640 PROCEEDINGS OF THE AMERICAN ACADEMY. 



the energy involved in producing a union between any two atoms is 

 the same for all cases of that particular union, excluding double and 

 treble ties, but without regard to the constitution of the rest of the 

 molecule, and with the aid of this theory he has woven the fabric of 

 his complicated thermochemical hypotheses. But, as I have shown in 

 another place,^ unless together with valence energy other considerable 

 energy relationships exist within the molecule, a possibility as yet un- 

 supported by fact, the energy of the tie between two atoms is variable, 

 and dependent upon the nature of every other atom of the molecule 

 and its position. This logical deduction was established by the con- 

 sideration of such processes as the substitution of hydrogen by hy- 

 droxyl to form primary alcohols or acids as the case may be. The 

 average values of the change in heat of formation of the molecules 

 in these two cases are 



R CH3 - R CH2 OH = 40 Cal. 

 R CHO - R COOH = 72 Cal. 



If we assume that the valence energy of a tie between two atoms is 

 constant for all cases of the union, and exclude the possibility of other 

 considerable energy relationships within the molecule, it is clear that 

 these two values, 40 Cal. and 72 Cal., should be identical. Therefore 

 the assumption is incorrect. 



This conception of varying valence energy is, indeed, old. It was 

 originally stated without experimental basis by Claus,^ and, again as a 

 theory, was developed by Werner ^ in his original discussion of his 

 stereochemical ideas. More recently ^ the significance of comparative 

 reactions has been carefully considered in this connection, and from 

 that point of view much evidence has been adduced in support of the 

 idea of variable valence energy. 



The influence of the nature of those atoms of a molecule which are 

 not directly concerned in a reaction of the molecule upon the heat of a 

 reaction, an influence which must manifest itself in part through va- 

 riations in the dependent valence energies of the reacting atoms, I 

 have already briefly discussed ; ^ with existing data little more can be 



1 Journal of Physical Chemistry, 9, 40-56 (1905). 



2 Berichte 14, 432 (1881). 



' Vicrtcljahrsschrift d. Ziiriehor Naturforschcr Gesollschaft (1801). 

 4 Werner, Bericlite, 39, lL'78 (1900). Fliirscheiin, Jour. f. prakt. Chcm., 66, 

 321 (1902) ; 71, 497 (1905) ; Berichte, 39, 497 (1905). 

 " 1. c. 



