Phyllogeiton trachybasis R.G.C.Boon & A.E.van Wyk, 2023

Boon, Richard G. C. & Van Wyk, Abraham E., 2023, A new species of Phyllogeiton (Rhamnaceae: Rhamneae) from Maputaland, South Africa, Phytotaxa 585 (3), pp. 193-209 : 195-206

Phyllogeiton trachybasis R.G.C.Boon & A.E.van Wyk , sp. nov. ( Figs 1–4 )

Diagnosis:— Morphologically most similar to Phyllogeiton discolor from which it differs particularly by the following characters: bark on the bole of mature trees almost black, very rough, exfoliating in large pieces; leaves fresh green and without a bloom, ± concolorous, apex acuminate; petals spreading to somewhat ascending at anthesis; fruit oblong-ellipsoid, up to 25 × 19 mm, exocarp often indistinctly longitudinally ribbed, bright orange-yellow; disc-shaped receptacle at point of fruit stalk’s attachment to base of fruit plane.

Type:— SOUTH AFRICA. KwaZulu-Natal: Hluhluwe, Bonamanzi Game Park [Reserve], game lodge car park, (2832AA/AB) [2832AB], 5 March 1996, Abbott 6929 (holotype PRU: 081091 ; isotypes NH, PCE).

“ Berchemia sp. nov. ” in Coates Palgrave (2002: 669); Boon (2010: 332); Van Wyk & Van Wyk (2013: 410); Du Randt (2018: 316). “ Berchemia sp. A ” in Schmidt (2018: 546).

Illustrations: Boon (2010: 333, Berchemia sp. nov. one photograph in second row from top, two in third row from the top); Van Wyk et al. (2011: 125, Berchemia sp. nov.); Du Randt (2018: 317, Berchemia sp. nov.); Schmidt (2018: 546, Berchemia sp. A ).

Semi-deciduous, erect tree up to 12(–15) m tall. Trunk single-stemmed and forking below 2.5 m or double-stemmed from near ground level, individual stems up to 0.55 m diam. at breast height (dbh). Bark very rough at the base of old trees, flaking in large pieces, flakes very often persistent, dark, almost black, sometimes with pale grey patches of crustose lichen, bark on upper bole and large branches thick, grey, tessellate, revealing dark bark below where peeling, fissured but fairly smooth on young branches. Heartwood (from dead branch ca. 130 mm in diam.) yellowish brown with a reddish or pinkish tinge towards the sapwood. Branchlets pendent, terete to ± longitudinally ribbed, smooth, lenticellate, initially green, becoming dark brown and then grey, glabrous. Stipules intrapetiolar, base broad, apex acute at first but fugacious, 2–3 mm long, green at first, basal remains becoming dark brown. Leaves opposite, sub-opposite (or occasionally alternate), distichous, simple, ovate to broadly elliptic, (35–)55–100(–115) × (20–)30–40(–50) mm, base rounded, apex broadly acuminate, occasionally acute, very occasionally obtuse, glabrous, firmly chartaceous, glossy, pale green at first, becoming fresh green, slightly paler but without a bloom below and very often with a yellowish tinge, margin entire, undulate, venation craspedodromous, usually slightly impressed above and slightly raised below, yellowish and paler than the lamina, principal lateral veins in (6–)7–9(–10) alternate pairs, tertiary venation minutely reticulate. Petiole canaliculate above, (7–)11–12(–20) mm long, glabrous. Inflorescences axillary (or terminal), flowers solitary or in 2–5(–7)-flowered sub-fascicles. Flowers bisexual, actinomorphic, ca. 8 mm diameter, pedicel ca. 10 mm long, receptacle flattish. Sepals 1-seriate, 5, free, ascending at anthesis, ovate to deltate, keeled on the adaxial surface, 3.0–3.5 × 2.0– 2.5 mm, pale green to yellow-green, glabrous. Petals 1-seriate, 5, free, alternate with sepals, somewhat spreading to ascending at anthesis, narrow and longitudinally folded upwards to clasp the filament, equal in length to sepals or slightly shorter, base unguiculate, apex cucullate, pale green, glabrous. Stamens 5, haplostemonous. Filaments inserted under the disc, ± equal in length to and clasped by the petals, whitish. Anthers basifixed, extrorse, whitish. Ovary superior, 2-locular, largely enveloped by the swollen, pale yellow disc, style apex 2-fid. Fruit a drupe, remains of disc and receptacle prominent at the base, oblong-ellipsoid, often faintly ribbed longitudinally, this being more pronounced in young and dry fruit, 18–25 × 16–19 mm, but often narrower (from ca. 12 mm) in dried material, receptacle at point of fruit stalk’s attachment to base of fruit disc-shaped and plane, exocarp glabrous, green ripening bright orange-yellow, mesocarp fleshy, endocarp woody, stone ellipsoid-ovoid, 15–20 × 10–12 mm.

