296 BOTANICAL GAZETTE [APRIL 
instance it was found that the low curve was caused by the soil 
water being reduced to the wilting coefficient. Such could not be 
true in this case, however, as the growth water was in excess all 
the time. The cause of depression at station J was found to be 
associated with abscission. This was likewise an influencing factor 
in the similar depression formerly referred to, being induced there 
by the inadequacy of water supply. 
WILTING COEFFICIENT.—The effect produced on transpiration 
when the soil water falls to the point of the wilting coefficient 
was not observed in any of the work on clay soils, for the growth 
-water content at the time of all readings was found to be adequate 
to meet the requirements of the plant. Its effect has been referred 
to, however, in connection with the dune data, where in the dune 
forest it develops in August. Transpiration under such conditions 
remains almost entirely cuticular, with the exception of a slight rise 
in the early morning, when the reserve water accumulated during 
the night is being utilized. 
RELATIVE HUMIDITY.—Relative humidity is considered one of 
the most potent of the atmospheric factors influencing the tran- 
spiration stream. It is the direct cause of the establishment of a 
diffusion gradient between the internal atmosphere of the leaf and 
the external atmosphere. Other physical factors initiate change 
in relative humidity, such as temperature, but because transpiration 
is a molecular diffusion problem, it should be interpreted as bearing 
its closest relation to those factors which initiate and directly 
influence this process. Relative humidity has this close relationship 
to the foliar water loss, and slight sudden changes in this factor are 
usually registered in the transpiration curve, even though the porous 
cup makes no record of it. Reference to fig. 9, graph 2, will illus- 
trate this relation. Curtis (2), in observing the effect of relative 
humidity, states that an increase of 8 per cent was followed by a 
pronounced reduction in the transpiration rate. 
EvaAporaTIon.-—The relative humidity and evaporation curves 
do not always show so close a parallel as might be expected (fig. 9, 
graph 3), so that in some instances there is greater correlation 
between those representing temperature and evaporation than 
between relative humidity and the latter. This divergence is very 
