23S Mi*. J. Miers on the Bi^nioniacese. 



from DeCandolle's account of the fruit of P. edulis, a description 

 framed entirely upon the drawing and descriptions of Mocino 

 and Hernandez. It is to be regretted that in the excellent 

 drawing of Parmentiera cerifera (Bot. Herald, pi, 32), no figure 

 of the structure of the ovary is given ; but it will be there seen 

 how remarkably that plant agrees in the peculiar shape of the 

 spathaceous calyx and the form of the corolla with Spathodea, 

 and scarcely less so in the shape of its cylindrical siliquose fruit, 

 which, according to that drawing, is evidently 2-valvular, with 

 numerous small apterous seeds, not enveloped in pulp, but fixed 

 to a greatly enlarged central dissepiment that nearly fills the 

 whole space within the valves, precisely as in the genus last 

 mentioned and in Stereospermum. Dr. Seemann mentions that 

 the fruits of P. cerifera are given as food to cattle, when mixed 

 with Guinea-grass and a kind of sweet potato, but does not say 

 which part of the fruit is eaten : this probably is the pericarpial 

 or valvular covering, which he defines as a " fructtis carnosus," 

 similar to that of P. edulis, described as being baccate and fleshy 

 like a cucumber, which it resembles in form : this agrees with the 

 fruit of Spathodea campanulata, which again offers much analogy 

 in its internal structure with that of Parmentiera cerifera, whose 

 fruit is said by Dr. Seemann to be " epulposa," — its seeds, like 

 small lentils, being figured as seated around the greatly enlarged 

 dissepiment, within the small annular space left between it and 

 the pericarpial covering. If, therefore, Parmentiera be found to 

 have a bilocular ovary with numerous ovules upon the dissepi- 

 ment, the genus ought at once to be consigned to the Bigno- 

 niacecB ; indeed its characters appear wholly at variance with the 

 CrescentiacecB. As its species form upright trees, it probably 

 belongs to the tribe Catalpece, and will find its place near Spa- 

 thodea (where DeCandolle was originally disposed to fix it), there 

 being a very close approximation in the form and structure of 

 the fruit in Parmentiera, Spathodea, and Stereospermum. Dr. 

 Seemann considers that the growth of the flowers upon its trunk 

 indicates its affinity with Crescentia; but we find the same mode 

 of floral development in Colea, and I have occasionally witnessed 

 the same in some species of Tecoma, where racemes grow out of 

 the old leafless axils of the stems. 



If these exclusions be adopted, the Crescentiacea would be 

 reduced to three genera, having for their characters an indehis- 

 cent fruit and apterous seeds imbedded in pulp, this last being 

 the chief distinguishing feature. These genera are Crescentia, 

 Kigelia, and Schlegelia; for Dr. Seemann affirms that Tripinnaria 

 belongs to Kigelia. The structure of the ovary in these genera 

 appears in no way different from that of the Eccremocarpea ; that 

 is to say, it is unilocular, with two opposite longitudinal parietal 



