394 Proceedings of the Roy a] Society of Edinburgh. [Sess. 
the saturated solution of the first solvent and the strongest solution of the 
second solvent, which is not precipitated by the first. 
If we examine Table II. it is evident that if the condition of equilibrium 
was the ratio of the saturated solutions, that ratio being 1 to 60, the 
E.M.F. zero would be shifted about — -027 volts. 
Taking curve (1), the shifting would be about — *04 volts, to give 
equilibrium at the saturation point of each solution. 
We may therefore express these results by saying that there is apparently 
a positive quantity of heat involved in the passage of KI from alcohol to 
water, which is subtracted from the latent heats of solution and which 
shifts the E.M.F. somewhere about ‘027 to ’04 volts, and which is not due 
apparently to the latent heats of solution of the salt but to some other 
cause. I discuss this more fully further on. 
Probably, of these figures the one obtained from the dilute solution is 
the more reliable. 
With reference to the small alteration made by the introduction of the 
glycerine between the electrodes, it must be remembered that ordinary 
glycerine was used, which doubtless contained water, and that therefore 
the alcohol was linked up through this contained water with the water 
electrode. The object of the introduction of the glycerine was to produce 
a big change in the surface contact conditions of the two solvents in order 
to test how far the slopes of concentration were affecting the results, 
according to the suggestions made by Abel. 
The E.M.F. which exists in the cell in which both the alcohol and 
water are saturated with KI seems to be only capable of explanation 
on the assumption that the ions are combined with water and also possibly 
combined with alcohol. For if the action of the cell be carefully considered, 
it is evident that as far as the two ends of the cell are concerned there 
is here no source of E.M.F., as at the two ends we have solid KI. Our 
only source, therefore, of electrical energy is in some way derived from the 
mixing of water and alcohol. Now, the mere diffusion of these two liquids 
into each other can hardly enter into the mechanism of the electrical 
process. By suitable arrangements the diffusion could be reduced to a 
minimum, and yet the electrical process would go on and must involve 
the using up of some energy within the cell. This difficulty is entirely 
explained if we suppose that the ions are combined with water and 
alcohol respectively, and carry their combined solvent into the other 
solvent, thus causing electrolytic mixing, the current flowing from the 
cell until this electrolytic mixing is complete, the interdiffusion of the 
two liquids simply resulting in a waste of some of the energy otherwise 
