Flora of New Zealand Series

Type locality: Campbell Id. Type: BD 49746, J. F. Findlay, Feb. 1945, Lyall Saddle. Not certainly noted on Auckland Is.

Rosettes up to c. 15 cm. diam., runners well-developed. Lamina thinner, duller, not at all polished, ± 5 cm. × 5 mm.; margins not or very slightly recurved; apex obtuse to subacute, not or very obscurely cucullate. Scapes up to c. 20 cm. long, ± flattened. Remaining parts very much as in the us. form, including achenes.

DIST.:

HYBRIDISM

C. pseudo-lyallii (Cheesem.) Ckn. in T.N.Z.I. 42, 1910, 314.  C. lyallii var. pseudo-lyallii Cheesem. Man. N.Z. Fl. 1906, 312.

Cheeseman's description is: "Leaves not so rigid, grooved and plaited above, beneath clothed with thick soft tomentum, which usually conceals both veins and midrib. Heads more cottony. Achenes nearly glabrous." No localities are given. Cockayne cited specimens from Mount Peel, Canterbury, remarking: "I have always considered this a valid species. Mr. T. Kirk referred my specimens to C. spectabilis, Hook. f. The leaf is altogether different from that of C. Lyallii . . . C. pseudo-lyallii has a much less stiff leaf which is tapering and acute, but not drawn out to a fine pungent point. There are no regular grooves and ridges, but the upper surface is plaited and dark--not yellowish-green; the midrib is keeled on the under-surface, which latter is thickly tomentose with dense silky tomentum."

Similar forms occur throughout the mountains of Canterbury in company with C. lyallii and C. spectabilis, and are almost certainly due to hybridism, as Cockayne later recognised (New Phytol. 22, 1923, 127).

C. compacta Cheesem. Man. N.Z. Fl. 1925, 955.

Stems prostrate, woody, much-branched; branchlets closely compacted; lvs ∞ in close rosettes at tips of branchlets. Lamina 30-50 × 5 mm., spreading, narrowly subulate-lanceolate, tapering to subacute apex and to very short petiole; margins strongly recurved, often almost to prominent midrib; surfaces densely clad in yellowish to buff tomentum. Sheath widening from petiole, thinly membr., glab. or nearly so, strongly grooved; margins with long whitish hairs. Scape up to c. 7 cm. long, stout, clad in dense buff tomentum; bracts linear. Capitula ± 25-40 mm. diam.; phyll. few, linear-spathulate, acute, tomentose. Achenes silky.

DIST.: S. Mount Arthur, Mount Ollivier Range, Arthur Pass. Apparently confined to fellfield and rocky places.

Cheeseman observes: "Allied to C. Hectori, and closely approaching it in size and habit, but differing in the laxer yellowish-buff tomentum, linear-subulate leaves with tightly appressed sheaths, strongly revolute leaf-margins, and prominent midrib." The specimens seen by me have all the appearance of belonging either to gracilenta × sessiliflora or insignis × sessiliflora.

C. poppelwellii Petrie in T.N.Z.I. 47, 1915, 50.

Low-growing, forming patches; stems rather stout, much-branched; branchlets spreading or ascending, ending in dense lf-rosettes. Lamina ± 2 cm. × 5 mm., linear to subspathulate, sub-coriac.; upper surface ± plaited, covered by rather loose pellicle of silvery-grey hairs; lower by loose grey tomentum, midrib prominent; margins entire, recurved, tapering to subacute apex and to short petiole, then abruptly expanded into a broad glab. or pellicled strongly nerved sheath. Scape ± 10-15 cm. long, slender, flexuous, ± floccose; bracts ∞, linear, ± floccose. Capitula c. 2-2.5 cm. diam.; phyll. linear-subulate. Ray-florets ∞. Achenes glab.

DIST.: S. Higher montane to subalpine grassland and fellfield: Eyre Mountains, Mount Cleughearn, Bold Peak, Lake Harris Saddle.

Petrie compares his sp. with C. haastii, but Cheeseman (Man. N.Z. Fl. 1925, 941) considered that "the branching stems and much smaller and narrower leaves do not show any close affinity with that plant." Cockayne and Allan (Ann. Bot., Lond. 49, 1934, 47) considered that the plants concerned were probably of origin C. haastii × hectori.

