the precipitation of the nickel by potass, renders its application far less advantageous.

2. OXIDE OF ZINC, PROTOXIDE OF NICKEL, AND PROTOXIDE OF MANGANESE, may also be separated from oxide of chromium, by means of carbonate of barytes, as will be found explained in § 124, 1, B. The digestion with the carbonate of barytes in excess should be continued for several hours, otherwise some chromium might remain in solution.

IV. SEPARATION OF THE OXIDES OF THE FOURTH GROUP FROM ONE ANOTHER.

$124.

1. PEROXIDE OF IRON FROM ZINC, MANGANESE, AND

NICKEL.

Of the two methods which I shall describe here, the first is to be preferred when the relative quantity of peroxide of iron present is considerable, the second when this quantity is small.

a. Separation by means of succinate of ammonia.

The acid solution is mixed with sal ammoniac, and subsequently neutralized with ammonia, so as to precipitate a very trifling portion of the peroxide of iron; solution of neutral succinate of ammonia (benzoate of ammonia will answer equally well) is then added, and the precipitated persuccinate of iron filtered off; the filtrate contains the other metals in solution. For the details of this process, vide § 85. The separation achieved by this method is complete, provided the process be properly and carefully performed.

B. Separation by means of carbonate of barytes.

This method is applicable only in cases where the metallic oxides present are combined with acids which form soluble com pounds with barytes; those acids should not be present in too great excess.

The acid solution is very gently heated, (to from 86° to 104°,) and mixed with artificially produced carbonate of barytes (still

moist) (vide § 37. 6) until this reagent manifestly predominates, even after protracted digestion of the mixture; (this point is marked by the cessation of disengagement of gas and its nonrenewal upon addition of a fresh portion of the carbonate). The precipitate which forms is allowed to subside, the fluid is filtered off, and the precipitate carefully washed. It consists of basic carbonate of the peroxide of iron, mixed with the excess of the carbonate of barytes added; it is dissolved in dilute hydrochloric acid, the barytes precipitated with sulphuric acid, the fluid filtered off from the precipitated sulphate of barytes, and the peroxide of iron in the filtrate finally precipitated with ammonia, (§ 85). The first filtrate contains, besides the salts of manganese, nickel, &c., the newly-formed salt of barytes; the barytes here is likewise removed by means of sulphuric acid, and the remaining metals are separated from one another by methods which we shall shortly have occasion to describe. The results afforded by this method are satisfactory. The digestion heat must not be carried beyond 104°, otherwise traces of manganese, &c., may be precipitated in conjunction with the basic persalt of iron. However, we obtain always rather a little too much of the iron at the expense of the other metals, than the reverse, since there is invariably a tendency for traces of the latter to precipitate with the basic persalt of iron. Artificially prepared carbonate of lime may be substituted for the carbonate of barytes, which latter is however to be preferred in most instances, since barytes is more readily removed both from the precipitated peroxide of iron and from the filtrate, than is the case with lime.

2. PEROXIDE OF IRON FROM COBALT.

The separation is effected by means of succinate of ammonia in exactly the same manner as § 124, 1, a. Complete separation of these two metals is not to be attained by carbonate of lime or of barytes.

3. PEROXIDE OF IRON FROM OXIDE OF ZINC.

The acetic solution, to which free acetic acid is to be added, and which must contain no other acid, is acted upon by sulphuretted hydrogen, (compare separation of alumina from oxide of

zinc, § 123, A., b., 1). Should the precipitated sulphuret of zinc not be of a pure white color-a sign that it contains some sulphuret of iron in admixture-the precipitated fluid is to be digested at a gentle heat until the precipitate acquires a pure white tint, and after cooling, a fresh amount of sulphuretted hydrogen is to be transmitted through it.

4. PROTOXIDE OF IRON FROM ZINC, COBALT, MANGA

NESE, AND NICKEL

The protoxide of iron is converted into peroxide, (§ 84), and the separation effected as directed § 124, 1.

5. PEROXIDE OF IRON FROM PROTOXIDE OF IRON. The separation of the peroxide from the protoxide of iron is a difficult task, and requires the strictest attention in the performance of the requisite process.

