106 BOTANICAL GAZETTE [AUGUST 
vulgaris the maximum diffusion occurred about 5 minutes after the 
leaf was severed from the stem, while in the case of the thick leaf of 
Prunus Laurocerasus nearly 20 minutes elapsed before the maximum 
stomatal opening was reached. KAMERLING (23) found that when 
Rhipsalis cassytha had lost 1 to 4 per cent of its normal water 
supply, the amount of transpiration per unit time increased, and 
later when the loss in weight had reached a certain point, varying 
from 6 to ro per cent, the transpiration again diminished. KAMER- 
LING is of the opinion that the increase in transpiration is due to 
the opening of the stomata. Ltioyp (31), on the other hand, failed 
to find this temporary opening. 
The evidences at hand support the conclusion that the stomata 
open for a short time during wilting. The time is short, however, 
and there is no evidence that the stomata ordinarily open up during 
the early stages of wilting and continue to be open until the plant 
has attained its permanent wilting point. This point is important 
in connection with argument presented for the break in the water 
columns. 
In order to prove that the stomata open only during the early 
stages of wilting, the porometer as modified and used by KNIGHT 
(24) was resorted to. The plant was attached to the aspirator and 
allowed to remain until the leaves were partially wilted. Tests 
were made upon plants grown in the greenhouse and later trans- 
ferred to the laboratory and plants which had been grown con- 
tinuously in the greenhouse. The plants in the laboratory were 
kept for 5 days before experimentation was begun. The tests 
of this series were begun on November 28, 1917, and continued 
until December 3, 1917; readings were made at three different 
times of the day. This conforms with the plan adopted in making 
the readings of the foliar transpiring power. Evaporation was 
recorded by means of a standardized cylindrical form of atmom- 
eter of the Livingston type. The data are given in table VI. 
In an examination of the data presented in table VII, readings 
were not taken until the 16th hour on November 29, 3 days after 
all watering had ceased. The time elapsing between two suc- 
cessive bubbles, as ascertained by means of a stop watch, was found 
to be 160 seconds. At the 11th hour it took 140 seconds, while 
