AMMONIA-COBALT BASES. 23 



chlorhydric acid in the cold, but the cold solution, after some hours standing, 

 deposited distinct crystals of chloride of Purpureocobalt. We believe that in this 

 case a combination of the oxide and chloride existed in the solution, so that, as we 

 have already suggested in speaking of the chloride of Roseocobalt, the oxidation 

 itself would be expressed by the equation 



6C0CI + 15NH 3 + 30= 5NH,.CoA+ 2 (5NH 3 .Co 2 Cl 3 ) . 

 If we admit that the oxide and chloride of Purpureocobalt as thus formed are 

 actually in combination so as to form a sort of oxychloride, and are not merely 

 mechanically mixed, we may perhaps explain, why no precipitates of salts of Pur- 

 pureocobalt are obtained in the oxidized liquid, since the reagents added might not 

 be able to overcome the affinity between the oxide and chloride. It is easy to see 

 that by boiling with chlorhydric acid, the combination of oxide and chloride will 

 give the chloride alone, since we may have the equation 



5NH 3 .Co 2 3 + 2 (5NH 3 .Co 2 Cl 3 ) + 3HC1= 3 (5NH 3 .Co 2 Cl 3 ) + 3H0. 



We have already mentioned that the chloride of Roseocobalt is readily converted 

 into chloride of Purpureocobalt, by boiling with chlorhydric acid, or even by 

 gently heating its solution. This circumstance explains why only the chloride of 

 Purpureocobalt is obtained by boiling a completely oxidized ammoniacal solution 

 of chloride of cobalt, even when this solution contains only Roseocobalt, or a mix- 

 ture of Roseocobalt and Purpureocobalt. It also enables us to understand why all 

 writers upon the ammonia-cobalt bases, up to the period of the present investiga- 

 tion, entirely overlooked the chloride of Roseocobalt, and consequently described 

 the salts of that base, as if they corresponded to, and contained the same radical 

 as chloride of Purpureocobalt. 



It is not necessary to add chloride of ammonium to the ammoniacal solution of 

 chloride of cobalt, in preparing the chloride of Purpureocobalt. It is, however, ad- 

 vantageous to do so, because, the oxide of Purpureocobalt is converted by it as fast 

 as formed into the chloride, and the formation of the oxychloride is prevented. To 

 prepare the chloride of Purpureocobalt in a pure state from the oxidized liquid, it 

 is only necessary to boil this with an excess of chlorhydric acid. A crimson 

 powder is thrown down, while the supernatant liquid becomes nearly colorless, 

 provided, at least, that a pure salt of cobalt was employed, and that the oxidation 

 was complete. The mother liquor is to be poured off and the precipitate dissolved 

 in a large quantity of boiling water, to which enough chlorhydric acid has been 

 added to render the solution distinctly acid. On cooling, the solution gives small 

 but beautiful crystals of the chloride, almost perfectly pure. A second crystalliza- 

 tion usually removes all traces of impurity. It is not, however, necessary to use a 

 pure chloride of cobalt in preparing the chloride of Purpureocobalt. Any com- 

 mercial oxide will answer, even when arsenic, nickel, iron, &c, are present. On the 

 other hand, it is easy to prepare a perfectly pure chloride of cobalt, by heating the 

 chloride of Purpureocobalt in a porcelain crucible until vapors of ammonia and 

 chloride of ammonium cease to be given off. The pure chloride of cobalt thus 

 obtained, is remarkable for its beauty of color, the anhydrous chloride forming pale 