The proportion of these two substances may, however, be likewise determined in an indirect manner, and this proceeding is, in general, preferable to direct separation.

a. Analysis by direct separation.

This method is applicable only when both oxides exist in a form in which they are soluble by hydrochloric acid, or when we have them in a solution which contains no acids, forming insoluble salts with barytes. The apparatus illustrated by Plate XL. serves for its execution.

The finely levigated substance is introduced into the flask A, and carbonic acid gas conducted into the flask through the tube d. As soon as the air is completely expelled, hydrochloric acid is poured into the flask through the funnel-tube c, a slow stream of carbonic acid gas being maintained all the while. The action of the hydrochloric acid is assisted

by the application of heat. The acid must not be added too greatly in excess. The solution obtained by this operation is diluted with boiled and still hot water, and, after cooling, pure recently precipitated carbonate of barytes, mixed with boiled water to a milky fluid, is added through the funnel tube until it predominates; the whole mixture is then digested for some time a little above the common temperature. The flask is then filled with boiled water nearly up to the end of the tube b, which is afterwards slowly and cautiously depressed (by turning) and immersed into the clear fluid as far as necessary. The clear fluid is then drawn off through this tube, which is subsequently raised again. The flask is once more filled with boiled water, and the fluid again drawn off through the tube b; the remaining contents of the flask are rinsed out with boiled water and placed upon a filter, where they are subsequently washed with boiled water, avoiding as much as possible the contact of air; the peroxide of iron in the washed precipitate is then finally determined. (Compare § 124, 1, B.) The whole of the dissolved protoxide of iron is contained in the liquid drawn off, and in the fluid filtered off from the precipitated peroxide. This solution is concentrated by evaporation, the protoxide is peroxidized, (§ 84,) and the peroxide of iron formed, precipitated, after previously removing the barytes, (§ 85). Should the analysed compound contain a constituent insoluble in hydrochloric acid, this is regained upon the subsequent solution in hydrochloric acid of the precipitate containing the basic persalt of iron; its weight is ascertained, and subtracted from the total weight of the analysed substance.

B. Indirect analysis.

This may be achieved in various ways; I will, however, content myself with giving the two following methods, which will be found sufficient under all circumstances. The process in both commences with the solution of the substance in hydrochloric acid, which may be effected either as directed sub. a., or in some other appropriate manner.

aa. The solution is effected in a weighed flask, (which we will designate a); the flask is then weighed with its contents, and about one-half of the latter poured into another flask (b), filled previously with carbonic acid, and provided with a ground-glass stopper; the quantity of the solution poured into this flask is ascertained from the loss of weight of a. The solution in b is then strongly acidified with hydrochloric acid, the flask filled nearly to the brim with boiled water, a weighed slip of clean sheet copper § 37, (16) is introduced into it, the flask stoppered, a piece of bladder tied over it, and the mixture digested in boiling water until the fluid grows colorless, or acquires a greenish hue,—a sign that the perchloride of iron is completely converted into protochloride. The slip of sheet copper is then taken out of the flask, dried and weighed. The total amount of peroxide of iron in a is determined after previous oxidation of the protoxide, (§ 84). The calculation is very simple. The quantity of peroxide of iron obtained in the last process, calculated upon the total amount of the solution, shows the total amount of iron operated upon. The diminution in the weight of the copper slip, calculated from the part to the whole, indicates the amount of chlorine consumed to convert the original protochloride of iron into perchloride, since every one equivalent of dissolved copper corresponds to one equivalent of chlorine; every one equivalent of chlorine converting two equivalents of protochloride of iron into perchloride, (2 Fe Cl+Cl=Fe, Cl,); it follows that every one equivalent of dissolved copper corresponds to two equivalents of perchloride of iron in the solution, or what amounts to the same, to two equivalents of peroxide of iron present in the analysed substance. (Fuchs.) This method is not applicable in presence of arsenic acid.

bb. The solution is effected in a flask, solution of chloride of gold and sodium is added in excess, the flask stoppered, and the precipitated metallic gold allowed to subside. The fluid is filtered off from the precipitated gold, and the amount of the latter determined as directed § 93. The total amount of iron present