128 Mr. W. Sutherland on the 



therefore, entirely on the average relative positions of the 

 molecules of A to those of B in the mixed transition -layer 

 between pure A and pure B. It may be negative or it may 

 be positive, or it may be zero. The distinction between 

 pressure and tension in the action of one body on another 

 represents a difference in the interlinkings of the orbits of the 

 surface molecules. The ordinary idea, that the pressure of a 

 fluid is due to repulsive actions during molecular impact, is 

 not competent to explain the tension that must act between 

 the top portion of the mercury and the glass, when the column 

 of mercury in a barometric tube is got to fill the tube and 

 stand at a height above that which corresponds to the atmo- 

 spheric pressure. A great array of facts supports the general 

 idea of the mingling of molecules at the surface of contact of 

 two bodies; for example, such an action as the rusting of iron 

 suggests the penetration of oxygen atoms to great molecular 

 depths; while Spring's experiments on chemical combination, 

 produced merely by the application of powerful pressure to 

 mixed powders of different elements, demonstrate molecular 

 interpenetration of bodies in a striking manner. Under 

 powerful enough pressure, a granule of sulphur and a granule 

 of lead diffuse into one another and combine as completely as 

 if their molecules had the freedom of the gaseous state. 



Amagat {Ann. de Chim. et de Phys. 5 serie, xxviii.) gives, 

 as a general result of his experimental studies, that at a given 

 volume the rate of variation of the pressure of a gas with 

 temperature is constant ; and Ramsay and Young (Nature, 

 13th Jan. 1887) have announced the same fact in connexion 

 with their studies of the vapours of ethylic alcohol, ether, and 

 methylic alcohol. Stating this result symbolically, we have 



and integrating, 





a type to which Van der Waals's equation belongs, but not 

 Rankine's or Clausius's. But this result is only a first 

 approximation for C0 2 , as a close study of Amagat's numbers 

 for C0 2 shows ; while Andrews has arranged his later experi- 

 mental determinations (Phil. Trans. 1875) to bring out the 

 "dp 



Thus f^ is not a function of volume only, but varies slowly 

 with temperature, and Van der Waals's form of equation 



variation of -^ with temperature in a marked manner. 



