JACQUES LOEB 



193 



through the membrane in the form of negatively charged molecules, 

 which according to our theory are attracted by the Al ion and repelled 

 by the anion, the repelHng action increasing with the valency of the 

 anion. The curves show that AI2CI6 attracts water very violently, 

 AI2 (804)3 much less, and that aluminium citrate acts like a solution of 

 a non-electrolyte, the attractive and repelKng action of the two tri- 

 valent ions, Al and citrate, balancing each other. But no negative 

 osmosis was noticed. This shows that the electrostatic effect of the 

 H ion upon the rate of diffusion of water is considerably smaller than 



t~i 



o 



o 

 h4 



+20 

 



-20 

 -40 

 -60 

 -80 

 -100 



HHilnri n n n n 



8192 4096 2048 1024 51Z 256 1E8 64 32 



Concentration 



n 



Fig. 12. Curves showing negative osmosis in the case of Ca(0H)2 and Ba(0H)2 

 occurring in that range of concentrations where the drop in Fig. 3 occurs. 



the electrostatic action of trivalent cations Kke Al, while the electri- 

 fying effect of the H ion upon the water seems to be more nearly equal 

 to the electrifying effect of the trivalent cation upon water. This sug- 

 gests that the electrifying effect of ions upon water and their electro- 

 static effect upon the rate of diffusion of water are due to different 

 quahties of the ion. 



The curves for aluminium salts in Fig. 13 show no drop. As a 

 matter of fact, a sKght drop can be demonstrated in the AICI3 curves 

 but in concentrations higher than those in Fig. 13, namely about m/32 

 or m/16, a phenomenon which finds its explanation perhaps in the 

 fact that the cation is trivalent and the anion only monovalent. 



