OF ARTS AND SCIENCES. 41 



strong argument in favor of doubling the formula given by Fremy, or 

 at least of regarding the cobalt as hexatomic in this salt. 



The above results appear to me to render it probable that the brown 

 solution formed bv the oxidation of an ammoniacal solution of cobaltic 

 chloride contains chiefly 



Co,(NH3)j„.0,.Cl,. 



According to Rose, the bi'own solution gives off oxygen by long 

 contact with the air, forming the well-known red liquid which yields, 

 l)j boiling with sal ammoniac or chlorhydric acid, chloride of purpureo- 

 cobalt. If we suppose that six atoms of oxygen are given off from 

 two molecules of the oxychloride, the salt, — 



Co,(NH3),„OCl„ 



will remain in solution, and it is easy to see that this, by boiling with 

 chlorhydric acid or ammonic chloride, will yield chloride of purpureo 

 cobalt, since we have — 



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



This is precisely the view taken by Genth and myself as regards the 

 nature of the red solution, though we did not trace its oi-igin to the 

 brown oxychloride. Genth and 1 stated in our paper that tlie pres- 

 ence of ammonic chloride was not necessary for the formation of 

 chloride of roseocobalt by the oxidation of an ammoniacal solution of 

 cobaltic chloride. We did not state, as Rose * appears to have under- 

 stood us, that it is a matter of iudiifereuce whether ammonic cliloride 

 is present or not. Rose has shown that in the presence of this salt a 

 much larger relative amount of chloride of purpureocobalt is formed. 

 Thus, as a mean of eight experiments, he obtained from one hundred 

 grams cobaltic chloride, oxidized in presence of sal ammoniac, 134-6 

 grams chloride of purpureocobalt, and 1 2*1 2 grams chloride of luteo- 

 cobalt. When no sal ammoniac was present, he obtained, as a mean 

 of eight experiments, 90*66% chloride of purpureocobalt, and 16-82 

 grams chloride of luteocobalt. Rose's results in no way disprove the 

 existence in the oxidized solution, after giving oiF oxygen to the air, of 

 the oxychloride Co2(NH3)i„.O.Cl^ ; and this view, which is perfectly 

 consistent with the facts, still gives the simplest exjilanation of them. 



Genth and I always obtained the largest relative quantity of luteo- 

 cobalt when the solution exposed to the air contained cobaltic chloride 



* Log. cit. p. 76. 



