( 273 ) 



after unmixing had taken place, the phase 3 disappeared on further 

 diminution of the volume, so that we might conchide tiiat we were 

 on the lefthaud side of tiie point 2. After this the pi-essiire began to rise 

 appreciably again, we always found ordinary end-condensation ; we 

 always finished wilh phase 2, so we were on the righthand side of 

 the liquid-vapour [)laitpoint. The quantity of phase 1 that was left 

 when 3 had just disappeared, at the temperature of + 30° a pretty 

 considerable quantity, rapidly decreased with rising T, so that this 

 quantity was very slight for 7^=: 39°. An increase of pressure of 

 0,10 KG per cm" was sufficient to make it disappear. At a somewhat 

 higher T J and 3 vanished simultaneously, so we knew that with 

 rising T the point 2 moves to the left and passes the point ^^ =r 0,015'' 

 at 39°.l. 



At higher T the gas-phase had disappeared before the viscous 

 liquid, so we were then on the righthand side of 2. At ascending pressure 

 this quantity decreased rapidly, and at last disappeared. So we were 

 on the left side of the liquid-liquid plaitpoint of that T. 



At T^39°.9 we found plaitpoint phenomena between vapour and 

 thin liquid phase in coexistence with phase 3. So this temperature gave 

 us the upper critical end-point. On further compression the phase 3 

 disappeared again and the mixture was again homogeneous. 



At temperatures above that of this end-point we found the retro- 

 grade condensation of the phase 3, which must always be the case 

 when we arc on the lefthand side of the plaitpoint. 



Then we proceeded to mixtures with smaller -x by adding car- 

 bonic acid. 



For mixture II, x = 0,0081 we were now quite on rhe left side 

 of the point 2 at all temperatures below that of the upper critical 

 end-point. We further found exactly the same phenomenon as we 

 observed in this case for our former mixture. At the temperature of 

 the upper critical end-point or at a temperature just a little lower 

 we happened just to come still in the point 2, which therefore almost 

 coincided with 1 there. So we may safely assume that this mixture 

 indicates the concentration of the upper critical end-point, the x of 

 the point of inflection with horizontal tangent, from which the loop 

 1.2 will have to form with descending temperature. So from this 

 observation and from that at 39.1 for the preceding mixture we 

 can determine the displacement of the point 2 with the tempera- 

 ture, and by approximation we find then for | -^^ \ the amount : 



dT 



0,015 — 0,008 0,007 



-^— — = 0,01. 



39,8—39,1 0,7 



19^ 



