Resolving nomenclatural ambiguity in South American Tephrosia (Leguminosae, Papilionoideae, Millettieae), including the description of a new species

Additional keywords: Fabaceae, lectotypification, synonymy, systematics, taxonomy.

Tephrosia Pers. (Leguminosae–Papilionoideae) comprises ~350 pantropically distributed species, occurring mainly in seasonally dry tropical woodlands, bushlands, thickets and grasslands, often in open and disturbed sandy or rocky areas (Schrire 2005). It comprises two subgenera, namely, Tephrosia subgenus Barbistyla Brummitt and Tephrosia subgenus Tephrosia. Tephrosia subg. Barbistyla has a pubescent style and a glabrous stigma, whereas T. subg. Tephrosia has a pubescent stigma (Brummitt 1980). According to Schrire (2005), the highest diversity for this genus is in Africa and Madagascar (~170 species), Australia (~90 spp.), Central and Tropical North America (~45 spp.) and Asia (~40 spp.). In South America, 18 taxa are recognised (Queiroz 2012).

Most treatments of the species of Tephrosia in South America are in regional floras, e.g. Argentina (Burkart 1952), Brazil (Bentham 1862; Queiroz and Tozzi 2009, 2011; Brazilian Flora Group 2015), Paraguay (Hassler 1919), Peru (Macbride 1943), Suriname (Amshoff 1939), and Venezuela (Pittier 1944). Queiroz (2012) produced the only complete account of this genus in South America, where Tephrosia subgenus Barbistyla is represented by four species (Tephrosia candida DC., T. nitens Benth., T. sinapou (Buc’hoz) A.Chev., and T. vogelii Hook.f.), and Tephrosia subg. Tephrosia comprises 14 taxa (T. adunca Benth., T. cinerea (L.) Pers., T. domingensis (Willd.) Pers., T. egregia Sandwith, T. fertilis R.T.Queiroz & A.M.G.Azevedo, T. guaranitica Chodat & Hassl., T. hassleri Chodat, T. macbrideana R.T.Queiroz, G.P.Lewis & A.M.G.Azevedo, T. marginata Hassl., T. noctiflora Bojer ex Baker, T. purpurea (L.) Pers. subsp. purpurea, T. senna Kunth, T. sessiliflora (Poir.) Hassl., and the new species described here, T. chaquenha R.T.Queiroz & A.M.G.Azevedo).

The species of Tephrosia in South America are shrubs or subshrubs with imparipinnate, multifoliolate, trifoliolate or unifoliolate leaves. Inflorescences are terminal, axillary or leaf-opposed pseudoracemes; the calyx is campanulate; the corolla is red, pinkish, purple, yellow, or white; and the ovary has 2–13 ovules. The fruit is a typical legume (Queiroz 2012). Bentham (1862) and Burkart (1952) noted that Tephrosia can be easily recognised by its linear, elliptical, oblong or obovate leaflets with numerous, parallel, oblique secondary veins.

Several species of Tephrosia are reported to have economic importance. According to Forbes (1948), ~22 species have been recorded as fish-poison plants. Wood (1949) noted that many species of this genus produce rotenone, a substance used to produce insecticides that are poisonous to invertebrates but not to most vertebrates, and Dzenda et al. (2007) reviewed the ethnomedical and veterinary uses of T. vogelii, and mentioned eight other species of Tephrosia with therapeutic properties. Compounds isolated from this genus include deguelin, rotenone, tephrosin, quercetin and rutin, which are reported as antimicrobial and potentially able to kill molluscs, fish, insects and helminths (Dzenda et al. 2007).

