314 PROCEEDINGS OP THE AMERICAN ACADEMY. 



in this direction is greater than the tendency to run backwards. The 

 reaction is, however, a cooling one, absorbing perhaps 8 kilojoules per 

 mole. The only available data concerning the heat capacity of the sys- 

 tem indicate that there is, however, a considerable gain, perhaps 15 mayers, 

 in this important quantity, as is usually the case when water is formed 

 in a reaction. If it were not for this gain, the reaction would probably 

 have been a warming reaction instead of a cooling one. Thus we have 

 the anomaly of an exothermic reaction (exothermic in the sense that chemi- 

 cal energy is converted into heat) which causes a cooling effect, because 

 the heat capacity is increased in yet greater measure. 



With strong electrolytes, the deviations are more marked. Here 

 dilution usually causes a continual decrease in heat capacity. But a 

 decrease in heat capacity seems to involve the expenditure of available 

 energy (P), therefore a diminution in (J£ — P and hence also in the " con- 



stant" equilibrium ratio In balanced against it. That is, an increase 



in the volume causes an increase in the "constant " In . This'cor- 



c ... 



responds qualitatively with the facts, for Rudolphi * found that upon 

 multiplying the normal equilibrium ratio by a direct function of the 

 volume (V^) a formula was obtained which represents the facts much 

 more closely. Bancroft f has pointed out the variable nature of the ex- 

 ponent — a variation which would only be expected according to the 

 present analysis. 



In the light of these considerations, it is evident that neither "mass- 

 law," vapor pressure, nor freezing point can give correct values for the 

 extent of electrolytic dissociation ; and since conductivity is affected by 

 varying migration, it also is unreliable. Hence it would appear that 

 each of the contestants in the recent unfortunately polemical discussion 

 upon this subject is partly right and partly wrong. Nevertheless, al- 

 though we have not yet data for determining the exact extent of disso- 

 ciation, it seems to me a great mistake to reject wholly this useful and 

 illuminating hypothesis. 



Until more satisfactory data concerning heat capacity are obtained, it 



is idle to pursue this aspect of the question further, important as it seems 



to be. 



On differentiating the equilibrium equation 



ln co n °... ar-p 



c 



RT 



* Rudolphi, Zeitsclir. phys. Chem., 17, 385 (1895). 

 t Bancroft, Ibid., 31, 188 (1899). 



