26 BOTANICAL GAZETTE [JANUARY 
side tend to curve the stem toward that side. But this is bal- 
anced by an equal tendency on the opposite side to curve in the 
Opposite direction, and the stem remains straight. The poten- 
tial energy of this normal tension is available for producing a 
curve as soon as the equilibrium of tension is disturbed. This 
equilibrium may be disturbed either by an increase of tension on 
one side, in which case that side will become concave, or by a 
decrease of tension on one side, in wh.ch case that side will 
become convex, or both these changes may occur. Kohl (33, 
p- 87) thinks the first of these alternatives is the true one, 
because when a curved stem is split in halves the convex half 
Straightens or curves in the direction opposite to that of the 
whole stem. Kohl does not make sufficient allowance for the 
tissue tensions which are present previous to stimulation. The 
convex half of the curved stem will curve more in the original 
direction when separated from the other half, only in the case 
in which the tension on the convex side is reversed from the 
normal, that is, where the cortex becomes longer than the pith. 
This probably seldom happens. Kohl’s experiments taken with 
those of Sachs (63), in which curved stems were separated into 
three strips, concave cortex, pith, and convex cortex, show very 
conclusively that the stimulus influences the tension on both sides of 
the curving stems, and that the tension is decreased on the con- 
vex side, and increased on the concave side. Sachs’ experi- 
ments also show that the deviation from the normal is greater on 
the concave side than on the convex side, that is, the increase of 
tension on the concave side is greater than the decrease on the 
‘convex side. These tension relations explain the results of Noll’s 
bending experiments, from which he draws the conclusion that 
the membranes of the convex side are more extensible than 
those of the concave side, because the stems bend more easily 
toward the latter side. However, increased tension between the 
pith and cortex of the side becoming concave, even without the 
decreased tension on the Opposite side, causes a greater resist- 
ance to bending toward the convex side than toward the con- 
cave. Hence Noll’s conclusion does not at all follow from his 
rib Re MBit sy ie 
ee Sen 
