the point, shorter than the scape: racemes erect,
flowers short pedicelled, supported by a rather large
scariose bract as long as the pedicel: sepals ovato-
lanceolate, all equal, and beardless: style about the
length of the stamens, capitate: capsule large, obso-
letely 3-angled, 3-sided, seed obovate, orbicular, compressed,
winged, shining black. Embryo about the
length of the albumen.
Sea Coast, station not stated.
This differs in some respects from Roxburgh’s
description, which unfortunately does not include any
account of the capsule and seed; I however, believe
it is his plant.
2064. U rgenia congesta (R. W.), leaves -linear
subulate, about the length of the scape: scape erect,
naked, raceme short, compact: flowers short pedicelled,
supported at the base by a short broadish
obtuse scariose bract: sepals lanceolate, the inner
slightly smaller: ovary conical: capsule sub-obovate
or globose, 3-celled: cells few- (3-4) seeded: seeds
orbicular, bound all round by a broad wing, shining
black. 6
Sea Coast, Malabar ? station not mentioned.
The specimens from which these drawings are taken
were not collected by me, hence the want of stations.
They are all referable to the very modern genus
Urgenia which was separated from Scilla on account
of its numerous much compressed, not few globose,
seeds, which is its distinguishing characteristic.
CoMMELTNACEJa
,.^ h is, in the most favourable circumstances, is a
difficult order to deal with as regards the discrimination
of species, and in giving representations of the
flower can only be done justice to from growing
plants, hence I infer our comparatively imperfect
acquaintance with its species. Having myself often
experienced this difficulty, I think it will be doing
a service, if I can, by giving representations of a
considerable number,, lighten the labours of others,
who may wish to undertake their investigation. I t
is rather unfortunate that I delayed entering on
their examination until this late date, as I have left
myself neither the time nor room required to do
them full justice, and what is worse, I have been
constrained to take many of my drawings from dried
plants in place of fresh ones. This I regret, but
such is now my position that it is unavoidable, unless
I leave them undone.^ I have, however, endeavoured
to compensate for this defect, by greater care, especially
as regards the analysis. In spite, however, of
all my care, the relative sizes of parts, as shown in the
magnified flowers, will sometimes be found defective
as in several instances they were necessarily taken
from young flowers artificially opened, and before
the petaloid series had attained their full development,
but the forms in these cases were as accurately
preserved as it was possible, so that I trust no very
striking discrepancy between the drawing and fresh
flower will in any case be found, and as regards the
outline of the plant I believe it is generally unexceptionable.
My materials for illustrating the order
are so considerable that I could easily have nearly
doubled the number of subjects represented. I may
here mention, for the encouragement of parties who
may have an opportunity of collecting specimens, that
I have learned in the course of their investigation,
that much more can be done with dried specimens
than I previously supposed possible, and would therefore
urge their collection, as I feel quite convinced
that the order is much richer in species than the
latest publications would lead one to suppose. Roxburgh
in his Flora Indica only describes 13, a very
small number, and only to be accounted for by the
insufficiency of the characters, as known a t the time
he wrote, for their discrimination.
At that time all the Indian species, indeed nearly
the whole order, were grouped under two genera ; one,
Commelym, having half the stamens sterile, the other,
Tradscantia, having them all fertile and the filaments
bearded. Brown struck off from the former, his
genus Aneilema, and subsequently Don his Cyanotis
from the latter. These separations, especially the
first, gave greater precision to the generic characters,
and have been followed since then by the addition of
several well-defined genera.
Aneilema has already become so over-grown (Kunth
enumerates 60 species) that it now requires sub-division.
This I have attempted in my genus Dictyo-
spermum, on the principle that, as in the true Anei-
lemas, the caJycine series of stamens are fertile and
the petaline sterile, so a departure from that arrangement,
indicates such a change of structure as to justify
generic separation where it occurs. In Dictyo-
sperrnum the anterior petaline stamen is polleniferous
and fertile, and the other two usually suppressed
along with th e . posterior calycine one. This is the
arrangement observed in Commelyna, which has 6
stamens divided into 2 sets, 3 anterior fertile, 3 posterior
sterile, not, as in Aneilema, alternately fertile
and sterile.
This arrangement of the stamens enables us to
divide the genera struck off from the old genus Commelyna
into two well defined groups, viz., anterior or
petaline stamen, fertile, Commelynce, all the petaline
stamens sterile, Aneilemee. Stamens all fertile and
anthers conformable, Tradescantece.
Following out that grouping, we have for the first,
Commelyna, Heterocarpus, Aclesia, T. inantia, Dictyo-
spermum and Dichorisandra ? ; for the second Aneilema,
and Dichspermum, and for the third, Callesia,
Pollia, Lamprocarpus, Dithyrocarpus, Tradescantia,
Spironema, Cyonotis, and Cartonema. I have separated
Dichspermum from Aneilema, on the ground of its
having two rows of seed in each cell, all the other
species having one only. This I believe forms a
good generic distinction. Heterocarpus is in like
manner separated from Commelyna on account of
difference of its fruit. In Commelyna the capsule
is 3-celled, in Heterocarpus it is reduced to one, the
other two aborting and shrivelling into a podocarp,
to which the fertile indéhiscent cell adheres. Of the
propriety of constructing a genus on such grounds
I feel less confident than on either of the preceding
instances, but still I think it a good genus, the more
so, as it does not rest on a solitary species, and is
moreover strengthened by the circumstance of the
two anterior sepals being connate.
I may here remark that Kunth in his Enumeratio,
describes the fertile stamens of Commelyna and others
of that group as posterior, while I describe them
as anterior. I do not know how he views the flower,
but I look at it from behind, and finding the odd
sepal next the axis call it posterior and as a matter
ot course, the odd petal, being on the opposite side
of the flower, must be anterior. In regard to the
lobes of the perianth, I may remark that, theoretically,