Phenology:— Flowering is mainly in spring from late October to early December at the same time that the new leaves are produced. Fruits are present from November to May and begin ripening in March.

Distinguishing characters:— The new species is frequently confused with Phyllogeiton discolor , but the two species’ ranges are allopatric with a gap of ca. 300 km between the closest populations. Phyllogeiton trachybasis is easily distinguished from P. discolor by usually growing on dry, sandy soils (vs. usually adjacent to rivers or on termitaria), bark on the bole of mature trees almost black, very rough, exfoliating in large pieces (vs. grey-brown, rough, tessellated), branchlets and leaves glabrous (vs. glabrous or minutely and densely pubescent; on leaves pubescence mainly on the abaxial surface, but the adaxial surface may also be minutely pubescent about the midrib), leaf blade fresh green, without a bloom, ± concolorous (vs. dark green, distinctly paler with a greyish bloom below), leaf apex acuminate (vs. acute to obtuse), petals spreading to somewhat ascending at anthesis (vs. spreading to reflexed), fruit oblong-ellipsoid (vs. slightly or markedly ovoid), up to 25 × 19 mm (vs. up to 20 × 11 mm), exocarp often indistinctly longitudinally ribbed (vs. plane), bright orange-yellow, even when overripe (vs. yellow but very often pale yellow; often brownish yellow when overripe), and disc-shaped receptacle at point of fruit stalk’s attachment to base of fruit plane (vs. concave).

Etymology:— The specific epithet “ trachybasis ” is derived from the Greek τραχυς, trachys, rough or shaggy, and βασις, basis, base or pedestal, referring to the bark towards the base of the trunk in old trees that becomes very thick, rough and flaking in large pieces ( Figs 1 & 2A–D ). The epithet is used here as a noun in apposition: “the shaggy base” and therefore does not necessarily agree in gender with the genus name.

Distribution:— Phyllogeiton trachybasis is only known with certainty from the southern part of the Maputaland Centre of Endemism, an area rich in restricted-range plants and animals ( Van Wyk 1996, Van Wyk & Smith 2001). The Maputaland Centre is at the southern end of the tropics in Africa ( Van Wyk 1996) and at the northern end of the Maputaland-Pondoland-Albany Hotspot, one of 36 global biodiversity hotspots ( Steenkamp et al. 2004).

Within the Maputaland Centre, P. trachybasis is known from 10 localities in KwaZulu-Natal ( Fig. 5 ). These localities range from Hluhluwe and Lake St Lucia in the south to Tembe Elephant Park and Ndumo Game Reserve just south of the South African and Mozambican border in the north.

There are no specimens of P. trachybasis at the Institute for Agricultural Research of Mozambique herbarium (LMA) and its presence in Mozambique needs confirmation. The species is, however, almost certainly present in the far south of Mozambique’s Maputo Province considering that trees of the new species were found in Tembe Elephant Park on the South African side directly next to the border between the two countries. A sight record (E. Schmidt, pers. comm.) and information supplied by a traditional healer (C. Hanekom, pers. comm.) suggest that it may occur at the Maputo Special Reserve south of Maputo, but confirmation is required.

