( Ö81 ) 



bria, not in principle different from those given by van der Waals* 

 theory. And very clear and special evidence must be brought forw^ard 

 (which is not the case (cf. § 2)) to show that the two above men- 

 tioned parts of an isotherm must not be united. 



Now, (cf. fig. 4) for a 



Fis;. 4. 



substance whicli exists in 

 the liquid and solid states, 

 call cd and e f the portions 

 of the connodal of the liquid 

 and solid ridges, gh and ik 

 portions of isotherms on the 

 liquid and solid ridges. It is 

 clear that i and h may be 

 joined by a continuous line. 

 For the formation of two 

 ridges it is clearly only neces- 

 sary, that two isotherms g'h', i'k', should incline to the t'-axis very 

 strongly, but still not differ much from the two preceding isotherms 

 gh, ik, and also that h'i', hi should not differ much. With such 

 small variations resulting from the above mentioned controlled changes 

 of the isotherms with the temperature on the Gibbs' surface, we are very 

 familiar since the a and b of van der Waals are taken as temperature 

 functions. The difference between ik and the isotherm ^A is analogous 

 to the difference between the true empirical isotherm gh and the simplest 

 form of isotherm given by van der Waals who has long shown 

 that h must necessarily be a volume function. The portion hi alone 

 appears to have received a somewhat greater change which we may 

 ascribe to a further change of b ^) with the volume. 



With this explanation of the cause for the displacement of the 

 isotherm on the Gibbs' surface we do not come into collision with 

 the assumptions of van der Waals, who assumes that b undergoes 

 a change in the fluid state owing to the compression of the molecule. 

 We thus only specify the possibility of a yet further change of the 

 same sort, which finally produces a new equilibrium between ri 

 and V in the solid state. Beforehand there can be no question of 

 explaining the solid state by referring it to the same processes as 

 those which exist in the liquid state. This can only be done when 

 the relation of the elasticity for instantaneous changes of^olumeand 

 the time of relaxation for the liquid and solid states are worked out 



1) We use the a and h of van der Waals in the most general manner given 

 by this physicist. 



