On the Law of Molecular Force. 169 



critical were depressed in capillary tubes of 2 millim. dia- 

 meter, and presented markedly convex surfaces. Thus, for 



these substances, E(- + — ) acquires a measurable negative 



value, and the whole expression — — I -+ — ) may acquire 



a large positive value at temperatures near the critical. We 

 thus see how the pressure of saturation of C0 2 for tem- 

 peratures below but near the critical temperature, as obtained 

 by experiments in capillary tubes, might differ greatly from 

 those obtained by experiments in large vessels. This, then, 

 illustrates the origin of the apparent conflict between the 

 experiments of Regnault and of Andrews. 



Some experiments of Kayser's ( Wiedemann's Ann. xv.) 

 bring out in a clear manner the difference between the laws 

 of compressibility and dilatation of a fluid as studied in 

 capillary tubes and in large vessels. When S0 2 is com- 

 pressed at 0° C. in a vessel containing a quantity of powdered 

 glass, it shows no pressure of saturation ; its isothermal is a 

 continuous curve representing quite a different relation be- 

 tween pressure and volume from that which holds when S0 2 

 is compressed free from capillary restraint. 



The full explanation of these facts will be forthcoming only 

 when the capillary theory of Laplace and Gauss receives com- 

 pletion in regard to the matter which, as it left their hands, it 

 has proved inadequate to explain, namely, the influence of 

 temperature in capillary phenomena. The equation of Gauss 

 for the potential energy of a mass of liquid enclosed in a solid 

 vessel contains three terms, representing the potential energy 

 of the liquid due to gravity, that which corresponds to the 

 mutual attractions of its molecules, and that which corre- 

 sponds to the attractions between the molecules of the liquid 

 and those of the solid ; four terms should be added to make 

 the equation complete as regards the whole mass of a fluid 

 partly liquid and partly vapour, — namely, a term to represent 

 the potential energy of the vapour due to gravity, that which 

 corresponds to the mutual attractions of the molecules of 

 vapour, that which corresponds to the actions between liquid 

 and vapour, and finally that which corresponds to the actions 

 between solid and vapour. 



Having discovered that the discarded equation is thus in 

 excellent agreement with Regnault' s work on 00 2 in bulk, 

 and that the differences between its results at 35°5 in the 

 neighbourhood of the critical point and those of Andrews and 

 Amagat are traceable entirely to the experimental circum- 

 stances, I resumed my study of it and proceeded to test its 



