30 MISSOURI BOTANICAL GARDEN. 
found that in certain cases the diffusing substance is less con- 
centrated inside than outside of the cell, although it readily 
passes into the protoplasm. Paine (14) found that on im- 
mersing yeast cells in alcoholic solutions varying in con- 
centration from five to twenty per cent, the ratio of the 
concentration of the alcohol within the cell to that of the 
solution outside was practically a constant, and independent 
of the absolute concentration. The alcohol outside of the 
cell was always more concentrated than the alcohol inside of 
the cell. Paine explained this observation by supposing that 
some of the water within the cell is combined in such a way 
that it is unable to act as a solvent for alcohol. 
The osmotic theory fails us most completely, however, when 
we come to cases of permanent plasmolysis of cells by sub- 
stances that readily enter the protoplasm. Lavison (7) found 
that sodium chloride passes into the protoplasm even from 
plasmolysing concentrations and causes permanent plas- 
molysis, although large quantities of the salt are taken up 
by the protoplasm. He further found that when he plas- 
molysed cells with an iron salt, the salt, nevertheless, entered 
the protoplasm, and that if the experiment was not continued 
too long the cells were not killed by such treatment. How is 
it possible for a salt that readily passes through the mem- 
brane to cause plasmolysis? So far as such salts are con- 
cerned the membrane does not exist, and yet we observe a 
phenomenon (plasmolysis) that is ordinarily explained as 
due to the action of a semi-permeable membrane. 
It must be obvious from the foregoing enumeration of facts 
recorded and conclusions reached by various workers that 
diffusion and osmosis do not explain absorption and secretion. 
There is great confusion regarding the nature of the mem- 
brane that has been assumed to surround cells. Lepeschkin 
(8) has considered it a protein membrane, Overton (13) a 
lipoid membrane, Nathansohn (11) a membrane in which 
lipoid particles are mixed with protein particles, while Czapek 
(2) looks on it as an emulsion of fat, each globule of which 
is surrounded by a thin layer of soap. All of these things 
cause us to doubt the value of the assumption that a semi- 
