizations 
 separated each fraction into portions melting, on the one hand, 
 always considerably higher, and, on the other hand, often 
 somewhat lower, and ordinarily no definite body of low 
 melting-point could be isolated. Twice, however, a nearly 
 complete separation seemed to have been obtained. Thus, 
 from fractions melting between 140 and 150, a few short, 
 thick prisms melting once at 142 and again at 144 sepa- 
 rated, leaving the compound which melts at 193. Of these 
 crystals there was not enough for a combustion, but an 
 analysis of a fraction melting almost completely at 137 gave 
 figures which show without doubt that this portion had the 
 same percentage composition as the body melting at 193. If 
 we are to regard the product as it is first obtained as a mixture 
 of two ethers, and consider that 142 -144 is the melting-point 
 of the lower body, we can account for the melting of fractions 
 below 142 on the ground that mixtures may have a lower 
 melting-point than either of the component substances. 
 These results were obtained repeatedly from three separate 
 portions of Kahlbaum's C. P. orthocresol, purchased at 
 different times, as well as from toluquinone melting at 67 
 C. and wholly volatile when exposed to the air. 
 
 When the silver salt of toluquinonemetaoxime acts upon 
 methyl iodide, the product is, likewise, not a single ether 
 but a mixture of ethers, softening at 55 and not melting 
 completely below 70. From it a body melting at 73 -74 
 can easily be separated, but no other compound of definite 
 melting-point or different crystal form could be obtained. 
 
 The acetyl compound of toluquinonemetaoxime presents 
 phenomena similar to those of the benzoyl ethers, but in 
 rather more marked degree. The product, as first obtained 
 by the action of acetyl chloride on the silver salt of the 
 oxime, or by acetic anhydride on the oxime itself, begins to 
 soften at 90 and melts completely at 110. Upon the first 
 recrystallization a distinction in crystal form appears, and a 
 
150 THE ETHERS OF TOLUQUINONEOXIME. 
 
 second crystallization of the separated portions gave short, 
 thick prisms melting at 112-113, and some smaller, spheri- 
 cally grouped crystals, melting at 85 -87. The extreme 
 difficulty of preparing the acetyl compound prevented further 
 investigation. 
 
 In like manner, the benzoyl ether of monobromtoluquinone- 
 metaoxime seems to be a mixture of isomeric bodies. Thus, it 
 was found possible to add two atoms of bromine to the ethers 
 of toluquinoneoxime forming colorless dibromides correspond- 
 ing to the dibrom addition-products of quinoneoxime, and 
 these dibromides, when boiled with dilute alcohol, split off 
 hydrobromic acid with the formation of colored monobrom 
 substitution-products. The benzoyl ether of monobromtolu- 
 quinonemetaoxime thus formed shows a variation in melting- 
 point similar to that of the ethers previously discussed. 
 
 The foregoing facts speak in favor of Kehrmann's theory of 
 space isomerism in the oximes so far as the metaoximes are 
 concerned. On the other hand, there is no evidence to show the 
 presence of isomers in the ethers of toluquinoneorthooxime : 
 the methyl, acetyl, and benzoyl ethers all act as simple 
 substances, each product melting completely at a definite 
 temperature. It is difficult to understand why isomerism 
 should be so much more evident in the ethers of toluquinone- 
 metaoxime than in the ethers of the orthooximes, unless, 
 possibly, the closer proximity of the side-chain to the 
 oximido-group prevents the formation of a space isomer. 
 There is, however, a remote possibility that isomeric bodies 
 may exist, so similar in properties that they cannot be detected 
 by the ordinary methods. 
 
 EXPERIMENTAL PART. 
 Preparation of the Oximes and their Salts. 
 
 The toluquinone-, ortho-, and metaoximes used in the 
 experiments to be described were made hi the following 
 manner : To a solution of 10 grams of cresol and 8 grams of 
 potassium nitrite in 900 cm 3 of water, a solution of 6 grams 
 
THE ETHERS OF TOLUQUINONE OXIME. 151 
 
 of concentrated sulphuric acid in 100 cm 3 of water was added 
 in small portions during the course of half an hour, care 
 being taken that both original solutions should be between 5 
 and 10, and that this temperature be maintained during the 
 mixing. Nearly all of the oxime separates out on standing 
 in ice-water for one to two hours, and after filtering and 
 washing with 200-300 cm 3 of ice-water, the amount of oxime 
 obtained by extracting the filtered solution with ether is so 
 small that it may be disregarded. The substances were 
 purified by dissolving in a saturated solution of sodium 
 carbonate and filtering into dilute sulphuric acid, cooled with 
 ice. At this stage it is generally ready for use, but if further 
 purification is desired, it may be accomplished by dissolving 
 the oxime in ether and shaking with animal charcoal. Upon 
 filtering and evaporating, the oxime crystallizes in long, 
 slightly colored needles. The yield is large in both cases, 
 that of the orthooxime being nearly theoretical. As given by 
 Beilstein, toluquinone-ra-oxime melts at 134 C. Toluquin- 
 one-o-oxime melts at 155 C. ; Beilstein gives 145-150. 
 
 The silver salt of the metaoxime was made in the following 
 manner : 5 grams of toluquinone-7?M)xime were dissolved in 
 a solution of sodium hydroxide, a little less than the quantity 
 calculated to form the sodium salt, and this solution was 
 filtered into 500 cm 3 of water containing 1J times the calcu- 
 lated amount of silver nitrate. The precipitate comes down 
 in a flocky, gelatinous condition, but goes over into a granular 
 form on heating to 50 in a water-bath. When dissolved in 
 the least possible amount of warm dilute ammonia, and the 
 solution precipitated with hydrochloric acid, 
 
 0.1006 gram of the substance, dried over H 2 S0 4 , gave 0.0590 
 gram AgCl. 
 
 Calculated for V^^A 
 
 C 7 H 6 N0 2 Ag. 
 
 Ag 44.23 44.14 
 
 The salt is light reddish-brown when first formed, but 
 turns darker on standing or heating. It decomposes when 
 heated to 100 and, when thoroughly dry, is spontaneously 
 
152 THE ETHERS OF TOLUQUINONEOXIME. 
 
 inflammable at a temperature above 60. It is a rather un- 
 stable body and cannot be kept long when at all impure. 
 
 The silver salt of toluquinone-o-oxime was made in the 
 manner described for the preparation of the silver salt of 
 the meta form. The salt falls in the cold as reddish-brown 
 crystals, tending to darken when exposed to light or heat. 
 
 0.2560 gram, dried over H 2 S0 4 , gave 0.1498 gram AgCl. 
 
 Calculated for ,, , 
 
 C 7 H 6 N0 2 Ag. Found ' 
 
 Ag 44.23 44.06 
 
 Although very similar in all its properties to the silver salt 
 of toluquinone-w-oxime, this salt is like all the ethers of the 
 ortho form, much more stable than its corresponding meta 
 isomer. 
 
 Toluquinone-m-oxime Methyl Ether. 
 
 From o- CresoL Of the silver salt of toluquinone-w-oxime, 
 2 or 3 grams were suspended in 10-15 cm 3 of ligroin and 
 twice the calculated quantity of methyl iodide added. After 
 standing for an hour with frequent shaking, the liquid was 
 filtered off and the residue extracted with a little hot ligroin. 
 The united ligroin solutions were allowed to evaporate spon- 
 taneously, and the methyl ether came out in large, dark- 
 yellow, hexagonal prisms. A little more may be obtained by 
 allowing the residue to stand for a week with methyl iodide. 
 The yield in any case is small, the best results apparently 
 being obtained by using not more than 2 or 3 grams of the 
 silver salt at one tune. After purifying with animal charcoal 
 and recrystallizing from ligroin, the product obtained softens 
 at 55 yet does not melt completely below 70. Portions 
 melting at 73-74 C. were separated by fractional crystalli- 
 zation, and, on analysis, 
 
 0.1101 gram of this body, dried over H 2 S0 4 , gave 0.2582 gram 
 C0 2 and 0.0599 gram H 2 0. 
 
 0.0881 gram of the substance gave 7.1 cm 8 N at 15 C. and 
 772 mm. pressure. 
 
THE ETHERS OF TOLUQUINONEOXIME. 153 
 
 Calculated for v> nm <{ 
 
 CgHoNO,. 
 
 C 63.53 63.96 
 
 H 6.00 6.05 
 
 C 9.29 9.59 
 
 Although fractions were often obtained melting from 55- 
 60, no other compound of very definite melting-point could 
 be separated. The methyl ether is very soluble in all organic 
 reagents ; in hot ligroin it is much more soluble than in cold, 
 from which it crystallizes in small bright-yellow prisms. 
 
 From Toluquinone. To a solution of 2 grams of tolu- 
 quinone in 800 cm 3 of water, the calculated amount of 
 methoxylamine hydrochloride was added. Yellow crystals 
 began to precipitate in the course of two hours, and at the 
 end of twelve hours the reaction was completed. The liquid 
 was filtered off and extracted with ether, which, upon evapo- 
 ration, left behind a yellow crystalline mass. The two 
 portions were united and recrystallized from ligroin. The 
 yield was very good, being 75 per cent of the theory. Even 
 after boiling in ligroin with animal charcoal and recrystallizing 
 several tunes, the substance acts like a mixture, softening at 
 58 and melting at 70. A portion, less soluble than any 
 other, was easily separated, which melted at 73-74 C. and 
 was identical in all respects with the methyl ether obtained 
 from the silver salt of toluquinone-m-oxime made from 
 0-cresol. 
 
 0.1263 gram, dried over H 2 S0 4 , gave 0.2962 gram C0 2 and 
 0.0672 gram H 2 0. 
 
 0.2952 gram gave 25 cm 8 N at 15 C. and 772 mm. pressure. 
 
 C 63.53 63.96 
 
 H 6.00 5.91 
 
 N 9.29 10.00 
 
 Toluquinonemetaoxime Acetyl Ether. 
 
 This ether can be made in two ways : By adding the cal- 
 culated amount of acetyl chloride, drop by drop, to 2-3 
 
154 THE ETHERS OF TOLUQUINONEOXIME. 
 
 grams of the silver salt suspended in 15-20 cm 8 of ligroin or 
 absolute ether, kept cool by ice-water, evaporating at once, 
 and extracting the residue with hot ligroin; or by heating 
 1 molecule of the oxime on the water-bath for an hour with 
 1.5 molecules of acetic anhydride, adding cold water, filtering 
 off the tarry mass which separates, and extracting it with hot 
 ligroin. The yield by either method is extremely poor, and 
 sometimes after purifying by boiling with animal charcoal, 
 the total product consisted of a few small crystals. By frac- 
 tional crystallization two portions were separated, the less 
 soluble composed of thick irregular prisms melting at 112- 
 113, and a much smaller fraction of minute, spherically 
 grouped crystals, melting at 85-87. In analyzing the 
 original product unf ractioned : 
 
 0.0953 gram, dried over H 2 S0 4 , gave 0.2090 gram C0 2 and 
 0.0444 gram H 2 0. 
 
 0.0848 gram gave 5.8 cm 8 N at 15 C. and 760 mm. pressure. 
 
 Calculated for - , 
 
 C 9 H 9 N0 3 . Found. 
 
 C 60.30 59.91 
 
 H 5.06 5.18 
 
 N 7.84 8.01 
 
 The acetyl ether is very soluble in alcohol and ether, much 
 less in ligroin, and very little soluble in water. 
 
 Toluquinone-m-oxime Benzoyl Ether. 
 
 From o-Cresol. This ether can be made from the silver salt 
 suspended in absolute ether, or, better, from the sodium salt 
 in alcohol solution. Slightly less than the amount of 
 sodium calculated to form a sodium oxime is dissolved in 100 
 cm 3 of alcohol, 2-5 grams of the oxime added, and into the 
 filtered solution slightly more than the theoretical quantity of 
 benzoyl chloride is dropped slowly, the solution being kept 
 cool. The benzoyl ether begins to separate immediately, and 
 after a few moments the alcohol can be filtered off and 
 rejected, as it contains little of the substance. After boiling 
 
THE ETHERS OF TOLUQUINONEOXIME. 155 
 
 in alcohol with animal charcoal, when the solution is submitted 
 to fractional crystallization, three-fourths of the crude product 
 can be readily separated in the form of bright-yellow needles 
 melting at 193. Upon concentrating the mother-liquor to 
 a small volume and cooling, nearly the theoretical quantity of 
 the benzoyl ether can be recovered. Nothing further was 
 ever obtained save a few flakes of benzoic acid formed by the 
 saponifying action of hydrochloric acid, produced by the 
 slight excess of benzoyl chloride acting on alcohol. When 
 the portion obtained upon concentration was repeatedly 
 fractioned, it could be separated into portions melting approxi- 
 mately at 193, and others melting almost completely from as 
 low as 135 to 155. There seemed to be a tendency, however, 
 for these lower fractions to liquefy at 142 -144, and once, 
 from an alcoholic solution of a fraction melting at 140-150, 
 that evaporated at ordinary temperature, short, thick, prismatic 
 crystals separated from the curved needles of the higher 
 melting (193) fraction. These few prisms melted at 144 
 without decomposition. An analysis of the body liquefying 
 at 193 C. gave the following figures: 
 
 0.2277 gram, dried over H 2 S0 4 , gave 0.5785 gram C0 2 and 
 0.0945 gram H 2 0. 
 
 0.2001 gram gave 10 cm 8 N at 15 C. and 760 mm. pressure. 
 
 C 69.68 69.30 
 
 H 4.60 4.61 
 
 N 5.82 5.85 
 
 An analysis of a fraction melting from 145 -165 gave 
 these percentages: 
 
 0.1421 gram, dried over H 2 S0 4 , gave 0.3635 gram C0 2 and 
 0.0565 gram H 2 0. 
 
 0.6726 gram gave 34.1 cm 8 N at 15 C. and 770 mm. pressure. 
 
 Calculated for uv*,,^ 
 
 n TT icrk iround. 
 
 C 14 H n NO 3 . 
 
 C 69.68 69.74 
 
 H 4.60 4.43 
 
 N 5.82 6.02 
 
156 THE ETHERS OF TOLUQUINONEOXIME. 
 
 The benzoyl ether is not at all soluble in water or ligroin, is 
 slightly soluble in cold alcohol, but dissolves readily in ether, 
 chloroform, and glacial acetic acid. The low-melting body 
 seems to be more soluble in alcohol than its higher-melting 
 isomer, but, as nine-tenths of the total product appeared to 
 be the body melting at 193, the small proportion of the low- 
 melting compound may account partially for the idea that the 
 latter is more soluble. 
 
 From Toluquinone. The oxime was made according to the 
 method of Goldschmidt and Schmidt,* by treating toluquinone 
 in aqueous solution with an excess of hydroxylamine hydro- 
 chloride, extracting with ether, and purifying with animal 
 charcoal. From the sodium salt of the oxime thus formed 
 the benzoyl ether was made in the manner previously described. 
 The product is identical with the benzoyl ether made from 
 o-cresol. The body melting at 193 C. was readily isolated 
 and analyzed: 
 
 0.1002 gram, dried over H 2 S0 4 , gave 0.2565 gram C0 2 and 
 0.0452 gram H 2 0. 
 
 0.3581 gram gave 17.8 cm 8 N at 15 C. and 760 mm. pressure. 
 
 Calculated for v , 
 
 C M H U N0 8 . Found. 
 
 C 69.68 69.81 
 
 H 4.60 5.01 
 
 N 5.82 5.77 
 
 By fractional crystallization, low-melting portions were 
 separated, exactly as hi the case of the benzoyl ether made 
 from o-cresol, except that some fractions were obtained melt- 
 ing partially as low as 129. In order to be certain that 
 benzoic acid (m. p., 120 C.) was not unduly lowering these 
 melting-points, these fractions were boiled with water and 
 filtered hot. The filtrate was not acid to litmus and contained 
 only a trace of organic matter. The ethers, when dried, gave 
 the same melting-point as before boiling with water, and, 
 upon recrystallizing from a little alcohol, did not exhibit any 
 
 * Ber. Dtsch. chem. Ges., xvii, 2063. 
 
THE ETHERS OF TOLUQUINONEOXIME. 157 
 
 change. Furthermore, the benzoyl ether obtained by the 
 usual method, when only 75 per cent, of the theoretical 
 quantity of benzoyl chloride was used, gave fractions begin- 
 ning to melt at 129. Obviously benzoic acid could not be 
 present in these instances. 
 
 By slow, spontaneous evaporation of the alcoholic solution 
 of a fraction liquefying at 140-150, a few crystals melting 
 at 142, similar to those obtained in the same manner from 
 the oxime made from 0-cresol, separated from the body 
 melting at 193. Since there was not enough of this com- 
 pound for a combustion, an analysis was made of a fraction 
 melting at 137 C. with the following results : 
 
 0.1131 gram, dried over HgSO^ gave 0.2901 gram C0 2 and 
 0.0480 gram H 2 0. 
 
 Calculated for _ , 
 
 C U H U N0 8 . Found - 
 
 C 69.68 69.96 
 
 H 4.60 4.72 
 
 Fractions were frequently obtained melting at about 177 
 C., and from these attempts were repeatedly made to isolate 
 some of the ortho isomer. None was ever detected under the 
 lens, and fractional crystallization always separated such 
 portions, principally into the body melting at 190, and a 
 small fraction melting much lower. 
 
 Since the benzoyl ether will decompose into a dark-brown 
 liquid with the evolution of brown fumes of nitrogen oxides 
 when heated above 160, it is only by rapidly heating that 
 193 can be observed as a melting-point. This applies also 
 to the bromine addition-products of the benzoates of both 
 the ortho- and metaoximes. 
 
 Dibromtoluquinone-m-oxime Benzoyl Ether. 
 
 The benzoyl ether was dissolved in chloroform and cooled 
 while the theoretical quantity of bromine was added in small 
 portions. After standing for an hour, the chloroform was 
 evaporated spontaneously, and the light-brown residue re- 
 
158 THE ETHERS OF TOLUQUINONEOXIME. 
 
 crystallized from glacial acetic acid. It can also be purified 
 by dissolving in fuming nitric acid and pouring into water. 
 On analysis : 
 
 0.1437 gram, dried over H 2 S0 4 , gave 0.2176 gram C0 2 and 
 0.0339 gram H 2 0. 
 
 0.1215 gram gave 0.1140 gram AgBr. 
 
 Calculated for ,, , 
 
 C M H u Br 2 N0 8 . Found - 
 
 C 41.90 41.30 
 
 H 2.76 2.62 
 
 Br 39.87 39.93 
 
 The dibromide is insoluble in water, somewhat soluble in 
 cold alcohol, and readily dissolves in chloroform and glacial 
 acetic acid, from which it crystallizes in white prisms melting 
 at 165 C. with decomposition. 
 
 Monobromtoluquinone-m-oxime Benzoyl Ether. 
 
 When the dibromide is boiled with alcohol hydrobromic 
 acid splits off, two hours being required to complete the 
 process, during which little or no saponification takes place. 
 After recrystallizing from alcohol, a mixture of monobrom 
 compounds is obtained, melting from 155 -170. A portion 
 melting with decomposition at 174 C., was separated and 
 analyzed : 
 
 0.1208 gram, dried over H 2 S0 4 , gave 0.2319 gram C0 2 and 
 0.0345 gram H 2 0. 
 
 0.0735 gram gave 0.0425 gram AgBr. 
 
 Calculated for Fminrl 
 
 C 14 H 10 BrN0 8 . 
 
 C 52.49 52.36 
 
 H 3.15 3.17 
 
 Br 24.98 24.63 
 
 It is very similar in its properties to toluquinone-ra-oxime 
 benzoyl ether, crystallizing from alcohol in bright>yellow 
 needles. 
 
THE ETHERS OF TOLUQUINONEOXIME. 159 
 
 Toluquinone-o-oxime Methyl Ether. 
 
 The methyl ether of toluquinone-o-oxime was made from 
 the silver salt and methyl iodide in the same manner as its 
 meta isomer, the yield being somewhat better. Once recrys- 
 tallized from ligroin: 
 
 0.1247 gram of the ether, dried over H 2 S0 4 , gave 0.2931 gram 
 C0 2 and 0.0675 gram H 2 0. 
 
 0.3484 gram gave 27 cm 8 N at 15 C. and 772 mm. pressure. 
 
 Calculated for , , 
 
 Found ' 
 
 C 63.53 64.10 
 
 H 6.00 6.01 
 
 N 9.29 9.22 
 
 Its properties are almost identical with the methyl ether 
 of the meta form. It crystallizes from ligroin in long yellow 
 needles, every portion of which, obtained by fractional crys- 
 tallization, melts at 69 C. 
 
 Dibromtoluquinone-o-oxime Methyl Ether. 
 
 The methyl ether was dissolved in chloroform, cooled, and 
 the calculated quantity of bromine added. The reaction is 
 completed in twenty minutes, and, on evaporation of the 
 chloroform, the dibromide is left behind as a dirty white 
 mass. Once recrystallized from ligroin, the substance is 
 ready for analysis: 
 
 0.1690 gram, dried over H 2 S0 4 , gave 0.1877 gram C0 2 and 
 0.0456 gram H 2 O. 
 
 0.0744 grain gave 0.0909 gram AgBr. 
 
 Calculated for &*,,** 
 
 C 8 H 9 Br 2 N0 2 . 
 
 C 30.87 30.29 
 
 H 2.92 3.00 
 
 Br 51.41 52.00 
 
 It is insoluble hi water but quite soluble in most organic 
 reagents. From ligroin it crystallizes in white prisms, melt- 
 ing at 112 C. 
 
160 THE ETHERS OF TOLUQUINONEOXIME. 
 
 Toluquinone-o-oxime Acetyl Ether. 
 
 This ether was made by Wurster and Riedel * from tolu- 
 quinone-0-oxime and acetic anhydride. It can be prepared 
 also from the silver salt and acetyl chloride, the yield being 
 very poor, only a trifle better than its meta isomer, which 
 it closely resembles in its properties. 
 
 0.1194 gram, dried over H 2 SO 4 , gave 0.2666 gram C0 2 and 
 0.0545 gram H 2 O. 
 
 0.2389 gram gave 15 cm 8 N. at 15 C. and 760 mm. pressure. 
 
 Calculated for - , 
 
 C 9 H 9 N0 8 . Found - 
 
 C 60.30 60.89 
 
 H 5.06 5.07 
 
 N 7.84 7.35 
 
 From ligroin it crystallizes in irregular yellow prisms, 
 melting at 92 C. 
 
 Toluquinone-o-oxime Benzoyl Ether. 
 
 This ether was made in alcohol solution from the sodium 
 salt in the manner already described for the benzoyl ether of 
 the meta form. The yield is practically theoretical, and, 
 after one recrystallization from alcohol, the product is ready 
 for analysis. 
 
 0.1589 gram, dried over H 2 S0 4 , gave 0.4074 gram C0 2 and 
 0.0686 gram H 2 0. 
 
 0.5101 gram gave 26.1 cm 8 N at 15 C. and 7.66 mm. pressure. 
 
 Calculated for im,! 
 
 C U H U N0 3 . 
 
 C 69.68 69.92 
 
 H 4.60 4.80 
 
 N 5.82 6.04 
 
 Although submitted to the most careful fractional crystal- 
 lization, every particle obtained melted sharply at 177 C., 
 with slight decomposition. From alcohol it crystallizes in 
 light, brownish-yellow blades, which have all the properties 
 
 * Ber. Dtsch. chem. Ges., xii, 1799. 
 
THE ETHERS OF TOLUQUINONEOXIME. 161 
 
 of the metaoxime benzoyl ether except that it is a little more 
 soluble in organic reagents. 
 
 Dibromtoluquinone-o-oxime Benzoyl Ether. 
 
 The benzoyl ether was dissolved in chloroform, bromine 
 added to the cooled solution, and the product recrystallized 
 from glacial acetic acid. 
 
 0.2111 gram, dried over H 2 S0 4 , gave 0.3208 gram C0 2 and 
 0.0564 gram H 2 0. 
 
 0.1003 gram gave 0.0924 gram AgBr. 
 
 Calculated for Found 
 
 C 41.90 41.44 
 
 H 2.76 2.97 
 
 Br 39.87 39.21 
 
 It is insoluble in ligroin and water and but little soluble in 
 cold alcohol. Fuming nitric acid dissolves it readily, and 
 from the solution it is precipitated unchanged by water. It 
 crystallizes from glacial acetic acid hi short, thick, ortho- 
 rhombic prisms melting at 159 C. with decomposition. 
 
 VOL. II. 11 
 
XXI 
 
 THE APPLICATION OF IODINE IN THE 
 ANALYSIS OF ALKALIES AND ACIDS. 
 
 BY CLAUDE F. WALKER AND DAVID H. M. GILLESPIE* 
 
 IT is well known that when a free mineral acid is added to a 
 neutral mixture of metallic iodate and iodide, the iodate is 
 reduced and iodine is liberated according to the equation : 
 
 EI0 8 + 5EI + 3H 2 S0 4 = 31, + 3E 2 S0 4 + 3H 2 0. 
 
 This reaction is complete and non-reversible under the 
 conditions of analysis, and it may therefore be applied to the 
 estimation of amounts of iodate, iodide or mineral acid present 
 in an unknown solution. A solution of iodate to be analyzed 
 is mixed with an excess of iodide and mineral acid, the 
 resulting free iodine estimated by directly titrating with 
 sodium thiosulphate or arsenious acid, and one-sixth of the 
 amount found taken as equivalent to the iodate originally 
 present.f Similarly, a solution of iodide to be analyzed is 
 mixed with an excess of iodate and mineral acid, the resulting 
 free iodine estimated by directly titrating in alkaline solution 
 with arsenious acid, and five-sixths of its amount taken as 
 equivalent to the iodide originally present.^ A solution of 
 mineral acid to be analyzed is mixed with an excess of iodate 
 and iodide, the resulting free iodine estimated by directly 
 titrating with sodium thiosulphate, and its entire amount 
 taken as equivalent to the amount of mineral acid originally 
 present. Groger has applied the last mentioned method to 
 
 * From Am. Jour. Sci., vi, 455. 
 
 t Rammelsberg, Fogg. Ann., cxxxv, 493 ; Walker, Am. Jour. Sci., iv, 235. 
 This volume, p. 52. 
 
 J Gooch and Walker, Am. Jour. Sci., iii, 293. This volume, p. 33. 
 
 Kjeldahl, Zeitschr. anal. Chem., xxii, 366 ; Furry, Am. Chem. Jour., vi, 
 341 ; Groger, Zeitschr. angew. Chem., 1894, 52. 
 
IODINE IN ANALYSIS OF ALKALIES, ETC. 163 
 
 the direct analysis of various mineral acids, and has obtained 
 results manifestly better than those afforded by the use of 
 vegetable indicators. Groger has also indirectly analyzed 
 solutions of alkali hydroxides and carbonates by adding the 
 solution to be analyzed to a measured volume of mineral acid, 
 previously standardized by the above method, and estimating 
 the small excess of free mineral acid that finally remains by 
 the same method. The only difficulty with the Groger 
 process lies in the fact that under the conditions present the 
 end-point of the final reaction between iodine and sodium 
 thiosulphate is somewhat obscured by a peculiar back-play 
 of color due to a continuous slow liberation of iodine in the 
 system. 
 
 When a solution of a metallic hydroxide is acted on by 
 iodine at a temperature high enough to decompose the small 
 amounts of hypoiodites that might otherwise be present, 
 the final action results in the formation of an exactly neutral 
 mixture of iodate and iodide, according to the equation : 
 
 6EOH + 3I 2 = BI0 8 + 5RI + 3H 2 0. 
 
 Phelps * has shown that in the case of barium hydroxide at 
 least this reaction is regular and complete under the conditions 
 of analysis, and is independent of the excess of iodine which 
 remains hi the neutral mixture unacted upon and may be 
 estimated by directly titrating with arsenious acid. Phelps 
 not only applies this principle of action to the standardization 
 of solutions of barium hydroxide by boiling with an excess of 
 iodine in a trapped flask, but also bases thereon a differential 
 method for determining carbon-dioxide, in which the liberated 
 gas is run into a measured amount of barium hydroxide, the 
 final excess of which is estimated by treating with iodine in 
 the presence of the precipitated barium carbonate. The good 
 result obtained by Phelps with barium hydroxide suggested 
 that the attempt be made to analyze alkali hydroxides, and 
 possibly carbonates, by a method, simpler than that devised by 
 Groger, based on the direct treatment of these compounds 
 
 * Am. Jour. Sci., ii, 70. Volume I, p. 369. 
 
164 APPLICATION OF IODINE IN THE 
 
 with iodine in hot solution. It also seemed possible to apply 
 the differential method not only to carbon dioxide but to any 
 acid or other compound that will act definitely and completely 
 with the metallic hydroxide employed, provided the soluble or 
 insoluble product formed will not be attacked when heated in 
 the presence of iodine. It was decided to modify the Phelps 
 process, however, in order to obviate the necessity of handling 
 large measured amounts of iodine in a flask trapped to prevent 
 mechanical loss by heating. The flask was therefore dispensed 
 with altogether, and the hydroxide solution to be analyzed 
 was mixed with an approximately measured excess of iodine 
 solution, in an Erlenmeyer beaker, the mouth of which was 
 lightly closed with a little trap to prevent loss by spattering. 
 The excess of iodine was then completely removed by boiling, 
 and the cooled colorless solution remaining, which contained 
 a neutral mixture of iodate and iodide, was acidified with a 
 mineral acid and the liberated iodine titrated with sodium 
 thiosulphate, the amount found being equivalent to the 
 amount of hydroxide taken for analysis. 
 
 The present investigation was undertaken to study the 
 limitations and possible applications in analysis of the 
 reactions between iodine on the one hand, and barium 
 hydroxide, potassium hydroxide and sodium carbonate on the 
 other. It was soon found that the reaction in the case of 
 sodium carbonate is entirely dependent on conditions of time, 
 mass, and temperature, and cannot be pushed to completion 
 except under conditions that make its application in analysis 
 impossible. In the case of barium and potassium hydroxides 
 both the original procedure of Phelps and the modification 
 above described were employed. The modified method was 
 found to be the more convenient and speedy of the two. The 
 results obtained in both cases agreed with one another, 
 but were invariably lower by a small nearly constant 
 amount than those obtained by both the gravimetric and the 
 Groger processes. This error of the Phelps process and its 
 modification is possibly due to the action of atmospheric carbon 
 dioxide on the hydroxide solution during the short tune it is 
 
ANALYSIS OF ALKALIES AND ACIDS. 165 
 
 exposed. While it will affect the value of the method as a means 
 of accurately determining the absolute amount of hydroxide 
 present in a given volume of solution, it cannot so affect the 
 accuracy of any differential method founded on the original 
 Phelps process or its modification. This is demonstrated by 
 the work of Phelps in the case of carbon dioxide, and by the 
 present investigation in the case of hydrochloric and sulphuric 
 acid. Analyses of these two acids were made by adding the 
 solution to be analyzed to a measured volume of barium or 
 potassium hydroxide, previously standardized by the modified 
 Phelps method. The small excess of hydroxide remaining 
 was then estimated by the same method, the results agreeing 
 with those already obtained by both the gravimetric and the 
 Gro'ger processes. It seems probable that other acids and 
 compounds for which there is now no rapid iodometric method 
 may be analyzed by a method similar to this, which has given 
 good results with carbonic, hydrochloric, and sulphuric acids. 
 
 Decinormal solutions of the alkali hydroxides were prepared, 
 and kept with great care in trapped bottles, from which por- 
 tions for analysis were measured by means of a self-feeding 
 burette, which was also fitted with a trap. All vessels and 
 water used were made as free as possible from carbon dioxide, 
 and the operations were conducted as rapidly as possible. 
 
 In the analyses by the Phelps method a carefully measured 
 excess of decinormal iodine was drawn into a small ether wash- 
 bottle, and the desired amount of alkali was rapidly run into 
 it. The stopper, to which had been sealed a Will and Varren- 
 trapp absorption bulb was placed in the bottle and the bulb 
 was charged with a 5 per cent solution of potassium iodide to 
 catch any escaping vapors of iodine. The apparatus was 
 placed over a low flame and the contents heated to boiling or 
 slightly longer, and then cooled hi a stream of water. The 
 contents of the bulb and connecting tubes were then washed 
 into the flask, and the excess of free iodine remaining was 
 titrated with arsenious acid, in the presence of 5 cm 3 of starch 
 emulsion. Blank analyses were made to insure against me- 
 chanical loss of iodine during boiling and to prevent any error 
 
166 
 
 APPLICATION OF IODINE IN THE 
 
 on account of the presence of carbonate or other impurity in 
 the solutions employed. 
 
 Some of the results obtained with barium hydroxide are 
 given in Table I. The variation in different analyses of the 
 same series is not large, and the results are independent of 
 the amount taken for analysis and of the excess of iodine 
 
 employed. 
 
 TABLE I. 
 
 ANALYSES OP * BARIUM HYDROXIDE SOLUTION. 
 (By boiling in a trapped flask with an excess of iodine.) 
 
 Exp. 
 
 Ba(OH) a 
 taken. 
 
 Iodine 
 taken. 
 
 Iodine 
 absorbed by 
 
 Ba(OH) 3 
 found. 
 
 Mean. 
 
 Variation. 
 
 
 
 
 Ba(OH) 2 . 
 
 
 
 
 
 cm 8 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm 
 
 (1) 
 
 10 
 
 0.13 
 
 0.1054 
 
 0.0712 
 
 0.0699 
 
 0.0013+ 
 
 (2) 
 
 10 
 
 0.14 
 
 0.1028 
 
 0.0692 
 
 0.0699 
 
 0.0007- 
 
 (3) 
 
 20 
 
 0.23 
 
 0.2072 
 
 0.1399 
 
 0.1398 
 
 0.0001+ 
 
 (4) 
 
 20 
 
 0.25 
 
 0.2074 
 
 0.1401 
 
 0.1398 
 
 0.0003+ 
 
 (5) 
 
 40 
 
 0.44 
 
 0.4143 
 
 0.2798 
 
 0.2796 
 
 0.0002+ 
 
 (6) 
 
 40 
 
 0.44 
 
 0.4148 
 
 0.2802 
 
 0.2796 
 
 0.0006+ 
 
 (7) 
 
 40 
 
 0.48 
 
 0.4160 
 
 0.2809 
 
 0.2796 
 
 0.0013+ 
 
 (8) 
 
 40 
 
 0.48 
 
 0.4126 
 
 0.2786 
 
 0.2796 
 
 0.0010- 
 
 (9) 
 
 40 
 
 0.51 
 
 0.4115 
 
 0.2779 
 
 0.2796 
 
 0.0017- 
 
 (10) 
 
 40 
 
 0.51 
 
 0.4136 
 
 0.2793 
 
 0.2796 
 
 0.0003- 
 
 The analyses of potassium hydroxide were made in the 
 same way as were those of barium hydroxide, and gave quite 
 similar results. They follow in Table II. 
 
 TABLE H. 
 
 ANALYSIS OP J POTASSIUM HYDROXIDE SOLUTION. 
 (By boiling in a trapped flask with an excess of iodine.) 
 
 Exp. 
 
 KOH 
 
 taken. 
 
 Iodine 
 taken. 
 
 Iodine 
 absorbed by 
 KOH. 
 
 KOH 
 
 found. 
 
 Mean. 
 
 Variation. 
 
 
 cm> 
 
 grm. 
 
 grm. 
 
 grin* 
 
 grm. 
 
 grm. 
 
 (1) 
 
 10 
 
 0.20 
 
 0.1621 
 
 0.0716 
 
 0.0717 
 
 0.0001- 
 
 (2) 
 
 10 
 
 0.23 
 
 0.1613 
 
 0.0715 
 
 0.0717 
 
 0.0002- 
 
 ( 3 ) 
 
 15 
 
 0.30 
 
 0.2404 
 
 0.1063 
 
 0.1076 
 
 0.0013- 
 
 4) 
 
 15 
 
 0.30 
 
 0.2429 
 
 0.1074 
 
 0.1076 
 
 0.0002- 
 
 5 
 
 16 
 
 0.34 
 
 0.2431 
 
 0.1076 
 
 0.1076 
 
 0.0001- 
 
 (6) 
 
 25 
 
 0.51 
 
 0.4089 
 
 0.1808 
 
 0.1792 
 
 0.0016+ 
 
 (7) 
 
 25 
 
 0.51 
 
 0.4058 
 
 0.1794 
 
 0.1792 
 
 O.OOOSH- 
 
ANALYSIS OF ALKALIES AND ACIDS. 167 
 
 The analyses by the modification of the Phelps method were 
 made by drawing into an Erlenmeyer beaker of convenient 
 size an approximately measured excess of decinormal iodine, 
 and rapidly running the desired amount of alkali into it. The 
 neck of the beaker was then closed by a little trap, made of 
 one of the halves of a double end calcium chloride drying 
 tube, to prevent appreciable loss by spattering. The beaker 
 was then placed over a low flame, and the contents boiled until 
 the last trace of the excess of iodine had volatilized from the 
 solution and the trap. The volume was carefully regulated 
 before and during the boiling, being kept as small as possible, 
 usually amounting to about 100 cm 3 at the start and 35 cm 3 
 at the close. In the case of barium hydroxide care had to be 
 taken to keep the dilution sufficient to prevent the separation 
 the crystalline barium iodate, which is soluble with difficulty. 
 To steady the ebullition a little spiral of platinum was intro- 
 duced into the beaker. After the boiling had ceased, the 
 colorless solution, containing a neutral mixture of iodate and 
 iodide, was cooled hi running water and treated with 10 cm 3 
 of dilute (1 : 3) hydrochloric acid or (1 : 3) sulphuric acid. 
 The liberated iodine was titrated directly with sodium thiosul- 
 phate, hi the presence of 5 cm 3 of starch emulsion. In the 
 case of barium hydroxide the iodine was liberated with dilute 
 (1 : 3) hydrochloric acid to save the inconvenience of working 
 in the presence of precipitated barium sulphate ; with potas- 
 sium hydroxide, however, dilute (1 : 3) sulphuric acid was 
 employed. In view of a statement by Pickering* that titra- 
 tions with sodium thiosulphate in the presence of acid involve 
 an error, a series of blank analyses was made which showed con- 
 clusively that no such error exists under the conditions which 
 obtain in the process under consideration. Care was also taken, 
 as in a former case, to guard against the possible presence of 
 carbonates or other impurities in the reagents employed. 
 
 In Table III are given the results of a series of analyses of 
 barium hydroxide by the modified method just described. 
 They agree fairly well with those of Table I. 
 * Jour. Chem. Soc., xxxvii, 134. 
 
168 
 
 APPLICATION OF IODINE IN THE 
 
 ANALYSES OF 
 
 TABLE III. 
 BAKIUM HYDROXIDE SOLUTION. 
 
 (By boiling with excess of iodine in an open beaker to decoloration, and 
 acidifying the residue.) 
 
 Exp. 
 
 Ba(OH), 
 taken. 
 
 Iodine 
 taken. 
 
 Iodine 
 absorbed by 
 Ba(OH),. 
 
 Ba(OH), 
 found. 
 
 Mean. 
 
 Variation. 
 
 
 cm** 
 
 grni. 
 
 grin. 
 
 gnu. 
 
 griii. 
 
 grin. 
 
 (1) 
 
 10 
 
 0.13 
 
 0.1023 
 
 0.0691 
 
 0.0695 
 
 0.0004- 
 
 2) 
 
 10 
 
 0.16 
 
 0.1020 
 
 0.0689 
 
 0.0695 
 
 0.0006- 
 
 3 
 
 16 
 
 0.18 
 
 0.1548 
 
 0.1046 
 
 0.1043 
 
 0.0003+ 
 
 4 
 
 15 
 
 0.20 
 
 0.1546 
 
 0.1045 
 
 0.1043 
 
 0.0002+ 
 
 (5) 
 
 20 
 
 0.23 
 
 0.2049 
 
 0.1384 
 
 0.1390 
 
 0.0006- 
 
 (6) 
 
 20 
 
 0.25 
 
 0.2058 
 
 0.1390 
 
 0.1390 
 
 0.0000 
 
 (7) 
 
 20 
 
 0.32 
 
 0.2065 
 
 0.1394 
 
 0.1390 
 
 0.0004+ 
 
 (8) 
 
 25 
 
 0.29 
 
 0.2567 
 
 0.1734 
 
 0.1738 
 
 0.0004- 
 
 (9) 
 
 25 
 
 0.32 
 
 0.2562 
 
 0.1730 
 
 0.1738 
 
 0.0008- 
 
 (10) 
 
 40 
 
 0.47 
 
 0.4120 
 
 0.2783 
 
 0.2780 
 
 0.0003+ 
 
 (11) 
 
 40 
 
 0.48 
 
 0.4119 
 
 0.2782 
 
 0.2780 
 
 0.0002+ 
 
 (12) 
 
 40 
 
 0.48 
 
 0.4152 
 
 0.2804 
 
 0.2780 
 
 0.0024+ 
 
 (13) 
 
 40 
 
 0.49 
 
 0.4109 
 
 0.2775 
 
 0.2780 
 
 0.0005- 
 
 The analyses of potassium hydroxide by the modified method 
 are given in Table IV, and are found to agree well with those 
 of Table II. 
 
 TABLE IV. 
 ANALYSES OF ^ POTASSIUM HYDROXIDE SOLUTION. 
 
 (By boiling with excess of iodine in an open beaker to decoloration, and 
 acidifying the residue.) 
 
 Eip. 
 
 KOH 
 taken. 
 
 Iodine 
 taken. 
 
 Iodine 
 absorbed by 
 KOH. 
 
 Ba(OH) 2 
 found. 
 
 Mean. 
 
 Variation. 
 
 
 cm 
 
 grm. 
 
 gnu. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 10 
 
 0.20 
 
 0.1624 
 
 0.0718 
 
 0.0721 
 
 0.0003- 
 
 (2) 
 
 10 
 
 0.23 
 
 0.1618 
 
 0.0715 
 
 0.0721 
 
 0.0006- 
 
 (3) 
 
 10 
 
 0.25 
 
 0.1622 
 
 0.0717 
 
 0.0721 
 
 0.0004- 
 
 (4) 
 
 15 
 
 0.30 
 
 0.2459 
 
 0.1087 
 
 0.1082 
 
 0.0005+ 
 
 (5 
 
 15 
 
 0.34 
 
 0.2473 
 
 0.1093 
 
 0.1082 
 
 0.0011+ 
 
 (6) 
 
 15 
 
 0.38 
 
 0.2441 
 
 0.1079 
 
 0.1082 
 
 0.0003- 
 
 (7) 
 
 20 
 
 0.41 
 
 0.3274 
 
 0.1447 
 
 0.1442 
 
 0.0005+ 
 
 (8) 
 
 20 
 
 0.46 
 
 0.3259 
 
 0.1441 
 
 0.1442 
 
 0.0001- 
 
 (9) 
 
 20 
 
 0.51 
 
 0.3269 
 
 0.1445 
 
 0.1442 
 
 0.0003+ 
 
 (10 
 
 25 
 
 0.51 
 
 0.4052 
 
 0.1791 
 
 0.1803 
 
 0.0012- 
 
 (11) 
 
 25 
 
 0.57 
 
 0.4082 
 
 0.1805 
 
 0.1803 
 
 0.0002+ 
 
 (12) 
 
 25 
 
 0.63 
 
 0.4080 
 
 0.1804 
 
 0.1803 
 
 0.0001+ 
 
ANALYSIS OF ALKALIES AND ACIDS. 
 
 169 
 
 A gravimetric analysis of the barium hydroxide solution in 
 which the barium was weighed as the sulphate, gave as a 
 result 0.1411 grm. Ba(OH) 2 for each 20 cm 3 taken. An anal- 
 ysis of the same solution by the Groger process gave for the 
 same volume 0.1420 grm. The result by the Phelps process, 
 however, was 0.1398 grm., and by the modified process 0.1390 
 grm. That the difference of 2 mg. between the results by the 
 gravimetric and the Groger processes on one hand, and the 
 Phelps process and its modification on the other, may be due 
 to atmospheric carbon dioxide, has already been pointed out. 
 A gravimetric analysis of the potassium hydroxide solution 
 by evaporating and weighing as KC1 gave 0.1111 grm. KOH 
 for each 20 cm 3 taken, agreeing with 0.1106 grm. obtained by 
 the Groger process. The analyses by the Phelps process and 
 its modification for the same solution gave 0.1076 grm. and 
 0.1082 grm. respectively. These results are strikingly in 
 accord with those obtained with barium hydroxide. 
 
 In the application of the modification of the Phelps process 
 to the indirect analysis of hydrochloric and sulphuric acids the 
 procedure was essentially the same as that detailed for the 
 analysis of barium and potassium hydroxides in Tables III and 
 IV. The acid solution to be analyzed was drawn into an 
 Erlenmeyer beaker, a measured excess of standardized alkali 
 
 TABLE V. 
 ANALYSES OF ^ HYDKOCHLOKIC ACID SOLUTION. 
 
 (By adding to excess of ^ Ba(OH) 2 , boiling with excess of iodine to decolora- 
 tion and acidifying the residue.) 
 
 Exp. 
 
 HCl 
 
 taken. 
 
 Ba(OH), 
 taken. 
 
 Ba(OH) 2 
 neutralized 
 
 HCl found. 
 
 Mean. 
 
 Variation. 
 
 
 
 
 by HCl. 
 
 
 
 
 
 cm 3 
 
 grill. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 15 
 
 0.17 
 
 0.1128 
 
 0.0480 
 
 0.0476 
 
 0.0004+ 
 
 2 
 
 15 
 
 0.17 
 
 0.1118 
 
 0.0475 
 
 0.0476 
 
 0.0001- 
 
 (3) 
 
 16 
 
 0.17 
 
 0.1112 
 
 0.0473 
 
 0.0476 
 
 0.0003- 
 
 (4) 
 
 25 
 
 0.26 
 
 0.1860 
 
 0.0791 
 
 0.0794 
 
 0.0003- 
 
 5) 
 
 25 
 
 0.26 
 
 0.1866 
 
 0.0794 
 
 0.0794 
 
 0.0000 
 
 6 
 
 35 
 
 0.34 
 
 0.2634 
 
 0.1120 
 
 0.1111 
 
 0.0009+ 
 
 (7) 
 
 36 
 
 0.34 
 
 0.2603 
 
 0.1107 
 
 0.1111 
 
 0.0004- 
 
170 
 
 APPLICATION OF IODINE IN THE 
 
 added, and the operation completed as described. It was 
 found that barium hydroxide and potassium hydroxide may be 
 applied with equal accuracy to the analysis of both hydrochloric 
 and sulphuric acids. Some of the results obtained are given 
 in Tables V, VI and VII. 
 
 TABLE VI. 
 
 ANALYSES OP ^ HYDROCHLORIC ACID SOLUTION. 
 
 (By adding to excess of ^ KOH, boiling with excess of iodine to decoloration, 
 and acidifying the residue.) 
 
 Exp. 
 
 HC1 
 
 taken. 
 
 KOH 
 
 taken. 
 
 KOH 
 
 neutralized 
 by HC1. 
 
 HC1 found. 
 
 Mean. 
 
 Variation. 
 
 
 cm 8 . 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 20 
 
 0.14 
 
 0.0972 
 
 0.0633 
 
 0.0633 
 
 0.0000 
 
 (2) 
 
 20 
 
 0.14 
 
 0.0975 
 
 0.0634 
 
 0.0633 
 
 0.0001+ 
 
 (3) 
 
 25 
 
 0.14 
 
 0.1222 
 
 0.0795 
 
 0.0791 
 
 0.0004+ 
 
 (4) 
 
 25 
 
 0.14 
 
 0.1207 
 
 0.0785 
 
 0.0791 
 
 0.0006- 
 
 TABLE VII. 
 
 ANALYSES OF J SULPHURIC ACID SOLUTION. 
 
 (By adding to excess of ^ Ba(OH) 2 , boiling with excess of iodine to decolora- 
 tion, and acidifying the residue.) 
 
 Eip. 
 
 as 
 
 BaOH 2 
 taken. 
 
 Ba(OH), 
 neutralized 
 by H 2 80 4 . 
 
 H 2 80 4 
 found. 
 
 Mean. 
 
 Variation. 
 
 
 cm 8 
 
 grm. 
 
 gTIH. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 10 
 
 0.21 
 
 0.0884 
 
 0.0506 
 
 0.0498 
 
 0.0008+ 
 
 (2) 
 
 10 
 
 0.21 
 
 0.0880 
 
 0.0503 
 
 0.0498 
 
 0.0005+ 
 
 (3) 
 
 15 
 
 0.30 
 
 0.1328 
 
 0.0754 
 
 0.0748 
 
 0.0006+ 
 
 (4) 
 
 15 
 
 0.30 
 
 0.1313 
 
 0.0751 
 
 0.0748 
 
 0.0003+ 
 
 (5) 
 
 25 
 
 0.43 
 
 0.2168 
 
 0.1239 
 
 0.1246 
 
 0.0007- 
 
 (6) 
 
 30 
 
 0.43 
 
 0.2600 
 
 0.1481 
 
 0.1495 
 
 0.0014- 
 
 An analysis of the hydrochloric acid solution by the Grb'ger 
 method, which was found to agree in every case with the 
 gravimetric determination, gave for each 25 cm 8 0.0801 grm. of 
 HC1, agreeing with 0.0794 grm. and 0.0791 grm. obtained by 
 the new method. An analysis of the sulphuric acid solution 
 
ANALYSIS OF ALKALIES AND ACIDS. 171 
 
 by the Groger method gave for each 25 cm 3 0.1241 grm. of 
 H 2 SO 4 agreeing with 0.1246 grm. obtained by the new method. 
 This investigation shows that the reaction between iodine 
 and hydroxides of the alkalies and alkaline earths in hot 
 solution is regular and complete under analytical conditions, 
 not being appreciably affected by the mass action of considerable 
 excesses of iodine. The reaction is best applied in analysis by 
 titrating the alkali with an excess of iodine, removing this 
 excess by boiling, and estimating the iodine in the residue. 
 While certain mechanical difficulties may effect the extreme 
 accuracy of the process as a direct means for analyzing alkalies, 
 the action is at all times regular and may be indirectly applied 
 with fair accuracy to the analysis of various acids and possibly 
 to other compounds. The reaction between iodine and alkali 
 carbonates on the contrary is irregular and cannot be made 
 the basis of any analytical process. 
 
XXII 
 
 THE ESTIMATION OF BORIC ACID. 
 
 BY F. A. GOOCH AND LOUIS CLEVELAND JOKES.* 
 
 THE estimation of boric acid by treating the salts of that 
 acid with sulphuric acid, distilling with methyl alcohol, 
 evaporating the distillate over magnesium oxide, igniting and 
 weighing, was proposed by Rosenbladt-t A little later, and 
 without knowledge of Rosenbladt's experience, a somewhat 
 similar process, J which consisted in the treating of the 
 compound of boric acid with acetic acid or nitric acid, 
 distillation with methyl alcohol, evaporation of the distillate 
 over calcium oxide, and ignition of the residue, was described 
 by one of us. In the course of the development of this process, 
 it transpired that the insolubility of magnesium oxide retards 
 the absorption of boric acid by that substance, and that the 
 more soluble calcium oxide retains boric acid more actively and 
 is therefore to be preferred. 
 
 Points in the treatment upon which special emphasis was 
 laid in the original description of this process were the choice 
 of a suitable apparatus for the distillation, the employment of 
 a loosely stoppered receiver for the reception of the distillate 
 upon slaked lime, the careful removal of water from the 
 substance in the retort before acidifying and treating with the 
 methyl alcohol, regulated use of acid, and care in the evapora- 
 tion and ignition. 
 
 The attainment of good results in this process depends upon 
 attention to details. Modifications have been suggested by 
 several investigators. Thus, instead of igniting the calcium 
 oxide in a large platinum crucible, transferring it to the 
 
 * From Am. Jour. Sci., vii, 34. t Zeitschr. anal. Chem., xxvi, 21. 
 
 J Am. Chem. Jour., ix, 23. 
 
ESTIMATION OF BORIC ACID. 173 
 
 receiver to hold the boric acid, and returning the calcium oxide 
 with the distillate to the same crucible for subsequent ignition 
 of the residue, as was originally proposed, Penfield* prefers 
 to ignite the calcium oxide in a small crucible, to collect the 
 distillate hi ammoniacal water, to evaporate the latter over the 
 calcium oxide in a large platinum dish, and to transfer this 
 residue back to the small crucible for the final evaporation and 
 ignition. Kraut f suggests a modification of form in the 
 apparatus with no other essential change hi conditions. 
 MoissanJ has suggested changes in the apparatus and avoids 
 a transfer of the calcium oxide collecting the distillate by 
 itself in a closed receiver, trapped with an ammonia bulb to 
 prevent the escape of the boric acid from the distillate; 
 furthermore, Moissan's process calls for the use of an amount 
 of calcium oxide from fifteen to twenty tunes greater than 
 that theoretically required. From our experience it seems 
 obvious that the demand for this amount of calcium oxide 
 arises from an excessive use of nitric acid in the retort and the 
 consequent modification of conditions in the distillate. For- 
 tunately this difficulty may be avoided by the use of a little 
 phenolphthalein as an indicator in the retort and care to limit 
 the addition of nitric acid to the amount required to produce 
 distinct acidity. The addition of a drop of the acid and 
 another of the indicator should be repeated once or twice 
 during the distillation to insure the return of the volatilized 
 acid to the salt slightly decomposed in the process. The 
 effect of much nitric acid is bad, not only because it neu- 
 tralizes the calcium oxide when it passes to the distillate, 
 but because when it is used a tendency is developed on the 
 part of the dried mixture of calcium hydroxide and borate to 
 puff explosively if the ignition is begun as soon as the residue 
 is dry. If the residue is heated gradually and as strongly as 
 possible over a radiator before the flame is actually applied 
 to the crucible, no such action takes place ; we are disposed to 
 attribute it to the effect of the nitrate and nitrite, produced by 
 
 * Am. Jour. Sci., xxxiv, 222. t Zeitschr. anal. Chem., xxxvi, 3. 
 
 t Comp. rend., cxvi, 1084. 
 
174 
 
 ESTIMATION OF BORIC ACID. 
 
 the absorption of nitrous fumes in the lime, upon the alcohol 
 or other organic matter retained by the lime hi the evaporation 
 and drying unless the latter process is prolonged at high 
 temperature. 
 
 That good results may be obtained with small amounts of 
 calcium oxide, provided care as to the use of nitric acid and 
 the conditions of ignition be taken, is shown by the figures 
 of the original description and by the following experiments, 
 in which phenolphthalein was employed as an indicator and 
 the residue heated strongly over the radiator before actual 
 ignition. 
 
 CaO taken. 
 
 B 2 O 8 taken. 
 
 B 2 3 found. 
 
 Error. 
 
 grin. 
 
 2.3405 
 1.7620 
 2.1757 
 2.5656 
 
 grm. 
 0.1788 
 0.1790 
 0.1824 
 0.1788 
 
 grm. 
 0.1792 
 0.1785 
 0.1840 
 0.1786 
 
 grm. 
 0.0004+ 
 0.0005- 
 0.0016+ 
 0.0002- 
 
 These results are accurate within reasonable limits. On 
 the other hand, without care to ignite gradually we have 
 noted errors of from 0.0030 grm. to 0.0060 grm. in the process 
 otherwise conducted similarly. Doubtless the use of large 
 amounts of calcium oxide as suggested by Moissan may serve 
 the purpose of diffusing the explosive mixture through a 
 mass of inert matter sufficient to prevent violent puffing, but 
 care to heat over the radiator as strongly as possible before 
 opening the flame directly to the crucible answers the same 
 end. The difficulty does not exist when acetic acid is used 
 in place of nitric acid, though even in this case it is safer to 
 use the radiator in the first stages of heating, thus avoiding 
 the danger of mechanical loss by too rapid ignition. 
 
 Following are determinations made by this method with the 
 use of acetic acid. The results of these experiments, as 
 well as those of the investigators mentioned, are a sufficient 
 answer to the criticism of Reischle,* that acetic acid and 
 nitric acid do not liberate boric acid in the distillation pro- 
 * Zeitschr. anal. Chem., xxvi, 512. 
 
ESTIMATION OF BORIC ACID. 
 
 175 
 
 CaO taken. 
 
 B 2 3 taken. 
 
 B,O 3 found. 
 
 Error. 
 
 gnu. 
 0.9977 
 1.0220 
 1.3717 
 1.1310 
 
 gnu. 
 0.2065 
 0.2067 
 0.2077 
 0.1791 
 
 gnu. 
 0.2062 
 0.2070 
 0.2075 
 0.1795 
 
 gnu. 
 0.0003- 
 0.0003+ 
 0.0002- 
 0.0004-1- 
 
 cess so that good results may be obtained. Moreover, it has 
 been shown by one of us * that even carbonic acid is strong 
 enough to bring about complete volatility of boric acid with 
 methyl alcohol. 
 
 The use of Calcium Oxide as a Retainer. 
 
 Quite recently Thaddeeff f has advocated the abandonment 
 of calcium oxide as an agent for holding boric acid in the 
 evaporation of alcoholic and aqueous solutions, on account 
 of the hygroscopic nature of the oxide and the consequent 
 difficulty of securing it in definite conditions for weighing, 
 and proposes, instead of using calcium oxide, to retain and 
 estimate boric acid in solution by converting it into the form 
 of potassium borofluoride. 
 
 In the final modification of ThaddeefFs method the proposal 
 is made to liberate the boric acid from its compounds by 
 sulphuric acid, to volatilize it in methyl alcohol with the 
 aid of a current of dry air, to catch the distillate in potassium 
 hydroxide, to treat the mixture of hydroxide and borate with 
 hydrofluoric acid in excess and evaporate on the steam bath, 
 to digest the residue of fluoride and borofluoride at normal 
 temperatures for two hours with 50 cm 3 of a potassium 
 acetate solution (sp. gr. 1.14) and for twelve hours more after 
 adding 100 cm 3 of ethyl alcohol (sp. gr. 0.805), to filter on 
 paper, wash the residue with 62-72 cm 3 of alcohol (sp. gr. 
 0.805), dry at 100 and weigh as potassium borofluoride, 
 after which the borofluoride is to be dissolved in boiling water 
 and tested with calcium chloride for possible contamination 
 
 * Jones, Am. Jour. Sci., v, 442. This volume, p. 100. 
 t Zeitschr. anal. Chem., xxxvi, 568. 
 
176 ESTIMATION OF BORIC ACID. 
 
 by the presence of a fluoride. Plainly ThaddeefFs procedure 
 presents at the outset difficulties; for besides the incon- 
 venience of conducting long digestions with reagents of 
 regulated strength, the difficulty of procuring hydrofluoric 
 acid free from silica, which if present (as it usually is hi the 
 so-called chemically pure hydrofluoric acid of commerce) 
 would be retained in the borofluoride as potassium fluosilicate, 
 the inaccuracy of the dried paper filter, and the obvious 
 uncertainty of success in an attempt to wash a mixture of 
 acid potassium fluoride and potassium borofluoride in potas- 
 sium acetate and alcohol so that the one shall be rendered 
 entirely soluble while the other remains sensibly unaffected, 
 besides these objections, there is the theoretical probability 
 that boric acid mnst be lost by volatilization during the 
 evaporation of the solution of the mixed salts in the presence 
 of free hydrofluoric acid. This last point was put to the 
 proof by submitting to distillation in a platinum retort a 
 mixture of equal quantities of borax and potassium hydroxide 
 with an excess of hydrofluoric acid, collecting the distillate 
 in potassium hydroxide, evaporating it to dryness and testing 
 it for the presence of boric acid. When this residue from the 
 evaporated distillate was treated with sulphuric acid and 
 methyl alcohol, the burning alcohol vapor gave plainly 
 the green flame of boric acid. Another portion showed 
 clearly the presence of boric acid when acidulated with 
 hydrochloric acid tested with turmeric paper. No boric acid 
 could be detected in any of the reagents used. It is plain, 
 therefore, that boric acid does volatilize upon the evapora- 
 tion of a mixture of potassium fluoride and borofluoride in 
 acid solution. The amount of such loss is disclosed hi the 
 record of the following experiment. Portions of a standard 
 solution of boric acid, prepared by dissolving a known weight 
 of anhydrous boric oxide in a liter of water, were mixed 
 with a solution of potassium hydroxide (free from silica and 
 standardized by conversion to the chloride) in the proportions 
 to form the potassium borofluoride, and an excess of hydro- 
 fluoric acid was added. The mixture was evaporated and 
 
ESTIMATION OF BORIC ACID. 
 
 177 
 
 the residue was dried and weighed at 100, the whole opera- 
 tion being conducted in platinum. 
 
 Exp. 
 
 HKP, 
 equivalent to 
 KOH taken. 
 
 taken. 
 
 KPBFs 
 theoretical 
 weight. 
 
 KFBF 8 
 found. 
 
 Error in 
 terms of 
 KFBF 8 . 
 
 Error in 
 terms of 
 8,03. 
 
 
 grm. 
 
 grin. 
 
 grm. 
 
 grin. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.3531 
 
 0.1582 
 
 0.5701 
 
 0.5580 
 
 0.0121- 
 
 0.0033- 
 
 (2) 
 
 0.3192 
 
 0.1430 
 
 0.5154 
 
 0.5100 
 
 0.0054- 
 
 0.0015- 
 
 (3) 
 
 0.3192 
 
 0.1430 
 
 0.5154 
 
 0.5030 
 
 0.0124- 
 
 0.0034- 
 
 (4) 
 
 0.3192 
 
 0.1430 
 
 0.5154 
 
 0.5088 
 
 0.0066- 
 
 0.0018- 
 
 (6) 
 
 0.3192 
 
 0.1430 
 
 0.5154 
 
 0.5114 
 
 0.0040- 
 
 0.0011- 
 
 In experiments (1) to (3) the volume of the solution 
 evaporated was about 50 cm 3 . In experiment (4) this 
 volume was reduced about one-half before acidifying with 
 hydrofluoric acid, while in experiment (5) the solution was 
 diluted about one-half before adding the hydrofluoric acid. 
 It is plain, therefore, that in this single step of Thaddeeff's 
 process there is a considerable error of deficiency. On the 
 other hand, the errors for the full process as laid down by 
 Thaddeeff have been in our experience invariably differences 
 of excess presumably because the loss due to volatilization 
 of boric acid has been overbalanced by the inaccuracy in 
 washing. It is plain that the process can give true indica- 
 tions only by the balancing of considerable errors. 
 
 If we take into consideration, therefore, the inevitable 
 inaccuracy and inconvenience of Thaddeeff's proposal, it 
 cannot be regarded as a desirable substitute for the process 
 according to which boric acid is absorbed and retained for 
 weighing with calcium oxide, especially since the difficulties 
 in the way of getting constant weights of that substance are 
 by no means insuperable. 
 
 Thus the following table shows the series of weights taken 
 in several experiments in bringing calcium oxide to a constant 
 weight in a 50 cm 3 platinum crucible ignited over a blast lamp, 
 as well as the weight taken after adding a known amount of 
 standard boric acid solution to the slaked oxide, evaporating, 
 and igniting. The results recorded are those of experiments 
 
 VOL. II. 12 
 
178 
 
 ESTIMATION OF BORIC ACID. 
 
 made on days not moist beyond the average and with the 
 greatest care to approach the limit of accuracy with which 
 calcium oxide and the boric acid held thereby can be weighed 
 under ordinarily favorable conditions. The first weight of 
 calcium oxide recorded under each experiment was taken after 
 a strong ignition over the blast lamp for about one-half hour. 
 The succeeding weights were taken after similar ignitions of 
 five minutes. In all cases the crucible was left to stand a 
 definite period in a sulphuric acid desiccator, and, after the 
 approximate value had once been obtained, the weights of 
 the preceding weighing were replaced on the balance before 
 the crucible was taken from the desiccator. The average of the 
 weights bracketed is the weight taken as constant for the 
 calculations. 
 
 Exp. 
 
 CaO taken. 
 
 B,0 8 
 taken. 
 
 CaO + B 2 8 
 taken. 
 
 CaO + B 2 O 8 
 found. 
 
 Error. 
 
 (1) 
 
 grm. 
 0.9505 
 0.9493 I ft 0400 
 0.9493 f U<W 
 
 grm. 
 
 0.2095 
 
 grm. 
 
 1.1588 
 
 grin. 
 
 1.1590 ) i 1591 
 
 1.1591 f L1591 
 
 grm. 
 
 0.0003+ 
 
 (2) 
 
 1.1319 
 1.1317 ) 
 1.1313 } 1.1315 
 1.1315 ) 
 
 0.2150 
 
 1.3465 
 
 1.3499 
 1.3474 ) 
 1.3475 } 1.3475 
 1.3476 ) 
 
 0.0010+ 
 
 (3) 
 
 0.8028 
 0.8025) ft o o 4 
 0.8024 f ' 8024 
 
 0.1184 
 
 0.9208 
 
 0.9205 
 0.9206 ) ft oonft 
 0.9206 f ' 9206 
 
 0.0002- 
 
 (4) 
 
 2.6980 
 2.6975 
 2.6973 ) 9 fiQ7 o 
 2.6973 I 2 
 
 0.2073 
 
 2.9046 
 
 2.9043 
 2.9049 Uqo4R 
 2.9044 ] ^ >yu4b 
 
 0.0002+ 
 
 Obviously calcium oxide may be weighed with accuracy, 
 with or without boric acid ; but the fact remains that a less 
 hygroscopic absorbent one requiring less care in the handling, 
 is to be desired. 
 
 The use of Sodium Tungstate as a Retainer. 
 
 In searching for a suitable material of less hygroscopicity to 
 replace calcium oxide as a retainer for boric acid, we have 
 found that sodium tungstate, fused with a slight excess of 
 
ESTIMATION OF BORIC ACID. 
 
 179 
 
 tungstic acid over that contained in the normal tungstate 
 (to insure its freedom from carbonate), answers this purpose 
 excellently. This substance is definite in weight, not hygro- 
 scopic, soluble in water, and recoverable in its original weight 
 after evaporation and ignition. To test its value as a retainer 
 for boric acid, portions of it 4 to 7 grm. were fused and 
 weighed in a 50 cm 3 crucible, the tungstate was dissolved in 
 water and to it was added a known amount of a standard 
 solution of boric acid. After diluting, mixing, evaporating, 
 and fusing the residue, the increase in weight should represent 
 the boric anhydride held by the tungstate. The results of the 
 accompanying table show how accurately the boric acid is 
 retained under these conditions. In experiments (3) to (7) 
 the tungstate, after its first weighing, was dissolved, transferred 
 to a larger platinum dish and mixed therein with the boric 
 acid. After evaporation to a suitable volume this solution of 
 tungstate and boric acid was transferred to the original crucible 
 for final evaporation and ignition. 
 
 Exp. 
 
 NajWOi + W0 3 
 taken. 
 
 B 2 8 taken. 
 
 B,0 3 found. 
 
 Error in B 2 3 . 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 6.6416 
 
 0.1784 
 
 0.1771 
 
 0.0013- 
 
 (2) 
 
 7.3134 
 
 0.1786 
 
 0.1773 
 
 0.0013- 
 
 (3) 
 
 5.5003 
 
 0.0950 
 
 0.0952 
 
 0.0002-f- 
 
 (4) 
 
 4.1394 
 
 0.0944 
 
 0.0944 
 
 0.0000 
 
 (5) 
 
 7.5037 
 
 0.2148 
 
 0.2149 
 
 0.0001+ 
 
 (6) 
 
 4.7744 
 
 0.2718 
 
 0.2702 
 
 0.0016- 
 
 (7) 
 
 6.6470 
 
 0.2503 
 
 0.2487 
 
 0.0016- 
 
 It is plain that though the sodium tungstate does not hold 
 the boric acid with absolute accuracy the errors are not 
 unreasonable 0.0008 grm. in the mean. Upon substituting 
 the tungstate for calcium oxide as a retainer in the distillation 
 process, the results were likewise highly favorable. 
 
 We used by preference the apparatus originally proposed, 
 excepting that the Erlenmeyer flask used as a receiver is fitted 
 tightly to the condenser and trapped with water bulbs. The 
 retort is made very easily from a 150 cm 3 pipette and has 
 
180 
 
 ESTIMATION OF BORIC ACID. 
 
 the special advantage that particles of the residue spattering 
 during distillation are easily washed from the walls of the 
 vessel by a slight rotary motion of the retort. It was found 
 that special care should be taken to give the tungstate ample 
 time for contact with the distillate before exposing the latter 
 to atmospheric evaporation. The distilkte was received, 
 therefore, in a dilute solution of sodium tungstate placed in 
 the receiver, cooled by ice and trapped with water, and the 
 mixture was well stirred, allowed to stand one half-hour, 
 evaporated to small volume in a large dish, and transferred 
 to the crucible in which the tungstate had been originally 
 weighed. After thorough drying the residue was ignited to 
 fusion and weighed. When acetic acid was employed in the 
 retort, care was taken in the ignition to expose the fused mass 
 freely to the air (by causing it to flow upon the sides of the 
 crucible) until the color of the cooled tungstate was white, in 
 order that the reducing effect of the acetate might be eliminated. 
 In the experiments recorded in the following table the tungstate 
 
 Na,W0 4 + WO S 
 
 B,0 3 taken. 
 
 B 2 3 found. 
 
 Error. 
 
 WITH NITRIC ACID. 
 
 gnu. 
 
 8.5516 
 4.9639 
 8.0033 
 
 grin. 
 
 0.1582 
 0.1329 
 0.1267 
 
 grm. 
 
 0.1572 
 0.1323 
 0.1256 
 
 grm. 
 
 0.0010- 
 0.0006- 
 0.0011- 
 
 WITH ACETIC ACID. 
 
 4.9658 
 6.0289 
 4.6797 
 4.0013 
 
 0.1434 
 0.1431 
 0.1589 
 0.1433 
 
 0.1418 
 0.1433 
 0.1587 
 0.1422 
 
 0.0016- 
 0.0002+ 
 0.0002- 
 0.0011- 
 
 WITH SULPHURIC ACID. 
 
 6.3439 
 8.8227 
 10.1516 
 6.5738 
 
 0.1582 
 0.1582 
 0.1265 
 0.1392 
 
 0.1579 
 0.1577 
 0.1264 
 0.1390 
 
 0.0003- 
 0.0005- 
 0.0001- 
 0.0002- 
 
 was used in the receiver to retain the boric acid distilled as 
 usual with methyl alcohol, from the borates treated with acetic 
 
ESTIMATION OF BORIC ACID. 181 
 
 acid, nitric acid or sulphuric acid, in amounts regulated by the 
 use of phenolphthalein as an indicator. 
 
 Excessive use of acid is disadvantageous, and this is 
 especially true in the case of sulphuric acid; for, if this 
 acid is carried over with the methyl alcohol, as it is at 100 
 if present in appreciable excess, a part of it, at least, is held 
 permanently by the tungstate to increase the apparent weight 
 of the boric acid to be estimated. 
 
 The manipulation of the tungstate presents no difficulties, 
 and the results obtained by its use are reasonably accurate. 
 
XXIII 
 
 A VOLUMETRIC METHOD FOR THE ESTIMATION 
 OF BORIC ACID. 
 
 BY LOUIS CLEVELAND JONES.* 
 
 WHEN boric acid and mannite are mixed in solution a peculiar 
 compound of strongly acid properties is the result. This com- 
 pound decomposes carbonates, and its acid taste is comparable 
 to that of citric acid, much stronger than that of boric acid 
 alone. Magnaninit has found that the product of such a 
 mixture of boric acid and mannite solutions shows greater 
 electrical conductivity and a lower freezing point than a simi- 
 lar molecular solution of either substance alone. Other poly- 
 atomic alcohols (but all to a less degree than mannite) and 
 some organic acids show this peculiar property of combining 
 chemically with boric acid to increase its acid qualities.^: Of 
 this reaction between boric acid and the polyatomic alcohols, 
 Thomson, Barthe,|| and Jb'rgensen^" have taken advantage 
 to develop methods for the volumetric estimation of boric acid. 
 Glycerine is used to form a combination with boric acid, suffi- 
 ciently acidic to give an acid reaction when used with a sensi- 
 tive indicator and make possible its titration with an alkaline 
 solution. Honig and Spitz** show that in the method of 
 Jb'rgensen a very large amount of glycerine must be used to 
 prevent the appearance of the indication of alkalinity with phe- 
 nolphthalein before all the boric acid is neutralized according 
 to the following equation, 2NaOH+B 2 O 3 = 2NaOBO+H 2 O ; 
 
 * From Am. Jour. Sci., vii, 147. 
 t Gaz. Chim., xx, 428-440; xxi, 134-145. 
 
 t Klein, Jour. Pharm. Chim., 4, vol. xxviii ; Lambert, Comp. rend., cviii, 
 1016-1017. 
 
 Jour. Soc. Chem. Ind., xv, 432. || Jour. Pharm. Chim., xxix, 163. 
 IT Zeitschr. angew. Chem., 1897, 5. ** Zeitschr. angew. Chem. (1896), 649. 
 
ESTIMATION OF BORIC ACID. 183 
 
 that in the presence of carbonates the solution must be boiled 
 to decompose bicarbonates and the escape of boric acid by vol- 
 atilization prevented by the use of a return condenser; and 
 that silica must be removed by the process of Berzelius, and 
 the solution then neutralized by the use of methylorange 
 before a titration of the boric acid can be made. 
 
 Vadam,* for the estimation of boric acid in butter makes 
 use of mannite, which, as he finds, gives sharper indication 
 with litmus than glycerine. According to this process, the 
 solution to be analyzed for boric acid is neutralized by the use 
 of litmus and a solution of sodium hydroxide. Mannite 
 (1-2 grm.) is then added, bringing about an acid reaction with 
 the boric acid present in free condition. The solution is then 
 titrated to alkalinity by sodium hydroxide. 
 
 None of the above methods with glycerine have, in my 
 experience, given anything but comparatively crude results. 
 The weak acidic properties of boric acid, the interference (and 
 difficulty of removal) of carbon dioxide with all organic indi- 
 cators sufficiently delicate to be used with boric acid, and 
 indeed, the procuring of a standard alkali containing no car- 
 bonate, together with the supposed detrimental influence of 
 silica and the lack of a convenient method for its removal, 
 have made the process of Gooch,f which involves distillation 
 and weighing with calcium oxide, the only means (though 
 requiring long time and exceeding care) in use for the accu- 
 rate separation and estimation of boric acid. Recently sodium 
 tungstate has been recommended from this laboratory^ as 
 a substitute for calcium oxide to retain the distilled boric acid. 
 The entire process, however, is one of the most exacting in 
 analytical chemistry, and for this reason a convenient, rapid 
 and at the same time accurate method for the estimation of 
 boron is especially desirable. The first step toward the 
 development of such a process must be the convenient prepa- 
 
 * Jour. Pharm. Chim. (6), viii, 109-111. 
 
 t Am. Chem. Jour, ix, 23-33; Moissan, Comp. rend., cxri, 1087; Kraut, 
 Zeitschr. anal. Chem. xxxvi, 165 ; Montemartini, Gaz. Chim. Ital., xxviii, 1, 
 344. 
 
 t Gooch and Jones, Am. Jour. Sci., rii, 34. This volume, p. 172. 
 
184 A VOLUMETRIC METHOD FOR THE 
 
 ration and the accurate estimation of the standard solution of 
 alkali to be used for neutralizing the boric acid. This has been 
 found to be easily accomplished by the process recommended 
 by Kusler.* This observer, in an extensive investigation of 
 the analytical methods for the volumetric estimation of alkalies 
 and alkali carbonates in solution, finds that both phenolphthal- 
 ein and methylorange are appreciably sensitive to carbonic acid, 
 but when this interfering agent is removed by precipitation 
 with barium chloride according to the process of Winkler,f the 
 remaining free alkali may be estimated with great accuracy by 
 phenolphthalein and decinormal hydrochloric acid. 
 
 Obviously, if the difficulties dependent upon the action of 
 carbon dioxide can be obviated, and if the acidity of the boric 
 acid can be increased to such an extent that a sufficiently sen- 
 sitive indicator will give with accuracy the neutralization 
 point with free alkali, and if the alkali and stronger acid can 
 be combined while boric acid alone remains free, then it 
 should be possible to estimate boric acid volumetrically. Ex- 
 periment has shown that barium chloride removes carbon 
 dioxide present in carbonates, and that mannite makes a com- 
 bination with boric acid strongly acidic to phenolphthalein. 
 
 To obtain the boric acid alone in free condition many 
 attempts have been made. Gladding,f Thaddeeff and Rosen- 
 bladt || have isolated the boric acid by distillation with methyl- 
 alcohol and a non-volatile acid. Many indicators theoretically 
 insensible to free boric acid have been used to indicate the 
 neutralization of the stronger acids. Honig and Spitz,*[f and 
 Thomson,** use methylorange, Morse and Burton,|t tropae- 
 olin 00, while Vadam JJ makes use of litmus. All these 
 indicators, however, have been found by experiment to be 
 more or less affected by boric acid in solution. On the other 
 hand, I have found in the well known reaction according 
 
 * Zeitschr. anorg. Chem., xiii, 124-150. t Massanalyse. 
 
 t Jour. Am. Chem. Soc., iv, 568. 
 
 Zeitschr. anal. Chem., xxxvi (9), 568. 
 
 || Zeitschr. anal. Chem., xxvi, 18. f Zeitschr. anorg. Chem. (18), 549. 
 ** Jour. Soc. Chem. Ind., xv, 432. ft Am. Chem. Jour., x, 154. 
 |J Jour. Pharm. Chim. (6) viii, 109-111. 
 
ESTIMATION OF BORIC ACID. 185 
 
 to which a stronger acid liberates regularly iodine from a 
 mixture of iodide and iodate, the solution of this difficulty. 
 If both the iodide and iodate are in excess of the acid the 
 entire amount of free acid will be neutralized and the cor- 
 responding amount of iodine liberated according to the 
 following equation: 
 
 5KI + KI03 + 6HC1 = 6KC1 + 3H 2 + 3I 2 . 
 
 This liberated iodine may be removed by sodium thiosulphate 
 and a solution obtained which is absolutely neutral containing 
 only neutral salts, potassium iodide, iodate, and tetrathionate. 
 The statements made by P. George vie* and Furry, f that 
 boric acid present in moderate amount in solution has not 
 the slightest action on a mixture of iodide and iodate, have 
 been experimentally verified. Therefore when this acid is 
 liberated by an excess of a stronger acid, and the iodine set 
 free destroyed by thiosulphate, it remains free in solution to 
 be titrated in any convenient manner possible. 
 
 Following along the lines suggested by the above reac- 
 tions, a volumetric process for the estimation of boric acid 
 has been developed. For a basis of the investigations, a 
 standard solution of boric acid was prepared by dissolving 
 in a liter of water about eight grams of carefully weighed 
 anhydrous boric oxide. This anhydrous boric oxide was 
 prepared from the several times recrystallized hydrous boric 
 acid, by long-continued fusion over a blast lamp. A solution 
 of approximately \ sodium hydroxide was prepared from the 
 ordinary sodium hydroxide of the laboratory. The free alkali 
 in this solution was estimated by the process of Winkler 
 recommended by Ktisler. The acid used to make this estima- 
 tion was hydrochloric, standardized by silver nitrate. 
 
 The full method for the estimation of boric acid as finally 
 elaborated is as follows : The solution is made clearly acid 
 to litmus by hydrochloric acid and 5 cm 3 of a solution (10%) 
 of barium chloride added. An amount of iodate and iodide 
 of potassium sufficient to liberate an amount of iodine at 
 
 * Jour, prakt. Chem., xxxviii, 118. t Am. Chem. Jour., vi, 341. 
 
186 A VOLUMETRIC METHOD FOR THE 
 
 least equivalent to the excess of hydrochloric acid in the 
 acidified solution is mixed with starch in a separate beaker, 
 and the iodine which is usually thrown out by this mixture, 
 is just bleached by a dilute solution of thiosulphate. 
 
