A taxonomic review of the austral genera Asterochaete and Carpha (Cyperaceae, Carpheae)

In Eastern Cape Province, South Africa (Fig. 4a). In marshes and on rocky river banks in montane grasslands, at 1300–1700 m in altitude, where it may be a co-dominant in its community (Brand et al. 2013).

The specific epithet reflects the distinctive shape of the involucral bracts, which are broadened and more or less ovate at the base, tapering gradually to the long-acuminate apex, one of the key distinguishing features of this species.

Asterochaete acuminata resembles A. capitellata, but differs in characters of the leaves, involucral bracts, inflorescences and nuts. Its leaves are thick, the proximal half thickly crescentiform in cross-section, the distal half thickly crescentiform to subcircular in cross-section, whereas the leaves of A. capitellata are thin and V-shaped in cross-section. Its bracts are ovate at the base with a long acuminate apex, 2–10 cm longer than the inflorescence or heads of spikelets, whereas the bracts of A. capitellata range from ovate to only slightly broadened at the base and are the same length as or only slightly longer than the heads of spikelets. The spikelets of A. acuminata are in one or few oblong heads rather than in ovoid heads as in A. capitellata. Its nutlet surface is finely reticulate (epidermal-cell outlines ~0.05 mm in diameter) and puncticulate, whereas the nutlet surface of A. capitellata is reticulate (cell outlines ~0.1 mm in diameter) and not or rarely puncticulate (Zhang et al. 2006, fig. 3a).

Asterochaete filifolia also has relatively long involucral bracts (the proximal bracts are usually 1.5–3 times longer than the inflorescence) and single-noded culms, but differs from A. acuminata in having narrow bases to the involucral bracts (including the bracts subtending the cluster(s) of spikelets); spikelets in one very dense conical cluster (rarely two clusters); leaves filiform, 0.5–0.8 mm in diameter; and nutlets 3.5–4.3 mm long.

Two specimens of this species, MacOwan 1616 and 2187 (see Zhang et al. 2006, fig. 14), were cited as Carpha bracteosa by Clarke (1898) when he described that species. The application of that name is fixed in this paper by lectotypifying it (see below under A. capitellata) on the third specimen cited by Clarke, which is A. bracteosa as traditionally understood. More recent specimens of this new species were also identified as C. bracteosa by other workers. These specimens were temporarily named as Carpha cf. bracteosa in our analyses (Zhang 2002; Zhang et al. 2006).

Of limited distribution but does occur in national parks; we suggest a status of ‘Data Deficient’ (IUCN Standards and Petitions Committee 2022).

South Africa: Eastern Cape: Uitenhage Division, Great Winterhoek Mountains, The Cockscomb, 30 Nov. 1958, E. Esterhuysen 28040 (K 000480712, PRE n.v.); in summo monte Boschberg, alt. 4800 ft [~1460 m], Somerset Division, Dec. [?year; received at Kew March 1871], P. MacOwan 1616 (K, Cibachrome print in NE, SAM 24258-1, SAM 24258-2); in paludosis summi Boschberg, 4500 ft [~1370 m], Nov.–Dec. [?year; received at Kew 1874], Somerset Division, P. MacOwan 2187 (K, Cibachrome print in NE); Amatole Mountains, below Gaikas Kop, 1 Dec. 1983, H.D. Furness 347 & P.B. Phillipson (K, UFH n.v.); Robertson Falls, 1500 m, 26 Nov. 1994, B. Sonnenberg 336 (NU); ridge W of Mount Thomas, 4500 ft [~1370 m], 11 Nov. 1948, R. Storey 3682 (K, PRE n.v.).

Tropical East Africa, on Mount Karisimbi, Mount Mgahinga, Mount Muhavura in the Virunga Mountains along the border between Uganda, Rwanda and the Democratic Republic of Congo; also on Mount Bigugu (SW Rwanda) and Mount Kahuzi (D.R. Congo) (Fig. 4b). In alpine bogs and heaths at ~2400‒3300 m in altitude.

Asterochaete angustissima was previously treated either as a species of Carpha (Chermezon 1935; Haines and Lye 1983) or as a variety of C. eminii (=Asterochaete eminii) (Kükenthal 1939b; Napper 1964), which is in the nearby Rwenzori Mountains in western Uganda. Our studies (Zhang 2002; Zhang et al. 2006) indicated that A. angustissima would be better treated as a species rather than a variety of A. eminii. It differs from A. eminii by usually having two flowers in each spikelet, with the axis elongated above the proximal fertile glume and adnate to the fertile glume base, the shorter perianth bristles (inner whorl (1.8–)3–4.4 mm long, and the outer whorl even shorter), and the leaf blades slender and usually canaliculate to thinly but stiffly crescentiform in cross-section, 0.7–1.6(–2.0) mm wide. A. eminii has one-flowered spikelets, with the axis not elongated above fertile glumes, longer perianth bristles (inner whorl 3.5–6.7 mm long), and the leaf blades V-shaped in cross-section, 2.0–4.0 mm wide at mid-length. Plants of A. angustissima are generally more slender, with few and shorter spikelets in the inflorescence, with shorter, paler glumes, and leaf sheaths that are more brightly coloured.

The third tropical montane African species, A. ulugurensis (q.v.), endemic to the Uluguru Mountains in Tanzania, is generally more robust in its parts than are A. angustissima and E. eminii, with darker glumes.

Enquiries at STR (where Chermezon’s main herbarium is held) failed to find either of the Humbert collections cited in the protologue (Françoise Deluzarche 2004, pers. comm.). In P, there are sheets of these collections that are annotated by H. Chermezon so it is reasonable to choose one of these as lectotype. The sheet of Humbert 7722 that is labelled by Chermezon as ‘Carpha angustissima H.Chermezon’ is chosen because it is labelled as having been in ‘Hb. Chermezon’ (printed label) and has been annotated by Chermezon. There is a probable isolectotype in B (10 0000961) that has not been annotated by Chermezon. In P, there is another sheet of Humbert 7722 (P 00742030) that differs in collecting details; according to the printed label it was collected ‘vers 3000 m’ in March (‘Mars’) 1929. Probable duplicates of this sheet are in B (10 0000960) and NSW (384362) but these differ in having the date as ‘Février – Mars’. Given these different details, we do not consider any of these variant sheets to be isolectotypes. The two sheets of Humbert 8586 in P have not been determined by Chermezon in his own handwriting, although on one sheet he has added habitat information. The sheet in B (10 0000959) has been determined by Chermezon.

Of limited distribution but some populations are in national parks; listed as ‘Endangered’ (Beentje 2017, as Carpha angustissima), which seems appropriate given that the high-altitude habitat is limited and predicted to become even more so with changing climate and increasing human impact.

Rwanda: Mount Bigugu, 2870 m, Prefecture Cyangugu, 4 Mar. 1972, P. Bamps 3297 (BR 13663056*, P 199387); Forêt de Nyungwe, sommet du Mont Ibigugu, alt. 2954 m, Prefecture Cyangugu, 1972, G. Bouxin 1360 (BR 13663117*, P 199385, L 486009); Mont Bigugu, Terr. Shangugu, 2900 m, 26 Mar. 1956, R. Christiansen 1642 (BR 13663063*, P 03471933); Parc des Volcans, piste vers la selle Ngahinga-Muhavura, Préf. Ruhengeri, 2840 m, 31 July 1974, P. Van der Veken 10449 (BR 13663094*, P 01948986).

Uganda: Kigezi District: crater of Gahinga Mountains, 3000 m, Bufumbira County, 24 Apr. 1970, A.B. Katende & K. Lye K207B (K); Gahinga-Muhavura Saddle, 3050 m, Bufumbira County, 24 Apr. 1970, K.A. Lye 5289 & A. Katende (EA, K, P 03471934); Muhenva-Mgahenga Saddle, 10,000 ft [~3000 m], Sep. 1946, J.W. Purseglove 2199 (BR 13662998*, EA, K); Western Province, Virunga-Kette, Gahinga, ~3400 m, 15 Nov. 1954, H.U. Stauffer 793 (BR 24800396*, K, P 03471932).

