. JACQUES LOEB 393 



His theory which has revolutionized colloid chemistry is expressed 

 in the following statement. 



"The primary factor of this electrification is always the action of the hydrogen 

 or the hydroxyl ions, which are pressed against the membrane in the same way, 

 no matter whether polyvalent ions are present or absent. But if polyvalent ions 

 of the opposite sign are present they are attracted towards the membrane. To be 

 more precise, let us suppose a liquid with a monovalent acid; hydrogen ions cover 

 the membrane with a positive charge according to the degree of acidity. Behind 

 them, at a distance which results from an equilibrium between the osmotic and 

 the electric forces are found the monovalent negative ions forming the sec- 

 ond stratum of the double layer. If we now add negative polyvalent ions, e.g. 

 Fe(CN)3, the osmotic forces acting on the tetravalent Fe(CN)6 remain of the 

 same order as before while the electric force is multiplied by four; the density of 

 the double layer will therefore diminish and as a consequence the p.d. of contact 

 and the amount of electrical endosmose. It only remains to explain in a precise 

 manner reversion of the sign of charge due to the presence of the necessary 

 amount of Fe(CN)6 ions."^ 



Perrin's view on the effect of electrolytes on the double layer does 

 not agree with our experiments with collodion membranes which 

 show unequivocally that the influence of electrolytes on the rate of 

 diffusion of water in the case of free osmosis is an additive effect of 

 the two oppositely charged ions of an electrolyte, and not the effect 

 of only one of the two ions. 



If we assume that the influence of ions is the same in the case of 

 free osmosis and in electrical endosmose an increase in the valency of 

 the anion, according to Perrin, should not increase the rate of diffusion 

 of positively electrified water in free osmosis, since in this case the 

 membrane has the same sign of charge as the anion. A glance at Fig. 

 1 shows, however, that when we separate a solution from pure v/ater 

 by a collodion membrane the initial rate of diffusion of water into the 

 solution increases in the lower concentrations of different potassium 

 salts with increasing valency of the anion when the water is positively 

 charged. In this case a watery solution of one of these salts was 

 put into a collodion bag connected with a manometer and the bag 

 was put into a beaker containing distilled water. The solutions of 

 NaCl, CaCl2, Na2S04, and Na4Fe(CN)6 were rendered alkaline by dis- 

 solving the neutral salt in m/1,024 or m/1,000 KOH. The abscissae 

 are the logarithms of the concentration and the ordinates are the rise 

 in the level of solution in the manometer after 20 minutes. 



