1900] MECHANISM OF ROOT CURVATURE 23 
of the cortex intact on two opposite sides. Some of the cortex 
necessarily had to be cut in getting to the axis, especially on 
the side where the scalpel was pushed in. A small amount of 
cortex was also cut on the side opposite to this, for the scalpel 
had to be pushed through till its point could be felt on the 
farther side. Part of these roots were branded on the side from 
which the scalpel was pushed in, part of them on the side oppo- 
site, and part of them go° from that side. A total of 304 roots 
were used, and of these 184 curved away from brand both above 
and below the stab in the root. That is, 60.5 per cent. showed 
the typical traumatropic curve more or less distinctly. 
The lot of sixteen tested November 16 was not looked at 
after branding for twenty-four hours, or the result might have 
been different. 1 cannot explain satisfactorily the low result on 
November 17 and 20, unless the stab was made too far from the 
tip. In most of the lots the stab was made between 2 and 3™™ 
from the tip, and the roots were looked at several times within 
the first eight hours after branding as well as twenty-four hours 
after. The results were taken when they showed the greatest 
number curved. The necessity of control experiments in which 
the roots should be stabbed and not branded, was obviated by 
branding on different sides as related to the position of the stab, 
Thirty-six control roots were branded, but not stabbed, and 
twenty-nine of them curved away from the brand, 80.5 per cent. 
In roots branded under the most favorable conditions only about 
5 per cent. fail to show the traumatropic curve. 
In order to be certain that the stimulus was not transmitted 
in parts of the axial cylinder which might not have been cut off, 
many of them were afterward examined microscopically. In an 
especially favorable lot of seventeen roots, of which sixteen 
showed the traumatropic curve both above and below the stab, 
twelve of them had the axial cylinder completely cut off, and 
those that had the axial cylinder partially intact were among 
those that curved least and straightened soonest. 
Six hours from the time of branding some of the roots that 
had curved somewhat were already beginning to straighten. 
