K.—BOTANY 193 
a rather horrific appearance. Above the entrance the massive wall 
protrudes as an overhang, and from the under side of this the door rises 
abruptly and not, as in U. cornuta, as an extension of the wall; and we 
have no reason for supposing that the overhang in U. capensis contributes 
to the action of the door, as in U. cornuta, in raising the angular divergence 
with the threshold, contributing to the thrust of the door edge. To 
compensate for this the inner zone region of the threshold has an upward 
slope, thus offering a resistant surface of sufficient angle to afford a 
resting-place for the door edge, while in the door itself there is a set 
transverse bend above the middle piece. 
The door, viewed as a flat object, has a rounded upper region, which 
is, as seen in the longitudinal sagittal section, relatively thin, with a much 
thinner outer course of cells than inner in its upper half and the reverse 
in its lower half. This is correlated with the set transverse flexure, the 
effect of which is to tilt the lower half of the door (the middle piece) with 
respect to the threshold, enlarging the angle of divergence. The lower 
region is differentiated into a large middle piece, whose cell courses are 
of nearly equal thickness, with thicker lateral regions having the same 
structure as the upper hinge region, i.e. thin outer course and thick inner 
course cells. As in U. cornuta, the lateral regions exert a thrust on the 
middle piece. Other features peculiar to U. capensis are, first, a tuft 
of rather large, clavate glands arising from the door upper region and, 
second, arising from the middle point of the upper limit of the middle 
piece, a single, curiously shaped, glandular trichome, which I have called 
the kriss trichome, as well describing the shape of the terminal cell. 
Its stalk (basal cell) is curved gracefully backward ; the mid-cell is short 
and oblique, holding the terminal cell in a backward-reaching position 
between the middle piece and the threshold. If this structure has any 
function, we do not know what it is. The absence of cuticle from the 
terminal cell suggests that, in common with the glandular trichomes, 
mucilage is excreted, but this does not throw light on the peculiar form, 
nor does it help us to know that in another species, U. puberula (New 
World), the door and general structure are similar in every detail but the 
absence of the kriss trichome, there being substituted therefor a pair of 
large, sessile, globular gland (capital) cells. Another species (Old World), 
U. Welwitschit, has, like U. capensis, a kriss cell, but it is more sharply 
curved, scimitar fashion. 
The actuation of the trap seems to be initiated by the contact of the prey 
with the short trichomes on the upper convex portion of the door surface. 
As in U. cornuta, the initial flexure thus caused is transmitted longi- 
tudinally to the middle piece, which, flexing along its midline, releases 
itself from the thrust of the lateral hinges and so the door opens before 
the moving water column. 
Almost identical in structure, as far as the entrance mechanism is 
concerned, is a group of species, the members of which have an elaborate 
guide complex; but here the radiating rows of glandular trichomes 
arise from a funnel-form elaboration of the front of the trap, while the 
upper sector of this funnel is drawn out into a long rostrum, as first 
described by Goebel (1891). The species, as far as known to me, are 
H 
