OF ARTS AND SCIENCES. 27 



7. I have rendered it, to say the least, extremely probable that there 

 are three distinct modifications of roseocobalt, yielding salts with 

 similar or identical empirical formulas, but differing in color and 

 solubility. Is it unreasonable to suppose that there may be a 

 fourth modification or possible variation in the arrangement of 

 the atoms constituting the molecule Co2(NH3)jq ? 



"With respect to the few known cases in which salts of roseocobalt 

 and purpureocobalt yield the same salts by double decomposition with 

 the same reagents, I have to say that there appears to be no reason 

 for doubtinij that in such cases there is a transformation of one modi- 

 ficatiou of the molecule Co2(NH3)jq to another, since we already 

 know that such transformation may be effected by heat alone. The 

 best instances of tliis transformation occur in the case of the reactions 

 of the two chlorides with potassic ferricyanide and cobalticyanide 

 mentioned by Genth and myself. 



LUTEOCOBALT. 



Rogojski * first noticed the existence of a salt of luteocobalt con- 

 taining both chlorine and sulphuric oxide, and having a formula which 

 we should now write Co2(NH3),2Clu-|-Co2(NH3)i.3(SOj3, but which 

 might also be written Co2(NH3),2(SO,).3Cl+Co,(NH3)„(SO,)Cl,. 

 In examining this salt, Genth and I found that the chloride and sul- 

 phate are capable of crystallizing together in all proportions. Dana 

 then showed that the two salts are isomorphous. Genth and I observed 

 further that mixtures of the chloride and sulphate gave peculiar crys- 

 talline salts with the chlorides of platinum and mercury. These we 

 naturally regarded simply as mixtures. Braun afterward obtained a 

 chloro-chromate which he regarded as a double salt, and which we 

 should now write Co2(NH3)]2(CrO^)^Cl2. The salt is easily formed 

 by mixing solutions of one molecule of the chloride and two of the 

 neutral chromate. As in the case of the corresponding salts of roseo- 

 cobalt, Krok t first showed that definite com]30unds are formed when 

 ammoniacal solutions of sulphate of luteocobalt are heated with iodine. 

 The resulting iodo-sulphate has the formula Co2(NHo),2(SOj2l>- By 

 the action of chlorine upon this salt the corresponding chloride is 

 formed, and this gives well-defined crystalline salts with platinic and 



* Compt. Rendus, xxxiv. 186. t Acta Univ. Lund. 1840. 



