270 Prof. Liebig on Cyanide of Potassium. 



this into (1) pure oxide of nickel, or the carbonate, which 

 is washed on a filter, dried and weighed, and (2) an alka- 

 line liquid wiiich contains the whole of the cobalt. The lat- 

 ter is evaporated to dryness, some nitre being added to it, and 

 the residuum ignited. On being treated with water, the oxide 

 of cobalt remains behind. 



This method is applicable in all analyses of cobalt ores in 

 ■which the amount of cobalt predominates. For nickel ores, 

 in which the quantity of cobalt amounts merely to a minimum, 

 the following precaution must be attended to: — a somewhat 

 considerable excess of muriatic acid must be taken to preci- 

 pitate the cyanides dissolved in the cyanide of potassium, and 

 the mixture must be kept boiling at least one hour. 



The precipitate contains in this case cyanide of nickel inter- 

 mixed, which is decomposed by potash into cyanide of potas- 

 sium and oxide of nickel ; this cyanide of potassium retains 

 however another portion of nickel in solution. On boiling 

 the precipitate with muriatic acid, the cyanide of nickel is de- 

 composed into chloride of nickel and hydrocyanic acid, which 

 last is removed by boiling, and no longer prevents the entire 

 precipitation of the nickel. The cobalti-cyanide of nickel is 

 not attacked by boiling hydrochloric acid, so that a complete 

 solution cannot be expected when any quantity of cobalt is 

 present. When the smell of hydrocyanic acid is no longer 

 perceptible, the boiling has continued long enough. 



Experiments made to separate the solution of the two cya- 

 nides in cyanide of potassium by boiling v.ith peroxide of 

 mercury, gave less certain results. The following points must 

 be attended to in this process: — as the cyanide of potassium 

 used contains a certain quantity of cyanate of potash, a por- 

 tion of an ammoniacal salt originates on its decomposition by 

 a mineral acid ; accordingly, after boiling and the addition of 

 caustic potash, ammonia is set free, which retains some oxide 

 of nickel in solution. This nickel is entirely thrown down by 

 boiling for a few minutes, or by a larger addition of potash. 



The same method may be followed for the separation of 

 manganese from cobalt, only in this case a complete solution 

 of the precipitate produced by the addition of the cyanide of 

 potassium to the mixture of the two metallic salts cannot be 

 expected; the greater portion of the protocyanide of manga- 

 nese remains undissolved. The residue is filtered, and the 

 liquid treated as if cobalt and nickel were to be separated. 



The cyanitle of potassium is not less applicable with advan- 

 tage for the separation of the oxide of chromium from prot- 

 oxide of iron. A mixture of the two is previously saturated 

 with ulphuretted hydrogen, to be certain that the iron is con- 



