AMMONIA-COBALT BASES. 61 



still greater in the case of the compounds of Xanthocobalt. We have, therefore, 

 no hesitation in rejecting Rogojski's theory as too limited in its application. 



We consider the ammonia-cobalt bases as conjugate compounds of sesquioxide, 

 sesquichloride, &c, of cobalt, the five or six equivalents of ammonia, or of ammonia 

 and deutoxide of nitrogen, forming the conjunct, and serving to give to the sesqui- 

 compound of cobalt the degree of stability which it possesses in this class of bodies. 

 Accordingly we should prefer to write the formula of chloride of Luteocobalt 



GNlCCo 3 .Cl 3 



employing the connecting circumflex, as Kolbe has suggested, as a symbol of conju- 

 gation. Adopting this view, we have the following conjugate radicals, which we 

 assume as existing in the ammonia-cobalt bases precisely in the same sense in 

 which we assume Co 2 as existing in the sesquichloride and sesquioxide of cobalt. 



Koseocobalt 

 Purpureocobalt 

 Luteocobalt 

 Xanthocobalt 



5NH,.Co 2 

 5NlCCo 2 

 6NlCCo 2 



noTsniCcos. 



We may, however, remark that while this view offers a satisfactory explanation 

 of the fact that two of the ammonia-cobalt bases are triacid, forming in all cases 

 neutral compounds with three equivalents of acid, two are biacid bases, and of 

 these latter, one, namely Purpureocobalt, forms acid compounds with four equiva- 

 lents of acid. 



In these cases we meet with the same difficulty which occurs with the ordinary 

 salts of sesquioxides; thus sesquioxide of iron may unite with one, two, or three 

 equivalents of acid, though there appear to be three in the neutral salts. While 

 some chemists assume that salts of sesquioxide of iron containing one or two 

 equivalents of acid are basic, others consider the different equivalents of oxygen 

 as united in different modes, so that the oxide may be, according to circumstances, 



Fe 2 3 

 Fe 2 2 .0 



Fe 2 0.0 2 . 



Peligot's theory of the constitution of sesquioxide of uranium is another case in 

 point ; as this oxide unites with but a single equivalent of acid, it may be con- 

 sidered, as Peligot has shown, as the oxide of a radical, U 2 2 .0, so that in this case 

 also, the rule that a base unites with as many equivalents of acid as the base itself 

 contains equivalents of oxygen is satisfied. 



To explain the biacid character of two of the ammonia-cobalt bases, we may 

 therefore assume in them the existence of secondary radicals containing oxygen. 

 Thus in Purpureocobalt and Xanthocobalt the primary radicals are 



5NfCCo 2 

 N075NH7Co 2 ; 