 To the now neutral solution of iodide and iodate a single 
 drop of the solution to be analyzed is transferred by a glass 
 rod. If a blue coloration is developed, the solution is acidic 
 with hydrochloric acid, and all the boric acid is in free 
 condition. The amount of iodide and iodate used depends 
 upon the acidity of the solution containing boric acid. 
 Usually 10 cm 3 of a 25 per cent solution of iodide and the 
 same amount of a saturated solution of iodate is sufficient. 
 Any larger excess of hydrochloric acid should be neutralized 
 by sodium hydroxide before the iodide and iodate mixture is 
 added. After the addition of the iodide and iodate solution, 
 containing starch, to the boric acid solution, the liberated 
 iodine should be carefully bleached by thiosulphate. Any 
 excess of thiosulphate in reasonable amount does not seem 
 to be detrimental, but in practice the starch iodide color is 
 clearly bleached, and no more then added. Carbonates pre- 
 vent a definite indication of the neutral point by thiosulphate 
 and starch iodide, therefore the barium chloride (about 5 cm 3 ) 
 should be added before this point in the process. The 
 mixture of iodide and iodate is not added to the solution to 
 be analyzed until after it is made acidic, for the reason that 
 when the neutral point is approached by the addition of 
 hydrochloric acid the starch iodide is thrown out locally by 
 the acid, and the small amount of sodium borate remaining 
 undecomposed does not again bleach the coloration produced 
 thus obscuring the neutral point which must be obtained 
 before titrating for boric acid. 
 
 The solution after the bleaching of the iodine by thiosul- 
 phate is colorless and contains only starch, neutral chloride, 
 potassium tetrathionate, iodide and iodate, and all the boric 
 acid present in uncombined condition. The carbonate lies out 
 of the sphere of action in insoluble form as barium carbonate. 
 A few drops of the indicator, phenolphthalein, are now added, 
 
ESTIMATION OF BORIC ACID. 187 
 
 and the alkaline solution run in until a strong red coloration is 
 produced. A pinch of mannite is then added, which bleaches 
 the phenolphthalein coloration, and the alkali solution again 
 run in to a faint indication, which if permanent on the addition 
 of more mannite, may be taken as the reading point. About 
 1-2 grm. of mannite are necessary for a determination. The 
 boro-mannite compound is sufficiently acidic to liberate iodine 
 abundantly, but it appears to be a time reaction, and at 
 the end of six hours only about 95 per cent of the theoretical 
 amount (considering B 2 O 3 as a bivalent acid) has been thrown 
 out. The combination of boric acid and mannite liberates in 
 the presence of iodide and iodate immediately only about 
 one-half the iodine required on the theory that B 2 O 8 under these 
 conditions acts as a bivalent acid, or with the neutralizing 
 power of metaboric acid, HOBO. When no mannite is present 
 phenolphthalein gives an alkaline indication when only about 
 one-half the amount of alkali theoretically necessary to form 
 the metaborate, NaOBO, has been added. Obviously, then, the 
 starch iodide coloration will not appear at all on the addition 
 of mannite, if about one-half the free boric acid is first 
 neutralized by the solution of alkali, and the remainder of the 
 alkali immediately added to complete neutralization. The 
 point at which the danger of the appearance of the iodide 
 coloration on the addition of mannite has been passed, is 
 roughly indicated before the mannite has been added by the 
 appearance of the strong alkaline indication of phenolphthalein. 
 This indicator would not need to be added at all, if the 
 boromannite compound quickly and regularly liberated iodine 
 from the iodide and iodate. The fact, however, that this 
 compound of boric acid and mannite as has been ascertained 
 by experiment liberates, on standing twelve hours, about 99 
 per cent of the theoretical amount of iodine, places the strength 
 of this acid above that of citric or tartaric acid as investigated 
 by Furry.* With phenolphthalein, however, the end reaction 
 is sharp and the small amount of carbonate present in the 
 standard solution of alkali is precipitated by the barium 
 * Am. Chem. Jour., vi, 341. 
 
188 A VOLUMETRIC METHOD FOR THE 
 
 chloride already in the solution. The calculation must 
 therefore be based on the amount of free hydroxide in the 
 standard solution of alkali used, according to the following 
 representation : 
 
 B 2 8 + 2NaOH = 2NaOBO + H 2 0. 
 
 The best results and the most definite indications are 
 obtained in cold solution of a volume not greater than 50 
 cm 3 . This fact accords with the observations of Magnanini * 
 that the relative electrical conductivity of the boro-mannite 
 solution is decreased by dilution and elevation of the 
 temperature. When silicates are present in solution, the 
 silicondioxide is liberated by the excess of hydrochloric acid, 
 and this oxide, whether in hydrous or anhydrous condition, 
 neither affects the indication with iodine nor phenolphthalein, 
 nor does it form with mannite a compound of acidic proper- 
 ties. Ammonium salts interfere with the indication given 
 by phenolphthalein and may be removed by boiling with 
 potassium hydroxide in excess, or an indicator used not 
 affected by them. 
 
 To test the action of fluorides in the process, several 
 experiments were made in which hydrofluoric acid (10 cm 3 of 
 T^ solution) was introduced into the solution containing salts 
 of sodium, free hydrochloric and boric acids. Barium chloride 
 was then added and the analysis for boric acid completed in 
 the usual way without the accuracy of the results being in any 
 way interfered with by the presence of hydrofluoric acid. 
 
 Table I contains the results of a series of analyses in which 
 the boric acid was first drawn into a excess of sodium hy- 
 droxide, then estimated according to the method described. 
 
 The standard solutions of boric acid used contained, in A, 
 7.153 grm., and in B, 7.706 grm. per liter. The solution of 
 free sodium hydroxide was 0.21427 normal. 
 
 Practical tests of the method upon specimens of crude 
 calcium borate and colemanite are recorded in Table II and 
 Table III. 
 
 * Gaz. chim. ital., xx, 428, and xxi, 134. 
 
ESTIMATION OF BORIC ACID. 
 TABLE I. 
 
 189 
 
 
 Exp. 
 
 B,0 8 SoL 
 
 taken. 
 
 NaOH SoL 
 required. 
 
 SSL 
 
 B,0, 
 
 found. 
 
 Errors on 
 B 2 S . 
 
 A 
 
 f(l) 
 j 2) 
 
 1(3) 
 
 ClUg 
 
 21.95 
 20.68 
 20.73 
 
 cm s . 
 21.02 
 19.65 
 19.63 
 
 grin* 
 
 0.1671 
 0.1479 
 0.1483 
 
 grin. 
 
 0.1577 
 0.1474 
 0.1473 
 
 grm. 
 0.0006+ 
 0.0005- 
 0.0010- 
 
 B 
 
 - 
 
 (4) 
 (5) 
 (6) 
 (7) 
 (8) 
 ,(9) 
 
 23.05 
 23.10 
 22.76 
 24.08 
 22.00 
 20.78 
 
 23.71 
 23.80 
 23.35 
 24.78 
 22.60 
 21.28 
 
 0.1776 
 0.1780 
 0.1754 
 0.1855 
 0.1695 
 ' 0.1601 
 
 0.1777 
 0.1783 
 0.1750 
 0.1857 
 0.1686 
 0.1695 
 
 0.0001+ 
 0.0003+ 
 0.0004- 
 0.0002+ 
 0.0009- 
 0.0006- 
 
 TABLE n. 
 ANALYSIS OP CRUDE BORATE OF LIME. 
 
 Bxp. 
 
 Ca borate 
 taken. 
 
 B 2 O 8 found. 
 
 B 2 0, 
 
 *% 
 
 
 grm. 
 
 grm. 
 
 
 (1) 
 
 0.4016 
 
 0.2289 
 
 56.99 
 
 (2) 
 
 0.4044 
 
 0.2302 
 
 56.92 
 
 (3) 
 
 0.4000 
 
 0.2285 
 
 57.11 
 
 TABLE m. 
 ANALYSIS OP COLEMANITE. 
 
 Bxp. 
 
 Mineral taken. 
 
 B 2 O 8 found. 
 
 %B,0 8 . 
 
 Average. 
 
 
 grm. 
 
 grm. 
 
 
 
 (1) 
 
 0.4034 
 
 0.2064 
 
 61.15] 
 
 
 
 (2) 
 
 0.4070 
 
 0.2069 
 
 60.80 
 
 
 
 (3 
 
 
 0.6004 
 
 0.3054 
 
 50.86 
 
 
 ff\ QQOf 
 
 (4 
 
 
 0.6006 
 
 0.3056 
 
 50.89 
 
 
 ou.yy% 
 
 (5 
 
 
 0.5059 
 
 0.2592 
 
 61.24 
 
 
 
 6 
 
 
 0.5092 
 
 0.2592 
 
 60.89 J 
 
 
 
 The finely-ground minerals were dissolved in hydrochloric 
 acid and the analyses proceeded with as described above. 
 
 An analysis for boric acid by this process can be completed 
 in five minutes and the results are obviously accurate within 
 the limits of ordinary analysis. 
 
 The usually interfering substances, fluorine, silica, and 
 carbon dioxide, have no detrimental influence on the results 
 of this process. 
 
XXIV 
 
 THE CONSTITUTION OF THE AMMONIUM 
 MAGNESIUM PHOSPHATE OF ANALYSIS. 
 
 BY F. A. GOOCH AND MARTHA AUSTIN.* 
 
 IN a recent paper from this laboratory f it has been shown 
 that the presence of ammonium chloride or other ammonium 
 salt is necessary in the precipitation of manganese as the 
 ammonium manganese phosphate by microcosmic salt in 
 order that the precipitate may have the ideal constitution 
 represented by the symbol NH 4 MnPO 4 . 
 
 It was also shown that the solvent effect of the ammonium 
 chloride upon the precipitated ammonium manganese phosphate 
 is not marked when an excess of the precipitant is present in 
 solution. 
 
 The relations disclosed in this paper suggest that the 
 chemical constitution of the precipitate rather than mechanical 
 contamination and varying solubility the explanations gene- 
 rally accepted, and, indeed, advocated by one of us in a former 
 paper J may be responsible for observed variations in the 
 weight of the residue derived by the ignition of the similar 
 salt of magnesium, the ammonium magnesium phosphate, 
 precipitated by an excess of a soluble phosphate from the 
 solution of a magnesium salt, or from the solution of a soluble 
 phosphate by an excess of a magnesium salt. 
 
 Precipitation by Excess of the Soluble Phosphate. 
 
 The precipitation of the magnesium salt by an excess of the 
 soluble phosphate was first studied. For this work a solution 
 of pure magnesium nitrate was prepared by dissolving the 
 
 * From Am. Jour. Sci., vii, 187. 
 
 t Am. Jour. Sci., vi, 233. This volume, p. 121. 
 
 J Am. Chem. Jour., i, 391. 
 
AMMONIUM MAGNESIUM PHOSPHATE. 
 
 191 
 
 pure magnesium oxide of commerce in a slight excess of pure 
 hydrochloric acid and boiling with more magnesium oxide. 
 After filtering off the excess of magnesium oxide and any trace 
 of iron or members of the higher groups, the solution was 
 precipitated by ammonium carbonate, the precipitate was 
 washed by repeated boilings and nitrations until silver nitrate 
 gave no precipitate in the solution acidified with nitric acid. 
 This precipitated carbonate was nearly dissolved in nitric acid 
 and the solution was boiled with an excess of the carbonate 
 (for the purpose of removing traces of barium, strontium, 
 and calcium) filtered, and diluted to definite volume. The 
 evaporation of a definite volume of the solution and strong 
 ignition of the residue would be a most natural method of 
 establishing a standard of the solution, were it not for the 
 fact, pointed out by Richards and Rogers,* that the oxide of 
 magnesium retains on ignition occluded nitrogen and oxygen 
 enough to increase its weight sensibly. For this reason the 
 nitrate was converted to the sulphate and weighed as such 
 either by evaporating to dryness in a weighed platinum 
 crucible a definite volume of the solution, igniting as oxide, 
 and changing to the sulphate by heating with sulphuric acid ; 
 or, by evaporating the magnesium nitrate directly with an 
 excess of sulphuric acid of half strength. In this treatment 
 the excess of acid was removed by heating the platinum 
 crucible upon a porcelain ring or triangle so placed within a 
 porcelain crucible that the bottom and walls of the inner 
 crucible were distant about one centimeter from the bottom 
 
 TABLE I. 
 
 MgS0 4 obtained by 
 converting ignited 
 MgO into the sulphate. 
 
 MgSO 4 obtained 
 directly from 50 cm 3 
 Mg(N0 3 ) s . 
 
 Theoretical 
 amount of MgO 
 in MgS0 4 . 
 
 grm. 
 0.5748 
 0.6739 
 
 grm. 
 
 0.5741 
 0.5750 
 
 grin. 
 
 0.1924 
 0.1923 
 0.1922 
 0.1925 
 
 * Amer. Chem. Jour., xvi, 567. 
 
192 CONSTITUTION OF THE AMMONIUM 
 
 and walls of the outer crucible. The excess of acid is easily 
 removed in this way, and the outer crucible may be heated 
 to redness without danger of breaking up the magnesium 
 sulphate. The results of this work, taking O = 16, Mg = 24.3, 
 N = 14.03, S = 32.06, are given in the accompanying table. 
 
 The magnesium oxide obtained by direct ignition of the 
 nitrate weighed on the average about 0.0010 grm. more than 
 the oxide theoretically present in the weighed sulphate from 
 equal portions of the solution. 
 
 Before proceeding to study possible chemical effects of 
 ammonium chloride in determining the constitution of the 
 ammonium magnesium phosphate, it is obviously necessary to 
 define the extent to which the ammonium salt may exert a 
 solvent action in presence of the precipitant. Fresenius 
 estimated that ammonium magnesium phosphate is soluble in 
 15293 parts of cold water, but the method of investigation 
 employed did not entirely preclude the possibility of counting 
 as ammonium magnesium phosphate soluble material included 
 and held in the original precipitate.* According to Kissel f 
 the phosphate, which dissolves in a mixture of ammonia and 
 water in the proportion of 0.0040 grams to the liter and in the 
 proportion of 0.0110 grams to the liter in a similar mixture 
 containing also 18 grm. of ammonium chloride, is practically 
 insoluble in the latter mixture if an excess of magnesia 
 mixture be added; and HeintzJ showed that the effect of 
 adding an excess of sodium phosphate in the solution is 
 similar. 
 
 So far as appears, no quantitative experiments have been 
 recorded hi which the behavior of a mixture of ammonium 
 chloride and magnesium salt and an insoluble phosphate in a 
 solution only slightly ammoniacal has been tested, though as a 
 matter of convenience the use of faintly ammoniacal solutions 
 and faintly ammoniacal washwater is to be preferred to the 
 mixture of strong ammonia and water [1 : 3] ordinarily 
 employed. As a preliminary step, therefore, in the work to be 
 
 * Fresenius, 6te Aufl., p. 805. t Zeitschr. anal. Chem., viii, 173. 
 
 J Zeitschr. anal. Chem., ix, 16. 
 
MAGNESIUM PHOSPHATE OF ANALYSIS. 193 
 
 described, experiments were made to find how small an amount 
 of magnesium could be detected in solution by precipitating 
 with microcosmic salt, either alone or in presence of ammonium 
 chloride in faintly ammoniacal solutions. The ammonium 
 chloride used for these tests (as well as in the similar quanti- 
 tative work following) was purified by boiling with a faint 
 excess of ammonia, filtering, digesting twelve hours with 
 microcosmic salt, and filtering again. The results are given 
 in Table II. 
 
 TABLE II. 
 
 Weight of 
 MgO taken as 
 the nitrate. 
 
 H(NH 4 )NaP0 4 . 
 4H,0 taken. 
 
 Volume. 
 
 NH 4 C1 
 taken. 
 
 Opalescent 
 precipitation. 
 
 grm. 
 
 grm. 
 
 cm 8 
 
 grm. 
 
 
 ( 0.0003 
 
 1.75 
 
 100 
 
 
 Marked. 
 
 1 0.0003 
 
 1.75 
 
 500 
 
 
 Marked. 
 
 (0.0003 
 
 1.75 
 
 100 
 
 10 
 
 Marked. 
 
 } 0.0003 
 
 1.75 
 
 500 
 
 10 
 
 Marked. 
 
 ( 0.0003 
 
 1.75 
 
 500 
 
 30 
 
 Faint. 
 
 0.0001 
 
 1.75 
 
 100 
 
 . 
 
 Marked. 
 
 ( 0.0001 
 
 1.75 
 
 100 
 
 10 
 
 Marked. 
 
 } 0.0001 
 
 1.75 
 
 500 
 
 10 
 
 Faint. 
 
 ( 0.0001 
 
 1.75 
 
 500 
 
 60 
 
 Faint. 
 
 The results of these tests show that even so little as 0.0001 
 grm. of magnesium oxide may be detected in five hundred 
 cubic centimeters of faintly ammoniacal water containing as 
 much as sixty grams of ammonium chloride, f It is plain that 
 strongly ammoniacal liquids are entirely unnecessary in the 
 precipitation of the ammonium magnesium phosphate under 
 the conditions. In nearly all the experiments to be detailed 
 use was made, therefore, of faintly ammoniacal solutions and 
 wash-water. 
 
 In Table III are given the results obtained in a study of the 
 effects of varying proportions of ammonium chloride and the 
 soluble phosphate upon the constitution of the precipitate. 
 All precipitates were gathered upon asbestos in the filtering 
 
 t It was found also, incidentally, that the presence of reasonable amounts of 
 ammonium oxalate (100 cm 8 of the saturated solution) does not interfere with 
 the precipitation of the ammonium magnesium phosphate by microcosmic 
 salt. 
 
 VOL. XI. 13 
 
194 
 
 CONSTITUTION OF THE AMMONIUM 
 
 crucible, washed in faintly ammoniacal water, and ignited as 
 usual. In every case the precipitation was practically com- 
 plete ; for, upon allowing the filtrates with the wash-water to 
 stand for several days after further addition of microcosmic 
 salt, nothing but insignificant traces of a precipitate not 
 exceeding 0.0001 grm. ever appeared. In the experiments 
 of section A precipitations were made in the cold by the action 
 of microcosmic salt in considerable excess upon the solutions 
 of magnesium nitrate containing varying amounts of ammonium 
 chloride. In experiments (1) to (5) the liquid was made 
 faintly ammoniacal after the addition of the precipitant and 
 the precipitate was filtered off immediately after complete 
 subsidence; in experiments (6) to (10) the precipitate first 
 
 TABLE in. 
 
 Exp. 
 
 M g2 P 2 7 
 corresponding 
 to Mg(N0 3 ) a 
 
 Mg 2 P 2 7 
 found. 
 
 Error in 
 terms of 
 Mg 2 P 2 7 . 
 
 Error in 
 terms of 
 MgO. 
 
 NH 4 C1 
 present. 
 
 HNaNH 4 P0 4 . 
 4H 2 O used. 
 
 Volume. 
 
 
 
 
 
 A. 
 
 
 
 
 
 grm. 
 
 grm. 
 
 grin. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 cm 3 . 
 
 (1) 
 
 0.5311 
 
 0.5418 
 
 0.0107+ 
 
 0.0038+ 
 
 
 
 2.5 
 
 150 
 
 2) 
 
 0.5311 
 
 0.5462 
 
 0.0151+ 
 
 0.0057+ 
 
 2 
 
 2.5 
 
 150 
 
 8 
 
 0.5311 
 
 0.5408 
 
 0.0097+ 
 
 0.0035+ 
 
 2 
 
 2.5 
 
 150 
 
 (4) 
 
 0.5311 
 
 0.5500 
 
 0.0189+ 
 
 0.0068+ 
 
 60 
 
 2.5 
 
 250 
 
 (5) 
 
 0.5311 
 
 0.5520 
 
 0.0209+ 
 
 0.0075+ 
 
 60 
 
 2.5 
 
 250 
 
 (6) 
 
 0.5311 
 
 0.5345 
 
 0.0034+ 
 
 0.0012+ 
 
 
 2.5 
 
 150 
 
 (7 
 
 0.5311 
 
 0.5371 
 
 0.0060 f 
 
 0.0022+ 
 
 
 2.5 
 
 150 
 
 (8) 
 
 0.5311 
 
 0.5384 
 
 0.0073+ 
 
 0.0026+ 
 
 
 2.5 
 
 150 
 
 (9) 
 
 0.5311 
 
 0.5386 
 
 0.0075+ 
 
 0.0027+ 
 
 
 2.5 
 
 150 
 
 (10) 
 
 0.5311 
 
 0.5415 
 
 0.0104+ 
 
 0.0037+ 
 
 
 2.5 
 
 150 
 
 
 
 
 
 B. 
 
 
 
 
 (11) 
 
 0.5311 
 
 0.5312 
 
 0.0001+ 
 
 0.0000 
 
 # 
 
 2.5 
 
 150,100 
 
 (12) 
 
 0.5311 
 
 0.5311 
 
 0.0000 
 
 0.0000 
 
 # 
 
 2.5 
 
 150,100 
 
 (13) 
 
 0.5311 
 
 0.5346 
 
 0.0035+ 
 
 0.0013+ 
 
 2 + 2 
 
 2.5 
 
 150,100 
 
 (14) 
 
 0.5311 
 
 0.5348 
 
 0.0037+ 
 
 0.0014+ 
 
 2+ 2 
 
 2.5 
 
 150,100 
 
 (15) 
 
 0.5311 
 
 0.5383 
 
 0.0072+ 
 
 0.0026+ 
 
 5+ 6 
 
 2.5 
 
 150,100 
 
 (16) 
 
 0.5311 
 
 0.5368 
 
 0.0057+ 
 
 0.0021+ 
 
 5+ 5 
 
 2.5 
 
 150,100 
 
 (17) 
 
 0.5311 
 
 0.5376 
 
 0.0065+ 
 
 0.0023+ 
 
 10 + 10 
 
 2.5 
 
 200,100 
 
 (18) 
 
 0.5311 
 
 0.5395 
 
 0.0084+ 
 
 0.0030+ 
 
 10 + 10 
 
 2.5 
 
 200,100 
 
 (19) 
 
 0.5311 
 
 0.5396 
 
 0.0085+ 
 
 0.0031+ 
 
 60+ 5 
 
 2.5 
 
 250,100 
 
 (20) 
 
 0.5311 
 
 0.5389 
 
 0.0078+ 
 
 0.0028+ 
 
 60+ 6 
 
 2.6 
 
 250,100 
 
 * Probably less than 1 grm. 
 
MAGNESIUM PHOSPHATE OF ANALYSIS. 195 
 
 thrown down was redissolved in a very little hydrochloric 
 acid and reprecipitated by dilute ammonia (the operation 
 being repeated several times) with a view to improving the 
 crystalline condition of the precipitate, and this treatment 
 introduced, of course, a small amount of ammonium chloride, 
 probably less than a gram. It will be observed that errors of 
 excess appear in all of these determinations, those being the 
 greatest in the experiments in which the largest amounts of 
 the ammonium salt were present. 
 
 In the experiments of section B the manipulation was so 
 changed that the supernatant liquid was poured off (through 
 the filtering crucible which was to be used subsequently to 
 collect the phosphate) after the precipitate had subsided and 
 the insoluble phosphate was dissolved in hydrochloric acid and 
 brought down again, after dilution, by the addition of a faint 
 excess of dilute ammonia. By thus removing the supernatant 
 liquid after the first precipitation, the excess of the precipitant 
 and the amounts of ammonium chloride orginally present were 
 reduced to relatively low limits, so that their effects in the 
 reprecipitation were at a minimum, and by adding varying 
 amounts of ammonium chloride, or none at all, before the 
 reprecipitation, it became possible to demonstrate the individual 
 effect of that reagent apart from that of an excess of the 
 microcosmic salt. It will be noted that in experiments (11) 
 and (12), in which no ammonium salt was added after the 
 decantation from the first precipitate, the results are ideal, and 
 that the errors of excess advance as the amounts of ammonium 
 salt present in the final precipitation increase. The quantity 
 of the ammonium salt present during the first precipitation 
 does not influence the error in the final precipitation unless it 
 is so large that a simple decantation of the supernatant liquid 
 would naturally leave an appreciable amount of it to act when 
 the second precipitation takes place. 
 
 It is plain that the errors of excess which appear when 
 either the ammonium chloride or the soluble phosphate is 
 present in considerable amount, must be due either to mechan- 
 ical inclusion on the part of the highly crystalline precipitate, 
 
196 CONSTITUTION OF THE AMMONIUM 
 
 or to variation in the ammonium magnesium phosphate from 
 the ideal constitution toward a condition represented by a 
 phosphate richer in ammonia and correspondingly deficient in 
 magnesium. If any appreciable amount of the ammonium 
 chloride present were held by the precipitate, it would natu- 
 rally be represented by magnesium chloride after ignition, but, 
 in no one of these experiments, even in those dealing with 
 sixty grams of ammonium chloride, did the residue, after dis- 
 solving in nitric acid, give with silver nitrate evidence of the 
 presence of more than a mere unweighable trace of chloride. 
 A special experiment, moreover, in which an attempt was 
 made to determine the silver chloride precipitated from the 
 solution hi nitric acid of an unignited precipitate thrown down 
 by microcosmic salt in presence of sixty grams of ammonium 
 chloride, confirms this conclusion : 'the precipitate was unweigh- 
 able. If ammonium chloride present in the solution to so 
 great an amount is not included in the precipitate in signifi- 
 cant quantity, it would seem to be unnatural that the micro- 
 cosmic salt should be included mechanically in any very great 
 amount. But unless the microcosmic salt was mechanically 
 included, the increase in weight must be due to the chemical 
 influence of the reagents that is, to the production of a 
 phosphate rich in ammonium and deficient in magnesium. 
 Berzelius* recognized the existence of such a phosphate of 
 magnesium ; but Wachf in following the work of Berzelius, 
 failed to find it. It would be natural to expect its formation, 
 if ever, when the precipitating phosphate is in excess and 
 ammonium salts are present in abundance, with free ammonia. 
 Obviously the natural effects of all these reagents would be 
 toward the production of a salt holding more ammonia and 
 more phosphoric pentoxide for a given amount of magnesium. 
 The results of the table seem to point strongly to such ten- 
 dencies, and, by inference, toward the existence of such a 
 compound. Thus in experiments (11) and (12), in which the 
 
 * Berzelius, Jahresbericht, 3. Jahrgang (1824), iibersetzt von C. G. Gmelin, 
 8.92. 
 
 t Schweigger, 1830, Band 29, s. 265. 
 
MAGNESIUM PHOSPHATE OF ANALYSIS. 197 
 
 greater part of this excess of microcosmic salt was removed 
 by decantation before the second precipitation, while no am- 
 monium chloride was present excepting the small amount 
 made by the solution and reprecipitation of the first precipi- 
 tate, the error is practically nothing. In experiments (13) and 
 (14), (15) and (16), (17) and (18), all similar to (11) and 
 (12) excepting that ammonium chloride was present, the aver- 
 age errors (+0.0036 grm. in terms of magnesium phosphate, 
 +0.0064 grm., +0.0074, respectively) increase as the ammonium 
 chloride is increased in the final precipitation. In experiments 
 (19) and (20), in which the ammonium chloride amounted to 
 sixty grams in the first precipitation and to five grams in the 
 second in addition to the amount that would naturally remain 
 after decanting the strong solution of the former precipita- 
 tion, the similarity of this error (+0.0082 in the mean) to 
 that of the experiments in which smaller amounts of the 
 ammonium chloride were used throughout goes to show that 
 only the amount of ammonium salt present in the final pre- 
 cipitation counts. Further, a comparison of corresponding 
 experiments of A and B shows very plainly that the treatment 
 which involves the removal of the large part of the micro- 
 cosmic salt, the solution of the precipitate, and reprecipitation 
 tends to reduce the higher indications. Thus, for example the 
 error in (2) and (3) is +0.0124 gram in terms of magnesium 
 pyrophosphate, while in (13) and (14), similarly carried out 
 except the decantation of the excess of the precipitant, solu- 
 tion and reprecipitation, the error is +0.0036 grm. 
 
 The special influence of free ammonia during precipitation, 
 was investigated in the following experiments. Definite 
 volumes of the magnesium nitrate solution were drawn from a 
 burette into a platinum dish, ammonium chloride 10 grm. 
 was added, the magnesium was brought down by dilute 
 ammonia in presence of microcosmic salt, and strong ammonia 
 equal to one-third the volume of the solution was added. The 
 solutions, after standing, were filtered off on asbestos under 
 pressure in a perforated crucible, and the precipitates were 
 washed with ammonia diluted to the proportion of three parts 
 
198 
 
 CONSTITUTION OF THE AMMONIUM 
 
 of water to one of ammonia, dried after moistening with a 
 drop of saturated solution of ammonium nitrate, ignited and 
 weighed. The results are given in experiments (1) and (2) of 
 Table IV. In these determinations the mean error reaches 
 +0.0193 gnn. in terms of magnesium pyrophosphate ; while 
 in experiments (3) and (4), made similarly excepting that the 
 supernatant liquid was decanted from the precipitate first 
 thrown down, the precipitate dissolved in hydrochloric acid, 
 and after dilution reprecipitated by dilute ammonia imme- 
 diately supplemented by enough strong ammonia to make one- 
 fourth the volume of the entire solution, the error amounts in 
 the mean to +0.0061 in terms of the pyrophosphate. 
 
 TABLE IV. 
 
 Exp. 
 
 Mg 2 P 2 7 
 corresponding 
 to Mg(NO 3 ) a 
 
 Mg 2 P 2 7 
 found. 
 
 Error in 
 terms of 
 M g2 P 2 7 . 
 
 Error in 
 terms 
 of MgO. 
 
 NH 4 C1 
 present. 
 
 HNaNH 4 P0 4 . 
 4H 2 used. 
 
 Volume. 
 
 (1) 
 
 gnn. 
 
 0.5311 
 
 grm. 
 0.5503 
 
 grm. 
 0.0192+ 
 
 gnn. 
 0.0069+ 
 
 grm. 
 
 10 
 
 grm. 
 2.5 
 
 cms 
 200 
 
 (2) 
 
 0.5311 
 
 0.5505 
 
 0.0194+ 
 
 0.0070+ 
 
 10 
 
 2.5 
 
 200 
 
 (3) 
 
 0.6311 
 
 0.5393 
 
 0.0082+ 
 
 0.0029+ 
 
 10, 
 
 2.5 
 
 200,100 
 
 (4) 
 
 0.5311 
 
 0.5351 
 
 0.0040+ 
 
 0.0017+ 
 
 10,- 
 
 2.5 
 
 200,100 
 
 In experiments (1) and (2) the precipitate was influenced by 
 an excess of microcosmic salt, ammonium chloride, and free 
 ammonia in large amount; in experiments (3) and (4), by 
 decanting in the manner previously described, by dissolving 
 the precipitate, and reprecipitating, the effects of an excess of 
 microcosmic salt and ammonium chloride are reduced to a mini- 
 mum, and, in a comparison of the results with those of experi- 
 ments (11) and (12) of Table III the tendency of the free 
 ammonia comes to view. The results discussed seem certainly 
 to point to a general tendency on the part of free ammonia, 
 ammonium choride and excess of the phosphate to produce a 
 salt rich in ammonia and deficient in magnesium, which for a 
 definite amount of magnesia precipitated must leave upon igni- 
 tion a residue weighing more than the normal phosphate. 
 If it be assumed that a salt of the symbol (NH 4 ) 4 Mg(PO 4 ) 2 
 
MAGNESIUM PHOSPHATE OF ANALYSIS. 199 
 
 the next natural step to the normal salt, NH 4 MgPO 4 
 is present in the precipitate, the residue which such a salt 
 would leave upon ignition would be the metaphosphate 
 Mg(PO 3 ) 2 . From the relations of the symbols for magnesium 
 pyrophosphate and magnesium metaphosphate the weight of 
 the residue obtained, and the weight of the pyrophosphate 
 theoretically derivable from the weight of magnesium salt 
 used, it is possible, of course, to calculate the proportionate 
 amounts of pyrophosphate and metaphosphate present in any 
 ignited residue. Proceeding in this manner, it appears that, 
 in order to account for the variations noted, it is necessary 
 to assume the presence in many cases of very considerable 
 amounts of the metaphosphate. Thus, in the case of those 
 results obtained according to the usually accepted method of 
 precipitating and washing with strongly ammoniacal liquids, 
 viz., in experiments (1) and (2) of Table IV, the proportion 
 of metaphosphate needed to account for the observed error 
 reaches ten per cent. 
 
 Precipitation ~by Excess of the Magnesium Salt. 
 
 The relations which obtain in the reverse process of pre- 
 cipitation the action of an excess of the magnesium salt 
 upon a soluble phosphate were studied in experiments to 
 be described. A solution of pure hydrogen disodium phos- 
 phate was prepared by carefully recrystallizing the pure salt 
 of commerce five tunes from distilled water in a platinum 
 dish, dissolving the crystals, and diluting to definite volume. 
 The standard of the solution was established by evaporating 
 to dryness in a weighed platinum crucible known volumes 
 of the solution, igniting the residue and weighing the sodium 
 pyrophosphate. Magnesia mixture, the precipitant, was pre- 
 pared by dissolving fifty-five grams of magnesium chloride 
 in as little water as possible and filtering, mixing with this 
 solution twenty-eight grams of ammonium chloride purified 
 by treating it in strong solution with bromine water and a 
 slight excess of ammonia, filtering, diluting to one liter, and, 
 after standing for some hours, filtering again. 
 
200 
 
 CONSTITUTION OF THE AMMONIUM 
 
 The tests of the following table show that the precipitation 
 of a soluble phosphate by the magnesia mixture is practically 
 complete in faintly ammoniacal solutions even when very 
 dilute and charged with large amounts of ammonium chloride, 
 provided the magnesia mixture is present in sufficiently large 
 excess. 
 
 TABLE V. 
 
 P 2 O, in HNa,PO 4 
 taken. 
 
 Magnesia 
 mixture. 
 
 Volume. 
 
 NHtCl. 
 
 Precipitation visible. 
 
 grm. 
 
 cms 
 
 cm 8 
 
 grm. 
 
 
 0.0005 
 
 10 
 
 100 
 
 t 
 
 At once throughout 
 
 0.0005 
 
 50 
 
 100 
 
 . 
 
 the liquid. 
 
 0.0005 
 
 10 
 
 100 
 
 10 
 
 it 
 
 0.0005 
 
 10 
 
 200 
 
 60 
 
 tt 
 
 0.0001 
 
 50 
 
 250 
 
 60 
 
 u 
 
 0.0001 
 
 10 
 
 100 
 
 
 
 
 0.0001 
 
 10 
 
 100 
 
 10 
 
 tl 
 
 0.0001 
 
 50 
 
 200 
 
 10 
 
 
 
 0.0001 
 
 10 
 
 250 
 
 60 
 
 
 
 0.0001 
 0.0001 
 
 50 
 50 
 
 300 
 500 
 
 60 
 60 
 
 After settling out. 
 
 This conclusion was further substantiated by an actual test 
 (by the molybdate method) of the ignited residue, obtained 
 by evaporating a filtrate from ammonium magnesium phos- 
 phate (equivalent to 0.8614 grm. of the pyrophosphate) 
 precipitated by a faintly ammoniacal solution of magnesia 
 mixture in presence of 60 grm. of ammonium chloride, 
 which gave a precipitate of ammonium phosphomolybdate 
 yielding 0.0002 grm. of magnesium pyrophosphate. It is 
 evident, therefore, that any considerable deficiencies of weight 
 of the magnesium phosphate obtained by precipitating equal 
 amounts of a soluble phosphate by magnesia mixture in 
 presence of varying amounts of ammonium chloride, cannot 
 be attributed to varying solubility of the magnesium phosphate 
 under changing proportions of the ammonium chloride. 
 
 The results recorded in section A of Table VI were obtained 
 by treating definite volumes of the pure solution of hydrogen 
 disodium phosphate with magnesia mixture, in slight excess 
 above the amount required to bring down the phosphate, and 
 
MAGNESIUM PHOSPHATE OF ANALYSIS. 
 
 201 
 
 making the solution distinctly ammoniacal. After thorough 
 subsidence, the precipitate was filtered off on asbestos under 
 pressure in a perforated platinum crucible, washed hi water 
 faintly ammoniacal, dried, ignited and weighed. In experi- 
 ments (1), (5) and (6), only the ammonium chloride present in 
 the magnesia mixture was used ; in the other cases weighed 
 portions were added. In the experiments of section B, the 
 precipitate was dissolved in hydrochloric acid after filtering off 
 the supernatant liquid, brought down again in dilute solution 
 by ammonia in distinct excess, and thereafter treated as in the 
 experiments of section A. The experiments of section C 
 were conducted similarly to (1), (5) and (6) of A excepting 
 that the magnesium mixture was introduced into the am- 
 
 TABLE VL 
 
 Exp. 
 
 Mg 2 P 2 7 
 corre- 
 sponding 
 to 
 
 taken. 
 
 found. 
 
 Error in 
 terms of 
 M gl P,0 7 . 
 
 Error in 
 terms of P. 
 
 Volume. 
 
 NH 4 C1 
 
 in mag- 
 nesia 
 mix- 
 ture. 
 
 NH^l 
 added. 
 
 MgCl, 
 6H 2 O 
 in mag- 
 nesia 
 mix- 
 ture. 
 
 A 
 
 (1) 
 (2) 
 (3) 
 (4) 
 (5) 
 (6) 
 7) 
 8) 
 9) 
 
 grm. 
 0.8615 
 0.8615 
 0.8615 
 0.8615 
 
 0.0852 
 0.0852 
 0.0852 
 0.0852 
 0.0852 
 
 gnu. 
 0.8613 
 0.8615 
 0.8602 
 0.8561 
 
 0.0862 
 0.0866 
 0.0847 
 0.0830 
 0.0811 
 
 grm. 
 
 0.0002- 
 0.0000 
 0.0013- 
 0.0054- 
 
 0.0010+ 
 0.0014+ 
 0.0005- 
 0.0022- 
 0.0041- 
 
 grm. 
 
 0.00005- 
 0.00000 
 0.00036- 
 0.00151- 
 
 0.00028+ 
 0.00039+ 
 0.00014- 
 0.00062- 
 0.00115- 
 
 cm 3 
 150 
 200 
 200 
 300 
 
 100 
 100 
 200 
 200 
 300 
 
 griii. 
 
 1.68 
 1.68 
 1.68 
 1.68 
 
 0.28 
 0.28 
 0.28 
 0.28 
 0.28 
 
 griii. 
 
 *20' 
 20 
 60 
 
 *20' 
 20 
 60 
 
 grm. 
 3.3 
 3.3 
 3.3 
 3.3 
 
 0.55 
 0.55 
 0.55 
 0.55 
 0.55 
 
 B 
 
 (10) 
 
 (11) 
 (12) 
 (13) 
 
 (14) 
 (15) 
 (16) 
 (17) 
 
 0.8111 
 0.8615 
 0.8615 
 0.8615 
 
 0.0852 
 0.0852 
 0.0852 
 0.0852 
 
 0.8114 
 0.8613 
 0.8578 
 
 0.8487 
 
 0.0855 
 0.0656 
 0.0853 
 0.0819 
 
 0.0003+ 
 0.0002- 
 0.0037- 
 0.0128- 
 
 0.0003+ 
 0.0004+ 
 0.0001+ 
 0.0033- 
 
 0.00008+ 
 0.00006- 
 0.00103- 
 0.00358- 
 
 0.00008+ 
 0.00011+ 
 0.00003+ 
 0.00092- 
 
 150,100 
 150,000 
 200,100 
 200,100 
 
 100,100 
 100,100 
 150,100 
 200,100 
 
 1.68 
 1.68 
 1.68 
 1.68 
 
 0.28 
 0.28 
 0.28 
 0.28 
 
 ! ,20 
 .,60 
 
 io] ! 
 
 10,10 
 20,20 
 
 3.3 
 3.3 
 3.3 
 3.3 
 
 0.55 
 0.55 
 0.55 
 0.55 
 
 C 
 
 (18) 
 (19) 
 
 0.8111 
 0.8111 
 
 0.8071 
 0.8052 
 
 0.0040- 
 0.0059- 
 
 0.00112- 
 0.00165- 
 
 120 
 120 
 
 1.4 
 
 1.4 
 
 
 
 2.75 
 2.75 
 
202 CONSTITUTION OF THE AMMONIUM 
 
 moniacal solution of the phosphate drop by drop from a 
 burette. The precipitations in A, B, and C were proved to 
 be practically complete ; for by treatment of the filtrates with 
 more magnesia mixture and standing, no more than a trace 
 0.0001 grm. at the most of the phosphate was found. 
 The ignited residues never contained more than a mere trace 
 of chlorine. 
 
 While the results are not entirely regular, the tendency of 
 the ammonium salt to produce errors of deficiency in propor- 
 tion to its amount is plain if we compare among themselves 
 the experiments of A upon similar amounts of phosphate, and 
 then those of B upon similar amounts of phosphate among 
 themselves ; and by a comparison of corresponding results in 
 A and B it is clearly shown that the presence of an excess of 
 magnesia mixture tends to counteract more or less completely 
 errors of deficiency due to the action of the ammonium chloride. 
 These facts are quite in harmony with the hypothesis that the 
 ammonium salt tends to produce an ammonium magnesium 
 phosphate richer in ammonia and phosphoric acid and poorer 
 in magnesia than the normal salt NH 4 MgPO 4 ; for, though 
 the production of such a salt in presence of an excess of the 
 soluble phosphate compels the combination of a definite 
 amount of magnesium with more than the normal amounts of 
 phosphoric acid and ammonia (as was the case in the former 
 series of experiments), when the supply of the soluble 
 phosphate is limited the amount of magnesium associated with 
 it must fall below the normal (as is the case in the present 
 series of experiments). Moreover, the behavior of the 
 precipitant is quite in accord with the hypothesis ; for, though 
 the influence of an excess of the soluble phosphate would 
 naturally tend (as was observed) in the same direction as that 
 of the ammonium salt and free ammonia, viz., to the production 
 of the phosphate deficient in magnesium, the tendency of an 
 excess of the magnesium salt must obviously be to increase 
 the amount of magnesium in the phosphate, as was observed 
 in the experiments of Table VI. The hypothesis fits the facts, 
 therefore, on both sides ; and, if precipitation is practically 
 
MAGNESIUM PHOSPHATE OF ANALYSIS. 203 
 
 complete (as was shown to be the case throughout) the 
 argument for the existence of an ammonium magnesium 
 phosphate poorer than the normal salt in magnesium 
 possibly the salt (NH^Mg^O^a seems to be strong. 
 
 The Practical Determination of Magnesium and Phosphoric 
 
 Acid. 
 
 In determining magnesium by the procedure in ordinary use, 
 the tendency isstrong as is shown in experiments (1) and(2) 
 of Table IV toward high plus errors, and the error is due 
 to the combined effects of excesses of the precipitant, the 
 ammonium salt, and free ammonia. The experiments (11) 
 and (12) of B, Table III, show conclusively that such 
 tendencies to error may be counteracted effectively by pouring 
 off the supernatant liquid (through the filter to be used 
 subsequently to collect the precipitate) as soon as the precipi- 
 tate subsides, dissolving the phosphate in the least amount 
 of hydrochloric acid, bringing it down again, after dilution, 
 by a faint excess of ammonia, filtering (best, we think, on 
 asbestos, under pressure), washing with faintly ammoniacal 
 water, and igniting as usual. 
 
 Many years ago* a method of precipitating the ammonium 
 magnesium phosphate was advocated by Professor Wolcott 
 Gibbs, which consists, essentially, in boiling the solution of the 
 magnesium salt with microcosmic salt and adding ammonia 
 after cooling, and by which most exact analytical results 
 were obtained. Our experience confirms completely that of 
 Gibbs, and we desire to direct attention again to a procedure 
 the advantage of which has, unfortunately, not been broadly 
 known and accepted. Even in the presence of considerable 
 amounts of ammonium chloride this process yields a phosphate 
 of nearly ideal constitution if only the boiling be prolonged 
 from three to five minutes. The greater part of the ammonium 
 magnesium phosphate about 90 per cent forms in this 
 process before free ammonia is added, and the ammonium 
 which enters the phosphate thus formed is derived from the 
 * Am. Jour. Sci. [3], v, 114. 
 
204 CONSTITUTION OF THE AMMONIUM 
 
 microcosmic salt, which must become correspondingly acidic. 
 Under these conditions, the tendency to form an insoluble 
 ammonium magnesium phosphate richer in ammonia and poorer 
 in magnesia than the normal salt, does not develop. In the 
 process of Gibbs, as well as in the modified precipitation 
 process in the cold, the use of the faintly ammoniacal solution 
 and wash-water is sufficient and advantageous. 
 
 In the precipitation of a soluble phosphate by magnesia 
 mixture the tendency of the precipitant and that of the 
 ammonium salt are antagonistic, so that the effect of the latter 
 salt is somewhat masked, though manifest. This opposition 
 of effects has been noted by Mahon,* who, though regarding 
 the actual attainment of an exact balance as uncertain, ventures 
 the opinion that accurate results should be attainable by the 
 careful relative adjustment of the proportions of the precipitant 
 and ammonium salt. Mahon claims to get the best results by 
 a very gradual addition of magnesia mixture to the ammoniacal 
 solution of the phosphate containing about sixteen per cent of 
 ammonium chloride, strong ammonia being added subsequently. 
 From our observations, however, recorded in section C of Table 
 VI, it appears that the method of introducing the magnesia 
 mixture gradually into the ammoniacal phosphate (taken in 
 quantity sufficiently large to give unmistakable indications) 
 produces a precipitate deficient in magnesium and so leads to 
 errors of deficiency in the phosphorus indicated. The use of 
 strong ammonia, moreover, we have shown to be both unnec- 
 essary and disadvantageous. Our experiments go to show that 
 good results may be expected when the solution of the phosphate 
 containing a moderate excess of the magnesium salt and not 
 more than five to ten per cent of ammonium chloride is 
 precipitated by making it slightly ammoniacal, the precipitate 
 being washed in slightly ammoniacal wash-water. In general, 
 however, and especially when more ammonium chloride than 
 this proportion, or more magnesium salt than twice the amount 
 theoretically necessary, is present, it is safer to decant the 
 supernatant liquid from the precipitate (through the filter to 
 
 * Jour. Am. Chem. Soc., xx, 445. 
 
MAGNESIUM PHOSPHATE OF ANALYSIS. 205 
 
 be used subsequently to hold the phosphate), to dissolve the 
 precipitate in a little hydrochloric acid, and reprecipitate by 
 dilute ammonia, washing with faintly ammoniacal wash-water. 
 Since our first publication of the work described above, 
 Neubauer* has called attention to the fact that the influence 
 of ammonia and ammonium salts upon the constitution of the 
 ammonium magnesium phosphate obtained in determining 
 phosphoric acid had been previously pointed out by him in a 
 paper | discussing methods for the estimation of that acid. 
 We take pleasure, therefore, in conceding to Neubauer full 
 priority in the observation of the effect which we have 
 endeavored to overcome in the determination of magnesium 
 and of phosphoric acid. 
 
 * Zeitschr. anorg. Chem., xxii, 162. 
 t Zeitschr. anorg. Chem., ii, 45. 
 
XXV 
 
 THE INFLUENCE OF HYDEOCHLOEIC ACID 
 IN TITEATIONS BY SODIUM THIOSULPHATE WITH 
 SPECIAL EEFEEENCE TO THE ESTIMA- 
 TION OF SELENIOUS ACIDS. 
 
 BY JOHN T. NORTON JR.* 
 
 IN the method of Norris and Fay f for the iodometric deter- 
 mination of selenious acid, advantage is taken of a direct 
 and unique action of sodium thiosulphate upon selenium 
 dioxide in the presence of hydrochloric acid. Most excellent 
 results are claimed for this method; but the explicit state- 
 ment of the originators of the method, that the amount of 
 hydrochloric acid present does not influence the result, pro- 
 vided the titration is made at the temperature of melting ice, 
 is so extraordinary in view of generally accepted ideas in 
 regard to the interaction of hydrochloric acid and sodium 
 thiosulphate, as to, suggest the necessity of careful investiga- 
 tion of this point. 
 
 Pickering, J in his investigation of the reaction between 
 iodine and sodium thiosulphate, has shown that more iodine 
 is required to oxidize the thiosulphate as the proportion of 
 hydrochloric acid increases. He ascribed this effect to the 
 formation of a sulphate, apparently, by the increased activity 
 of the iodine, but the more rational explanation is that, 
 although some sulphate is ultimately formed, the thiosul- 
 phate is first partially decomposed into free sulphur and 
 
 * From Am. Jour., Sci. vii, 287. 
 
 t Am. Chem. Jour., vol. xviii, p. 703. 
 
 t Jour. Chem. Soc., vol. xxxvii, p. 135. 
 
HYDROCHLORIC ACID IN TITRATIONS. 
 
 207 
 
 sulphur dioxide. Finkener* and Mohrf also mention the 
 decomposing effect of free acid upon sodium thiosulphate. 
 
 The sodium thiosulphate used in the following experiments 
 was taken in nearly decinormal solution and was standardized 
 by running it into an approximately decinormal solution of 
 iodine, the value of which had been determined by comparison 
 with decinormal arsenious acid made from carefully resub- 
 limed arsenious oxide. In the experiments of Table I the 
 solutions were stirred continuously and kept at a temperature 
 of from to 5 C., while the thiosulphate ran into the 
 acidified liquid. The volume of the solution, though fixed at 
 the beginning as given in the table was considerably increased 
 during the operation by the melting of the ice. Titrations 
 
 TABLE I. 
 
 Volume of 
 liquid at 
 beginning of 
 titration. 
 
 Na^Og 
 
 approx- 
 imately 
 ^ taken. 
 
 Volume of ^ iodine used in titration. 
 
 HC1 = none. 
 
 =: 1 cms. 
 
 = 5 cm 3 . 
 
 = 10 cm 3 . 
 
 cm 3 
 
 cm 3 
 
 cm 
 
 cm 3 
 
 cm 3 
 
 cm 8 
 
 100 
 
 30 
 
 30.251 
 
 
 30.75 
 
 30.76 
 
 31.20 
 
 200 
 
 30 
 
 30.22 
 
 
 30.21 
 
 30.56 
 
 31.40 
 
 300 
 
 30 
 
 30.20 
 
 mean = 
 
 on 99 
 
 30.22 
 
 31.03 
 
 30.90 
 
 400 
 
 30 
 
 30.21 
 
 OvwDS 
 
 30.20 
 
 30.20 
 
 30.55 
 
 500 
 
 30 
 
 30.20 
 
 
 30.20 
 
 30.21 
 
 30.55 
 
 100 
 
 25 
 
 25.29' 
 
 
 25.32 
 
 25.98 
 
 25.70 
 
 200 
 
 25 
 
 25.28 
 
 
 25.34 
 
 25.40 
 
 25.45 
 
 300 
 
 25 
 
 25.29 
 
 mean = 
 
 fc"*C OT 
 
 25.41 
 
 25.38 
 
 25.83 
 
 400 
 
 25 
 
 25.27 
 
 25.27 
 
 25.24 
 
 25.30 
 
 25.63 
 
 500 
 
 25 
 
 25.22 . 
 
 
 25.23 
 
 25.40 
 
 25.30 
 
 100 
 
 20 
 
 20.15 
 
 
 20.17 
 
 20.33 
 
 20.23 
 
 200 
 
 20 
 
 20.20 
 
 
 20.13 
 
 20.27 
 
 20.23 
 
 300 
 
 20 
 
 20.21 
 
 mean =: 
 
 9O 1 1\ 
 
 20.15 
 
 20.20 
 
 20.17 
 
 400 
 
 20 
 
 20.20 
 
 Zv.lO 
 
 20.10 
 
 20.27 
 
 20.07 
 
 500 
 
 20 
 
 20.10 
 
 
 20.10 
 
 20.17 
 
 20.13 
 
 were conducted as rapidly as possible to avoid the separation 
 of sulphur, which is likely to occur, especially when the acid 
 and thiosulphate are present in large quantities. A perusal 
 of the table shows that the influence of the hydrochloric acid 
 upon the thiosulphate depends chiefly upon the amount of 
 
 * Anal. Chem., 6. Aufl., p. 620. t Titrirmethode, 6. Aufl., p. 279. 
 
208 
 
 INFLUENCE OF HYDROCHLORIC ACID IN 
 
 the thiosulphate present and afterwards upon the degree of 
 dilution and its own absolute quantity. Thus when 30 cm 3 
 of sodium thiosulphate were employed the effect of 10 cm 3 
 of acid is marked at all dilutions within the range of the 
 experiments; the effect of 5 cm 3 of acid is inappreciable only 
 at a dilution of from 400 to 500 cm 3 , and when 1 cm 3 of acid is 
 employed the effect is only perceptible at a volume of 100 cm 3 . 
 When 25 cm 3 of the thiosulphate were used the influence of 
 the acid is less marked; for at a dilution of 500 cm 3 the 
 effect of 10 cm 8 of acid is not seen, and 20 cm 3 of the thio- 
 sulphate may be present at any dilution down to 100 cm 3 in 
 the presence of as much as 10 cm 3 of the acid, and even 
 considerably more, as experiments not included in the table 
 indicated. 
 
 The slight discrepancies which appear occasionally in the 
 table were due, no doubt, to unavoidable differences in the 
 tune of action. 
 
 This influence of time upon the reaction between sodium 
 thiosulphate, iodine, and hydrochloric acid comes out clearly in 
 the following series of experiments, in which the thiosulphate 
 was run into the acidified water, cooled to a temperature of 
 from to 5 C. by means of ice, the solution being allowed 
 to stand 5, 10, and 15 minutes. Sulphur was thrown down 
 in nearly every case. 
 
 TABLE IL 
 
 Volume of 
 the liquid at 
 beginning of 
 titration. 
 
 HC1 
 (sp. gr. 1.12) 
 present. 
 
 Na 2 S,0 8 
 approxi- 
 mately. 
 taken. 
 
 Volume of ^ iodine used in titration after standing. 
 
 5 minutes. 
 
 10 minutes. 
 
 15 minutes. 
 
 cm 8 
 
 cm 8 
 
 cm 8 
 
 cm 8 
 
 cm 8 
 
 cm 8 
 
 200 
 
 10 
 
 30 
 
 30.80 
 
 31.30 
 
 32.32 
 
 200 
 
 10 
 
 25 
 
 25.50 
 
 26.00 
 
 26.30 
 
 200 
 
 10 
 
 20 
 
 20.30 
 
 20.70 
 
 20.68 
 
 The results of the table emphasize sufficiently the necessity 
 of proceeding as rapidly as possible with the titration of 
 sodium thiosulphate by iodine in presence of hydrochloric 
 
TITRATIONS BY SODIUM THIOSULPHATE. 
 
 209 
 
 acid, when the thiosulphate is present in considerable amount. 
 As would be expected, the effect of temperature upon the 
 reaction is also marked. In the following experiments the 
 sodium thiosulphate was run into the acidified water, pre- 
 viously heated to the temperature indicated, and then titrated 
 
 with iodine. 
 
 TABLE m. 
 
 Volume of 
 
 
 
 
 Volume of ^ 
 
 liquid at 
 beginning 
 of 
 
 HCl 
 (sp. gr. 1.12) 
 
 Temperature 
 Centigrade. 
 
 approxi- 
 mately. 
 
 iodine used 
 in titratious at 
 different 
 
 titration. 
 
 
 
 is 
 
 temperatures. 
 
 cm 3 
 
 cm 
 
 C. 
 
 cm 8 
 
 cm 3 
 
 400 
 
 10 
 
 6 
 
 25 
 
 23.52 
 
 400 
 
 10 
 
 22 
 
 25 
 
 23.73 
 
 400 
 
 10 
 
 34 
 
 25 
 
 24.35 
 
 400 
 
 10 
 
 42 
 
 25 
 
 24.5 
 
 400 
 
 10 
 
 54 
 
 25 
 
 25 
 
 400 
 
 10 
 
 64 
 
 25 
 
 26.1 
 
 From these results it is plain that the conditions under which 
 considerable amounts of sodium thiosulphate are titrated in 
 presence of hydrochloric acid must be carefully guarded 
 when accuracy is a consideration. It is also apparent that 
 in all cases the temperature should be reduced as nearly to 
 C. as possible and rapidity of titration by the iodine is 
 an essential. So long as the thiosulphate present does not 
 exceed 20 cm 3 of the T ^ solution, rapid titration in cold 
 solution proceeds with fair regularity in presence of hydro- 
 chloric acid up to 10 cm 3 of the acid of sp. gr. 1.12. When, 
 however, the amount of thiosulphate is greater than 20cm 3 
 of the -fy solution, care as to the restriction of the acid and 
 dilution of the solution becomes a necessity. Fortunately, 
 in most analytical processes involving the use of the thio- 
 sulphate it is possible to add that reagent from the burette 
 to the solution to be acted upon, so that it is destroyed 
 normally as fast as it is introduced and the danger of inter- 
 action with the acid does not occur. In the process of Norris 
 and Fay, however, the method involves the addition of an 
 excess of the thiosulphate to the solution of selenious and 
 
 VOL. II. 14 
 
210 INFLUENCE OF HYDROCHLORIC ACID IN 
 
 hydrochloric acids, and thus the conditions prevail which 
 demand care as to the relation of the acid, the thiosulphate 
 and the degree of dilution. I have experimented, therefore, 
 with this process under varying conditions. 
 
 The process of Norris and Fay for the iodometric determi- 
 nation of selenious acid consists briefly in treating the solution 
 of that acid in ice water, in the presence of hydrochloric acid, 
 with an excess of a T * ff solution of sodium thiosulphate and 
 titrating back the excess of the thiosulphate with iodine. 
 Four molecules of sodium thiosulphate act, apparently, upon 
 one molecule of selenious acid according to a reaction which 
 the authors propose to study. 
 
 The selenium dioxide used was made by dissolving presum- 
 ably pure selenium in nitric acid and evaporating to dryness. 
 The residue was then treated with water, and a little barium 
 hydroxide was added to remove any sulphate which might be 
 present. The solution was then filtered and the filtrate 
 evaporated to dryness. The residue was mixed with four or 
 five times its volume of dried pulverized pyrolusite, and 
 the whole was put into a porcelain crucible and heated. The 
 sublimate of selenium dioxide was carefully collected on a dry 
 watch-glass and put into a drying bottle as quickly as possible. 
 The pyrolusite prevents any reduction of the selenium dioxide 
 to selenium and the product consisted of beautiful long white 
 needles. This method of preparing the selenium dioxide, 
 which has been used for some time in this laboratory, avoids 
 contamination of the selenium dioxide by nitric acid or water, 
 resulting from the decomposition of the latter, which would 
 be possible in case this reagent were employed in the final 
 sublimation, as is recommended by Norris and Fay. The 
 hydrochloric acid used was of a sp. gr. 1.12, as recommended 
 by the originators of the process. For the experiments of Table 
 IV the dilution at the beginning was fixed at 400 cm 3 , and this 
 was increased in every case by the melting of the ice used to 
 cool the liquid. A glance at the preceding part of this paper 
 will show that at this degree of dilution the hydrochloric acid 
 present has the least effect. 
 
TITRATIONS BY SODIUM THIOSULPHATE. 211 
 
 TABLE IV. 
 
 Exp. 
 
 Amount 
 SeO, 
 taken. 
 
 HCl(sp. 
 gr. 1.12) 
 taken. 
 
 Volume 
 at begin- 
 ning of 
 titration. 
 
 Excess 
 Na,S 2 3 
 employed. 
 
 SeO, found. 
 
 Error. 
 
 
 grin. 
 
 cm 3 
 
 cm 8 
 
 cm 3 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.0616 
 
 10 
 
 400 
 
 2.28 
 
 0.0625 
 
 0.0009+ 
 
 (2) 
 
 0.0628 
 
 10 
 
 400 
 
 7.11 
 
 0.0631 
 
 0.0003+ 
 
 (3) 
 
 0.0508 
 
 10 
 
 400 
 
 11.4 
 
 0.0511 
 
 0.0003+ 
 
 (4) 
 
 0.0587 
 
 10 
 
 400 
 
 12.8 
 
 0.0594 
 
 0.0007+ mean. 
 
 (5) 
 (6) 
 
 0.0807 
 0.0633 
 
 10 
 10 
 
 400 
 400 
 
 15.3 
 
 20.85 
 
 0.0813 
 0.0638 
 
 0.0006+ 0.00005+ 
 0.0005+ 
 
 (7) 
 
 0.0682 
 
 25 
 
 400 
 
 1.11 
 
 0.0685 
 
 0.0003+ 
 
 (8) 
 
 0.0779 
 
 25 
 
 400 
 
 1.35 
 
 0.0788 
 
 0.0009+ 
 
 (9) 
 
 0.0465 
 
 25 
 
 400 
 
 18.93 
 
 0.0469 
 
 0.0004+ 
 
 These results, while not so good as those obtained by Norris 
 and Fay, are satisfactory and show that at this degree of 
 dilution the process is accurate. These results accord closely 
 with those contained in Table I. At a dilution of 400 cm 3 or 
 in the presence of only 20 cm 3 of sodium thiosulphate in 
 excess the hydrochloric acid present had no perceptible effect. 
 Of course, it must be kept in mind that the hydrochloric acid 
 acts only upon the excess of thiosulphate which is not taken 
 up by the selenium dioxide. The slight constant plus error 
 in these results cannot be accounted for by errors in the 
 standards; they were all carefully determined. Another 
 preparation of selenium dioxide was made, starting with pure 
 selenium carefully precipitated by sulphurous acid, before 
 putting it through the course of treatment previously described, 
 and the results obtained by the action of the sodium 
 thiosulphate recorded in Table V agree closely with those 
 of the preceding table. 
 
 TABLE V. 
 
 Exp. 
 
 Amount 
 SeO, 
 taken. 
 
 HC1 
 (sp. gr. 
 1.12.) 
 
 H ? Oat 
 beginning. 
 
 NajSjOg in 
 excess. 
 
 SeO, found. 
 
 Error. 
 
 
 grm. 
 
 cm* 
 
 cm 8 
 
 cm 8 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.0562 
 
 10 
 
 400 
 
 9.52 
 
 0.0566 
 
 0.0004+ 
 
 (2) 
 
 0.0651 
 
 25 
 
 400 
 
 11.20 
 
 0.0655 
 
 0.0004+ 
 
 The next step was to determine the effect of diminishing the 
 
212 INFLUENCE OF HYDROCHLORIC ACID IN 
 
 dilution and of varying the strength of acid. The following 
 table gives the results of my experiments. 
 
 TABLE VI. 
 
 Exp. 
 
 Amount 
 of SeO 2 
 taken. 
 
 Volume 
 of H 2 O 
 at begin- 
 ning. 
 
 HC1 
 
 (sp. gr. 
 i:i2). 
 
 Excess of 
 NajSjOj. 
 
 SeO, 
 taken. 
 
 Error. 
 
 
 grin. 
 
 cm 3 
 
 cm 3 
 
 cm* 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.1042 
 
 200 
 
 5 
 
 24.16 
 
 0.1041 
 
 0.0001- 
 
 (2) 
 
 0.0611 
 
 200 
 
 10 
 
 13.3 
 
 0.0611 
 
 0.0000 
 
 (3) 
 
 0.0850 
 
 200 
 
 10 
 
 21.9 
 
 0.0828 
 
 0.0022- 
 
 (4) 
 
 0.0757 
 
 200 
 
 25 
 
 13.07 
 
 0.0749 
 
 0.0008- 
 
 (5) 
 
 0.0540 
 
 200 
 
 25 
 
 21.02 
 
 0.0522 
 
 0.0018- 
 
 (6) 
 
 0.0674 
 
 300 
 
 5 
 
 10.04 
 
 0.0679 
 
 0.0005-1- 
 
 (7) 
 
 0.2416 
 
 400 
 
 5 
 
 15.9 
 
 0.2424 
 
 0.0008+ 
 
 It is apparent that at the dilution of 200 cm 3 we run into 
 difficulties, and the greater the excess of thiosulphate present 
 the greater is the error. When the amount of sodium 
 thiosulphate exceeds 20 cm 3 a reduction in the amount of acid 
 to 5 cm 3 is plainly of advantage, as is shown in a comparison 
 of Exps. (1), (3), and (5), and is not disadvantageous at 
 larger dilutions and with smaller amounts of the thiosulphate, 
 as shown in Exps. (6) and (7). The necessity of placing 
 some limits on the method of Norris and Fay has now, I 
 think, been established. The excess of the thiosulphate must 
 be carefully regulated, as well as the temperature. If one has 
 knowledge of the approximate amount of selenious acid in 
 solution, this is not a matter of great difficulty, and things 
 should be so arranged that no more than 20 cm 3 of the ^ 
 thiosulphate should ever be present in excess of the amount 
 necessary to reduce the selenious acid. If this limit 
 amounting to 0.0400 cm 8 of SeO 2 is placed upon the 
 thiosulphate, so much as 10 cm 3 of hydrochloric acid (sp. gr. 
 1.12) may be present without endangering the accuracy of the 
 process, provided the solution is diluted to 400 cm 3 at the 
 outset; if only 5 cm 8 of hydrochloric acid are present, 
 the volume at the beginning may be reduced with safety to 
 200 cm 3 . At all events, 5 cm 3 of the hydrochloric acid are 
 
TITRATIONS BY SODIUM THIOSULPHATE. 213 
 
 amply sufficient to bring about the reaction between the 
 thiosulphate and the selenium at any dilution within the range 
 of my experiments. With these precautions taken, the process 
 of Norris and Fay is simple, rapid, and accurate ; without them, 
 as the experimental results indicate, errors of considerable 
 amount may enter. 
 
 According to the method of Muthmann and Shafer,* the 
 determination of selenious acid is effected by the simple 
 addition of potassium iodide to the acidulated solution of 
 selenious acid, and the iodine set free is titrated with sodium 
 thiosulphate. In this procedure the thiosulphate is taken up 
 by the iodine as it is added to the solution, so that the danger 
 of any action between the thiosulphate and the acid is out of 
 the question. It was shown in a former paper from this 
 laboratory f that this simple method is inaccurate on account 
 of the incompleteness of reduction in the cold and in presence 
 of the iodine evolved. In a later article also from this 
 laboratory J it was shown that selenium may be completely 
 precipitated and determined with accuracy gravimetrically 
 provided the amount of potassium iodide employed is enor- 
 mously in excess of that theoretically required. This suggests 
 naturally the trial of very large excesses of potassium iodide 
 in the procees of Muthmann and Shafer. The details of 
 experiments made in this manner are given in the following 
 table. 
 
 TABLE VII 
 
 
 
 
 Vol. of 
 
 HCl 
 
 
 
 Exp. 
 
 sa 
 
 KI. 
 
 solu- 
 tion. 
 
 (sp. gr. 
 1.12). 
 
 found. 
 
 Error. 
 
 
 grm. 
 
 grlii. 
 
 cm 3 
 
 cm 3 
 
 grui. 
 
 grm. 
 
 (1) 
 
 0.0553 
 
 10 
 
 150 
 
 10 
 
 0.0558 
 
 0.0005+ 
 
 (2) 
 
 0.0574 
 
 5 
 
 150 
 
 10 
 
 0.0567 
 
 0.0007- 
 
 (3) 
 
 0.0683 
 
 5 
 
 150 
 
 10 
 
 0.0683 
 
 0.0000 
 
 (4) 
 
 0.0487 
 
 5 
 
 150 
 
 10 
 
 0.0484 
 
 0.0003- 
 
 (5) 
 
 0.2617 
 
 10 
 
 150 
 
 10 
 
 0.2589 
 
 0.0028- 
 
 * Ber. Dtsch. chem. Ges., xxyi, 1008. 
 
 t Gooch and Reynolds, Am. Jour. Sci., 1, 254. Volume I, p. 310. 
 
 t Peirce, Am. Jour. Sci., i, 1896, 416. Volume I, p. 365. 
 
214 HYDROCHLORIC ACID IN TITRATIONS. 
 
 It is obvious that for small quantities of selenium dioxide the 
 accuracy of the process is very much increased by the use of 
 large amounts of iodide, though, of course, the difficulty in 
 reading the end reaction due to the presence of precipitated 
 red selenium still remains ; but the process is still inaccurate 
 when large amounts of selenium dioxide are employed. 
 
XXVI 
 
 THE VOLATILIZATION OF THE IRON CHLOR- 
 IDES IN ANALYSIS AND THE SEPARATION 
 OF THE OXIDES OF IRON AND ALUMINUM. 
 
 BY F. A. GOOCH AND FRANKE STUART HAVENS * 
 
 IT is well known that metallic iron is easily acted upon by an 
 excess of chlorine at moderately elevated temperatures with 
 the formation of ferric chloride, and that the product of the 
 action of hydrochloric acid gas upon the metal is ferrous 
 chloride. Out of contact with air, or moisture, both chlorides 
 may be volatilized at appropriate temperatures the ferric 
 chloride below 200 C. ; the ferrous chloride at a bright red 
 heat. If water vapor, or oxygen, or air be present during 
 the heating, both chlorides are partially decomposed with 
 the formation of non-volatile residues, ferric oxide or ferric 
 oxy-chloride. 
 
 Analytical processes involving the volatilization of iron at 
 temperatures more or less elevated, in an atmosphere of chlo- 
 rine or hydrochloric acid, have been the object of considerable at- 
 tention. Thus, Fresenius,f Drown and Shimer,J and Watts, 
 have heated crude iron in chlorine to remove the metal and 
 leave the non-volatile constitutents ; and Sainte-Claire Deville || 
 has employed hydrochloric acid to volatilize iron from mix- 
 tures of that metal with alumina (obtained by heating the 
 mixed oxides of iron and aluminum in hydrogen according to 
 Rivot),^[ exposing the mixture, contained in a porcelain boat 
 and placed within a porcelain tube, to the bright red heat 
 of a charcoal furnace an operation which was bettered by 
 
 * From Am. Jour. Sci., yii, 370. t Zeitschr. anal. Chem., iv, 72. 
 
 t Jour. Inst. Min. Eng., viii, 613. Chem. News, xlv, 279. 
 
 || Ann. China. [3], xxxviii, 23. IT Ann. Chim. [3], xxx, 188. 
 
216 THE VOLATILIZATION OF THE 
 
 Cooke's* use of a tube of platinum instead of the porcelain 
 tube and a gas blowpipe in place of the charcoal furnace. 
 Sainte-Claire Devillef showed, further, that ferric oxide may 
 be converted to ferric chloride and volatilized at the heat of 
 the charcoal furnace if the current of hydrochloric acid is 
 sufficiently rapid ; but the curious effect was observed that in 
 a sufficiently limited current of the acid no chloride whatever 
 was volatilized, while the amorphous oxide was converted to 
 the highly crystalline oxide of the same composition a phe- 
 nomenon which gave rise to a theory of the natural formation 
 of specular iron in volcanic regions. 
 
 Quite recently, Mover has made record of an unsuc- 
 cessful attempt (in the course of experimentation upon the 
 volatility of certain chlorides at comparatively low tempera- 
 tures) to convert ferric oxide completely to ferric chloride by 
 the action of gaseous hydrochloric acid at about 200 C. At 
 this temperature the greater part of the iron sublimed, but a 
 residue remained, which, volatilizing neither on long heating 
 at 200 nor upon considerable elevation of the temperature, 
 proved upon examination to be ferrous chloride. In the 
 experiments to be described we have acted with gaseous 
 hydrochloric acid upon ferric oxide made by igniting the 
 nitrate prepared from pure iron deposited electrolytically by 
 high currents passing between electrodes of platinum in a 
 strong solution of ammonio-ferrous sulphate. The oxide, 
 contained in a porcelain boat, was heated within a roomy glass 
 tube over a small combustion furnace. The hydrochloric acid 
 (generated by dropping sulphuric acid into a mixture of 
 strong hydrochloric acid and salt, and dried by calcium chlor- 
 ide) entered one end of the tube and passed out at the other 
 through a water trap. In early experiments a high-reading 
 thermometer was inserted through the stopper in the exit 
 end of the tube so that its bulb was above and immediately 
 adjacent to the boat carrying the oxide. In this way the 
 actual temperatures of the vapors about the boat were fixed 
 
 * Am. Jour. Sci., xlii, 78. t Compt. rend., Hi, 1264. 
 
 J Jour. Am. Chem. Soc., xviii, 1029. 
 
IRON CHLORIDES IN ANALYSIS. 217 
 
 with considerable accuracy ; later, after a little experience in 
 gauging the effect of the burners, it was found that the tem- 
 peratures could be regulated very closely without actually 
 depending upon the thermometer. We found, as did Moyer, 
 that ferric oxide, submitted to the action of dry hydrochloric 
 acid gas, volatilizes partially as ferric chloride at low tempera- 
 tures 180 to 200 C. leaving ultimately a crystalline 
 residue which does not change visibly when heated for an 
 hour or two at 200, or even at 500, in the pure dry acid. 
 According to our experience, this residue is generally slightly 
 reddish or salmon-colored ; but sometimes, especially after a 
 second heating, the boat having been withdrawn from the 
 tube or exposed to the atmosphere (and so to moisture), the 
 residue is white. When it is white it dissolves in water, 
 yields 'the characteristic reaction for a ferrous salt with potas- 
 sium ferricyanide, gives no reaction with potassium sulphocy- 
 anide, and upon treatment in weighed amount with potassium 
 permanganate destroys the amount of that reagent theoretically 
 required for its oxidation upon the supposition that it is fer- 
 rous chloride. The slightly colored residue when treated with 
 water yields a solution showing the reaction of a ferrous salt 
 only, but when treated with hydrochloric acid and then tested 
 shows the presence of a trace of iron hi the ferric condition. 
 Doubtless the coloration of the residue is due to an included 
 trace of ferric oxide or oxychloride, which after exposure of 
 the containing crystals to slight atmospheric action, is more 
 easily reached in the second heating by the gaseous acid. The 
 amount of residue is somewhat variable, but approximates 
 under the conditions of our work to from five to ten per cent 
 of the oxide taken : thus, in one typical experiment 0.1 grm. 
 of ferric oxide left a residue which (withdrawn after cooling) 
 weighed 0.0115 grm. 
 
 The greater portion of the ferric oxide volatilizes when 
 submitted to the action of the gaseous acid at 200 quickly 
 and abundantly in the form of the greenish vapor of ferric 
 chloride, and if the operation is interrupted at this stage the 
 residue which remains is nearly black, insoluble in water, 
 
218 THE VOLATILIZATION OF THE 
 
 slightly soluble in cold hydrochloric acid, and readily soluble 
 in hot hydrochloric acid with the formation of ferric chloride. 
 It is probably something analogous to the oxychloride which 
 Rousseau* identifies as the product of the action of water 
 upon ferric chloride at 275 to 300. This dark residue 
 yields to the action of the hydrochloric acid at 180 to 200 
 only slowly, but ultimately only the residue which is essen- 
 tially ferrous chloride remains ; thereafter little volatilization 
 occurs within the range of temperature of our experimentation 
 200 to 500. 
 
 It is obvious that a reduction of iron in the ferric condition 
 to iron in the ferrous condition takes place under the con- 
 ditions of our work, and it in difficult to see how this can 
 occur otherwise than by the direct dissociation of ferric 
 chloride under the low partial pressure conditioned by the 
 brisk current of hydrochloric acid gas. The temperature of 
 formation, 180 to 200, is far below that at which such 
 dissociation is supposed to begin. Thus, Gruenewald and 
 Meyer f found, after cooling, no evidence of the dissociation 
 of ferric chloride which had been heated in the Victor Meyer 
 vapor-density apparatus to 448 in contact or partial mixture 
 with nitrogen; but ten per cent of the residue obtained by 
 heating to 518 was in the ferrous condition. Friedel and 
 Crafts,! however, did see crystals of ferrous chloride at 440 
 on the walls of a Dumas container filled with the vapor of 
 ferric chloride and nitrogen, the former exerting a partial 
 pressure of 0.75; while ferric chloride volatilized into an 
 atmosphere of chlorine without evidence of dissociation. It 
 seems rather surprising, therefore, to find so large a percen- 
 tage of dissociation as that shown in our experiments at a 
 temperature so low 180 to 200. Curiously, too, we find, 
 on repeating the experiment of heating ferric oxide in gaseous 
 hydrochloric acid, that if the temperature of the oxide is 
 450 to 500 when the brisk current of acid begins to act, 
 the whole mass of oxide is converted and volatilizes without 
 
 * Compt. rend., cxvi, 118. t Ber. Dtsch. chem. Ges., xxi, 687. 
 
 | Compt. rend., cvii, 301. 
 
IRON CHLORIDES IN ANALYSIS. 219 
 
 residue. It is hardly to be supposed that the degree of 
 dissociation at 450 to 500 can be less than that at 180 to 
 200, and a test of the sublimate, after cooling, shows that 
 it contains a ferrous salt. Plainly, ferrous chloride (formed 
 by dissociation) has volatilized, and inasmuch as the ferrous 
 chloride constituting the residue formed at 180 to 200 does 
 not volatilize in the hydrochloric acid even at 500 it is plain 
 that the volatility of the former is not determined by the 
 presence of the latter. Apparently, the cause of the com- 
 pleteness of volatilization must be sought in its rapidity ; and 
 this is not an unreasonable hypothesis, if one considers that 
 an action sufficiently rapid to keep above the boat an atmos- 
 phere of ferric chloride and its products of partial dissociation, 
 might naturally provide the very condition which would be 
 effective in counteracting the tendency of the residue to 
 dissociate before it volatilizes. If this hypothesis is correct, 
 it is plain that the introduction of chlorine gas, the active 
 product of dissociation, into the atmosphere of hydrochloric 
 acid ought to bring about the volatilization of the residue 
 of ferrous chloride, formed at 180 to 200, which refuses 
 to volatilize in the acid alone. As a matter of fact, we find 
 by experiment that if a little manganese dioxide is added 
 to the contents of the generator, so that the hydrochloric 
 acid may carry with it a little chlorine, every trace of ferric 
 oxide is volatilized from the boat at 180 to 200; and 
 the residue of ferrous chloride found at 180 to 200 when 
 the hydrochloric acid is used alone is likewise volatilized 
 at the same temperature, when the admixture of chlorine 
 is made. 
 
 These facts, that ferric oxide is completely volatile hi 
 hydrochloric acid gas applied at once at a temperature of 
 450 to 500 C., and at 180 to 200 if the acid carries a 
 little chlorine, open the way to many analytical separations 
 of iron from substances not volatile under these conditions. 
 In the experiments of the following table we have applied 
 these methods to the separation of intermixed iron and 
 aluminum oxides. The ferric oxide employed was made, as 
 
220 
 
 THE VOLATILIZATION OF THE 
 
 before, by ignition of the nitrate prepared from iron deposited 
 electrolytically by a strong current passing between platinum 
 electrodes in a solution of ammonio-ferrous sulphate.* The 
 aluminum oxide was made by igniting to a constant weight 
 the carefully washed hydroxide precipitated by ammonia from 
 a pure hydrous chloride thrown down from the solution of 
 a commercially pure chloride by hydrochloric acid.f The 
 hydrochloric acid gas was made by dropping sulphuric acid 
 into strong hydrochloric acid mixed with salt, and a little 
 manganese dioxide was added when the mixture with chlorine 
 was desired. The experimental details are given in the 
 table. 
 
 Feo0 3 
 taken. 
 
 A1 2 3 
 taken. 
 
 found. 
 
 Error. 
 
 Time. 
 
 Temperature. 
 
 Atmosphere. 
 
 grm. 
 0.1000 
 0.2000 
 0.1020 
 0.2145 
 
 gnu. 
 
 o!ibi5 
 
 0.1006 
 
 grm. 
 
 o!ibi5 
 
 0.1008 
 
 grm* 
 0.0000 
 0.0000 
 0.0000 
 0.0002+ 
 
 hrs. 
 
 1 
 
 C. 
 450-600 
 450-500 
 450-500 
 450-500 
 
 HC1 
 HC1 
 HC1 
 HC1 
 
 0.1000 
 0.1000 
 0.1072 
 0.2045 
 0.1050 
 0.2008 
 
 o!l032 
 0.1013 
 0.1032 
 0.1023 
 0.1007 
 0.1087 
 
 o!l032 
 
 0.1015 
 0.1033 
 0.1019 
 0.1006 
 0.1087 
 
 0.0000 
 0.0000 
 0.0002+ 
 0.0001+ 
 0.0004- 
 0.0001- 
 0.0000 
 
 1 
 
 H 
 1 
 
 180-200 
 180-200 
 180-200 
 180-200 
 450-500 
 450-500 
 450-500 
 
 HC1 + CL. 
 HC1 + C1 2 . 
 HC1 + C1 2 . 
 HCi + C1 2 . 
 HC1 + CL. 
 HCI + C1 2 . 
 HCI + C1 2 . 
 
