Dr. E. J. Mills on certain Cobaltamines. 249 



gives off ammonia, and throws down a brown or black precipi- 

 tate. This precipitate is said by Fremy* to consist (if potassic 

 hydrate has been present) of hydrous dicobaltic teroxide ; and 

 Gibbs and Genthf assign to the product of the action of 

 water on the roseochloride the formula (Co 3 4 . 3H 2 0), whose 

 anomaly they do not leave unnoticed. It is difficult, indeed, to 

 reconcile this, 



2[6NH 3 . CoCl 3 ] + (3 + n)H 2 = Co 2 O 3 . nW 



+ 6(NH 3 .HC1) + 6NH 3 , 



which seems at first sight to be the only probable equation, with 

 the actual results of experiment. 



If the action be carried out in sealed tubes at 70°-100°, there 

 are obtained, after several hours, a colourless liquid and a black 

 precipitate. For complete decomposition, the chloride requires 

 from thirty to forty times its weight of water ; and the glass 

 tube, whether hard or soft, is very much corroded, often becom- 

 ing white and opake on cleaning. The black precipitate con- 

 tains cobaltic peroxide, silica, and water. 



0*1661 grm. of the above precipitate, dried over oil of vitriol 

 for a month, was treated with pure hydric sulphate, the liquid 

 evaporated, and the residue gently ignited. Water extracted 

 from this 0-2543 grm. cobaltic sulphate, which was weighed 

 after ignition in a crucible ; the insoluble portion consisted of 

 silica, and weighed 0*0226 grm. 



0*2018 grm. substance, heated below visible redness in a cur- 

 rent of carbonic dioxide, did not change its appearance, and lost 

 0*0167 grm. water. 



The percentage results are as follow : — 

 Found. 



Cobalt . 



A 



58*881/75*37 

 19-24 J \ 24-63 



(VO 5 . Co 3 4 . Co 2 3 . 

 75*00 73-77 71*43 



Oxygen . 



25-00 26-23 2857 



Silica 



13-61 





Water . 



8-27 





100-00 100*00 100-00 100-00 10000 



This result is in apparent disagreement with the numbers 

 obtained by Fremy and Gibbs and Genth ; for there does not 

 seem to be any reason why the employment of pressure, of po- 

 tassic hydrate, or of a penta- instead of the hexammoniochloride 

 should make any difference in the composition of the product in 

 question. Moreover the appearance of tricobaltic tetroxide or 

 of tetracobaltic pentoxide is quite unintelligible in the circum- 



* Pelouze et Fremy, vol. iii. 1. p. 566. 



t Silliman's American Journal, vol. xxiii. p. 256. 