Two other possible localities for the new species should be mentioned. The first is ca. 60 km from Caia in Sofala Province, Mozambique, on the EN1 route to Gorongosa (grid 1834BA), where a sterile specimen (J.E. & S.M. Burrows 8807 in Herb. BNRH) was collected by botanists familiar with the tree. This location is nearly 1000 km north of confirmed localities in KwaZulu-Natal. The second locality is near Siteki in Scarp Forest at an elevation of 450 m in the Lebombo [Lubombo] Mountains of Eswatini (formerly Swaziland), about 40 km north of Ndumo Game Reserve (grid 2631BD). A specimen collected here (P. & L. Loffler s.n. in Herb. BNRH) and photographs supplied by L. Loffler appear to match P. trachybasis , but the five trees seen were all sterile.Another collection from this same locality is also sterile (Schmidt 4054 in Ernst Schmidt Private Herbarium). Given that these specimens lack reproductive material, the presence of the species in Mozambique and Eswatini still needs to be confirmed.

Ecology:— Phyllogeiton trachybasis is a sub-canopy or canopy tree found mainly on well-drained, deep sandy soils at low elevations to about 100 m above sea level. Trees grow mainly in Tembe Sandy Bushveld (SVl 18) in open to closed woodland (codes of vegetation types follow Mucina & Rutherford 2006 and SANBI 2006–2018). They also grow in Sand Forest (FOz 8) patches in a mosaic with Tembe Sandy Bushveld ( Fig. 1 ). The species may also occur in Western Maputaland Sandy Bushveld (SVI19) at Ndumo Game Reserve, but information recorded on collecting labels is insufficient to confirm its presence in this vegetation type.

Phyllogeiton trachybasis has also been collected between Cape Vidal and Lake Sibaya in dune forest, which is included in the broader category Northern Coastal Forest (FOz 7). The most recent of these collections was in 1997, but the demographics of the species in dune forest are unknown .

Besides its normal sandy habitat, the species was collected once (Moll 5291 in Herbs K, NH, PRE) in 1971 from riverine forest embedded in Zululand Lowveld (SVl 23). This was on the Mansiya River in the Hluhluwe-iMfolozi Park where the presence of eight trees growing in the river’s floodplain was recently confirmed (F. du Randt, S. Louw, and S. Mabongwa, pers. comm.). The new species also apparently occurs occasionally in Scarp Forest (FOz 5). It was collected in this forest type in 1959 in the Hluhluwe-iMfolozi Park on a steep, southwest-facing rocky slope at an elevation of about 500 m (Ward 2947 in Herb. NU). A visit by F. du Randt, S. Louw, and S. Mabongwa on 18 May 2021 to try and relocate this plant or plants was unsuccessful.

Flowers are visited by honey bees. The fruit is eaten by birds, notably Trumpeter Hornbills, Bycanistes bucinator Temminck (1824 : livr. 48, pl. 284), at Nibela and Vervet Monkeys, Chlorocebus pygerythrus Cuvier (1821: 2) ( Pooley 1978, C. Hanekom, pers. comm.), although at Nibela it is reported that monkeys and people don’t feed on the fruit (G. Linforth, pers. comm.).

African Elephants, Loxodonta africana Blumenbach (1797: 125) , damage trees and this has killed at least one tree at Tembe Elephant Park (C. Hanekom, pers. comm.). Nyala antelope, Tragelaphus angasii Angas (1849: 89) , quickly removed fresh leaves from a cut branch at Bonamanzi Game Reserve (D. Bishop, pers. comm.) and are therefore suspected to browse on low-growing plants of the species.

No young plants are known in the wild, but the species is, despite the hard endocarp of the stone, easy to grow without any special treatment besides removal of the fleshy part of the pericarp (G. Nichols, pers. comm.). Possible reasons for the apparent lack of recruits are herbivory, disease or that conditions suitable for germination and establishment on sandy soils are rare.

The tree is an occasional host to the epiphytic orchids Ansellia africana Lindley (1844 : sub t. 12) and Polystachya concreta ( Jacquin 1760: 30) Garay & Sweet (1974: 206) .

