( 489 ) 



We have 



{Vs — Vf) — (.'•. — .';/•) ( ^^-' 



/■/pT 



Vsf. 



Vsf denotes the decrease of volunic pei- moleenhir quantity, when 

 an inlijiitely sniall (luantity of the solid |(liasc passes into the co- 

 existing phase at constant pressure and constant temperature. 



By substitution we get : 



d/'/ 



,ƒ . dvf 



Ovf.OWf Ox f 



dxf 



or 



diu 



^'sf 



Van uer Waals has hitely demonstrated that Vsf- can twice 

 become zero when Cs. is smaller than ly-, in consequence of which 



dvf 



dXf 



becomes twice intinitely large. 



Further ^— ^ can also twice become zero, but this does not give 

 ov/ 



drf d'\p 



rise to an infinitely large value for , as, ^ .^ being zero, and 



(/■/■/■ Orf 



we being therefore in D or 7J' (Fig. 2 van dkk Waals), b\^f=z<x) 



dv f 

 and — has therefore a tinite value. 



dxj 



If ?\ is largei' than Vf, which may also occur, then only one vertical 

 tangent is [«ossiblc. This is atteudcil by a cliango also in the course 

 of the lower |)art of llie line chi'li. In the aI)o\e ligui'c fli i-iiiis to 

 the left for smaller \olumes, but then this cuivc must directly rim 

 to the right, which means that the solubility of II in .1 inci-eases 

 for smaller xolumes (larger pressures), a behaviour which may also 

 be expected theoretically for smaller xolumes when initially /•ƒ• ^ r.,, 

 whereas the reverse, so the usual course is found for larger \ olnmes. 

 If, however, v,^ ^ Vf the course must be the abnormal one fi'om the 

 beginning. 



For the better understanding of fig. 3 I shall add a few words 

 about each of the ditferent regions. 



Let us assume that we have a mixture of the concentration ,r, 

 for a volume ./■, /", ; we are then in tiie region of L -\- (J. If we 

 draw the nodal line n v^ n^ through i\, n denotes the mol. vols, and 



