RICHARDS AND CUSHMAN. — ATOMIC WEIGHT OF NICKEL. 331 



scribed in sufficient detail in other papers,* so that no further words 

 need be wasted upon these points. It is almost unnecessary to state that 

 in these simple operations no loophole was left open through which an 

 error might creep in to destroy the value of the more dithcult under- 

 taking before us. We are indebted to the Cyrus M. Warren Fund for 

 Chemical Research in Harvard University for some of our more expen- 

 sive pieces of apparatus. 



Thb Method of Analysis. 



At first many attempts were made to determine nickel by electrolysis, 

 with the hope that nickelous bromide might be analyzed in this simple 

 and direct fashion. In order to test the metliod, weighed amounts of the 

 purest spongy metal were dissolved and reprecipitated electrolytically. 

 The spongy metal had been prepared by boiling the purest platinum- 

 made amraonio-nitrate with much water, igniting and reducing the pre- 

 cipitate with pure ammonia, and heating the metal in a vacuum. The 

 weight of nickel deposited by electrolysis always exceeded that of the 

 pure nickel taken, hence the electrolytic method was abandoned as un- 

 suitable for work of the highest accuracy. The excess of weight, which 

 was noticeable even when the film was heated to 120° before weighing, 

 and often exceeded two tenths of one per cent when it was dried at 50° 

 after the method of Winkle i",t was traced to inclusion of mother liquor 

 between the film and the dish, and to the probable presence of occluded 

 hydrogen in the nickel. t Since the deposit was beautifully metallic 

 and coherent in appearance, one might well have expected a better 

 result. It is possible that these observations may help to explain Wink- 

 ler's high values for the atomic weight of nickel and cobalt, since he 

 used the electrolytic method. On the other hand, the spongy metal 

 which had been used in our experiments was probably purer than that 

 prepared in any other way, for solid impurities had been rigorously 

 excluded, and the traces of gas present had been pumped out. 



These preliminary experiments showed that the best method of deter- 

 mining the amount of nickel in the bromide would be to reduce it in a 

 stream of hydrogen, provided that the reduction could be accomplished 

 without the loss of any of the bromide by volatilization. Following in 

 the footsteps of Mr. Baxter's work with cobalt, it was found that moist 



* These Proceedings, XXXIII 106 et seq. 

 t Zeitschr. Anorg. Chem., IV., 22. 



X Raoult, Compt. Rend., LXIX. 826 ; Bottger, Dingler's Polytech. Journ., 

 CCI. 80 (1871). 



