Introduction

Work by the author towards an account of the Compositae tribe Heliantheae, for a forthcoming part of the Flora Zambesiaca treatment of the family, has necessarily covered what Orchard (2013) has called ‘the Wollastonia/Melanthera/Wedelia generic complex’. Orchard, likewise, is preparing a treatment of the tribe for the Flora of Australia and, during his recent visit to Kew, covered many of the problems with the author. The problems faced by both myself and Orchard were essentially the widely differing opinions in generic delimitation within a group of genera in the subtribe Ecliptinae Less. Of most concern to the present author were the African species of Melanthera Rohr covered in many floras and treatments to date (e.g. Bentham & Hooker 1873; Oliver & Hiern 1877; Wild 1965; Wagner & Robinson 2002; Mesfin Tadesse 2004; Beentje & Hind 2005), often in an ever-expanding concept.

When Melanthera was described by Rohr (1792), he only cited the locality in which it grew (‘Habitat prope oppidum divi Petri in Martinica, legi & descripsi April 24 anno 1784.’), to which Vahl added Bidens nivea L. (=Melanthera nivea (L.) Small) based on a citation by Swartz (1791) (a point apparently omitted by Wagner & Robinson in their discussion introducing Melanthera — Wagner & Robinson 2002: 550), although no combination was made by Vahl. Bentham & Hooker f. (1873) considerably expanded the genus adding Echinocephalum Gardner, Lipotriche R. Br., Psathurochaeta DC. and Wuerschmittia Sch. Bip. ex Walp., although it was some time before all of the necessary combinations were made; much later Wollastonia DC. ex Decne. (Wild 1965) and Aphanopappus Endl. (as Lipochaeta DC. sect. Aphanopappus (Endl.) Benth. & Hook. f. — the earliest apparent valid combination) were added (Wagner & Robinson 2001). Wagner & Robinson (2002) recognised 35 species.

In its narrow sense, Melanthera contains only a handful of species and, in the sense of Parks (1973), is restricted to the southern USA and the Caribbean with the type (Melanthera nivea (L.) Small) extending down the Andean countries into South America (south to Bolivia, based on Gutiérrez et al. 535 — MO, US, USZ!). As Orchard (2013) stated, they share the diagnostic characters of possessing discoid capitula, white 5-lobed corollas, anthers with black thecae, conduplicate oblanceolate entire scarious greenish paleae lacking an apical appendage, obcompressed achenes with truncate or slightly concave, setuliferous apices, and a pappus of few to many barbellate caducous awns. This restricted concept excludes the radiate taxa found in Echinocephalum, Wollastonia and Lipotriche.

In Melanthera sensu lato, Wild (1965) included all of the African taxa (Wollastonia, Lipotriche, Psathurochaeta and Wuerschmittia). If the widespread Wollastonia (sensu Orchard 2013) is removed, the remaining entities are considered congeneric and, under the earliest generic name available, Lipotriche, easily separated from Melanthera. Liptoriche species can be recognised by their heterogamous radiate capitula, female or neuter ray florets, yellow (to orangish) ray limbs, hermaphrodite disc florets, black anther thecae, conduplicate linear to oblanceolate, entire or 3-lobed appendaged yellowish paleae, and obcompressed, angled achenes with truncate or rounded setuliferous apices, and a pappus of up to 10 caducous retrosely barbed awns.

In examining the African taxa involved, by Orchard’s own admission, ‘Only a selection of taxa required to discuss illustrations’ were cited and ‘incomplete synonymy’ provided (Orchard 2013: 382) for the limited material he examined; he considered Lipotriche to be a genus of ‘about 14 species’.

The synopsis below provides a complete alphabetical listing of taxa considered by the present author to belong to Lipotriche. A total of 12 species, and five infraspecific taxa are recognised (including autonymic subspecies). Full synonymy, type citations and the location of known types, together with relevant notes and commentary on those types, is provided. Extensive use has been made of digitised material available in several virtual herbaria and a close examination of the IDC microfiche of the de Candolle herbarium; barcodes of accessions have been provided where available, although it is realised that this is an ongoing process and many more duplicate types will most probably be added in time; clearly material in K has been examined but is still being barcoded. The new combinations are proposed so that all relevant taxa now have names in the genus and will allow any future revision to be based on these names; in a similar vein, it was felt unnecessary to lectotypify many names except where clarification (by second-step lectotypification) was deemed necessary. Several glaring historical errors are highlighted and corrected, and an index to names provided. A key is provided.

In providing an indication of Conservation Status I have tended to err on the side of caution. This synopsis is a herbarium-based exercise using the material and literature available. As such, the label data is frequently insufficient to assess risks of extinction to many of the taxa, regardless of how showy some of them may be. Under the current ‘Guidelines’ (IUCN 2014) providing ‘DD’ (Data Deficient) for a taxon signifies that there is simply ‘inadequate information to make a direct, or indirect, assessment of its risk of extinction based on its distribution and/or population status.’ The ‘Guidelines’ also state that ‘Taxa listed in these categories [of NE and DD] should not be treated as if they were not threatened.’ Merely providing ‘DD’ should prompt more field work and this synopsis should be a guide to those taxa that remain poorly known.

A synopsis of Lipotriche R. Br.

Lipotriche R. Br., Observ. Comp.: 118 (Brown pre-Sept. 1817a); Trans. Linn. Soc. London 12 (1): 118 (Brown 1817b) [Feb. 1818]. Type: not originally cited. Lipotriche brownii (as brownei) DC. = Lipotriche scandens (Schumach.) Orchard

Psathurochaeta DC., Prodr. 5: 609 (de Candolle 1836). Type: Psathurochaeta dregei DC. = Lipotriche scandens (Schumach.) Orchard subsp. dregei (DC.) Orchard

Wuerschmittia [as Würschmittia] Sch. Bip., Flora 24 (1), Intell. 1, 2: 27 (Schultz Bipontinus 1841), nom. nud.

