1891.] Some Suggestions regarding Solutions. 307 



represent a similar continuous change from saturated solution to 

 dissolved substance on increase of concentration. 



Mr. Aitken's experiments on the cooling of air containing water- 

 vapour have shown us that it is possible to realise a portion of the 

 curve AB ; the phenomenon of " boiling with bumping " constitutes 

 a practical realisation of a portion of the curve DE ; and we may 

 profitably inquire what conditions determine such unstable states 

 with solvent and solvent. 



Regarding the portion of the curve AB, I think that no reasonable 

 doubt can be entertained. It precisely corresponds to the condition 

 of supersaturation. In the liquid-gas curve, the volume is decreased 

 constant temperature without separation of liquid ; in the solvent- 

 jlvent curve the concentration is increased without separation of the 

 jlvents. Dr. Nicol has shown that it is possible to dissolve dry 

 >dium sulphate in a saturated solution of sodium sulphate to a very 

 jnsiderablo extent without inducing crystallisation ; and here we 

 we a realisation of the unstable portion of the curve AB. In the 

 3-liquid curve pressure falls with formation of a shower of drops ; 

 the solvent-solvent curve crystallisation ensues, and the solvents 

 jparate. The phenomena are, however, not completely analogous ; 

 ae complete analogy would be if the temperature were so low that 

 the substance in the liquid-gas couple were to separate in. the solid, 

 lot in the liquid, state. This, so far as I am aware, has not been 

 tperiinentally realised, but one sees no reason why it should not be 

 sible. 



I have some hesitation in offering speculations as to the state of 

 itter at the portion of the continuous curve DE. It may be that 

 corresponds to a syrupy or viscous state. Cane-sugar at a 

 loderate temperature dissolves water ; indeed it is possible to obtain 

 solution of 1 per cent, of water in molten cane-sugar. And such a 

 alution, if quickly cooled, remains a syrup. But it can be induced 

 crystallise by the presence of crystals. Thus, in such a mixture 

 if sugar and water, a few grains of crystalline sugar cause the 

 fhole mass to crystallise, and water saturated with sugar and sugar 

 sparate into two layers. Here, again, a complete analogy fails us, 

 it is a solid which separates. As we know nothing of the osmotic 

 ssure of a syrup, the analogy is a defective one ; but it is probable 

 that a dilute solution of sugar would pass continuously into a syrup 

 :>f pure sugar by evaporation of the solvent, and analogy would lead 

 the supposition that the syrup coincides with the unstable state of 

 le liquid. I would, therefore, offer the analogy between the syrupy 

 md the supercooled states as a tentative one ; it lacks foundation in 

 3th cases. 



One point remains to be mentioned. I have for the past nine 

 lonths, in conjunction with Mr. Edgar Perman, been determining 