Common names:— English and Afrikaans names in use include “sand ivory” and sandivoor respectively. These are also the names recommended by the Dendrological Society of South Africa ( Von Dürckheim et al. 2014).

There are several Zulu names recorded for Berchemia discolor in literature pertinent to KwaZulu-Natal ( Pooley 1980, Moll 1981, Pooley 1993, Hutchings et al. 1996). These names are likely to be applicable to P. trachybasis . The name ubalatsheni or ubaletsheni is recorded in all of these publications and means “marks” or “make marks on the stone or rock” (A. Koopman, pers. comm.). Semantic links to the names umuma and umumu ( Moll 1981, Pooley 1993, Hutchings et al. 1996), uvuka ( Pooley 1993, Hutchings et al. 1996), and umbenduza (Tinley 416B in Herbs NH, NU, PRE) are not clear (A. Koopman, pers. comm.). A herbarium specimen collected at Lake Sibaya bears the name umadlozane (Cunningham 2200 in Herb. NU), which is derived from the word amadlozi and refers to ancestral spirits (A. Koopman, pers. comm.).

The name umhungu is used at Nibela, which is at the northern end of Lake St. Lucia near Hluhluwe (G. Linforth, pers. comm.). This name and the similar umhungulo (Gerstner 5222 in Herb. PRE) and umhlungulo ( Hutchings et al. 1996) probably mean enticing, alluring, and deceiving (A. Koopman, pers. comm.).

In far northern KwaZulu-Natal an informant, M. Tembe, reports that the names vukakwabafileyo or vukakwabafile are used (C. Hanekom, pers. comm.). These names mean “to wake up at the place of those who have died” or “arise from the cemetery”, but A. Koopman (pers. comm.) suggests that their correct form may be vusabafileyo, which means “wake up the dead or unconscious”.

Conservation status: —The EOO of P. trachybasis was estimated at 5951 km 2. The AOO calculated was 60 km 2 for the 2 km cell width recommended by the IUCN Standards and Petitions Committee (2019) and 350 km 2 for a 5 km cell width. The actual AOO probably lies at the lower end of this range, because even the 2 km 2 cell width includes aquatic environments and unsuitable, degraded or transformed terrestrial habitats. The distinct bark of P. trachybasis also means that accessible specimens are unlikely to be overlooked.

The species occurs in 10 severely fragmented populations as defined by the IUCN (2012). Phyllogeiton trachybasis may occur in patches too small to support a viable population in the long-term and sub-populations are separated from other suitable habitat by large distances, thus genetic or demographic exchange seems unlikely.

There are 70 known mature individuals in the six sub-populations surveyed during field work, namely at Bonamanzi Game Reserve (35 trees), Tembe Elephant Park and surrounds (15 trees), Hluhluwe-iMfolozi Park (8 trees), Nibela Peninsula (8 trees), False Bay Park (2 trees), and Phinda Private Game Reserve (2 trees). At least eight of these trees are in poor condition and appear to be dying and no young plants (<10 years old) are known. The smallest (and presumably youngest) known plant is at Bonamanzi Game Reserve and is 4 m tall and has a dbh of 70 mm. The subpopulation sizes at Ndlozi Peninsula, Lake Bhangazi South, Sodwana Bay, and Ndumo Game Reserve are unknown. Even if known trees represent only about 10% of the species’ total population, there are likely to be fewer than 700 trees in the wild.

Phyllogeiton trachybasis appears to be almost entirely confined to protected areas. Bark of the few trees growing outside of these areas is heavily used for medicinal purposes, the removal of which kills individuals or limits growth (Cunningham 2200 in Herb. NU; L. Loffler, G. Linforth, C. Hanekom, pers. comm.). Even within some protected areas trees may not be safe from harvesting ( Groenewald 2010) or they may be damaged or killed by elephants (C. Hanekom, pers. comm.). It is probable that the current scarcity of trees resulted, at least in part, from the excessive stripping of bark for use in traditional medicine (locally referred to by the generic term umuthi, also spelled muti or muthi) over many years (see under “Uses” below).

