ILLUSTRATIONS OF INDIAN BOTANY.
185
purpose, of showing in several species the origin of what I have called the spurious partition. The formation of a true partition or septum is shown in the lower fig. of No. 3, and in No. 4. In these examples the inflexed margins of the carpels actually meet in the centre and coalesce, forming a single placenta in the axis. In all the others the placentae are parietal with a process stretching from the one to the other, which, as shown in the upper figure of No. 2, enlarges with the fruit and becomes the septum. But the placentae do not enlarge in the same way, they continue parietal throughout, hence the seed in the nearly mature capsule are always found attached to the sides of the capsule along the edge of the partition and covering it with their broad wings. Such being the position of parts it results, that it is an error to describe the position of the placentae as axile, and to assume that "the central or axile position of the placenta is an indispensable character of this natural order." It equally follows that Eccremocarpus is only exceptional to the extent of wanting the spurious partition, shown to exist in nearly all the others, not in regard to its parietal placentation.
EXPLANATION OF PLATE 160-61.
c, , , , „ „ ,, , -nr, . 9. An ovary a little further advanced, divided in the
bpathodea adenophylla (Wall. Alph. DC). line of future dehiscence, and opened to show more
1. 2. Flowering branch and ripe capsule. clearly the parietal placentation.
3. Flower-bud. 10. A portion of 2 placentas detached from the capsule.
4. Corolla split open. 11. A detached seed.
5. Stamens, back and front views. 12. Same, testa removed, cotyledons in situ.
6. Calyx and ovary. 13. A cotyledon and plumule seen from within.
7. Ovary cut transversely. 14. Section of a mature capsule after dehiscence had
8. Longitudinally. commenced.
EXPLANATION OF PLATE 161-b.
1. Bignonia xylocarpa (Roxb.). 6. Millingtonia hortenses (Linn, pi.)
2. Calosanthes Indica (Blume, DC). 7. Tecoma capensis (Lind.).
3. Spathodea, species undermentioned. 8. Tecoma Jasminoides (Lind.).
4. Sterospermum Checonoides (DC). 9. Pajanelia Rheedii (R. W.).
5. Bignonia amaena (Wall, an Spathodasa ? DC).
Sub-order Schrebereae:. R. W.
Under Jasminece I alluded to the genus Schrebera, referred by Alph. De Candolle and Fenzl to that order on the faith of Roxburgh's incorrect figure and imperfect description, but whose affinity with it I doubted. It had hitherto been placed in Bignoniacece. While writing the article Bignoniaceae, I reexamined the plant with more attention and had a drawing, and as careful an analysis as my rather imperfect materials enabled me to make, prepared. This examination has tended to confirm my former conjectures as to its being distinct from Jasminece, and Bignonal in its characters, but distinguished from all the genuine members of that group by its regular corolla and diandrous flowers. In these respects it approaches some Acanthacece as well as in its 2-valved, 2-celled capsules, the dehiscence of which is in accordance with those of that family, that is, loculicidal with contrary partitions. It however differs in its arborious habit, compound leaves, winged, pendulous seed, and plicate or sub-chrysoloid, fleshy cotyledons. Towards Pedaliacece it stands in nearly the same relationship as towards Acanthacece, but does not enter that order ; neither will it enter Bignoniacece, though approaching it in so many points as to have led to its being hitherto placed in that order, and, to my mind, more naturally than it now is in Jasminece, to which it has been removed on account of its diandrous flowers, and supposed erect seed. The flowers correspond but not the position of the seed ; and the cap- sule, both as to structure and dehiscence, is so remote from that fruit of Jasmineae as altogether to over-balance the single character of diandrous flowers in a didynamous order, in which the abortion of at least one stamen is so constant. Taking this view of its affinities, I have finally determined to leave it in the Bignonal group, raising it for the present to the rank of a sub-order only, until it has been better examined, or perhaps, now that its structure is better known, relations are found to raise it and them to the first rank.
