K. capitata (Sm.) Heynh. (fideToelken 1981b)

Creeping shrubs, shrubs, small or tall trees with or without lignotubers and rhizomes. Leaves mostly alternate, opposite in a few species. Inflorescences reduced conflorescences (botrya) usually pseudoterminal, globose to spiciform or cylindrical bearing sessile to subsessile flowers, otherwise corymbiform to elongate, with pedicellate flowers rarely reduced to solitary. Flowers 5-merous, red, pink, purple, yellow or white, free part of hypanthium usually exceeding the ovary summit. Calyx persistent in fruit. Petals free often much reduced. Stamens mostly numerous, in one or more series, exceeding petals or included; filaments finely striated, anthers versatile. Ovary mostly 2–3-locular sometimes up to 5–6-locular; placentation axillary and ovules spreading, numerous, to apical with few larger pendent ovules. Fruit a capsule, usually loculicidal, mostly dry, rarely indehiscent or fleshy, not persisting. Chromosome number: 2n = 22 based on x = 11 (Dawson 1987; de Lange and Murray 2004).

Australia: c.54 spp. (all endemic) New Zealand: 10 spp. (all endemic).

This key requires material with active new growth, buds, flowers, and ideally seedlings. In some species, such as K. amathicola, K. linearis and K. triregensis, the inflorescence condition can be easily determined in the absence of flowers from fruiting specimens, as fruits in these species are especially persistent. Use young growth only to determine branchlet indumentum, and examine the hairs produced 10–20 mm back from the branchlet tip–this is important as some species produce sparse, deciduous, antrorse-appressed hairs at the base of the actively growing branchlet apices. This key will not resolve hybrids, but these may be recognised by the hybridism notes given for each species. Geographic and ecological information is included in this key as a further aid to identification. For example, on Moutohora (Whale Island), Bay of Plenty, North Island, New Zealand, K. salterae can very occasionally (only one specimen with this condition seen) have glabrescent to almost fully glabrous branchlets, and so would key out to K. ericoides. In these very rare instances, in the absence of flowers and fruits, such specimens could only reliably be identified by their location.

(here designated)(Fig. 2). ‘Leptospermum ericoides nob. N^lle Zélande’ Herbarium Richard, Ex. Herbier E. Drake, P! Specimen labelled ‘TYPE’ in bold red lettering and bearing two handwritten labels by W. Harris dated 26 July 1989.

Figure 2. 

Lectotype of Leptospermum ericoides A.Rich. (held at l’Herbier du Laboratoire de Phanérogamie du Muséum National d’Histoire Naturelle (P)).

(here designated)(Fig. 3). ‘Leptospermum ericoides A.Rich. fl N^lle Zél 338 N^lle Zélande (Astrolabe)’ P214999!

Figure 3. 

Paralectotype of A Leptospermum ericoides (held at l’Herbier du Laboratoire de Phanérogamie du Muséum National d’Histoire Naturelle (P)). B Enlargement of paralectotype label showing Achille Richard’s distinctive handwriting.

Allan (1961; p. 322), Thompson (1983) and Harris and Cadic (1998; p. 36) published partial lectotypifications of Leptospermum ericoides A.Rich (see Article 9, of the International Code of Nomenclature (McNeill et al. 2012). The first two authors did this through their citation of elements of the protologue and by their indication of where type material was lodged, while Harris and Cadic (1998) also published statements of their intent to typify, noting also the location of type material and stating that they had seen ‘Richard’s type’. However, because there are at least two syntypes of Leptospermum ericoides A.Rich. in l’Herbier du Laboratoire de Phanérogamie du Muséum National d’Histoire Naturelle (P) (Figs 2, 3) that bear the distinctive handwriting of Achille Richard (Fig. 2, 3; see also Burdet 1978) and match the protologue with respect to their morphological condition and collection details, and because none of these author’s publications specified which one of these was the lectotype, a further lectotypification is necessary. Of the two sheets that I have seen, both are without date or collector. One (Fig. 2) from the ‘Herbarium Richard’ has two labels. The first of these is in Richard’s hand and reads ‘Leptospermum ericoides nob. N^lle Zélande’. The second label is printed ‘HERB. MUS. PARIS’ and ‘HERBIER E. DRAKE’. This sheet has also been stamped by the herbarium staff (P. Morat pers. comm.) ‘TYPE’ and bears two handwritten labels by W. Harris. The upper label states ‘Lectotype designation on the basis of ex Herbier E. Drake ex Herbarium Richard (in red) and Leptospermum ericoides Rich.’ and the next reads ‘Designated Lectotype Leptospermum ericoides A.Rich. Essai. Fl. N.Z. 1832, 338, syn. Kunzea ericoides (A. Richard) Joy Thompson Telopea 2(4) 378, 1983 W. Harris 26 July 1988’. The second herbarium sheet (Fig. 3) also bears two labels (Fig. 2A, B). The first of these (see Fig. 3B) is written in Richard’s hand, and reads ‘Leptospermum ericoides A.Rich. fl. Nov. Zél. N^lle Zélande (Astrolabe)’, the second (Fig. 3A) is printed ‘HERB. MUS. PARIS’. I designate as lectotype, the sheet labelled by Richard ‘Leptospermum ericoides nob.’ because this matches his protologue (p. 338) with respect to the usage of the possessive Latin abbreviation ‘nob.’ i.e. nobis, meaning ‘to us, of us’ in the sense of ‘this is my [choice of] name’ (Stearn 1992; R. O. Gardner pers. comm.). Further, it is the only sheet clearly identified as part of Richard’s herbarium. This was also the sheet designated lectotype by Harris. However, because Harris did not publish this, his lectotypification, although accurate cannot be upheld.

I conservatively designate as paralectotype the second sheet at P. This is because it is without date or collection notes so it is impossible to tell if it was part of the same gathering as the lectotype.

The specific epithet ericoides alludes to the similarity in the growth form of Kunzea ericoides to Erica arborea L. i.e., ‘Ericam arboream habitu referens’ (Richard 1832; p. 338).

(Figs 4, 5, 6). Growth habit mostly trees up to 18 m, sometimes (such as on ultramafic rocks and soils) decumbent and trailing forming shrubs up to 2 × 1 m. Plants with tree-habit usually rather slender and gracile with a somewhat spreading canopy; those in exposed conditions branching at or close to the trunk base, while those growing in dense stands or sheltered sites usually with the lower half devoid of branches. Plants with a decumbent habit usually heavily branched, not rooting on contact with soil. Trunk 1(–4) arising from the ground, 0.10–0.60(–0.85) m d.b.h., mostly erect, slender, weakly flexuose; often basally buttressed, mature trees usually devoid of branches for the first 1–2(–4) m; decumbent plants with scarcely discernible trunk due to branches arising from or close to the base; basal portion of trunks covered with layers of somewhat firm to semi-detached, weakly tessellated, short to long, ± irregularly tabular lengths of subcoriaceous brown-grey to greyish-white bark. Bark early bark chartaceous to subcoriaceous, brown to grey-brown, ± elongate, usually bearing a few transverse cracks (especially on branch flanges and decurrent leaf bases) otherwise remaining firmly attached, margins elongate sinuous, ± entire with scarcely any flaking; old bark similar though more distinctly corky subcoriaceous, often coarsely tessellated or broken in long elongate sections, otherwise remaining firmly attached, if detaching then usually doing so along transverse cracks, and peeling inwards and upwards to leave distinct layers of elongate to coarsely tabular, chartaceous, flakes that are centrally attached, with sinuous margins; upper bark surface usually with much secondary peeling, these flakes similar to primary flakes but more distinctly chartaceous, smaller, narrowly elongate with widely sinuous margins; bark usually crumbling readily in hand, and breaking readily if pulled hard into numerous, small, ± tabular to distinctly irregular flakes. Branches depending on growth habit and situation, numerous, initially arising from close to or at trunk base but as plants mature basally thinning such that branches are retained only in the upper half of the tree; usually rather slender, initially ascending but soon spreading, with apices often distinctly pendulous, branch bases mostly clean, sometimes congested by partially decorticated bark; branchlets numerous, usually rather slender, gracile, initially ascending, soon spreading, terminal growth erect or pendulous; initially bright green or bronze green, sometimes red, ± quadrangular to subterete, glabrescent; new growth sericeous, indumentum initially copious, soon sparse, deciduous, hairs divergent 0.02–0.05 mm long, hyaline to translucent (appearing silvery-white when young maturing silver-grey), apices straight not curled or curved; leaves of branchlets densely crowded along stems and brachyblasts; brachyblasts usually closely spaced, though in vigorous new growth they are sometimes quite widely spaced. Vegetative buds inconspicuous, usually obscured from view by surrounding leaves; at resting stage 0.5–0.8 mm diam. narrowly to broadly ovoid; scales often persistent; (0.4–)1.1 mm long, dark red-brown, broadly ovate, ovate-lanceolate grading through to lanceolate, rostrate to cuspidate; midrib strongly keeled, with one row of 4–8 oil glands on either side of midrib; scales glabrous except for the margins and apex; these densely invested in white, silky hairs. Leaves homophyllous; sessile, lamina surfaces glabrous, margins and the basal, adaxial portion of the midrib hairy (especially on young leaves); densely crowded (particularly toward apices) along branchlets and brachyblasts; initially obliquely ascending, mostly suberect to spreading when mature; lamina (4.0–)13.5(–25.0) × (0.5–)1.1(–1.8) mm, bright green to yellow-green, rarely dark green, adaxial surface often glossy when fresh, drying dull, abaxial surface paler; lamina linear, linear-lanceolate, to narrowly lanceolate, straight or with distal quarter weakly recurved, apex acute, sometimes cuspidate, base attenuate; adaxial lamina surface flat to weakly concave, without obvious oil glands, midrib very slightly raised near base, otherwise scarcely evident, basal portion finely and sparsely covered with deciduous, antrorse-appressed sericeous hairs; abaxial surface flat to weakly convex, glandular punctate, oil glands up to 200; midrib glabrous, usually not evident when fresh, sometimes weakly raised just near base, often not evident when dry but sometimes discernible as a slight groove for entire length; lamina margins initially very finely sericeous, becoming glabrate or glabrous; hairs when present antrorse-appressed, forming a fine, often discontinuous band failing just short of lamina apex, otherwise decurrent along leaf bases. Perules deciduous or persistent, (0.6–)0.8(–1.5) mm, initially squamiform, becoming foliose toward first flower, dark red-brown, broadly ovate, ovate-lanceolate grading through to lanceolate, rostrate to cuspidate; midrib strongly keeled, with one row of 4–12 oil glands on either side of midrib; glabrous except for the margins and apex; these densely invested in white, sericeous hairs. Inflorescence mostly a compact corymbiform to shortly elongate (3–)8(–15)-flowered botryum up to 60 mm long; usually on brachyblasts with the terminal shoot corymbiform or extending as a slightly longer (up to 80 mm long) 6–15-flowered, elongate botryum with flowers usually crowded, terminal portion usually bearing undeveloped flowers and active vegetative growth. Inflorescence axis densely invested with short, weakly divergent silky hairs. Pherophylls foliose ± persistent, 1 per flower; lamina (3.0–)6.7(–7.8) × (0.9–)1.1(–1.4) mm, leaf-like pherophylls bright green (rarely dark green) elliptic, lanceolate to narrowly lanceolate, apex acute, base attenuate; adaxial surface weakly concave to flat, oil glands scarcely evident up to 10; midrib scarcely evident at base only, surface glabrous; abaxial surface weakly convex or flat, oil glands up to 30; midrib scarcely evident at base only, lamina margin glabrescent, hairs as for leaf margins. Pedicels (1.6–)2.7(–3.8) mm long at anthesis, usually elongating slightly after anthesis, terete, usually glabrous, very rarely sparsely covered with divergent to weakly sericeous hairs. Flower buds pyriform to narrowly obconic, apex of mature buds weakly domed to flat, calyx lobes distant, not touching. Fresh flowers when fully expanded (4.1–)6.3(–8.3) mm diam. Hypanthium (1.4–)2.1(–3.2) × (1.9–)2.9(–4.1) mm, with free portion (0.4–)0.6(–1.0) mm long, bright green, bronze-green or yellow-green mottled with red; sharply obconic, apex terminating in a usually dark pink or crimson chartaceous rim bearing five persistent suberect to spreading calyx lobes (rim usually drying dark maroon to maroon-black); external hypanthium surface smooth, glabrous (very rarely glabrescent with basal quarter finely and sparsely covered with minute weakly antrorse hairs); oil glands, conspicuous, ± colourless; ribs not evident when fresh, conspicuous (along with venation) when dry. Calyx lobes 5, suberect to spreading, subcoriaceous, (0.4–)0.7(–1.0) × (0.4–)0.8(–1.0) mm, persistent, orbicular, obtuse to broadly deltoid, red-green, pink or crimson, keel not evident in fresh material, becoming prominent when dried, oil glands conspicuous, ± colourless, margins glabrous or finely ciliate; cilia white. Receptacle green or pink at anthesis, darkening to crimson or dark magenta after fertilisation. Petals 5, (1.4–)2.2(–2.6) × (1.5–)2.2(–2.9) mm, white (often drying yellow), orbicular, suborbicular to narrowly ovate, spreading, apex rounded, margins often incurved, entire or very finely denticulate, oil glands usually not evident when fresh, ± colourless. Stamens (10–)18–24(–34) in 1–2 weakly defined whorls, arising from receptacle rim, filaments white. Antipetalous stamens (2–)3(–5), antisepalous 2–3(–4). Antipetalous stamens outcurved usually with distal portion slightly incurved, on filaments 1.6–2.8 mm long, inner stamens if present, confined to the bases of the outermost antipetalous pair, 0.8–1.2 mm, incurved. Antisepalous stamens shorter than outermost antipetalous stamens, 0.6–1.2 mm, weakly to strongly incurved, rarely erect or outcurved, often in mixtures of both. Anthers dorsifixed, 0.35–0.48 × 0.16–0.24 mm, broadly ellipsoid, latrorse. Pollen white (14.1–)14.6(–17.3) μm. Anther connective gland prominent, pink or pinkish-orange when fresh, drying red to orange, ± spheroidal to pyriform, ± immersed to half of length between anthers, ± coarsely papillate. Ovary (4–)5 locular, each with 16–21(–24) ovules in two rows on each placental lobe. Style 1.5–2.2 mm long at anthesis, elongating slightly after anthesis, white, rarely basally flushed pink; stigma capitate, about 1¼ × the style diam., flat, cream or white, flushing pink after anthesis, surface very finely granular-papillate. Fruits rarely persistent, (1.9–)2.7(–3.4) × (1.8–)2.8(–3.9) mm, glabrous, initially dark green to reddish-green, maturing brown to grey-brown to grey-black; in all types fading with age to pale greyish-white, cupular, barrel-shaped, shortly cylindrical to hemispherical, calyx valves usually erect with the apices incurved, splits concealed by dried, erect, free portion of hypanthium. Seeds 1.00(–1.05) × 0.32(–0.50) mm, usually curved near apex, laterally compressed, 2–3-angled with convex to flattened faces, apex rounded to subacute; base oblique, ± flattened; testa semi-glossy, orange-brown to dark brown, obovoid, oblong, oblong-ellipsoid, or cylindrical and ± curved, surface coarsely reticulate. FL: (Nov–)Dec–Jan(–Mar). FT: Feb–Apr(–Aug). Chromosome Number n = 11_II, 2n = 22 (see de Lange and Murray 2004).

Figure 4. 

Distinguishing features of Kunzea ericoides. A Flowering branchlet (ex cult. AK 289138) B Fruiting branchlet (ex cult. AK 289138) C Vegetative bud and branchlet indumentum (ex cult. AK 289138) D Adaxial leaf surface (ex cult. AK 289138) E Abaxial leaf surface (ex cult. AK 289138) F Adaxial leaf apex (ex cult. AK 289138) G Leaf margin indumentum (ex cult. AK 289138) H Leaf variation: (H1) Abel Tasman National Park, Astrolabe Roadstead, (AK 253380), (H2) Knuckle Hill (AK 289160) I Flower (top view) (AK 289138) J Flower and hypanthium (side view) (AK 289138) K Flower cross section showing anther, style and ovules (AK 289138) L Style and stigma (AK 289138) M Stamen (AK 289138) N Dehisced fruit (AK 253380). Scale bars: (A, B, H) 10 mm; (C–F, I–N) 1 mm; (G) 0.5 mm.

Figure 5. 

Scanning Electron Micrographs of Kunzea ericoides. A–G Branchlet indumentum (AK 253380; AK 289161); Seeds H–K (HR 3766 K Testa surface showing reticulum (CHR 3766). Scale bars: (A, C) 500 μm; (B, D, G, H, J–K) 100 μm; (E, F, I) 1 mm.

Figure 6. 

Kunzea ericoides. A K. ericoides trees colonising old burn, South Island, north-west Nelson, Abel Tasman National Park, Awapoto River (photo: P. J. de Lange) B Flowering decumbent plant on windswept ridge line, South Island, north-west Nelson, Wakamarama Range, Knuckle Hill (photo: P. J. de Lange) C Young K. ericoides tree, South Island, north-west Nelson, Marahau (photo: P. J. de Lange) D Mature K. ericoides tree, South Island, north-west Nelson, Canaan Downs (photo: G. M. Crowcroft) E Flowering branchlets, South Island, eastern Nelson, Richmond Range, Hackett Creek (photo: G. M. Crowcroft) F Bark, South Island, eastern Nelson, Richmond Range, Hackett Creek (photo: G. M. Crowcroft) G Bark, South Island, north-west Nelson, Canaan Downs (photo: G. M. Crowcroft) H Flowering branchlet showing brachyblasts, leaves and few-flowered corymbiform botrya, South Island, north-west Nelson, Golden Bay, Pupu Springs (photo: M. D. Wilcox) I Close up of flowers, South Island, north-west Nelson, Golden Bay, Bishops Saddle (photo: G. M. Crowcroft).

