728 ELECTRIFICATION AND DIFFUSION OF WATER 



the presence of SO4 ions retards the rate of diffusion of water in com- 

 parison with CI or NO3 ions and this supports our conclusion that 

 HoO acts towards the solution of salts of trivalent cations as if its 

 molecules were negatively charged. 



The reverse is true when the cation of a salt is bivalent. When 

 the outside solution was m/128 MgS04 or m/128 CaS04 it required 

 about 75 minutes for the pressure head of H2O inside the bag to fall 

 from 120 mm. to about 30 mm. When the sulfates were replaced by 

 the chlorides it required much higher concentrations to produce the 

 same rate of diffusion. In the case of MgCl2 and CaCl2 it required 

 a concentration of approximately m/8. This is to be expected when 

 the molecules of water are positively charged since in this case the 

 attractive power of SO4 should be superior to that of the chlorides. 

 These statements are supported by Figs. 5 and 6 illustrating the dif- 

 ference here discussed, 



///. Modification of the Method. 



We had measured in the preceding experiments the rate of diffusion 

 of distilled water under an initial pressure head of about 110 mm. of 

 water against solutions of various electrolytes. When we reverse the 

 procedure, putting the solution of electrolyte inside of the bag and 

 the distilled water outside, the latter will diffuse into the bag and at 

 the same time the solution will commence to diffuse outside. Since, 

 however, the concentration of the solute is at the beginning greater 

 inside than outside, water will at first diffuse into the bag and the 

 liquid in the manometer tube will rise. After 1 hour or more the 

 maximum height in the glass tube is reached and the level of liquid 

 in the glass tube commences to fall again. By comparing the height 

 to which the liquid in the tubes rises in about 70 minutes we shall 

 get not a quantitative measure, but a qualitative indication of the 

 relative influence of the various types of ions on the rate of diffusion 

 of water; for we shall show in the next chapter that the differences ob- 

 served are the expression of differences in the rate of diffusion of water 

 into the solution and not the expression of differences in the rate of 

 diffusion of the solute into the distilled water. We used m/128 

 solutions, since we found that m/64 and even m/32 solutions of cane 



