( 180 ) 



Fig. 3. 

 from which appears that now too numerator and denominator are 



negative, so 1 — z=.pos. 

 "" dT 



Now it is clear that when we pi'oceed fnrtlier in the same direction 



with onr concentration, the point of intersection G will coincide with 



the point wdiere the vapour bi-anch of the border-lii^e possesses a 



vertical tangent, so where vig = 0. At this moment, too, numerator 



and denominator are negative, so T —— is positive. 



Beyond this concentration virj will have become positive, in conse- 

 quence of which, however, the fraction considered here, viz. 



- + - + 



IVlfj Vsf — 'U\,f . Via 



(.«s— .r^) vi(j 



+ ■ + 



(6) 



will certainly not change its sign in the neighbourhood, because now 

 the tirst member of the numerator and the second of the denominator 

 predominate now. 



Now we know that comparati\'ely far below the first critical end- 

 point p the positive value of v,/ becomes so great,' that soon iv,/ 

 passes through zero, and becomes also positive. 



Then we get the following value . 



- + + + 



{.Vs — .'c^) Vlfj- 

 + ' + 



{Wl X,,) Vsf 



+ ■ 4- 



(7) 



from which it appears that numerator and denominator ha\ e remained 

 negative and the vapour line has kept the same direction. 



If lUsf is positive in the point G, this means that the maximum 

 pressure point of the line solid-fluid has now also got into the stable 

 region, as fig. 4 indicates. 



