Ammoniacal Compounds of Cobalt. 259 



The study of this and other allied compounds of cobalt which 

 exist will no doubt greatly extend our views respecting the com- 

 pound ammonias. 



The chlorine of the original chloride may also be eliminated 

 by any silver salt, an analogous cobalt salt containing the acid of 

 the silver salt being formed and remaining in solution. In this 

 way I have been enabled to obtain a sulphate, nitrate, oxalate, 

 acetate, and carbonate of the new base. From the carbonate I 

 have prepared the bromide and iodide which have the octohedral 

 form of the chloride, are equally sparingly soluble in water, and 

 of a still darker ruby colour. The bromide was found to con- 

 tain 61 '15 per cent, bromine, the calculated amount being 61'8 

 per cent. 



Tlie insolubility of this ammoniacal compound of cobalt in 

 boiling hydrochloric acid may be advantageously turned to ac- 

 count in the preparation of chemically pure cobalt, and also in 

 the qualitative examination of substances containing small quan- 

 tities of cobalt. The pulverized ore or its oxide to be purified is 

 dissolved in nitro-hydrochloric acid, diluted with water, and 

 filtered in order to separate any gangue or insoluble residue. 

 Chloride of ammonium is now added in large excess, and the 

 liquid saturated with ammonia ; it is then poured into a glass 

 bottle, and oxidated in the way I have already described in the 

 preparation of the new salts. During the oxidation a certain 

 quantity of the new compound is deposited, especially when the 

 solutions are rather concentrated, on account of its insolubility 

 in a strong solution of chloride of ammonium. The solution still 

 retains a certain quantity of cobalt salt ; it is therefore boiled with 

 a considerable excess of hydrochloric acid, which causes the total 

 precipitation of the new compound, dissolving at the same time 

 any oxide of iron or other oxides thrown down by the ammonia. 

 When cold the clear liquid is decanted ofi", and the deposit well 

 washed with acidulated water and then dried. By heating this 

 compound to low redness it is decomposed, leaving for residue 

 protochloride of cobalt slightly decomposed, but absolutely free 

 from any other metal. This may be reduced by hydrogen gas 

 giving pure metallic cobalt. 



By these means I have been able to prepare perfectly pure 

 cobalt directly from the grey cobalt ore of Tunaberg, which is 

 an arsenio-sulphuret of cobalt, and also to detect small quantities 

 of cobalt in different samples of oxide of nickel. 



The preceding results embody the most definite conclusions of 

 an investigation of the ammoniacal salts of cobalt v/hich I have 

 had in hand for the last two or three years. M. Fremyhas also 

 lately announced that he is occupied with an extended inquiry 

 into the same class of compounds, respecting which he has pub* 



