(here designated) (Fig. 44). Leptospermum sinclairii T.Kirk, Stud. Fl. N.Z. Mount Young, T. Kirk 959, 20/11/67, WELT SP029323! (piece labelled ‘A’)

Paralectotypes (here designated). WELT SP029323! (piece labelled ‘B’), WELT SP029323C! (piece labelled ‘C’), WELT SP29321!, WELT SP29322!

Kirk visited Aotea (Great Barrier Island) between 15 November and 19 December 1867 (Hamlin 1965) during which time he visited a number of locations where Kunzea sinclairii (≡ Leptospermum sinclairii) is known to occur. In his protologue (1899; p. 158) he gave a brief description, attributing its original discovery to Dr Andrew Sinclair, and noted its distribution as ‘NORTH Island [sic]: Great Barrier Island, Hutton and Kirk. Sea Level to 1,800 ft’ but he did not specify locations or specimens. Allan (1961; p. 323) typified the name by his statement ‘Type: W[ELT], Herb. Kirk, Hutton and Kirk’. This action constitutes lectotypification under Article 9 of the International Code of Nomenclature (McNeill et al. 2012). However, there are no specimens in WELT collected jointly by Hutton and Kirk, but there are five specimens lodged there, and labelled by Kirk as Leptospermum sinclairii. Four of these were collected by Kirk and one possibly by Hutton (Fig. 45). Therefore I suggest that Allan’s typification was based either on a literal interpretation that Kirk (1899; p. 158) was citing a jointly collected specimen or that he was repeating Kirk’s protologue and indicating where type material might be found (which accords with remarks on typification made in the introduction to Allan 1961). Either way, Kirk’s statement may also have meant that he and Hutton collected specimens, but not necessarily together. Because of this uncertainty and because Allan did not specify which of the five specimens in WELT is the Lectotype, I regard his lectotypification as incomplete.

A major problem in typifying Kirk taxa named in the Students’ Flora is that this publication, was published some six years after Kirk’s death and is incomplete, comprising mostly those sections regarded as finished by an unknown individual or individuals tasked with assembling Kirk’s Students’ Flora for the then Education Department, Wellington, New Zealand in 1899 (see ‘Introductory Note’ in Kirk 1899). This has meant that typification of taxa published within the Students’ Flora requires thorough examination of the exact wording used by Kirk in each instance, in conjunction with the specimens he used (in this regard see also comments with respect to the typification of Lepidium oleraceum varieties established by Kirk (1899) in de Lange et al. (2013a)). Although Kirk had no type concept for his New Zealand Flora treatment, he was obviously aware of the need to indicate which specimens were representative of the taxa he was naming or treating for his flora. From critical study of his herbarium material, it seems that for his Student Flora of New Zealand, Kirk showed this by annotating those specimens he had finished working with ‘Stud. Fl. N.Z.’ (Fig. 44), and, as a further measure, somewhat unorthodoxly discarding the original label details, an action which often included effacing the actual collector’s name in whose place he usually wrote his own (see comments by de Lange and Gardner 2002; de Lange et al. 2013a). Therefore, I have used this pattern of labelling as an indication of his intent when naming new taxa.

Of the five specimens labelled by Kirk as “Leptospermum sinclairii”, and gathered from Mt Young, only three, WELT SP029321 (T. Kirk 960), WELT SP029322 (T. Kirk 958) and WELT SP029323 (T. Kirk 959), are clearly annotated ‘Leptospermum sinclairii T.Kirk Stud. Fl. N.Z.’. On each sheet there are several pieces, some of sufficient morphological difference to have come from different plants. In all cases one piece on each sheet has a small jeweller’s label attached to it bearing a collection number written by Kirk that matches the number on his specimen labels. WELT SP044298 from the ‘Private Herb. Kirk’ is the only specimen to bear a label in what seems to be Captain F. W. Hutton’s handwriting (Figs 45, 46A–B). Hutton’s assumed label is on blue paper and is mounted on the left hand side of the sheet. It is written in pencil and reads ‘Leptospermum ericoides prostrate hoary form M^t Young 19/11/67 Captⁿ Hutton’. Kirk has annotated the label in Indian ink, with ‘v. Sinclairii’, changed the date to ‘20/11/67’ and crossed out Hutton’s name (Fig. 46A). A second label (Fig. 46B) written by Kirk in Indian ink on the paper he usually used, is mounted to the right of Hutton’s label and reads ‘Leptospermum Sinclairii Kiwiriki, Mount Young, Great Barrier Isd T. K. 20/11/67'The name ‘Kiwiriki’ though legible, is struck out, and ‘Mount Young’, the date and Kirk’s initials have been added at a later date as is evident by the lighter coloured ink. The final specimen, WELT SP044299, also from the ‘Private Herb. Kirk’ is labelled by Kirk ‘Leptospermum Sinclairii Great Barrier Island 20/11/67’. Significantly all five sheets are mounted on paper different from that on which the original Kirk labels were written. This probably happened when what has come to be called the ‘Kirk Herbarium’, and which was originally held loosely in newspapers (N. M. Adams and F. Pitt pers. comm.) was eventually mounted, a process that was undertaken at various times from the 1930s onwards, and often by summer students employed for the task (J. E. Braggins pers. comm.). Thus it is possible that some specimen labels may not necessarily match the associated plant specimens. This is a major problem when undertaking typifications. It could also explain the different labels on the specimen I have attributed to Hutton, but is something I have been unable to resolve.

