48 TRANSPIRATION IN A DESERT PERENNIAL. 
and a new curve was made by connecting the points of intersection. Figure 
23 gives these resulting curves one space below their actual position in order 
to avoid numerous intersections with the original curves. This method of 
obtaining curves comparable in time is based on the assumption that both 
the relative transpiration and the water content change at a constant rate 
between the successive readings. 
DISCUSSION. 
The curves resulting from the plotting of water content determinations 
of the two trees, I and II, are fairly consistent in general shape. In the 
case of the end twigs there is a minimum water content about noon followed 
by a subsequent rise. The decrease begins in one case about 8" 30™ a. m. 
and in the other at 10" 30™ a. m.; in the former case recovery is slower than 
in the latter. In the case of branches a meter from the ends it may be said 
that in general the slopes of the curves are in opposite directions to the 
slopes of the twig curves. The branches 2 meters from the ends behave 
differently in the two trees; in Isthe curves are, in general, parallel to the I. 
curve, while curve III; is nearly flat or only slightly lower in the middle of 
the day than at the two ends of the day. 
The curve of relative transpiration from experiment III (see p. 20) is 
plotted at the top of fig. 22. Although-experiment III was performed in 
March and experiment XVIII in August, the conditions of soil moisture and 
evaporating power of the air at the two dates make it seem fair to compare 
the two sets of readings. The soil moisture at a depth of 30 cm., as given by 
weekly determinations made by Dr. Forrest Shreve, was 15 per cent on 
March 12 and 15.7 per cent on August 6. The summer rains which had 
intervened allowed for recovery from the extreme conditions of drought 
obtaining during the arid fore-summer. It may be, of course, that trans- 
piration conditions were different in the two cases, but the probability is that 
they were much the same. 
The following theory for the behavior shown is suggested. The relative 
‘transpiration rate increases until about 9° 30™ a. m., when the demand for 
' water made upon the plant by the evaporating power of the air exceeds the 
supply of water furnished by the transfer from the lower portions of the plant, 
and there results a slight drying out of the walls of the cells which form the 
true transpiring surface, both internal and external. This operates in two 
directions: the one, a lessening of the relative transpiration rate because the 
evaporating surface of the intercellular spaces has receded into the pores of 
the cell walls; the second, an assumption of the tensile state of the water in 
the vessels and cell walls, which results in a “pull” of water from the lower 
portions of the same stem and, as an immediate result of this, a transfer 
of water from lower to higher portions faster than the supply still farther 
down can stand, and therefore an increase in the size of the air bubbles in 
the vessels caused by the vacuum formed and hence a lower water content 
