Ozoroa bhangazica R.G.C.Boon & A.E.van Wyk, 2025

Ozoroa bhangazica R.G.C.Boon & A.E.van Wyk , sp. nov. ( Figs 1–4 View FIGURE 1 View FIGURE 2 View FIGURE 3 View FIGURE 4 )

Diagnosis:— A member of Ozoroa Group A; distinguished from all other species of Ozoroa in KZN Province, South Africa, by being a clonal geoxylic suffrutex not exceeding 1.5 m in height ( vs. non-clonal shrubs or trees> 2 m tall). Morphologically resembling O. barbertonensis in also being a geoxylic suffrutex with linear to narrowly oblong leaves on new growth following a fire, but differing as follows: lamina of post-fire leaves 80–110(–130) mm long [ vs. shorter, (23–)55–80(–90) mm], (6–) 9–18 mm wide [ vs. narrower, (2–)4–7(–8.5) mm], adaxially glabrous except for midrib and secondary veins ( vs. ca. uniformly hairy), with margin plane ( vs. revolute); petiole 5–18 mm long ( vs. shorter, 1.5–3.0 mm); mature drupe when dry 10 × 7 mm ( vs. smaller, 4.5–5.0 × 6.0– 6.5 mm); confined to Maputaland Centre of Endemism where it grows on nutrient-poor Quaternary coastal sands ( vs. a serpentine endemic confined to Barberton Centre of Endemism).

Type:— SOUTH AFRICA. KwaZulu-Natal: Sodwana State Forest, Chief’s Kraal , (2732DA), 12 September 1985, Gordon 235 ( holotype NH100853!; isotype CPF0005274!) .

“ Ozoroa View in CoL sp. nov. ” in Scott-Shaw (1999: 9); Mucina & Rutherford (2006: 577–578); “ Ozoroa sp. ” in Boon (2010: 262).

Illustrations: — Scott-Shaw (1999: 9, Ozoroa View in CoL sp. nov., line drawing, ramet showing leaves and mature fruit); Mucina & Rutherford (2006: 577, Ozoroa View in CoL sp. nov., photograph, plant in habitat); Boon (2010: 263, Ozoroa sp. , photograph bottom far-right, shoot with mature fruit).

