672 
MOLECULAR FORCES IN THE PLANT. 
pellicles have only micellar interstices ; and indeed these latter are, according to 
him, smaller than the molecules of the precipitate of which the pellicle is composed. 
But in spite of the greater density, the endosmose is quicker than with all other 
membranes, because they are thinner. The pellicle becomes firmer (stiffer ?) when lead 
acetate or copper sulphate is added to the ß gelatine. As soon as the micellae of the 
stretched pellicle have become so far separated by the pressure of the cell-contents 
which have increased in quantity by the action of endosmose that their interstices allow 
the passage of the two substances from which the pellicle is formed, these substances 
must obviously again at once mutually react upon one another at those points, and must 
cause the production of new micellae of pellicle, which are deposited between those 
already in existence. Growth therefore takes place by intussusception, and is caused by 
the stretching of the pellicle, which stretching is on its part occasioned by endosmose. 
That the growth takes place not only by stretching but also by deposition Traube 
proved by replacing the tannic acid by water. As soon as this was done {i.e. as soon 
as the formation of new molecules of the precipitate in the pellicle was prevented, the 
endosmose still continuing) the growth ceased. 
As long as the concentration of the contents of the artificial cell is everywhere the 
same, the pellicle remains everyv^^here equally thick, and the cell retains its spherical 
form. But when the contents become diluted, a denser solution is formed in the lower 
part of the cell, a more watery solution in the upper part. The pellicle becomes 
in consequence thinner above, because the diiference of concentration is smaller there, 
and therefore more extensible ; hence the pellicle becomes more strongly stretched above 
and increases more rapidly in superficies, and protuberances directed outwards are not 
unfrequently formed. This may be expressed shortly by saying that endosmose takes 
place principally in the lower part of the cell, growth in the upper part. The diflference 
however in the concentration in the interior of the cell which causes this is the con- 
sequence of the water which penetrates by endosmose not mixing at once uniformly 
with all parts of the interior solution, so that layers of different specific gravity lie 
one over another. 
Further experiments showed that growing pellicle-precipitates having the form of 
cell- walls are produced also by mixing colloids with crystalloids^; e.g. tannic acid with 
copper and lead acetates, gelatinous silica with the same substances or with copper 
chloride, or finally crystalloids with one another, as potassium ferro-cyanide with copper 
acetate or chloride. Traube came to the conclusion that every precipitate the inter- 
stices of which are smaller than the molecules of its components must assume the form of 
a pellicle when the solutions of its components come into contact with one another. 
Since the pellicle-precipitates, as has already been mentioned, contain micellar interstices 
but no perforations, they are peculiarly well adapted for the study of endosmotic pro- 
cesses. They behave in this respect very differently from other membranes, being 
themselves often perfectly impermeable to the most diffusible substances, but allowing 
other chemical compounds to pass through them ; and every kind of pellicle has in this 
respect its own peculiarities. Independently of the fact that every pellicle-precipitate 
is impermeable to the fluids from which it is itself produced, the ß gelatine tannate is, 
moreover, impermeable for example also to potassium ferro-cyanide, but permeable to 
ammonium chloride, barium nitrate, or water. The pellicle of copper ferro-cyanide 
which is formed round a drop of copper chloride in potassium ferro-cyanide is imper- 
meable to barium chloride, calcium chloride, potassium sulphate, ammonium sulphate, 
diffusion-current equal over the whole surface of the membrane, separate threads of the solution of 
salt are seen to sink down into the water. These experiments show how little dependence is to 
be placed on the researches hitherto made on diffusion with membranes. 
^ [The term ' crystalloid' is here used in the sense in which it was first employed by Graham, 
to indicate those substances — as opposed to ' colloids ' — which may be susceptible of crystallisation, 
and which are endowed with the power of diffusion through a porous septum,] 
