408 BOTANICAL GAZETTE - [DECEMBER 
one made by Plateau (3), who used as test objects animals, 
mainly invertebrates. His experiments turned out unfortunately 
on account of the fundamental error made in using as compara- 
tive solutions percentage trations. Recently, Davenport 
(2) has outlined a method of analysis, in itself correct, but likely 
to lead to error in many cases through the choice of sodium 
chlorid as a standard substance assumed to exert a_ purely 
osmotic effect. . 
METHOD. 
_In attempting to analyze the injurious action of a compound 
into its osmotic and its toxic constituents, a standard solution 
which has a known osmotic action, and is not open to the sus- 
picion of being in any degree toxic, may be selected for put- 
poses of comparison. 
As such a standard substance, I have selected cane sugar. 
It occurs normally in plant cells, often in high concentration, 
twenty per cent. having been observed in the sugar beet (4): 
As has been often proved, it is a most excellent plant food when 
supplied in cultures. In searching through the results of inves- 
tigators who have made use of this substance in cultures, I have 
failed to find in any case toxic action attributed to cane sugar. 
In higher concentrations it proves harmfui on exposure of plants 
to its action for long periods of time, an effect that seems to be 
amply accounted for by the osmotic properties of such solutions. 
Moreover, it diffuses rather slowly, and is less liable to injure the 
organism by the sudden withdrawal of water than is the cas€ 
with substances of more rapid osmotic action. A limited oppo! 
tunity for the cell to accommodate itself to the changing con 
centration is gained in the more slowly diffusing substances. — 
Having chosen cane sugar as a standard of pure osmotic 
effect, solutions were made up on the basis of a molecular weight 
of substance in the requisite numbers of liters required to give the 
concentrations desired. Tufts of Spirogyra filaments, first rinsed 
thoroughly in duplicate solutions to prevent dilution of those 
used in the experiments, were placed in the desired concentra- 
