( 272 ) 



then. In the same way some mixtures are wanting between A'=r 0,476 

 and ,r = 0,121. It is further worth mentioning that for the last 

 tubes we have replaced the Cailletet-cement, used by us formerly to 

 cement the Cailletet-tubes into the brass flanged tube, by ordinary 

 sealing wax, which melts at much higher temperature, for it had 

 happened a few times at the very high temperatures of the room 

 of this year that the gas had escaped from the tube in consequence 

 of the cement melting in spite of uninterrupted cooling with water. 

 We have experienced no disadvantages in consequence of this change. 



Further we have put in practice what we alluded to in our pre- 

 ceding communication, viz. the modified form of the Cailletet-tube. 

 The upper part of the tube was widened, and had a capacity of 

 more than 2 cm' for a length of 2 a 3 cm., i. e. about double the 

 capacity of the other part of the tube, which was 4: 30 cm. long. 

 Our mixture IV e.g. has been investigated in such a tube ; it appeared, 

 however, to be attended with some experimental difficulties, so that 

 we can consider the experiment as only very partially successful. 

 Temperature and pressure were obtained and observed in the same 

 way as in the investigaticm of the system carbonic acid and urethane. 



Our first mixture contained a slight quantity of nitrobenzene 

 + 30 mg. X = O.OIS'*; for we intended in the beginning to demon- 

 strate the double retrograde condensation. 



The system proved to come up to the requirement ^) that the three- 

 phase pressures, A, lie below the pressures of carbonic acid at the 

 same temperatures. For at 31°. 2 we find A = 71.50 KG. p. cm., 

 while we found 75.30 KG. for the critical pressure of carbonic acid 

 at r=31°.05. 



As the concentration of this mixture was not so very small yet, 

 we could still easily distinguish the different phases. At each of the 

 obser\ed temperatures there was always a slight amount of the pale 

 yellow liquid fphase 3) present by the side of the vapour phase, if 

 we made the volume as large as possible. So we were all the time 

 on the righthand side of the vertical tangent of the liquid branch, 

 the point for which Vg, = 0. On compression we brought about 

 unmixing at all temperatures lying between two critical end-points. 



At 7^= 29°. 9 we found the lower critical end-point, where the 

 loop 2 — 3 has contracted to a point of inflection with horizontal 

 tangent (plaitpoint phenomena between the two liquid phases in 

 coexistence with phase 1 (vapour). At somewhat higher temperature 



1) Double retrograde condensation can occur even if this condition is not fulfilled, 

 but it presents other properlies then; cf. van der Waals. These Proc. Vol. XI 

 p. 828. One of us will soon discuss these systems more at length elsewhere. 



