726 ELECTRIFICATION AND DIFFUSION OF WATER 



But this is only true as long as we deal with neutral salts with uni- 

 valent and bivalent cation. When we deal with neutral salts with 

 trivalent or tetravalent cations the rule is suddenly reversed. Chlo- 

 rides or nitrates of salts with trivalent or tetravalent cations attract 

 water considerably more than the salts with bivalent or even with 

 monovalent cation. This discontinuity of effect with increase of 

 the number of charges of the cation led the writer to the conclusion 

 that water behaves towards salts with trivalent and tetravalent cations 

 as if its molecules were negatively charged. On account of the high 

 charge of the trivalent and tetravalent cations the attraction for water 

 should be very powerful, as it indeed is. Of course, many of these 

 solutions are strongly acid on account of the hydrolytic dissociation, 

 but this is true only to a slight degree for Ce2Cl6 or La2Cl6. Yet a 

 m/512 solution of these two salts attracts water almost as powerfully 

 as does m/192 Na2S04 of the same hydrogen ion concentration. 



That water is attracted by salts with trivalent and tetravalent 

 cations as if its molecules were negatively charged becomes obvious 

 from a comparison of the effects of the chlorides and sulfates of these 

 salts. Since the anion of a salt should repel negatively charged water 

 — and the more so the greater the valency of the anion — the attrac- 

 tive power of a trivalent cation for water should be less in the case of 

 a sulfate than of a chloride. Hence m/512 CrzClg should have a 

 greater attraction for water than m/320 Cr2(S04)3. This is indeed 

 the case. It was found that water in the bag diffused under a pres- 

 sure head of about 110 mm. of water more rapidly towards m/512 

 AlaCle or CraCle than towards m/320 AlaCSO^s or Cr2(S04)3. In a 

 m/2048 solution of AI2CI6 it took only 20 minutes for the water in 

 the manometer to fall from a pressure head of 105 to that of 10 mm.; 

 while a m/32 to m/16 solution of Al2(S04)3 was required to bring about 

 the same rate of filtration. 



It required about 40 minutes for the water to fall from the pressure 

 head of 105 mm. to that of 20 mm. when the outside solution con- 

 sisted of a M/2048 solution of CrgCle. When Cr2(S04)3 was used in 

 the beaker, even a m/16 solution of the sulfate did not act as strongly 

 as a m/2048 solution of the chloride, although the latter was at least 

 partly in suspension. The nitrates of these salts behaved like the 

 chlorides. It is, therefore, obvious that in the case of trivalent cations 



