Surface Forces in Fluids, 345 



the most easily reconciled with the phenomena of evapora- 

 tion ; and though either hypothesis will lead us to the same 

 results, we shall find no reason fcr supposing that the pres- 

 sure at any point near the surface of a liquid, whether within 

 the liquid or in the vapour contiguous to it, is not equal in all 

 directions, and shall therefore keep before us the latter of the 

 two hypotheses ; though the reader should bear in mind that 

 in the conclusion we have now arrived at — that in a non- 

 volatile liquid exposed to a vacuum there must be a diminu- 

 tion of density as the surface- is approached from within — 

 the other interpretation of the phrase " diminution of density " 

 is possible. Now, if we desired to diminish the density of the 

 interior of the mass of liquid, keeping its temperature all the 

 time the same, so as to reduce its molecules to the same 

 condition as at any given very small depth below the surface, 

 we could do it by stretching the liquid. But to say that the 

 condition of the molecules near the surface is that into which 

 we should reduce the molecules of the interior by establishing 

 among them a state of tension, is equivalent to saying that 

 the surface-layers consist of liquid which differs from the 

 liquid in the interior by being in what we call a state of ten- 

 sion ; and it is evident that the tension increases as the 

 surface is approached. 



And it is further obvious that this tension will be exerted 

 on any solid wall cutting the surface. For if we imagine a 

 homogeneous cube of solid or liquid matter held in a state of 

 isotropic tension by adhesion to solid walls, we perceive that 

 the molecular condition which corresponds to equilibrium 

 between the wall and the stretched mass is one of a tension on 

 the wall. A solid wall, therefore, which cuts through the 

 liquid surface so that part is above and part below the surface 

 will only be in equilibrium with the layers near the surface 

 when the portion in contact with them is in a state of ten- 

 sion ; and the argument holds equally well if liquid wall be 

 substituted for solid. (It is also to be observed that a plane 

 solid surface laid horizontally upon the liquid surface will, if 

 we adopt the more general supposition, also experience a 

 tension at the first moment of contact.) 



Two solid walls connected by a liquid surface will there- 

 fore be drawn together, and, if capable of motion, will abso- 

 lutely move towards each other. Their motion involves the 

 diminution of the liquid surface between them, which in turn 

 requires that many molecules which have previously been so 

 near the surface as to be separated by greater distances than 

 those in the interior are now brought into the mass of the 

 liquid, and their distance apart diminished. Now the exten- 



Fhil. Mag. S. 5. Vol. 18. No. 113. Oct. 1884. 2 A 