Under these circumstances, and in consideration of all the available evidence, I designate one element of WELT SP029323 (T. Kirk 959) the lectotype of Leptospermum sinclairii Kirk (Fig. 44). WELT SP029323 matches Kirk protologue as to location, condition and collector, is one of three gatherings labelled by Kirk in his hand ‘Leptospermum sinclairii T.Kirk’ and significantly it is also annotated by him ‘Stud. Fl. N.Z.’ showing in my opinion (see above) that Kirk considered it representative of his new species for his publication. Because the specimen comprises three pieces (which I have labelled in pencil A, B, and C) and due to the history of the curation of the ‘Kirk Herbarium’, which was received, largely uncurated in several consignments dating from Thomas Kirk’s death in 1898 until the 1930s (N. M. Adams and F. Pitt pers. comm.) I cannot be sure if these three pieces came from the same plant or even the same gathering. I also harbour grave doubts over the validity of the claim that those Kirk specimens currently held in WELT constitute the ‘Kirk Herbarium’ as Kirk’s collections, including a great many specimens that may be considered as types occur throughout the world’s herbaria, and it was Kirk’s practice as a private collector (in modern terms a ‘botanical consultant’) to carry his herbarium with him from which he routinely parcelled up and gifted (or traded) examples of his new species (effectively type material) to colleagues and herbaria throughout the world. Therefore, one can never sure if Kirk collections held at WELT are truly the syntypes on which he founded his names, or even parts thereof. Ideally, typification of Kirk names requires the reassembling of his specimens which are usually scattered worldwide—a difficult prospect indeed. Thus, for all these reasons, I conservatively designate the piece labelled ‘A’, and bearing a small jeweller’s label numbered ‘959’ in Kirk’s hand, as the lectotype because that number is cross-referenced to Kirk’s specimen label mounted on the same sheet. The other two pieces I designate paralectotypes because they are part of the type collection but because of the way in which the Kirk herbarium was curated, their relationship to the lectotype is unclear. Similarly WELT SP029322 (T. Kirk 958), WELT SP029321 (T. Kirk 960) are designated paralectotypes because they are part of the type collection matching the protologue as to collector and/or they show the naming author’s intent with regard to publication of the name. WELT SP044299, although labelled by Kirk has no reference to his place of publication (Stud. Fl. N.Z.). Therefore I exclude it from this typification. Further, because of the uncertainty over the labels, I also exclude WELT SP044298 (Fig. 45) from this typification because I cannot be sure that Hutton’s adulterated label or even, for that matter, Kirk’s label are related to the same gathering, and, as Hutton’s label carries Kirk’s manuscript name ‘v. Sinclairii’ (Fig. 46A) and Kirk’s label is not annotated ‘Stud. Fl. N.Z.’, I believe Kirk did not intend this specimen to be ‘representative’ of his new species. WELT SP029316 in ‘Herb. D. Petrie’ also deserves mention. This collection comprises a small flowering piece of Kunzea sinclairii and bears an ‘Herb. T. F. Cheeseman’ label which is written on by both Cheeseman and Petrie. Cheeseman’s Indian ink label reads ‘Leptospermum Sinclairii T. Kirk, Great Barrier Island, T. Kirk’. To this Petrie has annotated the top of the label in blue ink ‘the Dominion (the word ‘Dominion’ is struck out but legible) Petrie Herbarium’ and at the bottom he has written ‘Dominion Museum, Wellington’. Attached to the stem of the specimen is a small paper slip also labelled by Petrie in blue ink ‘Type T. Kirk’. As there is nothing to associate this specimen with the original Kirk type collections I regard WELT SP029316 as having no nomenclatural status.

The specific epithet sinclairii honours Dr Andrew Sinclair (1794–1861) who, according to Kirk (1869, 1899), first discovered the species during a brief visit to the island. I have been unable to determine how Kirk came by that information, or when Sinclair’s visit to Aotea (Great Barrier Island) happened, or why such an assiduous plant collector as Sinclair, so prominent in the early annals of early New Zealand botany (Hooker 1867; Cheeseman 1906, 1925), failed to procure specimens.

= Leptospermum ericoides var. pubescens Kirk in T.P.N.Z.I. 1 (1869), 146–147.

(here designated) (Fig. 47). ‘Leptospermum ericoides A.Rich. v. pubescens - Great Barrier Island. T.K[irk]. Sometimes 3ft high - but usually prostrate, sometimes closely appressed to the rocks. Flowers fragrant in immense profusion’. AK 5515! Paralectotypes (here designated). AK 11437!, K (T. Kirk 176)!, WELT SP029465!