Taking into account that the overall number of mature individuals is small, the known sub-populations are very small and the population is declining due to natural mortality, harvesting and lack of recruitment, application of the IUCN Red List Categories and Criteria ( IUCN 2012) suggest that the species should be classified as Endangered (EN) B2(v); C2a(i); D. There is an urgent need to survey populations more fully and to try and establish and address the reasons for the lack of recruitment.

Uses:— On the oldest known herbarium specimens of the species (Gerstner 5222 in Herb. PRE), collected at False Bay Park in 1944, it is recorded that a tree had 90% of its bark removed by traditional healers. The label of another specimen collected in 1993 in the corridor between Ndumu Game Reserve and Tembe Elephant Park (Van Wyk BSA 748 in Herbs PRE and PRU) reads: “This is a ‘muti’ tree. The old trunk (about 1 m high) is regularly stripped of its bark—in patches”. A collection from dune forest at Lake Sibaya (Cunningham 2200 in Herb. NU) also records that the bark is heavily utilised for medicinal purposes. Trees which may belong to this species observed at Siteki, Eswatini (see under “Distribution” above) were debarked. Bark and the roots are heavily utilised east of Tembe Elephant Park and according to a local resident are mixed with material from other species to treat chest ailments (C. Hanekom, pers. comm.). There are few surviving trees in the community area surrounding Nibela at Lake St. Lucia because they have all been debarked and the majority have already died (G. Linforth, pers. comm.). At Nibela evil spirits are believed to be extracted by boiling and steaming bark, and fruit is used to treat impotence (G. Linforth, pers. comm.). At nearby False Bay Park, bark is used as a love charm ( Hutchings et al. 1996).

There are no records of the fruit being eaten by humans, which is surprising as the fruits of other Phyllogeiton species are popular (e.g. Pooley 1980, Van Wyk & Gericke 2000), with fruit of P. discolor even having potential in terms of commercialisation and domestication ( Lusepani 1999).

