NAMED KAFfLESIA. 385
analogous to the concentric series of processes in the apex of the {olunni of Ilafflcsia.
In all these genera of Amrince and in Braf/anlla of Loureiro, which is also referable to the same order, the flowei's are liermaphrodite ; but in Cijtinn^, which, if not absolutely belonging to this order, is at least very nearly related to it, they are diclinous.
The affinity is also in some degree contirmed by the appearance of the inner surface of the tube of the perianthiuni of some Asarince, especially ArisfolocJda grandijiora, and by the thickening or annular projection of the faux in the :~i9 same plant, as ^vell as in a new species of Brcif/antia dis- covered in Java by Dr. Horsfield.
It may also be noticed in support of it, that some of the largest flowers which were known before the discovery of Bajlesia belong to Asarince, as those of Aristolochia (jrandi- Jlora, and particularly AristoIocJtia cordiflora of j\Iutis, which, according to M. Bonpland. are sixteen inches in diameter, or nearly half that of our plant.-^
The first objection that occurs to this approximation is the ternary division of the perianthium in the regular flowered genera of Asarince, opposed to its quinary division in Baffletsia ; but in Cytinus it is divided into four segments, a number more G;enerallv connected in natural families with five than with three.'"
A second objection would exist, if it be considered more ])robable that the ovariiuii of Bafflesia is superior, or free, than inferior, or coheiing with the tube of the perian- thium.
There is indeed nothing in the structure of the column itself indicating the })articular position of the ovarium. But if it be admitted, that a base of a form equally calcu- lated for support should exist in the female flower, as is found in the male, it might perhaps be considered some- what more probable that such a base should be connected with a sujierior than with an inferior ovarium.
Even admitting this objection, however, it would be con- siderably weakened, on the one hand, by allowing that
' Ilnmboldt BoHpl. et Ktmih Xov. Gen. et Sp. ii, p. US.
J>5
�� �
