249 



Aqueous solutions of NiiNOj)^. 



Average : 445.4 



16.06 



Before we draw conclusions from the results obtained their accuracy 

 must be tested. The error in the value used for Xw"to is not greater 

 than 37oo at most"), from tliese, 2"/„ are a consequence of the error 

 in tlie measurement of the normal electromagnetic field. The propor- 

 tion of the susceptibility of the solution to that of the water, how- 

 ever, is independent of the error in the field ; as in this research 

 the proportion of the susceptibility of the solution to that of the 

 water has really been determined, it is only the inaccuracy in the 

 determination of the ascension, which was l"/»» at most, which con- 

 sists in that proportion obtained, while in the final results the error 

 of "/.water remains as well. From the results of the analyses it is 

 evident, that the error in the concentration always remains below lYoo- 



Thus within the limits of experimental accuracy the value of •/" 

 and also the number of magnetons seems to be independent of the 

 nature of the salt and of the concentration of the solution. This 

 result agrees with that of Cabrera''), who from his research about 

 the aqueous solutions of nickelsalts also concluded the atomsuscepti- 

 bility to be independent of the concentration and the nature of the 

 salt. For the number of magnetons of tlie nickelatom in solutions 

 of NiSO,,, NiCI,, and Ni(NOj)j he respectively gives the numbers 

 16,07, J 6,03 and 16,02. 



The number of magnetons of nickel in dissolved nickelsalts thus 

 seems to be a whole number within the limits of experimental 

 accuracy and as such supports the magneton theory. 



§ 3. Then the ammoniacal solutions of nickelsalts were investigated. 



') The analyses gave 37,164% and 37.131"/,,, the former value has been talien, 

 because the second is less reliable. 

 -) A. PicCARD, ibid. p. 53. 

 3) Cabreka, Moles et Guzman, Arch, de Geneve T. XXXVll, p. 330, 1914. 



