( 745 ) 
Yet this definition requires some further elucidation, because the 
number of molecules under consideration reaches a bounding plane 
of the phase, which does not exercise any attraction on those par- 
ticles, whereas on the particles discussed above and whose number has 
been calculated by Var per WAALS, viz. those which pass from the one 
phase into the other, a force does work directed to the other phase. But 
this difference is in my opinion, only apparent. Also in the equations 
arrived at by van per Waars, one member refers exclusively to one 
phase, the other to the second phase; there are no terms in them 
consisting of factors, one of which refers to the first phase, another 
to the second. That we had to arrive at that result, may be easily 
understood, for the thermodynamic potentials themselves refer either 
to the one or to the other phase and are quite determined by the 
condition of that phase. 
That at least in the definition of the thermodynamical potential 
one number may be put instead of the other, appears as follows. 
Let us consider a liquid in equilibrium with its vapour. The number 
of particles that now passes, per unit of area, through the bounding 
laver is that which Var per Waars treats of; let us now place 
on this liquid a layer of a substance which does not attract the 
molecules; let this layer be thick with respect to the spheres of 
action and provided with narrow channels. The number of particles 
that penetrates into these channels on either side is the number, 
which we used in our definition. Now I assert that the introduction 
of this layer cannot disturb the equilibrium of the homogeneous 
phases *), i.e. their pressure and concentrations will not change. For 
if this had been the case we should have been able to construct 
with the aid of such a layer a so-called perpetuum mobile of the 
second kind, and should have come in conflict with the second law 
of the theory of heat. From this follows that equality of the number 
1) The equilibrium in the non-homvgeneous, capillary layer és disturbed by 
introducing such a wall. For, as vay per Waars has shown (cf. the footnote p. 735) 
the equilibrium in a plane of such a layer is only stable in consequence of the 
attractive forces exercised by the surroundings. When introducing the solid layer 
in question the condjlion in the transition layers will be considerably modified, 
which might also be anticipated. This does not affect our reasoning, for by the 
word “homogeneous” we have positively excluded these transition layers in our 
definition. That this was necessary in any case appears already from the fact, to 
which we have already called attention above, that the thermodynamic potential 
for such layers is no lenger the quantity which determines the equilibrium, but 
that it is replaced by the total potential. We must therefore certainly not have 
recourse to such layers, in order to get acquainted with the thermodynamic poten- 
tial in its simplest signification. 
