28 



Mr W. N. Shaw, On the relation between [Nov. 25, 



We see therefore that in spite of the present tendency to 

 reduce the electrolytic action to convection by atomic ions if 

 possible, there are many cases in which the ions are aggregations 

 of atoms, if not strictly molecular, and in some cases molecules 

 are associated with a moving atom in electrolysis. It seems 

 therefore not altogether unreasonable to assume that the decom- 

 position of a complex molecule may not be exceptional but the 

 general rule. Let us therefore take the venturesome step of 

 considering that the whole result of electrolysis consists in the 

 separation of complex molecular aggregates of salt and water each 

 into two parts, each part containing one dissociated atom or its 



Table II. 



chemical representative, and suppose that the complete result of 

 electrolysis, including the transference of liquid known as endos- 

 mose, the migrations of ions, and the deposition on the electrodes 

 may be accounted for by this splitting up of the complex molecule 

 into two parts and the transference of the separated parts in 

 opposite directions, so that each separated atom would be loaded 

 with a number of water molecules or salt molecules or both. If 

 this be assumed, the effect of a porous diaphragm in a solution 

 would be not to cause the transference of liquid but to prevent 

 it slipping back again, and the friction against the plug which 

 would prevent the slip back would be very similar in its mathe- 

 matical expression to the force supposed by von Helmholtz to 

 cause the transference. 



The reasoning by which Wiedemann, on p. 592 of vol. II. of his 

 Electricitat, shews that electric endosmose is independent of 

 migration seems to me to be as follows. The total gain of cation 



