OF ARTS AND SCIKNCKS. 



47 



or to RcX.^Y^.Z.j. Now the salts of oitlier of tlicse types may Ix' <'(|iially 

 well referred to either of the striietural formulas jriveu alfovc 'I'd 

 which of thfin are we to refer salts of rose(Jcol)alt, and to u liicli >alts of 

 pnri)ureocobalt ? 



Blomstraiul gets over tlic dilliciilly in part hy the slioi-t and easy 

 method of declaring that chloride of roseocobalt is only the hy- 

 drate of chloride of pnrpnreocohalt. I think I have shown ccn- 

 cliisivi'ly that this view is wholly untenahle. lint, even if we admit its 

 correctness, to which of the four structural formulas shall we assign 

 chloride of ])urpureoct)balt alone, since all four explain its relations to 

 other chlorides, and its products of replacement equally well ? 



The case is the same with tlie other cobaltamines, though in these 

 we have no isomerisms to explain, in the present state of our 

 knowledge at least. The dodecamin and octarain series exhibit the 

 two types LcX^, OcX^, and LcX^Yj, OcX^Yj, with perfect distinctness ; 

 but in additiou we have in the case of the dodecamin series the type, 

 LcX.Y.., as, for instance, in the phosphate described by Brauu, 

 Co2(NHg),^,(P20^)-f-80H2, and perhaps also in the pyrophosphate 

 Co.,(NH..),2P^Oj.,-{-*jC)H2, though this salt may be referred to the type, 

 LcX,., if we consider (PjOj.,) as a hexatomic complex. Luteocobait 

 may therefore be saturated, if I may so speak, by ones as in LcCl,., 

 by twos as in Lc(S0^)^l2, and by threes as in Lc(PO^)„. It would 

 seem as if we in this manner arrive most naturally at the structural 

 formulas for the dodecamin series, — 



Co., i 



f a— a— CI 

 a— a— CI 

 a— a— CI 

 a— a— CI 

 a— a— CI 

 a— a— CI 



Co., 



a — a — O 

 a—a— O 

 a — a — 1 

 a— a— I 

 a— a— O 

 a — a- 



-O 



>S0., 



>S0., 



Co., 



a— a— O ) 



a— a— O y PO 



a—a— O ) 

 f a— a— () ) 

 I a— a— O PO 

 [ a— a— ) 



in which the six ammonia groups have equal weights or qualitative 

 values, so that the six units of affinity of Coj must also be of the same 

 kind and the same iuten ity. But, if we examine the list of salts of 

 roseocobalt which I have given, we find that Krok has described a 

 salt with the f .rn.ula, — 



Co,(Nn3),,(N03)3Cl3+3HgCl, ; 