Trigonotheca Sch. Bip., Flora 27 (2): 672 (Schultz Bipontinus 1844), nom. inval., non Hochst. (Hochstetter 1841)(=Gymnosporia (R. Wight & Arn.) Hook. f., nom. cons. [celastraceae], based on the herbarium name Trigonotheca natalensis Sch. Bip., Flora 27 (2): 672 (Schultz Bipontinus 1844), nom. nud. pro syn. of Lipotriche brownii DC. and Psaturochaeta [sic!] dregei DC. [=Lipotriche scandens (Schumach.) Orchard].

Wuerschmittia [as Würschmittia] Sch. Bip. ex Walp., Repert. Bot. Syst. [Walpers] 6: 161 (Walpers 3 – 5 Sept. 1846); [as Wurschmittia] in Richard, Tent. Fl. Abyss. 1: 413 (Feb. 1848). Type: Wuerschmittia abyssinica Sch. Bip. ex Walp. = Lipotriche abyssinica (Sch. Bip. ex Walp.) Orchard

Key to species of Lipotriche

  • 1. Styles present in ray florets…..........................................................................................................2

  • Styles absent in ray florets (i.e. neuter)…..........................................................................................................5

  • 2. Pappus setae absent; leaf margins regularly serrate-crenate; phyllaries 8 – 12, ovate…..........................................................................................................richardsiae

  • Pappus setae 1 – c. 15; leaf margins often irregularly toothed; phyllaries numerous, usually narrowly ovate…..........................................................................................................3

  • 3. Ray limb c. 6 mm long; lamina elliptic, base cuneate or cuneiform; venation prominent beneath…..........................................................................................................elliptica

  • Ray limbs >10 mm long; lamina ovate, or broadly deltate-ovate, base cordate, truncate or cuneate; venation scarcely prominent or conspicuous…..........................................................................................................4

  • 4. Pappus setae 1 – 4 (5); leaf venation scarcely prominent; ray florets 10 – 12, ray limb c. 11 mm long…..........................................................................................................marlothiana

  • Pappus setae 8 – 15; leaf venation very conspicuous beneath; ray florets to c. 20, ray limb c. 17 mm long…..........................................................................................................scandens

  • 5. Leaves sessile or subsessile; much-branched herb to 2 m…..........................................................................................................abyssinica

  • Leaves conspicuously petiolate; few-branched herbs 0.3 – 2.5 m…..........................................................................................................6

  • 6. Stems with dense pale-brown woolly indumentum…..........................................................................................................felicis

  • Stems short-hispid to scabrid, or with patent hairs and lacking dense pale-brown indumentum…..........................................................................................................7

  • 7. Stems and leaves with long patent hairs; apices of paleae conspicuously elongated…..........................................................................................................gambica

  • Stem and leaves short-hispid to scabrid; apices of paleae abruptly acuminate or triangular…..........................................................................................................8

  • 8. Ray florets 7 – 8…..........................................................................................................9

  • Ray florets ≥12;…..........................................................................................................10

  • 9. Leaves (3- or) 5-lobed…..........................................................................................................tithonioides

  • Leaves rhomboid and unlobed…..........................................................................................................rhombifolia

  • 10. Ray florets c. 12…..........................................................................................................11

  • Ray florets to 20…..........................................................................................................pungens

  • 11. Pappus setae 1 – 4 per achene; most leaves (in dry material) flat, although sometimes conduplicate beneath inflorescence, membranaceous; primary leaf venation concolorous with lower leaf surface; mid-stem internodes to 13 cm…..........................................................................................................robinsonii

  • Pappus setae 7 – 9 (– 13) per achene; most leaves (in dry material) conspicuously conduplicate and arching, coriaceous; primary leaf venation paler than lower leaf surface; mid-stem internodes 6 – 7 cm…..........................................................................................................scaberrima

Lipotriche abyssinica (Sch. Bip. ex Walp.) Orchard, Nuytsia 23: 383 (2013).

Würschmittia abyssinica Sch. Bip., Flora 24 (1), Intell. 1, 2: 27 (Schultz Bipontinus 1841), nom. nud.

Wuerschmittia [as Würschmittia] abyssinica Sch. Bip. ex Walp., Repert. Bot. Syst. [Walpers] 6: 162 (Walpers 3 – 5 Sept. 1846); Richard, Tent. Fl. Abyss. 1: 413 (Feb. 1848). Types: [ethiopia:] ‘Herb. Schimp. it. Abyss. sect. I. no. 334 et 1533. In declivibus australibus regionis mediae montis Abyssinia Scholoda.’ Syntype: Schimper 334, B(100024123), BM(000924415, 000924416), BR(8362797), GH(00274778), GOET(001821), HBG(504884), JE(00019607), K(×2 — s.n., 000410453), LG(90024887), M(0105208), P(0072980, 00072981, 00072982), S07-16902, S07-16903, TUB(005508), WAG(0025131). Syntype: Schimper 1533, BR(8363121, 8876430), K (×2 – s.n., 000410454), M(0105206, 0105207), MO(694330), ?P, TUB(005507). Lectotype (selected by Wagner & Robinson 2002: 556): Schimper 334, P. This is considered as first-step lectotypification since they neither saw the sheet, nor were apparently aware of the presence of other duplicates. Second-step lectotypification (provided here): P(00072980 — consisting of two sheets clearly originally pinned to one another, one with profuse notes by Schultz Bipontinus and only a small capsule of material referred to [Schimper] 332, the second a flowering specimen simply labelled ‘Herb. E. Cosson, 18/Herb. Mus. Paris’. The second sheet is the material chosen as the lectotype, not the first!). The barcode system unfortunately fails here, since one barcode applies to two sheets which have been databased as being Schimper 334 — mentioning nothing about Schultz Bipontinus’s annotations on the two capsules as ‘332’. note: i) the basioym is often cited as Wuerschmittia abyssinica Sch. Bip. ex A. Rich., Tent. Fl. Abyss. 1: 413’ yet it is quite clear that Schultz Bipontinus’ name was used in Walpers’ Repertorium at least a year earlier as a descriptio generico-specifica; Walper also used Schultz Bipontinus’s manuscripts for both many names and descriptions. To this end, an extra ‘syntype’, of Quartin Dillon, is mentioned, but is of no nomenclatural value — the Schimper collections are the only syntypes; ii) Orchard (2013: 383) appears to have been unaware of the Schimper 1533 syntype, although one such isosyntype was mentioned under Schimper 334; no mention was made of the lectotype.