 The residual alumina tested in several experiments by 
 fusion with sodium carbonate, solution in hydrochloric acid, 
 and addition of potassium sulphocyanide gave no indication 
 of the presence of iron. 
 
 The separation of the iron is obviously complete at 450 
 to 500 when the mixed oxides are submitted at once to 
 the action of hydrochloric acid gas, or at 180 to 200 when 
 
 * The use of an anode of commercially pure iron wire naturally facilitates 
 the operation, but in our experience the deposit thus obtained is likely to 
 carry traces of impurity. In an attempt, too, to prepare pure ferric oxide 
 from the oxalate thrown down out of ferrous sulphate with all precautions, 
 the material obtained still held traces of silica, and possibly alumina, amount- 
 ing to 0.0004 grm. in 0.1 grm. of the oxide. 
 
 t From Am. Jour. Sci., ii, 416. This volume, p. 20. 
 
IRON CHLORIDES IN ANALYSIS. 221 
 
 chlorine is mixed with the hydrochloric acid. Plainly, the 
 extremely high temperatures employed by Deville are un- 
 necessary if the mixed oxides are submitted at once to the 
 action of hydrochloric acid at 450 to 500 without previous 
 gentle heating in the acid atmosphere. We prefer, however, 
 to use the mixture of chlorine and hydrochloric acid, not 
 only because the temperature of the reaction is lower, but 
 because it needs no regulation, while the danger of error 
 arising from the liability of ferric chloride to dissociate, or 
 from deficiency of oxidation in the oxide treated, or from 
 mechanical loss due to rapid volatilization, is avoided. 
 
XXVII 
 
 THE TITRATION OF OXALIC ACID BY POTAS- 
 SIUM PERMANGANATE IN PRESENCE OF 
 HYDROCHLORIC ACID. 
 
 BY F. A. GOOCH AND C. A. PETEES.* 
 
 LOWENTHAL and LENSSEN f were the first to show that the 
 titration of a ferrous salt by potassium permanganate in the 
 presence of hydrochloric acid, according to the process of 
 Margueritte J is vitiated by the evolution of chloride outside 
 the main reaction, and to point out that a remedy for the 
 difficulty is to be found in the titration of the ferrous 
 salt in divided portions, other equal volumes of the ferrous 
 solution being added to the liquid in which the first titration is 
 accomplished until the amount of iron indicated by successive 
 titrations becomes constant. 
 
 Kessler showed the restraining influence of certain sulphates, 
 of manganous sulphate in particular, upon the irregular and 
 undesirable interaction of the permanganate and hydrochloric 
 acid, and Zimrnermann, || in apparent ignorance of Kessler's 
 forgotten proposal, advocated the introduction of a manganous 
 salt, best the sulphate, into the ferrous salt to be determined, 
 thus accomplishing the purpose of the empirical procedure of 
 Lowenthal and Lenssen. 
 
 The tendency toward evolution of chlorine in the oxidation 
 of a ferrous salt by permanganate, as compared with the 
 absence of such tendency in the similar oxidation of oxalic 
 acid, in presence of hydrochloric acid, was explained by 
 
 * From Am. Jour. Sci., vii, 461. t Zeitschr. anal. Chem., i, 329. 
 
 t Ann. Chim. Phys. [3], xviii, 244. 
 
 Ann. Phys., cxciv, 48 (1863) ; CXCT, 225 (1863). 
 
 |i Ann. Chem., ccxiii, 302. 
 
TITRATION OF OXALIC ACID, ETC. 223 
 
 Zimmermann on the hypothesis that an oxide of iron higher 
 than ferric oxide is formed as an intermediate product, and 
 that this unstable oxide is sufficiently active to break up 
 hydrochloric acid as well as to oxidize more of the ferrous 
 salt. Quite recently, Wagner * finds explanation of the 
 sensitiveness of the hydrochloric acid solution of the ferrous 
 salt in the probable formation of chlor-f errous acid (analogous 
 to chlor-platinic and chlor-auric acids), which suffers oxidation 
 more readily than hydrochloric acid under the action of the 
 permanganate. The protective influence of the manganous 
 salt turns apparently, as Zimmermann suggested, upon the 
 initiation of Guyard's reaction, according to which the per- 
 manganate and manganous salt interact to form a higher 
 oxide of manganese of a constitution approaching the dioxide 
 more or less closely this oxide being capable of oxidizing 
 the ferrous salt, but slow to act upon the hydrochloric acid, 
 or the chlor-ferrous acid of Wagner. According to Volhard,f 
 the reaction of Guyard is favored and hastened by heat and 
 concentration of the solution, while it is delayed by acidity 
 and dilution ; but even in solutions containing very little 
 manganous salt and a considerable quantity of free acid the 
 faint rose color developed by the careful addition of perman- 
 ganate ultimately vanishes until every trace of the manganous 
 salt is precipitated. When a considerable amount of the salt 
 is present interaction follows immediately the introduction of 
 the permanganate. Zimmermann advocates the use of 4 grams 
 of manganous sulphate uniformly in titrations of a ferrous 
 salt by permanganate, a procedure to which Wagner gives 
 acquiescence, though pointing out that a ninth of that amount 
 is all that he finds to be necessary. The excess of the 
 manganous salt can do no harm so long as the higher oxide, 
 the product of interaction of the manganous salt and the 
 permanganate, is immediately reduced by even traces of a 
 ferrous salt, and this appears to be the case at least within 
 the limits proposed by Zimmermann and Wagner. Thus we 
 
 * Massanalytische Studien, Habilitationsschrift, Leipzig, 1898. 
 t Ann. Chem., cxcviii, 318, 1879. 
 
224 
 
 TITRATION OF OXALIC ACID 
 
 find, as shown in results of the accompanying table, that so 
 much as five grams of the sulphate may be present in 135 cm 3 
 of the liquid, containing about 5 cm 3 of hydrochloric acid of 
 full strength, without interfering with the regularity of the 
 titration ; and the effect is trivial even when the amount of 
 manganous sulphate reaches ten grams. We find also practical 
 regularity of working when manganous chloride is substituted 
 for the sulphate, and in this respect our results accord with 
 those of Zimmermann and differ from those of Wagner.* 
 
 Total volume 
 at beginning 
 of titration. 
 
 HCl 
 
 (sp.gr. 1.09). 
 
 FeCl,. 
 
 KMn0 4 ^. 
 
 MnSO 4 . 5H 2 O. 
 
 MnCl 2 .4H,O. 
 
 cm 3 
 
 cm 8 
 
 cm 8 
 
 cm 8 
 
 grams. 
 
 grams. 
 
 135 
 
 10 
 
 25 
 
 21.70 
 
 1 
 
 
 135 
 
 10 
 
 25 
 
 21.70 
 
 3 
 
 ( 
 
 135 
 
 10 
 
 25 
 
 21.70 
 
 5 
 
 
 135 
 
 10 
 
 25 
 
 21.75 
 
 7 
 
 
 
 135 
 
 10 
 
 25 
 
 21.75 
 
 10 
 
 . 
 
 145 
 
 20 
 
 25 
 
 21.75 
 
 , 10 
 
 p 
 
 175 
 
 50 
 
 25 
 
 21.75 
 
 10 
 
 
 135 
 
 10 
 
 25 
 
 21.70 
 
 t 
 
 1 
 
 135 
 
 10 
 
 25 
 
 21.70 
 
 m 
 
 2 
 
 145 
 
 20 
 
 25 
 
 21.70 
 
 t 
 
 2 
 
 155 
 
 30 
 
 25 
 
 21.76 
 
 
 3 
 
 165 
 
 40 
 
 25 
 
 21.70 
 
 
 
 4 
 
 In all cases, however, in which the larger amounts of manga- 
 nous salt are present, the end reaction is marked by the advent 
 of a brownish-red precipitate rather than the clear pink of 
 the soluble permanganate, and it is obvious that in case the 
 solutions to be oxidized were not active enough to act with 
 rapidity upon the product of the Guyard reaction, difficulty 
 might follow the failure to adjust the conditions more 
 particularly. 
 
 It has been stated by Fleischerf and Zimmermann J that 
 hydrochloric acid interferes in no way with the titration of 
 oxalic acid by potassium permanganate. This statement, 
 however, is not in accord with our experience; for we find 
 that in such titrations there is a small though real waste of 
 
 * Loc. cit, p. 104. t Volumetric Analysis ; trans, by Muir, p. 71. 
 
 J Loc. cit. 
 
BY POTASSIUM PERMANGANATE. 
 
 225 
 
 permanganate proportionate to the amount of hydrochloric 
 acid present. This fact is brought out clearly in the comparison 
 of experiments of section A in the following table, in which 
 no hydrochloric acid was present, with experiments B in which 
 hydrochloric acid was present. 
 
 Temperature at beginning, about 80 C. 
 
 Approximate 
 volume at 
 beginning of 
 
 titration. 
 
 H,80 4 1 : 1. 
 
 HCl 
 (sp. gr. 1.09.) 
 
 Ammonium 
 
 oxalate 
 
 * 
 
 KMn0 4 . 
 
 Variation from 
 mean of A 
 taken as 
 standard. 
 
 
 
 
 A.. 
 
 
 
 200 
 200 
 200 
 200 
 200 
 200 
 
 cm 8 
 5 
 5 
 10 
 10 
 25 
 25 
 
 cm 
 
 cm 3 
 50 
 50 
 50 
 50 
 50 
 50 
 
 cm 
 47.50 
 47.50 
 47.50 
 47.50 
 47.50 
 47.50 
 
 cm 3 
 0.00 
 0.00 
 0.00 
 0.00 
 0.00 
 0.00 
 
 
 
 ] 
 
 B. 
 
 
 
 f 150 
 I 150 
 1 150 
 1 150 
 (500 
 }500 
 (500 
 
 10 
 10 
 10 
 10 
 5 
 10 
 10 
 
 2.5 
 2.5 
 5.0 
 10.0 
 
 ib.b 
 
 10.0 
 
 25 
 25 
 25 
 25 
 25 
 25 
 25 
 
 23.80 
 23.90 
 23.90 
 24.00 
 23.80 
 24.00 
 24.10 
 
 0.05+ 
 0.15+ 
 0.15+ 
 0.25+ 
 0.05+ 
 0.25+ 
 0.35+ 
 
 From these results it is evident that, though the error intro- 
 duced by the presence of the hydrochloric acid during the 
 action of the permanganate upon the oxalic acid is small, it is 
 plainly appreciable. The questions arise, therefore, first, as to 
 whether the secondary action of the permanganate upon the 
 hydrochloric acid may be prevented by the presence of a 
 suitable amount of a manganous salt, and, secondly, as to 
 whether in this event the reducing agent the oxalic acid 
 is sufficiently active, like the ferrous salt, to prevent the 
 premature establishment of an end color due to the Guyard 
 reaction. The latter question must naturally be settled before 
 the former can be taken up. In the accompanying table are 
 recorded the effects of varying amounts of manganous salt in 
 
 VOL. IX. 15 
 
226 
 
 TITRATION OF OXALIC ACID 
 
 presence of different amounts of sulphuric acid in the reaction 
 of permanganate upon oxalic acid. 
 
 Temperature at beginning, about 80 C. 
 
 Volume 
 at 
 beginning. 
 
 HjSO 4 1 : 1. 
 
 Ammonium 
 oxalate 
 
 ff 
 TV' 
 
 MnS0 4 . 5H,O. 
 
 KMnO 4 . 
 
 Variation 
 from 
 
 standard. 
 
 cm* 
 
 cm' 
 
 cm* 
 
 grm. 
 
 cm 3 
 
 
 f 130 
 
 5 
 
 25 
 
 
 
 23.75 
 
 0.00 
 
 
 130 
 
 6 
 
 25 
 
 0.0008 
 
 23.75 
 
 0.00 
 
 
 130 
 
 5 
 
 25 
 
 0.0032 
 
 23.76 
 
 0.00 
 
 
 130 
 
 5 
 
 25 
 
 0.0160 
 
 23.75 
 
 0.00 
 
 . 
 
 130 
 
 5 
 
 25 
 
 1 
 
 23.70 
 
 0.05- 
 
 
 130 
 
 5 
 
 25 
 
 2 
 
 23.75 
 
 0.00 
 
 
 130 
 
 5 
 
 25 
 
 2.5 
 
 23.60 
 
 0.15- 
 
 
 130 
 
 5 
 
 25 
 
 3.0 
 
 23.40 
 
 0.25- 
 
 
 130 
 
 5 
 
 25 
 
 4.0 
 
 23.60 
 
 0.15- 
 
 
 500 
 
 5 
 
 25 
 
 
 23.80 
 
 0.05+ 
 
 
 500 
 
 6 
 
 25 
 
 0.0008 
 
 23.80 
 
 0.05-f 
 
 
 500 
 
 5 
 
 25 
 
 0.0032 
 
 23.80 
 
 0.05+ 
 
 . 
 
 500 
 
 5 
 
 25 
 
 1 
 
 23.70 
 
 0.05- 
 
 
 500 
 
 5 
 
 25 
 
 2 
 
 23.40 
 
 0.35- 
 
 
 500 
 
 5 
 
 25 
 
 3 
 
 23.50 
 
 0.25- 
 
 
 500 
 
 5 
 
 25 
 
 4 
 
 23.30 
 
 0.45- 
 
 
 130 
 
 no 
 
 25 
 
 1 
 
 23.80 
 
 0.05+ 
 
 
 130 
 
 ho 
 
 25 
 
 2 
 
 23.75 
 
 0.00 
 
 
 130 
 
 jio 
 
 25 
 
 3 
 
 23.65 
 
 0.10- 
 
 
 130 
 
 1 10 
 
 25 
 
 4 
 
 23.50 
 
 0.25- 
 
 
 130 
 
 (15 
 
 25 
 
 2 
 
 23.75 
 
 0.00 
 
 " 
 
 130 
 
 )l6 
 
 25 
 
 4 
 
 23.70 
 
 0.05- 
 
 
 130 
 
 1 16 
 
 25 
 
 5 
 
 23.50 
 
 0.25- 
 
 
 130 
 
 (30 
 
 25 
 
 2 
 
 23.75 
 
 0.00 
 
 
 130 
 
 ?30 
 
 25 
 
 4 
 
 23.70 
 
 0.05- 
 
 
 130 
 
 <30 
 
 25 
 
 5 
 
 23.75 
 
 0.00 
 
 From the results given it is evident that the persistence of 
 the Giiyard reaction is liable to interfere with the end reaction 
 of oxidation of oxalic acid unless an adjustment is made 
 between the quantity of the manganous salt, the amount of 
 acid, and the dilution. In hot solutions of a total volume of 
 130 cm 3 at the beginning, no more than 2 grms. of the 
 manganous sulphate should accompany 5 to 10 cm 3 of the 
 1 : 1 sulphuric acid ; when the total volume at the beginning 
 reaches 500 cm 3 no more than a single gram of the salt should 
 be present with 5 cm 8 of the 1 : 1 sulphuric acid. The amount 
 of manganous salt may, however, be increased considerably if 
 the quantity of acid is increased. 
 
 OF THE 
 
 UNIVERSITY 
 
BY POTASSIUM PERMANGANATE. 
 
 227 
 
 As Kessier has noted, a sufficiency of the manganous salt, 
 acting no doubt as the medium of transfer of oxygen, may 
 bring about interaction between the permanganate and the 
 oxalic acid at atmospheric temperatures without the tedious 
 delay ordinarily encountered in the attempt to consummate 
 that action in cold solutions. It would seem natural that the 
 manganic hydroxide formed in the Guyard reaction at low 
 temperatures should yield more readily to the reducing action 
 of the oxalic acid than the more anhydrous form to be expected 
 in hot solutions, so that at such temperatures the limits as to 
 proportions of manganous salt, acid, and dilution, within 
 which favorable action may take place, should be wider; 
 moreover, the undesirable action of the permanganate upon 
 
 Temperature 20-26 C. 
 
 Number 
 of 
 experi- 
 
 BMOfc* 
 
 Volume 
 at begin- 
 ning of 
 
 titration. 
 
 W 
 
 HCl 
 
 Ammo- 
 nium 
 oxalate 
 
 KMn0 4 . 
 
 MnSO 4 . 
 5H 2 0. 
 
 MnCl 2 . 
 4HjO. 
 
 Variation 
 from 
 standard. 
 
 
 
 cm 3 
 
 cm' 
 
 cm 
 
 cm 
 
 grin. 
 
 grin. 
 
 cm 
 
 (ij 
 
 130 
 130 
 
 
 10 
 10 
 
 25 
 
 25 
 
 23.90 
 23.90 
 
 0.0040 
 0.0120 
 
 
 0.15+ 
 0.15+ 
 
 (3) 
 
 130 
 
 
 10 
 
 25 
 
 23.80 
 
 0.0250 
 
 
 0.05+ 
 
 (4) 
 
 130 
 
 
 10 
 
 25 
 
 23.75 
 
 0.0400 
 
 
 0.00 
 
 (5) 
 
 130 
 
 
 10 
 
 25 
 
 23.76 
 
 0.0500 
 
 
 0.01+ 
 
 (6) 
 
 130 
 
 
 10 
 
 25 
 
 23.70 
 
 0.1000 
 
 
 0.05- 
 
 (7) 
 
 130 
 
 
 10 
 
 25 
 
 23.75 
 
 0.2000 
 
 
 0.00 
 
 (8) 
 
 130 
 
 
 10 
 
 25 
 
 24.20 
 
 
 0.0200 
 
 0.45+ 
 
 (9) 
 (10) 
 
 130 
 130 
 
 
 10 
 10 
 
 25 
 25 
 
 23.95 
 23.80 
 
 
 0.0200 
 0.0400 
 
 0.20+ 
 0.05+ 
 
 (11) 
 
 130 
 
 
 20 
 
 25 
 
 23.75 
 
 
 0.0400 
 
 0.00 
 
 (12) 
 
 130 
 
 
 30 
 
 25 
 
 23.75 
 
 
 0.0400 
 
 0.00 
 
 (13) 
 
 130 
 
 
 10 
 
 25 
 
 23.75 
 
 1.0000 
 
 t 
 
 0.00 
 
 (14) 
 
 130 
 
 
 10 
 
 25 
 
 23.75 
 
 2.0000 
 
 > 
 
 0.00 
 
 (15) 
 
 130 
 
 
 10 
 
 25 
 
 23.75 
 
 3.0000 
 
 
 0.00 
 
 (16) 
 
 130 
 
 1 
 
 
 25 
 
 23.72 
 
 
 1.0000 
 
 0.03- 
 
 (17) 
 
 130 
 
 1 
 
 
 25 
 
 23.74 
 
 
 2.0000 
 
 0.01- 
 
 (18) 
 
 130 
 
 1 
 
 
 25 
 
 23.72 
 
 
 3.0000 
 
 0.03- 
 
 (19) 
 
 130 
 
 2 
 
 
 25 
 
 23.70 
 
 
 0.5000 
 
 0.05- 
 
 (20) 
 
 130 
 
 3 
 
 
 25 
 
 23.76 
 
 
 0.5000 
 
 0.00 
 
 Temperature about 80. 
 
 (21) 
 
 145 
 
 10 
 
 10 
 
 25 
 
 23.90 
 
 0.5000 
 
 
 0.15+ 
 
 (22) 
 
 145 
 
 10 
 
 10 
 
 25 
 
 23.70 
 
 1.0000 
 
 
 0.05- 
 
 (23) 
 
 500 
 
 10 
 
 10 
 
 25 
 
 23.75 
 
 1.0000 
 
 
 0.00 
 
 (24) 
 
 500 
 
 m 
 
 10 
 
 25 
 
 28.70 
 
 1.0000 
 
 
 0.05- 
 
 (25) 
 
 500 
 
 
 
 10 
 
 25 
 
 24.10 
 
 0.5000 
 
 
 0.35+ 
 
228 TITRATION OF OXALIC ACID 
 
 hydrochloric acid, when that acid is present, should be less 
 appreciable at lower temperatures. In our experiments, 
 therefore, upon the oxidation of oxalic acid by potassium 
 permanganate in presence of hydrochloric acid, we have studied 
 the effect of varying the proportions of the manganous salt 
 both at atmospheric temperatures and the higher temperatures 
 generally employed. 
 
 From these results it appears that the presence of a suit- 
 able amount of manganous salt either the sulphate (4-7), 
 (13-15), (22-24) or the chloride (10-12), (16-20) is 
 capable, either in cold solution (1-20) or in hot solution 
 (22-24) of preventing the action of the permanganate upon 
 the hydrochloric acid. It appears, also, that, for a given 
 dilution and strength of acid, less manganous salt is needed 
 in the cold solution (4-7) than in the hot solutions (22-24). 
 Thus, in the hot solution, at a dilution at 145 cm 3 to 500 cm 3 
 1 grm. of manganous sulphate must be present with 5 cm 3 . 
 of strong hydrochloric acid, with or without sulphuric acid ; 
 while in the cold solution 0.04 grm. of either the sulphate 
 or chloride is enough to secure adequate protective effect. 
 Experience showed, however, that 0.5 grm. or 1.0 grm. of the 
 manganous salt should be present in order to push the re- 
 action with reasonable speed in cold solutions. 
 
 Wagner* has made record of the increased evolution of 
 chlorine in oxidations of ferrous chloride by potassium 
 permanganate in presence of various salts, of which barium 
 chloride was the most active. We have made some experi- 
 ments, therefore, to determine whether such action would 
 appear in the oxidation of oxalic acid in cold solutions 
 containing certain salts, and, if so, whether it would be 
 preventable by the presence of the manganous salt under our 
 conditions of working. From the results given in the 
 accompanying table, it is plain that the evolution of chlorine 
 in cold solutions is less in the presence of these salts than 
 when hydrochloric acid is used without them, and that such 
 evolution may be entirely prevented (within the proportions 
 
 * Loc. cit. 
 
BY POTASSIUM PERMANGANATE. 
 
 229 
 
 of our work) by the presence of 0.5 grm. to 1 grm. of 
 manganous chloride. 
 
 Finally, it appears as the result of an investigation, that the 
 titration of oxalic acid by potassium permanganate in presence 
 of hydrochloric acid is ordinarily attended with some inaccu- 
 racy due to liberation of chlorine from the hydrochloric acid ; 
 that this tendency may be overcome by the presence of a 
 manganous salt either the sulphate or chloride; that 1 grm. 
 of the manganous salt is enough to so affect the conditions of 
 equilibrium that titrations in moderate volumes (100 cm 3 to 
 500 cm 3 ) and in presence of hydrochloric acid (5 cm 3 to 15 cm 3 
 of the strong acid) may be conducted with safety and reasonable 
 rapidity, either with or without sulphuric acid, at the ordinary 
 atmospheric temperature. 
 
 Volume at Beginning of Titration = 140 cm 8 . 
 Temperature = 20-24 C. 
 
 Ammonium 
 oxalate. 
 
 HCl. 
 
 strongest. 
 
 MnCl 2 . 
 4H 2 0. 
 
 BaClo. 
 
 SrCl,. 
 
 CaCla- 
 
 MgCl,. 
 
 KMn0 4 
 used. 
 
 Error. 
 
 cm 8 
 25 
 
 cm 8 
 
 5 
 
 grm. 
 
 0.5 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 C1113 
 
 26.05 
 
 cm 3 
 0.00 
 
 25 
 25 
 25 
 25 
 25 
 
 5 
 5 
 5 
 5 
 5 
 
 
 
 
 
 
 
 
 
 
 
 27.45 
 26.50 
 26.53 
 26.36 
 26.13 
 
 1.40+ 
 0.45+ 
 0.48+ 
 0.35+ 
 0.08+ 
 
 25 
 25 
 25 
 25 
 
 5 
 5 
 6 
 5 
 
 0.5 
 0.5 
 0.5 
 0.5 
 
 2 
 
 
 
 
 
 
 
 26.05 
 26.10 
 26.10 
 26.05 
 
 0.00 
 0.05+ 
 0.05+ 
 0.00 
 
 25 
 25 
 25 
 25 
 
 10 
 10 
 10 
 10 
 
 1.0 
 1.0 
 1.0 
 1.0 
 
 2 
 
 V .. 
 
 V 
 
 V 
 
 26.10 
 26.05 
 26.06 
 26.11 
 
 +0.05 
 0.00 
 0.01+ 
 0.06+ 
 
XXVIII 
 
 THE ESTIMATION OF IKON IN THE FERRIC STATE 
 BY REDUCTION WITH SODIUM THIOSULPHATE 
 AND TITRATION WITH IODINE. 
 
 BY JOHN T. NORTON, JB.* 
 
 THE action of sodium thiosulphate on ferric iron has long 
 been known and depends upon the following reaction: 
 
 2FeCl 3 + 2Na2S 2 O 8 = SFeCL, + Na 2 S 4 O 6 + 2NaCl. 
 
 As early as 1859 Schererf proposed a method for the 
 estimation of ferric iron depending on the above reaction. 
 Scherer's method of procedure was to act upon a solution of 
 ferric chloride with sodium thiosulphate until the purple 
 color produced by the interaction of these two salts just 
 vanished. Mohr's $ experimental tests of this process were 
 not successful. A year or two later Kremer and Landolt, 
 after a careful investigation of Scherer's process, recommended 
 it with the modification that any free hydrochloric acid pres- 
 ent should be neutralized by sodium acetate until the solution 
 assumed a red color, just enough hydrochloric acid added to 
 destroy this red color, and sodium thiosulphate run into the 
 solution in slight excess. When the liquid became perfectly 
 colorless and gave no reaction for ferric iron with potassium 
 sulphocyanide, the excess of sodium thiosulphate was titrated 
 back with iodine and starch. The authors also state that 
 the ferric iron should not be present in concentrated solution. 
 Very good results were claimed for this process, but it 
 apparently gained but slight recognition. 
 
 * From Am. Jour. Sci., viii, 25. 
 
 t Gel. Anzeig. k. Bayrisch. Acad., Aug. 31, 1859. 
 
 $ Ann. Chem. Pharm., cxiii, 260. Zeitschr. anal. Chem., i, 214. 
 
ESTIMATION OF IRON IN THE FERRIC STATE. 231 
 
 Oudemanns,* who was the next to study the action of ferric 
 iron and sodium thiosulphate, claimed that the addition of a 
 small quantity of cupric salt to the iron solution hastened the 
 reducing action of the sodium thiosulphate. Mohr,f however, 
 condemned this method also as unreliable, both because the 
 sodium thiosulphate acted upon the copper as well as the iron 
 and also because the potassium sulphocyanide, added as an 
 indicator of the completeness of the reduction, produced a 
 precipitate of cupric sulphocyanide which interfered with 
 the reaction. In a second paper Oudemanns J reiterated 
 his former statement as to the accuracy of his method but 
 advised the use of a smaller quantity of the cupric salt. 
 An improvement on Oudemanns' process was proposed by 
 Haswell, who mixed the moderately acid solution of ferric 
 chloride in the presence of a cupric salt with a few drops of 
 sodium salicylate and then reduced with sodium thiosulphate 
 previously standardized upon a known quantity of iron by 
 the same process and estimated the excess by potassium 
 dichromate. Bruel || modified this process by operating 
 without the copper solution, relying merely on the discharge 
 of the violet color in a boiling solution by sodium thiosul- 
 phate standardized on a ferric solution of known strength. 
 
 Although considerable work has been done on the reaction 
 between ferric iron and sodium thiosulphate, no process 
 depending upon this reaction has obtained acceptance. In 
 view, therefore, of previous work on the action of hydro- 
 chloric acid upon sodium thiosulphate If and with the idea that 
 a careful control of the dilution and quantity of acid present 
 might greatly better the accuracy of the method, it has 
 seemed to me to be desirable to study this process again in 
 detail. 
 
 The ferric oxide employed in the experiments was prepared 
 with great care by the ignition of ferrous oxalate obtained by 
 
 * Zeitschr. anal. Chem., vi, 129. t Titrirmethode, 5 et Aufl., 294. 
 J Zeitschr. anal. Chem., ix, 362. 
 Kepertorium der analytischen Chem., i, 179. 
 || Compt. rend., xcvii, 954. 
 1 Am. Jour. Sci., vii, 287. This volume, p. 206. 
 
232 ESTIMATION OF IRON IN THE FERRIC STATE 
 
 acting with oxalic acid on pure ammonium ferrous sulphate. 
 To ascertain, however, if this oxide contained any impurity, 
 about 0.5 of a grm. was put into a porcelain boat and sub- 
 mitted to the action of a current of hydrochloric acid gas and 
 chlorine at a temperature of about 280 C. (according to a 
 process recently described from this laboratory *) until all the 
 ferric salt is volatilized hi the form of ferric chloride. A 
 residue of 0.0010 grm. for every 0.5 of a grm. of the oxide 
 was found, and this correction, small for the amounts 
 generally used, has been applied in the following determina- 
 tions. The sodium thiosulphate used was taken in nearly ^ 
 solution and was standardized against an approximately 
 decinormal solution of iodine which had been determined by 
 comparison with decinormal arsenious acid made from care- 
 fully resublimed arsenious oxide. 
 
 In those experiments which deal with amounts of ferric 
 oxide not exceeding 0.2 of a grm., measured portions of a 
 solution of ferric chloride made of known strength by dis- 
 solving about 2 grms. of the pure carefully weighed ferric 
 oxide in 20 cm 3 of strong hydrochloric acid and diluting to 
 one liter, were drawn from a burette. In the case of the 
 krger quantities of ferric oxide the salt was weighed out, 
 dissolved in hydrochloric acid and brought to the required 
 dilution. The ferric chloride, either drawn from the burette 
 or prepared directly from the weighed oxide, was diluted with 
 water, a drop of potassium sulphocyanide added to serve as 
 an indicator and an excess of sodium thiosulphate was run 
 in until, after standing for a few minutes, the solution became 
 perfectly colorless, and the excess of sodium thiosulphate 
 was then titrated back with decinormal iodine after the 
 addition of starch. 
 
 Several sources of error are, plainly, possible in the process : 
 incompleteness in the reduction of the ferric salt ; decomposition 
 of the thiosulphate by the acid, resulting in the subsequent 
 over-run of iodine; the possible tendency of the ferric salt 
 under concentration to oxide the thiosulphate to the condition 
 * Gooch and Havens, Am. Jour. Sci., vii, 370. This volume, p. 215. 
 
BY DEDUCTION WITH SODIUM THIOSULPHATE. 233 
 
 of the sulphate rather than to that of the tetrathionate ; and 
 finally the oxidizing action of the air, which may tend to 
 keep up progressive oxidation of the iron salt and excessive 
 expenditure of thiosulphate. The first three sources of diffi- 
 culty tend to produce errors of deficiency; the fourth an 
 error of excess. 
 
 The first step in the experimental study of the process was 
 to determine the effect of varying dilution upon the estimation 
 of a given quantity of iron reduced by sodium thiosulphate, 
 taken in practically uniform excess above the amount theo- 
 retically required, in the presence of 1 cm 3 of hydrochloric 
 acid. 
 
 TABLE I. 
 
 Exp. 
 
 Fe 2 3 
 taken. 
 
 Fe 2 3 
 corrected. 
 
 Dilution. 
 
 HC1. 
 
 NaAO, 
 
 in excess. 
 
 Fe 2 8 
 found. 
 
 Error. 
 
 
 grm. 
 
 grm. 
 
 cm* 
 
 cm 8 
 
 cm 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.1000 
 
 0.0998 
 
 100 
 
 1 
 
 18.08 
 
 0.0957 
 
 0.0041- 
 
 2 
 
 0.1000 
 
 0.0998 
 
 200 
 
 
 20 
 
 0.0966 
 
 0.0032- 
 
 (3) 
 
 0.1000 
 
 0.0998 
 
 300 
 
 
 17.56 
 
 0.0995 
 
 0.0003- 
 
 (4) 
 
 0.1000 
 
 0.0998 
 
 400 
 
 
 17.16 
 
 0.0998 
 
 0.0000 
 
 M 
 
 0.1000 
 
 0.0998 
 
 600 
 
 
 17.76 
 
 0.0996 
 
 0.0002- 
 
 (6) 
 
 0.1000 
 
 0.0998 
 
 800 
 
 
 17.65 
 
 0.0993 
 
 0.0005- 
 
 (7) 
 
 0.1000 
 
 0.0998 
 
 1000 
 
 
 18.02 
 
 0.0988 
 
 0.0010- 
 
 (8) 
 
 0.1000 
 
 0.0998 
 
 1200 
 
 
 17.95 
 
 0.0977 
 
 0.0021- 
 
 (9) 
 
 0.1000 
 
 0.0998 
 
 1400 
 
 
 17.99 
 
 0.0965 
 
 0.0033- 
 
 (10) 
 
 0.1000 
 
 0.0998 
 
 1600 
 
 1 
 
 18.01 
 
 0.0947 
 
 0.0051- 
 
 (11) 
 
 0.2001 
 
 0.1997 
 
 400 
 
 2 
 
 27.05 
 
 0.2029 
 
 0.0032+ 
 
 (12) 
 
 0.2001 
 
 0.1997 
 
 800 
 
 2 
 
 15.95 
 
 0.1998 
 
 0.0001+ 
 
 (13) 
 
 0.4998 
 
 0.4988 
 
 1000 
 
 2 
 
 22.36 
 
 0.5104 
 
 0.0116+ 
 
 (14) 
 
 0.5051 
 
 0.5041 
 
 1800 
 
 4 
 
 15.27 
 
 0.5026 
 
 0.0015 
 
 (15) 
 
 0.4002 
 
 0.-3994 
 
 1500 
 
 4 
 
 27.29 
 
 0.3996 
 
 0.0002+ 
 
 (16) 
 
 0.7502 
 
 0.7487 
 
 1000 
 
 1 
 
 9.73 
 
 0.7572 
 
 0.0085+ 
 
 (17) 
 
 0.7029 
 
 0.7015 
 
 2000 
 
 4 
 
 12.67 
 
 0.7004 
 
 0.0011- 
 
 This table shows plainly that with quantities of ferric oxide 
 present up to 0.1 grm. the dilution can vary from 400 cm 3 to 
 1000 cm 3 for each cm 3 of strong hydrochloric acid and still 
 give excellent results. At a dilution greater than 1000 cm 3 
 the action of the thiosulphate is evidently incomplete, and at 
 a smaller dilution than 400 cm 3 the decomposing action of the 
 acid on the thiosulphate becomes noticeable. When larger 
 quantities of iron oxide are dealt with, it appears that the 
 
234 ESTIMATION OF IRON IN THE FERRIC STATE 
 
 dilution ought to be increased proportionally with the quantity 
 of ferric oxide present as well as with that of the acid. This 
 is illustrated in experiments 9-15 of the table. On this 
 account it seems necessary, assuming that the quantity of 
 acid present is always kept within the maximum strength 
 mentioned, 1 cm 3 to 400 cm 3 , to regulate the dilution from the 
 approximate quantity of the iron so that not less than 400 cm 3 
 of water shall be used to every 0.1 grm. of iron oxide present. 
 Under properly regulated conditions of dilution as regards 
 acid and the iron salt, the reduction is completed in from five 
 to ten minutes. 
 
 Great excesses of acid, however, contrary to the statement 
 of Kremer,* retard the reduction greatly, and, hi spite of the 
 tendency of the thiosulphate to decomposition and the pro- 
 duction of errors of deficiency under such circumstances, plus 
 errors due to partial oxidation come to light. This fact appears 
 in the following table, which records the results of processes 
 lasting many hours. 
 
 TABLE II 
 
 Exp. 
 
 Fe 2 s 
 taken. 
 
 Fe 2 8 
 corrected. 
 
 DUution. 
 
 HC1. 
 
 Na 2 S 2 0, 
 in excess. 
 
 Fe 2 8 
 found. 
 
 Error. 
 
 
 grm. 
 
 grm. 
 
 cm 8 
 
 cm* 
 
 cm 3 
 
 grm. 
 
 grm. 
 
 (18) 
 
 0.5012 
 
 0.6002 
 
 1700 
 
 10 
 
 25.99 
 
 5308 
 
 0.0306+ 
 
 (19) 
 
 0.7512 
 
 0.7497 
 
 1200 
 
 16 
 
 57.8 
 
 7685 
 
 0.0188+ 
 
 (20) 
 
 0.7520 
 
 0.7505 
 
 2000 
 
 16 
 
 56.4 
 
 7983 
 
 0.0478+ 
 
 (21) 
 
 0.7520 
 
 0.7505 
 
 1700 
 
 15 
 
 27.2 
 
 7627 
 
 0.0122+ 
 
 As to the temperature at which the reduction should be 
 made, my experience, contrary to that of Kremer, goes to show 
 that no elevation above atmospheric conditions is necessary ; 
 under the conditions of acidity and dilution laid down, the 
 process of reduction is complete within ten minutes after the 
 introduction of the thiosulphate ; moreover, former experience f 
 shows clearly the danger of submitting mixtures of sodium 
 thiosulphate and acid to temperatures much above the ordinary. 
 On the other hand, artificial reduction of temperature tends 
 
 * Zeitschr. anal. Chem., i, 214. 
 
 t Am. Jour. Sci., vol. vii, 287. This volume, p. 206. 
 
EY REDUCTION WITH SODIUM THIOSULPHATE. 235 
 
 to retard the action to an impossible degree. Thus, in an 
 experiment it took five minutes to reduce 0.0500 of ferric 
 oxide at 21 C. completely at a dilution of 200 cm 3 and in 
 the presence of |- cm 3 of hydrochloric acid ; under conditions 
 otherwise precisely similar excepting that the temperature was 
 lowered to C., the action lingered forty-five minutes. 
 
 Lastly, the question as to the excess of thiosulphate necessary 
 to complete the reduction within a reasonable time must be 
 considered. In nearly all previously recorded experiments 
 the excess of thiosulphate was not less than 15 cm 3 of the 
 ^ solution. The following table shows the effect of dimin- 
 ishing this excess. 
 
 TABLE HI. 
 
 Kxp. 
 
 Fe,0 8 
 taken. 
 
 Fe,0 s 
 corrected. 
 
 Dilution. 
 
 HC1. 
 
 NaAO, 
 
 in excess. 
 
 aa 
 
 Error. 
 
 
 grin* 
 
 grm. 
 
 cm 3 
 
 cm 3 
 
 cm s 
 
 grm. 
 
 grm. 
 
 (22) 
 
 0.0250 
 
 0.0250 
 
 400 
 
 
 12.2 
 
 0.0241 
 
 0.0009- 
 
 (23) 
 
 0.0500 
 
 0.0499 
 
 400 
 
 
 12.2 
 
 0.0495 
 
 0.0004- 
 
 (24) 
 
 0.0500 
 
 0.0499 
 
 400 
 
 
 13.66 
 
 0.0493 
 
 0.0006- 
 
 (25) 
 
 0.1000 
 
 0.0998 
 
 400 
 
 1 
 
 7.31 
 
 0.0984 
 
 0.0014- 
 
 (26) 
 
 0.1000 
 
 0.0998 
 
 400 
 
 1 
 
 7.63 
 
 0.0972 
 
 0.0026- 
 
 (27) 
 
 0.1001 
 
 0.0999 
 
 400 
 
 1 
 
 12.88 
 
 0.1007 
 
 0.0008+ 
 
 (28) 
 
 0.1498 
 
 0.1495 
 
 600 
 
 1* 
 
 11.97 
 
 0.1475 
 
 0.0020- 
 
 (29) 
 
 0.1996 
 
 0.1992 
 
 800 
 
 2 
 
 12.43 
 
 0.1980 
 
 0.0012- 
 
 From the above experiments taken in connection with those 
 of Table I it is clear that there should always be present an 
 excess of at least 15 cm 3 of the T ^ solution of sodium thiosul- 
 phate. If the quantity of hydrochloric acid is kept very low 
 there is no reason why this excess of thiosulphate could not 
 be considerable without producing any disturbing effect. 
 Practically, however, the presence of an excess between the 
 limits of 15 cm 8 and 35 cm 3 of the ^ solution has been found 
 to give the most satisfactory results. 
 
 To recapitulate, then, it has been shown that the dilution 
 must be at least 400 cm 3 for each 0.1 of a grm. of iron oxide 
 present, that the quantity of acid should never exceed 1 cm 3 
 of the strong acid to each 400 cm 3 of water, that the time of 
 reduction must be short to avoid progressive oxidation, that 
 
236 ESTIMATION OF IRON IN THE FERRIC STATE 
 
 the temperature of the solution should be kept at the normal 
 temperature of the atmosphere, and finally that the excess of 
 sodium thiosulphate present should never be less than 15 cm 3 
 of the f solution. In the case of large dilution the use of 
 freshly boiled water is recommended so as to avoid the 
 reoxidizing effect of the air upon the reduced iron. In the 
 experiments included in the following table, the above 
 precautions were closely adhered to and manifestly satisfactory 
 results were obtained. 
 
 TABLE IV. 
 
 Exp. 
 
 Fe 2 3 
 taken. 
 
 Fe 2 3 
 corrected. 
 
 Dilution. 
 
 HC 
 
 i. 
 
 Excess 
 Na,S 2 3 
 
 found. 
 
 Error. 
 
 
 grin. 
 
 grm. 
 
 cm 8 
 
 cm 
 
 3 
 
 cm* 
 
 grm. 
 
 grm. 
 
 (30) 
 
 0.0125 
 
 0.0125 
 
 200 
 
 
 
 23.5 
 
 0.0125 
 
 0.0000 
 
 (31) 
 
 0.0250 
 
 0.0250 
 
 400 
 
 
 
 21.98 
 
 0.0250 
 
 0.0000 
 
 (32) 
 
 0.0250 
 
 0.0250 
 
 400 
 
 
 
 17 
 
 0.0250 
 
 0.0000 
 
 (33) 
 
 0.0250 
 
 0.0250 
 
 400 
 
 
 
 17 
 
 0.0250 
 
 0.0000 
 
 (34) 
 
 0.0500 
 
 0.0499 
 
 400 
 
 
 
 24 
 
 0.0498 
 
 0.0001- 
 
 (35) 
 
 0.0500 
 
 0.0499 
 
 400 
 
 
 
 19 
 
 0.0498 
 
 0.0001- 
 
 (36) 
 
 0.0500 
 
 0.0499 
 
 400 
 
 
 
 15.1 
 
 0.0497 
 
 0.0002- 
 
 (37) 
 
 0.0500 
 
 0.0499 
 
 400 
 
 
 
 19 
 
 0.0498 
 
 0.0001- 
 
 (38) 
 
 0.1001 
 
 0.0999 
 
 400 
 
 
 
 23.1 
 
 0.0993 
 
 0.0006- 
 
 (39) 
 
 0.1001 
 
 0.0999 
 
 400 
 
 
 
 17.93 
 
 0.0997 
 
 0.0002 
 
 (40) 
 
 0.1001 
 
 0.0999 
 
 400 
 
 
 
 22.92 
 
 0.0997 
 
 0.0002- 
 
 (41) 
 
 0.1001 
 
 0.0999 
 
 400 
 
 
 
 18 
 
 0.0997 
 
 0.0002- 
 
 (42) 
 
 0.1001 
 
 0.0999 
 
 400 
 
 
 
 16 
 
 0.0996 
 
 0.0003- 
 
 (43) 
 
 0.1498 
 
 0.1495 
 
 600 
 
 j 
 
 - 
 
 23.26 
 
 0.1493 
 
 0.0002- 
 
 (44) 
 
 0.1498 
 
 0.1495 
 
 600 
 
 ^ 
 
 - 
 
 16.66 
 
 0.1493 
 
 0.0002- 
 
 (45) 
 
 0.1498 
 
 0.1495 
 
 600 
 
 
 - 
 
 26.87 
 
 0.1475 
 
 0.0020- 
 
 (46) 
 
 0.1996 
 
 0.1992 
 
 800 
 
 2 
 
 
 22.38 
 
 0.1990 
 
 0.0002- 
 
 (47) 
 
 0.1996 
 
 0.1992 
 
 800 
 
 2 
 
 
 17.29 
 
 0.1999 
 
 0.0007+ 
 
 (48) 
 
 0.1996 
 
 0.1992 
 
 800 
 
 2 
 
 
 22.20 
 
 0.1991 
 
 0.0001- 
 
 (49) 
 
 0.4045 
 
 0.4037 
 
 1600 
 
 4 
 
 
 16.03 
 
 0.4042 
 
 0.0005+ 
 
 (50) 
 
 0.4045 
 
 0.4037 
 
 1600 
 
 4 
 
 
 16.2 
 
 0.4023 
 
 0.0014- 
 
 (51) 
 
 0.4018 
 
 0.4010 
 
 1600 
 
 4 
 
 
 16.34 
 
 0.4007 
 
 0.0003- 
 
 (52) 
 
 0.5051 
 
 0.5041 
 
 1800 
 
 4 
 
 
 15.27 
 
 0.5026 
 
 0.0015- 
 
 As seen in the table this process is very accurate, especially 
 in the use of small amounts of ferric oxide. The introduc- 
 tion of cupric sulphate as recommended by Oudemanns, or of 
 sodium salicylate according to Haswell's method, seems to be 
 unnecessary and only complicates the process. 
 
 In treating ferric oxide, the following method of procedure 
 is recommended. Dissolve an amount not exceeding 0.2 grm. 
 of the oxide in hydrochloric acid, evaporate to a pasty mass 
 
BY REDUCTION WITH SODIUM THIOSULPHATE. 237 
 
 dilute to about 800 cm 3 with freshly boiled water, add a drop of 
 potassium sulphocyanide, and into this solution run 50 cm 3 of 
 approximately ^ sodium thiosulphate; allow the liquid to 
 stand until perfectly colorless and determine the excess of 
 thiosulphate by ^ iodine and starch. For quantities of iron 
 oxide up to 0.2 of a gram this process is quick and most accu- 
 rate; when care is taken to preserve the relations of acidity 
 and dilution, twice the amount of ferric oxide mentioned 
 above may be handled. 
 
XXIX 
 
 THE DETERMINATION OF TELLUROUS ACID IN 
 PRESENCE OF HALOID SALTS. 
 
 BY F. A. GOOCH AND C. A. PETEKS.* 
 
 THE estimation of tellurous acid by oxidation with excess of 
 potassium permanganate (either in acid or alkaline solution), 
 destruction of the higher oxides of manganese or the manga- 
 nate by standard oxalic acid in presence of sulphuric acid, and 
 titration of the residual oxalic acid by more permanganate, has 
 been shown by Braunerf to be feasible. The tendency of the 
 permanganate to throw off too much oxygen when the oxida- 
 tion is made in solutions strongly acidified with sulphuric acid 
 (as must be the case if the tellurous oxide is to be held perma- 
 nently in solution by sulphuric acid) necessitates the applica- 
 tion of a considerable correction.:]: Fortunately, however, as 
 has been shown in a former paper from this laboratory,! when 
 the tellurous oxide is dissolved originally in an alkaline hydrox- 
 ide and the solution made acid only to a limited degree with 
 sulphuric acid either before or after oxidation by the perman- 
 ganate, no correction appears to be necessary. Thus, when 
 an excess of permanganate is added to the alkaline solution, 
 followed by an excess of oxalic acid and sulphuric acid to an 
 amount not exceeding 5 cm 3 of the [1 : 1] mixture with water, 
 the titration of the residual oxalic acid by more permanganate 
 (after heating to 80 C.) leads to results which give no indi- 
 cation of over-decomposition of the permanganate; so also, 
 when the process is similarly conducted excepting that before 
 addition of the permanganate the original alkaline solution is 
 acidified with sulphuric acid [1 : 1] to an amount 1 cm 3 in excess 
 
 * From Am. Jour. Sci., viii, 122. t Jour. Chem. Soc., lix, 238. 
 
 J Loc. cit, p. 249. 
 
 Gooch and Danner, Am. Jour. Sci., xliv, 301. Volume I, p. 145. 
 
DETERMINATION OF TELLUROUS ACID, ETC. 239 
 
 of that necessary to redissolve the first precipitate, the results 
 are theoretically accurate, and in close agreement with those 
 obtained by the former procedure. 
 
 In the presence of free hydrochloric acid the action of the 
 permanganate upon tellurous acid has been shown by Brauner* 
 to be irregular and excessive, and the irregularity cannot be 
 corrected (as in the titration of ferrous salts in presence of 
 hydrochloric acid) by the addition of a manganous salt accord- 
 ing to the well-known procedure of Kesslerf and Zimmer- 
 mann4 So far as appears, however, there should be nothing 
 to prevent the accurate determination of tellurium in tellurous 
 compounds in the presence of chlorides by the permanganate 
 process providing the first oxidation is made in alkaline solu- 
 tion, and the second oxidation carried out with such precau- 
 tions as are necessary to a correct determination of oxalic acid 
 by permanganate hi presence of hydrochloric acid ; for the 
 special danger of over-action on the part of the permanganate 
 cannot exist while the solution is alkaline, and has passed 
 when the tellurite has become a tellurate and before the solu- 
 tion is made acid. As to the proper conditions for the titra- 
 tion of oxalic acid by permanganate we have shown recently 
 that the presence of a manganous salt is necessary and suf- 
 cient to secure regularity of action when a considerable 
 amount of hydrochloric acid is in the solution; when the 
 amount is small so much as would be formed in the decom- 
 position of a gram or two of halogen salt of tellurium the 
 disturbing effect under ordinary conditions of work is prob- 
 ably inappreciable, but even in such a case it is better to work 
 in the presence of a manganous salt for the reason that the 
 titration of the oxalic acid may then be made at the ordinary 
 atmospheric temperature. 
 
 In the following table are gathered the results of experi- 
 ments made with, and without, the addition of the manganous 
 salt. 
 
 * Loc. cit., p. 241. t Ann. Phys. cxviii, 48 ; cxix, 225, 226. 
 
 | Ann. Chera. (Liebig), ccxiii, 302. 
 
 Am. Jour. Sci., vii, p. 461. This volume, p. 222. 
 
240 
 
 DETERMINATION OF TELLUROUS ACID 
 
 TABLE I. 
 
 O = 16, Te = 127.5. 
 
 Volume at beginning, 150 cm 8 . 
 
 Temperature of titration, 60-80 C. 
 
 TeO, 
 
 taken. 
 
 NaCl. 
 
 5f2- 
 
 MnCl 2 .4H 2 0. 
 
 TeOg 
 
 found. 
 
 Error. 
 
 grm. 
 
 0.1000 
 0.1000 
 0.1000 
 0.1000 
 0.0650 
 
 grm. 
 
 0.4 
 0.4 
 0.4 
 1.0 
 1.0 
 
 cm 8 
 5 
 5 
 5 
 5 
 5 
 
 grill. 
 
 grm. 
 
 0.1003 
 0.1000 
 0.1004 
 0.1003 
 0.0653 
 
 grm. 
 
 0.0003+ 
 0.0000 
 0.0004+ 
 0.0003+ 
 0.0003+ 
 
 B. 
 
 Temperature of titration, 20-26 C. 
 
 0.0700 
 0.0700 
 0.0700 
 0.1000 
 
 0.4 
 0.4 
 0.4 
 0.4 
 
 5.7 
 5.7 
 5.7 
 5.7 
 
 1.0 
 1.0 
 0.5 
 
 0.5 
 
 0.0705 
 0.0698 
 0.0701 
 0.1008 
 
 0.0005+ 
 0.0002- 
 0.0001+ 
 0.0008+ 
 
 The tellurium dioxide, made by the careful ignition of the 
 crystallized basic nitrate obtained by oxidizing tellurium with 
 nitric acid, was dissolved in a small amount of sodium hydrox- 
 ide, the halogen salt was added to the amount shown, the per- 
 manganate standardized against ammonium oxalate was run 
 in until its characteristic color appeared, standard ammonium 
 oxalate was added in excess of the quantity required to reduce 
 the excess of permanganate, manganate, and higher oxides, 
 and the solution was heated with enough sulphuric acid [1:1] 
 to neutralize the alkaline hydroxide and have an excess of 
 about 5 cm 3 . In the experiments of Section A the liquid was 
 heated to 60 -80 C. to dissolve the oxides at the final titra- 
 tion begun at that temperature ; in those of Section B, 
 manganous chloride (0.5 to 1 gram) was added, so that the 
 reduction of the higher oxides of manganese and the final 
 titration of the excess of oxalic acid might take place at the 
 ordinary temperature of the room. 
 
 Plainly the presence of the chloride does not interfere 
 materially in the determination of the tellurium by this 
 process whether the titration is made at a high or low 
 temperature. 
 
IN PRESENCE OF HALOID SALTS. 
 
 241 
 
 It appears, also, upon putting the matter to the test, that 
 fairly good determinations of tellurous acid may be made 
 similarly in the presence of a bromide, provided the titration 
 is made at the atmospheric temperature in the presence of a 
 sufficiency (0.5 gram to 1 gram) of a manganous salt and of 
 an excess of sulphuric acid limited to about 5 cm 3 or less of the 
 12.5 per cent mixture. At the higher temperatures bromine 
 is liberated at once from the acid solution by the permanganate. 
 The experimental results are given in Table II. 
 
 TABLE II. 
 
 = 16, Te = 127.5. 
 
 Volume at beginning, 150 cm 8 . 
 
 Temperature of titration, 24-26 C. 
 
 Te0 2 
 taken. 
 
 NaCl 
 
 KBr. 
 
 Sfc 
 
 MnCl 2 . 
 4H 2 O. 
 
 TeO, 
 found. 
 
 Error. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 cm s 
 
 grm. 
 
 gnn. 
 
 grm. 
 
 0.1000 
 
 . . 
 
 0.5 
 
 20 
 
 1.0 
 
 0.1022 
 
 0.0022+ 
 
 0.3000 
 
 . . 
 
 1.5 
 
 25 
 
 1.0 
 
 0.3030 
 
 0.0030+ 
 
 0.0650 
 
 
 0.5 
 
 1 
 
 1.0 
 
 O.OG61 
 
 0.0011+ 
 
 0.0650 
 
 
 0.5 
 
 1 
 
 1.0 
 
 0.0647 
 
 0.0003- 
 
 0.1000 
 
 
 0.5 
 
 1 
 
 1.0 
 
 0.1002 
 
 0.0002+ 
 
 0.3000 
 
 . . 
 
 0.5 
 
 5 
 
 0.5 
 
 0.3010 
 
 0.0010+ 
 
 0.0650 
 
 0.5 
 
 0.5 
 
 1 
 
 1.0 
 
 0.0661 
 
 0.0011+ 
 
 It is obvious, therefore, that tellurous acid may be deter- 
 mined with a fair degree of accuracy by the permanganate 
 method in the presence of chlorides and bromides, provided 
 the first oxidation is made in alkaline solution and the final 
 titration of the residual oxalic acid is made at ordinary 
 temperatures in the presence of a manganous salt and restricted 
 amounts of free sulphuric acid. 
 
 In the presence of an iodide, however, the case is different. 
 Upon acidifying the mixture of iodide and the higher oxygen 
 compounds of manganese, produced in the action of the 
 permanganate upon the solution, iodine is at once set free, and 
 oxalic acid does not suffice to reconvert it. In the presence of 
 an excess of potassium iodide the higher manganic compounds 
 are completely reduced with rapidity and the iodine liberated 
 is the measure of the excess of permanganate over that 
 
 VOL. II. 16 
 
242 DETERMINATION OF TELLUROUS ACID 
 
 required to oxidize the tellurous acid ; the difference between 
 the amount of permanganate thus indicated and that originally 
 introduced should determine the amount of the tellurous acid. 
 It is upon this basis that Norris and Fay * have founded their 
 excellent iodometric determination of tellurous acid. This 
 process consists in treating the alkaline solution of tellurous 
 acid with standard permanganate until the meniscus of the 
 liquid shows a deep pink color, then diluting the solution with 
 ice-water, adding potassium iodide and sulphuric acid, and 
 titrating with sodium thiosulphate. The results are excellent. 
 
 It is plain that any agent capable of converting the iodine 
 to hydriodic acid without at the same time reducing telluric 
 acid should be capable of measuring the excess of the 
 permanganate, and so the amount of tellurous acid originally 
 present. We find that the standard arsenite made, as usual, 
 by dissolving 4.95 grams of pure resublimed arsenious oxide 
 to the liter of water containing potassium bicarbonate answers 
 the purpose admirably, and possesses the further advantage of 
 fixing at once the entire standard of the process, the strength 
 of the permanganate (approximately -^ being determined by 
 running a definite volume of its solution into water containing 
 potassium iodide (1 gram) with 2 to 3 cm 3 of dilute sulphuric 
 acid and titrating by the standard arsenite the iodine (set 
 free by the action of the excess of permanganate and higher 
 oxides) after neutralization with acid potassium bicarbonate. 
 In this titration of iodine by the arsenite we find it best to 
 dispense with the starch solution usually employed to secure 
 the end reaction. The color of the free iodine itself is 
 sufficiently definite, even at a dilution so much as 300 cm 8 , 
 and its disappearance under the action of the arsenite is much 
 sharper than that of the blue starch iodide. 
 
 In Table III are recorded results obtained by adding the 
 alkaline solution of tellurous oxide to 100 cm 8 of water 
 containing 0.5 gram or 1 gram of potassium iodide, introducing 
 the standardized potassium permanganate until the green color 
 of the manganate appears (about 30 cm 3 of the ^ solution for 
 
 * Am. Chem. Jour., xx, 278. 
 
IN PRESENCE OF HALOID SALTS. 
 
 243 
 
 every 0.1 gram of TeO 2 ), adding a few cubic centimeters of 
 dilute sulphuric acid, followed, when the solution has cleared 
 and separated iodine, by an excess of acid potassium carbonate, 
 and titrating to the destruction of color with the standard 
 solution of arsenic. It is essential, in order that oxygen may 
 not go to waste in the breaking down of the oxides, that more 
 than enough iodide should be present when the solution is 
 acidified to complete the reduction of the manganese oxides, 
 or else, that the arsenious acid should be present in suitable 
 amount before the sulphuric acid is put in. This latter 
 procedure may be used in case, for any reason, it is preferred 
 not to introduce more iodide into the solution than may be 
 present originally : when, for example, a direct determination 
 of the iodine present is to follow. 
 
 TABLE III. 
 O - 16, Te = 127.5. 
 
 Te0 2 
 taken. 
 
 NaCL 
 
 KBr. 
 
 KI. 
 
 Total 
 volume 
 
 NaOH 
 present 
 during 
 
 TeO, 
 found. 
 
 Error. 
 
 
 
 
 
 
 oxidation. 
 
 
 
 griii* 
 
 grin* 
 
 grm. 
 
 grm. 
 
 cm 3 . 
 
 grm. 
 
 grm. 
 
 grm. 
 
 0.1000 
 
 
 
 0.5 
 
 160 
 
 0.1 
 
 0.1005 
 
 0.0005+ 
 
 0.1000 
 
 
 
 0.5 
 
 160 
 
 0.1 
 
 0.1001 
 
 0.0001+ 
 
 0.1000 
 
 . 
 
 
 0.5 
 
 160 
 
 0.1 
 
 0.1003 
 
 0.0003+ 
 
 0.1000 
 
 
 
 1.0 
 
 250 
 
 0.1 
 
 0.1007 
 
 0.0007+ 
 
 0.2000 
 
 
 
 1.0 
 
 250 
 
 0.2 
 
 0.1997 
 
 0.0003+ 
 
 0.1000 
 
 0.6 
 
 0.5 
 
 0.5 
 
 250 
 
 0.1 
 
 0.1000 
 
 0.0000 
 
 0.2100 
 
 1.0 
 
 1.0 
 
 1.0 
 
 225 
 
 0.2 
 
 0.2105 
 
 0.0005+ 
 
 0.1000 
 
 
 
 
 0.5 
 
 160 
 
 1.0 
 
 0.1011 
 
 0.0011+ 
 
 0.2000 
 
 
 
 1.0 
 
 300 
 
 2.0 
 
 0.2009 
 
 0.0009+ 
 
 These results are reasonably good. Like those of Table I 
 they would be brought practically in the average to the figure 
 demanded by theory if the value of the Committee of the 
 German Chemical Society, Te = 127, were to be taken instead 
 of Te = 127.5, the value of Clarke and of Richards. 
 
XXX 
 
 AN IODOMETRIC METHOD FOR THE ESTIMA- 
 TION OF BORIC ACID. 
 
 BY LOUIS CLEVELAND JONES.* 
 
 IN a recent article,! I have described a process for the alka- 
 limetric estimation of boric acid, depending upon the forma- 
 tion of a strongly acidic compound when boric acid and a 
 polyatomic alcohol are placed together in solution. The 
 method in brief consists in destroying the free mineral acid 
 in a solution containing boric acid, by means of a mixture of 
 potassium iodide and iodate, bleaching the liberated iodine by 
 sodium thiosulphate, adding the indicator phenolphthalein 
 and sufficient standard solution of caustic soda to give a 
 faint alkaline coloration, bleaching by a small amount of 
 mannite and adding caustic soda again to alkalinity, and 
 thus alternating with mannite and alkali until the alkaline 
 coloration produced is permanent. The amount of sodium 
 hydroxide used represents the amount of acidity developed 
 by the influence of the mannite upon the boric acid present, 
 according to the hypothesis that the molecule B 2 O 8 acts as 
 two molecules of a univalent acid, HOBO. 
 
 On making further study of this reaction, I have found 
 that the acid developed by the combination of boric acid and 
 mannite is, under certain definite conditions, sufficiently 
 strong to liberate, quantitatively, from a mixture of potassium 
 iodide and iodate, the amount of iodine required on the 
 supposition that each molecule of metaboric acid (HOBO) 
 acts in a manner similar to a univalent mineral acid under 
 the same conditions. (5KI + KIO 8 + 6HOBO = 3I 2 + 
 
 * From Am. Jour. Sci., viii, 127. 
 
 t Am. Jour. Sci., vii, 147. This volume, p. 182. 
 
THE ESTIMATION OF BORIC ACID. 245 
 
 6KOBO + 3H 2 O.) Obviously, this reaction depends upon 
 the behavior of the acidic boromannite compound as a strong 
 acid, stronger than acetic, tartaric, or citric acid; for these 
 acids have been found by Furry * to be incapable of liberat- 
 ing iodine regularly from a mixture of iodide and iodate. 
 Conditions which tend to increase the acidic activity of this 
 compound are concentrated solutions and moderately low 
 temperatures.f 
 
 Glycerine acts in general like mannite to produce acidic 
 compounds with boric acid; and hi a preliminary way, the 
 relative acidity of the products formed by these two poly- 
 atomic alcohols with boric acid may be indicated by the results 
 of two experiments in which the iodine liberated from a 
 mixture of potassium iodide and iodate, proportionately to 
 the time required for the liberation, is taken as a measure of 
 the strengths of the acids developed. 
 
 Equal amounts (10 cm 3 ) of a standard solution of boric 
 acid, prepared from the anhydride, J were drawn into separate 
 Erlenmeyer flasks and a neutral solution of iodide and iodate 
 added to each in an amount sufficient to liberate iodine in 
 quantities corresponding to the acid used. One solution 
 was treated with glycerine enough to constitute one-half the 
 entire volume of the liquid: mannite (about 5 grms.) was 
 added to the other. The thiosulphate required immediately 
 and after definite periods of tune, is shown for each solution 
 in the following table. 
 
 The solution of boric acid contained 7.706 grm. per liter. 
 The thiosulphate was 0.0999 normal. According to theory, 
 the amount of thiosulphate required for 10 cm 3 of the boric 
 
 * Am. Chem. Jour., vi, 341. 
 
 t Magnanini, Gaz. chim. Hal. xx, 428, xxi, 134 ; and Lambert, Compt. rend., 
 CTiii, 1016, 1017. 
 
 J The recrystallized hydrous boric acid should be fused in a platinum dish 
 and, after cooling and breaking into small pieces, the desired amount placed 
 in a small weighed platinum crucible and again fused until no more water 
 escapes. After cooling and weighing, the boric oxide may be separated from 
 the crucible, or with it placed in warm water, dissolved and made up to a 
 definite volume. 
 
246 
 
 AN IODOMETRIC METHOD FOR 
 
 TABLE I. 
 
 Bj0 8 solution (10 cm) 
 with maimite. 
 
 Time. 
 
 B 2 S solution (10 cm 8 ) 
 with glycerine. 
 
 Thiosulphate required. 
 
 
 Thiosulphate required. 
 
 cm* 
 
 
 cm' 
 
 18.60 
 21.30 
 
 Immediately. 
 After 15 minutes. 
 
 8.48 
 10.50 
 
 22.00 
 
 After 30 minutes. 
 
 11.15 
 
 22.05 
 
 After 2 hours. 
 
 11.60 
 
 acid solution is 22.02 cm 3 . From these data we may observe 
 that at the end of 30 minutes, in the solution containing 
 mannite, practically the theoretical amount of thiosulphate 
 had been used, while only about 50 per cent of that amount 
 had been required to bleach the iodine liberated by the 
 glycerine compound. Obviously, mannite forms with boric 
 acid a more acidic compound than glycerine does, and, from 
 the indication given in the above experiments, may be relied 
 upon, under certain conditions, to liberate the theoretical 
 amount of iodine. If, from the iodide and iodate used to 
 destroy the excess of mineral acid already present, the boric 
 acid, upon the addition of mannite does liberate iodine regu- 
 larly as the previous experiments seem to indicate this 
 liberated iodine should form a most convenient measure of 
 the boric acid present. 
 
 On studying the conditions requisite for the complete 
 liberation of iodine according to theory, several important 
 points have come to light. 
 
 It has not been found possible under any conditions to rely 
 upon the immediate liberation of the full amount of iodine: 
 a certain period of time is required for the completion of 
 the reaction. When the solution is of small volume and 
 saturated with mannite, the reaction goes to the end most 
 quickly sometimes almost immediately but there is this 
 limitation, which must be made emphatic, viz.: that if the 
 solution of boric acid is too concentrated near saturation 
 the boric acid alone, when the iodate and iodide are added to 
 destroy any other free acid present, throws out some iodine 
 
THE ESTIMATION OF BORIC ACID. 247 
 
 and on bleaching with thiosulphate a starting-point is ob- 
 tained at which some of the boric acid has already entered 
 into combination. The amount of iodine thus liberated by 
 the boric acid is, however, not large, and if upon the addition 
 of the iodide and iodate, the iodine thrown out by the free 
 hydrochloric acid present is immediately bleached by thio- 
 sulphate and the analysis proceeded with from this as the 
 neutral point, even in concentrated solutions the error is 
 almost inappreciable. If, however, considerable time inter- 
 venes between the adding of the iodide and iodate and the 
 determination of the neutral point by thiosulphate, as much 
 as several milligrams of boric acid may have liberated its 
 amount of iodine and is, therefore, not capable of being 
 registered by thiosulphate after the addition of mannite. 
 This difficulty was not met with in those experiments in 
 which the iodide and iodate were added at a dilution little 
 greater than that of the standard solution used (7.738 grm. 
 per liter), but in an attempt to estimate the boric acid in 
 colemanite, where the solution was kept as concentrated as 
 possible, hoping in this way to decrease the time required for 
 the complete liberation of iodine, low values were obtained; 
 that is, a false starting point was used. 
 
 The dilution found most convenient at the time of adding 
 the iodide and iodate is not less than 25 cm 3 for each decigram 
 of boric acid (B 2 O 3 ) present and should not be much greater 
 than two or three times that amount. This limitation as 
 regards volume is equally applicable, whether after obtaining 
 the neutral point and treating with mannite, the boric acid is 
 to be measured by a standard solution of alkali as before 
 described or as here by the iodine liberated. As has been 
 suggested, a large volume, even though saturated with mannite, 
 prolongs the time of standing necessary and increases the 
 effect of carbon dioxide upon the iodide and iodate present, for 
 carbon dioxide, whether derived from the atmosphere or 
 existing dissolved in the solution, upon standing, slowly 
 liberates iodine. The amount, however, is small, and, in the 
 time required for the completion of the process, has never been 
 
248 AN IODOMETRIC METHOD FOR 
 
 found equivalent to more than a single drop of the solution of 
 thiosulphate used. Even if the material to be analyzed 
 contains carbonates, after acidifying in concentrated solution 
 and shaking vigorously, the small amount of uncombined 
 carbon dioxide remaining has almost an inappreciable effect 
 upon the results. The length of tune required for the 
 liberation of the theoretical amount of iodine hi a solution of 
 the volume suggested above, is 20 to 45 minutes, and at the end 
 of 45 minutes standing in a solution saturated with mannite the 
 reaction may be considered complete. During this period, how- 
 ever, it is well to keep the solution cool at zero will do no harm 
 and shake occasionally to insure thorough mixture. The free 
 iodine would tend to escape upon standing unless kept in a 
 closed flask, but it is more convenient, immediately after the 
 addition of mannite, to treat with an excess of the standard 
 solution of thiosulphate 8 or 10 cm 3 more than the amount 
 required to bleach the iodine liberated, and at the expiration 
 of 40 to 60 minutes titrate back with ^ iodine. The strength 
 of the thiosulphate solution found most convenient is -", while 
 the use of iodine of one-half this strength (^) enables the 
 error of reading to be correspondingly diminished. In solutions 
 of the volume recommended the addition of starch to give the 
 indication with iodine is unnecessary and even detrimental, 
 since a single drop of one-twentieth normal iodine in excess is 
 sufficient to give a strong lemon coloration, while in the 
 presence of starch an indefinite dirty red first appears and 
 remains until the blue is brought out by the further addition 
 of iodine. 
 
 With these observations in mind, a series of experiments was 
 made in which the standard solution of boric acid was drawn 
 into an Erlenmeyer flask, containing a small amount of free 
 hydrochloric acid and made up to a definite volume. To 
 bring the conditions to those of an actual analysis about 1 grm. 
 of crystalline calcium chloride in solution was also added. 
 Potassium iodate (5-10 cm 3 of a 5 per cent solution) and 
 iodide (3-5 cm 3 of a 40 per cent solution) were added, and the 
 iodine liberated by the hydrochloric acid, barely bleached and 
 
THE ESTIMATION OF BORIC ACID. 
 
 249 
 
 again brought to coloration by iodine. Mannite was added to 
 saturate the solution, an excess of standard thiosulphate put 
 in, and the solution set aside for various periods of time, at 
 the end of which the excess of thiosulphate was titrated by 
 iodine and the amount of unrecovered thiosulphate taken as a 
 measure of the boric acid present. 
 
 The thiosulphate used was 0.198 normal and the iodine 
 0.0996 normal. The solution of boric acid contained 7.733 
 
 grm. per liter. 
 
 TABLE II. 
 
 B 3 8 
 
 taken. 
 
 Thio. 
 taken. 
 
 Iodine 
 taken. 
 
 Time of 
 standing. 
 
 Volume. 
 
 S&. 
 
 B 2 3 
 
 found. 
 
 Error. 
 
 A 
 
 cm 3 
 28.00 
 27.03 
 27.02 
 
 cm 3 
 32.00 
 32.00 
 31.97 
 
 cm 8 
 1.88 
 4.37 
 4.04 
 
 hrs. 
 
 0.30 
 0.27 
 1.00 
 
 cm 3 
 28 
 27 
 
 27 
 
 grm. 
 0.2165 
 0.2090 
 0.2089 
 
 gnu. 
 
 0.2168 
 0.2081 
 0.2090 
 
 gnu. 
 
 0.0003+ 
 0.0009- 
 0.0001+ 
 
 B 
 
 
 27.06 
 27.02 
 27.04 
 
 32.04 
 32.02 
 31.72 
 
 3.88 
 4.40 
 3.39 
 
 1.00 
 1.00 
 1.00 
 
 50-60 
 50-60 
 50-60 
 
 0.2093 
 0.2089 
 0.2091 
 
 0.2101 
 0.2081 
 0.2096 
 
 0.0008+ 
 0.0058- 
 0.0005+ 
 
 C 
 
 27.01 
 26.05 
 
 31.53 
 31.01 
 
 2.88 
 4.01 
 
 2.00 
 3.00 
 
 50-60 
 50-60 
 
 0.2089 
 0.2014 
 
 0.2100 
 0.2025 
 
 0.0011+ 
 0.0011+ 
 
 D 
 
 27.00 
 27.00 
 26.01 
 27.03 
 27.05 
 26.07 
 27.00 
 
 31.00 
 32.00 
 32.02 
 31.01 
 31.89 
 31.02 
 32.04 
 
 2.12 
 4.05 
 6.20 
 2.21 
 3.81 
 4.14 
 4.30 
 
 0.30 
 0.30 
 0.30 
 0.48 
 0.45 
 0.40 
 0.40 
 
 50-60 
 50-60 
 50-60 
 50-60 
 50-60 
 50-60 
 60 
 
 0.2088 
 0.2088 
 0.2011 
 0.2090 
 0.2092 
 0.2016 
 0.2088 
 
 0.2089 
 0.2092 
 0.2018 
 0.2087 
 0.2093 
 0.2020 
 0.2086 
 
 0.0001+ 
 0.0004+ 
 0.0007+ 
 0.0003- 
 0.0001+ 
 0.0004+ 
 0.0002- 
 
 These results are so regular that the method seems worthy of 
 high commendation, and especially since the standard solutions, 
 thiosulphate and iodine, upon which the process depends, are 
 so easily prepared and generally at hand. 
 
 The full method of procedure recommended is as follows : 
 The borate is dissolved in as small volume and as little 
 
250 
 
 AN IODOMETRIC METHOD FOR 
 
 hydrochloric acid as possible, shaking well to remove free 
 carbon dioxide and diluting so that, at the time of adding 
 potassium iodide and iodate, there shall be approximately 25- 
 50 cm 3 of solution for each decigram of boric anhydride 
 present. The greater part of the excess of hydrochloric acid 
 in the solution is destroyed by sodium hydroxide and the use 
 of litmus paper, leaving the solution distinctly acid in reaction. 
 Potassium iodide (3-5 cm 3 of a 40 per cent solution), and 
 iodate (5-10 cm 3 of a 5 per cent solution) are added in excess 
 of that required to liberate iodine in an amount corresponding 
 to the hydrochloric acid and the boric acid present. The 
 iodine liberated by the free hydrochloric acid is bleached by a 
 small amount of a strong solution of thiosulphate, and after 
 agitating to insure thorough mixture, iodine is added to faint 
 coloration. Sufficient mannite is now used to saturate the 
 solution about 10-15 grm. for a volume of 50 cm 3 and 
 sodium thiosulphate added in standard solution 8-10 cm 8 in 
 excess of that required to bleach the iodine immediately thrown 
 out by the mannite. The solution is again brought to 
 saturation, if necessary, by mannite and after standing in a 
 cool place for 40-60 minutes, titrated with decinormal iodine 
 to determine the excess of thiosulphate present. In the 
 manner described, specimens of crude calcium borate and 
 crystals of colemanite were analyzed with the results given 
 below. 
 
 TABLE III. 
 CALCIUM BORATE. 
 
 Mineral. 
 
 Thio. 
 taken. 
 
 Iodine 
 taken. 
 
 Time 
 stand- 
 ing. 
 
 Volume 
 of solu- 
 tions. 
 
 B 2 0. 
 
 found. 
 
 Per cent. 
 
 grm. 
 0.4015 
 0.4010 
 
 cms 
 35.05 
 35.34 
 
 cm 8 
 4.75 
 5.23 
 
 hrs. 
 1.00 
 2.00 
 
 cm 8 
 
 40 
 45 
 
 grm. 
 0.2280 
 0.2283 
 
 56.92 
 56.94 
 
 COLEMANITE. 
 
 0.4002 
 0.2513 
 0.4007 
 
 32.00 
 32.01 
 33.03 
 
 5.50 
 7.36 
 
 7.72 
 
 1.30 
 1.00 
 0.50 
 
 50 
 40 
 65 
 
 0.2043 
 0.1279 
 0.2036 
 
 51.04 
 50.91 
 50.81 
 
THE ESTIMATION OF BORIC ACID. 251 
 
 The solution of thiosulphate used was 0.19939 and the 
 iodine 0.0996 normal. 
 
 These results show little variation and in the case of 
 colemanite correspond closely to the theory 50.97 per cent. 
 The process is convenient, generally applicable, and accurate 
 within the ordinary limits of analysis. 
 
XXXI 
 
 THE DOUBLE AMMONIUM PHOSPHATES OP 
 
 BERYLLIUM, ZINC, AND CADMIUM IN 
 
 ANALYSIS. 
 
 BY MARTHA AUSTIN.* 
 
 IT has been shown f that the composition of the phosphate 
 of manganese thrown down by microcosmic salt from the 
 solution of a pure manganous salt contains more manganese 
 than belongs to the ideal ammonium manganese phosphate 
 NH 4 MnPO 4 ; and, further, that by acting with ammonium 
 chloride in proper proportion the phosphate of manganese 
 thrown down by microcosmic salt may be completely converted 
 to the ideal ammonium manganese phosphate. Ammonium 
 chloride, likewise, in the case of magnesium phosphate f tends 
 to cause the replacement of the metal by ammonia. Indeed, 
 the replacement here is readily carried so far beyond the point 
 corresponding to the normal ammonium magnesium phosphate, 
 NH 4 MgPO 4 , that the tendency to form a salt richer in 
 ammonia and poorer in magnesium perhaps something like 
 Mg(NH 4 ) 4 (PO 4 ) 2 must be recognized. 
 
 These facts suggested an investigation into the constitution 
 of certain other ammonium phosphates with reference to their 
 utility in analytical processes. Of the elements of Mendele'eff's 
 second group, beryllium, magnesium, zinc, cadmium, and mer- 
 cury are capable of yielding double ammonium phosphates, 
 while no such compounds of calcium, strontium and barium 
 have been described. The solubility in ammonia of the double 
 ammonium phosphates of the elements of the former category 
 
 * From Am. Jour. Sci., viii, 206. 
 
 t Am. Jour. Sci., vol. yi, 233. This volume, p. 121. 
 
 } Am. Jour. Sci., vol. vii, 187. This volume, p. 190. 
 
DOUBLE AMMONIUM PHOSPHATES IN ANALYSIS. 253 
 
 appears to increase as the elements of which they are 
 compounds are removed in the series from the beryllium, 
 and, while the same is true of the simple phosphates of 
 members of the latter category, the extent of such solvent 
 action is slight comparatively. According to the work recorded 
 in the literature, calcium, barium, and strontium form individu- 
 ally a neutral tribasic phosphate or acid phosphates of greater 
 or less degree of acidity according to the conditions of 
 precipitation. In my experience where salts of these elements 
 were precipitated either with ammonium phosphate or 
 microcosmic salt hi presence of varying amounts of ammonium 
 chloride, or ammonia, or both, only the recognized phosphates 
 were obtained. The effect of ammonium salts in presence of 
 ammonia seemed to promote the formation of the tribasic salt 
 in the case of calcium and strontium; barium tends to form 
 the barium acid phosphate almost exclusively even in the 
 presence of ammonium salts and free ammonia. No double 
 ammonium phosphate of either calcium, strontium, or barium 
 was produced under any condition. As is well known, 
 mercury does form an ammonium mercuiy phosphate, but the 
 salt is soluble to so great a degree in ammonia, ammonium 
 chloride, and even in the precipitant itself, that nothing of 
 any value for analytical work seemed likely to come from its 
 study. 
 
 The Ammonium Beryllium Phosphate. 
 
 The ammonium beryllium phosphate has beeen described 
 by Roessler * as a crystalline salt produced by boiling some 
 time in ammoniacal solution the phosphate precipitated by 
 ammonium phosphate, though the best results of this treatment 
 failed to yield the ideal constitution of this salt, NH 4 BePO 4 . 
 This same precipitate cannot be obtained, Roessler further 
 states, by using a sodium salt as the precipitant. In order to 
 follow out this work of Roessler, a solution of berryllium 
 chloride for use was prepared as follows : The pure beryllium 
 chloride of commerce was dissolved in as little water as 
 * Fresenius, Zeitschr. anal. Chem., xvii, 148. 
 
