104 PROCEEDINGS OF THE AMERICAN ACADEMY. 



The initial decrease of C„ may very possibly be an association effect. 

 A word should be said about this association effect. The reason that 

 an association from single to double molecules, for example, reduces 

 the specific heat to one half, neglecting the effect of the internal 

 degrees of freedom, is that a rise of temperature of one degree corre- 

 sponds to a definite increase in the kinetic energy of each molecule, 

 and when there are half as many molecules, half the additional energy 

 is needed to increase the energy by an amount corresponding to one 

 degree. This means that association has an effect on specific heat as 

 long as temperature remains a molecular affair. This is true for a 

 gas. But the law of atomic heats for solids suggests very strongly 

 that in solids temperature is no longer an affair of the molecule, but 

 has become a matter of the atom. We naturally expect somewhere 

 a transition from the one state to the other, and the liquid is the 

 natm*al place to look for it. That is, if a liquid temperature is on 

 its way from being an affair of the molecule to becoming an affair of 

 the atom. It is not likely, therefore, that the existence of large mole- 

 cular complexes in a liquid, as in the case of association, will modify 

 very much the behavior with respect to temperature, and in particular 

 to the specific heats. Such effect as there is should be looked for at 

 low pressures. 



A word should be said about the curves for the average of Cp and C^ 

 over the entire temperature range. The average Cp is the total heat 

 absorbed at constant pressure between 20° and 80° divided by GO. 

 C, is the average of the four values at 20°, 40°, 60°, and 80°. These 

 two averages are not always equivalent when there are large varia- 

 tions with temperature, but they are always approximately equal. 

 In cases of question, the average Cp corresponding to C^ may be found 

 from the curves for Cj, at the four temperatures. To find the average 

 0, corresponding to the average Cp would involve a more complicated 

 procedure. 



General Discussion of the Bearing of the Results on a 

 Theory of Liquids for High Pressures. 



It is proposed to discuss here the nature of the problems which 

 confront one at high pressures. No attempt is to be made to develop 

 a new theory, but merely to indicate some of the directions in which 

 the data of this paper suggest modifications of conceptions which 

 have hitherto worked at low pressures. 



Perhaps the most far reaching modification is in regard to our 