C. linearis J. B. Armst. in T.N.Z.I. 13, 1881, 337.

Low-growing herb with stout us. much-branched stems, often forming rather large patches; lvs in dense rosulate tufts at tips of branchlets. Lamina (3)-5-9-(12) cm. × 3-4-6·5 mm., coriac., narrow-linear; both surfaces covered by persistent pellicle or in some forms with dense soft silky tomentum, midrib ± keeled below, apex acute to subacute; margins entire, strongly recurved, sts to midrib; sheath ± 3-4 cm. × 4 mm., indumentum as in lamina but us. slighter. Scape 10-25 cm. long, rather stout, clad in floccose white or greyish hairs; bracts ∞, lamina very narrow-linear up to 25 mm. long, lower with conspicuous broad sheathing base. Capitula 25-30-(40) mm. diam.; phyll. c. 12 mm. long, linear-subulate, margined by long hairs, somewhat indurated in lower half. Ray-florets ∞, c. 15 mm. long, limb narrow-oblong; disk-florets funnelform, c. 7 mm. long, teeth minute. Achenes strongly grooved, c. 3 mm. long, clad in dense ascending hairs on ribs; pappus-hairs slender, up to c. 6 mm. long, very finely barbellate.

DIST.: S., St. Boggy ground. Armstrong's specimens came from "Canterbury Provincial District", but appear to be lost. In herbaria, under this name, is included a somewhat diverse assemblage of forms from the Mount Cook district, Mount Maungatua, Mackinnon Pass, Haast Pass, Routeburn, Garvie Mountains, as well as Stewart Id.

Cheeseman (Man. N.Z. Fl. 1925, 955) remarks: "Most of the Stewart Island specimens in my herbarium have longer and narrower leaves, with much more revolute margins, the tomentum is thinner and more appressed, and the scape is longer. Possibly more species than one are included under the description." The group would repay closer study; there is fairly good field evidence that most forms are of origin C. gracilenta × sessiliflora. C. hectori may be involved also.

C. mollis Ckn. in T.N.Z.I. 31, 1899, 423.  C. spectabilis var. mollis (Ckn.) Cheesem. Man. N.Z. Fl. 1925, 946.

Tufted herb with densely imbricate lvs around stock. Lamina lanceolate, rather thin, c. 10-25 × 2-3 cm., unequally longitudinally furrowed; upper surface with scattered white hairs; lower surface with ± floccose pale yellow matted tomentum, shorter towards apex; margins entire, slightly recurved, clad in brownish tomentum; nerves evident, midrib purple, hairy. Petiole ± = lamina in length; sheath broad, purplish, ± hairy without. Scape up to ± 25 cm. long, purplish, densely clad in long brown hairs; bracts linear, purplish, ± pilose. Capitula 2-3 cm. diam.; phyll. ∞, with shaggy yellowish hairs without, glab. within. Ray-florets up to c. 15 mm. long, apex 2-3-toothed; disk-florets with ± hairy tube.

The original specimens came from "Hill's Peak, Canterbury, at altitude of 1,200 m.; L.C. (1898). Flowers in early January." Cockayne observes: "It is very probably a hybrid between C. petiolata and C. spectabilis." There is a specimen, from the original gathering, at K labelled No. 1924. There is hardly a doubt that the hybrid origin assigned by Cockayne is correct.

× C. christensenii Ckn. in T.N.Z.I. 47, 1914, 115.

Tufted herb with densely imbricate lf-sheaths surrounding stock. Lamina oblong to sublanceolate-oblong, ± 10 × 3 cm.; upper surface green, ± sulcate, glab. or nearly so; lower clad in pale brownish soft rather lax tomentum, midnerve keeled, purple (concealed by hairs); margins entire with fringe of sordid-white to brownish hairs, tapering to short broad petiole, then widening to purplish silky sheath c. 4·5 × 2 cm. Scape long, stout, clad in silky white appressed hairs; bracts linear-spathulate, tomentum as in laminae. Capitula c. 4 cm. diam.; phyll. linear, with white to brownish hairs. Ray-florets c. 25 mm. long, 4-toothed.

