therefore hope our error should it prove one will be speedily corrected, and at the same time the other to which I have alluded. That the circumstances which induced us to constitute this genus the type of a distinct order may be fairly stated, I shall reprint here, from Jameson's Philosophical Journal, our introductory observations, and also Lindley's remarks explanatory of his views.
"The simple generic character of Millingtonia given by Roxburgh, in his Flora Ind. vol. i. p. 102., although sufficiently exact for the Linnean classification, in which those parts only are accounted stamens that have pollen, conveys little information as to the real structure of parts. The nectarial bodies opposite the petals, are of a very singular shape. The apex (which Roxburgh erroneously represents free) is incurved, and attached in front, similar to the petals of some umbelliferous plants, leaving two large hollows, one on each side, as if for the reception of the cells of an anther. Indeed, their whole appearance is that of abortive stamens, in which light we feel disposed to view them. The bifid scales, at the back of the fertile stamens, are of a very different texture, and these, we believe, are abortive petals. Thus, we have both stamens and petals heteromorphous; the imperfect forms of the one set of organs opposite to the perfect ones of the other. The calyx we have always found to consist of two interior sepals, and three exterior, one of which, and sometimes, but rarely all, are similar in size to the interior, and alternating with them : there are in some species in addition, small close-pressed bracteolae. The mode in which the calyx is placed is well figured by De Candolle (Organ. Veg. t. 37. f. 12. p.) We have, then, a calyx, a corolla, and androecium, each of five parts, placed apparently in a double scries; the one dissimilar to the other, and alternate with it; thus analogically shewing, that the hypogynous disk must be viewed as an outer series of the gymnoecium, the bidentate angles alternating with the two cells of the ovary. At first, also, it would appear that the two outer parts of each organ alternate with the inner of the next, but this is only in appearance; for, if that were the case, the angles of the hypogynous scale would be opposite to the three larger petals, whereas they alternate with them. The real disposition of parts, therefore, will be better understood, if we suppose each organ to be of only one series, and of five parts; the petals alternating with the calyx, the stamens opposite to the petals, and the pistilla alternating with both stamens and petals. That this is the true explanation, is confirmed by the fact, that, in no known plant, where any organ consists of a double series of parts, do the component parts of one series differ in number from those of the other. The asstivation will thus be imbricate and quincuncial; and in such, two or three (as may happen) parts of the same organ are interior. It is. however, remarkable to find them of so very different a structure as occurs in this genus.
The Affinities of Millingtonia have not, so far as we know, been pointed out. The habit is much that of Semecarpus, Mangifera, and Buchanania, and, like the Terebinthaceae, the embryo is campulitropal. The genus Sabia, also, has the stamens opposite the petals, the ovarium bilocular, two ovules in each cell, the one placed above the other; but the petals are likewise opposite to the sepals, and the habit is different: moreover, it is by no means certain that Sabia ought to be referred to the Terebinthaceae; and the characters of all the other genera of the order present little in common with Millingtonia. Our friend Dr. Hooker has suggested an affinity with Sapindaceae; and with different genera of that order, it has several points in common,—as the fleshy disk, the two superposed ovules in each cell, the indehiscent fruit, with part of it abortive; the absence of albumen, and the curved embryo; but that order has usually stamens twice as numerous as the petals, and, in addition, scales or tufts of hair at the base of the petals; so that if, as in Millingtonia, these scales were to be viewed as abortive stamens, the whole number of stamens would much exceed that of the petals. In Sapindaceae, too, the hypogynous disk is fleshy, and is, we believe, the torus : here it is quite free from the receptacle, except at the point of attachment, and appears to be formed by the union of an outer series of styles. Although, therefore, we cannot agree to place it among the true Sapindaceae, we can see but little objection to its forming the type of a new order next them."
The following remarks on the Affinities of this order are extracted from Dr. Lindley's Natural System of Botany. "The plants belonging to this assemblage are looked upon by Wight and Arnott as forming a family distinct from, but closely related to, Sapindaceae. The principal differences pointed out. by those authors are. that in the latter the stamens are usually twice as numerous as the petals, which have scales or tufts of hair at their base; and the hypogynous disk is fleshy. Other points are, indeed, adverted to, but they are either unimportant, or not clearly explained. These authors do not take the same view of the structure of the genus as Roxburgh.
