seeds remain distinct from each other but separated from the seeds forming two membranous plates between the seeds, or remaining attached to the nucleus or to the whole seed leaving the seeds separate, each with a double or wingle wing.—Shrubs, often low or flowering near the base. Leaves alternate, very rarely entire, usually either sinuate and prickly-toothed, or pinnatifid or pinnate with numerous small regular lobes or segments, usually smooth and veinless on the upper surface, white-tomentose or marked with parallel transverse veins underneath. Flowers sessile, in pairs, in dense terminal or lateral heads in an involucre of numerous imbricate scale-like bracts and usually surrounded by a ring of floral leaves similar to the stem leaves; receptacle flat or convex, densely villous or woolly, with narrow-linear villous or woolly bracts or paleæ subtending each pair of flowers, sometimes very small or deficient at least in the centre of the head. Perianth usually yellow, the short entire base glabrous or villous towards the divided part, the remainder of the tube or claws usually pubescent or villous, the limb occasionally, the whole perianth very rarely, glabrous. Ovary almost always hairy. Capsules usually villous, but the hairs very readily rubbing off, and in some species apparently glabrous from the first.
The genus is endemic in West Australia. It is readily distinguished from Banksia by the involucre, by the flat or nearly flat receptacle, and by the fruit; but the structure of the flowers is so uniform that it is very difficult to establish any definite sections. The differences in the foliage correspond but very little with those in inflorescence, and both are variable in some species. Meissner has founded his groups on the former, I have preferred the inflorescence, which appears to me more characteristic. With regard to the sections founded upon the differences in the so-called dissepiment of the capsule (the plate intervening between the seeds), I have adopted them upon the supposition that these differences are constant, but the seeds remain to be examined in a considerable number of species. If it should prove that these species, here arranged according to their apparent affinity with those whose seeds are known, have been misplaced, all practical utility in these sections will be lost, and some other principle of division must be sought for, although no good one has as yet suggested itself.
| Sect. 1. Eudryandra.—Outer integuments of the inner faces of the two seeds united in a bifid plate separating from them Involucres various, the bracts narrow or very rarely rather broad. | |
| Series 1. Armatæ.—Flower-heads usually large, mostly terminal, enclosed in floral leaves longer than the flowers. Involucres broad. Perianths above 1 in. long. Leaves with prickly teeth or lobes. | |
Involucre (2 in.) as long as the flowers. Leaves obovate-oblong, deeply prickly-toothed, not white underneath. |
1. D. quercifolia. |
Involucre about half as long as the flowers. | |
Leaves obovate or oblong-cuneate, prickly-toothed. | |
Leaves white underneath |
2. D. præmorsa. |
Leaves green on both sides |
3. D. cuneata. |
Leaves pinnatifid, with flat pungent-pointed lobes | |
Perianth-limb glabrous. Fruit 1-seeded |
4. D. falcata. |
Perianth-limb more or less hairy. Fruit 2-seeded |
5. D. armata. |
Leaves divided to the midrib or nearly so into small rigid segments with revolute margins. | |
Leaves 6 in. to 1 ft. long, the lobes lanceolate or triangular. |
6. D. longifolia. |
Leaves 2 to 4 in. long, the segments linear, distant. |
7. D. Fraseri. |
