338 Dr. G. Bakker on the 



Whereas the equations (•!) and (5) are true likewise for a 

 curved capillary layer, the equation (6) may be used equally 

 well for a curved as for a plane capillary layer. That means: 

 if a plane capillary layer could exist in stable equilibrium 

 with vapour not having the ordinary pressure, the relalion 

 between p and p would be the same as for a carved capillary 

 layer in contact ivitli the same vapour *. 



If, for instance, in fig. ± 9 C 3 denotes the pressure and the 



value of — for the vapour, which touches a concave surface 



P 

 of liquid (consider a spherical bubble of vapour in the inte- 

 rior of a liquid), I have found for the relation between p and 



- a curve such as C 3 B S A 3 . Likewise the relation p=/( -I 



is presented for instance by the curve C r B G A 6 , when the state 

 of vapour, which touches now a spherical drop of liquid, is 

 given by the point C 6 &c. 



If a plane complete capillary layer exists in contact 

 with the vapour, the state of which is given by the point C 4 , 



the relation between p and - is, likewise as for a curved 



P 

 capillary layer, given by the curve A. 1 B 4 C 4 . Now, for the 

 homogeneous phases which limit a complete capillary layer, 

 the potential of the forces of cohesion is given by the 

 formula of Gauss-van der Waals : 



V=-2ap. 



The equation (5) gives, therefore, > ct 1 — /x = ; that means: 

 the thermodynamical potentials in the two homogeneous 

 phases which limit the complete capillary layer must have 

 the same value. The states of these two homogeneous phases 

 are thus given by the points C 4 and A 4 . We have, however, 

 still a second condition for the stability of our plane capillary 

 layer : the pressure in the points A 4 and C 4 must have the 

 same value. The latter condition being contrary to the first, 

 we see that the plane capillary layer might not be complete. 



If we consider the possibility of an incomplete plane capil- 

 lary layer, the relation between p and - would be given hy 



P 

 ■a part of the curve C4B4A4 ; and for a film which could 

 exist of two incomplete capillary layers which touch each 



* The supposition is of course that the thermic pressure 9, being a 

 pure function of the density, has the same value for the two correspond- 

 ing points in the two considered capillary layers. 