DIST.: S. Occurs on both Mount Charon and Mount Isobel. "The plant here dealt with is almost certainly a hybrid between C. spectabilis Hook. f. and C. Traversii Hook. f. It strongly resembles C. mollis Cockayne . . . But C. mollis is undoubtedly a hybrid between C. spectabilis Hook. f. and C. petiolata Hook. f. . . . × C. christensenii is at once distinguished from C. spectabilis by the softer slightly rust-coloured tomentum, the purple midrib and petiole, and matted rather dirty-white hairs somewhat rusty coloured at their bases, which form a fringe round the margin of the leaf."

Specimens I have seen in the field match Cockayne's description very well, and appear to be sterile.

× C. morrisonii Ckn. in T.N.Z.I. 47, 1914, 116.

Tufted herb with lvs densely imbricate around stock. Lamina oblong, ± 12 × 3·5 cm., subcoriac., upper surface pale green, hidden by thin pellicle; lower clad in appressed slightly yellow-flushed white tomentum, ± sulcate; margins distantly mucronately toothed, clad in white hairs; midnerve prominent, keeled, purple, ± hidden by tomentum; petiole up to 2 × 1·5 cm., tomentose, expanding into white-hairy sheath c. 4 × 2-5 cm. Scape rigid, purple below tomentum of silky white hairs; bracts ∞, ± linear, up to 7·5 mm. long. Capitula c. 6 cm. diam.; phyll. linear, scarious at apex. Ray-florets c. 3 cm. long.

DIST.: S. First collected on Mount Miromiro near Hanmer, but occurs on other mountains in the area.

The forms at present included under this name, and almost certainly of the same origin, coriacea × traversii, as seen by me in the field, fall into 3 groups: (a) with laminae broad-lanceolate, of nearly coriacea shape, with white pellicle above, midrib with white tomentum. Scapes floccose; (b) with laminae more of traversii shape, tomentum darker, margins lanuginose, midrib pilose below; (c) laminae slightly broader than those of coriacea, with soft white appressed not at all buff tomentum below, midrib white. What the breeding status of the different forms may be is not known.

C. lanigera Petrie in T.N.Z.I. 48, 1916, 187.  C. coriacea var. lanigera (Petrie) Cheesem. Man. N.Z. Fl. 1925, 950.

Tufted herb; lf-sheaths densely imbricate around stock. Lamina ± 15-20 × 4 cm., narrow-lanceolate, subcoriac., upper surface woolly, lightly plicate; lower with white close tomentum, midrib not conspicuous; margins entire, ± recurved, tapering to acute apex and oblique obscurely cordate base. Scape up to ± 30 cm. long, rather slender, floccose; lowermost bracts up to ± 8 cm. long, linear, floccose. Capitula ± 4 cm. diam.; phyll. narrow-linear, thin, densely hairy on back. Ray-florets ∞, narrow. Achenes linear, glab. or nearly so, strongly ribbed.

DIST.: S. Garvie Mountains, apparently common.

Petrie remarks: "This appears to be a well-marked species, of which I have seen only three or four rather indifferent specimens. In general look it recalls C. coriacea Hk. f., but it is readily distinguished from it by the smaller narrower woolly or villous leaves, the slender scapes, and the glabrous achenes." Cockayne and Allan (Ann. Bot., Lond. 48, 1934, 47) refer to the name as belonging to forms of C. petiolata × spectabilis. Petrie refers specimens from the Takitimu Mountains to his lanigera, but they may possibly be of origin C. coriacea × petiolata.

C. boweana Petrie in T.N.Z.I. 44, 1912, 182.

Tufted herb with lvs densely imbricating around stock. Lamina ± 15-25 × 1-1·5 cm., strict, narrowly lanceolate; upper surface glab. or nearly so, coarsely sulcate, somewhat yellowish green; lower surface clad, except distinct glab. midrib, in laxly appressed somewhat yellowish hairs; margins entire, ± recurved, gradually narrowed to acute apex and to striate membr. sheath ± 6 cm. long. Scape rather slender, up to ± 30 cm. long, flexuous, clad in lax yellowish-white hairs; bracts linear us. ∞, tomentose except on midrib. Capitula ± 2.5 cm. diam.; phyll. ∞, tomentose. Achenes glab. or nearly so.

