Electromotive Force of Diphasic Liquid Systems. 99 
systems. By a curious misinterpretation he regards their theories as 
mutually exclusive, and attempts to decide between them ; whereas both 
recognize the possibility of potential differences arising at all three boun¬ 
daries, and both attempt, by an appropriate adjustment of the conditions of 
experiment, to eliminate certain of these—Haber the diffusion potential at 
C by the interpolation of a phase of constant composition : Cremer the 
interphase potentials at A and B by similarity of composition in the two 
aqueous solutions. How far these devices are successful is of course a 
matter for criticism, and Beutner justly points out that the results of 
measurements of potential in Cremer’s nitrobenzol-picric acid system do not 
bear out his theory, as is recognized by Cremer himself. 1 
Beutner denies the existence of a diffusion potential in the oil systems 
with which he deals on the ground that the addition of such a substance as 
salicyclic acid to the oil phase increases the e.m.f., whereas on the hypothesis 
of a diffusion potential the presence of excess of electrolyte on both sides 
should tend to obliterate it. In the light of the considerations discussed 
above, it will be seen that this argument has little force, and on theoretical 
grounds Beutner’s general proposition, that no diffusion potential exists 
between solutions of different concentrations, is entirely inadmissible. 
The total e.m.f. of ‘oil ’ systems such as the above must therefore be 
the algebraic sum of the following : 
1. The interphase potentials at A and B. These will be determined 
by the relative concentrations of electrolytes in the layers adjacent to each 
boundary, and in certain cases their sum may be expressed by the formula 
C 
R 7 log —7, where C 15 C 2 represent the concentration in ‘ oil ’ of salts present 
in equivalent concentration in the two aqueous solutions. 
3. The diffusion potential due to difference of concentration and to 
difference in the nature of the ions on each side. Where the same electro¬ 
lyte is present in different concentrations this potential can be calculated by 
1 Cremer found the system 
A C B 
M /5 NaCl 
Nitrobenzene 
Nitrobenzene 
sat. with nitrobenzene 
sat. with NaCl and 
sat. with NaCl 
and picric acid 
picric acid 
M /5 NaCl 
sat with nitrobenzene 
to be positive on the acid side. He regards ihe interphase potentials at A and B as equal and 
opposite, and ascribes the effect to diffusion potential at C in spite of the difficulty of maintaining 
this explanation. Beutner ascribes it to a difference of interphase potential at A and B arising irom 
the greater solubility of sodium chloride in nitrobenzene containing acid, and quotes conductivity 
measurements to show that the solubility ol sodium chloride in nitrobenzene is increased by the 
presence of picric acid. This argument is entirely inconclusive, for hydrochloric acid must be 
formed by the reaction of sodium chloride and picric acid, and a great part of the increased con¬ 
ductivity must be due to this cause. In the opinion of the writer, the effect is in all probability due 
to the fact that while picric acid is only slightly soluble in water, hydrogen ions are highly soluble. 
On this account an excess of hydrogen ions may easily be present in the aqueous phase even when 
the nitrobenzene phase contains only a very small quantity of dissociated acid. 
