466 BEPOET— 1886. 



formation of peculiar hydrates containiDg less water, and more soluble 

 than the normal salt. On the other hand, Lowel, in his last memoir 

 (' Ann. Chim.' [3] xlix.), recants his earlier belief, and definitely expresses 

 himself in favour of the opinion that a supersaturated solution (referring 

 specially to sulphate of sodium) contains the anhydrous salt, and that no 

 solution can be properly called supersaturated. 



' Dans toutes les dissolutions, quelque riches qu'elles soient, qui ne 

 sont pas en contact avec un exces de cristaux a 10 HO ou a 7H0, les 

 molecules salines dissoutes restent a I'etat de sel anhydre, malgre les 

 variations de temperature, si elles sont preservees de cette action 

 mysterieuse de contact que ' Fair atmospherique et d'autres corps ont 

 la propriete d'exercer sur elles en determinant la formation de cristaux a 

 lOHO, et si leur temperature ne tombe pas a un degre suffisamment bas 

 pour determiner la formation spontanee de cristaux a 7H0 ' ('Ann. 

 Chim.' [3] xlix. p. 56). 



Tomlinson, de Coppet, and other writers have since adopted the same 

 ■views. 



De Coppet showed (' Compt. Rend.' Ixxiii. 1324) that a supersaturated 

 solution of sodium sulphate may be formed by dissolving in cold water the 

 anhydrous salt, provided the latter had been heated above 33°, and preserved 

 from contact with the dust of the air. He also succeeded in preparing 

 supersaturated solutions of sodium carbonate, and magnesium sulphate 

 by dissolving the dehydrated or partly dehydrated salt in cold water. 



Nicol (' Phil. Mag.' June and Sept. 1885) has made similar observa- 

 tions, and has also shown from density determination of sodium sulphate 

 and thiosulphate of various strengths that in passing the ordinary satu- 

 ration point there is nothing to indicate any change in the constitution 

 of the solution. He therefore concludes that a so-called supersaturated 

 solution is merely a solution saturated or non-saturated of the anhy- 

 drous salt, and that in this respect it differs in no way from an ordinary 

 solution which is not capable of supersaturation. Nicol also considers, 

 as Lowel seems to have done, that any solution of a hydrated salt con- 

 tains no hydrate, but that combination between the salt and the water 

 takes place at the moment of crystallisation. 



I thought at one time that supersaturation resulted from dissociation 

 of the dissolved salt into water and the anhydrous salt, owing to the 

 action of the higher temperature — above 33° — to which, by the ordinary 

 process of making supersaturated solutions, the liquid is exposed. Bat 

 before the experiments of De Coppet and Nicol had come to my know- 

 ledge I had satisfied myself that this was not so ; for on saturating a 

 solution at the temperature of the air with crystals of sodium sulphate, 

 and then filtering the solution into a stoppered bottle and cooling to about 

 0°, the cold solution exhibits all the phenomena of supersaturation so long 

 as it is kept at that lower temperature. This is difficult to explain on any 

 hypothesis of dissociation. 



Supersaturated solutions are at present only known to be formed by 

 hydrated salts, or by anhydrous salts only at such low temperatures that 

 hydrates are formed, e.g., common salt. I am inclined to the belief 

 that this is due to their much greater fusibility. If we compare the melt- 

 ing-points of those salts which are known to give supersaturated solutions 

 readily with the melting-points of those which, although easily soluble, do 

 not give supersaturated solutions under any conditions that have yet been 

 tried, we see this difference plainly. 



