352 Mr. A. M. Worthington on the 



at which the two curves are again parallel. Now it is very 

 interesting to remark that for any extension between those 

 indicated by N and Q the equilibrium is unstable; for between 

 N and the ordinate to the second intersection a diminution of 

 volume is attended by an increase of attraction, while between 

 the ordinate to the intersection and Q an increase of volume 

 is attended by an increase of repulsion. The first corresponds 

 to the oscillation of liquid molecules outwards, the second to 

 the oscillation of gaseous molecules towards the liquid. 



Hence we see that, when a liquid is in contact with its 

 vapour, there is a limit, on the one hand, to the rarefaction 

 possible for the former at the surface, and, on the other hand, 

 to the condensation possible to the latter — a result which 

 explains the sudden transition from the density of the liquid 

 to that of the vapour. Within these limits the rarefaction 

 and condensation will evidently still take place ; for the con- 

 ditions of the investigations which we gave for the contact of 

 two liquids are, so far as the denser liquid is concerned, 

 unaltered, except by the presence of a mechanical pressure, 

 which, as we have seen, will not affect the character of the 

 result ; while the vapour which is substituted for the less dense 

 liquid of the investigation fulfils the one essential condition 

 that the pressure-curve shall slope more rapidly than the 

 attraction-curve — the fact that it, too, is in a state of pressure 

 not affecting the result. If the dilatation near the surface of 

 the liquid only took place parallel to, and not also normally 

 to, the surface, the equilibrium of the liquid molecules for 

 outward oscillations would be much more stable ; but we 

 know that when a liquid is passed up into a barometric 

 vacuum the evaporation is extremely rapid, a fact which indi- 

 cates that the surface-layers are very near the limit of stability 

 with respect to outward oscillations, and Avhich confirms the 

 supposition that the dilatation takes place in all directions. 



We are thus led to the conclusion that a solid wall cutting 

 the surface-layers of a liquid transversely will experience a 

 pressure due to the vapour and a tension due to the liquid 

 near the surface, the resultant action being the difference 

 between the two. 



12. Since a solid whose surface is exposed to a gas does 

 not exhibit the phenomenon of a surface-tension — for its 

 edges and angles are not rounded as they would be were the 

 surface-tension efficient in modifying the figure of the solid — 

 we must conclude that the internal friction among its mole- 

 cules permits a state of strain to exist between the surface- 

 molecules and those in the interior. Thus if we were 

 suddenly to expose the interior of a mass of water to the air, 



