Surface Forces in Fluids. 351 



themselves at uniform distances apart, and will exercise a 

 pressure on the surface-molecules of the liquid, which, by 

 causing them to approach the interior molecules more closely, 

 will render their position one of stability for indefinitely small 

 oscillations. For more considerable oscillations, however, the 

 position of the remaining surface-molecules will still be one 

 of instability : hence, if the molecules are subject to such 

 oscillations, still more will evaporate, and the pressure of the 

 vapour above the liquid will increase, and this will go on till 

 an equilibrium is attained, as it evidently ultimately will be, 

 both on account of the increase of pressure and of density in 

 the vapour. 



Let us now suppose the volume to be diminished, so that 

 the pressure is increased. The molecules of the vapour can 

 readjust themselves to such increased pressure, so also can 

 those of the liquid ; and the new equilibrium of liquid and 

 vapour under increased pressure will obviously be stable, 

 provided that the molecules of the gas are not in their turn sub- 

 ject to oscillations. If, however, a molecule of the gas be 

 carried to a position nearer the liquid than that to which a 

 liquid molecule may oscillate and yet be in equilibrium, it 

 will evidently not return, and there will be a condensation of 

 vapour into liquid. Hence, by admitting that the gaseous 

 molecules also are liable to sufficiently great oscillations, we 

 see that a reduction of pressure to the original amount will 

 take place. If the reduction goes beyond this the rarefaction 

 of the surface-layers will increase, so that a smaller oscillation 

 of a liquid molecule will suffice to secure its escape, which 

 points to a condition of stable equilibrium. It may be re- 

 marked also that the pressure of a different gas above the 

 liquid will, by its pressure, increase the stability of the sur- 

 face-molecules, and thus retard evaporation. 



11. If we now return to the graphical representation of 

 attractive and repulsive forces, which we have seen reason to 

 believe to hold good for liquids as well as for solids, w r e find 

 that at the extension corresponding to the breaking-strain the 

 attraction-curve is above the repulsion-curve, and begins to 

 slope more rapidly. Hence, if this condition continues, the 

 curves must again cut one another, and the repulsion-curve 

 be uppermost. In other words, a state will ensue in which 

 repulsion exceeds attraction. Now this is what we find in a 

 gas, but, further, in a perfect gas repulsion diminishes with 

 an increase of volume, a condition which can only be repre- 

 sented by the repulsion-curve sloping more rapidly than the 

 attraction-curve. Hence, if the gaseous and liquid states are 

 continuous, there must be some distance, such as OQ (fig. 3), 



