628 Albert P. Mathews 



different from those computed by Walden from Stefan's con- 

 clusion, Walden 's values being in fact about two-thirds of 

 my values. The figures given for the internal pressures of 

 liquids by Walden, Davies, and Traube are certainly far too 

 low and are erroneous. 



Summary 



The internal pressures of liquids, or rather the value "a" 

 of van der Waals' equation, has been computed from the 

 surface tension, assuming that the depth of the surface layer 

 is (T C /T C T))f /3 molecular diameters; from the law of Eotvos 

 and T. Young; from van der Waals' equation, assuming that 

 b c is always 2V C /S, S being equal to RT C /V C P C ; from the in- 

 ternal latent heat of vaporization close to the critical tempera- 

 ture, and from the molecular weight and the number of va- 

 lences. All of these methods give practically the same results. 

 The values of "a" thus computed are constant in the case of 

 pentane and ether over a wide range of temperature, indicating 

 that barring association "a" is constant; the values are uni- 

 formly higher than those computed by others with the excep- 

 tion of some computed recently by Lewis from the latent 

 heat of expansion of liquids. The values recently given by 

 Traube, Walden and Davies are too low and incorrect in other 

 ways. The results confirm my conclusion that the molecular 

 cohesion is a function of the molecular weight and the num- 

 ber of valences in the molecule. The formula: a = 

 27^/64 X 273 2 P C gives values about 14 per cent, too low 

 for ordinary substances and very much too low for simple 

 diatomic gases. It should be replaced by the formula a = 

 (S 2 S + 2)T*R 2 /(S 2 (S 2)P C ). These results show, also, 

 that Stefan's conclusion that half the work in vaporization is 

 done in moving a particle into the surface is incorrect. 1 



University of Chicago 



1 APPENDIX. By combining (23) and (26) we have the following formulas 

 for calculating S and d c . M is the molecular weight: 



(35) S 2 = (dt ^)T 4 / 3 /(Tc T)V 3 MP C 



(36) d c = MPc(d e ck>)/RTc /3 (Tc T) x /3. 



