226 MR. R. A. ROLFE ON THE APOSTASIER. 
latter with but two, together with other important differences 
pointed out elsewhere. Neuwiedia at once breaks up into about 
half a dozen species; but Apostasia first bifureates into two 
marked sections :—JMesodactylus, with the third stamen repre- 
sented as a narrow staminode, adnate to the back of the style, and 
the anthers versatile, with their bases unequal; and Adactylus, with 
the third stamen entirely suppressed, the anthers basifixed, with 
their bases quite equal. These groups then break up ; the former 
into three, the latter into two (known) species. Notwithstanding 
the marked difference between the two sections of Apostasia, still 
in habit and general appearance they are so thoroughly identical, 
that I do not think it advisable to consider them as genera ; though 
they are at least as distinct as some others so separated, and had 
each given rise to a large number of species, they might perhaps 
have been so distinguished. The one negative character correlated 
with the floral differences above mentioned is, that in both the 
known species of the section Adactylus the peduncle is a little 
elongated and covered with a series of lanceolate imbricating 
bracts, which are not present on the more sessile one of the other 
section. 
Returning now to the Apostasiee, the point in dispute with the 
different botanists who have treated of the group is not so much 
their characters (though some of these have been somewhat mis- 
understood), as the particular rank in the system of classification 
to which those characters entitle them. Those who treat the 
group as a distinct Order, at the same time uniting Cypripediee 
with Orchidee, take a view which, in my opinion, is wholly at 
variance with the structural peculiarities of the respective groups ; 
while, on the other hand, to regard both as distinct Orders would 
atleast render a similar subdivision of the Monandra necessary. 
In fact the difference between .4postasiee and Cupripediec is 
simply a developmental one, the latter group being a more highly 
specialized form, or development, of the same structural plan. 
Nor do I think Diandre and Monandre should be considered 
as more than distinct Suborders, for the amount of agreement 
between them is far closer than that between Orchidee and Bur- 
manniacee, the latter itself by no means a homogeneous group, 
though not so markedly subdivided as is the Orchidew *. 
* In Burmanniacee the perianth-segments are united at their bases into a 
tube, the upper part being free and divided into six segments. The inner whorl 
of three is generally smaller than the outer (or rarely quite suppressed), both 
