﻿640 On Molecular Thermodynamics. 



data will determine (differences inj Rlogc s — J fi , that is, 



c/ 



Rlogc/— J/, In the former, therefore, -^ as well as C 



is a measured quantity, and when obtained (graphically) as 

 a series of integral powers of C gives the quantities t, A, . .., 

 directly. 



But without (and independently of the precision of) this 



- — . (y 



determination of t, h, .,., the measurement of ^r makes cj a 



\j 



measured quantity, so that a combination of solvent-separa- 

 tion data and solute-separation data gives us J s ' directly as a 

 measured quantity, and this when obtained (graphically) as 

 a series of integral powers of C gives us t S9 h s , ... ., referring 

 to the solute " separated/' at once. 



It is, of course, in the obtaining of the series to express, 

 as accurately as possible, the data, that the essential ad- 

 vantage, the rapid convergency, attained by the use of the 

 "convenience-concentrations" (c/, C y ), appears. The sim- 

 plification, however, of the coefficients of the series (as 

 compared with (62)) will probably mean greater accuracy 

 in the evaluation of t, h, ..., t s , A s , .... 



Summary of Papers II. and III. 



Paper II., Section I. A postulate, a rider to the definition 

 of " a molecule/' is proposed as the basis of rigorous 

 molecular thermodynamics, and its significance is discussed. 



Section ii. The theoretical basis of the method and the 

 nature of the general problem are outlined. 



Section in. Analysis of the problem of " complex sol- 

 vents. " 



Paper III. (the present) contains some analysis of the 

 problem of "solvation of solutes/' to which is attached a 

 discussion of the results with some brief preliminary 

 illustrations. 



Balliol College, 

 March] 922. 