In his paper on the botany of Aotea (Great Barrier Island), Kirk (1869, pp. 146–147) published a brief description of Leptospermum ericoides var. pubescens which is given here in full. ‘Leptospermum ericoides A. Rich. var. pubescens – A prostrate or sub-erect shrub, sometimes 3 feet high, at others appressed to the rock, like an alpine plant; leaves more or less pubescent and ciliate; flowers fragrant, produced in immense profusion, sometimes concealing the leaves; pedicels and calyx downy. This would make a valuable bedding plant for the culturist. It was originally observed on the island by the late Dr Sinclair, but I am not aware of its occurrence elsewhere’. Later in the same publication Kirk provides a list of the vascular plants he saw on the island, in which he records it as ‘Leptospermum ericoides var.’ (Kirk 1869; p. 150). Although no specimens were cited or a location given, Kirk’s description is both valid and effective under the terms and conditions of the International Code of Nomenclature (McNeill et al. 2012). There are at least three herbarium specimens that are annotated by Kirk ‘Leptospermum ericoides var. pubescens’, WELT SP029465!, AK 5515!, and AK 11437! These I regard as part of his type collections. A further gathering held at K! (T. Kirk 176) is also labelled ‘Leptospermum ericoides var. pubescens’, though not in Kirk’s hand although it is clear it was sent by Kirk to Kew as a gift because it carries the same Kirk collection number (176) as AK 11437. I therefore regard it also as part of his original type collection. Indeed it is also the only clear duplicate among these specimens. One of the AK specimens, T. Kirk s.n. AK 5515 (Fig. 47), was used as the basis for the illustration of Leptospermum sinclairii by Kew botanical illustrator Matilda Smith in Cheeseman (1914, pl. 47). This sheet carries three labels; the first is by Kirk which reads ‘Leptospermum ericoides A.Rich. v. pubescens - Great Barrier Island. T.K. Sometimes 3ft high - but usually prostrate, sometimes closely appressed to the rocks. Flowers fragrant in immense profusion’ (Fig. 48A). The second (Fig. 48B), written by Cheeseman, states ‘Leptospermum sinclairii Hk.f. Great Barrier Island’, and the third, also in Cheeseman’s hand (Fig. 48C), records its use for his illustrated flora (Cheeseman 1914) thus: ‘Illustrations of the N.Z. Flora, Plate 44 Leptospermum sinclairii T. Kirk. Great Barrier Island, T. Kirk’ (the image ultimately became Plate 47). I designate this sheet Lectotype of Leptospermum ericoides var. pubescens Kirk because Kirk’s label details match his protologue (Kirk 1869; P. 146–147) more closely than the other specimens available, with respect to plant height (‘3 feet’), its growth habit (‘appressed to rocks’), flower scent (‘fragrant’) and abundance (‘in immense profusion’). Although some may feel it is desirable to select a specimen from the naming author’s herbarium, Kirk at the time of his death (8 March 1898) had no official herbarium (Brown 1968). Further, as I have noted above, the subsequent placement of the majority of his collections in WELT as the ‘Kirk Herbarium’ does not by default indicate that this is, therefore, the Kirk herbarium. The ‘Kirk Herbarium’ at WELT came about partly because of the circumstances surrounding Kirk’s death during the preparation of the Student Flora of New Zealand (see Introductory Note, Kirk (1899)) and the later desire of his descendants to gift what they had retained of their father’s herbarium (distinguished at WELT as the ‘Private Herb. T. Kirk’) to the place where he was mostly working at the time of his death, WELT (Moore 1973; F. Pitt and J. Fox pers. comm.).

The remaining specimens (AK 11437, K (T. Kirk 176), WELT SP29465) I designate Paralectotypes. This is because, although part of the type collection, I cannot determine what date they were collected, or even whether they came from the same plant.

The varietal epithet pubescens, though not elaborated on by Kirk (1869), probably refers to leaves, which he described in his paper as ‘more or less pubescent’.