During his revision of Tephrosia in South America, the first author analysed more than 3000 accessions in North American, Brazilian and European herbaria. Materials from the following herbaria were studied: B, BAB, BHCB, BM, BR, C, CEN, CGMS, COL, E, EAC, EAN, ESA, F, FHI, G, GH, GOET, HAL, HAS, HB, HEPH, HNBU, HRB, HRCB, HRJ, HST, HUEFS, HUFMS, HUFU, IAC, IBGE, ICN, INPA, IPA, JPB, K, L, LP, M, MA, MBM, MG, MICH, MO, MOSS, NX, OUPR, OXF, NY, P, PH, PAMG, PEUFR, RB, S, SI, SING, SP, SPF, SPFR, TCD, TEPB, UB, UC, UEC, UFG, UFMT, UFRJ, UFRN, UFP, UPCB, U, US, W and WU (codes follow Index Herbariorum, New York Botanical Garden’s Virtual Herbarium, see http://sweetgum.nybg.org/science/ih, accessed 22 February 2017).

This work has brought several nomenclatural issues to light. As a result, we are here describing one new species, proposing nine lectotypes, one neotype, and five new synonyms. In addition, an identification key to the taxa occurring in South America is presented.

Type: Tephrosia villosa (L.) Pers.

Cracca L., Sp. pl. 2: 752 (1753), nom. rej., non Hill (1756), nec Medik. (1787), nec Benth. (1853).

Type: Cracca virginiana L.

Colinil Adans., Fam. Pl. 2: 327 (1763).

Type: none.

Erebinthus Mitch., Diss. Brev. Bot. Zool. 32 (1769), nom. rej.

Type: Tephrosia spicata (Walter) Torr. & A. Gray (fide C.Wood, Rhodora 51: 292 (1949)).

Needhamia Scop., Intr. Hist. Nat. 310 (1777), nom. rej.

Type: Vicia littoralis Jacq.

Reineria Moench, Suppl. Meth. 44 (1802), nom. rej.

Type: Reineria reflexa Moench.

Kiesera Reinw. ex Blume, Catalogus, 93 (1823).

Type: Kiesera sericea Reinw.

Xiphocarpus C.Presl, Symb. Bot. 1: 13 (1830).

Type: Xiphocarpus martinicensis C.Presl.

Apodynomene E.Mey., Comm. Pl. Afr. Austr. 111 (1836).

Type: Apodynomene grandiflora (L’Hér. ex Aiton) E.Mey.

Macronyx Dalzell, Hooker’s J. Bot. Kew Gard. Misc. 2: 35 (1850).

Type: Macronyx strigosus Dalzell.

Type: Brazil. Minas Gerais: in desertorum pascuis ad S. Isabel, s. dat., Pohl s.n. (lecto, here designated: K 502599; isolecto: NY 623410). Cracca adunca (Benth.) Kuntze, Revis. Gen. Pl. 1: 174 (1891).

Tephrosia penicillata Benth., Ann. Nat. Hist. 3: 431 (1839).

Type: Guyana [British Guiana], s. dat., R.Schomburgk 678 (holo: K 502605; iso: BR 5792504, F 59923F, F 360659F, G 367963, G 367977, G 367990, NY 33832, TCD 4328, TCD 4329).

Tephrosia rufescens Benth., Linnaea 22: 513 (1849); Cracca villosa var. rufescens (Benth.) Kuntze, Revis. Gen. Pl. 1: 174 (1891); Tephrosia adunca var. rufescens (Benth.) Hassl., Repert. Spec. Nov. Regni Veg. 16: 165 (1919), syn. nov.

Type: Brazil. Minas Gerais, 1848, A.F.Regnell 78 (lecto, here designated: K 502608; isolecto: R 64353, R 64353a, M (not seen), MO 1737286 (image seen). Remaining syn: Brazil, Nov. 1833, Riedel 1565 (P 2949532). Probable syn: Brazil., Minas Gerais. Caxoeira do Campos, 1839, P.Claussen & B.Delessert 4 (K 502607).

Tephrosia adunca var. intermedia Chodat & Hassl., Bull. Herb. Boissier sér. 2, 4: 839 (1904).

Type: Paraguay, Canindeyú, Yerbales. In regione Yerbalium de Maracayú Paraguaria euro-astra, Serra de Maracayú, E.Hassler 4643 (lecto, here designated: G 448507; isolecto: BM 538080, G 448505, G 448506, G 448506a, G 448507-a, G 448507-b, G 448507-c, G 448507-days, G 448508, G 448508-a, K 502603, UC 934893).

