NICKEL AND COBALT. 175 



are best handled by employing the ratios between the carbon 

 dioxide and the metal. Accordingly ratios (1) and (2) give 

 a single value for Ni, and ratios (3) and (4) another. In all, 

 we have nine values for the atomic weight in question : 



From (i) and (2) Ni = 57.907, zh .0379 



" (3) " (4) "=57.926,^.0654 



" (6) " =57.884,^.0396 



" (7) " = 57.947, zb .0467 



" (11) " ^58.170,^3.0829 



" (5) "=58.607,^.0139 



" (9) " =58.634,^.0165 



" (8) " = 58.899, ± .0339 



" (10) " =59.120,^.0376 



General mean " = 58.547, rt .0089 



If = 16, Ni = 58.682. 



In the foregoing result it will be seen that the two sets of 

 figures due to Russell receive very great weight. This is 

 because the one set is referred directly to hydrogen, without 

 the intervention of the probable error of any other element ; 

 while the second set involves only the atomic weight of 

 oxygen, of which the probable error is small. As regards 

 accuracy of methods, however, and certainty concerning the 

 purity of material, Russell's work is no better than Schnei- 

 der's, and probably inferior to Lee's. Now values one to 

 five in the above table represent the tolerably concordant 

 results of Schneider, Lee, and Sommaruga. They, combined 

 by themselves, give a general mean of Ni = 57.928, ± .0215 ; 

 or, if = 16, of Ni = 58.062. This value, taking every- 

 thing into account, I cannot but regard as more likely to 

 prove correct than the larger mean deduced from all the 

 ratios. At all events, the atomic weight of nickel needs 

 further careful investigation. 



For cobalt these ratios are available : 



(l.) Per cent, of Co in CoC20^.2H20, 32.5555, zb .0149 



(2.) " CO2 from " 47.7475, zb .0213 



(3.) " Co in CoO, 78.592, zb .0023 



(4.) " " purpureo-cobalt chloride, 23.6045, zb .0018 



(5.) " " phenylammonium cobalticyanide, 1 1.8665, zb .0124 



(6.) " " ammonium " 21.943, zb .029 



