44 BOTANICAL GAZETIE [ JANUARY 
times before the direction of the curve was finally determined. 
At the end of twenty-four hours more than half were turned 
down at the tips, some of these tips being in a vertical position, 
others only slightly below the horizontal. This behavior indi- 
cates beyond doubt that there is a shortening of the cortex in 
relation to the axial cylinder on the side of the root that 
becomes concave, for in these roots with the upper half of the 
tip removed there can be no compression of the cortex on the 
lower side by the increased growth or extension of the cortex 
on the upper side, and the normal tension between the cortex 
and axial cylinder of a half root is such as to cause it always 
to curve with the cut face concave. It has already been shown 
that a half root stimulated on the outer side of the tip curves 
with the cut face concave not only more rapidly than a whole 
root so stimulated, but also more rapidly than an unstimulated 
half root (fig. 7). We are forced to the conclusion that a stim- 
ulus to the root increases the tissue tension between axial cylin- 
der and the cortex on the side that becomes convex, and 
decreases or reverses the tension between the axial cylinder and 
the cortex on the side that becomes concave. These changes 
on the opposite sides of a curving root are independent of each 
other, and neither side of the root is passive in the curvature, 
but both are active in producing it. Those theories, therefore, 
which attempt to account for the curvature by a change on only 
one side of the root, whether on the convex or concave side, 
cannot be accepted. 
I have made numerous attempts to find a difference in the 
turgor of cells on the concave and convex sides of the curving 
root, but was not able to demonstrate such a difference. Solu- 
tions of KNO, of varying strength were used, and the two 
sides always plasmolyzed in the same solution and in the same 
time. Ina few cases it was noticed that the cells of stimulated 
roots plasmolyzed with a weaker solution than straight unin- 
jured ones that had been under precisely the same conditions. 
The point was not pursued far enough to show whether or not 
the difference was merely an individual variation. 