Tephrosia adunca f. glabrior Chodat & Hassl., Bull. Herb. Boissier, sér. 2, 4: 839 (1904).

Type: Paraguay. In dumeto pr Bellavista in regione cursus superioris fluminis Apa, Oct. 1901–1902, E.Hassler 7843 (holo: G 448504).

Tephrosia rufescens var. paraguayensis Ulbr., Repert. Spec. Nov. Regni Veg. 2: 12 (1906); Tephrosia adunca f. paraguayensis (Ulbr.) Hassl., Repert. Spec. Nov. Regni Veg. 16: 165 (1919).

Type: Paraguay, 20–25 Feb. 1903, K.Fiebrig 903 (lecto, here designated: G 400050; isolecto: K 502609, M 233384).

Tephrosia adunca var. subglabrata Hassl., Repert. Spec. Nov. Regni Veg. 16: 165 (1919).

Type: Paraguay. Zwischen Rio Apa und Rio Aquidaban, Nov. 1908–1909, K.Fiebrig 4229 (lecto, here designated: G 448503; isolecto: G 448503-a, G 440502, G 440502-a, G 440502-b, G 440502-c. Remaining syn: Paraguay. Cordillera de Altos, 1903, K. Fiebrig 938 G 448500, G 448500-a, G 448501, G 448501-a).

Tephrosia adunca f. pseudomarginata Hassl., Repert. Spec. Nov. Regni Veg. 16: 165 (1919).

Type: Paraguay. Prope Caaguazú in campis combustis, 1905, E.Hassler 8908 (holo: G 448441).

Tephrosia adunca has a wide geographical distribution (Argentina, Boliva, Brazil, the Guianas, Venezuela and Uruguay) and presents a large morphological variation. It is morphologically similar to T. fertilis, but the latter can be recognised by its thick, curved branches (v. straight and thinner branches in T. adunca). The indumentum of T. adunca is always rust-coloured. The leaf has a curved rachis when the leaflets are closed, and the leaflets are variable in form and covered by a sparse to medium sericeous indumentum. A diagnostic character of the species is that the pedicel is approximately twice the length of the bract.

The protologue of T. adunca cites a syntype collection: ‘Dr Pohl in the desert pastures of S. Isabel’, and a second gathering from Minas Gerais. We have located only two specimens collected by Dr Pohl from S. Isabel (K 502599 and NY 623410). As the specimen at K includes an original Pohl label, we are designating it as the lectotype of Tephrosia adunca Benth.

In the protologue of Tephrosia rufescens, Bentham (1862, p. 513) cited three syntypes: ‘Ad Caldas prov. Miňas Geräes A. Regnell. ser. 2. n.78; Caxoeiras do Campos, Claussen; ad San Paulo, Riedel’. According to Stafleu and Cowan (1976), George Bentham’s herbarium and types are housed at K, where we found two collections: Brazil, Minas Gerais, Caxoeira dos Campos, 1839, P.Claussen & B.Delessert 4 (K 502607) and Brazil, Minas Gerais, ad Caldas, 1848, A.F.Regnell 78 (K 502608). Both specimens are on the same sheet. An additional specimen, A.F.Regnell II-78 (K 858670, M 0233383), was also located; however, the label data indicate that this specimen was collected in Uberaba municipality, instead of Caldas, as cited in the protologue. Duplicates of Regnell 78 collected at Minas Gerais, Caldas, were located at R (R 64353 and R 64353a). A probable syntype, L.Riedel 1565, is housed at P (P 2949532). Because Bentham cited the number 78 for the Regnell collection in the protologue, and did not cite either the co-collector B. Delessert or the number 4 for the Claussen collection (therefore, we are not sure whether the Claussen specimen is indeed even a syntype), we are here designating K 502608 as the lectotype of Tephrosia rufescens Benth.

