PRESIDENTIAL ADDRESS — SECTION B. 97 



state) is like the substance A in the B layer. As temperature 

 rises the liquid diminishes in total quantitj-, the vapour in- 

 creasing ; but the specific volume of the liquid increases, while 

 that of the vapour decreases. The residual liquid is, in fact, 

 constantly encroaching on the space of its vapour, just as the 

 residual substance A in the A layer is constantly absorbing 

 the solvent B from the B layer. Finally, in either case, the 

 specific volume of the substance will become identical in both 

 layers, which means that the layers themselves will become 

 homogeneous and indistinguishable. Our system will then 

 have reached its critical temperature, — the temperature of 

 infinite volatility in the one case and of infinite solubility in 

 the other. 



So much for hypothesis. Are there any facts in support 

 of it ? Well, in the first place the hypothesis demands that (in 

 the absence of chemical change) increase of solubility with rise 

 of temperature shall be as general a law as increase of vapour- 

 pressure ; and we find that this agrees with the known facts, 

 more especially since Tildeu and Sheustone (Phil. Trans., 188i) 

 cleared up certain doubtful cases. Secondly, the hypothesis 

 seems to demand some connection between the true melting- 

 points of salts and the rates of their increase of solubility ; and 

 such a relation has, in a general way, been established by the 

 same observers. Thirdly, we have the fact, in complete ac- 

 cordance with the hypothesis, that, while no case is known of 

 a solid body ha\dng, as such, infinite solubility in any simple 

 solvent, several cases are known of liquids of infinite solubility 

 and also of solids luliich, after they have melted in inesence oj 

 their oiun solution, become at some higher tevq^erature infinitely 

 soluble. This last statement refers to the cases investigated 

 by Alexeeff (Wiedemann's Annalen, 1886), of which I must say 

 a good deal more directly. It would seem to apply also to the 

 case of silver-nitrate, which Tilden and Shenstone described as 

 dissolving in water to the extent of 18-25 parts to 1 at so 

 low a temperature as 130° C. The true melting-point of the 

 salt is 217° ; and I have seen it stated (but have been unable 

 to find the published account) that Shenstone has himself 

 shown it to be fusible in water and of infinite solubility at 

 quite reachable temperatures. 



With regard to substances that are liquid under ordinary 

 conditions, we have the well-known fact that some pairs are 

 infinitely soluble in one another, while others exhibit the 

 phenomenon of only partial solubility. The hypothesis would 

 draw no hard-and-fast distinction between these cases, except 

 the practically important one that such a mixture as that of 

 ether and alcohol, which belongs to the first class, is usually 

 above its critical solution - point, while such a one as ether 

 and water, which belongs to the second class, is usually 

 7 