Dioecious, clonal, pyrogenic geoxylic suffrutex, with few to ca. 80 erect or ascending simple, slender, sub-woody, cylindrical stems (ramets) sprouting from large subterranean lignotuber or extensive system of spreading woody axes (stems/roots), usually 250–500(–800) mm tall, but woody, branched and up to 1.0(–1.5) m tall when long unburnt; individual plants/clones mostly 1–2 m in diam., clones sometimes reaching ca. 11 m in diam.; latex milky; indumentum of young parts dense, with short, < 0.5 mm long, white, eglandular, appressed, antrorse, unicellular hairs, becoming glabrescent to glabrous on older parts; newly sprouting ramets following a fire appear silvery-white (due to reflective indumentum on abaxial leaf surface) from a distance. Bark on unburnt plants and older stems, smooth, grey to grey-brown, lenticellate; lenticels plane to raised and conspicuous, pale grey-brown to red-brown. New stems fawn to olivegreen, young parts with indumentum partially obscuring lenticels. Stipules absent. Leaves on new growth following fire (or other disturbance?) ( Figs 2B View FIGURE 2 & 3A View FIGURE 3 ) irregularly spirally arranged (appearing almost three-whorled), simple, erect, spreading later, linear, narrowly elliptic or narrowly oblong, ca. 80–110(–130) × (6–) 9–18 mm when mature, base broadly tapered, apex obtuse or broadly acute (to acute), with a slender mucro 0.5–2.0 mm long; margin plane, slightly thickened, yellowish, entire, may appear crenulate when dry; adaxial surface glabrous except for the midrib and secondary veins, abaxially with dense silvery-white indumentum ( Fig. 2B View FIGURE 2 ) of short, appressed, white hairs, principal lateral veins less hairy and appearing darker than rest of abaxial surface, chartaceous when young, becoming coriaceous with faintly velvety texture abaxially, discolorous on account of indumentum, glossy, mid- to dark green adaxially, becoming yellowish green with age, bluish green to cinereous below when young, midrib yellowish below; venation pinnate, craspedodromous, midrib slightly sunken on adaxial surface, strongly raised on abaxial surface, secondary (principal lateral) veins parallel, ca. 80 o to midrib, straight, unbranched, or once- or twice-forked near margin, ca. 33–50 pairs, secondary veins conspicuous below, impressed (slightly raised when dry), tertiary veins reticulate, reticulation obscure and slightly raised. Petiole proximally terete, distally shallowly canaliculate, 5–10 mm long, densely hairy. Leaves on plants not burnt for>1–several years ( Figs 1D View FIGURE 1 & 3B View FIGURE 3 ) elliptic, 45–95 × 20–26 mm, apex obtuse, rounded or emarginate, base broadly acute, principal lateral veins ca. 24–26 pairs. Petiole 8–15 mm long. Inflorescences axillary in upper leaf axils or terminal, paniculate, to ca. 55 mm long, with occasional small leaf in branch axils, axes terete, densely hairy. Flowers unisexual, regular, 5-merous, subtended by one or more caducous bracts ca. 1.5–2.0 × 0.25 mm; pedicel usually 1–2 mm long, articulated near apex. Male flowers with s epals free, broadly ovate-triangular, ca. 1.5–1.75 × 1.0 mm, off-white, apex acute, densely hairy abaxially; petals free, oblong, ca. 2.50–3.25 × 1.0– 1.5 mm, white or pale cream, apex obtuse or rounded, revolute and slightly reflexed at apex, hairy abaxially but less so than sepals; disc annular with wavy margin; stamens 5, filaments 0.50–0.75 mm long, anthers 0.75–1.00 × 0.5 mm; gynoecium absent or vestigial. Female flowers with sepals, petals and disc as in male flowers; staminodes 5; filaments ca. 1 mm long, antherodes ca. 0.5 × 0.25 mm; ovary globose but distinctly axially flattened, ca. 1.5 mm in diam., ca. 0.75 mm high, 1-locular with single laterally attached ovule, styles 3, united at very base, stigmas somewhat capitate. Fruit a fleshy drupe, transversely ellipsoid, younger and nearly mature ones slightly laterally compressed, shiny green with scattered sunken spots, each containing modified stoma, distinctly compressed with age, ca. 10 × 7 × 3 mm when dry, green to purplish red-spotted or -tinged when young, glabrous, glossy black and irregularly rugose when mature.

Phenology:— Flowering is from September–February(–May), with a peak in November. Timing is probably influenced by the time elapsed since the last grassland fire. Mature fruit has been observed between November and June. Fruit probably takes about six months to develop after flowering. Collections of plants with mature fruit in spring and early summer were probably taken in grasslands that did not burn in the preceding dry season.

Etymology:— The specific epithet “ bhangazica ” refers to the fact that the species is currently known only from the vicinity of Lake Bhangazi North and Lake Bhangazi South in the iSimangaliso Wetland Park World Heritage Site in KZN Province, South Africa.

The name of the lakes derives from the Zulu word umbhangazi, which refers to various quick-growing trees— Zulu names of geographical features are often named for the dominant vegetation there (A. Koopman, pers. comm.).

Distribution:— Ozoroa bhangazica is known only from KZN Province in South Africa ( Fig. 4 View FIGURE 4 ); more specifically 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, O. bhangazica occurs in two separate, small areas in the iSimangaliso Wetland Park, which is a UNESCO World Heritage Site (Nomination 914). One area, referred to as the Eastern Shores, stretches from the town of St Lucia in the south to Cape Vidal and Lake Bhangazi South in the north. The second area is about 60 km north of the first and lies between Lake Bhangazi North and Sodwana Bay, which is about 10 km north of the Lake.

It is possible that the species occurs in suitable habitat in between, but it is difficult to access the area as there are no roads and numerous wetlands. Ozoroa bhangazica has not yet been recorded in the Mozambican part of the Maputaland Centre.