In the protologue of Tephrosia adunca var. intermedia, Chodat and Hassler (1904, p. 834) cited ‘in campo pr. flumen Jejui guazu, Sept., no. 4643’. Ten duplicates of Hassler 4643 were found at G. The following five of these specimens are annotated as the holotype: G 448507, G 448507-a, G 448507-b, G 448507-c, G 448507-d; and the other five as isotypes: G 448505, G 448506, G 448506-b, G 448508-a, G 448508-b. All the specimens fit the protologue and are eligible for selection as the lectotype. Five of these sheets carry a specimen label (G 448505, G 448506, G 448507, G 448507-d and G 448508) and are equally informative. We opted to follow the G labels, which annotate G 448507 as the holotype, and we designate this as the lectotype of the name T. adunca var. intermedia.

In the protologue of Tephrosia adunca var. subglabrata, Hassler (1919, p. 165) cited the following three syntypes: ‘Paraguay: Hassler 7843; Fiebrig 938, 4229’. According to Stafleu and Cowan (1979), the largest set of Émile Hassler’s Paraguayan specimens is housed at G, the second largest set at K, and other major sets at B (no longer extant), P, S and W. Material of all cited syntypes is held at G. To avoid confusion, material of Hassler 7843 was excluded from consideration for lectotype selection, as it is type material of the name Tephrosia adunca f. glabrior Chodat & Hassl. Four duplicates of Fiebrig 938 and six of Fiebrig 4229 are held at G. Fiebrig 4229 has flowers and fruits, whereas Fiebrig 938 has only flowers. Among the Fiebrig 4229 duplicates, G 448502, G 448502-a, and G 448503 have flowers and fruits and we, here, designate K.Fiebrig 4229 (G 448503) as the lectotype of Tephrosia adunca var. subglabrata Hassl. In the protologue of T. rufescens var. paraguayensis, Ulbrich (1906¸ p. 165) cited ‘K. Fiebrig, Plantae Paraguayenses no. 903’. Ulbrich’s herbarium is housed at B (partly extant; Stafleu and Cowan 1976). We did not locate any specimens that could be considered original material of Tephrosia adunca var. paraguayensis Hassl. at B, but three specimens were located at G, M, and K. The specimen located at M (M 233384) is sterile. Both K 502609 and G 400050 have flowers and fruits, in accordance with the protologue. Because G 400050 is the most informative specimen (it has mature fruits, whereas the other specimen has immature fruits), we are designating this as the lectotype of Tephrosia rufescens var. paraguayensis Ulbr.

Type: Paraguay, Ciges Couchées, Villa Rica, sur les collines inculte, 6 Oct. 1874, B.Balansa 1539 (holo: BM 538072; iso: BR 5020184238698) (Fig. 1, 2).

Tephrosia cinerea f. pseudoadunca Hassl., Repert. Spec. Nov. Regni Veg. 16: 166 (1919).

Type: Paraguay, Gran Chaco, Loma Clavel, E.Hassler 2591 (holo: G 229579; iso: G 229578, G 229580, G 400028, MPU 024206, P 02950680 US 00003906).

Haec species Tephrosiae cinereae (L.) Pers. similis, sed ab ea praecipue foliolis adaxialitre glabris, floribus longioribus in partibus omnibus majoribus atque petalis organis secretoriis praeditis distinguitur.

