( 246 ) 



tribroointoluols, this ahvays points to difference in size of the molecules 

 in tiie liqnid and solid phase '). In fact Jaegek, found that his 

 isomers are very likely hl-molecidar in the suVkI phase '). 



2. We may now ])ut the question : When will the niininium at 

 D, which will disa]»peai' in any case for values of ,^' siimller than 

 those for which tig. 3 holds, disappear before the case of tig-. 6, so 

 that a course as in fig. 5 becomes possible ; and when will it disappear 

 ajter the case of fig. 6, as has been assumed iji our figures 6 to 8. 



To answer this question, we shall first state for what values of /3' 



the case of fig. 6 occurs. 



d»g' 



The point It/i lying then on the top of the curve r-7- = at 



x' = V, '), ^^e have, besides the equations (2) for .r'^^^l, (see p. 153 1. c), 



ö*S' RT 



also the relation 3-— = or — ; 2 a' = 0, i. e. with U = 2 



ox^ x{\ — x) 



the relation T = u'.c'Cl—x'). 



The condition sought is accordingly : 



r = 



TAl- 



3' 



KT 0,0 

 1 H '- lo., -^- 



1 H lo,i -^ 



T 1i <^'' 



for which with regard to the fundamental equations, some simplifying 

 hypotheses permissible for our pur[)ose have been made, which may 

 be found on page 152 of the paper mentioned. 

 Now we can solve (R = 2): 



log 



0,5 V 4 ' 4: T ^ 



1-.Ï 



/5' 



0,5 



lOff = 



X 



2 1- 



LI1 y ^Ll±ili' 



'1 a' 



hence, as (1 — ./•'. -|- .v = 1 : 



\-'^ 



L^V... ^' 



^-1 



2, (1) 



1) See p. :208 and 209 of the "Proefschrift", where Jaeger gives the proof of 

 this thesis, which I had communicated to him in a letter. 



2) See p. 208 and 194 of the "Proefschrift". 



*) Only if we assume x\ — x.^ (so 6^ = b.^), tliis pai'abolic curve will be sym- 

 metric and its top will be exactly at x' — 7o. 



