the Branches in Shrubs and Trees . 
641 
a slight straightening and hence outward movement, which might in the 
uninjured plant remain constant, such a cause was wholly insufficient to 
account for the entire amount of movement. That the swelling of air in 
the stem did not produce the result was proven by forcing air powerfully 
into the stem with a foot pump, a process always without appreciable 
result. 
Having thus to abandon this hypothesis I turned to another, more 
than once taken up and dropped in the earlier part of the study, that the 
movement was in some way connected with the quantity of water present 
in the stem. This was, indeed, very strongly indicated by two facts : 
(a) the Broussonetia earlier referred to (p. 634) showed a continuous in- 
curving of its branches after the plant was dead, which incurving was 
apparently correlated with the drying out of the branches ; and (b) invariably 
during the experiments a drying out of any branch was accompanied by 
an incurving, that is, by an inward movement. The incurved, or extreme 
inward position, is evidently the natural position of the dry tissues, and it 
seemed probable, therefore, that the outward movement might be corre- 
lated with, and proportional to, the amount of water in the stem. This 
supposition could evidently be readily submitted to experiment. Accord- 
ingly on certain days showing extreme outward and inward movement, and 
therefore of extreme high and low temperature, during the winter, I cut 
from each of these shrubs ten healthy branches each 10 cm. long, tied them 
in bunches, and immediately weighed the latter. They were then dried for 
several months in a dry room, and subsequently for some days in a water 
bath. They were then again weighed, and the percentage of water in the 
original branches was thus readily determined. The results were as 
follows : — 
Date. 
Temp. 
Plant. 
Original 
Dry 
A mount 
Percentage 
Weight. 
Weight. 
of Water. 
of Water. 
Jan. 19 
-18 
j Linder a 
3*991 
2*552 
i *439 
36.0 
( Salix 
2-960 
1*572 
1.388 
46.8 
Jan. 22 
8 
i L indera 
\ Salix 
3.628 
2-613 
2.281 
I * 37 I 
i *347 
1-242 
37 *i 
47*5 
Feb. 20 
-11 
{ Linder a 
4.027 
2.465 
1.562 
38*7 
( Salix 
2.785 
1.490 
1.295 
46-4 
Feb. 21 
-15 
Lind era 
4-032 
2.494 
1*538 
38 *i 
Feb# 24 
( Lind era 
3*572 
2.220 
i* 35 2 
35 *° 
5 
\ Salix 
2.884 
t * 5 2 3 
1.361 
47.1 
Mar. 13 
15 
i Linder a 
3.820 
2.072 
1.748 
45*7 
1 Salix 
2.870 
!. 4 83 
1.387 
48*3 
Comparison of these figures with the amount of movement at the 
corresponding dates (as shown on Fig. 57) will show at once that in Salix 
the agreement between amount of water and amplitude of movement is 
very close, a fact graphically illustrated by the polygon at the foot of 
