214 
CHEMISTRY: J. LOEB 
Proc. N. a. S. 
In the case of metal gelatinate the particles of water diffusing through 
the membrane are positively charged and attracted by the gelatin anion 
and repelled by the cation with which the gelatin is combined. This re- 
pelling effect is greater in the case of bivalent cations (Mg, Ca, Ba, etc.), 
than in the case of monovalent cations (e.g., Li, Na, K, NH4). Hence 
the initial rate of diffusion of water into solutions of metal gelatinates is 
greater in the case of sodium gelatinate than in the case of calcium gela- 
tinate. When we add an electrolyte to the solution of a gelatin salt the 
increase in concentration of electrolyte brings about the depressing effect 
of that ion which has the same sign of charge as the electrified watery phase, 
as expressed in Rule c. 
The permanent osmotic pressure of a solution separated from pure 
water by a semipermeable membrane is that pressure which has to be ap- 
plied to the solution side in order to cause equal numbers of particles of 
water to pass simultaneously through the membrane in opposite directions. 
It follows from this that the relative influence of electrolytes must be the 
same on the permanent osmotic pressure as on the initial velocity of diffu- 
sion through the membrane from pure water to solution, and this was found 
to be the case with the solutions of gelatin salts. 
IV. The terms "attractive and repulsive effect" of the ion on the charged 
particles of water served the purpose of simplifying the presentation 
of the facts observed. In reality the "attractive and repulsive effect" 
of ions as expressed in Rules a, h, and c are the expression of an additive 
effect of the oppositely charged ions of an electrolyte on the density of 
charge of the double layer formed at the boundary of collodion mem- 
brane and watery phase. This was proved by experiments on electrical 
endosmose. When the collodion membrane is bounded on both sides by 
identical solutions, an equal number of particles of water will diffuse in 
opposite directions through the collodion membrane. When, however, 
an external potential difference is produced oti the two sides of the mem- 
brane, an electrical endosmose will be established to one of the two elec- 
trodes according to the sign of the electrification of the watery phase in 
contact with the collodion membrane. By this method I ascertained the 
influence of electrolytes on the sign of charge on the watery phase which 
was expressed in Rules a, b, and c. 
Cases of abnormal of osmosis, e.g., the fact that liquid diffuses 
through parchment membranes from oxalic acid into pure water had been 
known for more than 50 years and it had been suggested by several 
authors, Girard, Bernstein, Bartell and Hocker, and Freundlich,^ that 
potential differences on both sides of the membrane caused these anomalies. 
To prove this view it was necessary to measure these potentials. Direct 
measurements of such potentials have been made by Girard, by Bartell 
and Hocker,^ by T. Hamburger,^ and by the present writer (as yet un- 
published), but the results have thus far not been satisfactory. One can, 
