120 PROCEEDINGS OP THE AMERICAN ACADEMY. 



solved in liydrochloric acid, and the cobalt was partially freed from al- 

 kalis by precipitation as sulphide. The sulphide was then treated with 

 an excess of nitric acid and was heated upon the steam-bath until the re- 

 sulting cobalt nitrate was free from every trace of chlorine. 



In order to free the cobalt from alkalis it was next to be precipitated 

 by electrolysis-^ ; and at first we hoped to convert the electrolytic film 

 directly into bromide. However, on passing a galvanic current through 

 the ammoniacal nitrate, an oxy-amine was precipitated in such quantities 

 as to render the use of this electrolyte impossible. Experiments were 

 then made with the ammoniacal sulphate in order to find the conditions 

 most favorable for the deposition of a blight coherent film. The nega- 

 tive electrode was a platinum dish, the positive electrode a flat spiral of 

 platinum wire. The first attempts, which were made in a dish holding 

 250 cubic centimeters and having an area of 150 square centimeters, with 

 currents from 0.3 to 0.7 ampere, and with solutions of widely varying 

 strengths, resulted in the deposition of dull black films. With a current 

 strength of more than one ampere, liowever, the film was always compara 

 tively bright. It was found impracticable to use this film directly for 

 conversion into bromide on account of the difficulty of stripping the film 

 from the dish. Accordingly, after being thoroughly washed, it was dis- 

 solved in the dish by nitric acid which had been distilled with a platinum 

 condenser ; and when the excess of acid had been driven off by evapora- 

 tion, a slight excess of pure ammonia (made by distilling ordinary pure 

 ammonia into pure water contained in a platinum dish) was added. The 

 resulting green precipitate was washed several times by decantation with 

 the purest water, filtered on a washed filter, dried in a steam oven and 

 finally, after separation from tlie filter [)aper, ignited to the black oxide in 

 a platinum crucible. During the ignition this cruciblei was contained 

 in a porcelain crucible which was heated by an alcohol lamp. Pure 

 metallic cobalt was prepared from this oxide by reduction in a stream of 

 ammonia gas, generated in an apparatus composed entirely of glass and 

 dried by passage over stick-soda. Metallic cobalt is thus obtained in a 

 spongy form, well adapted for its following treatment with bromine, and 

 certainly free from all such impurities as alkalis and silica. The cobalt 

 bromide, prepared from this metal in the manner already described, was 

 labelled No. I. 



The second method of preparing pure cobalt was by precipitation as 

 one of its amines. The purpureo-chloride seemed best suited for the 

 present purpose, both on account of the ease with which it can be pre- 

 pared and because of its comparative insolubility in cold ammoniacal or 