DIST.: S. Montane to lower subalpine grassland and herbfield, Sealey Range and elsewhere in the Mount Cook district.

Field evidence leaves little doubt that specimens placed under this name are of origin C. coriacea × spectabilis and are probably sterile.

HYBRID GROUPS. Apart from those otherwise discussed the following groups have been recorded or suggested, some with good to fair field evidence, occ. supported by herbarium specimens, others more speculative (see Cockayne and Allan in Ann. Bot., Lond. 48, 1934, 45-47): allanii × incana, × discolor; angustifolia × bonplandii, × brevifolia, × spectabilis, × walkeri; argentea × gracilenta, × sessiliflora; armstrongii × coriacea, × gracilenta, × lyallii, × petriei, × spectabilis; bonplandii × brevifolia, × du-rietzii; brevifolia × discolor, × ramulosa, × walkeri; coriacea × gracilenta, × lyallii, × petiolata, × rigida, × verbascifolia; dallii × traversii; densiflora × prorepens; discolor × du-rietzii, × incana, × walkeri; dubia × parva, × spectabilis; du-rietzii × incana, × prorepens; gracilenta × monroi, × sessiliflora, × spectabilis, × verbascifolia; haastii × hectori, × sessiliflora; hectori × sessiliflora; lyallii × petiolata, × petriei, × viscosa; monroi × spectabilis; petiolata × sessiliflora, × verbascifolia.

The hybrid groups are distributed among the sections as follows:

A. Within the LIGNOSAE 23. Within the IMBRICATAE 6. Within ROSULATAE 9 (7 within Discoloratae). Between ROSULATAE and IMBRICATAE 8 (all with Discoloratae as one member).

B. Within the HERBACEAE 28. With spectabilis as one parent 8. With a member of PETIOLATAE as one parent 9. With a member of PELLICULATAE as one parent 2. Within STOLONIFERAE 1. With a member of ANGUSTATAE as one or both parents 8.

C. Between LIGNOSAE and HERBACEAE 8.

The spp. producing the most interspecific hybrid groups (all with ample opportunity) are: spectabilis with 10 other spp., coriacea with 9, gracilenia with 8, petiolata with 7, sessiliflora with 6, and brevifolia, discolor, du-rietzii, armstrongii, lyallii each with 5.

SPECIMENS WITH BRANCHED SCAPES

Kirk (Stud. Fl. 1899, 288) under C. coriacea remarks: "Under cultivation this plant varies to a remarkable extent. Two forms may be specially mentioned:―(a) foliacea. Scape with crowded foliaceous bracts often 6 in. long. (b) corymbifera. Scape much branched above. Heads forming a loose corymb." Of his var. ensata he says " Scapes very slender, irregularly branched " and doubts its occurrence in the wild state. Cockayne (T.N.Z.I. 49, 1917, 58) collected a number of branched scapes from a wild plant of C. gracilenta " on the Port Hills, Banks Peninsula ". I have seen branched forms in several localities, and also of C. coriacea. The phenomenon is not at all restricted to cultivated plants.

OBSERVATIONS

Du Rietz (Svensk bot. Tidskr, 24, 1930, 356) says of C. petiolata Hook. f. : " Of this species, which plays a prominent rôle in a good part of the mountain-chain, I had the opportunity of examining at least three different subspecies in the field (one in the upper Waimakariri region at about lat. 43°, one in the Mount Cook region at about lat. 43° 40' and one in the Routeburn-Humboldt Mts region at the western end of Lake Wakatipu at about lat. 44°40'). Each of these subspecies was composed by a large number of forms, the general composition of the population appearing to change gradually from the northernmost subspecies of the Waimakariri region to the southernmost one of the Northern Fiord District, which may or may not be connected with Celmisia verbascifolia of the Southern Fiord District by a gradual transition."

