JACQUES LOEB 565 



has the same sign of charge as the membrane and being diminished 

 by that ion which has the opposite sign of charge.^ Since the collo- 

 dion membrane is negatively charged in contact with the solutions 

 mentioned in Fig. 1, the density of charge on the membrane is in- 

 creased by the anion and diminished by the cation of the electrolyte. 



The relative influence of the oppositely charged ions of an electro- 

 lyte is, however, not the same for different concentrations. When the 

 membrane is negatively charged the influence of the anion on the 

 charge of the membrane increases in lower concentrations more rap- 

 idly with increasing concentration than the depressing effect of the 

 cation until the concentration is m/512 or m:/256. When this point is 

 reached, the depressing influence of the cation on the negative charge 

 of the membrane increases more rapidly with increasing concentration 

 than the influence of the anion. 



This explains why the curves representing the initial rate of dif- 

 fusion of water from pure water through a collodion membrane to a 

 solution of an electrolyte rise at first with increasing concentration 

 until the concentration is about m/512 or m/256 and then generally 

 drop with a further increase in concentration.' The rise of the curves 

 is higher the higher the valency of the anion and less the higher the 

 valency of the cation. In the case of CaCl2 the strong depressing 

 effect of Ca prevents a rise by the CI ion. 



It follows from the theoretical discussion in the preceding paper^ 

 that if this explanation of the influence of the concentration of the 

 electrolyte is correct, the rate of diffusion of water through a col- 

 lodion membrane must vary in the same sense with the concentration 

 of the electrolyte in the case of electrical endosmose as it does in the 

 case oifree osmosis. This is shown to be true by the curves in Fig. 2. 

 These curves represent the relative transport of water in the case of 

 electrical endosmose through a collodion membrane in the presence 

 of the same electrolytes as those used in Fig. 1. 



In these experiments the collodion flasks were filled with a solution 

 of an electrolyte and were dipped into a beaker containing an iden- 

 tical solution. A large disc of platinum at the bottom of the beaker 



2 Loeb, J., J. Gen. Physiol., 1919-20, ii, 387. 



' The second rise beginning with m/16 is due to the gas pressure effect and will 

 not be discussed in this paper. 