Melanthera abyssinica (Sch. Bip. ex Walp.) Vatke, Linnaea 39 (6): 496 (1875). note: Attribution is made by Oliver & Hiern (1877) and much later by Wild (1965) and Mesfin Tadesse (2004), that this combination was made by Bentham & Hooker (1873: 377); no such combination was made, but only inferred. Aké Assi (1964: 342) appears to have been the first to provide the correct authority.

Amellus abyssinicus (Sch. Bip. ex Walp.) Kuntze, Revis. Gen. Pl. 1: 306 (1891).

Melanthera abyssinica (Sch. Bip. ex Walp.) Vatke var. angustifolia Chiov., Annuario Reale Ist. Bot. Roma 8 (2): 183 (Chiovenda 1904). Type: [eritrea:] ‘Mensa: Aba Maitan-Dadà, 1600 – 1400 m., 8. I. 1893 (T[erracciano &] P[appi] n. 1975 bis).’ Holotype: FT(003757).

Melanthera djalonensis A. Chev., Explor. Bot. Afrique Occ. Franc. 1: 372 (Chevalier 1920), nom. nud. (based on Chevalier 18699 & 18829).

Melanthera sokodensis Muschl. ex Hutch. & Dalziel, Fl. W. Trop. Afr. [Hutchinson & Dalziel] 2 (1): 146 (1931). Type: ‘Togo: Sokode Farm, 1,000 ft. (Oct.)! Exsicc. Schröder 73.’ Holotype: ?B†. note: Wild (1965) suggested that the holotype was in K. Repeated searches have failed to find the material in K suggesting it might not have been there in the first place — which is highly likely. It might possibly have been in B, now destroyed, or is in K mounted with some other collection and curated elsewhere.

distribution . Sierra Leone, Guinea, Togo, Cameroon, Central African Republic, Congo (Kinshasa), Sudan, Ethiopia, Eritrea, Yemen, Tanzania, Angola, Zambia, Zimbabwe.

conservation status . A common and widespread species: LC (Least Concern).

Lipotriche elliptica (O. Hoffm.) D. J. N. Hind, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77141635-1

Melanthera elliptica O. Hoffm., Bot. Jahrb. Syst. 24 (4): 474 (Hoffmann Feb. 1898). Types: [togo:] ‘Togoland: Bismarckburg, Ketschenki (Kling n. 47, 162, 163; Büttner n. 34 – Juli–Oct.); Konkoa (Büttner n. 678).’ Syntypes: B†.

distribution. Togo, Ghana, Cameroon, Nigeria, Central African Republic.

conservation status. A widespread species: LC (Least Concern).

Lipotriche felicis (C. D. Adams) D. J. N. Hind, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77141636-1

Melanthera felicis C. D. Adams, J. W. Afr. Sci. Assoc. 8 (2): 135 (1964). Type: ‘Guinea. Faranna, October 1937, Jacques-Félix 1868’. Holotype: K; isotypes: P(016247, 016248, 016249).

distribution. Guinea, Mali.

conservation status. Most certainly under-collected: DD (Data Deficient).

Lipotriche gambica (Hutch. & Dalziel) Orchard, Nuytsia 23: 383 (2013).

Melanthera gambica Hutch. & Dalziel, Fl. W. Trop. Afr. [Hutchinson & Dalziel] 2 (1): 146 (1931). Type: ‘Gambia: ? locality! [1928] Exsicc. Hayes 586.’ Holotype: K(000410457).

distribution. Gambia, Guinea-Bissau, Mali, Senegal, Sierra Leone.

conservation status. A relatively widespread species, but clearly under-collected: LC (Least Concern).

Lipotriche marlothiana (O. Hoffm.) D. J. N. Hind, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77141638-1

Melanthera marlothiana O. Hoffm., Bot. Jahrb. Syst. 10 (3): 277 (Hoffmann 1888). Type: [namibia:] ‘Hereroland, Okahandja, in fruticetis, alt. 1200 m (Marloth n. 1332). — Florif. m. Majo 1886.’ Holotype: B†

Wedelia triternata Klatt, Bull. Herb. Boissier 4 (12): 839 (Dec. 1896). Type: [mozambique:] ‘Sambesigebiet: Nhaondue; nicht selten an feuchten Stellen, Menyhart, Juli 1891, No. 735.’ Holotype: GH(00014085).

