544 
CHEMISTRY: W. D. HARKINS 
cohesive forces to the extent which is necessary in the formation of 
the surface. However, as has already been stated, all of this energy 
does not come from the molecular motion, but a part of it is introduced 
by the action of an external force, which does the work necessary to 
produce what is called the free energy of the surface. The energy 
which is not supplied by the molecular motion, must be supplied in 
the form of work, so the amount of energy supplied by the molecular 
motion is not uniquely determined by the cohesive forces. When a 
liquid vaporizes in an ordinary experiment, the whole of the energy 
necessary for the separation of the molecules against the cohesive forces, 
is supplied by the energy of molecular motion, except for the amount 
which may be supplied by changes in the potential energy of the mole- 
cules themselves, and the latter is probably not a very important factor. 
It is obvious that the contribution of the molecular motion to the 
formation of a surface may be largely independent of the cohesive 
forces involved, and this is made probable by the vaHdity of the entropy 
principle in this case. The cohesion enters so directly into the heat of 
vaporization that it would seem doubtful if for it the entropy principle 
would hold. It would seem surprising, too, if the principle should 
hold for melting, for sublimation, or for sublimation combined with 
dissociation. 
Nevertheless, a study of the literature shows that while the entropy 
principle itself has hitherto remained unrecognized, there are a number 
of empirical relations, which if they hold, involve the validity of a 
much more general entropy principle. Without discussing to what 
extent I believe such a principle is valid, I will state it in two different 
forms as the theoretical basis of all of these empirical relations. In 
order to give this statement I v^^ill define a region as a phase, surface, 
or interface. When in a system consisting of one component, a molecule 
moves from one region into another, the average molecular kinetic energy 
which is converted into the form of molecular potential energy depends in 
general only on the change of state, that is on the region from which the 
molecule comes and the one into which it goes, provided that when a vapor 
phase is involved one condition which enters is the state of the vapor 
with reference to one variable. According to Trouton this condition is 
that the pressure of the vapor shall be the same in all cases, and this 
has been modified by Hildebrand, with an increase in accuracy, to the 
condition that the molecular concentration of the vapor phase shall be 
the same in all cases. In the second form, the principle states that the 
entropy of the change depends only on the change of state, and not on 
the individual nature of the molecules. In the application of this 
