Page:Illustrations of Indian Botany, Vol. 1.djvu/438

216

ILLUSTRATIONS OF INDIAN BOTANY. country, and also in Ceylon, they are very abundant, and many of them most magnificent and showy plants.

The species are either trees, shrubs, or herbaceous plants, with opposite, exstipulate, entire leaves ; usually without pellucid dots and marked by three or more thick longitudinal nerves or ribs. The flowers are usually bi-sexual, regular, often panicled, rarely solitary, the panicles or cymes usually contracted. The most remarkable peculiarity of this order is the position of the stamens in aestivation. The filaments are inserted near the orifice of the calyx, and the anthers are bent down into its tube, occupying the vacant space between it and the ovary, after the expansion of the flower they ascend. A somewhat similar arrangement is observable in Memecyleae with this difference, that the ovary is there altogether inferior and the anthers fill the cup of the calyx. The relative position of the ovary in the two orders generally affords a good discriminating mark between them, out is not always to be depended on as some Melastomaceae resemble Memecylon in this respect.

" Calyx with 3-5 teeth or divisions, which are more or less deep, or are sometimes united and separated from the tube like a lid. Petals equal to a segment of the calyx, perigynous, twisted in aestivation. Stamens either equal in number to the petals and alternate with them, or usually twice as many, the alternate ones of a different shape and perhaps never with fertile pollen : filaments in aestivation, bent downwards towards the bottom of the calyx : anthers long, 2-celled, bursting usually by one or two terminal pores, rarely longitudinally. Ovarium with several cells, rarely completely combined with the tube of the calyx, very rarely entirely free from it, usually cohering with it more or less by means of 3-10 longitudinal nerves, thus forming as many cases as the anthers which they contain during aestivation : ovules indefinite : style 1 : stigma simple, entire, capitate or reduced to a mere point. Placentae in the axis. Fruit pluri- locular : either free and then capsular, valvate and loculicide ; or adherent, baecate (a balausta), and indehiscent. Seeds numerous, minute. Albumen none. Embryo straight or curved : radicle pointing to the hilum: cotyledons equal or unequal. — Leaves opposite, undivided, not dotted, 3-9-nerved."

My acquaintance with this very extensive order being slight, and my means of extending it very limited, I refrain from attempting to offer any opinion of my own on this head, but that this article may not be, by so much, deficient I shall introduce the whole of the valuable remarks of Dr. Lindley on their affinities for the benefit of those of my readers who may not have an opportunity of consulting the original.

" The family of Melastomaceae"" remarks DeCandolle, in an excellent memoir upon the subject, " although composed entirely of exotic plants, and established at a period when but few species were known, is so well characterized, that no one has ever thought of putting any part of it in any other group, or even introducing into it genera that do not rightly belong to it." These distinct characters are, the opposite leaves, with several great veins or ribs running from the base to the apex, something as in Monocotyledonos plants, and the long beaked anthers; to which peculiarities combined there is nothing to he compared in other families. Permanent, however, as these characters undoubtedly are, yet the cause of no uncertainty having been yet found in fixing the limits of the order, is rather to be attributed to the small number of species that have been examined, than to the want of connecting links : thus Diplogenea has traces of the dots of Mijrtaceae, which were not known to exist in Mdastomaceae until that genus was described; several genera are now described with a superior ovary, a structure which was at one time supposed not to exist in the order ; and, finally, in the remarkable genus Sonerila, the leaves are sometimes not ribbed.

The greatest affinity of Melastomaceae is on the one hand with Lijthraceae, on the other with Mijrtaceae and their allies ; from the former they differ in the aestivation of their calyx not being valvate, from the latter in having the petals twisted before expansion and no dots on the leaves, and from both, and all others to which they can be compared, in their long anthers bent down parallel to the filaments in the flower, and lying in niches between the calyx and ovary; with the exception of Memecylaceae, in which, however, the union between the calyx and ovary is complete, and which have leaves destitute of the lateral ribs that so strongly point out Melastomaceae. The structure of the seeds of Memecylaceae is also different.