254 DOUBLE AMMONIUM PHOSPHATES OF BERYLLIUM, 
 
 possible and treated for the precipitation of aluminum by 
 ethereal hydrochloric acid.* After filtering and evaporating 
 from the filtrate the ether and a part of the hydrochloric acid, 
 the beryllium was precipitated with ammonia, filtered to 
 remove any members of the magnesium group, and washed 
 free from ammonium chloride. The larger part of the 
 precipitate was dissolved in hydrochloric acid in slight excess, 
 and boiled with the reserved portion. After filtering, the 
 solution was diluted to definite volume and standardized by 
 precipitating measured portions of the solution with ammonia, 
 filtering on asbestos under pressure in a perforated platinum 
 crucible, igniting the residue and weighing as the oxide. The 
 results recorded in section A of the following table were 
 obtained by precipitating definite volumes of the pure solution 
 of beryllium chloride with ammonium phosphate in a platinum 
 dish, dissolving the precipitate in hydrochloric acid in faint 
 excess, and while hot precipitating slowly with dilute ammonia, 
 boiling (while the solution was kept distinctly ammoniacal) 
 until the flocky precipitate was entirely converted to a fine, 
 powdery, semi-crystalline, rapidly subsiding mass. A quarter 
 to a half-hour is necessary under the most favorable conditions 
 to cause this conversion. After cooling, the precipitate was 
 filtered off on asbestos under pressure in a perforated platinum 
 crucible, washed carefully with distilled water, dried, ignited 
 and weighed. The filtrate was tested for beryllium by boiling 
 with ammonia. None was found in these cases, nor in any of 
 the following work. Faint traces of chloride were found in 
 the residues after ignition after dissolving in nitric acid and 
 testing with silver nitrate. 
 
 The results are in every case in excess of the theory for the 
 pyrophosphate derived by ignition of the ammonium beryl- 
 lium phosphate, possibly because the ammonium chloride 
 present may have a tendency to form a salt too rich in ammo- 
 nium (as was shown to be the case with the magnesium salt), 
 consequently giving too much phosphoric acid in the ignited 
 residue ; or, because of inclusion of the chloride and phosphoric 
 
 * Am. Jour. Sci., ir, 111. This volume, p. 111. 
 
ZINC, AND CADMIUM IN ANALYSIS. 255 
 
 acid. It might reasonably be expected that some phosphoric 
 acid may be held, since a trace of chloride was found. Either 
 or both of these substances may have been held mechanically, 
 or in combination. 
 
 It was found that on boiling for some tune the solution of 
 beryllium chloride with microcosmic salt (6) section B of 
 the table and precipitating in the same manner as when 
 ammonium phosphate was used the same sort of powdery 
 mass remained as was obtained by the ammonium phosphate. 
 The residue being tested for sodium according to the method 
 brought out by Kreider and Brecken^idge,* showed sodium 
 present to the amount of 0.0062 grm. reckoned as sodium 
 phosphate. It may be reasonably supposed that the presence 
 of the sodium was due to one of two causes, inclusion of 
 the soluble phosphate, or a tendency on the part of the beryl- 
 lium to form an ammoniumf sodium beryllium phosphate or a 
 sodium J beryllium phosphate, both of which are known to 
 exist. Long boiling of the precipitates is tedious, and, unless 
 great care is taken, may involve small losses of material ; hence 
 if the same results could be obtained with less boiling such 
 treatment would be decidedly advantageous. The results in 
 section C of the table were obtained by adding microcosmic 
 salt to the hot solutions of the chloride, boiling five minutes, 
 cooling, filtering off on an ashless filter because of the flocky 
 condition of the precipitate treating as usual before igniting 
 the residue in a platinum crucible. The results compare well 
 with those obtained by long boiling of the precipitated beryl- 
 lium although all are in excess of the theory. That ammo- 
 nium chloride here, as in cases above, has a marked effect in 
 changing the constitution of the phosphate precipitated by 
 microcosmic salt is not readily seen. It is obvious that the 
 presence of an excess of the soluble phosphate is essential to 
 precipitate the beryllium as the double ammonium phosphate 
 from the results recorded in section D of the table, where, 
 
 * Am. Jour. Sci., ii, 263. Volume I, p. 401. 
 
 t Persoz, Ann. Chem. (Liebig), Ixv, 174 ; Atterberg, Bulletin Soc. Chim., 
 xxiv, 358. 
 
 J Scheffer, Ann. Chem. (Liebig), cix, 144. 
 
256 DOUBLE AMMONIUM PHOSPHATES OF BERYLLIUM, 
 
 after the precipitate of beryllium phosphate had subsided and 
 the supernatant liquid had been poured off, the precipitate dis- 
 solved in hydrochloric acid was brought down again at the 
 boiling temperature with ammonia either alone or in presence 
 of ammonium chloride. The results obtained show that the 
 salt approaches the constitution of the tribasic phosphate, 
 when it is precipitated in presence of a faint excess of phos- 
 phoric acid, even though ammonium chloride in large amount 
 be present. 
 
 TABLE I. 
 
 Exp. 
 
 Be 2 P 2 O 7 corresponding 
 to BeCl,. 
 
 Be 3 P,O 8 corresponding 
 to BeCl,. 
 
 (NH 4 ) 3 P0 4 . 
 
 NB^Cl. 
 
 Taken. 
 
 Found. 
 
 Error. 
 
 Taken. 
 
 Found. 
 
 Error. 
 
 A. 
 
 (1) 
 (2) 
 (3) 
 (4) 
 (5) 
 
 grm. 
 0.3578 
 0.3578 
 0.3578 
 0.3578 
 0.3578 
 
 grin. 
 
 0.3613 
 0.3808 
 0.3707 
 0.3640 
 
 0.3680 
 
 grin. 
 
 0.0035+ 
 0.0230+ 
 0.0129+ 
 0.0062+ 
 0.0102+ 
 
 grm 
 
 grm. 
 
 grm. 
 
 grm. 
 2 
 2 
 2 
 2 
 2 
 
 grm 
 
 ' 'so 
 
 B. 
 
 HNaNH 4 PO 4 
 .4H 2 6. 
 
 
 (6) 
 
 0.3578 
 
 0.3697 | 0.0119+ | ... 
 
 ... | ... 
 
 C. 
 
 (7) 
 (8) 
 (9) 
 (10) 
 
 0.3578 
 0.3578 
 0.3578 
 0.3578 
 
 0.3618 
 0.3680 
 0.3729 
 0.3631 
 
 0.0040+ 
 0.0102+ 
 0.0151+ 
 0.0053+ 
 
 
 
 
 1.2 
 1.2 
 1.2 
 1.2 
 
 ' *10 
 60 
 
 D. 
 
 (11) 
 (12) 
 (13) 
 (14) 
 
 
 
 
 0.2700 
 0.2700 
 0.2700 
 0.2700 
 
 0.2589 
 0.2989 
 0.2936 
 0.2507 
 
 0.0111- 
 0.0289+ 
 0.0236+ 
 0.0193- 
 
 0.5 
 0.5 
 0.5 
 0.5 
 
 10 
 5-60 
 60 
 
 From the work described it is clear that the ammonium 
 beryllium phosphate is not obtained in ideal condition by pre- 
 cipitating a solution of the chloride with ammonium phosphate. 
 Roessler's own results were likewise only approximately cor- 
 rect, as he states. It is also plain that hydrogen sodium 
 ammonium phosphate precipitates the ammonium beryllium 
 
ZINC, AND CADMIUM IN ANALYSIS. 257 
 
 phosphate in a condition as nearly ideal as does the ammonium 
 phosphate, while the effect of the ammonium chloride in either 
 case is not marked in producing a phosphate containing 
 ammonia. Of most importance in obtaining the ammonium 
 salt is an excess of the soluble phosphate, for when the amount 
 of the precipitant is reduced to a little more than the theo- 
 retical amount the condition of the phosphate coincides almost 
 exactly with the theory for the tribasic phosphate, even though 
 a large excess of ammonium chloride be present. When there 
 is an abundance of the precipitant the results are all in excess 
 of the theory, which may be accounted for on the supposition 
 that foreign material is included the chloride of ammonia and 
 the soluble phosphate to a greater or less extent by the pre- 
 cipitate. The formation of a phosphate of beryllium contain- 
 ing too much ammonia and phosphoric acid, or, in case of the 
 precipitations by microcosmic salt, sodium by the formation 
 of a sodium ammonium beryllium phosphate and sodium beryl- 
 lium phosphate (known salts), is not definitely proved. 
 
 The Ammonium Zinc Phosphate. 
 
 Debray,* Bette f and Heintz J separately found that am- 
 monium zinc phosphate is formed by boiling a solution of 
 zinc sulphate with ammonium phosphate. This salt was 
 investigated later by A. Guyard (Hugo Tamm), who found 
 that if to a solution of a zinc salt of an organic or a mineral 
 acid supersaturated with ammonia until all the zinc oxide 
 is dissolved and made faintly acid with hydrochloric acid, 
 sodium phosphate be added, a flocky precipitate resulted, 
 which on being kept near the boiling point for some seconds 
 was converted to crystalline zinc ammonium phosphate, which 
 filtered readily and was washed free from impurities with the 
 greatest facility. He found that all the zinc in solution was 
 thrown down as the ammonium zinc phosphate, which on 
 ignition yielded the zinc pyrophosphate. With care in 
 handling this process to avoid an excess of the precipitant, 
 
 * Compt. rend., lix, 40. t Ann. Chem. (Liebig), xv, 129. 
 
 J Ann. Chem. (Liebig), cxliii, 156. Chem. News, xxir, 148. 
 VOL. ix. 17 
 
258 DOUBLE AMMONIUM PHOSPHATES OF BERYLLIUM, 
 
 and the presence of sodium and potassium salts (on account 
 of the danger of occlusion) the precipitation of the ammo- 
 nium zinc phosphate, ignition, and weighing as the pyro- 
 phosphate made, Guyard believed, an ideal process for the 
 estimation of zinc. Although there was slight solubility of 
 the salt, it made an insignificant loss when the process was 
 handled properly. Acids present, or certain alkalies to any 
 great extent, increased the solubility of the salt so much 
 that the loss became appreciable. Another source of error 
 was to Guyard's mind loss of zinc during the ignition 
 of the zinc ammonium phosphate with the paper on which 
 the precipitate had been collected. Garrigues* found, in 
 estimating zinc in a practical way, that this process advocated 
 by Guyard gives in solutions of zinc free from salts of all 
 metals, even alkaline salts solutions that from previous 
 steps in analysis, however, must have contained ammonium 
 chloride in large amount as satisfactory results as Guyard 
 claimed for it. Garrigues' method of procedure was to add 
 acid diammonium phosphate to a warm solution of zinc 
 exactly neutralized with either hydrochloric acid or ammonia, 
 so that the weights of zinc ammonium phosphate and that 
 of the diammonium phosphate added should be as one to 
 five respectively, to heat until the flocky precipitate becomes 
 crystalline and subsides, filtering off on asbestos, drying at 
 100 C. and weighing preferably, although the residue may 
 be ignited without loss, since the filtration is made on asbestos 
 in a perforated crucible. Langmuir f modifies the method by 
 destroying with dilute acetic acid any free ammonia that may 
 be left in the solution after boiling. 
 
 In the work that follows, in which an attempt was made to 
 show what precipitate is formed from a solution of zinc by the 
 action of a soluble phosphate, also what effect ammonium 
 chloride has upon the precipitate, a solution of zinc chloride 
 prepared as detailed below was employed. The pure zinc 
 chloride of commerce was treated with zinc carbonate, filtered 
 and precipitated with ammonium sulphide. This precipitate 
 
 * Jour. Am. Chem. Soc., xix, 936. t Jour. Am. Chem. Soc., xxi, 115. 
 
ZINC, AND CADMIUM IN ANALYSIS. 
 
 259 
 
 was boiled in a slight excess of hydrochloric acid until all the 
 hydrogen sulphide was removed, and then was precipitated 
 with sodium carbonate. After washing carefully until all 
 the chloride was removed, the greater part of the carbonate 
 was dissolved in sulphuric acid in slight excess, boiled with 
 the remaining portion of the carbonate and filtered. This 
 solution diluted to definite volume was standardized as sul- 
 phate by evaporating the solution to dryness in a platinum 
 crucible and heating the residue.* The heating is carried 
 on safely by so placing the platinum crucible in a radiator 
 (consisting of a crucible and a triangle) that the bottom of 
 the platinum crucible was held about one centimeter above 
 the bottom of the outside crucible. Constant weights were 
 obtained in successive treatment with a few drops of sulphuric 
 acid and heating over the radiator. The results obtained in 
 this manner were a trifle higher, though in fair agreement 
 (when the nature of the carbonate process is taken into 
 consideration) with determinations of the zinc in the solu- 
 tions as oxide after precipitating with sodium carbonate with 
 the usual precautions, filtering off on asbestos under pressure 
 in a perforated platinum crucible, washing with distilled 
 water, drying, and igniting. Results are given in Table II 
 showing the amount of zinc sulphate found in five different 
 portions each of forty cubic centimeters of the solution 
 of zinc sulphate, and, for comparison, the results of de- 
 terminations as zinc oxide by the carbonate processes are 
 
 included. 
 
 TABLE II. 
 
 ZnSO 4 found in 
 40 cm of 
 solution. 
 
 Mean value of ZnO 
 corresponding to 
 ZnSO 4 in 40 cms 
 of solution. 
 
 ZnO found in 40 cm 8 
 of solution by 
 precipitation as 
 the carbonate. 
 
 grm. 
 
 grim* 
 
 gnu. 
 
 0.5386 ' 
 
 
 0.2712 
 
 0.2691 
 
 0.5385 
 
 
 
 0.2685 
 
 0.5387 
 
 
 
 0.2711 
 
 0.5387 
 
 
 t 
 
 
 0.5390 j 
 
 
 
 
 Rose-Finkener, Analytische Chemie, 6te Auflage, vol. ii, 117. 
 
260 DOUBLE AMMONIUM PHOSPHATES OF BERYLLIUM, 
 
 Definite portions of the solution of zinc sulphate were 
 carefully drawn from a burette into a platinum dish, heated and 
 treated with ammonium phosphate until the solution turned 
 red litmus paper blue. The whole was heated until the flocky 
 precipitate became crystalline and fell to the bottom of the 
 dish. The solution after standing as recorded in section A of 
 the table was filtered off on asbestos under pressure in a 
 perforated platinum crucible, and the precipitate was washed 
 with distilled water, dried, ignited and weighed. The filtrate 
 in each case, as in all following cases, was tested for zinc with 
 sulphuretted hydrogen. The results recorded in section B of 
 the table were obtained in the same manner as those of section 
 A, with microcosmic salt substituted for the ammonium salt as 
 the precipitant. The results are below the theory for the 
 pyrophosphate, but no appreciable amount of zinc appeared in 
 the filtrates. Neither ammonium phosphate nor ammonium 
 sodium phosphate seems to precipitate the ideal ammonium 
 zinc phosphate under these conditions; and the time of 
 standing appears to be without effect. 
 
 The results recorded in section C were obtained by precipi- 
 tating the warm solution of the zinc in presence of large 
 amounts of ammonium chloride by adding microcosmic salt 
 until the solution was alkaline to litmus. From these results 
 it seems that the presence of ammonium chloride is essential 
 for the conversion of the zinc phosphate precipitated by 
 hydrogen sodium ammonium phosphate to the ammonium zinc 
 salt. As a matter of fact the solutions employed by Guyard 
 and those in which estimations are made by practical workers 
 do contain ammonium chloride formed in previous steps of the 
 analysis. The proportion of zinc to phosphate suggested by 
 Garrigues 1 : 5 is the amount of soluble phosphate neces- 
 sary to turn red litmus blue after the zinc is precipitated. In 
 order to find out whether the presence of so large an amount 
 of the soluble phosphate is necessary in presence of ammonium 
 chloride, the solution of zinc sulphate was precipitated in 
 presence of the necessary amount of ammonium chloride by 
 the microcosmic salt in small excess above the equivalent of 
 
ZINC, AND CADMIUM IN ANALYSIS. 
 
 261 
 
 the zinc salt, and the solution was made just ammoniacal 
 to litmus with a few drops of dilute ammonia both before 
 and after heating to convert the precipitate to crystalline 
 condition. Experiment (15) shows that precipitation is not 
 complete under these conditions. The zinc left in the solution 
 was precipitated at once as sulphide, and estimated as the 
 oxide, after dissolving in hydrochloric acid and precipitating 
 
 TABLE III. 
 
 Exp. 
 
 Zn 2 P 2 O 7 
 corre- 
 sponding 
 to 
 ZnS0 4 . 
 Taken. 
 
 Pound. 
 
 Error. 
 
 > 
 
 Error 
 in terms 
 of Zinc. 
 
 Zn 2 P 2 7 
 corre- 
 sponding 
 to Zn 
 left in the 
 filtrate. 
 
 (NH 4 )^0 4 . 
 
 NH 4 CL 
 
 Time of 
 stand- 
 ing* 
 
 A. 
 
 $ 
 
 (3) 
 
 grm. 
 0.6355 
 0.6355 
 0.6355 
 
 grin. 
 
 0.6206 
 0.6254 
 0.6300 
 
 grin. 
 
 0.0149- 
 0.0101- 
 0.0055- 
 
 grm. 
 0.0060- 
 0.0040- 
 0.0022- 
 
 grm. 
 Trace. 
 Trace. 
 Trace. 
 
 grm. 
 
 3.13 
 3.13 
 3.13 
 
 grm. 
 
 hrs. 
 
 iJ* 
 
 16 
 
 B. 
 
 (4) 
 (5) 
 
 0.6355 
 0.6355 
 
 0.6271 
 0.6256 
 
 0.0084- 
 0.0099- 
 
 0.0034- 
 0.0040- 
 
 Trace. 
 None. 
 
 HNaNH 4 P0 4 
 . 4H a O. 
 
 0.5 
 0.5 
 
 1 
 
 20 
 
 4.47 
 4.47 
 
 C. 
 
 (6) 
 
 | 
 
 (9). 
 (10) 
 
 (11) 
 (12) 
 (13) 
 (14) 
 
 0.6355 
 0.6355 
 0.6355 
 0.6355 
 0.6355 
 0.6355 
 0.6355 
 0.6355 
 0.6367 
 
 0.6285 
 0.6304 
 0.6295 
 0.6335 
 0.6381 
 0.6379 
 0.6386 
 0.6393 
 0.6355 
 
 0.0070- 
 0.0051- 
 0.0060- 
 0.0020- 
 0.0026+ 
 0.0024+ 
 0.0031+ 
 0.0038+ 
 0.0012+ 
 
 0.0028- 
 0.0020- 
 0.0024- 
 0.0008- 
 0.0010+ 
 0.0009+ 
 0.0012+ 
 0.0014+ 
 0.0005+ 
 
 None. 
 None. 
 None. 
 None. 
 None. 
 None. 
 None. 
 None. 
 None. 
 
 4.47 
 4.47 
 4.47 
 4.47 
 4.47 
 4.47 
 4.47 
 4.47 
 4.47 
 
 10 
 10 
 10 
 10 
 
 20 
 20 
 20 
 20 
 30 
 
 16 
 
 ,> 
 
 j 
 
 D. 
 
 (15) 
 (16) 
 
 0.6355 
 0.6355 
 
 0.6172 
 10.6227 
 II 0.0040 
 
 0.0183- 
 0.0098- 
 
 0.0072- 
 0.0039- 
 
 0.0108 
 None. 
 
 0.894 
 10.894 
 II 3.576 
 
 20 
 20 
 
 3 
 * 
 
 E. 
 
 (17) 
 (18) 
 (19) 
 
 0.6355 
 0.6355 
 0.6355 
 
 0.6270 
 0.6125 
 0.6303 
 
 0.0085- 
 0.0230- 
 0.0052- 
 
 0.0034- 
 0.0093- 
 0.0021- 
 
 None. 
 0.0148 
 0.0020 
 
 4.47 
 4.47 
 4.47 
 
 10 
 
 3 
 18 
 18 
 
262 DOUBLE AMMONIUM PHOSPHATES OF BERYLLIUM, 
 
 with sodium carbonate. In (16) of the table the first nitrate 
 was treated with an excess of microcosmic salt, and boiled. 
 Another portion of the ammonium zinc phosphate was 
 precipitated, and was filtered off and estimated. No zinc 
 was found by sulphuretted hydrogen in the second filtrate. 
 
 From the results it seems obvious also that an excess of the 
 soluble phosphate is necessary to complete the precipitation of 
 the zinc as the ammonium zinc phosphate instead of partly 
 ammonium zinc phosphate and partly tribasic phosphate. 
 
 In section E of the table are recorded results where the 
 precipitation was made in presence of an excess of the precipi- 
 tant either alone or in presence of ammonium chloride, the 
 solution being made faintly acid to litmus with acetic acid, 
 according to the manner in which Langmuir recommends to 
 conduct the precipitation. All the results by the method are 
 low. The condition of the ammonium zinc phosphate most 
 nearly approximating to the ideal is obtained as shown in ( 9) 
 to (14) by precipitating in presence of ammonium chloride in 
 large amount. Microcosmic salt is added until the solution 
 containing the ammonium salt is alkaline and the whole is 
 heated until the mass subsides in crystalline condition. The 
 amount of ammonium chloride should be twenty grams if the 
 filtration is to be made as soon as the solution cools. One-half 
 the amount will do if the liquid stands a number of hours. 
 Larger amounts tend to give a salt too rich in ammonia. The 
 time of standing seems to be a less important factor than either 
 the excess of microcosmic salt or ammonium chloride. 
 
 The Ammonium Cadmium Phosphate. 
 
 According to S. Drewsen* the cadmium ammonium phos- 
 phate is precipitated by allowing a solution of cadmium 
 sulphate to stand twenty-four hours with ammonium phosphate. 
 It is very soluble both in acids and alkalies. No further 
 preparation of this seems to have been recorded. For the 
 work on this salt to be given below, done with reference to 
 the constitution of the salt precipitated by hydrogen sodium 
 
 * Gmelin-Kraut, 6te Auflage, iii, 74. 
 
ZINC, AND CADMIUM IN ANALYSIS. 263 
 
 ammonium phosphate, the effect of ammonium chloride in the 
 precipitation, and the value of the salt for quantitative work, 
 the solution of cadmium chloride employed was prepared as 
 follows : A solution of cadmium sulphate acidulated with 
 hydrochloric acid was precipitated with sulphuretted hydrogen, 
 filtered and washed, and the precipitated sulphide was dissolved 
 in hydrochloric acid and filtered from possible traces of copper 
 and lead. The solution of the sulphide in hydrochloric acid 
 was boiled until all the sulphuretted hydrogen was expelled, 
 and filtered on asbestos in a perforated crucible of platinum 
 under pressure. The cadmium in the filtrate precipitated with 
 ammonium carbonate in excess was washed free from chloride, 
 dissolved in hydrochloric acid and diluted to definite volume. 
 It was standardized as oxide* after precipitating with sodium 
 carbonate with the necessary precautions. 
 
 The standard solution of cadmium chloride was drawn 
 carefully from a burette into a platinum dish, and, while hot, 
 was precipitated by adding hydrogen sodium ammonium 
 phosphate until the solution was alkaline to litmus. After 
 heating until the solution became crystalline, the whole stood 
 three hours in case of (1) of the table and sixteen hours in 
 case of (2) and (3), before filtering. In experiments (4) to 
 (12), inclusive, recorded in the table, precipitation was made 
 in the same manner as in (1) to (3) in presence of varying 
 amounts of ammonium chloride, and the precipitates were 
 filtered after standing as stated below in the table. It is clear 
 from the results that the cadmium separates out completely on 
 long standing only. Moreover, the ideal condition of the 
 ammonium cadmium, phosphate is obtained only when an 
 abundance of ammonium chloride is present; but large 
 amounts of ammonium chloride dissolve this salt. In (14), 
 where ammonia was added after precipitation was complete, 
 the salt dissolved somewhat ; also in (15), where the solution 
 was left faintly acid with acetic acid, a large part of the salt 
 was dissolved. These weights of cadmium dissolved in the 
 filtrate were obtained by treating the filtrates with sulphuretted 
 * Browning, Am. Jour. Sci., xlvi, 280. Volume I, p. 226. 
 
264 DOUBLE AMMONIUM PHOSPHATES OF BERYLLIUM. 
 
 TABLE IV. 
 
 
 Cd 2 P,0 7 
 
 
 
 
 Cd s P 2 7 
 corre- 
 
 
 
 Time 
 
 Exp. 
 
 corre- 
 sponding 
 to CdCl,. 
 
 Found. 
 
 Error. 
 
 Errorin 
 terms of 
 Cadmium. 
 
 sponding 
 toCd 
 found in 
 
 HNaNEUPO* 
 .iHsO. 
 
 NH 4 C1. 
 
 of 
 stand 
 ing. 
 
 
 Taken. 
 
 
 
 
 the nitrate. 
 
 
 
 
 
 griu. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 hra. 
 
 (1) 
 
 0.6972 
 
 0.6201 
 
 0.0771- 
 
 0.0434- 
 
 0.0059 
 
 4.5 
 
 t 
 
 3 
 
 (2) 
 
 0.6972 
 
 0.6135 
 
 0.0837- 
 
 0.0471- 
 
 None. 
 
 4.5 
 
 . . 
 
 16 
 
 (3) 
 
 0.6972 
 
 0.6134 
 
 0.0838- 
 
 0.0471- 
 
 None. 
 
 4.5 
 
 . . 
 
 16 
 
 (4) 
 
 0.6972 
 
 0.6792 
 
 0.0180- 
 
 0.0101- 
 
 Trace. 
 
 4.6 
 
 1 
 
 16 
 
 (5) 
 
 0.6972 
 
 0.6831 
 
 0.0141- 
 
 0.0079 
 
 0.0113 
 
 4.5 
 
 10 
 
 2 
 
 ( 6 ) 
 
 0.6972 
 
 0.6976 
 
 0.0004+ 
 
 0.0002-f 
 
 Trace. 
 
 4.5 
 
 10 
 
 16 
 
 (7) 
 
 0.6972 
 
 0.6969 
 
 0.0003- 
 
 0.0002- 
 
 Trace. 
 
 4.5 
 
 10 
 
 18 
 
 ( 8 ) 
 
 0.6972 
 
 0.6962 
 
 0.0010- 
 
 0.0006- 
 
 Trace. 
 
 4.5 
 
 10 
 
 16 
 
 9 
 
 0.6972 
 
 0.6891 
 
 0.0081- 
 
 0.0045- 
 
 0.0191 
 
 4.5 
 
 20 
 
 16 
 
 (10) 
 
 0.6972 
 
 0.6972 
 
 0.0000 
 
 0.0000 
 
 Trace. 
 
 4.5 
 
 20 
 
 16 
 
 (11) 
 
 0.6972 
 
 0.6942 
 
 0.0030- 
 
 0.0016- 
 
 Trace. 
 
 4.5 
 
 20 
 
 16 
 
 (12) 
 
 0.6972 
 
 0.6737 
 
 0.0235- 
 
 0.0132- 
 
 0.0304 
 
 4.5 
 
 30 
 
 16 
 
 (13) 
 
 0.6972 
 
 0.5655 
 
 0.1317- 
 
 0.0741- 
 
 0.1378 
 
 4.5 
 
 30 
 
 16 
 
 (14) 
 
 0.6972 
 
 0.6922 
 
 0.0050- 
 
 0.0023- 
 
 0.0088 
 
 4.5 
 
 10 
 
 16 
 
 (15) 
 
 0.6972 
 
 0.3209 
 
 0.3763- 
 
 0.2117- 
 
 0.2449 
 
 4.5 
 
 
 
 16 
 
 hydrogen, dissolving the sulphide in nitric acid, and weighing 
 as oxide after precipitating with sodium carbonate. 
 
 The ammonium cadmium phosphate is obtained in ideal 
 condition by precipitating with microcosmic salt in presence of 
 10 grm. ammonium chloride in a total volume of 100 cm 3 to 
 150 cm 3 shown in (6), (7), and (8) filtering after standing 
 some time. On drying and igniting the pyrophosphate is left. 
 Very large amounts of ammonium chloride 30 grm. 
 dissolve the salt, and seem to tend to cause the formation of a 
 phosphate too rich in ammonia. Either acid or ammonia in 
 small amount dissolves the salt, as is shown in (14) and (15). 
 
 The results of this investigation as to the analytical 
 application of the double ammonium phosphates of beryllium, 
 zinc, and cadmium may be summarized briefly as follows : It 
 is impossible to estimate beryllium with accuracy as the 
 pyrophosphate obtained by igniting the double ammonium 
 phosphate precipitated from beryllium solutions by microcosmic 
 salt or ammonium phosphate in presence of ammonium chloride. 
 In presence of the proper amount of ammonium chloride (10 
 grm. to 20 grm. in 100 cm 3 -200 cm 3 of liquid) zinc ammonium 
 
ZINC, AND CADMIUM IN ANALYSIS. 265 
 
 phosphate can be obtained in the ideal condition, which on 
 ignition yields the pyrophosphate. This method may serve, 
 therefore, for the accurate estimation of zinc. 
 
 Cadmium may be estimated with accuracy as the pyrophos- 
 phate if the precipitate by microcosmic salt in the nearly 
 neutral solution containing ammonium chloride in the 
 proportion of ten grams to one hundred cubic centimeters is 
 allowed to stand several hours before filtering. In this way 
 all cadmium separates out from the solution as a beautiful 
 crystalline mass of cadmium ammonium phosphate of ideal 
 constitution. The conditions, must, however, be preserved 
 with care ; there must be no excess of ammonia, no free acid, 
 and no excess of ammonium salt beyond the quantity indicated, 
 while that amount is necessary. 
 
XXXII 
 
 SEPARATION OF IRON FROM CHROMIUM, ZIRCONIUM, 
 AND BERYLLIUM, BY THE ACTION OF GASEOUS 
 HYDROCHLORIC ACID ON THE OXIDES. 
 
 BY FRANKE STUART HAVENS AND ARTHUR FITCH WAY.* 
 
 IT has been shown in a former paper from this laboratory f 
 that iron oxide may be completely volatilized as chloride 
 by a strong current of hydrochloric acid gas acting at a 
 temperature of 450 -500, and also that the addition of a little 
 free chlorine to the gaseous hydrochloric acid renders this 
 action complete at lower temperatures, 180-200, without the 
 danger of error arising from the liability of ferric chloride to 
 dissociation, or from deficiency of oxidation in the oxide 
 treated, or mechanical loss due to too rapid volatilization. 
 It has also been shown that this reaction can be employed for 
 the separation of iron and aluminum, taken as the oxides, and 
 its application to the separation of iron from other metallic 
 oxides has been suggested. 
 
 The oxides of chromium, zirconium, and beryllium, like 
 aluminum oxide, are not acted upon by a current of dry 
 hydrochloric acid gas at the temperatures before mentioned, 
 and these oxides also can be entirely freed from iron by this 
 reaction, as the experiments to be described will show. The 
 procedure was the same in each case and analogous to 
 that employed for the separation of iron from aluminum. A 
 mixture of a weighed portion of one of these oxides with a 
 weighed portion of ferric oxide, contained in a porcelain boat 
 and placed within a roomy glass tube supported in a small 
 
 * From Am. Jour. Sci., viii, 217. 
 
 t Gooch and Havens, Am. Jour. Sci., vii, 370. This volume, p. 215. 
 
SEPARATION OF IRON FROM CHROMIUM, ETC. 267 
 
 combustion furnace, was submitted to the action of a dry 
 current of hydrochloric acid gas and chlorine generated by 
 dropping sulphuric acid upon a mixture of strong hydrochloric 
 acid, common salt, and a small amount of manganese dioxide. 
 The gas was admitted at one end of the combustion tube and 
 passed out at the other through a water trap, while the required 
 temperature, from 200 -300, was maintained by regulating 
 the various burners of the furnace. The time of action varies 
 somewhat with the condition of the oxide to be volatilized, and 
 the temperature; generally an hour's heating at 200, proves 
 sufficient for the complete removal of 0.1 grm. of iron. At 
 higher temperatures the action is more rapid ; but the lighter 
 oxide, the beryllium especially, is liable to mechanical loss 
 through the too rapid volatilization of the iron, as experiment 
 (17), where a temperature of 500 was used, will show. It is 
 better, therefore, to use lower temperatures, raising the heat 
 for a few minutes when the action is apparently complete to 
 ensure the removal of the last traces of iron. Tests showed 
 
 Exp. 
 
 Fe 2 3 taken. 
 
 Cr,O 3 taken. 
 
 Cr 2 O 8 found. 
 
 Error. 
 
 (1) 
 
 (2) 
 (8) 
 (4) 
 (5) 
 (6) 
 
 grm. 
 
 0.1007 
 0.1007 
 0.1010 
 0.1019 
 0.2007 
 
 grm. 
 0.1008 
 0.1006 
 0.1000 
 0.1005 
 0.1006 
 0.1003 
 
 grm. 
 0.1008 
 0.1006 
 0.1002 
 0.1003 
 0.1005 
 0.0999 
 
 grm. 
 
 0.0000 
 0.0000 
 0.0002+ 
 0.0002- 
 0.0001- 
 0.0004- 
 
 
 
 ZrO, taken. 
 
 ZrO, found. 
 
 
 (7) 
 (8) 
 (9) 
 (10) 
 
 (11) 
 
 0.1053 
 0.1204 
 0.1236 
 0.2150 
 
 0.1516 
 0.1010 
 0.1519 
 0.1516 
 0.1517 
 
 0.1516 
 0.1010 
 0.1523 
 0.1517 
 0.1519 
 
 0.0000 
 0.0000 
 0.0004+ 
 0.0001+ 
 0.0002+ 
 
 
 
 BeO taken. 
 
 BeO found. 
 
 
 (12) 
 (13) 
 (14) 
 (16) 
 (16) 
 (17) 
 18) 
 
 0.0997 
 0.1045 
 0.1215 
 0.1510 
 0.0230 
 
 0.1309 
 0.1285 
 0.0456 
 0.1099 
 0.1080 
 0.1305 
 0.1081 
 
 0.1311 
 0.1285 
 0.0457 
 0.1099 
 0.1081 
 0.1290 
 0.1083 
 
 0.0002+ 
 0.0000 
 0.0001+ 
 0.0000 
 0.0001+ 
 0.0015- 
 0.0002+ 
 
268 SEPARATION OF IRON FROM CHROMIUM, ETC. 
 
 the residual oxides from which the ferric oxide had been 
 removed in this manner to be entirely free from iron. 
 
 The separation of iron from chromium, zirconium, and 
 beryllium by this method is obviously complete within very 
 satisfactory limits of error. 
 
XXXIII 
 
 THE IODOMETRIC DETERMINATION OF GOLD. 
 
 BY F. A GOOCH AND FREDERICK H. MORLEY.* 
 
 IN a recent attempt to measure small amounts of gold in 
 solution by titrating with sodium thiosulphate the iodine set 
 free in the action of an excess of potassium iodide upon auric 
 chloride, Petersonf has been led to conclude that, on the aver- 
 age, one-half more thiosulphate is used up in changing the 
 characteristic starch iodide blue to the faint rose color which 
 precedes entire bleaching than is called for upon the theory 
 that the thiosulphate is simply converted to the tetrathionate 
 in the usual manner. Peterson explains the anomaly upon the 
 hypothesis that, besides acting upon the free iodine, the thio- 
 sulphate is used up coincidently by interaction with the aurous 
 salt, formed in the reduction, with formation of a gold sodium 
 thiosulphate on the type of the well-known silver sodium thio- 
 sulphate. The reaction of this hypothesis is in the nature of 
 things most improbable, since there is no reason to suppose 
 that the soluble double thiosulphate could resist the action of 
 the free iodine which is present to the end the appearance 
 of the rose color, and our study of the reaction of sodium thio- 
 sulphate upon the mixture of gold chloride and potassium 
 iodide, the account of which follows, discloses no evidence of 
 the consumption of more thiosulphate than is demanded by 
 the usual theory, which postulates the simple formation of the 
 tetrathionate by the interaction of the thiosulphate and free 
 iodine. 
 
 It appeared in the course of our preliminary experimenta- 
 
 * From Am. Jour. Sci., viii, 261. 
 t Zeitschr. anorg. Chem., xix, 63. 
 
270 
 
 IODOMETRIC DETERMINATION OF GOLD. 
 
 tion that, while practically similar results were obtained by 
 adding the thiosulphate until the blue of the starch iodide had 
 changed to rose, the indications were somewhat more con- 
 cordant when the final rose color was developed by adding 
 iodine to the solution from which the blue had been bleached 
 to colorlessness by a slight excess of the thiosulphate. 
 
 It appeared, also, that the reduction of the auric salt, with 
 the consequent liberation of iodine, is conditioned by the vol- 
 ume of the solution, the mass of the iodine present, and the 
 time of action. 
 
 The following statement, in which each result is the average 
 of several titrations in close agreement, shows the effect upon 
 the immediate evolution of iodine brought about by adding 
 varying amounts of water to the gold solution before introduc- 
 ing the iodide, and the effect of different amounts of iodide 
 at different dilutions. 
 
 
 Potassium iodide. 
 
 Gold 
 chloride. 
 
 Volume 
 before the 
 
 
 
 
 addition of 
 
 
 
 
 
 
 
 
 the thio- 
 
 
 0.01 grm. 
 
 0.02 grm. 
 
 0.05 grm. 
 
 0.1 grm. 
 
 0.2 grm. 
 
 0.00087 grm. 
 
 sulphate. 
 
 .sF . 
 
 fO.81 
 
 0.81 
 
 0.81 
 
 0.82 
 
 0.84 
 
 0.00087 
 
 cm 3 
 15 
 
 If 1 ?"!. 
 
 0.77 
 0.74 
 
 0.78 
 0.72 
 
 0.80 
 0.78 
 
 0.81 
 0.79 
 
 0.81 
 
 0.80 
 
 0.00087 
 0.00087 
 
 25 
 60 
 
 11 1 S 
 
 0.61 
 
 0.61 
 
 0.68 
 
 0.76 
 
 0.79 
 
 0.00087 
 
 100 
 
 2 fl 
 
 [0.45 
 
 0.49 
 
 0.60 
 
 0.72 
 
 0.75 
 
 0.00087 
 
 200 
 
 It is evident that for the smaller amounts of iodide the lib- 
 eration of iodine decreases rapidly with the dilution. The 
 larger amounts at the highest concentration show readings 
 a trifle above the normal perhaps because the well-known 
 effect of concentrated solutions of a soluble iodide upon the 
 delicacy of the starch end-color begins to appear. At vol- 
 umes lying between the limit of 25 cm 3 and 50 cm 3 0.1 
 grm. of potassium iodide is an appropriate amount to use; 
 at a volume of 15 cm 3 , 0.01 grm. to 0.05 grm. of the iodide 
 will do the work; and at lower dilutions, as will appear in 
 
IODOMETRIC DETERMINATION OF GOLD. 271 
 
 the tabular statements to follow, even less of the iodide is 
 effective. 
 
 In the series of experiments of which the details are given 
 in Table I, use was made of a solution of pure gold chloride 
 containing 0.8710 grm. to the liter as determined by careful 
 precipitation in the usual manner by ferrous sulphate, and by 
 an alkaline solution of formaldehyde according to the method 
 of Vanino.* A nearly centinormal solution of iodine was pre- 
 pared by diluting to a liter 100 cm 3 of nearly decinormal iodine 
 in potassium iodide carefully standardized against exactly 
 decinormal arsenious acid. A nearly centinormal solution of 
 sodium thiosulphate (containing 1.7012 grm. of Na 2 S 2 O 8 to 
 the liter) was made by diluting to a liter 100 cm 3 of a nearly 
 decinormal solution of that reagent which had been standard- 
 ized carefully against the standard iodine prepared as described. 
 The solution of potassium iodide employed contained 10 grm. 
 of that salt in the liter. 
 
 In conducting the experiments, a convenient amount of the 
 solution of gold chloride was drawn from a burette, potassium 
 iodide was introduced in the amounts indicated (always several 
 times the theoretical equivalent of the gold, and more than 
 enough to dissolve the aurous iodide precipitated at first), a 
 sufficiency of clear starch indicator was added, the starch blue 
 was bleached by the thiosulphate, and the iodine was added 
 until the liquid assumed a faint rose color. Upon the theory 
 that potassium iodide sets free two atoms of iodine for every 
 molecule of auric chloride (or every atom of gold) present, 
 and that the thiosulphate acts only upon the free iodine to 
 form the tetrathionate in the usual manner, every cubic centi- 
 meter of the thiosulphate solution used in the reaction after 
 deducting the amount equivalent to the iodine introduced to 
 get the end-color, should represent 
 
 1Q7 3 
 
 X 0.0017012 = 0.001061 grm. of gold. 
 
 2(158.22) 
 
 Ber. Dtsch. chem. Ges., xxi, 1763. 
 
272 IODOMETRIC DETERMINATION OF GOLD. 
 
 TABLE I. 
 
 
 Gold chloride = 0.8710 to 1 liter. 
 
 
 Sodium thiosulphate, nearly ~ = 1.7012 " " 
 
 
 Iodine, nearly ^ =1.3697 " " 
 
 
 Volume at beginning of titration, approximately 50 cm 8 . 
 
 
 AuCl 3 
 taken. 
 
 KI 
 
 taken. 
 
 Na 2 S 2 8 
 used. 
 
 Gold found. 
 
 Theory 
 for gold. 
 
 Error. 
 
 Per cent. 
 
 
 
 cm 3 
 
 grm. 
 
 cm 3 
 
 grm. 
 
 grm. 
 
 grin. 
 
 
 (1) 
 
 5 
 
 0.05 
 
 4.02 
 
 0.00426 
 
 0.00435 
 
 0.00009- 
 
 2.1 
 
 (2) 
 
 6 
 
 0.05 
 
 4.01 
 
 0.00425 
 
 0.00435 
 
 0.00010- 
 
 2.3 
 
 
 3) 
 
 5 
 
 0.05 
 
 4.06 
 
 0.00431 
 
 0.00435 
 
 0.00004- 
 
 0.9 
 
 
 4) 
 
 5 
 
 0.05 
 
 4.07 
 
 0.00432 
 
 0.00435 
 
 0.00003- 
 
 0.7 
 
 5) 
 
 5 
 
 0.05 
 
 4.04 
 
 0.00428 
 
 0.00435 
 
 0.00007- 
 
 1.6 
 
 
 6) 
 
 10 
 
 0.08 
 
 8.17 
 
 0.00867 
 
 0.00871 
 
 0.00004- 
 
 0.5 
 
 
 7) 
 
 10 
 
 0.08 
 
 8.15 
 
 0.00864 
 
 0.00871 
 
 0.00007- 
 
 0.8 
 
 
 8) 
 
 10 
 
 0.08 
 
 8.16 
 
 0.00865 
 
 0.00871 
 
 0.00006- 
 
 0.7 
 
 
 9 
 
 10 
 
 0.08 
 
 8.15 
 
 0.00864 
 
 0.00871 
 
 0.00007- 
 
 0.8 
 
 (10) 
 
 10 
 
 0.08 
 
 8.19 
 
 0.00869 
 
 0.00871 
 
 0.00002- 
 
 0.2 
 
 (11) 
 
 10 
 
 0.08 
 
 8.46 
 
 0.00897 
 
 0.00871 
 
 0.00026+ 
 
 3.0 
 
 (12) 
 
 10 
 
 0.08 
 
 8.24 
 
 0.00874 
 
 0.00871 
 
 0.00003+ 
 
 0.3 
 
 Plainly, these results accord reasonably with the theory that 
 two molecules of the thiosulphate are the equivalent in this 
 reaction of two atoms of iodine and one atom of gold. There 
 is no evidence whatever of the excessive action affirmed by 
 Peterson. 
 
 The strength of the standard solutions used in the experi- 
 ments described was such that an error of 0.01 cm 3 in reading 
 the volumes used would correspond to an error of 0.00001 
 grm. of gold. It is not to be expected that such readings can 
 be trusted ordinarily to a higher degree of accuracy than 0.02 
 cm 3 . In case all three solutions should be read to this limit 
 of accuracy with the errors of all lying in the same direction, 
 the summation of error would correspond to 0.00006 grm. of 
 gold. 
 
 In the following experiments, therefore, solutions obtained 
 by properly diluting those of the previous series were em- 
 ployed. The use of a more dilute solution of gold obviated 
 the necessity for diluting the mixture of gold chloride and the 
 iodide before titrating with the thiosulphate. It was found, 
 
IODOMETRIC DETERMINATION OF GOLD. 273 
 TABLE II. 
 
 A. 
 
 Gold chloride = 0.0871 to 1 liter. 
 
 Sodium thiosulphate, nearly ^ = 1.7012 " " 
 
 Iodine, nearly j =1.3697 " 
 
 Solution of gold chloride not diluted before mixing with potassium iodide. 
 
 Exp. 
 
 AuCl s 
 taken. 
 
 Kl 
 
 taken. 
 
 JSra,BJO, 
 used. 
 
 Gold 
 taken. 
 
 Gold 
 found. 
 
 Error. 
 
 
 cm 8 
 
 grin. 
 
 cm 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 10 
 
 0.01 
 
 0.83 
 
 0.00087 
 
 0.00088 
 
 0.00001+ 
 
 2) 
 
 10 
 
 0.01 
 
 0.83 
 
 0.00087 
 
 0.00088 
 
 0.00001+ 
 
 3) 
 
 10 
 
 0.01 
 
 0.80 
 
 0.00087 
 
 0.00085 
 
 0.00002- 
 
 4) 
 
 10 
 
 0.02 
 
 0.84 
 
 0.00087 
 
 0.00089 
 
 0.00002+ 
 
 5) 
 
 10 
 
 0.02 
 
 0.88 
 
 0.00087 
 
 0.00093 
 
 0.00006+ 
 
 6) 
 
 10 
 
 0.02 
 
 0.82 
 
 0.00087 
 
 0.00087 
 
 0.00000 
 
 (7) 
 
 10 
 
 0.02 
 
 0.88 
 
 0.00087 
 
 0.00093 
 
 0.00006+ 
 
 (8) 
 
 10 
 
 0.02 
 
 0.83 
 
 0.00087 
 
 0.00088 
 
 0.00001+ 
 
 (9) 
 
 10 
 
 0.10 
 
 0.80 
 
 0.00087 
 
 0.00085 
 
 0.00002- 
 
 (10) 
 
 10 
 
 0.10 
 
 0.82 
 
 0.00087 
 
 0.00087 
 
 0.00000 
 
 (11 
 
 10 
 
 0.01 
 
 0.83 
 
 0.00087 
 
 0.00088 
 
 0.00001+ 
 
 (12) 
 
 9 
 
 0.01 
 
 0.73 
 
 0.00078 
 
 0.00077 
 
 0.00001- 
 
 (13) 
 
 8 
 
 0.01 
 
 0.65 
 
 0.00070 
 
 0.00069 
 
 0.00001- 
 
 (14) 
 
 7 
 
 0.01 
 
 0.58 
 
 0.00061 
 
 0.00061 
 
 0.00000 
 
 (15) 
 
 6 
 
 0.008 
 
 0.51 
 
 0.00052 
 
 0.00054 
 
 0.00002+ 
 
 (16) 
 
 6 
 
 0.008 
 
 0.41 
 
 0.00043 
 
 0.00044 
 
 0.00001+ 
 
 (17) 
 (18) 
 
 4 
 3 
 
 0.005 
 0.005 
 
 0.35 
 0/24 
 
 0.00035 
 0.00026 
 
 0.00037 
 0.00026 
 
 0.00002+ 
 0.00000 
 
 (19) 
 
 2 
 
 0.003 
 
 0.21 
 
 0.00017 
 
 0.00022 
 
 0.00005+ 
 
 (20) 
 
 1 
 
 0.003 
 
 0.10 
 
 0.00009 
 
 0.00011 
 
 0.00002+ 
 
 B. 
 
 Gold chloride =0.0871 to 1 liter. 
 
 Sodium thiosulphate, nearly ~ = 0.17012 " " 
 
 Iodine, nearly ^ =0.13697 " " 
 
 (21) 
 
 10 
 
 0.01 
 
 8.39 
 
 0.000871 
 
 0.000890 
 
 0.000019+ 
 
 (22) 
 
 9 
 
 0.01 
 
 7.45 
 
 0.000784 
 
 0.000790 
 
 0.000006+ 
 
 (23) 
 
 8 
 
 0.01 
 
 6.30 
 
 0.000697 
 
 0.000668 
 
 0.000029- 
 
 (24) 
 
 7 
 
 0.008 
 
 5.50 
 
 0.000610 
 
 0.000583 
 
 0.000027- 
 
 (25) 
 
 6 
 
 0.008 
 
 5.12 
 
 0.000523 
 
 0.000543 
 
 0.000020+ 
 
 (26) 
 
 5 
 
 0.005 
 
 4.23 
 
 0.000435 
 
 0.000449 
 
 0.000014+ 
 
 (27) 
 
 4 
 
 0.005 
 
 3.38 
 
 0.000348 
 
 0.000358 
 
 0.000010+ 
 
 (28) 
 
 3 
 
 0.003 
 
 2.55 
 
 0.000261 
 
 0.000270 
 
 0.000009+ 
 
 (29) 
 
 2 
 
 0.003 
 
 1.71 
 
 0.000174 
 
 0.000181 
 
 0.000007+ 
 
 (30) 
 
 1 
 
 0.003 
 
 0.90 
 
 0.000087 
 
 0.000095 
 
 0.000008+ 
 
 however, that when the T ^- solution of iodine is employed a 
 correction of 0.1 cm 3 for volumes not exceeding 30 cm 3 be- 
 
 VOL. II. 
 
 18 
 
274 IODOMETRIC DETERMINATION OF GOLD. 
 
 comes necessary the amount required to bring out the rose 
 color in blank tests containing no gold. After the introduc- 
 tion of 0.1 cm 3 of 1 -j W5 iodine into a mixture of potassium 
 iodide and starch indicator of volume not exceeding 30 cm 3 , a 
 single drop of the gold solution equivalent to 0.000002 grm. 
 of gold gave a distinct rose color : before such adjustment 
 of the solution five drops equivalent to 0.000010 of gold 
 were needed to develop the same color. 
 
 These results run on the whole as regularly as could be 
 expected, and the use of the dilute standard solutions is 
 obviously of advantage. 
 
 In the practical application of any such process for the 
 determination of gold, the elementary form of that metal is 
 the natural starting-point. To get the metal into solution 
 with chlorine water or mixed hydrochloric and nitric acids is 
 an easy matter, but the removal of the excess of the oxidizer 
 by evaporation without reducing some auric chloride to the 
 aurous form is difficult. We have found, however, that the 
 free chlorine may be removed from a solution of auric 
 chloride, without reducing the auric salt, by treatment of the 
 solution with ammonia in excess, boiling gently, acidifying 
 with hydrochloric acid and heating if necessary to redissolve 
 the precipitate by ammonia, again treating with ammonia and 
 heating, and once more acidifying. On the second addition 
 of ammonia no precipitation usually takes place with the 
 amounts of gold which we have thus handled, perhaps because 
 enough ammonium chloride has been found to hold it up. 
 
 The following table contains determinations made with 
 such a solution of pure gold leaf tested gravimetrically as 
 to purity. 
 
 Obviously, this method, which rests upon the hypothesis 
 that sodium thiosulphate acts in the normal manner only 
 upon the iodine set free by the interaction of gold chloride 
 and potassium iodide, offers trustworthy means for the 
 determination of small amounts of gold. 
 
IODOMETRIC DETERMINATION OF GOLD. 275 
 TABLE III. 
 
 Gold chloride made by dissolving 0.0104 grm. of pure gold in the manner 
 
 described and diluting to 200 cm 8 . 
 
 Sodium thiosulphate, nearly j^ = 0.17012 to 1 liter. 
 
 Iodine, nearly ~ =0.13697 " " 
 
 Potassium iodide = 10 grm. " " 
 
 Portions were treated with the potassium iodide without previous dilution. 
 
 Eip. 
 
 AuCl. 
 taken. 
 
 KI 
 
 taken. 
 
 &' 
 
 Gold 
 taken. 
 
 Gold 
 found. 
 
 Error. 
 
 
 cm 3 
 
 grm. 
 
 cm** 
 
 grm. 
 
 grm. 
 
 gnn. 
 
 (1) 
 
 1 
 
 0.005 
 
 0.55 
 
 0.000052 
 
 0.000058 
 
 0.000006+ 
 
 (2) 
 
 1 
 
 0.005 
 
 0.55 
 
 0.000052 
 
 0.000058 
 
 0.000006+ 
 
 (3) 
 
 2 
 
 0.005 
 
 1.06 
 
 0.000104 
 
 0.000112 
 
 0.000008+ 
 
 (4) 
 
 2 
 
 0.005 
 
 1.08 
 
 0.000104 
 
 0.000114 
 
 0.000010+ 
 
 (5) 
 
 5 
 
 0.01 
 
 2.45 
 
 0.000260 
 
 0.000260 
 
 0.000000 
 
 (6 
 
 5 
 
 0.01 
 
 2.50 
 
 0.000260 
 
 0.000265 
 
 0.000005+ 
 
 7) 
 
 5 
 
 0.01 
 
 2.45 
 
 0.000260 
 
 0.000260 
 
 0.000000 
 
 (8) 
 
 5 
 
 0.01 
 
 2.50 
 
 0.000260 
 
 0.000265 
 
 0.000005+ 
 
 (9) 
 
 5 
 
 0.01 
 
 2.50 
 
 0.000260 
 
 0.000265 
 
 0.000005+ 
 
 (10) 
 
 10 
 
 0.02 
 
 4.86 
 
 0.000520 
 
 0.000515 
 
 0.000005- 
 
 (11) 
 
 10 
 
 0.02 
 
 4.85 
 
 0.000520 
 
 0.000517 
 
 0.000003- 
 
 (12) 
 
 10 
 
 0.02 
 
 4.90 
 
 0.000520 
 
 0.000520 
 
 0.000000 
 
 (13) 
 
 10 
 
 0.02 
 
 4.80 
 
 0.000520 
 
 0.000512 
 
 0.000008- 
 
 (14) 
 
 10 
 
 0.02 
 
 4.84 
 
 0.000520 
 
 0.000516 
 
 0.000004- 
 
XXXIV 
 
 THE ACTION OF ACETYLENE ON THE OXIDES 
 OF COPPER. 
 
 BY F. A. GOOCH AND DEFOREST BALDWIN.* 
 
 IN a recent paper by Erdmann and Kothnerf an account is 
 given of the formation of a peculiar, light-brown, highly 
 voluminous substance by the action of acetylene below 250 
 C. upon cuprous oxide, or even (though more slowly) upon 
 copper. The product obtained by passing acetylene during 
 eighteen hours over 1 grm. of cuprous oxide (prepared 
 from copper sulphate, grape sugar, and sodium hydroxide) 
 amounted to 7 grm. and filled a space of nearly 300 cm 3 . 
 At higher temperatures a black carbonaceous mass is the 
 result, and at red heat (400-500 C.) carbon is deposited in 
 graphitic condition. The light-brown fluffy material yielded 
 cuprous chloride to hydrochloric acid, a distillate from its 
 mixture with zinc dust possessing the characteristics of 
 naphthene or, at higher temperature and under rapid heating, 
 aromatic compounds among which naphthalene and a kresol 
 were indicated. Erdmann and Kothner classify this body as 
 a very complex but non-explosive copper acetylene (acetylen- 
 kupfer,), and from their analyses deduce the formula C4 4 H 64 
 Cu s . Apart from the unusual constitution of this symbol, 
 its most striking peculiarity is that it implies a loss of carbon, 
 rather than hydrogen, from the acetylene in the reaction with 
 cuprous oxide a condition of affairs which would be most 
 remarkable in the light of Campbell's experience,^ according 
 to which acetylene passed over palladinized copper oxide 
 
 * From Am. Jour. Sci., viii, 354. t Zeitschr. anorg. Chem., xviii, 49. 
 t Amer. Chem. Jour., xvii, 690. 
 
ACTION OF ACETYLENE ON OXIDES OF COPPER. 277 
 
 yielded water at 225-230 and carbon dioxide only when 
 the temperature rose to 315-320 with the formation of a 
 black deposit. Upon scrutinizing the figures of Erdmann and 
 Kbthner with care, however, it appears that the formula 
 given by these investigators rests upon some oversight in 
 calculation: the ratio of carbon atoms to hydrogen atoms 
 proves to be actually, according to the data given, 6.45 : 5.70 ; 
 which means, of course, that the new product is deficient, as 
 would be expected, in hydrogen (not in carbon) as compared 
 with acetylene. 
 
 As to the content of the new substance in copper, the 
 analytical data are unfortunately ambiguous; for we note 
 the weights found of copper oxide converted into percentages 
 of copper without preliminary reduction. If the fault is 
 typographical and in the analytical data, the calculated 
 percentages of copper being correct, the average percentage 
 of copper amounts to 15.43 : if, on the other hand, the ana- 
 lytical data are right, the error being in their reduction, the 
 percentage of copper amounts to 12.92. In the one case the 
 summation of the analysis leaves a deficiency of about 1.5 
 per cent, and in the other of about 4 per cent, which hi either 
 case may really represent oxygen hi the substance. This 
 condition of matters leaves the " acetylen-kupfer " of Erd- 
 mann and Kbthner in uncertain standing. 
 
 More than thirty years ago it was noticed by Berthelot * 
 that acetylene is polymerized by heat or decomposed partially 
 into carbon and hydrogen, and that such action takes place 
 more readily and at lower temperatures in presence of metallic 
 iron with production of carbon, hydrogen and compounds 
 different from those formed by heat alone. 
 
 Moissan and Moureu f have observed the incandescence of 
 acetylene passed over finely divided iron, cobalt, nickel, or 
 platinum at the ordinary temperature, with production of 
 carbon, hydrogen, and pyrogenic compounds, and have found 
 the occasion of such behavior in the porosity of the metals 
 employed. 
 
 * Ann. Chim. [4], ix, 448. t Compt. rend., cxxii, 1240. 
 
278 THE ACTION OF ACETYLENE 
 
 It would seem natural, however, that the presence of oxygen, 
 free or combined, may also play a considerable part in such 
 phenomena, just as appears to be the case in the peculiar action 
 recorded by Gruner J of carbon monoxide upon iron reduced 
 by hydrogen, which, as Moissan has shown, is produced pure 
 only with the greatest precaution and generally carries a 
 large proportion of ferrous oxide. The fact that the " acety- 
 len-kupfer " of Erdmann and Kothner is produced more easily 
 by the action of cuprous oxide upon acetylene than by the 
 action of metallic copper upon acetylene, suggests that it may 
 be the oxidizing power of the cuprous oxide which gives 
 to this reagent its peculiar activity. The question arises, 
 therefore, as to whether the copper is in reality an essential 
 constituent of the compound of Erdmann and Kothner. 
 
 In our experiments upon the action of acetylene upon the 
 oxides of copper (and other elements) we have conducted the 
 gas (made in the ordinary way by the action of water on 
 calcium carbide, and kept over water) over the oxide contained 
 in a porcelain boat placed within a glass tube, 2 cm. in diameter 
 and 50 cm. long, which was heated over a small combustion 
 furnace. The glass tube was fitted at each end with a rubber 
 stopper, one carrying a smaller tube for the introduction of the 
 acetylene and a high-temperature thermometer so held that its 
 bulb rested horizontally immediately over the boat containing 
 the oxide, while the other was fitted with a water-trap. In the 
 preh'minary experiments no attempt was made to purify the 
 acetylene employed other than to keep it over water, or, since 
 water is a product of its action upon oxides, to dry it : in later 
 experiments to secure products for careful analysis it was dried 
 and purified with care. 
 
 We found that 225 C. is the temperature most favorable 
 for the formation of the voluminous product obtained by acting 
 with acetylene upon cuprous oxide as described by Erdmann 
 and Kothner. At this temperature the tube is choked rapidly 
 with the fluffy product and water forms, but, as Campbell 
 found in his experiments upon palladinized copper oxide, no 
 
 t Ann. Chim. [4], xxvi, 6. Ann. Chim. [5], xxi, 199. 
 
ON THE OXIDES OF COPPER. 279 
 
 appreciable amount of carbon dioxide is produced. The 
 content of the product in copper varies in the sample and in 
 different experiments, our results lying between 1.54 per cent 
 and 24.21 per cent of the substance taken for ignition. 
 
 It appeared, also, that the action of acetylene upon cupric 
 oxide is precisely similar to that upon cuprous oxide excepting 
 the evident reduction of the former oxide early in the action. 
 The amount of copper in the product of such action varied in 
 our experiments from 6.53 per cent to 21.30 per cent. In one 
 case the experiment of re-submitting to the action of acetylene 
 a product containing 9.34 per cent of copper was made with 
 the result that a new growth of the substance formed which 
 on analysis yielded 3.87 per cent of copper. 
 
 A roll of copper gauze carefully reduced in hydrogen and 
 then oxidized at one end in the outer flame of a Bunsen 
 burner gave, when acted upon by acetylene at 225-250 C., 
 the characteristic deposit upon the oxidized end only, the 
 unoxidized end being merely discolored. 
 
 These results go to show that, while metallic copper may at 
 comparatively high temperatures induce the polymerization of 
 acetylene, it is an oxidiiing action which starts at moderately 
 low temperatures the formation of the peculiar derivatives 
 under consideration. Thus we find that ferric oxide heated in 
 acetylene at temperatures varying from 150 to 360, accord- 
 ing to circumstances, darkens, glows, and gathers with evolution 
 of heat a dark carbonaceous deposit. In the products of such 
 action we have found the content of iron varying from 2.80 
 per cent to 5.86 per cent. 
 
 Silver oxide, too, acts upon acetylene: thus, in one 
 experiment, action was evident at the ordinary atmospheric 
 temperature, and a violent explosion, which completely shat- 
 tered the boat and scattered metallic silver upon the sides of 
 the glass tube, followed before the temperature reached 100. 
 
 In the locally violent explosion of the last experiment 
 we have evidence of the formation in the early stage of an 
 acetylide which is decomposed later when the temperature of 
 dissociation is reached. In the experiments with the oxides of 
 
280 
 
 THE ACTION OF ACETYLENE 
 
 copper and iron the temperature at which the acetylene begins 
 to act is evidently above the point at which sensitive acetylides 
 would naturally dissociate, and we have in the observed 
 phenomena no evidence of the formation of such compounds of 
 copper and iron under the conditions of experimentation. 
 
 In experiments (1) to (3) of the following table are given 
 the results of the analysis of several products obtained by 
 conducting acetylene (purified by passing through a solution 
 of mercuric chloride in hydrochloric acid and dried over caustic 
 potash) over pure cuprous oxide. The temperature was kept 
 in these experiments at 225, and in the course of a half-hour 
 the tube was choked completely by material compacted by 
 the pressure to (1) a spongy mass of light-brown color on the 
 exterior next the walls of the tube, (2) darker within and (3) 
 nearly black in the bottom of the boat, where the cuprous 
 oxide lay originally. 
 
 Ezp. 
 
 Weight 
 of sub- 
 stance 
 taken. 
 
 Found. 
 
 Calculated. 
 
 CO, 
 
 H 2 O 
 
 CuO 
 
 C 
 
 H 
 
 Cu 
 
 Oby 
 differ- 
 ence. 
 
 (1) 
 (2) 
 (3) 
 
 grin. 
 
 0.1170 
 0.2247 
 0.1096 
 
 grnit 
 
 0.3978 
 0.7489 
 0.3678 
 
 gnu. 
 0.0673 
 0.0979 
 0.0488 
 
 gnu. 
 0.0022 
 
 o!o045 
 
 gnu. 
 0.1085 
 0.2042 
 0.1003 
 
 grm. 
 0.0075 
 0.0109 
 0.0054 
 
 grm. 
 0.0018 
 
 grm. 
 
 0.0008- 
 
 0.0036 
 
 0.0003 
 
 (4) 
 (5) 
 
 0.1360 
 0.1188 
 
 0.4116 
 0.3098 
 
 0.0579 
 0.0461 
 
 0.0182 
 0.0317 
 
 0.1123 
 0.0845 
 
 0.0064 
 0.0051 
 
 0.0146 
 0.0253 
 
 0.0027 
 0.0039 
 
 Per cent of carbon 
 Pet cent of hydrogen 
 Per cent of copper 
 Per cent of oxygen 
 
 (1) 
 92.74 
 
 (2) 
 90.88 
 
 (3) 
 91.51 
 
 (4) 
 
 82.57 
 
 (6) 
 71.13 
 
 6.41 
 
 4.85 
 
 4.93 
 
 4.71 
 
 4.29 
 
 1.54 
 
 
 3.29 
 0.27 
 
 10.74 
 1.98 
 
 21.30 
 3.20 
 
 100.69 
 
 100.00 100.00 100.00 
 
 In experiments (4) and (5) the substances analyzed repre- 
 sent the products of the action of acetylene (not specially 
 purified) on cupric oxide. 
 
 The oxygen present in these products is obviously pro- 
 portional to the amount of copper and is never more than 
 
ON THE OXIDES OF COPPER. 281 
 
 enough to be completely accounted for upon the supposition 
 that some of the original oxide taken stills holds its oxygen. 
 So far as the analyses show, the product of lightest color (1) 
 contains very little copper and no oxygen ; the darkest prod- 
 uct (3) obtained from the cuprous oxide contains oxygen 
 corresponding to a mixture of two parts of copper with three 
 parts of cuprous oxide; the oxygen in the products of (4) 
 and (5) obtained by acting upon cupric oxide is approximately 
 enough to correspond to a mixture of cuprous and cupric 
 oxides in equal proportions. This fact, taken in connection 
 with the great range of variation in proportion and the 
 minimum to which the copper falls in the product, which 
 would be least likely to include contaminating metal or 
 oxide, suggests very strongly the probability that the oxygen 
 present is in union with copper and that the copper is held 
 mechanically as metal or oxide and is not the essential 
 constituent of an organic compound. Leaving out of con- 
 sideration, therefore, the copper and copper oxides, and 
 calculating the composition of the products assumed to 
 consist essentially of carbon and hydrogen, we derive the 
 following statement: 
 
 (1) (2) (3) (4) (5) 
 
 Per cent of carbon 93.54 94.93 94.88 94.60 94.31 
 
 Per cent of hydrogen 6.46 5.07 5.12 5.40 5.69 
 
 100.00 100.00 100.00 100.00 100.00 
 
 These figures correspond to symbols varying from Ci 2 H 10 to 
 nearly Ci 6 Hi , with an average approximating CnHi , the 
 symbol of anthracene or paranthracene. The analytical data 
 of Erdmann and Kothner point in the average to a product 
 corresponding more nearly to the first of these symbols than 
 to either of the others. The product is doubtless variable 
 with the temperature and the activity of oxidation. Thus, 
 in one experiment in which acetylene was passed over ferric 
 oxide the action began at 365 with incandescence, as de- 
 scribed by Moissan and Moureu,* and the analysis of the 
 
 * Loc. cit 
 
282 ACTION OF ACETYLENE ON OXIDES OF COPPER. 
 
 product (carbon 91.53, hydrogen = 1.36, Fe = 5.85, O = 
 1.26) indicates a proportion of carbon to hydrogen about 
 four times as great as that of the average product of action 
 at 225 on the oxides of copper. 
 
 Finally, we find no evidence that the product of the action 
 of acetylene on the oxides of copper under the conditions of 
 our experimentation is other than a mixture of a hydro- 
 carbon or hydrocarbons with metallic copper or an oxide of 
 copper, and, probably, in the darker preparations, some free 
 carbon. 
 
XXXV 
 
 NOTES ON THE SPACE ISOMERISM OF THE 
 TOLUQUINONEOXIME ETHERS. 
 
 BY WM. CONGER MORGAN.* 
 
 IN an article on the "Ethers of Toluquinoneoxime and 
 their Bearing on the Space Isomerism of Nitrogen,"! pub- 
 lished from this laboratory, it was stated that the methyl, 
 acetyl, and benzoyl ethers of toluquinonemetaoxime, whether 
 formed by the action of hydroxylamine on the quinone or by 
 nitrous acid on the corresponding cresol, showed evidence of 
 existing in isomeric forms. Of these bodies the benzoyl ether 
 received the most careful investigation, and by fractional 
 crystallization, from the crude reaction-product, two portions 
 were obtained, one readily separating from an alcoholic solu- 
 tion in the form of yellow crystals melting at 193, while 
 a second body, melting approximately at 144, was never 
 obtained in a state of purity. The fact that from this 
 lower-melting fraction, on recrystallization, a portion of the 
 higher-melting body was always obtained, suggested the 
 possibility that the solvent might have a tendency to cause a 
 transition from one isomer to the other. Since, however, 
 after repeated crystallizations from boiling alcohol a low- 
 melting fraction was obtained, and therefore such rearrange- 
 ment was evidently incomplete, the action of alcohol under 
 pressure on the different fractions was investigated. From a 
 portion of the ether melting at 139, heated for three hours in 
 a closed tube to 120, the product melting at 193 crystallized 
 in a characteristic form and no crystals were obtained melting 
 
 * From Am. Chem. Jour., xxii, 402. 
 
 t Bridge and Morgan, Am. Chem. Jour., xx, 761. This volume, p. 145. 
 
284 NOTES ON THE SPACE ISOMERISM 
 
 lower than 180. On treating the higher-melting isomer in 
 the same way, no action was observed until the tempera- 
 ture was raised above 150, when complete decomposition 
 ensued. 
 
 In order to ascertain, if possible, a direct method of tran- 
 sition from one form into the other, the action of alkalies, 
 among other reagents, was tried. No change other than 
 a hydrolytic cleavage was observed, and this was readily 
 completed on warming. By saponifying a fraction melting 
 completely at 142 and treating the isolated oxime in the 
 usual way with benzoyl chloride, the benzoyl ether was again 
 obtained, liquefying at 193, and, although there was a 
 trifling irregularity in the melting-point of some of the crys- 
 tals, no low-melting fraction was isolated. On similar treat- 
 ment of a portion melting at 193 the original ether was 
 obtained, but none of the low-melting isomer. 
 
 From these facts it is obvious that the isomer existing 
 in a much smaller proportion in the crude reaction-products, 
 present as the principal constituent in the low-melting frac- 
 tions, must be regarded as the labile form, tending to go over 
 into the stable form under the influence of boiling alcohol 
 or during the process of a chemical reaction. This action 
 of alcohol increases the difficulty of isolating the labile form 
 and naturally suggests the idea of using other solvents ; but 
 imperfect as was the separation of the two isomers, better 
 results were obtained from an alcoholic solution than by 
 any other means. 
 
 The phenomena described in the article to which reference 
 was previously made, have been reproduced completely in the 
 ethers made by the action of acid chlorides on the sodium salt 
 of the oxime (to be described later) produced by the action 
 of pure amyl nitrite on the sodium salt of the cresol. In the 
 resulting metathesis, as hi the product of the reaction of 
 hydroxylamine on the quinones, there is not the possibility of 
 the formation of a nitro-body as there is in the action of free 
 nitrous acid on the cresols. Consequently the interfering 
 action of an admixture of such a body with the oxime ethers 
 
OF THE TOLUdUINONEOXIME ETHERS. 285 
 
 cannot influence the results as obtained, yet the presence of a 
 low-melting body existing hi a larger proportion in this product 
 than in the substances formed by the agency of free nitrous 
 acid according to the type fraction, was indisputably evidenced. 
 The identity of the observed phenomena under varying 
 conditions and methods of formation, thus removing the 
 probability of an admixture of an impurity and at the same 
 tune of a structural difference, seems to establish the hypothesis 
 of a space isomerism in the case of the ethers of toluquinone- 
 metaoxime. 
 
 Inasmuch as the monohalogen derivatives of the quinone- 
 oxime ethers are beautifully crystalline bodies, it was thought 
 advisable to prepare the monochlor- and monobrombenzoyl 
 ethers of toluquinoneorthooxime in the hope of obtaining 
 from these well-characterized products additional evidence as 
 to the existence of isomeric phenomena in the orthooxime 
 ethers. Although there was the possibility of both " space" 
 and " place " isomerism neither body offered any indication of 
 the presence of isomers of any kind, but each appeared to be 
 an entirely homogeneous and simple substance, liquefying 
 sharply at a definite melting-point. 
 
 EXPERIMENTAL PART. 
 The Sodium Salt of Toluquinone-m-oxime. 
 
 The sodium salt was prepared according to the general 
 method suggested by Walker.* To a molecule of sodium 
 alcoholate freshly prepared by dissolving metallic sodium in 
 as little alcohol as possible, a molecule of the orthocresol was 
 added and the solution treated with slightly more than the 
 theoretical quantity of amyl nitrite, the whole being thoroughly 
 mixed together. On standing in a vacuum over sulphuric 
 acid, the sodium salt separates in fine purple crystals, which, 
 on washing carefully with ether to remove amyl alcohol and 
 excess of nitrite, is ready for use. It may be recrystallized 
 from dilute alcohol if further purification is desired. This 
 
 * Walker, Ber. Dtsch. chem. Ges., xvii, 399. 
 
286 NOTES ON THE SPACE ISOMERISM 
 
 salt is extremely soluble in water, much less in alcohol, and 
 insoluble in most other organic liquids. On standing in the 
 air it tends to decompose, turning almost black. On analysis : 
 
 0.1231 gram, dried over H 2 S0 4 , gave 0.0541 gram Na 2 S0 4 . 
 
 Calculated for Fmind 
 
 C 7 He0 2 NNa. Found. 
 
 Na 14.48 14.26 
 
 Monobromtoluquinone-o-oxime Benzoyl Ether. 
 
 This ether was obtained from the dibromide previously 
 described* by boiling with dilute 75 per cent alcohol, during 
 which process hydrobromic acid is split off. Being much less 
 soluble in alcohol than the dibromide, the monobrom-body 
 separates from the solution as soon as formed and may be 
 obtained as a yellow crystalline powder by filtration of the 
 cooled liquid. On analysis, 
 
 0.1450 gram, dried over H 2 S0 4 , gave 0.2774 gram C0 2 , and 
 0.0415 gram H 2 0. 
 
 0.1102 gram gave 0.0643 gram AgBr. 
 
 Calculated for w , 
 
 C M H 10 BrN0 8 . 
 
 C 52.49 52.17 
 
 H 3.15 3.18 
 
 Br 24.98 24.83 
 
 On crystallizing from alcohol the ether was readily obtained 
 in two, apparently unlike, modifications, one being long 
 prismatic crystals, the other appearing as broad, thick, mono- 
 clinic plates. It was at first believed that this distinction in 
 crystal form was due to the presence of isomeric bodies, but 
 no difference in melting-point could be found since each 
 portion liquefied sharply at 184. Under the lens, moreover, 
 the plates are seen to be striated parallel to one edge and have 
 all the appearance of consisting of a number of the simple 
 crystals united to each other, since both forms are plainly of 
 
 * Bridge and Morgan, Am. Chem. Jour., xx, 776. 
 
OF THE TOLUQUINONEOXIME ETHERS. 287 
 
 the same system. Slow cooling was found to be productive 
 of the massed crystals. 
 
 Toluquinone-o-oxime Benzoyl Ether Dichloride. 
 
 The benzoyl ether was dissolved in a small amount of 
 chloroform and dry chlorine gas passed into the solution. 
 The action is very rapid and after fifteen minutes the liquid 
 may be allowed to evaporate and the white product, crystallized 
 once from glacial acetic acid, melts sharply to a colorless 
 liquid at 149 without decomposition. Fractional crystalliza- 
 tion does not change the melting-point, except as the action of 
 the solvent causes a slight formation of the monochlor-body. 
 On analysis, 
 
 0.1241 gram, dried over H 2 S0 4 , gave 0.2431 gram C0 2 , and 
 0.0424 gram H 2 O. 
 
 0.2105 gram gave 0.1930 gram AgCl. 
 
 Calculated for ,, . 
 
 Found - 
 
 C 53.84 53.42 
 
 H 3.55 3.80 
 
 Cl 22.72 22.67 
 
 The dichloride, like the dibromide, is slightly soluble in 
 alcohol, readily soluble in chloroform, glacial acetic acid, and 
 fuming nitric acid. Water precipitates the ether entirely 
 unchanged from the two last-mentioned solvents. In crystal 
 form it resembles very closely the dibromide, separating from 
 a boiling acetic acid solution in short, thick, colorless, almost 
 microscopic prisms, suggesting the orthorhombic system. In 
 point of stability this ether, as well as the monochlor-body, 
 far surpasses the corresponding bromine compounds. 
 
 Monochlortoluquinone-o-oxime Benzoyl Ether. 
 
 Analogous to the dibromide, by the action of dilute alcohol 
 on the dichloride, hydrochloric acid is split off and a yellow 
 monochlor-substitution-product is formed, one hour being 
 sufficient to complete the reaction. This ether closely resem- 
 
288 ISOMERISM OF TOLUQUINONEOXIME ETHERS. 
 
 bles the monobrom-derivative in properties and in crystal 
 form. The product as obtained from an alcoholic, solution 
 will decompose and melt at a temperature varying ordinarily 
 from 185-193, depending on the rapidity with which the 
 heat is applied. " Dipped " for fifteen seconds, the crystals 
 melt without decomposition at 200. Although carefully 
 fractioned from alcohol, ligroin, and benzol, the substance 
 appeared to be entirely homogeneous, and no variation in the 
 melting-point was observed. On analysis, 
 
 0.1209 gram, dried over H 2 S0 4 , gave 0.2688 gram C0 2 , and 
 0.0424 gram H 2 0. 
 
 0.1035 gram gave 0.0533 gram AgCl. 
 
 Calculated for -p rt ,,j 
 
 C M H 10 C1N0 8 . Found< 
 
 C 60.97 60.63 
 
 H 3.66 3.90 
 
 Cl 12.87 12.75 
 
XXXVI 
 
 ON THE VOLUMETRIC ESTIMATION OF 
 CERIUM. 
 
 BY PHILIP E. BROWNING.* 
 
 SOME forty years ago Bunsenf showed that the oxide 
 obtained by the ignition of cerium oxalate might be estimated 
 volumetrically by bringing it in contact with potassium iodide 
 and strong hydrochloric acid and determining the iodine set 
 free. This method may be briefly described by a translation 
 of part of the original article : " The substance to be deter- 
 mined is weighed out in a glass flask of from ten to fifteen 
 cubic centimeters capacity, a few crystals of potassium iodide 
 are added, and the neck of the flask is drawn out by the aid of 
 a blowpipe to a narrow opening. The flask is filled almost to 
 the narrowing of the neck with hydrochloric acid which is free 
 from chlorine or iron chloride, and a little sodium carbonate is 
 added in order to displace the last trace of air by carbon 
 dioxide. The flask is then closed by sealing off the neck in 
 the blowpipe and warmed in a water bath until the cerium 
 compound is completely dissolved, and the quantity of iodine 
 set free is determined by iodometric analysis." 
 
 The anhydrous dioxide prepared by the ignition of the 
 oxalate or hydroxide is very slowly acted on by acids, 
 especially when pure.f For this reason the method which 
 Bunsen described has remained the only one adapted to the 
 satisfactory volumetric estimation of the ignited dioxide. 
 
 Two portions of the dioxide were prepared by treating the 
 crude cerium chloride in concentrated solution with gaseous 
 
 * From Am. Jour. Sci., viii, 451. 
 
 t Ann. Chem. Phar., cv, 49. 
 
 $ Rose, Handbuch der analytischen Chemie, Band i, 219. 
 
 VOL. II. 19 
 
290 VOLUMETRIC ESTIMATION OF CERIUM. 
 
 hydrochloric acid * to saturation to remove the iron. The 
 cerium chloride was then dissolved hi water, potassium 
 hydroxide added in excess and chlorine gas passed until the 
 precipitate became distinctly orange in color and the solution 
 gave a strong odor of chlorine.! This operation was repeated 
 until a portion of the precipitate dissolved in acid showed 
 no didymium absorption bands when examined before the 
 spectroscope. The whole precipitate of the dioxide was then 
 dissolved in hydrochloric acid and the oxalate precipitated by 
 ammonium oxalate in large excess. The precipitated oxalate 
 was then washed thoroughly with hot water until the washings 
 gave no test for hydrochloric or oxalic acids and ignited to the 
 dioxide. Another portion of the dioxide was later prepared 
 by precipitating a solution of pure cerium chloride by means 
 of ammonium oxalate, washing and igniting as described. 
 The dioxide in all three cases was of a light chamois color, 
 and uniform results were obtained from the three portions. 
 
