284 
Journal of Agricultural Research 
Vol. IX, No. 9 
loss of water from a large body of water, heat energy being stored in both 
systems during the day. The transpiration graph of alfalfa may be as¬ 
sumed to represent approximately the distribution of the hourly loss 
of water from any actively growing vegetative cover. The hourly loss 
from large bodies of water thus appears to be more nearly uniform 
throughout the 24-hour period, while the loss from land areas covered 
with vegetation is confined almost wholly to the daylight hours and 
largely to the midday period. 
influence; of wind velocity and sol,ar-RADIation intensity on 
EVAPORATION 
It is of interest to consider the departure of the various evaporation 
graphs from the transpiration graph in relation to the intensity of the 
several climatic factors. It will be noted that all the atmometer graphs 
showed a relatively low evaporation from 10 a. m. to 4 p. m. on July 20 
compared with either the transpiration or the shallow-tank evaporation. 
The explanation of this may be found in the wind velocity, which is 
relatively low during this period, being less than 2 meters per second 
(4 miles per hour). The relatively high evaporation from the atmom- 
eters from 2 to 6 p. m. on the following day (July 21) is also ex¬ 
plainable on the basis of the high wind velocity during this period. 
The maximum evaporation from the atmometers on the third day also 
coincides with the period of maximum wind velocity. It is evident, 
therefore, that the evaporation from the atmometers is affected to a 
much greater degree by variations in wind velocity than either the trans¬ 
piration or the evaporation from the shallow tank. 
On the other hand, the atmometers are less sensitive to changes in 
solar radiation than either the plant or the shallow tank, as is shown 
by the afternoon readings on July 22. Both the transpiration graph and 
the shallow-tank evaporation graph show maxima at 11 a. m. and 3 p. m. 
on this date, due to a temporarily clouded sky from 12 noon to 2 p. m. 
The radiation on July 22 was also deficient from 7.45 a. m. to 11 a. m. 
due to the passing of cumulus clouds. This reduced the transpiration 
and shallow-tank evaporation and brought the transpiration graph 
more nearly into conformity with the atmometer graphs during this 
period. 
atmometer calibration and observation errors 
Since four instruments of each type were used in the atmometer 
readings, it is possible from the results to determine the calibration 
error and the probable error of a single observation for each individual 
atmometer. To each determination was applied the calibration cor¬ 
rection which accompanied the instrument (see footnote, p. 278). 
The ratio of the individual observation of each instrument to the mean 