Additional collections (paratypes): — SOUTH AFRICA. KwaZulu-Natal: Ndumu [ Ndumo ] Game Reserve , (2632CD), 9 May 1956, Hancock 9 ( NU!); Ndumu [ Ndumo ] Game Reserve , (2632CD), 23 February 1959, Tinley 416B ( NH!, NU!, PRE!); 3 miles S of Makane’s Drift, (2632CD), 28 February 1968, Ross & Moll 1820 ( PRE!); Ingwavuma District, Ndumu [ Ndumo ] Game Reserve , Matini Forest , E. area, (2732AA), 8 February 1964, Tinley 943 ( NH!, NU!); Makatini Flats, Makani’s [Makane’s] Pont, (2732AA), 26 February 1968, Venter 4614 ( PRE!); Corridor between Ndumu [ Ndumo ] and Tembe , (2732AB), 15 March 1993, Van Wyk BSA748 ( PRE!, PRU!); Jobe’s Kraal, (2732 AA), 16 May 1999, Schmidt 2406 (Ernst Schmidt Herbarium!); Sihangwane Nature Conservation Camp, (2732AB), 9 November 1981, Cunningham 518 ( NU!); Tembe Elephant Park, (2732AB), 23 April 1995, Van Wyk 12786 ( PRU!); Tembe Elephant Reserve [ Park ], (2732AB), 23 October 1995, Nichols , Kruger & Symmonds s.n. ( NH!); Tembe Elephant Park, (2732AB), 13 February 1996, Prozesky 50 ( PRU!); Tembe Elephant Park, (2732AB), 10 October 1997, Matthews 1304 ( PRU!); Tembe Elephant Reserve [Park], (2732AB?), 17 January 1999, Schmidt 2185 (Ernst Schmidt Herbarium!); Tembe Elephant Park, (2732AB), 30 March 2011, Wright 1 ( PRU!); Amanzimnyama Pan area, Lake Sibayi [ Sibaya ], (2732BC), 15 January 1958, Tinley 59 ( NH!, PRE!); E. side Lake Sibayi [ Sibaya ], (2732BC), 28 November 1967, Nicholson 621 ( NH!); Sibaya Lake, (2732BC), 28 November 1967, Strey & Moll 3977 ( K!, NH!, PRE!); E. Shores Lake Sibayi [ Sibaya ], (2732BC), 26 October 1971, Moll & Nel 5618 ( K!, NH!, PRE!); Lake Sibayi [ Sibaya ], (2732BC), 20 January 1987, Cunningham 2200 ( NU!); False Bay Park, (2732CD), 28 June 1971, Ward 7114 ( NH!, NU!, PRE!, UDW!); False Bay, birds sanctuary, (2732DC), December 1944, Gerstner 5222 ( PRE!); Sodwana State Forest, (2732DC), 16 June 1985, Gordon 203 ( NH!); Sodwana State Forest, south of Air Force camp, (2732DC), 22 December 1985, Gordon 259 ( NH!, PRE!); False Bay Park, near the end of the Mphophomeni Trail , (2732DC), 30 April 2014, Boon 63 ( NH!); Hluhluwe Game Reserve, (2832AA), 7 January 1956, Ward 2947 ( NU!); Hluhluwe Game Reserve, (2832AA), 28 November 1959, Ward 3327 ( PRE!, NU!); Hluhluwe Game Reserve, Monsia [ Mansiya ] River , (2832AA), 2 March 1971, Moll 5291 ( K!, NH!, PRE!); Bonamanzi Game Ranch [ Reserve ], 10 km S of Hluhluwe , (2832AA), 22 February 1994, Abbott 6317 ( PCE!, PRU!); Bonamanzi Game Ranch [ Reserve ], 10 km ESE of Hluhluwe , (2832AA) [2832AB], 5 November 1994, Abbott 6498 ( PCE!, PRU!); Hluhluwe, Bonamanzi Game Park [ Reserve ], on left of road 150 m beyond Game Lodge turnoff, (2832AA), 5 March 1996, Abbott 6930 ( PCE!, PRU!); Hluhluwe, Bonamanzi Game Ranch [ Reserve ], (2832AA), 5 March 1996, Abbott 6931 ( PRU!); Hluhluwe, Bonamanzi Game Ranch [ Reserve ], (2832AA), 5 March 1996, Abbott 6932 ( PRU!); Hluhluwe, Dlozi (or Dhlozi ) Plantation , Hlabisa District , (2832AB), 16 January 1963, Dutton 88 ( NH!, NU!); North Coast, St. Lucia system (sensu lato), Lake Bhangazi (South) area, eastern side of lake, 15 February 1997, (2832BA), Ward 13915 ( NH!).

Notes and key to the species of Phyllogeiton : — Phyllogeiton trachybasis is morphologically most similar to P. discolor ( Fig. 6 ), but the species’ known ranges are allopatric and P. trachybasis is for now considered to be endemic to the Maputaland Centre of Endemism, as well as to KwaZulu-Natal. Phyllogeiton discolor has not been collected from KwaZulu-Natal nor has it been recorded from Eswatini ( Loffler & Loffler 2005, L. Loffler, pers. comm.). A specimen at Herb. SDNH (Compton 28939) tentatively identified as Berchemia discolor is of P. zeyheri ( Fig. 7 ). There is one correctly labelled P. discolor specimen at LMA from the southern part of Mozambique (covered by the provinces of Maputo, Gaza and Inhambane). This specimen was collected at Massingir in Gaza (Lousã & Rosa 301), which is about 325 km north of the nearest known P. trachybasis plants in KwaZulu-Natal. Specimens at Herb. LMA labelled B. discolor from Catuane, which is in Mozambique just north of Ndumo Game Reserve in KwaZulu-Natal, and Namaacha (west of Maputo near the border with Eswatini) are misidentified collections of P. zeyheri .

