jan. 3 ,1916 Hourly Transpiration Rate on Clear Days 
647 
their maximum between noon and 2 p. m., after which they fell approxi¬ 
mately with the radiation. 
The transpiration during the night at Akron if very low, being only 
3 to 5 per cent of the transpiration during the daylight hours. 
The radiation graphs are practically symmetrical with respect to noon, 
showing that the days selected were relatively dear. When all the 
mean hourly values are expressed as a percentage of the maximum, the 
radiation intensity rises in advance of the transpiration (and in advance 
of all the other environmental factors as wdl) and falls either in advance 
of the transpiration or with it, depending on the plant considered. Radia¬ 
tion then may be looked upon as the primary causative factor in the 
cyclic changes. 
The air temperature and wet-bulb depression graphs are very similar 
in form, since the latter can be determined from the former on days 
in which the absolute humidity of the air is not changing. The transpira¬ 
tion graphs usually rise and always fall in advance of air temperature. 
The evaporation graph from the shallow, blackened tank (water 
approximately 1 cm. in depth) is similar in form to the graph representing 
the vertical component of radiation. This is to be expected, since only 
the vertical component would strike the horizontal water surface. The 
evaporation graph rises and falls with, or slightly later than, the vertical 
component of radiation. 
Computation of the correlation coefficients between transpiration and 
the various environmental factors shows the radiation, air-temperature, 
and wet-bulb depression to be correlated with transpiration approximately 
to the same degree. The correlation coefficients of transpiration with 
radiation range from 0.82 to 0.89; with temperature from 0.77 to 0.86; 
and with wet-bulb depression, from 0.75 to 0.85. These figures show 
the intercorrelations existing among the environmental factors, since 
the sum of the squares of the coefficients of independent causative factors 
influencing transpiration can not exceed unity. If radiation is taken as 
the primary causative factor, the correlation coefficients show that 0.67 to 
0.77 of the transpiration on clear days under Akron conditions is deter¬ 
mined by the radiation intensity. 
If the environmental factors are considered as independent, their rela¬ 
tive influence on transpiration may be determined by the method of least 
squares. In the case of alfalfa and amaranthus, the vertical component 
of radiation and the temperature enter into the determination of transpira¬ 
tion in the ratio of 1 to 1, approximately; and the corresponding ratios 
for vertical radiation and saturation deficit are approximately the same. 
On the other hand, in the case of rye, the radiation by this method of re¬ 
duction shows less influence than either temperature or saturation deficit 
on the transpiration rate, which from 9 a. m. to 2 p. m. shows a marked 
departure from the graph indicated by the transpiration during the 
early morning hours. 
17208°—16 - 5 
