Page:Illustrations of Indian Botany, Vol. 1.djvu/367

Rh work, is, I believe, as nearly correct as our present imperfect knowledge of the tribe enables us to approach. The whole order certainly requires revision, and it is to be hoped some European Botanist, favourably situated for the performance of the work will undertake to supply this Botanical desideratum. In the mean time, it may be observed, that, the order as left by Kunth and DeCandolle has been completely broken by Barrling and Lindley, while it has been preserved entire by Meisner, though, with the works of both these latter authors before him. Bartling does not depart very widely from DeCandolle for he groups the whole under his class, : the definition of which however can be of little use in practice as its negative and positive characters nearly balance each other. It has the segments of the calyx imbricated (they are valvate in Burseraceae which he unites with Amyrideae) petals and stamens definite, hypogynous or sub-perigynous : ovaries 1-5 distinct or united (Ochna one of his Terebinthaceae has sometimes as many as 10.) Pericarp double, the laminae separating spontaneously or divisible : seeds albuminous or exalbuminous. Embryo straight or curved. Leaves compound or simple, often with pellucid glandular points—every positive has here its saving negative, hence it can be of little value, in doubtful cases, in practice, even though, with one or two exceptions, the orders ranged under it have many very striking affinities, and form upon the whole, in their properties a natural group—they are—Ochnacece, Semarubece, Zantlioxylacece, Diosmeae, Butaceae, Zygophylleae, 'Auraniiaceae, Amyrideae, (our Burseraceae) Connaraceae, Cassuvieae, (our Anacardiaceae) and doubtfully, Juglandeae —(the walnut family). This arrangement has the advantage of forming in some respects a very natural group, but has the disadvantage, of bringing together a number of orders, the Botanical characters of which are so far at variance as to render them in other respects badly associated. Dr. Lindley's distribution seems to me to be subject to similar objections, but on this point I would speak with much reserve, for, from having been long accustomed to view their connections in one light I may not be able to recognize them when placed in a different.

Meisner retains the order as left by Kunth and DeCandolle with the exception of a few genera referred to Zanthoxylaceae. He distributes the order under two primary divisions, namely—1st, those having a single ovary or at least containing only one perfect cell : and 2dly, those with a many-celled ovary or with several distinct perfect ovaries.

The first of these divisions he again divides into three subdivisions or tribes, Juglandeae, Amyrideae, and Cassuvieae. The last, which is the only one of the three having reference to the Botany of this portion of India, is distinguished : by having several ovaries but all except one aborting, the perfect one having but one ovule : flowers usually unisexual, rarely apetalous; stamens perigynous ; filaments free or united at the base ; indehiscent, drupaceous fruit, and pendulous, or rarely ascending, exalbuminous seed. The leaves, except in Bouea, (an Indian genus) alternate, simple or compound, exstipulate.

This tribe is divided into two subtribes Sumachineae and Anacardieae—the first is distinguished by having foliaceous or flat, not thick fleshy, cotyledons, the last by the cotyledons being thick and fleshy. To the former of these tribes Odina and Rhus of our flora belong, to the latter Sorindeja, Pegea, (a genus omitted in our Prodrornus as not being a native of the peninsula) Solenocarpus, Holegarna, Bouea, Meisner (a native of India but so far as I am aware not of the peninsula) Mangifera, Semecarpus, Buchanania, Melanorrhaea (Wall.) and Anacardium.

The second primary division is also subdivided into three tribes Spondiaceae, Burseraceae, and Connaraceae. In Spondiaceae, the cells of the ovary have one ovule—in Burseraceae they have two collateral pendulous ovules, and in Connaraceae the cells themselves are distinct with two ovules in each, but they have been already disposed of as a distinct order. Spondias is the only genus found in this part of India referable to the first of these tribes : to the second Boswellia Idea Balsamodendron, (second section of our Protium) Protium Garuga and Casarium belong. Such is a brief but I fear imperfect exposition of the order as it relates to Indian Botany, given in the most recent work to which I have access, from which it will be seen, on a comparison with our distribution, that it is nearly the same throughout, so far as our work extends.

With the slender materials in my possession I cannot venture to give a complete character of an order, embracing according to Meisner, six tribes and upwards of 60 genera, I shall therefore content myself with copying from our Prodrormus for the benefit of those who may not have access to the work itself, our characters of the three suborders under which we have