New Theories of Solution. 363 



the a priori standpoint, Mr. Pickering goes on to say (loc. cit. 

 p. 26) : " It does not seem necessary, however, to imagine 

 the presence of free atoms to explain the phenomena of elec- 

 trolysis (cf. Lodge, Brit. Assoc. Report, 1887, p. 338)." At 

 the place cited, and more fully in his ( Modern Views of 

 Electricity,' p. 80, Prof. Lodge gives as his opinion that 

 " the addition of the idea of double-decomposition and inter- 

 change to the original hypothesis of Grotthus explains ail 

 that is required by the facts, viz. a virtual or potential disso- 

 ciation, a momentary state of hovering and indecision, without 

 the need for any continuous and actual dissociation." 



It appears to me that such a potential dissociation is not 

 quite sufficient to explain the known facts of the behaviour 

 of electrolytic solutions. The state of hovering or indecision 

 can only be brought about, according to the theory, by the 

 approach of two molecules of the electrolyte to such a distance 

 that they can interchange radicals. The number of molecules 

 whose radicals are in this undecided state is therefore depend- 

 ent on the number of molecular collisions per unit time, the 

 word collision being used to express a conjunction of mole- 

 cules close enough for atomic interchange to take place. On 

 the other hand, the conductivity depends directly on the num- 

 ber of potential interchanges, and therefore on the number of 

 collisions. Now in very dilute solutions the speed of the ions 

 cannot be supposed to be altered by any increase in the dilu- 

 tion, and therefore the conductivity will be regulated by the 

 number of collisions alone. In such salt-solutions as Kohl- 

 rausch investigated, where 1 gram-molecular-weight of salt 

 was dissolved in as much as 100,000 litres of water (i. e. where 

 there was only one molecule of salt to 5,500,000 molecules of 

 water *), it was found that the conductivity is proportional to 

 the concentration. But halving the concentration (i. e. the 

 number of salt-molecules in a given volume) would diminish 

 the number of collisions of the salt-molecules to a quarter of 

 the former value. We should, therefore, expect on the 

 potential-dissociation hypothesis that in dilute solutions the 

 conductivity should be proportional, not to the concentration, 

 but to the square of the concentration, which is not the case. 

 On this ground, then, we must reject, I think, the idea of 

 virtual dissociation brought about by atomic interchange. It 

 might be contended that collisions of the salt-molecules with 

 the water-molecules would be sufficient to cause the potential 

 dissociation ; but if that were so, sodium hydrate and hydro- 

 chloric acid would always exist in a solution of common salt, 



* It is here assumed that the molecule of liquid water is the same as 

 that of water-vapour — probably it is twice as large. 



