370 KEYES 



ART. J 



where R, the universal gas constant, is equal to the product of 

 tti, 02, ... and the corresponding molecular weights. Here 

 El, Ci and Hi are also assumed to have been multiplied by the 

 corresponding molecular weights. 



II. Inferences in Regard to the Potentials in Liquids and 

 Solids (Gihbs, I, 164, 165) 



There might be included under this heading a large portion 

 of the principles and doctrine which have found application in 

 physical chemistry in the last half-century. The fact that a 

 comparatively simple basis of fact could have such general 

 applicability was well known to Gibbs, as is indicated by the last 

 sentence of the section (7th line from bottom, p. 165). Indeed a 

 few empirically discovered facts interrelated thermodynamically 

 suffice to form the theory of those liquid mixtures wherein the 

 masses of one or several constituents are very small relative to the 

 mass of one of the components*^. The principle of the equality 

 of the potentials of a component in equilibrium in the coexisting 

 gaseous and liquid or solid phases affords the means of deter- 

 mining the potentials of the condensed phases. Because of this a 

 full knowledge of the properties of pure gases and their mixtures 

 is of fundamental importance in extending the range of applica- 

 bility of the general theory. Thus it becomes clear that great im- 

 portance attaches to a knowledge of the constants of the equation 

 of state for different substances, and the rules for combining 

 these constants, in order that the constants for the equations 

 for mixtures may become available. On the other hand" 

 given sufficient data for pure substances and their mixtures, the 

 required thermodynamic quantities may be accurately com- 

 puted empirically, using the assumption that the ideal gas laws 

 hold rigorously in the limit of low pressures. It is evident, 

 however, that on this basis an almost prohibitive amount of 

 experimental data would be required to satisfy the needs of the 

 science, and therefore continuous effort should be made to 

 develop a rational form of equation of state with the aid of 

 statistical mechanics. It is, indeed, apropos to add that the 

 correlations of physico-chemical facts by thermodynamics can 



