ILLUSTRATIONS OF INDIAN BOTANY.
sive arrangement of the whole order, showing its several suborders and tribes as developed in detail in his Memoirs and Prodromus. This distribution, has hitherto been adopted by all suc- ceeding Botanists, but now I believe, in course of being somewhat modified by Mr. Bentham, (Labiatarum genera et species auclor) who lately published a Memoir entitled « Commentationes de Leguminosarum generebus" embracing however only a small portion of this vast family, but which, he informs me, is but the precursor of a complete revision of the whole order, similar to his work on the Labiatse, a work, pronounced by universal consent a model of perfection in Botanical literature. Leaving for the present his views, not yet sufficiently developed to be touched upon here, I shall proceed to explain, as briefly as possible, the arrangement of DeCandolle which we have adopted in our Prodromus, after a few brief remarks on the affinities of the order, as a whole.
Affinities. Papilionaceous flowers and leguminous fruit will always mark any plant hav- ing them as a member of this order, many cases however occur where one or other of these marks are wanting, and some where both are absent, rendering it most difficult to distinguish by written characters this order from the section Amygdaleae, of Romceae. The Mimoseae have all regular flowers but at the same time bear leguminous fruit. The Caesalpineae have flowers which nearly approach the regular rosaceous forms ; but they also have leguminous fruit. Hence no difficulty is there experienced but the Detariums are apetalous with a four lobed calyx and have drupaceous fruit. " Simple therefore as the diagnosis of the order usually is, Brown is perfectly correct in asserting that until he indicated the difference of the position of the odd lobe of the calyx in Leguminosae and Rosaceae, (§ Amygdaleae) no positive character had been discovered to distinguish the one from the other." And even this character in many cases can only be made out theoretically, since nothing is more common than the union of the two upper sepals leaving a4-lobed calyx or one without any odd sepal, but then, the upper one in these cases is much larger than the opposed sepal. The drupaceous fruit in Detarium a leguminous genus, is thus explained by Dr. Lindley. " Detarium instead of a legume bears a fruit not distinguishable from a drupe. This last circumstance is easily to be understood if we bear in mind that a legume and a drupe differ more in name than in reality, the latter being formed on precisely the same plan as the former but with this modification, that its perecarp is thickened, more or less fleshy on the outside and stony on the inside, 1 -seeded and indehiscent." The following remarks which I extract from Lindley's Natural System of Botany, merit careful consideration.
" In many respects this order is one of the most important which the Botanist can study, but especially as it serves to show how little real importance ought to be attached to dehiscence of fruit in determining the limits of natural orders. What may be called the normal fruit of Leguminosae is a legume, that is to say, a dry simple ovary, with a suture running along both its margins, so that at maturity it separates through the middle of each suture into two valves ; but every conceivable degree of deviation from this type occurs : the Arackis and many more are indehiscent ; Detarium is drupaceous ; in Carmichaelia the valves separate from the suture, which remains entire, like the replum of Cruciferae ; in all lomentaceous genera, such as Orni- thopus, the valves are indehiscent in the line of the suture, but separate transversely ; in Entada a combination of the peculiarities of Carmichaelia and Lomentaceae occurs ; and, finally, in Haematoxylon the valves adhere by the suture and split along the axis. The divi- sions which have been proposed in this extensive order are of unequal value ; it is possible that two of them, namely, Mimoseae and Caesalpinieae may deserve, as Brown seems to think, the rank of suborders ; for they really appear to be of the same importance with reference to Papilionaceae, as Amyridaceae, Connaraceae, Anacardiaceae, and Burseraceae, with respect to each other."
DeCandolle divides the order into two principal sections, those namely in which the ra- dicle is curved back on the commisure or margin of the cotyledons, Curvembryce (Plates 78-79-81-82-84) and those in which the radicle is straight, Rectembryce (Plates 80 and 82).
Cukvembryje are again divided into two suborders Papilionacece and Swartziece.
The Papilionacece are again divided into two sections, one distinguished by having thin almost foliaceous cotyledons, Phyllolobece : the other by having them thick and fleshy Sarco- lobece. To the former of these subsections belong the tribes Sophorece, Lotece and Hedysarece, to the latter Vicieae, Phasioleae, and Dalbergieae. The suborder Sivartzieae, being altogether of American origin are not noticed here. The Rectembryce are in like manner divided into two
