56 BOTANICAL GAZETTE [ JANUARY 
roots. The loss of water due to greater permeability of the pro- 
toplasm accounts for the decrease or the reversal of the tension 
between the cortex and axial cylinder on the concave side; and 
it is especially satisfactory in explaining the actual shortening of 
the concave side which has been so often observed by different 
men. in grass nodes, dicotyledonous stems, and roots. In 
answer to the objection that mechanical compression may also 
shorten the concave side, I have already shown (p., 44) that 
there is at least a relative shortening of the concave side under 
conditions where compression cannot possibly be exerted by the 
convex cortex in the case of roots; and Noll’s Hippuris stems _ 
show the same thing for dicotyledonous stems. The objection 
that greater elasticity of the cell membranes may cause the 
shortening is answered on p. 51; but this elasticity is not’ 
excluded absolutely as one of the factors in the curvature. 
Noll has pointed out that in wilting plants the cell sap may 
become more concentrated and its osmotic equivalent greater 
while turgor falls; hence the greater permeability of the proto- 
plasm in the cells of the concave side is not contradictory to 
those observations of Kraus and Kohl which show an increased 
amount of osmotic substances on the concave side. Indeed it 
may be that the very activity by which the protoplasm becomes 
more permeable results in an increased production of osmotic 
substances, as Vines has pointed out. 
The loss of water through greater permeability of the proto- 
plasm on the concave side is not disproved by the fact that the 
concave side often lengthens more or less during the curvature. 
Turgor might sink quite appreciably from the loss of water and 
the cells still be capable of growth in length ; and in those cases 
which lengthen, this growth more than counteracts the shortening 
effect. If the cells did not grow in length the walls could still 
grow in thickness, and the thickened walls observed by Wort- ; 
mann are thus accounted for. : 
The time required for root curvatures cannot be regarded as 
evidence against the increased permeability of protoplasm in| 
them, for MacDougal has found that tendrils coil in response to @ 