(99 sheets seen).New Zealand (South Island). Wakamarama Range, Knuckle Hill, P. J. de Lange 4953, 10 Jan 2001, (AK 253376, Duplicate: AD, CHR); Whanganui Inlet, P. J. de Lange 4952, 10 Jan 2001, (AK 289162, Duplicate: AD); Wakamarama Range, Mt Burnett, South Peak, P. J. de Lange 4946, 10 Jan 2001, (AK 289159, Duplicate: AD, MEL); Aorere River, H. Talbot s.n., 15 Dec 1959, (CHR 300594); Golden Bay, Wainui Inlet, Takapou Point, P. J. de Lange 4994, 12 Jan 2001, (AK 289173, Duplicate: AD); Waitui Stream, upper Takaka, W. D. Burke s.n., 23 Nov 1979, (WELTU 13480); Kahurangi National Park, Mt Peel, near source of Trilobite Creek, P. J. de Lange 6329 & G. M. Crowcroft, 16 Jan 2001, (AK 289179); Abel Tasman National Park, Astrolabe Roadstead, Adele Island; P. J. de Lange 5001, 13 Jan 2001, (AK 253380, Duplicate: AD, P); Moutere E.D., Upper Moutere, J. F. F. Hobbs s.n., 18 Aug 2002, (NZFRI 25012); Wangapeka River Road, Rolling River, P. J. de Lange 5082, 21 Jan 2001, (AK 253381, Duplicate: AD); Hope River, Sandy Creek, P. J. de Lange 5085, 21 Jan 2001, (AK 287548, Duplicate: AD); Ngatimoti, Haycock's Hill, R. Wilson s.n., Nov 1964, (OTA 13619); Bryant Range, Hackett Creek, near Whispering Falls, P. J. de Lange 5031 & G. M. Crowcroft, 17 Jan 2001, (AK 286127, Duplicate: AD); Golden Downs, Wakefield, Faulkner’s Bush, P. J. de Lange 5073, 21 Jan 2001, (AK 289193, Duplicate: AD); Nelson Lakes National Park, Speargrass Creek, Speargrass Track, P. J. de Lange 5066, 23 Jan 2001, (AK 289190, Duplicate: AD); Owen Valley East Road, Carrol Creek, P. J. de Lange 5136, 21 Jan 2001, (AK 289201, Duplicate: AD); Buller River, near Owen Junction, P. J. de Lange 5137, 21 Jan 2001, (AK 289202, Duplicate: AD, CHR); Lower Buller Gorge, Buller River, P. J. de Lange 4787 & P. I. Knightbridge, 7 Dec 2000, (AK 288294, Duplicate: AD); D'Urville Island, north of Attempt Hill, P. J. de Lange 5053 & G. M. Crowcroft, 19 Jan 2001, (AK 289185); Pelorus Sound, Mahakipawa Inlet, Moenui, P. J. de Lange 4904, 8 Jan 2001, (AK 288400, Duplicate: AD); Mt Freeth, Queen Charlotte Sound, W. R. B. Oliver s.n., 5 Apr 1931, (WELT SP029535); Cloudy Bay, Rarangi - Port Underwood Road, top of Rarangi Zig Zag Track, P. J. de Lange 5116, 23 Jan 2001, (AK 289198, Duplicate: AD).

(Fig. 7). Endemic, New Zealand, South Island (sea level–1600 m a.s.l.). Kunzea ericoides is endemic to the northern South Island north of and including the Wairau and Buller River catchments.

Figure 7. 

Distribution of Kunzea ericoides, K. salterae, K. sinclairii, K. tenuicaulis, K. toelkenii, K. triregensis and K. “Lottin Point”.

Kunzea ericoides can be easily recognised by its glabrescent, often terminally pendent branchlets (Fig. 6D). The branchlet hairs (Fig. 5A–D, G) are deciduous, consistently divergent, short (up to 0.05 mm), and usually sparse. The leaves of K. ericoides are typically bright green, linear, linear-lanceolate, to narrowly lanceolate, glabrate and mostly crowded on brachyblasts (Fig. 6H). Often the brachyblasts are widely spaced along the branchlets. The hypanthium of K. ericoides is sharply obconic, glabrous (very rarely sparsely hairy), and copiously dotted with conspicuous colourless oil glands. The fruits of K. ericoides are usually glabrous (very rarely sparsely hairy near the base) and are mostly cupular, barrel-shaped, or shortly cylindrical in shape (Fig. 4N). In growth habit K. ericoides varies from a trailing, decumbent shrub on ultramafic substrates to a tall tree (up to 18 m tall) but, irrespective of stature, it typically retains a somewhat spreading to slightly pendulous openly branching habit (Fig. 6B–D). Cytologically K. ericoides has a similar chromosome complement to K. amathicola, K. triregensis and K. sinclairii (de Lange and Murray 2004). Molecular evidence (rDNA ITS) grouped Kunzea ericoides with K. linearis, K. triregensis, K. robusta (all samples), K. serotina, and K. toelkenii (Table 2; see also de Lange 2007; de Lange et al. 2010). However, the rDNA ETS sequence of K. ericoides had one unique cytosine/thiamine mix, shared with K. linearis a cytosine, and, with K. linearis, K. robusta (Mt Egmont samples only), K. salterae, K. serotina and K. toelkenii, a guanine/cytosine mix (Table 2; see also de Lange 2007).

Kunzea ericoides is sympatric with K. amathicola, K. robusta and K. serotina and also forms hybrids with them. Of the three, Kunzea ericoides is most frequently sympatric with K. robusta. Both species naturally meet in the Marlborough Sounds where K. robusta, though mainly coastal, is locally common (another occurrence at Totaranui, Abel Tasman National Park, results from the naturalisation from plantings of K. robusta within a camp ground). On the south-eastern coast of D’Urville Island and from about the Tory Channel south toward Rarangi, both species are abundant. The form of K. robusta present in the north-eastern South Island is easily distinguished from K. ericoides as it has extremely hairy branchlets, with the long, silky, antrorse-appressed hairs easily seen by the naked eye or with a 10× hand lens. Aside from branchlet hairs both species have different foliage colour and leaf shape; K. robusta mostly has dark green leaves that are oblanceolate, broadly oblanceolate to broadly lanceolate with irregularly long sericeous hairs on the margins. Kunzea ericoides has bright green, linear-lanceolate leaves whose margins are sparsely hairy trending toward glabrous (Fig. 4D–H). The hypanthia of both species also differ. In K. robusta the hypanthium is broadly obconic to almost turbinate with the external faces copiously covered in short, antrorse-appressed, sericeous hairs. In K. ericoides the hypanthium is glabrescent to completely glabrous, and sharply obconic. The fruits show much the same differences, K. robusta being broadly obconic to turbinate with the external faces distinctly hairy and K. ericoides mostly cupular, barrel-shaped to shortly cylindrical or sometimes hemispherical and glabrous.

In the north-western part of its range K. ericoides is frequently sympatric with K. amathicola following a line running from the Aorere River west to the Paturau River. Again, both species are easily distinguished due to the very different branchlet indumentum. Kunzea amathicola is distinctly more hairy, with the branchlet hairs long, sericeous, and antrorse-appressed, a marked contrast to the glabrescent branchlets, and short divergent hairs diagnostic of K. ericoides. The leaves, inflorescence type, flowers and fruits of both species are also very different and further distinctions are given in more detail under K. amathicola and in Table 1. However, two critical characters are briefly mentioned here. The inflorescences of K. amathicola are always elongate botrya, while those of K. ericoides are usually corymbiform (Fig. 4A) though mixtures of corymbiform and shortly elongate botrya are sometimes produced in shaded, stressed or late summer flowering specimens. The pherophylls of K. ericoides are always in mixtures of squamiform and foliose, with the foliose ones ranging from elliptic, lanceolate to narrowly lanceolate, they are never oblong to oblong-obovate like those of K. amathicola. The other major difference is that the leaf margins of K. amathicola are covered in a conspicuous thick, continuous band of long antrorse-appressed, sericeous hairs which reaches the leaf apex. In K. ericoides although marginal hairs are usually present these are hardly conspicuous, tend to be shed as the leaf matures, and they normally form a rather thin, often discontinuous band which does not reach the leaf apex (Fig. 4F–G).

Along the Buller River, at the northern end of Nelson Lakes and in the upper Wairau, Kunzea ericoides is occasionally syntopic with K. serotina. Both species are easily distinguished (Table 1), especially as K. serotina has such a distinctive bark type and a pyramidal, columnar, growth habit with obliquely ascending, fastigiate branches. Further, the usual leaf shape of K. serotina is shortly linear-oblanceolate to obovate rather than the longer linear-lanceolate condition of K. ericoides. However, on the ultramafics of the Red Hills and on Mt Dun, K. ericoides tends to be more stunted and in these forms the leaves tend to be half their usual size. While this condition does not appear to have a genetic basis, field distinction can become confusing, especially when such forms grow amongst similarly dwarfed K. serotina specimens. In these situations the branchlet hairs are diagnostic: in K. serotina they are copious, up to 0.08 mm long with curved or curled apices, whereas the branchlets of K. ericoides are glabrescent, with the hairs up to 0.05 mm long and with straight apices (Fig. 5B, D). Flowering material readily separates both species because K. serotina has deciduous, mostly spathulate to spathulate-orbicular, rather than persistent, elliptic, lanceolate to narrowly lanceolate pherophylls, and the oil glands in the petals are yellow, rather than ± colourless.

Although much has been written about the ecology of Kunzea ericoidessens. lat. (Burrell 1965; Burrows 1973; Wardle 1991; Smale 1994; Smale et al. 1995) most of that information is based on K. amathicola, K. linearis, K. robusta, K. salterae, K. serotina, and K. toelkenii. My observations suggest that Kunzea ericoides is ecologically most similar to K. robusta in that it can be found in a wide variety of habitats from early stage seral shrubland (Fig. 6A) through to tall forest, and it has a similar growth habit. However, it attains a much higher altitudinal limit than K. robusta, having been collected during this study at 1600 m a.s.l. on the northern slopes of Mt Peel, Kahurangi National Park, north-west Nelson. In montane areas it is frequently found within north-facing canopy gaps developed within montane black beech (Fuscospora solandri (Hook.f.) Heenan et Smissen) and mountain beech (F. cliffortioides (Hook.f.) Heenan et Smissen) forests, often close to the upper tree limit for these species in the Richmond Range and parts of north-west Nelson. The species is also commonly encountered in coastal forest, though rarely on sand dunes, and it can be a conspicuous tree of lowland areas, especially along river flats and on outwash gravels within the Waimea Plain and on sections of the northern bank of the Buller River.

No obvious substrate requirement is evident. It appears to avoid permanently waterlogged soils and peat and it is scarce from wetlands, though it may at times be common in the vegetation bordering these habitats. Free draining soils and recent alluvium is commonly colonised, as is open ground within lowland to upper montane forests. On the extensive karstfield of north-west Nelson K. ericoides is often prominent, especially in places where there has been a history of logging, mining, farming or frequent fires. Kunzea ericoides is also a common component of the ultramafic areas of D’Urville Island, Mt Dun, the Red Hills and upper Takaka.

Kunzea ericoides is an important primary tree coloniser of formerly cleared ground in many parts of its range. In these situations, perhaps more than any other, the dense leaf litter produced by the often closely growing trees is ideal for a wide range of terrestrial orchids (especially of the genera Acianthus R.Br., Caladenia R.Br., Corybas Salisb., Gastrodia R.Br., and Pterostylis R.Br.), and fungi (see McKenzie et al. 2006), while the bark is often colonised by mosses such as Macromitrium Brid. spp., Cryphaea tenella Müll.Hal., Distichophyllum pulchellum (Hampe) Mitt., and Weymouthia cochlearifolia (Schwägr.) Dix, and by liverworts of the genera Frullania Raddi, Lejeunea Lib. and Metzgeria Raddi. Kunzea ericoides bark also supports a diverse array of lichens in the following genera; Coccocarpia Pers., Heterodermia Trevis., Pannaria Delise ex Bory, Parmotrema A.Massal., Pseudocyphellaria Vain., Sticta (Schreb.) Ach. and Ramalina Ach. The upper branches and branchlets are also frequently parasitised by the dwarf mistletoe Korthalsella salicornioides (A.Cunn.) Tiegh. and, less commonly, green mistletoe (Ileostylus micranthus (Hook.f.) Tiegh.). Kunzea ericoides-dominated shrubland and forest also provides an important and at times critical habitat for a range of geckos in the endemic genera Mokopirirakau Nielsen et al. 2011, Naultinus (Gray, 1842), and Woodworthia Garman 1901 (R. Hitchmough pers. comm.).

Kunzea ericoides naturally hybridises only with K. amathicola, K. robusta and K. serotina. The most common of these hybrids is K. ericoides × K. robusta, which is abundant on the south-eastern side of D’Urville Island (especially around Katherine Bay), and along the eastern side of the Marlborough Sounds, particularly east of Picton and south of the Tory Channel to about Rarangi. These areas comprise some of the first sites of European settlement in New Zealand, prior to which they were once heavily populated by Maori (Buick 1976; de Lange et al. 2005). Both cultures extensively modified these landscapes and even in modern times much of this area has been repeatedly burned. These are ideal conditions for hybridism, and in the case of both K. ericoides and K. robusta it has helped create and maintain an extensive introgressed hybrid swarm.

Recognition of K. ericoides × K. robusta in the field and in the herbarium is generally easy because both species have such different branchlet hairs. Hybrid plants can usually be recognised by the presence of mixtures of long, silky, antrorse-appressed and very short divergent hairs. Further, K. ericoides × K. robusta hybrids often have a distinctly pale glaucous sheen to their usually yellow-green leaves. Similar plants were produced in the experimental F₁ crosses K. ericoides^♀ × K. robusta^♂ and K. robusta^♀ × K. ericoides^♂ and these proved to be fully fertile (de Lange et al. 2005). In most instances the hybrid can be easily recognised but because hybrids are fully fertile, an often bewildering array of introgressants can be found in sites of prolonged disturbance, in some cases almost to the exclusion of either parent. Recognition of hybrids within such populations can at times be difficult. For example, in those hybrids trending toward K. ericoides there is a progressive loss of the long, antrorse-appressed sericeous branchlet hairs diagnostic of K. robusta. In some examples where the branchlet hairs are virtually dominated by short divergent hairs, a hybrid ancestry may still be elucidated by diligent searching, particularly along the branchlet axis immediately opposite active leaf buds and emergent leaves where a few of the longer, antrorse-appressed sericeous type diagnostic of K. robusta are usually retained as sparse patches. In pure K. ericoides these are mostly all soon shed. One of the last K. robusta traits to be lost is the presence of antrorse-appressed, silky hairs on the external surface of the hypanthia, allowing the hybrid origin of specimens in all other respects matching K. ericoides to still be elucidated. Branchlet hairs, or rather their relative abundance also serves to help distinguish introgressants trending toward pure K. robusta such that apparently pure specimens of K. robusta prove on careful examination of the branchlets to be glabrescent and completely dominated by short, divergent hairs. For K. ericoides at least, the branchlet hairs seem to be a long-lived trait traceable in any hybrid swarm it may form. In north-west Nelson from about Golden Bay and the Whanganui inlet north, K. ericoides and K. amathicola are commonly syntopic and hybrids are frequent in the more heavily modified lowlands and roadsides of this area. Like the K. ericoides × K. robusta hybrid, the branchlet indumentum enables recognition because it comprises mixtures of both hair types. The inflorescences of K. amathicola × K. ericoides hybrids are also in mixtures of mostly elongate (typical of K. amathicola) and some subcorymbiform to corymbiform botrya (typical of K. ericoides). The pherophylls tend to be rather variable in length and size but they are mostly lanceolate to linear-lanceolate like K. ericoides. The leaf margins of K. amathicola × K. ericoides are also distinctive. Initially they are hairy but, like K. ericoides, the hairs are progressively shed during leaf maturation. Further, hairs rarely (if ever) meet at the leaf apex.

Kunzea ericoides × K. serotina is much less common, because the ranges of both species rarely overlap. This hybrid is mostly confined to the upper Buller River and within the Barnicoat Range and Hackett areas of eastern Nelson where the usually montane K. serotina extends down into lowland areas and so abuts the main habitats of K. ericoides. These are also areas where road construction and maintenance has significantly altered the surrounding indigenous vegetation allowing these species to meet and hybridise. Kunzea ericoides × K. serotina is easily recognised by the erect, somewhat open, sub-pyramidal growth habit, short, weakly fastigiate branchlets, and bright yellow-green, red-tinged leaves (all features of K. serotina). However, as with K. ericoides, the leaves tend to be linear-lanceolate, though sometimes quite broadly so. Branchlet indumentum is not particularly useful because both parent species have divergent hairs, though in most examples the hybrid tends to be, like K. serotina, distinctly hairy. However, rare glabrescent examples suggest that introgressive hybridisation toward K. ericoides is occurring at some sites. The pherophylls of K. ericoides × K. serotina, like those of K. serotina, are not persistent but are mostly shed during flower maturation. Further, the pherophylls of the hybrid are pandurate to elliptic (never spathulate like K. serotina or narrowly elliptic, lanceolate to narrowly lanceolate like those of K. ericoides). The flowers of K. ericoides × K. serotina offer another point of recognition. Like K. serotina they tend to have more obvious oil glands than is usual for K. ericoides, and while those of K. ericoides are normally colourless, those of the hybrid vary from colourless through to pale yellow, the latter of which is the usual condition of K. serotina.

Although now universally known as ‘kanuka’, the name recorded for Kunzea ericoides from the Nelson area during Dumont d’Urville’s second voyage was ‘manuoea’ (Richard 1832). Based on other specimens lodged at P(!) and elsewhere this species was also known as ‘manuka’, ‘titire’ and ‘atitire’ (for a discussion on the names ‘titire’ and ‘atitire’ see K. robusta).

Kunzea ericoides is a widespread and abundant species throughout its northern South Island range and is lited as ‘Not Threatened’ by de Lange et al. (2013b).

A K. ericoide habitu in arboribus juvenalis et juveni-adultis, erecto columnare fastigiato vel pyramidale, transformante in arboribus veterrimis habitu laxe ramoso divergente apice applanata; cortice maturo crisparenti circinanti et decorticanti promte partibus parvis irregularibus chartaceis, habitu fastigiato, pherophyllis spathulatis, hypanthio urceolato vel campanulato, ovariis saepe trilocularibus, glandibus oleosis flavis petalorum differt.

(Fig. 8).New Zealand: North Island, 3.5 km north-east of Rangitaiki River, 38°55'26"S, 176°26'1"E, 743 m a.s.l. ‘Abundant at interface between frost flat and upland Podocarp forest. Growing in skeletal soil overlying Taupo Pumice (186 A.D. ejecta), in association with dense Dracophyllum subulatum’. P. J. de Lange 6695, 18 Feb 2005, AK 297548! Isotypes: AD! BM! CHR! MEL! WELT!

Figure 8. 

Holotype of Kunzea serotina de Lange et Toelken (P. J. de Lange 6695, AK 297548).

The specific epithet serotina refers to the later flowering habit of the species (peaking January–February) when compared to K. robusta and K. tenuicaulis, earlier flowering species (peaking October–December) with which K. serotina is sympatric in the North Island.

