382 SCIENCE PROGRESS. 



molecular conductivity becomes less. The work of the 

 founders of the " Dissociation " theory has at all events 

 convinced us that in such cases a part only of the salt 

 present is actively concerned in conveying the current, 

 whatever be the physical cause of the inactivity of the rest. 

 We are thus prevented from supposing simply that the ionic 

 velocities are reduced by the increased viscosity of the solu- 

 tion — an idea which an examination of viscosity data would 

 also upset. We must take a dynamical view of the ionisa- 

 tion equilibrium and (using for the sake of convenience the 

 language of the dissociation theory) imagine each ion some- 

 times free and therefore active, and sometimes combined 

 with an opposite ion and rendered inert. The ratio of the 

 time during which each ion is on the average free, to the 

 time during which it is combined, measures, as will at once 

 be seen, the fractional number of the molecules which are 

 at any instant dissociated. While it is free each ion is 

 urged forward by the electric forces with a velocity equal 

 to its specific velocity at infinite dilution (except for a small 

 charge introduced by the increased viscosity). While it is 

 combined it is at rest. We thus get an average velocity 

 which is reduced to the value required on Kohlrausch's 

 theory by the diminished molecular conductivity. 



The process of electrolysis in strong solutions of those 

 salts which have a low conductivity, and are therefore only 

 slightly "dissociated," is probably in reality a rather com- 

 plicated phenomenon. In order to explain the abnormal 

 migration constant of Cadmium iodide in alcohol, observed 

 by Hittorf, the idea of complex ions has been introduced. 

 In the case of this electrolyte, more salt is taken from the 

 solution round the kathode than is decomposed, so that the 

 anode solution actually gets stronger. This is explained by- 

 supposing that the ions in this case are Cd and I 2 (C^I 2 ), a 

 molecule of undecomposed salt being attached to the anion. 

 While measurements were being made with the solutions 

 of cobalt chloride and cobalt nitrate in alcohol mentioned 

 above, it was found that good results could only be obtained 

 with solutions whose strength was less than about 0*06 gram- 

 equivalent per litre. In the case of stronger solutions vari- 



