SURFACES OF DISCONTINUITY 511 



weight of a superincumbent column of liquid is in general the 

 most important portion of the thrust on the enclosing vessel. 

 Yet it only complicates the situation we are discussing to bring 

 this in at all. It is best to conceive the liquid to be free from 

 gravity, as Gibbs actually does in a great part of his treatise. We 

 may, if we wish, consider it to be contained in a vessel which it 

 touches everywhere, and which can be regarded as fitted with a 

 piston so that a thrust can be applied if required, — a thrust 

 which by Pascal's law is distributed at all parts of the surface in 

 proportion to the size of each part, or is exerted normally across 

 any conceptual dividing surface in the interior, again in pro- 

 portion to its extent. Or we may think of the hquid as a 

 spherical mass subject to the pressure of a surrounding gas and 

 for the moment regard the sphere as so large that any small 

 portion of the surface is practically plane. If now the pressure 

 of the surrounding gas were zero the pressure would also vanish 

 in the liquid. (Actually the pressure cannot be less than that 

 of the saturated vapor.) The reader who has studied the 

 earlier portion of Article K of this volume (pp. 395 to 429) will 

 realize that this would be just a special case of an unstressed 

 state of a body. Yet in the interior of the liquid there must 

 be a relatively enormous pressure in the sense in which that 

 word is used in connection with a gas; "kinetic" pressure we 

 shall call it. In the liquid there exists a thermal motion of the 

 molecules, and on account of the much larger density of the 

 liquid the rate of transference of momentum across an interior 

 conceptual surface is very great indeed. Clearly this internal 

 kinetic pressure cannot be the quantity which is denoted by 

 the symbol p in our equations; for that, as we have seen, would 

 practically vanish when the stress in the liquid produced by the 

 thrust of an external gas or piston in an enclosing vessel dis- 

 appears. Of course at the surface there is the well known 

 inward pull on each molecule in the layer whose thickness is 

 equal to the radius of molecular attraction. This has the effect 

 of turning inwards all but a small fraction of the molecules 

 moving through this layer towards the surface, and in conse- 

 quence the actual kinetic pressure at the surface is enormously 

 reduced below the kinetic pressure which exists in the interior. 