Ecology:— Ozoroa bhangazica is confined to Maputaland Wooded Grassland ( Fig. 1A View FIGURE 1 ), a vegetation type that extends along the coastal plain from southern Mozambique south to Richards Bay in KZN ( Mucina & Rutherford 2006, South African National Biodiversity Institute 2006–2024). This grassland is particularly notable for its richness in geoxylic suffrutices, rhizomatous dwarf shrubs and forbs ( Mucina & Rutherford 2006). From a phytosociological perspective, this distinctive grassland was originally classified by Myre (1964) as the Themedo-Salacietum association. However, Matthews et al. (1999) advocated for the term “woody grassland” to describe it, with “woody” underscoring the abundance of geoxylic suffrutices—likely representing the highest concentration of this growth form in any South African vegetation type. The term “wooded”, as applied in the name Maputaland Wooded Grassland by Mucina & Rutherford (2006), might be misinterpreted to suggest a grassland type with scattered trees, which is not the case here. In fact, “woody” more accurately reflects the dominance of underground woody structures that account for much of the plant biomass in this ecosystem.

The landform of the coastal plain is gently undulating with low rises and inter-dune depressions ( Fig. 1A View FIGURE 1 ). The latter are inundated at times of high rainfall because of a relatively shallow water table. Ozoroa bhangazica is absent from the depressions and instead grows on the dune crests and gentle slopes above hygrophilous grasslands and wetlands. Matthews et al. (1999) also found that geoxyles are absent from inter-dune depressions at the nearby Sileza Nature Reserve, north of Sodwana Bay. Ozoroa bhangazica grows at low elevations, probably not exceeding 40 m above sea level. The soils occupied by Maputaland Wooded Grassland are moderately- to well-drained, acidic Quaternary sands and the water table is ca. 1.5–2.0 m below the surface and sometimes more ( Matthews et al. 1999, Mucina & Rutherford 2006). The climate is subtropical/tropical and average annual rainfall is about 1200 mm, with weak seasonality ( Matthews et al. 1999, Mucina & Rutherford 2006). There is no frost ( Van Wyk & Smith 2001).

Maputaland Wooded Grassland is a fire-prone ecosystem subject to and dependent on frequent wildfires and prescribed burns ( Van Wyk & Smith 2001). The geoxyle habit of O. bhangazica makes it a fire-adapted species ( Scott-Shaw 1999), like many other taxa found in this vegetation type ( Van Wyk & Smith 2001). Similar to cooccurring geoxyles, O. bhangazica is an obligate resprouter and it produces several to many new stems (ramets) from a lignotuber and/or subterranean woody axes following fire ( Fig. 1B, C View FIGURE 1 , 3A View FIGURE 3 ). Plants flower best several months after a fire and fruit usually matures before the next fire season, when fire will likely consume all or most aerial phytomass in the grassland. Plants growing in unburnt grassland flower much less prolifically. The grassland at Ozabeni is burnt annually by neighbouring communities and the Eastern Shores grasslands to the south are usually burnt biennially by conservation staff (C. Myhill, pers. comm.). When one of us (RGCB) visited in October 2024, most of both known localities were burnt a few months previously. At the time of the visit, many plants were flowering, but none had fruit, including the few individuals found in grassland that was not burnt in the previous dry season.

The current fire regimes at the two areas where O. bhangazica occurs appear to favour the persistence of adult plants while also limiting the encroachment of taller native and introduced species. However, successful seedling recruitment probably requires the uncommon combination of bare ground, adequate moisture and a sufficient gap between fires, and it is not known how often young plants are able to become established. Nonetheless, as clonal plants, geoxyles also reproduce vegetatively, and their clones are effectively “immortal,” with individual clonal colonies undoubtedly capable of surviving for very long periods.