(Figs 49, 50, 51). Growth habit mostly decumbent, trailing, silvery grey to grey, reddish-grey or grey-green, shrubs up to 3 × 1 m, very rarely forming a small tree up to 6 m tall; irrespective of stature, branches widely spreading and densely leafy, sometimes rooting on contact with soil or rock. Trunk 1(–4 or more), usually shortly erect between 0.2–1.0 m tall before branching but sometimes indistinguishable due to branches arising at ground level, 0.05–0.12(–0.16) m d.b.h.; basal portion of trunks covered with layers of somewhat firm to loose, stringy, pale grey to light brown chartaceous bark. Bark early bark dark brown to brown; firmly stringy, chartaceous to subcoriaceous, ± elongate, lying in numerous overlapping strips; usually bearing a few transverse and many longitudinal cracks (especially on branch flanges and decurrent leaf bases), otherwise firmly attached; margins elongate, sinuous, ± entire with scarcely any flaking; old bark initially dark brown to grey-brown, soon becoming covered in crustose lichens and sparse liverwort growth; coarsely stringy to tessellated and distinctly corky-coriaceous, usually remaining firmly attached, if detaching, then usually doing so along transverse cracks; flakes usually centrally attached, margins somewhat tabular with entire margins and coarsely frayed apices; upper surface of bark flakes coarsely tessellated, sometimes flaking secondarily as small tabular shards; upper trunk bark breaking into smaller pieces in hand but not crumbling. Branches numerous, usually present from close to or at trunk base, prostrate and widely spreading, new growth subscandent (in rare tree forms this habit is retained resulting in arching, pendulous branches); branchlets numerous, widely spreading to subscandent, often coarsely interwoven, initially red, ± quadrangular to subterete, leaves usually densely crowded along stems and brachyblasts, though in vigorous new growth sometimes widely spaced; branchlets sericeous, indumentum copious, silky, hairs antrorse-appressed, weakly flexuose up to 0.06 mm long, hyaline to translucent (appearing silvery-white when young, maturing silver-grey). Vegetative buds inconspicuous, usually obscured from view by surrounding leaves; at resting stage 0.3–0.8 mm diam. narrowly ovoid to ellipsoid; scales deciduous; (0.3–)1.2 mm long, pale yellow-brown to reddish brown, broadly to narrowly ovate-lanceolate grading through lanceolate to narrowly lanceolate; midrib ± keeled, sometimes prolonged to apiculate tip, otherwise apex obtuse to subacute or acute; oil glands inconspicuous, sparse, scattered in irregular lines either side of midrib; lamina initially completely invested by long silky silvery-white hairs, becoming glabrate, with hairs progressively confined to scale margins, midrib, and keel prolongation (if present). Leaves heterophyllous, weakly to strongly spicy-scented when crushed, mostly sessile, sometimes shortly petiolate (up to 1.6 mm long). Seedling and juvenile leaves dark green to glaucous, glabrous up to 25.0 × 3.5 mm, oblanceolate to lanceolate, apex acute, often shortly mucronate, base attenuate, lateral veins evident, especially on seedling leaves, both surfaces distinctly gland-dotted, oil glands up to 480 on either surface. Mature leaves soon developing (depending on degree of exposure), densely crowded along branchlets and brachyblasts, particularly toward apices, initially obliquely ascending, soon suberect to widely spreading usually weakly recurved in distal 30%; lamina (5.6–)14.5(–20.6) × (2.0–)3.2(–4.5) mm, initially appearing silvery-white (due to dense hair covering), maturing silvery-grey to reddish grey (as some hairs are shed), usually paler beneath; lamina broadly lanceolate, elliptic to obovate, rarely oblong-obovate, apex sharply acute, often cuspidate, base attenuate; adaxial lamina surface flat to weakly concave, glandular punctate, with oil glands scarcely evident when fresh due to dense hair covering, becoming more obvious in old leaves and in dried specimens, up to c.380, midrib slightly raised for c. 70% of leaf length; abaxial surface flat to weakly convex, usually densely covered in hairs, sometimes glabrate in old leaves, glandular punctate, oil glands up to 300; midrib raised for entire length; lamina margins distinctly less hairy than lamina surface; hairs of midribs and margins converging at leaf apex. Perules deciduous (shedding very early in inflorescence maturation), (0.8–)1.2(–1.4) × (0.8–)1.0(–1.2) mm, orange brown to amber with a broad pale brown margin (this reducing in thickness toward apex); broadly ovate grading through to ovate-lanceolate, apex cuspidate; lamina 6–8-nerved with poorly defined midrib and bearing up to 10–20 oil glands between nerves; lamina surface initially sparsely covered in deciduous long silky silvery-white hairs, soon becoming