Supersaturation of Salt-Solutions. 459 
Again, J.8. Thompson*, in experimenting with potash-alum, 
found that a supersaturated solution, prepared by saturating 
water at 90° C. with the ordinary hydrate, deposited no 
crystals on cooling to the ordinary temperature ; but that a 
solution saturated at 95° C. deposited small clear crystals. 
On breaking the flask in ice-cold water and washing the 
erystals, he found on analysis that they were ordinary alum 
with 24 molecules of water. This was of course to be ex- 
pected as the natural result of his mode of experiment, for the 
anhydrous salt or a lower hydrate would at once take up 
water to form the normal hydrate ; while it is equally evident 
that the crystals could not have been the normal hydrate, or 
they would have caused crystallization of the supersaturated 
solution in which they were formed. 
There is therefore, it appears to me, little doubt that the 
so-called supersaturated solutions furnish a further proof of 
my contention that a salt exists in solution not as a hydrate 
definite or indefinite, but in the anhydrous state. This may 
appear a curious statement, but I fail to see what other 
explanation can be given of the water of crystallization 
possessing the same volume as the solvent water, as has been 
proved by the experiments by Ostwald and myself on the 
molecular volumes of solutions of nearly two hundred salts of 
various metals. This will not appear so strange when it is 
remembered that it is already pretty generally admitted that 
double salts as a rule exist only in the solid state, combina- 
tion taking place only at the moment of crystallization from 
solution. I have already (loc. cit.) pointed out the fallacy 
involved in the argument from the heat of hydration ; and 
the only difficulty that remains is the colour of certain salts in 
solutions of various degrees of concentration, or in the solid 
state. Thus anhydrous copper sulphate or the monohydrated 
salt is colourless, while the pentahydrate and the solution is 
blue. Again, cobalt chloride is red when hydrated and solid, 
blue when dehydrated; on the other hand, a strong solution 
of this salt is red when cold, blue when hot. This has 
always been regarded as a conclusive proof of the existence 
of the hydrated salt in at least dilute and cold solutions. But 
it is more than probable that the whole secret of the above 
colour-changes lies in the difference in the aggregations of 
molecules and not in the amount of hydration of the salt. 
We do not in the least know what multiple of the usual 
formula for a molecule forms the individual in the solid state 
or in solution ; and it would be by no means a forced explana- 
tion of the above colour-changes to attribute them to the 
* Journ. Chem. Soc. 1882, vol. xli. p. 382. 