Various authors have suggested explanations for the evolution of the geoxyle habit in southern Africa, including the interaction of high rainfall and frequent fire ( Maurin et al. 2014, Davies et al. 2016), frost ( Burtt Davy 1922, Finckh et al. 2016), herbivory ( Steenkamp et al. 2001), and edaphic factors ( White 1976). Although White (1976) admitted that fire played an important role, he attributed the rise of the geoxyle habit mainly to edaphic factors, in particular seasonally waterlogged soils. As previously mentioned, in the coastal grasslands of Maputaland geoxyles, including O. bhangazica , are typically situated on relatively elevated, well-drained landforms such as dune crests and slopes, where the surface soils are never waterlogged ( Matthews et al. 1999). These plants are also conspicuously absent from the lower-lying inter-dune depressions, which are the only parts of the landscape that experience clear seasonal waterlogging.

While recognising that any site condition that reduces growth rates and prevents individuals reaching fire-proof sizes favours the geoxyle lifeform, Maurin et al. (2014) and Davies et al. (2016) argue that high precipitation, which leads to increased grass fuel loads, along with frequent fires, are the primary evolutionary drivers of the geoxyle habit in African savannas. Supporting this, a study by Courtenay et al. (2024) examining the habitat preferences of various geoxyles and their arborescent relatives in Africa found that Ozoroa geoxyles typically occupy environments with the highest fire frequencies and generally high rainfall. Ozoroa bhangazica does not appear to be browsed by native or introduced mammalian herbivores, does not grow where soils are waterlogged, and is not subjected to frost. However, it does grow on low-nutrient soils and persists in areas characterised by high precipitation and frequent fires.

Common names:— The new species does not have a widely-used English common name and we recommend the name Maputaland Dwarf Resintree given by Scott-Shaw (1999). The name translates to maputalanddwergharpuisboom in Afrikaans. We do not know the Zulu name, but other Ozoroa species in KZN are called isifice or isifico, and it is likely that O. bhangazica has the same name. The name isifico is derived from the verb fica and there are a number of meanings, the most likely of which is “find” or “come upon” (A. Koopman, pers. comm.). The link to the plant is unclear to us.

Conservation status:— Ozoroa bhangazica is apparently confined to two areas of Maputaland Wooded Grassland in the iSimangaliso Wetland Park (South African National Biodiversity Institute 2006–2024). Nearly half of the original extent of Maputaland Wooded Grassland has been transformed, mainly for timber plantations, and 17% is conserved in the iSimangaliso Wetland Park ( Mucina & Rutherford 2006).

The iSimangaliso Wetland Park is a World Heritage Site stretching from Maphelane and St Lucia in the south to Kosi Bay at the border with Mozambique (iSimangaliso Wetland Park 2025). It was established under the World Heritage Convention Act, No 49 of 1999 and the South African National Environmental Management: Protected Areas Act, 2003 (iSimangaliso Wetland Park 2025), and thus has a high level of statutory protection.

Ozoroa bhangazica occurs in two known subpopulations as defined by the IUCN (2012). These subpopulations are about 55 km apart and genetic or demographic exchange seems unlikely. It is possible that the species occurs in between the two known localities in less accessible grassland away from management tracks and tourist roads. However, the area has a long history of conservation, which predates the establishment of the iSimangaliso Wetland Park in 1999, and it is probable that experienced botanists have collected in the intervening area, at least occasionally, over several decades.

The southern subpopulation is in a part of the Park that has been developed for nature-based tourism and there is a high level of management and control over public access. This population was probably impacted when a tourist road was constructed many years ago. The population experiences limited encroachment by native and non-native woody plant species, which are not typically found in Maputaland Wooded Grassland. This is currently a minor threat because regular ecological burns, conducted by conservation staff, seem sufficient to control most of these encroaching weeds.

The plants near Lake Bhangazi North grow where management controls are harder to implement. The grassland usually burns annually when fires are lit in the Park or spread from adjacent communal lands. Adult O. bhangazica plants do not seem to be harmed by the fires, and they do not appear to be browsed by native or introduced herbivores, which are attracted to fresh grass following fires. It is unknown whether the fires have a negative impact on seedling establishment, although that seems likely. The threat from encroachment by woody plants is similar in the north to that experienced by plants in the south. There are several off-road vehicle tracks between Sodwana Bay and Lake Bhangazi North. New tracks are created when old tracks are eroded or the sand becomes too soft. When tracks are formed O. bhangazica plants may be impacted.