glabrate, except for a distinct stout weft on the cuspidate apex. Inflorescence mostly a compact, corymbiform (4–)9(–20)-flowered botryum 7.0–20.0 mm long and usually terminated by a tuft of leaves and a semi-dormant vegetative bud; inflorescences initially present on brachyblasts in the distal one-third of the active branchlets, increasing in abundance and soon dominating all the distal terminal and lateral growth; on occasion inflorescences may extend to elongated botrya on late season’s vegetative growth. Inflorescence axis densely invested with antrorse-appressed, weakly flexuose, silky hairs. Pherophylls deciduous (shed early during bud maturation), rarely present at flowering, foliose or squamiform, basal portion tightly clasping pedicel base; (1.0–)1.2 × (0.2–)0.4 mm; foliose pherophylls pale green to red-green, oblong to oblong-lanceolate, very rarely broadly spathulate, cuspidate, deeply concave in cross section, with the abaxial surface copiously invested in sericeous, antrorse-appressed hairs; oil glands scarcely evident, up to 10 (usually less); midrib not evident; squamiform pherophylls tightly clasping pedicels, 0.3–1.0 × 0.4–0.8 mm, red-brown to brown, broadly to narrowly ovate or lanceolate, apex acute, subacute to obtuse, weakly keeled, margins and distal portion of keel finely ciliate. Pedicels (2.8–)5.7(–7.3) mm long at anthesis, scarcely elongating after anthesis, terete, initially invested with silky, antrorse-appressed, weakly flexuose, hairs becoming glabrate. Flower buds (2.3–)3.8(–4.9) × (2.1–) 3.1(–4.2) mm, ovoid to pyriform, apex flat to weakly domed prior to bud burst with calyx lobes held flat across surface, rarely meeting. Fresh flowers when fully expanded (5.7–)8.1(–10.2) mm diam. Hypanthium (1.9–)2.6(–3.6) × (2.1–)3.1(–4.2) mm, with free portion 0.4–0.7 mm long, silvery-white to silvery grey or reddish-grey due to copious covering of hairs; narrowly obconic to obconic or cupular, terminating in a scarcely defined chartaceous rim bearing five persistent calyx lobes; hypanthium surface smooth when fresh becoming irregularly wrinkled when dry, somewhat finely glandular punctate, oil glands scarcely evident due to dense covering of long, silky, antrorse-appressed silvery hairs, ribs not evident. Calyx lobes 5, initially erect to suberect, sometimes spreading, submembranous, (1.1–)1.3(–1.6) × (0.9–)1.2(–1.8) mm, broadly obtuse, red-green to pale green with a white or pink membranous margin, not obviously keeled, sparsely and finely gland-dotted, oil glands ± colourless; lobe margins finely ciliate, hairs eglandular, central portion of lobes densely covered in short silky, antrorse-appressed hairs. Receptacle greenish pink or pink at anthesis, darkening to crimson after fertilisation. Petals 5(–6), (2.0–)2.9(–3.6) × (2.1–)2.7(–3.3) mm, white, very rarely basally flushed pink, broadly ovate, suborbicular to orbicular, rarely ± cuneate-truncate, spreading, upper third often weakly recurved, apex rounded, margins ± finely and irregularly crumpled or frayed, oil glands not evident in fresh or dried material. Stamens 18–38(–46) in 1–2 weakly defined whorls, arising from receptacular rim, filaments white. Antipetalous stamens (2–)3(–4), antisepalous (2–)4(–6). Outermost antipetalous stamens outcurved, widely spreading, more rarely slightly incurved, on filaments 2.0–3.6 mm long, inner stamen if present, 0.4–0.9 mm, outcurved or incurved. Antisepalous stamens usually shorter than outermost antipetalous stamens, 0.6–3.6 mm, weakly incurved or outcurved, usually in mixtures of both. Anthers dorsifixed, 0.06–0.1 × 0.06–0.09 mm, broadly ellipsoid to scutiform, latrorse. Pollen white (11.9–)15.4(–19.9) μm. Anther connective gland pale pink when fresh, drying pale orange, spheroidal, coarsely papillate. Ovary (3–)4(–5) locular, each with 18–22(–34) ovules in two rows on each placental lobe. Style 1.8–2.5(–3.0) mm long at anthesis, elongating slightly after anthesis, white basally flushed pink or pale pink; stigma narrowly capitate, as wide as or scarcely wider than style, ± flat, greenish-pink or pink, flushing red after anthesis, surface finely granular-papillate. Fruits long persistent, copiously covered in short, silky, antrorse-appressed hairs; (2.2–)3.0(–3.6) × (2.7–)3.2(–3.9) mm, initially graphite grey, maturing to charcoal at dehiscence, and in old dehisced capsules fading to greyish-white; narrowly obconic to obconic, rarely cupular, calyx valves persistent, incurved, somewhat chartaceous, splits concealed by dried, erect, free portion of hypanthium. Seeds 0.52–1.04(–1.09) × 0.38–0.58(–0.72) mm, obovoid, oblong, or oblong-ellipsoid, usually curving toward 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, surface coarsely reticulate. FL: (Sep–)Nov–Jan(–Mar). FT: Feb–May(–Jul). Chromosome Number n = 11_II, 2n = 22 (see de Lange and Murray 2004).

