396 CAUSE OF INFLUENCE OF IONS. I 



Moreover, according to Perrin, an increase in the valency of the 

 cation should not increase the rate of diffusion of water from pure 

 water into a slightly acid solution of salts, yet a glance at Fig. 2 

 shows that the initial rate of diffusion of water through a collodion 

 membrane into a slightly acid solution increases considerably with 

 the increase in the valency of the cation of the salt added. 



In former papers,'- ^'^ further proof of these statements can be 

 found and in addition the writer has shown that in neutral solutions 

 (where the water diffusing through the collodion membrane is posi- 

 tively electrified) the rate of diffusion of water into the solution in- 

 creases with the valency of the anion of the electrolyte in the solution. 



This discrepancy between the actual observations concerning the 

 influence of electrolytes on the rate of diffusion of water through a 

 collodion membrane in free osmosis on the one hand, and the theory 

 of Perrin concerning the influence of electrolytes on the rate of dif- 

 fusion of liquid in the case of electrical endosmose on the other indi- 

 cates that either the influence of electrolytes is not the same in both 

 cases or that the theory of Perrin is not the correct expression of the 

 facts in the case of electrical endosmose, at least for collodion mem- 

 branes. It seemed, therefore, necessary to test the influence of elec- 

 trolytes on the rate of transport of water through collodion membranes 

 by electrical endosmose as a first step towards a theory of the influence 

 of electrolytes on free osmosis. 



III. 



In the experiments on electrical endosmose we used the collodion 

 bags which served for the experiments on free osmosis. These collo- 

 dion bags were cast inside an Erlenmeyer flask of a volume of about 

 50 cc. The collodion flask was closed with a rubber stopper which 

 was perforated by a glass rod serving as a manometer. The bag 

 was filled at the beginning of the experiment with the solution whose 

 influence on the osmotic transport was to be investigated and was 

 put into a beaker containing the identical solution. The bag was 

 then so adjusted that the upper level of the rubber stopper was at the 

 surface of the liquid in the beaker and that a column of liquid of about 

 30 mm. in the manometer was above the level of the liquid in the 

 beaker. The manometer was a glass tube with a bore of about 2 



