was then taken from the filter, air-dried, and analyzed (analysis b). Through the remainder (perhaps two teaspoonfuls) of the precipitate five litres of water were passed; another portion was then dried at 100° and analyzed (analysis c). When dried at 100° the precipitate is much less readily soluble in dilute nitric acid, than when not so heated, but the difficulty is easily overcome. The method of analysis was that described above.

A comparison of these three determinations teaches that the abstraction of chromic acid by washing with cold water stops at no definite point, and probably has no limit except the complete change of the original precipitate into chromic oxide. The composition of the substance a must be a close approximation to the composition of the original precipitate, for the slight rinsing which was intended merely to remove the chromates belonging to the filtrate could hardly have abstracted much of its combined chromic acid. In the chromate of chromium whose formula would be Cr2O, CrO3, the ratio of the chromic oxide to the chromic acid would be that of 3: 2; and we find that the ratio of the chromic oxide to the chromic acid in analysis a is but little larger than that of 3: 2. The substance a is probably therefore a chromate of chromium of the formula Cr2O3 CrO3, from which a little of the chromic acid has been removed by washing.

Assuming for the present that the precipitate produced by the reaction of chrome alum on monochromate of potash is in truth this chromate of chromium, let us inquire into the nature of the reaction between these two salts by which this precipitate could be formed, and bichromate of potash left in the filtrate. Concentrated solutions of pure chrome alum and of pure monochromate of potash were prepared of known strengths, in order to determine the amount of either solution required to produce a precipitate in the other. A single drop of chrome alum solution produces an immediate and permanent precipitate in the chromate solution; but if the process be reversed and the solution of monochromate of potash be dropped into the alum solution, no permanent precipitate is formed till an apparently large excess of chromate

of potash has been added. By accurate quantitative experiments we have proved that five equivalents of chromate of potash must be added to one equivalent of chrome alum to effect precipitation. If a less quantity be added, any precipitate which may form will at once redissolve when the mixture is made complete by agitation.* The reaction between the two salts may therefore be expressed by the formula,

5 (KO CrO3) + KO SO,, Cr2O, 3 SO3=4 (KO SO3) + 2 (KO 2 CrO3) +Cr303 Cro3,

and if to one equivalent of chrome alum more than five equivalents of chromate of potash be added, the excess above five remains inactive.

x (KO CrO3)+KO SO3, Cr2O, 3 SO3=4 (KO SO3) +2 (KO 2 CrO3) +Cr2O3 Cro3+ (x-5) (KO CrO3).

If any other normal salt of chromic oxide be used, a similar formula will express the reaction; thus:

Cr2Cl2+5 (KO CrO3) = 3 (K Cl) +2 (KO 2 CrO3) + Cr2O3 Cro3. CrO3 3 SO3+5 (KO CrO3) = 3 (KO SO2) + 2 (KO 2 CrO3) + Cr2O3 Cro3.

With the exact proportions of chrome alum and chromate of potash which are by the formula necessary for the precipitation of chromate of chromium we prepared a quantity of the precipitate for a second series of analyses in corroboration of analyses a, b, and c. The precipitate thrown upon a filter was very slightly rinsed with cold water, a portion of it was taken off, pressed between folds of filter-paper under heavy weights, and air-dried for analysis (analysis d); the rest of the precipitate was washed for six hours till no trace of the filtrate could possibly be supposed to be retained by the precipitate, when another portion was taken out for analysis, pressed, and air-dried (analysis e) ; lastly five litres of water were passed through the remaining precipitate, which was then dried at 100° and submitted to analysis (analy

* In this connection we would call attention to the inaccuracy of the statement made by Berzelius in his Traité de Chimie, (Paris, 1846, Vol. II. 307), to the effect that a compound whose formula is Cr2O3 3 CrO3 is precipitated, when a few drops of the solution of monochromate of potash are added to the solution of a neutral chrome salt. No permanent precipitate whatever is produced under these circumstances, as any one may satisfy himself by repeating the simple experiment described above.

sis f). The process of analysis was the same described above, and the results were as follows:

This set of analyses corroborates the first series in every respect, and the ratio of the chromic oxide to the chromic acid of the substance d is almost precisely the ratio 3: 2 of the chromic oxide to the chromic acid in the chromate of chromium Cr2O, CrO3. We again see that there is no definite limit to the removal of chromic acid by prolonged washing, and that the substance may be readily made more basic than any imaginable definite compound of chromic oxide and chromic acid would be.

