442 



SMITHSOXIAX MISCELLAXEOUS COLLECTIONS 



VOL. 54 



The second of these ratios was also stndiecl by Winkler, and the two 

 series combine as follows : 



Winkler 20.864, ± .0050 



Hempel and Thiele 20.556, ± .0043 



General mean 20.687, ± .0033 



Hempel and Thiele apply to it a correction for silver chloride retaiiird 

 in solution, but its amonnt is small and not altogether certain. For 

 present purposes the correction may be neglected. 



The atomic weight of nickel was determined by Eichards and Cush- 

 man ' from analyses of nickel bromide. This salt, as first prepared, con- 

 tained traces of oxide, which are to be deducted from the halide compound. 

 In a preliminary series of experiments the following figures were ob- 

 tained, representing vacuum weights : 



NiBr^. AgBr. Insoluble, mg. Ratio. 



2.26113 3.88769 3.22 58.161 



2.80668 4.82431 7.08 58.178 



1.41317 2.42880 3.05 58.184 



1.71759 2.95307 .88 58.163 



2.48565 4.273.57 5.24 58.163 



4.32997 7.44280 15.83 58.177 



2.18072 3.74856 .... 58.175 



Mean, 58.172, 



.002c 



In the second set of analyses, both ratios were determined, namely, 

 with silver and with silver bromide, by the standard methods. The 

 data follow : 



This value for the AgBr ratio, combined with the preliminary series, 

 gives a general mean of 55.1738, ±.0007. 



From Ag ratio, Xi = 58.661. 

 From AgBr ratio, ^1 = 58.661. 

 And Ag:Br:: 100: 74.082. 



' Proc. Amer. .\cad., 33, 97. 1897. 