In a note Du Rietz adds: (a) Waimakariri population―"I saw only forms with distinctly petiolate, rednerved and more or less serrulate leaves with rounded base"; (b) Wakatipu population―" only forms without distinct petioles, with leaves gradually attenuated at the base, not serrulate, or very indistinctly, and with the nerve either green or red; (c) Mount Cook population―" distinctly kept in the middle between these two extremes, consisting partly of unpetiolate partly of more or less petiolate forms, with the nerve red or green and more or less indistinct serrulation."

On the whole my own observations support this grouping (substituting "mainly" for "only" and "purple" for "red"), but much closer study remains to be made. Du Rietz (loc. cit. 357) after discussing C. verbascifolia remarks: "Further field-studies must decide whether C. verbascifolia and the southern subspecies of C. petiolata are connected only by a narrow hybrid-zone (to which then probably some of the forms in my Wakatipu-population should be referred) or by a broad zone of more continuous intergrading. In the latter case it may prove necessary to include C. petiolata and C. verbascifolia into one heterofacial species." My own, quite incomplete, observations suggest that neither hypothesis is correct, the hybrids occurring here and there where the two species meet.

Du Rietz (loc. cit. 411) remarks that " in the genus Celmisia brown and yellow hairs showed a remarkable frequency in the northern part of South Island, being found in species widely differing in other respects (C. Traversii, C. cordatifolia, C. Macmahoni, C. Dallii, and C. hieracifolia), but of much rarer occurrence in other parts of New Zealand." Species with distinctly ferruginous indumentum (C. traversii, C. cordatifolia, C. praestans, C. rigida) range throughout South Island and reach Stewart Island.

While hybrids have been met with in areas where the vegetation shows little sign of disturbance there is evidence that man's interference has facilitated the increasing occurrence of hybrids. A number of spp. withstand fire and increase in numbers of individuals and range. Intermediate habitats are formed in which spp. previously isolated come together and hybridize. For instance, there are several localities where Nothofagus forest in the South Id has been burnt, and the new vegetation is of a herbfield or tussock-grassland character. C. spectabilis. C. lyallii, C. coriacea notably form hybrid groups under these conditions, e.g. C. pseudo-lyallii, C. lanigera. An example was given early on by Cockayne (New Phytol. 22, 1923, 117) "I have seen hundreds of × C. Morrisonii as a member of an indigenous-induced herb-field on Mount Miromiro (Northeastern Botanical District), at about 3000 ft. altitude, where the Nothofagus forest had been burnt."

INCERTAE SEDIS

1. C. campbellensis Chapman in T.N.Z.I. 23, 1891, 407.  C. chapmanii Kirk in Gdnrs' Chron. 9, 1891, 731, f. 146. Chapman's description is: "A low-growing glabrous species. Not tufted. Leaves rosulate, 3 in.-6 in. long, 1/2 in.-1 in. wide, lanceolate, obtuse or acute, serrate, coriaceous. Above, glabrous; below, thinly clad with loose hairy tomentum. Remarkably deeply ribbed. Sheathing. Scapes numerous, 12 in. high. Slightly tomentose. Bracts numerous, 1 in.-2 in. long, large, sheathing, glabrous or slightly tomentose, serrate. Heads very similar in size and colour and the form of involucral scales to C. vernicosa. Corolla-tube pilose; pappus 1/6 in. long; achene hispid. Hab. Campbell Island." About 12 plants were found "in the space of an acre".

The specimen in K is labelled: "This is Mr. Chapman's type plant, and all that he has of it. He sends it to Kew at my request. D.P." There is one If and 5 scapes in poor order. The following description is drawn up from this material and further pieces in Kirk's herbarium at W.

Lamina ± 9 × 2 cm., elliptic-lanceolate, ± coriac., deeply longitudinally furrowed; upper surface glab. or nearly so, lower clad in floccose tomentum; gradually narrowed to obtuse apex and to petiole ± 15 × 10 mm.; margins rather distantly sharply serrate; sheath ± 20 × 15 mm. Scape c. 20-30 cm. long, ± floccose; bracts ∞, lowermost lflike, upper linear, glab. or nearly so. Capitula 4-5 cm. diam.; phyll. linear, up to 15 mm. long, gland-dotted, glab. or nearly so. Ray-florets with tube c. 8 mm. long, limb c. 10 mm. long. Disk-florets purple, tube pilose, abruptly flaring into ovate, acute teeth ± 2 mm. long. Achenes c. 5 mm. long, compressed, densely clad in long slender ascending hairs; pappus-hairs 6-7 mm. long, distinctly barbellate.

