Permeability 193 
change in permeabilit}' of a membrane at different temperatures is 
nothing more nor less than a measure of the change of viscosity of 
the liquid. 
There remain to be considered the various theories put forward 
to explain the differential permeability of membranes. There are 
two well-defined theories which are usually described as the sieve 
theory and the solution theory. 
The sieve theory was first put forward by Traube (1867) who 
attempted to explain the semi-permeability of his precipitation 
membranes by supposing them to act as “ atom-sieves ” (more 
properly “molecule-sieves”), so that molecules below a certain size 
pass through, while those above this limiting size are kept back. 
The experiments of Tammann and Walden already cited, as well 
as those of Kahlenberg (1906), do not wholly favour the sieve theory. 
Nevertheless, the theory has found in recent years a considerable 
number of supporters, e.g. Sutherland (1897), Perrin (1900). In an 
investigation on the passage of dyes through parchment paper Biltz 
(1910) found that the rate of passage of dye through the membrane 
depended very definite^ on the dimensions of the molecule of dye. 
Thus when the number of atoms in the molecule did not exceed 45 
the dye passed readily through the membrane; as the number of 
atoms in the molecule increases from 45 to 70 the rate of penetration 
through the membrane is slower and slower, and when the number 
of atoms in the dye molecule is greater than 70, the membrane is 
impermeable to the dye. 
Also Bayliss (1915) points out that the difficulties in the way of 
the theory presented by Tammann’s experiments in which the order 
of penetration of different substances varies with the membrane, can 
to a large extent be explained away. 
Contrasted with the sieve theory is the solution theory of 
permeability, which may be traced back to the writings of Liebig 
(1849) an( l L’Hermite (1855). According to this theory substances 
pass through the membrane if they are soluble in the membrane, 
and are held back if they are insoluble in it. L’Hermite illus¬ 
trated the theory with the “three liquid layers” experiments, 
one of which may be quoted as an example. Layers of water, 
castor-oil and alcohol were superposed in a cylinder, the castor-oil 
thus forming the layer separating water and alcohol. The alcohol 
is soluble in the castor-oil and passed through the middle layer 
composed of it (equivalent to a membrane) into the water. Similar 
experiments were performed with other systems of three liquids. The 
application of these experiments to the permeability of membranes 
5—5 
