390 Proceedings of the Royal Society of Edinburgh. [Sess. 
The ionisation of the KI is determined by interpolating on Kohlrausch’s 
latest results. It will be noted that the agreement between the found and 
calculated is very fair. 
I also tried the effects of introducing in the cell between the stoppered 
vessels a solution of potassium iodide in glycerine. This very slightly 
affected the E.M.F. readings, small changes being caused by altering the 
concentration of the potassium iodide in the glycerine. 
It is evident from these results that the concentrations which are in 
equilibrium are about 1 to 20. Now the solubility of potassium iodide in 
alcohol is *096 molecules, and in water it is 6'2 molecules, so that there is 
here a big departure from the partition coefficient which might have been 
expected. 
To throw further light on this discrepancy, experiments were next 
made with concentrated solutions. 
The saturated solution of potassium iodide in alcohol was prepared by 
dissolving potassium iodide in excess in alcohol at a higher temperature, 
letting it cool in the thermostat, and shaking the solution up with the 
deposited crystals at intervals for several days, the solution being kept in 
the thermostat. This solution poured off the crystals was saturated with 
silver iodide at the same temperature. 
Another alcohol electrode was prepared by taking this solution and 
packing the stoppered electrode with dried potassium iodide and silver 
iodide and then introducing the same solution. 
This electrode usually gave against the first an E.M.F. of '015 volts. 
This E.M.F. is worthy of further investigation. 
The preparation of the water electrode was more difficult. 
If crystals of potassium iodide and silver iodide are introduced into a 
saturated potassium iodide solution, both dissolve, a liquid being obtained 
on which potassium iodide crystals float, and further addition of both 
substances finally produces a stiff paste. 
It is therefore difficult to define what is meant by a saturated potassium 
iodide solution under these circumstances. 
I determined, therefore, to prepare a saturated solution by repeated 
shaking in the thermostat with a large excess of potassium iodide crystals, 
then pouring off this solution and introducing an excess of silver iodide into 
it. This is the solution used in the following experiments. 
The paste above described gave a slightly higher E.M.F. than the above 
solution. The following curves represent the E.M.F. obtained by reading 
a series of water solutions against the two alcohol electrodes, described 
above, and also with the interposition of glycerine. 
