1900] MECHANISM OF ROOT CURVATURE 27 
results. Without at present going into the question as to how 
stimuli cause the changes in tensions of normal stems, it seems 
very certain that changes of an opposite kind are produced on 
opposite sides of curving stems. Those writers, therefore, who 
attempt to explain stimulation-curvatures by a change on only 
one side of the curving organ are overworking a half truth. 
On taking up the study of roots the question presents itself 
as to what the tissue tensions of roots are and how stimuli 
affect them. Hofmeister asserted that there were no tissue ten- 
sions in roots. More recently Pfeffer has written (53, p. 31): 
“In the zone of roots in which growth in length takes place, 
in spite of differentiation of tissue, only a slight longitudinal 
tension is shown.” .... “A certain negative tension is to be 
assumed for that save of the axial vascular cylinder which is still 
capable of growth, since a gradual curve results with the con- 
cavity toward this part, if a root split longitudinally is kept under 
proper conditions.” Pfeffer refers to the experiments of Sachs 
(62, p. 435) for the truth of his statement. My own experiments 
lead me to believe that the tissue tensions of roots are more 
important than has been supposed, and that they play a consider- 
able part in the formation of curvatures. That the tension 
between the axial cylinder and cortical parenchyma in roots can 
become considerable is shown by the fact that in Vicia faba roots 
growing rapidly in water the axial cylinder is frequently ruptured 
transversely by the more rapidly growing cortex. These breaks 
in the axial cylinder become visible only when the root is cut 
into longitudinal sections, and I have found as many as six breaks 
in the same root within a piece of the root 2-3 long. Besides 
the method used by Sachs to show tissue tensions of roots, 
namely, splitting them into longitudinal halves, I have also used 
the method he and others have applied to stems. If the cortex 
is carefully separated from the axial cylinder and the separated 
lamellae left attached to that part of the root above the zone of 
separation, the lamellae of the cortex project beyond the axial 
cylinder. If the tip of the root is cut off square before the sepa- 
ration, the difference in length after separation becomes visible 
