104 PROCEEDINGS OF THE AMERICAN ACADEMY. 



tion of our sample. The nickel was twice fractionally precipitated as 

 hydroxide by means of pure sodic hydroxide, the mixture being thor- 

 oughly boiled. As far as our test could show it, all the cobalt remained 

 in the filtrate, the last precipitate being contaminated only with a small 

 amount of alkali. 



When ammonia is added to a solution of nickelous bromide, a beautiful 

 violet crystalline compound is formed, having the formula KiBro . 6NH3, 

 according to our analyses as well as those of Kammelsberg.* Since this 

 compound is characteristic of nickel, and similar compounds are not 

 formed by cobalt or most other metals under similar circumstances, and 

 since it is soluble in strong hot ammonia water, but almost insol- 

 uble in cold ammonia, it affords a very convenient and effectual means 

 of purifying nickel preparations. Our purified oxide was hence dis- 

 solved in pure hydrobromic acid, ammonia was added in excess, and 

 the mixture, contained in a platinum flask, was cooled to zero. Ihe 

 beautiful purple precipitate was collected upon pure filter paper and was 

 washed with strong ammonia. All the material used in the analyses 

 was passed through this treatment at least once, although the various 

 samples were subsequently subjected to different methods of further 

 treatment which will be described in each case. 



The violet compound made from our first sample of purified nickel was 

 treated with an excess of water and boiled in a platinum dish, a proceed- 

 ing which completely precipitated the nickel as hydroxide. The greenish 

 mass was thoroughly washed, and was then dissolved in hydrobromic acid. 

 The nickelous bromide thus obtained was dried in a vacuum desiccator 

 over dry soda ; but even after this treatment it was found to have retained 

 considerable quantities of water, an impurity which greatly increased the 

 difficulties of sublimation. Hence this sample, numbered I., was used 

 only for two preliminary analyses. 



Another portion of this same sample of the violet compound was re- 

 crystallized several times in succession by cooling its hot ammoniacal 

 solution. The resulting magnificent crystals were dissolved in water and 

 'the solution was boiled to drive off the ammonia. The precipitation 

 being thus accom|)lished the basic hydroxide was collected, carefully 

 washed, dried, and ignited over an alcohol flame in a porcelain vessel. 

 The resulting nickel oxide was reduced to the metal by igniting this 

 material, held by a porcelain boat in a combustion tube through which a 

 current of pure dry ammonia gas was passing. The spongy nickel thus 



* Pogg, Anal., LV. 243. 