Most collections of Ozoroa bhangazica lack precise coordinates and, where precise information was recorded by collectors, the collections were made close to roads and tracks. If only exact localities are used to calculate the EOO and AOO for the new species, the areas are very small. Instead, satellite imagery available on GeoCAT was used to produce a coarse-scale map of what appears to be suitable habitat in the vicinity of precisely known localities. This will likely have produced EOO and AOO calculations that exceed the actual figures.

The EOO of Ozoroa bhangazica was estimated at 600 km 2. The AOO calculated was 165 km 2 for the 2 km cell width recommended by the IUCN Standards and Petitions Committee (2024). The actual AOO is probably less than 165 km 2, because the 2 km 2 cell width includes wetlands and unsuitable, degraded, or transformed terrestrial habitats. Although the AOO and EOO for O. bhangazica meet the thresholds for the IUCN Red List Endangered (EN) Category under Criterion B., only one of the three conditions are met (two are required), i.e. Condition (a) as the population may be severely fragmented and there are ≤5 locations (i.e. threat-based areas sensu IUCN Standards and Petitions Committee 2024). However, there is insufficient evidence to show that the species is experiencing a continuing decline (although that is possible) or extreme fluctuations (very unlikely) in the EOO, AOO, number of locations or subpopulations, or mature individuals (Conditions (b) and (c)).

The size of the two subpopulations is unknown. When estimating the number of mature individuals in the case of clonal organisms, the smallest entity capable of both independent survival and (sexual or asexual reproduction), i.e. the ramet, is considered a mature individual ( IUCN Standards and Petitions 2024). Based on the AOO and field observations of the density of genets (and ramets), it seems possible that the total population of the species is <10 000 mature individuals. Therefore, the species might qualify for Criterion C. (small population), but there is no evidence to suggest that the thresholds for C1. (rate and size of decline) or C2. (size of subpopulations and % of mature individuals in one subpopulation) are met.

Whereas O. bhangazica occurs in only two widely separated and isolated subpopulations and the total population is relatively small, both populations are included in a managed protected area and there is no evidence that the current overall population is declining. Perhaps the most plausible threat to O. bhangazica is that of encroachment by surrounding communities into the grasslands occupied by the northern subpopulation. The risk of encroachment is unknown, but other protected areas in KwaZulu-Natal have been impacted by this threat in the past (e.g. Groenewald 2010). As the species is close to qualifying as Endangered in Criterion B. and the northern population may be under threat or declining, our recommendation is that the species be classified as Near Threatened (NT). This is consistent with the guidance provided by the Standards and Petitions Committee (2024).

Notes:— Ozoroa bhangazica is a range-restricted species, which occurs only on Quaternary coastal sands in Maputaland Wooded Grassland in the iSimangaliso Wetland Park, South Africa. No other Ozoroa species is found in this vegetation type. However, O. obovata ( Oliver 1868: 437) Fernandes & Fernandes (1965: 161) var. obovata View in CoL is found nearby and is a common constituent of dune and coastal forest margins and scattered bush clumps in the grassland matrix. It does not grow in open grassland and it is easily distinguished from O. bhangazica by being a large shrub or small tree with leaves arranged in whorls of three ( vs. irregularly arranged to occasionally subverticillate). Ozoroa obovata var. obovata View in CoL has elliptic to obovate leaves with a broadly rounded or broadly tapered apex. On fresh growth, O. bhangazica has linear to narrowly elliptic leaves ( Figs 2B View FIGURE 2 & 3A View FIGURE 3 ), quite unlike those of O. obovata var. obovata View in CoL . Leaves on long unburnt individuals of O. bhangazica ( Figs 1D View FIGURE 1 & 3B View FIGURE 3 ) are more similar to those of O. obovata var. obovata View in CoL , but are irregularly and sparsely arranged ( vs. in whorls of three and densely arranged). Even when not burnt for several years O. bhangazica is always less than 1.2(–1.5) m tall ( Fig. 1D View FIGURE 1 ).

