have been considered a convenient receptacle for all manner of refractory plants, having opposite entire polished leaves, and square branches, however much they may depart from the characters of the order in the form and structure of the organs of re-production.
According to the arrangement I have proposed the Asiatic portion of the order, all that I undertake to investigate, might be thus distributed.
Guttiferae. Floral envelopes ranged in a binary order, (two and its multiples).
Suborder 1st.— Garcinieae — Ovary 4-6-8 or more celled : cells, usually, with one. rarely several, ovules attached to the inner angle of the cell next the axis of the fruit. Style short or wanting. Stigma spreading, lobed ; lobes corresponding in number with the cells of the ovary. Flowers axillary, solitary, or fascicled, peduncles 1-flowered, very rarely more. Garcinia, Gynotroches.
Suborder 2d. — Calophylleae — Ovary 1-2 celled : ovules solitary, or several, attached to the base, ascending. Style elongated. Stigma radiato-peltate. Peduncles axillary, 1 -flowered, or racemose, or forming terminal panicles.
This suborder may again be divided into two sections, or might perhaps be advantageously removed altogether to form a distinct order, on account of the difference observed in the ovary and fruit, but for the present I, in accordance with all former practice, Professor Martius excepted, allow it to remain as a section of the order.
1st. — Mesueae — Ovary 2-celled : ovules several in each cell. Peduncles axillary, 1-flowered — Mesua.
2d. — Calophylleae — Ovary 1 -celled: ovules solitary or several, erect. Flowers racemose, or panicled — Calopkyllum, Apoterium, Kayea.
The genus Xanthochymus I exclude from the order on account of the quinary arrangement of its flowers, and for the present refer it to Hypericineae, as being the order most nearly akin, in which that structure prevails. It may be objected to this proposal, that many of the Hypericineaehave quaternary flowers, but then, the ternary or quinary fascicles of stamens, and the 3 or 5-celled ovaries show, that that is not their normal structure, but the effect of abortion of parts. Taking number therefore as the basis of our classification, we can no longer experience the difficulty which has hitherto been felt in distinguishing the species referable to one or other of these orders, and however closely allied in all other respects, this character alone, preserves a clear and well marked line of demarcation between them.
If the precedent established by Dr. Graham in the formation of his genus Hebradendron be followed, we may, I fear, soon expect to see the off-sets from Garcinia about as numerous as its species now are, since that genus is separated on account of a variation in a single point of structure, and without reference to analogous forms met with in other species. The only point in which it differs from Garcinia, as defined in our Prodromus is — in having 1 -celled circumcissile anthers — while the more usual form in that genus is to have them two celled, with introse longitudinal dehiscence. Should this be considered a satisfactory reason for its removal, then G. Kydiana Roxb. which has a four-sided connectivum with a polliniferous cell in each face, must equally be separated from the genus, as well as another species of which I possess specimens from Mergui, the anthers of which are 1-celled, dehising transversely across the apex. Another variation of structure which has been long observed in a few species of the genus will equally demand separation, as being of at least equal generic value, I allude to those in which the stamens of the male flower are united into four thick fleshy androphores, with a highly developed sterile pistil in the centre. Here then, assuming that we are justified in assigning generic value to such variations of structure, limited as they are to the male organization, are four distinct genera, and all, so far as such artificial characters can make them, equally stable. I confess that I have an objection to this kind of excessive sub-division, in as much as, whatever rule holds good with respect to genera must equally apply to orders, and must inevitably lead to the elevation of half our present species to the rank of genera, and an equal proportion of genera to natural orders ; both of which might be avoided by a slight extension of our characters, and still better by a careful and comprehensive investigation of groups of allied species and genera, before attempting their disunion in the formation of new genera and orders. In support of these views I think I may safely cite the recorded opinion of