Sorensen (D.S.I.R. Cape Exped. Ser. Bull. 7, 1951, 34), concerning C. campbellensis, says: "Searched for thoroughly in the place from which it was originally obtained, but not located. One solitary plant found on Mt. Lyall in a rock crevice by W. Brockie early in 1947 and is undoubtedly the so-called C. campbellensis. Considered [as suggested by W. R. B. Oliver] to be a hybrid between Pleurophyllum speciosum and Celmisia vernicosa. . . . Mt. Lyall and originally found at Venus Bay."

2. C. novae-zelandiae (Buchan.) Cheesem., also considered of uncertain status, is discussed under C. angustifolia Ckn.

3. C. polyvena Simpson et Thomson in T.R.S.N.Z. 72, 1942, 34. The description of the authors is: "A small laxly-tufted plant forming patches to ·5 m. diam. Stems stout, 1-2 cm. diam. with the leaves on. Leaves spreading, 6-9 cm. long, 3-4 mm. broad, linear, narrowed to an obtuse tip, wrinkled above, thickened at the base, subcoriaceous, densely cottony tomentose on both surfaces; with the midrib branching in veins to the margins, prominent; sheaths 3-4 cm. long, 8-10 mm. broad, thin, woolly tomentose. Scapes 10-15 cm. long, slender, with many linear, lanceolate, silvery tomentose bracts. Heads not seen. Habitat:. Peat slopes with Oreobolus and other bog vegetation. . . . In recording this plant as C. linearis, Cockayne (1909, p. 63) mentioned the conspicuously veined under-surface of the leaves and the 'closely leaved flower bud'; and he recorded another closely allied but smaller plant on Mount Rakiahua. C. polyvena is probably an endemic of Stewart Island, and its characteristic veining separates it from any other species we are acquainted with. We have not seen the smaller plant recorded by Cockayne."

The type, from Table Hill, has not been located in BD. Cockayne (Rep. Bot. Surv. St. Id 1909, 63) remarks: "There are two distinct forms [of C. linearis] in Stewart Island, one much smaller than the other, and only noted on Mount Rakiahua. The large form differs from that of the Southern Alps in its open-leaved, not sharp-pointed, and close-leaved flower-bud, rather blunt leaves which have much-recurved margins, conspicuous veins on under-surface, upper surface longitudinally wrinkled, tomentum not quite silvery, but with a brownish tinge, and leaf-sheath only half length of blade. The above differences lead to leaves of very different appearance."

4. Fenwick (New Zealand Native Flora. 1.(Celmisias. Dunedin, p. 27) writes: "C. Nobilis is a species that requires special notice no less because of the striking features of its handsome leaves than for the reason that doubts have been cast on its rights to the name (given to it by the late Sir J. D. Hooker, F.R.S., to whom specimens had been sent by Mr. H. H. Travers, who found it growing on one of the Hanmer mountains in 1916, and also as to the colour of its flower." The colour was said to be brownish yellow. Plants were later collected and distributed to several gardens. "The leaves reach a length of 12 to 15 inches, and are, somewhat like those of C. coriacea in appearance, but with a much softer texture and with the under surface, which is of a cream colour, quite velvety to the touch. The outer surface is silvered with a loose filament". At K, under the unpublished name of C. nobilis Hutchinson are 4 lvs and 2 flowering scapes. A letter from Travers to Hooker, dated March 31, 1911, says "I cannot help thinking it is a natural hybrid of these species" [C. coriacea × traversii]. The lvs are up to 30 × 6 cm. on petioles c. 12 cm. long; phyll. mostly ± 4 cm. Travers' suggestion seems correct.

Fenwick (loc. cit. 22) says of C. nigrescens : "This is another remarkable variety and has leaves nearly a foot long of purplish-brown colour as intense as the purple beech. The plant stands out prominently from its congeners and has been specially admired in my garden. Mr. Darton informs me that the leaves of the plants at Weatherstones get darker with age. It has not flowered either at Weatherstones or with me."