Permeability 153 
this case the facts of distribution differ very considerably from those 
in the case of blood serum and corpuscles, for example. Thus, if, for 
instance, the sodium salt of a protein is separated from pure water 
by a membrane of parchment paper, the diffusible sodium ions can 
only diffuse to a position in which their osmotic pressure is balanced 
by the electrostatic attraction of the non-diffusible oppositely 
charged protein ion. Now if the pure water of the external medium 
is replaced by a solution of a freely diffusible salt, as for example 
potassium chloride, the potassium ions and chlorine ions can both 
diffuse through the membrane, and some of the potassium ions can 
replace the sodium ions in relation to the negatively charged protein 
ions, and this process will proceed until the sodium and potassium ions 
arc distributed throughout both solutions in the same relative propor¬ 
tions. But the absolute concentrations of the two kations will be 
different on the two sides of the membrane. Now in the case of rabbit 
blood (Abderhalden, 1898) the potassium bears a different ratio to 
sodium in the plasma from that which it bears in the corpuscles, so 
that the presence of a colloidal salt within a membrane permeable 
to crystalloids does not account for the distribution of ions in this 
particular case. 
In reviewing the evidence given by a consideration of the relation 
of solutes to the cell in favour of, or against, the presence of a semi- 
permeable plasma-membrane, it is again necessary to distinguish the 
cases of the vacuolated and non-vacuolated cell. Failure to do this 
in the past has undoubtedly led to the drawing of conclusions im¬ 
properly. It has already been indicated that the protoplasm of the 
vacuolated cell fulfils the conditions necessary for a membrane 
between the vacuole and the external medium, and one which may 
exhibit differential permeability to different substances. The reten¬ 
tion of sugar and the red pigment by the cells in slices of root of red 
beet can be explained satisfactorily on these grounds. Similarly the 
deplasmolysis of vacuolated cells such as those of Spirogyra and of 
the staminal hairs and epidermis of Tradescantia when plasmolysed 
in glycerol or solutions of certain salts is to be explained by the 
gradual passage of the solute through the protoplasm into the vacuole. 
In order to obtain evidence in relation to the presence of a plasma- 
membrane we shall do well to devote special attention to data derived 
from a study of the relation of solutes to non-vacuolated cells. 
Although Pfeffer (1900) admitted that no absolutely convincing 
proof had been brought forward, he gave reasons for supposing that 
the limiting membranes have different diosmotic properties from the 
