Remarks on Genera and Species. The genera of this order are few, but require much nicety, and attention to minute organs, for their accurate discrimination; the structure of the ovary, which in most is very small, and the number and position of the ovules in its cells, affording the marks on which the greatest dependence can be placed, accordingly form the basis of our characters. The following synoptical table which I extract from our Prodromus will place this in a clear point of view. It contains an analysis of all those having stamina twice as numerous as the petals and a definite number of ovules in each cell of the ovary.
Ovules solitary, or 2 collateral in each cell. LIMO-NEÆ. Ovary cells as many as the petals :
Filaments combined 1. Atalantia.
Filaments distinct.
Stamens 6; , 2. Triphasia.
Stamens 8-10.
Style elongated (ovules 1-2 in each cell) 3. Limonia.
Style short, scarcely distinguishable : (ovules solitary.) 4. Glycosmis. Ovary cells fewer than the petals :
Style short, scarcely distinguishable; ovules 1-2 pendulous from the apex of the cells 5. Sclerostylis.
Style elongated; ovules solitary, peritropal, attached to the middle of the axis......................... 6, Bergera.
Ovules in pairs, one above the other. CLAUSENEÆ
Filaments distinct : Ovary cells fewer than the petals 7. Murraya.
Ovary cells as many as the petals;
Ovary hirsute; berry full of liquid ; flower-buds angled 8. CooKia. Ovary glabrous ; berry dryish ; flower-buds not angled; Stamens 8, dilated part of filaments concave ; petals oval 9. Clausena.
Stamens 10, filaments flat ; petals linear-lanceolate... Micromelum. Filaments (8) combined to the middle: (anthers linear : ovarium 2-3-celIed : style long: berry 2-3-celled With a resinous pulp : calyx tubular: petals 4.) Luvunga.
Of each of these genera, except the two last, detailed characters are given in the work; of them no Peninsular species have, so far as I am aware, been yet found. The plant which I have here, I now find erroneously, referred to Micromelum, differs widely in habit, though in structure it associates in many points, as may be seen by comparing the figure, with, the character given in the table of the ovary and flowers of that genus, — but I shall add to this account of the order the detailed generic character of Micromelum as given by Blume. Though I have not yet found Luvunga I will also, to facilitate its recognition should it be met with, extract Roxburgh's description of the flowers of his Limonia scandens, which is referred to this genus. Of the genus Sclerostyles, of which it was remarked when we wrote that Dr. Berry was the only person who had found it in the Peninsula, I have now specimens received from Malabar, Courtallum, Ceylon, and the Circars. Whence it would appear that so far from being so rare a genus as we then supposed, it is on the contrary but little if at all less frequent than Atalantia ; to which it bears so close a resemblance, that it is quite impossible to distinguish them with certainty otherwise than by an examination of the ovary or fruit, the 2 or 4 cells of which immediately marks the difference. (See Icones No. 71,) Of the genus Glycosmis I now possess one or two species in addition to those described in our Prodromus, but for which I find it difficult to assign characters suited to distinguish them from the preceding ones. This, of all the genera of the order, next perhaps to Citrus, is the most difficult to distinguish by written characters. To the genus Atalantia I now add one species and take away another : our Atalantia racemosa being in truth a Sclerostylis. The new species is distinguished from the old, by wanting the style and having the broad stigma adherent to the apex of the ovary; in all other respects it resembles the other. In his " Pugillus Plantarum Indiæ Orientalis" Dr. Arnott has given the character of a new genus of Aurantiaceœ under the name of Rissoa. This genus, judging from the characters only, I fear I cannot adopt, as I see no sufficient mark by which to distinguish it from Sclerostylis. Luvunga differs from both in having superposed not collateral ovules, but in other respects the characters are very nearly alike. Respecting the genus Citrus I have nothing to add, to what we have already said in the Pro-
