JACQUES LOEB 259 



at first rise sharply with an increase in concentration and the more 

 so the higher the valency of the anion. The maximum of the curves 

 is reached at about m/256, then the curves fall with a further increase 

 in the concentration until a minimimi is reached again at about m/16, 

 and then another rise begins. This second rise expresses the gas 

 pressure effect of the solute. When we use membranes treated with 

 gelatin, we get a similar system of curves for the same solutions,^ 

 and the explanation of the curves is the same for both kinds of mem- 

 brane. The particles of water diffuse through the membrane as if 

 they were positively charged, being attracted by the anion of the salt 

 and repelled by the cation, the attraction increasing with the valency 

 of the anion. That the repulsion increases with the valency of the 

 cation is shown in Fig. 2 where the lower curve represents the rate of 

 diffusion of water in 20 minutes into MgCl2 solutions and the upper 

 curve the initial rate of diffusion of water in 20 minutes into solutions 

 of LiCl through membranes not treated with gelatin. These curves 

 are also practically identical with those obtained for the same salt 

 solutions when the collodion membrane had previously been treated 

 with gelatin. 1 Hence Rule 1 holds in all essentials equally for mem- 

 branes treated and not treated with gelatin. 



The situation is altogether different for solutions of those elec- 

 trolytes whose influence is described in Rule 2; namely, acids and 

 neutral or acid solutions of salts with trivalent or tetravalent cation. 



When we separate solutions of different concentrations of AI2CI6 

 from H2O by collodion bags treated with gelatin, water diffuses very 

 rapidly into the solution and the level of Uquid in the manometer 

 rises steeply with an increase in concentration, as is shown by the 

 upper curve in Fig. 3. Water is negatively charged and is powerfully 

 attracted by the trivalent cation Al. When we repeat the same 

 experiment with membranes not treated with gelatin (lower curve of 

 Fig. 3), we notice that in that range of concentrations of AloCls where 

 the diffusion is determined chiefly (or exclusively) by electrical forces 

 no rise occurs until the concentration of the solution of AI2CI6 is about 

 m/64; at about this concentration the gas pressure effect of solutions 

 of cane sugar begins to be noticeable. It, therefore, looks as if solu- 

 tions of AI2CI6 showed no electrical but only the gas pressure effect 

 when separated from pure water by a membrane not treated with 