Democratic Republic of Congo: Mount Kahuzi (Kivu), 3000 m, 25 May 1958, A. Meurillon 626 (BR 13663124*, K); Mt Kahuzi, 3300 m, Province Kivu, 29 Dec. 1971, G. Troupin 14282 (BR 13663148*, LWI 444303451 n.v., P 00199386); Kivu, Parc National du Kahuzi-Biega, Mont Kahuzi, alt. 3200 m, 29 Dec. 1971, P. Van der Veken 8882 (AAU, BR 25390414V*).

Endemic to the Mascarene Island of Réunion (previously called ‘Ile Bourbon’), in the Indian Ocean (Fig. 4b). In montane shrubland in deep clay-loam soil on basalt at high altitude (~1950–2300 m).

Asterochaete borbonica is more slender than A. nitens, the other species on Réunion (also endemic there). It differs from that species principally in being shortly rhizomatous, with more or less filiform leaves crescentiform in cross-section and with shorter perianth bristles. Some specimens are morphologically intermediate (e.g. Bruhl 2984 and 2986), with leaves U-shaped in cross-section and up to ~2 mm wide without prominent midrib abaxially, and with longer perianth bristles (up to 5 mm long). Further study is needed to determine whether these names represent extremes of one taxon.

Asterochaete borbonica is a slender species with nearly filiform leaf blades that are thickly channelled in cross-section, 0.5–0.7 mm wide and lacking an obvious midrib abaxially. The uppermost leaf at the base of a culm (and the lowest primary involucral bract) has the junction of the leaf sheath and blade marked by an elongated V-shaped opening in the sheath opposite the attachment of the blade, with the margins of the V often slightly overlapping towards its base or narrowly U-shaped.

The inflorescences in A. borbonica are slender and short (up to 11 cm long). The perianth bristles are shorter than or subequal to the fruit body, 1.8–2.5 mm long, often with some bristles less than half the length of the body. The bristles are more or less terete in cross-section, covered densely by short, rigid hairs and of equal diameter throughout their length, tapering abruptly to the relatively much-enlarged, short, smooth basal section. The nutlets are obovoid to ellipsoid, and the persistent stylebase is short (<0.5 mm long) and not clearly differentiated from the upper part of the style (Fig. 1c).

Asterochaete nitens is a generally taller and more robust species with leaf blades broadly V-shaped in cross-section, 2.5–5 mm wide and with an obvious midrib abaxially. The opening at the sheath–blade junction of the uppermost leaf is mostly deeply U-shaped, often asymmetrically so, but V-shaped openings, with the margins overlapping or not, are occasionally found on the same plant as the more common U-shape (e.g. J.J. Bruhl 2976). The inflorescences of A. nitens are slender but usually more elongated (20–50 cm long). The perianth bristles are either all much longer than the fruit body (up to twice as long), 2.8–5.0 mm long, or with the outer three bristles occasionally shorter than the body. The bristles are more or less flat in cross-section towards the base, covered with more or less sparse longer, softer hairs. They are filiform near the apex, expanding gradually over their length to be relatively much broadened and flat for the lower third of their length above the short, smooth basal section, which is more obvious than in A. borbonica. The nutlet is ellipsoid to obovoid, with the stylebase thickened compared with the upper part of the style and approximately twice as long (up to ~0.7 mm long) as in A. borbonica. The nutlet is usually slightly larger than in A. borbonica but with similar reticulate surface (Fig. 1c).

These differences are slight but appear to hold for most specimens seen. However, there are a few slender specimens that resemble A. borbonica but have bristle and nutlet features similar to those in A. nitens (e.g. Bruhl 2979), and conversely a few coarser specimens similar to A. nitens but with bristle and nutlet features similar to those in A. borbonica (e.g. some of the five plants collected as Bruhl 2953a‒e from a population on the plateau at Saint Joseph, Foc-Foc, SW of Piton de la Fournaise). Further investigation is needed to clarify whether this indicates hybridisation between the two species or that there is a single variable species on the island.

Both type sheets bear labels in Steudel’s script. The lectotype bears the usual ‘Herbarium Steudel’ label. The isolectotype has a label bearing a manuscript diagnosis in Steudel’s script.

Of limited distribution but occurs in national parks; we suggest a status of ‘Vulnerable’ (IUCN Standards and Petitions Committee 2022), pending further investigation of its taxonomic status, and considering that the high-altitude habitat is limited and predicted to become even more so with changing climate.

Réunion: Bourbon, [no date; received at Kew Oct. 1875], I.B. Balfour s.n. (K, fragment B 10 0000966 p.p.); overlooking Piton de la Fournaise, Nez Coupé de Sainte Rose, 1942 m, 14 Jan. 2011, J.J. Bruhl 2977, J. Hivert & S. Pichillou (BOL, BRI, K, MO, NE 104059, NSW, P, REU); Grand Bénard, environs du Maido, 2100 m, 22 Nov. 1968, T. Cadet 1461 (P 00592876); Plaine des Salazes, 2200–2400 m, 14 Feb. 1969, T. Cadet 1953 (P 592875); Ile de la Réunion, 21 July1875, G. de l’Isle 410 bis (K, P 00592874); Plaine des Caffres, Ile Bourbon, 1837, J.M.C. Richard 660 (NSW 538940).

At high altitude in Zimbabwe (only one record, at ~1900 m in the Mutasa District) and Madagascar (1200–2600 m from several massifs); at a wide range of altitudes (180–1500 m) in southern Africa, from KwaZulu-Natal to the Western Cape Province (Fig. 4b). In marshy places.

Asterochaete capitellata is a variable, widespread species and needs more detailed study. Clare Archer (née Reid) noted on some specimens in NU (e.g. Robinson 1976 from Zimbabwe and Getliffe, Hearne & Kerfoot 56 from the Transvaal) that these more northern specimens of this species have three of the perianth bristles much longer than the other three, whereas in Cape specimens the bristles are more or less equal in length. We have not found this to be a consistent difference in the material examined.

The variation in plant size and in glume colour over the geographic range of this species needs to be examined more closely too. The sole Zimbabwean specimen seen, Robinson 1976, is particularly slender, especially its inflorescence branches (Browning et al. 2020, fig. 14.64). Two slender specimens from KwaZulu-Natal (Linder 5516, BOL 60368) and Western Cape (Esterhuysen 18381, K) have less dense spikelet clusters and are generally less robust and rigid, but the fruit and bristles are similar to those of A. capitellata. Molecular study of populations will probably be needed to understand all of the variation in this species.

One obvious variant with very broad bases to the inflorescence bracts, found in the Cape Province of South Africa, was described as C. bracteosa by Clarke. Plants of this form can be found growing with more typical plants, e.g. Bruhl 1725 (broad bases) and 1726 (typical), and can be distinguished by their bracts and somewhat more robust appearance.

Carpha bracteosa was treated as a variety of C. capitellata by Kükenthal (1939a) and later as conspecific with C. capitellata by Levyns (1950) and Archer (2000), whereas other authors (Clarke 1894, 1898; Reid and Arnold 1984; Browning and Guthrie 1994) have recognised it as a distinct species. Archer (2000) treated C. bracteosa as a synonym of A. capitellata. In our phenetic analyses (Zhang 2002; Zhang et al. 2006), the specimens of C. bracteosa and C. capitellata always formed a group, but with the specimens of C. bracteosa consistently forming a subgroup. These results suggest that C. bracteosa is only slightly different from typical C. capitellata, differing in having involucral bracts that are ovate towards the base rather than leaf-like bracts that are only slightly broader near the base (Zhang et al. 2006, fig. 13). These two forms are not isolated geographically and apparently have the same habitat in the Cape Province. No other character variation seems to correlate with the broad bracts of C. bracteosa, so it seems unwise to recognise this variant formally before further investigation. Robust plants are found also in Madagascar.

