482 Hill. — The Floral Morphology of the Genus Sebaea. 
stigmatic nature does not appear to have been suspected, and it was not 
until living plants could be examined that these bodies were definitely 
proved to be efficient stigmas. 
In some species the secondary stigmas are much more in evidence 
than in others, and as a rule they are not well defined in the newly-opened 
flower. While the flower remains open, however, they become more 
prominent, and in withered flowers they are often seen to be much enlarged 
and covered with pollen, which seems to have been brought into contact 
with the stigmatic surfaces largely by the constriction of the throat of the 
corolla tube due to torsion. Some pollen, no doubt, is shed on the secondary 
stigmas when the anthers open in the bud, and may eventually germinate ; 
but it seems probable that the somewhat late development of the lower 
stigmas may be related to the time of ripening of the terminal stigma, and 
be so arranged that self-fertilization by way of the lower stigmas shall not 
take place until the pollination of the terminal stigma has become impossible. 
The arrangement of three stigmatic surfaces disposed in this way on the 
same style is, I believe, unique in the Vegetable Kingdom, and in order to 
understand their relationship to each other and to the stigmatic arrange- 
ments typical of the Gentians, it is necessary in the first place to give 
a short account of the typical terminal stigma and its modifications. In the 
simplest cases the terminal stigma would appear to be a bilabiate structure, 
corresponding to the two carpels, with the stigmatic papillae borne only 
on the inner surfaces of the lobes. Stigmas of this type are well shown 
by 5 . ambigua (PI. XXXV, Fig. 18) and in most specimens of S. aurea 
(Fig. 5). Many species are very variable, and though the bilabiate 
character is apparent the lips do not as a rule open widely, but are pressed 
closely together, leaving only a small cleft at the apex. The edges of the 
stigmatic lobes 9 however, are often somewhat rolled over, so that there 
is a fairly large surface exposed over the top of the stigma and down its 
sides as far as the base of the cleft or primitive lobes. 
A further stage is represented by those species in which the stigma 
is truly clavate and all trace of the bilabiate structure is obliterated, the 
stigmatic area being more or less disposed in a band passing over the top 
of the stigma and continued for a certain distance on either side (PI. XXXV, 
Figs. 13, 20). 
The relation of the primary terminal stigma to the secondary lateral 
stigmatic areas may now be considered. In most Sebaeas the terminal 
organ is widely separated from the lateral ones, but in a few species the 
secondary stigmas may be practically confluent with the terminal one, or 
very close to it. It is the species of the latter type which furnish the clue 
to the probable origin of the secondary stigmas. The arrangement found 
in X. khasiana is of particular interest, as there is no line of separation 
between the lobed terminal stigma and the lateral secondary patches, and 