(111 sheets seen). Aotea (Great Barrier Island). Kaiarara Plateau, R. C. Lloyd s.n., Nov 1950, (CHR 92451); Perry's Hill, R. C. Lloyd s.n., Oct 1973, (AK 133455, 133456); South slope of Mt Hobson, C. C. Ogle 470, 13 Apr 1980, (CHR 367190); Mt Hobson - Windy Canyon Track, P. J. de Lange 4537 & B. G. Murray, 7 Mar 1999, (AK 287195, Duplicates: CANB, HO, NSW, Z); Overton (Harataonga) Access Road, P. J. de Lange 4562 & G. M. Crowcroft, 7 Oct 2000, (AK 255943, Duplicate: AD); Mt Heale, Eastern Side, P. J. de Lange 4867, 18 Dec 2000, (AK 286128, Duplicates: AD, MEL, P); Peach Tree Track, P. J. de Lange 6266 & D. A. Norton, 18 Apr 2002, (AK 288493, Duplicate: CHR); Motukaikoura Island, North side of main ridge, south-west of Bradshaw Cove, E. K. Cameron 14105, 17 Dec 2006 (AK 298033).

(Fig. 7). Endemic, New Zealand, Aotea (Great Barrier Island) (20–510 m a.s.l.). On Aotea (Great Barrier Island) found mostly on the rhyolite outcrops of the central high points and western slopes (de Lange and Norton 2004).

As discussed under Kunzea triregensis, there has been much confusion over the exact circumscription and identification Kunzea sinclairii. Nevertheless, as defined here Kunzea sinclairii is endemic to Aotea (Great Barrier Island). It differs from all other New Zealand Kunzea by its usually prostrate trailing habit (Figs 49A, 51C), and its consistently silvery-grey, uniformly hairy leaves (see Table 1). It is also the only Kunzea endemic to rhyolitic rock (de Lange and Norton 2004). In the field, Kunzea sinclairii is sympatric (and often syntopic) with Kunzea robusta and, occasionally, Kunzea linearis, particularly along track sides, forestry roads, and formerly forested, now open clay pans and associated rock exposures. In these places it forms extensive hybrid swarms, mostly with Kunzea robusta but also, rarely, with Kunzea linearis and, apparently uniquely amongst New Zealand Kunzea, with Leptospermum scoparium (de Lange et al. 2005).

Kirk (1869) quickly recognised the novelty of Kunzea sinclairii. Many of his herbarium specimens are annotated with detailed enthusiastic descriptions of the species. For example, T. Kirk 75 (K!) carries the following observations ‘suckering or prostrate – producing its flowers in such immense abundance as often to hide the leaves. Often closely pressed to the rocks like an alpine plant and resembling in foliage Phlox verna! & [sic] might easily be taken for an herbaceous plant – small patches from 3 inches broad in the faces of the rocks were simply patches of flowers. Flowers fragrant’. While on the label of a specimen in MEL(!) (T. Kirk 399) Kirk waxes lyrical about the flowers, their profusion and concludes that the species is ‘a capital bedding plant! However, as is evident from his herbarium specimens at WELT and those he lodged overseas, he remained uncertain as to what rank seemed most appropriate for the plant until his final 1899 posthumous publication of it at species rank (Kirk 1899).

This uncertainity is reflected in Kirk’s vague distinctions. For example in his original diagnosis (Kirk 1869) where he treated the species as Leptospermum ericoides var. pubescens, he provided few distinguishing characters. His subsequent diagnosis of Leptospermum sinclairii (Kirk 1899; p. 158), although more detailed, still offered few discriminating characters (‘...distinguished at sight by the white silky leaves, larger flowers, longer pedicels. The ovary is sunk fully one-third below the narrow calyx-tube, while the sepals and petals are narrower, and the style extremely slender. The flowers are deliciously fragrant.’). This admittedly vague diagnosis may have prompted Cheeseman’s standard cynical remark, which he used ‘politely’ when dealing with taxa he thought dubious (see de Lange et al. 2014) that it was ‘a somewhat critical’ species (Cheeseman 1914). Whatever his misgivings, Cheeseman nevertheless still accepted the species, as did all subsequent authors until the perfunctory treatment of Thompson (1983) who merged it (along with several other distinct Australia species and New Zealand varieties) in the oldest available name Leptospermum ericoides A.Rich. as Kunzea ericoides (A.Rich.) Joy. Thomps.

Harris (1987) did not agree with the circumscription of Kunzea ericoides offered by Thompson (1983) and he made the necessary combination at species rank in Kunzea for Leptospermum sinclairii. However, in that account he again offered little basis for that decision. Subsequently, Harris et al. (1992) discussed the species in more detail, though, oddly, as neither he nor his co-authors had seen the species in the wild, their perceptions of its ecology, and statements about its growth habit were based entirely on the behaviour of cultivated specimens supplemented by field observations gleaned from herbarium labels and comments made by other people who had actually seen it in the wild (B. P. J. Molloy pers. comm.). Although their conclusions as to rank are accepted here, I provide a more detailed description of the species based on direct field observations and wild collected specimens. Unique defining characters of Kunzea sinclairii include its usually prostrate widely spreading habit (Figs 49A, 51C–E), tendency for the branches to layer on contact with the soil, broadly lanceolate, elliptic to obovate, uniformly sericeous hairy leaves (Fig. 49E–G), and the unusual inflorescence type, which is an aggregated corymbiform conflorescence (see Briggs and Johnson 1979). The leaves and young branchlets of Kunzea sinclairii differ from all other New Zealand Kunzea in that they are often strongly red-pigmented, particularly along the lamina margins, and it is this dark colour which, when overlaid by the copious light reflecting antrorse-appressed hairs (Fig. 49A–D), helps to impart the characteristic silvery colour of the plant. As Kirk (1869, 1899) observed, the new growth appears white, due to the dense covering of hairs, which are partially shed as the leaf matures, thereby causing the white ‘colour’ to fade first to silver, and ultimately to the more usual silver-grey typical of mature foliage. Although the flowers of Kunzea sinclairii were referred to by Kirk as larger than Kunzea ericoides, borne on longer pedicels and ‘deliciously fragrant’ is not confirmed. In fact the flower size of Kunzea sinclairii falls within the range commonly seen in Kunzea robusta, Kunzea triregensis and Kunzea amathicola (see Table 1), species which, in Kirk’s time were included in Kunzea ericoides. Nevertheless, they are larger than the range usually seen in Kunzea ericoides s.s. (see Table 1). Similarly, pedicel length is not diagnostic, the length varying with the growing conditions of the plant. Finally, the majority of the fresh flowering specimens of Kunzea sinclairii, in common with the rest of Kunzea ericoides group treated here, had little or no discernible scent. However, in one respect the flowers of Kunzea sinclairii are distinctive: the external hypanthium surface, calyx-lobes and receptacular disc are usually strongly pigmented red or pink, a background colour which helps highlight the copious covering of silky hairs on the external face of the hypanthium.

