ii I^^TROI»tJCTION.
of cellular tissue, and of which the prolonged caudicles are enclosed in forward-projecting^
tubes, there being apparently no rostellum. These structural arrangements are seen in a
still more pronounced form in Diplomeris MrsiUa. They are also found in very many of
the South African OpJirydeae, and they are very clearly shown in the excellent
drawings of Mr. Bolus.* They culminate perhaps in TTahenaria Bonatea, a remarkable
South African species, the structure of the flowers of which has been figured with
marvellous skill and delicacy by Pranois Bauer in his well-known and magnificent
drawings. The incompatibility of the monantherous theory with the arrangements in
this and in other species of Eahenaria appears to have struck Mr. Darwin, who
refers to it in his book on the Fertilization of Orchids (pages 297 and 302, first
edition), but he did not follow up the subject.
I n my opinion the Sikkim species of Ophnjdeae have really two anthers, one
cell of each of which is fertile and the other cell infertile. The infertile cell invariably
occupies a position on the outer surface of the fertile one, and is the body
usually described as a "staminode".t The two fertile anthers belong in my opinion to
the inner whorl, the infertile anther being merged in the column. In the Sikkim
Ophrj/deae one of the three stigmas is infertile, and the two lateral are fertile.
These fertile stigmas are in many species quite distinct from each other; in other
species thsy are conjoined into a simple or bi-lobed mass. The infertile stigma
in many of the species is obscure; in others it forma a thickened and usually
curved line running between the bases of the antlier-cslb. In IMenaria stencmtha
and E. genicuUta, it takes the form of a triangular concave plate occupying the
lower part of the broad space betwen the bases of the anther-cells, but on a plane
anterior to the web of cellular tissue by which these cells are connected with each
other. The infertile stigma acquires its greatest development in Diplomeris hirsiita,
in which {see description on page 33S and figs. 1 and 2 on plate it forms a
large concave hood placed weU in front of the anther-cells, and (in a front view)
hidls the whole of the latter, except the extremities of the caudicles and their tubes.
I n this remarkable plant the two fertile pistils consist of two elongated parallel
bodies, stigmatic at the apex, which overhang the claw of the lip and point downwards
like the lip. The arrangement closely resembles that which obtains in the
South African species of Eahenaria of the section Bonatea.
The taxonomic matters in which there is in the following pages divergence
from the current views, consist {d) in the restoration of Lindley's tribe Malaxideae,
which has by most recent writers been merged in Epidendreae; [l] in the
restitution to the tribe Vandeae of a few genera (hereafter mentioned in detail)
wliich have of late been included in Epidendreae i (c) in the breaking up of the
* Orchidi of iioatk Afrint. arid Orc'iids of the CJDS Peninsula.
t I bare exnmincd many flowjrs nod dr.itrinsi oE species of Ophrjideae in tlio hopo of rmding a species in ivhich
flie infertile ceUs or stnminDdes ars atttiched on the inner- side, of ilic ferfilo cells. But the only case in which 1 hare
found such an nppearance is in Mr. Bolus's drawing of PUrygodium carnoiuv,, Lindl. (Orchids of the Cape Peninsula,
t. 12, fig. 51. It the slmctiircs Ibore represented os OT-jid rngulose lodiea really aro stamincdes, the iacl affords a
strong ronfirraation of my theory.
INTHODUCTION.
tribe NeoitiecB into two tribes, which we have named Lisiereae and Goodyereae.
I n all these changes Mr. Pantling and myself arc in agreement. We arc both
strongly of opinion that, in limiting the larger groups, it is (as far as the Sikkim
species go) the safest course to assign a supreme value to the structure and
appendages of the pollen masses. In working with fresh specimens we have found
that, in the genera which we have included in the old tribe Malaxideae, the
pollinia have absolutely no caudicle and no gland; and that, in the majority of
cases, they are perfectly free from each other; in only a very few, viz., in three
species of Cirrhopefalum and in the single species of Biglypliosa, being attached by
their bases to a ball of translucent substance. In the genera which we have
arranged in the tribe Epidendreae the pollen masses, on the other hand, arc attached,
either by their backs or by their bases, to a single or double translucent appendage,
but have no gland derived from the stigma. It may be that this limitation will not
be found to answer in the South American Epidendreae; and it is also possible that,
from the extreme delicacy of the translucent appendages, the character may be one
difficult to work in the Herbe^rium. However, as the present work professes to
be based on dissections of living, and not on those of dried specimens; and, moreover,
as the poUiniar differences are accompanied by certain differences in the fades of
the vegetative organs, wc feel bound to give full prominence to them. And in
doing so we are, as abeady indicated, only returning to the lines followed by that
great orchidologist, John Lindley. The tribe Epidendreae passes into Vandeae by a
transition suite of five genera {Tiptolaria, Monomeria, AcroeJmne, Agrostophyllum,
and a new one which we have named Ritaia), in all of which the poUiniar appendages
are attached to a gland the origin of which from the stigma is doubtful. The
transfer of CryptocUlas, Calanthe and OreoreUs to Vandeae was forced upon us by
the fact that the gland terminating the caudicle is, in all three, unmistakeably derived
from the stigma. Tor this reason we have also restored to Vandeae the genus lone
which had been merged in Bidhophjlhm,—a genus to which it has undoubtedly a
great resemblance in its vegetative organs.
I t only remains to say a few words in explanation of the re-arrangement of
the genera formerly referred to Neottieoe. "We found that these genera may, as we
believe, be arranged in two groups by following the characters afforded by the pollen.
I n one group the pollen is powdery (very rarely granular); in the other it is
always granular or sectile. The plants with powdery pollen have a terminal anther,
and their pollinia cohere by their sides, but they have neither caudicles nor glands.
These we have formed into a tribe under the name Listereae. The genera with
granular or sectile pollen have a posticous anther, and the poUiuia are attached to a
single gland by either one or two caudicles. These we have formed into our tribe
Goodyereae. The location in Listereae of Epipactis and CephalantJiera appears rather
anomalous; and there is no doubt that, in their vegetative organs, they resemble
no other genus in the group; but their pollen is .unmistakably powdery.
Om- study of the Sikkim species convinces us that the fertilization of orchids by
insect agency is by no means so universal as is sometimes supposed. "We have found