Older collections of O. bhangazica have most frequently been identified as O. paniculosa var. paniculosa , although more recent herbarium specimens have been recognised as a new species. Distributions of these two taxa are allopatric. Ozoroa bhangazica is found in grassland on the coastal plain at low elevations no more than 10 km from the sea. In contrast, O. paniculosa has a much wider range than O. bhangazica and is found in Botswana, Eswatini, Mozambique, Namibia, and the northern provinces of South Africa, as well as KZN, but excluding the coastal areas of Maputaland. Across its wide range, it grows in bushveld savanna, often on hillsides, and apparently does not occur closer to the coast than the Lubombo (Lebombo) Mountains (iNaturalist 2025). This mountain range forms the western boundary of the Maputaland Centre and is oriented approximately north-south about 40 km inland of the KZN coastline. The distribution map supplied for O. paniculosa var. paniculosa in Boon (2010) is incorrect and is a reminder that care needs to be taken in identifying Ozoroa plants in the field and herbarium.

Ozoroa paniculosa var. paniculosa is always a shrub or small tree ( vs. a geoxylic suffrutex). Leaves are usually arranged in whorls of three or are subverticillate ( vs. usually irregularly spirally arranged), they are relatively shorter and broader than in O. bhangazica , the adaxial surface is puberulous to glabrescent ( vs. essentially glabrous), the leaf margin is somewhat thickened and undulate ( vs. slightly thickened and ca. plane), and the principal lateral veins are raised and prominent abaxially ( vs. not raised and only conspicuous because they appear darker than the remainder of the abaxial surface) ( Fernandes & Fernandes 1966).

Three other described Ozoroa geoxyles are currently recognised in the FSA region. Ozoroa schinzii and O. albicans have distinctly different leaf morphology and occur in Namibia and South Africa’s Limpopo Province (with O. albicans extending slightly into Mpumalanga Province), respectively. These two species are unlikely to be confused with O. bhangazica . The third species, O. barbertonensis , is morphologically the most similar, sharing linear to narrowly oblong leaves. However, it is a restricted-range serpentine endemic found only in the Barberton Centre of Endemism ( Van Wyk & Smith 2001), where it is confined to grassland on rocky mountain slopes within Lowveld Sour Bushveld ( Retief 1990), thus occupying a markedly different habitat from that of O. bhangazica . Additionally, their distribution ranges are separated by approximately 250 km. Further distinguishing features between O. barbertonensis and the new species are outlined in the diagnosis preceding the description above.