Carpha perrieri was described by Chermezon (1922) from Madagascar. He regarded it as close to C. capitellata but differing from that species by its slightly bigger glumes, its longer perianth bristles, its larger stylebase, and its narrower nutlet. We found that glume, perianth bristle and stylebase length and nutlet diameter of C. perrieri overlap with those of typical C. capitellata (Zhang et al. 2006, table 7). Specimens of the two species grouped together in the phenetic analyses, and C. perrieri does not have any unique characters to distinguish it from C. capitellata; therefore C. perrieri is here included in C. capitellata.

The CGE sheet of Asterochaete capitellata is the only sheet seen of the Forbes and Ecklon syntypes mentioned in the protologue and is here selected as the lectotype. The specimen consists of a complete tuft with leaves, inflorescence and a young rhizomatous shoot at the base. It has been annotated by Nees as ‘Asterochaete capitellata N ab E’. Ecklon’s specimens are found in numerous herbaria, but we have seen no specimen of this species labelled (by a code number, as detailed by Drège 1847; also Glenn and Germishuizen 2010) as coming from the Zwellendam (now Swellendam) district.

We have not seen any specimen recognisable as a type of Asterochaete glomerata var. β minor, but from the protologue description, this is probably merely a smaller plant of A. capitellata. Both Clarke (1898, p. 270) and Kükenthal (1939a, p. 116) cited this name as a synonym of A. capitellata. At the same time, Kükenthal cited an Ecklon specimen from ‘Waterfall bei Tulbagh’ as Carpha capitellata var. β bracteosa, so the Nees varietal name may apply to this form of A. capitellata with broad involucral bracts (see below).

Typification of Asterochaete tenuis is problematic. Kunth’s private herbarium ended up in B (Stafleu and Cowan 1979, p. 693; Hiepko 1987) but at least part of it seems to have been lost in World War II, and we have not seen any material of this species that is obviously recognisable as coming from Kunth’s herbarium or has been annotated by Kunth. We are not choosing a lectotype because further study could find authentic material from Kunth’s herbarium and meantime there is no ambiguity about the taxon to which the probable syntypes belong. There is a fragmentary specimen (loose glumes and nuts) of this taxon in B, Drège 2449, (‘zum Elefant rivier’, 10 Nov 1820) B10 0000968a, mounted on a sheet with fragments (a spikelet cluster and loose glumes and nutlets) from a more recent specimen of this species, Meebold 11724. This Drège material may have come from Kunth’s herbarium, but the label on the fragment was written by Kükenthal, so it is impossible to tell. He does not cite this specimen in his paper (Kükenthal 1939a). We have seen various sheets of C. F. Drège collections of this species, some numbered (e.g. Drège 1840 (in K) and 2031 (P 00461843) from Zuurberg Range, Alexandria Division, 2000–3000 ft (~600–900 m), and Drège 2449 (P00461845 and the fragmentary specimen B10 0000968) from Rooderberg 4000–5000 ft (~1200–1500 m) and some sheets without collection numbers, e.g. in K (2 sheets) and P 00461844. The latter sheets, along with the sheet of Drège 1840 in K, have the name ‘Asterochaete tenuis Kunth a’ written on them. Note that Drège codified his collecting localities by the letters a, b, c, etc. following the taxon name; see the full list of his collecting localities in South Africa in Drège (1843).

A Drège specimen in K (Zuurberg Range, Alexandria Division, 2000–3000 ft, ~609–914 m, C.F. Drège 1840, no date, stamped as coming from Bentham’s herbarium) has been annotated by C. B. Clarke as ‘Carpha capitellata forma tenuis; this differs from type capitellata in having 1 nut only’. Contrary to this statement, we found one or two nutlets per spikelet in the specimens of Drège 1840.

Asterochaete ludwigii has sometimes been regarded as a nomen nudum but the description in a long footnote in Krauss (1845) is more than adequate for valid publication. Krauss attributes the name and description to Hochstetter.

The lectotypification of the name Carpha bracteosa effected here preserves the traditional usage of that name. When Clarke (1898) published his new species C. bracteosa, he cited three specimens (Bolus 2867, MacOwan 1616 and 2187), all with involucral bracts broadened towards the base. Of the three syntypes, only Bolus 2867 belongs to this taxon as traditionally recognised and, therefore, Bolus 2867 is selected as lectotype here to maintain the application of the name. The other two syntypes, MacOwan 1616 and 2187 (Zhang et al. 2006, fig. 3b, 14), are considered to belong to the new species Asterochaete acuminata. C. capitellata differs from A. acuminata in having leaves with a keel, V-shaped in cross-section, involucral bracts more or less equal in length to the inflorescence, spikelets in ovoid clusters, and the nutlet surface reticulate.

Hans Heinrich Pfeiffer worked in Bremen for part of his career and his herbarium is at least partly in BREM, as are some of Schlechter’s African collections (Metzing 1999) but no Schlechter 8970 sheet is present in BREM (Michaela Grein, pers. comm.). We are not choosing a lectotype for Ficinia trinkleriana because it is possible that a specimen annotated by Pfeiffer will be found and, furthermore, all the syntypes and protologue agree as to which taxon they represent so there is no ambiguity in the application of the name. The type collection belongs to the broad-based form of A. capitellata.

There are two sheets in P of Perrier de la Bâthie 14555 with the same collection details and labelled as Carpha perrieri; one is from Chermezon’s herbarium (P 00199383), the other (P 00199384) has a general Paris herbarium label. Both bear adequate specimens of this same taxon, but neither has been annotated by Chermezon. The sheet from Chermezon’s herbarium has the more obvious connection with him and is therefore chosen as the lectotype of Carpha perrieri. None of the apparent replicates seen in other herbaria has been annotated by Chermezon. The sheets GOET 010743 and NSW 491742 of Perrier de la Bâthie 14555 bear the date February 1923, which may have been a slip of the pen. The material looks similar to that on the other sheets of this number dated February 1922, and may have been part of the same original collection.

Widespread in South Africa, including in national parks, so listed as ‘Least Concern’ (Contu et al. 2013, as Carpha capitellata). However, it has very restricted known distributions in Madagascar and Zimbabwe, so we suggest a status of ‘Data Deficient’ or possibly ‘Vulnerable’ (IUCN Standards and Petitions Committee 2022) for populations in those countries.

Madagascar: Massif de l’Andringitra (Iratsy), vallées de la Riambava et de l’Antsifotra et montagnes environnantes, 2000–2500 m, 27 Nov.–8 Dec. 1924, H. Humbert 3878 (B10 0000970, P 01880322, P 01948583); crête et rochers siliceux du sommet, Massif de l’Andohahelo, 1800–1979 m, 21–22 Oct. 1928, H. Humbert 6146928 (BM, BR 24800914*, K, P 00199390, P 00199391, P 00199393, US 1528711), 1200–1800 m, 18–26 Oct. 1928 (P 00199394), 1700–1850 m, 21–22 Oct. 1928 (B 10 0000972); Massif de Tsaratanana, 2300‒2800 m, Nov.–Dec. 1937, H. Humbert 18404 (P 01880323, P01880324, P 03471827); Massif de Tsaratanana, 2600 m, Nov. 1966, P. Morat 2307 (NSW 996706, P 01948582, P 00199389).

