438 BOTANICAL GAZETTE [JUNE 
and D with K,Fe(CN),; fill B and & with distilled water, taking 
care that the liquids stand at the same height in D and £; fill 
A with CuSO,, and close it. The membrane containing K,Fe(CN), 
is promptly filled with a precipitate of Cu,Fe(CN), which is practically 
impervious to CuSO,, and any leak in this precipitation membrane 
will be immediately mended. The other membrane, of parchment 
paper alone, is relatively easily passed. The pressure against it 
is accordingly less. I have worked the experiment using 6 per cent. 
CuSO, and 3 per cent. K,Fe(CN),. After 2 hours 15 minutes, the 
K,Fe(CN), column had fallen 1™, and the column of water had 
risen the same distance. Wishing to stop while the water was as 
pure as possible. I poured it out then, and was unable to detect 
any copper in it with K,Fe(CN),; it may have been there in extreme 
dilution, or it may have passed into but not yet through the membrane. 
What had been done then was taking water out of a solution of appre- 
ciable osmotic strength, and forcing it over into practically pure water. 
The energy to perform this work is furnished by the loss of CuSO, 
from the T tube, as is easily seen if the experiment is continued longer. 
The concentration of the solution that is forced through the parch- 
ment membrane must obviously be at first slightly in excess of that of 
the K,Fe(CN), that the water is drawn from. 
The experiment is interesting in botany, because root pressure 
must be caused by exactly this same process. Under certain very 
unlikely conditions, differences in temperature, without any loss of 
dissolved matter, might cause guttation, but they could not cause root 
pressure, because all parts of the root must be at too nearly the same 
and if the solute escapes it is unnecessary to imagine rhythmic or any 
other variations. As Pfeffer suggests, pure water might be forced from 
a cell by local internal differences in concentration, but it would 
require energy whose source is unsuggested, to set up and maintain 
these differences. 
In order that root pressure may be caused in the way that the 
experiment with this artificial cell illustrates, the protoplasm must be 
permeable to the osmotically active matter of the cell sap in different 
measure in different parts of itselfi—which is not much to expect of 
protoplasm, which shows a finer development of the power of internal 
