Urania. LEGUMINOSJE. 223 
ments triangular ; lower deeply 3-cleft, segments lanceolate-acuminated : le- 
gume 2-4-jointed, much longer than the calyx.— Wight! cat. n. 986, 791, b. 2 
—Hedysarum styracifolium, Linn. ! sp. p. 1052 (not of most authors).—H. 
retroflexum, Linn. ! mant. p. 103; Spr. syst. 3. p. 317.—H. procumbens, 
Roxb.? fl. Ind. 3. p. 645.—H. pilosum, Rowb.? in E. I. C. mus. tab. 305.— 
Nicolsonia styracifolia, Desv. in ann. sc. nat. 9. p. 418.—Desmodium retro- 
flexum, DC. prod. 2. p. 336.—D. orbiculatum, Wall.! L. n. 5695.—D. rotun- 
difolium, Wall.! L. n. 5696. 
. Wight, cat. n. 791. b, probably from growing in a poor soil, has a quite 
different appearance from the usual forms: it is merely suffruticose, with pro- 
strate, diffuse, slender branches; the branches are only pubescent, and the 
leaves, although pubescent, are not covered with long white hairs on the un- 
der side ; the calyx is however white with hairs, and the raceme is the same 
as In the type of the species: this agrees so well with Roxburgh’s figure of his 
H. pilosum (H. procumbens of the Flora Indica), that, although the pedicels 
are represented erect, we can hardly doubt of their identity.— The Linnean 
specimens are both indifferent: on the one the leaves are all simple, on the 
other all trifoliolate ; but they agree in every other respect : and Dr Wallich's 
and our own Specimens show both kinds of leaves on the same individual. 
The reflexed pedicels are very remarkable ; they become so, soon after the en- 
veloping bracteas have fallen off. With regard to the genus, we have had 
Some difficulty : the legume is by no means straight as in Desmodium, nor yet 
18 it folded up when mature, as is usual in Uraria : when about half ripe, it 
may, however, be observed considerably folded ; and it is probably this cir- 
cumstance that has already induced Dr Graham, in Wallich’s List, to refer it- 
doubtfully as we have done. It agrees also with Uraria in the rigid nature of 
Its pedicels, and the hairy calyx ; but it differs from all the other hitherto de- 
scribed species by the pedicels being deflexed, and not horizontal with the 
extremity incurved: but in the U. retrofracta, Wall. L. n. 5678 (if a specimen 
m Dr Wight’s herbarium be correctly so named), the pedicels are rigid, be- 
tween erect and spreading, with the extremity recurved : in that species, one 
of the most curious of the genus, the branches, petioles, and leaves, are gla- 
brous 3 the leaves reverse-reniform, retuse at the base ; the bracteas subulate, 
ngid, and with the pedicels sprinkled with spreading hairs; and the flowers 
are very minute. 
XXXI. DESMODIUM. Desv.; DC. 
_ Calyx with two usually caducous bracteoles at its base, cleft to the middle 
into two lips ; upper lip 2-cleft ; lower 3-partite. Corolla papilionaceous, 
inserted into the base of the calyx: vexillum roundish: keel obtuse, but not 
truncated. Stamens diadelphous (9 and 1), or monadelphous from the base 
to the middle and usually diadelphous upwards: ‘filaments somewhat persis- 
tent. Ovary with several ovules. Style filiform. Stigma capitate. Le- 
ume compressed, composed of several joints : joints 1-seeded, separating at 
maturity.—Herbaceous or suffrutescent plants or small trees. Leaves either 
Pinnately trifoliolate, or simple by the abortion of the lateral leaflets. Par- 
tial stipules 2 at the base of the terminal leaflet, solitary at the base of the 
R ones. Flowers usually racemose, sometimes umbelled, sometimes on 
simple peduncles, axillary or more usually terminal. Flowers purplish, blue, 
or white, 
Sub-gen, I. DENDROLOBIUM (W. & A.) Calyx with two lanceolate subu- 
late caducous bracteoles, campanulate, 4-cleft; upper segment split at the 
apex. Stamens monadelphous: filaments of equal length: anthers oblong, 
versatile. Joints of the legume coriaceous, indehiscent, somewhat elliptical, 
