( 477 ) 



uroimd wliv llic "lioiiioiiciicdiis' crilical «•iiinc should luuc a loon 

 as well as llic |>lailj)oiiilciirv('. 



1 am slfciiiilheiied iii this opinion l»y ihc fact, that (ncii in ihc 

 "synunctricar" caso Prof. Kortkw lu; lias conic across similar pconliarilics 

 in the jtlail and il follows, that thcorolically even (liosc niixlni-cs 

 which ohcy van dkk W.vai.s's c'(|na(ion may haxc a crilical cnrve 

 w ith a looj) in il. 



l do "nol mean to maintain that the lionioü-ejicons cni'\c in our 

 case does not jxtssess a loo|) or 1o deny the j)rohaI»ilil\ of I*i-of. 

 \AN DKii Waai.s's liy|)olh(\sis accordinu' to which the association of 

 the alcohols plays a pai'l in pi'odncinii- the aliiioi'nialit\ . What I 

 want lo point oiil is ihal it has not liceii ])ro\('d thai with normal 

 mixlnres the ahnormality cannot occur, allhoiiiih il is xcrx probable 

 that this ahnoi-mality and iji ucneral the formal ion of two iicpiid Uwei-s 

 -~ while theoretically possible foi' normalniixlnres w ith sjx'cial \;diies 

 of the cojislanls — in reality oecui's with associatinji' siil)stajices ojdv '). 



The crilical cnr\-e in the y;, /'-diagram havijig been coin|)leted in (lie 

 WMV desci-ibed one te(ds inclined lo say w itii Pr<>f". \ an dkh Waals 

 that the case is one of a modilied cross-jjlait and not of a li-iie litpnd- 

 plail. According- lo van dkr Lke's experimenls -) the li(piid-j)lait for 

 mixlnres of phenol and water has its ])laitj)oints Inrned in the dii-eclion 

 of the ])ositive volume-axis and above a certain lempei-atnre is entijvly 

 indej)endenl of the cross-|dait. Pi'of. van dkr Waai.s seems inclined 

 lo look n[>oii those })roperlies as charactei'islic of ihe liipiid-plait and 

 lo wilhhohl this name from that pai't of the plait which in onr case 

 is Inrned towards the ./-axis. It will appear j)rcsenlly that this view 

 cannot \<'ry well be suslaijied so that at any rale this ground for draw ing 

 the distinction in (piestion disai)[)ears. Still the pecnlijirily i-emains '') Ihat 

 the ci'ilical cur\'e is a continuous curve, at least when no account is 

 taken of the objecli(»ii iii-ged abo\e against joining iIummh-vcs beyond 

 />. K\<'n then how cxcr it will be observed that on the »|'-surface Iw o 

 independent plaits exist comj)lelely or j)arlly inside each other and 

 thus when Ihe i}'-s>fr/<fcr itsi'lf is considered the contrast between 

 our case and one in w Inch a Iriu' litpiid j)lail would exisl disa|)pears. 

 Moreover ihe abnormality is asci-ibcd 1<» the same cnuse — association — 

 as the foi-mation of the licpiid plait ; if bolh are due lo the same 

 cause, one feels cncu less inclined lo maintain a dislini'lion in the 

 iKniienclalure. 



1) (loiiliiiuilül II |). 17i». 1 iliiiild llio possihilily of (IcdiiciiiLr ;i r('i;ili(in liclwcfii 

 t/jo on the OIK' side ami (i\\ and '/._,.. «m the olhiT. 

 -) Zeilschr. Pliyriik. Clicmic, 33. p. (j^'2, 630. 

 3) Gontinuitat II p. 188. 