South Africa: Eastern Cape Province: Hopewell, ~600 ft [~180 m], Bathurst District, 30 Oct. 1964, J.P.H. Acocks 23507 (PRE); Zuurberg Range, Alexandria Division, [no date], C.F. Drège 1840 (K, two sheets); marshy places near Komgha, 2000 ft [~600 m], Sep. ?year [received at K Aug 1893], Komgha Division, H.G. Flanagan 920 (K); Humansdorp, 650 ft [~200 m], Oct. 1930, H.G. Fourcade 4476 (K); 7 miles [~11 km] E of Humansdorp along P.E. road, 21 Dec. 1943, H.G. Fourcade 6336 (BOL 63166); in humidis pr. [prope] Howisons Point, 1800 ft [~1490 m], [no date], P. MacOwan 351 (MEL 1543862).

Mpumalanga Province: Wakkerstroom, Oshoek District, ~1800 m, 3 Nov. 1979, N.J. Devenish 1821 (US); Mount Sheba Nature Reserve, 1960 m, Mpumalanga, Oct. 1975, F. Getliffe, [initial unknown] Hearne & O. Kerfoot 56 (NU). Lydenburg District, Steenkampsberg Pass near layby, 3 Dec. 1992, C. Reid 1807 (US). Western Cape Province: off R323, 200 m along Langkloof Road, 6 Dec. 1996, J.J. Bruhl 1718 & F.C. Quinn (BOL, CANB, GRA, K 486710, MO, NE 76468, NSW 599812, PRE); Franschhoek Pass, Route 45 between Villiersdorp and Franschhoek, 305 m, 7 Dec. 1996, J.J. Bruhl 1726 & F.C. Quinn (BOL, EIU, GENT, K, MO, NE 76340, NSW, NU, P, PRE); Worcester Division, SE slopes on Fonteintjietberg, 5500 ft [~1675 m], 26 Nov. 1944, E. Esterhuysen 10977 (BOL 63169); Kenilworth, Sep. 1938, M. R. Levyns 6846 (P 0371832); Cape Infanta, Oct. 1947, M.R. Levyns 8391 (K); Swartberg, 4900 ft [~1490 m], 7 Jan. 1975, M.F. Thompson 2282 (PRE).

Zimbabwe: Gairesi Ranch, 6 miles [~10 km] N of Troutbeck, Inyanga, 6200 ft [~1890 m], 21 Nov. 1956, E.A. Robinson 1976 (BR 24800884*, IRLCS n.v., K n.v., LISC n.v., NRGH n.v., NU, SRGH n.v.).

South Africa: Eastern Cape: valley N side of Outeniquas near Joubertina, Uniondale Division, 5 Nov. 1944, E. Esterhuysen 10611 (BOL 63167, PRE). Western Cape Province: Franschhoek Pass, Route 45 between Villiersdorp and Franschhoek, 305 m, 7 Dec. 1996, J.J. Bruhl 1725 & F.C. Quinn (BOL, CANB, GENT, K, MO, NE 76339, NSW 834199, NU); Du Toits Kloof, 19 Oct. 1947, M.R. Levyns 8624 (BOL 63165); Somerset West, Stellenbosch Division, 25 Oct. 1942, R.N. Parker 3744 (K); Jakkalsvlei, Jonkershoek, ~1300–2300 ft [~390–700 m], Stellenbosch District, 26 Sep. 1963, H.C. Taylor 5220 (K, PRE), ~1300 ft [~390 m], 14 Oct. 1964, H.C. Taylor 5988 (NU).

In southern parts of South Africa (KwaZulu-Natal and Eastern and Western Cape provinces) (Fig. 4c). In marshes, on stream banks and low-lying flats in near-coastal regions, at 90–500 m in altitude (Bruhl 1706 and Sonnenberg 387 seem to be at exceptionally high altitude for this species, at 1249–1500 m near Robertson Falls). This species was reported for Mount Uluguru in Tanzania by Haines and Lye (1983) and Hoenselaar et al. (2010), but that refers to the taxon here published as Asterochaete ulugurensis (q.v.).

This is the biggest species in the genus and has many more dense clusters of spikelets in the inflorescence than any other species. It resembles the somewhat smaller A. schlechteri, and the two species share a synapomorphy in their unusual perianth bristles, with the upper 2/3 or so of each bristle being terete and antrorsely scabrous, whereas the lower part of the bristle is much more slender and smooth. The upper part is readily broken off. Both species have a more elongated apical appendage on the anthers than in the other species.

There have been different views on how to distinguish A. dactyloides (universally referred to as C. glomerata or A. glomerata in past literature) and A. schlechteri, essentially based on the former being more robust than the latter. When Clarke (1904, p. 37) published his new species Carpha schlechteri, he described it as follows: ‘C. glomerata Nees affinis; differt: foliis angustioribus (vix 5 mm. latis); inflorescentia graciliore, spiculis minoribus; nuce grosse conspicue hexagono-reticulata levi (neque, ut in C. glomerata, minute obscure reticulata)’. He did not mention the actual size of the spikelet and nutlet of A. schlechteri in his description but he did say that the spikelets were smaller. Later, Clarke (1908) noted that A. schlechteri had leaves 4–5 mm wide, and nutlets 2–2.5 mm long. Examination of the type collection (R. Schlechter 10010) found the leaves to be 4.5–6 mm wide at their mid-length. Occasional specimens such as J. Browning 823 match the type and have nutlets 2.2–2.4 mm long but with leaves up to 8 mm wide.

Leaf width is a useful measure of the robustness of C. dactyloides, given that complete herbarium specimens are very rare because of the large size of the plants. The leaf width of A. dactyloides was first described by Boeckeler (1874, p. 265) as ‘7–12 lin. lat.’ (~14.8–25.4 mm wide), whereas Clarke (1898) described it as 1/4 inch (~6.4 mm) or more. Kükenthal (1939a) considered that A. dactyloides had leaves 5–12 mm wide, spikelets 6–8 mm long and nutlets 3.5 mm long, whereas A. schlechteri had leaves 4–5 mm wide, spikelets 4 mm long and nutlets 2.5 mm long. Levyns (1950) described the leaves of A. dactyloides as 2 cm wide, and Gordon-Gray (1995) as 15–28 mm wide. Reid and Arnold (1984) distinguished A. dactyloides by its more robust stature, and its leaves being 15–28 mm wide, whereas the leaves of A. schlechteri were 5–8 mm wide. Our examination of specimens found leaves to be usually 11‒25 mm wide (rarely some leaves as narrow as 7 mm wide) in A. dactyloides v. 4.5‒6 mm wide (rarely some leaves up to 8 mm wide) in A. schlechteri.

Our study also found spikelet characters useful for distinguishing A. dactyloides and A. schlechteri (Zhang 2002; Zhang et al. 2006). Asterochaete dactyloides has two bisexual flowers per spikelet, both of which mature a nutlet, and the axis is elongated above the proximal fertile glume, whereas A. schlechteri has spikelets with a proximal male flower, one bisexual flower, and the axis does not elongate above the proximal fertile glume. A. dactyloides has longer spikelets (6.2–9 mm long), longer inner-whorl bristles (3.1–3.8 mm long) and longer nutlets (2.5–3.4 mm long) than those of A. schlechteri (spikelets 4–5 mm long, inner-whorl bristles 2.3–3.4 mm long, nutlets 2.2–2.4 mm long).

Geographically speaking, A. dactyloides is much more widespread, occurring from near sea level to ~1400 m across south-eastern and southern South Africa, whereas A. schlechteri is found only at high altitude (900–1370 m) in the Koue Bokkeveld Mountains of the Western Cape Province. They are not known to grow together; however, further fieldwork is needed to investigate their distributions and habitats.

