458 BOTANICAL GAZETTE [JUNE 
It was found that different Chamaedaphne plants from the 
Chamaedaphne association (fig. 2) transpired at virtually the same 
rate during the same experiment. Consequently, that plant was 
taken as the basis for all comparisons and was included in practically 
every set of experiments. 
With the close of the experiment the leaves were detached, 
placed side by side on white paper, covered with a thin piece of 
glass, and their outlines traced with a polar planimeter, by which 
means the leaf area was obtained. After a little practice with this 
instrument it was found that successive determinations of the 
same set of leaves did not vary by as much aso.3 percent. Accord- 
ingly the average of two determinations was used throughout the 
work. When stomates were also present on the upper surfaces 
(the rare exception in the plants used), the area thus obtained was 
oubled. 
With the data so recorded, the results of each experiment were 
calculated with the aid of a slide rule to a standard basis, the rate 
of transpiration in grams per hour per 100 sq. cm. of leaf surface. 
These results were plotted on cross-section paper and the com- 
parison made. As more than one determination for each plant 
was made, the resulting curves should approach a general simi- 
larity. Under the same conditions the similarity of the graphs was 
striking. Through dissimilarity of the resulting graphs, it was 
possible both to demonstrate the effect of change of experimental 
conditions and to weed out aberrant plants. As the greater part 
of the work was concerned with relative values, the continued 
reappearance of the same result in the graphs was taken to uphold 
the contention and no contention not thus uniformly upheld is 
presented in this paper. 
During a part of the year 1912 the volume of the leaves was also 
determined by ascertaining the amount of alcohol they displaced. 
Alcohol was used in place of water on account of the large amount 
of air coating the leaves submerged in the latter. The results were 
calculated on the basis of water loss per hour per 1 cc. of volume. 
This was done in order. to incorporate the results obtained from 
plants, the difficulty of determining the leaf surface of which would 
otherwise have rendered it virtually impossible. The correlation 
