CONNARACEZ. ? 
within whose coats is found a fleshy albumen, at whose apex is a 
pretty long embryo, with its radicle superior. Sometimes the seed 
has no aril; sometimes on the contrary this organ is represented by 
a sort of fleshy frill near the hilum, with its superior edge irregu- 
larly divided.’ Cnestis consists of bushy shrubs, often sarmentose, 
with alternate, imparipinnate, exstipulate leaves; the flowers are 
in racemes, simple or composed of cymes, axillary or terminal, or 
more rarely grouped in numbers on peculiar short woody branches. 
About a dozen species are known, natives of tropical Asia? and 
Africa,’ the Indian Archipelago, the Mascarene Islands, and Mada- 
gascar and the neighbouring islands.* 
Cnestidium® is a New World type, closely analogous to Cnestis. 
The perianth and androceum are nearly the same, but the valvate 
calyx has sometimes only three or four sepals instead of five. The 
petals are longer than the sepals, tapering at the base and imbri- 
cated in the bud. There are ten stamens, of which the five oppositi- 
petalous are the smaller; they all cohere at the base into a very 
short ring, above which the slender filaments become free and taper 
towards the reflexed apex, ending in introrse two-celled anthers, 
also finally reflexed. The carpels are sessile, the ovaries being as 
in Cnestis ; but the style is long, slender and reflexed, with an entire 
or two-lobed, dilated, stigmatiferous head. The fruit is sessile, 
velvety outside, glabrous within; the seed possesses a fleshy aril. 
Only one species of this genus is known,’ a tree from Mexico and 
the north of Colombia. It has velvety imparipinnate leaves, with 
the leaflets symmetrical at the base; the flowers are numerous, in 
multiple ramified racemes of cymes, axillary to the leaves or termi- 
nating the branches.* 
and Poildgratier to several species of Cnestis,such 
as C. glabra LaMx., from Bourbon and Mauritius ; 
it appears to be due not only to the mechanical 
action of the hair, which easily comes off and 
remains sticking in the skin, but perhaps also to a 
brownish liquid which it contains and which fills 
its cavity more or less completely in the dry 
herbarium specimens. 
1 In C. polyphyila Lamx., for instance, this 
- frill surrounds the lowest quarter of the seed, 
which tapers in this part. Thus botanists are 
wrong in characterizing Cnestis as exarillate. 
2 Roxpuren (Cat. Hort. Cale., 34) only 
describes a single species in this country ; namely 
C. monadelpha (DC., n. 5); but the genus is 
certainly represented by other species in India 
and the neighbouring countries. 
3 Bentu., Niger, 290.—Px., in Linnea, 
xxiii, 440.—H. Bw., loc. cit., 242, not. 1— 
Baker, in Outv. Fl. Trop. Afr., i. 460.—WaLpP., 
Ann., ii. 306. 
4H, By., loc. cét., 244, not. 1. 
5 Pr, in Linnea, xxiii. 488.—B. H., Gen, 
433, n. 7. 
§ And in that case they are often unequal. 
7 ©. rufescens’ Pu., loc. cit——Watp., Ann., 
ii. 305. 
8 The genus Teniochlena (Hooxr., F., Gen., 
438, n. 10) comes extremely near to Cnestidium 
and Cnestis, and we doubt whether it ought to 
be separated from the latter genus. It is dis- 
tinguished chiefly by the three following cha- 
racters. 1st. The form of its floral receptacle, 
which is nearly hemispherical, owing to the 