Melanthera monochaeta Hiern, Cat. Afr. Pl. (Hiern) 1 (3): 581 (1898). Types: [angola:] ‘Loanda. — An annual, subscandent herb. In thickets near Quicuxe; fl. and fr. July 1858. [Welwitsch] No. 3547. An apparently annual herb; stem suberect or frequently quasi-scandent among bushes; leaves bright green in the living state. In dry and also in damp thickets between Teba and Quicuxe; fl. and fr. August 1858. [Welwitsch] No. 3548. Benguella. — By thickets along the river Cotumbella on the side of the town of Benguella, sporadic; fl. and fr. June 1859. [Welwitsch] No. 3550. Mossameded. — On gravelly bushy places along the banks of the river Bero, at Quipola; fl. and fr. June 1860. [Welwitsch] No. 3551.’ Hiern added the following observation: ‘The above Nos. from the district of Loanda differ from the others, which must be considered the type of the species, by the rather smaller capitula, also by less truncate achenes (No. 3548), or by the presence of a hairy ring at the top of the otherwise glabrous ovary in addition to the solitary arista (No. 3547); they perhaps belong to distinct species.’ pro parte excl. spec. Welwitsch 3547 et 3548. — Mendonça, Contrib. Conhec. Fl. Angol. 1, Compositae: 96 (1943). Lectotype (effectively selected by Wild, 1965: 10): Welwitsch 3551: BM(000924424); isolectotypes: COI(00005608), K(000410440), LISC(219579). Syntype: Welwitsch 3548, COI(00005609), K(000410441), LISU(219577). Syntype: Welwitsch 3550, BM(000924425), LISU(219578). Since Hiern was working on the ‘Study set’ of the Welwitsch collections in BM it seems unnecessary to provide a second-step lectotypification.

Melanthera varians Hiern, Cat. Afr. Pl. (Hiern) 1 (3): 580 (1898). Types: [angola:] ‘Barra de Bengo. — In swampy places by the river Bengo; fl. and fr. Dec. 1853. [Welwitsch] No. 3544. Golungo Alto. — By thickets and in the less dense forests; roadway near Menha-Lula; fl. and fr. July 1855. [Welwitsch] No. 3545. In bushy wooded parts of Sobato de Mussengue; fl. and fr. May 1856. [Welwitsch] No. 3546. Mossamedes. — Florets golden-yellow. In sandy thickets at the river Bero; fl. July 1859. The ovary of the disk-florets is more rounded than in the three previous Nos. [Welwitsch] No. 3543.’ Syntype: Welwitsch 3543, BM(000924429), LISU(219571). Syntype: Welwitsch 3544, BM(000924427), BR(8875341), K(000410443), LISU(219572), M(0105204, 0105205), P(00093209). Syntype: Welwitsch 3545, BM(000924428), LISU(219573). Syntype: Welwitsch 3546, BM(000924426), K(000410442), LISU(219574, 219575).

Melanthera baumii O. Hoffm. in Warburg, Kunene-Samb.-Exped. Baum: 418 (Hoffmann 1903). Type: [angola:] ‘Zwischen Goudkopje und Kakele, auf torfigem Boden, 1238 m ü. M. ([Baum] Nr. 197, blühend am 3. Oktober 1899.)’ Holotype: B†; isotypes: BM(000924430), COI(00005607), E(00239254), K(000410444), W(19010009468).

Melanthera schinziana S. Moore, Bull. Herb. Boissier, sér. 2, 4 (10): 1018 (1904). Type: [namibia:] ‘Südwest-Afrika: Amboland, Unkuanyama, Omupanda, Wulfhorst.’ Holotype: BM(000924423).

Melanthera rivicola Muschl. ex Dinter, Repert. Spec. Nov. Regni Veg. 19: 96 (1923), nom. nud. (based on Dinter 95).

Melanthera seineri Muschl. ex Dinter, Repert. Spec. Nov. Regni Veg. 19: 96 (1923), nom. nud. (based on Dinter I. 225).

Melanthera triternata (Klatt) Wild, Kirkia 5 (1): 9 (1965).

Lipotriche triternata (Klatt) Orchard, Nuytsia 23: 385 (2013).

note. All too often, ‘Melanthera triternata (Klatt) Wild’ has been cited as the accepted name for this taxon. Wild (1965: 9) originally made the combination clearly stating the later publication date of Klatt’s name (1896), citing Hoffmann’s earlier name (from 1888) immediately after in synonymy; I can find no explanation for this lapse in providing the relevant combination in Lipotriche — Wild (1967) simply repeated this error, and Orchard (2013: 385) followed suit, accepting Wild, one assumes.

distribution. Angola, Zambia, Zimbabwe, Mozambique, South Africa.

conservation status. A widespread species: LC (Least Concern).

Lipotriche pungens (Oliv. & Hiern) Orchard, Nuytsia 23: 383 (2013).

Melanthera pungens Oliv. & Hiern, Fl. Trop. Afr. [Oliver et al.] 3: 382 (1877). Type: [sudan:] ‘Nile Land. Djur-land, Schweinfurth!’ Material in K and BM give the following handwritten label data: ‘Buschwaldung bei Seriba Ghattar. 28 Juni 1869, Schweinfurth 1990’; the top and bottom of the label are preprinted ‘Reise nach Central Africa im Auftrage der Humboldt Stiftung./im Lande der Djur ges. v. G. Schweinfurth.’ Holotype: K(000410451, 000410452); isotypes: BM(000924413), M(0105203), P(00073056, 00073057). note: Beentje in Beentje & Hind (2005) noted only that the holotype was in K; one of the two sheets bears the full printed and annotated label, the duplicate has the briefest information written on the sheet — both are marked ‘bis’; the handwriting on the second sheet (and the ‘bis’ on the first sheet) is clearly that of Oliver. Orchard (2013: 383) was of the opinion that the duplicates were to be considered as syntypes.

