356 Knight. — ‘ Relative Transpiration as a Measure of 
Experiment IT. 
The method was similar to that used in Exp. I. The plant, 
Eupatorium adenophorum , and the atmometers 1 were placed in the air-flue 
(2) in a dark room, and a stream of air was drawn through at a speed of 
7 metres per minute. Weighings of the plant and the atmometers were 
made at 30-minute intervals, for which purpose each was removed from the 
flue for two minutes. When the rates of transpiration and evaporation had 
become practically constant, the temperature of the room was raised by 
means of an electric radiator, and the rates of transpiration and evaporation 
were again determined at intervals (90-minute periods in the experiment 
given).- The temperature continued to rise slowly during the rest of the 
experiment, and simultaneously the relative humidity tended to decrease. 
An attempt was made to keep the latter constant by watering the floor of 
the dark room, but with the changing temperature the humidity was found 
somewhat difficult to regulate with any degree of accuracy. 
Readings were made of the absorption of water by the plant from the 
potometer as in Exp. I, in order to determine whether there was any 
tendency towards wilting. 
The results appear in Table II. 
Table II. 
Temp. 
0 C. 
Relative 
Riant 
Loss from 
Time. 
Humidity 
per cent. 
Absorptio, 
JL I'LVfl'lr 
n. Transpiration. 
Atmometers. 
I II 
mg. 
mg. 
mg. 
mg. 
11.30 a.m. 
12 noon 
12.30 p.m. 
15*9 
15-9 
16*0 
83*5 
83*5 
83 
420 
39 ° 
420 
4 i 5 
407 
411 
1 77 
176 
180 
153 
149 
158 
1.0 p.m. 
16*0 
83 
16*7 
82 
17*2 
82 
1400 
1402 
604 
525 
2.30 p.m. 
17-6 
81 
17-8 
80 
i8-o 
80*5 
I 54 ° 
j 536 
671 
578 
4.0 p.m. 
l8*2 
80*5 
Dividing the experiment into three 90-minute periods, representing 
three different average temperatures, we find that the change of rate of 
1 The atmometers used in this experiment and in Exp. Ill, quoted later, were paper ones of 
a modified Piche type (see Livingston, 7). Atmometers of porous earthenware have also been 
used in the present experiment and give similar results, but there is a distinct lag in the response to 
temperature changes, presumably owing to the slowness with which the relatively large mass of 
water and earthenware assumes the changed temperature of the surrounding air. 