(Figs 9, 10, 11).Growth habit erect, initially columnar to tightly pyramidal, fastigiate, densely branched shrubs or trees 3–20 × 2–4 m developing with time into less densely branched, open pyramidal crowns; in very stable conditions growth habit changing in aged specimens to a more openly branched, flat-topped, divergent crown, with branches restricted to upper half to third of trunk. Trunk usually single, very rarely 2–3 arising from base, 0.10–0.45(–0.86) m d.b.h., often basally buttressed, with basal portion of trunk covered in numerous, curled, chartaceous bark remnants. Bark early bark chartaceous, greyish-white to pinkish-white, ± elongate, initially with few transverse cracks but soon becoming heavily cracked (especially on branch flanges and decurrent leaf bases), often crumpled, soon detaching as in-rolled, curled, sinuous, irregular pieces, with ± frayed margins, detached pieces often congregating in branch forks and forming deep drifts at trunk base; old bark similar though more distinctly chartaceous-corky, upper surface with much secondary peeling, transverse cracking and crumpling; detaching readily, with flakes often hanging in loose curled masses beneath main branch forks and along trunk, margins rather irregular, sinuous, very rarely somewhat tabular; early and old bark flakes crumbling readily in hand. Branches numerous, usually arising at or near trunk base; short and stout, obliquely ascending, fastigiate; branchlets numerous, very leafy with many brachyblasts, quadrangular, indumentum copious, sericeous; persistent, divergent 0.05–0.06(–0.08) μm, hyaline to translucent (appearing white when young maturing grey) apices curved or slightly curled. Vegetative buds usually inconspicuous and obscured by surrounding foliage; at resting stage 0.2–0.5(–1.3) mm diam.; scales scarious, deciduous, (0.3–)0.8(–1.3) mm long, yellow-brown to red-brown, ovate, ovate-deltoid to broadly deltoid-rostrate; midrib prominent, strongly keeled in upper half, prolonged to long cuspidate tip, lateral veins absent, oil glands usually absent, upper half of scale margins, keel, and keel apex ciliate. Leaves heterophyllous; seedling, subadult leaves and those of reversion shoots, spreading to patent; lamina (0.8–)5.2(–7.8) × (0.6–)0.8(–1.2) mm, red-green or pale green suffused with red, rarely bright green, linear-lanceolate to lanceolate; flat or involute, apex acute to obtuse, finely cuspidate; adult leaves usually densely aggregated along brachyblasts of branchlets, initially obliquely ascending to suberect, spreading; lamina (2.0–)3.7(–6.3) × (0.8–)1.1(–1.8) mm, dark glossy green or bronze-green, margins and base often flushed red, linear-oblanceolate, oblanceolate to obovate; strongly recurved from about half of total length, apex initially acute to subacute, maturing obtuse to rounded, often cuspidate; base attenuate; adaxial surface concave or flat, glandular punctate; oil glands up to 780, more evident when dry, midrib slightly raised near base, otherwise not evident for rest of length, glabrous, very rarely with fine antrorse hairs near base; abaxial surface convex, glandular punctate, oil glands up to 180, more evident when dry; midrib flush with surface or depressed, glabrous, very rarely with a fine covering of silky, antrorse-appressed, antrorse to patent hairs near base; lamina margin sparsely hairy, hairs finely silky, flexuose, appressed to spreading, antrorse to subantrorse up to 0.08 mm long, hyaline to translucent, appearing as white to naked eye, aligned in one usually interrupted row failing well short of cuspidate leaf apex. Perules scarious, persistent, (0.3–)0.6(–1.0) mm; basal ones amber-brown to dark brown, broadly ovate, ovate-oblong, ± rostrate, apex acute, margins flat to involute especially in upper third, midrib weakly keeled, usually prolonged as a very short, deciduous, cuspidate apex, with one row of 4–12 oil glands on each side of midrib, glabrous except for finely ciliate margin and apex; intermediate perules deciduous, chartaceous, (0.2–)0.6(–0.8) mm long, initially pale brown to orange, upper perules usually pinkish-white when fresh, drying amber-brown to amber-orange, ovate to ovate-oblong, apex obtuse often appearing acute due to apical infolding, ± cuspidate, glabrous except for sparsely ciliate margin, weakly keeled, keel ± prolonged. Inflorescence a compact (1–)8(–12)-flowered corymbiform botryum up to 25 mm long, mostly borne on alternate, distinctly spiralled, basally densely leafy brachyblasts up to 15 mm long, each often bearing a terminal tuft of pherophylls and emergent leaves at anthesis; brachyblasts near branchlet apex usually subopposite; inflorescences at the ultimate branchlet terminus, uncommon, if present then often rather elongated and bearing well developed terminal vegetative growth. Inflorescence axis densely invested with divergent hairs. Pherophylls deciduous (falling very early), mostly foliose (rarely squamiform), 0.9–2.5 mm long, green to bronze-green, spathulate, spathulate-orbicular, rarely pandurate or lanceolate, margins and apex finely ciliate, grading into leaves at inflorescence axis apex. Pedicels (3.0–)3.5(–4.8) mm long at anthesis, usually elongating slightly after anthesis, terete, copiously invested in short, divergent to subantrorse, silky hairs. Flower buds clavate to pyriform, apex flat to weakly domed prior to bud burst with calyx valves not or scarcely meeting. Fresh flowers when fully expanded (2.8–)5.2(–8.8) mm diam. Hypanthium (1.6–)2.0(–3.4) × (1.5–)1.9(–3.8) mm, with free portion 0.4–0.8 mm long, dark green or red-green, if green then basally flushed with red when fresh, drying brown-green to red-brown; urceolate to campanulate terminating in a distinctly thicker rim bearing five persistent calyx lobes; surface smooth, copiously dotted with red oil glands, finely puberulent to ± glabrescent, with weakly defined ridges leading up to calyx lobes (these becoming more distinct upon drying); hairs if present, very short, divergent. Calyx lobes 5, upright (not spreading), firmly fleshy, (0.8–)1.0(–1.2) × (0.7–)1.0(–1.2) mm, persistent, ovate to broadly ovate, weakly keeled (keel evident only in dried specimens, where it is seen as a slightly thicker, often pale yellow, green or pink, central ridge), central portion of lobe pale green or yellow-green, with margins usually cream to pale pink, surface glandular punctate, oil glands usually pink in exposed situations otherwise ± colourless, glabrous except for distinctly spreading, ciliate margins. Receptacle usually pink at anthesis, consistently darkening to dark crimson magenta after fertilisation. Petals 5(–6), 1.4–1.6(–2.0) × 1.2–1.6(–2.0) mm, white, sometimes basally flushed pink, narrowly orbicular to broadly ovate or cuneate, apex obtuse to rounded, margins ± frayed to finely and irregularly toothed, oil glands yellow when fresh, when dried very pale yellow to colourless. Stamens 20–26(–38) in 1(–2) weakly defined whorls, arising from the receptacle rim, filaments white occasionally tinged rose-pink toward base. Antipetalous stamens (2–)3(–4), antisepalous stamens (1–)3(–6). Outermost antipetalous stamens usually weakly incurved, on filaments 0.7–1.9 mm long, inner stamen if present, 0.3–0.8 mm, strongly or weakly incurved, sometimes strongly outcurved, very rarely a further 1–2 strongly incurved stamens, 0.3–0.6 mm long, may be present at the base of the outermost antipetalous pair. Antisepalous stamens much shorter than outermost antipetalous stamens, 0.2–0.6 mm, usually incurved, rarely outcurved or in mixtures of both. Anthers dorsifixed, 0.04–0.06 × 0.02–0.04 mm, testiculate to ellipsoid, latrorse. Pollen white (11.1–)12.4(–13.7) μm. Anther connective gland prominent, orange often flushed with rose when fresh, drying dark orange-brown or purple, spheroidal, distinctly papillate. Ovary 3–4(–5) locular, each with 10–18(–23) ovules in two rows on each placental lobe. Style 0.6–0.8(–1.2) mm long at anthesis, often elongating slightly after anthesis, white; stigma capitate, scarcely wider than style, usually flat to very weakly domed along margins with a basal central depression, greenish-white, cream or pale pink, surface finely papillate. Fruits rarely persistent (1.2–)2.1(–3.0) × (1.2–)2.1(–3.4) mm, light brown to grey, finely hairy, urceolate to shortly-campanulate, rarely cupular, splits concealed by dried, suberect to erect, free portion of hypanthium and incurved calyx lobes. Seeds 0.60–0.90(–1.00) × 0.48–0.50(–0.60) mm, narrowly oblong, oblong, oblong-obovate, curved near apex, laterally compressed, 2–3-angled with convex to flattened faces, apex rounded to subacute; base cuneate to oblique, ± flattened; testa semi-glossy, orange-brown to dark brown, surface coarsely reticulate. FL: (Nov–)Jan–Feb(–May) FT: Jan–Dec. Chromosome Number n = 11_II, 2n = 22 (see de Lange and Murray 2004).

Figure 9. 

Distinguishing features of Kunzea serotina. A Flowering branchlet (ex cult. AK 282217) B Fruiting branchlet (ex cult. AK 282217) C Vegetative bud and branchlet indumentum (ex cult. AK 282217) D Seedling (AK 286184) E Adaxial leaf surface (ex cult. AK 282217) F Abaxial leaf surface (ex cult. AK 282217) G Adaxial leaf apex (ex cult. AK 282217) H Leaf margin indumentum (ex cult. AK 282217) I Leaf variation within the same individual (ex cult. AK 282217) J Adaxial surface of pherophyll (ex cult. AK 282217) K Abaxial surface of pherophyll (ex cult. AK 282217) L Side view of pherophyll (ex cult. AK 282217) M Adaxial pherophylls apex (ex cult. AK 282217) N Pherophyll margin indumentum (ex cult. AK 282217) O Pherophylls variation within the same individual (ex cult. AK 282217) P Leaf variation: (P1) Mangatoetoenui Stream (AK 288142), (P2) Te Porere Redoubt (AK 288140), (P3) Medbury Scientific Reserve, (AK 288543), (P4) Maruia Springs (AK 289968), (P5) Lewis Pass (AK 287555), (P6) Bendigo Scenic Reserve (AK 289978) Q Flower (top view) (ex cult. AK 282217) R Flower and hypanthium (side view) (ex cult. AK 282217) S Flower cross section showing anther, style and ovules (ex cult. AK 282217) T Style and stigma (ex cult. AK 282217) U Stamens (ex cult. AK 282217) V Dehisced fruit (ex cult. AK 282217). Scale bars: (A, B, D, I, O, P) 10 mm; (C, E–G, J–M, Q–V) 1 mm; (H, N) 0.5 mm.

Figure 10. 

Scanning Electron Micrographs of Kunzea serotina. A–E Branchlet indumentum (AK 285217) F–I Seeds (AK 289978) I Testa surface showing reticulum (AK 289978). Scale bars: (A, D, G) 1 mm; (B) 500 μm; (C, E, F, H) 100 μm; (I) 50 μm.

Figure 11. 

Kunzea serotina. A K. serotina trees on ridge line, North Island, Kaweka Range, upper Makahikatoa Stream (photo: P. J. de Lange) B K. serotina shrubs and trees in frost flat, North Island, Hauhangaroa Range, Moerangi (photo: P. J. de Lange) C K. serotina tree, South Island, Medbury Scientific Reserve (photo: P. J. de Lange) D K. serotina tree showing columnar growth habit and fastigiate, obliquely ascending branches, North Island, Rangitaiki Frost Flats, near Iwitahi (photo: P. J. de Lange) E Upper trunk bark and branching pattern, North Island, Rangitaiki Frost Flats, near Iwitahi (photo: P. J. de Lange) F Lower trunk bark, North Island, Rangitaiki Frost Flats, near Iwitahi (photo: P. J. de Lange) G K. serotina showing characteristic branching pattern of the species, North Island, Rangitaiki Frost Flats, near Iwitahi (photo: P. J. de Lange) H Flowering branches, North Island, Tongariro Forest, Te Porere redoubt, (photo: P. J. de Lange) I Close up of flowering branchlet, North Island, Tongariro National Park, Oturere Stream (photo: J. E. Braggins).

(155 sheets seen):New Zealand (North Island). Waikato River, Lake Waipapa, P. J. de Lange 1197, 6 Jan 1992, (AK 207188, Duplicates AD, CHR); Lake Rerewhakaaitu, W. D. Burke s.n., Jan 1960, (WELTU 3052); Te Kuiti, D. A. Franklin s.n., 1953, (WELTU 3053); Paeroa Range, Te Kopia Road, P. J. de Lange 4701, 16 Nov 2000, (AK 288232, Duplicate: AD); Pureora Forest, Link Road, Mihanga Stream headwaters, P. J. de Lange 4608 & R. O. Gardner, 18 Oct 2000, (AK 288230, Duplicate: AD); Waione Frost Flats, Waione Stream, P. J. de Lange 6440 & P. B. Cashmore, 12 Apr 2005, (AK 289785); Hauhangaroa Range, Waituhi Saddle, P. J. de Lange 4271 & P. de Lange, 31 Jan 2000, (AK 288110, Duplicate: AD); Rangitaiki, Otamatea Plains, A. J. Healy s.n., 19 Feb 1948, (CHR 62349); Puketitiri, Balls Clearing, A. P. Druce s.n., Mar 1973, (CHR 208702); Kaweka Ranges, south-west of Tutaekuri River Gorge, L. R. Perrie 3016, L. Shepherd & M. Shepherd, 14 Dec 2003, (AK 289511); Te Porere Redoubt Historic Reserve, above Waewaeru Stream, P. J. de Lange 4244 & N. J. D. Singers, 27 Jan 2000, (AK 288140, Duplicate: AD, CHR, NSW); Mt Ruapehu, Tukino, upper Mangatoetoenui Stream, P. J. de Lange 5142, 25 Jan 2001, (AK 288142); Kaimanawa Range, Waipahihi Road, P. J. de Lange 5981, 25 Jan 2001, (AK 286070, Duplicate: AD); Rangipo, Waikato Stream, P. J. de Lange 4247 & N. J. D. Singers, 27 Jan 2000, (AK 288133, Duplicates: AD, P); Moawhango – Napier Road, upper Makahikatoa Stream, P. J. de Lange 4382, 10 Aug 2000, (AK 286077, Duplicate: AD); North-West Ruahine Range, Pokopoko Stream, W. D. Burke s.n., 13 Dec 1966, (WELTU 3049); Near Rangiwahia, C. I. Pemberton Memorial Reserve, P. J. de Lange 4376, 10 Aug 2000, (AK 288234, Duplicate: AD); Puketoi Range, near Makuri, P. J. de Lange 6506, 28 May 2001, (AK 289945). New Zealand (South Island). Wangapeka Valley, R. Mason s.n., 22 Dec 1946, (CHR 58114); Hope River, Sandy Creek, P. J. de Lange 5086, 21 Jan 2001, (AK 287549); Lake Rotoiti, West Bay, W. Harris s.n., 30 Jan 1987, (CHR 437953); St Arnaud Range, Nocatchem Stream, P. J. de Lange 5129, 23 Jan 2001, (AK 286973, Duplicate: AD); Wairau Valley, between Coldwater and Judges Creeks, A. P. Druce APD1263, Jan 1991, (CHR 471855); Upper Buller River, Dellows Bluff, P. J. de Lange 4792 & P. I. Knightbridge, 7 Dec 2000, (AK 288290, Duplicate: AD); Maruia Springs, Calf Paddock, P. J. de Lange 6509 & P. I. Knightbridge, 7 Dec 2000, (AK 289968, Duplicate: AD); Red Hills, Maitland Creek, P. J. de Lange 5135, 23 Jan 2001, (AK 288548, Duplicate: AD); Awatere River, between the Hodder and Limestone Rivers, L. B. Moore s.n., 16 Feb 1962, (CHR 129190); Lewis Pass, St James Walkway Shelter, P. J. de Lange 5095, 21 Jan 2001, (AK 286185, Duplicate: AD); Hanmer Forest Park, Waterfall Trail, P. J. de Lange 4285 & B. P. J. Molloy, 8 Feb 2000, (AK 286136, Duplicates: AD, CHR); Culverden, Lowry Peaks Road, Culverden Scientific Reserve, P. J. de Lange 4284 & B. P. J. Molloy, 8 Feb 2000, (AK 2888540, Duplicates: AD, NSW, WAIK); Eyrewell Scientific Reserve, B. P. J. Molloy s.n., 30 Jan 1970, (CHR 201642). OTAGO: The Neck, between Lakes Hawea and Wanaka, B. P. J. Molloy s.n., 13 Apr 2006, (AK 296424); Dunstan Range, near Crippletown, Rocky Point (Tarras –Cromwell Road), Bendigo Reserve, B. P. J. Molloy s.n., 15 Mar 2001, (AK 289978, Duplicate: AD); Lake Roxburgh Bluffs, Clutha Valley, K. J. M. Dickinson s.n. & B. D. Rance, 24 Mar 1986, (OTA 43677); McCraes, Nenthorn Region, Manuka Creek, J. P. Burrell s.n., 18 Jan 1962, (OTA 7352).

(Fig. 12). Endemic, New Zealand, North and South Islands (30–2000 m a.s.l.). In the North Island present from about Te Kuiti, the Paeroa Range, Mt Tarawera and Kaingaroa Plain south through the Central Volcanic Plateau to the northern Aorangi Range. Absent from the high country west of Tongariro Forest, including Mt Taranaki/Egmont. In the South Island, present in the east from the upper Wairau River, and west from Karamea and the Wangapeka Valley inland along the upper Buller River and Nelson Lakes area, south through the main axial ranges to Sumner and the upper Hurunui catchment. Extending east into North Canterbury, particularly in the inland Hanmer, Emu and Amuri plains, thence present as isolated remnant stands on the Canterbury Plains. Otherwise apparently absent until Lakes Hawea and Wanaka from where it is locally present through portions of eastern Central Otago to about Roxburgh on the Clutha River and Nenthorn.

Figure 12. 

Distribution of Kunzea serotina.

Kunzea serotina is recognised by a combination of characters including growth habit, bark type, branchlet hair, leaf, floral and fruit characters. No other member of the K. ericoides complex has the same distinctive narrowly columnar to pyramidal growth habit, with obliquely ascending, fastigiate branches (Fig. 11A–D, G, H), a multitude of such densely leafy brachyblasts, spathulate pherophylls (Figs 9A, J–O, 11G, H), and petals copiously spotted with yellow oil glands (Fig. 9Q–R). The species also generally flowers later than most other species, with flowering tending to peak in January–February, and continuing as late as May.

Kunzea serotina was first recognised as distinct by W. Colenso who collected specimens in 1849 from the headwaters of the Makaroro River, eastern Ruahine Range (Colenso 1618, K, WELT SP022869—both mixed sheets, with the gatherings cited here being those labelled ‘A’) and who sent these to J.D. Hooker at Kew recommending their formal recognition with the suggested manuscript name “Leptospermum pulchrum”. Kunzea serotina was also recognised by Allan (1961) who described it briefly as a ‘thicket-forming variant’ confined to Marlborough, and it was mentioned by Harris (1996) who described it as ‘a small-leaved variant possibly worthy of taxonomic segregation’, and ‘occupying stations south of about Latitude 38°S’. As is often the case for New Zealand vascular plants (see Heenan 1998; Heenan et al. 2001; Heenan and de Lange 2004; Heenan and de Lange 2006) this species’ distinctiveness, so long overlooked by biosystematists, has long been appreciated by horticulturists who, in the North Island especially, have grown it widely, usually under the informal horticultural name Kunzea ‘Central North Island’ (R. Mains pers. comm.).

