386 
Proceedings of the Boyal Society of Edinburgh. [Sess. 
Substituting this value in equation (1), 
<fe = (V-\) Ta ~ Tl ,V-X = t'^. 
Thus we find that the source of energy in this cell when the temperature 
is changed is equal to the difference between the latent heats of the salt in 
the two solutions. We can now consider in detail the effect of temperature 
on such a cell. 
Let us suppose that in raising the temperature of the cell the 
solubility of S' increases faster than S. Then it will be necessary to 
transfer salt from S to S' to restore the equilibrium, and there will be an 
electromotive force in the direction necessary to transfer salt from A to B. 
But if the temperature coefficient of S' is greater than that of S, then 
X' is greater than X, and therefore during the running of the cell heat is 
absorbed = X' — X, and the cell is endothermic. 
If the temperature, however, is lowered below T v then, since S' has the 
higher temperature coefficient, its solubility will now diminish quicker than 
the solubility of S, and the current will be reversed and flow the other way, 
d e 
so that now heat will be set free instead of absorbed, and ^ has a 
negative value ; that is, the E.M.F. diminishes with rise and increases with 
fall of temperature. 
I shall now proceed to describe in more detail the experiments made 
with the alcohol, water, potassium iodide cells, with silver silver iodide 
electrodes. 
It will be remembered that the general investigation given above will 
probably only apply with a modification, owing to the diffusion of alcohol 
into water and salt precipitation. 
The type of cell used was a modification of the type used in my 
experiments on iodine concentration cells, little vessels about two cms. long 
and half a cm. broad, stoppered at one end and with a fine open tube at 
the other for the introduction of the silver wire, being used. These fine 
tubes were passed through a cork, the stoppered ends dipping into a bottle, 
as shown in the diagram on p. 387 : — 
The iodised silver wires were then introduced and cemented in with a 
little paraffin wax. 
As in my former experiments, a Dolezalek electrometer was used to 
measure the E.M.F., the E.M.F. of the cell to be measured being balanced 
against the E.M.F. of a storage cell through a resistance, and the storage 
cell being standardised from time to time against a standard cadmium 
cell. 