 A modification of the method of Bunsen (with G. A. 
 FOKD and F. J. HALL). 
 
 Weighed portions of the pure cerium dioxide were placed in 
 small glass stoppered bottles of about 100 cm 3 capacity, together 
 with a gram of potassium iodide free from iodate and a few 
 drops of water to dissolve the iodide. A current of carbon 
 dioxide was passed into the bottle for about five minutes to 
 expel the air, 10 cm 3 of pure strong hydrochloric acid were 
 added, the stopper inserted and the bottle heated gently upon 
 a steam radiator for about one hour until the dioxide dissolved 
 completely and the iodine was set free. After cooling the 
 bottle, to prevent loss of iodine upon removing the stopper, 
 the contents were carefully washed into about 400 cm 3 of 
 water and titrated with standard sodium thiosulphate to 
 determine the amount of iodine liberated according to the well 
 known reaction 
 
 2Ce0 2 + 8HC1 + 2KI = 2CeCl s + 2KC1 + 4H 2 + I 2 . 
 
 * Dennis and Magee, Zeitschr. anorg. Chem., iii, 260. 
 
 t Mosander, Phil. Mag., xxviii, 241 ; Dennis, Zeitschr. anorg. Chem., yii, 252. 
 
VOLUMETRIC ESTIMATION OF CERIUM. 
 
 291 
 
 A few blank determinations were carried through in the bottles 
 without the presence of the cerium dioxide to determine the 
 amount of iodine set free under these conditions. The amount 
 obtained was uniformly equal to 0.04 cm 3 of the ^ iodine 
 solution which was taken as the correction and applied to all 
 the determinations. The results follow in Table I. 
 
 TABLE I. 
 
 Exp. 
 
 CeO, taken. 
 
 CeOj found. 
 
 Error. 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.1000 
 
 0.0994 
 
 0.0006- 
 
 (2) 
 
 0.1032 
 
 0.1034 
 
 0.0002+ 
 
 (3) 
 
 0.1016 
 
 0.1017 
 
 0.0001+ 
 
 (4) 
 
 0.1054 
 
 0.1041 
 
 0.0013- 
 
 (5) 
 
 0.2010 
 
 0.2021 
 
 0.0011+ 
 
 (6) 
 
 0.1104 
 
 0.1109 
 
 0,0005+ 
 
 (7) 
 
 0.1914 
 
 0.1907 
 
 0.0007- 
 
 (8) 
 
 0.1604 
 
 0.1603 
 
 0.0001- 
 
 (9) 
 
 0.2146 
 
 0.2145 
 
 0.0001- 
 
 (10) 
 
 0.1108 
 
 0.1099 
 
 0.0009- 
 
 (11) 
 
 0.1346 
 
 0.1347 
 
 0.0001+ 
 
 (12) 
 
 0.1540 
 
 0.1534 
 
 0.0006- 
 
 13 
 
 0.1976 
 
 0.1968 
 
 0.0008- 
 
 (14) 
 
 0.1230 
 
 0.1240 
 
 0.0010+ 
 
 (15) 
 
 0.1199 
 
 0.1202 
 
 0.0003+ 
 
 (16) 
 
 0.1524 
 
 0.1528 
 
 0.0004+ 
 
 (17) 
 
 0.1212 
 
 0.1211 
 
 0.0001- 
 
 (18) 
 
 0.1528 
 
 0.1543 
 
 0.0015+ 
 
 In order to obtain a further check upon the accuracy of the 
 method, portions of the cerium dioxide were weighed out and 
 placed in a distillation apparatus previously employed for 
 similar purposes and described in former articles from this 
 laboratory, viz.: a Voit flask, serving as a retort, sealed to 
 the inlet tube of a Drexel wash-bottle, used as a receiver, 
 the outlet tube of which was trapped by sealing on Will and 
 Varrentrapp absorption bulbs. In the retort the cerium 
 dioxide together with 15 cm 3 of water, 1 grm. of potassium 
 iodide and 10 cm 3 of pure strong hydrochloric acid were placed. 
 In the receiver were 100 cm 3 of water and 2 to 3 grm. of 
 potassium iodide, and in the bulbs a dilute solution of potas- 
 sium iodide. Before adding the hydrochloric acid a current of 
 carbon dioxide was passed through the apparatus for some 
 
292 
 
 VOLUMETRIC ESTIMATION OF CERIUM. 
 
 minutes. After adding the acid, the liquid was boiled in the 
 current of carbon dioxide * to a volume of 15 cm 3 , when the 
 free iodine had almost completely left the retort and passed 
 into the receiver, and the apparatus was allowed to cool. 
 
 The iodine in the receiver was titrated directly with sodium 
 thiosulphate, and that in the retort after dilution of the 
 residue to about 400 cm 3 , the later amount seldom exceeding 
 the equivalent of a few drops of -^ iodine solution. 
 
 The results f oUow in Table II. 
 
 Here also blank determinations were made but no correction 
 was found to be necessary. 
 
 An attempt early in the work to titrate by an alkaline arse- 
 nite the iodine liberated, after neutralizing the hydrochloric 
 acid, brought out some curious results which seem worthy of 
 mention. 
 
 TABLE H. 
 
 Exp. 
 
 CeO, taken. 
 
 CeOj found. 
 
 Error. 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 
 1) 
 
 0.1028 
 
 0.1013 
 
 0.0015- 
 
 
 2) 
 
 0.2060 
 
 0.2055 
 
 0.0005- 
 
 
 3) 
 
 0.2014 
 
 0.2012 
 
 0.0002- 
 
 
 4 
 
 0.1716 
 
 0.1711 
 
 0.0005- 
 
 5) 
 
 0.0974 
 
 0.0972 
 
 0.0002- 
 
 (6) 
 
 0.1600 
 
 0.1587 
 
 0.0013- 
 
 (7) 
 
 0.1268 
 
 0.1254 
 
 0.0014- 
 
 . 
 
 8) 
 
 0.1276 
 
 0.1268 
 
 0.0008- 
 
 
 9 
 
 0.1620 
 
 0.1612 
 
 0.0008- 
 
 (10) 
 
 0.1016 
 
 0.1011 
 
 0.0005- 
 
 (H) 
 
 0.1648 
 
 0.1543 
 
 0.0005- 
 
 (12) 
 
 0.1352 
 
 0.1342 
 
 0.0010- 
 
 In these experiments the contents of the bottles after the 
 cerium had dissolved were carefully washed into a Drexel 
 wash bottle upon the inlet tube of which was fused a thistle 
 tube with a stop-cock and to the outlet tube a Will and Var- 
 rentrapp absorption trap. In the trap a solution of potassium 
 iodide was placed and through the thistle tube a saturated 
 
 * The carbon dioxide gas was furnished by a Kipp generator from marble 
 and hydrochloric acid of one-half strength, both of which had been boiled 
 previously to remove all air. 
 
VOLUMETRIC ESTIMATION OF CERIUM. 
 
 293 
 
 solution of potassium bicarbonate was added to complete neu- 
 tralization of the acid. Any iodine carried mechanically by 
 the carbon dioxide should be held by the potassium iodide 
 solution in the trap. After neutralization the free iodine was 
 titrated by standard arsenious oxide solution. The results 
 appear in Table III. 
 
 TABLE IIL 
 
 Exp. 
 
 CeO 2 taken. 
 
 CeO, found. 
 
 Error. 
 
 
 grin. 
 
 gnn. 
 
 gnn. 
 
 (1) 
 
 0.1000 
 
 0.0987 
 
 0.0013-] 
 
 2 
 (3) 
 
 0.1005 
 0.1030 
 
 0.0981 
 0.1009 
 
 0.0024-1 T 
 0.0021- [* 
 
 (4) 
 
 0.1500 
 
 0.1475 
 
 0.0025- 1 
 
 (6) 
 
 0.1030 
 
 0.1005 
 
 0.0025-' 
 
 
 (6 
 
 0.1010 
 
 0.0988 
 
 0.0022- 
 
 
 7 
 
 0.1510 
 
 0.1608 
 
 0.0002- 
 
 TT 
 
 (8) 
 
 0.1530 
 
 0.1485 
 
 0.0045- 
 
 11. 
 
 (9) 
 
 0.2045 
 
 0.2011 
 
 0.0034- 
 
 
 (10) 
 
 0.2000 
 
 0.1958 
 
 0.0042- 
 
 
 (11) 
 
 0.1334 
 
 0.1302 
 
 0.0032- 
 
 
 (12) 
 
 0.1354 
 
 0.1330 
 
 0.0024- 
 
 
 (13) 
 
 0.1312 
 
 0.1294 
 
 0.0018- 
 
 
 (14) 
 (15) 
 
 0.1308 
 0.1060 
 
 0.1277 
 0.1042 
 
 0.0031- 
 0.0018- 
 
 m. 
 
 (16) 
 
 0.1602 
 
 0.1567 
 
 0.0035- 
 
 
 (17) 
 
 0.1504 
 
 0.1488 
 
 0.0016- 
 
 
 As will be seen by the table, an average error of about 2 per 
 cent runs through the entire set. The natural conclusion 
 would be that the cerium dioxide contained some impurity; 
 but, as the first, second, and third samples, very carefully pre- 
 pared, gave the same results, it seemed necessary to look else- 
 where for an explanation. Two possible causes suggested 
 themselves : first, mechanical loss during the process of neu- 
 tralization, and second, the possible formation, under the condi- 
 tions, of iodine chloride, which if formed would in the process 
 of neutralization probably take the form of potassium chloride, 
 iodide, and iodate, and thus some of the originally free iodine 
 would be withdrawn from the amount titrated. To test these 
 theories, portions of the ^ iodine solution roughly equiva- 
 lent to the amounts of iodine set free by 0.1 and 0.2 grms. of 
 CeO a , were drawn off into bottles previously rilled with carbon 
 
294 
 
 VOLUMETRIC ESTIMATION OF CERIUM. 
 
 dioxide, treated with the usual amount of strong hydrochloric 
 acid (10 cm 3 ), and after standing from thirty to forty-five 
 minutes, neutralized and titrated as already described. The 
 results were most interesting and seemed to show a loss of 
 iodine closely equivalent to that shown by the results of 
 Table III, and proportional to the amount of iodine originally 
 present. A few determinations were carried through in the 
 same way except that the neutralization was omitted and 
 dilution and titration with thiosulphate substituted. These 
 showed a loss of iodine well within the limits of such a proc- 
 ess. The results follow in Table IV. 
 
 TABLE IV. 
 WITH ARSENIOUS OXIDE. 
 
 Exp. 
 
 Iodine *>. 
 taken. 
 
 Iodine , 
 found. 
 
 Error. 
 
 Equivalent error 
 on Ce0 2 . 
 
 (1) 
 (2) 
 (3) 
 (4) 
 5) 
 (6) 
 (7) 
 
 cm 3 
 
 5.22 
 6.09 
 5.10 
 6.66 
 5.10 
 10.22 
 10.21 
 
 cms 
 6.07 
 
 4.97 
 4.97 
 6.47 
 4.97 
 9.97 
 9.97 
 
 cm 8 
 0.15- 
 0.12- 
 0.13- 
 0.19- 
 0.13- 
 0.25- 
 0.24- 
 
 grm. 
 
 0.0026- 
 0.0021- 
 0.0022- 
 0.0033- 
 0.0022- 
 0.0043- 
 0.0041- 
 
 WITH SODIUM THIOSULPHATE. 
 
 g| 
 
 i 
 
 5 
 5 
 10 
 10.08 
 
 5.01 
 4.99 
 10.02 
 10.12 
 
 0.01+ 
 0.01- 
 0.02+ 
 0.04+ 
 
 0.0002+ 
 0.0002- 
 0.0003+ 
 0.0007+ 
 
 The action of arsenious oxide upon Cerium Dioxide (with 
 WM. D. GUTTER). 
 
 The fact that cerium dioxide is reduced by hydriodic acid 
 suggested the possibility of the application of arsenious acid 
 in acid solution to the same end according to the reaction 
 
 4Ce0 2 + As 2 8 = 2Ce 2 3 -f As 2 6 . 
 
 The extreme difficulty with which the ignited cerium dioxide 
 when pure dissolves in acids has already been mentioned, and 
 for this reason it was found practically impossible to obtain 
 
VOLUMETRIC ESTIMATION OF CERIUM. 
 
 295 
 
 any results by this method. Weighed portions of the dioxide 
 were placed in Erlenmeyer beakers with an excess of a solu- 
 tion of arsenious oxide , 10 cm 3 of (1 : 1) sulphuric acid 
 were added, and the boiling continued until the fuming point 
 of the acid was reached ; but even at this point only a partial 
 solution of the dioxide had taken place. 
 
 The dark brown powder obtained by igniting the carefully 
 washed oxalates, precipitated in acid solution by treating a 
 solution of crude cerium chloride with ammonium oxalate or 
 oxalate acid is very fairly soluble in acids. Mengel* has 
 recently shown that this product contains a dioxide of praseo- 
 didymium which acts as does cerium dioxide toward reducing 
 agents. This fact makes the results recorded in the treatment 
 of this ignited mixture of oxides of no value analytically, but 
 of interest in the comparative study of the two reducing 
 agents, arsenious oxide and hydriodic acid. Two portions of 
 this mixture of oxides gave the following results, which agree 
 fairly well with those of Mengel. 
 
 Ezp. 
 
 Amount of 
 substance taken. 
 
 Ce0 2f -HPr0 2 ?) 
 
 Calculated on 
 0.1000 grm. 
 
 8 
 
 grm. 
 0.1037 
 0.1034 
 
 grm. 
 
 0.0530 
 0.0538 
 
 gnu. 
 0.0511 
 0.0520 
 
 The average of these results was taken as a standard 
 0.0515 grm. CeO 2 , etc., to every 0.1000 grm. of material. 
 Three carefully weighed portions of this same material were 
 placed in Erlenmeyer beakers with 10 cm 3 of arsenious 
 oxide solution and 10 cm 3 of dilute (1 : 4) sulphuric acid and 
 boiled until complete solution had taken place. The liquid 
 was then cooled, neutralized with potassium bicarbonate and 
 titrated with standardized iodine to determine the amount of 
 arsenious oxide remaining, and from it the amount used in the 
 reduction of the dioxide according to the reaction given above. 
 The results obtained follow. 
 
 * Zeitschr. anorg. Chem., xix, 67. 
 
296 
 
 VOLUMETRIC ESTIMATION OF CERIUM. 
 
 Bxp. 
 
 Amount) 
 taken. 
 
 Amount CeO s 
 
 found. 
 
 CeO, calculated 
 for 0.1000 grm. 
 
 
 gnu. 
 
 grin. 
 
 grm. 
 
 (1) 
 
 0.1005 
 
 0.0493 
 
 0.0491 
 
 (2) 
 (3) 
 
 0.1016 
 0.1005 
 
 0.0494 
 0.0486 
 
 0.0487 
 0.0484 
 
 As will be seen, the results obtained by this method fall 
 about 0.0030 grm. below the standard as obtained by the dis- 
 tillation method, which seems to show that the arsenious oxide 
 does not effect the complete reduction of the cerium dioxide 
 from CeO 2 to Ce 2 O 8 . 
 
 In order to study this point a little more fully and upon the 
 pure dioxide, definite portions of a standard solution of pure 
 cerium chloride were precipitated by ammonia in the presence 
 of hydrogen dioxide and boiled to reduce the CeO 8 formed to 
 the conditions of CeO 2 . The precipitated hydrated dioxide 
 was filtered off and carefully washed until the washings gave 
 no indication of hydrogen dioxide. The moist precipitate was 
 then washed into a beaker, one gram of potassium iodide added 
 and 10 cm 3 of strong HC1. The precipitate dissolved quite 
 readily in the cold and the iodine liberated was determined by 
 standard sodium thiosulphate. The results appear in Table V. 
 
 TABLE V. 
 
 Bxp. 
 
 CeO 2 taken. 
 
 CeO 2 found. 
 
 Error. 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 
 1) 
 
 0.1142 
 
 0.1140 
 
 0.0002- 
 
 
 2) 
 
 0.1142 
 
 0.1147 
 
 0.0005+ 
 
 
 i 
 
 0.1142 
 0.1142 
 
 0.1152 
 0.1159 
 
 0.0010+ 
 0.0017+ 
 
 (5) 
 
 0.1142 
 
 0.1152 
 
 0.0010+ 
 
 
 0.1142 
 
 0.1156 
 
 0.0014+ 
 
 Another series of these precipitates prepared in the same 
 way was boiled with a definite amount of arsenious acid in 
 acid solution, as previously described in the case of the 
 ignited dioxide. The results which are recorded in Table VI 
 
VOLUMETRIC ESTIMATION OF CERIUM. 
 
 297 
 
 show, as in the case of the ignited dioxide, an insufficient 
 reduction of the cerium by the arsenious acid. 
 
 TABLE VI. 
 
 Exp. 
 
 CeO 2 taken. 
 
 Ce0 8 found. 
 
 Error. 
 
 
 gnn. 
 
 grin. 
 
 grin. 
 
 (1) 
 
 0.0881 
 
 0.0370 
 
 0.0011- 
 
 2) 
 
 0.0381 
 
 0.0361 
 
 0.0020- 
 
 (3) 
 
 0.1142 
 
 0.1077 
 
 0.0064- 
 
 (4) 
 
 0.1060 
 
 0.1002 
 
 0.0058- 
 
 The Estimation of Cerium Oxalate ly Potassium Perman- 
 ganate (with LEO A. LYNCH). 
 
 Stolba* has stated that cerium oxalate may be estimated 
 volumetrically after the same manner as calcium oxalate by 
 treating the washed precipitate, suspended in warm water, to 
 which a moderate amount of sulphuric acid has been added, 
 by potassium permanganate. As the titration proceeds the 
 precipitate disappears and the end reaction is sharp. He also 
 finds that the permanganate does not oxidize the cerium 
 from the lower to the higher condition. So far as we have 
 been able to discover, no experimental evidence has been 
 presented to prove the correctness of Stolba's statement, and 
 the work to be described was undertaken to furnish such 
 evidence. 
 
 The solutions used were prepared and standardized as 
 follows : The cerium solutions were made by dissolving 
 10 grams of pure cerium chloride in one liter of water, and 
 standardized by precipitating measured and weighed portions, 
 in a faintly acid solution, with ammonium oxalate, filtering, 
 washing, igniting, and weighing as the dioxide (CeO 2 ). A 
 solution of potassium permanganate was prepared and stan- 
 dardized by titration against weighed amounts of ammonium 
 oxalate. A solution of ammonium oxalate was made and 
 its value determined by titrating measured amounts against 
 
 * Sitzungsber. d. kgl. bohm. Gesellsck. d. Wissenschaften v. 4. Juli, 1879; 
 Zeitschr. anal. Chem., xix, 194. 
 
298 VOLUMETRIC ESTIMATION OF CERIUM. 
 
 potassium permanganate. Definite portions of the cerium 
 solution were drawn from a burette and after diluting with 
 water from 100 to 200 cm 3 a definite amount of ammonium 
 oxalate was added, care being taken to have an excess over 
 the amount necessary, and the whole warmed to insure a 
 more crystalline precipitation.* The precipitate was then 
 filtered off on paper and carefully washed, the filtrate and 
 washings being collected in a liter Erlenmeyer flask and set 
 aside for future use. The precipitate was treated with about 
 10 cm 3 of hot (1 : 4) sulphuric acid, which dissolved it 
 completely, if not at first, by running it through the filter 
 a few times, and the solution and washings were collected hi 
 another liter flask. The total volume of liquid was made 
 up to about 500 cm 3 , warmed to about 70 C. to 80 C. and 
 titrated with potassium permanganate to the appearance of 
 the faint blush of color showing the complete oxidation of the 
 oxalic acid. The filtrate from the cerium oxalate containing 
 the excess of oxalic acid was diluted to 500 cm 3 , acidified 
 with 10 cm 3 of dilute (1 : 4) sulphuric acid, one gram of 
 manganous sulphate added to prevent the interfering action 
 of the free hydrochloric acid upon the estimation of the oxalic 
 acid,f and titrated with potassium permanganate after the 
 same manner as the dissolved precipitate. A definite quantity 
 of ammonium oxalate having been originally taken, it became 
 possible, by subtracting from it the amount obtained, to 
 derive the measure of the oxalate used in the precipitation 
 of the cerium oxalate. By this procedure, it will be observed 
 a check was made upon the results obtained by the titration 
 of the precipitate. In experiments (1) to (6) the cerium 
 oxalate was thrown down in neutral solution, in experiments 
 (7) to (10) in acid solutions. The treatment of the filtrate 
 in experiment (1) was made without the presence of the 
 manganous sulphate. The results recorded in Table VII 
 seem to uphold the statement of Stolba. 
 
 * As shown by the table, the precipitation was sometimes in neutral, some- 
 times in faintly acid solution. 
 
 t Gooch and Peters, Am. Jour. Sci., vii, 461. This volume, p. 222. 
 
VOLUMETRIC ESTIMATION OF CERIUM. 299 
 TABLE VII. 
 
 Bxp. 
 
 
 Amount found. 
 
 
 Amount found. 
 
 
 Amount taken. 
 
 Calculated as 
 
 Error. 
 
 Calculated as 
 
 Error. 
 
 Calculated 
 
 CeCl 3 . 
 
 Calculated as 
 
 CeCl 8 . 
 
 Calculated as 
 
 as CeCl 3 . 
 
 Treatment of 
 
 CeCl 3 . 
 
 Treatment of 
 
 CeCl 8 . 
 
 
 precipitate. 
 
 
 filtrate. 
 
 
 
 
 grm. 
 
 grm. 
 
 grlu. 
 
 grm. 
 
 grm. 
 
 (1 
 
 
 [0.1091 
 
 0.1087 
 
 0.0004- 
 
 0.1023 
 
 0.0068-] 
 
 (2 
 
 
 0.1091 
 
 0.1103 
 
 0.0012+ 
 
 . 
 
 g 
 
 3 
 
 
 0.1091 
 
 0.1087 
 
 0.0004- 
 
 0.1087 
 
 0.0004- 
 
 4 
 
 
 0.1364 
 
 0.1373 
 
 0.0009+ 
 
 0.1391 
 
 0.0027+ 
 
 1: 
 
 0.1364 
 0.2182 
 
 0.1367 
 0.2202 
 
 0.0003+ 
 0.0020+ 
 
 0.1367 
 0.2206 
 
 0.0003+ 
 0.0024+ 
 
 (7) 
 
 0.1091 
 
 
 
 0.1087 
 
 0.0004- 
 
 (8) 
 
 0.1519 
 
 o.is35 
 
 0.0016+ 
 
 0.1535 
 
 0.0016+ 
 
 (9) 
 
 0.1364 
 
 0.1367 
 
 0.0003+ 
 
 0.1367 
 
 0.0003+ 
 
 (10) 
 
 
 0.2182 
 
 0.2183 
 
 0.0001+ 
 
 0.2183 
 
 0.0000 
 
XXXVII 
 
 ON THE ESTIMATION OF THALLIUM AS THE 
 CHROMATE. 
 
 BY PHILIP E. BROWNING AND GEORGE P. HUTCHINS.* 
 
 CROOKES has shown f that the chromate precipitated by the 
 addition of potassium dichromate to an alkaline solution of a 
 thallous salt has the constitution of a neutral salt and is very 
 insoluble in water 100 parts of water at 100 C. dissolving 
 about 0.2 parts and at 60 C. about 0.03 parts. He has also 
 made use of this reaction J to effect a rough separation of 
 thallium from cadmium. 
 
 The object of this paper is to describe some work directed 
 toward a study of the application of this reaction to the gravi- 
 metric estimation of thallium and the best conditions under 
 which to effect the precipitation. For the work a solution 
 of thallous nitrate was made by dissolving 10 grms. in water 
 and making up to a liter. The standard was determined 
 by taking measured and weighed portions from a burette, 
 precipitating with a slight excess of potassium iodide, agitat- 
 ing to bring about a good separation of the thallous iodide, 
 and allowing to stand until the supernatant liquid was 
 clear. The iodide was then filtered off upon an asbestos felt 
 contained in a perforated platinum crucible, the whole having 
 been previously ignited and weighed, washed with a mixture 
 of alcohol and water, dried over a low flame and weighed 
 to a constant weight. The filtrate, which together with 
 the washings seldom amounted to more than 50 cm 3 , was 
 evaporated to dryness on a water bath, a few drops of water 
 
 * From Am. Jour. Sci., viii, 460. t Chem. News, viii, 255. 
 
 J Chem. News, vii, 145. 
 
ESTIMATION OF THALLIUM AS CHEOMATE. 301 
 
 added and thus the small amount of thallous iodide which 
 had been dissolved recovered. This small insoluble residue, 
 which seldom amounted to one milligram in weight, was 
 filtered off, washed and weighed as previously described. 
 Baubigny * has shown this method to give very satisfactory 
 results, and the uniformity of our determinations certainly 
 confirms his statements. 
 
 For convenience in the calculations of results to be de- 
 scribed later, a solution of potassium dichromate of definite 
 strength was made. Portions of the thallium solution were 
 drawn from a burette into test tubes of about 100 cm 3 capacity 
 and weighed as a check on the burette reading. The solution 
 was heated to about 70 C. to 80 C. and a few drops of 
 ammonia or potassium carbonate solution added to distinct 
 alkalinity. A definite amount of the potassium dichromate 
 in solution was delivered from a burette, care being taken to 
 have an excess, and the contents agitated to bring about a 
 good separation of the precipitated chromate. After the 
 precipitate had completely settled out and the solution had 
 become cold the chromate was filtered upon asbestos, as 
 described above, dried over a low flame and weighed to a 
 constant weight. The filtrates from several determinations 
 were evaporated to a small volume and in one or two cases 
 a residue amounting to a few tenths of a milligram was 
 obtained, but no appreciable quantity of dissolved chromate 
 was thus recovered. It was found that when the precipita- 
 tion was made in the cold the chromate did not flock well, 
 but remained partly in a finely divided condition which would 
 run through the felt and require repeated filtration. The 
 addition of ammonium nitrate before precipitation prevented 
 this largely, even in the cold, bat the best results were 
 obtained by warming the solution before precipitation and 
 using potassium carbonate rather than ammonium hydroxide. 
 The results follow in Table I. 
 
 An attempt was made to estimate the thallium volumetri- 
 cally by determining the amount of chromate in the filtrate 
 * Chem. News, Ixiv, 239. 
 
302 ESTIMATION OF THALLIUM AS CHROMATE. 
 
 TABLE I. 
 
 Brp. 
 
 TINOs taken. 
 Calculated as 
 11,0. 
 
 TliCrO 4 found. 
 Calculated as 
 Tl a O. 
 
 Error. 
 Calculated as 
 T1 2 0. 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.0796 
 
 0.0791 
 
 0.0005- 
 
 (2) 
 
 0.0792 
 
 0.0788 
 
 0.0004- 
 
 (3) 
 (4) 
 
 0.0792 
 0.1188 
 
 0.0786 
 0.1177 
 
 0.0006- 
 0.0011- 
 
 (5) 
 
 0.1192 
 
 0.1186 
 
 0.0006- 
 
 (6) 
 
 0.1185 
 
 0.1178 
 
 0.0007- 
 
 (7) 
 
 0.1190 
 
 0.1185 
 
 0.0005- 
 
 (8) 
 
 0.1189 
 
 0.1183 
 
 0.0006- 
 
 (9) 
 
 0.1196 
 
 0.2000 
 
 0.0004+ 
 
 (10) 
 
 0.1196 
 
 0.2005 
 
 0.0009+ 
 
 (ID 
 
 0.1173 
 
 0.1173 
 
 0.0000 
 
 (12) 
 
 0.1171 
 
 0.1163 
 
 0.0008- 
 
 from the thallous chromate, and by difference (the potassium 
 dichromate originally added being known) the amount 
 combined with the thallium in the precipitate. The method 
 used to determine the standard of the dichromate solutions 
 and also the chromate remaining in the filtrate was described 
 by one of us in a previous paper from this laboratory.* 
 According to this procedure the filtrate from the thallous 
 chromate containing the excess of alkali chromate was acidified 
 with sulphuric acid, a definite amount of a solution of arsenious 
 oxide, previously standardized, was added and the whole was 
 allowed to stand a few moments until the change from the 
 yellow to the bluish green showed the complete reduction of 
 the chromic acid. Potassium bicarbonate was added to distinct 
 alkaline reaction and the arsenious oxide remaining was 
 determined by titration with -standard iodine solution. The 
 amount of the arsenious oxide oxidized is of course the 
 measure of the chromate in the solution. The amount of 
 chromate in the original solution used being known, by 
 subtracting the amount thus determined in the filtrate the 
 chromate in combination with the thallium may be readily 
 found, and from it the thallium estimated. Filtrates from 
 certain precipitates, of which the determinations are given in 
 
 * Am. Jour. Sci., i, 35, 1896. Volume I, p. 344. 
 
ESTIMATION OF THALLIUM AS CHROMATE. 303 
 
 Table I, were treated in this way, and the results, indicated by 
 corresponding numbers, follow in Table II. 
 
 TABLE II. 
 
 Bxp. 
 
 TINO 3 taken. 
 Calculated as 
 
 Tl 8 CrO 4 found. 
 Calculated as 
 
 Error. 
 Calculated as 
 
 
 TlaO. 
 
 Tl,0. 
 
 TljO. 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (5) 
 
 0.1192 
 
 0.1198 
 
 0.0006+ 
 
 (8) 
 
 0.1189 
 
 0.1205 
 
 0.0016+ 
 
 (9) 
 
 0.1196 
 
 0.1180 
 
 0.0016- 
 
 (10) 
 
 0.1196 
 
 0.1192 
 
 0.0004- 
 
 (11) 
 
 0.1173 
 
 0.1182 
 
 0.0009+ 
 
 (12) 
 
 0.1171 
 
 0.1190 
 
 0.0019+ 
 
 The method cannot be very accurate on account of the high 
 molecular weight of thallium oxide as compared with that of 
 the chromic acid determined, but the results check fairly well 
 with the gravimetric method. 
 
xxxvm 
 
 THE ETHERS OP ISONITROSOGUAIACOL IN 
 
 THEIR RELATION TO THE SPACE 
 
 ISOMERISM OF NITROGEN. 
 
 BY JOHN L. BRIDGE AND WM. CONGEE MORGAN* 
 
 WHEN the presence of isomerism in the quinoneoximes was first 
 noted by one of us,f and the phenomenon exhibited by the 
 two isomeric ethers then described was shown by Kehrmann J 
 to be due to no structural differences, but to necessitate the 
 assumption of a spatial arrangement about the nitrogen atom, 
 according to the theory of Hantzsch and Werner, the plan was 
 adopted of studying the various substituted quinoneoximes, 
 by means of their acyl and alkyl ethers, with reference to 
 this phenomenon. Accordingly we have investigated the 
 toluquinoneoximes, both ortho and meta, producing them 
 by the action of nitrous acid on the cresol as well as by 
 hydroxylamine on toluquinone, and found that, whereas there 
 is abundant evidence for the existence of stereoisomeric bodies 
 in the metaoxime ethers, in the derivatives of the orthooxime, 
 all such indication is wanting. The significance of this 
 observation is furthermore increased by the fact that all 
 oximes, in which isomerism has been reported, may be 
 considered as derivatives of metasubstituted quinones, and it 
 seemed not improbable, therefore, that these observations 
 might be formulated into a general rule regarding the 
 appearance of isomerism in the quinoneoximes. 
 
 In the course of his investigation of the properties and 
 reactions of isonitrosoguaiacol, among other derivatives, Pfob || 
 
 * From Am. Chem. Jour., xxii, 485. t Ann. Chem. (Liebig), cclxxvii, 79. 
 
 t Ann. Chem. (Liebig), cclxxix, 27. 
 
 Am. Chem. Jour., xx, 761 ; xxii, 402. This volume, pp. 145, 283. 
 
 II Monatsh. Chem., xyiii, 467. 
 
THE ETHERS OF ISONITROSOGUAIACOL 305 
 
 made the methyl and acetyl ethers, but did not announce the 
 observation of any cases of isomerism. Isonitrosoguaiacol may 
 be considered as the metamethoxyquinoneoxime, hence this 
 metasubstituted quinoneoxime presented conditions differing 
 widely from the other members of the same series above 
 mentioned. Moreover, because of its close relationship to 
 toluquinonemetaoxime, from which it differs only by the 
 interposition of an oxygen atom between the ring and the 
 methyl group, it seemed possible that isomeric modifications 
 of the ethers might exist, which had been overlooked by the 
 former investigator. When, furthermore, Rupe * made no 
 mention of such appearance in his research on isonitrosoguaia- 
 col, it seemed advisable to undertake anew the investigation of 
 this body with the special purpose of discovering such 
 isomerism, if possible, and to couple with it an investigation 
 of the orthomethoxyquinoneoxune or isonitroso derivative of 
 the monomethyl ether of resorcin, in order by this means to 
 be able to parallel in these closely analogous bodies the 
 experiments with the ortho- and metacresols. 
 
 With this idea, the work of Pfob and Rupe was carefully 
 repeated, so far as it pertained to the question in hand, 
 but, aside from minor differences, our results served only to 
 corroborate the testimony of these investigators. New 
 derivatives, to be described later, were prepared in the hope 
 that these bodies might show some variations leading to the 
 discovery of isomeric modifications, but each appeared to be 
 entirely homogeneous and no evidence for isomerism could be 
 found. These results are of course only negative and do not 
 disprove the existence of space isomers in the same bodies, 
 yet the same methods, which gave very positive evidence of 
 their presence hi other quinoneoxime ethers, were used to 
 detect them in this instance. 
 
 Aside from the difficulty of obtaining the pure monomethyl 
 ether of resorcin in any considerable quantity, because of poor 
 synthetical processes and inefficient methods of separation, 
 inasmuch as preliminary experiments pointed to a multiplicity 
 
 * Ber. Dtsch. chem. Ges., xxx, 2444. 
 VOL. ii. 20 
 
306 THE ETHERS OF ISONITROSOGUAIACOL. 
 
 of products in the reaction with nitrous acid such as Kietaibl * 
 found with the monoethyl ether, it was thought inadvisable in 
 the light of the results obtained with the meta body, to con- 
 tinue the work on the ethers of orthomethoxyquinoneoxime. 
 Work along the general line will be continued, and the results 
 of experimentation with mononitrosoresorcin will soon appear. 
 
 EXPERIMENTAL PART. 
 Isonitrosoguaiaeol and Salts. 
 
 The isonitrosoguaiacol used in the investigation was pre- 
 pared both by the method of Pfob, working with nascent nitrous 
 acid in alcoholic solution, and also by the general method for 
 the formation of the sodium salts of isonitroso bodies sug- 
 gested by Walker, f To a concentrated alcoholic solution of 
 sodium alcoholate, guaiacol is added in sufficient quantity to 
 form the sodium salt by the resulting metathesis, then, to this 
 solution of sodium guaiacol, slightly more than the theoretical 
 quantity of amyl nitrite is added and, after thorough mixing, 
 the liquid is allowed to stand over sulphuric acid for twenty- 
 four hours, when the bright olive-green crystalline sodium salt 
 of isonitrosoguaiacol separates. After washing thoroughly 
 with ether, pulverizing and rewashing, the salt may be used 
 directly for the preparation of derivatives, or, if further purifi- 
 cation is desirable, it may be dissolved in water, acidified with 
 hydrochloric acid, and the filtered and dried product dissolved 
 in ether and shaken with animal charcoal, when, upon evapo- 
 ration, the pure isonitrosoguaiacol crystallizes. Of the above 
 methods of preparation the latter is much to be preferred, 
 although Rupe mentions it unfavorably because of poor yields 
 and impure products. He advocates the use of ethyl nitrite 
 in a closed tube ; but on trial we were unable to obtain the 
 quantitative yields which he reports and the seventy per cent 
 yield which the amyl nitrite gives, makes this method quite 
 equal, in efficiency as well as purity of reaction-product, to 
 the other more tedious process. 
 
 The silver salt, formed from the sodium salt by treating the 
 aqueous solution with a slight excess of silver nitrate, comes 
 
 * Monatsh. Chem., xix, 536. t Ber. Dtsch. chem. Ges., xvii, 399. 
 
THE ETHERS OF ISONITROSOGUAIACOL. 307 
 
 down as a brown gelatinous precipitate, which becomes crystal- 
 line on gently warming, or may be obtained in crystalline form 
 at once by heating the separate solutions to 50 C. before mix- 
 ing. It is a very unstable salt, the dry product decomposing 
 with a very gentle heating. 
 
 Isonitrosoguaiacol Benzoyl Ether. 
 
 The sodium salt formed as above was dissolved in as little 
 water as possible and four or five tunes its volume of alcohol 
 added. This solution was thoroughly shaken with a slight 
 excess of benzoyl chloride, added drop by drop. The reaction 
 is immediate and the ether soon begins to come out of the 
 solution in almost pure condition. Recrystallized from alco- 
 hol, it separates in straw-colored, branching crystals, which 
 melt sharply at 188 C. when " dipped " for ten seconds. 
 Heated gradually from normal temperatures, it begins to de- 
 compose at 175 C. and liquefies at 185-188 C., the temper- 
 ature depending on the rapidity with which heat is applied. 
 
 Fractional crystallization from alcohol or other solvents did 
 not essentially change the melting-point, nor were any differ- 
 ent phenomena observed when the isonitrosoguaiacol was made 
 by the acid reaction. This ether dissolved readily in chloro- 
 form and glacial acetic acid, much less in benzene and ligroin, 
 and is practically insoluble in ether and carbon disulphide. 
 On analysis : 
 
 0.1100 gram of the substance, dried over H 2 S0 4 , gave 0.2632 
 gram C0 2 , and 0.0428 gram H 2 0. 
 
 0.1205 gram of the substance gave 5.45 cm 8 K at 15 C. and 
 772 mm. pressure. 
 
 Calculated for d 
 
 C 14 H U 4 N. Foun<L 
 
 C 65.34 65.26 
 
 H 4.31 4.32 
 
 N 5.46 5.38 
 
 Isonitrosoguaiacol Benzoyl Ether Dibromide. 
 
 Three grams of the benzoyl ether were dissolved in 25 cm 3 
 of chloroform and 2 grams of bromine were added, the mix- 
 
308 THE ETHERS OF ISONITROSOGUAIACOL. 
 
 ture being kept cool by running water. The solution was 
 allowed to stand for from one to two hours and then to evapo- 
 rate spontaneously. The dibromide is left behind as a light- 
 brown substance. Attempts to crystallize from glacial acetic 
 acid proved unsatisfactory since the boiling solvent caused a 
 decomposition of the ether; nor did a solution in fuming 
 nitric acid, precipitated by water, give a pure white product. 
 The best results were obtained by washing with warm dilute 
 alcohol when the product becomes yellowish-white, and melts 
 at 153-154 C., browning considerably above 140 C. An 
 analysis of the substance purified in this manner shows a low 
 percentage of bromine. 
 
 0.1052 gram of the substance, dried over H 2 S0 4 , gave 0.1547 
 gram C0 2 , and 0.0284 gram H 2 0. 
 
 0.0509 gram of the substance gave 0.0449 gram AgBr. 
 
 Calculated for Found. 
 
 C * 40.29 40.10 
 
 H 2.66 3.00 
 
 Br 38.34 37.52 
 
 Monobromisonitrosoguaiacol Benzoyl Ether. 
 
 By boiling the white dibromide with 60 per cent alcohol for 
 three-quarters of an hour, the halogen acid and a bright-yel- 
 low monobromine compound are formed. This ether crystal- 
 lizes from boiling alcohol in the characteristic prismatic form 
 of such bodies and melts with decomposition at 178 C. 
 Repeated recrystallizations did not change this melting-point, 
 and no indication was noted suggesting the possibility of a 
 mixture. On analysis: 
 
 0.1317 gram of the substance, dried over H 2 S0 4 , gave 0.2410 
 gram C0 2 , and 0.0368 gram H 2 0. 
 
 0.2101 gram of the substance gave 0.1172 gram AgBr. 
 
 Calculated for ,, . 
 
 C M H 10 4 NBr. Found ' 
 
 C 49.99 49.91 
 
 H 3.00 3.11 
 
 Br 23.79 23.75 
 
XXXIX 
 
 THE CONSTITUTION OF THE AMMONIUM 
 MAGNESIUM ARSENIATE OF ANALYSIS. 
 
 BY MARTHA AUSTIN.* 
 
 THE striking analogy between the phosphates and the arse- 
 niates led Levolf to undertake the separation of an ammonium 
 arseniate corresponding to the ammonium magnesium phos- 
 phate, the composition of which Berzelius had given. Levol 
 states that ammonium magnesium arseniate of the composition 
 NH 4 MgAsO 4 . 10H 2 O is obtained by adding a solution of a 
 double ammonium magnesium salt to arsenic acid, and that it 
 is a salt possessing about the same degree of solubili ty in water, 
 in ammoniacal water, and hi ammoniacal water containing 
 magnesium salt, as the corresponding phosphate. Further, he 
 found that by heating this salt to red heat after carefully 
 drying, magnesium pyroarseniate was given, from which 
 arsenic can be estimated readily. 
 
 WachJ and H. Rose obtained the ammonium magnesium 
 arseniate containing six molecules of water of crystallization by 
 precipitating arsenic acid with magnesia mixture and then add- 
 ing an excess of ammonia, and by drying at 100 C. were able 
 to estimate the arsenic present as the ammonium-magnesium 
 arseniate containing one-half molecule of water. This method 
 seemed to offer an advantage over the method of estimation as 
 the pyroarseniate, for results obtained below the theoretical 
 amount of arsenic present gave rise to the suspicion that 
 during ignition arsenic was reduced by the ammonia driven off. 
 Rose attempted hi another way to avoid this loss by igniting in 
 a current of oxygen ; and later Reichelt|| ignited the residue 
 
 * From Am. Jour. Sci., ix, 65. t Ann. Chim. in, xvii, 501 (1846). 
 
 t Schweigger, Jour. f. Ch. u. Pbys., lix, 297. 
 
 Ann. Phys., clii, 20 (1849). || Zeitschr. anal.Chem., xx, 89. 
 
310 CONSTITUTION OF THE AMMONIUM 
 
 after carefully saturating it with ammonium nitrate and nitric 
 acid, and drying at 100 C. Rammelsberg * believed that 
 it was safer to ignite after drying at 120 C. because 
 drying at 100 -110 C. caused a loss of ammonia before ignition. 
 Kaiser f dried the residue in a current of air. 
 
 A second source of error discussed by H. Rose4 Fresenius, 
 and others is due to the solubility of the ammonium magnesium 
 arseniate in water, in ammoniacal water, and ammoniacal water 
 containing magnesium salts. Wood attempted to avoid 
 this by precipitating the ammonio-magnesium arseniate with 
 an alcoholic magnesia mixture, and by washing the precipitate 
 with an alcoholic solution, by weighing the residue ignited in 
 a crucible held in a second protecting crucible, after treating 
 with nitric acid and ammonium nitrate, he obtained concordant 
 results. Brauner || followed this method with success. 
 
 As has been shown by Neubauer,l[ ammonium chloride 
 tends to form a phosphate of magnesium too rich in ammonia, 
 and in papers ** from this laboratory the tendency of ammonia 
 and ammonium chloride to influence the proportion of the 
 metal in the ammonium phosphate of manganese, of magnesium 
 and of other metals in the second group of Mendele*eff has 
 been pointed out. The marked similarity as to behavior 
 between the phosphates and the arseniates led to the investi- 
 gation of the constitution of the ammonium magnesium 
 arseniate under the usual conditions imposed in analysis, and, 
 further, to the effect of ammonium chloride on the salt of 
 ideal constitution. 
 
 In the table which follows are recorded a set of qualitative 
 tests, hi which the hot filtrates were tested by hydrogen 
 sulphide in presence of hydrochloric acid. These tests were 
 made to show under what conditions of volume, given amounts 
 of arsenic acid can be entirely removed by magnesia mixture 
 
 * Ber. Dtsch. chem. Ges., vii, 544. t Zeitschr. anal. Chem., xiv/250. 
 
 t Zeitschr. anal. Chem., iii, 206. Am. Jour. Sci. Ill, vi, 368. 
 
 || Zeitschr. anal. Chem., xvi, 57. 
 
 1 Zeitschr. anorg. Chem., ii, 45 ; iv, 251 ; x, 60. Zeitschr. angew. Chem., 
 1896, 435. Jour. Am. Chem. Soc., xiv, 289. 
 
 ** Am. Jour. Sci., vi, 233 ; vii, 187 ; viii, 206. This volume, pp. 121, 190, 252. 
 
MAGNESIUM ARSENIATE OF ANALYSIS. 
 
 311 
 
 from ammoniacal solutions alone, or from ammoniacal solutions 
 containing ammonium chloride. The ammonium chloride for 
 this work was carefully purified by heating it to boiling 
 temperature in concentrated solution 1 grm. to 3 cm 3 
 with ammonia in slight excess. The magnesia mixture was 
 prepared by dissolving one hundred and ten grams of the 
 crystallized magnesium chloride in a small volume of water, 
 filtering and adding to it fifty-eight grams of ammonium 
 chloride in solution, purified by adding bromine water and 
 bleaching with ammonia, filtering, and then diluting to a 
 volume of two liters, adding enough ammonia 10 cm 3 to 
 make this solution smell distinctly of ammonia. 
 
 The results (1) to (3) recorded in the table were obtained 
 by precipitating the arsenic present in solution by means of 
 the magnesia mixture prepared as described and rendering the 
 
 TABLE I. 
 
 Exp. 
 
 Volume. 
 
 Arsenic 
 present 
 in terms 
 ofA8 a 5 . 
 
 Magnesia 
 mixture. 
 
 NH 4 OH. 
 
 NH^Cl. 
 
 Indications of 
 arsenic in 
 the filtrate 
 byH 2 S. 
 
 
 cm* 
 
 grm. 
 
 cm 3 
 
 cm' 
 
 grm. 
 
 
 (1) 
 
 100 
 
 0.05 
 
 20 
 
 Slight excess. 
 
 t 
 
 None. 
 
 (2) 
 
 200 
 
 0.05 
 
 20 
 
 Slight excess. 
 
 m 
 
 None. 
 
 (3) 
 
 300 
 
 0.05 
 
 120 
 
 Slight excess. 
 
 . . 
 
 i Present. 
 
 
 
 
 H20 
 
 
 
 ii Present. 
 
 
 
 
 i"20 
 
 
 
 m None. 
 
 8 
 
 200 
 300 
 
 0.5 
 0.5 
 
 50 
 50 
 
 Slight excess. 
 Slight excess. 
 
 
 None. 
 Present. 
 
 (6) 
 
 300 
 
 0.5 
 
 50 
 
 2 
 
 
 Present. 
 
 (7) 
 
 200 
 
 0.5 
 
 60 
 
 4 
 
 
 Present. 
 
 (8) 
 
 200 
 
 0.5 
 
 30 
 
 2 
 
 
 Present. 
 
 (9) 
 
 200 
 
 0.5 
 
 130 
 
 4 
 
 
 1 Present. 
 
 
 
 
 ii 10 
 
 
 
 ii None. 
 
 (10) 
 
 100 
 
 0.5 
 
 130 
 
 2 
 
 
 1 Present. 
 
 
 
 
 ii 10 
 
 
 
 ii None. 
 
 (11) 
 
 130 
 
 0.5 
 
 60 
 
 Excess. 
 
 10 
 
 Present. 
 
 (12) 
 
 200 
 
 0.5 
 
 50 
 
 Excess. 
 
 10 
 
 Present 
 
 (13) 
 
 150 
 
 0.5 
 
 75 
 
 Excess. 
 
 10 
 
 None. 
 
 (14) 
 
 250 
 
 0.5 
 
 76 
 
 Excess. 
 
 10 
 
 Present. 
 
 (15) 
 
 300 
 
 0.5 
 
 75 
 
 2 
 
 10 
 
 Trace. 
 
 (16) 
 
 400 
 
 0.5 
 
 76 
 
 2 
 
 10 
 
 Present. 
 
 (17) 
 
 285 
 
 0.5 
 
 100 
 
 Excess. 
 
 10 
 
 None. 
 
 (18) 
 
 215 
 
 0.5 
 
 100 
 
 Excess. 
 
 20 
 
 None. 
 
 (19) 
 
 335 
 
 0.6 
 
 100 
 
 Excess. 
 
 60 
 
 Present. 
 
 (20) 
 
 360 
 
 0.5 
 
 125 
 
 Excess. 
 
 60 
 
 Present. 
 
 (21) 
 
 360 
 
 0.5 
 
 150 
 
 Excess. 
 
 60 
 
 None. 
 
312 
 
 CONSTITUTION OF THE AMMONIUM 
 
 solution distinctly ammoniacal. After standing until the 
 precipitate subsided, the solution was filtered and tested for 
 arsenic by hydrogen sulphide. Within certain limits of 
 volume all the arsenic is removed from solution by the magnesia 
 mixture. In case of (3) two additional amounts of the 
 precipitant had to be made in order to remove the arsenic, the 
 solution being ammoniacal. Results (4) to (10) of the table 
 were obtained in the same manner as (1) to (3), the amount 
 of arsenic acid present being increased ten times. In (9) and 
 (10), where the smaller amounts of magnesia mixture were 
 used, additional portions had to be added before all the arsenic 
 was removed from solution. When the larger amounts of 
 ammonia were used the results did not seem to be influenced, 
 (7) and (9). The results recorded in (11) to (21) were 
 obtained by precipitating the arsenic in solution hi presence of 
 ammonium chloride by the magnesia mixture. It is evident 
 that the presence of ammonium chloride causes some arsenic 
 to be dissolved and further that this solvent effect is overcome 
 by the magnesia mixture added in larger amounts, even when 
 the ammonium chloride present amounts to sixty grams in 
 weight, as in (21). 
 
 In order to find how much arsenic is dissolved from the 
 ammonium magnesium arseniate once precipitated so that no 
 arsenic is left in solution, experiments were made in which the 
 arsenic in the filtrate was weighed after precipitating by hydro- 
 gen sulphide in hot acid solution, filtering off on asbestos 
 under pressure, washing successively with water, alcohol, car- 
 bon disulphide, alcohol, and water, and drying at 100 C. 
 
 TABLE II. 
 
 H,O containing 1 cm 3 . 
 NB 4 Ofl in 200 cm. 
 
 NH 4 OH 
 
 (sp. gr. 0.96). 
 
 As 8 5 
 digested. 
 
 As-jOrj found 
 as As 2 S 8 . 
 
 grm. 
 
 cm 8 
 
 grm. 
 
 grm. 
 
 
 100 
 
 0.5 
 
 0.0019 
 
 
 100 
 
 0.5 
 
 0.0026 
 
 100 
 
 . 
 
 0.5 
 
 0.0003 
 
 100 
 
 
 
 0.5 
 
 0.0005 
 
 10 
 
 . 
 
 0.5 
 
 0.0002 
 
 10 
 
 . . . 
 
 0.5 
 
 0.0004 
 
MAGNESIUM ARSENIATE OF ANALYSIS. 313 
 
 It is evident from these qualitative tests that, so far as con- 
 cerns the amount of arsenic dissolved from the ammonium 
 magnesium arseniate, it is safe to use a faintly ammoniacal 
 wash water in small amounts less than 100 cm 3 to remove 
 traces of reagents from the residue after it is gathered upon 
 the asbestos felt. Usually 25-50 cm 3 of wash water were used 
 in rinsing off the precipitate hi the experiments about to be 
 given below. 
 
 The solution of arsenic employed in this work was prepared 
 by dissolving ten grams of pure arsenious oxide, carefully 
 resublimed, in a platinum dish in an excess of pure nitric acid 
 and evaporating on the waterbath to dryness, dissolving the 
 arsenic acid produced in water and diluting to a volume of 
 one liter in a standard flask. 
 
 Definite portions of this solution were drawn from a burette 
 into a platinum dish and precipitated with magnesia mixture, 
 prepared as described, in the proportion shown necessary by 
 the qualitative tests for the complete removal of the arsenic 
 from solution, and the solution was made distinctly ammoni- 
 acal. The precipitate, dissolved in hydrochloric acid in slight 
 excess, was brought down again by ammonia in distinct excess. 
 After standing until the precipitate had completely subsided, 
 the precipitate was gathered on an asbestos felt in a perforated 
 platinum crucible, making use of the filtrate to remove the 
 last portions of the precipitate to the felt, before washing any 
 reagents from it on the felt with faintly ammoniacal water. 
 After carefully drying, the residue was ignited. The results 
 shown hi section A of Table III fall so far below the theory for 
 magnesium pyroarseniate that it seems evident that the arsen- 
 iate shows a tendency here to form a salt richer hi ammonia 
 than the ideal MgNH 4 AsO 4 , and yielding on ignition some 
 meta-arseniate instead of the normal pyroarseniate. In section 
 B the magnesium ammonium arseniate was precipitated by add- 
 ing to the solution of arsenic acid magnesia mixture contain- 
 ing no free ammonia in the proportion necessary (50 cm 3 ) to 
 remove the arsenic from solution, and then making the solution 
 distinctly ammoniacal. After the precipitate had subsided, it 
 
314 
 
 CONSTITUTION OF THE AMMONIUM 
 
 was filtered off on asbestos under pressure in a perforated 
 platinum crucible washed on the felt with ammoniacal water, 
 dried and ignited. These results are also below the theory 
 for the pyroarseniate. Evidently the conditions here are even 
 better for the formation of the salt too rich in ammonia than 
 they were in the first case. No arsenic was found by hydro- 
 gen sulphide in any case either in the nitrate or in the wash 
 water after acidifying and heating. 
 
 TABLE IIL 
 
 A. 
 
 Mg 2 As 2 7 corresponding to A&jO-. 
 
 As 2 5 found 
 byBLjSin 
 the filtrate. 
 
 Taken. 
 
 Found. 
 
 Error. 
 
 grm. 
 0.7843 
 0.7843 
 
 grm. 
 0.7800 
 0.7794 
 
 grm. 
 0.0043- 
 0.0049- 
 
 None. 
 None. 
 
 B. 
 
 0.7843 
 0.7843 
 
 0.7772 
 0.7769 
 
 0.0071- 
 0.0074- 
 
 None. 
 None. 
 
 In section A of Table IV the results recorded were obtained 
 by precipitating definite portions of arsenic acid drawn from 
 a burette into a platinum dish with the distinctly ammoniacal 
 magnesia mixture in proper proportion, and afterward adding 
 a little more ammonia, filtering off on asbestos under pressure 
 in a perforated platinum crucible as soon as the precipitate 
 subsided, washing off on the felt with ammoniacal water any 
 of the reagents left on the precipitate in transferring it to the 
 felt by using the filtrate, drying, and igniting. No arsenic 
 was found in any case in the filtrate or in the wash water by 
 hydrogen sulphide. The conditions of precipitation here 
 prove to be such, as the results show, that the salt of ideal 
 constitution is formed. Comparing these results with those 
 of Table III, it seems that the conditions under which the salt 
 of ideal composition is formed are such that at the moment 
 
MAGNESIUM ARSENIATE OF ANALYSIS. 
 
 315 
 
 and in the locality of precipitation the amount of magnesium 
 chloride in a certain volume of solution must be large in pro- 
 portion to the amount of ammonia present as the chloride 
 and the hydroxide ; otherwise the ammonium of the ammonium 
 arseniate, naturally formed first, does not suffer sufficient dis- 
 placement by magnesium to produce the normal ammonium 
 magnesium arseniate. At all events, the treatment applied in 
 the experiments of A resulted practically in the complete pre- 
 cipitation and in the production of a precipitate of nearly 
 ideal constitution. 
 
 TABLE IV. 
 
 A. 
 
 Exp. 
 
 Mg2As 2 O 7 corresponding to AsjO 6 . 
 
 Magnesia 
 mixture. 
 
 NH 4 C1. 
 
 Taken. 
 
 Found. 
 
 Error. 
 
 R 
 
 | 
 
 grm. 
 
 0.7843 
 0.7843 
 0.7843 
 0.7843 
 
 grm. 
 
 0.7830 
 0.7849 
 0.7841 
 0.7843 
 
 gnu. 
 0.0013- 
 0.0006+ 
 0.0002- 
 0.0000 
 
 cm 8 
 50 
 60 
 50 
 50 
 
 grm. 
 
 B. 
 
 (5) 
 (6) 
 
 11! 
 
 (9) 
 (10) 
 
 (11) 
 (12) 
 
 0.7843 
 0.7843 
 0.7843 
 0.7843 
 0.7843 
 0.7843 
 0.7843 
 0.7843 
 
 0.7763 
 0.7762 
 0.7832 
 0.7838 
 0.7784 
 0.7810 
 0.7849 
 0.7846 
 
 0.0080- 
 0.0081- 
 0.0011- 
 0.0005- 
 0.0059- 
 0.0033- 
 0.0006+ 
 0.0003+ 
 
 75 
 75 
 100 
 100 
 100 
 100 
 150 
 150 
 
 10 
 10 
 10 
 10 
 20 
 20 
 60 
 60 
 
 In section B of the same table the results recorded show the 
 effect of increased amounts of ammonium chloride on the 
 constitution of the ammonium magnesium arseniate. Enough 
 magnesia mixture was used in each case to remove the arsenic 
 completely from solution. The precipitates were gathered on 
 asbestos in a perforated platinum crucible and treated like 
 those described in section A of the table. It is evident as 
 shown by (5) and (6) of the table that ammonium chloride 
 causes a replacement of some of the metal by ammonia in the 
 
316 AMMONIUM MAGNESIUM ARSENIATE. 
 
 ammonium magnesium arseniate (to form, possibly, a salt of 
 the constitution Mg(NH 4 ) 4 AsO 4 ) though the solvent effect 
 of the ammonium chloride is overcome by the addition of a 
 sufficient amount of magnesia mixture. No arsenic appeared 
 in the filtrates, and, further, experiments (7) and (8) show 
 that increasing the amount of magnesia mixture present will 
 cause the formation of the salt of ideal constitution even in 
 presence of considerable amounts of ammonium chloride. 
 Indeed this is possible where as large an amount as sixty 
 grams of the salt is present, as results (11) and (12) show. 
 Obviously, ammonium chloride in any amount above what is 
 required for the magnesia mixture tends to dissolve the 
 precipitate, but this solvent effect may be neutralized by 
 increasing the amounts of magnesia mixture even though the 
 precipitate formed is richer in ammonia than the ideal salt. 
 
 In no one of the many precipitates tested by silver nitrate 
 for included chlorides was more than an inappreciable trace 
 found. 
 
 Evidently, when ammoniacal magnesia mixture, amounting 
 to about thirty cubic centimeters in excess of the theoretical 
 amount necessary to precipitate all the arsenic as the ammo- 
 nium magnesium arseniate, is added to the faintly acid 
 solution of arsenic acid (carrying no ammonium salts) in a 
 volume not exceeding two hundred cubic centimeters, the 
 precipitate appears to fall in ideal condition. If the precipi- 
 tated salt is transferred to the filtering crucible by the aid of 
 portions of the filtrate used as the washing liquid and finally 
 washed on the asbestos with about twenty-five cubic centimeters 
 of faintly ammoniacal water an amount which is quite 
 sufficient after the transfer has been made no arsenic gets 
 into solution. The weight of the carefully dried and ignited 
 pyroarseniate indicates with accuracy the amount of arsenic 
 present. 
 
XL 
 
 ON THE ESTIMATION OF THALLIUM AS THE 
 ACID AND NEUTRAL SULPHATES. 
 
 BY PHILIP E. BROWNING.* 
 
 CEOOKES t has shown that the salt obtained by heating 
 thallous chloride with sulphuric acid until the excess of the 
 latter is expelled and then raising the heat to redness has the 
 constitution of a neutral sulphate, 
 
 He also found that continued heating did not result in 
 any essential loss of weight, and suggested the possibility 
 of applying this method of treatment to the estimation of 
 thallium. 
 
 CastanjenJ in a recent paper discusses thoroughly the 
 compounds of thallium and confirms essentially the statements 
 of Crookes in regard to the neutral sulphate, adding, however, 
 the observation that on strong ignition in the air this salt 
 tends to lose sulphuric acid. He also mentions in the same 
 paper the acid sulphate, and states that on heating it first 
 melts and on continued heating gives off sulphuric acid, 
 leaving the neutral sulphate. 
 
 The work to be described in this paper was undertaken to 
 determine under what conditions the formation of these salts 
 may be applied to the estimation of thallium. For the work a 
 solution was made by dissolving a given amount of the nitrate 
 in water and making up to a liter. The value of the solution 
 was determined by precipitating measured and weighed 
 amounts of this solution both as the iodide and chromate, as 
 described in a previous paper. Closely agreeing results by 
 
 * From Am. Jour. Sci., ix, 137. t Chem. News, viii, 243. 
 
 t Jour, prakt. Chem., cii, 131. 
 
 Am. Jour. Sci., yiii, 460. This volume, p. 300. 
 
318 
 
 ESTIMATION OF THALLIUM AS THE 
 
 both methods were taken as the standard. Measured amounts 
 of this solution were drawn from a burette into weighed 
 platinum crucibles, and the weight taken as a check on the 
 burette reading. To the solution in the crucible a few drops 
 of sulphuric acid were added and the water removed by 
 evaporation over a steam bath. The crucible was then removed 
 to a radiator, consisting of a conical iron cup, and heated 
 at a temperature ranging from 220 C. to 240 C., until 
 fuming ceased and the weight after half-hour periods of 
 heating remained constant. The crucibles were placed in the 
 radiator upon a pipe stem triangle so that they were about 5 
 cm. from the bottom, which was heated at low redness. A 
 thermometer, hung so that the bulb occupied the same position 
 as the crucible, gave the reading mentioned above. 
 
 As will be seen, the results obtained by this treatment agree 
 closely with the calculated amounts of acid sulphate which 
 should be formed. In several experiments this salt was dis- 
 solved in water and the sulphuric acid present in combination 
 precipitated by barium nitrate. The results obtained agreed 
 closely with the formula of the acid sulphate of thallium. 
 Having obtained by the method described the acid sulphate, 
 the crucibles were removed and heated carefully over a free 
 flame to low redness, when, after a considerable Devolution of 
 sulphuric acid fumes, the weight again became constant, and 
 the results showed a condition closely approximating to that of 
 the neutral sulphate. In several of these experiments the 
 
 Exp. 
 
 T1HS0 4 
 calculated. 
 
 T1HS0 4 
 found. 
 
 Error. 
 
 T1 2 S0 4 
 calculated. 
 
 T1 2 S0 4 
 found. 
 
 Error. 
 
 
 grin. 
 
 grm. 
 
 grin. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.1605 
 
 0.1596 
 
 0.0009- 
 
 0.1344 
 
 0.1346 
 
 0.00024- 
 
 (2) 
 
 0.1611 
 
 0.1608 
 
 0.0003- 
 
 0.1349 
 
 0.1346 
 
 0.0003- 
 
 (3) 
 
 0.1608 
 
 0.1608 
 
 0.0000 
 
 0.1347 
 
 0.1352 
 
 0.0005-f 
 
 (4) 
 
 0.1 6T2 
 
 0.1600 
 
 0.0012- 
 
 0.1350 
 
 0.1346 
 
 0.0004- 
 
 5) 
 
 0.1602 
 
 0.1596 
 
 0.0006- 
 
 0.1341 
 
 0.1346 
 
 0.0005-j- 
 
 (6) 
 
 0.1608 
 
 0.1596 
 
 0.0012- 
 
 . 
 
 . 
 
 . 
 
 (7) 
 
 0.1617 
 
 0.1604 
 
 0.0013- 
 
 t t 
 
 4 
 
 
 (8) 
 
 0.1608 
 
 0.1592 
 
 0.0016- 
 
 0.1347 
 
 0.1358 
 
 0.00114- 
 
 (9) 
 
 0.1609 
 
 0.1590 
 
 0.0019- 
 
 0.1348 
 
 0.1346 
 
 0.0002- 
 
ACID AND NEUTRAL SULPHATES. 319 
 
 sulphuric acid present in combination was determined, and 
 showed amounts closely agreeing with the constitution of the 
 neutral sulphate. 
 
 These results would seem to show that thallium may be 
 estimated either as the acid sulphate or as the neutral sulphate 
 by careful attention to the proper conditions of temperature. 
 
XLI 
 
 THE SEPARATION AND DETERMINATION OF 
 MERCURY AS MERCUROUS OXALATE. 
 
 BY C. A. PETERS.* 
 
 IT is stated in the literature! that oxalic acid, neutral and 
 acid oxalates of the alkalies, precipitate mercurous salts, and 
 that oxalic acid and the double oxalates of potassium produce 
 no precipitate with mercuric chloride solution. Starting with 
 these facts, the attempt was made to estimate mercurous salts : 
 volumetrically, by precipitating with ammonium oxalate and 
 determining the oxalic acid by potassium permanganate ; and 
 gravimetrically by direct weighing of the precipitate. 
 
 The Volumetric Estimation. 
 
 The mercurous nitrate solution used was standardized by 
 the battery, and contained about 12 grm. of metallic mercury 
 to the liter. To obviate the tendency of the mercury salt 
 to break down and form basic salts J upon the addition of a 
 large amount of water if no nitric acid is added, the solution 
 was prepared in the following manner. About 20 grm. of 
 mercurous nitrate were ground in a mortar, transferred to a 
 flask, and 200-300 cm 3 water added. After shaking well, 
 the solution was filtered and the nitrate diluted to one liter. 
 Five cubic centimeters of this solution when precipitated with 
 a sodium chloride solution gave a nitrate from which only a 
 very slight darkening in color could be obtained, even upon 
 
 * From Am. Jour. Sci., ix, 401. 
 
 t Rose-Finkener, Handbuch der analytischen Chemie, i, 319. 
 
 J Graham-Otto, Handbuch, iii, 1102. 
 
SEPARATION OF MERCURY. 
 
 321 
 
 several hours' standing, when treated with hydrogen sulphide, 
 thus showing the absence of a mercuric salt. 
 
 A solution made in the above manner had not changed 
 its standard after a period of eight weeks. The potassium 
 permanganate solution (approximately ^) was standardized 
 against lead oxalate. 
 
 It was first attempted to estimate the mercurous salts as 
 follows. The mercurous oxalate was precipitated cold by 
 means of ammonium oxalate, stirred well, and allowed to 
 settle, the completion of the precipitation being determined 
 by addition of more ammonium oxalate. The precipitate was 
 collected on asbestos, washed once or twice with cold water, 
 and (still in the crucible) treated in a beaker with 5 cm 8 of 
 strong hydrochloric acid. To the solution diluted to 100-200 
 cm 3 1 grm. of a manganous salt was added, and the oxalic 
 acid was titrated with permanganate at the ordinary tem- 
 perature of the room. The end color was not stable and 
 was hard to determine. Three experiments, using 0.1217 of 
 mercury hi form of mercurous nitrate, gave plus errors of 
 0.0011 grm., 0.0017 grm. and 0.0028 grm. respectively, or 1.5 
 per cent. The precipitate when dissolved in sulphuric acid 
 
 
 
 
 Excess of 
 
 
 
 
 
 
 
 
 ammonium 
 
 
 
 
 
 Exp. 
 
 Hg taken as 
 
 Hg 2 (N0 8 ), 
 
 Hg found. 
 
 oxalate 
 approxi- 
 
 
 MnCl, . 4H 2 O. 
 
 H 2 S0 4 
 
 Error 
 asHg. 
 
 
 
 
 mately^. 
 
 ' 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 A. 
 
 
 
 
 
 
 grm. 
 
 grm. 
 
 cm 8 
 
 cm 
 
 grm. 
 
 cm 3 
 
 grm. 
 
 (1) 
 
 0.1825 
 
 0.1823 
 
 0.90 
 
 5 
 
 0.5 
 
 
 0.0002- 
 
 (2) 
 
 0.1217 
 
 0.1218 
 
 0.93 
 
 5 
 
 0.6 
 
 
 0.0001+ 
 
 (3) 
 
 0.1217 
 
 0.1206 
 
 0.99 
 
 5 
 
 0.5 
 
 
 0.0011- 
 
 (4) 
 
 0.1217 
 
 0.1210 
 
 4.05 
 
 5 
 
 0.5 
 
 
 0.0007- 
 
 (5) 
 
 0.3042 
 
 0.3034 
 
 4.97 
 
 5 
 
 0.6 
 
 . . 
 
 0.0008- 
 
 
 
 
 B. 
 
 
 
 
 
 (6) 
 
 0.1217 
 
 0.1220 
 
 0.93 
 
 t 
 
 
 5 
 
 0.0003+ 
 
 (7) 
 
 0.1217 
 
 0.1211 
 
 0.97 
 
 , 
 
 g 
 
 5 
 
 0.0006- 
 
 (8) 
 
 0.1825 
 
 0.1827 
 
 0.89 
 
 
 
 5 
 
 0.0002+ 
 
 (9) 
 (10) 
 
 0.3042 
 0.1217 
 
 0.3040 
 0.1202 
 
 0.82 
 4.10 
 
 * 
 
 
 
 5 
 5 
 
 0.0002- 
 0.0015- 
 
 TOL. II. 21 
 
322 SEPARATION AND DETERMINATION OF 
 
 and titrated gave no better results. To obviate this difficulty 
 the ammonium oxalate solution was matched on the per- 
 manganate and the oxalic acid in the nitrate determined. 
 The results obtained by this method, given in the preceding 
 table, are quite accurate. 
 
 In the experiments recorded in section A of the table the 
 filtrate was titrated in the presence of hydrochloric acid and 
 a manganous salt, at a temperature of 20 -40.* In the 
 experiments under section B sulphuric acid was added and 
 the solution heated in the usual manner. An excess of 
 ammonium oxalate, as shown hi experiments (4), (5) and 
 (10), interferes in no way. 
 
 The separation of the mercurous salt from small quantities of 
 mercuric salts, by the means of dilute nitric acid (sp. gr. 1.15), 
 was next attempted. It is stated f that mercurous oxalate is 
 insoluble in cold dilute nitric acid, while mercuric oxalate is 
 more or less soluble in the same reagent. Before attempting 
 any separation, however, there are three factors with reference 
 to the action of the nitric acid which need to be deter- 
 mined first, the maximum amount of nitric acid that may 
 be present in the titration of an oxalate without interference ; 
 second, the maximum amount of nitric acid that may be 
 present in a precipitation of mercurous oxalate without hav- 
 ing any perceptible solvent action upon the same ; and, third, 
 the amount of mercuric oxalate which will be held in solution 
 by given amounts of nitric acid. 
 
 To determine the amount of nitric acid that may be present 
 in the titration of an oxalate, 10 cm 3 of ^ ammonium oxalate 
 were titrated with permanganate at a dilution of 100 cm 3 
 with sulphuric acid at 80 C. The event proved that 10 cm 3 
 of nitric acid (sp. gr. 1.15) may be present without appearance 
 of interfering action. The maximum amount of nitric acid 
 which may be present without action upon the mercurous salt 
 was determined as shown in the following experiments. 
 
 * Gooch and Peters, Am. Jour. Sci., vii, 461, 1899. This volume, p. 222. 
 t Souchay and Lenssen, Ann. Chem. (Liebig), cii, 43. 
 
MERCURY AS MERCUROUS OXALATE. 
 
 323 
 
 
 Excess of 
 
 
 
 
 
 Hg taken as 
 
 Hg 2 (N0 3 ),. 
 
 ammonium 
 oxalate 
 approximately 
 
 TO' 
 
 HNO, 
 
 "?:$: 
 
 Volume 
 at 
 precipitation. 
 
 Hg found. 
 
 Error. 
 
 gnn. 
 
 era. 3 
 
 cm 8 
 
 cm 8 
 
 grin. 
 
 grm. 
 
 0.1122 
 
 1.64 
 
 10 
 
 100 
 
 0.1087 
 
 0.0035- 
 
 0.1122 
 
 2.62 
 
 6 
 
 100 
 
 0.1098 
 
 0.0024- 
 
 0.1122 
 
 1.59 
 
 6 
 
 100 
 
 0.1096 
 
 0.0026- 
 
 f 0.1122 
 
 1.53 
 
 6 
 
 100 
 
 0.1109 
 
 0.0013- 
 
 J 0-1122 
 
 1.40 
 
 5 
 
 100 
 
 0.1134 
 
 0.0012+ 
 
 1 0.1122 
 
 1.50 
 
 5 
 
 100 
 
 0.1115 
 
 0.0007- 
 
 10.1122 
 
 1.70 
 
 4 
 
 100 
 
 0.1116 
 
 0.0006- 
 
 0.1122 
 
 1.59 
 
 8 
 
 200 
 
 0.1096 
 
 0.0026- 
 
 0.1122 
 
 6.62 
 
 8 
 
 200 
 
 0.1111 
 
 0.0011- 
 
 0.1122 
 
 7.72 
 
 8 
 
 200 
 
 0.1108 
 
 0.0014- 
 
 ( 0.1122 
 
 2.59 
 
 5 
 
 200 
 
 0.1107 
 
 0.0015- 
 
 \ 0.1010 
 
 2.07 
 
 5 
 
 200 
 
 0.1003 
 
 0.0007- 
 
 Working under the conditions stated in the above table, it 
 is plain that 5 cm 3 of nitric acid (sp. gr. 1.15) may be used 
 before its solvent action is sufficient to interfere with the 
 accuracy of the process. 
 
 To determine the amount of mercuric salt that would be 
 held up by 5 cm 3 of nitric acid (sp. gr. 1.15), the following 
 experiments were made. 
 
 Hg taken as 
 Hg 2 (N0 3 ),. 
 
 HNO 3 
 
 Tiff 
 
 Ammonium 
 oxalate^ 
 in excess. 
 
 Volume. 
 
 Time before precipitation. 
 
 gms. 
 
 cm** 
 
 cm 8 
 
 cm 8 
 
 
 0.0335 
 
 4 
 
 0.76 
 
 100 
 
 6 hours. 
 
 0.0095 
 
 4 
 
 1.00 
 
 100 
 
 No ppt., 20 hours. 
 
 0.0143 
 
 5 
 
 6.6 
 
 100 
 
 No ppt., 20 hours. 
 
 0.0238 
 
 5 
 
 1.5 
 
 100 
 
 Slight ppt. 20 hours. 
 
 0.0238 
 
 5 
 
 5.5 
 
 100 
 
 26 minutes. 
 
 Five cm 3 of dilute nitric acid (sp. gr. 1.15) will prevent the 
 precipitation of small amounts of mercuric salt, 10-20 mgrm. 
 calculated as mercury, depending upon the amount of am- 
 monium oxalate present in excess. This amount of nitric acid 
 has no apparent solvent action on a precipitate of mercurous 
 oxalate under conditions already stated and does not interfere 
 
324 
 
 SEPARATION AND DETERMINATION OF 
 
 with the titration of an oxalate by permanganate as already 
 shown. 
 
 Carrying out the process of separation of mercurous salts 
 from mercuric salts, the precipitation was made as described for 
 the estimation of mercurous salts alone, excepting that nitric 
 acid and the mercuric salt were added. The experiments in A, 
 B, and C, of the accompanying table show the amounts of mer- 
 curic salt from which the mercurous oxalate may be separated 
 with 2 cm 3 of nitric acid. In experiments in section A the 
 results are quite accurate, but an excess of ammonium oxalate 
 tends to increase the results a little as seen in experiments 
 under B. An increase in the amount of mercuric salt present 
 causes also, as shown in section C, the results to be a trifle 
 high. 
 
 In experiments D, using 4 cm 3 of nitric acid even in the 
 
 Hg taken as 
 
 JJg^JfOaU. 
 
 Hg(N0 3 ), 
 present 
 calculated 
 asHg. 
 
 Ammonium 
 oxalate. 
 approxi- 
 mately^ 
 in excess. 
 
 HNO 8 
 
 Volume 
 at 
 precipi- 
 tation. 
 
 Hg found. 
 
 Error. 
 
 grm. 
 
 gnu. 
 
 cm 
 
 cm 3 
 
 grm. 
 
 grm. 
 
 grm. 
 
 ( 0.1217 
 
 0.0067 
 
 0.86 
 
 2 
 
 100 
 
 0.1232 
 
 0.0015+ 
 
 1 0.1217 
 
 0.0067 
 
 0.92 
 
 2 
 
 100 
 
 0.1220 
 
 0.0003+ 
 
 A 
 
 0.1217 
 
 0.0067 
 
 0.97 
 
 2 
 
 100 
 
 0.1218 
 
 0.0001+ 
 
 I 0.1217 
 
 0.0067 
 
 0.90 
 
 2 
 
 100 
 
 0.1224 
 
 0.0007+ 
 
 1 0.1217 
 
 0.0067 
 
 0.91 
 
 2 
 
 100 
 
 0.1213 
 
 0.0004- 
 
 
 0.1221 
 
 0.0067 
 
 3.92 
 
 2 
 
 100 
 
 0.1237 
 
 0.0016+ 
 
 B 
 
 0.1217 
 
 0.0067 
 
 3.93 
 
 2 
 
 100 
 
 0.1235 
 
 0.0018+ 
 
 
 0.1242 
 
 0.0067 
 
 8.75 
 
 2 
 
 100 
 
 0.1263 
 
 0.0021+ 
 
 
 0.1217 
 
 0.0134 
 
 0.87 
 
 2 
 
 100 
 
 0.1230 
 
 0.0013+ 
 
 
 0.1217 
 
 0.0134 
 
 0.86 
 
 2 
 
 100 
 
 0.1232 
 
 0.0015+ 
 
 
 ; 0.1217 
 
 0.0134 
 
 3.93 
 
 4 
 
 100 
 
 0.1218 
 
 0.0001+ 
 
 " 
 
 0.1217 
 
 0.0134 
 
 3.90 
 
 4 
 
 100 
 
 0.1211 
 
 0.0006- 
 
 T? 
 
 ! 0.2244 
 
 0.0067 
 
 1.88 
 
 4 
 
 115 
 
 0.2244 
 
 0.0000 
 
 E - 
 
 0.2244 
 
 0.0067 
 
 1.91 
 
 4 
 
 130 
 
 0.2240 
 
 0.0004- 
 
 
 I 0.2244 
 
 0.0141 
 
 2.98 
 
 4 
 
 100 
 
 0.2230 
 
 0.0014- 
 
 
 0.2244 
 
 0.0141 
 
 2.94 
 
 4 
 
 100 
 
 0.2241 
 
 0.0003- 
 
 
 i 0.2244 
 
 0.0067 
 
 8.88 
 
 4 
 
 100 
 
 0.2289 
 
 0.0045+ 
 
 
 0.2244 
 
 0.0067 
 
 8.90 
 
 4 
 
 100 
 
 0.2285 
 
 0.0041+ 
 
 
 [ 0.2424 
 
 0.0067 
 
 1.75 
 
 4 
 
 200 
 
 0.2432 
 
 0.0008+ 
 
 G 
 
 ! 0.2424 
 
 0.0067 
 
 2.00 
 
 4 
 
 200 
 
 0.2414 
 
 0.0010- 
 
 i 
 
 ' 0.2424 
 
 0.0067 
 
 2.96 
 
 4 
 
 200 
 
 0.2421 
 
 0.0003- 
 
 
 0.2241 
 
 0.0134 
 
 1.74 
 
 4 
 
 200 
 
 0.2271 
 
 0.0026+ 
 
 u. 
 
 0.2245 
 
 0.0134 
 
 1.81 
 
 4 
 
 200 
 
 0.2256 
 
 0.0011+ 
 
 I 
 
 0.1122 
 
 0.0144 
 
 6.62 
 
 5 
 
 200 
 
 0.1121 
 
 0.0001- 
 
 
 (0.1122 
 
 0.0240 
 
 6.54 
 
 5 
 
 200 
 
 0.1136 
 
 0.0014+ 
 
 
 i 0.1122 
 
 0.0240 
 
 6.57 
 
 5 
 
 200 
 
 0.1130 
 
 0.0008+ 
 
MERCURY AS MERCUROUS OXALATE. 325 
 
 presence of an excess of ammonium oxalate, the results are 
 accurate. In experiments E the amount of mercurous salt 
 was doubled and the results are accurate. In EE the amount 
 of mercuric salt was also doubled and the results are still 
 fairly accurate ; but when a large excess of ammonium oxa- 
 late is present as in experiments F, even with the smaller 
 amount of mercuric salt, the results are high. At a dilution 
 of 200 cm 3 the results are normal as seen in experiments G ; 
 but the introduction of more mercuric salt, as in experiments 
 H, causes a plus error. 
 
