Affinity and Fleat. 313 
the period of solution and extension, will reappear and be 
given out during concentration and crystallisation. Hence, 
it might be said that these solutions contain (and that will 
only be true potentially) all the heat which water and the 
salt have absorbed during the development of the pheno- 
menon. 
Thus, a body which cools spontaneously does really be- 
come heated by all the latent heat which its molecules fix ; 
and it may be conceived that in an almost indefinite state 
of extension it might even be decomposed by the heat 
which it has absorbed at each addition of the solvent. In 
this manner is to be explained the decomposition by diffu- 
sion produced by Mr. Graham, and which serves as the 
basis of his admirable method of dialysis. 
If you put bisulphate of potash in the internal diffusion- 
vessel (two cylindrical concentric vessels, Graham’s first 
apparatus), you know that the illustrious physicist shows you 
sulphuric acid separating from neutral sulphate of potash 
and passing into the external vessel. 
Then we have a true decomposition which is necessarily 
accompanied by the absorption of a certain quantity of 
heat. Let us inquire to what itis due. If we add sul- 
phuric acid to sulphate of potash, both in dilute solution, 
the liquid will become heated; but the quantity of heat 
thus produced will always be less than the heat of contrac- 
tion, calculated in the manner I have shown. In fact, there 
will be a transformation of the heat which should have 
become sensible during the contraction, into latent heat; 
and when this has become considerable enough, decompo- 
sition by diffusion will take place. 
In order that the inverse of the original effect may take 
place, the combination which is destroyed must take up 
the quantity of heat which became sensible at the time it 
was effected. It is the latent heat, stored up during solu- 
tion which furnishes it; but it is inadequate for completely 
producing this effect, and thus the phenomenon is only 
partial, as Mr. Graham has well observed. 
The phenomena of decomposition by diffusion take 
place therefore in a continuous manner, like the decompo- 
sition of gases by heat, and all that I am about to say re- 
specting dissociation and its tension (which might here be 
replaced by the ratios of weights) is absolutely applicable 
to them. Whether diffusion takes place in two concentric 
vessels as I have supposed, or in Graham’s new membrane 
apparatus, the process is the same. 
Dy be2 
