60 



PROCEEDINGS OF THE AMERICAN ACADEMY 



Pt ^ 



CI 



a — a — CI 



a—a— CI 

 CI 



and which explains the formation of the different salts at least equally- 

 well ? In this case, as in the cases of the platinamines generally, we 

 have precisely the same difficulties which meet us in applying the 

 theory to the cobaltamines. We reason throughout from perfectly ar- 

 bitrary fundamental assumptions. Our only fixed points are the atomi- 

 cities of platinum and cobalt. All else is purely speculative. In the 

 present state of our knowledge we are not able to say whether a chain 

 of atoms of ammonia, like — NH^ — NHg — NH3 — , has more or less 

 powerful affinities than the single divalent atom, — NH, — , or even 

 whether, — NHg — CI, is more or less chlorous than a single chlorine 

 atom. But these points are of fundamental importance in the applica- 

 tion of the theory of atomicities to the metal-ammonias. Certainly 

 the great majority of chemists maintain that there is a perfect equiva- 

 lence of value in the units of affinity of the different elements, while 

 admitting that in complex molecules, as in the benzol ring, positional 

 differences may result from peculiarities of structure. 



Blomstrand, on the other hand, maintains not merely that the four 

 units of affinity in tetratomic platinum are qualitatively different two 

 by two, but even that it is possible to determine in the case of which 

 pair the most powerful affinities are exerted. Thus he asserts that 

 if we write for platinum, — 



r« 



a 

 b 

 d 



the two "points of attack," a and b, act differently from c and d. 

 With logical consistency he extends this view to nitrogen, cobalt, and 

 carbon, standing, so far at least as carbon is concerned, wholly alone, I 

 believe, in opinion, the question of a difference between the four units 

 of affinity in carbon having been long since discussed, and by common 

 consent decided in favor of their perfect equivalence. 



While then I write the structural formulas of the chlorides of pur- 

 pureocobalt and of roseocobalt respectively : — 



Pt 



Co., \ 



a— CI 

 a— a— CI 

 a—a- CI 

 a— a— CI 

 a— a— CI 

 a— CI 

 Chloride of purpureocobalt. 



Co 



< 



CI 



a— a— CI 



:z:z:zg+20H, 



a—a— Cl 

 CI 



Chloride of roseocobalt. 



