( 169 ) 



From this (I'r^ mnj he f'ouiul. These values, calenhited, arc given 

 ill Table V : 



TABLE V. 



The values given for ,i\ = 0,7, 0,8 and 0,9, however, arc not 

 very reliable, on account of the small xalne which .t'.^ — .t\ then has, 

 by which the tirst member is to be divided. For the same 

 reason (i" ^-^ cannot be given for i\, smaller than 0,1. 



Evidently all the values of ft".,j are negative and as to their 



absolute value, they are smaller than 4. This in accordance with the 



/ d% \ 

 fact, that for stable phases, .r-- nmst be larger than 0. 



Froju the course of the curve ft'.,, =./'('t^^i) appears that the diffe- 

 rential quotient of about ,<■ =i 0.2 to ,(,'=: 0.6 must have a constant 

 value. And this is really found for (-i"x^ . 



With regard to fi".,^ we may still remark, that its numerical \aliie 

 may be found also for the .t'., of the inflection point and the corre- 

 sponding A'j. Van der Waals has namely derived : ^) 



d^p tip <l.i\ 



dx 



d.> 



d.,-. 



2 (6>' ■•'■'-!) 



2 .V, e 



1 



(^ .'1 



— .'*, 



•'1 



/'■' — 1 



•'1 ^ 



l-.t'i+.Vje "^ 1— .Vi-l-.Vi« ■ e 



d'^p dp d,i\ 



— ^ being for the point of inflectio]i, the factor of - — 



must be 0. 



dx„ dx. 



1) These Pioc. ill p. 172. 



