242 Mr. D. L. Hammick on Latent Heat 



then the area o£ such a laver will be -, , where d is the 



pV d 



molecular diameter. The product* '-j- will then be approxi- 

 mately the "true" molecular surface energy, since every 

 molecule in V c.c. is contributing to it and we expect to find 

 that the relationship 



d ' J 6 [1) 



will hold (J is the mechanical equivalent of heat, L x being 

 the internal latent heat per gramme molecule in calories), 

 though in view of the assumptions made great precision 

 cannot be looked for. It is also fairly certain that the 

 density change on passing from liquid to vapour is not 

 abrupt, but that the molecules must pass through a 

 "capillary layer." It would seem reasonable, however, to 

 take the potential energy (i. e. surface energy) associated 

 with the molecules in the actual surface layer as a correct 

 measure of the work done in getting there, even though that 

 work has been done by passing through a succession of 

 surface layers (cf. Matthews, loc. cit.). But the surface 

 energy as measured is the potential energy at the boundary 

 between liquid and vapour, which has also a " surface 

 energy" due to its own internal pressure. The surface 

 energy of a liquid not in the presence of its vapour would 

 thus be greater than when in equilibrium with vapour. At 

 present it does not seem possible to compute or measure 

 this vapour surface tension; hence equation (1) must be 

 tested at temperatures where the vapour effect is likely to 

 be a minimum, i.e. at temperatures as low as possible. Low 

 temperatures are also necessary if the assumption of approxi- 

 mate contiguity of the molecules is not to cause very large 

 error. Unfortunately, most of the latent heat data refer to 

 the boiling-point of the liquid. 



A further difficultv lies in the determination of the 

 molecular diameter d. There is no satisfactory method 

 available for deriving d from measurements of properties of 

 the liquid itself. The diameters used in testing (1) have 

 therefore been obtained by taking the mean of the values 

 calculated from measurements of various physical properties 

 of the vapour, such as thermal conductivity, refractive index 

 (or dielectric constant), the critical data in the form of 

 van der Waal's "£," and the limiting density of the most 

 dense form. Values derived from viscosity measurements 



