104 
Walter Stiles 
where and c 2 are the concentrations of the ion in the two respective 
phases, n the valency of the ion, F the electric charge of a monovalent 
gram-ion, and k a constant. It is outside the scope of this work 
to describe how these formulae are derived; those interested should 
consult the original work of Nernst (1889, 1892) and Haber (1908) 
and the general account of electrical phenomena at surfaces given 
by Michaelis (1914). For the application of these formulae to bio¬ 
logical phenomena reference may be made to the papers of Beutner 
(1912, 1913) and Loeb and Beutner (1912). 
It should be emphasized that the difference of potential (phase 
potential) arises from the unequal partition coefficients of the two 
ions between two phases, and is not connected with the different 
mobilities of the two ions. A difference of potential due to this 
latter cause (diffusion potential) arises when two solutions containing 
the same ions but in different concentrations come into contact. 
Diffusion takes place and if the mobilities of the two ions are 
different one ion will diffuse faster than the other and a difference of 
potential will thus result. 
It appears however that the sign of the charge should be the same 
in whichever of the two ways the difference of potential arises, for in 
the latter case the solution takes the charge of the more mobile ion 
and in the former case it takes the charge of the more soluble one, 
and it appears that the more soluble ions are also the more mobile 
(Michaelis, 1914). Probably the electric charge at most surfaces can 
be accounted for in one of the preceding ways, but there are cases, 
as for instance that at the surface of aniline in contact with water 
(Ellis, 1912) where the aniline is negatively charged, although one 
would expect it to be positively charged as it feebty dissociates into 
the slightly mobile aniline ion C 6 H 5 .NH 3 and the very mobile hy¬ 
droxyl ion. Although explanations have been offered of such cases 
(Lewis, 1910) the problem cannot be regarded as solved. 
We may now pass on to a consideration of how adsorption is 
affected by electrical phenomena. 
In the first instance we may consider a case of an electrolyte 
partly dissociated into its constituent ions. There will then be in 
solution the kation, the anion and the undissociated molecule all 
with their characteristic constants in regard to the adsorption equa¬ 
tion, so that they tend to be adsorbed to different extents. Any 
difference in adsorption of the kation and anion must however be 
very slight as this would result in a potential difference between the 
surface and the interior of the liquid. An equilibrium position would 