Throughout its range Kunzea serotina is sympatric with K. ericoides, K. robusta, and, in some thermal areas (notably around Lake Taupo), with K. tenuicaulis. In these areas despite the widespread sympatry, K. serotina mostly forms pure populations with little evidence of hybridism (see below).

Kunzea serotina is distinguished from K. robusta, by its columnar, pyramidal growth habit and singular bark-type (Fig. 11A–H). However, in very old specimens of K. serotina the columnar, pyramidal growth habit tends to be less marked or lost altogether because the lower branches thin with age, leaving trees with usually flat-topped branches crowded toward the top. In these situations, the bark–type will distinguish both species when it is not possible to observe foliage or flowers. Kunzea serotina, uniquely of the New Zealand species, has bark that is very easily detached, chartaceous, greyish-white to pinkish-white in colour, with highly irregular and sinuous margins (Fig. 11E–F). The bark is often found hanging partially detached as inrolled or curled up lengths within the branch forks, or it forms piles of curled up ‘wood shavings’ around the trunk base. In contrast, the bark of K. robusta has a distinctly coriaceous texture which detaches basally first then peels up the trunk forming long (up to 4 m) broad to narrowly tabular strips with ± smooth, ± entire margins.

The branchlet hairs of K. serotina consistently differ from those of most K. robusta populations by their smaller (up to 0.08 mm) divergent, apically curved or curled (Fig. 10B, C–E) rather than longer (up to 0.38 mm), weakly flexuose, antrorse-appressed hairs. However, in some parts of the eastern Central Volcanic Plateau the length of the branchlet hairs of K. robusta falls within the range cited for K. serotina, although in these situations the hairs are still antrorse rather than divergent. In the upper Rangitikei, distinction of seedlings and saplings of K. robusta from K. serotina can be difficult because in that area young plants of K. robusta usually have short divergent hairs of the same size range and form as K. serotina. However, adult trees of K. robusta from the same area (and those raised from these sites in cultivation) possess the more usual antrorse hairs. In these areas the growth habit of both species immediately distinguishes them. Floral and fruit characters also separate both species. In particular the spathulate to spathulate-orbicular pherophylls of Kunzea serotina (Fig. 9L–O) are unique among the New Zealand species and are a marked contrast to the narrowly deltoid, lanceolate, elliptic, oblanceolate to broadly oblanceolate, squamiform and/or foliose pherophylls typical of K. robusta. Further, the petals of K. serotina are uniquely dotted with yellow rather than the colourless or rose-pink oil-glands typical of other New Zealand species. Although the distributions of both species frequently overlap, K. serotina is ecologically distinctive, being a species of inland basins, upland river flats, frost flats, tussock grassland, ridge crests of mountain ranges, and subalpine scrub, favouring sites prone to extremes of climate and often with recent or skeletal soils. Kunzea robusta is much more generalist in its habitat preferences but as a rule it is more of a coastal, lowland to montane species favouring better developed soils, less extreme climates, and conditions that tend to suit the development of mixed lowland forest types. While K. serotina has been found at an upper altitudinal limit of 2000 m a.s.l. on Mt Ruapehu, K. robusta is rarely seen above 800 m a.s.l. Thus, the main areas of sympatry are generally places where K. serotina has colonised suitable lowland habitats abutting those dominated by K. robusta, e.g., the pseudo-karst ignimbrite country around and north of Lake Taupo toward Te Kuiti and Tokoroa, or the steep-sided river valleys of the Kaimanawa and Kaweka Ranges.

The growth habit of K. serotina (Fig. 11A–D) readily distinguishes it from the prostrate/decumbent shrub or multi-trunked widely spreading, flat-topped tree habit of K. tenuicaulis. In addition, both species can be distinguished by their branching pattern with the obliquely ascending, shortly fastigiate branches of K. serotina a marked contrast to the slender, widely spreading to pendulous branches of K. tenuicaulis. The pherophylls are also distinctive; those of K. tenuicaulis are mostly oblong, oblong-obovate to oblanceolate while those of K. serotina are characteristically spathulate. The calyx lobes of K. serotina are weakly keeled and flush with the rest of the hypanthium whereas in K. tenuicaulis they are distinctly thickened toward the base, and the external junction with the hypanthium is marked by a faint to prominent groove. The flower petals of K. serotina readily distinguish the species due to their being uniquely furnished with yellow rather than colourless oil glands. The fruit of both species is also diagnostic, with those of K. serotina mostly urceolate to campanulate (rarely cupular) (Fig. 9V), while the fruits of K. tenuicaulis tend to be barrel-shaped to cupular. Kunzea serotina and K. tenuicaulis are generally separated by their ecology, with the latter endemic to geothermal habitats. Sympatry occurs mostly in the vicinity of LakeTaupo where the ranges of both species overlap, and where on occasion K. serotina will colonise the more quiescent ground or cold ‘inliers’ of geothermal fields. Induced sympatry also occurs along the urbanised northern shoreline of Lake Taupo and at Wairakei, Karapiti and Tokaanu, where modification of the geothermal field for tourism has allowed K. serotina (and K. robusta) to more freely colonise habitats more usually occupied by K. tenuicaulis.

The columnar to pyramidal growth habit of Kunzea serotina readily separates it from Kunzea ericoides. Sometimes, as in shaded, damp or upland areas, the old bark of K. ericoides may have secondary peeling, some of which curls in a manner reminiscent of K. serotina. In these situations the foliage separates the species from each other. The leaves of K. ericoides are much longer (up to 25 mm long), mostly bright green and mostly linear, whereas K. serotina has distinctly shorter (up to 6.3 mm long), bronze-green to dark green, oblanceolate to obovate leaves. Both species have divergent branchlet hairs, but in K. ericoides they are sparse, shed early in branchlet maturation, smaller (up to 0.05 mm long) and have straight apices; in K. serotina, they are copious, persistent, longer (up to 0.08 mm long), with curved or curled apices. The flowers and fruits also differ. The petals of K. ericoides are usually sparsely dotted with ± colourless oil glands, and the fruits are slightly larger (up to 3.4 × 3.9 mm), mostly glabrous, typically 5-locular, and mostly cupular, barrel-shaped, shortly cylindrical or hemispherical. The petals of Kunzea serotina are copiously dotted with yellow oil glands and the fruits are smaller (up to 3.0 × 3.4 mm), finely hairy, 3–4 locular, and urceolate to shortly-campanulate (Fig. 9V).

Although the other ‘small-leaved’ Kunzea species, K. salterae and K. toelkenii, are allopatric from K. serotina, they could be confused in the herbarium. Kunzea serotina leaves (up to 6.3 × 1.8 mm) are smaller than the leaves of K. salterae (up to 18 × 2 mm), and in K. serotina they are oblanceolate to obovate rather than linear-lanceolate to narrowly oblanceolate. Kunzea serotina has spathulate pherophylls, and urceolate to campanulate, very rarely cupular, rather than cupular to sub-campanulate fruits. Herbarium material of K. serotina differs from K. toelkenii by its consistently divergent branchlet hairs rather than admixed large, antrorse-appressed, weakly flexuose, and small, divergent, curled hairs. Kunzea serotina also has smaller leaves (up to 6.3 × 1.8 mm) than K. toelkenii (up to 8.5 × 2.5 mm) and K. serotina lacks functionally male late season flowers. Other differences are summarised by Table 1.

Molecular evidence (rDNA ITS) grouped Kunzea serotina with K. ericoides, K. linearis, K. triregensis, K. robusta, K. sinclairii and K. tenuicaulis (Table 2; see also de Lange 2007; de Lange et al. 2010) while the rDNA ETS sequence of K. serotina shares a guanine/cytosine mix with K. ericoides, K. linearis, K. robusta (Mt Egmont samples only), K. salterae and K. toelkenii, and an adenine with K. salterae, K. tenuicaulis and K. toelkenii (Table 2; see also de Lange 2007; de Lange et al. 2010). Although there are no unique characters present in either of these marker regions, the inferred relationship of K. serotina with K. salterae, K. tenuicaulis and K. toelkenii, was already noted karyologically by de Lange and Murray (2004). Those data, coupled with the morphology of these ‘small-leaved’ species suggests that they are probably related and may form a natural group.

Kunzea serotina is an important component of the vegetation of the central North Island ranges. In the South Island K. serotina attains local prominence in the montane vegetation of the Marlborough, the northern Southern Alps, and the dry intermontane basins of north Canterbury and eastern Central Otago.

Kunzea serotina has the highest altitudinal limit of all the New Zealand Kunzea, frequently growing above 1000 m a.s.l., and reaching its maximum recorded elevation as stunted shrubs at 2000 m a.s.l. on steep, sparsely vegetated slopes growing amongst other alpine herbs and grasses in the upper Mangatoetoenui Stream, on the eastern flanks of Mt Ruapehu. This habitat is, however, rather unusual. More typically, in the North Island K. serotina grows along the ecotone between tall forest and tussock grassland, wetlands, frost flats or within grey scrub. In some places, most notably within the ignimbritic pseudo-karst of the north and eastern Volcanic Plateau, K. serotina forms a distinct, often ecotonal forest along the boundaries of the frost flats formed by numerous low aspect ratio ignimbritic eruptions, mostly sourced from the Taupo Volcanic Centre, (Houghton et al. 1995; Wilson et al. 1995). The South Island equivalent of this habitat seems to be the dry intermontane basins east of the main divide, especially those of north Canterbury, where remnant stands of K. serotina are still locally common. In these montane and inland basin habitats, K. serotina appears to have a long-term presence, and, if left undisturbed by fire, it probably would form the natural climax woody vegetation, particularly in areas prone to summer drought and extreme frosts and snow falls during winter.

Although Kunzea serotina is sympatric with K. ericoides, K. robusta and K. tenuicaulis. It appears to hybridise most frequently with K. robusta.

Hybrids with K. ericoides are easily recognised in the field and in herbaria because of the obvious differences in branching pattern, leaf colour, size and shape. This hybrid is probably uncommon because the distributions of K. ericoides and K. serotina rarely overlap. Putative K. ericoides × K. serotina has been been sparingly collected from Mt Dun, the Barnicoat Range and Hackett areas of eastern Nelson, and from the upper Buller and Sabine Rivers in southern Nelson are recognised. These are all areas where for various reasons the normally more montane and cold tolerant K. serotina extends outside its usual range.

Kunzea ericoides × K. serotina is easily recognised by the erect, somewhat open, sub-pyramidal growth habit, short, weakly fastigiate branchlets, and bright yellow-green, red-tinged leaves which are broadly lanceolate to linear-lanceolate. In most cases the branchlets tend to be distinctly hairy, though they can also be glabrescent. The inflorescences vary from corymbiform to widely spaced, elongate botrya. The pherophylls are deciduous and mostly pandurate to elliptic, and usually intermediate in length between either of the parents.

Recognition of the hybrid K. serotina × K. tenuicaulis is difficult, and examples of this hybrid have probably gone undetected during this study in herbaria (see K. tenuicaulis). The difficulty arises not so much in the field where the intermediate growth habit assists with hybrid recognition, but rather with herbarium material, from which hybridism can be difficult to infer if the accompanying notes are scant. Irrespective, it does seem that this hybrid is genuinely scarce, probably in part because it is only in the southern two-thirds of the Taupo Volcanic Centre that the ranges of either parent species overlap, and then mostly because human disturbance has induced novel habitats that both species (and the ubiquitous K. robusta) have utilised. Nevertheless, putative hybrids are represented by three herbarium specimens, from Karapiti (Craters of the Moon), Taupo on the shores of Lake Taupo, and at Tokaanu. Constant human activity at these places has allowed the ranges of these species to overlap and this disturbance has maintained the conditions for them. Beyond these modified habitats, K. serotina though frequently found near sites of geothermal activity appears to avoid colonising them. Even in modified geothermal habitats, K. serotina remains scarce, favouring local cold spots often on higher ground, well away from the main areas of thermal activity. However, because K. tenuicaulis frequently colonises cold spots within thermal areas, it is not unusual to find isolated shrubs or trees of K. serotina surrounded by and often partially smothered by K. tenuicaulis growth. Under such conditions it is surprising that hybridism between these two species seems to be so uncommon, especially as the flowering times of both species overlap. The scarcity of hybrids possibly reflects something deeper. In the experimental cross K. serotina^♀ × K. tenuicaulis^♂, and its reciprocal, the progeny all had 2n = 23 chromosomes (de Lange and Murray 2004; de Lange et al. 2005). Few of these plants flowered. Pollen stainability from these plants was markedly reduced (12–23%). These results suggest that there might be some reproductive barrier in place which could explain the scarcity of this hybrid in the wild. Further investigation into this is needed.

The most commonly encountered hybrid is Kunzea robusta × K. serotina because the ranges of both species frequently overlap. This is particularly the case in the North Island, around the northern Central Volcanic Plateau, southern Kaweka Ranges, and upper Rangitikei catchment. Hybrid swarms are common in these places, especially in grossly disturbed habitats such as reverting farmland or roadsides. Field recognition of hybrids is aided by the effect of the distinctive growth habit of K. serotina, which results in somewhat openly branched, sub-pyramidal to widely spreading, sparingly branched, small spindly trees. The branchlets may be fastigiate or spreading but are usually shorter than in K. robusta. The distinctive bark of K. serotina is often present to some degree, though it tends to peel off in long tabular strips like K. robusta but with much secondary peeling. It is the secondary peels that tend to curl up as K. serotina bark normally does. Branchlet hairs are particularly diagnostic because K. robusta has antrorse-appressed and K. serotina has divergent hairs, with hybrids showing obvious mixtures of both. The pherophylls of K. robusta, though very variable in shape and length, are mostly deltoid to lanceolate, and are usually present in mixtures of foliose and squamiform types. In K. serotina the pherophylls are usually foliose and spathulate, spathulate-orbicular, only rarely pandurate or lanceolate. Thus, hybrids tend to show mixtures of sparse deciduous, pandurate, and lanceolate foliose and frequent squamiform pherophylls.

Artificial hybrids involving K. serotina as staminate or pistillate parent and other New Zealand members of the K. ericoides complex were easily produced and showed no obvious reduction in fertility except for crosses involving K. tenuicaulis, which were effectively sterile (see de Lange and Murray 2004; de Lange et al. 2005).

Kunzea serotina is (or was) locally known to Maori inhabiting the Central North Island as ‘makahikatoa’. Makahikatoa is said to mean ‘white kahikatoa’ referring to the wood colour, and is a name intended to distinguish white-wooded K. serotina from the red-wooded ‘warrior-wood’ of kahikatoa (Leptospermum scoparium) (W. Kawhaki pers. comm.). A few herbarium gatherings made during the 1800s also refer to this species as ‘manuka’.

Kunzea serotina is an abundant and widespread species which should not be regarded as threatened or at risk (a more detailed account of its conservation status, using the criteria of Townsend et al. (2008) at both national and regional levels is provided by de Lange (2007).

A K. ericoide habitu fruticis decumbentis ad serpentis vel arboris erectae sed patentis multicaulis ad pendentibus, caulibus et ramibus multis tenuibus vel ramulibus numerosis patentibus ad pendentibus tenuissimis, ramulibus juvenibus dense tomentosis pilis multis brevibus patentibus, foliis brevioribus oblanceolatis ad obovatis; hypanthio minore cupuliformi campanulato gracile puberulenti; lobis calicis distincte incrassatis basipetale et sulco ad basim, fructibusque minoribis doliiformibus differt. Chromosomatibus constanter parvis et equatis seriebusque singularibus rDNA ITS et ETS differt.

(Figs 13, 14, 15)(Spread over three sheets).New Zealand: North Island, Central Volcanic Plateau, Paeroa Range, Te Kopia Thermal Area, 38°24'S, 176°13'E, 420m a.s.l. ‘Dominant shrub and small tree on active geothermal field. Abundant around vents. Growing with Leucopogon fasciculatus, Leptospermum scoparium and Dracophyllum subulatum, Seedling to adult collections from same site, over 3 sheets (AK 288088!, AK 288171!, AK 288172!)’. P. J. de Lange 4702A, B, C 16 Nov 2000, AK 288172 (Adult branch in bud, and branch with flowers and buds), AK 288088 (three seedlings), and AK 288171 (One sapling). Isotypi: AD, CHR.

Figure 13. 

Holotype of Kunzea tenuicaulis de Lange (P. J. de Lange 4702A, AK 288172).

Figure 14. 

Holotype of Kunzea tenuicaulis de Lange (P. J. de Lange 4702B, AK 288088).

Figure 15. 

Holotype of Kunzea tenuicaulis de Lange (P. J. de Lange 4702C, AK 288171).

The holotype gathering of K. tenuicaulis comprises six specimens spread over three sheets and lodged at the same herbarium. The two adult specimens mounted on AK 288172, de Lange 4702A, come from the same plant, while the three seedlings (AK 288088, de Lange 4702B) and sapling (AK 288171, de Lange 4702C) were collected from the ground directly beneath that plant.

The manner in which I have collected and designated these sheets as holotype is in accordance with the International Code of Nomenclature (McNeill et al. 2012) Article 8.2 ‘….the holotype may consist of a single plant, parts of one or several plants, or of multiple plants…’ Thus in accordance with Article 8.3, Example 4 of the International Code of Nomenclature (McNeill et al. 2012), my collecting numbers reflects my intended association of these gatherings as type.

Kunzea tenuicaulis most probably equates with Kunzea ericoides var. microflora (G.Simpson) W.Harris which is based on Leptospermum ericoides var. microflorum G.Simpson (Simpson 1945). This variety was described from garden specimens sent to George Simpson by Norman Potts of Opotiki, and which, according to Simpson, came from ‘Rainbow Mountain, Nelson’ where its habitat was said to be the ‘Mineral Belt, Nelson’. As noted by Allan (1961; pp. 322–323), no plants matching Simpson’s description occur in the Nelson area or, indeed, the rest of the South Island. Furthermore, there is no Rainbow Mountain in the South Island. However, plants similar to Simpson’s description have been found in active geothermal areas within the Taupo Volcanic Zone of the North Island from a geothermal site near Lake Rotoiti (Tikitere) in the north west and Kawerau in north east to just south of Tokaanu, near Turangi at the southern end of Lake Taupo. These are the plants referred here to Kunzea tenuicaulis.

Allan (1961; p. 322) had also recognised this, and sought to rectify what he considered to be a genuine geographic mistake in Simpson’s protologue, by his statement that Simpson’s type locality was in fact ‘Rainbow Mountain near Waiotapu’. However, because Allan’s opinion cannot be matched with any supporting statement or evidence from Simpson, we simply cannot be certain where Simpson’s type (based on Norman Pott’s garden specimens) really came from. Further, the only specimen that I can unequivocally say is labelled by Simpson as Leptospermum ericoides var. microflorum is in extremely poor condition (AK 22886; Fig. 16A, B). The all-important diagnostic branchlet hairs are scant, there are only eight flowers left that are so shrivelled and damaged as to make a proper reconstruction impossible, and most of the foliage is missing. While I am convinced that Simpson’s type matches what I have named K. tenuicaulis, I have elected to remove any further ambiguity by redescribing the Kunzea endemic to the geothermal fields of the Taupo Volcanic Zone at the rank of species, with a full and detailed description, and an unambiguous wild type and duplicates gathered from a wild legally protected site where the species is uniform and widespread, and hybridism is not evident.