Amellus pungens (Oliv. & Hiern) Kuntze, Revis. Gen. Pl. 1: 306 (1891).

subsp. caudata (Wild) D. J. N. Hind, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77141640-1

Melanthera albinervia O. Hoffm. subsp. caudata Wild, Kirkia 5 (1): 12 (1965). Type: [zambia:] ‘N. Rhodesia. Mufulira, Cruse 224’. Holotype: K (×2). note: Cruse 224 at K was mounted as ‘Sheet 1’ and ‘Sheet 2’ marked clearly on the labels. However, only ‘Sheet 1’ (a single flowering specimen complete with rootstock) was sent on loan to Wild (upon which the description was based), and it bears the full collecting label and Wild’s det slip indicating ‘type’; ‘Sheet 2’ has a poor transcription onto a ‘Herb. Hort. Bot. Reg. Kew.’ label, and the material is of four flowering shoots. ‘Sheet 2’ could be considered part of the holotype, or an isotype sheet.

distribution. Congo (Kinshasa), Zambia.

conservation status. A subspecies of apparently restricted distribution and certainly under-collected: DD (Data Deficient).

note. Whilst my concept of the species essentially follows that of Beentje (in Beentje & Hind 2005), but without recognising his varieties I have, however, considered that Wild’s M. albinervia subsp. caudata still worth recognising as a subspecies. All of the collections examined have the same short habit, consistently small leaves and a small xylopodiaceous rootstock with numerous wiry roots.

subsp. pungens

Melanthera albinervia O. Hoffm., Bol. Soc. Brot. 13: 30 (Hoffmann 1896). Type: ‘Angola, Quidumbo (Anchieta [49], [12] dezembro 1887).’ Holotype: COI(00005605); isotype: LISU.

Aspilia zombensis Baker, Bull. Misc. Inform., Kew 1898 (139): 152 (1898). Type: [malawi:]‘british central africa. Mount Zomba, alt. 4000 – 6000 ft., Whyte [s.n.]’ Holotype: K(000410438); isotypes: BM, SAM(0017630).

Melanthera acuminata S. Moore, J. Linn. Soc., Bot. 35 (245): 344 (1902). Type: [kenya:] ‘Hab. British East Africa, Kavirondo; G. F. Scott Elliot, no. 7052. [Ruwenzori Expedition — 1893-94].’ Holotype: BM(000924418); isotype: K(000410449 — bearing the original jeweller’s tag).

Aspilia zombensis Baker var. longifolia S. Moore, J. Linn. Soc. Bot. 35 (245): 345 (1902). Types: [malawi:] ‘Nyassaland, 1895; J. Buchanan, no. 24. [Ruwenzori Expedition, 1893 – 1894.] Shire country; [Dec.] G. F. Scott Elliot, no. 8555.’ Syntype: Buchanan 24, BM(000924422), MO(391642, 391643), SAM(0038658). Syntype: Scott Elliot 8555, BM(000924421), K(000410439 — bearing the original jeweller’s tag, which reads ‘855/Shire Highlands/Ndirani/Dec’, referable to Ndirande).

Aspilia eylesii S. Moore, J. Bot. 45 (530): 45 (1907). Type: [zimbabwe:] ‘Hab. Sebakwe; F. Eyles, 164.’ Holotype: BM(000924420); isotype: SRGH.

Melanthera ugandensis S. Moore, J. Bot. 54 (649): 256 (1916). Type: ‘Hab. Uganda, grassland at Uamananzi; [4000 ft, June 1915] Dümmer, 2564.’ Holotype: BM; isotype: K(000410450).

Melanthera gossweileri S. Moore, J. Bot. 65, Suppl. 2, Gamopet.: 58 (1927). Type: ‘Hab. Angola: in “Mumua” woods near Forte P[rinceza]. Amelia, Cubango, [December 1906, Gossweiler] 4184.’ Holotype: BM(000924419); isotype: COI(00005604). note: The COI isotype is quite different in the label data: ‘Vivas; multicaule; Disseminada pela [?Hiemisilva] na Vila da Ponte, Ganguelas, Huila, Desembro de 1906.’

Melanthera letestui Philipson, J. Bot. 77 (915): 87 (1939). Type: [central african republic:] ‘french equatorial africa: Oubangui-Chari; Upper Kotto Yalinga, 4 Aug. 1921, Le Testu 3055 (Type in herb. Le Testu; duplicate in Brit. Mus. Herb.)’. Types: BM(000924411, 000924414), P(00073054, 00073055), US(00125270). note: There seems to be great confusion as to where Le Testu’s private herbarium is preserved, although it is most probably the material in P; even though these sheets are marked as isotypes it is likely that P(00073054) could serve as lectotype; Philipson clearly worked with the material, and frequently stated that the duplicates were in BM.

Melanthera albinervia O. Hoffm. subsp. acuminata (S. Moore) Wild, Kirkia 5 (1): 12 (1965).