Vahl’s Schoenus dactyloides has been overlooked as the correct name for this species despite his providing a full description, probably because he cited S. glomeratus Thunb. as a synonym, and therefore his name has been incorrectly treated as illegitimate. However, Thunberg’s name (dating from 1794) is a later homonym of S. glomeratus L. (= Rhynchospora glomerata (L.) Vahl) from 1753. Vahl obviously recognised that two different taxa were involved because he included both Schoenus dactyloides and Rhynchospora glomerata in his publication (Vahl 1805, pp. 224 and 234 respectively). Under Art. 6.13 of the International Code of Nomenclature for algae, fungi, and plants (Turland et al. 2018), Vahl’s name could be treated either as publication of a new species or as a replacement name. We choose to treat this as a new species because the protologue description is clear and typification is clearcut. The holotype collection from the Cape of Good Hope from Burman’s herbarium that was given to Vahl is now in C as sheet C10001535. Carpha glomerata Nees is illegitimate because it cites the earlier name Schoenus dactyloides as a synonym.

Widespread, including in national parks; listed as ‘Least Concern’ (Foden and Potter 2010, as Carpha glomerata).

South Africa: Eastern Cape Province: 23 km from Uitenhage on road to Groendaldam, Oct. 1975, T.H. Arnold 1065 (PRE); Hogsback Range, N side of Robinsons Fall Creek, at top of falls, 1240 m, 1 Dec. 1996, J.J. Bruhl 1706 & F.C. Quinn (NE 76453, NU); along R102, N of Witelsbos, 210 m, 4 Dec. 1996, J.J. Bruhl 1711 & F.C. Quinn (K, MO, NE 76464A, NSW 6586779, NU); along R102, S of Witelsbos, 0.5 km N of Elands R crossing, 220 m, 4 Dec. 1996, J.J. Bruhl 1712 & F.C. Quinn (BOL, CANB, CHR, GENT, GRA, K, MO, NE 76460, NSW 599810, NU, PRE); Port Elizabeth to Storms River Road, past Humansdorp turn-off, 12 Feb. 1982, F. Getliffe 1142 (NU); Kabusie Forest, Stutterheim, 20 Oct 1980, O.M. Hilliard and B.L. Burtt 13204 (NU); 7 km uphill from Robertson Falls, 1500 m, 26 Nov. 1994, B. Sonnenberg 387 (NU); Witelsbosch SAFCOL Forests, 1 Feb. 1996, B. Sonnenberg 458 (NU). KwaZulu-Natal Province: Vernon Crookes Nature Reserve, 400‒500 m, Port Shepstone District, 1 Oct. 1989, J. Browning 228 (NU); Alexandra District, Station Dumisa, Farm Friedenau, Umgaye, 600 m, 22 June 1909, H. Rudatis 661 (NSW 996707, P 01948981, P 03471866, PRE n.v.); Farm Etheldale, adjacent Mtamvuna Forest Reserve, ~370 m, Port Shepstone District, 5 Sep. 1971, C.J. Ward 7196 (NU 54603, PRE n.v.). Western Cape Province: 32 km from Riversdale, 12 km from turnoff along road to Langekloof, Cape, Oct. 1975, T.H. Arnold 1041 (PRE); Umgegend von Hopefield, Weg nach Theefontein, Oct. 1885, F. Bachmann 866 (Z 80381*); Simonstown, Kogelberg Forest Reserve, junction of Louw and Palmiert Rivers, 300 ft [~90 m], Cape, 21 Nov. 1969, C. Boucher 911 (PRE); Devon Valley, Fransmanskraal Farm, 14 Nov 1978, C. Boucher 4091 (US); Verlorenvlei, Clanwilliam, 24 Jan. 1996, J. Browning 803 (NU); Franschhoek Pass, Route 45 between Villiersdorp and Franschhoek, 320 m, 7 Dec. 1996, J.J. Bruhl 1729 & F.C. Quinn (BOL, NE 76343, NSW 834200).

Known only from one collection from a damp cliff face below a shale band on the eastern side of Milner Peak at 1370 m in altitude, in the Western Cape Province of South Africa (Fig. 4a).

The specific epithet is from the manuscript name written on the two isotype sheets in K by South African botanist Trevor H. Arnold (1947‒). It reflects the conspicuous mixed colours of the glumes, one of the key distinguishing features of this species.

This species resembles Asterochaete capitellata and A. dactyloides in the structure of the inflorescence, spikelets and flowers, but differs in its leaves, which are thickened, channelled and more or less trigonous in cross-section, the fibrous old leaf bases around the tufts, the conspicuously coloured glumes, and the perianth bristles, which are plumose in the lower half (similar to those of Carpha sens. str. but the hairs are shorter and less dense), and antrorsely scabrous in the upper half. The leaves and longer involucral bracts become more or less uncinate with age and curl somewhat over their whole length. Its distribution overlaps with that of A. schlechteri but the species are not known to be sympatric. It differs in many features from A. schlechteri, including stature (A. discolor much more slender and smaller), leaf lamina cross-section (thick and channelled to hemispherical to trigonous in A. discolor; V-shaped in cross-section in A. schlechteri); inflorescence with one or two nodes in A. discolor; usually many-noded in A. schlechteri); spikelets length (~10 mm long in A. discolor; 4–5 mm long in A. schlechteri); perianth and nutlet morphology (cf. Fig. 1g, k).

Known only from the type locality; we suggest a status of ‘Data Deficient’ or possibly ‘Vulnerable’ (IUCN Standards and Petitions Committee 2022).

South Africa: type collection only.

Endemic to the Rwenzori (formerly Ruwenzori) Mountains straddling the border of the Western Province of Uganda and the Democratic Republic of Congo (Fig. 4b). In mountain bogs and on stream banks at high altitude (~2700‒3750 m).

Kükenthal (1939b) treated Carpha angustissima as a variety of A. eminii, and this was followed by Napper (1964), but our analyses (Zhang 2002; Zhang et al. 2006) indicated that it should be maintained as a separate species, in agreement with Chermezon (1935) and Haines and Lye (1983).

These two species are allopatric; Asterochaete eminii is found on the Rwenzori Mountains (west of Kampala) in western Uganda and A. angustissima on the Mgahinga Mountains in south-western Uganda. A. eminii differs from A. angustissima in being slightly more robust and by usually having 1-flowered spikelets with the axis not elongated above the fertile glumes and not adnate to the fertile glume base, and the glumes pale to dark yellow-brown; the leaf blades V-shaped to nearly flat in cross-section, 2.0–4.0 mm wide; and longer perianth bristles (the inner whorl 3.5–6.7 mm long). A. angustissima has two flowers per spikelet with the axis elongated above the fertile glumes and adnate to the fertile glume base, and the glumes pale yellow-brown; the leaf blades thinly crescentiform in cross-section, 0.7–1.6 mm wide; and shorter perianth bristles (the inner whorl (2.0–)3–4.4 mm long).

The third African tropical montane species is A. ulugurensis (q.v.), found in the Uluguru Mountains of the Eastern Arc in Tanzania.

The fragments (glumes, anthers and an immature nutlet with associated bristles much longer than the fruit) in B 10 0000975 were apparently removed by Kükenthal from the main type specimen for examination during his work on this group. The label of this sheet is in Kükenthal’s script. This material is identifiable as A. eminii. The main specimen was probably destroyed in B during World War II. A complete duplicate of Stuhlmann 2439 may yet be found in another herbarium, so we refrain from designating the fragments in B as lectotype.

Of very restricted distribution but occurs in conservation reserves; listed as ‘Vulnerable’ (Luke 2010) but ‘Endangered’ may be more appropriate given that the high-altitude habitat is limited and predicted to become even more so with changing climate and increasing human impact.

