164 MR. JOHN MIERS ON THE GENUS CRESCENTIA. 



finds to be unilocular, with two large placentae in the centre almost touching, but not 

 united,supported on two opposite longitudinal semidissepiments, so that in dried specimens 

 the ovary appears 2-locular, as it does also in the ripe fruit. This structure is analogous to 

 that seen in the baccate species of Cyrtandra\ but it is attended by a very anomalous 

 circumstance ; for not unfrequently the ovary presents three similar semidissepiments, 

 with as many placenta? almost meeting in the centre, but not united; and in such case, 

 the style is always trificl. Swartz, in a remark upon the same species (El. Ind. Occ. 

 p. 1054), though he figures the fruit as being completely 2-locular, says that he has often 

 seen it to be 3-locular. This structure is quite unparalleled in the JBignoniacece and all 

 its subfamilies, and points to a different affinity ; for the normal development of its com- 

 ponent carpels, according to the views suggested in my " Observations on the Mgno- 

 niacece" must be different from that in the other cases. 



In Kigelia I had fancied that the ovary is 1-locular, or rendered sub-2-locular by the 

 approximation of two opposite parietal placentas, as described by Bojer, who had frequent 

 opportunities for examining the living plant. M. Bureau has, however, shown me his 

 analyses, where it is distinctly 2-locular, with the ovules congregated in three or four se- 

 ries on the margins, as in the true Bicjnoniaceaii not in the middle of the dissepiment ; and 



1 have since clearly verified this structure in a specimen in the British Museum. The 

 genus will therefore class with Amphitecna, Colea, and several others, distinguished by 

 their fleshy subindehiscent fruit — among which JParmentiera will probably find its place. 



It appears clear, therefore, that we know no genus that can class with Crescentia, which 

 consequently will constitute a tribe near the JacarandecR and the CyrtandracecB, as they 

 also have 1-locular ovaries. The former differ in their dry, 2-valvate, dehiscent capsules 

 with winged seeds : the latter disagree in their free revolute placentae attached to semi- 

 dissepiments, the cell filled with hyaline mucilage, and numerous very minute foveatcd 

 seeds suspended by a long funicle, and the embryo enclosed in very sparse albumen. 



We may distinguish two very different forms of Crescentia : in one section the leaves 

 are always fasciculated, out of a number of scale-like woody pseudostipules which fringe 

 the axillary nodes, and which, upon the older branches, form a warty excrescence ; the 

 fruit is not umbonated at the summit ; the pericarp is hard and very firm, even when it 

 is very thin ; the contained pulp, when much shrunk in drying, always retains, even after 

 a period of fifty years, its soft subviscid consistence, and is composed of innumerable 

 fine laminar membranes agglutinated together, containing spiral nourishing-vessels, and 

 leaving numerous hollow dry cells polished inside, each provided with a seed ; the seeds 

 are \ or £ of the size of those of the following section, are very compressed, flat on the 

 opposite faces, cordate at the summit, cuneate to the base, where by a minute hilum 

 they are attached to the cell : the outer integument is hard and coriaceous, with a 

 peculiar surface ; the inner integument is very membranaceous, cordate at the summit, 

 and suddenly contracted at the base into a long thread-like tube ; the embryo is greatly 

 flattened, is white and waxy, never becomes black in drying ; the short radicle fills the 

 sinus, and is not concealed by the auricular lobes of the cotyledons. In the other section 

 (Enallagma), which may probably resolve itself into a distinct genus when the plants are 

 bettor known, the leaves are always alternate, single, petiolated, articulated in the axil, 



