Similar in habit to A. ferox. Lvs pinnate (2-(3) pairs,) up to c. 8 dm. long; sheaths thick and coriac., c. 5 cm. wide at base, narrowing to 3 cm.; stipules simple, ± (10)-15-(25) cm. long, pungent. Petioles thick, coriac., ribbed; margins cartilaginous, serrulate; internodes short. Primary pinnae up to c. 4 dm. × 2-(3) cm., very coriac., finely striate, midrib obscure, us. strongly curved, lanceolate, margins serrulate, apex. pungent. Stems ± 1·5 m. tall, stout, strongly grooved, flowering through most of their length; infl. broad. Bracts ∞, the lower empty, or with ill-developed umbels. Bract-sheath ± 55 × 35 mm.; stipules up to 75 × 7 mm., distinctly ribbed; lamina simple, lanceolate, up to c. 185 × 15 mm., narrowing from widest part ± ⅔ way. Umbels ± = bract-sheaths; rays ∞, up to 3 cm. long. Umbellules of male plants on long slender rays, of females on shorter rays. Involucral bracts lanceolate. Fr. c. 4 mm. long; mericarps with 2 or 3 broad wings.

DIST.: S. Higher montane to subalpine, mostly along and west of divide from c. lat 41º southwards.

Type locality: Alec's Knob, Franz Josef Glacier. Type: W, W. R. B Oliver.

POLYMORPHY AND HYBRIDISM

The spp. are for the most part still too imperfectly known for a satisfactory treatment, and the key provided will need considerable amendment as knowledge increases; there are probably a number of spp. still undescribed. Stipular characters are useful, but an individual plant may show considerable differences in the length and branching. Accessory lflts are often developed on the pinnae. Dwarfed specimens at high altitudes are often difficult to place, especially in the "monroi" complex. The arrangement of the umbels and umbellules is often very irregular, and in general fl. and fr. details have been inadequately studied. owing to the lack of well-collected material.

Cockayne and Allan (Ann. Bot., Lond. 48, 1934, 33) suggested 6 hybrid groups, one with doubt, including A. colensoi (sens. strict.) × intermedia. Of this group they remark; "Recent collections on Tararua Mountains by Mr. N. Elder appear to put this group beyond doubt." Specimens supporting the view that hybrids occur are few, and the question needs restudy. Plants of a number of spp. have been grown in gardens without difficulty. Oliver (loc. cit. 14) mentions that A. P. Druce, who has studied the Ruahine and Tararua florulas intensively in the field and his garden, suggests that A. intermedia is based on hybrid plants or origin A. colensoi × dissecta.