Hill . — The Floral Morphology of the Genus Sebaea. 487 
pollinated by a minute thrips, but beyond his observations we know nothing 
of the insects which may visit the attractive flowers of Sebaeas under 
natural conditions. Should his observations be found to hold good through- 
out the genus, it would seem probable that the insect adapted to fertilize 
these flowers may have ceased to visit them, since one assumes that flowers 
so constructed and so conspicuous should be pollinated by fairly large 
insects and not by microscopic thrips. The suggestion may also be 
hazarded that the ‘short-circuiting’ of the pollination process rendered 
possible by the development of the secondary stigmas may be regarded in 
the nature of a response by the plant to its changed biological conditions. 
It is no doubt the case that these creeping thrips serve as excellent pollen 
carriers to the secondary stigmas, and that in their winged condition they 
may take pollen from one flower to another, but it seems very doubtful 
whether they would pollinate the terminal stigma except under fortuitous 
circumstances. The genus Pleurogyne appears to afford the only com- 
parable case to Sebaea. Here the apical stigma is functionless, and self- 
fertilization is no doubt rendered particularly easy by the stigmatic bands 
which occur on the sides of the ovary itself (PI. XXXV, Figs. 1 and 2). 
The floral arrangements in Sebaea might be described as an illegitimate 
and condensed form of heterostylism, since, though stigmas are provided at 
two different levels, there is no corresponding arrangement of anthers 
differing in length ; self- rather than cross-fertilization would thus appear 
to be the object in view. Whether a case of true heterostylism exists in 
Sebaea is somewhat uncertain with our present knowledge, but it seems not 
improbable that S. T ho dean a, Gilg., 1 and X. spathulata , Steud., may repre- 
sent the long and short styled forms respectively of one and the same 
species (PL XXXV, Figs. 24 and 25). 
The specimens are very similar in general appearance, and the anthers, 
with their large black apical glands, are identical in structure ; moreover, 
they have been gathered in the same localities. Should this supposition 
prove to be correct, X. spatlndata will stand as the name for the dimorphic 
species. 
The only other aberrant species of the genus, with a style shorter 
than the anthers, is X. Thomasii , 2 a very distinct species. It may con- 
ceivably be conspecific with X. Marlothii , though this supposition is open 
to grave doubt because the two species are somewhat dissimilar as regards 
their vegetative characters. There is also, on the other hand, the possibility 
that the long-styled form of X. Thomasii may not yet have been discovered. 
The genus Exochaeniump which is closely related to Sebaea , is of 
interest in this connexion, since in E. grande long-styled, short-styled, and 
1 Fl. Capensis, iv, i, pp. 1091, 1092, and Kew Bull., 1908, p. 334. 
2 1 . c., p. 1902, and Kew Bull., 1908, p. 335. 
3 See Kew Bull., 1908, pp. 336-41 with plate. 