Uganda: Bujuka, Mount Ruwenzori, 11,500 ft [~3500 m], Aug. 1931, C.W.L. Fishlock & G.L.R. Hancock 109 (K); Mount Ruwenzori, near Nyamileju Hut, 10,500 ft [~3200 m], 30 Dec. 1968, R.W. Haines 277 (BR 13662936*, K); Bujuku Valley, near Bigo camp, Mt Ruwenzori, 3400 m, 24 Mar. 1948, O. Hedberg 435 (EA, K, LD); Mijusi Valley, Mount Ruwenzori, 3850 m, 30 Mar. 1948, O. Hedberg 598 (EA, K); Kasese, Busongora County, above Bigo camp, 3050 m, 21 Jan. 1981, A. Katende 3003 (BR 13662929*, MO, PRE, US 3398837); Muteinda, Mount Ruwenzori, 11,900 ft [~3625 m], 25 Dec. 1961, J.P. Loveridge 120 (K); near Nyamileju (2nd) Hut, Mount Ruwenzori, 3300 m, 30 Dec. 1968, K.A. Lye 1249 (EA); above Kichucha, Mount Ruwenzori, 3000 m, 31 July 1953, H.O. Osmaston 3210 (K); Butahu Valley, Kitandara, Mount Ruwenzori, 3500 m, Aug. 1953, H.O. Osmaston 3788 (13662912*, K); Mount Ruwenzori, 12,500 ft [~3810 m], Aug. 1938, J.W. Purseglove P270 (K); Mount Ruwenzori, 12,500 ft [~3810 m], 21 Feb. 1906, A.F.R. Wollaston (BM 000092183).

Democratic Republic of Congo [sic on label]: affluent de la Mososa (Cascade), E de Mahungu (Ruwenzori), 3180 m, 23 May 1953, G.F. De Witte 9104 (BR 13662950, K, P 01724486).

Restricted to the sandstone and basalt formations of the upper Karoo System in the Drakensberg Mountains of South Africa (Eastern Cape, KwaZulu-Natal, Free State and Mpumalanga provinces) and Lesotho (Fig. 4d; Reid and Arnold (1984, fig. 1). In marshes and along stream banks at high altitude (1800–2800 m) in the Drakensberg Grassland (Bastervoetpad) (Archer 2019). Co-dominant with Bulbostylis schoenoides (Kunth) C.B.Clarke in one community (Brand et al. 2013).

The features of this species are compared with the other South African species by Reid and Arnold (1984), and a key is given there. This species differs from other South African species in being tufted and having the following features: unnoded or single-noded culms; inflorescence of one very dense obconical cluster (rarely two clusters) of spikelets; filiform leaves (0.5–0.8 v. 1.5–24 mm wide in other southern African species), and a longer nutlet (3.3–4.3 v. 2.0–3.5 mm long in other southern African species). The overall colour of plants of this species is distinctively golden brown, particularly the rather shiny or glossy leaf sheaths. Some tussocks show obvious signs of regenerating after fire.

Its glumes are distichous to subdistichous in their arrangement rather than spiral as Reid and Arnold (1984) indicated, and spikelets have two bisexual flowers without an empty uppermost glume (the uppermost glume, described by Reid and Arnold (1984) as sometimes subtending a male flower or empty, was not found by Browning and Guthrie (1994) – also not found in this study). The specimen Devenish 1067 is unusual in the inner whorl of bristles being short-plumose for ~1/3 of their length, rather than only near the base of these bristles.

Although we did not examine the holotype, its image is on Global Plants JSTOR, the protologue has a clear illustration, and we examined other specimens cited in the protologue, such as Devenish 1067 and 1821, Hilliard & Burtt 9788, 13462, 16258 and 16677, Hoener 2138, and Killick & Vahrmeijer 3595.

Widespread, including in national parks, so not considered to be at risk: listed as ‘Least Concern’ (Raimondo et al. 2009, as Carpha filifolia).

South Africa: Eastern Cape: Elliot–Maclear district boundary, Bastervoetpad, ~7200 ft [~2195 m], 15 Feb. 1983, O.M. Hilliard & B.L. Burtt 16677 (K, NU 3500285). KwaZulu-Natal: Giants Castle Nature Reserve, Drakensberg Mountains, near top of Giants Ridge track, 2200 m, 24 Nov. 1996, J.J. Bruhl 1701 & F.C. Quinn (CANB, K, MO, NE 76458, NSW 599809, NU); path from Loteni Nature Reserve to Redi, Mpendhle District, 7700 ft [~2345 m], Sani Pass, ~8000 ft [~2435 m], Underberg District., 22 Mar. 1977, O.M. Hilliard & B.L. Burtt 9788 (K, 2 sheets, NU 3500280, E, P 03471846, PRE); Cobham Forest Reserve, ~6800 ft [~2070 m], Underberg District, 16 Feb. 1979, O.M. Hilliard 12509 & B.L. Burtt (K, NU 3500282, PRE 0557011-0); path from Loteni Nature Reserve to Redi, 26 Dec. 1982, O.M. Hilliard & B.L. Burtt 16098 (E n.v., K, P); Highmoor Forest Reserve, SE of Giants Castle,7700 ft [~2345 m], Mpendhle District, 6 Jan. 1983, O.M. Hilliard & B.L. Burtt 16258 (E n.v., K, NU 3500279); vicinity of Tarn Cavek above Bushman’s Nek, ~8000 ft [~2435 m], Underberg District, 20 Nov. 1983, O.M. Hilliard & B.L. Burtt 16813 (E n.v., K, PRE n.v.); vicinity of Tarn Cavek above Bushman’s Nek, ~8000 ft [~2435 m], Underberg District, 22 Jan. 1984, O.M. Hilliard & B.L. Burtt 17489 (E n.v., K); Little Berg, Highmoor Forest Station, 6500 ft [~1980 m], 14 Jan. 1966, D.J.B. Killick & J. Varhmeijer 3595 (PRE). Mpumalanga Province: Wakkerstroom, Oshoek District, 6600 ft [~2010 m], 14 Nov. 1963, N.J. Devenish 1067 (K, PRE); Wakkerstroom, Oshoek District, ~1800 m, 3 Nov. 1979, N.J. Devenish 1821 (K, PRE n.v.).

Lesotho: Base of Matadeng Pass, Sehlabathebe National Park, 25 Jan. 1995, J. Browning 696 (NE 87962, NU 3500299); between Pocha and Tsoelikana River, ~2300 m, Sehlabathebe National Park, 10 Jan. 1979, F.K. Hoener 2138 (K, NU 23103-0*, PRE); Sehlabathebe National Park, 5 Jan. 1990, C. Schwabe 0171 (K – 2 sheets, NU 3500296).

Endemic to the island of Réunion in the Indian Ocean (Fig. 4b). Growing in dry low shrubland, in skeletal soils on laval rock pavements, and by roadsides on moist rock faces, at middle to high altitude (~1200–2300 m). The plants in the more exposed habitats are not as tall as those from more protected habitats but are similar in their leaves and other features.

This is one of two species growing on Réunion, both at high altitude. A. nitens differs most obviously from A. borbonica in having broader leaf blades that are keeled and broadly V-shaped to nearly flat in cross-section, a longer inflorescence, and with inner perianth bristles longer than in A. borbonica (2.8–5.0 v. 1.8–2.5 mm). See further discussion under A. borbonica, including the status of these taxa. Apparent morphological intermediates are listed below.

The type of Asterochaete nitens is labelled as coming from Mauritius (‘île Maurice’ in French) rather than Réunion, which, together with Rodrigues, are known collectively as the Mascarene Islands. However, this was presumably a mistake in labelling, as noted by Jean Raynal on the isotype in P, which is labelled ‘Ins. Borbon’ and ‘Ile de Bourbon’ (the old name for Réunion). The type is very much more likely to have come from Réunion, which has peaks rising to 3000 m in altitude, whereas Mauritius is a much lower island, with its highest peaks reaching only 800 m in altitude. All other known specimens come from high-altitude areas on Réunion.

