XI 
to the pollen masses, but it is also, in numerous other genera, at all times 
inseparable from it. In Bonatea, in Habenaria, and in some other 
genera of Ophrydeae, there are two arms to the upper edge ot the 
stigma, each arm being channelled for the reception of the caudicula 
of a pollen-rnass, and terminating in a separable gland; between these 
lies a membrane, very variable in size, sometimes merely a connecting 
web, sometimes a distinct plicature or lobe, and occasionally fornicate 
and extended in the middle into a tnuero. In these genera, on the lower 
edge of the stigmatic excavation are two processes, long, narrow, chan¬ 
nelled, lined with stigmatic hairs, tree at the points, united to the 
labellum at the base. Dr. Brown regards the two upper arms and 
intervening web of the superior edge of the stigmatic excavation as 
belonging to one carpel, and the two imperfect processes adhering to 
the base of the labellum, as the styles of the two other carpels; this 
view of their nature is the necessary consequence of assuming that the 
carpels of which the ovary is composed, are placed opposite the sepals. 
But it appears more probable that in such plants the double placenta? 
are connected with a two lobed style to each carpel; a structure which 
would be analogous to what occurs in this and in other Iridaceous 
plants, to which Orchidaceae are very nearly allied. Upon this sup¬ 
position one arm of each of the fertile styles must be regarded as sup¬ 
pressed, or united with the contiguous imperfect arm into a plicature or 
mucro, and the two sterile processes will belong to the carpel which 
is opposite the labellum. That the stigmatic arms are opposite the 
petals is manifest in most of the genera where they are found, and 
especially in such plants as Disa, Penthea, &c. in which the third and < .*> j > 
sterile stigma forms a tubercle at the base of the labellum. That the 
connecting web of the fertile arms is of a double nature is also shewn by r 
its lobed condition in Diplomeris hirsuta, Bonatea flexuosa, and others. 
It results from the preceding statements, that the theoretical structure 
of an Orchidaceous flower is as follows :—Sepals 3, usually suppressed. 
Petals 3, commonly called sepals. Stamens 6, in two series, of which 
the outer is sterile, and commonly called petals, and the inner is conso¬ 
lidated into a column : the two stamens of this series, which are opposite 
the so called sepals, being usually abortive. Ovary inferior, composed 
of 3 carpels, having double polyspermous placentae in their axis, and 
