156 ` CELASTRINEJE. TunPINIA. 
only differs from the other Celastrinee by the compound leaves and bony seeds, the 
other characters proposed being common to both : we therefore do not think it ne- 
cessary to retain the section. 
I. TURPINIA. Ventn.—Dalrymplea. Roxb. 
Flowers polygamous or bisexual. Calyx 5-partite, persistent. Petals 5. 
Torus discoid, with a free 10-crenulated margin. Stamens 5, inserted under 
the margin of the disk, alternate with the petals: anthers ovate, dehiscing lon- 
gitudinally. Carpels 3, follicular, at first distinct or separable, soon combining 
into one ovary, lower part immersed in the disk: ovules 2-8 in each carpel. 
Styles 3, separable. Stigmas patent, flat, cuneate. Fruit baceate, 3-celled 
(or with fewer cells by abortion). Seeds 1-3 in each cell, bony and shining, 
truncated at the hilum, fixed along the axis or to its apex. Albumen fleshy. 
—Trees or shrubs. Leaves opposite, unequally pinnated; leaflets coriaceous, 
glabrous, stalked, ovate or oblong, acuminated, serrated. Flowers white, 
panicled: branches of the panicle alternate (in the American species), oF 
opposite (in the Indian). 
When the fruit of the American or West Indian species is better known, characters 
will probably be found to distinguish Roxburgh’s genus Dalrymplea. 
491. (1) T. nepalensis (Wall.:) leaflets 3-5, oblong-lanceolate, acuminated, 
coriaceous : branches of the panicle opposite: styles almost quite distinct : 
ovules 3 or occasionally 2 in each cell: berry (immature) scarcely fleshy, 
marked on the outside above the middle with 3 small distant points (the re- 
mains of the styles), about 3-seeded: seeds pendulous: radicle superior.— 
Wall.! L.n. 4277 ; Wight! cat. n. 396.——Mountainous districts. 
We can see no difference between ours and Dr Wallich's specimens from- 
Nepal, but the latter are only in flower: if characters be found to separate 
them, the Peninsular one may called T. microcarpa, the fruit not exceeding the 
size of a large pea: the pericarp or fleshy part is very thin, and dry, so that 
it might be termed a dry-berry. We have not observed more than three seeds 
in each fruit, one in each cell when all the cells are perfect, occasionally two 
in one cell and one in another when one of the cells becomes abortive. 
IL ELZEODENDRON. Jacq.?; Roxb. 
Calyx 5-parted. Petals 5, expanding, linear-oblong. Torus 5-angled, 
very thick, fleshy. Anthers 5, inserted into the margin of the torus: fila- 
ments at length recurved: anthers with a thick connectivum at the ‘back, 
roundish, opening longitudinally. Ovary immersed in the disk, 2-celled: 
ovules 2 in each cell. Style short, conical. Stigma simple, obtuse. Fra} 
drupaceous, dry or pulpy : nut 1-2-celled. Seeds usually solitary, rarely in 
pairs, with a membranaceous or spongy integument, erect. Albumen pm 
Cotyledons thick, fleshy: radicle inferior.—Small trees. Leaves opposite, 
entire, glabrous. Peduncles axillary, branching dichotomously. 
„We have no opportunity at present of examining the fruit of E. orientale, the ori- 
ginal species of the genus: if Gertner be correct in describing it as a 3-celled drupe, 
and the seeds with a fleshy albumen and remarkably thin membranaceous cotyle- — 
dons, the Indian species must be again removed, and the genus Schrebera of Retz 
(not Roxb.) retained for them : or as this might now cause much confusion, Rube 
Comm. and Juss. might be restored for Æ. orientale, while the Indian € 
may constitute E/eodendron: this last would be preferable to givin the name 0 — 
Neerija to z m -— rage The American and Now Holland species 
require much elucidation, but they do not appear to belong to the genus. P U^ ^ 
Ham.! Wall.! L. n. 4326), appears to us 19 be allied to Holi E and Mangifera 
j canens, Roxb. (not, however, a Mangifera), among the erebinthacee: its mare — 
only are known. 