The probable isotype in P is labelled as coming from the ‘Herbier Petit-Thouars’ but with the locality given as ‘Ile de Bourbon’ (=Réunion). An unknown person has pencilled ‘Carpha aubertii’ on the sheet, and Jean Raynal annotated it ‘Double de l’échantillon Asterochaete ?nitens hb. Willdenow 1119 qui porte l’indication ‘Maurice’ érronée’. That is, Raynal believed that the specimen had not come from Mauritius. The fragments in B were apparently removed by Kükenthal from the parent specimen for examination during his work on this group. The label on the B 10 0000965 sheet is in Kükenthal’s script; it states that the fragments were from du Petit-Thouars’ herbarium, suggesting the sheet in P was the source.

Some have considered that Hooker (1883, in Icon. Pl. 15: t. 1434) made a new combination Cyclocampe nitens (Kunth) Hook.f., but his note discussing the relationships of Costularia arundinacea stated that he had not seen Asterochaete nitens and merely ‘assumed’ (his word) that it belonged in Cyclocampe. This does not seem adequate for a new combination.

The lectotype sheet of Rhynchospora boivinii in P was examined by KLW in May 2006. It was determined earlier by Jean Raynal as A. borbonica but KLW re-determined it as A. nitens on the basis that it has the habit of A. nitens and the perianth bristles are slightly less than 3 mm long and longer than the nutlet, which is a feature of A. nitens rather than A. borbonica. There is a fragmentary possible isotype or isolectotype of R. boivinii in B: Insel Bourbon, Boivin 999 on sheet B 10 0000966. These fragments (glumes and one fruit with perianth bristles exceeding the fruit) in a paper packet were probably removed by Kükenthal from the collection in P for examination during his study of this group. The label on this B sheet is in Kükenthal’s script, ending with ‘6.II.1935 vid. G. Kükenthal’. There is another small packet on the sheet B 10 0000966 that contains glumes and nutlets of A. borbonica (with bristles shorter than the fruit). It has been annotated by Kükenthal as ‘leg. I.B. Balfour’. These fragments are probably from a Balfour specimen in K that was annotated by Kükenthal in October 1935, in which case the fragments were presumably removed by him then.

Of limited distribution but some populations are in national parks. We suggest a status of ‘Least Concern’ (IUCN Standards and Petitions Committee 2022).

Réunion: Pelouse altimontaine du Plateau Kerveguen, 16 Nov. 1973, F. Badré 964 (P 03471835, P 00592878); Bas de Bellecombe, 20 Dec. 1973, F. Badré 1116 (P 01948936); Bourbon [Réunion], [no date; received at Kew Oct. 1875], I.B. Balfour s.n. (K); Roches Plates path, Rivière des Remparts, 7000 ft [~2130 m], 16 Nov. 1967, C. Barclay 501 (K); Rivière des Remparts, Roches Plates path, 5100 ft [~1550 m], 20 Nov. 1968, C. Barclay 1251 (K); Piton Fougères east of Dos D’ane [sic], 4400 ft [~1340 m], 21 Nov. 1970, C. Barclay 1920 (K); Rampe de la Grande Montée, 4800 ft [~1360 m], 22 Nov. 1970, C. Barclay 1966 (K); Gite du vulcan, 2200–2300 m, June 1957, J. Bosser 11441 bis (P 03471834); SW of Piton de la Fournaise, 2255 m, 13 Jan. 2011, J.J. Bruhl 2949, J. Férard, L. Gigord, H. Hivert & K.L. Wilson (BOL, CANB, K, NE 101298, NSW, NU, P, PRE, REU); SW of Piton de la Fournaise, 2263 m, 13 Jan. 2011, J.J. Bruhl 2953, J. Férard, L. Gigord, H. Hivert & K.L. Wilson (BOL, BRI, CANB, CHR, EA, GENT, HO, K, M, MEL, MO, NE 101958, NSW, NY, P, PRE, REU, SI); SW rim of Piton de la Fournaise, 2265 m, 13 Jan. 2011, J.J. Bruhl 2955, J. Férard, L. Gigord, H. Hivert & K.L. Wilson (BOL, CANB, EA, K, NE 104037, NSW, P, PRE, REU); La Plaine des Cafres, Route du Volcan, 1904 m, 13 Jan. 2011, J.J. Bruhl 2963, J. Férard, L. Gigord, H. Hivert & K.L. Wilson (BOL, BRI, CANB,K, MEL, NE 104045, NSW, P, PRE, REU); La Plaine des Cafres, Route du Volcan, 1904 m, 13 Jan. 2011, J.J. Bruhl 2964, J. Férard, L. Gigord, H. Hivert & K.L. Wilson (BOL, BRI, CANB, K, MEL, NE 104046, NSW, P, PRE, REU); La Plaine des Cafres, Route du Volcan, 1904 m, 13 Jan. 2011, J.J. Bruhl 2966, J. Férard, L. Gigord, H. Hivert & K.L. Wilson (BOL, CANB, K, NE 104048, NSW, NY, P, PRE, REU); La Plaine des Cafres, Route du Volcan, 1904 m, 13 Jan. 2011, J.J. Bruhl 2967, J. Férard, L. Gigord, H. Hivert & K.L. Wilson (BOL, BRI, CANB, EA, K, NE 104049, NSW, NY, P, REU); La Plaine des Cafres, Route du Volcan, 1904 m, 13 Jan. 2011, J.J. Bruhl 2968, J. Férard, L. Gigord, H. Hivert & K.L. Wilson (K, NE 104050, NSW, P, REU); La Plaine des Cafres, Route du Volcan, 1904 m, 13 Jan. 2011, J.J. Bruhl 2972, J. Férard, L. Gigord, H. Hivert & K.L. Wilson (K, NE 104054, NSW, P, REU); La Plaine des Cafres, near Col Bellevue, 1627 m, 13 Jan. 2011, J.J. Bruhl 2976, J. Férard, L. Gigord, H. Hivert & K.L. Wilson (K, NE 104058, NSW, P, REU); above Piton de la Fournaise, Nez Coupé de Sainte Rose, 14 Jan. 2011, J.J. Bruhl 2985, J. Hivert & S. Pichillou (BOL, BRIP, CANB, EA, K, NE 104067, NSW, P, PRE, REU); above Piton de la Fournaise, Nez Coupé de Sainte Rose, 14 Jan. 2011, J.J. Bruhl 2986, J. Hivert & S. Pichillou (BOL, EA, K, NE 104068, NSW, P, PRE, REU); Plateau des Basaltes (sur la Plaine des Remparts), 2400 m, 31 May 1966, T. Cadet 452 (P 03471842, P 03471843); Plaine des Palmistes, bord du la Route N3, 1400 m, 16 Dec. 1968, T. Cadet 1787 (P 00592868); Plaines des Salazes, 2400 m, 14 Feb. 1970, T. Cadet 1955 (P 00592869); Cirque de Cilaos, au dessous couteau Kerveguen, 2000 m, 16 Nov. 1973, T. Cadet 4507 (P 00592870); Sentier de la Roche Ecrite, above Saint Denis, ~1200 m, 25 Nov. 1973, M.J.E. Coode 4186 & F. Badré (K, P 03471845); hauts de St Leu, 22 Oct. 1875, G. De L’Isle 619 (P 0059287); Rempart de Bellecombe, 3 Dec. 1970, F. Friedmann 870 (P 00592879); Piton de la Fournaise, Volcano, Plaine des Sables, ~2200 m, 9 Nov. 1966, H-J. Schlieben 10904 (PRE).

Réunion: above Piton de la Fournaise, Nez Coupé de Sainte Rose, 14 Jan. 2011, J.J. Bruhl 2979, J. Hivert & S. Pichillou (BOL, K, NE 104061, NSW, P, PRE, REU [arching, slender tussock perennial herb; perianth segments exceeding nutlet, typical of A. nitens.]; above Piton de la Fournaise, Nez Coupé de Sainte Rose, 14 Jan. 2011, J.J. Bruhl 2980, J. Hivert & S. Pichillou (BOL, BRI, CANB, CHR, EA, K, MEL, NE 104062, NSW, P, PRE, REU [much more slender than usual for A. nitens, but with nutlets more typical of A. nitens]; above Piton de la Fournaise, Nez Coupé de Sainte Rose, 14 Jan. 2011, J.J. Bruhl 2983, J. Hivert & S. Pichillou (BOL, CANB, K, MO, NE 104065, NSW, P, PRE, REU) [as slender as typical A. borbonica, but with nutlets more typical of A. nitens].

