22 RESEARCHES ON THE 



which is common to nearly all of them, and which is very different from the com- 

 paratively dull red of the salts of Roseocobalt. They are in general somewhat less 

 soluble than the compounds of Roseocohalt, and crystallize, for the most part, in 

 well defined crystals. When neutral they have a purely saline, non-metallic taste. 



Heat readily decomposes the salts of this base ; the final products of the decom- 

 position are the same as in the case of the salts of Roseocobalt, but intermediate 

 products are often formed. The neutral solutions are readily decomposed by boil- 

 ing, the products of the decomposition being a black or dark-brown oxide of cobalt 

 and a salt of ammonium, free ammonia being given off. In some cases, however, 

 salts of Luteocobalt are intermediate products of this decomposition. 



All the salts of Purpureocobalt by long boiling with an excess of chlorhydric 

 acid yield the chloride. 



CHLORIDE OP PURPUREOCOBALT. 



The substance which we shall describe under the name of chloride of Purpureo- 

 cobalt is the same as that to which Fremy gave the name of chloride of Roseo- 

 cobaltiaque. In the course of our investigations it at length became clear that, under 

 the name of salts of Roseocobalt, the compounds of two perfectly distinct bases have 

 hitherto been confounded. It became, therefore, necessary to devise a new name. 

 The purple color of the salts which correspond to the chloride now to be de- 

 scribed, led us to adopt the name of Purpureocobalt for the radical of these salts, 

 as more appropriate than Roseocobalt, which we have retained for most of the salts 

 to which it was originally applied. Such a change is to be regretted; it could not, 

 however, have been avoided, without an introduction of two entirely new names. 



We have already stated that the chloride of Purpureocobalt is often a product 

 of the direct oxidation of an ammoniacal solution of the chloride of cobalt exposed 

 to the air. In these cases it is sometimes mixed with chloride of Roseocobalt, and 

 sometimes forms the entire product of the oxidation. We believe that the tempe- 

 rature at which the process of oxidation goes on is the condition which determines 

 the character and the amount of the chloride which is formed during the oxidation. 

 The chloride of Roseocobalt is, in our view, the first product of the oxidation 

 under all circumstances. At a moderately high temperature, however, the chloride 

 passes as fast as it is formed into chloride of Purpureocobalt, which may thus be 

 the only final product of the oxidation, or may be mixed with variable proportions 

 of the chloride of Roseocobalt. 



In one experiment made during the summer season, and in which chemically 

 pure chloride of cobalt and ammonia were employed, the process of oxidation 

 went on very slowly, without the precipitation of any trace of sesquioxide of 

 cobalt. The liquid had a dull purple color, and gave with reagents no precipitates 

 or reactions to indicate the presence of chloride or oxide of Purpureocobalt or 

 Roseocobalt. On boiling with chlorhydric acid, however, the chloride of Purpureo- 

 cobalt was thrown down in abundance, and no other substance could be detected 

 in the supernatant liquid. In this case the oxidized liquid gave no precipitate with 