The ITS and ETS sequence data obtained for Kunzea sinclairii is unremarkable (de Lange 2007; de Lange et al. 2010). A relationship to Kunzea amathicola and Kunzea robusta is inferred in the ETS sequence at alignment position 269 (Table 2) where all three taxa share a guanine (de Lange 2007). Kunzea sinclairii also has the same chromosome karyotype as that seen in Kunzea amathicola and Kunzea triregensis (de Lange and Murray 2004). Irrespective, Kunzea sinclairii is easily distinguished from all three species (Table 1). The confusion of this species with Kunzea triregensis has already been discussed (see above). More puzzling by far is that Kunzea sinclairii is frequently confused with the very different with Kunzea linearis. Notably, Webb and Simpson (2001; p. 239, pl. 96), despite accurately recording the distribution of Kunzea sinclairii, based their seed description, illustration and voucher on a Bay of Islands specimen of Kunzea linearis (WELT SP029320!) that had been incorrectly identified by its collector, D. Petrie as Kunzea sinclairii. Kunzea sinclairii is easily distinguished from Kunzea linearis by its usually prostrate, trailing rather than erect tree habit, and uniformly hairy, silvery-grey, broadly lanceolate, oblanceolate to elliptic leaves (Fig. 48A–B, D–G, 51B–E). The flowers of Kunzea sinclairii are produced in mostly corymbiform rather than spiciform inflorescences (Fig. 48A, 51D–E), and are distinctly pedicellate rather than sessile to subsessile. The pherophylls are deciduous, usually short, broadly rhomboid, elliptic to lanceolate rather than long, linear to linear falcate, and persistent (as in Kunzea linearis). Further details are provided in Table 1.

In its natural habitat the only Kunzea commonly found in association with Kunzea sinclairii is Kunzea robusta. Because Kunzea sinclairii is usually a prostrate shrub, and has uniformly, silvery-grey hairy leaves, confusion with the usually arborescent (up to 30 m tall) dark to light green, glabrescent leaved Kunzea robusta is unlikely. Additional distinctions are provided in Table 1.

The ecology of Kunzea sinclairii is described by de Lange and Norton (2004). They found that the species is naturally confined to rhyolitic rock of the central Hirakimata massif of Aotea (Great Barrier Island). Here it is primarily a rock tor, cliff and gorge endemic (Fig. 51A–C) occupying an estimated 90.5 ha (or 0.3% of the extent of Aotea (Great Barrier Island).

Kunzea sinclairii is usually the dominant woody plant within its preferred rock habitat. It rarely grows on well-developed soils, preferring fissures within rock outcrops, and their associated saprolite and/or skeletal soils (Fig. 51A–C). These habitats are characterised by being extremely infertile and with low species diversity (Harris et al. 1992; de Lange and Norton 2004). In their study de Lange and Norton (2004) found that in the main rock habitats Kunzea sinclairii generally grew in association with stunted kahikatoa (Leptospermum scoparium agg.), Phormium cookianum subsp. hookeri (Hook.f.) Wardle, and another Aotea (Great Barrier Island) rhyolite endemic, Olearia allomii Kirk, the lichen Cladia retipora (Labill.) Nyl. and the moss Ptychomnion aciculare (Brid.) Mitt.

The bark and twigs of well-established plants of Kunzea sinclairii are often festooned with corticolous and epiphyllous liverworts, particularly from the genus Frullania, and the trunk bases and exposed roots are often encrusted by the liverworts Cuspidatula monodon (Lehm.) Steph., Jamesoniella colorata (Lehm.) Schiffn., and Harpalejeunea filicuspis (Steph.) Mizut. Occasionally the branches and branchlets of Kunzea sinclairii are heavily parasitised by the dwarf mistletoe Korthalsella salicornioides.

