﻿634 Mr. Bernard Cavanagh on 



It was shown in the first of these papers that in the case 

 of a mixture of ions, all of the same valency, theory alone is 

 able to predict, for (sufficiently) dilute solution, a simple 

 limiting form 



G=-R<z> / a 3 ' 2 , ..... (99) 



where 2C, is the total ion-concentration *. The range of 

 validity of the "point-charge" assumption, on which this 

 form depends, should be greatest in the case of the univalent 

 ions of simple structure, and might here extend to, or even 

 above, tenth-normal concentration. 



Assuming this, and supposing that, while other solutes 

 (typified by s) besides the ions (typified hjj) are present, 

 there are no other " true " general terms at these low 

 concentrations, we have 



^ = </> + R[C-Xc s logc,-^log Cj ]+J-R</)'a 3 / 2 , (100) 



Mo 



or 



~ i n 3 / 2 



log 7,= "g J* 



log7/=-|j,-r-#'0^ 2 



where %Cj is 2G\-, and C is 2c s + 1cj. 



At the concentrations considered, J, of course, reduces to 

 one term, and we have 



i 0g7s = (Ut s )c, ( ' ' ' (102) 



\ gy j = (t + tj)G+U'G i ^,\ 



In practice, log y,- will always* be a mean quantity for the 

 two ions o£ the electrolyte j separated, so that 



log 7i ( = iIog 7j ,. 7ia ) = (i+it je+ i tja )G+WG^, (103) 



where tj c and t ja are peculiar to cation and anion respectively. 

 For a single binary electrolyte by itself 



log?., = 2(t jc + t Ja )G i + i<t>' C,-w . 



<p-i =2(^ + ^)C,+ #'C," 2 )' ' ' • 



* See equations (88), (89), (92) of 1st paper. 



