88 THE ROYAL SOCIETY OF CANADA 



TV U 



Rudorf gets —r = 302 and — =240. As Rudorf adds "the agreement 



between 302 and 240 can hardly be considered good." 



The agreement between the values for water are even worse; 

 TF/c? = 1 133, Li/6 = 1699. A point that crops up here is whether 

 the surface tension should be taken or the surface energy. In defining 

 T he states it to be the surface energy in ergs per sq. cm. but in the 

 table the surface tension in dynes per cm. is quoted. Now the surface 

 tension of water at 0°C. is about 75.6 dynes per cm. but the surface 

 energy equals 115.7 ergs per cm^, and if we take the latter value, 

 we get for TV/d not 1133 but 1735 which is not so very far off 1699. 

 In Phil. Mag. Vol., 39, Jan. 1920, Hammick carried on his work 

 still further and obtains many interesting comparisons. Hammick 

 treats of plane surfaces just as Waterston did. 



Mr. Wilson Taylor in a paper published in Phil. Mag., Vol. 41, 

 1921, applies the formulae to spheres. Starting with a gram-molecule 

 of liquid spheres of the same size he finds how small they must be at 

 the start in order that the total energy yielded up by the shrinkage 

 of the surfaces as they condense into a single sphere may be equal 

 to the internal latent heat of vaporisation. From the original size 

 he deduces the number 



n 



^\f/ 36^ 



Where L = latent heat of vaporisation per gram a1 6° A. 

 7" = the tension of the envelope at ^°A. 

 m— the molecular weight. 

 p = the density of the liquid at 6° A. 

 Taylor suspects that n should be Avogardro's number and selecting 

 water as a test substance, by a happy selection of the internal latent 

 heat of vaporisation at 373°A, viz 498, the density of water at 277°A 

 viz. 1, and an extrapolated value of the surface tension of water at 

 0°A viz. 133.6, he arrives at n= 6.05X10^^ which is very near the 

 accepted value of the number of molecules per gram molecule. That 

 his formula cannot be true is evidenced by the values of n he calcu- 

 lates for a number of other substances. Thus for mercury he gets 

 0.21 XIO^'' and for methyl alcohol W.^'SXW. 



Anyone who looks up the tables of surface tensions will see how 

 varied are the results obtained for T by different experimenters and 

 it seems going beyond the bounds of reason to deduce by extrapola- 

 tion a value of T for a solid, "ice," at a temperature 273° below the 

 temperatures at which T has been measured for the liquid, water. 



