Theory of Surface Forces. 213 



travel along the straight line from G towards E under very 

 peculiar conditions. Apart from the action of the walls of 

 the containing vessel, and of suspended nuclei, the path from 

 G to F must be followed. The path from Gr to E implies that 

 the vapour at Gr is in contact with the liquid in mass. This 

 is by supposition not the case ; and the passage in question 

 could only be the result of foreign matter whose properties 

 happened to coincide with those of the liquid. If the walls 

 attract the vapour less than the vapour attracts itself, they 

 cannot promote condensation, and the path H G F must be 

 pursued. In the contrary case condensation must begin 

 before G is reached, although it may be to only a limited 

 extent. Probably the latter is the state of things usually met 

 with in practice. So soon as the walls are covered with a 

 certain thickness of liquid, the path coincides with a portion 

 of GE 0, and the angle at Gis only slightly rounded off. 



Similar considerations apply at the other end of the straight 

 course. If the liquid be expanded through C, it will not, in 

 general, pass along C E, but will continue to pursue the curve 

 C D, and will even attain the limit D, if the attraction of the 

 walls upon the liquid be not less than that of the liquid upon 

 itself. In the contrary case separation will suddenly occur at 

 a point upon the wall, a bubble of vapour will be formed, and 

 a point on the straight line C E will be attained. It is thus 

 scarcely conceivable that a fluid should follow the broken 

 course A B E G H without some rounding of the corners, or 

 else of overshooting the points G, G, with subsequent precipi- 

 tation upon the line G E G. 



A very important question is the position of the line C G. 

 Maxwell * showed that inasmuch as the area of the curve 

 represents work performed at a constant temperature, it must be 

 the same for the complete course as for the broken one. The 

 line C G is therefore so situated as to cut off equal areas above 

 and below. 



This discussion is of course quite independent of the precise 

 form of the relation between p and v. All that is necessary 

 is such a modification of Boyle's law at great densities as will 

 secure the fluid against indefinite collapse under the influence 

 of its self -attraction. 



We will now pass to the question of the transition from 

 liquid to vapour, still supposing the strata to be plane. This 

 is a problem considered by Maxwell in his article upon 

 "Capillary Action" in the Encyclopcedia Britannica f; but 



* ' Nature,' vol. xi. p. 358, 1875 ; Reprint, vol. ii. p. 418. 

 t Reprint, vol. ii. p. 560. 