The habitat of the new species and P. discolor also differs. Phyllogeiton trachybasis prefers sandy soils of Early to Late Pleistocene coastal dunes ( Van Wyk & Smith 2001), although it has also been collected from Scarp Forest and once from riverine forest. Phyllogeiton discolor also occurs in bushveld (savanna), but is usually found on river banks or termitaria (Coates Palgrave 2002, Schmidt 2018). The two species differ in a number of morphological features including bark, leaf shape, sepal and petal position at anthesis, and fruit size, shape and colour. In addition, P. trachybasis has ± concolorous leaves without any bloom and the fruit are often faintly longitudinally ribbed, whereas P. discolor has leaves that are distinctly paler due to the presence of a greyish bloom and the fruit surface is plane. Further distinctions, also from P. zeyheri , are supplied in the identification key below.

Phyllogeiton discolor ( Fig. 6 ) and P. zeyheri ( Fig. 7 ) differ markedly in the colour and fluorescence of their hardwood ( Dyer 1988, Dyer et al. 2016), a distinction reflected in their English and Afrikaans common names. Known as “red ivory”/ rooi-ivoor, the heartwood of P. zeyheri ranges in colour from pale pink to various shades of bright red and is much sought-after for making furniture and curios ( Dyer et al. 2016). On the other hand, the heartwood of P. discolor (“brown ivory”/ bruinivoor) is yellowish brown to deep reddish brown. Heartwood of both species fluoresce when viewed under longwave ultraviolet light (ca. 365 nm), a distinction that has been successfully used to confirm the identification of commercial consignments of these two woods ( Dyer 1988). Phyllogeiton zeyheri has a characteristic orange fluorescence, while P. discolor fluoresces with a green colour. Because of its rarity, the colour and fluorescense of the mature heartwood of P. trachybasis still needs to be established with certainty. Heartwood in a transverse section of a long-dead branch (ca. 130 mm in diameter) of this species ( Fig. 2E ) is yellowish brown with a reddish tinge towards the sapwood, thus bearing a close resemblance to the heartwood of P. discolor . The narrow sapwood in this particular branch was bleached to a greyish colour by exposure to the elements and was also extensively damaged by wood boring insects.

Two herbarium collections of P. trachybasis refer to ripe purple fruit (Moll 5291 in Herbs K, NH, PRE) and pink fruit (Gordon 259 in Herbs NH, PRE). These observations are curious because in all other collections where fruit colour is mentioned it is recorded as yellow, which matches known images and our own field experience with the species.

The conservation outlook for P. trachybasis seems poor. There are few known trees, the overall population is fragmented, most sub-populations consist of few individuals, a number of trees are dead or in poor health, and there is an apparent lack of recruitment. Given these population demographics and the excessive and destructive harvesting of bark, traditional biocultural uses of the species are clearly unsustainable. Conservation actions should include ensuring that trees in protected areas are not harvested, supplementing populations with cultivated specimens, and testing whether plants will recruit where browsing mammals are excluded. There is also a need to assess the current situation with trees in the dune forest between Lake Sibayi and Cape Vidal and the population at the Dlozi Peninsula in the iSimangaliso Wetland Park.

Identification key to the species of Phyllogeiton

1. Leaf blade usually to 40 × 25 mm; flowers with ovary not immersed in disc; fruit ripening red (very often yellowish in immature fruit), to ca. 13 × 5 mm; receptacle at point of fruit stalk’s attachment at base of fruit markedly convex (see Figs 4G & 7C ) .......... .............................................................................................................................................................................................. P. zeyheri

- Leaf blade usually to 100 × 40 mm; flowers with ovary immersed in disc; fruit ripening pale yellow, yellow, orange-yellow or brownish yellow,>13 × 5 mm; receptacle at point of fruit stalk’s attachment at base of fruit plane or markedly concave (see Figs 4E & F ) ..............................................................................................................................................................................................2.

