270 DIFFUSION, OSMOSIS, AND FILTRATION 
Physiological methods of estimating osmotic pressure have also been 
devised. The method of de Vries l is based upon the plasmolysis of the 
protoplasts of vegetable cells. The cells filled with coloured sap from 
the middle nervure of the leaf of Tradescantia discolor are useful 
for the purpose, sections of this part being allowed to soak for three to 
five hours in the solutions whose osmotic pressures are to be determined. 
If the cells are plasmolysed, i.e. if the protoplasts are found on 
examination to have shrunk from the cell walls, the osmotic pressure 
of the solution producing this effect is above that of the cell sap, for 
water has passed from the latter to the former, as evidenced by the 
diminution in volume. By investigating a series of solutions with 
sections from the same leaf, it is of course possible to find two of 
slightly differing concentration of the substance under investigation, 
one of which just causes plasmolysis, while the other (weaker) does 
not. A solution of concentration equal to the mean of these two is 
said to be isotonic with the cell sap. 
De Vries, on preparing a number of solutions of different substances, 
all isotonic with the same batch of cells, and expressing their 
concentrations in gramme-molecules to the litre, found that it required 
a lower gramme-molecular concentration of some substances than of 
others to obtain isotony. The term " water extracting power " 
( Wasscranzichmigsvcrmdgcn) was used to express this peculiarity which 
is obviously related to what has above been termed dissociation. 
Taking OT grm. molecule to the litre of saltpetre as a standard, and 
giving it a magnitude of 3, the relative value (as regards plasmolysis 
of vegetable cells) of a molecule, of a number of substances was 
expressed in terms of that of a molecule of saltpetre, and the numbers 
expressing this ratio called isotonic coefficients of the substances. 
Thus, to barium chloride (BaCL + 2Aq = 244) is given the 
isotonic coefficient 4, which means that f'244 parts by weight of 
barium chloride in aqueous solution exert the same plasmolysing 
action as 101(KNO :1 = 101) parts by weight of saltpetre i.e. a 183 
per cent, solution of crystallised barium chloride is isotonic by the 
method, with a 1*01 per cent, solution of potassium nitrate. 
Since cane sugar on this system is given the value of 2 for its 
isotonic coefficient, and since, being a non-electrolyte, it is not 
dissociated in solution, it is merely necessary to divide the isotonic 
coefficients of de Vries by 2, in order to obtain ordinary dissociation 
coefficients. 
It is obvious that the substances in solution must exert no 
deleterious action on the protoplast of the cell, and must, moreover, 
be quite unable to diffuse through it, if the method is to be exact. 
Here, again, we are met with the difficulty, that the protoplast is 
not a strictly semipermeable membrane. It must let certain substances 
pass, otherwise the cell sap could not have any other constituent than 
water ; and it is only because the permeability to certain substances is 
so far below that to water, that it is possible to obtain fairly approximate 
measures of osmotic pressure by this method. With other substances 
the permeability is so great that the values are far too low. 
Thus with sodium chloride, by this method, the dissociation coefficient 
is reckoned as 1*5 (de Vries' isotonic coefficient 3), but by lowering of 
1 Jahrb f. icis*. Botanik, 1884, Bd. xiv. S. 427; Ztschr. f. physi&al. CJiem., Leipzig, 
1888, Bd. ii. S. 415. 