Figure 16. 

Lectotype of Leptospermum ericoides var. microflorum G.Simpson (N. Potts s.n., AK 22886). A Specimen B George Simpson’s handwriting on piece of newspaper mounted on AK 22886 over which two AK herbarium labels mounted.

The specific epithet tenuicaulis refers to the numerous very fine and slender branchlets produced by this species, irrespective of its overall growth habit.

= Leptospermum ericoides var. microflorum G.Simpson in T.R.S.N.Z. 75, (1945), 189.

≡ Kunzea ericoides var. microflora (G.Simpson) W.Harris in N.Z.J.Bot. 25, (1987), 134.

(here designated)(Figs 16, 17). “Leptospermum ericoides var. microflorum, [R]ainbow Mt, Potts” AK 22886! (label in George Simpson’s hand, written in pencil on back of piece of printed paper (possibly newspaper) (Fig. 16B)).

Figure 17. 

Lectotype of Leptospermum ericoides var. microflora G.Simpson (N. Potts s.n., CHR 48079) as selected by Allan (1961). This sheet has no nomenclatural status as it cannot unequivocally be shown to have been handled by the naming author George Simpson. The label details are written in the hand of T. Rawson then technician to H. H. Allan (P. B. Heenan pers. comm.).

Leptospermum ericoides var. microflorum was briefly described by Simpson (1945) and for most of his protologue his intent is clear. Simpson’s type collections were based on a specimen that had been cultivated by Mr N. Potts of Opotiki. Potts’s plant was said to have come from Rainbow Mountain (Simpson 1945). The fact that

Potts, a North Island botanist, wrote to Simpson stating “the plant occupies craters on Rainbow Mountain” strongly suggests that his plant had come from Rainbow (Maungakakaramea) Mountain (38°19'S, 176°23'E), an old volcanic vent and active geothermal field just to the north east of the small settlement of Waiotapu on the western marches of the Kaingaroa State Forest in the North Island.

Despite this, Simpson explicitly states ‘Habitat: Mineral belt, Nelson’ (a South Island location), and further ‘Specimens from a plant in cultivation by Mr N. Potts’ garden at Opotiki, collected at Rainbow Mountain, Nelson’. Subsequently, Allan (1961; p. 322) stated that the variety is ‘known with certainty only from the type locality: Rainbow Mountain near Waiotapu’ and that Simpson has ‘inadvertently given the locality as “Mineral belt, Nelson”’. This, while probably true (see above) is not strictly correct, as Simpson twice cited Nelson as the source of the plants, and as far as I can determine never clarified the apparent ambiguity in subsequent literature or correspondence with Allan or anyone else—though the handwriting on CHR 48079, ‘Note published as Mineral Belt by mistake. cf. Miss L.B.Moore’ suggests that L. B. Moore might have obtained some comment from someone about the original source of Potts’s specimens. Irrespective, based on available evidence we cannot be sure of what Simpson’s statements about the source of his type material really meant.

Simpson (1945) also explicitly states that the ‘specimens’ (my emphasis) on which he based his name were lodged in the ‘Herbarium Plant Research Bureau, Wellington’ then known as BD (Botany Division) and now CHR. Allan (1961; p. 322) lectotypified the name (from a specimen in the former Botany Division Herbarium that he evidently considered was part of the type gathering) in the following manner ‘Type: BD 48079, from a plant cultivated by N. Potts’. That specimen (Fig. 17), now CHR 48079, is clearly labelled ‘TYPE’ in red ink, in what is probably the hand of T. Rawson, then Technician to Allan (P. B. Heenan pers. comm.). However none of the associated handwriting on the label is in Simpson’s hand (see Heenan 1995) and, beyond the filing of the specimen in the ‘Simpson Herbarium’ there is nothing to clearly identify this gathering as one that Simpson had actually handled, let alone anything to show that Potts gathered the specimen. For these reasons I reject Allan’s typification.

In AK I have located a further gathering attributed to Simpson, AK 22886, bearing several labels (Fig. 16A, B). The main label is an official one of the type used at AK between 1929 and about the early 1970s (E. K. Cameron pers. comm.) and is labelled in blue ink ‘Rainbow Mt., Nelson, N. Potts’ and ‘cotype’ in handwriting that is most likely that of the herbarium curator at AK toward the end of the 1940s, B. E. Molesworth. Beneath that label, on the back of a small scrap of what appears to be newspaper, is a pencil annotation in Simpson’s distinctive handwriting (Fig. 16B; see also Heenan 1995), ‘Leptospermum ericoides var. microflorum …[R]ainbow Mt…Potts’. As the official AK label was inadvertently glued over part of Simpson’s handwriting, what I regard as the letter ‘R’ has been partially erased by glue and dirt and is thus no longer fully legible (Fig. 16B). There is no doubt in my mind however, that the locality on the label is Rainbow Mountain.

Although Simpson makes it clear that specimens of his new variety were to be found in what is now the Allan Herbarium (CHR), I cannot now find any specimens there that unequivocally show this, and yet, as Simpson indicates that he had lodged ‘specimens’, I feel it unwise to regard the AK specimen as a holotype because other collections may exist. For this reason I select AK 22886 as lectotype. I consider CHR 48079 as having no nomenclatural status because there is no evidence that Simpson ever handled or in anyway used it to describe his variety.

The varietal epithet ‘microflorum’ was adopted by Simpson (1945) in the mistaken belief that his Leptospermum ericoides var. microflorum had smaller flowers than the type variety.

(Figs 18, 19, 20).Growth habit decumbent, trailing subshrubs, shrubs or small trees 0.1–6.0(–8.0) × 2.0–6.0(–8.0) m. For specimens with a tree habit, crown widely spreading, often arching to somewhat pendulous. For specimens found around active fumaroles or on open, geothermally heated ground, growth habit varying from completely decumbent and densely branched, with stems sprawling across ground, to semi-erect, densely branched, widely spreading, often pendulous. Trunk in tree forms (1–)4–6 arising from base, 0.1–0.6 m d.b.h., these branching from close to base, with branches thinning in close canopies only; in decumbent plants trunk virtually indistinguishable, 0.01–0.10 m diam., trailing to semi-erect, curved and somewhat sinuous, in erect plants at first erect, soon widely spreading and curving to somewhat sinuous. Bark early bark greyish brown to brown, initially firm, somewhat sinuous-fluted, elongate, over time cracking transversely (especially on branch flanges and decurrent leaf bases), and with margins gradually detaching and rolling-in to present as easily detached, papery, narrowly short to long, somewhat irregular-margined flakes; old bark grey-brown to grey, chartaceous to mildly corky, flaking readily in short to long, usually narrow and slightly sinuous to irregular, tabular shards, these usually remaining attached in several places with the spaces between detached, cracked and more or less raised, upper bark surface often with much secondary peeling and transverse cracking, crumbling in hand easily. Branches numerous, rather narrow and long, often weakly flexuose, in decumbent plants prostrate, trailing, otherwise initially ascending, soon suberect to widely spreading, and arching, often pendulous; branchlets numerous, very leafy, rather slender, quadrangular, sericeous, with dense, silky indumentum; hairs persistent, divergent, weakly flexuose, 0.03–0.06(–0.08) mm, hyaline to translucent (appearing white when young maturing grey), hair apices more or less straight. Vegetative buds inconspicuous, usually obscured by surrounding foliage; at resting stage 0.5–1.0(–1.6) mm diam.; scales scarious, deciduous, (0.3–)0.8(–1.3) mm long, red-brown to dark brown, initially broadly ovate grading through to broadly lanceolate; midrib prominent, strongly keeled, prolonged to cuspidate tip, with 1–2 lateral veins either side, and two prominent rows of 3–8 oil glands straddling midrib, margins and keel apex ciliate. Leaves heterophyllous, seedling and subadult leaves flat or involute, ± spreading to recurved; 0.9–3.0(–4.5) × 0.2–0.4(–0.6) mm, red-green or pale green suffused with red, rarely bright green; lamina finely linear-lanceolate, long persistent in stressed habitats (in damaged plants reversion shoots bearing juvenile foliage frequent); adult leaves ± spreading to patent; lamina (1.1–)4.0(–10.0) × (0.8–)1.3(–2.8) mm, dark glossy green, red-green, to bronze-green, narrowly oblanceolate, oblanceolate, obovate to obovate-rostrate; usually recurved from about half of total length, apex usually obtuse, rounded, rarely subacute, cuspidate; base attenuate; adaxial surface convex, finely glandular punctate; oil glands up to 590, more evident when dry, midrib slightly raised to depressed near base, otherwise depressed for entire length, glabrous, very rarely with fine antrorse hairs near base; abaxial surface slightly concave, finely glandular punctate, oil glands less obvious, up to 280, these more evident when dry; midrib depressed, finely and sparsely covered with sericeous, deciduous, antrorse-appressed hairs, these increasing in density toward base; lamina margin sparsely to densely, finely sericeous, hairy; hairs weakly flexuose, appressed to weakly spreading, antrorse to subantrorse, up to 0.1 mm, hyaline to translucent, appearing as white to naked eye, aligned in 1 row not quite meeting at cuspidate leaf apex. Perules scarious, basal ones usually persistent, these 0.4–1.0 mm long, pale brown to brown, broadly oblong to oblong-lanceolate, margins involute especially in upper third, midrib strongly keeled, prolonged as a cuspidate apex, with one row of 4–8 oil glands on each side of midrib, glabrous except for finely ciliate margin and apex; remaining perules deciduous, chartaceous, (0.6–)0.8(–1.4) mm long, pink to pinkish-white when fresh, drying apricot to apricot-brown, ovate to broadly oval, apex obtuse often appearing acute due to apical infolding, ± cuspidate, glabrous except for sparsely ciliate margin, strongly keeled, keel ± prolonged. Inflorescence usually a compact, (1–)6(–10)-flowered corymbiform botryum up to 25 mm long, borne on alternate brachyblasts up to 15 mm long, with those near branchlet apex usually subopposite; inflorescences at the ultimate branchlet tips rarely elongated, in which case these are invariably surmounted with terminal vegetative growth. Inflorescence axis densely invested with divergent hairs. Pherophylls deciduous (falling very early), tightly clasping pedicels to ± spreading, 0.5–1.0 mm long, initially foliose soon squamiform; foliose pherophylls pale green, oblong, oblong-obovate to oblanceolate, margins and apex finely ciliate; squamiform pherophylls brown or pink, drying apricot-brown or amber, broadly deltoid to oblong-ovate, margins involute especially in upper one-third, midrib strongly keeled, prolonged as cuspidate apex, with one row of 4–8 oil glands on each side; glabrous except for the finely ciliate margin and apex; similar to perules in size and shape at apex. Pedicels (1.0–)2.1(–2.4) mm long at anthesis, elongating slightly after anthesis, terete, copiously invested in slightly flexuose, antrorse to subantrorse sericeous hairs. Flower buds clavate to pyriform, apex distinctly domed (due to thickened calyx lobes) prior to bud burst with calyx valves ± meeting. Fresh flowers when fully expanded (3.3–)5.5(–9.0) mm diam. Hypanthium (1.8–)2.5(–3.3) × (1.7–)2.4(–3.1) mm, with free portion 0.3–0.8(–1.0) mm long, dark green often basally mottled or flushed with red when fresh, drying brown to grey; narrowly cupular to campanulate terminating in a slightly thicker rim bearing five persistent calyx lobes; surface smooth, finely gland-dotted, and puberulent, with weakly defined ridges leading up to calyx lobes (these becoming more distinct upon drying); hairs shortly subantrorse to antrorse. Calyx lobes 5, upright (not spreading), firmly fleshy, 0.4(–0.8) × 0.4(–1.0) mm, persistent, oblong, oblong-ovate to broadly triangular, in longitudinal-section distinctly thicker at base, ± subtended by a faint to prominent groove at the external junction with the hypanthium, otherwise tapering to apex, scarcely keeled (the keel if evident recognisable as a darker green or pink, thicker central prolongation of the hypanthium ridges), margins cream to pale pink, gland-dotted, oil glands usually colourless sometimes pink; otherwise glabrate except for ciliate margins; cilia widely spreading. Receptacle green or pale pink at anthesis, darkening to crimson-red or magenta after fertilisation. Petals 5(–6), 1.4–1.6(–2.0) × 1.4–1.6(–2.0) mm, white or pinkish white, usually basally flushed pink, very rarely completely pink, orbicular, sometimes cuneate, apex obtuse to rotund, margins plane or finely crimped 3–12 times, oil glands not evident when fresh, drying colourless. Stamens 10–24(–32) in 1(–2) weakly defined whorls, arising from receptacular rim, filaments white often tinged rose-pink toward base. Antipetalous stamens 2(–3), antisepalous 1(–4). Outermost antipetalous stamens usually weakly to strongly incurved, on filaments 0.9–2.2 mm long; inner stamen, if present, 0.6–0.8 mm, strongly incurved; very rarely a further 1–2 strongly incurved stamens, 0.4–0.7 mm long, may be present at the base of the outermost antipetalous pair. Antisepalous stamens much shorter than outermost antipetalous stamens, 0.3–0.8 mm, incurved, rarely outcurved or in mixtures of both. Anthers dorsifixed, 0.04–0.08 × 0.02–0.04 mm, testiculate, latrorse. Pollen white (12.8–)14.7(–16.6) μm. Anther connective gland prominent, orange when fresh, drying pale brown, spheroidal, distinctly papillate. Ovary (3–)4(–5) locular, each with 15–18(–22) ovules in two rows on each placental lobe. Style 2.0–2.6(–3.6) mm long at anthesis, often elongating slightly after anthesis, white basally flushed with pink; stigma capitate, scarcely wider than style, domed along margins with a central depression, pale cream to pink, surface papillate to distinctly rugulose. Fruits ± persistent, (1.0–)2.3(–3.3) × (1.6–)2.2(–3.2) mm, light brown to grey, usually barrel-shaped, rarely cupular, splits concealed by dried, suberect to erect, free portion of hypanthium. Seeds 0.80–1.00 × 0.45–0.50 mm, narrowly oblong, oblong, oblong-obovate to falcate-oblong, curved near apex, laterally compressed, 2–3-angled with convex to flattened faces, apex rounded, base cuneate to oblique, ± flattened; testa semi-glossy, orange-brown, surface coarsely reticulate. FL: (Aug–)Sep–Oct(–Mar) FT: Jan–May(–Nov). Chromosome Number n = 11_II, 2n = 22 (de Lange and Murray 2004).

Figure 18. 

Distinguishing features of Kunzea tenuicaulis. A Flowering branchlet (ex cult. AK 284554) B Fruiting branchlet (ex cult. AK 284554) C Vegetative bud and branchlet indumentum (ex cult. AK 284554) D Seedling (no voucher, self sown from AK 284554) E Adaxial leaf surface (ex cult. AK 284554) F Abaxial leaf surface (ex cult. AK 284554) G Leaf variation within the same individual (ex cult. AK 284554) H Adaxial leaf apex (ex cult. AK 284554) I Leaf margin indumentum (ex cult. AK 284554) J Flower (top view) (ex cult. AK 284554) K Flower and hypanthium (side view) (ex cult. AK 284554) L Flower cross section showing anther, style and ovules (ex cult. AK 284554) M Style and stigma (ex cult. AK 282217) N Stamens (ex cult. AK 282217) O Dehisced fruit (ex cult. AK 282217). Scale bars: (A, B, D, G) 10 mm; (C, E, F, J–O) 1 mm; (I) 0.5 mm.

Figure 19. 

Scanning Electron Micrographs of Kunzea tenuicaulis. A–F all AK 288105) Branchlet indumentum G–J Seeds (AK 286159, AK 288105). Scale bars: (A, G) 1 mm; (B–F, H–J) 100 μm.

Figure 20. 

Kunzea tenuicaulis. A K. tenuicaulis habitat, North Island, Karapiti (Craters of the Moon) (photo: P. J. de Lange) B K. tenuicaulis habitat, North Island, Maungakakaramea (Rainbow Mountain) Scenic Reserve, Crater area (photo: P. B. Cashmore) C K. tenuicaulis decumbent form on heated ground, Maungakakaramea (Rainbow Mountain) Scenic Reserve, Crater area (photo: P. B. Cashmore) D K. tenuicaulis tree form showing multi-trunked growth habit and widely spreading, narrow branchlets, North Island, Rotorua, Kuiarau Park (photo: P. B. Cashmore) E K. tenuicaulis tree form showing pendulous growth habit, North Island, Rotorua, Kuiarau Park (photo: P. B. Cashmore) F K. tenuicaulis tree form showing multi-trunked, widely spreading, pendulous growth habit, North Island, Rotorua, Kuiarau Park (photo: P. B. Cashmore) G K. tenuicaulis showing characteristic, flexuose trunks and widely spreading branches; North Island, Tokaanu, Tokaanu Geothermal Reserve, (photo: P. J. de Lange) H K. tenuicaulis trunk and bark, North Island, Tokaanu, Tokaanu Geothermal Reserve, (photo: P. J. de Lange) I K. tenuicaulis trunk and bark, North Island, Paeroa Range, Te Kopia Geothermal Reserve (photo: P. J. de Lange); (J) K. tenuicaulis branches showing distinctive widely spreading, fine, pendulous branchlets, North Island, Tikitere (Hell’s Gate) Thermal Park, (photo: P. J. de Lange) K K. tenuicaulis flowering branchlet showing, compact corymbiform botrya, North Island, Waiotapu Geothermal Park (photo: G. M. Crowcroft).

