THE ELEOTRTO AHD LITMINIFER,OTJ>S MEDllTiVr. 2^9 
tion exerted on the surface molecules in the layer of transition. It follows that the 
pressure of molecular attraction across an internal interface, which is the virial per 
unit volume with changed sign, is equal to Laplace’s intrinsic pressure K in the liquid 
arising from the inward attraction of the surface molecules. This equality is easily 
seen to involve the consequence that the layer of transition at the free surface is very 
thin compared with the radius of molecular attraction, an important conclusion of 
which the bases are here the statistical stability of the liquid state, the dynamical 
principle of the virial, and the hypothesis that the range of sensible molecular attrac¬ 
tion extends over a considerable number of molecules in the liquid state. In the 
condensation ol a vapour there is degradation of internal energy into sensible heat of 
amount equal to the latent heat of condensation diminished by the work of condensa¬ 
tion of the vapour and increased by the volume of liquid thereby produced multiplied 
by the Laplaciax pressure K. 
55. Consider two fluids, one the pure solvent and the other a solution, separated 
by a rigid porous partition, with extraneous pressure applied on the side of the 
solution to balance the osmotic pressure and so to produce equilibrium as regards 
transpiration through the partition. Now let a slight amount of transpiration occur 
by very slightly reducing this extraneous pressure ; thereby work is done against 
that pressure, equal to its intensity multiplied by the change of volume owing to 
transpiration of the solvent into the solution. Tlie operation takes place steadily 
under conditions of equilibrium, so that it can be reversed either by a known process 
or, as we might assume, by some process not yet discovered—in this case merely by 
reversing the pressure, or it may be cyclically by evaporation : thus the work is done 
at the expense of an equivalent of available energy, partly thermal, and partly of a 
molecular type which would otherwise run down into heat of mixing of the liquids. 
Hence the osmotic pressure between two fluids is equal to the whole amount of free 
or available (not total) energy that would be degraded when unit volume of the pure 
solvent is mixed with an indefinitely great volume of the solution into which it 
transpires, supposing that there is no sensible change of volume in that process; if 
there is change of volume this value must be altered in the ratio of the final to the 
original volume of the transpired material: so long as the dissolved molecules are out 
of each others’ range of influence, the change of volume, if any, must be independent 
of concentration. This proposition will be exactly true if the pores in the partition 
are so narrow, that the cross-sections of the filaments of fluid contained in them each 
involve so few molecules that the mutual energy of the molecules of fluid in the pores 
is negligible compared with that of an equal mass of fluid in bulk. Inasmuch as to 
excite the osmotic pressure, pores or tubes of molecular fineness have to be employed, 
it follows that it is not an ordinary transmitted mechanical pressure; and the energy 
which is associated with it is not merely the organized energy from which the 
mechanical forcive is derived, but the whole amount of energy thermodynamically 
available. If the pores are wider, the mutual energy of the molecules in them ceases 
