368 Knight. — Further Observations on the Transpiration , 
from the frequency of the bubbles can be calculated the rate at which air is 
being drawn through the stomata. This gives a relative measure of the 
size of the stomatal aperture. The success of the method depends, of 
course, on the equality in size of the bubbles. It is found that within 
reasonable limits these are of the same size as long as the apparatus remains 
undisturbed, although the depth of the water in the aspirator certainly 
does slightly influence the size. The variation in the bubble size due to 
this cause is very small and is also gradual, and is not a serious drawback. 
The regularity of the bubbles may, however, be considerably interfered with 
by the dripping of water from the end of the outflow tube, and for this 
reason it is advisable that this tube should dip into an overflow vessel (e), 
in order to prevent the periodical jerks on the bubble as the drops fall. 
During the formation of a bubble the puli on the air passing through the 
leaf decreases slightly as the mercury surface is depressed, returning again 
to its original magnitude on the release of the bubble. This change of 
pressure is very small, and of course the cycle of change is identical for all 
bubbles. Necessary adjustments of the central electrode and of the height 
of the outflow tube may be made by means of rubber sleeves (f). 
In the experiments under consideration the recorder was placed in the 
flue and attached by means of T-pieces to two or three leaf-chambers, each 
on a different leaf. Thus an average reading of stomatal aperture on 
different parts of the plant was obtained. In order to avoid the necessity 
of disconnecting the recorder for every weighing to be made, a two feet 
long piece of very flexible small-bore rubber tubing was used to connect 
the apparatus to the leaf-chambers. This was allowed to hang in a loop, 
and tests showed that with a balance sensitive to about 5 mg. this length 
of tubing, although somewhat damping the oscillation of the beam, did not 
seriously affect the accuracy of the weighings. After setting up the 
potometers and the recording apparatus, observations were made at short 
intervals of temperature, relative humidity, loss of weight of the potometer, 
absorption of water by the plant, and rate of evaporation from the atmo- 
meter. The record of stomatal aperture was taken continuously throughout 
the experiment. After a period of about an hour, during which about ten 
sets of observations were made, it was possible by comparison of transpira- 
tion and water absorption figures to determine whether the water-content 
of the plant was decreasing under the conditions of the experiment. If no 
decrease in water-content could be demonstrated, it was assumed that the 
records of the preliminary period were representative of the normal progress 
of the functions of the plant. At this stage the water-content of the plant 
was experimentally diminished. This was accomplished by closing a stop- 
cock at the base of the potometer, thereby cutting off any further supply of 
water to the cut end of the stem. Observations were continued as before 
over a period as long as the results warranted, usually one or two hours. 