(80 sheets seen).New Zealand (North Island). Kawerau, Ruruanga Stream (Parimahana Geothermal Field), P. J. de Lange 4628, 7 Nov 2000, (K 288085); Tikitere (Hell’s Gate) Geothermal Field, Upper Waiohewa Stream, P. J. de Lange 4628, 7 Nov 2000, (AK 288086); Rotorua, Kuirau Park, P. J. de Lange 4627, 7 Nov 2000, (AK 286156); Whakarewarewa Park, L. Cockayne s.n., 29 Dec 1905, (WELT SP029450); Waimangu Thermal Area, near Lake Rotomahana, R. J. Chinnock s.n., 17 Oct 1967, (WELTU 9731); Maungakakaramea (Rainbow Mountain), P. J. de Lange 4223, 26 Jan 2000, (AK 286186); Te Kopia Geothermal Area, P. J. de Lange 4700, 16 Nov 2000, (AK 288099, Duplicates: AD, MEL, MO); Paeroa Range, Waikite Geothermal Reserve, P. J. de Lange 4713 & R. O. Gardner, 19 Nov 2000, (AK 286168, Duplicates: AD, WELT); Te Kopia – Waihunuhunu Road, Waihunuhunu Stream, P. J. de Lange 4699, 16 Nov 2000, (AK 288087, Duplicates: AD, FI, HO, MSC, P); Waikato River, Lake Ohakuri, Orakeikorako, P. J. de Lange 4693, 16 Nov 2000, (AK 286170); Waikato River, Wairakei, P. J. de Lange 4683, 10 Nov 2000, (AK 288084, Duplicate: AD); Wairakei Geothermal Field, Karapiti, Craters of the Moon, P. J. de Lange 5765, 10 Nov 2003, (AK 286152, Duplicates: CANB, CANU, MSC, NSW, Z); Lake Taupo, The Spa, D. Petrie s.n., Dec 1895, (WELT SP029561); Tokaanu, Tokaanu Geothermal Reserve, P. J. de Lange 4582, 19 Oct 2000, (AK 288103, Duplicates: AD, HO, MEL, WELT).

(Fig. 7). Endemic, New Zealand, North Island, Bay of Plenty to the Central North Island (40–580 m a.s.l.). Confined to active geothermal fields (i.e. those with surface expression) of the Taupo Volcanic Zone (for geology see Healy 1992; Houghton et al. 1995; Wilson et al. 1995; Neall 2001) from the vicinity of Kawerau (Parimahana Geothermal Field) and Lake Rotoiti (Tikitere) south to Tokaanu and the hills above Waaihi, Lake Taupo (Figs 20A–C, 21).

Kunzea tenuicaulis is recognised by a combination of growth habit, branchlet hair and floral characters (Figs 19, 20; see also Table 1) supplemented by cytological and molecular differences. The ITS and ETS sequence data (Table 2) show that K. tenuicaulis is the most diverged of the New Zealand members of K. ericoides complex (de Lange 2007; de Lange et al. 2010). Uniquely within the K. ericoides complex, the ITS sequence of K. tenuicaulis possesses two adenine nucleotides rather than the guanine common to all other members of the complex at ITS-1 alignment position 639, and at ITS-2 alignment position 994 (Table 2; see also de Lange 2007). Otherwise it shares a guanine/thiamine mix with K. salterae (de Lange 2007). The ETS sequence (Table 2) showed two further unique characters; a thiamine at alignment position 18 (whereas all other members of the complex possess a cytosine), and a cytosine at alignment position 202 (whereas all other members of the complex possess an adenine) (de Lange 2007). Otherwise the aligned ETS sequence of Kunzea tenuicaulis has an adenine at position 269 in common with the other ‘small-leaved’ New Zealand Kunzea, K. salterae, K. serotina and K. toelkenii (de Lange 2007). In view of the geologically recent (estimated to be a maximum of 2 million years old (Neall 2001; Briggs et al. 2005)) habitats this species occupies, this molecular divergence from all other members of the K. ericoides complex is considered remarkable (de Lange 2007; de Lange et al. 2010).

Kunzea tenuicaulis seems to be most closely allied to K. salterae and K. serotina, and, based on the results obtained from experimental hybridisations (de Lange et al. 2005), K. tenuicaulis may have had a role in the evolution of K. salterae and K. toelkenii through hybridisation with K. linearis and K. robusta. With the exception of K. salterae which has linear-lanceolate leaves, K. serotina, K. tenuicaulis and K. toelkenii all possess small oblanceolate to obovate leaves. Branchlet hairs in all four species tend to be copious, short (up to 0.08 mm), divergent, and persistent (Table 1). Further, as reported by de Lange and Murray (2004) and de Lange et al. (2005) K. tenuicaulis shares with K. serotina and K. toelkenii uniformly small chromosomes (0.9–1.0 μm). Kunzea salterae, has a similar chromosome complement, though this was not reported previously because that species had not yet been recognised when those papers were published. Despite this Genomic In Situ Hybridisation experiments determined that, alone of those taxa analysed, K. tenuicaulis has the most diverged genome (de Lange et al. 2005).

As circumscribed here, K. tenuicaulis includes a range of prostrate to erect plants found in close proximity to active geothermal vents and fields (Fig. 20D–G, J). Although in the past, (probably because of the past poor circumscription of this plant) field workers had formed the impression that Kunzea ericoides var. microflorum applied only to the flat, prostrate to small decumbent shrubs found near active geothermal vents (e.g., Harris et al. 1992; Harris 1996; Smale 1994). There seems little point in trying to separate these prostrate plants from the erect plants they resemble in all respects except stature. Indeed, stature itself often presents as a gradient from thermally heated to thermally quiescent ground. Cultivation experiments showed that while some seedlings raised from seed sampled from prostrate/decumbent plants retained that growth habit, the majority grew into multi-trunked erect small trees with flat topped, spreading to pendulous crowns typical of K. tenuicaulis as defined here. Irrespective of growth habit, all forms of K. tenuicaulis are consistently unified by their many, fine, slender branchlets; copious, short, divergent branchlet hairs (Fig. 19A–F); small oblanceolate to obovate leaves (Fig. 18C–G); small cupular to campanulate, finely puberulent hypanthium (Fig. 18K–L); by the calyx lobes which are thickened toward the base and there subtended by a faint to prominent groove along the external junction with the hypanthium; and by the smaller, barrel-shaped fruit (Fig. 18O). Furthermore the ITS and ETS sequence data obtained from multiple samples spanning the range and variation of this species were consistent, and readily distinguished K. tenuicaulis from the rest of the K. ericoides complex (de Lange 2007; de Lange et al. 2005).

Provided that care is taken to note the growth habit, collect old bark, new season’s growth, and emergent flowers K. tenuicaulis is easily separated from the other New Zealand Kunzea. This is important for although K. tenuicaulis is distinguished ecologically because it is endemic to active geothermal fields (Figs 20A–C, 21), within these habitats it may be found sympatric (or even syntopic) on associated ‘cool’ sites with K. serotina and K. robusta.

Kunzea tenuicaulis can be distinguished from K. serotina by its growth habit, which is either prostrate/decumbent or multi-trunked, with widely spreading pendulous, mostly flat-topped, branches, producing numerous spreading, slender, long branchlets (Fig. 20D–G, J; Table 1). This contrasts with the strictly upright columnar to pyramidal growth habit, with short, obliquely ascending, fastigiate branches of K. serotina. The bark of both species is also diagnostic. The old bark of K. tenuicaulis is grey-brown to brown, readily detached, somewhat corky-chartaceous, flaking in narrow, short to long, slightly sinuous to irregular, tabular shards (Fig. 20H–I; Table 1). The old bark of K. serotina is greyish-white to pinkish-white, and presents as inrolled and curled ‘wood shavings’, these having little if any discernible shape, with highly irregular, sinuous, often frayed margins. Although both species have similar leaves, in K. serotina the leaves are less widely spaced, and tend to be densely clustered around the branchlets. Branchlet hairs however are not overly diagnostic as both species have similarly sized divergent hairs. In general though, those of K. serotina tend to have more curly apices and K. tenuicaulis less so (Fig. 19C–F). The pherophylls, if present, serve to separate both species: those of K. tenuicaulis are mostly foliose oblong, oblong-obovate to oblanceolate while those of K. serotina are characteristically spathulate to spathulate-orbicular. The calyx lobes of K. tenuicaulis are especially diagnostic when fresh because they are distinctly thickened toward the base below which is a faint to prominent groove at the external junction with the hypanthium. Those of K. serotina are slightly keeled and flush with the rest of hypanthium. The petals of K. tenuicaulis are broadly orbicular and have colourless oil glands (often not evident until the petals have dried) while those of K. serotina are narrowly orbicular to broadly ovate and typically have pale yellow oil glands. The mature fruits of K. tenuicaulis tend to be barrel-shaped to cupular, those of K. serotina urceolate to campanulate or, rarely, cupular (Table 1).

Kunzea tenuicaulis is distinguished from K. robusta by its smaller size (up to 8 m cf. up to 30 m in K. robusta) and growth habit (see Table 1). Kunzea robusta is mostly an arborescent species, and so usually forms a single-trunked tall tree, with a broad trunk, stout, ascending to spreading branches, and a very wide, spreading, multi-tiered canopy. However, occasionally K. robusta can be a low (up to 2 m tall) shrub with prostrate to pendulous branches, or be a tree with entirely pendulous branches, in which case leaf size, shape and branchlet hair type serve to distinguish it from K. tenuicaulis. The bark of K. robusta is particularly distinctive, being very coriaceous, long persistent, typically detaching with age as long (up to 4 m), broad to narrowly tabular strips, with ± smooth, ± entire margins, which when deliberately bent and snapped have ± entire or weakly frayed broken surfaces. In contrast, the bark of K. tenuicaulis is not long persistent, readily detaches, and is distinctly corky-chartaceous, flaking as rather narrow, much shorter (up to 100 mm long) tabular shards with slightly sinuous to irregular margins. The oblanceolate to lanceolate leaves of K. robusta, up to 20.1 × 3.0 mm, are usually much larger and broader than the leaves of K. tenuicaulis which grow to 4.5 × 0.6 mm. A key difference between these species is the branchlet hairs. In K. robusta populations that occur within the range of K. tenuicaulis, branchlet hairs are mostly antrorse-appressed, larger (up to 0.38 mm) and straight to weakly flexuose. Kunzea robusta also has larger inflorescences containing more flowers (up to 30, more usually 12) than K. tenuicaulis. The inflorescences of K. robusta typically progress from a compact corymbiform botryum at the onset of flowering to an elongated botryum as the flowering season progresses due to activation of the apical vegetative bud. In K. tenuicaulis this very rarely happens, and then only to the terminal inflorescence, those of the brachyblasts tending to remain as compact corymbiform botrya bearing far fewer (up to 10, more usually six) flowers (Figs 18A, 20K). The flowers of K. robusta tend to have a greater overall diameter (up to 12.0 mm, more usually 7.7 mm cf. up to 9.9 mm, more usually 5.5 mm) and in the field more stamens (up to 60 cf. up to 32 in K. tenuicaulis) (see Table 1). The fruits of both species are also rather different, those of K. tenuicaulis tending to be smaller (up to 3.3 × 3.2 mm )and barrel-shaped to cupular (Fig. 18B, O), while those of K. robusta are much larger (up to 4.6 × 5.3 mm), and mostly obconic, broadly obconic to ± turbinate (see Table 1).

Although Kunzea salterae and K. toelkenii are allopatric from K. tenuicaulis, they could be confused in the herbarium. The oblanceolate to obovate leaves of K. tenuicaulis, which grow up to 4.5 × 0.6 mm, are smaller than the 18 × 2 mm, linear-lanceolate to narrowly oblanceolate leaves of K. salterae. Furthermore, the fruits of K. tenuicaulis are barrel-shaped to cupular rather than cupular to subcampanulate (see Table 1). The consistently divergent branchlet hairs of K. tenuicaulis are distinct from the admixed large, antrorse-appressed, weakly flexuose, and small, divergent, curled hairs of K. toelkenii. The leaves of K. tenuicaulis are also smaller than those of K. toelkenii, which grow to 8.5 × 2.5 mm, and K. tenuicaulis lacks functionally male late season flowers, unlike K. toelkenii (see Table 1).

Kunzea tenuicaulis is the dominant woody plant on the active geothermal fields within the Taupo Volcanic Zone (Fig. 20A–C) where it colonises not only heated ground but quiescent and/or ‘cool’ ground associated with each geothermal field. In these ‘cool’ peripheral situations it may be dominant, though it usually associates with Leptospermum scoparium, Weinmannia racemosa L.f., Knightia excelsa R.Br. and Kunzea robusta. The understorey of this peripheral vegetation is usually dominated by shrubs and ferns such as Leptecophylla juniperina (J.R.Forst. et G.Forst.) C.M.Weiller, Leucopogon fasciculatus (G.Forst.) A.Rich., Dracophyllum subulatum Hook.f., Pteridium esculentum (G.Forst.) Cockayne, Histiopteris incisa (Thunb.) J.Sm., and Lycopodiella cernua (L.) Pic.Serm. Toward the active geothermal vents, where surface temperatures can abruptly rise to as much as 90 °C (Burns 1997) and most woody vegetation becomes scarce, Kunzea tenuicaulis is the dominant macro-vegetation. In these habitats, which may include extensive areas of steam field, heated mud pools and hot springs, active and quiescent hydrothermal explosion craters, and fumaroles, the low shrub or prostrate trailing form of K. tenuicaulis is best developed (Fig. 20A, C, J). It is this form which may be found flowering at less than 40 mm tall and which is the plant referred to in past literature as K. ericoides var. microflora (see Given 1980a; Harris et al. 1992; Smale 1994; Harris 1996; Burns 1997). However, in many geothermal areas (e.g., Tikitere, Kuiarau Park andTokaanu) this decumbent form of K. tenuicaulis is mostly replaced by multi-trunked, erect to suberect trees of K. tenuicaulis (see Fig. 20D–G) identical to those seen growing in peripheral ‘cold’ and/or thermally quiescent areas in more active fields. Further study is needed to determine why the low shrub form of K. tenuicaulis seems to be favoured in the more unstable geothermal systems, though physiological stress and aluminium toxicity has been suggested as a partial explanation (Burns 1997). Irrespective, in these geothermally more active habitats K. tenuicaulis frequently associates with the ferns Dicranopteris linearis (Burm.f.) Underw., Nephrolepis flexuosa Colenso, Cheilanthes distans (R.Br.) Mett. and C. sieberi Kunze subsp. sieberi, sparse, stunted Leucopogon fasciculatus and Dracophyllum subulatum shrubs, the lilies Dianella nigra Colenso, D. haematica Heenan et de Lange and the exotic love grass Eragrostis brownii (Kunth) Wight. Underneath K. tenuicaulis shrubs a ground cover of liverworts and mosses dominated by Chiloscyphus semiteres (Lehm.) Lehm. and Campylopus pyriformis (Schultz) Brid. is usually present. In these habitats and also in the peripheral cool soils, ectomycorrhizal fungi of the genus Pisolithus Alb. et Schwein. (see Burns 1997; Moyersoen and Beever 2004) have been found in exclusive association with K. tenuicaulis (McKenzie et al. 2006 as K. ericoides var. microflora). Kunzea tenuicaulis is also occasionally, and at times rather heavily, parasitised by the dwarf mistletoe Korthalsella salicornioides.

The putative hybrids K. robusta × K. tenuicaulis and K. serotina × K. tenuicaulis have been collected throughout the range of K. tenuicaulis. However, of these hybrids, only K. robusta × K. tenuicaulis is commonly encountered, because K. robusta is more frequently sympatric with K. tenuicaulis along the margins of that species’ geothermal habitats, and in the plantation forests abutting many of the geothermal fields within the Rotorua Volcanic Centre (for geology see Briggs et al. 2005; Neall 2001). Putative gatherings of K. serotina × K. tenuicaulis are less common because the ranges of those species rarely overlap (except around Karapiti and Tokaanu). These hybrids are discussed in detail under K. serotina.

Hybrids involving K. robusta are best recognised by the presence of mixtures of long, appressed, weakly flexuose and shortly divergent branchlet hairs. However, in the field they can be distinguished by their general tendency to form single trunked, weakly spreading trees, with fewer branches and branchlets that are often somewhat slender and semi-pendulous to pendulous. Kunzea robusta × K. tenuicaulis is usually present as introgressive hybrid swarms because most of the geothermal habitats are now extensively modified.

Artificial hybrids involving K. tenuicaulis as staminate or pistillate parent and other New Zealand members of the K. ericoides complex were easily produced, and showed no obvious reduction in fertility, except for crosses involving K. serotina, which were sterile (de Lange and Murray 2004; de Lange et al. 2005).

No specific Maori name for this species seems to have been recorded.

Currently, as K. ericoides var. microflora, K. tenuicaulis is appropriately listed as ‘At Risk/Naturally Uncommon’ qualified ‘RR’ [Range Restricted] by de Lange et al. (2013b).

A K. tenuicauli foliis constanter longioribus angustioribus lineo-lanceaceolatis, hypanthio maiore glabrato anguste obconico vel infundibuliforme, stigmate plano anguste capitato, lobisque antherae profunde sulcatis non testiculatis differt.

(Fig. 21).New Zealand: North Island, Bay of Plenty, Moutohora (Whale Island), McEwans Bay, 37°51'26"S, 176°58'57"E, 20m a.s.l. ‘Occasional on sand dunes well away from active or senescent thermal areas’. P. J. de Lange 6471 & P. B. Cashmore, 15 Apr 2005, AK 289816! Isotype: AD!

Figure 21. 

Holotype of Kunzea salterae (P. J. de Lange 6471 & P. B. Cashmore, AK 289816).

The specific epithet salterae commemorates my colleague and botanical illustrator for this monograph Josh Salter (1946–), whose critical attention to detail when illustrating specimens of K. salterae proved invaluable in deciding on an appropriate taxonomic rank.

