posed by continued heating of the acidified fluid, completely neutralise the excess of acid with soda solution free from carbonic acid, allow to cool, add the usual quantity of oxalate of soda, and proceed as above.

If no soda solution free from carbonic acid is at hand, place the oxalate of soda or oxalic acid (about 3 grm.) in a small tube, and suspend this in the flask A by means of a thread, fastened by the cork. When the apparatus is tared, and it has been proved to be air-tight, release the thread, and proceed as above.

In the decomposition flask place the ore and some dilute sulphuric acid, and add a solution of oxalic acid through the funnel tube; if necessary, also dilute sulphuric acid. If the ore contains alkaline earthy carbonates, their carbonic acid may be determined in a convenient manner by this process, before the oxidation of the oxalic acid is commenced.

BLOWPIPE REACTIONS OF MANGANESE.

SULPHIDE OF MANGANESE.-Alone, in the matrass, undergoes no change.

In the open tube roasts slowly, but gives no sublimate. The roasted surface takes a bright green tinge.

On charcoal, after complete roasting, behaves with the fluxes like pure oxide of manganese.

PEROXIDE OF MANGANESE.-Alone, in the matrass, when pure, undergoes no sensible alteration, but in general it contains more or less hydrate of manganese, the water of which may be driven off by means of heat. The more water the heated matter gives off, the less available oxide of manganese it contains. On charcoal it becomes reddish-brown in a good reducing flame.

With borax and microcosmic salt it dissolves with a brisk effervescence, produced by disengagement of oxygen; it then behaves as oxide of manganese.

OXIDE OF MANGANESE.-Alone, the protoxide is not fusible, but becomes brown in a strong flame.

With borax it forms a transparent glass, having the colour of amethyst, which becomes colourless in the reducing flame.

If much oxide be present, the glass must be pressed on a cold body, at the instant the blast ceases. The colour returns by

a slow cooling.

With microcosmic salt it fuses readily, forming a transparent glass, which is colourless in the reducing flame, and amethystine in the oxidising flame. If the glass produced by the union of oxide of manganese with phosphoric acid contain so little of the former as to give no sensible reaction, it can be rendered evident by plunging into the bead a crystal of nitre. The bead swells and foams, and the froth becomes on cooling an amethystine or pale rose tint, according to the quantity present.

With soda, the oxide fuses on platinum foil or wire, forming a transparent green glass, which becomes on cooling a bluish-green. This assay is best made on platinum foil. One-thousandth of oxide of manganese gives a very perceptible colour with soda.

662

CHAPTER XXII.

ASSAY OF COBALT AND NICKEL ORES.

ALTHOUGH cobalt and nickel usually accompany each other, yet it will be more convenient to give the ores of both sepa rately, commencing with those of cobalt.

Assay for Cobalt.-The analysis of cobalt ores is the most tedious, with the exception of those of platinum, of any that fall under the assayer's notice-the greatest difficulty being in the separation of cobalt and nickel. The following process, however, is the most ready that has yet been devised. Very carefully roast, in a porcelain capsule or crucible, 100 or more grains of the sample to be examined. (In case, however, any of the rich ores are under assay, 25 to 50 grains will suffice.) When no more vapours of arsenious acid are evolved, add a little finely-powdered charcoal, and again roast, and so on until no arsenical smell is perceptible. Allow the roasted mass to cool, and then gently heat it in a

flask with hydrochloric acid until all but silica is dissolved; evaporate to dryness; allow to cool; moisten with hydrochloric acid; let stand for an hour: then add water, boil, and filter. To the cold filtered solution add a little hydrochloric acid, and pass into this acidulated solution sulphuretted hydrogen gas until in great excess; allow the solution so saturated with gas to remain at rest for two or three hours, then filter it, add a little nitric acid to the filtered solution, and boil so as to peroxidise all the iron present: this point must be carefully attended to, and may be recognised by the addition of a few drops of nitric acid to the hot solution giving no dark tinge. Allow the solution to cool, and if not quite bright, filter it. To the filtered solution add excess of carbonate of baryta. Iron and alumina will be removed after a digestion of three or four hours. Again filter, and to the solution add sulphide of ammonium in excess, gently warm and filter, wash the precipitate, dissolve it in hydrochloric acid; if not bright, filter, and to the filtered solution add cyanide of potassium in excess, and boil. To the boiling solution add a little carbonate of soda -this will precipitate manganese if present-and filter. The solution now contains nothing but cobalt and nickel. These may be separated as follows:-Warm the solution. and add to it excess of pulverised peroxide of mercury this decomposes the potassio-cyanide of nickel, and the whole of the nickel precipitates, the cobalt alone remaining in solution. Remove the nickel by filtration, and neutralise as nearly as possible the filtered solution containing the cobalt by the aid of nitric acid; then add neutral nitrate of mercury solution as long as a white precipitate forms: this cipitate is cyanide of mercury and cobalt. It is collected on a filter, well washed, dried, and then ignited, with free access of atmospheric air, to convert it into black peroxide of cobalt, which is weighed. The nickel precipitate collected on the filter is treated in the same manner: every 100 parts of oxide of nickel correspond to 787 parts of metallic nickel. It may be here mentioned, that cobalt is always estimated commercially as oxide, and nickel as metal.

pre

A method of separating these metals, given some years since by Liebig, consists in boiling the mixed double cyanides of nickel and potassium and cobalt and potassium with oxide of mercury. Oxide of nickel is precipitated, while an equivalent quantity of mercury is dissolved as cyanide. The method certainly gives good results, but is not free from objection. Long boiling is necessary before the precipitation is complete, and it is difficult to prevent bumping during ebullition. The excess of oxide of mercury must be separated from the oxide of nickel by a special operation, and the nickel afterwards again precipitated by caustic alkali.

According to Wolcott Gibbs, these inconveniences may be completely avoided by employing, instead of the oxide alone, a solution of the oxide in the cyanide of mercury. When this solution is added to a hot solution of the double cyanide of nickel and potassium, the whole of the nickel is immediately thrown down as a pale green hydrate of the protoxide, Under the same circumstances cobalt is not precipitated from the double cyanide of cobalt and potassium. Mr. W. N. Hill, who has repeatedly employed this method and carefully tested it, has found that the separation effected is complete. No cobalt can be detected in the precipitated oxide of nickel by the blowpipe, nor can the nickel be detected in the cobalt (finally separated as oxide) by Plattner's process with the gold bead. The solution of oxide of mercury is easily obtained by boiling the oxide with a strong solution of the cyanide, and filtering. According to Kuhn, the cyanide formed in this manner has the formula HgCy+3HgO. The hydrated oxide of nickel precipitated may be filtered off, washed, dried, ignited, and weighed. The cobalt is more readily and conveniently determined by difference, when, as it is always possible, the two metals have been weighed together as sulphates. I am not prepared to say that this modification of Liebig's method of separating nickel and cobalt gives better results than Stromeyer's process by means of nitrite of potassium, but it is at least very much more convenient, and requires much less

* Chemical News, March 17, 1865.