PRESIDENTIAL ADDRESS. 289 
thick. It precludes any temptation to view molecules as shooting across from one 
liquid to the other through some kind of peepholes in the membrane. 
The advantage of a thin membrane in practice is simply that the necessary 
moisture is rapidly applied to the active surface, thus enabling the pressure on 
the side of the solution to rise quickly, but it has no effect on the ultimate 
equilibrium. ; 
As far as that goes, the semi-permeable membrane or saturated medium 
might be infinitely thick, or, in other words, there need be no receptacle or place 
for holding the pure solvent outside the membrane at all. In fact, the function 
of the receptacle containing the pure solvent is only to keep the medium moist, 
and is no more or no less important than the vessel of water supplied to the 
gauze of the wet-bulb thermometer. It is merely to keep up the supply of water 
to the medium. 
The real field where the phenomenon of osmosis takes place is the surface of 
separation between the saturated semi-permeable medium and the solution. 
Imagine a large mass of colloidal substance saturated with water and having a 
cavity containing a solution. The pressure will now tend to rise in the cavity 
until it reaches the osmotic pressure—that is, until there is established an equi- 
librium of surface transfer of molecules from the solution into the medium and 
back from the medium into the solution. 
No doubt, the phenomenon as thus described occurs often in Nature. It is 
just possible that the high-pressure liquid cavities which mineralogists find in 
certain rock crystals have been formed in some such manner in the midst of a 
mass of semi-permeable medium; the pure solvent in this case being carbon 
dioxide and the medium colloidal silica, which has since changed into quartz 
crystal. 
In considering equilibrium between a saturated semi-permeable medium 
and a solution there seems to me to be a point which should be carefully con- 
sidered before being neglected in any complete theory. That is, the adsorption 
layer over the surface of the semi-permeable medium. We have seen that solu- 
tions are profoundly modified in the surface layers adjoining certain solids, 
through concentration or otherwise of the salts in the surface layer, so that the 
actual equilibrium of surface transfer of water molecules is not between the 
unmodified solution and the semi-permeable medium, but between the altered 
solution in the absorption layer and the saturated medium. Actual determina- 
tions of the adsorption by colloids are much wanted, so as to be able to be quite 
sure of what this correction amounts to or even if it exists. It may turn out to 
be zero. If there is adsorption, however, it may possibly help to account for part 
of the unexpectedly high values of the osmotic pressure observed at high con- 
centrations of the solution, the equilibrium being, as we have seen, between the 
turated medium and a solution of greater concentration than the bulk of the 
iquid, namely, that of the adsorption layer. In addition, when above the 
critical adsorption point, there may be a deposit in the solid state. This may 
produce a kind of polarised equilibrium of surface transfer in which the molecules 
which discharge from the saturated medium remain unaltered in amount, but 
those which move back from the adsorption layer are reduced owing to this 
deposit, thus necessitating an increase in pressure for equilibrium. If either 
or both of these effects really exist, it would seem to require that the pressure 
should be higher for equilibrium of the molecular surface transfer than if there 
were no adsorption layer and the unaltered solution were to touch the medium, 
but at the same time it should be remembered that there is a second surface 
where equilibrium must also exist—that is, the surface of separation of the 
adsorption layer and the solution itself. It is just possible that the two together 
cancel each other’s action. 
Quantitative determinations of absorption by solid media from solution are 
hard to carry out, but with a liquid medium are not so difficult. Ether con- 
stitutes an excellent semi-permeable medium for use with sugar solution, because 
it takes up or dissolves only a small quantity of water and no sugar. <A series 
of experiments using these for medium and solution has shown (1) that the absorp- 
tion of water from a solution diminishes with the strength of the solution; and 
(2) that the absorption of water for any given strength of solution increases with 
the pressure. This increase with pressure is somewhat more rapid than if it 
1914. o 
