234 MR. H. BOLUS ON SOME CAPE ORCHIDS. 
In fig. 2 is represented the column of this species. A is 
the rostellum viewed anteriorly, showing two incurved arms 
holding a single gland g, the tip d Fig. 2. 
of a dorsal process holding the 
gland in its place, 2 the anther 
turned back from the rostellum, 
and s the stigma, which is dis- 
tinetly two-lobed. At B is shown 
ihe rostellum, with the gland and 
anther removed. At Ca posterior 
view of the same, with the anther- 
ease removed, showing the attach- 
ment of the pollinia. D the de- 
tached gland showing a median de- 
pression which receives the dorsal 
process d; but in the flowers I 
examined I observed no tendency 
in the gland to become divided 
along this depression into two Column of Disa barbata, Sw. 
glands. At E is shownan oblique (sn D 
view of the rostellum, every other part of the column being 
removed. 
Subsequently (in February ofthe present year) I had an oppor- 
tunity of examining living plants of Disa graminifolia, Ker,— 
Herschelia celestis, Lindl., and found the structure of the column 
to be closely similar to that of Disa barbata, Sw. 
Fig. 3. 
Column of Disa graminifolia, Ker, = Herschelia celestis, Lindl. (enlarged). 
Fig. 3 showsthe column of this species seen from different points 
of view :—A, anterior view, 7 the single gland, » the rostellum, s the 
two-lobed stigma ; B, posterior view, showing the dorsal appen- 
dage d; C, oblique view, with the anther removed ; D, gland, pos- 
terior view. In this species the dorsal process of the rostellum 
is bilobed at the apex. 
