Vol, 6, 1920 
CHEMISTRY: J. LOEB 
215 
however, test the idea indirectly by experiments on electrical endosmose. 
This method of verification seemed to be confronted with a serious diffi- 
culty. The influence of ions on electrical endosmose has been investigated 
by Perrin^^ who found that only one of the oppositely charged ions of an 
electrolyte, namely, the one which has the same sign of charge as the 
watery phase, has any influence (namely, a retarding one) on the velocity 
of endosmose, while the ion with the opposite charge has no influence. 
In my experiments on common osmose both ions of an electrolyte influ- 
enced the rate of diffusion of water simultaneously but in an opposite sense 
(as stated in Rules a, 6, and c). I have investigated the influence of ions 
on the rate of electrical endosmose and found that Perrin's rule does not 
hold for collodion membranes, but that the Rules a, 6, and c express 
not only the influence of ions on the transport of water through collodion 
membranes in common osmosis but also in the case of electrical endosmose. 
According to the formula of Helmholtz as modified by Perrin, we know 
that in the case of electrical endosmose the following relation holds : 
where v is the quantity of liquid carried electro-osmotically, e is the potential 
difference between the two strata of the double layer, E the external 
electromotive force, D the dielectric constant of the medium, r; the coeffi- 
cient of internal friction, and / the distance of the external electrodes. 
In our experiments all quantities occurring in the formula except v 
and € were kept approximately constant. If the theory of Helmholtz is 
correct, we must, therefore, attribute the influence of ions on the rate of 
transport of water in electrical endosmose to their influence on e, i.e., 
the density of charge on the watery phase; and since, moreover, my ex- 
periments show that the influence of ions is the same for electrical and for 
free endosmose through collodion membranes we must conclude that the 
influence of ions on the initial rate of diffusion and on the osmotic pressure 
of a solution is due to the influence of ions on the density of charge in the 
double layer at the boundary of watery phase and membrane. 
V. This permits us to define more accurately the influence of ions on 
osmotic pressure. At the boundary of a membrane and liquid an elec- 
trical double layer is formed. In the case of collodion membranes the 
latter usually assumes a negative charge while the watery phase assumes 
a positive charge. When the collodion membrane is bounded on one side 
by pure water on the other by a solution of an electrolyte the anions of 
the latter increase the negative charge on the membrane and the cations 
diminish it, both effects increasing with the valency and another property 
of the ions. At lower concentrations of the electrolyte the influence of 
the anion increases more rapidly with increasing concentration of the elec- 
trolyte than the depressing effect of the cation, while at higher concentra- 
