646 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



It is clear in all the above cases that the effect of a group upon the 

 heat of reaction is greatest when it occupies the a position, and that 

 when it is sufficiently distant from the reacting group, this necessary 

 distance varying with the nature of the case, it has no appreciable 

 effect. Evidently, too, in most cases, before the effect disappears it is 

 reversed, as in the y, 8, and e positions on the above curve and in the 

 /3 position in those cases where the effects of the groups COOK, 

 CONH2, etc., upon the reaction CH — C CH 3 are considered. 



In most cases an atom or atom group does not appreciably influence 

 the heat of a reaction when the reacting group is further away from 

 it than the /3 or y position, and it may be noticeable only when it oc- 

 cupies the a position, as in the case represented by the accompanying 

 curve. 



Cal.r 



160 



150 



I > 



Curve II. 



In the case represented by the first curve, however, an effect is still 

 notable when the two groups are in the e position to each other, pro- 

 vided the data are nearly correct, and it is important to consider that 

 the data for this series of compounds are perhaps the best of all 

 Stohmann's admirable determinations. Accordingly it is not impos- 

 sible that curve I represents at least one general case of variation in 

 heat of reaction with variation in position of an inactive group ; at 

 any rate the data here represented is not inconsistent with such a 

 conclusion, and for the most part the curves would be of similar shape. 

 These regularities in the variation of heats of reaction which 

 have been here considered, as well as those previously pointed 

 out, lead naturally to the conclusion that the valence energy 

 of a tie between two atoms must be capable of continuous varia- 

 tion^^ within certain limits, not yet to be defined ; otherwise we must 

 assume the existence of large forces of unknown nature Avithin the 

 molecule. AcanxUng to this conclusion and the above con^tiderations the 

 valence energies of two ties between like atoms will under like conditions 

 be equal, under tmlike conditions different. Thus in the compound CR4 



^^ Or, what is less probable, through muuerous distinct niagnitiules discoa- 

 tinuously. 



