May 28, 1917 
Transpiration Rate of Medicago sativa 
283 
The deep tank gives little indication of the hourly transpiration rate, 
since the average departure in this case amounts to about 90 per cent 
of the mean evaporation rate. In other words, the evaporation from 
the deep tank at different hours of the day is not proportional to the 
transpiration of the plant. 
The average hourly departure in the case of the white cylinders amounts 
to about 50 per cent of the mean hourly transpiration. In other words, 
the average error in predicting the amount of transpiration at each hour 
in the day from the evaporation when the ratio of daily transpiration 
of the plant to the daily evaporation of the atmometer is known would 
amount to one-half the mean hourly value of the transpiration. When 
the brown cylinder, Bellani plate, or white sphere is used, this error is 
reduced to about 40 per cent. In the case of filter-paper evaporimeter 
the error is reduced to about 30 per cent. The best results were obtained 
with the shallow, blackened tank, the error being less than 20 per cent. 
Since the night values of transpiration and evaporation are low and the 
greater part of the evaporation and transpiration occurs during the day¬ 
light hours, the writers have also computed the mean departure of each 
of the different instruments on the basis of the daylight hours only. 
The departures, with the exception of the deep tank, are, in this case, 
about two-thirds of those for the total period. The evaporation rate of 
the shallow, blackened tank during the daylight hours again follows the 
transpiration graph more closely than that of any of the other instru¬ 
ments, the mean departure being only about two-fifths that observed in 
the case of the atmometers. The filter-paper evaporimeter also shows 
a lower mean departure than the atmometers. Of the latter instruments, 
the brown cylinder, white sphere, and Bellani plate are practically identi¬ 
cal and show a departure of about 30 per cent, while the white cylinder 
shows a somewhat greater mean departure. The similarity of the results 
from the three first-named types is remarkable when the difference in 
form and color are considered. 
The evaporation from the deep tank shows practically no correlation 
with the transpiration when the hourly values are considered. Attention 
has already been called to this discrepancy, which results from the storage 
of heat energy in the great mass of water during the day and its slow dis¬ 
sipation through evaporation during the night. The maximum in the 
evaporation graph from the deep tank occurs in the late afternoon or 
early evening. 
The fact that an evaporating surface shows a low correlation with the 
hourly transpiration does not necessarily imply a correspondingly low 
correlation on a daily basis. This is demonstrated in the case of the deep 
tank, which showed in 1914 (a dry year) a correlation with transpiration 
of o.63±o.oi (Briggs and Shantz, 1916a), while the correlation of the 
shallow tank for the same period was 0.72 ±0.01. The evaporation from 
the deep tank probably represents approximately the distribution of the 