(Figs 22, 23, 24).Growth habit shrubs to small trees 0.1–6(–10) × 2–4(–6) m with broad, spreading to somewhat pendulous crowns, rarely plants completely decumbent, sprawling across ground. Trunk usually multi-trunked from base, up to 0.3 m d.b.h., these mostly widely spreading to suberect, flexuose, often basally buttressed, branches frequent from base in exposed sites, otherwise naturally thinning in the lower half of the trunk. Bark early bark brown, initially firm, somewhat sinuous-fluted, elongate, over time cracking transversely (especially on branch flanges), and with apices gradually detaching and raising to present as small lunate (in profile) flakes, old grey-brown bark flaking readily in small, somewhat irregular tabular shards, often with small lunate secondary peeling; somewhat corky to chartaceous. Branches Two to many, suberect to widely spreading, rarely ascending, mostly pendulous, branchlets numerous and very leafy, rather slender, initially subterete soon becoming quadrangular; sericeous, indumentum initially copious rarely glabrate to glabrous, hairs on young rapidly growing apices, copious, sericeous, straight, antrorse-appressed up to 0.55 mm, these soon falling; other mostly divergent hairs long persistent, (especially opposite leaf buds and expanding foliage), 0.04–0.08(–0.1) mm, hyaline to translucent (appearing white when young maturing grey), apices ± curled, often admixed (particularly toward branchlet apices and near decurrent leaf bases) with deciduous antrorse-appressed, straight to somewhat sinuous hairs up to 0.28 mm. Vegetative buds inconspicuous at resting stage 0.5–1.0 mm diam.; scales deciduous; (0.6–)1.2–2.3 mm long, stramineous to pale brown, initially broadly ovate to ovate-lanceolate grading through broadly lanceolate to narrowly lanceolate, midrib strongly keeled, prolonged to apiculate tip, with one prominent row of 4–10 oil glands on either side of midrib, margins, apex, apiculus and keel finely ciliate. Leaves ± spreading to patent; lamina (4–)10(–18) × (0.6–)1.2(–2.0) mm, bright glossy green, yellow-green, bronze-green to dark green; linear-lanceolate to narrowly oblanceolate, flat not recurved, apex acute to subacute, cuspidate, rarely obtuse to rounded; base attenuate; adaxial surface slightly concave to flat, finely glandular punctate; oil glands 180(–280), more evident when dry; midrib slightly raised to depressed near base otherwise depressed for entire length, initially densely covered in fine, antrorse-appressed silky hairs up to 0.22 mm, becoming glabrescent; abaxial surface slightly convex, finely glandular punctate, oil glands less obvious when fresh than when dry, up to 100, with the larger glands aligned longitudinally along midrib; midrib slightly raised, usually glabrous, sometimes with a fine weft of silky, deciduous, antrorse-appressed hairs near base; lamina margin sparsely to densely, finely sericeous, hairs mostly antrorse-appressed, up to 0.5 mm, hyaline to translucent, appearing as white to naked eye; hairs in 1–2 somewhat irregular rows just failing to meet short of cuspidate leaf apex. Perules scarious, basal ones usually persistent, 1.2–1.4 mm long, stramineous to brown, broadly to narrowly lanceolate, involute, midrib strongly keeled prolonged as a cuspidate apex, with one row of 4–8 oil glands on either side of midrib, lower two-thirds glabrous, upper one-third finely ciliate; remaining perules deciduous, chartaceous, 0.6–1.4 mm long, pale pink to pinkish-white when fresh, drying apricot to apricot-pink, broadly oval, ovate to rhomboid, finely and copiously ciliate, strongly keeled, keel prolonged, apiculate, margins and keel more distinctly ciliate. Inflorescence a (2–)4(–8)-flowered corymbiform botryum up to 45 mm long, usually on brachyblasts, rarely on long shoots in which case invariably terminal (only very rarely with terminal vegetative growth). Inflorescence axis densely invested with mostly divergent hairs. Pherophylls deciduous (falling very early), mostly squamiform, rarely foliose, spreading, 0.6–1.8 mm long; squamiform pherophylls brown or amber, sometimes pink, drying apricot-brown, broadly deltoid to oblong-ovate, margins involute especially in upper one-third, midrib strongly keeled prolonged as cuspidate apex, with one row of 4–8 oil glands on each side of midrib; glabrous except for the finely ciliate margin and apex; foliose pherophylls bright green, linear, margins and apex finely ciliate; both types grading into chartaceous, into perules and/or leaves at inflorescence terminus. Pedicels (1.1–)2.6(–3.0) mm long at anthesis and elongating slightly after anthesis, terete, finely invested in divergent to subantrorse sericeous hairs. Flower buds pyriform to clavate, apex domed with calyx valves not or scarcely meeting. Fresh flowers when fully expanded up (9–)10(–12) mm diam. Hypanthium (2.1–)2.2(–3.8) × (1.8–)2.2(–3.2) mm, with free portion 1.0–1.6 mm long, reddish-brown when fresh, drying resinous brown to grey; narrowly obconic to funnelform terminating in a slightly thicker rim bearing five persistent calyx lobes; surface smooth, finely glandular punctate, sparsely hairy to glabrate, with five rather weakly defined ridges leading up to calyx lobes (these becoming more distinct upon drying); hairs scattered, subantrorse to antrorse, flexuose. Calyx lobes 5, upright (not spreading), 0.6(–0.9) × 1.1(–1.3) mm, persistent, broadly to narrowly triangular, weakly and broadly keeled (the keel though ill-defined in fresh specimens recognisable as a dark pink to red, thicker central prolongation of the hypanthium ridges), margins cream to pale yellow, gland-dotted, subcoriaceous, glabrate except for distinctly ciliate apex. Receptacle dark red at anthesis. Petals 5, spreading, 1.4–1.6 × 1.4–1.6 mm, white, rarely basally flushed pink, orbicular to suborbicular, apex obtuse to rotund, margins usually finely crimped, oil glands colourless or rose-pink, scarcely evident when fresh. Stamens 28–36(–38) in 1–2 weakly defined whorls, adnate to receptacular rim, filaments white rarely tinged rose-pink toward base. Antipetalous stamens 3(–5) antisepalous 3(–4). Outermost antipetalous stamens strongly outcurved, on filaments 2.5–3.25 mm long, inner stamen 1.8–2.2 mm, outcurved, on occasion a further 1–2 incurved or outcurved, stamens 0.8–1.0 mm long, positioned at the base of the outermost antipetalous pair. Antisepalous stamens much shorter than antipetalous, 0.6–0.9(–1) mm, incurved, outcurved or in mixtures of both. Anthers dorsifixed, 0.11–0.16 × 0.10–0.14 mm, scutiform to ovoid, latrorse, each anther deeply and longitudinally furrowed, with one anther lobe in each pair fused at right angles along inner margin with adjoining anther lobe to form a prominent ‘pinched’ longitudinal ridge. Pollen white, (10.2–)14.7(–16.6) μm. Anther connective gland prominent, pale orange to pink when fresh, drying orange-brown, spheroidal, finely papillate, somewhat farinose. Ovary (3–)4 locular, each locule with 8–10 ovules in two rows on each placental lobe. Style 2.1–3.2 mm long at anthesis, white basally flushed with pink; stigma capitate, up to 1× style diam., flat, abruptly broadened, pale cream, finely papillate rugulose. Fruits rarely persistent, (2.0–)2.2(–2.7) × (2.0–)2.9(–4.0) mm, light brown to grey, cupular to suburceolate, splits concealed by dried, erect, free portion of hypanthium. Seeds 0.80–1.00 × 0.45–0.48 mm, narrowly oblong, oblong, oblong-obovate to falcate-oblong or elliptic, curved near apex, laterally compressed, 2–3-angled with convex to flattened faces, apex rounded; base cuneate to oblique, ± flattened; testa semi-glossy, orange-brown; surface coarsely reticulate, ridges prominent, central portion of some cells furnished with short, tubular-spiny, protuberances. FL: Aug–Apr FT: Aug–Sep. Chromosome Number n = 11_II, 2n = 22 (AK 283253, P. B. Cashmore s.n., AK 298088, P. J. de Lange.

Figure 22. 

Distinguishing features of Kunzea salterae. A Flowering branchlet (AK 289816) B Vegetative bud and branchlet indumentum (AK 289816) C Adaxial leaf surface (AK 289816) D Abaxial leaf surface (AK 289816) E Adaxial leaf apex (AK 289816) F Leaf margin indumentum (AK 289816) G Leaf variation, all from Moutohora (Whale Island): (G1) (AK 185215), (G2) Boulder Bay (AK 288250), (G3–5) Sulphur Bay (AK 284105, AK 283253, AK 289814), (G6) Summit Hill Saddle (AK 289815), McEwans Bay (AK 289816) H Flower (top view) (AK 289816) I Flower and hypanthium (side view) (AK 289816) J Flower cross section showing anther, style and ovules (AK 289816) K Style and stigma (AK 289816) L Stamens (AK 289816) M Dehisced fruit (AK 289816). Scale bars: (A, G) 10 mm; (B–E, H–M) 1 mm; (F) 0.5 mm.

Figure 23. 

Scanning Electron Micrographs of Kunzea salterae. (A–E all AK 284105) Branchlet indumentum F–L Seeds (AK 283253, AK 289815) K–L Close up of reticulum showing spines. Scale bars: (A, C, F) 1 mm; (B, E) 500 μm; (D, F–J) 100 μm; (K, L) 50 μm.

Figure 24. 

Habitats of Kunzea salterae on Moutohora Island (photos: P. B. Cashmore). A Sand dunes and early stage successional forest leading to Summit Hill B Stable sand dunes and early stage forest surrounding Department of Conservation Hut and ride line leading to Summit Hill C Active geothermal vents within Sulphur Valley.

(15 Sheets seen).Moutohora (Whale Island): P. Hynes s.n., 28 Aug 1970, (AK 185215); Sulphur Bay, Geothermal Area, P. B. Cashmore s.n., 4 Sep 2002, (AK 297561); Boulder Bay, P. B. Cashmore s.n., 4 Sep 2002, (AK 283250); Sulphur Bay, Thermal Area (Active), P. J. de Lange 6469 & P. B. Cashmore, 15 Apr 2005, (AK 289814); Pa Hill/Summit Hill Saddle, P. J. de Lange 6469 & P. B. Cashmore, 15 Apr 2005, (AK 289815); Summit Hill (southern slopes), P. J. de Lange 6472 & P. B. Cashmore, 15 Apr 2005, (AK 289817, Duplicate AD).

(Fig. 7). Endemic, New Zealand, North Island, Bay of Plenty, Moutohora (Whale Island) (sea level to 220 m a.s.l.).

Kunzea salterae is recognised at species rank because it forms a true-breeding, morphologically stable population, recognised here by a combination of growth habit, branchlet hair and floral characters (see Table 1) as well as minor but consistent DNA sequence differences in the ETS marker region (Table 2; see also de Lange 2007; de Lange et al. 2010). It is further distinguished ecologically by its preference for sand dune and geothermal habitats (Fig. 24A–C), and also by its sympatry/syntopy with K. robusta, from which it is isolated morphologically, and from which I saw no field evidence of hybridism (see below). The presence of K. salterae on Moutohora, a small (143 ha) volcanic island estimated to be 36 000 years BP (see Ramsay and Hayward 1971) is as remarkable as it is unexpected. Whether the species evolved in situ, is a remnant population that persisted there following the extinction of other populations that had colonised the lowered shore line of the Bay of Plenty prior to the sea level rise that occurred at the end of the last glacial maximum, or has recently colonised the island from another as yet unrecognised mainland location remains to be determined.

In past literature Kunzea salterae has usually been recorded as K. ericoides (e.g., Parris 1971 (as Leptospermum ericoides); Ogle 1990; Smale 1994). However, plants found growing within the geothermally active part of the island have also been referred to K. ericoides var. microflora (= K. tenuicaulis of this revision) (Wildlands Consultants Limited 2005) probably because of their low stature, apparent habitat preferences, and the widely held but largely mistaken belief (see K. tenuicaulis) that any decumbent Kunzea found near active fumaroles was that variety.

Kunzea salterae has some similarity to K. tenuicaulis. In particular the ability to grow in geothermal habitats (Fig. 24C), the characteristically multi-trunked growth habit, broadly spreading canopy, and numerous rather fine, often pendulous branches and branchlets are typical of both species, while the abundance of short, divergent branchlet hairs (Fig. 23A–E) is shared otherwise only with the allied K. serotina. Kunzea salterae, like K. tenuicaulis, has a tendency to produce numerous semi-erect, somewhat trailing or completely decumbent plants in the vicinity of or around active fumaroles (Fig. 24C). From Kunzea tenuicaulis, K. salterae is distinguished by its longer (up to 18 mm cf. up to 10 mm), linear-lanceolate (Fig. 22A, C–G) rather than oblanceolate to obovate leaves, by its slightly larger (range 2.1–3.8 mm long) and glabrate, rather than small (range 1.8–3.1 mm long) and puberulent, narrowly obconic to funnelform (Fig. 22I–J) rather than cupular to campanulate hypanthium, by its flat, narrowly capitate rather than slightly domed centrally depressed stigma (Fig. 22K), and by the non-testiculate, deeply furrowed thecae (Fig. 22L). Only one mitotic count was obtained from K. salterae and this matched K. tenuicaulis, K. serotina and K. toelkenii in having uniformly small chromosome complements. Further observations using different plants are needed to confirm this.

Aside from K. tenuicaulis, the narrowly linear-lanceolate foliage of K. salterae is similar to that of K. linearis and K. ericoides. However, the shorter glabrate leaves of Kunzea salterae are distinct from the leaves of K. linearis, and branchlet hairs of K. salterae are short and divergent rather than long, silky and antrorse. Further, the inflorescences of K. salterae are corymbiform rather than spiciform, and the individual flowers are distinctly pedicellate, never sessile to subsessile. Both species are also allopatric, the nearest occurrence of K. linearis to K. salterae being the Tairua Peninsula on the eastern side of the Coromandel Peninsula some 145 km north-west of Moutohora.

Kunzea salterae can be distinguished from K. ericoides by its copiously hairy branchlets furnished with much longer divergent hairs than are on K. salterae branchlets. Further both species are allopatric and have different ITS and ETS sequences (Table 2).

On Moutohora, K. salterae is sympatric with K. robusta, (e.g., P. J. de Lange 6473 & P. B. Cashmore (AK 289818)), which grows locally on the southern slopes of Summit Hill. The typically multitrunked, pendulous growth habit, consistently narrow linear-lanceolate leaves, and abundance of short, divergent branchlet hairs easily distinguish K. salterae from K. robusta, which has long, silky, antrorse-appressed, branchlet hairs. Kunzea robusta is the less common of the two species on Moutohora, and is absent from sites of geothermal activity there.

Kunzea salterae appears to combine the narrow linear leaves typical of K. linearis, with the growth habit of K. tenuicaulis. Interestingly, artificial F₁ hybrids (see de Lange et al. 2005) using K. tenuicaulis as the pistillate parent, (e.g. P. J. de Lange 5816 (AK 285268)), are a close morphological match for K. salterae. Based on current herbarium and field evidence, K. linearis is not known from the Bay of Plenty (see above). Nevertheless, the extremely similar morphology exhibited between the aforementioned F₁ hybrids and K. salterae is rather striking. Further research into the possible hybrid origin of K. salterae, particularly whether it is derived from past hybridism between K. linearis and K. tenuicaulis, would be worthwhile.

The rDNA ITS and ETS sequence data (Table 2) showed that K. salterae was most similar to K. tenuicaulis (de Lange 2007). Otherwise, despite its narrow linear-leaves, it consistently clustered with the other ‘small-leaved’ Kunzea, K. tenuicaulis, K. toelkenii and K. serotina (de Lange 2007). ETS sequence data also indicated a relationship with K. ericoides, K. linearis, K. tenuicaulis, K. toelkenii, K. serotina and Mt Egmont samples of K. robusta all of which share a guanine/cytosine mix (de Lange 2007; de Lange et al. 2010). Otherwise Kunzea salterae differs from all other Kunzea taxa within the K. ericoides complex by having a unique cytosine/thiamine mix in its ITS-2 sequence (Table 2; see also de Lange 2007; de Lange et al. 2010).

The identity of Kunzea on Tuhua (Mayor Island) was discussed by de Lange (2007) who noted that one collection from the ‘crater rim’ of that island (AK 262432, G. W. Mason s.n.), approached K. salterae in general branching habit, leaf shape and size, and by the numerous small, corymbiform inflorescences. Although this specimen was in poor condition, de Lange (2007) noted that the branchlet indumentum comprised mainly fine, somewhat wispy, appressed antrorse hairs, and that the anthers lacked the deep furrow and fused ‘pinched’ ridge typical of K. salterae. Subsequent field work on that island has shown that the Kunzea there forms a uniform population matching the ‘eastern North Island variant’ of K. robusta which is common on the adjacent eastern side of the Coromandel Peninsula (see below) (Wilcox et al. 2012).

Kunzea salterae is a widespread and at times dominant woody shrub or tree of the coastal forest, geothermal field, cobble beach and sand dune vegetation of Moutohora (Fig. 24A–C). As Kunzea ericoides var. ericoides, Smale (1994) described in detail the vegetation associations, population structure and dynamics of Kunzea salterae. He concluded (p. 441) that this species ‘may replace itself indefinitely on the unstable dunes on Whale [Moutohora] Island, where the community is still expanding’. Smale’s study was confined to Kunzea ‘heaths’ developed over sand dunes, and so he did not appraise the associations formed by K. salterae within the geothermally active parts of Moutohora. From observations outside the sand dune habitat, I suggest that K. salterae, being a species evidently favouring frequent disturbance, will also have a long standing presence in the geothermally active parts of Moutohora, where it is the dominant vascular plant species. Indeed, in its abundance and growth within the thermal areas on this island, it behaves very much as K. tenuicaulis does in similar mainland habitats within the Taupo Volcanic Zone of the North Island (Burns 1997). Kunzea salterae is currently often the dominant canopy cover outside the sand dune and geothermal areas of Moutohora (Fig. 24B), but it is expected to decline as the coastal forest regenerates and other larger, coastal forest species attain local dominance.

Smale (1994) observed that the sand dune vegetation dominated by K. salterae was species poor, recording 29 vascular plant taxa within what he regarded as ‘older’ stands (i.e. ≥ 27 years of age). The same is the case for the thermal areas, where, aside from dense coverings of the mosses Isopterygium albescens (Hook.) A.Jaeger, Campylopus pyriformis, Dicranella dietrichiae (Müll.Hal.) A.Jaeger, Philonotis tenuis (Taylor) Reichardt and Hypnum cupressiforme Hedw. var. cupressiforme (Beever and Brownsey 1990), vascular plants (other than Kunzea) are extremely scarce. Smale observed that inland from the dune systems his “Kunzea ericoides var. ericoides” stands changed from the multi-stemmed semi-prostrate growth habit (the K. salterae of this treatment) to erect single-stemmed trees. From my observations this transition is not nearly as clear cut as he described, with K. salterae growing in most situations across the island with a consistently multi-stemmed habit. However, occasional larger erect trees do occur toward the back of the dune systems at Boulder Bay and these are not K. salterae but K. robusta, a species that avoids thermal areas, and favours more stable habitats, overlying better developed soils, within the more mature successional coastal forest on the island.

No putative wild hybrids have been observed on Moutohora, and putative hybrids were not evident in the 15 Kunzea herbarium specimens examined from that island. The distinctiveness of K. salterae was recognised too late in this revision to include it in hybridisation experiments (de Lange et al. 2005).

No specific name for K. salterae has been recorded.

Currently the species, as K. aff. ericoides var. microflora has been appropriately assessed by de Lange et al. (2013b) as ‘At Risk / Naturally Uncommon’ qualified ‘IE’ (Island Endemic) and ‘OL’ (One Location).

A K. tenuicaulis habitu late expanso (usque ad 6 m lato), brevi (usque ad 4 m alto), valido multicauli, caulibus pertortis torsivis et flexis; surculibus frequentibus perfecte prostratibus ad 4 m e basi trunci expositis; ramis et ramulis superis pendulis, faragine pilorum longorum leniter flexuorum antrorum et brevium divergentium crisporum circinatorum; seriebus rDNA ITS et ETS differt.

(Fig. 25).New Zealand (North Island). Bay of Plenty, State Highway 2, near Thornton (Wahieroa Dunes), Walker Road, 37°58'27"S, 176°50'11"E, 10 m a.s.l. ‘Dominant, growing with Muehlenbeckia complexa, Lupinus arboreus, boxthorn and Pyrrosia eleagnifolia as sparse associates. Multi-trunked shrubs to small trees up to 4 × 4 m’. P. J. de Lange 5322 & R. O. Gardner, 25 Oct 2001, AK 255350! Isotypes: AD! BM! CHR! NZFRI! P!

