RICH.VIIDS. — SIGNIFICANCE OF CHANGING ATOMIC VOLUME. 585 



in one of sodinm, and a greater change in one of sodium than in 

 one of lithium. If we use heats of formation as indices of tlie free- 

 energy change, or the work which may be done by afHnities (a pro- 

 ceeding wliich is permissible in the case of these simple binary salts 

 because of the small change of heat-capacity which occurs when they 

 are formed from the elements*), we find that this prediction is fully 



* Richards, These Proceedings, 38, 293 (1902). In the last paper upon this sub- 

 ject, it should have been made clear that the statement concerning the possible 

 application of the heat-capacity relation to electrolytic dissociation is applied to 

 very dilute solutions only by extrapolation, since no data e.xist for the heat- 

 capacities of solutions more dilute than half normal. Professor Ncrnst has 

 kindly pointed out to me that in case this relation really holds in very dilute 

 solutions the coefficients obtained by him and by Jahn would acquire a special 

 significance, their application to the facts remaining unchanged. This circum- 

 stance has nothing to do with the cases considered in the present paper; it is 

 alluded to now only because I regret that in the previous discussion a possible 

 inconsistency was suggested, wliich does not really e.xist. 



t These results are due to careful experiments verj^ kindly made in this labora- 

 tory by Dr. G. P. Baxter. The results of F. W. Clarke are undoubtedly too low, 

 since the salts are hard to prepare pure. The value of the specific gravity of 

 lithium, 0.552, is due to experiments made here by Richards and Bonnet. 



t Because the heats of formation of chlorides as given in the text-books always 

 refer to gaseous chlorine, there must be subtracted from them the latent heat of 

 evaporation of liquid chlorine. The molecular latent heat of evaporation of chlo- 



