CONNARACEZL. 5 
alternate imparipinnate,’ and the flowers are axillary to the leaves, 
as in Connarus. . 
A distinct genus has been made of Byrsocarpus, in which the 
calyx, instead of being closely applied to the base of the fruit, 
diverges more or less, or even becomes spreading at maturity. But 
this character is often ill-marked,* and is, moreover, of so very little 
value that it will only allow us to consider Byrsocarpus as a section 
of the genus Rowrea, of which it has altogether the floral and vege- 
tive organs.‘ This little group contains seven or eight African 
species, some from the west coast,’ and others from the east coast 
and Madagascar.’ 
So we have been unable to exclude from the genus Rourea 
the Brazilian species Bernardinia fluminensis,’ in which the calyx 
falls off before the fruit is ripe.* Thus we admit three sections’ 
in the genus Rowrea, often difficult of clear discrimination by these 
characters drawn from the calyx. 
II. CNESTIS SERIES. 
Cnestis” (figs. 9-11) has hermaphrodite or polygamous flowers. 
In the former the receptacle is the same as in Connarus. The calyx 
consists of five free sepals, valvate in the bud, while the alternating 
petals, of the same number as the sepals but usually shorter," have 
1 Sometimes reduced to three leaflets, or even 
to a single one; these variations may be met 
with on one and the same plant, as indicated by 
the specific name of R. heterophylla. 
2 Scoum. & Tuonn., Beskr., 226.—B. H., 
Gen., 431, n. 1.—H. By., in Adansonia, vii. 
229. 
3 «Tn the series of species from Madagascar 
we find every intermediate stage in this respect 
between the Bengal species of Byrsocarpus, with 
spreading sepals, and those mimosoid Rowreas 
from Tropical Africa, where the calyx is more or 
less markedly constricted.” (See H. Bn., loc. 
cit., 229.) 
4 And again, we have observed, “If Byrso- 
carpus were considered as a section of the genus 
Rourea, it would be very difficult to separate 
this section from Hurowrea, which would con- 
tain Rowrea proper.” 
5 Pp, in Linnea, 412.—Hoox., Wiger., 
290.—Baxer, loc. cit., 452.—WaLp., Ann., ii. 
294. 
6 H. By., loc. cit., 230-234. 
7 Px, in Linnea, xxiii, 412.—B. H., Gen., 
431, n. 2.—WatLp., Anm., ii. 295. 
8 See Adansonia, vii. 232. It is not usual 
to separate those species of Connarus in which 
the calyx thus comes off from the base of the 
fruit, from the rest of the genus. 
9 I. Eurouwrea, 2. Byrsocarpus, 3. Bernar- 
dinia. 
0 J., Gen. 874.—Lamx., Dict., iti. 23; 
Suppl., ii. 828; Z2/., t. 387.—R. Br., Congo, 
423; Misc. Works, ed. Buyy., i. 113.—DC., 
Prodr., ii. 86.—K., in Ann. Se. Nat. sér. 1, 
ii. 859.—ENDL., Gen., n. 5950.—B. H., Gen, 
433, n. 8.—H. By., in Adansonia, vii. 2.40., 
1 ‘Their breadth is often nearly equal to their 
length, and the apex is rounded or emarginate, 
but in some species they are more elongated 
like ribbons. In C. corniculata Laux. (Dict., 
iii. 23, n. 3;—Agelea pruriens SOLAND., herb. ; 
—Spondioides pruriens SMratTuM., herb.), the 
petals may exceed the scpals in length by a 
variable extent. So too in C. polyphylla Lax. 
(Dict., loc, cit., n. 2). 
