■94 EEPOET 1891. 



misingly. They accept the methods of inquiry and many of 

 the facts of the van't Hoff school — how could they do other- 

 wise ? — but decline to accept the theory that is advanced to 

 explain them. Most of their arguments appear to me to have 

 but little weight, being based upon misconceptions of the 

 theory they are criticizing ; while some of the facts put forward 

 prove, on examination, to tell against themselves. Moreover, 

 we find that the hydrate theory which they advocate leads to an 

 alarming intimacy with such compounds as IOOH2SO4 . H.^O 

 (i.e., a hydrate containing only 0-0018 per cent, of water); 

 and chemists are justified in demanding the very strongest 

 testimonials as to character before they consent to make such 

 acquaintances. On the other hand, it must be confessed that 

 one or two arguments against the gaseous theory appear to be, 

 in the meantime at any rate, real stumbling-blocks, which 

 want removing. 



One difficulty has struck me that has not, so far as I am 

 aware, been raised. If it be true that dissolved matter is 

 really in the state of vapour, and that such a salt as KI is 

 dissociated by solution into its ions, K and I, then these ions 

 should obey the gaseous laws, and, among them, Graham's 

 law of gaseous diffusion. Of course the absolute rates of dif- 

 fusion will be slow, on account of the friction-causing character 

 of the space occupied; but the ratio of these rates should be 

 in accord with the law. There must be, to start with, equal 

 concentrations, equal osmotic pressures, and, of course, equal 

 temperature ; but the masses of the two ions will be as 19 to 

 127. Should not, then, a partial separation by diffusion be 

 possible, which would prove the dissociation in the same way 

 that it has been proved in the case of actual vapours ? This is 

 not borne out by experiment. The rates of diffusion of salts 

 are roughly proportional to their molecular conductivities. 

 These are additively composed of the separate ionic values, 

 which represent the rates of migration of the ions during 

 electrolysis — itself a kind of diffusion. Now, these values 

 have been determined, and it has been shown that they are 

 practically equal in the case not merely of K and CI (which 

 we could understand), but of K and I. It would appear, then, 

 that among the gaseous laws obeyed by dissolved matter 

 Graham's law of diffusion cannot be included. This, probably, 

 admits of explanation ; but, as it has puzzled myself, I may be 

 pardoned for stating my difficulty. 



I wish now to ask your attention more particularly to the 

 actual process of dissolving, and then to lay before you a 

 hypothesis w^hich, as it seems to me, is a logical consequence 

 of the general theory. 



Imagine, then, a soluble solid in contact with water at a 

 fixed temperature. The substance exercises a certain pressure, 