Additional specimens examined ( paratypes):— SOUTH AFRICA. KwaZulu-Natal: Northern Zululand , illala [ilala] veld, [2732DA], July 1924, Boocock s.n. (PRF 5715 in PRE!) ; N. [Northern] Zululand , 5 miles W of Sordwana [Sodwana] Bay, [2732DA], July 1924, Miller & Boocock s.n. (PRF 5715 in PRE!) ; Lower Mbazwane, Kwa-Mbila area , 27 o 38’S 32 o 33’E, (2732DA), 14 December 1971, Ward 7473 (K002226574!, NH0064634-0!, UDW10957! & UDW10958!) GoogleMaps ; Sodwana State Forest, grassland east[west] of Lake Bhangazi [North], (2732DC)[2732DA], 5 November 1985, Van Wyk 611 (CPF0005264!, NH0100696-0!) ; Sodwana State Forest, Ozabeni, nr. Lake Bangazi [Bhangazi] North, (2732DA), 25 February 1997, Scott-Shaw 8543 (CPF0005276!) ; Sodwana State Forest, Ozabeni, nr. Lake Bangazi [Bhangazi] North, (2732DA), 25 February 1997, Scott-Shaw 8558 (CPF0005277!) ; Osabeni [Ozabeni], close to Lake Bangazi [ Bhangazi ) North , (2732DA), 5 October 1997, Matthews & Welman 1 (PRE!, PRU!) ; ISimangaliso wetland park [iSimangaliso Wetland Park], 2.3 km from Lake Bhengazi [Lake Bhangazi North], next to road ( Sodwana Bay ), 27.61114S 32.63185E, (2833AA) [2732DA], 16 June 2016, Parbhoo S & Doarsamy 183 (NH0145456-0!) GoogleMaps ; ISimangaliso [iSimangaliso] Wetland Park, Sodwana Bay, Ozabeni Grassland , ± 7 km north of Bangazi Lake [Lake Bhangazi North], 27.58363S 32.641578E, (2733CC) [2732DA], 13 November 2016, Mtshali 154 (NH0145399- 0!) GoogleMaps ; Isimangaliso [iSimangaliso] Wetland Park, Ozabeni section, 27°34’45.26”S 32°38’44.02”E, (2732DA), 15 October 2024, Boon 184 (NU0138914!) GoogleMaps ; Isimangaliso [iSimangaliso] Wetland Park, Ozabeni section, 27°34’45.26”S 32°38’44.02”E, (2732DA), 15 October 2024, Boon 185 (NU0138915!) GoogleMaps ; Isimangaliso [iSimangaliso] Wetland Park, Ozabeni section, 27°34’45.26”S 32°38’44.02”E, (2732DA), 15 October 2024, Boon 187 (NH!) GoogleMaps ; Mpangazi , [2732DC], 10 January 1964, Strey 5084 (NH0051495-0!, PRE0272996-0!) ; Sand Flats, Bazwana, Mpangazi , [2732DC], 10 January 1964, Strey 5047 (K002226682!, NH0051516-0!, PRE0270506-0!) ; St Lucia Park, W [west] of Bangazi Lake [Lake Bhangazi South], [2832AB/BA?], 12 December 1960, Ward 3631 (NU0032838!, NH0049280-0!, PRE0462805- 0!) ; St Lucia Park, west of Bangazi Lake [Lake Bhangazi South], [2832AB/BA?], 12 December 1960, Ward 3634 (NU0032796!) ; Lake St Lucia, East Shore [Eastern Shores], (2832AB), 11 January 1975, Taylor 279 (NU0032797!) ; St Lucia Estuary Game Park, E [east] of Vidal Road , [2832AB/AD?], 9 November 1977, Pooley 1935 (NU0032836!) ; St Lucia, E ( Eastern) Shores , 30 March 1981, MacDevette 22b (CPF0005263!) ; Western [Eastern] shores, Sibomvini [Ezibomvini], (2832AB), 10 March 1982, Nicholas 1257 (CPF0005260!, K002226583!) ; Eastern Shores State Forest, Sibomvini [Ezibomvini] grassland, (2832AB), 11 November 1985, MacDevette 349 (CPF0005262!, NH0085858-0!) ; Eastern Shores, Sibomvini [Ezibomvini], monitoring site MWG3, (2832AB), 10 December 2002, Scott-Shaw s.n. (CPF0005271!) ; Eastern Shores, Sibomvini [Ezibomvini], monitoring site MWG3, (2832AB), December 2003, Scott-Shaw s.n. (CPF0005272!) ; Isimangaliso [iSimangaliso] Wetland Park, St Lucia E., Sibomvini [Ezibomvini] South , 28°00’42”S 32°29’52”E [ca. 28°8’55”S 32°29’40”E], (2832AB), 16 February 2006, Ward, Jewitt & Taylor 16438 (CPF0005246!) GoogleMaps ; Isimangaliso [iSimangaliso] Wetland Park, Eastern Shores , 28°9’21.54”S 32°29’58.20”E, (2832AB), 13 October 2024, Boon 186 (NH0154846-0!) GoogleMaps ; Isimangaliso World Heritage Site [iSimangaliso Wetland Park], Eastern Shores, south-west of Lake Bangazi [Lake Bhangazi South], 28°9’3.42”S 32°0’6.74”E [ 28°9’4.90”S 32°30’6.76”E], (2832BA), 22 June 2019, Boon & Church 152 (NH0152513-0!) GoogleMaps .

Feely, Tinley & Ward 27 (K002226577!, NH048745-0!, CPF0032795! & CPF0032841!) is a gathering of O. bhangazica , but there is insufficient information on the label to determine near which Lake Bhangazi it was collected. The details of the gathering are: West of Lake Bengazi [Bhangazi], 16 January 1960.