May 28,1917 
Transpiration Rate of Medicago sativa 
291 
for the day was then determined. For the shallow tank the mean 
hourly departure for the 24-hour period was 17 per cent; for the filter- 
paper evaporimeter 31 per cent; for the brown cylinder, white sphere, 
and Bellani plate about 40 per cent; for the white cylinder about 50 per 
cent; and for the deep tank about 90 per cent of the mean hourly 
transpiration. The corresponding departures for the daylight hours 
from 6 a. m. to 6 p. m. were as follows: For the shallow tank, 12 per 
cent; filter-paper evaporimeter, 22 per cent; brown cylinder, white 
sphere, and Bellani plate atmometers, about 30 per cent; white cylinder 
atmometer, 38 per cent; and the deep tank, 93 per cent. 
Since the hourly evaporation graphs of the various evaporation systems 
employed differ widely in form, it does not seem justifiable to attribute 
the discrepancy between the observed hourly transpiration and that 
calculated from the evaporation rate of any particular system to a 
change in the transpiration coefficient of the plant during the day, unless 
it can be shown that under less extreme conditions the transpiration rate 
is in accord with the evaporation rate. The plant may not be responding 
freely to its environment, but a departure in its relative transpiration 
rate from the evaporation rate of an arbitrarily chosen physical system 
does not necessarily establish this fact. A close correspondence does not 
appear to exist between the hourly transpiration rate of normal alfalfa 
plants and the hourly evaporation rate of any of the systems employed 
in this investigation. The best agreement in this instance was obtained 
with the shallow, blackened evaporation tank. 
The departure of the hourly evaporation rate of the porous-cup 
atmometer from the hourly transpiration rate of alfalfa is due largely (1) 
to the marked increase in the evaporation over transpiration during the 
night hours; (2) to the marked response of the atmometers to changes 
in wind velocity, which were not accompanied by corresponding changes 
in the transpiration rate; and (3) to the lack of a proportionate response 
on the part of the atmometers to changes in solar radiation. 
It should be emphasized in this connection that the failure of an 
evaporating surface to show a high correlation with the hourly transpira¬ 
tion rate does not necessarily imply a correspondingly low correlation 
on the daily basis. This is strikingly illustrated by the hourly evapora¬ 
tion rate from the deep tank, which, in these experiments, shows practi¬ 
cally no correlation with the hourly transpiration rate, but which on a 
daily basis was found in 1914 to be correlated with the daily transpira¬ 
tion rate to the extent of 0.63 ±0.01. 
