1910-11.] Temperature Coefficient of Concentration Cells. 389 
lower value, but a higher reading for this dilution was never obtained, 
though a slightly higher value was indicated for a still greater dilution. 
The results for higher dilutions were, however, very unstable, and 
readings for greater concentrations gave lower E.M.F.s. 
These results are to be expected when we remember that potassium 
iodide ionises far more rapidly on dilution in water than in alcohol, the 
ionising power of alcohol being supposed to be about one-third that of 
water. 
On the assumption that a ’0001 solution in alcohol might be regarded 
as completely ionised, I calculated back the ionisation of the alcoholic 
solution and found the results in good general agreement with the observa- 
tions of others, giving for *01 molecules 34% ; *005, 39%; ‘001, 67%. 
If the experimental difficulties could be overcome, these cells would 
be of value as giving another method of determining the ionisation of 
organic solvents. 
Starting with a cell consisting of •001 alcohol solution and *001 water 
solution, I altered the water concentrations, and the results are given in 
the following table : — 
Table II. 
Molecules 
K.I. 
in 1000 cc. 
Alcohol. 
Molecules 
K.I. 
in 1000 cc. 
Water. 
E.M.F. 
found. 
E.M.F. 
calculated. 
•001 
•001 
= 
•080 
•001 
•0025 
= 
•049 
•057 
+ ve 
•001 
•005 
= 
•034 
•039 
+ ve 
•001 
•0075 
= 
•025 
•029 
+ ve 
•001 
*01 
= 
•017 
•021 
+ ve 
•001 
•02 
= 
•o 
■005 
+ ve 
•001 
•04 
= 
•016 
•013 
- ve 
•001 
•06 
•027 
•023 
- ve 
•001 
•08 
- 
•033 
•030 
- ve 
•001 
T 
= 
•041 
•041 
- ve 
T.C. 
for -001 al. - 
•001 water = 
•00092 + ve 
T.C. 
for '001 al. - 
•04 
water — 
•00072 -ve 
The calculations were made from the equation 
E = RT (log C x x - log C') 5 
where x is a constant depending on the partition coefficient, and treated as 
a concentration. 
A correction is made for the ionisation of the water solution of the 
potassium iodide. 
