which case the precipitate consists of pure sesquicyanide of cobalt and nickel, Cy Co2Ni3. If a larger proportion of nickel be present, the precipitate contains also cyanide of nickel, which by prolonged boiling in hydrochloric acid is decomposed into hydrocyanic acid and soluble chloride of nickel. If, on the other hand, a larger proportion of cobalt be present, the precipitate consists of sesquicyanide of cobalt and nickel, and some cobalt remains in solution as sesquicyanide of cobalt and potassium. The pre

cipitate of sesquicyanide of cobalt and nickel (when all the cyanide of potassium and cyanide of nickel present have been decomposed by boiling in hydrochloric acid) is decomposed by caustic potash into hydrate of nickel which is precipitated, and sesquicyanide of cobalt and potassium which remains in solution. This reaction is applied as follows for the detection of nickel in presence of cobalt. The solution (which must not contain iron or manganese) is slightly acidulated with hydrochloric acid, excess of cyanide of potassium added, and the whole heated to boiling if a precipitate is produced by the addition of dilute sulphuric or hydrochloric acid, nickel is present. A solution containing cobalt as sesquicyanide of cobalt and potassium remains clear when mixed with an alkali and saturated with chlorine in the cold: if a trace of nickel be present an inky-black solution is produced. Ferrocyanide of potassium gives with cobalt-salts a green, ferricyanide, a red-brown precipitate, both insoluble in hydrochloric acid. Alkaline phosphates and arsenates give a peach-red precipitate, readily soluble in acids. Oxalic acid gives a pale-rose precipitate of oxalate of cobalt, which forms slowly: it is soluble in ammonia, and the oxalate separates out again on exposure to the air, but much more slowly than the oxalate of nickel. Alkaline carbonates precipitate a peach-red basic carbonate, readily soluble to a red solution in carbonate of ammonium, very slightly soluble in carbonate of potassium or sodium. Potash precipitates all the cobalt as a blue basic salt, which, if the air be excluded, is gradually converted into rose-red hydrate, CoHO; this change is facilitated by heat. In contact with the air it becomes olive-green, owing to the formation of an intermediate oxide. Ammonia precipitates a blue basic salt, soluble in excess

to a red-brown solution. If this solution mixed with chloride of ammonium be allowed to absorb oxygen, it becomes brown; and on boiling it in strong hydrochloric acid the whole of the cobalt. is precipitated as a peculiar compound, having the formula N5H16 Cl3Co2. (Claudet. Chem. Soc. Qu. J. iv. 355.) If the cobalt solution contains free acid or an ammoniacal salt, ammonia produces only a red-brown solution, in which potash gives no precipitate. All soluble cobalt-salts are decomposed by heat, and after ignition are not completely soluble in water.

b. Sesquioxide of Cobalt, Co103 - A black powder, obtained by gently igniting the nitrate of cobalt. When heated it loses oxygen, and forms an intermediate oxide. The sesquihydrate is a black mass, obtained by passing chlorine through water in which the hydrate is suspended. Sesquichloride of cobalt is formed when a very dilute solution of cobalt containing no acid but hydrochloric is completely saturated with chlorine in the cold: from this solution carbonate of barium precipitates sesquihydrate. It is decomposed by most acids, with formation of a protosalt, and evolution of oxygen or chlorine. Acetic acid dissolves it without decomposition, forming a dark-brown solution, whence potash precipitates the sesquihydrate.

Blowpipe reactions. -Cobalt-salts are reduced when heated on charcoal with carbonate of sodium in the inner flame, cobalt being separated as a grey magnetic powder. (c) With borax or microcosmic salt they give in both flames a dark-blue bead, the colour of which is scarcely affected by the presence of any other metal. Separation and estimation of Cobalt. -The best reactions for the qualitative detection of cobalt are the blowpipe reaction, and the difficult solubility of its sulphide in dilute hydrochloric acid, in which it is less soluble than the sulphide of nickel. The least trace of cobalt and nickel may be detected in iron-ores or other minerals by dissolving the substance in hydrochloric acid, saturating the acid solution with hydrosulphuric acid (to remove the metals of Group I.), adding sulphide of potassium to the filtrate, and washing the precipitated sulphides with cold dilute hydrochloric acid. The black residue (sulphides of nickel and cobalt) is tested for cobalt by the blowpipe, and for nickel by cyanide of

potassium in its acid solution. If a large quantity of iron be present, it is best to remove it first by nearly saturating the acid solution with carbonate of sodium, and boiling with acetate of sodium. The general method for analysing cobalt-ores is the same as that described in the case of nickel (p. 143).

The separation of cobalt from nickel is very difficult. The best method is that given by H. Rose. The metals, or their oxides, are dissolved in hydrochloric acid, the solution very largely diluted with water, and chlorine passed through it in the cold for several hours till it is thoroughly saturated. The chloride of cobalt is thus converted into sesquichloride, the chloride of nickel remaining unchanged. Carbonate of barium is then added, and the whole allowed to stand for twelve or eighteen hours, being shaken up from time to time: and then filtered. The precipitate is washed with cold water, dissolved in boiling hydrochloric acid, freed from barium by sulphuric acid, and the cobalt precipitated by potash. The green filtrate is also treated with sulphuric acid to remove any barium it may contain, and the nickel then precipitated by potash.

