SURFACES OF DISCONTINUITY 577 



It has already been stated that recent research shows the 

 existence in the case of numerous substances of surface films in 

 different phases, liquid, solid, gaseous, which can occur under 

 stable conditions at definite temperatures. If this is so the 

 surface layer may not always be a single phase of the whole 

 system ; we may have at times to consider it as a system of 

 phases and treat them thermodynamically just as we treat the 

 different homogeneous bulk phases. The usual procedure 

 would show that the potential of each component would have the 

 same value in each surface phase, still agreeing with the value 

 of this component in each of the homogeneous bulk phases. 

 Actually, in the subsection of Gibbs' treatise which deals with 

 the stability of surfaces of discontinuity, he considers the 

 possibility of a part of the surface being changed in nature while 

 the remaining part remains as before, and the entropy, total 

 masses and volumes of the whole system remain unchanged. 

 The changed part is to be uniform in nature and still to be in 

 equilibrium with the adjacent bulk phases. We shall com- 

 ment on this presently; but in the meantime we may antic- 

 ipate and mention the conclusion to which Gibbs comes (page 

 240). If two films of the same components can exist between 

 the same homogeneous masses, having the same temperature and 

 potentials as the homogeneous masses have for the components 

 in those masses, and the same potentials for components only 

 existing in the surface, then the film which is most stable is the 

 one with the smaller tension. Consequently in a stable film 

 consisting of two or more surface phases the surface tension for 

 each must be the same, for if one phase had a greater surface 

 tension than the other it would disappear on the slightest dis- 

 turbance of equilibrium. Suppose, therefore, that with increas- 

 ing bulk concentration there comes a point when a part of a 

 hitherto gaseous film begins to condense into a liquid film. This 

 seems to be a natural way of imagining the creation of a close- 

 packed unimolecular layer. A small addition of the solute to 

 the bulk phase would not result in an increase of bulk con- 

 centration; all the material would go to the surface gradually 

 increasing the extent of the liquid surface phase which en- 

 croaches on the gaseous; during this period there would be no 