 Using 5 cm 3 of nitric acid, the larger amount of mercuric 
 salt together with a large excess of ammonium oxalate, as 
 recorded in experiments K, the error is raised a trifle; but 
 with a smaller amount of mercurous salt, as in experiment I, 
 the result is normal. 
 
 In precipitating mercurous salts by ammonium oxalate (^) 
 it is an easy matter to keep the excess of the precipitant 
 within the limits of 1 or two cm 3 , because the mercurous oxa- 
 late, when properly stirred, settled very rapidly. 
 
 The Gravimetric Estimation. 
 
 All the conditions described above in the volumetric estima- 
 tion of mercurous oxalate, for the separation of mercurous 
 from mercuric salts, may be applied to the gravimetric estima- 
 tion of mercurous oxalate. The precipitate is collected on a 
 weighed asbestos filter, washed two or three times with cold 
 water and dried over sulphuric acid to a constant weight. 
 Amounts of mercurous oxalate equivalent to 0.1217 and 0.2244 
 grm. of metallic mercury dried to a constant weight over sul- 
 phuric acid in about 15 hours ; a larger amount equivalent to 
 0.3 grm. of metallic mercury, required about 2 days to dry to a 
 constant weight. Souchay and Lenssen* state that mercurous 
 oxalate breaks up at 100 ; consequently this temperature 
 cannot be used for drying. For example, a precipitate con- 
 taining 0.1122 grm. mercury as the oxalate which when 
 
 * Ann. Chem. (Liebig), ciii, 308. 
 
326 
 
 SEPARATION AND DETERMINATION OF 
 
 brought to a constant weight at the ordinary temperature over 
 sulphuric acid weighed 0.1371 grm., when heated weighed as 
 follows : 
 
 After 7 hours at 110, weight = 0.1338 grm. 
 2 " " " 0.1328 " 
 
 7 ( ( 0.1302 
 
 The result shows a loss of 0.0069 grm. for 16 hours, heating, 
 and agrees with the statement of Souchay and Lenssen. 
 
 The following experiments give the results of the gravi- 
 metric work, in which the drying was effected, by exposure 
 for 15 hours or less, at ordinary temperatures over sulphuric 
 acid. The larger amounts of nitric acid present in the separa- 
 tions cause the precipitate to be more granular and aid in the 
 filtering process. 
 
 
 
 Excess of 
 
 
 
 
 
 
 
 ammonium 
 
 
 
 
 
 Hg taken as 
 
 Hg present 
 
 oxalate 
 
 HN0 3 
 
 Volume at 
 
 
 
 Hg 2 (N0 3 ) 3 
 
 asHg(N0 8 ) 2 
 
 present 
 
 
 precipi- 
 
 Hg found. 
 
 Error. 
 
 
 
 approxi- 
 
 1.15). 
 
 tation. 
 
 
 
 
 
 mately ^. 
 
 
 
 
 
 grm. 
 
 grm. 
 
 cm 8 
 
 cm 8 
 
 cm 3 
 
 grm. 
 
 grm. 
 
 
 0.1217 
 
 . 
 
 2-4 
 
 , , 
 
 100 
 
 0.1217 
 
 0.0000 
 
 K 
 
 0.1217 
 
 . . . 
 
 2-4 
 
 . . 
 
 100 
 
 0.1217 
 
 0.0000 
 
 
 0.1122 
 
 . . . 
 
 2-4 
 
 . . 
 
 100 
 
 0.1124 
 
 0.0002+ 
 
 
 0.1122 
 
 0.0067 
 
 0.93 
 
 2 
 
 100 
 
 0.1130 
 
 0.0008- 
 
 
 0.1122 
 
 0.0067 
 
 0.93 
 
 2 
 
 100 
 
 0.1112 
 
 0.0010- 
 
 M " 0.1122 
 
 0.0067 
 
 4.40 
 
 2 
 
 100 
 
 0.1124 
 
 0.0002+ 
 
 N 0.1122 
 
 0.0135 
 
 0.72 
 
 4 
 
 100 
 
 0.1125 
 
 0.0003+ 
 
 
 0.2244 
 
 0.0071 
 
 1.68 
 
 4 
 
 100 
 
 02253 
 
 0.0009+ 
 
 
 0.2244 
 
 0.0071 
 
 2.46 
 
 4 
 
 100 
 
 0.2241 
 
 0.0003- 
 
 Tj 
 
 0.2244 
 
 0.0048 
 
 0.54 
 
 4 
 
 200 
 
 0.2248 
 
 0.0004+ 
 
 
 0.2244 
 
 0.0048 
 
 2.44 
 
 4 
 
 200 
 
 0.2245 
 
 0.0001+ 
 
 In section K are experiments showing the accuracy of the 
 process where a mercurous salt is precipitated in the absence 
 of a mercuric salt. A small amount of mercuric salt was 
 introduced in experiments L, and an excess of ammonium 
 oxalate in experiment M, a still larger amount of mercuric 
 salt was present in experiment N, and a larger amount of 
 mercurous salt in experiments under O. In experiments in 
 section P a dilution of 200 cm 3 was employed both with and 
 
MERCURY AS MERCUROUS OXALATE. 327 
 
 without an excess of ammonium oxalate. All the results are 
 within reasonable limits of error. 
 
 The work may be summed up briefly as follows : Mercurous 
 nitrate may be estimated volumetrically by precipitating as 
 the oxalate and determining the excess of the precipitant with 
 permanganate. 
 
 The precipitated mercurous oxalate may also be estimated 
 gravimetrically by drying it over sulphuric acid and weighing 
 directly. 
 
 In solutions containing 2-5 per cent dilute nitric acid, sp. gr. 
 1.15, mercurous salts may be separated quantitatively as the 
 oxalate from small quantities of mercuric salts. 
 
 If about 0.12 grm. of mercury is present as the nitrate in 
 100 cm 3 of water, about 12 per cent of that amount of mercury 
 as the mercuric salt may be present without interfering with 
 the accuracy of the estimation, and even 20 per cent may be 
 present before an appreciable rise in the result is apparent. If 
 the amount of mercurous salt present is doubled, the amount 
 of mercuric salt which may be present is cut down about 
 one-half. 
 
XLII 
 
 THE TITRATION OF MERCURY BY SODIUM 
 THIOSULPHATE. 
 
 BY JOHN T. NORTON, JR.* 
 
 ACCORDING to J. J. Schererf mercurous nitrate, mercuric 
 nitrate and mercuric chloride may be estimated by direct 
 titration with sodium thiosulphate, Hg 2 S, 2HgS . Hg(NO 3 ) 2 , 
 and 2HgS . HgCla being the precipitates obtained in each 
 case. I have been unable to obtain access to Scherer's original 
 publication, but Suttonf gives the following very general 
 directions for this process : 
 
 " (a) Mercurous salts. The solution containing the metal 
 as a protosalt only is diluted, gently heated and the thiosul- 
 phate delivered in from the burette at intervals, meanwhile 
 well shaking until the last drop produces no brown color. The 
 sulphide settles freely and allows the end of the reaction 
 to be easily seen. One cm 3 of the ^V normal solution of 
 thiosulphate = 0.02 grm. Hg or 0.0208 grm. Hg 2 O. 
 
 " (5) Mercuric nitrate. The solution is considerably diluted, 
 put into a stoppered flask, nitric acid added and the thiosul- 
 phate cautiously added from the burette, vigorously shaken 
 meanwhile, until the last drop produces no further precipitate. 
 Scherer recommends that when the greater part of the metal 
 is precipitated the mixture should be diluted to a definite 
 volume, the precipitate allowed to settle and a measured 
 quantity of the clear li quid taken for titration ; the analysis 
 may then be checked by a second titration of the clear liquid if 
 needful. One cm 3 of ^ normal thiosulphate = 0.015 grm. Hg or 
 0.0162 grm. HgO. 
 
 * From Am. Jour. Sci., x, 48. 
 
 t Scherer's Lehrbuch der Chemie, i, 513. 
 
 J Volumetric Analysis, p. 220. 
 
THE TITRATION OF MERCURY. 329 
 
 " (<?) Mercuric chloride. With mercuric chloride the end of 
 the process is not so easily seen. The very dilute solution is 
 acidified with hydrochloric acid, heated nearly to boiling, and 
 the thiosulphate cautiously added so long as a white precipitate 
 is seen to form ; any great excess of the precipitant produces 
 a dirty-looking color. Filtration is necessary to distinguish 
 the exact ending of the reaction. One cm 3 of fa normal 
 thiosulphate = 0.015 grin. Hg or 0.0162 grm. HgO." 
 
 Fresenius * gives practically the same directions, but omits 
 all mention of that portion of the process dealing with 
 mercurous nitrate. 
 
 In view, therefore, of the scant information available on the 
 subject and of the apparent difficulty of working the process 
 accurately according to the directions given, an attempt was 
 made to ascertain whether the careful regulation of tempera- 
 ture, dilution, and amount of acid present might not produce 
 beneficial results. 
 
 That portion of the process dealing with mercuric chloride 
 was first taken up. The mercuric chloride used was pulverized, 
 dried at 100 and its purity proved by several determinations 
 as mercuric sulphide. The sodium thiosulphate was made 
 up of approximately fa normal strength and standardized 
 on decinormal iodine, which in turn was titrated against 
 decinormal arsenious acid made from pure resublimed arsenious 
 oxide. 
 
 For the action of sodium thiosulphate upon the mercuric 
 chloride Scherer gives the equation, 
 
 3HgCl 2 + 2Na 2 S 2 3 + 3H 2 = 2HgS . HgCl 2 + 2Na 2 S0 4 + 4HC1. 
 
 According to my experience, the action results in the formation 
 of a dense white precipitate which refuses to settle either by 
 shaking or standing, thus making it impossible to fix the end 
 reaction by reading the first drop of thiosulphate which produces 
 no further white precipitate in the solution containing the 
 mercuric chloride. Recourse must be had therefore to filtering. 
 By far the quickest and neatest method is to use the asbestos 
 
 * Quantitative Analysis. 
 
330 
 
 THE TITRATION OF MERCURY 
 
 filter deposited on a large perforated platinum cone.* This 
 cone is set in a glass funnel by means of a rubber connector 
 and the funnel is passed through the stopper of a large side- 
 necked Erlenmeyer connected with an exhaust pump. A little 
 asbestos fiber shaken in the liquid to be filtered was found to 
 be very beneficial in preventing the precipitate from running 
 through the filter. In all the following experiments the 
 thiosulphate was run into the solution containing the mercuric 
 chloride in excess, the whole shaken up with asbestos fiber, 
 filtered and the excess of thiosulphate determined by ^ 
 normal iodine. This procedure seems to be far preferable 
 to attempting to catch the end of the reaction by run- 
 ning hi the thiosulphate until the last drop produces no 
 precipitate. In the experiments shown in Table I no at- 
 tention was paid to the temperature of the solution, and the 
 thiosulphate was run in until the liquid turned brown. In 
 every case the solution was allowed to stand until there was 
 no further visible change of color. 
 
 TABLE I. 
 
 Exp. 
 
 HgCl 2 taken, 
 calc'd as 
 
 Na 2 S 2 3 in 
 
 Volume 
 at 
 
 HgCl 2 found, 
 calc'd as 
 
 Error. 
 
 
 Hg. 
 
 excess. 
 
 beginning. 
 
 Hg. 
 
 
 
 grm. 
 
 cm 8 
 
 cm 8 
 
 grin. 
 
 grin. 
 
 
 1) 
 
 0.0446 
 
 46.28 
 
 200 
 
 0.0343 
 
 0.0103- 
 
 
 
 0.0354 
 
 46.28 
 
 400 
 
 0.0326 
 
 0.0028- 
 
 
 3) 
 
 0.0356 
 
 44.97 
 
 400 
 
 0.0225 
 
 0.0131- 
 
 
 4) 
 
 0.0345 
 
 44.5 
 
 100 
 
 0.0308 
 
 0.0037- 
 
 
 5) 
 
 0.0354 
 
 44.43 
 
 50 
 
 0.0326 
 
 0.0028- 
 
 (6) 
 
 0.0382 
 
 22.59 
 
 50 
 
 0.0354 
 
 0.0028- 
 
 
 7) 
 
 0.0375 
 
 8.58 
 
 50 
 
 0.0385 
 
 0.0010+ 
 
 
 8) 
 
 0.0371 
 
 1.84 
 
 50 
 
 0.0304 
 
 0.0067- 
 
 9) 
 
 0.0731 
 
 2.28 
 
 50 
 
 0.0774 
 
 0.0043+ 
 
 (10) 
 
 0.1486 
 
 9.34 
 
 50 
 
 0.1489 
 
 0.0003+ 
 
 A glance at the table shows that the results are most 
 irregular. In Table II is seen the result of regulating the 
 temperature and the length of standing after the addition of 
 the sodium thiosulphate. 
 
 * Amer. Chem. Jour., i, 321. 
 
 OF THE 
 
 UNIVERSITY 
 
 OF 
 
 s*UFC 
 
BY SODIUM THIOSULPHATE. 
 TABLE II. 
 
 331 
 
 Exp. 
 
 HgCl, 
 taken as 
 Hg. 
 
 Volume 
 at be- 
 
 ginning* 
 
 Temper- 
 ature. 
 
 Stand- 
 ing. 
 
 Na^O, 
 excess. 
 
 HgCl, as 
 Hg 
 found. 
 
 Error. 
 
 
 gnn. 
 
 cm 3 
 
 C. 
 
 minutes. 
 
 cm 3 
 
 grin. 
 
 grin. 
 
 (1) 
 
 0.0738 
 
 50 
 
 36 
 
 40 
 
 16.68 
 
 0.0494 
 
 0.0244- 
 
 2) 
 (3) 
 
 0.0741 
 0.0741 
 
 60 
 75 
 
 70 
 70 
 
 15 
 12 
 
 15.42 
 16.07 
 
 0.0738 
 0.0733 
 
 0.0003- 
 0.0008- 
 
 (4) 
 
 0.0744 
 
 50 
 
 70 
 
 10 
 
 14.6 
 
 0.0755 
 
 0.0011+ 
 
 (5) 
 
 0.0764 
 
 50 
 
 72 
 
 7 
 
 6.77 
 
 0.0771 
 
 0.0007+ 
 
 (6) 
 
 0.0762 
 
 50 
 
 75 
 
 10 
 
 8.54 
 
 0.0799 
 
 0.0037+ 
 
 ( 7 ) 
 
 0.0756 
 
 60 
 
 73 
 
 15 
 
 9.99 
 
 0.0815 
 
 0.0059+ 
 
 8 
 
 0.0774 
 
 50 
 
 68 
 
 15 
 
 10.84 
 
 0.0767 
 
 0.0007- 
 
 (9) 
 
 0.0745 
 
 76 
 
 69 
 
 7 
 
 6.62 
 
 0.0805 
 
 0.0060+ 
 
 (10) 
 
 0.0736 
 
 50 
 
 68 
 
 5 
 
 15.82 
 
 0.0714 
 
 0.0022- 
 
 These results, although better than those of Table I, are 
 still very uncertain. On the supposition that the change 
 from white to black, which takes place in the solution after 
 the addition of an excess of sodium thiosulphate more or less 
 quickly according to the temperature, was due to an increased 
 amount of HgS in the compound 2HgS . HgCl 2 , the next 
 step was to ascertain whether this could be avoided by 
 stopping the addition of the thiosulphate at the first indica- 
 tion of a change of color in the white precipitate, diluting 
 the solution with a large amount of cold water and immed- 
 iately throwing it on the filter. Table III shows the result 
 of the experiments. 
 
 TABLE III. 
 
 Exp. 
 
 HgCl 2 
 taken as 
 Hg. 
 
 Volume 
 at begin- 
 ing. 
 
 Tempera- 
 ture. 
 
 Na 2 S 2 3 
 in 
 excess. 
 
 HgCl, 
 found as 
 Hg. 
 
 Error. 
 
 
 gnn. 
 
 cm 3 
 
 C. 
 
 cm 8 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.0749 
 
 50 
 
 70 
 
 4.15 
 
 0.0751 
 
 0.0002+ 
 
 (2) 
 
 0.0749 
 
 50 
 
 75 
 
 0.72 
 
 0.0728 
 
 0.0021- 
 
 (3) 
 
 0.0756 
 
 50 
 
 72 
 
 1.46 
 
 0.0759 
 
 0.0003+ 
 
 (4) 
 
 0.0753 
 
 60 
 
 70 
 
 2.57 
 
 0.0750 
 
 0.0003- 
 
 (5) 
 
 0.0390 
 
 50 
 
 70 
 
 3.43 
 
 0.0396 
 
 0.0005+ 
 
 (6) 
 
 0.0388 
 
 60 
 
 72 
 
 8.19 
 
 0.0390 
 
 0.0002+ 
 
 (7) 
 
 0.0380 
 
 60 
 
 76 
 
 2.03 
 
 0.0393 
 
 0.0013+ 
 
 (8 
 
 0.1494 
 
 50 
 
 78 
 
 4.99 
 
 0.1498 
 
 0.0004+ 
 
 (9 
 (10) 
 
 0.1489 
 0.1480 
 
 150 
 60 
 
 78 
 70 
 
 4.38 
 0.52 
 
 0.1512 
 0.1438 
 
 0.0023+ 
 0.0042- 
 
 (11) 
 
 0.1498 
 
 50 
 
 78 
 
 1.47 
 
 0.1540 
 
 0.0042+ 
 
 (12) 
 
 0.1484 
 
 50 
 
 71 
 
 2.09 
 
 0.1517 
 
 0.0033+ 
 
 (13) 
 
 0.1480 
 
 76 
 
 72o 
 
 1.59 
 
 0.1509 
 
 0.0029+ 
 
332 
 
 THE TITRATION OF MERCURY 
 
 In the case of quantities of mercuric chloride up to 0.1 
 grm. the results shown in Table III are very satisfactory, but 
 when larger amounts of mercuric chloride are used the errors 
 again become prominent. In Table IV, the effect of lowering 
 the temperature to 60 C. and of increasing the dilution to 
 100 cm 3 is shown. 
 
 TABLE IV. 
 
 Exp. 
 
 HgCl 2 
 taken as 
 
 Volume 
 at begin- 
 
 Temper- 
 ature. 
 
 Na 2 S 2 O s 
 in 
 
 HgCL 
 found as 
 
 Error. 
 
 
 Hg. 
 
 ing. 
 
 
 excess. 
 
 Hg. 
 
 
 
 grm. 
 
 cm 3 
 
 C. 
 
 cm 3 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.0759 
 
 100 
 
 60 
 
 3.06 
 
 0.0766 
 
 0.0007+ 
 
 (2) 
 
 0.0384 
 
 100 
 
 60 
 
 2.81 
 
 0.0387 
 
 0.0003+ 
 
 (3) 
 
 0.1492 
 
 100 
 
 60 
 
 1.1 
 
 0.1500 
 
 0.0008+ 
 
 w 
 
 0.1503 
 
 100 
 
 60 
 
 1.63 
 
 0.1506 
 
 0.0003+ 
 
 5 
 
 0.1479 
 
 100 
 
 60 
 
 2.41 
 
 0.1480 
 
 0.0001+ 
 
 n 
 
 0.1489 
 
 100 
 
 60 
 
 2.12 
 
 0.1503 
 
 0.0014+ 
 
 (7) 
 
 0.2244 
 
 100 
 
 60 
 
 2.63 
 
 0.2259 
 
 0.0015+ 
 
 (8 
 
 0.1490 
 
 100 
 
 60 
 
 2.33 
 
 0.1484 
 
 0.0006- 
 
 (9 
 (10) 
 
 0.0758 
 0.0383 
 
 100 
 100 
 
 60 
 60 
 
 2 
 2.58 
 
 0.0762 
 0.0379 
 
 0.0004+ 
 0.0004- 
 
 From this table it is plain that Scherer's process for the 
 estimation of mercury in the form of mercuric chloride is 
 capable of yielding accurate results if carried out under 
 certain fixed conditions. These conditions, which must be 
 closely adhered to, are as follows: The solution containing 
 the mercury in the form of mercuric chloride is placed in a 
 liter flask, diluted to 100 cm 3 and heated to a temperature 
 of 60 C. The sodium thiosulphate in ^ normal solution 
 is run in from a burette until the white precipitate formed 
 begins to take on a brownish tinge. The solution is then 
 diluted with cold water, some asbestos fiber added to coagu- 
 late the precipitate and the whole is quickly thrown on the 
 filter. After careful washing of the precipitate, the filtrate 
 is diluted to a definite volume, 3 grm. of potassium iodide 
 added and the excess thiosulphate titrated with iodine and 
 starch solution. The duration of the process need not 
 exceed 15 minutes. It is worthy of note that there is no 
 necessity of using any hydrochloric acid in addition to that 
 
BY SODIUM THIOSULPHATE. 
 
 333 
 
 formed in the reaction. This certainly eliminates one prob- 
 able source of error the interaction of hydrochloric acid 
 and sodium thiosulphate. 
 
 In dealing with the estimation of mercury in the form of 
 mercurous nitrate the same procedure was employed as in the 
 case of mercuric chloride. A solution of mercurous nitrate 
 was prepared by dissolving as much as possible of 20 grm. 
 of the salt hi about 200 cm 3 of water, filtering off the clear 
 liquid and diluting to a definite volume. The standard of 
 the solution was determined by precipitation as metallic 
 mercury by means of the electric current. Contrary to the 
 statement made in Sutton, the brown precipitate of Hg 2 S, 
 formed as shown in the equation, 
 
 Hg 2 (N0 8 ) 2 + Na 2 S 2 8 = Hg 2 S + Na 2 S0 4 + N 2 6 , 
 
 does not settle and leave a clear supernatant liquid, but the 
 solution remains cloudy and it is impossible to see any end 
 reaction. Although the conditions of dilution, temperature 
 and amount of acid present were carefully considered, no 
 arrangement or adjustment of these conditions was found 
 
 TABLE V. 
 
 Exp. 
 
 Hg 2 (N0 3 ) a 
 taken 
 
 as Hg. 
 
 HNO S 
 1:3. 
 
 Volume 
 at be- 
 ginning. 
 
 Temper- 
 ature. 
 
 Na.8,0, 
 
 in excess. 
 
 Hg 2 (NO s ) 2 
 found 
 asHg. 
 
 Error. 
 
 
 
 grin. 
 
 cms 
 
 cm* 
 
 C. 
 
 cm 3 
 
 grm. 
 
 grm. 
 
 
 (1) 
 
 0.0148 
 
 None. 
 
 60 
 
 60 
 
 4.28 
 
 0.0129 
 
 0.0019- 
 
 
 (2) 
 
 0.0148 
 
 None. 
 
 75 
 
 60 
 
 4.45 
 
 0.0117 
 
 0.0031- 
 
 
 ( 3 ) 
 
 0.2976 
 
 None. 
 
 300 
 
 60 
 
 12.67 
 
 0.2760 
 
 0.0216 
 
 
 (4) 
 
 0.1488 
 
 None. 
 
 100 
 
 40 
 
 2.19 
 
 0.1386 
 
 0.0102- 
 
 
 (5) 
 
 0.1488 
 
 None. 
 
 100 
 
 50 
 
 6.92 
 
 0.1378 
 
 0.0110- 
 
 
 6) 
 
 0.1488 
 
 None. 
 
 200 
 
 50 
 
 0.78 
 
 0.1388 
 
 0.0100- 
 
 
 7 
 
 0.0744 
 
 None. 
 
 100 
 
 65 
 
 0.73 
 
 0.0636 
 
 0.0108- 
 
 
 8 
 
 0.0744 
 
 1 
 
 100 
 
 55 
 
 0.49 
 
 0.0733 
 
 0.0011- 
 
 
 9 
 
 0.0744 
 
 2 
 
 100 
 
 40 
 
 1.16 
 
 0.0660 
 
 0.0084- 
 
 
 10) 
 
 0.0744 
 
 1 
 
 100 
 
 65 
 
 1.35 
 
 0.0685 
 
 0.0059- 
 
 
 11) 
 
 0.0744 
 
 1 
 
 100 
 
 65 
 
 0.82 
 
 0.0686 
 
 0.0058- 
 
 
 12) 
 
 0.0744 
 
 i 
 
 200 
 
 55 
 
 1.71 
 
 0.0686 
 
 0.0059- 
 
 
 13 
 
 0.0744 
 
 
 100 
 
 40 
 
 1.77 
 
 0.0649 
 
 0.0095- 
 
 
 14) 
 
 0.0744 
 
 5* 
 
 100 
 
 40 
 
 1.34 
 
 0.0645 
 
 0.0099- 
 
 
 15) 
 
 0.0744 
 
 1 
 
 100 
 
 45 
 
 1.71 
 
 0.0654 
 
 0.0090- 
 
 16) 
 
 0.0744 
 
 10 
 
 100 
 
 45 
 
 1.55 
 
 0.0669 
 
 0.0075- 
 
 (17) 
 
 0.1488 
 
 1 
 
 100 
 
 40 
 
 0.73 
 
 0.1391 
 
 0.0097- 
 
334 THE TITRATION OF MERCURY 
 
 under which satisfactory results could be obtained. Table V 
 gives the result of experiments. 
 
 The errors in experiments (1), (3) to (6), (10) to (12), 
 and (17) are, proportionally to the amount of material handled, 
 practically the same and this fact caused me to make a careful 
 revision of all standards; but no mistake could be found. 
 The reaction upon which the process depends requires the 
 formation of Hg 2 S, but this mercurous sulphide breaks down 
 immediately into mercuric sulphide and mercury. The latter 
 is probably acted upon by the free nitric acid present to 
 form mercuric nitrate, which in turn is transformed into the 
 compound 2HgS . Hg(NO 3 ) 2 by the action of the thiosulphate. 
 At any rate, with an error so large, whatever its source may 
 be, the process is plainly impracticable. 
 
 The third step in Scherer's process deals with the action of 
 sodium thiosulphate on mercuric nitrate. In the following 
 experiments a solution of mercuric nitrate was prepared 
 either by dissolving as far as possible 20 grm. of mercuric 
 nitrate hi about 200 cm 3 of cold water, filtering off the 
 supernatant liquid and diluting to a definite volume (l)-(8), 
 or by dissolving the salt in strong nitric acid and diluting 
 (9)-(19). The standard of the solution was obtained by 
 precipitation as metallic mercury by means of the electric cur- 
 rent. The yellow precipitate, formed according to Scherer's 
 reaction, 
 
 3Hg(N0 3 ) 2 + 2Na 2 S 2 O s = 2HgS . Hg(ST0 3 ) 2 + 2Na 2 S0 4 + 2N a 6 , 
 
 on adding the sodium thiosulphate settles much better than 
 in the case of either mercuric chloride or mercurous nitrate ; 
 but, as the supernatant liquid takes on a permanent yellow 
 color towards the end of the reaction, it is impossible to see 
 when the thiosulphate produces no further precipitation. On 
 this account, therefore, the same procedure was adopted as 
 in the case of the mercuric chloride and mercurous nitrate, 
 i. e., filtration and titration of the excess of sodium thio- 
 sulphate with iodine and starch solution. The result of the 
 experiments is shown in the following table. 
 
BY SODIUM THIOSULPHATE. 
 TABLE VL 
 
 335 
 
 Bxp. 
 
 Hg(N0 3 ), 
 taken as 
 
 Volume 
 at begin- 
 
 HN0 8 
 1 :3 
 
 Temper- 
 
 NA 2 S 2 3 
 in 
 
 HgfNOs), 
 found as 
 
 Error. 
 
 
 Hg. 
 
 ing. 
 
 
 
 excess. 
 
 Hg. 
 
 
 
 grm. 
 
 cms 
 
 cm 3 
 
 C. 
 
 cm 8 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.1167 
 
 200 
 
 None. 
 
 60 
 
 0.46 
 
 0.1384 
 
 0.0217+ 
 
 (2) 
 
 0.1167 
 
 100 
 
 1 
 
 60 
 
 3.63 
 
 0.1348 
 
 0.0181+ 
 
 (3) 
 
 0.1167 
 
 100 
 
 2 
 
 60 
 
 0.23 
 
 0.1375 
 
 0.0208+ 
 
 (4) 
 
 0.1167 
 
 100 
 
 None. 
 
 60 
 
 0.17 
 
 0.1360 
 
 0.0193+ 
 
 (5) 
 
 0.1167 
 
 300 
 
 None. 
 
 21 
 
 0.12 
 
 0.1232 
 
 0.0065+ 
 
 (6) 
 
 0.1167 
 
 300 
 
 None. 
 
 21 
 
 0.15 
 
 0.1375 
 
 0.0208+ 
 
 (7) 
 
 0.1167 
 
 200 
 
 None. 
 
 21 
 
 11.8 
 
 0.1461 
 
 0.0294+ 
 
 (8) 
 
 0.1167 
 
 200 
 
 20 
 
 21 
 
 15.97 
 
 0.1403 
 
 0.0236+ 
 
 (9) 
 
 0.1278 
 
 200 
 
 None. 
 
 21 
 
 5.23 
 
 0.1647 
 
 0.0369+ 
 
 (10) 
 
 0.1278 
 
 200 
 
 None. 
 
 21 
 
 1.35 
 
 0.1653 
 
 0.0375+ 
 
 (11) 
 
 0.1278 
 
 100 
 
 None. 
 
 21 
 
 1.43 
 
 0.1662 
 
 0.0384+ 
 
 (12) 
 
 0.0752 
 
 200 
 
 None. 
 
 21 
 
 2.09 
 
 0.0996 
 
 0.0244+ 
 
 (13) 
 
 0.0255 
 
 100 
 
 None. 
 
 21 
 
 2.01 
 
 0.0280 
 
 0.0025+ 
 
 (14) 
 
 0.0256 
 
 200 
 
 5 
 
 210 
 
 3.44 
 
 0.0264 
 
 0.0009+ 
 
 (15) 
 
 0.0255 
 
 200 
 
 10 
 
 21 
 
 0.85 
 
 0.0334 
 
 0.0079+ 
 
 (16) 
 
 0.0255 
 
 300 
 
 None. 
 
 21 
 
 1.96 
 
 0.0264 
 
 0.0009+ 
 
 (17) 
 
 0.0255 
 
 200 
 
 None. 
 
 21 
 
 1.9 
 
 0.0287 
 
 0.0032+ 
 
 (18) 
 
 0.0255 
 
 200 
 
 None. 
 
 21 
 
 1.76 
 
 0.0290 
 
 0.0035+ 
 
 (19) 
 
 0.0639 
 
 200 
 
 None. 
 
 60 
 
 1.53 
 
 0.0831 
 
 0.0192+ 
 
 These results seem to show the impossibility of obtaining 
 accurate results according to Scherer's method for the deter- 
 mination of mercuric nitrate by direct titration with sodium 
 thiosulphate. The constant plus error cannot be accounted 
 for on the hypothesis that the nitric acid present decomposes 
 the sodium thiosulphate, for in that case the error would 
 lie in the other direction. It is more probable that the 
 constitution of the compound 2HgS . Hg(NO 3 ) 2 is not definite 
 enough to make it the basis for an analytical process. 
 
XLIH 
 
 THE IODOMETRIC ESTIMATION OF ARSENIC 
 
 ACID. 
 
 BY F. A. GOOCH AND JULIA C. MOKRIS.* 
 
 THE interaction of a soluble arseniate and a soluble iodide in 
 a suitably acidulated solution results, as is well known, in the 
 reduction of the arsenic acid (more or less completely according 
 to conditions of temperature and proportions of reagents and 
 solvents) with the corresponding liberation of two atoms of 
 iodine for every molecule of arsenic acid (H 8 O 3 AsO) reduced. 
 Inasmuch, however, as the reaction of this process is reversible, 
 it is necessary, hi order that the reduction may be complete, 
 to nullify the oxidizing action of the iodine liberated. 
 Theoretically this end may be accomplished in either of two 
 ways, by volatilizing the free iodine bodily or by destroying 
 the oxidizing power of the iodine by converting it to hydriodic 
 acid. The former method was followed in a processs devised 
 for the estimation of arsenic acid and elaborated in this 
 laboratory.! This method, as originally put forward, consisted 
 in adding to the solution of the arseniate potassium iodide in 
 excess of the amount theoretically indicated, with 10 cm 3 of 
 sulphuric acid of half strength, and so arranging the dilution 
 that the total volume of the liquid should be about 100 cm 3 , 
 boiling until the volume decreased to 40 cm 3 , bleaching by 
 the cautious addition of sulphurous acid the trace of free 
 iodine still held by the hydriodic acids, diluting, cooling, 
 neutralizing with acid potassium carbonate, and titrating with 
 iodine, after adding the starch indicator. This process, 
 depending upon the removal by volatilization of all but the 
 
 * From Am. Jour. Sci., x, 151. 
 
 t Gooch and Browning, Am. Jour. Sci., xl, 66. Volume I, p. 30. 
 
ESTIMATION OF ARSENIC ACID. 337 
 
 last traces of liberated iodine and the conversion of this 
 minute residue by sulphurous acid, involves no secondary- 
 reactions of a sort likely to influence the main effect. It is 
 exact and fairly rapid. 
 
 The method of Williamson,* brought forward more recently, 
 depends upon the conversion of the liberated iodine to 
 hydriodic acid. The interaction at ordinary temperatures of a 
 suitably strong acid, hydrochloric or sulphuric acid, upon the 
 mixture of the arseniate and iodide sets free iodine, and the 
 liberated iodine is converted to hydriodic acid by the action of 
 sodium thiosulphate, the end point being the disappearance 
 of the iodine color. 
 
 According to Williamson's directions, 25 cm 3 portions of the 
 solution of the arseniate are treated with potassium iodide and 
 mixed with an equal volume of hydrochloric acid of sp. gr. 
 1.16. The precaution is recommended that the strength of 
 the solution of the arseniate shall not exceed the decinormal 
 value, in order that the dilution consequent upon titration by 
 the thiosulphate may not be too great the reducing action 
 brought about by the action of the strong acid upon the 
 arseniate and iodide being reversible upon the dilution of liquid 
 with water. This procedure thus limits the process to the 
 determination of about 0.18 grm. of arsenic acid in 25 cm 3 of 
 the solution to be treated with an equal volume of hydrochloric 
 acid of sp. gr. 1.16. Obviously, however, the process should, 
 so far as the reduction is concerned, be applicable to larger 
 amounts of arsenic provided the strength of the acid is kept up 
 proportionately. It is essential that the liquid at the end of 
 the titration should contain approximately ten per cent of its 
 mass of absolute hydrochloric acid or about one-third of its 
 volume of the aqueous acid of sp. gr. 1.16. 
 
 The arsenic acid is measured either by the amount of 
 standard thiosulphate required to bleach the iodine or by the 
 amount of iodine required to reoxidize the reduced arsenious 
 acid, after neutralizing with acid potassium carbonate. If the 
 former alternative is followed, the end-reaction must be the 
 
 * Jour. Soc. Dyers and Colorists, 1896, 86-89. 
 VOL. ii. 22 
 
338 
 
 THE IODOMETRIC ESTIMATION 
 
 disappearance of the yellow color of the iodine, since in 
 solutions so strongly acid it is impossible to place dependence 
 upon the starch indicator; in using the latter alternative, the 
 starch indicator is, of course, permissible and preferable. 
 
 In the direct titration of the iodine by thiosulphate two 
 sources of error present themselves as possibilities ; first, the 
 excessive liberation of iodine by the action of air upon the 
 strongly acidulated iodide; and second, the liability of the 
 thiosulphate,* if present even in momentary or local excess 
 during the process of titration, to break down under the action 
 of strong acid, thus changing its capacity to convert iodine to 
 hydriodic acid. The latter contingency should be remote in 
 proportion to the caution used in adding the thiosulphate and 
 in keeping the liquid well stirred ; the former must of necessity 
 vary with the acidity of the solution containing the iodide, the 
 time of exposure to atmospheric action, and the degree of 
 contact with the air incidental to stirring. We have thought 
 it desirable, therefore, to see how far each of these possibilities 
 is likely to interfere in the practical conduct of an ordinary 
 analysis. 
 
 HCl 
 (sp. gr. 1.16) 
 taken. 
 
 KI 
 
 taken. 
 
 Total 
 volume. 
 
 Na-jSaOg added at 
 once. In terms 
 of H 3 O 3 AsO. 
 
 Na 2 S 2 3 added 
 after 5 minutes. 
 In terms of 
 H 3 3 AsO. 
 
 Na 2 S 2 3 added 
 after stirring 
 5 minutes. In 
 terms of 
 H 3 O 3 AsO. 
 
 cms 
 
 grin. 
 
 cm 8 
 
 grm. 
 
 grin. 
 
 grm. 
 
 25 
 
 2 
 
 50 
 
 0.0013 
 
 . 
 
 . 
 
 25 
 
 2 
 
 75 
 
 0.0004 
 
 t 
 
 
 25 
 
 2 
 
 50 
 
 
 0]0035 
 
 
 25 
 
 2 
 
 75 
 
 
 0.0019 
 
 
 25 
 
 2 
 
 50 
 
 , 
 
 . 
 
 0^0042 
 
 25 
 
 2 
 
 75 
 
 
 . 
 
 0.0021 
 
 50 
 
 2 
 
 100 
 
 o!ooi7 
 
 . 
 
 ( 
 
 60 
 
 2 
 
 150 
 
 0.0004 
 
 . 
 
 
 
 50 
 
 2 
 
 100 
 
 
 o!o035 
 
 t 
 
 50 
 
 2 
 
 150 
 
 
 0.0019 
 
 ( 
 
 50 
 
 2 
 
 100 
 
 
 
 . 
 
 o!o035 
 
 50 
 
 2 
 
 150 
 
 
 . . . 
 
 0.0014 
 
 The effects likely to result simply from the strong acidifica- 
 tion of the solution containing potassium iodide and their 
 variation for conditions of dilution representing the beginning 
 * Norton, Am. Jour. Sci., vii, 287. This volume, p. 206. 
 
OF ARSENIC ACID. 
 
 339 
 
 and the end of a titration on the lines laid down are shown in 
 the preceding table. The solution of potassium iodide was 
 diluted as indicated before the addition of the acid and the 
 iodine set free was titrated by thiosulphate. The proportion- 
 ate strength of acid and the time before titration are, obviously, 
 the essential factors. The absolute amount of acid present 
 and the stirring seem to make little difference. 
 
 As to the action of the hydrochloric acid on small amounts 
 of the thiosulphate, we have the evidence of the experiments 
 detailed in the following statements in which 1, 2, and 5 cm 3 
 of nearly ^ thiosulphate are exposed to the action of 25 cm 3 
 hydrochloric acid, sp. gr. 1.16, without dilution or diluted with 
 an equal volume of water, were titrated with nearly ^ iodine. 
 The condition of acidity when the volume of 50 cm 3 contains 
 25 cm 3 of hydrochloric acid, sp. gr. 1.16, is that of the 
 beginning of titration of Williamson's process. In order that 
 the effect of error due to such action upon the determination 
 of arsenic acid may appear immediately, the thiosulphate and 
 iodine used are expressed in terms of that acid. 
 
 
 
 
 
 Error of 
 
 Iodine to 
 
 Error of 
 
 HCl 
 
 Tiff 
 
 Volume 
 before 
 titration. 
 
 Na-jS-jOg nearly 
 ^ in terms 
 of H 3 O 3 AsO. 
 
 Iodine to 
 color with- 
 out dilution, 
 in terms of 
 
 titration 
 without 
 dilution, 
 in terms of 
 
 color after 
 diluting 
 to 75 cm 8 , 
 in terms of 
 
 titration 
 after 
 dilution, in 
 terms of 
 
 
 
 
 H 8 O 3 AsO. 
 
 H 3 O 3 AsO. 
 
 H 3 O 3 AsO. 
 
 H 8 8 AsO. 
 
 cms 
 
 cm 3 
 
 cm3 
 
 gnn. 
 
 grm. 
 
 grm. 
 
 gnu. 
 
 grm. 
 
 25 
 
 26 
 
 1 
 
 0.0071 
 
 0.0062 
 
 0.0009- 
 
 0.0071 
 
 0.0000 
 
 25 
 
 50 
 
 1 
 
 0.0071 
 
 0.0071 
 
 0.0000 
 
 0.0071 
 
 0.0000 
 
 25* 
 
 50 
 
 1 
 
 0.0071 
 
 0.0079 
 
 0.0008+ 
 
 0.0079 
 
 0.0008+ 
 
 25 
 
 50 
 
 2 
 
 0.0141 
 
 0.0146 
 
 0.0005+ 
 
 0.0146 
 
 0.0005+ 
 
 25* 
 
 50 
 
 2 
 
 0.0141 
 
 0.0157 
 
 0.0016+ 
 
 0.0157 
 
 0.0016+ 
 
 25 
 
 30 
 
 5 
 
 0.0353 
 
 0.0336 
 
 0.0017- 
 
 0.0374 
 
 0.0021+ 
 
 25 
 
 60 
 
 6 
 
 0.0353 
 
 0.0359 
 
 0.0006+ 
 
 0.0359 
 
 0.0006+ 
 
 25* 
 
 50 
 
 5 
 
 0.0353 
 
 0.0411 
 
 0.0058+ 
 
 0.0411 
 
 0.0058+ 
 
 These two sources of error, the one due to a liberation of 
 iodine and the other due to decomposition of the thiosulphate, 
 would naturally tend to overcome one another, but the com- 
 pleteness of such neutralization would naturally be largely a 
 
 * In these experiments the acid stood in contact with the thiosulphate 
 5 minutes before titration. 
 
340 
 
 THE IODOMETRIC ESTIMATION 
 
 matter of chance in the varying conditions of actual analysis. 
 The experiments of the following table, however, in which 
 ^ thiosulphate, to the amount of 1, 2, and 5 cm 8 , was added 
 to the liquid, 50 cm 3 and 75 cm 3 , containing 25 cm 3 acid, and 
 titrated with iodine at once, and after five minutes, were 
 made to test the matter for the conditions of dilution at the 
 beginning and at the end of a titration. 
 
 HCl 
 (sp. gr. 
 1.16). 
 
 KI. 
 
 Volume. 
 
 Na 2 S 2 O a nearly 
 JC. in terms 
 of H 3 3 AsO. 
 
 Iodine in 
 terms of 
 H 3 O s AsO, 
 at once. 
 
 Iodine in 
 terms of 
 H 3 O 3 AsO, 
 after 5 min. 
 
 Error 
 in terms of 
 H 3 O s AsO. 
 
 cm 3 
 
 grm. 
 
 cm 8 
 
 cm 
 
 grm. 
 
 grm. 
 
 grm, 
 
 grm. 
 
 25 
 
 2 
 
 50 
 
 1 
 
 0.0071 
 
 0.0057 
 
 
 0.0014- 
 
 25 
 
 2 
 
 76 
 
 1 
 
 0.0071 
 
 0.0071 
 
 . 
 
 0.0000 
 
 25 
 
 2 
 
 50 
 
 2 
 
 0.0141 
 
 00131 
 
 t 
 
 0.0010- 
 
 25 
 
 2 
 
 76 
 
 2 
 
 0.0141 
 
 00143 
 
 . . 
 
 0.0002+ 
 
 25 
 
 2 
 
 50 
 
 5 
 
 0.0353 
 
 00332 
 
 f 
 
 0.0021- 
 
 25 
 
 2 
 
 75 
 
 5 
 
 0.0353 
 
 00357 
 
 . 
 
 0.0004+ 
 
 25 
 
 2 
 
 50 
 
 1 
 
 0.0071 
 
 
 0.0028 
 
 0.0048- 
 
 25 
 
 2 
 
 76 
 
 1 
 
 0.0071 
 
 
 0.0067 
 
 0.0004- 
 
 25 
 
 2 
 
 50 
 
 2 
 
 0.0141 
 
 t 
 
 0.0116 
 
 0.0025- 
 
 25 
 
 2 
 
 75 
 
 2 
 
 0.0141 
 
 t 
 
 0.0139 
 
 0.0002- 
 
 25 
 
 2 
 
 60 
 
 6 
 
 0.0353 
 
 
 0.0312 
 
 0.0041- 
 
 25 
 
 2 
 
 75 
 
 6 
 
 0.0353 
 
 
 
 0.0361 
 
 0.0008+ 
 
 It is clear that under the conditions covered by the experi- 
 ments of the two preceding tables the decomposition of the 
 thiosulphate is likely to occur in greater or less degree, and 
 that when the acid of sp. gr. 1.16 is not much diluted, the 
 products of decomposition are not oxidized by the iodine 
 completely. The latter observation is quite hi harmony 
 with the fact that sulphur dioxide bleaches iodine in strong 
 hydrochloric acid only slowly and incompletely. In such 
 cases dilution favors further action of the iodine, but results 
 obtained by titration with iodine in the acid solution diluted 
 with an equal amount of water are unmodified by further 
 dilution. 
 
 In the following tables are recorded actual determinations 
 of arsenic according to Williamson's process. To each 25 
 cm 3 of the arseniate were added 1, 2, or 3 grms. of potassium 
 iodide and 25 cm 3 hydrochloric acid, sp. gr. 1.16. The iodine 
 
OF ARSENIC ACID. 
 
 341 
 
 was bleached by nearly decinonnal thiosulphate without 
 addition of the starch indicator, which loses all delicacy in 
 the presence of strong acid. The time occupied by each 
 titration was about five minutes. The standards of the 
 arseniate were determined by the vaporization process,* the 
 purity of reagents employed in that process having been 
 proved by trying the process in the estimation of a solution 
 of arsenic acid made by oxidizing pure decinormal arsenious 
 acid by iodine. 
 
 
 
 Volume at 
 
 Volume 
 
 H 2 KAs0 4 
 
 
 
 HCL 
 
 KI. 
 
 beginning 
 of 
 
 at 
 
 end of 
 
 m 
 terms of 
 
 H s 3 AsO 
 found. 
 
 Error. 
 
 
 
 titration. 
 
 titration. 
 
 H 3 O 3 AsO. 
 
 
 
 cm 3 . 
 
 grin. 
 
 cm 8 . 
 
 cm s . 
 
 grin. 
 
 grm. 
 
 grm. 
 
 25 
 
 2 
 
 50 
 
 51 
 
 0.0062 
 
 0.0085 
 
 0.0023+ 
 
 25 
 
 2 
 
 50 
 
 52 
 
 0.0125 
 
 0.0156 
 
 0.0031+ 
 
 25 
 
 2 
 
 50 
 
 55 
 
 0.0312 
 
 0.0350 
 
 0.0038+ 
 
 25 
 
 2 
 
 50 
 
 55 
 
 0.0624 
 
 0.0666 
 
 0.0042+ 
 
 25 
 
 2 
 
 50 
 
 73 
 
 0.1559 
 
 0.1588 
 
 0.0029+ 
 
 25 
 
 2 
 
 60 
 
 73 
 
 0.1559 
 
 0.1587 
 
 0.0028+ 
 
 25 
 
 2 
 
 60 
 
 73 
 
 0.1559 
 
 0.1591 
 
 0.0032+ 
 
 25 
 
 2 
 
 50 
 
 73 
 
 0.1559 
 
 0.1595 
 
 0.0036+ 
 
 25 
 
 3 
 
 50 
 
 73 
 
 0.1559 
 
 0.1595 
 
 0.0036+ 
 
 25 
 
 1 
 
 50 
 
 73 
 
 0.1559 
 
 0.1581 
 
 0.0022+ 
 
 25 
 
 2 
 
 50 
 
 73 
 
 0.1559 
 
 0.1581 
 
 0.0022+ 
 
 25 
 
 2 
 
 50 
 
 73 
 
 0.1559 
 
 0.1588 
 
 0.0029+ 
 
 The range of error in these results is from -f- 0.0022 grm. 
 to + 0.0042 grm. with a mean of + 0.0031 grm. not very 
 different from what might be expected from the effect of the 
 interaction of the strong hydrochloric acid and the iodide 
 alone. The counter-effect due to the decomposition of the 
 thiosulphate is not large, yet it is probably real, as will appear 
 in the sequel. In the following series of determinations, 
 made with new solutions and new standards throughout, the 
 arsenic acid was determined, first, by titrating the iodine set 
 free by 25 cm 3 of hydrochloric acid, sp. gr. 1.16 and 3 grms. 
 potassium iodide, the solution having a total volume of 50 cm 8 
 at beginning and of 75 cm 3 at the end of titration and, 
 secondly, the arsenious acid produced in the first reaction 
 was titrated, after being neutralized with acid potassium car- 
 bonate by iodine in the presence of the starch indicator. 
 * Gooch and Browning, loc. cit. 
 
342 
 
 THE IODOMETRIC ESTIMATION 
 
 H 2 KAs0 4 
 
 taken, in 
 terms of 
 H 3 O 3 AsO. 
 
 H 3 3 A80 
 found by the 
 thioeulphate. 
 
 Error. 
 
 H 3 3 A80 
 
 found by 
 titration of 
 H S O 3 A8 with 
 iodine. 
 
 Error. 
 
 grm. 
 
 gnn. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 0.1767 
 
 0.1798 
 
 0.0031+ 
 
 0.1776 
 
 0.0009+ 
 
 0.1767 
 
 0.1798 
 
 0.0031+ 
 
 0.1777 
 
 0.0010+ 
 
 0.1767 
 
 0.1795 
 
 0.0028+ 
 
 0.1785 
 
 0.0018+ 
 
 0.1767 
 
 0.1793 
 
 0.0026+ 
 
 0.1785 
 
 0.0018+ 
 
 0.1767 
 
 0.1794 
 
 0.0027+ 
 
 0.1780 
 
 0.0013+ 
 
 0.1767 
 
 0.1798 
 
 0.0031+ 
 
 0.1785 
 
 0.0018+ 
 
 The average error of the first operation is 0.0029 grm., not 
 far from that of the previous series ; the error of the second 
 operation, the titration of the arsenious acid, amounts on the 
 average to 0.0014 grm. In the second operation the error 
 due to over-use of the thiosulphate by iodine set free outside 
 the main reaction is obviously eliminated. The tetrathionate 
 present after neutralization with acid potassium carbonate is 
 unaffected by iodine, as we have found by titrating 25 cm 3 ^ 
 iodine mixed with 25 cm 3 hydrochloric acid, sp. gr. 1.16, by 
 the thiosulphate, neutralizing with acid potassium carbonate,* 
 adding starch and getting the starch blue with a single drop 
 of -^ iodine. The average error of this process, therefore, 
 0.0014, is probably due to the products of decomposition of 
 the thiosulphate in the first operation. 
 
 From the foregoing experiments it is clear that an arbitrary 
 correction of about 0.0030 grm. must be deducted from the 
 indications of Williamson's process of direct titration by 
 thiosulphate, made with the greatest care under the conditions 
 mentioned ; and that a correction varying from one-half that 
 amount (0.0015 grm.) to nothing (according to the amount 
 of arsenious acid present) when the determination is made 
 by iodine after neutralization with acid potassium carbonate. 
 
 * It is worthy of note, that, as we have found by experience, it is not 
 possible to substitute an alkaline hydroxide for the carbonate in the early 
 stages of the process of neutralization, on account of the decomposing effect 
 of the former reagent upon the tetrathionate. This effect is in proportion to 
 the heating of the solution, but is never wholly absent even when ice is 
 intermixed with the liquid and the greatest care taken to prevent a rise of 
 temperature. 
 
OF ARSENIC ACID. 
 
 343 
 
 After making these arbitrary corrections in the results of the 
 preceding table, the individual variations fall within reason- 
 able limits. 
 
 On the other hand, the vaporization process, in which 
 the arseniate is reduced by boiling with sulphuric acid and 
 potassium iodide in the manner described,* gives indications 
 reasonably regular and accurate without the application of 
 an arbitrary correction. This process, moreover, may be 
 shortened by restricting the volume at which heating begins 
 so that the boiling need not be extended beyond five or six 
 minutes. According to this slight modification, the solution 
 of the arseniate is heated in an Erlenmeyer flask with potas- 
 sium iodide to an amount about 0.5 grm. in excess of the 
 amount theoretically required and 10 cm 3 of sulphuric acid 
 of half strength in a total volume of between 50 cm 3 and 
 75 cm 3 . The liquid is boiled till the iodine vapors are no 
 longer visible in the flask above the liquid, the iodine color 
 in the still hot liquid is bleached by the cautious addition of 
 sulphurous acid, the whole is diluted with cold water, and 
 cooled quickly. The solution is nearly neutralized with 
 potassium hydroxide and the neutralization is completed with 
 acid potassium carbonate. The reduced acid is titrated with 
 iodine after adding the starch indicator. By this procedure 
 the results of the following table were obtained. 
 
 Volume. 
 
 H 8 8 ABO 
 taken. 
 
 H 8 3 AsO 
 found. 
 
 Error. 
 
 cm 8 . 
 
 grm. 
 
 grm. 
 
 grm. 
 
 35 
 
 0.1559 
 
 0.1559 
 
 0.0000 
 
 35 
 
 0.1559 
 
 0.1560 
 
 0.0001+ 
 
 40 
 
 0.1559 
 
 0.1559 
 
 0.0000 
 
 65 
 
 0.1559 
 
 0.1559 
 
 0.0000 
 
 50 
 
 0.2495 
 
 0.2499 
 
 0.0004+ 
 
 50 
 
 0.2557 
 
 0.2449 
 
 0.0008- 
 
 60 
 
 0.3119 
 
 0.3117 
 
 0.0002- 
 
 60 
 
 0.3119 
 
 0.3120 
 
 0.0001+ 
 
 75 
 
 0.3119 
 
 0.3124 
 
 0.0005+ 
 
 75 
 
 0.3119 
 
 0.3132 
 
 0.0013+ 
 
 75 
 
 0.3119 
 
 0.3121 
 
 0.0002+ 
 
 75 
 
 0.3119 
 
 0.3115 
 
 0.0004- 
 
 75 
 
 0.3119 
 
 0.3124 
 
 0.0005+ 
 
 * Loc. cit 
 
XLIV 
 
 ON THE QUALITATIVE SEPARATION OF NICKEL 
 FROM COBALT BY THE ACTION OF AMMO- 
 NIUM HYDROXIDE ON THE FERRICYANIDES. 
 
 BY PHILIP E. BROWNING AND JOHN B. HARTWELL* 
 
 SOME years ago F. W. Clarke f suggested a method for the 
 separation of nickel from cobalt depending upon the solvent 
 action of ammonium hydroxide upon the precipitated fern- 
 cyanides. The method may best be described by quoting 
 from the original article : " To the slightly acid solution 
 containing the two metals, I first add an excess of ammonium 
 chloride. This causes the cobalt precipitate, which otherwise 
 would run through the filter, to fall in a denser state, and 
 also of a much darker color, often nearly black. I then add 
 the potassium ferricyanide until the precipitation is complete, 
 and afterwards agitate strongly with a considerable excess of 
 ammonia. Upon filtering, all the cobalt remains upon the 
 filter, being recognized by the characteristic color of the 
 precipitate, and the nickel is readily detected in the filtrate, 
 by means of ammonium sulphide. If, upon filtering, the 
 portion at first running through is turbid, it may be disre- 
 garded, or returned to the filter, that which filters through 
 subsequently being almost invariably clear." 
 
 In making a study of this method we found two serious 
 objections; first, the practical impossibility of obtaining a 
 good filtration from the cobalt ferricyanide, even in the pres- 
 ence of the ammonium chloride, and, second, the large amount 
 of sulphur thrown down when ammonium sulphide was 
 
 * From Am. Jour. Sci., x, 316. 
 t Am. Jour. Sci., xlviii, 67. 
 
SEPARATION OF NICKEL FROM COBALT. 345 
 
 added to the filtrate containing the nickel with the excess 
 of ferricyanide. 
 
 Our first attempt was to secure, if possible, a complete 
 separation of the precipitated cobalt ferricyanide and the 
 dissolved nickel by filtration. This we were able to accom- 
 plish by the addition of a small amount of a solution of an 
 aluminum salt to the original solution which held back the 
 cobalt, and, as experiment showed, allowed the complete 
 solvent action of the ammonium hydroxide upon the nickel 
 salt. Amounts of nickel as small as 0.0001 grm. were 
 detected, when mixed with the aluminum salt, by precipi- 
 tating as ferricyanide, extracting with ammonium hydroxide, 
 and testing in the manner to be described. 
 
 On turning our attention to a possible improvement hi the 
 method for the detection of the nickel, a reaction first dis- 
 cussed by Allen * was applied. When the ammoniacal solution 
 of the nickel ferricyanide was treated with strong sodium or 
 potassium hydroxide solution, in the presence of an excess 
 of potassium ferricyanide, a black flocky precipitate formed 
 which gave no test for ferro- or ferricyanide, and gave every 
 indication of being nickelic hydroxide. This reaction we 
 found to afford us a most delicate test for nickel. 
 
 The method as modified by us may be described as follows : 
 Dissolve not more than 0.1 grm. of the salts of the two ele- 
 ments hi about 5 cm 3 of water, add a few drops of a saturated 
 solution of alum, destroy any free mineral acid by neutraliz- 
 ing with ammonium hydroxide, and make faintly acid with 
 acetic acid. To this solution add about 0.5 grm. of potassium 
 ferricyanide and agitate to effect the solution of the ferri- 
 cyanide and the complete precipitation of the nickel and cobalt 
 salts. Then add about 5 cm 3 of strong ammonium hydroxide 
 and filter. To the filtrate, which should have no reddish 
 color, add a piece of sodium or potassium hydroxide about 
 the size of a pea and boil. The appearance of a black pre- 
 cipitate, showing first as a dark coloration in case of very 
 small amounts, indicates nickel. 
 
 * Chein. News, xxxiii, 290. 
 
346 
 
 SEPARATION OF NICKEL FROM COBALT. 
 
 The tables following give a record of the experimental 
 results. With the precautions indicated, this method may 
 be applied very satisfactorily. 
 
 
 
 
 
 I. 
 
 
 
 
 Exp. 
 
 CoSOJHjjO. 
 
 NiS0 4 . 
 7H 8 0. 
 
 KA1(S0 4 ) 2 
 saturated 
 solution. 
 
 K 3 FeC 8 N 6 
 
 NH 4 OH 
 
 concen- 
 trated. 
 
 NaOH solid. 
 
 Result. 
 
 
 griu. 
 
 gnn. 
 
 cm 
 
 grin. 
 
 cms 
 
 
 
 (1) 
 
 . . . 
 
 0.0100 
 
 2 
 
 0.5 
 
 5 
 
 About size 
 
 Heavy ppt. 
 
 
 
 
 
 
 
 of a pea. 
 
 
 (2) 
 
 . . , 
 
 0.0050 
 
 2 
 
 0.5 
 
 5 
 
 (i 
 
 Heavy ppt. 
 
 (3) 
 
 . 
 
 0.0010 
 
 2 
 
 0.5 
 
 5 
 
 
 
 Heavy ppt. 
 
 (4) 
 
 
 0.0003 
 
 2 
 
 0.5 
 
 5 
 
 <t 
 
 Distinct. 
 
 (5) 
 
 . . . 
 
 0.0001 
 
 2 
 
 0.5 
 
 5 
 
 a 
 
 Plain. 
 
 
 
 
 
 II. 
 
 
 
 
 (1) 
 
 0.10 
 
 . . . 
 
 2 
 
 0.5 
 
 5 
 
 About size 
 
 None. 
 
 
 
 
 
 
 
 of a pea. 
 
 
 (2) 
 
 0.10 
 
 0.0100 
 
 2 
 
 0.5 
 
 5 
 
 
 Heavy. 
 
 (3) 
 
 0.10 
 
 0.0050 
 
 2 
 
 0.5 
 
 5 
 
 
 Distinct. 
 
 (4) 
 
 0.10 
 
 0.0030 
 
 2 
 
 0.5 
 
 5 
 
 
 Very faint. 
 
 (5) 
 
 0.05 
 
 
 2 
 
 0.5 
 
 5 
 
 
 None. 
 
 (6) 
 
 0.05 
 
 0.0100 
 
 2 
 
 0.5 
 
 5 
 
 
 Heavy. 
 
 (7) 
 
 0.05 
 
 0.0050 
 
 2 
 
 0.5 
 
 5 
 
 
 Distinct. 
 
 (8) 
 
 0.05 
 
 0.0030 
 
 2 
 
 0.5 
 
 5 
 
 
 Plain. 
 
 (9) 
 
 0.05 
 
 0.0010* 
 
 2 
 
 0.5 
 
 5 
 
 
 Faint. 
 
 * Equivalent to 0.0002 of the metal. 
 
XLV 
 
 THE VOLUMETRIC ESTIMATION OF COPPER AS 
 THE OXALATE, WITH SEPARATION FROM 
 CADMIUM, ARSENIC, TIN, IRON, AND ZINC. 
 
 BY CHARLES A. PETERS* 
 
 IT is a well known fact that copper oxalate is insoluble in 
 water and scarcely attacked by moderate amounts of dilute 
 nitric acid.f Upon this fact Bornemann ^ has recently 
 based a method for the separation of copper from cadmium 
 by precipitating copper as the oxalate in the presence of 
 nitric acid, filtering hot, and estimating the copper after 
 ignition, by any of the well known gravimetric methods. Six 
 to ten grams of copper, as the oxide, were used for a single 
 determination, and the errors were large. Bornemann does 
 not recommend this process as an accurate analytical method. 
 Classen describes a method for the separation of metals as 
 oxalates by adding to the solution of the salt of the metals 
 a dilute solution of the potassium oxalate (1 : 6) and con- 
 centrated acetic acid to 80 per cent of the total volume. 
 Regarding copper salts in particular, Classen states that 
 precipitation takes place only in dilute solution and then not 
 completely. 
 
 It has been the experience of the writer, that the precipi- 
 tation of copper oxalate from solutions containing at least 
 0.0128 grm. of the oxide and saturated with the oxalic acid is 
 practically complete. The filtrate in such cases gives no 
 blue color with ammonia, looking down on a column of liquid 
 
 * From Am. Jour. Sci., x, 359. 
 
 t Storer, Dictionary of Chemical Solubilities, p. 463. 
 
 J Chem. Ztg., xxiii, 565. Ber. Dtsch. chem. Ges., x, 6, 1316. 
 
348 
 
 VOLUMETRIC ESTIMATION OF COPPER. 
 
 in a test-tube, and only a faint brown color is developed 
 when the nitrate is neutralized, made acid with acetic acid, 
 and tested with potassium ferrocyanide. It is the object of 
 this paper to show that moderate amounts of copper may be 
 determined quantitatively as the oxalate by precipitation with 
 oxalic acid and titration of the precipitate by potassium 
 permanganate, and also to show that moderate amounts of 
 copper may be separated from other metals in the presence 
 of nitric acid, by the addition of considerable amounts of 
 oxalic acid. 
 
 Before attempting the quantitative separation of copper 
 from solution by the addition of oxalic acid a few qualitative 
 experiments upon the precipitation of varying amounts of 
 copper sulphate by varying amounts of oxalic acid were tried 
 at different dilutions. In all the experiments the mixtures 
 stood 16-20 hours, and were filtered from 2 to 4 times 
 through four filters folded together, and the nitrates were 
 tested both with ammonia and with potassium ferrocyanide. 
 In cases where the filtrate gave no blue color with ammonia 
 
 TABLE I. 
 
 Dilution 50 cm 8 . 
 
 Ezp. 
 
 CuO 
 
 taken as 
 CuSO 4 . 
 
 Oxalic acid added in solution. 
 
 Oxalic acid added in crystalline 
 form. 
 
 Filtrate 
 treated 
 with NH 4 OH. 
 
 Filtrate 
 treated with 
 
 K 4 FeC 6 N 6 . 
 
 Filtrated 
 with 
 NH 4 OH. 
 
 Filtrate 
 treated with 
 E^FeCeNa. 
 
 2.0 grm. 
 oxalic 
 acid 
 present. 
 
 1.0 grm. 
 oxalic 
 acid 
 present. 
 
 0.5 grm. 
 oxalic 
 acid 
 present. 
 
 
 
 grm. 
 
 r o.ois 
 
 0.031 
 0.051 
 0.064 
 0.018 
 0.031 
 0.051 
 0.064 
 0.094 
 ' 0.018 
 0.031 
 0.051 
 0.064 
 0.094 
 
 Blue color. 
 Trace ' 
 
 Abundant ppt. 
 Abundant ppt. 
 Evident ppt. 
 Trace ppt. 
 Abundant ppt. 
 Abundant ppt. 
 Abundant ppt. 
 Evident ppt. 
 Trace ppt. 
 Abundant ppt. 
 Abundant ppt. 
 Abundant ppt. 
 Evident ppt. 
 Trace ppt 
 
 Blue color. 
 Trace " 
 
 Abundant ppt. 
 Evident ppt. 
 Evident ppt. 
 Trace ppt. 
 Abundant ppt. 
 Evident ppt. 
 Evident ppt. 
 Evident ppt. 
 Trace ppt. 
 Abundant ppt. 
 Abundant ppt. 
 Evident ppt. 
 Trace ppt. 
 Trade ppt 
 
 
 
 Blue color. 
 Blue color. 
 Trace " 
 
 Blue color. 
 Trace " 
 
 
 
 
 Blue color. 
 Blue color. 
 Trace " 
 
 Blue color. 
 Trace " 
 
 
 
 
 
 
VOLUMETRIC ESTIMATION OF COPPER. 
 
 349 
 
 and only a slight precipitate with ferrocyanide the precipita- 
 tion was considered practically complete and the conditions 
 were regarded suitable for the trial of the method quantita- 
 tively. In the following table is recorded the work upon the 
 precipitation of copper sulphate by 0.5 grm., 1.0 gnn., and 2.0 
 grm. of oxalic acid in 50 cm 3 of solution. 
 
 It will be seen readily by comparison of the right and left 
 hand sides of the table above that somewhat smaller amounts 
 of copper may be precipitated completely by the addition of 
 crystallized oxalic acid than by the same amount of oxalic 
 acid already in solution. Thus, when dissolved oxalic acid 
 is added to the solution of 50 cm 8 , amounts of copper sulphate 
 less than 0.040-0.050 grm. are not precipitated completely, 
 while under conditions otherwise the same excepting that the 
 oxalic acid is added in crystalline form, the precipitation of 
 amounts as small as 0.030 grm. is practically complete. The 
 amount of oxalic acid in solution necessary for the complete 
 
 TABLE H. 
 
 
 
 
 A. 
 
 
 CuO 
 
 taken as 
 CuSO 4 . 
 
 Oxalic 
 acid 
 added in 
 solution. 
 
 Volume 
 at pre- 
 cipita- 
 tion. 
 
 Filtrate treated with 
 NH 4 OH. 
 
 Filtrate treated 
 with K4FeC 6 N 8 . 
 
 gnn. 
 0.031 
 0.031 
 0.031 
 0.031 
 0031 
 
 gnn. 
 0.5 
 1.0 
 2.0 
 3.0 
 35 
 
 cm 8 
 
 50 
 50 
 50 
 60 
 50 
 
 Blue color. 
 Blue color. 
 Trace blue color. 
 Slight trace blue color. 
 
 Abundant ppt. 
 Abundant ppt. 
 Abundant ppt. 
 Abundant ppt. 
 Evident ppt. 
 
 0.0128 
 0.0064 
 
 5.0 
 5.0 
 
 50 
 50 
 
 No blue. 
 Blue color. 
 
 Trace ppt. 
 Abundant ppt. 
 
 
 
 
 B. 
 
 
 0.0064 
 0.0064 
 0.0064 
 00064 
 
 0.5 
 0.5 
 0.5 
 05 
 
 20 
 15 
 10 
 5 
 
 Faint blue. 
 Faint blue. 
 Faint blue. 
 
 Abundant ppt. 
 Abundant ppt. 
 Faint ppt. 
 Trace ppt. 
 
 00003 
 
 05 
 
 5* 
 
 
 
 00003 
 
 01 
 
 If 
 
 
 
 
 
 
 
 
 * Precipitate redissolved. 
 
 t Precipitate remained. 
 
350 VOLUMETRIC ESTIMATION OF COPPER. 
 
 precipitation (after 16 to 20 hours) of this minimum amount 
 of copper, 0.031 grm. of copper oxide taken as the sulphate, 
 appears, as shown in Table II, A, which precedes, to be about 
 3.5 grm. in 50.0 cm 3 . If the amount of oxalic acid is in- 
 creased to 5 grm., making the solution saturated for that 
 substance, using the same volume of liquid, the minimum 
 amount completely precipitable is reduced to 0.0128 grm. but 
 not to one-half that amount. 
 
 It appears from the experiments of Table II, B, that the 
 volume of liquid in which precipitation takes place influences 
 the complete precipitation of the copper oxalate. Thus the 
 precipitation of 0.0064 grm. of copper oxide taken as the 
 sulphate by 0.5 grm. of oxalic acid is complete in 5 cm 3 of 
 liquid. The precipitate which falls from 0.0003 grm. of the 
 oxide taken as the sulphate dissolves in 5 cm 3 of liquid, but 
 remains visible in 1 cm 3 . 
 
 As a result of the preluninary experiments, it may be said 
 that the presence of a certain minimum amount of copper, 
 varying with the conditions, is essential to complete precipi- 
 tation. Thus, at a dilution of 50 cm 3 a saturated solution of 
 oxalic acid will precipitate with practical completeness copper 
 taken as the sulphate in amounts exceeding the equivalent 
 of 0.0128 grm. of copper oxide; 2.0 grm. of oxalic acid 
 will precipitate almost completely for the same volume of 
 solution the equivalent of 0.03 grm. of copper oxide; and 
 1.0 grm. or 0.5 grm. of oxalic acid will precipitate the equiva- 
 lent of 0.064 grm. of the oxide. 
 
 In the quantitative separation of copper as the oxalate 
 the method of treatment was in general as follows. Copper 
 sulphate hi 50 cm 3 of water was thrown down by the addition 
 of dry oxalic acid to the hot solution, and, after standing 
 over night, the precipitate was filtered on asbestos and washed 
 two or three times with small amounts of cold water. The 
 precipitate, still in the crucible, was returned to the beaker 
 in which precipitation took place, 5 or 10 cm 3 of dilute sul- 
 phuric acid (1 : 1) were then added together with a convenient 
 amount of water, and, after heating the liquid to boiling, the 
 
VOLUMETRIC ESTIMATION OF COPPER. 
 
 351 
 
 oxalic acid was titrated with permanganate, the oxalate of 
 copper dissolving readily as fast as the excess of oxalic acid 
 is removed by the permanganate. The precipitate may also 
 be dissolved in 10 cm 3 of strong hydrochloric acid,* and, after 
 adding 0.5 grm. manganous chloride, titrated at 30-50. 
 Experiments (4) and (5) were conducted after this manner. 
 In Table III, A, which follows, are recorded results of the 
 quantitative tests of the method. 
 
 TABLE III. 
 
 Exp. 
 
 CuO 
 taken as 
 CuSO 4 . 
 
 Oxalic 
 acid. 
 
 Volume 
 at precipi- 
 tation. 
 
 CuO 
 found. 
 
 Error. 
 
 A. 
 
 
 grm. 
 
 grm. 
 
 cm 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.0372 
 
 0.15 
 
 100 
 
 0.0286 
 
 0.0086- 
 
 (2) 
 
 0.1860 
 
 0.50 
 
 125 
 
 0.1831 
 
 0.0029- 
 
 (3) 
 
 0.0398 
 
 0.50 
 
 50 
 
 0.0376 
 
 0.0022- 
 
 (4) 
 
 0.1860 
 
 1.0 
 
 150 
 
 0.1834 
 
 0.0026- 
 
 (5) 
 
 0.1860 
 
 0.5 
 
 50 
 
 0.1864 
 
 0.0004+ 
 
 (6) 
 
 0.1860 
 
 0.5 
 
 60 
 
 0.1866 
 
 0.0006+ 
 
 (7) 
 
 0.1860 
 
 0.5 
 
 50 
 
 0.1866 
 
 0.0006+ 
 
 (8) 
 
 0.1860 
 
 1.0 
 
 50 
 
 0.1866 
 
 0.0006+ 
 
 (9) 
 
 0.0398 
 
 1.0 
 
 50 
 
 0.0391 
 
 0.0007- 
 
 In experiments (l)-(4), deficiencies are found in the 
 amounts of oxalate precipitated at different degrees of dilution 
 and by different amounts of the precipitant which are in 
 agreement with the results obtained in the preliminary work ; 
 the results of experiments (5)-(9), in which 0.5 grm. and 1.0 
 grm. of oxalic acid act in a total volume of 50 cm 3 , show 
 the precipitation to be essentially complete under these 
 conditions. 
 
 To study the insolubility of the copper oxalate in nitric 
 acid the experiments in Section B of the table were made. 
 
 In experiments (10)-(13) amounts of oxalic acid varying 
 from 0.5 grm. to 3.0 grm. appear to precipitate the copper 
 completely in the presence of 5 cm 3 of strong nitric acid. In 
 experiment (14) the amount of oxalic acid used was not suffi- 
 
 Gooch and Peters, Am. Jour. Sci., vii, 461. This volume, p. 222. 
 
352 
 
 VOLUMETRIC ESTIMATION OF COPPER. 
 
 cient to throw down all the copper in the presence of 10 cm 3 
 of nitric acid, but the copper does come down completely in 
 the presence of the large amount of the nitric acid upon the 
 addition of more oxalic acid, as seen in experiments (15) and 
 (16). In experiments (17) and (18) with a larger volume of 
 water and a larger absolute amount, though approximately 
 the same percentage, of nitric acid present as in experiments 
 (10)-(13), there is a slight loss of copper; but hi experiments 
 (21) and (22) when the amount of nitric acid is reduced to 
 5 cm 3 in the larger total volume the results are normal. 
 Experiments (19) and (20) show the increased loss when still 
 larger amounts of nitric are present. These facts would make 
 it seem best to limit the absolute amount of nitric in solution 
 to about 5 cm 3 . 
 
 TABLE III (continued). 
 
 
 CuO 
 
 
 HNOg 
 
 Volume 
 
 
 
 Ezp. 
 
 taken as 
 CuS0 4 . 
 
 acid. 
 
 <!y? 
 
 at precipi- 
 tation. 
 
 CuO found. 
 
 Error. 
 
 B. 
 
 
 grm. 
 
 gnn. 
 
 cm 3 
 
 cm 
 
 gnn. 
 
 grm. 
 
 (10) 
 
 0.1860 
 
 0.5 
 
 5.0 
 
 55 
 
 0.1859 
 
 0.0001- 
 
 (11) 
 
 0.1860 
 
 0.5 
 
 5.0 
 
 65 
 
 0.1860 
 
 0.0000 
 
 (12) 
 
 0.1990 
 
 2.0 
 
 5.0 
 
 66 
 
 0.1989 
 
 0.0001- 
 
 (13) 
 
 0.1990 
 
 3.0 
 
 6.0 
 
 55 
 
 0.1990 
 
 0.0000 
 
 (14) 
 
 0.1990 
 
 2.0 
 
 10.0 
 
 60 
 
 0.1971 
 
 0.0019- 
 
 (15) 
 
 0.1990 
 
 3.0 
 
 10.0 
 
 60 
 
 0.1987 
 
 0.0003- 
 
 16 
 
 0.1990 
 
 3.0 
 
 10.0 
 
 60 
 
 0.1985 
 
 0.0005- 
 
 (17) 
 
 0.1990 
 
 5.0 
 
 12.0 
 
 130 
 
 0.1977 
 
 0.0013- 
 
 (18) 
 
 0.1990 
 
 5.0 
 
 12.0 
 
 130 
 
 0.1975 
 
 0.0015- 
 
 (19) 
 
 0.1990 
 
 5.0 
 
 25.0 
 
 130 
 
 0.1837 
 
 0.0153- 
 
 (20) 
 
 0.1990 
 
 5.0 
 
 25.0 
 
 130 
 
 0.1831 
 
 0.0159 
 
 (21) 
 
 0.1990 
 
 6.0 
 
 5.0 
 
 130 
 
 0.1983 
 
 0.0007- 
 
 (22) 
 
 0.1990 
 
 6.0 
 
 5.0 
 
 130 
 
 0.1988 
 
 0.0002- 
 
 C. 
 
 (23) 
 
 0.1990 
 
 2.5 
 
 5.0* 
 
 65 
 
 0.1961 
 
 0.0029- 
 
 (24) 
 
 0.1990 
 
 2.0 
 
 5.0 
 
 65 
 
 0.1971 
 
 0.0019- 
 
 One observation may well be noted here : namely, that while 
 one-half gram oxalic acid is all that is needed for the complete 
 
 * About 9 grm. of ammonium nitrate present in addition to the 5 cm 3 of 
 nitric acid. 
 
VOLUMETRIC ESTIMATION OF COPPER. 
 
 353 
 
 precipitation of the copper in the presence of 5 cm 3 strong 
 nitric acid, still the oxalic acid may be added up to the point 
 of saturation of the solution. More than this causes difficulty 
 owing to the fact that a large amount of water is necessary to 
 wash the precipitated oxalate. About 2.0 grin, of oxalic acid 
 to 50 cm 3 of water is a convenient proportion. 
 
 In experiments (23) and (24), 5 cm 8 of nitric acid were 
 neutralized with ammonium hydroxide before adding the 5 cm 3 
 strong nitric acid in excess. The results show the solubility 
 of copper oxalate in ammonium nitrate and exclude the 
 possibility of such a procedure in this work. 
 
 TABLE HI (continued). 
 
 Exp. 
 
 CuO 
 
 taken as 
 CuS0 4 . 
 
 Oxalic 
 acid. 
 
 HN0 3 
 (sp. gr. 
 1.40). 
 
 Volume 
 at precip- 
 itation. 
 
 CuO 
 
 found. 
 
 Error. 
 
 Details 
 of filtration. 
 
 D. 
 
 
 grin* 
 
 gnu* 
 
 cm 3 
 
 cms 
 
 gnu. 
 
 grin. 
 
 
 (23) 
 
 0.1990 
 
 2.0 
 
 
 50 
 
 0.1984 
 
 0.0006- 
 
 Filtered hot 
 immediately. 
 
 (24) 
 
 0.2030 
 
 2.0 
 
 
 50 
 
 0.2025 
 
 0.0005- 
 
 Filtered hot 
 immediately. 
 
 
 
 
 
 
 
 
 t Filtered after 
 
 (25) 
 
 0.1990 
 
 1.0 
 
 . . 
 
 50 
 
 0.1990 
 
 0.0000 
 
 ? cooling ; stood 
 
 
 
 
 
 
 
 
 ( 15 minutes. 
 
 
 
 
 
 
 
 
 Filtered after 
 
 (26) 
 
 0.1990 
 
 1.0 
 
 . . 
 
 50 
 
 0.1987 
 
 0.0003- 
 
 cooling; stood 
 
 
 
 
 
 
 
 
 15 minutes. 
 
 E. 
 
 
 
 
 
 
 
 
 ( Filtered after 
 
 (27) 
 
 0.1990 
 
 2.0 
 
 5.0 
 
 55 
 
 0.1943 
 
 0.0047- 
 
 < cooling ; stood 
 
 
 
 
 
 
 
 
 ( 15 minutes. 
 
 (28) 
 (29) 
 
 0.1990 
 0.1990 
 
 2.0 
 2.0 
 
 5.0 
 6.0 
 
 55 
 55 
 
 0.1969 
 0.1973 
 
 0.0021- 
 0.0017- 
 
 Stood 2 hours. 
 Stood 6 hours. 
 
 (30) 
 
 0.1990 
 
 2.0 
 
 5.0 
 
 55 
 
 0.1989 
 
 0.0001- 
 
 Stood 16 hours. 
 
 Some experiments were made to show the time necessary 
 for the complete precipitation, both in the presence and absence 
 of nitric acid. Above is the record of such work. 
 
 The results in section D would seem to show that a solution 
 containing copper may be precipitated hot as the oxalate and 
 filtered either hot or after cooling with a very slight loss. 
 VOL. ii. 23 
 
354 
 
 VOLUMETRIC ESTIMATION OF COPPER. 
 
