( 280 ) 

 given for the variability of b, holding onlv in the neighbourhood of 



the critical point, \-i/.. -- = 1 — a\ ^\ , now a theoretical for- 



b,, \v J 



inula can be derived for this variability, which just as my theory 



for the Solid State, has a purel}^ pliysiccd foundation. 



Already in my Solid State YII I derived (p. 98—100) from the 



general formula an approximate formula, viz. — =1 — '/( — ~\ , 



b,j \v — bj 



which in mv opinion is preferable to van dkr Waals' empiric for- 

 mula, though of course oui- approximaHve formula is not accurate 

 either. But iu what follows we shall make use of the original quite 

 accurate formula b =:: f {o,T). 



The same quantities /> and Lb, which play so important a part in 

 the transition of the liquid state to the solid state, and the reverse — 

 so that we may safely say : no solid state without these quantities — 

 must also necessarily play a ])art in the theor}' of the liquid (and 

 of the solid) state considered in itself. 



So this furnishes one cause, both for the deviations of the beha- 

 viour of liquids from the original ideal equation of van dek Waals, 

 and for the solid state appearing at lower temperatures. And so in 

 this way the ivhole behaviour of a substance, also the appearance of 

 the three states of aggregation with their gradual transition at cr/^/cvz/ 

 teuiperatures, can be brought under one point of \iew. 



This solves at the same time the question repeatedly put by van 

 DER Waals in his last paper but one (These Proc, April 1911) on 

 the critical quantities (see among other p. 1212 at the bottom; 

 p. 1222 in the middle; p. 1228 at the bottom): "What is, after all, 

 the cause of the variability of b". 



As principal causes he seems still to accept the real diuiinuliou 

 by compressibility (p. 1212 loc. cit), and the apparent diminution in 

 consequence of the partial overlapping of the distance spheres (see 

 p. 1225 and 1226 where the coefficient « == '/s occurring in this 

 case is mentioned). The so-called quasi-association would play only 

 a negligible part (at least at die critical temperature) (see p. 1213 

 at the bottom). 



In our theory, on the other hand, the association, with which the 

 quasi-association is practically identical (see p. 93 — 94 of my last 

 paper on the solid state), is the only factor — and it will appear 

 from what follows that the critical quantities are also perfectly 

 accurately determined by the sole assumption of association, with the 

 variation of volume Ai accompanying it. We shall find that at the 

 critical point the compound molecules are decomposed to an amount 



