remaining adnate to the seeds leave no loose plate between them, or separating from the seeds and forming two parallel plates between them. Involucres large, with numerous broad bracts.
As far as known the carpological differences between the two sections of Dryandra appear to be constant, but there are several species of both in which the seed has not been examined, and the characters they furnish are very little available for practical purposes. The involucres, however, give to the species here included in Aphragmia a different aspect from all others of the genus. The structure of the seeds is perhaps not so different in the two as would at first appear. In both the nucleus has a double integument, whilst the wing is apparently formed of a prolongation of the outer integument, only with a different venation in the inner and outer layer (the prolongation of the inner and outer faces of the seed) which occasions the ready separation of the two layers when ripe. In Eudryandra, as in Banksia, this outer integument, wing-like, detaches itself from the inner face of the seed, becomes or remains connate with the corresponding integument of the other seed, to the extent of the nucleus, the wing-like prolongations forming the two wings or lobes to the plate thus interposed between the ripe seeds, the wing-like prolongation of the outer integument on the outer face forming the simple wing to the seed. In Aphragmia the outer integument either remains adherent to the nucleus on both faces, the wing-like prolongations forming a double wing of which the external layer is deciduous and has been called an appendicular membrane, although the homologue of the wing in Eudryandra, or on the inner faces of the two seeds the respective outer integuments separate from the nucleus bearing with them their respective wing-like prolongations and forming two plates between the seeds. The species in which the latter peculiarity has been observed, D. bipinnatifida, has been separated on that account into a distinct section, Diplophragma, but in the few seeds that I have been able to examine, the separation of the integument from the nucleus when not consolidated with the corresponding integument of the other seed has not appeared to me to be at all constant. The whole question requires further investigation on the part of those who may have a sufficient supply of good fruits of the several species.
41. D. tenuifolia, R. Br. in Trans. Linn. Soc. x. 215, Prod. 398. A robust shrub, sometimes low and procumbent, sometimes erect bushy and attaining 3 or 4 ft., the branches nearly glabrous, with few narrow scales at the base of each year's shoot. Leaves very narrow, often 6 to 8 in. long, with closely revolute margins, tomentose underneath, rarely all entire, frequently toothed towards the end or in the upper half only, or in the typical forms regularly divided for more than half the length or quite to the base into short recurved lobes or teeth. Flower-heads large, lateral without any or with very few small linear floral leaves. Involucres at first ovoid, at length very broad, black and glabrous or when young slightly woolly, 1½ to 2 in. long; outer bracts broad, sometimes with short subulate points, inner ones broadly linear, obtuse. Perianths not exceeding the involucre, villous above the glabrous face, pubescent or glabrous towards the end, the limb very narrow, 3 lines long. Style not exceeding the perianth, with a slightly furrowed but not thickened stigmatic end. Capsule above ½ in. broad. Seeds (in the fruit examined perhaps not quite ripe) entirely separating without leaving any intermediate plate, the wing very thin though formed of two separable layers.—Meissn. in Pl. Preiss. i. 597, and in DC. Prod. xiv. 478; Bot. Mag. t. 3513; D. uncata, A. Cunn. Herb.
W. Australia. King George's Sound or to the eastward, R. Brown, Baxter, Drummond, 3rd coll. n. 294; Beaufort river, Preiss, n. 505; Tone river, Oldfield.—In Drummond's n. 294 the involucres are some of them as large as in D. proteoides.