 Tests of the filtrates made with potassium ferrocyanide con- 
 firmed these results. When nitric acid is present, however, 
 the mixture must stand after the addition of the precipitant. 
 In section E the gradual decrease of the minus error is noticed, 
 as the time of standing is extended, the precipitation being 
 practically complete upon standing over night. 
 
 Separation from Cadmium. 
 
 Bornemann * has used nitric acid for a rough separation of 
 copper from cadmium. This method was tried for a quantita- 
 tive separation in the presence of 6-10 per cent strong nitric 
 acid. The results are found in section F of the table to follow. 
 Experiments (33)-(35) stood six hours before filtering. Ex- 
 periments (36) and (37) stood over night. Copper is separated 
 from more than twice its weight of cadmium, and the results 
 are accurate. 
 
 TABLE III (continued). 
 
 Exp. 
 
 CuO 
 
 taken as 
 CuS0 4 . 
 
 Element 
 from which 
 copper was 
 separated. 
 
 Oxalic 
 acid. 
 
 HN0 3 
 
 (sp. gr. 
 1.40). 
 
 Volume 
 at precip- 
 itation. 
 
 CuO 
 found. 
 
 Error. 
 
 
 CdO taken 
 as CdS0 4 . 
 
 F. 
 
 (33) 
 (34) 
 (35) 
 (36) 
 (37) 
 
 gTDl. 
 
 0.1990 
 0.1990 
 0.1990 
 0.1990 
 0.1990 
 
 grm. 
 0.10 
 
 0.20 
 0.30 
 0.40 
 0.50 
 
 griii. 
 
 2.0 
 2.0 
 2.0 
 2.0 
 2.0 
 
 cm 
 5.0 
 5.0 
 5.0 
 5.0 
 6.0 
 
 cm 8 
 60 
 65 
 70 
 76 
 80 
 
 grm. 
 0.1983 
 0.1987 
 0.1987 
 0.1994 
 0.1996 
 
 gnu. 
 
 0.0007- 
 0.0003- 
 0.0003- 
 0.0004+ 
 0.00064- 
 
 
 As 2 O 3 taken 
 as NagAsOs. 
 
 G. 
 
 (38) 
 (39) 
 (40) 
 (41) 
 
 (42) 
 (43) 
 
 0.1990 
 0.1990 
 0.1990 
 0.1990 
 0.1990 
 0.1990 
 
 0.10 
 0.20 
 0.50 
 0.10 
 0.20 
 0.60 
 
 2.0 
 2.0 
 2.0 
 2.0 
 2.0 
 2.0 
 
 5.0 
 5.0 
 5.0 
 
 65 
 60 
 75 
 
 60 
 
 75 
 
 85 
 
 0.1991 
 0.1987 
 0.1986 
 0.1994 
 0.1992 
 0.1995 
 
 0.00014- 
 0.0003- 
 0.0004- 
 0.00044- 
 0.00024- 
 0.00054- 
 
 
 As,O B taken 
 as H 2 KAsO 4 . 
 
 H. 
 
 (44) 
 (45) 
 (46) 
 (47) 
 (48) 
 (49) 
 
 0.1990 
 0.1990 
 0.1990 
 0.1990 
 0.1990 
 0.2030 
 
 0.10 
 0.20 
 0.10 
 0.20 
 0.30 
 0.30 
 
 2.0 
 2.0 
 2.0 
 2.0 
 2.0 
 3.0 
 
 5.0 
 5.0 
 5.0 
 
 5.0 
 
 60 
 70 
 65 
 75 
 
 85 
 85 
 
 0.1985 
 0.1990 
 0.1990 
 0.1992 
 0.1985 
 0.202f> 
 
 0.0005- 
 0.0000 
 0.0000 
 0.00024- 
 0.0005 
 0.0004- 
 
 * Loc. cit. 
 
VOLUMETRIC ESTIMATION OF COPPER. 
 
 355 
 
 TABLE m (continued). 
 
 Exp. 
 
 Cu 
 taken as 
 CuSO 4 . 
 
 Element 
 from which 
 copper was 
 separated. 
 
 Oxalic 
 acid. 
 
 HN0 3 
 
 'So 8 )'' 
 
 Volume 
 at precip- 
 itation. 
 
 Cu 
 found. 
 
 Error. 
 
 
 Sn taken as 
 SnClj + HCl. 
 
 I. 
 
 grm. 
 (50) 
 (51) 
 (51 a) 
 (52) 
 (53) 
 (54) 
 (55) 
 
 grin. 
 0.1590 
 0.1590 
 0.1590 
 0.1590 
 0.1590 
 0.1590 
 0.1590 
 
 grin. 
 
 0.0468 
 0.0936 
 0.0936 
 0.0936 
 0.1873 
 0.2809 
 0.2809 
 
 grm. 
 2.0 
 2.0 
 2.0 
 2.0 
 2.0 
 2.0 
 3.0 
 
 cm 
 5.0 
 6.0 
 5.0 
 5.0 
 5.0 
 5.0 
 5.0 
 
 cm 8 
 65 
 60 
 60 
 60 
 65 
 70 
 75 
 
 grm. 
 0.1681 
 0.1603 
 0.1591 
 0.1694 
 0.1603 
 0.1914 
 0.1988 
 
 grm. 
 0.0009- 
 0.0013+ 
 0.0001+ 
 0.0004+ 
 0.0013+ 
 0.0324+ 
 0.0398+ 
 
 
 Sn taken 
 
 as SnCl 4 . 
 
 K. 
 
 (56) 
 (57) 
 (58) 
 (59) 
 
 0.1590 
 0.1590 
 0.1590 
 0.1590 
 
 0.10 
 
 0.10 
 0.20 
 0.50 
 
 2.0 
 2.0 
 2.0 
 2.0 
 
 5.0 
 
 5.6 
 6.0 
 
 65 
 55 
 55 
 60 
 
 0.1581 
 0.1565 
 0.1577 
 0.1562 
 
 0.0009- 
 0.0026- 
 0.0013- 
 0.0028- 
 
 
 CuO taken 
 as CuSO 4 . 
 
 FeO 3 taken 
 as Fe(N0 3 ) 3 . 
 
 L. 
 
 CuO found. 
 
 
 (60) 
 (61) 
 (62) 
 (63) 
 (64) 
 (65) 
 (66) 
 
 0.1990 
 0.1990 
 0.1990 
 0.1990 
 0.1990 
 0.1990 
 0.1990 
 
 0.136 
 0.272 
 0.364 
 0.544 
 0.272 
 0.544 
 0.218 
 
 2.0 
 2.0 
 2.0 
 2.0 
 2.0 
 2.0 
 2.0 
 
 6.0 
 6.0 
 5.0 
 5.0 
 
 2.6 
 2.0 
 
 60 
 60 
 60 
 65 
 60 
 60 
 65 
 
 0.1987 
 0.1983 
 0.1988 
 0.1971 
 0.1995 
 0.1998 
 0.1999 
 
 0.0003- 
 0.0007- 
 0.0002- 
 0.0019- 
 0.0005+ 
 0.0008+ 
 0.0009+ 
 
 
 ZnO taken 
 as ZnSO 4 . 
 
 M. 
 
 (67) 
 (68) 
 (69) 
 (70) 
 
 0.1990 
 0.1990 
 0.1990 
 0.1990 
 
 0.028 
 0.057 
 0.057 
 0.085 
 
 2.0 
 2.0 
 2.0 
 2.0 
 
 5.0 
 6.0 
 6.0 
 6.0 
 
 60 
 65 
 65 
 70 
 
 0.2007 
 0.2008 
 0.2008 
 0.2036 
 
 0.0017+ 
 0.0018+ 
 0.0018+ 
 0.0045+ 
 
 Separation from Arsenic, in Both Conditions of Oxidation. 
 
 For the separation of arsenic, arsenious oxide dissolved in 
 sodium carbonate, and di-hydrogen sodium arseniate were the 
 forms of arsenic used. The results are accurate and are given 
 in sections G and H of the table. In experiments (38)-(40) 
 and (44)-(45) no nitric acid was added. While the presence 
 of the nitric acid is not necessary for the separation of the 
 copper from the arsenic, still the filtration in the absence of 
 
356 VOLUMETRIC ESTIMATION OF COPPER. 
 
 the nitric acid is so slow as to be objectionable. The presence 
 of the nitric acid causes the precipitate to come down in a 
 coarser condition, and in such condition it filters easily and is 
 capable of being washed quickly. 
 
 Separation from Tin, in Both Conditions of Oxidation. 
 
 For the separation of copper from tin a preparation of 
 stannous chloride (20 cm 3 giving 0.3746 grm. metallic tin by 
 the battery) containing sufficient hydrochloric acid to prevent 
 deposition of oxy-salts was used. The solution of stannic 
 chloride contained 1.0 grm. metallic tin to every 10 cm 3 , and 
 was used without hydrochloric acid. The results of the work 
 are found in sections I and K of the table. The experiments 
 go to show that while copper may be separated from small 
 amounts of tin as stannous chloride yet there is a limit to the 
 amount of tin which may be present. One-tenth of a gram of 
 metallic tin is the largest amount that can be present, with 
 0.15 gm. copper oxide taken as the sulphate, without significant 
 error. Practically the same statement can be made of the 
 separation of copper from tin taken as stannic chloride. 
 Experiment (57) shows a greater loss of copper when the 
 nitric acid is omitted. 
 
 Separation of Copper from Iron. 
 
 A solution of ferric nitrate was used for the work on the 
 separation of copper from iron. Low results were obtained 
 when a solution of ferrous or ferric sulphate was used as the 
 source of iron. The results of the experiments are recorded 
 in section L of the table, and show that 0.20 grm. copper oxide 
 as the sulphate may be separated from 0.2-0.3 grm. iron oxide 
 taken as the nitrate. In experiment (64) a good result was 
 obtained when no nitric acid was present, save that added in 
 combination with the iron. A comparison of experiments 
 (63) and (65) shows that it is best to avoid the use of large 
 amounts of nitric acid when the larger amounts of ferric 
 nitrate are present. 
 
VOLUMETRIC ESTIMATION OF COPPER. 357 
 
 For a practical application of the above separation of copper 
 from iron a convenient amount of finely ground chalcopyrite 
 (0.5 gnn.) was roasted 2-3 hours in a porcelain crucible until 
 all sulphur was driven off, washed into a beaker, strong nitric 
 acid, about 5 cm 3 , was added and, with the beaker covered, 
 allowed to evaporate slowly on a hot plate, nearly to dryness. 
 A little dilute nitric acid was added, the solution was filtered, 
 the residue was washed with water containing dilute nitric 
 acid, the filtrate, about 50 cm 3 in volume, was precipitated 
 with 2.0 grm. oxalic acid, and the precipitate was estimated 
 after standing 12-16 hours, as previously described. The 
 washing with water acidified with nitric acid is important, 
 because the finely ground ferric oxide remaining undissolved 
 passes through the filter when washed with water alone, but 
 gives no trouble if the water be acidic. The results of two 
 estimations are here given. 
 
 Chalcopyrite. 
 
 Copper found 
 by battery. 
 
 Copper found by 
 oxalate method. 
 
 Difference. 
 
 grm. 
 
 % 
 
 % 
 
 % 
 
 0.5000 
 
 31.00 
 
 30.92 
 
 0.08- 
 
 1.0000 
 
 31.00 
 
 31.25 
 
 0.25+ 
 
 Separation of Copper from Zinc. 
 
 The separation of copper from zinc was not altogether 
 successful owing to the tendency of the zinc oxalate to come 
 down with the copper oxalate. Some experiments are given 
 in section M of the table. 
 
 The separations of copper from bismuth and antimony were 
 unsuccessful. 
 
 The work may be briefly summarized as follows : Copper 
 exceeding in amount the equivalent of 0.0128 grm. of the 
 oxide to 50 cm 3 of solution as the sulphate may be separated 
 completely, even in the presence of a moderate amount of 
 strong nitric acid, by the addition of sufficient amount of 
 oxalic acid. 
 
 Copper may be separated from cadmium, arsenic, iron, and 
 small amounts of tin, when precipitated by oxalic acid in 
 
358 VOLUMETRIC ESTIMATION OF COPPER. 
 
 a volume of 50 cm 3 containing 5 cm 3 strong nitric acid. 
 Inasmuch as the completeness of precipitation of the copper 
 depends upon the presence of a certain minimum amount 
 of the copper salt this method is not applicable when the 
 amount of copper falls below 0.0128 grm. of the oxide to 50 
 cm 3 of solution. 
 
XLVI 
 
 THE SULPHOCYANIDES OF COPPER AND SIL- 
 VER IN GRAVIMETRIC ANALYSIS. 
 
 BY K. G. VAN NAME * 
 
 Cuprous Sulphocyanide. 
 
 As early as 1854 attention was drawn by Rivotf to the 
 possibility of estimating copper gravimetrically by weighing 
 as cuprous sulphocyanide, and to the advantages which the 
 process afforded in separating copper from other metals. 
 Rivot's procedure consisted in dissolving the substance to 
 be analyzed in hydrochloric acid, reducing copper with hypo- 
 phosphorous or sulphurous acid, and precipitating with potas- 
 sium sulphocyanide. The precipitate dried at a moderate 
 temperature was weighed as cuprous sulphocyanide and then 
 as a control converted by ignition with sulphur into cuprous 
 sulphide and weighed in that condition. 
 
 In his well known work upon quantitative analysis Fre- 
 senius in one place J denies the practicability of the direct 
 weighing of copper as cuprous sulphocyanide on account of 
 the tendency of the latter to hold water even when heated to 
 the temperature of incipient decomposition. As authority 
 for this statement he cites Claus, who found 3 per cent of 
 water in the precipitate after drying at 115, and Meitzendorff, 
 who gave the percentage of water under the same conditions 
 as 1.54. 
 
 On a later page of the same volume, || however, Fresenius, 
 after a trial of the process which gave 99.66 per cent of the 
 
 * From Am. Jour. Sci., x, 451. t Compt. rend., xxxviii, 868. 
 
 J Quant. Anal., 6. Aufl., i, 187. L. Gmelin, Handbuch, iv, 472. 
 
 || Quant. Anal, 6. Aufl., i, 335. 
 
360 SULPHOCYANIDES OF COPPER AND SILVER 
 
 theory for copper, concludes that the method is practicable 
 although apt to give low results, particularly in the presence 
 of free acid. 
 
 The process was again recommended in 1878 by Busse,* 
 who had employed it for the estimation of copper, both 
 alone and in the presence of iron, nickel, zinc, and arsenic, 
 obtaining results very near the theory and plainly comparable 
 with the figures obtained by afterwards igniting the cuprous 
 sulphocyanide with sulphur in hydrogen. 
 
 In spite of the evident advantages for certain purposes of 
 Rivot's method over other modes of determining copper, it 
 has never come into general use. The chief reason for this 
 has apparently been the difficulty and inaccuracy attendant 
 upon the weighing of the precipitate upon dried paper 
 filters, a process which can hardly be depended upon unless 
 managed with extreme care. 
 
 In the experiments to be described this difficulty was 
 avoided by performing the filtering and weighing upon asbes- 
 tos in a perforated platinum crucible. The method of con- 
 ducting a determination was as follows : A suitable quantity 
 of a standard copper sulphate solution was run from a burette, 
 diluted to a convenient volume, a few cubic centimeters of a 
 saturated solution of ammonium bisulphite added, and the 
 copper precipitated by an excess of ammonium sulphocyanide. 
 The precipitate was allowed to settle, collected upon asbestos 
 in a weighed crucible, washed with cold water, and dried at 
 110 until no further loss of weight took place. 
 
 In Table I are given the results of a number of determi- 
 nations made in this way. The copper sulphate solution 
 was made up exactly decinormal and the standard confirmed 
 electrolytically. As the ammonium sulphocyanide solution 
 was slightly above decinormal, 13 cm 3 represent a small excess 
 (about one' cubic centimeter) above the amount theoretically 
 required to precipitate 25 cm 3 of the copper sulphate solution. 
 The ammonium bisulphite, which had been recently prepared 
 by saturating aqueous ammonia with sulphur dioxide, was 
 
 * Zeitschr. anal. Chem., xvii, 63. 
 
IN GRAVIMETRIC ANALYSIS. 
 
 361 
 
 always used in sufficient quantity to give the liquid a strong 
 and permanent odor of the latter. 
 
 TABLE I. 
 
 25 cm 8 of ^ CuSO 4 solution, equivalent to 0.0795 grm. Cu, taken for each 
 
 experiment. 
 
 Exp. 
 
 H 2 S0 4 . 
 concentrated. 
 
 HNH 4 SO S 
 sat. sol. 
 
 NH 4 SCN 
 approz. 
 
 ff 
 
 Final 
 volume. 
 
 Time of 
 standing. 
 
 Cu 
 found. 
 
 Error. 
 
 
 
 
 TIT 
 
 
 
 
 
 
 cms 
 
 cm 3 
 
 cm 3 
 
 cm 3 
 
 hrs. 
 
 grui. 
 
 grm. 
 
 (1) 
 
 None. 
 
 5 
 
 13 
 
 68 
 
 | 
 
 0.0795 
 
 0.0000 
 
 (2) 
 
 None. 
 
 3 
 
 13 
 
 66 
 
 48 
 
 0.0793 
 
 0.0002- 
 
 (3) 
 
 None. 
 
 3 
 
 25 
 
 78 
 
 i 
 
 0.0796 
 
 0.00014 
 
 (4) 
 
 None. 
 
 3 
 
 25 
 
 78 
 
 12 
 
 0.0796 
 
 0.0001+ 
 
 (5) 
 
 1.5 
 
 10 
 
 13 
 
 85 
 
 12 
 
 0.0792 
 
 0.0003- 
 
 (6) 
 
 1.6 
 
 8 
 
 13 
 
 105 
 
 48 
 
 0.0785 
 
 0.0010- 
 
 (7) 
 
 1.5 
 
 3 
 
 25 
 
 85 
 
 4 
 
 0.0783 
 
 0.0012- 
 
 (8) 
 
 1.5 
 
 5 
 
 25 
 
 85 
 
 21 
 
 0.0795 
 
 0.0000 
 
 (9) 
 
 5 
 
 5 
 
 25 
 
 85 
 
 3 
 
 0.0797 
 
 0.0002+ 
 
 (10) 
 
 15 
 
 10 
 
 25 
 
 115 
 
 21 
 
 0.0793 
 
 0.0002- 
 
 
 HCl 
 
 
 
 
 
 
 
 
 concentrated. 
 
 
 
 
 
 
 
 (11) 
 
 10 
 
 6 
 
 25 
 
 100 
 
 20 
 
 0.0795 
 
 0.0000 
 
 (12) 
 
 25 
 
 10 
 
 25 
 
 100 
 
 28 
 
 0.0784 
 
 0.0011- 
 
 When there is no free acid present the time of standing 
 before nitration and the amount of the excess of ammonium 
 sulphocyanide are practically without effect, as experiments 
 (1) to (4) of the table show. 
 
 Experiments (5) to (10) were carried out in the presence of 
 various amounts of free sulphuric acid up to 12 per cent of 
 the total volume of liquid. The acid, at least within this 
 limit, does not exert a sufficient solvent effect upon the cuprous 
 sulphocyanide to interfere materially with the accuracy of the 
 process, but it retards the precipitation, making it necessary to 
 increase the time of standing before filtering in proportion to 
 the amount of acid present. In several of these determina- 
 tions the precipitation was visibly incomplete even after sev- 
 eral hours' standing. This effect of the acid, however, hardly 
 shows in the results of the table because the standing was 
 always prolonged until the copper appeared to be all down 
 before filtering. 
 
362 SULPHOCYANIDES OF COPPER AND SILVER 
 
 The low result of Experiment (7) was probably due chiefly 
 to incomplete precipitation, although (9) shows that even with 
 a much larger amount of acid precipitation may be complete 
 within three hours. In general, however, it is safer to allow 
 ample time (twelve hours or more) for the precipitation when 
 there is much free acid present. 
 
 Comparison of Experiments (5) and (6), for which only a 
 bare excess of ammonium sulphocyanide was used, with (7) 
 to (12) shows an apparent advantage in the larger excess in the 
 presence of acid. Hydrochloric acid, judging from the results 
 of (11) and (12), has no greater disturbing influence than sul- 
 phuric acid, although in (12), where the concentrated acid 
 constituted one-fourth of the entire volume, there was appar- 
 ently a slight solvent action. The filtrate from this determi- 
 nation when concentrated to about 25 cm 8 and treated with 
 potassium ferrocyanide gave a strong test for copper, as did 
 also the filtrate from (6). Several of the other nitrates were 
 tested in the same way, but none showed more than an insig- 
 nificant trace of copper. The nitrate of (7), however, was 
 not tested. 
 
 Table II contains the results of a series of experiments con- 
 ducted as before, except that larger amounts of copper were 
 employed. The copper sulphate solution was approximately 
 f and standardized by the battery. The solution of ammo- 
 nium sulphocyanide was the same previously used and a con- 
 siderable excess was employed in every determination. More 
 
 TABLE II. 
 
 Bxp. 
 
 Cu 
 taken. 
 
 H 2 S0 4 
 concen- 
 trated. 
 
 NH 4 SCN 
 approx. 
 ff 
 TS' 
 
 Final 
 volume. 
 
 Cu 2 S 2 (CN) 2 
 found, calcu- 
 lated as Cu. 
 
 Error. 
 
 Cuin 
 filtrate. 
 
 | 
 
 (3) 
 (4) 
 
 grm. 
 
 0.3175 
 0.3176 
 0.3175 
 0.3175 
 
 cms 
 
 None. 
 None. 
 None. 
 10 
 
 cm 8 
 60 
 60 
 60 
 100 
 
 cm 8 
 500 
 500 
 500 
 500 
 
 grm. 
 0.3176 
 0.3177 
 0.3176 
 0.3176 
 
 grm. 
 0.0001+ 
 0.0002+ 
 0.0001+ 
 0.0000 
 
 None. 
 None. 
 None. 
 None. 
 
 (5) 
 
 0.3175 
 
 HC1 
 cone. 
 
 100 
 
 600 
 
 0.3165 
 
 0.0010- 
 
 Distinct. 
 
 20 
 
IN, GRA VIMETRIC ANAL YSIS. 363 
 
 than twice the amount theoretically required was used in every 
 case where free acid was present, and at least twenty hours 
 were allowed for the precipitation, which was made in cold, 
 and as the table shows, rather dilute solutions. If the solu- 
 tion is too concentrated the copper is apt to be thrown down 
 in a finely divided condition, making it hard to filter. 
 
 The time required to dry the cuprous sulphocyanide at 110 
 is in general from two to three hours. Heating much longer 
 than this is not to be recommended, as a gradual increase in 
 weight begins to take place, as is shown by the following 
 example, which gives a series of weights of the same precipi- 
 tate at different stages. 
 
 Cu,S 8 (CN),. Calculated as Cu. 
 gnu. gnu. 
 
 After 2 hours at 110 . . . 0.6060 0.3167 
 
 4 " ... 0.6059 0.3167 
 
 19 " . . . 0.6067 0.3171 
 
 23 " ... 0.6069 0.3172 
 
 This tendency to increase in weight is, however, usually less 
 marked than in the above example, and in any case need not 
 interfere materially with the accuracy of the process unless the 
 drying is prolonged far beyond the necessary length of time. 
 
 The method is easily handled and, as the results of Tables I 
 and II show, is capable of considerable accuracy. From the 
 nature of the process it is evident that it is much less likely to 
 be interfered with by the presence of other metals than the 
 other gravimetric methods for copper, and may therefore be 
 directly applied with good results in many cases where the use 
 of the electrolytic or the oxide method would involve a pre- 
 vious separation. 
 
 Silver Sulphocyanide. 
 
 The sulphocyanide of silver, unlike that of copper, is solu- 
 ble in an excess of ammonium or alkali sulphocyanides and 
 this fact prevents the use of the latter to precipitate silver for 
 gravimetric estimation. The reverse process, however, the 
 precipitation of a soluble sulphocyanide by an excess of silver 
 
364 SULPHOCYANIDES OF COPPER AND SILVER 
 
 nitrate, as will be shown by the experiments to be described, 
 furnishes a convenient means of standardizing sulphocyanide 
 solutions and in general for estimating sulphocyanic acid. 
 
 When freshly precipitated the sulphocyanide of silver resem- 
 bles the chloride in appearance, but when allowed to stand a 
 few hours becomes finely granular and is very easily filtered 
 and washed. It may be safely dried to a constant weight upon 
 an asbestos filter at 110 -120, but at a somewhat higher tem- 
 perature is decomposed, leaving a residue of silver sulphide. 
 
 The determinations which are tabulated below were made as 
 follows. Portions of 25 cm 3 of an approximately decinormal 
 solution of ammonium sulphocyanide were measured from a 
 burette, diluted with 100 cm 3 of water and silver nitrate added 
 in excess. The precipitate was collected upon asbestos in a 
 platinum crucible, washed with cold water and dried to a con- 
 stant weight at 115 the drying requiring usually between two 
 and three hours. 
 
 The filtering is facilitated by allowing a few hours for the 
 precipitate to settle ; but this is by no means essential, as it is 
 easy with a little care to obtain a clear filtrate even when the 
 filtering is performed at once. 
 
 The solution of ammonium sulphocyanide was prepared 
 from a pure salt, especially tested and found free from choride. 
 This point is of importance, as chlorine is a common impurity 
 and its presence in any considerable quantity will vitiate the 
 results. 
 
 In order that the effect of varying the excess of silver 
 might be investigated, an approximately decinormal solution 
 of silver nitrate was titrated against the ammonium sulpho- 
 cyanide and the ratio between the two solutions determined. 
 This silver nitrate solution was used for the first five determi- 
 nations of Table III. For the rest the quantity of silver nitrate 
 was not measured but regulated by the eye alone, thus making 
 the conditions the same as would be the case in practical use 
 of the method. 
 
 These results are as uniform as could be expected, considering 
 the variations which would be produced by even very small 
 
GRAVIMETRIC ANALYSIS. 
 
 365 
 
 TABLE in. 
 
 Final volume of liquid 150 cm 8 . 
 25 cm 8 of NH 4 SCN sol. equivalent to 25.15 cm 8 of AgN0 8 sol. 
 
 Exp. 
 
 NH 3 SCN. 
 
 AgN0 8 . 
 
 Excess of 
 AgNO s . 
 
 AgSCN 
 found. 
 
 
 cm 3 
 
 cm 3 
 
 cm 3 
 
 grm. 
 
 (1) 
 
 25 
 
 25.3 
 
 0.15 
 
 0.4372 
 
 (2) 
 
 25 
 
 25.3 
 
 0.15 
 
 .0.4376 
 
 8 
 
 25 
 25 
 
 25.4 
 25.4 
 
 0.25 
 0.25 
 
 0.4373 
 0.4375 
 
 (5) 
 
 25 
 
 30.4 
 
 5.25 
 
 0.4382 
 
 (6) 
 
 25 
 
 Kough excess. 
 
 0.4366 
 
 (7) 
 
 25 
 
 Rough excess. 
 
 0.4381 
 
 (8) 
 
 25 
 
 Rough excess. 
 
 0.4873 
 
 (9) 
 
 25 
 
 Rough excess. 
 
 0.4372 
 
 (10 
 
 25 
 
 Rough excess. 
 
 0.4369 
 
 errors in measuring out 25 cm 3 of decinormal sulphocyanide 
 solution. It is moreover clearly shown that there is no 
 difference in the results whether a bare excess or a moderately 
 large excess of the silver nitrate is used. 
 
 The mean of the values in the last column is 0.4374, which is 
 equivalent to 0.2006 grm. of ammonium sulphocyanide for 
 every 25 cm 3 of the solution. 
 
 The standard of the sulphocyanide solution was also 
 determined volumetrically by Volhard's process. The mean of 
 four titrations carried out with great care against a standard 
 silver nitrate solution gave as the standard 0.2003 grm. of 
 ammonium sulphocyanide for 25 cm 3 of solution. This differ- 
 ence between the standards as determined by the two methods 
 (one part in 670) is much less than the variations which 
 frequently appear between successive determinations by 
 Volhard's method, under like conditions as to strength of 
 solutions and amounts used. It is about equal to the error 
 that would be produced in a single volumetric determination 
 by a mistake of one drop in measuring one of the solutions, or 
 of one-half drop in the same direction on each. 
 
 It is therefore evident that the standard of a sulphocyanide 
 solution obtained in the above way may be applied directly to 
 
366 SULPHOCYANIDES OF COPPER AND SILVER 
 
 the estimation of unknown amounts of silver by Volhard's 
 method without sensible error. 
 
 To remove a possible doubt as to whether the silver 
 sulphocyanide dried at 115 was entirely free from water, a 
 number of electrolytic determinations of the silver contained 
 in the previously weighed precipitates of Table III were made 
 in the following way. 
 
 The perforated platinum crucible containing the silver 
 sulphocyanide and asbestos was hung in a loop of heavy 
 platinum wire and served as the anode. For the cathode a 
 deep platinum dish of about 200 cm 3 capacity was used. An 
 ammoniacal solution of potassium cyanide was employed as the 
 electrolyte and gave the best results when made up by dissolving 
 2 grm. of potassium cyanide in 15 cm 3 of strong ammonia and 
 15 cm 3 of water. The crucible which served as the anode was 
 filled with this solution in full strength, and the remainder 
 was put into the platinum dish and diluted to the required 
 volume with water. In this medium the silver sulphocyanide 
 is slowly dissolved and diffuses through the asbestos felt into 
 the space between the electrodes where the silver is deposited 
 in the usual way. This diffusion, is, however, aided but little 
 if at all by the current, and there is a tendency for traces of 
 the silver to remain behind in the crucible. The current density 
 employed was about 0.0012 ampere per square centimeter 
 of cathode surface and the time about twelve hours. After 
 weighing the silver deposited, it was dissolved in nitric acid, 
 precipitated by hydrochloric acid and weighed again as the 
 chloride, giving a check upon the results. 
 
 Seven of the ten determinations of Table III were thus 
 treated, but owing to the imperfections of the process the 
 results were all slightly low, the worst showing a deficiency 
 of 0.0025 grm. of silver, an error of less than 0.9 per cent. The 
 results of the two best of these determinations given below are, 
 however, sufficient to prove the point hi question, namely 
 that the silver sulphocyanide dried at 115 has the theoretical 
 constitution and contains no water. The numbers are those 
 under which the determinations appear in Table III. 
 
/.AT GRAVIMETRIC ANALYSIS. 
 
 367 
 
 Bxp. 
 
 AgSCN 
 taken. 
 
 Calculated 
 as Ag. 
 
 Ag found 
 by battery. 
 
 Error. 
 
 Weighed 
 AgCl. 
 
 Calculated 
 asAg. 
 
 Error. 
 
 (4) 
 (10) 
 
 grm. 
 
 0.4375 
 0.4369 
 
 gnu. 
 0.2844 
 0.2840 
 
 grm. 
 0.2839 
 0.2838 
 
 grm. 
 0.0005- 
 0.0002- 
 
 grm. 
 0.3765 
 0.3761 
 
 grm. 
 0.2834 
 0.2831 
 
 grm. 
 0.0010- 
 0.0009- 
 
 It is clear therefore that the estimation of sulphocyanides by 
 precipitation with silver nitrate and direct weighing of the 
 precipitate is wholly permissible. The method is extremely 
 simple and, as has been shown, the results are quite accurate. 
 
XLVII 
 
 ON THE ESTIMATION OF CAESIUM AND EUBIDIUM 
 AS THE ACID SULPHATES, AND OF POTASSIUM 
 AND SODIUM AS THE PYROSULPHATES. 
 
 BY PHILIP E. BROWNING. 
 
 BUNSEN * is authority for the statement that the acid sulphate 
 of rubidium does not lose sulphuric acid at a heat approaching 
 redness. It is stated in the literature f that the acid sulphates 
 of caesium and rubidium when subjected to a low red heat 
 pass into the form of the pyrosulphates. 
 
 R. Weber J found that by treating the dry sulphates of 
 potassium, caesium, rubidium, and thallium with sulphuric 
 anhydride in a closed tube and heating on a water bath two 
 layers separated. In the lower layer he obtained crystalline 
 bodies which proved to have the constitution R 2 O . 8SO 3 . On 
 strong heating he obtained from these substances, bodies of 
 the form R 2 . 2SO 3 and finally R 2 O . SO 3 . He also notes that 
 in the case of the caesium salt the removal of the excess of 
 the sulphuric anhydride was attended with greater difficulty. 
 
 Baum states that the pyrosulphates of the alkalies may be 
 obtained by heating the acid sulphates under atmospneric 
 pressure at low redness, or under diminished pressure at a 
 temperature between 260 C. and 320 C. 
 
 In a recent paper || from this laboratory I have shown that 
 thallium may be estimated as the acid sulphate by evaporating 
 a thallous salt in solution with an excess of sulphuric acid and 
 bringing the residue to a constant weight at a temperature of 
 about 250 C. 
 
 * Ann. Chem. (Liebig), cxix, 110. 
 
 t Graham-Otto, Lehrbuch d. Chem., iii, 269, 278. 
 
 | Ber. Dtsch. Chem. Ges., xvii, 2497. Ber. Dtsch. Chem. Ges. xx, 752. 
 
 || Am. Jour. Sci., ix (1900), 137. This volume, p. 317. 
 
ESTIMATION OF CAESIUM AND RUBIDIUM. 369 
 
 The similarity which thallium bears in some of its combi- 
 nations to the alkaline metals suggested the study of the 
 sulphates of these elements under the same general conditions 
 of procedure. 
 
 My first experiments were made with a pure caesium salt 
 as follows : A weighed amount of the nitrate was placed in 
 a previously weighed platinum crucible and treated with an 
 excess of sulphuric acid. The crucible was then placed upon 
 a steam bath until the water and nitric acid were largely 
 expelled and then removed to a radiator, consisting of a 
 porcelain crucible fitted with a pipe-stem triangle so arranged 
 that the bottom of the platinum crucible would be about 
 midway between the top and bottom of the porcelain crucible. 
 This improvised radiator was set in an iron ring and a 
 thermometer so placed that the mercury bulb would be on 
 a level with the bottom and close to the side of the platinum 
 crucible. An ordinary Bunsen burner served as the source 
 of heat and the temperature was kept so far as possible 
 between 250 C. and 270 C. After the fuming attending 
 the removal of the large excess of sulphuric acid ceased, the 
 crucible and contents were removed to a desiccator, and, after 
 being allowed to cool, weighed. This process of heating 
 was continued for half-hour periods until the weights were 
 constant. The results shown in Table I were obtained by 
 this method of treatment. In experiments (1), (4), and (9) 
 
 TABLE I. 
 
 Ezp. 
 
 CsN0 8 
 taken. 
 
 CsHS0 4 
 calcu- 
 lated. 
 
 First 
 constant 
 weight. 
 
 Second 
 constant 
 weight. 
 
 Error 
 CsHS0 4 . 
 
 Cs 2 SO 4 
 calcu- 
 lated. 
 
 Cs 2 S0 4 
 found. 
 
 Error 
 Cs 2 S0 4 . 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grin. 
 
 grm* 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.1706 
 
 0.2013 
 
 0.2054 
 
 0.2020 
 
 0.0007+ 
 
 
 
 
 (2) 
 
 0.1706 
 
 0.2013 
 
 0.2010 
 
 
 0.0003- 
 
 t 
 
 
 
 (3) 
 
 0.1032 
 
 0.1217 
 
 0.1201 
 
 
 0.0016- 
 
 
 
 
 (4) 
 
 0.1032 
 
 0.1217 
 
 0.1252 
 
 0.1222 
 
 0.0005+ 
 
 0.0961 
 
 0.0948 
 
 0.0013- 
 
 (5) 
 
 0.1218 
 
 0.1437 
 
 0.1458 
 
 ( 
 
 0.0021+ 
 
 0.1130 
 
 0.1118 
 
 0.0012- 
 
 (6) 
 
 0.1214 
 
 0.1435 
 
 0.1430 
 
 
 0.0005- 
 
 ( 
 
 
 
 (7) 
 
 0.1214 
 
 0.1435 
 
 0.1422 
 
 , 
 
 0.0013- 
 
 ( 
 
 
 
 (8) 
 
 0.1150 
 
 0.1356 
 
 0.1330 
 
 . 
 
 0.0026- 
 
 
 
 
 (9) 
 
 0.1050 
 
 0.1245 
 
 0.1272 
 
 0.1248 
 
 0.0003+ 
 
 
 
 
 (10) 
 
 0.1056 
 
 0.1245 
 
 0.1252 
 
 
 0.0007+ 
 
 
 
 
 VOL. ii. 24 
 
370 ESTIMATION OF CESIUM AND RUBIDIUM. 
 
 it will be noticed that the weights were constant somewhat 
 above the condition of the acid sulphate, a fact which would 
 go to show a tendency on the part of the caesium salt to hold 
 an excess of sulphuric acid over the amount necessary to 
 form the ordinary acid sulphate. The results show that by 
 regulating the heat at a temperature between 250 C. and 
 270 C. caesium may be brought, with a fair degree of cer- 
 tainty to the condition of the acid sulphate. As a check upon 
 the results, the acid sulphate was, in a few cases, treated with 
 a little ammonium hydroxide, the excess of this was removed 
 upon a steam bath and the neutral sulphate obtained by 
 ignition at a red heat to a constant weight. These determi- 
 nations agreed fairly well with the theory. The same pro- 
 cedure was followed with rubidium, a pure rubidium chloride 
 having been chosen as the starting-point. The results are 
 
 TABLE II. 
 
 Exp. 
 
 RbCl 
 taken. 
 
 RbHS0 4 
 calculated. 
 
 RbHS0 4 
 found. 
 
 Error. 
 
 Rb 2 S0 4 
 calculated. 
 
 Rb 2 S0 4 
 
 found. 
 
 Error. 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.1252 
 
 0.1889 
 
 0.1878 
 
 0.0011- 
 
 . . 
 
 
 
 (2) 
 
 0.1212 
 
 0.1829 
 
 0.1840 
 
 0.0011+ 
 
 0.1460 
 
 0.1460 
 
 0.0000 
 
 (3) 
 
 0.1230 
 
 0.1856 
 
 0.1850 
 
 0.0006- 
 
 
 
 
 (4) 
 
 0.1230 
 
 0.1856 
 
 0.1858 
 
 0.0002+ 
 
 0.1357 
 
 0.1350 
 
 0.0007- 
 
 (5) 
 
 0.1610 
 
 0.2430 
 
 0.2416 
 
 0.0014- 
 
 0.1777 
 
 0.1772 
 
 0.0005- 
 
 (6) 
 
 0.1360 
 
 0.2052 
 
 0.2032 
 
 0.0020- 
 
 0.1501 
 
 0.1490 
 
 0.0011- 
 
 given in Table II. No tendency was observed on the part 
 of this element to hold sulphuric acid in excess of the amount 
 necessary for the formation of the acid sulphate. When the 
 same method was applied to sodium and potassium salts, 
 pure chlorides being used as the starting-point, a somewhat 
 different result was obtained, in that the weight of the 
 final product appeared to indicate the formation of the 
 pyrosulphate. The results given in Tables III and IV, in 
 which the sodium and potassium salts are calculated as 
 pyrosulphates, are sufficiently satisfactory for purposes of 
 quantitative estimation. As in the case of the caesium and 
 rubidium salts, a number of determinations as the neutral 
 
ESTIMATION OF CAESIUM AND RUBIDIUM. 371 
 TABLE m. 
 
 Exp. 
 
 KC1 
 taken. 
 
 K 2 S,0 7 
 calculated. 
 
 K,8 2 7 
 found. 
 
 Error. 
 
 K,S0 4 
 calculated. 
 
 Sffi 
 
 Error. 
 
 
 gnu. 
 
 grm. 
 
 grm. 
 
 grin. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.2172 
 
 0,3704 
 
 0.3698 
 
 0.0006- 
 
 
 
 
 2 
 
 0.1706 
 
 0.2909 
 
 0.2886 
 
 0.0023 
 
 0.1993 
 
 0.1972 
 
 0.0021- 
 
 (3) 
 
 0.1192 
 
 0.2032 
 
 0.2022 
 
 0.0010- 
 
 0.1393 
 
 0.1381 
 
 0.0012- 
 
 (4) 
 
 0.1074 
 
 0.1830 
 
 0.1823 
 
 0.0007- 
 
 
 
 
 (5) 
 
 0.1096 
 
 0.1868 
 
 0.1860 
 
 0.0008- 
 
 
 
 
 
 TABLE IV. 
 
 Exp. 
 
 NaCl 
 taken. 
 
 Na 2 8 2 7 
 calculated. 
 
 Na 2 S 2 7 
 found. 
 
 Error. 
 
 Na.jS0 4 
 calculated. 
 
 NajSO* 
 found. 
 
 Error. 
 
 (1) 
 
 (2) 
 (3) 
 (4) 
 
 grin* 
 
 0.1042 
 0.1028 
 0.1093 
 0.1402 
 
 grm. 
 
 0.1978 
 0.1952 
 0.2075 
 0.2662 
 
 grm. 
 0.1972 
 0.1952 
 0.2065 
 0.2651 
 
 grm. 
 
 0.0006- 
 0.0000 
 0.0010- 
 0.0011- 
 
 gnu. 
 
 0.1266 
 
 0.1328 
 0.1703 
 
 grm. 
 
 0.1254 
 
 0.1320 
 0.1696 
 
 grm. 
 
 0.0012- 
 
 0.0008- 
 0.0007- 
 
 sulphate were made by ignition of the sodium and potassium 
 pyrosulphates, with results which are recorded. In Table V, 
 two determinations are recorded, in one of which the caesium 
 and rubidium salts were treated together and in the other 
 the sodium and potassium salts. 
 
 TABLE V. 
 
 Exp. 
 
 RbCl + CsNOj 
 taken. 
 
 RbHS0 1 + 
 CsHS0 4 
 calculated. 
 
 RbHS0 4 + 
 CsHS0 4 
 found. 
 
 Error. 
 
 Rb 2 S0 4 + 
 Cs 2 SO 4 
 calculated. 
 
 Rb 2 S0 4 + 
 CB*S0 4 
 found. 
 
 Error. 
 
 (1) 
 
 grm. 
 
 (RbCl 0.1428 ) 
 {CsN0 3 0.1264 f 
 
 grm. 
 
 0.3646 
 
 grm. 
 
 0.3666 
 
 grm. 
 0.0020+ 
 
 grm. 
 0.2749 
 
 grm. 
 
 0.2752 
 
 griii. 
 0.0003+ 
 
 
 NaCl + KC1 
 taken. 
 
 ^80,+ 
 MLftfff 
 calculated. 
 
 Na 2 S,0 7 + 
 K 2 S 2 O 7 
 found. 
 
 Error. 
 
 ^A 
 
 calculated. 
 
 Na28 g 4 ik 
 
 K 2 SO 4 
 found. 
 
 Error. 
 
 (2) 
 
 (NaCl 0.1233 ) 
 |KCI 0.1340 ] 
 
 0.4627 
 
 0.4630 
 
 0.0003+ 
 
 0.3062 
 
 0.3040 
 
 0.0022- 
 
 An application of this general method to a lithium salt 
 gave no evidence of the existence of a stable acid sulphate 
 or pyrosulphate. 
 
372 ESTIMATION OF CAESIUM AND RUBIDIUM. 
 
 The results may be summed up as follows : Csesium and 
 rubidium salts of volatile acids when treated with sulphuric 
 acid in excess and brought to a constant weight at a tem- 
 perature between 250 C. and 270 C. form acid salts of 
 the type RHSO 4 and the neutral salts of the type R 2 SO 4 on 
 ignition. 
 
 Some tendency of the caesium salt to hold more sulphuric 
 acid than corresponds to the formation of the acid sulphate 
 RHSO 4 was apparent at temperatures between 258 C. and 
 270 C., but upon raising the temperature above 300 C. the 
 loss was excessive and showed a tendency on the part of the 
 acid sulphate to pass, at this temperature, toward the con- 
 dition of the pyrosulphate. 
 
 Sodium and potassium salts, when heated under the con- 
 ditions described, give pyrosulphates of the type R 2 S 2 O 7 which 
 on ignition go into the neutral sulphate of the form R 2 SO 4 . 
 Lithium gives neither salts of the type RHSO 4 nor R 2 S 2 O 7 
 under the conditions of these experiments. 
 
XLVIII 
 
 THE ESTIMATION OF CALCIUM, STRONTIUM, 
 AND BARIUM AS THE OXALATES. 
 
 BY CHARLES A. PETERS. 
 
 A FORMER article from this laboratory* describes the condi- 
 tions under which oxalic acid may be titrated by potassium 
 permanganate in the presence of hydrochloric acid, and states 
 that the extra consumption of permanganate which ordinarily 
 takes place when oxalic acid is titrated by permanganate in 
 the presence of hydrochloric acid, may be prevented by the 
 addition of a manganous salt. This fact led to the idea of 
 effecting the solution of the alkaline earth oxalates in hydro- 
 chloric acid and titrating the free oxalic acid with perman- 
 ganate in the presence of a manganous salt, and so to the 
 study of the conditions under which precipitates of strontium 
 and barium oxalates could be obtained sufficiently insoluble 
 for quantitative purposes, the conditions under which calcium 
 oxalate is insoluble being already well known. 
 
 The permanganate solution for this work was standardized 
 against freshly recrystallized ammonium oxalate, and on 
 oxalic acid, the standards agreeing. 
 
 Calcium Oxalate. 
 
 It is well known that calcium may be estimated by treating 
 the precipitated oxalate with sulphuric acid and titrating by 
 permanganate the oxalic acid set free.f In the work 
 described in the present article, the precipitate of calcium 
 oxalate has been dissolved in hydrochloric acid and the oxalic 
 
 * Gooch and Peters, Am. Jour. Sci., vii, 461. This volume, p. 222. 
 t Mohr, Titrirmethode, 6. Aufl., S. 227. 
 
374 
 
 ESTIMATION OF CALCIUM, STRONTIUM, 
 
 acid titrated by permanganate in the presence of a manganous 
 salt. The process was as follows : The boiling hot solution 
 of calcium chloride was precipitated with ammonium oxalate, 
 allowed to stand 12 hours, and the supernatant liquid de- 
 canted on asbestos. The precipitate was washed two or three 
 tunes by decantation with 50-100 cm 3 of cold water and 
 brought on the felt. The crucible containing the precipitate 
 was returned to the beaker, 100-200 cm 3 of water were 
 added, together with 5-10 cm 3 of strong hydrochloric acid and 
 0.5-1.0 grin, of manganous chloride, and the oxalic acid 
 titrated at a temperature of 35 -45. The results, given in 
 Table I, are obviously excellent. 
 
 TABLE I. 
 
 CaO taken as 
 CaClj. 
 
 Ammonium 
 oxalate. 
 
 Volume at 
 precipitation. 
 
 CaO found. 
 
 Error. 
 
 grm. 
 
 grm. 
 
 cm 3 
 
 gnu. 
 
 grm. 
 
 0.0656 
 
 0.3 
 
 100 
 
 0.0657 
 
 0.0001+ 
 
 0.0656 
 
 0.3 
 
 100 
 
 0.0656 
 
 0.0000 
 
 0.0656 
 
 0.3 
 
 150 
 
 0.0658 
 
 0.0002+ 
 
 0.0656 
 
 0.3 
 
 100 
 
 0.0655 
 
 0.0001- 
 
 0.0985 
 
 0.5 
 
 175 
 
 0.0981 
 
 0.0004- 
 
 0.1313 
 
 0.6 
 
 150 
 
 0.1315 
 
 0.0002+ 
 
 0.1313 
 
 0.6 
 
 200 
 
 0.1315 
 
 0.0002+ 
 
 Extended washing with hot water, however, is to be avoided 
 after the precipitant, ammonium oxalate, has been removed. 
 In one experiment, for example, in which the precipitate, on 
 the felt, was washed fourteen times with portions of about 
 50 cm 3 each of hot water, each portion bleached from 2-6 
 drops of approximately ^ permanganate, making a total loss 
 of 0.0034 grm. of calcium oxide. 
 
 Strontium Oxalate. 
 
 Souchay and Lenssen * state that strontium oxalate is solu- 
 ble in 12,000 parts of water. This fact would seem sufficient 
 to warrant the study of the quantitative separation of stron- 
 tium as the oxalate. In the work which follows strontium 
 
 Ann. Chem. (Liebig), cii, 35. 
 
AND BARIUM AS THE OXALATE S. 
 
 375 
 
 oxalate has been precipitated both in alcoholic solution and 
 in water solution, and for convenience these two conditions 
 of precipitation will be discussed separately. 
 
 All the strontium salts, of established purity, were stan- 
 dardized by precipitation with sulphuric acid in a solution 
 containing at least one-half its volume of alcohol, and with 
 some solutions confirmatory standards were also obtained by 
 evaporation with sulphuric acid. 
 
 Precipitation in Alcoholic Solution. To determine the com- 
 pleteness of the precipitation in alcoholic solution strontium 
 nitrate was precipitated by ammonium oxalate in a solution 
 containing one-third of its volume of alcohol, the mixture was 
 allowed to stand over night, the liquid was filtered off on 
 asbestos, and the precipitate was treated in the capped filter- 
 ing crucible with sulphuric acid, ignited, and weighed as the 
 sulphate. The results are given in Table II. It is plain from 
 
 TABLE IL 
 
 SrO taken as 
 Sr(N03) 2 . 
 
 Ammonium 
 oxalate. 
 
 Volume at 
 precipitation. 
 
 Volume of 
 alcohol. 
 
 SrO found 
 
 as SrS0 4 . 
 
 Difference. 
 
 grm. 
 
 grm. 
 
 cm* 
 
 
 griii. 
 
 griii. 
 
 0.2434 
 
 0.8 
 
 180 
 
 
 
 0.2440 
 
 0,0006+ 
 
 0.2434 
 
 0.8 
 
 180 
 
 
 
 0.2437 
 
 0.0003+ 
 
 0.0022 
 
 0.2 
 
 100 
 
 
 
 0.0022 
 
 0.0000 
 
 0.0013 
 
 0.2 
 
 100 
 
 
 
 0.0014 
 
 0.0001+ 
 
 0.0004 
 
 0.04 
 
 100 
 
 
 
 0.0004 
 
 0.0000 
 
 the results in this table that the precipitation of even small 
 amounts of the strontium salt from a solution containing one- 
 third of its volume of alcohol is practically complete. 
 
 To determine the minimum amount of alcohol necessary for 
 the complete precipitation of the strontium oxalate, experiments 
 were made using varying proportions of 85 per cent alcohol 
 with different amounts of ammonium oxalate, and the filtrates 
 from such experiments were tested for strontium by the addi- 
 tion of more alcohol. The results given in Table III show 
 that when a moderate excess of ammonium oxalate is present, 
 a volume of 85 per cent alcohol, amounting to one-fifth of 
 
376 ESTIMATION OF CALCIUM, STRONTIUM, 
 
 TABLE III. 
 
 SrO present 
 as Sr(N0 3 ) 3 . 
 
 Ammonium 
 oxalate. 
 
 Volume of 
 liquid. 
 
 Proportion 
 of 85 per cent 
 
 alcohol. 
 
 SrO found in 
 filtrates, weighed 
 as SrSO 4 . 
 
 grin. 
 
 grm. 
 
 cm 3 
 
 
 grm. 
 
 0.1 
 
 0.4 
 
 100 
 
 1 
 
 0.0000 
 
 0.1 
 
 0.4 
 
 100 
 
 JL 
 
 0.0000 
 
 0.1 
 
 0.4 
 
 100 
 
 1 
 
 0.0004 
 
 0.1 
 
 0.2 
 
 100 
 
 1 
 
 0.0000 
 
 0.1 
 
 0.2 
 
 100 
 
 JL 
 
 0.0009 
 
 0.1 
 
 0.2 
 
 100 
 
 JL 
 
 0.0020 
 
 0.1 
 
 0.1 
 
 100 
 
 * 
 
 0.0002 
 
 the whole, is sufficient to complete the precipitation of the 
 strontium as the oxalate. 
 
 The conditions under which strontium oxalate is insoluble 
 having been determined, the process for the volumetric esti- 
 mation of strontium was carried out as follows: The hot 
 solution of a strontium salt was precipitated with ammonium 
 oxalate, 85 per cent alcohol, amounting to from one-fifth to one- 
 third the total volume, was added, the mixture was allowed to 
 stand over night, and the clear liquid was decanted on an 
 asbestos filter. The precipitate was washed with a mixture 
 of equal parts of 85 per cent alcohol and water, transferred to 
 the filter, dried in the filtering crucible over a flame to free 
 it from alcohol, returned to the beaker previously dried, treated 
 with sulphuric acid, or with 5-10 cm 3 of hydrochloric acid 
 (in the latter case 0.5-1.0 grm. of a manganous salt being 
 added) and the liberated oxalic acid was titrated by perman- 
 ganate. The results obtained by this method are accurate 
 and are given in Table IV. 
 
 In the last experiment in which a comparatively large 
 amount of strontium salt was present and the dilution low, 
 there is a slight tendency towards a minus error, due probably 
 to the occlusion of some oxalic acid by the strontium sulphate 
 formed. This phenomenon would favor titration at greater 
 dilution when sulphuric acid is used to liberate the oxalic 
 acid from large amounts of strontium oxalate. 
 
 Precipitation in Water Solution. In order to determine the 
 
AND BARIUM AS THE OXALATES. 
 
 377 
 
 TABLE IV. 
 VOLUME DURING TITBATION 150-250 CM S . 
 
 SrO taken 
 as 
 
 Ammonium 
 
 Volume at 
 precipita- 
 tion. 
 
 Propor- 
 tion of 
 85 per cent 
 alcohol. 
 
 Acid pres- 
 ent during 
 titratiou. 
 
 SrO 
 
 found. 
 
 Error. 
 
 grm. 
 0.0974 
 
 grm. 
 0.4 
 
 cm 3 
 100 
 
 
 
 HC1 
 
 grm. 
 0.0973 
 
 grm. 
 0.0001- 
 
 0.0974 
 
 0.4 
 
 100 
 
 
 
 HC1 
 
 0.0983 
 
 0.0009+ 
 
 0.0974 
 
 0.4 
 
 100 
 
 
 
 HC1 
 
 0.0975 
 
 0.0001+ 
 
 0.0974 
 
 0.8 
 
 100 
 
 
 
 HC1 
 
 0.0981 
 
 0.0007+ 
 
 0.1948 
 
 0.4 
 
 200 
 
 
 
 HC1 
 
 0.1943 
 
 0.0005- 
 
 0.1948 
 
 0.8 
 
 200 
 
 
 
 HC1 
 
 0.1942 
 
 0.0006- 
 
 0.0974 
 
 0.4 
 
 100 
 
 
 
 H 2 S0 4 
 
 0.0970 
 
 0.0004- 
 
 0.0974 
 
 0.4 
 
 100 
 
 
 
 H 2 S0 4 
 
 0.0977 
 
 0.0003+ 
 
 0.0974 
 
 0.4 
 
 100 
 
 ] 
 
 
 H. 2 S0 4 
 
 0.0976 
 
 0.0002+ 
 
 0.1948 
 
 0.6 
 
 150 
 
 
 
 H 2 S0 4 
 
 0.1938 
 
 0.0010- 
 
 degree of precipitation of strontium salts in water solution, 
 strontium oxide, taken as the nitrate, was precipitated by 
 ammonium oxalate, the mixture was allowed to stand over 
 night, filtered on asbestos, the precipitate was washed with 
 water containing one-half its volume of 85 per cent alcohol, 
 treated hi the capped crucible with a few drops of sulphuric 
 acid, ignited, and weighed as the sulphate. The result gave 
 0.0973 grm. of strontium oxide instead of 0.0974 grm. taken. 
 The precipitation, therefore, of strontium oxalate, in water 
 solution with a sufficient excess of ammonium oxalate present, 
 is practically complete. 
 
 To determine the amount of ammonium oxalate necessary 
 for the precipitation of strontium salts in water solution, 
 experiments were made in which strontium oxalate was pre- 
 cipitated in the presence of varying amounts of ammonium 
 oxalate, allowed to stand over night, the clear liquid was 
 decanted on asbestos, and the precipitate was washed twice 
 with 10-20 cm 3 of cold water. The results obtained by 
 the estimation of the oxalic acid by permanganate show 
 that an amount of ammonium oxalate several times larger 
 than that required for the theoretical formation of strontium 
 oxalate is necessary for the separation of the strontium oxalate. 
 The experiments are recorded in Table V. 
 
378 ESTIMATION OF CALCIUM, STRONTIUM, 
 
 TABLE V. 
 
 
 
 
 
 
 
 SrO, 
 taken as 
 Sr(NOa),. 
 
 Ammonium 
 
 oxalate. 
 
 Volume 
 at precipi- 
 tation. 
 
 Acid present 
 during 
 titratiou. 
 
 SrO 
 found. 
 
 Error. 
 
 gnn. 
 
 grm. 
 
 cms 
 
 
 grm. 
 
 grm. 
 
 0.0487 
 0.0487 
 
 0.064 
 0.0768 
 
 100 
 100 
 
 H 2 S0 4 
 H 2 SO 4 
 
 0.0441 
 0.0465 
 
 0.0046- 
 0.0022- 
 
 0.0487 
 
 0.16 
 
 100 
 
 H 2 SO 4 
 
 0.0488 
 
 0.0001+ 
 
 0.0974 
 
 0.128 
 
 100 
 
 H 2 S0 4 
 
 0.0939 
 
 0.0025- 
 
 0.0974 
 
 0.16 
 
 100 
 
 H 2 S0 4 
 
 0.0959 
 
 0.0015- 
 
 0.0974 
 
 0.32 
 
 100 
 
 H 2 S0 4 
 
 0.0976 
 
 0.0001+ 
 
 The solvent action of a large amount of water on a pre- 
 cipitate of strontium oxalate was tested by washing a pre- 
 cipitate equivalent to 0.0974 grm. of the oxide with 150 cm 3 
 of cold water. The precipitate, when weighed as the sulphate, 
 showed a loss of 0.0033 grm., as the oxide, which amount was 
 subsequently recovered from the filtrate by the addition of 
 ammonium oxalate and alcohol. Plainly excessive washing 
 with water is to be avoided. In the estimation, therefore, of 
 strontium precipitated as the oxalate in water solution, the 
 amount of water used in washing was limited. It was found 
 that 40-50 cm 3 of water judiciously applied was sufficient to 
 wash out the ammonium salt without producing appreciable 
 solvent effect upon the strontium oxalate. The process of 
 treatment was similar to that used in the precipitations from 
 alcoholic solution, excepting that no alcohol was added to the 
 solution, that the washing was effected with a limited amount 
 of water, and that, there being no alcohol present to effect the 
 titration, the precipitate was not dried before treatment with 
 permanganate. The results are given hi Table VI. 
 
 In the results recorded in section A of Table VI, the stron- 
 tium oxalate was treated with sulphuric acid and titrated at 
 80, the volume being 200-300 cm 3 ; while hi the experiments 
 given in section B, the precipitate was treated with hydro- 
 chloric acid and titrated at 35-45, at a volume of 100-200 cm 3 , 
 after the addition of 0.5-1.0 grm. of manganous chloride. The 
 results show that 0.1 grm. of strontium salt, calculated as the 
 
AND BARIUM AS THE OXALATES. 
 
 379 
 
 oxide, may be estimated as the oxalate with a fair degree of 
 accuracy when precipitated in 100-250 cm 3 of water by a 
 
 TABLE VI. 
 
 SrO, 
 taken as 
 Sr(NOs),. 
 
 Ammonium 
 oxalate. 
 
 Volume 
 at precipi- 
 tation. 
 
 Acid present 
 during 
 titration. 
 
 SrO 
 found. 
 
 Error. 
 
 A. 
 
 grm. 
 
 0.0974 
 0.0974 
 0.0974 
 0.0974 
 0.0974 
 0.0974 
 0.0974 
 0.0974 
 0.0974 
 
 grm. 
 0.5 
 0.5 
 0.5 
 0.5 
 0.8 
 0.8 
 1.0 
 2.0 
 2.0 
 
 Cfflg 
 
 100 
 100 
 100 
 100 
 100 
 100 
 100 
 100 
 100 
 
 H2SO 4 
 H 2 S0 4 
 H 2 S0 4 
 H 2 S0 4 
 H 2 S0 4 
 H 2 S0 4 
 H 2 S0 4 
 H 2 S0 4 
 H 2 SO 4 
 
 grm. 
 0.0966 
 0.0985 
 0.0977 
 0.0963 
 0.0981 
 0.0966 
 0.0965 
 0.0963 
 0.0970 
 
 grm. 
 0.0008- 
 0.0011+ 
 0.0003+ 
 0.0011- 
 0.0007+ 
 0.0008- 
 0.0009- 
 0.0011- 
 0.0004- 
 
 SrO, 
 taken as B. 
 8rCl 2 . 
 
 
 0.0778 
 0.0778 
 0.0778 
 0.0778 
 
 0.5 
 0.5 
 0.5 
 0.5 
 
 100 
 100 
 100 
 100 
 
 H 2 S0 4 
 H2S0 4 
 H 2 SO 4 
 H 2 S0 4 
 
 0.0792 
 0.0767 
 0.0776 
 0.0776 
 
 0.0014+ 
 0.0011- 
 0.0002- 
 0.0002- 
 
 SrO, 
 taken as 
 Sr(N0 8 ) 2 . 
 
 0.0974 
 0.0974 
 0.0974 
 0.0974 
 0.0974 
 0.0974 
 0.0974 
 0.0974 
 
 0.8 
 2.0 
 0.8 
 0.8 
 0.8 
 0.8 
 0.8 
 0.8 
 
 250 
 250 
 100 
 100 
 100 
 100 
 100 
 100 
 
 H 2 S0 4 
 H 2 S0 4 
 HCI 
 HCI 
 HCI 
 HCI 
 HCI 
 HCI 
 
 0.0973 
 0.0975 
 0.0971 
 0.0980 
 0.0975 
 0.0980 
 0.0973 
 0.0978 
 
 0.0001- 
 0.0001+ 
 0.0003- 
 0.0006+ 
 0.0001+ 
 0.0006+ 
 0.0001- 
 0.0004+ 
 
 C. 
 
 0.2425 
 0.2436 
 0.2436 
 0.2436 
 0.2436 
 0.2436 
 
 0.384 
 0.384 
 0.64 
 0.8 
 2.0 
 2.0 
 
 125 
 125 
 125 
 125 
 125 
 125 
 
 H 2 S0 4 
 H 2 S0 4 
 H 2 S0 4 
 H 2 S0 4 
 H 2 S0 4 
 H 2 S0 4 
 
 0.2376 
 0.2402 
 0.2411 
 0.2367 
 0.2376 
 0.2402 
 
 0.0049- 
 0.0034- 
 0.0025- 
 0.0069- 
 0.0060- 
 0.0034- 
 
 D. 
 
 0.2436 
 0.2436 
 0.2436 
 0.2436 
 
 0.8 
 0.8 
 2.0 
 2.0 
 
 250 
 250 
 250 
 250 
 
 H 2 SO 4 
 H 2 S0 4 
 H 2 S0 4 
 H 2 S0 4 
 
 0.2443 
 0.2446 
 0.2440 
 0.2431 
 
 0.0007+ 
 0.0010+ 
 0.0004+ 
 0.0005- 
 
380 ESTIMATION OF CALCIUM, STRONTIUM, 
 TABLE VI (continued). 
 
 SrO, 
 taken as 
 Sr(N0 8 ),. 
 
 Ammonium 
 
 oxalate. 
 
 Volume 
 at precipi- 
 tation. 
 
 Acid present 
 during 
 titratiou. 
 
 SrO 
 found. 
 
 Error. 
 
 E. 
 
 grui. 
 
 grm. 
 
 cm 8 
 
 
 grm. 
 
 grm. 
 
 0.2436 
 
 0.8 
 
 500 
 
 H 2 S0 4 
 
 0.2396 
 
 0.0040- 
 
 0.2436 
 
 2.0 
 
 500 
 
 H 2 S0 4 
 
 0.2403 
 
 0.0033- 
 
 0.2436 
 
 2.0 
 
 500 
 
 H 2 S0 4 
 
 0.2413 
 
 0.0023- 
 
 0.2436 
 
 4.0 
 
 600 
 
 H 2 S0 4 
 
 0.2410 
 
 0.0026- 
 
 0.2436 
 
 8.0 
 
 500 
 
 H 2 S0 4 
 
 0.2407 
 
 0.0029- 
 
 0.4872 
 
 2.0 
 
 500 
 
 H 2 SO 4 
 
 0.4837 
 
 0.0035- 
 
 0.4872 
 
 4.0 
 
 500 
 
 H 2 S0 4 
 
 0.4855 
 
 0.0017- 
 
 0.5430 
 
 5.0 
 
 500 
 
 HnS0 4 
 
 0.5422 
 
 0.0008- 
 
 0.4579 
 
 10.0 
 
 500 
 
 H 2 SO 4 
 
 0.4554 
 
 0.0025- 
 
 0.7307 
 
 4.0 
 
 500 
 
 HC1 
 
 0.7262 
 
 0.0045- 
 
 sufficient excess of ammonium oxalate. In the experiments 
 recorded in section C, in which the amount of strontium salt 
 in 125 cm 3 of water is increased, a negative error is intro- 
 duced, which is not diminished by the presence of a large 
 amount of ammonium oxalate, but when the dilution is in- 
 creased to 250 cm 3 , as is the case in experiments given in 
 section D, so that the conditions correspond more nearly to 
 those recorded in sections A and B, the errors fall to a 
 minimum. In the experiments recorded in section E, in 
 which the dilution is increased to 500 cm 3 , an error is in- 
 troduced which is not prevented by the presence of a large 
 excess of ammonium oxalate and which is independent of the 
 amounts of strontium salt used. 
 
 Eight of the water filtrates and wash waters obtained in 
 the experiments recorded in Table VI were tested for traces 
 of strontium by the addition of alcohol, and in all cases a 
 small amount of strontium was found, amounting, in the 
 average, to 0.0010 grm. in 100 cm 3 of water. 
 
 Barium Oxalate. 
 
 Barium oxalate according to Souchay and Lenssen * is 
 soluble in 2590 parts of cold water, and according to Berg- 
 
 * Ann. Chem. (Liebig), xc, 102. 
 
AND BARIUM AS THE OXALATES. 
 
 381 
 
 man * is scarcely at all soluble in alcohol. The attempt was 
 made to estimate barium by precipitation with ammonium 
 oxalate in a mixture containing alcohol. It was found that 
 in filtrates from oxalate precipitations in which 0.1-0.2 grm. 
 of barium oxide, taken as the nitrate, had been precipitated in 
 volumes of 100 cm 8 , containing 30 cm 3 of absolute alcohol, 
 and allowed to stand over night, treatment with sulphuric 
 acid gave barium sulphate amounting in the average to no 
 more than 0.0001 grm. of barium oxide. The insolubility of 
 barium oxalate under these conditions, therefore, is practically 
 complete. 
 
 The process for the estimation of barium was as follows : 
 Ammonium oxalate was added to a solution of a barium salt, 
 
 TABLE VII. 
 
 BaO taken 
 as Ba(N0 3 ) 2 . 
 
 Ammonium 
 oxalate. 
 
 Volume at 
 precipita- 
 tion. 
 
 Acid present 
 during 
 titration. 
 
 BaO 
 
 found. 
 
 Error. 
 
 A. 
 
 grm. 
 
 0.1165 
 0.1165 
 0.1165 
 0.1165 
 0.1165 
 0.1165 
 0.1165 
 0.2330 
 0.2330 
 0.2330 
 
 grin. 
 
 0.2 
 0.2 
 0.2 
 0.2 
 0.2 
 0.2 
 0.2 
 0.4 
 0.4 
 0.4 
 
 cm 3 
 100 
 100 
 100 
 100 
 100 
 100 
 100 
 100 
 100 
 100 
 
 HCI 
 HCI 
 HCI 
 HCI 
 HCI 
 HCI 
 HCI 
 HCI 
 HCI 
 HCI 
 
 gnu. 
 
 0.1177 
 0.1170 
 0.1164 
 0.1151 
 0.1165 
 0.1176 
 0.1164 
 0.2319 
 0.2335 
 0.2342 
 
 grm. 
 0.0012+ 
 0.0005+ 
 0.0001- 
 0.0014- 
 0.0000 
 0.0011+ 
 0.0001- 
 0.0011- 
 0.0005+ 
 0.0012+ 
 
 BaO taken 
 as BaCl 2 . 
 
 0.4 
 
 0.4 
 0.4 
 0.4 
 0.4 
 
 100 
 100 
 100 
 100 
 100 
 
 HCI 
 HCI 
 HCi 
 HCI 
 HCI 
 
 0.0952 
 0.0939 
 0.0941 
 0.1893 
 0.1892 
 
 0.0010+ 
 0.0003- 
 0.0001- 
 0.0009+ 
 0.0008+ 
 
 0.0942 
 0.0942 
 0.0942 
 0.1884 
 0.1884 
 
 B. 
 
 0.0942 
 0.1884 
 0.0942 
 
 0.2 
 0.4 
 0.2 
 
 200 
 200 
 500 
 
 H 2 S0 4 
 H 2 S0 4 
 H 2 S0 4 
 
 0.0858 
 0.1732 
 0.0857 
 
 0.0084- 
 0.0152- 
 0.0085- 
 
 * Bergman's Essays, i, 320. 
 
382 ESTIMATION OF CALCIUM, STRONTIUM, 
 
 containing 30 per cent of its volume of absolute alcohol, the 
 mixture was allowed to stand over night, filtered on asbestos, 
 the precipitate was washed by decantation with 100-200 cm 3 
 of water containing 30 per cent of absolute alcohol, and dried 
 over a flame to insure the removal of the alcohol. The cruci- 
 ble containing the precipitate was returned to the beaker also 
 previously dried over a flame, 100-200 cm 3 of water, 5-10 
 cm 3 of strong hydrochloric acid, and 0.5-1.0 grm. of manganous 
 chloride were added, and the solution was titrated at 35 -45 
 with permanganate. The results of the experiments, given in 
 Table VII, A, show that barium, either as the nitrate or 
 chloride, may be estimated in the manner described with a fair 
 degree of accuracy. 
 
 In the experiments given in section B of Table VII, the 
 precipitate of barium oxalate was treated with sulphuric acid 
 after the addition of the stated amount of water. The results 
 show a large loss of oxalic acid probably due to the occlusion 
 of some of the oxalic acid by the barium sulphate. This fact 
 must prevent the use of sulphuric acid in an analytical 
 process which depends upon the liberation of oxalic acid from 
 barium oxalate. 
 
 Gravimetric Estimation of the Oxalates of Strontium 
 and Barium. 
 
 It is well known that calcium may be weighed as the 
 carbonate after a careful ignition of the oxalate, and it would 
 seem probable that strontium might also be weighed as the 
 carbonate. Precipitates of strontium oxalate, on asbestos, 
 were ignited in a capped crucible from 2-8 minutes in the 
 flame of a Bun sen burner and weighed as the carbonate, and 
 in a single case the carbonate thus produced was converted by 
 treatment with sulphuric acid to the sulphate and weighed as 
 such. The results are given in Table VIII, and while they 
 all show a very slight loss, which amounts in experiment (3) 
 to one milligram, when one-fourth of a gram of strontium 
 oxide taken as the nitrate was used, still the results are fairly 
 accurate. 
 
AND BARIUM AS THE OXALATES. 
 TABLE VUL 
 
 383 
 
 Ezp. 
 
 BrO taken 
 as Sr(NO s ),. 
 
 SrO calculated 
 from SrC0 3 
 found. 
 
 SrO calculated 
 from SrSO 4 
 found. 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 (2) 
 
 0.1120 
 0.1120 
 
 0.1113 
 0.1116 
 
 .... 
 
 (3) 
 
 0.2435 
 
 0.2425 
 
 0.2437 
 
 Precipitates of barium oxalate were also ignited from 5-10 
 minutes and weighed as the carbonate. The results are given 
 in Table IX, and are fairly accurate. 
 
 TABLE IX. 
 
 Exp. 
 
 BaO taken 
 as Ba(N0 3 ) 2 . 
 
 BaO calculated 
 from BaCO 8 
 found. 
 
 Difference. 
 
 (1) 
 (2) 
 (3) 
 
 griii. 
 
 0.2912 
 0.2912 
 0.2912 
 
 grm. 
 0.2909 
 0.2901 
 0.2901 
 
 grm. 
 0.0003- 
 0.0011- 
 0.0011- 
 
 The results of this work may be summarized as follows: 
 In the estimation of calcium by titration of the oxalate with 
 permanganate, accurate results may be obtained when hydro- 
 chloric acid (with a manganous salt) is used as the solvent. 
 Strontium salts may be precipitated by ammonium oxalate 
 with practical completeness in a solution containing one-fifth 
 of its volume of 85 per cent alcohol, and with approximate 
 completeness from water solutions at a dilution not exceeding 
 250 cm 3 . Furthermore strontium oxalate may be titrated by 
 permanganate with accuracy when either sulphuric acid or 
 hydrochloric acid (with a manganous salt) is used to liberate 
 the oxalic acid. Barium may be precipitated with practical 
 completeness by ammonium oxalate in a solution containing 
 30 per cent of alcohol, and the barium oxalate thus obtained 
 may be dissolved in hydrochloric acid and titrated by per- 
 manganate after the addition of a manganous salt. Strontium 
 and barium oxalates may be converted to carbonates by 
 ignition, and weighed as such. 
 
XLIX 
 
 THE ACTION OF SODIUM THIOSULPHATE ON 
 SOLUTIONS OF METALLIC SALTS AT HIGH 
 TEMPERATURES AND PRESSURES. 
 
 BY JOHN T. NORTON, JR. 
 
 THE use of sodium thiosulphate as a substitute for hydro- 
 gen sulphide in effecting precipitations and its application 
 in the case of arsenic, antimony, copper, and platinum was 
 suggested by Himly * before the middle of the present cen- 
 tury. Thirteen years later Vohl f and Slater, independently $ 
 drew attention to this use of sodium thiosulphate and ex- 
 tended the investigation to salts of tin, mercury, silver, gold, 
 lead, bismuth, and cadmium. Slater in addition studied the 
 action of sodium thiosulphate upon chromic acid, molybdates, 
 ferrous and ferric ferrocyanides, ferric sulphocyanides and 
 potassium permanganate. Following out these lines, the 
 precipitation of copper, together with arsenic antimony, by 
 treating with sodium thiosulphate the hot solution contain- 
 ing sulphuric acid, and the separation of these elements from 
 tin, zinc, iron, nickel, cobalt, and manganese has been advo- 
 cated by Westmoreland ; and quite recently Faktor || has 
 studied the action of sodium thiosulphate upon neutral salts 
 of several of the elements mentioned, as well as the modi- 
 fying influence of ammonium chloride and other salts upon 
 the course of the reaction. 
 
 Subsequently to the work of Himly, Vohl, and Slater, 
 Chancel ** developed his well known method for the precipi- 
 
 * Ann. Chem. (Liebig), xliii, 150. 
 t Ann. Chem. (Liebig), xcvi, 237. 
 
 t Chemical Gazette, 1855, p. 369. Jour. Soc Chem. Ind., v, 61. 
 
 || Chem. Centralblatt, 1900, ii, 20, 67, 239, 594. 
 * Comp. rend., xlvi, 987. 
 
ACTION OF SODIUM THIOSULPHATE, ETC. 385 
 
 tation of aluminum as the hydroxide and its separation from 
 salts of iron by boiling with sodium thiosulphate the nearly 
 neutral solution, containing the salt of aluminum and iron, 
 at suitable dilution ; and upon an extension of the principle 
 of Chancel's separation of aluminum from iron Stromeyer * 
 founded his well known processes for the separation of titan- 
 ium and zirconium from iron. The latter process appears 
 to be fairly trustworthy ; but of Chancel's method, although 
 it has met with wide acceptance, it was shown by Wolcott 
 Gibbs, very soon after its announcement,! that it fails to bring 
 about complete separation of alumina within a reasonable 
 period of boiling, and this result has been confirmed by Zimmer- 
 man, J who has shown that the boiling must be continued fifteen 
 hours in order to complete the precipitation of the alumina. 
 
 It was shown by Dr. Gibbs that when the treatment of salts 
 of aluminum by thiosulphate was carried on in sealed tubes 
 under pressure at 120 C., the precipitation of alumina was 
 complete, and further that the precipitation of sulphides of 
 nickel, cobalt, and iron, though partial under ordinary atmos- 
 pheric pressure, was made complete by heating in sealed tubes 
 to 120-140 C. 
 
 In repeating the experiments of Dr. Gibbs qualitatively and 
 extending them, I have made use of the well known Pfungst 
 tube to secure the necessary pressure. In each experiment 
 a test-tube containing the mixture of an excess of sodium 
 thiosulphate with the salt whose action was studied was 
 placed within the Pfungst tube containing some water, the 
 cover of the latter was set in place and firmly bolted upon 
 a washer of lead, and the whole was submitted to tempera- 
 tures varying from 140 to 200 C. for an hour by immersing 
 in a bath of paramne. After cooling, the test-tube was taken 
 out, the precipitate was filtered off, and the filtrate tested by 
 appropriate reagents to determine the completeness of pre- 
 cipitation. The following table records the details of these 
 experiments. 
 
 * Ann. Chem. (Liebig), cxiii, 127. 
 
 t Zeit. anal. Chem., iii, 389. t Inaug. Diss. Berlin, 1887. 
 
 VOL. u. 25 
 
386 
 
 ACTION OF SODIUM THIOSULPHATE ON 
 
 TABLE I. 
 ACTION OP NA 2 S 2 O 3 ON SALTS UNDER PRESSURE. 
 
 Salts used. 
 
 Precipitates. 
 
 Degree of 
 precipitation. 
 
 SULPHIDES. 
 
 NiSO 4 . 
 CoSO 4 . 
 FeCl 3 . 
 ZnSO 4 . 
 Pb0 2 (C 2 H 8 0) a . 
 Hg(N0 3 V 
 AgN0 8 . 
 CuSO 4 . 
 CdS0 4 . 
 KSbC 4 H 4 (X. 
 Bi(N0 8 ) 8 . 
 
 NiS + S. 
 CoS + S. 
 FeS + S. 
 ZnS + S. 
 PbS + S. 
 HgS + S. 
 Ag 2 S + S. 
 CuS,Cu 2 S, + S. 
 CdS + S. 
 Sb 2 S 8 + S. 
 Bi 2 S 8 + S. 
 
 Complete. 
 Complete. 
 Complete. 
 Complete. 
 Complete. 
 Complete. 
 Complete. 
 Complete. 
 Complete. 
 Complete. 
 Complete. 
 
 HYDROXIDES. 
 
 (NH 4 )A1(S0 4 ) 2 .12H 2 0. 
 K 2 Cr 2 O 7 . 
 K 2 ZrF 6 . 
 KoTiFa. 
 Th(N0 8 ) 4 . 
 
 A10 8 H 3 + S. 
 Cr0 3 H 3 + S. 
 Zr0 4 H 4 + S. 
 Ti0 4 H 4 + S. 
 Th0 4 H 4 + S. 
 
 Complete. 
 Complete. 
 Complete. 
 Complete. 
 Complete. 
 
 ELEMENTS. 
 
 SeO 2 . 
 Te0 2 . 
 
 Se + S. 
 Te + S. 
 
 Complete. 
 Complete. 
 
 SULPHIDES. 
 
 MnS0 4 . 
 AuCl 8 . 
 (NH 4 ) 2 Mo0 4 . 
 
 MnS + S. 
 Au 2 S + S. 
 MoS 8 (?) + S Red liquid. 
 
 Partial. 
 Partial. 
 Partial. 
 
 HYDROXIDES. 
 
 BeCLj. 
 
 Be0 2 H 2 + S. 
 
 Partial. 
 
 UNDETERMINED. 
 
 (NH 4 ) 2 U 2 7 . 
 K 2 PtCle. 
 CeCl 8 . 
 CaCl 2 . 
 SrCL. 
 Bad-. 
 MgS0 4 . 
 NH 4 VO 8 . 
 H 2 KAsO 4 . 
 
 Black. 
 Gray, reddish brown liquid. 
 White, yellow liquid. 
 White, yellow liquid. 
 White, yellow liquid. 
 White, yellow liquid. 
 
