38 
BROWN. 
that a similar curvature might be produced in dead pulvini if 
the proteids of the cells could be coagulated around the osmot- 
ically active substances thus forming osmotic cells. A reduc- 
tion in the pressure within the cells could then be obtained by 
replacing the water in them with some liquid in which the 
osmotically active substances would not dissolve. 
In order to test this hypothesis a large number of leaves were 
killed in boiling water. The pulvini used were those of the 
leaflets as these could be killed more quickly than the larger ones 
of the petiole. When leaves were killed after the production of 
curvature in the pulvini the pulvini remained curved; but at- 
tempts to kill the pulvini, before bending had taken place, were 
only partially successful as they were always stimulated, to some 
extent, by the treatment before losing the power of responding. 
However, thirty-eight leaves were obtained in which the curv- 
ature of the pulvini was only about half as much as it would have 
been if completed. In order to remove the water from the cells 
of the pulvini the leaves were run up through several grades of 
alcohol to absolute alcohol after which they were transferred 
to xylene. Since sugars are practically insoluble in this it would 
seem probable that there could be little or no osmotic pressure 
in the cells of the pulvini after they had been placed in it. Those 
pulvini which had been killed after complete curvature showed 
no change in shape, while in those in which the curvature was 
only partial, it had been completed. Since in these later cases 
the effect of replacing the water with xylene was to reduce the 
pressure in all the cells of the pulvini, it would seem that the 
completion of curvature could have been caused only by a reduc- 
tion in the osmotic pressure in the cells of the concave or reacting 
half of the pulvini, without the aid of pressure due to the expan- 
sion of the cells of the convex half. Concave is used here to 
designate the half of the pulvinus toward which bending takes 
place, and convex the opposite half. These terms are substituted 
for upper and lower as applied to the pulvini of the petioles, since 
physiologically the upper half of the pulvini of the leaflets, with 
which we are dealing, corresponds to the lower half of the pulvini 
of the petioles. The production of curvature in the pulvini placed 
in xylene would indicate that the cell walls of the convex half were 
less elastic or more rigidly placed than those of the concave half, 
for an equal contraction of both halves would not produce curva- 
ture; and the osmotic pressure must have been largely removed 
from both. In this connection it is interesting to note that a 
longitudinal section of a pulvinus shows that on the concave half 
