282 
Journal of Agricultural Research 
Vol. IX, No. 9 
The ti^nspiration graph has also been superimposed on each of the 
evaporation graphs in figure 1, choosing the scale of ordinates in each 
instance so that the total area under the transpiration graph for the three 
days is equal to the area under the evaporation graph. This affords a 
graphical comparison of the fidelity with which the transpiration is 
reflected in the hourly evaporation rate from the different instruments. 
&<>;/&/& c, 
& //ASf & // s //s4St & /s & //A/# & sf 
Fig. 2.—Ratio of the transpiration rate to the evaporation rate for each of the various types of porous- 
cup atmometers, plotted hour by hour. 
MEAN HOURLY DEPARTURE OF EVAPORATION AND TRANSPIRATION 
A quantitative expression of the relative value of the several instru¬ 
ments in predicting the hourly transpiration may be obtained by deter¬ 
mining the average departure of the evaporation graph from the trans¬ 
piration graph, the mean departures being expressed in percentage of 
the mean hourly evaporation in each instance. The calculation has been 
based both on the total period covered in the measurements and on the 
three daylight periods from 6 a. m. to 6 p. m. These computations are 
presented in Table II. The error involved in predicting the transpira¬ 
tion rate from the evaporation measurements is indicated by the fig¬ 
ures in the last column of the table. 
Table II .—Average hourly departure of evaporation rate from transpiration rate in per¬ 
centage of the mean evaporation rate for various types of atmometers and for free water 
surfaces 
Type of atmometer. 
Day peri¬ 
ods, 6 a. m. 
to 6 p. m. 
Total pe¬ 
riod. 
White cylinder. 
Per cent, 
38 
29 
3* 
30 
22 
Per cent. 
49 
41 
43 
4i 
31 
17 
89 
Brown cylinder. 
White sphere. 
Bellani plate.-.:. 
Filter-paper evaporimeter. 
Shallow tank. 
12 
Deep tank. 
93 
