310 - BOTANICAL GAZETTE [OCTOBER 
the cells, when first set up, with half-strength sea water in the upper 
cell and sea water in the lower. Thirty minutes was ample for the 
establishment of a steady diffusion gradient through the tissue 
between the two solutions. The upper solution was then replaced 
by fresh half-strength sea water, after which regular readings were 
taken. In order to eliminate the error due to exosmosis from the 
protoplasm, such as might be occasioned by toxic salts, 3 controls 
out of each set of 11 to 13 simultaneous experiments had the more 
dilute solution in both cells. At the end of the experiment the 
average conductance of the solution in the upper cells of the controls 
was taken as a standard of measurement, the average conductance 
of all the other upper solutions being divided by this figure in order 
to obtain the percentage which expresses their gain as compared 
with the control. The figures which were obtained in this manner 
measure the amount of salt which has passed through the tissue, 
while the errors due to exosmosis from the protoplasm as well as 
those due to diffusion from the intercellular substance are elimi- 
nated. 
5. We must eliminate the error due to variations in the thick- 
ness and maturity of the disks of tissue from different fronds, and 
also that due to variations in the area of tissue through which salt 
can pass (such as might be introduced by unavoidable smearing 
of the cement over the surface of the disks).4 In order to eliminate 
all of these errors, controls were established in the following man- 
ner. After a preliminary half-hour with half sea water in the 
upper and sea water in the lower cell, the upper solution was 
replaced with 5.4 cc. of fresh half sea water, and the rate of change 
of conductivity determined at the end of 2 hours. Both upper and 
lower solutions were now replaced with solutions of the salt to be 
investigated (the fresh solutions having the same conductivity as 
those which they replaced), and the rate of change of conductivity 
determined after a further period of 2 hours. By dividing the 
4 A single experiment was conducte d to determine the influence of frond thickness. 
The results were entirely negative. This is in accord with the results secured by ABEL 
(Aset, J. J., Rowntree, L. G., and Lurver, B. B., On the removal of diffusible 
substances from the circulating blood of living animals by dialysis. Jour. Pharm. 
and Exp. Ther. 5:275. 1914.), who found that onaeor at electrolytes through 4 
collodion membrane was independent of the thickness of 