Liebig's method is based on the reaction with cyanide of potassium above described. The solution of the two metals is treated with excess of potash and hydrocyanic acid (or cyanide of potassium perfectly free from cyanate) till the precipitate first formed is redissolved. The solution is boiled to expel free hydrocyanic acid, and mixed while warm with finely divided oxide of mercury, which precipitates all the nickel as hydrate and cyanide: the precipitate, when washed and ignited, yields pure oxide of nickel. The filtrate, which contains all the cobalt as sesquicyanide of cobalt and potassium, is neutralised by nitric acid, and precipitated by a neutral solution of subnitrate of mercury: the precipitate is washed, ignited, and heated in a stream of hydrogen, when pure metallic cobalt is left. Or the solution of the cyanides in cyanide of potassium may be saturated with chlorine in the cold, with occasional addition of caustic potash or soda: the whole of the nickel is thus precipitated as sesquihydrate, while the cobalt remains in solution.

Cobalt is best estimated in the metallic state. It is precipi

tated by potash as a basic salt, which is converted into the hydrate by heat: the hydrate is washed with hot water, dried, and ignited in a stream of hydrogen, and the reduced metal weighed.

3. Manganese. Mn. Atomic weight, 26.

Found as sesquioxide (braunite); binoxide (pyrolusite); and an intermediate oxide (Hausmannite): as sulphide, carbonate, silicate, &c. and in most iron-ores. It is a greyish-white, brittle, difficultly fusible metal. It is readily oxidised by exposure to moist air at the ordinary temperature, forming a black powder.

OXIDES OF MANGANESE. Protoxide, Mn2O. Sesquioxide, Mn403. Binoxide, Mn2O2. Manganic Anhydride, Mn2O3. Permanganic Anhydride, Mn407.-There are also two intermediate oxides, Manganoso-manganic Oxide, Mn604 (Mn2O+Mn403), found native as Hausmannite; and Mn8O7 (Mn1O3+2Mn2O2), found native as Varvicite.

a. Protoxide of Manganese. Manganous Oxide.Mn2O.—A green powder, obtained by igniting any of the higher oxides in a stream of hydrogen. It readily absorbs oxygen from the air, turning brown. It is not reduced by ignition in a stream of hydrogen. It is readily soluble in acids, forming protosalts of manganese, or manganous salts, which are either colourless or of a pale rosecolour: they are mostly soluble in water, all in hydrochloric acid, forming colourless solutions, which are not converted into manganic salts by exposure to the air or boiling with nitric acid. All the higher oxides of manganese, when heated with hydrochloric acid, evolve chlorine, and are converted into the chloride, MnCl: hence any manganese compound, after treatment with hydrochloric acid, exhibits the reactions of a manganous salt. Manganous salts are isomorphous with magnesium-salts. Hydrosulphuric acid does not precipitate a neutral solution of a manganous salt containing a mineral acid: even acetate of manganese is precipitated very slowly and imperfectly, and not at all if free acetic acid be present. (c) Sulphide of ammonium precipitates flesh-coloured sulphide, Mn2S, readily soluble in acetic acid, absorbing oxygen and turning brown on exposure to the air. Yellow sulphide of ammonium gives a yellowish-white precipi

tate, which gradually becomes flesh-coloured by exposure to the air, or more rapidly on heating. Cyanide of potassium gives a whitish precipitate, soluble in a large excess to a brown solution, which is not precipitated by sulphide of ammonium. Ferrocy anide of potassium gives a reddish-white precipitate, soluble in acids: ferricyanide of potassium, a brown-yellow precipitate, insoluble in acids. Alkaline phosphates, arsenates, oxalates, and carbonates, give white precipitates. Potash precipitates white hydrate, MnHO, insoluble in excess: it absorbs oxygen from the air and turns brown, and is then not completely soluble in chloride of ammonium. Ammonia precipitates white hydrate from neutral solutions: in solutions containing free acid or ammoniacal salts, it gives no precipitate: but, if sufficient ammonia be added, the solution, on exposure to the air, gradually deposits all the manganese as brown sesquihydrate. (c) The least trace of a manganous salt is detected by heating the solution with a little binoxide of lead (or minium) and nitric acid (which must be free from hydrochloric acid), when an intense purple-red colour is produced, owing to the formation of permanganic acid: the colour is readily perceptible when the excess of lead-oxide has subsided. This is the most delicate test for manganese in the wet way.

b. Sesquioxide of Manganese. Manganic Oxide. Mn403.- A black powder, soluble in some acids without decomposition, forming manganic salts, which are isomorphous with sesquisalts of iron and aluminium. Their solution is red: potash precipitates from it black sesquihydrate, unless chloride of ammonium be present. They are very unstable, being reduced to manganous salts merely by heating the same reduction is effected, and the solution decolorised, by hydrochloric, sulphurous, or nitrous acid, by any organic compound, &c. When any oxide of manganese, or a decomposable manganese-salt is heated in the air, manganoso-manganic oxide, Mn604, is formed.

c. Binoxide of manganese, Mn2O2.-This is the most important native ore of manganese. It is a black earthy powder: the hydrate is brown, and is obtained when a manganous salt is precipitated by a hypochlorite. When heated alone, or with sul