88 PROCEEDINGS OF THE AMERICAN ACADEMY 



same after reci'ystallization. A solution of potassic dichromate gives, 

 with one of chloride of purpureocobalt, a dark red crystalline precipi- 

 tate, the analyses of which pointed to a mixture of Co2(NH3)jQ(Cr20-)3 

 and Co2(NH3)jQ.CI,(Cr20.)2. I did not succeed in obtaining the basic 

 dichromate Co^(NH„)ju.O.(Cr20-)2. In all the chromates of the 

 cobaltamines which I have studied, the direct determination of the 

 water of crystallization has been effected with peculiar difficulty, in 

 consequence of the tenacity with which these salts retain water up to 

 temperatures very near to those at which mcipient decomposition 

 occurs. 



{To be continued.) 



