328 REPORT—1890. 
1,000 miles deep surrounding the world with a membrane in it, say, 100 
miles deep, through which the water could go, but the world could not 
because the holes were only, say, about a square mile in area, we see at 
once that, if this membrane were made of a material lighter than water, 
i.e. less attracted by the world than water, it would tend to burst out 
with a great force, i.e. it would float out from the world because the 
pressure in the water near the earth was much greater than at a distance 
from it. This shows where van ’t Hoff’s argument fails. He has 
neglected the difference of pressure in the solvent near and far from the 
salt, or at least has assumed that this difference of pressure could not act 
upon his semipermeable membrane because the membrane is permeable 
to the solvent. It is, however, quite evident that the water can press 
very hard even on a membrane permeable to it, as is explained by the 
example I have just mentioned. Considering the complex nature of the 
problem, I think it is quite too soon to assume that the state of affairs 
assumed by van ’t Hoff is at all like reality. I would much rather look 
for an explanation in the direction I have pointed out in this year’s report 
of the Committee on Electrolysis. The argument there tends to show 
that the distances between molecules would arrange themselves so that 
the forces due to different kinds of molecules would be independent of 
their kind and depend on their numbers, and this would lead to the laws 
of osmotic pressure. It seems to me much more likely that a state of 
affairs such as I have supposed existing near the earth is the one existing 
in a liquid. 
As regards the argument for the independent mobility of the ions 
founded on the laws of electrolysis, I think that just as in the case of 
osmotic pressure this does require a certain kind of independent mobility, 
but just as in that case I do not see that the required amount of indepen- 
dence cannot be attained without supposing a complete independence. 
There seems no doubt that conductivity and double decomposition are 
essentially connected with the same quality in the solution, and this 
property I have proposed to call ‘ionisation.’ Now, Williamson’s hypo- 
thesis as to the nature of double decomposition and Clausius’ as to the 
nature of electrolytic conduction only require that the ions shall be so 
far free as that they shall be frequently exchanging partners; neither 
hypothesis requires that they shall be during a finite time without 
partners, which I consider to be an essential condition of any right use 
of the term dissociated ions. If during the time the ions are paired they 
can move independently within the little chinks they have to move in 
between the molecules of the solyent—and be it observed that this is the 
same conditicn as for the extra osmotic pressure, 7.e. if their chinks are 
small compared with the variation of force between the ions——then there 
seems quite sufficient independence for any theory of electrolysis, if, 
whenever two molecules were within the same chink, there were, as there 
would be, sufficient independence for an exchange of partners. Thus 
these two phenomena would be explicable upon the same hypothesis, and 
that withont assuming that ionisation was a true dissociation. I have 
already explained that even those who insist most strongly upon the 
dissociation hypothesis yet guard themselves from its being supposed 
that this dissociation is an actually complete independence of the ions 
from one another. On all these grounds then I would appeal against 
the use of the word dissociation in this connection. Professor Ostwald 
says that there will result two theories leading to the same result. I 
