12 Walter Stiles 
that the osmosis of salt solutions is markedly affected by the presence 
of acids and alkalies. 
The explanation of negative osmosis put forward by Girard (1908, 
1909, 1910 a, 1910 b, 1914) and also by Bartell and his collaborators 
(1914, 1916 a, 1920 a , b) ascribes the phenomenon to the action of 
electrical forces. It has already been noted (Chapter III) that at 
the surface of separation of two liquids there exists a potential 
difference. But Bartell and Madison quote the results of Briinings 
(1903, 1907), Lillie (1911), Beutner (1913 e), Girard and Bartell and 
Hocker to show that permeable membranes of almost any material 
separating water and a solution exhibit a difference of potential on 
the two sides of the membrane which is different from the potential 
difference arising from contact of the two liquids. Girard supposes 
that in addition to the difference of potential between the two sur¬ 
faces of the membrane there is also an electric charge on the walls 
of the capillaries of the membrane, while the liquid in the capillaries 
possesses a charge opposite to that on the walls of the capillaries, 
there being produced a Helmholz double layer (Helmholz, 1879). 
This charge on the walls of the capillaries is supposed to be deter¬ 
mined by the preponderance of a small excess of hydrogen or 
hydroxyl ions as suggested by Perrin (1904). Such a state of affairs 
would tend to the movement of negatively charged water, for 
example, in the direction of the surface of the higher potential, and 
vice versa. 
The explanation of negative osmosis offered by Bartell and his 
collaborators is essentially similar. The tendency to normal osmosis, 
determined by the differences in salt concentration on the two sides 
of the membrane will always be present, but superimposed on these 
is the passage of water brought about by the presence of electrical 
factors. These are as Girard suggested (1) the electric charge on the 
membrane with respect to the liquid layer in the capillaries of the 
membrane, that is, the value of the Helmholz double layer between 
membrane and liquid, and (2) the polarisation of the membrane, 
that is, the difference of potential between the two faces of the 
membrane. The value of the former we have noted is ascribed by 
Girard to the influence of an excess of hydrogen or hydroxyl ions. 
Bartell and his collaborators throw doubt on the adequacy of this 
explanation and suppose three factors to be active here, namely 
(a) the extent of the diffusion of the electrolyte through the mem¬ 
brane (phase potential), (b) the relative migration velocities of the 
ions (diffusion potential), and ( c) the extent of selective adsorption 