Endemic to the Koue Bokkeveld mountains of the Ceres District of the Western Cape Province, South Africa (Fig. 4d; Reid and Arnold 1984; Archer and Muasya 2012). Only known from a few collections, so the distribution may be broader than currently known. In marshes and on sandy flats near streams, at 900–1370 m in altitude.

Carpha schlechteri was first described by Clarke (1904, p. 37) as ‘C. glomerata Nees affinis; differt: foliis angustioribus (vix 5 mm. latis); inflorescentia graciliore, spiculis minoribus; nuce grosse conspicue hexagono-reticulata levi (neque, ut in C. glomerata, minute obscure reticulata)’. That is, he considered it to differ from A. dactyloides in its narrower leaves, its fewer clusters of spikelets in the inflorescence and larger reticulations on the nutlet surface. We found that the reticulations on the nutlet surface of C. schlechteri are not obviously larger than those of A. dactyloides, and that spikelet cluster number in the two species overlaps slightly. The nutlet surface in A. schlechteri is unique in the genus in having the outer epidermal cell walls reticulate and colliculate, forming a layer that sloughs off readily as a single sheet to expose the inner walls of the epidermal cells as a reticulate pattern as seen in the other species in this genus. The epidermal layer is similar in A. dactyloides but not as obvious nor as readily sloughed off as in A. schlechteri.

As mentioned above under Asterochaete dactyloides, that species and A. schlechteri are similar morphologically, including having the same unusual form of perianth bristles, but A. schlechteri has the proximal flower male and produces only one nutlet per spikelet; the axis is not elongated above the proximal fertile node; spikelets are 4–5 mm long; the proximal fertile glume is 3.5–4.2 mm long and 1.2–1.8 mm wide, whereas the second fertile glume is 3.3–4.3 mm long and 0.5–1.5 mm wide; the nutlets are 2.1–2.4 mm long; and the inner-whorl bristles are 2.3–3.4 mm long. By contrast, A. dactyloides has two bisexual flowers in each spikelet, both developing nutlets; the axis is elongated above the proximal fertile node; the spikelets are 6.2–9 mm long; the proximal fertile glume is 5.8–7.5 mm long and 1.8–2.8 mm wide and the second fertile glume is 5.5–7 mm long and 1.5–2 mm wide; the nutlets are 2.5–3.4 mm long; and inner-whorl bristles 3.1–3.8 mm long.

Asterochaete schlechteri is found only at high altitude in the Koue Bokkeveld Mountains of the Western Cape Province, whereas A. dactyloides has a much wider distribution, often at a lower altitude, across southern South Africa.

The sheet in BR, annoted by Clarke with a draft description, is selected as lectotype. This specimen is labelled as being originally in the herbarium of the Hortus Thenensis (‘Horti Thenensis Herbarium’) in the town of Tienen (Tirlemont in French and latinised as Thenen), which was owned by Léon van den Bossche. As van den Bossche wrote in the preface to part 1 (de Wildeman 1904), he commissioned Emile de Wildeman (director of the Jardin Botanique de l’Etat de Bruxelles) to write the Plantae novae vel minus cognitae ex Herbario Horti Thenensis, which in practice meant that de Wildeman also edited contributions by specialists in particular groups, including Clarke (1904). In the preface, van den Bossche stated that the Schlechter specimens being used to describe the new species were in his personal herbarium. This means that, in the case of Carpha schlechteri, it could be argued that the BR specimen is the holotype of the name. However, if one takes the publication literally, Clarke’s protologue text does not state explicitly which herbarium the type collection is in, so Article 9.1 (Turland et al. 2018) does not apply. As well as the sheet in BR, Clarke also annotated the sheets BOL 63205 and the K sheet, indicating that he had seen more than one of the sheets of Schlechter 10010, which implies that these are syntypes (Art. 9.6) and that it is therefore appropriate to select a lectotype.

Collected only twice in the past 80 years, most recently in 1996; it is regarded as ‘Data Deficient’ (Raimondo et al. 2009, as Carpha schlechteri), but it is probably ‘Endangered’ (IUCN Standards and Petitions Committee 2022).

South Africa: Western Cape Province: Cederberg, Kromriviere, O. Niewoudt Farm, Worcester, 29 Jan. 1996, J. Browning 823 (NU, 3 sheets); Elands Kloof, Ceres Division, 3000 ft [~910 m], 13 Dec. 1946, M.R. Levyns 8098 (BOL 63206, K, PRE).

Endemic to the Lukwangule Plateau in the Uluguru Mountains, Morogoro District of Tanzania (Fig. 4b). In alpine peat bogs or marshes (at 2300–2500 m in altitude) in grassland.

The specific epithet refers to the locality of this species, and is the manuscript name written on a specimen in K (E.M. Bruce 742; see Fig. 7) by the late Ernest Nelmes, a cyperologist at the Royal Botanic Gardens Kew. The type was labelled as ‘Carpha ulugurensis Nelmes’ by the late Sheila Hooper, and another specimen in K (Bidgood et al. 232) was also so labelled by David Simpson in 1987.

The specimens of this species were provisionally named as Carpha ulugurensis and C. cf. nitens in our previous analyses (Zhang 2002; Zhang et al. 2006). This species resembles A. nitens (which is endemic on Réunion) but differs from A. nitens in inflorescence features, number of culm nodes, and nutlet length. Its culms are one-noded or unnoded, whereas A. nitens has (1–)2–3-noded culms. Its nutlets are longer than those of A. nitens (2.0–2.6 v. 1.5–1.8 mm), with a very short persistent stylebase. The epidermal cells of the fruits of A. ulugurensis are mostly slightly longitudinally elongated, whereas those of A. nitens are isodiametric or slightly horizontally elongated. The stylebase is very short and glabrous, whereas it is better developed in A. borbonica and A. nitens, with a hispidulous surface. It was illustrated as Carpha glomerata in Haines and Lye (1983, fig. 664, 665). These authors stated that the illustrations were based on specimen E.M. Bruce 742 from Mount Uluguru.

This is one of three tropicalpine species in eastern Africa, each endemic to high peaks in a different mountain range. It differs from A. eminii in having more numerous spikelets, with one or two flowers per spikelet, shorter glumes (4.1–5.2 v. 5.5–7.5 mm in A. eminii) and nutlets (2.0–2.6 mm long v. 2.8–3 mm in A. eminii), inner bristles shorter than to slightly exceeding the nutlet body (much longer than the body in A. eminii), and anthers 1.4–2.2 mm long (2.8–4.0 mm in A. eminii).

The isotype bears more complete collecting information than does the holotype. The collectors and number are given on the isotype as ‘T. Pócs, P. & K. Csontos, BJH [B.J. Harris] 3766’, and the habitat is given as ‘peat bogs with Sphagnum’ at an altitude of 2385 m.

Of limited distribution but some populations are in national parks; we suggest a status of ‘Vulnerable’ (IUCN Standards and Petitions Committee 2022), but ‘Endangered’ may be more appropriate given that the high-altitude habitat is limited and predicted to become even more so with changing climate and increasing human impact.

Tanzania: South Ulugurus, Lukwangule Plateau, 2400 m, 14 Mar. 1986, S. Bidgood, J. Lovett, J.L. Paul & T. Poćs 232 (K); Ulugurus, Lukwangule Plateau, 8200 ft [~2500 m], 30 Jan. 1935, E.M. Bruce 742 (BM, BR 24800907*, K, NU (fragment with photocopy of K sheet and notes by J. Browning), P).