﻿628 Mr. Bernard Cavanagh on 



limitations imposed upon J and Gr by their respective physical 

 significances. 



It seems reasonable, however, that when all possible has 

 been done in this direction, if doubt still remains as to 

 whether all or some portion of the observed effects should 

 be assigned to J or to Gf, a " casting-vote " should be 

 given (temporarily) in favour of J, since J does not re- 

 present real departure from perfect behaviour, properly 

 defined. 



It is possible that G may be completely determined on 

 theoretical grounds alone in certain cases, the nearest 

 approach to this, so far, being in the case of electrolytes, or 

 rather of a mixture of ions all of the same valency, for 

 which the (probable) form and order of magnitude of Gr, in 

 (sufficiently) dilute solution, was obtained in the first of 

 these papers, on the basis of a calculation of Milner's. 



With regard to J, besides the fact that it takes the form 

 of a series of ascending integral powers of C, the total solute 

 concentration, a good deal can be deduced from the form of 

 the coefficients £/, t u etc., and the physical significance 

 of the three series of quantities, the " solvent-constants " 

 (a u a 2 , a 3 ), the a-series (« , j3, y, E), and the a s -series (a So , 



fis, Js, $ s ). 



The Form of the Coefficients t h t{, etc. 



(59), (61), (63) , (64) give us these coefficients as functions, 

 at first sight rather complex, of these three series of quantities, 

 but for practical purposes this complexity is only apparent, 

 for each successive term of the J s -series introduces just one 

 further term from each of the three series of physically 

 significant quantities, and introduces it in the first degree, 

 while in J M the same is the case except that the a 5 -series 

 does not enter. 



These coefficients can be obtained in a much simpler and 

 more practical form by introducing the quantities t s , Ji s , k SJ .. ., 

 and the corresponding mean quantities t, h, k, ..., such that 



t s — a So — t — a =—±a 1 "^ 



K — f3 s = h — fi = ±a 2 — \a? ^ ^ a x 2 

 k s — <y s — ±a x li s = 7c — y-^aili 



h (85) 



