92 NATURAL HISTORY OF PLANTS. 
and the foot of the gyneceum. This last is nearly or quite central. 
Its ovary, borne on a pretty long foot, ends in a slender style which 
is dilated truncate and stigmatiferous at the tip.’ The fruit is a 
large bivalve pod, with a thick straight or bowed compressed woody 
pericarp, containing large exalbuminous seeds with fleshy embryos, 
often: deformed by mutual compression. Campsiandra consists of 
three or four species’ of unarmed trees from tropical America; their 
leaves are alternate and pinnate, and their flowers form axillary or 
terminal simple or much ramified racemes. 
Phyllocarpus,’ an abnormal genus of this group, has a concave re- 
ceptacle lined by glandular tissue, with four free imbricated sepals, 
and three petals, of which one is posterior and overlapped by the 
two lateral ones. The stamens are diadelphous, as in many of the 
Papilionacee,‘ nine being united into a sheath split open above, with 
the tenth free. All possess a versatile introrse two-celled anther of 
longitudinal dehiscence. The gyneceum consists of a central stipi- 
tate pauciovulate ovary surmounted by a style which is at first con- 
volute and ends in a truncate scarcely dilated stigmatiferous apex. 
The fruit is a straight or slightly bowed compressed leaf-like inde- 
hiscent (?) oblong pod, whose ventral suture is produced into a 
narrow wing. Only one species of this genus is known,’ a tall un- 
armed tree from tropical Brazil. It has alternate pinnate leaves or 
numerous leaflets. Its flowers form short racemes, solitary or fascicled 
in the leafless nodes of last year’s leaves. 
IV. AMHERSTIA SERIES. 
The magnificent Asiatic tree named Amherstia’ (figs. 65-67) is 
considered as the type of a distinct series, because the very deep 
narrowly-tubular floral receptacle bears the gynzceum close to its 
mouth (R, fig. 67), and not in the bottom of its cavity. The rim of 
1 This style usually becomes much exserted, 
like the stamens. In certain flowers, however, 
it remains very short, included and straight, and 
is much dilated at its apex, which does not 
‘extend beyond the mouth of the floral recep- 
tacle. The ovules also then remain ill developed, 
‘so that these flowers must probably be considered 
as male through abortion of the gyneceum. In 
this case Campsiandra would be polygamous. 
2 Watp., Rep., v. 568. 
3 RIEp., ex TUL., in Ann. Se. Nat., sér, 2, xx. 
142 ; in Arch. Mus., iv. 171, t.10.—EnDL., Gen., 
n. 6720 1._B. H., Gen., 564, n. 305. 
* To which group Phyllocarpus should perhaps 
be referred. 
5 P. Riedeli Tun., loc. cit—WAtp., Rep., v. 
546. 
§ Wau. Pl. Asiat. Rar. i. 1, t 1, 2— 
Envi, Gen., n. 6793.—B. H., Gen., 578, n. 
340. 