Melanthera pungens Oliv. & Hiern var. albinervia (O. Hoffm.) Beentje in Beentje & Hind, Fl. Trop. E. Afr. Compositae pt. 3: 742 (2005).

distribution. Mali, Congo (Kinshasa), Senegal, Central African Republic, Sudan, Kenya, Uganda, Tanzania, Angola, Caprivi Strip, Zambia, Zimbabwe, Malawi, South Africa.

conservation status. Widespread: LC (Least Concern).

note. Beentje (in Beentje & Hind 2005) was of the opinion that Melanthera albinervia was nothing more than a variety of M. pungens, there being few consistent differences between Wild’s subspecies. I have re-examined the types and other material concerned and agree with this view. M. albinervia subsp. caudata was only mentioned in the note under M. pungens var. albinervia — it does not belong there. However, I do not agree with many of the descriptions of the leaf pubescence provided in the literature. In much material the scabrid hairs are essentially adpressed, the basal portion often conspicuously verrucose and the apical cell (usually short- or long-acute) is smooth. In some material, though, the hairs are quite obviously erect and much less stocky, whilst still possessing the verrucose basal portion and with a much longer, finer smooth apical cell. There appears to be no geographical distribution associated with the different hair types, nor any apparent link with other characters.

There are three sheets in K from Zimbabwe and South Africa (‘TV’) that bear the determination, by Burtt Davy, of ‘Wedelia holubii’ and are variously marked as ‘co-type’ or ‘type’; the material is undoubtedly Lipotriche albinervia. In the public records that exist at Kew, the file on Burtt Davy’s Flora of the Transvaal [PRO 1/F/5, cat. no. 112492] indicates that the manuscript for part three of this work (parts I and II were published under the title A manual of the flowering plants and ferns of the Transvaal …) was largely complete; it was never published (although it exists in part as galley proof and in part as type script), hence any new names in the family that were to appear in that part remained unpublished by Burtt Davy — along with the names of many dozens of other prospective new taxa. Burtt Davy’s generic concept of Wedelia Jacq. was quite broad!

Lipotriche rhombifolia (O. Hoffm. & Muschl.) D. J. N. Hind, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77141641-1

Melanthera rhombifolia O. Hoffm. & Muschl., [Bull. Soc. Bot. France 57] Mém. Soc. Bot. France 2 (8c): 117 (Hoffmann & Muschler 1910). Type: [mali:] ‘haut-niger. — Tiédiana, sous les Beobabs autour du village, 15 juin 1899, no 999 (échantillon unique) [Aug. Chevalier].’ Holotype: P(00073058); isotype: K(000410456 — two leafy inflorescence branches and one separate leaf).

distribution. Mali, Ivory Coast, Nigeria, Togo.

conservation status. A relatively widespread species, although certainly under-collected: LC (Least Concern).

Lipotriche richardsiae (Wild) D. J. N. Hind, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77141642-1

Melanthera richardsiae [as richardsae] Wild, Kirkia 5 (1): 9 (1965). Type: [zambia:] ‘n. rhodesia. — Abercorn, Chilongowelo, [alt. 4900 ft., 21.4.1952] Richards 1491.’ Holotype: K (Sheet I, 000410446); isotypes: BR(8877147), EA, K (Sheet II 000410445). note: the epithet correction, noted by Wagner & Robinson (2002: 557), is allowed under the Code (Art. 60.11, Note 4) as the species is named in honour of Mrs H. M. Richards.

distribution. Uganda, Tanzania, Rwanda, Zambia.

conservation status. A relatively widespread species, although still clearly under-collected: LC (Least Concern).

Lipotriche robinsonii (Wild) D. J. N. Hind, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77141643-1

Melanthera robinsonii Wild, Kirkia 5 (1): 15 (1965). Type: [zambia:] ‘n. rhodesia. Namwala, 9.i.1957, Robinson 2099’. Holotype: SRGH; isotype: K.

distribution. Congo (Kinshasa), Zambia, ?Tanzania.

conservation status. Other than the type there are very few other collections, and it is most certainly under-collected: DD (Data Deficient).

note. Beentje (in Beentje & Hind 2005: 738) determined Lazarus Thomas 26 from Tanzania as Lipotriche pungens.

Lipotriche scaberrima (Hiern) D. J. N. Hind, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77141644-1

Melanthera scaberrima Hiern, Cat. Afr. Pl. (Hiern) 1 (3): 582 (1898). Type: [angola:] ‘pungo andongo. In the drier thickets between Quilanga and Luxillo; fl. Feb. 1857. [Welwitsch] No. 3556.’ Holotype: BM(000924417); isotype: K(000410436), LISU(219580). note: The manuscript label of the holotype in BM is marked clearly ‘Würschmittia scaberrima Welw.’ in Welwitsch’s hand — a name also appearing (in blue pencil) on the LISU isotype.

distribution. Angola.

conservation status. Apparently endemic to Angola and represented by very few collections and the type — it is most probably under-collected: DD (Data Deficient).

Lipotriche scandens (Schumach.) Orchard, Nuytsia 23: 383 (2013).

Buphthalmum scandens Schumach., Beskr. Guin. Pl.: 392 (Schumacher 1827), non Vell. (Vellozo 1831)[=Tilesia baccata (L.) Pruski]. Type: not cited, mentioning only flowering and fruiting time: ‘Almindelig, blomstrer i den frngtbare [sic!] Aarstid.’ There are 2 Thonning 52 sheets in C, C(10003431, 10003432) from ‘Guinea’ (=Ghana), the latter specimen was determined by Adams in 1955. Orchard (2013: 384) was of the opinion that the two sheets were syntypes.

Lipotriche brownii [as brownei] DC., Prodr. 5: 544 (de Candolle 1836). Type: [?central african republic:] ‘in Africa aequin. ad ripas fluminis Congo legit infeliciss. Chr. Smith. Caet. ign.’ Holotype: G-DC; isotypes: BM, K(000410455), P(00073153).

Melanthera brownii (DC.) Sch. Bip., Flora 27 (2): 673 (Schultz Bipontinus 1844).

Amellus scandens (Schumach.) Kuntze, Revis. Gen. Pl. 1: 306 (1891).

