( 3G0 ) 



posilion wliic'li lias l)ct'ii slaled as possihlo, llie coiilacl could only 

 bc brought about by drawing the dotted lines higher. In ilie lirst 

 ])laee this proves that the last [toint which the» two (^>, 7',, r) surfaces 

 have in common, lies on the uppei- sheet of the eqnilibrinms 

 between the fluid phases and secondly, that the section of the 

 surface of the equilibriums of the solid phase foi- the \a\uQ of ./• of 

 tlie point of cont<aet must have besides two vei-tical tangents, also 

 two horizontal tangents. This is in so far in concordance with what 

 has been observed on p. 240 of the preceding communication, that 

 a similar course for such a curve has been given there. l>ut in so 

 far dilferent, that on acctMinl of an inconi|)lete investigation the 

 opinion was e.xjjressed there, that the two liori/.ontal tangents are 



subjected to the condition - ^0. For, if \Uv\ wei-e subjected to 



Ihis condition, they would exist theoretically^ but they could not be 

 realized. 1 shall continue ami complete the investigation of p. 240 

 prescjitlv, when it Avill appear, that the shape derived in fig. 7 for 

 a section can occur at constant .r and that it can really show a 

 maximum and a miinmum in the realizable part. 



l)ut let us now proceed to examine tig. 8. There the |»ai-ticnlari- 

 ties of the coidact of the curves are dra\vu in the neighboui'hood 

 of the .r of the second critical phase which can coexist Avith the solid 

 bo(h', viz. that which is I'icher in anthraquinone. Again for the ,/• of 

 the critical phase the ( [t/l] curxe has been di-awii of the equilibriums 

 bet^veen the fluid ])liases, and the plaitpoint I* on this cui-\e has 

 been chosen left of the maximum pi-essure. If the coui-se of this 



plaitpoint ciirxc should be such that --- is negatixe tor this plaitpoint, 



(IT ^ 



we ought to have chosen it right of the maximum pressure. But 

 for our pur}K)se the place where /* is chosen, w liether right or left 

 of the maximum })ressure, is of no account. Ojdy /^ nnist not be 

 chosen on the lower sheet, as Avould be the case for //. (\ If. Also 

 the i)rqjection of the three-])hase-j)ressurv? has been drawn this time, 

 in concordance with the fact, that j) decreases witli increasing value 

 of T. Now that P is chosen 0]i the left, the three-phase-pressure 

 need not decrease so rapidly, as would be the case, \\\\q\\ I* was 

 chosen on the right. The points V and Q of this figure are now two 

 points of the section of the two {p, l\.v) surfaces for the >alue of 

 ,/', which we may denote l)y ('V,)'- Foi" the poijit of contact of the 

 two surfaces we must know the circumstances at .r.^, which \'alue 

 is lai'gcr than (./•;.)(• Now we shall be able to l)i'iug about the contact, 

 which is assumed to take [)lace in the figure in point //. by i-aising 



