CONDENSATION AND CRITICAL PHENOMENA. 213. 



that were mentioned before as disobeying the law of the 

 diameter, and for which an association of molecules is 

 assumed. The capillary constants obtained for these bodies 

 allow of a calculation of the amount of association by means 

 of the law of corresponding states, a law which will be con- 

 sidered later on. This will be explained in Dr. Shields' 

 article quoted above. 



The law could not be tested up to the critical point as 

 the experiments did not go far enough. Verschaffelt has 

 measured the constant for carbonic acid up to 30° C. and 

 finds that w is higher than i'25 near the critical point. But 

 his results depend greatly on Amagat's values for the densi- 

 ties of CO., near the critical point, which at temperatures 

 near the critical point were found by graphical interpolation 

 from observations at lower temperatures ; this method un- 

 happily leaves uncertainties which make a decision between 

 I "5 and, say, i "3 impossible. Suppose, however, the law 



o- = r [ I -^ I ^''' was not confirmed, this would not show that 



the idea of continuity in the surface-layer ought to be given 

 up. The way in which the formula is arrived at seems 

 to show that it depends on Van der Waals' equation of 

 condition mentioned before, which, though of great im- 

 portance, has been shown not to agree numerically with 

 observed facts. 



The difference between the formula (p + ~\ {v-d) = K t 



and observation is too absolute to be treated as a mere cor- 

 rection. In fact, though at the time it was impossible to 

 say in how far the formula would agree with facts, one thing 

 was pointed out by Van der Waals himself, viz., that the 

 formula would certainly not hold beyond certain densities. 

 Its applicability has since been proved to be even smaller 

 than was anticipated and might have been expected after 

 the successful calculations v/ith regard to carbonic acid. 



For one thing, it leads to a value for the critical volume 



-1 / 

 v,= -x b = (approximately) - — ~ — "- — which relation has never 

 ^ 8 + 273 /, 



been confirmed, v^ always being a great deal less than 



