'• 4S7 ) 



to xjiiiisli '): llic iiitli'cl ion point, liowcxcr. conlinnc^ lo exist"). In 

 niv fii-sl |);i|H'i' I was led hv llic (v\|M'riin('nls ol' \'iM.\Kn to acci'|>f 

 tlu' |iossil»ilit V ol" a rcli'oiifcssion of llic //-./-curve, w liich lor a sysleiii 

 of the tv|tc cllicr-aiilln-a(|ninonc rcpfcscnls the •lluid" jdiascs coc.xistinu 

 with solid /> and now wc sec that this is iiccessai'ilv so foi' the 

 iuin»e(lia1c neiuhliourhood of // and //■ '"^" '' ''■i'^ hecn ascertained 

 iheorelically. that wc can make solid /,' c\a|)orate somewhat in \\\v 

 iiiiniediale neiuhhonrhood of y/ aiul //. If foi' a svsleni ol' tlic tvpe 

 elher-aidhracininonc an entire \"olalili- sation of orhltrdrij (jiidiifil'ies 

 of ]> were obserxcd, this wonid point to a /y-./-loop as is found for 

 the system o.xyu-en-hronnnc at 17 . oi' in other words to eon/in/n'/i/ 

 between the sn/it/ and the "tluid"" phases. 



As I said hefore it is necessary to e.xaininc the /■-./-sections for 

 dillereiit t(Mnperatur(^s in or<hM- to iiain a hettei' underslandinu uf 

 the question. 



Let ns lirst considcM- the usual case, that the lhre(>-phase-cui'\e 

 lies entirely below the plailpoiiil-curve. 



For a temperature Ixdow the critical temperatnre of A. we <»el 

 th(^ follow inu' /•-./•-section. (Fi.u,'. Hi. 



Ad i'epres(Mits the mol. V(^l. of th(> saturated xapoiir of J., h/ that 

 of li(pii(l J. />'/• is the mol. \'ol. of the saturated \apour of /J and /i/" 

 the mol. ^■oI. of soli<l />. 



(ill denotes the mol. \ols. of the \aponr nnxlnres . !-[-/>, eoe.xistinii 



Fit;;. 3, 



^1 loc. (il. 



~) This is llic case wlicii llic ciuvc has originally only one inllcclion poin!. It' 

 it has two inlk'Clioii points, wliicli is most likely also possijjle, lliey can both 

 disappear. 



