42 PROCEEDINGS OF THE AMERICAN ACADEMY 



and sulphate, ammonia and coarsely powdered amraonic chloride iii 

 large excess. I have obtained the same results iu frequent rei^etitions 

 of the process. According to Rose, this result is due not to the forma- 

 tion of a greater amount of luteocobalt salts iu consequence of the 

 presence of sal ammoniac, but to the fact that the precipitation of the 

 sulphato-chloride, Co2(NH3)j2(SO^)2Cl2.. as fast as it is formed, prevents 

 its further decomj^osition. Rose's own experiments, cited above, show 

 that the larger quantity of chloride of luteocobalt was formed, when no 

 sal ammoniac was present, when only cobaltic chloride was employed. 

 He sugi^ests that the quantity of luteocobalt formed depends upon the 

 longer action of a concentrated solution of ammonia upon the oxidized 

 solution. If this be the case, the luteocobalt must be formed by the 

 direct oxidation of the solution, and not by the decomposition of the 

 brown salt, whatever that may prove to be. Yet Rose assumes that all 

 the other cobaltamines are formed by the decompcsition of this brown 

 salt. 



The results of Fremy, in connection with those of Rose, aj^pear to 

 show that in the oxidation of an ammoniacal solution of cobaltic 

 chloride at least two brown salts are formed. These are the chloride 

 of oxy-cobaltia, Co2(NH,J]|,O^Cl^, and the chloride of fuscocobalt 

 (octamin oxy-chloride), Co.,(NH3)yOCl^, the last named being in rela- 

 tively small quantity. By the action of chlorhydric acid upon each of 

 these salts the chlorides of luteocobalt and purjDureocobalt are formed. 

 This appears in the case of the octamin salt ft-om the experiments of 

 Fremy,* SchitF, f and Braun t ; in the case of the salt of oxy-cobaltia 

 (tetroxy-decamin), from those of F. Rose. Iu Rose's experiments 

 relatively small quantities of the hexamin and octamin chlorides, 

 Co2(NH.,)yCl|;, and Co._,(XH3)yClg, were always found in the mother 

 liquor after the precipitation of the chlorides of purpureo- and luteo- 

 cobalt by chloi-hydric acid. It seems at least probable that the octa- 

 min chloride is formed from the brown oxy-chloride or fuscocobalt 

 salt of Fremy, since we may with great probability expect the reac- 

 tion expressed by the equation, — 



Co,(NH3),OCl,+2HCl = Co,)NH3),Cl,+OII,. 



Rose also obtained in his experiments a nearly black crystalline salt, 

 the analyses of which, however, did not lead to any rational formida. lie 



* Ann. de Cliiniie et de Physique [3] T. xxxv. 28G. 

 t Ann. der Clicniie und Pliarmacie, cxxi. 124, cxxiii. 1. 

 t Ann. der Chemie und Pharmacie, cxl. ii., p. 60. 



