MR. R. A. ROLFE ON THE APOSTASIEZ. 227 
The Diandre and Monandre evidently represent the two great 
diverging branches along which the Order has been evolved, the 
more ancestral Diandre having developed but two marked tribes, 
while the highly specialized Monandre has multiplied enormously, 
and given rise to several well-marked tribes and a large number 
of genera; all connected together by a very strong thread of 
affinity, and many of them separated from each other by very 
slight differences. 
With regard to the Oypripediee, a very curious point presents 
itself. The genus Selenipedium has retained the ovarian charac- 
ters of the more ancestral Apostasiee, while Cypripedium has a 
unilocular ovary with parietal placentation as in the Monandre. 
This cannot of course be held to constitute any affinity with the 
Monandr@, as Cypripedium clearly represents the culminating 
point of development of the Diandre. The trilocular ovary with 
axile placentation obviously represents the ancestral condition of 
the Order, and the development of a unilocular ovary with 
parietal placentation in each of the two diverging branches may 
possibly be an adaptation for saving room to accommodate the 
enormous number of seeds produced. In floral characters Seleni- 
pedium so clearly agrees with Cypripedium, that horticulturists 
generally treat the two as constituting a single genus, though 
from the above-named important difference, correlated with a few 
minor ones, I am convinced that Selenipedium should be regarded 
as a sufficiently distinct genus *. 
Affinities with Hypoxidee have been pointed out in the Apos- 
tasie@; but these are nothing but developmental analogies, for 
being regular, except that in the tribe Corsiee the median segment of the outer 
whorl (not the inner one, the lip, as in Orchids) is larger than the rest. The 
stamens are situated on the perianth-tube, six in number, both whorls being 
present, or the outer whorl suppressed in the tribe Zuburmanniee, when the 
stamens are but three, opposite the inner perianth-segments. The ovary is uni- 
locular with parietal placentation, except in two genera of Euburmanniee, where 
it is trilocular with axile placentation. In this respect the Order is in a transi- 
tion state; for in some genera the placent® intrude considerably, the ovary 
being three-celled at the extreme base, but only one-celled above. So that 
here a considerable range of variation is seen. 
* There are other genera which, neglecting some constant and important 
character, are equally difficult to separate. For instance Zria and Dendrobium, 
the one with eight, the other with but four pollen-masses ; also Octomeria and 
Pleurothallis, the one with eight, the other with but two, cannot be absolutely 
separated without reference to these highly important characters; yet every one 
considers these as good and sufficiently distinct genera. 
LINN. JOURN.—BOTANY, VOL. XXV. s 
