RICHARDS AND CUSHMAN. — ATOMIC WEIGHT OP NICKEL. 107 



was used. For some of the platinum and other apparatus we are in- 

 debted to the Cyrus M. Warren Fund for chemical research in Harvard 

 University. 



The Method of Analysis. 



Turning now to the method of analyzing the carefully prepared nick- 

 elous bromide, it is obvious that the first point to be considered is the 

 accurate determination of the weight of the salt to be analyzed. This 

 process was effected by means of apparatus similar to that devised for the 

 drying and weighing of magnesic chloride, and described in a recent paper 

 by Mr. H. G. Parker and one of us,* upon the atomic weight of magnesium. 

 In this apparatus, constructed wholly of glass by Mr. Baxter, the bro- 

 mide under consideration, contained in a platinum boat, was ignited at 

 about 400° in a stream of mixed nitrogen and hydrobromic acid until 



Fig. 2. Bottling Apparatus, Horizontal Section. 



A = weighing bottle. B = stopper of bottle. CC = hard glass tube. 

 D = Platinum boat containing nickelous bromide. 



constant in weight. It was then allowed to cool in a stream of pure dry 

 nitrogen ; and when cool it was pushed in pure dry air into its weighing 

 bottle, which was immediately closed by a mechanical device. In this 

 fashion it is possible to dry and weigh accurately the most hygroscopic of 

 substances ; and repeated ignitions of the same specimen have shown that 

 perfect constancy in weight may thus be obtained. It is hard to believe 

 that any water is retained by nickelous bromide at 400° ; and certainly 

 none could be absorbed during the cooling, for the whole apparatus was 

 shut off from the outside air, and all the gases admitted were first passed 

 through phosphoric oxide. f 



The bromide in question was then weighed by substitution, using as 

 the tare to be substituted a weighing bottle precisely like the one con- 

 taining the platinum boat and substance. In this way alone can the 

 weight of a large bottle be determined within the fraction of a tenth of a 



* These Proceedings, XXXII. 58. 



t A detailed description of this apparatus will be given in the following paper, 

 upon Cobalt. 