Melanthera scandens (Schumach.) Roberty, Bull. Inst. Franc. Afr. Noire, Ser. A, 16: 68 (Jan. 1954).

Melanthera scandens (Schumach.) Brenan, Mem. New York Bot. Gard. 8: 480 (Feb. 1954).

distribution. Sierra Leone, Liberia, Ivory Coast, Ghana, Cameroon, Bioco, Ghana, Nigeria, Gabon, Congo (Kinshasa), Angola, Ethiopia, Sudan, Kenya, Uganda, Tanzania, Botswana, Zambia, Zimbabwe, Malawi, Mozambique, Madagascar, South Africa, Swaziland, Namibia.

conservation status. A common and widespread species: LC (Least Concern).

Lipotriche scandens (Schumach.) Orchard subsp. madagascariensis (Baker) D. J. N. Hind, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77141645-1

Melanthera madagascariensis Baker, J. Linn. Soc., Bot. 21 (137): 418 (1885). Types: [madagascar:] ‘Baron 2344! 2534! Humblot 410!’. Syntype: Baron 2344, K(000410435), P(00442917). Syntype: Baron 2534, K, P(00442916). Syntype: Humblot 410, K, LD(1217175), TAN(000160). note: Lisowski (1991: 208) and Mesfin Tadesse (2004: 299) both indicated Baron 2344 as the ‘holotype’, although most subsequent authors have suggested only part of Baron 2344 is this entity; in this instance, since there is only one specimen of Baron 2344 (marked ‘in part’ by Baker) at K, this is considered the lectotype.

Melanthera cuanzensis Hiern, Cat. Afr. Pl. (Hiern) 1 (3): 583 (1898). Type: [angola:] ‘pungo andongo. By wooded thickets along the right bank of the river Cuanza; fl. and fr. Dec. 1856. [Welwitsch] No. 3554.’ Holotype: BM(000924434); isotypes: K(000410447), LISU(219581, 219582), P(00073176).

Melanthera cuanzensis Hiern var. altior Hiern, Cat. Afr. Pl. (Hiern) 1 (3): (1898). Type: [angola:] ‘pungo andongo. By streams in the wooded parts of Sobato Cabanga, growing in masses; fl. and fr. Jan. 1857. [Welwitsch] No. 3555.’ Holotype: BM(000924435); isotypes: K(000410448), LISU(219583). Note: The mimeographed specimen label on BM(000924435) is marked ‘Lipotriche cuanzensis Welw. β’, and ‘(1/3)’ appears after the collector’s number; the manuscript label in LISU reads the same.

Amellus madagascariensis (Baker) Kuntze, Revis. Gen. Pl. 1: 306 (1891).

Melanthera swynnertonii S. Moore, J. Linn. Soc. Bot. 40 (275): 115 (1911). Type: [mozambique:] ‘[Gazaland] Hab. Chibabava, Lower Buzi, 400 ft.; fl. Nov.; [Swynnerton] n. 1882.’ Holotype: BM(000924433).

Melanthera scaberrima Hiern var. angustifolia S. Moore, J. Bot. 65, Suppl. 2, Gamopet.: 58 (1927). Types: ‘Angola: in wet places at Malange, [Gossweiler] 1204; among Glumaceae in flooded marshes of R. Cuito, [Gossweiler] 3194. Angola.’ Syntype: Gossweiler 1204, BM(000924431), K, P(00073175). Syntype: Gossweiler 3194, BM(000924432), BR, COI(00005606), K.

Melanthera scandens (Schumach.) Roberty subsp. subsimplicifolia Wild, Kirkia 5 (1): 6 (1965), synon. nov. Type: ‘Cameroons, Likomba, xii.1928, Milbraed 10752’. Holotype: K.

Melanthera scandens (Schumach.) Roberty subsp. madagascariensis (Baker) Wild, Kirkia 5 (1): 7 (1965).

Melanthera angustifolia Gilli, Ann. Naturhist. Mus. Wien 78: 158 (1974), nom. illeg. non A. Rich. (Richard 1850), synon. nov. Type: [tanzania:] ‘[Gilli] 626. Lichter Wald von Brachystegia manga De Wild. nordöstlich von Lupingu, 870 m, 2.VIII.1958, soz. Aufn. XI, fr.’. Holotype: W(19730001155).

Liptoriche scandens (Schumach.) Orchard subsp. subsimplicifolia (Wild) Orchard, Nuytsia 23: 384 (2013), synon. nov.

distribution. Sierra Leone, Liberia, Ivory Coast, Ghana, Cameroon, Bioco, Nigeria, Congo (Kinshasa), Angola, Ethiopia, Sudan, Kenya, Uganda, Tanzania, Botswana, Zambia, Zimbabwe, Malawi, Mozambique, Madagascar, Namibia.

conservation status. A widespread and common subspecies: LC (Least Concern).

note. Although Beentje (in Beentje & Hind 2005) suggested that subsp. subsimplicifolia ‘seems to be more like a form’ I cannot find a good division between the two taxa, certainly not that suggested by Wild (1965). I have merged the two into a slightly broader concept of subsp. madagascariensis. Melanthera scaberrima var. angustifolia has also been included in synonymy, although the leaves are conspicuous simply because of the absence of basal lobes — part of the variation of this subspecies.

Lipotriche scandens (Schumach.) Orchard subsp. scandens

distribution. Sierra Leone, Ghana, Nigeria, Gabon, Congo (Kinshasa), Angola.

conservation status. A relatively widespread subspecies: LC (Least Concern).

Lipotriche scandens (Schumach.) Orchard subsp. dregei (DC.) Orchard, Nuytsia 23: 384 (2013).

