128 REPORTS ON THE STATE OF SCIENCE.—1918. 
these pores were small enough, we have the possibility of the pro- 
duction of a membrane with the structure that Traube held that of 
the copper ferrocyanide membrane to be, namely, a sieve whose holes 
were of dimensions such that water molecules could pass through, 
while larger molecules, such as those of calcium chloride or potas- 
sium sulphate, could not. Here we meet with the disputed question 
of the nature of the semi-permeable membrane in general. It seems 
to the writer that the evidence in favour of Traube’s original view 
is very strong, although further evidence on the point is much 
required. In the case of membranes such as those of parchment 
paper or collodion, which are permeable to water and the small 
molecules of crystalloids, impermeable to the large molecules or 
aggregates of colloids, the degree of permeability appears to be strictly 
in proportion to the dimensions of the particles of the solutes. In 
the case of the copper ferrocvanide membrane, some observers hold 
that the passage of water depends on the different degrees of bydra- 
tion of the colloid material of the membrane on its two faces. Even 
Morse (1914, p. 87) appears to hold this view, but the degree of 
hydration of colloids follows quite a different law from that of 
osmotic pressure. While the osmotic pressure of a solution is 
directly proportional to the molar concentration, not the chemical 
nature, of the solute, the hydration by salt solutions follows the 
series known as that of Hofmeister, where sodium and lithium salts, 
for example, behave differently from one another. Some inves- 
tigators, moreover, state that non-electrolytes have no influence on 
the amount of water taken up by colloids. Objection also may be 
made to the theory of the production of osmosis by surface tension 
effects. Surface tension is not a function of molar concentration 
only, whereas osmotic pressure is. But, however this may be, 
Tinker’s (1916) photographs show a definitely porous structure in 
the copper ferrocyanide membrane. The measurements given by 
him of the dimensions of the pores must be received with caution, 
since the published photographs show obvious diffraction. Indeed, 
the method of illumination used, a narrow cone of light, is not in 
agreement with the recognised methods of “ critical” illumination, 
with a wide angled cone, adopted by the microscopists. The con- 
clusions drawn as to adsorption of water molecules o1 the walls of the 
pores are also open to question. Further discussion of the problem 
is beyond the scope of this report, but we have seen how a sieve-like 
membrane might be formed on the surface of protoplasm, whereas a 
membrane which owed its semi-permeable character to its behaviour 
as a solvent for some solutes and not for others is much moré difficult 
to imagine. 
At this place a few words are desirable with respect to the Over- 
ton (1899) theory of the membrane as composed of lipoid material. 
It is a striking fact that the cell membrane is always permeable to 
those substances which are soluble in fats, such as the alcohols are. 
But, when we extend this statement and make it to apply generally, 
we meet with difficulties. For instance, certain dyes, such as 
methylene blue, are found to penetrate the cell membrane, while 
others, aniline blue, do not. Methylene blue is insoluble in chloro- 
form, but if a solution of kephalin (a lipoid allied to lecithin) in 
