﻿630 Mr. Bernard Cavanagh on 



(and so permit accurate evaluation of a So , a , etc.) we shall 

 require an entirely unsolvated solute (such as an inert gas in 

 water), for which t s will be (— Jc^), and so on *. 



That only the solute s need be unsolvated is an important 

 point, since the solubility of the more inert gases (for 

 instance) is small. 



Physical Significance and Limitations of the 

 Quantities. 



The quantities a , a 5o , ft, etc., though more difficult to 

 estimate than the solvent-constants, are not, of course, merely 

 arbitrary parameters by any means 



(49), for instance, shows that if only one solvate is formed 

 by the solute s, then 



U-&)=M S , ..... (87) 



so that the composition of the solvate is determinable from 

 u So and ft, and, of course, M s will have to be an integral 

 number of times the molecular weight of the solvent. 

 Similarly, if two solvates are formed and the amount of (s) 

 remaining unsolvated is very small, then, on the assumption 

 of perfect solution a g0 , ft, and <y s will suffice to determine the 

 composition of both solvates, according to the analysis men- 

 tioned, but not inserted, above. If the unsolvated residue is 

 not negligible, 8 S will be required. 



A more general limitation is that so long as only "positive"" 

 solvation is in question (see first pages of this paper) a So 

 cannot be negative, and ft cannot be positive, the latter 

 being demonstrable from " Le Chatelier's Principle. " 



In other words, t s may be a fairly large positive quantity, 

 but only a quite small negative quantity, ranging, in fact, 

 from — 1% to the highest probable value for M s . Similarly, 

 h s can range from (very nearly) y^i 2 to the highest probable 

 value for £M S 2 . When the solvent is water a x may be 60, 

 possibly even higher near 0° C, and M s might easily be 180 

 (decahydrate), while, of course, hydrates much higher than 

 this are known even in the solid state with low aqueous 

 vapour-pressure, so that a value of M s as high as 300 would 

 not be improbable in certain cases. 



* Compare here 2nd paper loc. cit., p. 241. 