2. Bark towards base of trunk on mature trees very rough, flaking in large patches, almost black (sometimes with pale crustose lichen); leaves fresh green and without a bloom, ± concolorous, apex acuminate; flowers with petals spreading to ascending at anthesis; fruit usually ellipsoid, very often slightly ribbed longitudinally, rich orange-yellow, to 25 × 19 mm; receptacle at point of fruit stalk’s attachment at base of fruit plane (see Fig. 4E ); currently known with certainty only from KwaZulu-Natal, South Africa (Maputaland Centre of Endemism) ............................................................................................................................... P. trachybasis

- Bark towards base of trunk on mature trees rough, grey-brown, if flaking then in small patches; leaves dark green, distinctly paler with a greyish bloom below, apex acute to obtuse; flowers with petals reflexed at anthesis; fruit usually ovoid, surface plane, pale yellow to yellow, often turning brownish yellow when overripe, to 20 × 11 mm; receptacle at point of fruit stalk’s attachment at base of fruit markedly concave (see Fig. 4F ); widespread, from southern through central, eastern and northeastern Africa to Arabia, also in Madagascar, but not yet recorded in KwaZulu-Natal, South Africa........................................................................ P. discolor

Figures

FIGURE 1. Phyllogeiton trachybasis, habit and habitat. Mature tree in Sand Forest at Bonamanzi Game Reserve, Hluhluwe, KwaZuluNatal, South Africa. Photograph: R.G.C. Boon.

FIGURE 2. Phyllogeiton trachybasis, bark morphology illustrating variation (A–D) and colour of the wood (E). A. Bark on trunk of old, mature tree. B. Bark on trunk of relatively young tree (sunny habitat). C. Bark on trunk of relatively young tree (shady habitat). D. Exceedingly rough and flaking bark on trunk of mature tree. E. Finely sanded, untreated wood in transverse section; from dead branch ± 130 mm in diameter. Photographs: R.G.C. Boon (A, D), F. du Randt (B, C), & D. Bishop (E).

FIGURE 3. Phyllogeiton trachybasis, leaf, flower and fruit morphology. A. Branchlet with leaves, the latter fresh green, without a bloom and ca. concolorous. B. Flowers, open and in bud. C. Almost ripe fruit. D. Mainly ripe fallen fruits underneath a tree, some with stone removed or otherwise damaged by wildlife. E. Cleaned stones. Photographs: R.G.C. Boon (A, C, D), G. Nichols (B), & E. Douwes (E).

FIGURE 4. Phyllogeiton trachybasis (A–E), with fruit of P. discolor (F) and P. zeyheri (G). A. Fruiting branchlet (fruit almost mature). B. Flower. C. Petal. D. Stamen. E–G. Comparison of the fruit (bottom) of the three species of Phyllogeiton, with corresponding enlarged persistent receptacle and fruit stalk (top). Scale bar = 10 mm (A), 1 mm (B–D), or 4 mm (E–G). Drawings based on photographs and liquid-preserved fresh material deposited in Herb. PRU. Artist: Daleen Roodt.

FIGURE 1. Phyllogeiton trachybasis, habit and habitat. Mature tree in Sand Forest at Bonamanzi Game Reserve, Hluhluwe, KwaZulu- Natal, South Africa. Photograph: R.G.C. Boon.

FIGURE 5. Topographical map showing the known distribution (black dots) of Phyllogeiton trachybasis in northeastern KwaZulu-Natal, South Africa. Map based on herbarium collections in BNRH, K, NH, NU, PCE, PRE, PRU, and UDW.The insert shows a map of southern Africa with names of countries; the grey rectangle indicates the area depicted by the topographical map.

FIGURE 6. Phyllogeiton discolor. A. Mature bark pattern. B. Flowers. C. Fruit; note leaves with upper surface of blade glossy and dark green. D. Leaves viewed from below, showing a greyish bloom. Photographs: R.G.C. Boon (A, C) & A.E. van Wyk (B, D).

FIGURE 7. Phyllogeiton zeyheri. A. Mature bark pattern. B. Flowers; note leaves with upper surface of blade dull and with a greyish bloom. C. Fruit. D. Leaves viewed from below, showing a greyish bloom similar to that of the upper surface. Photographs: R.G.C. Boon (A) & A.E. van Wyk (B–D).