 , Brown liquid. 
 
 Partial. 
 Partial. 
 Partial. 
 Partial. 
 Partial. 
 Partial. 
 None. 
 None. 
 None. 
 
SOLUTIONS OF METALLIC SALTS. 387 
 
 A perusal of this table brings to light several interesting 
 facts. It appears that salts of nickel, cobalt, iron, zinc, lead, 
 mercury, silver, copper, cadmium, antimony and bismuth are 
 completely precipitated as sulphides by sodium thiosulphate 
 under the prevailing conditions of temperature and pressure. 
 In the case of manganese precipitation is only partial, and 
 arsenic does not seem to be precipitated from an arsenate with- 
 out the addition of acid. Tin, curiously enough, is not thrown 
 down as the sulphide from a stannous salt, but gives a dirty 
 white precipitate of uncertain composition. Salts of aluminum, 
 chromium, titanium, zirconium and thorium are completely 
 precipitated as the hydroxides ; but in the case of beryllium, 
 which one might expect to act similarly, the precipitation as 
 the hydroxide is incomplete. Salts of selenium and tellurium 
 are reduced, and the elements are precipitated. The precipi- 
 tates obtained with barium, strontium, and calcium were 
 white in a bright yellow liquid, but no study was made of the 
 constitution of either precipitate or liquid. In the case of 
 magnesium there was no precipitate. Salts of molybdenum, 
 vanadium and uranium gave dark colored liquids. Thallium 
 yielded a white spongy mass which on compression was re- 
 duced to a very small bulk without disintegrating. Salts of 
 gold and platinum gave slight dark precipitates, presumably 
 sulphides, surrounded by dark colored liquids. 
 
 The apparatus used in these experiments and described 
 above is easily handled and answers sufficiently well for 
 qualitative purposes. But, obviously, the introduction into 
 precipitates of foreign matter caused by the action of water 
 on the glass of the test-tube and porcelain lining of the 
 Pfungst tube, precludes the possibility of an exact quanti- 
 tative study of the reactions involved. For the subsequent 
 experiments, therefore, conducted upon the same general 
 lines, a digester with an interior cylindrical cavity of about 
 12 cm. in depth by 5 cm. in diameter, and provided with a 
 pressure gauge was employed. As a container for the solu- 
 tions to be tested, use was made of a platinum cylinder, 4 cm. 
 in diameter and 10 cm. deep, provided with a loose cover. 
 
388 
 
 ACTION OF SODIUM THIOSULPHATE ON 
 
 With this apparatus the following quantitative experiments 
 which deal with those elements which are precipitated as 
 hydroxides namely, aluminum, beryllium, chromium, zir- 
 conium, and titanium were made. 
 
 In each case a weighed quantity of the salt taken for the 
 experiment was dissolved in 50 cm 3 of water in the platinum 
 vessel and to this a known amount of sodium thiosulphate 
 was added. The vessel was placed in the digester, and the 
 latter was heated by a Bunsen burner in the customary way 
 until the required pressure was shown on the gauge. The 
 apparatus was then cooled and the platinum vessel removed 
 from the digester. The precipitate was filtered off on ashless 
 paper, ignited, and weighed. 
 
 Experiments with a Salt of Aluminum. 
 
 In a series of experiments made according to the method of 
 Chancel, the results of which are shown in Table II, the 
 solution in water of a weighed portion of pure ammonium 
 alum was treated with an excess of sodium thiosulphate and 
 boiled vigorously for periods varying from ten minutes to 
 half an hour. 
 
 TABLE II. 
 
 Ezp. 
 
 Amount of 
 Alum taken 
 
 Amount of 
 
 A1 2 3 
 found. 
 
 Error. 
 
 
 as A1,O 3 . 
 
 
 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.0537 
 
 Large excess. 
 
 0.0471 
 
 0.0066- 
 
 (2) 
 
 0.0537 
 
 Large excess. 
 
 0.0397 
 
 0.0140- 
 
 is 
 
 0.1083 
 
 Large excess. 
 
 0.0931 
 
 0.0152- 
 
 (4) 
 
 0.1137 
 
 5 grin. 
 
 0.0979 
 
 0.0158- 
 
 (5) 
 
 0.1139 
 
 2 grm. 
 
 0.1002 
 
 0.0137- 
 
 The results substantiate the observations of Gibbs * and of 
 Zimmerman f and show clearly that the boiling of solutions of 
 the aluminum salt and sodium thiosulphate for a reasonable 
 time does not effect the complete precipitation of aluminum as 
 the hydroxide. 
 
 Table III shows the result of submitting solutions of 
 
 * Loc. cit. t Loc. cit. 
 
SOLUTIONS OF METALLIC SALTS. 
 
 389 
 
 ammonium alum treated with varying quantities of sodium 
 thiosulphate to a pressure of 20 atmospheres in the digester. 
 It usually required about 40 minutes to raise the pressure to 
 the limit set; but this limit once reached, the digester was 
 allowed to cool slowly. The duration of an experiment was 
 about two hours. 
 
 TABLE III. 
 
 Alum taken 
 as A1 2 3 . 
 
 Amount of 
 Na,S 2 3 used. 
 
 A1 3 3 found. 
 
 Error. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 0.0565 
 
 5 
 
 0.0633 
 
 0.0068+ 
 
 0.1132 
 
 10 
 
 0.1154 
 
 0.0022-f 
 
 0.1153 
 
 5 
 
 0.1186 
 
 0.0033-f- 
 
 0.1128 
 
 3 
 
 0.1129 
 
 0.0001-f- 
 
 0.1126 
 
 3 
 
 0.1142 
 
 0.0016-f- 
 
 0.1128 
 
 2 
 
 0.1120 
 
 0.0008- 
 
 0.1136 
 
 2 
 
 0.1121 
 
 0.0015 
 
 0.1128 
 
 2.5 
 
 0.1136 
 
 0.0008-f- 
 
 0.1124 
 
 2.5 
 
 0.1127 
 
 0.0003+ 
 
 0.1134 
 
 2.25 
 
 0.1133 
 
 0.0001- 
 
 This table shows that sodium thiosulphate precipitates 
 aluminum completely as the hydroxide when pressure is 
 employed. The high results seen in some of the experiments 
 appear to be due to the difficulty of removing by ignition the 
 large amounts of sulphur found in the action, as well as to the 
 salts mechanically included in the precipitate. The amounts 
 of sulphur and contaminating salts present depend upon the 
 amount of thiosulphate taken ; therefore this should be as 
 small as possible, 2-3 grm. being sufficient to precipitate 
 all the alumina in a gram of alum. When the amount of 
 thiosulphate is reasonably restricted the weights of alumina 
 obtained accord fairly well with the theory. 
 
 Experiments with a Salt of Chromium. 
 
 Up to the tune of the completion of this work nothing 
 appears to have been done upon the precipitation of chromium 
 as the hydroxide by means of sodium thiosulphate. Slater * 
 and Rose f make mention of the action of sodium thiosulphate 
 
 * Loc. cit. t Traite de Chimie Analytique, vol. i, p. 479. 
 
390 
 
 ACTION OF SODIUM THIOSULPHATE ON 
 
 upon chromic acid, bichromates, and neutral chromates, but 
 give no quantitative data. Recently, however, F. Faktor * has 
 studied the action of sodium thiosulphate on chromium 
 compounds. This investigator has found that if aqueous 
 solutions of potassium bichromate and sodium thiosulphate 
 are boiled together a brown precipitate of hydrated Cr 2 O 8 , 
 CrO 8 separates out and the liquid turns yellow owing to the 
 formation of normal chromate. A solution of potassium 
 chromate is unaffected by boiling with thiosulphate but in 
 presence of ammonium or of magnesium chloride the chromium 
 is separated rapidly and completely in the same form as with 
 the bichromate, and after continued boiling with an excess of 
 thiosulphate all the chromium present is precipitated. Faktor 
 also found that a solution of chromic chloride is completely 
 decomposed by continued boiling with thiosulphate, chromic 
 hydroxide and sulphur being precipitated. 
 
 In the experiments shown in Table IV a weighed quantity 
 of pure potassium bichromate was dissolved in water, a known 
 amount of sodium thiosulphate added, and the whole submitted 
 to a pressure of 20 atmospheres in the digester. After cooling, 
 the precipitate was filtered off on an ashless paper, ignited and 
 weighed as Cr 2 O 3 . 
 
 TABLE IV. 
 
 Exp. 
 
 K,Cr 2 O 7 taken 
 as Cr 2 O 3 . 
 
 Amount of 
 Na 2 S,0 6 . 
 
 Cr 2 O s found. 
 
 Error. 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.1330 
 
 3 
 
 0.1341 
 
 0.0011+ 
 
 (2) 
 
 0.1330 
 
 2.5 
 
 0.1326 
 
 0.0004- 
 
 (3) 
 
 0.1322 
 
 2.5 
 
 0.1318 
 
 0.0004- 
 
 (4) 
 
 0.1303 
 
 2 
 
 0.1303 
 
 0.0000 
 
 (5) 
 
 0.1301 
 
 2 
 
 0.1310 
 
 0.0009+ 
 
 6) 
 
 0.1320 
 
 2 
 
 0.1322 
 
 0.0002+ 
 
 The results of these experiments are very satisfactory, and 
 show that under pressure sodium thiosulphate precipitates 
 chromium rapidly and completely as the hydroxide. It is 
 advisable to use as small a quantity of thiosulphate as possible 
 
 Zeitschr. anal. Chem., 1900, xxxix, 345. 
 
SOLUTIONS OF METALLIC SALTS. 391 
 
 in order to prevent the presence of much free sulphur in the 
 precipitate. 
 
 Experiments with a Salt of Beryllium. 
 
 In experiments dealing with beryllium the salt used was the 
 chloride, a certain amount of which was dissolved in water 
 diluted to a liter and the amount of beryllium present deter- 
 mined by precipitating with ammonia and weighing as the 
 oxide. Measured quantities of this solution were drawn from 
 a burette as required. "When a solution of a salt of beryllium 
 and sodium thiosulphate are merely boiled together nearly all 
 the beryllium remains in solution. It was expected that the 
 use of pressure Would 'throw out all the beryllium, but, 
 curiously enough, when solutions of beryllium chloride and 
 sodium thiosulphate were submitted in the digester to pressures 
 ranging from 10 to 80 atmospheres only a partial precipitation 
 of the hydroxide took place. 
 
 Experiments with Salts of Zirconium. 
 
 To prepare a standard solution of the salt of zirconium it 
 was found to be most convenient to heat the double fluoride 
 of potassium and zirconium with sulphuric acid, evaporate to 
 dryness in platinum, dissolve the zirconium sulphate remaining 
 in water and enough sulphuric acid to prevent the precipitation 
 of the basic salt, and dilute to standard volume. Measured 
 portions of the solution were taken from a burette as required 
 for the experiments. The presence, however, of so large an 
 amount of sulphuric acid as was necessary to keep the 
 zirconium salt in solution tends to decompose sodium thio- 
 sulphate so rapidly that it was found necessary to nearly 
 neutralize the solution with ammonium carbonate before 
 adding the sodium thiosulphate. The solution of zirconium 
 sulphate was standardized by precipitating with ammonia and 
 weighing as the oxide. 
 
 In experiment (1) of Table V, the solutions of zirconium 
 sulphate and sodium thiosulphate were boiled together for a 
 few minutes and then the precipitate filtered off, ignited, and 
 
392 
 
 ACTION OF SODIUM THIOSULPHATE ON 
 
 weighed as the oxide. In experiments (2)-(5) inclusive similar 
 solutions were submitted to a pressure of 20 atmospheres in 
 the digester. 
 
 TABLE V. 
 
 Exp. 
 
 Zr0 2 
 taken. 
 
 Na 2 S 2 3 
 taken. 
 
 ZrO. 
 found. 
 
 Error. 
 
 
 grin. 
 
 grru. 
 
 grm. 
 
 grm. 
 
 (1) 
 
 0.0658 
 
 3 
 
 0.0651 
 
 0.0007- 
 
 2 
 
 0.0658 
 
 3 
 
 0.0676 
 
 0.0016+ 
 
 (3) 
 
 0.0666 
 
 2 
 
 0.0670 
 
 0.0004+ 
 
 (4) 
 
 0.0641 
 
 2 
 
 0.0648 
 
 0.0007+ 
 
 (5) 
 
 0.0641 
 
 2 
 
 0.0645 
 
 0.0004+ 
 
 These results clearly show that sodium' thiosulphate precipi- 
 tates zirconium completely as the hydroxide either with or 
 without the aid of pressure. 
 
 Experiments with a Salt of Titanium. 
 
 The solution of the salt of titanium was obtained by treating 
 the double fluoride of potassium and titanium with sulphuric 
 acid, evaporating to dryness, and dissolving the residue in 
 sulphuric acid and water. The solution was standardized by 
 precipitating the titanium hydroxide with ammonia and then 
 adding an excess of acetic acid as recommended by Gooch.* 
 This method of procedure avoids the tendency to excessive 
 weight observed when the titanium hydroxide is precipitated 
 by ammonia in presence of salts of the alkalies. 
 
 In the following table is shown the effect of treating a 
 solution of titanium sulphate with sodium thiosulphate. 
 
 TABLE VI. 
 
 Exp. 
 
 Ti0 2 
 taken. 
 
 Na 2 S 2 
 taken. 
 
 TiO 2 
 
 found. 
 
 Error. 
 
 
 grm. 
 
 grm. 
 
 grm. 
 
 gnn. 
 
 (1) 
 
 0.0240 
 
 2 
 
 0.0237 
 
 0.0003- 
 
 (2) 
 
 0.0240 
 
 2 
 
 0.0240 
 
 0.0000 
 
 (3) 
 
 0.0240 
 
 2 
 
 0.0240 
 
 0.0000 
 
 * Am. Chem. Jour., vii, 285. 
 
SOLUTIONS OF METALLIC SALTS. 393 
 
 Experiment (1) was conducted by merely boiling a solution of 
 the reagents named above, filtering off the precipitate and 
 weighing as the oxide. In experiments (2) and (3) the 
 solution of titanium sulphate and sodium thiosulphate was 
 submitted to a pressure of 20 atmospheres in the digester. 
 
 These results show that titanium is completely precipitated 
 by sodium thiosulphate either with or without the aid of 
 pressure. 
 
 To recapitulate : I have shown that sodium thiosulphate 
 will completely precipitate aluminum, chromium, zirconium 
 and titanium as the hydroxides with the aid of high tempera- 
 ture and pressure. Beryllium is only partially precipitated 
 under similar conditions. Mere boiling for a reasonable time 
 will not precipitate aluminum and chromium, but it is suffi- 
 cient in the case of zirconium and titanium. 
 
SYSTEMATIC INDEX. 
 
 LABORATORY APPLIANCES AND PREPARATIONS. 
 
 Laboratory Apparatus (Gooch), I, 141. Generation of Chlorine (Gooch and 
 Kreider), 1, 260. Preparation of Perchloric Acid (Kreider), 1, 282. Labora- 
 tory Apparatus (Kreider), I, 306. 
 
 INORGANIC CHEMISTRY. 
 
 Interaction of Potassium Permanganate and Sulphuric Acid (Gooch and Dan- 
 ner), I, 145. Reducing agents on lodic Acid (Roberts) 1,250. Existence 
 of Selenium Monoxide (Peirce), I, 385. Condition of Oxidation of Man- 
 ganese precipitated by the Chlorate Process (Gooch and Austin), II, 85. 
 Action of Carbon Dioxide on Soluble Borates (Jones), II, 100. Action of 
 Acetylene on Oxides of Copper (Gooch and Baldwin), II, 276. Action of 
 Sodium Thiosulphate on Solutions of Metallic Salts at High Temperatures 
 and Pressures (Norton), II, 384. 
 
 ORGANIC CHEMISTRY. 
 
 Blue Iodide of Starch (Roberts) I, 236. Action of Urea and Sulphocarbanilide 
 on Acid Anhydrides (Dunlap), I, 355. Action of Urea and Primary Amines 
 on Maleic Anhydrides (Dunlap and Phelps), 11,42. Ethers of Toluquinone- 
 oxime and Space Isomerism of Nitrogen (Bridge and Morgan), II, 145. 
 Space Isomerism of Toluquinoneoxime Ethers (Morgan), II, 283. Ethers of 
 Isonitrosoguaiacol and Space Isomerism of Nitrogen (Bridge and Morgan), 
 11,304. 
 
 MINERALOGICAL CHEMISTRY. 
 
 Rhodochrosite from Franklin Furnace (Browning), I, 57. So-called Perofskite 
 from Magnet Cove (Mar), I, 60. 
 
 ANALYTICAL CHEMISTRY. 
 QUALITATIVE ANALYSIS. 
 
 Detection of Iodine, Bromine and Chlorine (Gooch and Brooks), I, 47. Detec- 
 tion of Strontium and Calcium (Browning), 1, 121. Detection of Arsenic 
 with Antimony and Tin (Gooch and Hodge), I, 231. Detection of Per- 
 chlorates (Gooch and Kreider), I, 246. Reduction of Arsenic Acid (Gooch 
 and Phelps), I, 265. Separation and Identification of Potassium and Sodium 
 (Kreider and Breckenridge), I, 401. Detection of Sulphides, Sulphates, 
 Sulphites, Thiosulphates (Browning and Howe), II, 134. Separation of 
 Nickel from Cobalt (Browning and Hart well), II, 344. 
 
396 SYSTEMATIC INDEX. 
 
 QUANTITATIVE ANALYSIS. 
 
 Colorimetric Methods. 
 
 Detection and Approximative Estimation of Minute Amounts of Arsenic in 
 Copper (Gooch and Moseley), I, 272. 
 
 Electrolytic Methods. 
 
 Determination of Halogens in mixed Silver Salts (Gooch and Fairbanks), I, 290. 
 Spectroscopic Methods. 
 
 Determination of Potassium (Gooch and Hart), I, 92. Determination of Ru- 
 bidium (Gooch and Phinney), I, 157. 
 
 Gravimetric Methods. 
 
 Determination of Chlorine in Alkaline Chlorides and Iodides (Gooch and Mar), 
 
 I, 18. Determination of Bromine in Alkaline Bromides and Iodides (Gooch 
 and Ensign), I, 37. Estimation of Barium as Sulphate (Mar), I, 63. Sepa- 
 ration of Strontium from Calcium (Browning), I, 107. Separation of Barium 
 from Calcium (Browning), I, 116. Determination of Barium in presence of 
 Calcium and Magnesium (Mar), I, 125. Separation of Barium from Stron- 
 tium (Browning), I, 168. Influence of Nitric Acid and Aqua Regia on the 
 Precipitation of Barium as Sulphate (Browning), I, 181. Treatment of 
 Barium Sulphate (Phiuney), I, 187. Separation of Copper from Cadmium 
 (Browning), I, 226. Determination of Potassium (Kreider), I, 282. De- 
 termination of Carbon Dioxide (Gooch and Phelps), I, 302. Determination 
 of Selenium (Peirce),!, 365. Estimation of Cadmium as Oxide (Browning 
 and Jones), 1, 409. Separation of Aluminum from Iron (Gooch and Havens), 
 
 II, 20. Separation of Aluminum and Beryllium (Havens), II, 47. Esti- 
 mation of Manganese as Sulphate and as Oxides (Gooch and Austin), II, 
 77. Estimation of Manganese separated as Carbonate (Austin), II, 96. 
 Separations of Aluminum by Hydrochloric Acid (Havens), II, 106. De- 
 termination of Manganese as Pyrophosphate (Gooch and Austin), II, 121. 
 Separation of Nickel and Cobalt by Hydrochloric Acid (Havens), II, 141. 
 Estimation of Boric Acid (Gooch and Jones), II, 172. Ammonium Magne- 
 sium Phosphate of Analysis (Gooch and Austin), II, 190. Volatilization 
 of Iron Chlorides and Separation of Oxides of Iron and Aluminum (Gooch 
 and Havens), II, 215. Double Ammonium Phosphates of Beryllium, 
 Zinc, Cadmium (Austin), II, 252. Separation of Iron from Chromium, 
 Zirconium, Beryllium, by Gaseous Hydrochloric Acid (Havens and Way), II, 
 266. Ammonium Magnesium Arseniate of Analysis (Austin), II, 209. 
 Estimation of Thallium as Acid and Neutral Sulphates (Browning), II, 
 317. Separation and Determination of Mercury as Oxalate (Peters), II, 
 325. Sulphocyanides of Copper and Silver (Van Name), II, 359. Estima- 
 tion of Caesium and Rubidium as the Acid Sulphates and of Sodium and 
 Potassium as the Pyrosulphates (Browning), II, 368. 
 
 Volumetric Methods. 
 
 Standard Solutions. Tartar Emetic (Gruener), I, 216. Potassium Permanganate 
 (Roberts), I, 269. 
 
 lodometric Processes. Determination of Iodine in Haloid Salts (Gooch and Brown- 
 ing), 1, 1. Reduction of Arsenic Acid (Gooch and Browning), I, 30. De- 
 
SYSTEMATIC INDEX. 397 
 
 termination of Antimony and its Condition of Oxidation (Gooch and 
 Gruener), 1, 73. Estimation of Chlorates (Gooch and Smith), I, 82. Sepa- 
 ration of Antimony from Arsenic (Gooch and Banner), I. 36. Determina- 
 tion of Nitrates (Gooch and Gruener), I, 132. Determination of Iodine in 
 Haloid Salts by action of Arsenic Acid (Gooch and Browning), I, 173. 
 Determination of Nitrates (Gruener), I, 193. Estimation of Chlorates and 
 Nitrates, and of Nitrites and Nitrates (Roberts), 1, 219. Estimation of Tellu- 
 ric Acid (Gooch and Rowland), I, 277. Reduction of Acids of Selenium by 
 Hydriodic Acid (Gooch and Reynolds), I, 310. Determination of Per- 
 chlorates (Kreider), I, 316. Reduction of Selenic Acid (Gooch and Evans), 
 
 I, 331. Reduction of Selenic Acid (Gooch and Scoville), I, 335. Deter- 
 mination of Selenious and Selenic Acids (Gooch and Peirce), I, 338. Inter- 
 action of Chromic and Arsenious Acids (Browning), I, 344. Separation of 
 Selenium from Tellurium (Gooch and Peirce), 1, 348. Determination of Car- 
 bon Dioxide (Phelps), I, 369. Estimation of Molybdic Acid (Gooch and Fair- 
 banks), I, 375. Determination of Phosphorus in Iron (Fairbanks), I, 391. 
 Reduction of Vanadic Acid (Browning), I, 397. Estimation of Vanadium 
 (Browning and Goodman), II, 4. Determination of Oxygen in Air and 
 Aqueous Solution (Kreider), II, 11. Estimation of Molybdenum (Gooch), 
 
 II, 27. Application of lodic Acid to Analysis of Iodides (Gooch and 
 Walker), II, 33. Titration of Sodium Thiosnlphate with lodic Acid (Walker), 
 II, 52. Determination of Molybdenum (Gooch and Norton), II, 111. 
 Analysis of Alkalies and Acids (Walker and Gillespie), II, 162. Influence 
 of Hydrochloric Acid in Titrations by Thiosulphate, and Estimation of Se- 
 lenious Acid (Norton), II, 206. Estimation of Iron in the Ferric State 
 (Norton), II, 230. Determination of Tellurous Acid in presence of Haloid 
 Salts (Gooch and Peters), II, 238. Estimation of Boric Acid (Jones), II, 
 244. Determination of Gold (Gooch and Morley), II, 269. Estimation of 
 Cerium (Browning), II, 289. Estimation of Thallium (Browning and 
 Hutchins), II, 300. Titration of Mercury by Sodium Thiosulphate (Norton), 
 II, 328. Estimation of Arsenic Acid (Gooch and Morris), II, 336. 
 
 Alkalimetric Processes. Estimation of Boric Acid (Jones), II, 182. 
 
 Oxidimetric Processes. Determination of Selenious Acid (Gooch and demons), 
 I, 297. Titration of Oxalic Acid in presence of Hydrochloric Acid (Gooch 
 and Peters), II, 222. Determination of Tellurous Acid in presence of Haloid 
 Salts (Gooch and Peters), II, 238. Separation and Determination of Mercury 
 as Mercurous Oxalate (Peters), II, 320. Estimation of Copper as Oxalate, 
 with Separations (Peters), II, 347. Estimation of Calcium, Strontium, and 
 Barium as the Oxalates (Peters), II, 373. 
 
 Precipitation Processes. Determination of Tellurium by Precipitation as the 
 Iodide (Gooch and Morgan), II, 1. 
 
 Gasometric Processes. Reduction of Nitric Acid by Ferrous Salts (Roberts), 
 I, 203. Estimation of Chlorates and Nitrates, and of Nitrites and Nitrates 
 (Roberts), I, 219. 
 
INDEX OF AUTHORS. 
 
 AUSTIN, MARTHA. Estimation of Manganese as Sulphate and Oxide 
 
 (with Gooch, F. A.) H, 77 
 
 Condition of Oxidation of Manganese precipitated by Chlorate 
 
 Process (with Gooch, F. A.) II, 85 
 
 Estimation of Manganese Separated as Carbonate II, 96 
 
 Determination of Manganese as Pyrophosphate (with Gooch, 
 
 F. A.) II, 121 
 
 Constitution of Ammonium Magnesium Phosphate of Analysis (with 
 
 Gooch, F. A.) II, 190 
 
 Double Ammonium Phosphates of Beryllium, Zinc, Cadmium, in 
 
 Analysis II, 252 
 
 Constitution of Ammonium Magnesium Arseniate of Analysis . . II, 309 
 BALDWIN, DEFOREST. Action of Acetylene on Oxides of Copper (with 
 
 Gooch, F. A.) II, 276 
 
 BRECKENRIDGE, J. E. Separation and Identification of Potassium and 
 
 Sodium (with Kreider, D. Albert) I, 401 
 
 BRIDGE, JOHN L. Ethers of Toluquinoneoxime, and Space Isomerism 
 
 of Nitrogen (with Morgan, Wm. Conger) II, 145 
 
 Ethers of Isonitrosoguaiacol and Space Isomerism of Nitrogen (with 
 
 Morgan, Wm. Conger) II, 304 
 
 BROOKS, F. T. Detection of Iodine, Bromine, and Chlorine (with 
 
 Gooch, F. A.) I, 47 
 
 BROWNING, PHILIP E. Determination of Iodine in Haloid Salts (with 
 
 Gooch, F. A.) I, 1 
 
 Reduction of Arsenic Acid in Analysis (with Gooch, F. A.) ... I, 30 
 
 Analysis of Rhodochrosite from Franklin Furnace I, 57 
 
 Quantitative Separation of Strontium from Calcium by Amyl 
 
 Alcohol on Nitrates I, 107 
 
 Quantitative Separation of Barium from Calcium by Amyl Alcohol 
 
 on Nitrates I, 116 
 
 Separation and Detection of Strontium and Calcium by Amyl 
 
 Alcohol on Nitrates . . 1,121 
 
 Quantitative Separation of Barium from Strontium by Amyl Alcohol 
 
 on Bromides I, 168 
 
 Determination of Iodine in Haloid Salts by Action of Arsenic Acid 
 
 (with Gooch, F. A.) I, 173 
 
 Influence of Nitric Acid and Aqua Regia on Precipitation of Barium 
 
 as Sulphate I, 181 
 
 Separation of Copper from Cadmium by Iodide Method .... I, 226 
 
 Interaction of Chromic and Arsenious Acids I, 344 
 
 Reduction of Vanadic Acid by Hydrobromic and Hydriodic Acids, 
 
 and Estimation by Iodine I, 397 
 
400 INDEX OF AUTHORS. 
 
 VOL. PAGE 
 
 Estimation of Cadmium as Oxide (with Jones, Louis C.) . . . . I, 409 
 Application of Organic Acids to Estimation of Vanadium (with 
 
 Goodman, Richard J.) II, 4 
 
 Detection of Sulphides, Sulphates, Sulphites, and Thiosulphates 
 
 (with Howe, Ernest) II, 134 
 
 Volumetric Estimation of Cerium (with Hanford, G. A. ; Hall, F. J. ; 
 
 Cutter, Wm. D. ; Lynch, Leo A.) II, 289 
 
 Estimation of Thallium as Chromate (with Hutchius, George P.) . II, 300 
 
 Estimation of Thallium as Acid and Neutral Sulphates 11,317 
 
 Qualitative Separation of Nickel from Cobalt by Ammonia on the 
 
 Ferricyanides (with Hartwell, John B.) II, 344 
 
 On the Estimation of Caesium and Rubidium as the Acid Sul- 
 phates, and of Potassium and Sodium as the Pyrosulphates . . II, 368 
 CLEMONS, C. F. Determination of Selenious Acid by Potassium Perman- 
 ganate (with Gooch, F. A.) I, 297 
 
 CUTTER, WM. D. Volumetric Estimation of Cerium (with Browning, 
 
 Philip E.) 11,294 
 
 DANNER, E. VV. Separation of Antimony from Arsenic by Hydrochloric 
 
 and Hydriodic Acids (with Gooch, F. A.) I, 86 
 
 Interaction of Potassium Permanganate and Sulphuric Acid (with 
 
 Gooch, F. A.) I, 145 
 
 DONLAP, FREDERICK L. Action of Urea and Sulphocarbanilide on Acid 
 
 Anhydrides 1, 355 
 
 Action of Urea and Primary Amines on Maleic Anhydride (with 
 
 Phelps, Isaac K.) 11,42 
 
 ENSIGN, J. R. Determination of Bromine in Alkaline Bromides and 
 
 Iodides (with Gooch, F. A.) I, 37 
 
 EVANS, P. S., JR. Reduction of Selenic Acid by Hydrochloric Acid 
 
 (with Gooch, F. A.) 1,331 
 
 FAIRBANKS, CHARLOTTE. Estimation of Halogens in Silver Salts (with 
 
 Gooch, F. A.) . . 1,290 
 
 lodometric Estimation of Molybdic Acid (with Gooch, F. A.) . . I, 375 
 
 lodometric Determination of Phosphorus in Iron I, 391 
 
 GILLESPIE, DAVID H. M. Iodine in Analysis of Acids and Alkalies 
 
 (with Walker, Claude F.) II, 162 
 
 GOOCH, F. A. Determination of Iodine in Haloid Salts (with Browning, 
 
 P.E.) I, 1 
 
 Determination of Chlorine in Alkaline Chlorides and Iodides (with 
 
 Mar, F. W.) I, 18 
 
 Reduction of Arsenic Acid in Analysis (with Browning, P. E.) . . I, 30 
 Determination of Bromine in Alkaline Bromides and Iodides (with 
 
 Ensign, J. R.) I, 37 
 
 Detection of Iodine, Bromine, and Chlorine (with Brooks, F. T.) . I, 47 
 Determination of Antimony and its Condition of Oxidation (with 
 
 Gruener, H. W.) I, 73 
 
 Estimation of Chlorates (with Smith, C. G.) I, 82 
 
 Separation of Antimony from Arsenic by Hydrochloric and Hydri- 
 odic Acids (with Danner, E. W.) I, 86 
 
 Detection and Determination of Potassium Spectroscopically (with 
 
 Hart, T. S.) I, 92 
 
 lodometric Determination of Nitrates (with Gruener, H. W.) . . I, 132 
 Laboratory Apparatus I, 141 
 
INDEX OF AUTHORS. 401 
 
 VOL. PAGE 
 
 Interaction of Potassium Permanganate and Sulphuric Acid (with 
 
 Banner, E. W.) I, 145 
 
 Quantitative Determination of Rubidium by the Spectroscope (with 
 
 Phinney, J. I.) I, 157 
 
 Determination of Iodine in Haloid Salts by Action of Arsenic Acid 
 
 ( with Browning, P. E.) 1,173 
 
 Detection and Separation of Arsenic with Antimony and Tin (with 
 
 Hodge, B.) 1,231 
 
 Detection of Alkaline Perchlorates with Chlorides, Chlorates, and 
 
 Nitrates (with Kreider, D. Albert) I, 246 
 
 Generation of Chlorine (with Kreider, D. Albert) I, 260 
 
 Reduction of Arsenic Acid by Hydrochloric Acid and Potassium 
 
 Bromide (with Phelps, I. K.) I, 265 
 
 Detection and Estimation of Minute Amounts of Arsenic in Copper 
 
 (with Moseley, H. P.) I, 272 
 
 lodometric Estimation of Telluric Acid (with Rowland, J.) ... I, 277 
 Estimation of Halogens in Silver Salts (with Fairbanks, Charlotte) I, 290 
 Determination of Selenious Acid by Potassium Permanganate (with 
 
 demons, C. F.) I, 297 
 
 Precipitation and Gravimetric Determination of Carbon Dioxide 
 
 (with Phelps, I. K.) I, 302 
 
 Reduction of Acids of Selenium by Hydriodic Acid (with Reynolds, 
 
 W.G.) 1,310 
 
 Reduction of Selenic Acid by Hydrochloric Acid (with Evans, P. S., 
 
 Jr.) 1,331 
 
 Reduction of Selenic Acid by Potassium Bromide in Acid Solution 
 
 (witb Scoville, W. S.) I, 335 
 
 lodometric Determination of Selenious and Selenic Acids (with 
 
 Peirce, A. W.) I, 338 
 
 Separation of Selenium from Tellurium by difference in Volatility 
 
 of Bromides (with Peirce, A. W.) I, 348 
 
 lodometric Estimation of Molybdic Acid (with Fairbanks, Char- 
 lotte) 1,375 
 
 Determination of Tellurium by precipitation as Iodide (with Mor- 
 gan, W. C.) II, 1 
 
 Separation of Aluminum from Iron (with Havens, F. S.) . . . . II, 20 
 
 Estimation of Molybdenum lodometrically II, 27 
 
 Application of lodic Acid to Analysis of Iodides (with Walker, 
 
 C. F.) II, 33 
 
 Estimation of Manganese as Sulphate and Oxide (with Martha 
 
 Austin) II, 77 
 
 Condition of Oxidation of Manganese precipitated by Chlorate 
 
 Process (with Austin, Martha) II, 85 
 
 lodometric Determination of Molybdenum (with Norton, John T., 
 
 Jr.) 11,111 
 
 Determination of Manganese as Pyrophosphate (with Austin, 
 
 Martha) II, 121 
 
 Estimation of Boric Acid (with Jones, Louis Cleveland) . . . . II, 172 
 Constitution of Ammonium Magnesium Phosphate of Analysis 
 
 (with Austin, Martha) II, 190 
 
 Volatilization of Iron Chlorides, and Separation of Oxides of Iron 
 
 and Aluminum (with Havens, Franke Stuart) II, 215 
 
 VOL. n. 26 
 
402 INDEX OF AUTHORS. 
 
 VOL. PAGE 
 
 Titration of Oxalic Acid by Potassium Permanganate in presence 
 
 of Hydrochloric Acid (with Peters, C. A.) II, 222 
 
 Determination of Tellurous Acid in presence of Haloid Salts (with 
 
 Peters, C. A.) II, 238 
 
 lodometric Determination of Gold (with Morley, Frederick H.) . II, 269 
 Action of Acetylene on Oxides of Copper (with Baldwin, De Forest) II, 276 
 lodometric Estimation of Arsenic Acid (with Morris, Julia C.) . . II, 236 
 GOODMAN, RICHARD J. Application of Organic Acids to Estimation of 
 
 Vanadium (with Browning, Philip E.) II, 4 
 
 GRUBNER, H. W. Determination of Antimony and its condition of Oxi- 
 dation (with Gooch, F. A.) I, 73 
 
 lodometric Determination of Nitrates (with Gooch, F. A.) ... I, 132 
 
 lodometric Determination of Nitrates I, 193 
 
 Stability of Standard Solutions of Tartar Emetic I, 216 
 
 HALL, F. J. Volumetric Estimation of Cerium (with Browning, 
 
 Philip E.) II, 290 
 
 HANFORD, G. A. Volumetric Estimation of Cerium (with Browning, 
 
 Philip E.) 11,290 
 
 HART, T. S. Detection and Determination of Potassium Spectroscopi- 
 
 cally (with Gooch, F. A.) I, 92 
 
 HARTWELL, JOHN B. Qualitative Separation of Nickel from Cobalt by 
 
 Ammonia on the Ferricyanides (with Browning, Philip E.) . . II, 344 
 HAVENS, FRANKE STUART. Separation of Aluminum from Iron (with 
 
 Gooch, F. A.) II, 20 
 
 Separation of Aluminum and Beryllium by Hydrochloric Acid . . II, 47 
 Further Separations of Aluminum by Hydrochloric Acid . . . . II, 106 
 
 Separation of Nickel and Cobalt by Hydrochloric Acid II, 141 
 
 Volatilization of Iron Chlorides, and Separation of Oxides of Iron 
 
 and Aluminum (with Gooch, F. A.) II, 215 
 
 Separation of Iron from Chromium, Zirconium, and Beryllium by 
 Action of Gaseous Hydrochloric Acid on the Oxides (with Way, 
 
 Arthur Fitch) II, 266 
 
 HODGE, B. Detection and Separation of Arsenic with Antimony and 
 
 Tin (with Gooch, F. A.) I, 231 
 
 HOWE, ERNEST. Detection of Sulphides, Sulphates, Sulphites, and 
 
 Thiosulphates (with Browning, Philip E.) II, 134 
 
 HOWLAND, J. lodometric Estimation of Telluric Acid (with Gooch, 
 
 F. A.) I, 277 
 
 HUTCHINS, GEORGE P. Estimation of Thallium as Chromate (with 
 
 Browning, Philip E.) II, 300 
 
 JONES, Louis CLEVELAND. Estimation of Cadmium as Oxide (with 
 
 Browning, Philip E.) I, 409 
 
 Action of Carbon Dioxide on Soluble Borates II, 100 
 
 Estimation of Boric Acid (with Gooch, F. A.) II, 172 
 
 Volumetric Estimation of Boric Acid 11,182 
 
 lodometric Estimation of Boric Acid . II, 244 
 
 KREIDER, D. ALBERT. Detection of Alkaline Perchlorates with Chlo- 
 rides, Chlorates, and Nitrates (with Gooch, F. A.) I, 246 
 
 Generation of Chlorine (with Gooch, F. A.) I, 260 
 
 Preparation of Perchloric Acid and Determination of Potassium . I, 282 
 
 Laboratory Apparatus I, 306 
 
 Quantitative Determination of Perchlorates I, 316 
 
INDEX OF AUTHORS. 403 
 
 VOL. PAGE 
 
 Separation and Identification of Potassium and Sodium (with 
 
 Breckenridge, J. E.) . . I, 401 
 
 Determination of Oxygen in Air and Aqueous Solution .... II, 1 1 
 LYNCH, LEO A. Volumetric Estimation of Cerium (with Browning, 
 
 Philip E.) II, 297 
 
 MAR, F. W. Determination of Chlorine in Alkaline Chlorides and 
 
 Iodides (with Gooch, F. A.) I, 18 
 
 So-called Perofskite from Magnet Cove I, 60 
 
 Estimation of Barium as the Sulphate I, 63 
 
 Determination of Barium in presence of Calcium and Magne- 
 sium I, 125 
 
 MORGAN, WM. CONGER. Determination of Tellurium hy precipitation 
 
 as Iodide (with Gooch, F. A.) II, 1 
 
 Ethers of Toluquinoneoxime and Space Isomerism of Nitrogen 
 
 (with Bridge, John L.) II, 145 
 
 Space Isomerisms of Toluquinoneoxime Ethers II, 283 
 
 Ethers of Isonitrosoguaiacol and Space Isomerism of Nitrogen 
 
 (with Bridge, John L.) II, 304 
 
 MORLEY, FREDERICK H. lodometric Determination of Gold (with 
 
 Gooch, F. A.) H, 269 
 
 MORRIS, JULIA C. lodometric Estimation of Arsenic Acid (with 
 
 Gooch, F. A.) II, 336 
 
 MOSELEY, H. P. Detection and Estimation of Minute Amounts of 
 
 Arsenic in Copper (with Gooch, F. A.) I, 272 
 
 NORTON, JOHN T., Jr. lodometric Determination of Molybdenum 
 
 (with Gooch, F. A.) II, 111 
 
 Hydrochloric Acid in Titrations by Sodium Thiosulphate, and Esti- 
 mation of Selenious Acid II, 206 
 
 Estimation of Iron in Ferric Condition by Sodium Thiosulphate 
 
 and Iodine II, 230 
 
 Titration of Mercury by Sodium Thiosulphate II, 328 
 
 The Action of Sodium Thiosulphate on Solutions of Metallic Salts 
 
 at High Temperatures and Pressures II, 384 
 
 PEIRCE, A. W. lodometric Determination of Selenious and Selenic 
 
 Acids (with Gooch, F. A.) I, 338 
 
 Separation of Selenium from Tellurium by difference in Volatility 
 
 of Bromides (with Gooch, F. A.) I, 348 
 
 Gravimetric Determination of Selenium I, 365 
 
 Existence of Selenium Monoxide I, 385 
 
 PETERS, CHARLES A. Titration of Oxalic Acid by Potassium Permangan- 
 ate in presence of Hydrochloric Acid (with Gooch, F. A.) ... II, 222 
 Determination of Tellurous Acid in presence of Haloid Salts (with 
 
 Gooch, F. A.) 11,238 
 
 Determination of Mercury as Mercurous Oxalate H, 320 
 
 Volumetric Estimation of Copper with Separation from Cadmium, 
 
 Arsenic, Tin, Iron, and Zinc II, 347 
 
 The Estimation of Calcium, Strontium, and Barium as the Oxalates. II, 373 
 PHELPS, ISAAC K. Reduction of Arsenic Acid by Hydrochloric Acid 
 
 Potassium Bromide (with Gooch, F. A.) I, 265 
 
 Precipitation and Gravimetric Determination of Carbon Dioxide 
 
 (with Gooch, F. A.) I, 302 
 
 lodometric Determination of Carbon Dioxide I, 369 
 
404 INDEX OF AUTHORS. 
 
 VOL. PAGE 
 
 Action of Urea and Primary Amines on Maleic Anhydride (with 
 
 Duulap, Frederick L.) II, 42 
 
 Combustion of Organic Substances in the Wet Way II, 62 
 
 PHINNEY, J. L Quantitative Determination of Rubidium by the Spec- 
 troscope (with Gooch, F. A.) I, 157 
 
 Treatment of Barium Sulphate in Analysis I, 187 
 
 REYNOLDS, W. G. Reduction of Acids of Selenium by Hydriodic Acid 
 
 ( with Gooch, F. A.) 1,310 
 
 ROBERTS, CHARLOTTE F. Reduction of Nitric Acid by Ferrous Salts I, 203 
 Estimation of Chlorates and Nitrates, and of Nitrites and Nitrates . I, 219 
 
 Blue Iodide of Starch I, 236 
 
 Action of Reducing Agents on lodic Acid I, 250 
 
 Standardization of Potassium Permanganate in Iron Analysis . . I, 269 
 SCOVILLE, W. S. Reduction of Selenic Acid by Potassium Bromide in 
 
 Acid Solution (with Gooch, F. A.) I, 335 
 
 SMITH, C. G. Estimation of Chlorates (with Gooch, F. A.) .... I, 82 
 VAN NAME, R. G. The Sulphocyanides of Copper and Silver in Gravi- 
 metric Analysis II, 359 
 
 WALKER, CLAUDE. Application of lodic Acid to Analysis of Iodides 
 
 (with Gooch, F. A.) II, 33 
 
 Titration of Sodium Thiosulphate with lodic Acid II, 52 
 
 Iodine in Analysis of Acids and Alkalies (with Gillespie, David 
 
 H. M.) II, 162 
 
 WAY, ARTHUR FITCH. Separation of Iron from Chromium, Zir- 
 conium, and Beryllium, by action of Gaseous Hydrochloric Acid 
 on the Oxides (with Havens, Franke Stuart) II, 266 
 
INDEX OF SUBJECTS. 
 
 VOL. PAGE 
 
 Acetylene, action of, on oxides of copper (Gooch and Baldwin) . . . 11,276 
 
 Acids, application of iodine in analysis of (Walker and Gillespie) . . II, 162 
 
 of selenium, reduction of, by hydriodic acid (Gooch and Reynolds) . I, 310 
 
 Acid anhydrides, action of urea and sulphocarbanilide upon (Dunlap) . I, 355 
 
 Alkalies, application of iodine to analysis of (Walker and Gillespie) . II, 162 
 
 Alkaline bromides, determination of bromine in (Gooch and Ensign) . I, 37 
 Alkaline chlorides, determination of chlorine in alkaline iodides mixed 
 
 with (Gooch and Mar) I, 18 
 
 Alkaline iodides, determination of chlorine in (Gooch and Mar) . . I, 18 
 
 determination of bromine in (Gooch and Ensign) ...... I, 37 
 
 Alkaline perchlorates, detection of, associated with chlorides, chlorates, 
 
 and nitrates (Gooch and Kreider) 1,246 
 
 Aluminum salts, action of sodium thiosulphate upon, at high tempera- 
 tures and pressures (Norton) II, 388 
 
 Aluminum, separation of, by hydrochloric acid, from iron (Gooch and 
 
 Havens) II, 20 
 
 separation of, by hydrochloric acid, from beryllium (Havens) . . II, 47 
 
 separation of, by hydrochloric acid, from bismuth, copper, and . . . 
 
 mercury (Havens) II, 109 
 
 separation of, by hydrochloric acid, from zinc (Havens) .... II, 107 
 
 Aluminum oxide, separation of oxides of iron from (Gooch and Havens) II, 215 
 
 Ammonium magnesium arseniate in analysis, constitution of (Austin) . II, 309 
 
 Ammonium beryllium phosphate in analysis (Austin) II, 253 
 
 Ammonium cadmium phosphate in analysis (Austin) II, 262 
 
 Ammonium magnesium phosphate in analysis (Gooch and Austin) . . 11,190 
 
 Ammonium zinc phosphate in analysis (Austin) II, 257 
 
 Amyl alcohol, use of, in detecting strontium and calcium (Browning) . I, 121 
 
 use of, in separating strontium and calcium (Browning) 
 
 use of, in separating barium and calcium (Browning) 
 
 use of, in separating barium and strontium (Browning) .... 
 
 Antimony, detection of arsenic associated with (Gooch and Hodge) . . 
 
 determination of, and its condition of oxidation (Gooch and Gruener) 
 
 separation of, from arsenic, by hydrochloric and hydriodic acids 
 
 (Gooch and Banner) 
 
 Antimonious chloride, decomposition of nitrates by (Gruener) .... 
 Antimouic acid, salts of, reduced by potassium iodide and sulphuric acid, 
 
 and estimated iodometrically (Gooch and Gruener) 
 
 Apparatus burette clip (Gooch) 
 
 121 
 
 116 
 
 ^168 
 
 231 
 
 73 
 
 86 
 199 
 
 73 
 141 
 
 ,264 
 ,308 
 
 chlorine generator (Gooch and Kreider) 
 
 force pump (Kreider) 
 
 hot filter (Kreider) I, 306 
 
 mercury washer (Gooch) 1,143 
 
 steam evaporator (Gooch) I, 142 
 
 support (Gooch) I, 142 
 
406 INDEX OF SUBJECTS. 
 
 VOL. PAGE 
 
 valve (Kreider) 1, 307 
 
 used in analysis of iodides by iodic acid (Gooch and Walker) ... II, 37 
 
 used in combustion of organic substances in the wet way (Phelps . II, 68 
 used in estimation of carbon dioxide gravimetrically (Gooch and 
 
 Phelps) 1,302 
 
 used in estimation of iodine in haloid salts (Gooch and Browning) . I, 12 
 
 used in estimation of molybdenum (Gooch and Norton) . . . . 11,114 
 used in estimation of molybdic acid (Gooch and Fairbanks) . I, 378, 382 
 
 used in estimation of molybdic acid (Fairbanks) I, 394 
 
 used in estimation of nitrates (Gooch and Gruener) I, 137 
 
 used in estimation of oxygen in air and aqueous solution (Kreider) II, 17 
 used in estimation of selenium iodometrically, by volatilization of 
 
 the bromide (Gooch and Peirce) I, 350 
 
 used in reduction of arsenic acid (Gooch and Browning) .... I, 33 
 
 used in reduction of nitric acid by ferrous salts (Roberts) . . . 1,208 
 Arsenic, detection of, associated with antimony and tin (Gooch and 
 
 Hodge) 1,231 
 
 detection and approximative estimation of, in copper (Gooch and 
 
 Moseley) 1, 272 
 
 separation of antimony from, by hydrochloric and hydriodic acids 
 
 (Gooch and Danner) I, 86 
 
 separation of copper as oxalate from (Peters) II, 347 
 
 Arsenious acid, action of, upon cerium dioxide (Browning and Cutter) II, 294 
 
 interaction of, with chromic acid (Browning) I, 344 
 
 Arsenic acid, determination of, by reduction with potassium iodide and 
 sulphuric acid, and titration by iodine in alkaline solution (Gooch 
 
 and Browning) I, 30 
 
 iodometric estimation of (Gooch and Morris) . II, 336 
 
 reduction of, in analysis (Gooch and Browning) II, 30 
 
 reduction of, by action of hydrochloric acid and potassium bro- 
 mide (Gooch and Phelps) . 1, 265 
 
 use of, to liberate iodine in quantitative estimation of iodides (Gooch 
 
 and Browning) I, 1 
 
 use of, in determination of iodine in haloid salts (Gooch and Brown- 
 ing) I, 1 
 
 Aqua regia, influence of, on the precipitation of barium as the sul- 
 phate (Browning) I, 181 
 
 Barium, determination of, in presence of calcium and magnesium (Mar) I, 125 
 
 estimation of, as oxalate (Peters) II, 373 
 
 points in estimation of, as sulphate (Mar) .... ... I, 63 
 
 precipitation of, as sulphate, in presence of nitric acid and aqua 
 
 regia (Browning) I, 181 
 
 quantitative separation of, from calcium by amyl alcohol on the ni- 
 trates (Browning) I, 116 
 
 quantitative separation of, from strontium by amyl alcohol on the 
 
 bromides (Browning) I, 168 
 
 Barium chlorides, precipitation and separation of, from calcium and mag- 
 nesium, by hydrochloric acid and ether (Mar) I, 125 
 
 Barium sulphate, influence of hydrochloric acid upon precipitation of 
 
 (Mar) I, 63 
 
 purification of, by crystallizing from sulphuric acid (Mar) .... I, 71 
 
 treatment of, in analysis (Phinney) I, 187 
 
INDEX OF SUBJECTS. 407 
 
 VOL. PAGE 
 
 Beryllium, separation of, from aluminum, by action of hydrochloric acid 
 
 (Havens) II, 47 
 
 separation of iron from, by action of hydrochloric acid (Havens and 
 
 Way) 11,266 
 
 Beryllium ammonium phosphate in analysis (Austin) II, 253 
 
 Beryllium salt, action of sodium thiosulphate upon, at high tempera- 
 tures and pressures (Norton) n, 391 
 
 Bismuth, separation of aluminum from (Havens) II, 109 
 
 Blue iodide of starch (Eoberts) I, 236 
 
 Borates (soluble), action of carbon dioxide on (Jones) . II, 100 
 
 Boric acid, estimation of (Gooch and Jones) II, 172 
 
 iodometric method for estimation of (Jones) II } 244 
 
 use of calcium oxide as a retainer for (Gooch and Jones) .... 11,175 
 
 use of sodium tungstate as a retainer for (Gooch and Jones) . . . II, 178 
 
 volumetric estimation of (Jones) II, 182 
 
 Bromine, detection of, in presence of chlorine and iodine (Gooch and 
 
 Brooks) I, 47 
 
 , determination of, in alkaline bromides and iodides (Gooch and 
 
 Ensign) I, 37 
 
 volatilization of, from aqueous solutions of bromide and chloride 
 
 by action of sulphuric acid and nitrous acid (Gooch and Ensign) . I, 43 
 
 Cadmium, estimation of, as oxide (Browning and Jones) I, 409 
 
 Separation of copper from, by the iodide method (Browning) . . I, 226 
 
 Separation of copper from, as oxalate (Peters) II, 354 
 
 Cadmium ammonium phosphate, in analysis (Austin) II, 262 
 
 Caesium, estimation of, as the acid sulphate (Browning) II, 368 
 
 Calcium, determination of barium in presence of (Mar) I, 125 
 
 estimation of, as oxalate (Peters) II, 373 
 
 quantitative separation of barium from, by action of amyl alcohol on 
 
 the nitrates (Browning) ' I, 1 1 6 
 
 separation of, from strontium, and detection of, by action of amyl 
 
 alcohol on the nitrates (Browning) 1,121 
 
 Calcium oxide, use of, as a retainer for boric acid (Gooch and Jones) . II, 175 
 
 Carbon dioxide, action of on soluble borates (Jones) II, 100 
 
 iodometric method for determination of (Phelps) 1,369 
 
 precipitation and gravimetric determination of (Gooch and Phelps) I, 302 
 Cerium, modified Bunsen method for determination of (Browning, Han- 
 ford, and Hall) II, 290 
 
 volumetric estimation of (Browning) II, 289 
 
 Cerium dioxide, action of arsenious acid upon (Browning and Cutter) . II, 294 
 Cerium oxalate, estimation of, by potassium permanganate (Browning 
 
 and Lynch) 11,297 
 
 Chlorates, detection of -perchlorates associated with (Gooch and Kreider) I, 246 
 
 estimation of (Gooch and Smith) I, 82 
 
 Chlorates and nitrates, estimation of, in one operation (Roberts) ... 1,219 
 Chlorate process, condition of oxidation of manganese precipitated in 
 
 (Gooch and Austin) II, 85 
 
 Chlorides, detection of perchlorates associated with (Gooch and Kreider) I, 246 
 Chlorine, detection of, in presence of bromides and iodides (Gooch and 
 
 Brooks) I, 47 
 
 direct determination of, in alkaline chlorides and iodides (Gooch and 
 
 Mar) I, 18 
 
408 INDEX OF SUBJECTS. 
 
 generation of, by hydrochloric acid and potassium chlorate (Gooch 
 
 and Kreider) I, 260 
 
 Chromium salt, action of sodium thiosulphate upon, at high tempera- 
 tures and pressures (Norton) II, 389 
 
 Chromium, separation of iron from, by gaseous hydrochloric acid (Ha- 
 vens and Way) II, 266 
 
 Chromic acid, interaction of, with arsenious acid (Browning) .... I, 344 
 
 use of, in combustion of organic substances in the wet way (Phelps) II, 67 
 
 Cobalt, separation of, from nickel (Havens) II, 141 
 
 separation of nickel from, by action of ammonium hydroxide on the 
 
 ferricyanides (Browning and Hart well) II, 344 
 
 Combustion of organic substances in the wet way (Phelps) II, 62 
 
 Copper, detection and approximate estimation of minute amounts of 
 
 arsenic in (Gooch and Moseley) I, 272 
 
 estimation of, as oxalate, with separation from cadmium, arsenic, 
 
 tin, iron, and zinc (Peters) II, 347 
 
 preparation of, free from arsenic (Gooch and Moseley) I, 275 
 
 separation of aluminum from (Havens) II, 109 
 
 separation of, from cadmium by the iodide method (Browning) . . I, 226 
 
 Copper oxides, action of acetylene on (Gooch and Baldwin) 11,276 
 
 Copper sulphocyanide in gravimetric analysis (Van Name) 11,359 
 
 Dibrommaleinamide, preparation of, from urea and dibrommaleic anhy- 
 dride (Dunlap) I, 358 
 
 Dibrommale'inuric acid, preparation of, from urea and dibrommaleic an- 
 hydride (Dunlap) 1, 358 
 
 Dibromtoluquinonemetaoxime beuzoyl ether (Bridge and Morgan) . . II, 157 
 Dibromtoluquinoneorthooxime benzoyl ether (Bridge and Morgan) . . II, 161 
 Dibromtoluquinoneorthooxime methyl ether (Bridge and Morgan) . . II, 159 
 Dichlormaleinimide preparation of, from urea and dichlormale'ic anhy- 
 dride (Dunlap) 1, 357 
 
 Dichlormale'inuric acid, preparation of, by action of urea on dichlorma- 
 
 leic anhydride (Dunlap) I, 356 
 
 Double ammonium phosphates of beryllium, zinc, and cadmium in ana- 
 lysis (Austin) II, 252 
 
 Electrolytic iron, use of, in standardizing permanganate solutions (Rob- 
 erts) 1,269 
 
 Ethers of toluquinoneoxime, and their bearing on the space isomerism 
 
 of nitrogen (Bridge and Morgan) II, 145 
 
 Ferric alum, use of, with nitric acid, to liberate iodine from haloid salts 
 
 (Gooch and Mar) I, 23 
 
 Ferrous salts, use of, in reduction of nitric acid (Roberts) I, 203 
 
 Gold, iodometric determination of (Gooch and Morley) II, 269 
 
 Halogens, estimation of, in mixed silver salts (Gooch and Fairbanks) . I, 290 
 Haloid salts, determination of iodine in (Gooch and Browning) ... I, 2 
 determination of tellurous acid in presence of (Gooch and Peters) . II, 238 
 Hydriodic acid, action of, with hydrochloric acid in separation of anti- 
 mony from arsenic (Gooch and Danner) I, 86 
 
 use of, in reduction of acids of selenium (Gooch and Reynolds) . . I, 310 
 
 use of, in reduction of vanadic acid ( Browning) I, 397 
 
 Hydrobromic acid, use of, in reduction of vanadic acid (Browning) . . I, 397 
 Hydrochloric acid, action of, with hydriodic acid, in separation of anti- 
 mony from arsenic (Gooch and Danner) I, 86 
 
INDEX OF SUBJECTS. 409 
 
 VOL. PAGE 
 
 influence of, upon the precipitation of barium sulphate (Mar) ... II, 63 
 
 influence of, in titrations by sodium thiosulphate, with special refer- 
 ence to the estimation of selenious acid (Norton) II, 206 
 
 use of, in reduction of seleuic acid {Gooch and Evans) I, 331 
 
 use of, in separation of aluminum from iron (Gooch and Havens) . II, 20 
 
 use of, in separation of aluminum from zinc, copper, mercury, bis- 
 muth (Havens) II, 106 
 
 use of, with ether, to precipitate barium chloride in presence of salts 
 
 of magnesium and calcium (Mar) 1,125 
 
 use of, with potassium bromide, in reducing and volatilizing arsenic 
 
 acid (Gooch and Phelps) I, 265 
 
 use of, with potassium bromide, in separating arsenic from copper 
 
 (Gooch and Moseley) I, 272 
 
 use of, with potassium chlorate, to generate chlorine (Gooch and 
 
 Kreider) 1, 260 
 
 use of, with potassium iodide in volatilizing arsenic (Gooch and 
 
 Hodge) 1, 231 
 
 titration of oxalic acid by potassium permanganate in presence of 
 
 (Gooch and Peters) 11,222 
 
 volatility of, in aqueous solutions containing sulphuric acid and so- 
 dium chloride (Gooch and Mar) I, 19 
 
 Hydrochloric acid (gaseous) use of, in separation of iron from chromium, 
 
 zirconium, and beryllium (Havens and Way) II, 266 
 
 lodic acid, action of iodine on, in presence of hydrochloric acid (Roberts) I, 257 
 
 action of reducing agents on, in presence of hydrochloric acid (Rob- 
 
 erts) I, 252 
 
 application to the analysis of iodides (Gooch and Walker) .... II, 33 
 
 use of, in absorption of nitric oxide (Roberts) 1,250 
 
 use of, in titration of sodium thiosulphate (Walker) II, 52 
 
 Iodides, application of iodic acid to the analysis of (Gooch and Walker) . II, 33 
 
 Iodide method, use of, in separating copper from cadmium (Browning) . I, 226 
 
 Iodine, action of, on iodic acid in presence of hydrochloric acid (Roberts) I, 257 
 
 application of, in analysis of alkalies and acids (Walker and Gilles- 
 
 pie) 11,162 
 
 detection of, in presence of chlorine and bromine (Gooch and 
 
 Brooks) I, 47 
 
 .determination of, in haloid salts, by action of arsenic acid (Gooch 
 
 and Browning) I, 1 
 
 liberation of, from haloid salts, by arsenic acid (Gooch and Brown- 
 ing) I, 1 
 
 liberation of, from haloid salts, by ferric alum with nitric acid (Gooch 
 
 and Mar) I, 23 
 
 liberation of, from haloid salts, by nitrous acid (Gooch and Mar) . I, 27 
 
 use of, in estimating iron reduced from the ferric state by sodium 
 
 thiosulphate (Norton) II, 230 
 
 lodometric determination of gold (Gooch and Morley) II, 269 
 
 of molybdenum (Gooch and Norton) 11,111 
 
 of nitrates (Gooch and Gruener) I, 132 
 
 of nitrates (Gruener) 1,193 
 
 of selenious and selenic acids (Gooch and Peirce) I, 338 
 
 lodometric estimation of alkalies and acids (Walker and Gillespie) . . II, 162 
 
 of antimonic acid (Gooch and Gruener) I, 73 
 
410 INDEX OF SUBJECTS. 
 
 VOL. PAGE 
 
 of antimony separated from arsenic (Gooch and Danner) .... I, 86 
 
 of arsenic acid (Gooch and Browning) I, 30 
 
 of arsenic acid (Gooch and Morris) II, 336 
 
 of boric acid (Jones) n, 244 
 
 of carbon dioxide (Phelps) I ? 359 
 
 of chlorates (Gooch and Smith) I, 82 
 
 of chromic acid (Browning) I ? 344 
 
 of cerium (Browning, Hanford, and Hall) II, 290 
 
 of gold (Gooch and Morley) II, 269 
 
 of iodides (Gooch and Walker) II, 33 
 
 of iodine in haloid salts (Gooch and Browning) I, i 
 
 of iron (Norton) II, 230 
 
 of mercury (Norton) II, 328 
 
 of molybdenum (Gooch) II, 27 
 
 of molybdenum (Gooch and Norton) 11,111 
 
 of molybdic acid (Gooch and Fairbanks) I, 375 
 
 of nitrates ( Gooch and Gruener) 1,132 
 
 of nitrates (Gruener) I, 193 
 
 of oxygen, in air and in aqueous solution (Kreider) II, 11 
 
 of oxygen, in perchlorates (Kreider) I, 316 
 
 of phosphorus in iron (Fairbanks) I, 391 
 
 of selenious acid (Gooch and Reynolds) I, 310 
 
 of selenious acid (Gooch and Peirce) 1,338 
 
 of selenic acid (Gooch and Reynolds) 1,314 
 
 of selenic acid (Gooch and Peirce) I, 338 
 
 of selenium associated with tellurium (Gooch and Peirce) .... I, 348 
 
 of tellurous acid (Gooch and Peters) II, 238 
 
 of vanadic acid (Browning) 1,397 
 
 of vauadic acid (Browning and Goodman) II, 4 
 
 lodometric method for the determination of carbon dioxide (Phelps) . I, 369 
 
 for the determination of phosphorus in iron (Fairbanks) .... I, 391 
 
 for the estimation of boric acid (Jones) II, 244 
 
 Iron, estimation of, in the ferric state by reduction with sodium thiosul- 
 
 phate and titration with iodine (Norton) II, 230 
 
 iodometric method for determination of phosphorus in (Fairbanks) I, 391 
 method for the separation of aluminum from (Gooch and Havens) . II, 20 
 separation of, from chromium, zirconium, and beryllium, by gase- 
 ous hydrochloric acid ( Havens and Way) II, 266 
 
 separation of copper oxalate from (Peters) II, 347 
 
 Iron analysis, standardization of potassium permanganate in (Roberts) . I, 269 
 
 Iron oxides, separation of, from aluminum oxide (Gooch and Havens) . II, 215 
 
 Isonitrosoguaiacol, and salts of (Bridge and Morgan) II, 306 
 
 Isonitrosoguaiacol benzoyl ether (Bridge and Morgan) II, 307 
 
 Isonitrosoguaiacol benzoyl ether dibromide (Bridge and Morgan) . . II, 307 
 Isonitrosoguaiacol, ethers of, in their relation to the space isomerism of 
 
 nitrogen (Bridge and Morgan) II, 304 
 
 Isomerism (space) of nitrogen, in ethers of isonitrosoguaiacol (Bridge 
 
 and Morgan) II, 304 
 
 in ethers of toluquinoneoxime (Bridge and Morgan) 11,145 
 
 Isomerism (space) of the toluquinoneoxime ethers (Morgan) .... II, 283 
 
 Laboratory apparatus (Gooch) I, 141 
 
 Laboratory apparatus (Kreider) 1,306 
 
INDEX OF SUBJECTS. 411 
 
 VOL. PAGE 
 
 Magnesium, determination of, by precipitation as ammonium magne- 
 sium phosphate (Gooch and Austin) 11,190 
 
 determination of barium in presence of (Mar) I, 125 
 
 Maleic anhydride, action of primary amines upon (Dnnlap and Phelps) II, 44 
 
 action of urea and primary amines upon (Dunlap and Phelps) . . II, 42 
 Maleiiric acid, preparation of, from urea and maleic anhydride (Dunlap 
 
 andPhelps) II, 42 
 
 Manganese, determination of, as the pyrophosphate (Gooch and Austin) II, 121 
 condition of oxidation of, precipitated by the chlorate process 
 
 (Gooch and Austin) II, 85 
 
 estimation of, as the sulphate and as the oxides (Gooch and Austin) II, 77 
 
 estimation of, separated as the carbonate (Austin) II, 96 
 
 Manganous chloride, use of, in hydrochloric acid, in detection of oxidiz- 
 ing agents (Gooch and Gruener) I, 134 
 
 use of, in estimating nitrates (Gooch and Gruener) I, 132 
 
 Mercury, gravimetric estimation of, as the oxalate (Peters) II, 325 
 
 separation of aluminum from ( Havens) 11,109 
 
 titration of, by sodium thiosulphate (Norton) II, 328 
 
 Mercurous oxalate, separation and determination of (Peters) .... II, 320 
 Metallic salts, action of sodium thiosulphate upon, in solution at high 
 
 temperatures and pressures (Norton) II, 384 
 
 Molybdenum, estimation of iodometrically (Gooch) II, 27 
 
 iodometric determination of (Gooch and Norton) 11,111 
 
 Molybdic acid, iodometric estimation of (Gooch and Fairbanks) ... I, 375 
 
 Monobromisonitrosoguaiacol benzoyl ether (Morgan) II, 308 
 
 Monobromtoluquinouemetaoxime benzoyl ether ( Bridge and Morgan) . II, 1 58 
 
 Monobromtoluquinoneorthooxime benzoyl ether (Morgan) II, 286 
 
 Naphthylmaleamic acid j8, preparation of, from 0-naphthylamine and 
 
 maleic anhydride (Dunlap and Phelps) II, 45 
 
 Nickel, separation of, from cobalt (Havens) II, 141 
 
 separation from cobalt by action of ammonium hydroxide on the 
 
 ferricyanides (Browning and Hart well) II, 344 
 
 Nitrates, action of phosphoric acid and potassium iodide upon (Gruener) I, 193 
 
 decomposition of, by antimonious chloride (Gruener) I, 199 
 
 detection of perchlorates associated with (Gooch and Kreider) . . I, 246 
 
 iodometric determination of (Gooch and Gruener) I, 132 
 
 odometric determination of (Gruener) I, 193 
 
 Nitrates and chlorates, estimation of, in one operation (Roberts) ... I, 219 
 
 Nitrates and nitrites, estimation of, in one operation (Roberts) ... I, 222 
 Nitric acid, influence of, in precipitation of barium as the sulphate 
 
 (Browning) I, 181 
 
 reduction of, by ferrous salts (Roberts) I, 203 
 
 Nitric oxide, absorption of, by iodic acid (Roberts) I, 250 
 
 Nitrogen, space isomerism of, and bearing of ethers of toluquinoneoxime 
 
 (Bridge and Morgan), (Morgan) II, 145, 283 
 
 space isomerism of, in ethers of isonitrosoguaiacol (Bridge and 
 
 Morgan), (Morgan) II, 304 
 
 Nitrous acid, use of, in liberating iodine (Gooch and Mar) I, 27 
 
 use of, in liberating iodine (Gooch and Ensign) I, 43 
 
 Oxalic acid, titration of, by potassium permanganate in presence of hy- 
 drochloric acid (Gooch and Peters) II, 222 
 
 Oxygen, amount of, required to oxidize an organic substance (Phelps) . II, 71 
 
412 INDEX OF SUBJECTS. 
 
 VOL. PAGE 
 
 determination, In air and in aqueous solution (Kreider) II, 11 
 
 Organic acids, application of, in estimation of vanadium (Browning and 
 
 Goodman) II, 4 
 
 Organic substance, amount of oxygen required for oxidation of (Phelps) II, 71 
 
 combustion of, by chromic acid in the wet way (Phelps) .... II, 67 
 
 combustion of, by potassium permanganate in the wet way (Phelps) II, 62 
 
 Perchlorates, quantitative determination of (Kreider) 1,316 
 
 Perchloric acid, application of, to the determination of potassium 
 
 (Kreider) I, 286 
 
 preparation of (Kreider) I, 282 
 
 Permanganate solutions, standardization of, by electrolytic iron (Roberts) I, 269 
 
 Perofskite (so-called), analysis of, from Magnet Cove, Ark. (Mar) . . I, 60 
 Phosphorus, iodometric method for the determination of, in iron 
 
 (Fairbanks) 1,391 
 
 Phosphoric acid, determination of, by precipitation as ammonium mag- 
 nesium phosphate (Gooch and Austin) II, 204 
 
 Phosphoric acid, use of, with potassium iodide, in determining nitrates 
 
 (Gruener) I, 193 
 
 Phthalanil, preparation of, from phthalic anhydride and sulphocar- 
 
 banilide (Dunlap) I, 361 
 
 Phthalauilic acid, preparation of, from phthalic anhydride and sulpho- 
 
 carbanilide (Duiifap) 1, 361 
 
 Phthalimide, preparation of, by action of urea on phthalic anhydride 
 
 (Dunlap) 1,355 
 
 Primary Amines, action of, on maleic anhydride (Dunlap and Phelps) . II, 44 
 Potassium, detection and determination of, spectroscopically ( Gooch and 
 
 Hart) I, 92 
 
 determination of, by perchloric acid (Kreider) I, 282 
 
 estimation of, as the pyrosulphate (Browning) II, 368 
 
 separation of, from sodium (Kreider and Breckenridge) .... 1,401 
 Potassium Spectrum, brightening of, by sodium chloride (Gooch and 
 
 Hart) I, 101 
 
 Potassium bromide, use of, in reduction of selenic acid (Gooch and 
 
 Scoville) 1,335 
 
 use of, in reduction of arsenic acid (Gooch and Phelps) .... I, 265 
 use of, with hydrochloric acid, in separation of arsenic from copper 
 
 (Gooch and Moseley) I, 272 
 
 Potassium perchlorate, decomposition of, by anhydrous zinc chloride 
 
 (Gooch and Kreider) I, 247 
 
 Potassium permanganate, estimation of cerium oxalate by (Browning 
 
 and Lynch) II, 297 
 
 action of sulphuric acid upon (Gooch and Danner) I, 145 
 
 standardization of, in iron analysis (Roberts) I, 269 
 
 titration of oxalic acid by, in presence of hydrochloric acid (Gooch 
 
 and Peters) II, 222 
 
 use of, in combustion of organic substances in the wet way (Phelps) II, 62 
 use of, in estimation of copper, with separation from cadmium, 
 
 arsenic, tin, iron, and zinc (Peters) II, 347 
 
 use of, in the volumetric estimation of mercury as the oxalate 
 
 (Peters) 11,320 
 
 use of, in the estimation of selenious acid (Gooch and demons) . . I, 297 
 
 use of, in the estimation of tellurous acid (Gooch and Danner) . . I, 154 
 
INDEX OF SUBJECTS. 413 
 
 VOL. PAGE 
 
 use of, in the estimation of tellurous acid (Gooch and Peters) . . II, 238 
 
 Reducing agents, action of, on iodic acid (Roberts) I, 250 
 
 Rhodochrosite, analysis of, from Franklin Furnace, N. J. (Browning) . I, 57 
 
 Rubidium, estimation of, as the acid sulphate (Browning) II, 370 
 
 quantitative spectroscopic determination of (Gooch and Phinney) . I, 157 
 
 Selenic acid, iodometric determination of (Gooch and Peirce) .... I, 341 
 
 reduction of, by hydrochloric acid (Gooch and Evans) I, 331 
 
 reduction of, by potassium bromide in acid solution (Gooch and 
 
 Scoville) l t 335 
 
 Selenious acid, determination of, by potassium permanganate (Gooch 
 
 and demons) I ? 297 
 
 influence of hydrochloric acid in thiosulphate titrations of (Norton) II, 206 
 
 iodometric determination of (Gooch and Peirce) 1,338 
 
 Selenium, gravimetric determination of (Peirce) I, 365 
 
 method for separation of, from tellurium (Gooch 'and Peirce) ... I, 348 
 
 reduction of acids of, by hydriodic acid (Gooch and Reynolds) . . I, 310 
 
 Selenium monoxide, on the existence of (Peirce) 1,385 
 
 Silver salts, electrolytic reduction of, in estimation of halogens (Gooch 
 
 and Fairbanks) I, 290 
 
 Silver sulphocyanide in gravimetric analysis (Van Name) II, 359 
 
 Sodium, estimation of, as the pyrosulphate (Browning) 11,371 
 
 separation of, from potassium (Kreider and Breckenridge) ... I, 401 
 
 Sodium chloride, brightening of potassium spectrum by (Gooch and Hart) I, 101 
 Sodium thiosulphate, action of, on solutions of metallic salts at high 
 
 temperatures and pressures (Norton) II, 384 
 
 influence of hydrochloric acid in titrations by, with special reference 
 
 to the estimation of selenious acid (Norton) II, 206 
 
 reduction of iron in ferric state by (Norton) II, 230 
 
 use of, in titration of mercury (Norton) II, 328 
 
 titration of, by iodic acid (Walker) II, 52 
 
 Sodium tungstate, use of, as a retainer for boric acid (Gooch and Jones) II, 178 
 Space isomerism of nitrogen, bearing of ethers of toluquiuoneoxime on 
 
 (Bridge and Morgan) (Morgan) 11,145,283 
 
 Spectroscopic determination of potassium (quantitative) (Gooch and 
 
 Hart) I, 92 
 
 Spectroscopic determination of rubidium (quantitative) (Gooch and 
 
 Phinney) I, 157 
 
 Standard solutions of tartar emetic, stability of (Gruener) I, 216 
 
 Standardization of potassium permanganate in iron analysis (Roberts) . I, 269 
 
 Starch, blue iodide of (Roberts) I, 236 
 
 Starch blue, conditions governing formation and decomposition of 
 
 (Roberts) . 1, 236 
 
 Strontium, estimation of, as oxalate (Peters) 11,374 
 
 quantitative separation of, from calcium, by action of amyl alcohol 
 
 on the nitrates (Browning) I, 121 
 
 separation of barium from, by action of amyl alcohol on the brom- 
 ides (Browning) 1,168 
 
 separation of, from calcium, by action of amyl alcohol on the nitrate 
 
 (Browning) I, 121 
 
 Succinanil, preparation of, from succinic anhydride and sulphocarban- 
 
 ilide (Dunlap) 1,363 
 
 Succinic anhydride, action of sulphocarbanilide upon (Dunlap) ... I, 363 
 
414 INDEX OF SUBJECTS. 
 
 VOl. PAGB 
 
 Succinimide, preparation of, from urea and succinic anhydride (Dunlap) I, 359 
 Sulphates, detection of, in presence of sulphides, sulphites, and thiosul- 
 
 phates ( Browning and Howe) 11,134 
 
 Sulphides, detection of, in presence of sulphates, sulphites, and thiosul- 
 
 phates (Browning and Howe) II, 134 
 
 Sulphites, detection of, in presence of sulphides, sulphates, and thiosul- 
 
 phates (Browning and Howe) II, 134 
 
 Sulphocarbanilide, action of, on certain acid anhydrides (Dunlap) . . I, 355 
 
 action of, on phthalic anhydride (Dunlap) I, 355 
 
 action of, on succinic anhydride (Dunlap) 1,359 
 
 Sulphocyanides of copper and silver in gravimetric analysis (Van Name) II, 359 
 Sulphuric acid, action of potassium permanganate upon (Gooch and Dan- 
 
 ner) I, 145 
 
 Tartar emetic, stability of standard solutions of (Gruener) I, 216 
 
 Tellurium, determination of, by precipitation on the iodide (Gooch and 
 
 Morgan) n, 1 
 
 method for the separation of selenium from (Gooch and Peirce) . I, 348 
 Telluric acid, iodometric method for the estimation of (Gooch and How- 
 land) 1, 277 
 
 Tellurous acid, determination of, by potassium permanganate (Gooch 
 
 and Danner) I, 154 
 
 determination of, in presence of haloid salts (Gooch and Peters) . . II, 238 
 
 Thallium, estimation of, as acid and neutral sulphates (Browning) . . II, 317 
 
 estimation of, as chromate (Browning and Hutchins) II, 300 
 
 Thiosulphates, detection of, in presence of sulphides, sulphites, and sul- 
 phates (Browning and Howe) II, 134 
 
 Tin, detection of arsenic associated with (Gooch and Hodge) .... I, 231 
 
 separation of copper as oxalate from (Peters) H, 347 
 
 Titanium salt, action of sodium thiosulphate upon, at high temperatures 
 
 and pressures (Norton) II, 392 
 
 Toluquinoneoxime, ethers of, and their bearing on the space isomerism 
 
 of nitrogen (Bridge and Morgan) II, 145 
 
 Toluquinoneoxime ethers, space isomerism of (Morgan) II, 283 
 
 Toluquinonemetaoxime acetyl ether (Bridge and Morgan) 11,153 
 
 Toluquinonemetaoxime benzoyl ether (Bridge and Morgan) II, 154 
 
 Toluquinonemetaoxime methyl ether (Bridge and Morgan) II, 152 
 
 Toluquinonemetaoxime, sodium salt of (Morgan) II, 285 
 
 Toluquinoneorthooxime acetyl ether (Bridge and Morgan) II, 160 
 
 Toluquinoneorthooxime benzoyl ether (Bridge and Morgan) . . . . II, 160 
 
 Toluquinoneorthooxime benzoyl ether dichloride (Morgan) II, 287 
 
 Toluquinoneorthooxime methyl ether (Bridge and Morgan) . . . . II, 159 
 Tolylmaleamic acid (o), preparation of, from maleic anhydride and 
 
 o-toluidine (Dunlap and Phelps) II, 145 
 
 Tolylmaleamic acid (p), preparation of, from maleic anhydride and 
 
 p-toluidine (Dunlap and Phelps) II, 44 
 
 Urea, action of, on certain acid anhydrides (Dunlap) I, 355 
 
 action of, on maleic anhydride (Dunlap and Phelps) II, 42 
 
 Vanadic acid, reduction of by hydriodic and hydrobromic acids (Brown- 
 ing) 1,397 
 
 estimation of, iodometrically (Browning) 1,397 
 
 Vanadium, application of certain organic acids to estimation of (Brown- 
 ing and Goodman) II, 4 
 
INDEX OF SUBJECTS. 415 
 
 VOL. PAGE 
 
 Volatilization of the iron chlorides in analysis (Gooch and Havens) . . 11,215 
 
 Volumetric estimation of mercury (Peters) 11,320 
 
 Zinc, separation of aluminum from (Havens) II, 107 
 
 separation of copper as oxalate from (Peters) II, 357 
 
 Zinc ammonium phosphate in analysis (Austin) II, 257 
 
 Zinc chloride (anhydrous), use of, in detecting perchlorates (Gooch and 
 
 Kreider) 1, 247 
 
 Zirconium, separation of iron from, by gaseous hydrochloric acid 
 
 (Havens and Way) 11,266 
 
 Zirconium salt, action of sodium thiosulphate upon, at high temper- 
 atures and pressures (Norton) II, 391 
 
UNIVERSITY OF CALIFORNIA LIBRARY 
 
 
 THIS BOOK IS DUE ON THE LAST DATE 
 STAMPED BELOW 
 
 JAN 31 1916 
 
 30m-l,'15 
 

 04198