Permeability 185 
Cavara, 1901, 1905; Nicolosi-Roncati, 1909; Dixon and Atkins, 1910, 
1912 a, 19126, 1912 c; Atkins, 1910; Marie and Gatin, 1912). A 
number of observations made by Dixon and Atkins among their 
earlier determinations led these authors to doubt the correctness of 
the assumption that the liquid so obtained represents an average 
sample of the sap. It had also been shown by Andre (1906 a, 1906 b, 
1907 a, 1907 b) and by Marie and Gatin (1912) and confirmed by 
Dixon and Atkins (1913 a) that if a number of samples of sap are 
successively pressed from the same tissue, the samples exhibit a pro¬ 
gressive increase in concentration and osmotic pressure. The explana¬ 
tion is that when pressure is first applied the protoplasm is not very 
permeable to the solutes in the vacuole, so that the first sample of 
liquid contains a much higher percentage of water than the vacuole. 
But the pressure results in the rupture of the cells and the passage 
out of the dissolved substances as well as the water. Consequently, 
as the pressure is increased to produce the successive samples of sap, 
more and more cells are burst so that the expressed sap becomes 
more and more concentrated. 
In order to obtain an average sample of cell sap it is therefore 
necessary to treat the tissue so that the protoplasm is rendered com¬ 
pletely permeable. Heat and toxic agents (toluene vapour and chloro¬ 
form vapour) were tried but it was decided that the time required to 
render the cell membranes completely permeable was so long that 
enzyme or other actions might have sufficient play in the cells to 
alter materially the composition of the sap and so vitiate the results. 
Dixon and Atkins came to the conclusion that the most satisfactory 
way of rendering the cells completely permeable is to immerse the 
tissue in liquid air, as at this low temperature enzyme actions and 
autolytic and other changes are completely negligible. After such 
treatment the sap from successive pressings of the same piece of 
tissue had practically the same osmotic pressure. Subsequent experi¬ 
ments of Gortner, Lawrence and Harris (1916) have confirmed these 
conclusions. 
For the determination of the osmotic pressure of the expressed 
liquid the method usually employed is that depending on the lower¬ 
ing of the freezing point of the solution. The thermo-electric method 
of determining the freezing point lowering evolved by Dixon (1911) 
may be mentioned in this connection. Harris and Gortner (1914) 
have published a useful table giving the values of the osmotic pres¬ 
sure for depressions of the freezing point between o-ooi° and 2-999° C. 
This method, of course, gives the mean osmotic pressure of all 
Phyt. XXI. IV. 
13 
