OF ARTS AND SCIENCES. 37 



0-6031 gr. gave 0'0747 gr. cobalt = 12-38 per cent. 



0-7101 gr. gave 1-1252 gr. CrO^Ba= 67-71 per cent. (Chromium = 



52-2) 

 0-6295 gr. lost, at 105° C, 0-0077 gr. water=l-22 per cent; at 120° 



C, 0-0118 gr.=:l-87 per cent; and at 133° C, 0.0166 gr.= 



2-64 per cent. 



At 133° C. the salt was slightly decomposed. Between 133° and 

 145° C. it lost 4-46 per cent with partial decomposition. These anal 

 yses con-espoud to the formula Co^,(NH3)jQ(Cr20-)3-|-On2. 



The salt was dried for two weeks in pleno over sulphuric acid- In 

 pi'eparing nitrate of purpureocobalt by Mr. Mills's process, in which an 

 ammoniacal solution of cobaltic nitrate is oxidized by potassic dichro- 

 mate, I obtained, besides the nitrate, a large quantity of beautiful 

 orange-red crystalline scales, with gold reflections. The crystals were 

 easily purified by recrystallization. They are readily soluble in hot 

 water, and crystallize from the solution almost completely on cooling. 

 The formula of this salt is Co2(NH3)jo(Cr20-)3-{-50H2, as the follow- 

 ing analyses show : — 



0-6366 gr. gave 0-0735 gr. cobalt= 11-54 per cent. 

 0-6447 gr. gave 0-2888 gr. Cr03= 63-31 per cent Cr^O-. 

 0-1740 gr. gave, up to 139° C, 0-0125 gr. wafer = 7-19 per cent. 

 0-0800 gr. gave, up to 145° C, 0-0082 gr. water = 10-25 per cent. 

 Mean, 8*72 per cent. 



In the last water determination the salt was slightly decomposed. The 

 formula requires 



Calculated. Found. 



Cobalt, 2 11-48 11-54 



Cvfij 3 63-20 63-31 



Water 5 8.76 8-72 (mean.) 



The difference in appearance and in the number of atoms of water in 

 the dichromate of purpureocobalt may possibly arise from the fact that, 

 in one case, a solution of the nitrate of purpureocobalt was poured into 

 one of potassic dichromate in excess ; in the other, the nitrate was pre- 

 sented to the dichromate as fast as formed, — in some sense in the 

 nascent state. But it is singular that the two hydrates are not the 



