234 



NATURAL HISTORY OF PLANTS. 



acute, while the claws of the latter are much shorter and narrower.' 

 The petals only approach the form of those of C. drasi/icme towards 

 the time of the complete expansion of the flower, remaining until 

 then very much like those of several true UnounH. But the sexual 

 orLj^ans and fruit (fig. 290)' are exactly those of Ct/mbopetalum. We 

 already know nine species of this genus,^ small American trees 

 found between Mexico and Brazil, with subsessile mem- 

 branous leaves, often somewhat unsymmetrical at the 

 base. The flowers are solitary and terminal, leaf-opposed 

 or extra-axillary, usually on very long peduncles,* 



Beside Cymbopctalum has been placed Enantia; in 

 which there are only six leaves to the perianth : three 

 sepals and three petals superposed to them. The former 

 are lanceolate and valvate ; the latter are much longer, 

 erect or slightly spreading, thick and coriaceous, flat, or 

 with slightly reflexed edges, and with a concave con- 

 tracted base. The convex receptacle bears an indefinite 

 number of linear-oblong stamens, whose dilated con- 

 nective is but little dilated above the anther-cells- 

 The carpels are also indefinite ; each ovary contains 

 a single erect ovule, and is surmounted by a short linear oblong 

 style, traversed by an internal longitudinal groove. E. chlorantha 

 Oliv., the only species known, is a tree from the west of tropical 

 Africa, with alternate membranous leaves, and solitary extra- 

 axillary flowers on short peduncles. 



Cymbopetalum 



oHusiJluyutn. 



Fio. 290. 



Fruit. 



' The calyx also becomes very different from 

 the inner petals, especially in thickness. It is nt 

 firut a membranous globular kic, completely sur- 

 roundini^ the corolla in the bud. 



- Here tiie fruit, though of quite the same 

 ii|)])earanrc as in C braxilietise apjjears tho- 

 roughly indehiscent. We see also from lig. 21M) 

 that the lo\verun)^t segment remains empty and 

 of Ninall size, hut is se)mrated from the rest of the 

 car|)el by n well-marked, nearly tnuisverso 

 furrow. 



^ To which we have referred (Adangonia, viii. 

 2U8) I'nona pemlulijlora Din. (tig. 28l>), riri- 

 dijlora Si-lito., ubtusijlora DC, and w ith some 

 doubt, IJ.ftuicata DC. Sect. Brachyci/mhium. 



* 1 hey may be either erect, or pendulous, as 

 in C. pendulijlonnn, and are sometimes even 

 thicker than the brunch from which they spring. 



* Olivku, Joiini. Linn. Sue, ix. 171. li. H., 

 Gen., y58, n. 2S(i. — H. H.n., Adansunia, viii. 

 313. 



