SOLUTIONS. MACGREGOR. 77 



It will thus be seen that Schrader's coefficients differ from 

 mine by various amounts up to about 10 per cent., being in 

 most cases smaller, but in some greater. It should be noted also, 

 that while my coefficients are in all cases greater for the iodide 

 than for the chloride, Schrader's are in three out of four cases 

 greater for the chloride than the iodide. 



(2) Kay 1 has recently employed an approximation method 

 based upon one previously used by Arrhenius. As a first approxi- 

 mation the ionization coefficients of the two electrolytes in a com- 

 plex solution were assumed to be the same as they would be in 

 simple solutions of concentration equal to the total concentration 

 of the complex solution. The total concentration of ions of the 

 complex solution (equal to the regional ionic concentrations of 

 the respective electrolytes) was then calculated, and gave a first 

 approximation to the value of the regional ionic concentration. 

 From curves plotted with ionization coefficients of simple solu- 

 tions as ordinates and ionic concentrations of the same solutions 

 as abscissae, the values of the ionization coefficients correspond- 

 ing to the first approximation to the regional ionic concentra- 

 tions were read off and formed second approximations to the 

 ionization coefficients required. Calculation of the second 

 approximation to the total ionic concentration and a repetition 

 ot the above procedure gave a third approximation to the 

 ionization coefficients In dealing with solutions containing 

 sulphuric acid and a neutral sulphate, he found that in general 

 the second approximation was so close to the first that a third 

 was not necessary ; and he seems to have found that the third 

 in no case differed appreciably from the second. 



As, in the case of electrolytes with a common ion, the varia- 

 tion of ionizatiou with dilution is in general not very different, 

 this method may be expected to give very closely approximate 

 results. By way of a test I have made a few determinations 

 for solutions containing zinc and potassium sulphates, using 

 Kohlrausch's conductivity data. The result is shown in the 

 following table in which zinc sulphate is indicated by 1 and 



l Proc. R. S. Edin., 22, 502, 1898-99. 



