194 SECTIONAL ADDRESSES 
U. albina, U. nivea, U. rosea and U. Warburgit, the last two having been 
described in general terms by Goebel. Three other collections, sent 
me by Mr. N. D. Simpson (his Nos. 9579, 9857, 9871) from Ceylon, 
though possibly differing specifically or varietally, have the same general 
structure. They are all Indian (with Ceylon) and found nowhere else, 
so far as known. In all these the glandular armature of the door is as 
in U. capensis, except for the absence of the kriss or other similarly placed 
trichome. The mechanics of the door and threshold are doubtless the 
same. 
Another small group of species (only two, so far as known) of minute 
plants with the habit of U. capensis harmonises with the above in regard 
to the entrance structures. The one is a Tasmanian, U. Jateriflora, 
briefly and inadequately described by Kamienski, the other a Ceylonese 
species (Figs. 13, 14), collected by Mr. N. D. Simpson (9482) and sent 
to me (my No. 131). It seems to be undescribed. My Jateriflora 
material has two derivations : one lot transmitted by Dr. Merl, but col- 
lected long ago by Rodway in Tasmania (my No. 85); the other, a well- 
preserved lot from Mr. Allan McIntyre, of Hobart (my No. 157). 
U. lateriflora has a very small trap (including the proboscis, 0-65 mm. 
long—a large one) with a huge conical downward-turned proboscis and 
with two short rows of glandular trichomes extending outward and 
obliquely downward from the lower angles of the entrance, increasing 
in size, the smallest at the entrance. They thus form two oblique shelves 
leading to the entrance, which is blocked in front by the proboscis. The 
capital cells of these trichomes are globose. ‘The door is quite like that 
of cornuta, but the armature of sessile glands is aggregated into a patch 
on the convex surface of the upper part. The articulation of the door 
with the wall is, however, like that in U. capensis. The external glands 
are sessile globose. 
The Ceylon plant (Figs. 13, 14) has a trap which, at a casual glance, 
is quite similar to the foregoing. It has, however, a second pair of rows 
of protuberances on each side and above the level of the entrance opening, 
but these are not glandular (Fig. 13). The basal cells of the other rows 
of trichomes are double-celled. The proboscis is smaller and may 
project straight forward or be more or less bent downward—in this 
there is no constancy. The threshold is composed of fewer and rela- 
tively larger cells, but is clearly zonate, the outer, middle and inner zone 
being readily recognisable. The glands of the outer surface have elongated 
capital cells. 
On the interior surface of the trap there are single, bifid and quadrifid 
absorbing trichomes, but these are very few in number and relatively 
large. In U. Jateriflora there are fourteen bifids in five rows on the 
inner flank of the threshold; six quadrifids above the inner entrance, 
three on each side; and two to six quadrifids in the interior, making 
about twenty-six in all. 
THe Type U. c@RuLea (Figs. 9, 11, 12). 
The mechanical principles prevailing in all the foregoing prevail also 
in a number of species, of which U. caerulea may be taken as a type. 
