CONNAUACEJE. 



carpels of unequal development, one or more of which may abort when 

 the flower has attained a variable age. 1 Each 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. 5 En 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, 3 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), 



i • i • !• •/ i i.i i -, ,,' Connarus africanus 



which is stipitate, with a more or less elongated dry 

 coriaceous pericarp, 5 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, 7 Africa, 8 and Asia, 9 

 and, in a few rare cases, Oceania. 10 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, 

 are axillary to the leaves, or terminate the branches. 



1 On this character alone was founded the 

 genus Omphalobium, whose flowers have often, 

 though not constantly, only a sir.gle 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. 



3 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 ansymmetrica] 

 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 Pi., in Linncea, xxiii. 429. — Guisfh.. /•"'. 

 Brit. W. Lid., 228.— KAEST., Fl. Columb., t. 

 137. — II. I5n., in Adansonia, ix. 151, n. 25. 



8 Schum. & Thonn., Beskr., 299.— Lamk., 

 Diet., ii. 95.— Guill. & Pekr., Fl. - tey., 

 Tent., 156. — H. Bn., in Adansonia, vii. 285. 

 Bakek, in Oliv. Fl. Trop. Afric, t. 156. 



9 W., Spec, iii. 692.— G.BBTN., F ... i. 

 27._Cav., Dissert., vii. 375 Pi., he. eii., 

 425.— Thw, Emm. PI. Zeyl., 80. 



»° 15l., Mus. Bot. Lugd.-Bat., 266.— 1 

 Fl. Tnd.-Bat., i. p. 2, 662; Suppl., i. 629. 

 A. <1k\y, in Unit. State* /' , '- E 

 375, t. 45.— Waxp., t»ri.,ii.300j iv. I 51. 





