1G0 MR. JOHN MIERS ON THE GENUS CRESCENTIA. 



tribe {Jacarandece) , differing from the rest of the family in having a unilocular ovary 

 and fruit, with parietal placentation : in that respect they offer some analogy with the 

 Crescentiece, from which they differ in their dry capsular fruit splitting into two valves, 

 containing winded seeds imbricately attached to a parietal median line of placentation 



■o " — o 



on each valve. In the structure of the flower and the ovary, there is little that can 

 distinguish the Jacarandece from the Crescentiece ; but subsequently a notable difference 

 arises, in the growth of the fruit, the former producing the development just described, 

 while in the Crescentiece we find a hard indehiscent 1-locular fruit, with numerous 

 apterous seeds imbedded in pulp. The Crescentiece have, therefore, scarcely higher claims 

 than the Jacarandece to be separated from the Bignoniacece, unless it could be proved 

 that the pulpy matter is formed by an arillus surrounding each seed, in which case there 

 would be more valid ground for the maintenance of Crescentiaccce as a distinct order. 



Finally (in 1864) M. Bureau published his valuable monograph of the Bignoniacece, in 

 which he suggests a new arrangement of the family, and where he constitutes the 

 Crescentiece the third tribe of the order. He follows the views of preceding botanists in 

 classing in this tribe Colea, jPhyllarthron, Farmentiera, and Kigelia, with Crescentia, 

 under the persuasion that Crescentia, like all the others, has a 2-locular ovary (/. c. p. 56); 

 but he confesses that he had not seen the ovary of any other species than C. macrophylla, 

 where it is unquestionably 2-locular ; and hence he naturally inferred that it was a mistake 

 to regard the ovary of Crescentia as being 1-locular : it will, however, presently be 

 shown that the species in question does not belong to that genus. 



The learned DeCandolle, in his ' Prodromus,' was evidently doubtful in regard to this 

 point of structure ; but his son, the present eminent Professor in Geneva, stated (in a note) 

 that the ovary is manifestly 1-locular in C. cujete, as Gardner had shown it to be in 

 C. cuneifolia. Dr. Seemann, in his monograph of the Crescentiacece, did not dispute this 

 fact, but considered it of such little importance that he classified indiscriminately with 

 Crescentia six other genera, all with 2-locular ovaries. I examined with great care the 

 ovary of Crescentia, from specimens preserved in spirits, in different stages of growth, 

 and found it to be most decidedly 1-celled, without the slightest trace of any dissepiment ; 

 and the generality of this structure was again confirmed in the fruits of several species 

 which I have since observed. In support of the same conclusion, Mr. Bentham has 

 added his valuable testimony. 



But there still remains another important desideratum, which is to ascertain the real 

 structure of the fruit of Crescentia; for this has not yet been determined with any 

 approach to accuracy. With this view, I procured from Brazil ripe fruits, which I was 

 able to examine within a month after they were gathered. This fruit, which I had often 

 seen growing in Rio de Janeiro, but to which I then paid little attention, considering it 

 to belong to the well-known type of the genus described by Linnams, pertains, however, 

 to a species hitherto undescribed. It is nearly the size of that of the last-mentioned type, 

 being regularly oval, 7J inches long, 6 inches in diameter, supported upon a thickened 

 peduncle 1J inch long, 4 lines thick ; the outer shell is rather thin in substance, of a 

 texture between nnrinpftnun ar 



een cori aceous and osseous, of the uniform thickness of 1 line, externally 

 polished, yellowish green, without any removeablc epidermis, marked by four equidistant, 