In these analyses the precipitates had been somewhat washed before any analysis was made; it remained to analyze the substance precipitated with so much of the adhering filtrate as could not be removed by pressure between folds of filter-paper without washing. One precipitate (analysis g) was prepared by mixing a solution of one equivalent of chrome alum with eight equivalents of chromate of potash in solution, and a second by mixing a solution of one equivalent of chrome alum with sixteen equivalents of chromate of potash in solution (analysis h). In both cases the solutions used were concentrated and the precipitates air-dried. The method of analysis which we had heretofore employed was open to one objection, -a trace of chromic oxide was liable to be dissolved in the excess of ammonia, to be again separated when the filtrate and wash-waters from the chromic oxide precipitate were concentrated by evaporation, and the process would obviously be altogether inadmissible in any case where the sulphates of the mother-liquor had not been removed by washing before submitting the precipitate to analysis, since sulphate of lead would be formed and vitiate the determination of the chromic acid. We therefore resorted to Rose's method of separating chromic oxide from chromic acid, by means of the nitrate of the suboxide of mercury.* The solution of the substance to be analyzed in dilute nitric acid was nearly neutralized with carbonate of potash, and when the carbonic acid had been allowed sufficient time to escape, nitrate of the suboxide of mercury was added,

*Handbuch der Analytischen Chemie, (Braunschweig, 1851,) II. 379.

the precipitate filtered off, washed, strongly ignited, and the chromic acid calculated from the pure chromic oxide which remained. To the filtrate from this first precipitate ammonia in excess was added, and the washed precipitate, strongly ignited, was the chromic oxide contained in the original substance.

It is quite evident from these analyses that the precipitate caused by mixing chromate of potash with a neutral chrome salt cannot contain more than one equivalent of chromic acid in combination with its chromic oxide, since in the above determinations the chromic acid is nearer one equivalent than two with reference to the chromic oxide found, in spite of the fact that the considerable amount of chromic acid which, combined with potash, has adhered mechanically to the precipitate, is thrown down by the nitrate of the suboxide of mercury, together with the chromic acid which was originally united to the chromic oxide. But the fact that the substance under examination cannot be purified by washing without altering its constitution, renders it impossible to ascertain the exact composition of the body by the methods of analysis heretofore used, or by any similar methods; for this purpose the process used by Vogel in analyzing this same substance, obtained by him from a different source and miscalled CrO2, is admirably adapted. The precipitate examined by this method was not washed at all, but was simply dried by pressure between folds of filter-paper and exposure to the air. A weighed portion was placed in the bulb of a reductiontube, with which a weighed chloride of calcium tube was connected; a current of dry air was then drawn through the apparatus, and the reduction-tube was cautiously heated till all the water of the precipitate had been absorbed by the weighed chloride of calcium tube, the heat finally rising to dull redness. The salts with which the precipitate under examination was contaminated were sulphate of potash, bichromate of potash, and a little chromate of potash, and since the solutions from which the precipitate was prepared were concentrated, it was to be

*Jour. pr. Ch., LXXVII. 484 (1859).

expected that a considerable quantity of these salts would adhere to the precipitate; but since all these salts are fixed at a low red heat, they were not altered by the heat to which the precipitate was exposed. The rest of the precipitate must have lost by the ignition all its water and all the oxygen over and above that necessary to the constitution of chromic oxide. By subtracting the weight of the water collected from the total loss by ignition, the weight of the oxygen expelled is obtained. The ignited residue was then washed out of the bulb of the reductiontube, digested in hot water, thrown upon a filter and washed with hot water till only pure chromic oxide remained; lastly, this oxide was ignited and weighed. By subtracting the weight of the chromic oxide from the weight of the whole ignited residue, the weight of the soluble salts which adhered mechanically to the precipitate was obtained. The following are the figures of an analysis by this method:

Weight of reduction-tube + precipitate

10.3848 gram.

9.6136 66

In the compound whose formula is Cr2O3 CrO3, the ratio of the number of atoms of oxygen to the number of atoms of chromium is that of 2: 1. By adding another equivalent of oxygen to the chromic oxide found, we shall make the ratio of the atoms of oxygen to the atoms of chromium that of the compound Cr2O, CrO3. Hence the proportion.