CONNARACEZL. 3 
carpels of unequal development, one or more of which may abort when 
the flower has attained a variable age.!. Hach carpel is formed of a 
one-celled ovary, tapering above into a style of variable length, which 
dilates at the tip into a stigmatiferous head. In the ventral angle 
of the ovary-cell, and somewhere near its base, is seen a placenta 
bearing two collateral ascending ovules, which are orthotropous, or 
nearly so,° so that the micropyle is quite superior. The fruit, which 
may be accompanied by the remains of the non-accrescent calyx,* 
consists of only a single fertile follicle (figs. 5 and 8), , aiezs 
which is stipitate, with a more or less elongated dry ‘ 
coriaceous pericarp,’ dehiscing over a variable extent, 
beginning at the ventral angle. It contains a single 
erect orthotropous or suborthotropous seed,° at whose 
base is a lobed fleshy umbilical aril of variable form 
and size (figs. 6 and 7). Within the seed coats is a large 
fleshy exalbuminous embryo, with a superior radicle 
and thick plano-convex cotyledons. The genus Connarus 
consists of half a hundred species of trees and shrubs 
from the tropical parts of America,’ Africa,’ and Asia,’ 
and, in a few rare cases, Oceania.” Their branches, which are some- 
times sarmentose, bear persistent alternate exstipulate leaves, impari- 
pinnate, or more rarely trifoliolate. The flowers are in racemes, simple 
or with cymose ramifications ; these racemes, usually many-flowered, 
Fria. 8. 
Fruit. 
are axillary to the leaves, or terminate the branches. 
1 On this character alone was founded the 
genus Omphalobiwm, whose flowers have often, 
though not constantly, only a single well-de- 
veloped carpel at anthesis, and have normally 
but one capsule in the ripe fruit. Some fruits 
of Connarus Patrisii are however exceptional, 
and consist of two carpels (fig. 1). 
2 In this genus, as in several others, the form 
of this dilatation is very variable—sometimes 
regular and subcircular, sometimes flattened and 
turned outwards, here entire, there more or less 
deeply two-lobed. 
> The hilum is not constantly basilar, and 
diametrically opposed to the micropyle; but is 
often some way up the side of the ovule, looking 
towards the ventral angle of the ovary. The 
first step towards the incomplete anatropy of 
the ovule, which we shall find in several genera ; 
and this shows how little real value should be 
attached to this character of orthotropy which, 
as we shall see, is not absolute, in all the 
genera of this order, and of several others. 
4 When the calyx persists, as is usually the 
case, its leaves are pretty closely applied to th® 
stalk of the fruit it surrounds. 
5 lways slightly oblique and unsymmetrical 
when we get its exact profile, looking at it so 
that the midrib of the pericarp is on the one 
side, and the ventral angle on the other. 
6 The hilum varies in situation just like the 
ovule. 
7 Px, in Linnea, xxiii. 429.—GRIsEB., Fl. 
Brit. W. Ind., 228.—Karst., Fl. Columb., t. 
137.—H. Bw., in Adansonia, ix. 151, u. 25. 
8 Scuum. & THonn., Beskr., 299.—Lamx., 
Dict. ii. 95.—Guiru. & Perr, Fl. Seneg., 
Tent., 156.—H. By., in Adansonia, vii. 235.— 
Baxer, in Outv. Fl. Trop. Afric., i. 456. 
9 W., Spee. iii. 692—Garrn., Fruct., i. 
27,.—Cav., Dissert., vii. 375.—Pu., loc. cit., 
425.—Tuw., Enum. Pl. Zeyl., 80. 
10 Br, Mus. Bot. Lugd.-Bat., 266.—Miq., 
Fl. Ind.-Bat., i. p. 2, 662; Suppl., i. 529.— 
A. Gray, in Unit, States Expl. Expd. Bot., 
875, t. 45.—Watp., Ann., ii. 300; iv. 451. 
B 2 