Figure 25. 

Holotype of Kunzea toelkenii (P. J. de Lange 5322 & R. O. Gardner, AK 255350).

The specific epithet toelkenii honours Australian Kunzea expert Hellmut Toelken (1939–).

(Figs 26, 27, 28).Growth habit shrubs up to 4 × 6 m; ‘juveniles’ usually prostrate and trailing up to 4 m diam., often flowering, taking 2–4 years to develop several, usually central, ascending branches; ‘adults’ forming widely spreading (up to 2 m diam.), characteristically flat-topped shrubs, with pendulous branches and branchlets; branches confined to upper 30–50% of shrub, basal 50–70% usually completely devoid of branches and vegetative growth, sometimes bearing completely pendulous growth; trunk bases usually bearing epicormic, prostrate growth spreading up to 4 m diam. from point of origin; this growth occasionally layering and producing further trunks. Trunk (1–)6(–10), ascending to suberect, serpentine, highly contorted, twisted, bent, and spiralled, 0.10–0.25(–0.40) m d.b.h.; mostly arising from the top of a broad rootstock, and also from layered masses of prostrate epicormic growth; in all cases basal portions of trunks covered with numerous semi-detached, sinuous, rather corky, lengths of bark. Bark early bark firmly coriaceous, grey or grey-brown, ± elongate, initially with few transverse cracks, soon becoming heavily cracked (into highly irregular pieces with rather sinuous margins (especially on branch flanges and decurrent leaf bases) but remaining firmly attached; old bark similar though more distinctly coriaceous-corky, upper surface often deeply corrugated and cracked but not peeling; detaching inwards readily but usually remaining centrally firmly attached; margins sinuous to lunate, often highly irregular and frayed, rarely shortly tabular; early and old bark flakes firm, scarcely crumbling in hand. Branches of trunks numerous, usually confined to the upper 30–50% of trunk; widely spreading, ± serpentine, flexuose, often pendulous and interwoven; branchlets numerous, slender, usually apically pendulous, very leafy, with few to many brachyblasts; those of epicormic growth, straight not flexuose or serpentine, prostrate or pendulous if arising from basal half of trunk, widely spreading; in all cases quadrangular, sericeous, indumentum copious; hairs persistent, of two types: long, appressed often flexuose hairs up to 0.26 mm long, and smaller divergent hairs, with strongly curled and spiralled apices 0.04–0.10(–0.18) mm hyaline to translucent (appearing white when young maturing grey). Vegetative buds conspicuous; at resting stage 1.2(–1.8) mm diam.; scales scarious, deciduous, (0.4–)0.9(–1.2) mm long, brown to red-brown, broadly ovate to ovate-deltoid, apex obtuse to rounded; midrib prominent, strongly keeled in upper half, occasionally prolonged to a long cuspidate tip, lateral veins usually absent, oil glands usually absent, upper half of scale margins, keel, and keel apex ciliate. Leaves well spaced along branchlets, spreading, patent to recurved; lamina (2.6–)5.7(–8.5) × (0.6–)1.6(–2.5) mm, dark glossy green or bright-green, margins and base usually flushed red; spreading, obovate, clavate, to broadly oblanceolate; weakly to strongly recurved from about 30–50% of total length; apex sharply acute to apiculate, base attenuate; adaxial surface concave very rarely flat, finely glandular punctate; oil glands up to 280, more evident when dry; midrib slightly raised near base, otherwise not evident for rest of length, finely covered in deciduous, sericeous, antrorse-appressed, hairs in lower half otherwise glabrous; abaxial surface convex, glandular punctate, oil glands up to 180, more evident when dry; midrib raised for most of length, glabrous; lamina margin finely to densely sericeous, hairs weakly flexuose, antrorse, subantrorse to spreading, up to 0.5 mm long, hyaline to translucent, appearing white to naked eye, aligned in 1–2 uninterrupted rows meeting just short of leaf apiculus. Perules scarious, persistent, (0.5–)1.0(–1.8) mm; basal ones dark brown to red-brown, broadly ovate, ovate, ovate-rostrate, to lanceolate, without oil glands, margins involute, ciliate, midrib strongly keeled with 1–2 usually finely ciliate lateral veins on each side, keel prolonged as a short to long, deciduous, obtuse-tipped, densely ciliate, cuspidate apiculus; remaining perules similar but smaller, chartaceous, (0.3–)0.8(–1.0) mm long. Inflorescence a compact, (1–)7(–10)-flowered corymbiform botryum up to 40 mm long, mostly borne on alternate, distinctly spiralled, basally densely leafy, brachyblasts up to 12 mm long; inflorescences at the ultimate branchlet terminus uncommon (except in trailing epicormic growth), if present, often rather elongated (up to 80 mm long) and bearing well developed terminal vegetative growth, often with the uppermost flowers in elongated shoots male. Inflorescence axis densely invested with divergent hairs. Pherophylls deciduous (falling very early), initially foliose, soon squamiform, tightly clasping pedicel or spreading, 0.4–1.6 mm long, foliose pherophylls green to bronze-green, shortly lanceolate to obovate, squamiform pherophylls amber-brown to brown, narrowly deltoid to ovate, both types adaxially deeply concave, margins and apex finely ciliate, grading into leaves at inflorescence axis apex. Pedicels (1.6–)2.9(–3.8) mm long at anthesis, usually elongating slightly after anthesis, terete, copiously invested with short, divergent to subantrorse, silky hairs. Flower buds bluntly clavate to obconic, rarely pyriform, apex flat prior to bud burst with calyx valves not meeting. Fresh flowers when fully expanded (3.6–)6.8(–9.0) mm diam., often functionally male toward end of flowering season. Hypanthium (1.7–)2.4(–3.2) × (2.8–)3.6(–4.3) mm, with free portion 0.6–0.9 mm long, green, dark green or red-green; obconic to funneliform, terminating in light-green to pink-green membranous rim bearing five persistent calyx lobes; surface smooth when fresh somewhat wrinkled when dry, with weakly defined ridges leading up to calyx lobes; sparingly dotted with pink or colourless oil glands otherwise with basal half finely and rather densely puberulent with areas leading to calyx lobes distinctly glabrescent; hairs silky, spreading, subantrorse to antrorse-appressed, often with smaller divergent hairs underlying larger appressed ones. Calyx lobes 5, upright (not spreading), submembranous, (0.8–)1.0(–1.2) × (0.7–)1.0(–1.2) mm, persistent, ovate, broadly ovate to ovate-deltoid, of uniform thickness in transverse section, without keel, often uniformly green, otherwise with central portion of lobe darker green or pinkish green, with margins usually pale green to green flushed with pink, surface somewhat glandular punctate, oil glands inconspicuous, ± colourless, otherwise glabrous except for distinctly spreading, ciliate margins. Receptacle usually pink at anthesis, consistently darkening to dark magenta or maroon-black after fertilisation. Petals 5(–6), 1.5–1.9(–2.8) × 1.5–1.9(–2.6) mm, white, orbicular to very broadly ovate, apex obtuse to rotund, margins ± entire, often finely folded or crimped 1–5 times, oil glands colourless. Stamens 20–36(–50) in 1(–3) weakly defined whorls, arising from receptacular rim, filaments white. Antipetalous stamens (2–)3(–6), antisepalous (1–)3(–8). Outermost antipetalous stamens weakly incurved or outcurved, on filaments 1.2–3.6 mm long, inner stamen if present, 0.8–1.2 mm, incurved or outcurved, a further 1–3 stamens, of similar length are very rarely present at the base of the outermost antipetalous pair. Antisepalous stamens usually shorter than outermost antipetalous stamens, sometimes of comparable length, generally 0.6–3.2 mm, mostly incurved, outcurved or in mixtures of both. Anthers dorsifixed, 0.06–0.09 × 0.05–0.08 mm, testicular-oval to testicular-ellipsoid, latrorse. Pollen white (12.2–)13.6(–17.8) μm. Anther connective gland prominent, pale lemon to pink when fresh, drying yellow to pale orange, spheroidal, finely papillate. Ovary absent in males flowers, otherwise 3–4(–5) locular, each with 12–20(–24) ovules in two rows on each placental lobe. Style absent in male flowers, otherwise 1.0–1.4(–1.8) mm long at anthesis, elongating slightly after anthesis, white; stigma capitate, scarcely wider than style, flat, greenish-white, cream or pale pink, surface papillate. Fruits rarely persistent, (2.1–)2.6(–3.0) × (2.5–)3.0(–3.7) mm, light brown to grey, obconic, broadly obconic, to cupular, splits concealed by dried, suberect to erect, free portion of hypanthium. Seeds 0.50–1.00(–1.02) × 0.52–0.60(–0.68) mm, oblong, oblong-obovate, curved near apex, laterally compressed, 2–3-angled with convex to flattened faces, apex rounded to subacute; base oblique, ± flattened; testa semi-glossy, amber, orange-brown to brown, surface coarsely reticulate. FL: (Sep–)Oct–Nov. FT: Oct–Sep. Chromosome Number n = 11_II, 2n = 22 (see de Lange and Murray 2004).

Figure 26. 

Distinguishing features of Kunzea toelkenii. A Flowering branchlet (ex cult. AK 284553) B Fruiting branchlet (ex cult. AK 284553) C Vegetative bud and branchlet indumentum (ex cult. AK 284553) D Adaxial leaf surface (ex cult. AK 284553) E Abaxial leaf surface (ex cult. AK 284553) F Leaf variation: (F1) Walker Road (AK 255350), (F2) Seacombes Canal (AK 287042) G Adaxial leaf apex (ex cult. AK 289816) H Leaf margin indumentum (ex cult. AK 284553) I Flower (top view) (ex cult. AK 284553) J Flower and hypanthium (side view) (ex cult. AK 284553) K Flower cross section showing anther, style and ovules (ex cult. AK 284553) L Style and stigma (ex cult. AK 284553) M Stamens (ex cult. AK 284553) N Dehisced fruit (ex cult. AK 284553). Scale bars: (A, B, F) 10 mm; (C–E, G, I–N) 1 mm; (H) 0.5 mm.

Figure 27. 

Scanning Electron Micrographs of Kunzea toelkenii (all AK 255350). A–C Branchlet indumentum. Scale bars: (A, B) 1 mm; (C) 100 μm.

Figure 28. 

Kunzea toelkenii at type locality, Walker Road (photos: P. J. de Lange). A Growth habit of K. toelkenii within sand dunes at type locality, note extensive suckering growth at base of shrub B Side view of the same K. toelkenii as (A) showed tortured growth and root suckers C Close up of distinctive branching pattern developed by mature K. toelkenii at type locality D–E Bark of K. toelkenii F Late season functionally male flowers of K. toelkenii G–H Flowering branchlet of K. toelkenii (an example with longer than usual stamens).

(16 sheets seen).New Zealand (North Island). State Highway 2, near Seacombes Canal, P. J. de Lange 5324 & R. O. Gardner, 25 Oct 2001, (AK 287042, Duplicate: AD); State Highway 2, Walker Road, P. J. de Lange 5314, 29 Sep 2001, (AK 287049); State Highway 2, Walker Road, P. J. de Lange 5323 & R. O. Gardner, 25 Oct 2001, (AK 287045, Duplicate: CHR); 1.75 km east of Rangitaiki River Mouth, Thornton Wildlife Management Reserve, eastern end of lagoon, P. B. Cashmore s.n., 7 Jun 2007, (AK 299633); 3.3 km East of Rangitaiki River Mouth, near Whakatane Airport Buildings, P. B. Cashmore s.n., 7 Jun 2007, (AK 299634); Whakatane, Piripai Spit, East of Coastlands Subdivision, P. B. Cashmore s.n., 12 Jul 2007, (AK 300903); Ohiwa Harbour, Whangakopikopiko (Tern Islet), P. J. de Lange 7247 & P. B. Cashmore, 5 Dec 2007, (AK 301682, Duplicate: CHR).

(Fig. 7). Endemic, New Zealand, North Island, Bay of Plenty (2–10 m a.s.l.). Kunzea toelkenii is known from a small strip of sand dune country (the Wahieroa Dunes) between the eastern bank of the Tarawera River mouth and the west bank of the Rangitaiki River mouth, near Thornton, and from another small population on a barrier sandspit island, Whangakopikopiko (Tern Islet), at the mouth of the Ohiwa Harbour. The current distribution is undoubtedly relict; the habitats occupied are remnants of indigenous woody sand dune vegetation that formerly extended as far west as Papamoa. Although I have been unable to find any supporting herbarium evidence, locals recollect that much of the sand country between Papamoa and Pikowai beach once supported dense ‘kanuka [Kunzea] shrublands’ (G. Wrigley pers. comm.). From the descriptions given by these people which include such phrases as ‘tortured growths....pendulous shrubs.......... suckering stems’ it is quite likely that K. toelkenii was the species involved, and that it was once a locally important species of the Bay of Plenty sand dune country.

Kunzea toelkenii is recognised by its uniquely suberect, sprawling growth habit, typically extensive suckering (Fig. 28A–C), by its mixed branchlet hairs (Fig. 27A–C), tendency to produce late season functionally male flowers (Fig. 28F), and also by its restriction to active sand dunes (Fig. 28A–B). Further differences are given in Table 1. The distinctiveness of K. toelkenii was probably first recognised in the mid 1980s by the late Mr Derek Gosling of Whakatane who cultivated it, while the unusual ecology of the species was first noted and described in detail by Smale (1994 as Kunzea ericoides var. ericoides).

Kunzea toelkenii is distinguished from the other New Zealand members of the K. ericoides complex by its unique growth habit, in particular the spiralled, often tortured multi-trunked (Fig. 28C) growth habit, and ability of the trunk base to produce numerous, completely prostrate, widely spreading epicormic branches. This growth habit has previously been interpreted as habitat induced (Smale 1994). However, cultivation trials initiated by Mr Gosling showed that the distinctive multi-trunked shrub habit has a genetic basis. In cultivation, seedlings are completely prostrate but within 2–4 years of germination most develop one or more suberect trunks. Interestingly, Smale (1994) described the same development (what he termed ‘semi-prostrate candelabra’ habit) in specimens as ‘early as 6 years’ of age. The ability of K. toelkenii to sucker from the trunk base is similar to the growth habit of five of the seven endemic Australian members of the K. ericoides complex (de Lange 2007). However, while these Australian species all possess a distinctly bulbous, fire-resistant lignotuber (with at least one of these species, K. leptospermoides F.Muell. ex Miq. also having a rhizomatous habit), such structures are absent in K. toelkenii.

The branchlet indumentum of K. toelkenii is also distinctive, comprising mixtures of sparse, long antrorse-appressed, somewhat flexuose hairs, and more numerous, somewhat shorter, divergent, often curled and spiralled hairs (Fig. 27C). No other New Zealand Kunzea species has such distinctive curled and spiralled divergent hairs, although mixtures of appressed and divergent hair types is a feature common to the hybrid Kunzea robusta × K. tenuicaulis. Although a hybrid origin for K. toelkenii seems likely, and experimental examples of Kunzea robusta × K. tenuicaulis (AK 286145) are very similar to it, experimental hybrids of this cross lacked the curled and spiralled branchlet hairs and distinctive growth habit of K. toelkenii.

Ecologically, Kunzea toelkenii is further distinguished as the only member of the K. ericoides complex truly endemic to sand dune systems (cf. K. amathicola described later). Within its sand dune habitat K. toelkenii is known to occur sympatrically only with K. robusta and even then scarcely so (e.g., Coastlands, Whakatane (P. B. Cashmore s.n., (AK 300902)) and Whangakopikopiko (Tern Islet), Ohiwa Harbour (P. J. de Lange 7248 & P. B. Cashmore, (AK 301683)). However, this pattern more probably reflects past land clearance patterns because elsewhere in the Bay of Plenty K. robusta is widespread, with a range that extends to sand dunes, e.g., Waihi Beach, Matakana Island.

A final peculiarity is the tendency of K. toelkenii, uniquely amongst the New Zealand species, to produce functionally male flowers (Fig. 28F). In most cases these flowers had almost vestigial, non-functional stigmas, though occasionally even these are absent. Such flowers have been described for New Zealand examples of Leptospermum scoparium (Primack and Lloyd 1980) and are now known from at least five other Australian species of that genus (Andersen 1990; O’Brien 1994) but as far as I am aware they have not been reported previously for Kunzea. As with the Leptospermum examples studied (see in particular Andersen 1990), functionally male flowers appear toward the end of the flowering season. In K. toelkenii they appear to be consistently produced in wild and cultivated plants, though in varying degrees and not necessarily on every plant.

Kunzea toelkenii has the same ITS and ETS sequences (Table 2) as K. serotina (de Lange 2007; de Lange et al. 2010). No variable sites are present in the ITS sequence, while the ETS sequence of both species, together with K. ericoides, K. robusta (Mt Egmont samples only) K. salterae and K. tenuicaulis, share a guanine/cytosine mix at ETS alignment position 232 (de Lange 2007). Otherwise the ETS asequence of K. toelkenii shares an adenine nucleotide with K. salterae, K. serotina, and K. tenuicaulis (de Lange 2007).

The ecology of K. toelkenii was described in detail by Smale (1994 as Kunzea ericoides var. ericoides). The extant consolidated and semi-consolidated foredune and dune swale habitats of K. toelkenii near Thornton are estimated to be less than 700 years old (Pullar and Selby 1971), while Whangakopikopiko (Tern Islet), a barrier-spit island at the Ohiwa Harbour mouth, is probably even younger. Kunzea toelkenii is the dominant woody species within these habitats and associated species are scarce. Smale (1994) recorded 17 vascular flora associates from the Thornton site, the majority of which were exotic naturalised species, and only five were woody trees or shrubs. Of the naturalised species at Thornton, Smale (1994) identified boxthorn (Lycium ferocissimum Miers) as potentially invasive and a possible threat to K. toelkenii. He also observed that the K. toelkenii population appeared to form a single distinct cohort with a mean age of 45 years, though specimens up to 70 years of age were occasionally encountered, and that the stands sampled appeared to have arisen through invasion of open sand dune vegetation, possibly after fire had removed the previous vegetation cover. Smale (1994) was of the opinion that the Kunzea stands (here K. toelkenii) would replace themselves indefinitely, partly because Thornton is so isolated from other seed sources of potential successors. A similar, though less dense population to that seen at Thornton is present at Whangakopikopiko (Tern Islet), and there the impression is of a very recent establishment. If so, this suggests that at some stage K. toelkenii was present on the adjoining sand spits of Ohope and Ohiwa, all now extensively developed and housed, and from where only scattered individuals of K. robusta are now known.

Within its habitat K. toelkenii very rarely associates with other Kunzea species and thus far wild hybrids have not been found. Nevertheless experimental hybrids were readily produced using K. toelkenii as pistillate or staminate parent (de Lange et al. 2005, as Kunzea aff. ericoides (d)).

No specific vernacular appears to be in use for K. toelkenii.