Psathurochaeta dregei DC., Prodr. 5: 609 (de Candolle 1836). Type: [south africa:] ‘in Africa australi ad Omsamcoubo, ferè ad maris altitudinem, detexit cl. Drege. … α latifolia … (v. s.) … β reticulata … (v. s.)’. note: de Candolle included within his new species two ‘varieties’ making no distinction as to which was typical. There appear to be two collections in G-DC, both numbered Drege 5121, each one representing one variety, ‘α latifolia — 5121 Omsamcoubo’, and ‘β reticulata – 5121, [s.loc.]’, both dated 1835. There are scattered Drege duplicates named as ‘β reticulata’ which may be syntypes, e.g. K(000410426, 000410428), P(00073165) and S07-16911. P(00073161) is labelled as both ‘α latifolia’ and ‘β reticulata’ on a slip label, however the main label reads as ‘α latifolia’ and is numbered ‘5121’. P(00073162) is somewhat differently labelled — ‘14/12 32. Zw. Gebüsch bei Omsamcoubo, 900–1000'/(V. b).’ — but is unnumbered. P(00073163) is labelled as α latifolia. P(00073164) is labelled as ‘β reticulata’ but bears both a date and locality: ‘25/2 32. Aüf [---] Grasfeldern am Gestade zw. Omtendo & Omsamcoubo 200' (V. c).’ — cf Meyer (1843: 154). Orchard (2013: 384 – 385) has cited a ‘holotype’ for the species, and for each of the varieties, which I disagree with on the basis of what is apparently in G-DC.

Trigonotheca natalensis Sch. Bip., Flora 27 (2): 672 (Schultz Bipontinus 1844), nom. nud. pro. syn. (based on Lipotriche brownii DC.)

Melanthera scandens (Schumach.) Roberty subsp. dregei (DC.) Wild, Kirkia 5 (1): 8 (1965).

distribution. Mozambique, South Africa, Swaziland.

conservation status. A common subspecies throughout its range: LC (Least Concern).

Lipotriche tithonioides (Aké Assi) D. J. N. Hind, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77141646-1

Melanthera tithonioides Aké Assi, Adansonia 4 (2): 338 (1964). Type: [ivory coast:] ‘Mont Tonkoui, 20 février 1957, fl. et fr., Mangenot et Aké Assi IA 4142’. Holotype: P(00073210); isotypes: P(00073211, 00418379).

distribution. Ivory Coast, Guinea.

conservation status. A full assessment has recently been provided by Cheek (2014 — in press): Endangered (EN, B1ab(i, ii, iii, iv, v) + 2ab (i, ii, iii, iv, v)) ‘since only five threat based locations are known, at two of which it is known to be threatened.’ The area of occupancy is estimated as 60 km2 and extent of occurrence is assessed as 1300 km2. The potential threat from mining is highlighted at the Pic de Fon locality in Guinea.

note. Adams (1963 — based on Roberty 15827, cited by Aké Assi as a paratype) and Wild (1965) considered this entity, sensu stricto (i.e. excluding the Easter Cape material — Pegler 420 and Rudatis 2791), to be no more than a form of Melanthera rhombifolia; I disagree with this and consider it a very distinctive species, but I do exclude the South African material from the concept.

Excluded taxa/names

Melanthera biflora (L.) Wild, Kirkia 5 (1): 4 (1965) = Wollastonia biflora (L.) DC.

Melanthera cinerea Schweinf. in Hoehnel, Zum Rudolf-See und Stephanie-See: 862 (Dec. 1892); Abh. Königl. Preuss. Akad. Wiss. 1891: 434 (1892), nom. nud. Type: [tanzania:] ‘massaihochland. Ndoro um 1930 m (v. Höhnel 71 (112)).’ Holotype: ?B†. = incertae sedis.

Melanthera elegans C. D. Adams, J. W. Afr. Sci. Assoc. 8 (2): 135 (1964). Type: ‘guinea. Hollande Tossékré, [1200 m d’alt., 15] October 1956, Adam 12728’. Holotype: K (000453489 — which has simply the locality, date and number, and no altitude); isotypes: IFAN, P(00072535). note: The holotype in K was laid away under Aspilia; there is an isotype in P(00072535) — determined as ‘Aspilia elegans (C. D. Adams) J.-P. Lebrun & Stork’! = Aspilia sp. [Note: In Lebrun & Stork (1997: 250) the apparent combination ‘Aspilia elegans (C. D. Adams) J.-P. Lebrun & Stork’, appears, with mention only of ‘F.T.W.A., éd. 2’ and citing the basionym as ‘Melanthera elegans C. D. Adams’, but no direct citation of the place of publication. Lebrun & Stork (1997) indicated new combinations with the abbreviation ‘comb. nov.’, so the entry under Aspilia elegans cannot be considered a new combination; unfortunately this work, and every other compilation list, does not provide the place of publication of this name. The binomial Aspilia elegans (Gardner) B. D. Jacks. represents a Brazilian plant.]

Extended synonymy of Tilesia baccata

In the treatment of Lipotriche scandens I have equated the later homonym of Buphthalmum scandens Schumach., B. scandens Vell. (Vellozo 1831: pl. 132), with Tilesia baccata (L.) Pruski. This is a new synonym, to which I am also adding B. equinum Vell. (Vellozo 1831: pl. 131). Sampaio & Peckolt (1943) provided no accepted name/s for Vellozo’s names, but a close examination of the descriptions in the full text of Flora Fluminensis provided by Netto et al. (1881: 340, 341), and a comparison of the two plates in the Florae Fluminensis Icones, suggest that they are none other than the widespread Tilesia baccata.