K.—BOTANY IQI 
thrust on the middle piece of the door (m.z., Fig. 8), which has quite 
a different structure, in that the two courses are of equal depth, the cells 
small and very densely studded with cellulose props, which appear, in the 
face view of the cells, as more or less regular transverse bands. These 
have the same function as elsewhere, but here are more numerous, because 
of the severe flexing which the middle piece is subjected to during initial 
opening of the door. ‘The properties of the middle piece are stiffness 
and capacity of flexure in either sense. When the door is closed, the 
middle piece is held firmly against the middle reach of the middle zone of 
the threshold (/.4., Fig. 5) by the downwardly thrusting lateral hinges 
Fig. 8). 
in sddition to the door proper, the beak wall cells also take part in the 
flexures of the door, and further exert a longitudinal thrust on the door, 
so that this is pushed backward and downward. The thrust of the door 
edge is then in the direction indicated by arrow c, Fig. 3, and is down- 
ward against the threshold middle zone. The thrust exerted by the beak 
is referable to the total flexures of the trap when exhausted ; when the 
trap has been sprung and is relaxed, the beak is not bent downward so 
sharply (d;, Fig. 3). 
By making photographic silhouettes of traps in the set, relaxed, and 
totally relaxed conditions, the difference in shape of the curvatures can 
be recorded. It has thus been possible to record the position of the door 
under these conditions. In the set condition the door, as betrayed by 
the longitudinal aspect, is bowed upward, and the transverse curvature at 
e and f, Fig. 3, will be flatter (Fig. 6). This has the effect of a more 
directly downward thrust of the door edge. After actuation the sil- 
houette of the door is as shown by the broken line d;, Fig. 3 ; the beak 
being less bent and the thrust due to the beak being eliminated. If total 
relaxation is obtained (as by puncturing the wall of the trap), the door 
takes the position d,—the dot-and-dash line in Fig. 3. These contours 
of relaxation serve to emphasise the significance of the curvatures seen in 
the set condition, when the lateral areas (/.h., Fig. 4) clamp the middle 
piece firmly on the threshold (Fig. 8), and the upper region of the door is 
under transverse tension, bowing the door longitudinally. The chink 
between the door and threshold is now filled with the membranes of the 
velum, making the whole watertight. 
Actuation is procured experimentally by slight pressure of a needle 
point (the operation must be without damage to the tissues) on the surface 
of the door in the region e-f, Fig. 3. Sometimes a very light touch will 
do the trick, but I get the impression that the mechanism is not so sensitive 
as that of U. purpurea or U. vulgaris. The smallness of the trap makes 
experimentation difficult. What happens in nature can only be inferred. 
It is to be noticed that there is a total lack of appendages supposed to act 
as guides and lures, the only equivalent being the gland below the 
‘entrance and the glandular trichomes on the door surface, aggregated 
chiefly in the region of actuation. But it is not difficult to make the 
inference that a small animal (Schimper mentions rotifers, worms and 
“ For a full discussion of the histology of the door, see my paper of 1932 in the 
Canadian Journal of Research. 
