§ I. Connectivum of the anthers 4-sided, with a polliniferous cell in each side, four angled at the apex.
Observation.— The extension of these cells, causing the absorption of two of the partitions, would produce the form which distinguishes the. next section, the more usual form of the genus, and if all the four partitions were absorbed, the 1 -celled anthers which constitutes the distinctive peculiarity of the 3 section would result : hence 4 cells may perhaps be considered the normal structure, and the other forms transitions caused by an excessive development of pollen from the operation of some vet unknown cause.
6. G. Kydiana (Roxb.) Hort. Bengalensis.G. Kydia Roxb. Fl. Ind. not W. and A. Prod.
This is I suspect the species to which Mr. Brown alludes in his letter to Dr. Graham, where he says, " but it is right to add, that approaches to this structure, and which serve to explain its analogy with the ordinary structure of the family, exist in Garcinia.'"
I II. Anthers oblong,2-celled,dehiscing longitudinally, introrse.
A. Fruit globose, not furrowed.
7. G. pedunculata, (Roxb.) Flowers terminal, long peduncled : male — fascicled ; female — solitary, or two or three from the same branch: fruit very large: (2 pounds weight) leaves obovate, cuniate, membranacious, marked with numerous prominent parallel veins.
A stately tree about 60 feet high. The leaves of this species differ from those of all the others of the genus.
8. G. paniculata, (Roxb.) Male — flowers panicled ; panicles axillary, many flowered ; female — racemose, subsessile; racemes terminal: fruit globose, small, 4-seeded.
The fruit of this species, raised in Calcutta, is represented as about the size of a cherry, that of native specimens received from Silhet about twice as large. The former greatly resembles that of G. morella, Gærtner.
9. G. purpuria, (Roxb.) Leaves obovate, lanceolate, acuminated: male — flowers longish pedicelled, aggregated, 4-8 congested in the terminal axils, 2-4 in the lateral ones: column of stamens short, capitate, filaments free, for a short distance at the apex ; anthers few, (12-20) occasionally one or two in the centre, simulating a rudimentary ovary: fruit globose, not furrowed, 4-8 seeded, whole fruit deep purple.
Roxburgh received specimens of this plant from Malabar, under the name of Mahi Mangostan. The specimens from which this character is taken, were communicated by Dr. Wallich, from the Calcutta Botanic garden, and of course identical with Roxburgh's. Rumph. Amb. — 3 to 32, may be cited as a figure of this plant, though a different species, except that his is the female, mine the male plant, but having full grown, detached, fruit, which seems to correspond in size and form with the Amboyna one.
10. G. lanceœfolia, (Roxb.) Leaves narrow lanceolate, acuminate : Male — -flowers axillary, and terminal, solitary, short pedicelled; stamens all united, anthers capitate; in the female about 20, filaments dilated at the base, and united, forming a ring round the base of the ovary, splitting irregularly into several fascicles, stigma 6-8 lobed, fruit somewhat obovate, 6-8 celled.
G. dioica. ? Blume, Bijd. 1, page 215.
Native of Silhet.
Roxburgh has figured the female plant only, I am indebted to Dr. Wallich for the specimens which have enabled me to characterize the male one — It is closely allied to the former but quite distinct.
11. ? G. Cochin-chinensis(Choisy). Leaves ovate, oblong, acute, flowers lateral, congested, white ; short peduncled : berry, reddish yellow, pear-shaped.
Hub. — China and India.
Rumphius, Herb. Amb. 3-32, is the authority for this species — it seems referable to this section, but is too imperfectly known to be referred to, with certainty.
B. Fruit globose or oval, furrowed.
To this section a long list of names belong, but I suspect very few species : at least, if each name really belongs to a distinct species I must confess my inability to find marks among the characters assigned by which to distinguish them. The following is the list of names referable to this section.
G. Cambogia, (Desrous) G. Cambogia, (Roxb.) G. Zeylanica, (Roxb.) G. Cowa, (Roxb.) G.Affinis, (W. and A.) G. Kydia, (W. and A. not Roxb.) ? G. Indica, Choisy.
These may be thus grouped and briefly defined.
12. G. Cambogia, (Desrous, Moon's Catalogue of Ceylon plants, not Roxb.) Fruit somewhat elongated, tapering a little at the ends, furrows broad, with angular edges, and intervening flattened, or but slightly rounded ridges, fruit yellow.
G. Kydia, (W. and A.) ? G. Indica, Choisy, and D. C. Rheede, Hort. Mai. I tab. 24.
This species I have now found at Courtallum, in Malabar, and in Ceylon : different specimens vary somewhat in the appearance of their foliage, and in the number and position of their flowers, but all agree in having the ridges and furrows alike square, as if cut artificially.
This to my mind is unquestionably the plant figured by Rheede, and therefore the Garcinia Cambogia of Desrousseaux, and all subsequent authors who have followed him, but is not the Cambogia Gutta described by Linnæus, Fl. Zeyl. No. 195 : neither is it Garcinia Cambogia Roxb. if his figure and description are correct, as both represent a plant having globose fruit, with narrow sloping furrows and intermediate semicircular ridges or costee like those of a melon. For these reasons I consider Roxburgh's plant a species distinct from Rheede J s, but not distinct (so far as I can judge from his figure and definition) from his own G. cowa and G. Zeylanica, in both of which, the fruit is described as spherical and torose (swelling over the seeds] which is indeed the only mark on which it appears to me the slightest dependence can be placed. These species therefore I unite, assigning one specific name, for the whole.
13. G. Rqxburgii (R. W.) Fruit globose, 6-8, furrowed ; furrows narrow, sloping towards the bottom: intermediate costæ or ridges rounded : male — flowers aggregated or solitary, axillary, or terminal : female — flowers usually, solitary, nearly sessile, sometimes, when terminal, two or three together.
G. Cambogia, Roxb. cor. pi. 3-298— Fl. Ind. 2-621, not Desrous: G. Zeylanica, Roxb. Fl. Ind. 2-621, G. Cowa, Roxb. Fl. Ind. 2-622, W. and A. Prod. 1-101. Garcinia affinis(W. and A.)
The depth of the furrows varies, they are deeper in G. Cambogia, less so in G. Zeylanica and Cowa, but in all totally different from those of the preceding.
I have not quoted Linnæus' Cambogia Gutta for either of these, though it seems the general opinion of Botanists that it belongs to the former. This opinion however, his brief description of the plant before him in the flora Zeylanica, shows to be erroneous, and proves almost to demonstration that that it is Dr. Graham's Hebradendron. The following are his words " Rami oppositi. Folia lanceolato-ovata, integerima, petiolata,
