( 31 ) 



Both tliese results may be deduced directly. If cj tends to zero, 

 uo differences of concentration of the free ions are possible and the 

 velocities must be equiiliseil by the action of electromagnetic and 

 electric forces alone. 



The equation 



(£"., U HAr E,) U r= {E^ V H — E.) V 



immediately gives the relation (3). 



In a completely dissociated solution on the contrary no molecules 

 of salt can go back by diffusion, so the velocities of the ions in the 

 direction of the axis of Z should both become equal to zero ; hence 



„ A dc-, A dc-, 



E^U H-}-E- i = and E^ V H — E, -^ := 



C] dz Cj dz 



from which the relations (4) follow immediately. 



As a matter of fact, the reaction-velocities are not infinitely great. 

 By supposing these velocities to be equal to zero we may however 

 perceive that the result is not very much altered by this. In this 

 case no decomposition or combination occur; the concentration of 

 the molecules of salt is not altered, these have therefore no influence 

 at all upon the phenomenon. So we then find, whatever the real 

 degree of dissociation may be, the same result as in the case of 

 complete dissociation. 



Hence it is very likely that the IlALLeffect in liquids may ge- 

 nerally be represented by 



§'=-/.//(f/- V) 



f( lying between 1 and ^. 



This result gives the rotation of the equipotential lines in the 

 inner part of the liquid. The question rises, whether this is what 

 has been observed in the experiments. In these namely the difference 

 of potential between two electrodes of metal is read ; if these elec- 

 trodes are immersed in liquids of different concentration, another 

 difference of potential will appear, the value of which may be 