It was the widely reported, putative, ‘rife’ hybridism with other Kunzea that was the basis for initial conservation concerns that Kunzea sinclairii might be at risk of extinction (Dopson et al. 1999; de Lange and Norton 2004). In the wild, Kunzea sinclairii is currently commonly sympatric (and syntopic) with Kunzea robusta and, much less frequently, Kunzea linearis. It is likely that the high level of sympatry between these species is not natural, and is the direct consequence of widespread habitat destruction that resulted from kauri logging of the island (Ogden 2001; de Lange and Norton 2004). Logging and the associated deliberate and accidental burning of the forest left large areas of bare land ideally suited to the establishment of primary colonisers, including Leptospermum scoparium, Kunzea linearis, Kunzea robusta and Kunzea sinclairii. Today, Kunzea robusta and Leptospermum scoparium form the dominant woody vegetation over most of the southern two thirds of Aotea (Great Barrier Island) (Ogden 2001). Widespread and frequent disturbance favours the proliferation of hybrids in Kunzea (de Lange et al. 2005), and this is especially so on Aotea (Great Barrier Island) where the putative hybrids Kunzea linearis × Kunzea sinclairii, Kunzea robusta × Kunzea sinclairii and the intergeneric ×Kunzspermum hirakimata W.Harris (Kunzea sinclairii × Leptospermum scoparium) have all been collected. All these hybrids were successfully generated by de Lange et al. (2005) who found that Kunzea sinclairii, whether as pistillate and staminate parent, readily crossed with seven New Zealand Kunzea species, and, only as the pistillate parent, with Leptospermum scoparium. No reduction in seed viability was recorded with those crosses involving Kunzea species but the intergeneric hybrids proved to be aneuploid and sterile. They also found that the progeny raised through two generations (F₂, F₃) from their F₁ hybrid, Kunzea sinclairii × Kunzea robusta (as Kunzea sinclairii × Kunzea aff. ericoides (b) in their paper), showed clear segregation toward both parents, such that F₃ individuals were scarcely distinguishable from the original parents used for the initial cross (de Lange et al. 2005). This suggests that not only can hybridism happen readily but that introgressed swarms can segregate quickly (within 5–8 years) to their constituent parents. In their ecological study de Lange and Norton (2004) concluded that hybridisation was not a threat to Kunzea sinclairii, which they considered to form ‘pure’ populations within its preferred rhyolitic rock outcrop habitats. They argued that hybrid swarms outside these rupestral habitats were the product of past human-induced disturbance and that over time they would be lost through succession to forest. While this appears to be happening over large parts of the island, field botanists are still faced with an often baffling array of hybrids, whose distinction from Kunzea sinclairii can be difficult.

Of the three naturally occurring hybrids involving Kunzea sinclairii recorded from Aotea (Great Barrier Island), the greatest difficulty is experienced recognising Kunzea robusta × Kunzea sinclairii. This is because the hybrid is present throughout large parts of the island and the adjoining islands as a complex array of introgressants. Further, Kunzea sinclairii has been inadequately circumscribed in the past, resulting in many herbarium specimens collected as this species proving on inspection to be mixed gatherings of Kunzea sinclairii and Kunzea robusta × Kunzea sinclairii, or just the hybrid Kunzea robusta × Kunzea sinclairii.

Kunzea robusta × Kunzea sinclairii is also far more wide ranging than Kunzea sinclairii. For example, it has been collected from Aiguilles Island at the northern end of Aotea (Great Barrier Island), some 50 km north of the northern-most known Kunzea sinclairii occurrence on Aotea (Great Barrier Island). In most cases the hybrid presents an intermediate growth habit, typically as an erect shrub to small tree, characteristically with glaucous green, pinkish-green or yellowish-green, oblanceolate to lanceolate leaves. The red or pink tones typical of the branchlets and foliage of Kunzea sinclairii are expressed in the hybrids as pale pink or even yellowish-pink. This is particularly evident in herbarium specimens where the hybrids tend to dry yellow-green or dark green, never the diagnostic greyish-white to silvery-white of Kunzea sinclairii. This colour change is accentuated by differences in the density of leaf indumentum which in the hybrid tends to be much less than is typical of Kunzea sinclairii though more copious than is usual for Kunzea robusta. Another difference is that, like those of Kunzea robusta, the hairs tend to be shed more freely as the leaf matures until they are mainly found on the abaxial surface; though in some examples they can form a sparse covering on the adaxial surface, particularly in the lower half to one-third of the leaf. Because the flowers of Kunzea robusta and Kunzea sinclairii are so similar, they offer little of diagnostic value. Similarly there is little consistent difference in fruit shape or size, indumentum cover or seed size.

The other two hybrids Kunzea linearis × Kunzea sinclairii and the intergeneric ×Kunzspermum hirakimata are very uncommon. Kunzea linearis × Kunzea sinclairii is discussed under Kunzea linearis while ×Kunzspermum hirakimata was described in detail by Harris et al. (1992) and Harris (1993).

No specific Maori name for this species has been recorded.

Kunzea sinclairii is appropriately assessed as ‘At Risk/Naturally Uncommon’ qualified ‘IE’ (Island Endemic) and ‘RR’ (Range Restricted) by de Lange et al. (2013b).