2 Litovp: LEAF WATER IN GOSSYPIUM 
the dry weight of the leaf increased during the day. The conclu- 
sions drawn rested more especially on the assumption that the 
leaf water at the close of day is not in material excess of that at 
the beginning, and upon the fact that during the afternoon there 
is a sharp rise in the amount of leaf water in spite of a constant, 
if not a still increasing, dry weight. The hope that during the 
summer of 1911 the more convincing data based upon constant 
leaf area could be obtained was not realized, and the evidence 
therefore still remains incomplete. 
In September 1910 a similar attack was made upon the cotton 
plant at Auburn, Alabama, in which analogous conditions with 
respect to leaf water were found. During the present year the 
question has been further studied, the material being obtained 
from a pure strain of Dixie cotton, the seeds of which were supplied 
me by the Bureau of Plant Industry. 
Plants were grown at Auburn and, under irrigation, at Tucson, 
Arizona, in the experimental grounds of the Desert Botanical 
Laboratory of the Carnegie Institution of Washington. My 
thanks are due to Dr. D. T. MacDougal for his courteous co- 
operation. The present paper reports upon a portion of the data 
obtained bearing upon the variation in the amount of leaf water 
and upon the behavior of the stomata. 
METHODS 
DRY WEIGHT AND LEAF WATER. The first two series of leaf 
samples, the figures for which are given in TABLE I, consisted of 
leaf material placed in tared, cork-stoppered bottles, weighed, 
oven-dried thoroughly and reweighed. The samples were taken 
at two hour intervals. The actual work was done by Mr. C. S. 
Ridgway, to whom my thanks are extended. The samples on 
which TABLE III is based were taken at assumed critical times, and 
treated in the same manner by myself. 
The data in the remaining tables were obtained by cutting 
circles of the lamina with a sharp cork borer, placing them forth- 
with in tared vials and weighing before and after thorough drying. 
The area of these circles, of which between fifty and sixty were 
taken for each sample, was calculated, and the exact number of 
circles taken determined a second time after drying. All data were 
