1900] . MECHANISM OF ROOT CURVATURE 4y 
with the cut face upward. The up and down nutation exhibited 
by roots in that condition can only be due to the varying tension 
between cortex and axial cylinder in the half root that becomes 
concave; and the fact that more than half of the roots experi- 
mented upon in that way finally curved downward proves that 
the tension may be often reversed in the concave half of the 
root. This conclusion is also indicated by the relative length 
of the separated tissues of curved roots, though in these cases 
compression of the concave cortex is not eliminated. Experi- 
ments similar to those just mentioned, but with different details, 
prove the increased tension on the convex side. 
6. Not only is the tension between the axial cylinder and the 
cortex on the side that becomes concave changed by the stimu- 
lus, but the different layers of the cortex itself change in their 
tensile relations to one another. In curving roots the outer lay- 
ers of the cortex on the concave side are under negative tension 
in relation to those layers lying more central. This is shown by 
the fact that when the tissues of curving roots are separated the 
concave cortex curves more than the whole root has done, often 
curving with a sudden jerk as soon as separated, and forming a 
curve with a very short radius. That this is not due to the nor- 
mal tension of the epidermis and root cap is shown by compari- 
son with the convex cortex and with the isolated cortex of 
unstimulated roots. It is not impossible that forcible compres- 
sion of the outer layers of the concave side may sometimes play 
apart here. Such compression necessarily takes place in the 
outer layers of this side notwithstanding its shortening, unless 
the outer layers are shortened: more than the inner ones as a 
direct result of the stimulus. If the greater shortening of the 
outer layers was merely the result of compression, we should 
expect them to lengthen when freed from that compression, and 
the isolated cortex would then straighten. Since the isolated 
cortex curves more strongly than before isolation, it seems prob- 
able that the change produced directly by the stimulus in the 
cells of the concave side is greater in the outer layers of that 
side than in the inner layers. 
Mo. Bot. Garden, 
1900. 