Decumbent subshrub up to 80 cm tall, indumentum sparsely grey sericeous to glabrous. Leaves imparipinnate, 5–19-foliolate; stipules 2–7 × ~0.5 mm, with 1–3 veins, narrowly triangular to linear; petiole 3–20 mm long, sulcate; rachis 5–50 mm long, sulcate; petiole, stipule, and rachis scattered sericeous; leaflets 10–45 × 3–7 mm, elliptic or narrowly elliptic to narrowly oblanceolate or linear; base acute to cuneate, apex rounded to acute; upper surface glabrous, lower surface scattered sericeous; secondary veins 7–9 pairs; all petiolules 1–2 mm long, sericeous. Inflorescence a terminal or axillary pseudoraceme 2.5–13 cm long; peduncle 3–5 cm long, cylindrical, scattered sericeous to glabrous; bracts 2–5 × 0.4–1 mm, narrowly triangular; pedicels 3–5 mm long. Flowers pink, 15–15.5 mm long; calyx 5–6 mm long, campanulate, bilabiate, scattered sericeous, teeth 2.8–4 × ~1 mm, narrowly triangular, subulate; standard petal 12–13 × 12–15 mm, broad ovate, retuse at apex, dorsal surface scattered sericeous to tomentose, claw 2–2.5 × 1–1.5 mm; wings 12–14 × 6–7 mm, asymmetrically obovate, apex rounded, claw 2–2.5 × 0.3–0.5 mm; keel petals 9–10 × 2–5 mm, connate, cucullate, apex acute, secretory structures present, claw 2.5–2.8 × 0.2–0.3 mm; staminal tube 8–10 × 2.5–3 mm, free portion of filaments 3–4 mm long; annular disk present at base of sessile ovary; ovary 7–9 mm long, oblong, indumentum sericeous, ovules 8–10; style ~6 mm long, laterally compressed, glabrous; stigma penicillate with translucent trichomes. Fruit a legume 43–60 × 3–4 mm, coriaceous, plane, linear, apex and base asymmetric, brown with scattered sericeous, greyish indumentum, and persistent calyx; seeds 7–10 × 2.8–3.8 × 1.8–2.2 mm, oblong, ochraceous, testa speckled, hilum excentric, aril absent.

Flowers and fruits are reported from November to January.

Tephrosia chaquenha is currently recorded from Argentina, Brazil and Paraguay. The extent of occurrence (EOO) for the species is 235 906.527 km² (which equates to a conservation status of Least Concern, LC). Its area of occupancy (AOO) is 24.000 km² (which would give it an Endangered status, EN). We have little information about the quality of the habitats where the species grows in each country, and because of the current paucity of specimens and habitat data, we assess Tephrosia chaquenha to be Data Deficient (DD) following IUCN 2017 criteria (IUCN 2017).

The species is named in reference to its distribution in chaco vegetation (in Brazil and Paraguay). The habitat of this species in Argentina is presently unknown.

Tephrosia chaquenha morphologically resembles T. cinerea and shares the same habit, grey-coloured indumentum and similar linear fruits. The species can be distinguished by the characters presented in Table 1. Moreover, another noteworthy character is that Tephrosia cinerea presents a reticulate-smooth testa, whereas it is reticulate-papillate in T. chaquenha (Fig. 2).

ARGENTINA. Corrientes: Lavalle, 2 km N de Cerrito sobre ruta 152, 23 Nov. 1979, A.Schinini 19043, R.Vanni & G.Norrmann (fl., fr.) (IAC); entre Rios: Colón, 13 Nov. 1979, N.S.Troncoso 2675 (fl.) (SI); Colón, Palmar, a 13 km E de R 14, barrancas del rio Uruguay hasta el arenal, lat. 32°0′S, long 58°10′W, 17 Jan. 1976, C.Romanczuk 23 (fl., fr.) (BAB). BRAZIL. Mato Grosso do Sul: Porto Murtinho, Fazenda das flores 21°43′01″S, 57°53′57″W, 15 Dec. 2009, R.T.Queiroz 1439 (fl.) (UEC). PARAGUAY. Paraguay Centralis, in regione lacus Ypacaray, Dec. 1913, E.Hassler 12410 (fl.) (BM, G, SI); Zwischen Rio Apa und Rio Aquidaban, 1909, K.Fiebrig 4229 (G, K).

Fig. 1.  Tephrosia chaquenha. A. Branch. B–D. Distinct leaflets. E. Bud. F. Calyx. G. Standard. H. Wings. I. Keel petals, J. Gynoecium, K. Staminal tube. L. Young fruit with persitent calyx. Voucher Schinini et al. 19043 IAC-24649. Drawn by Kley Souza.

Fig. 2.  Scanning electron micrographs of the testa of two species of Tephrosia. A. T. chaquenha (simple-reticulate papillate testa). B. T. cinerea (simple-reticulate smooth testa).

Table 1.  Characters used to distinguish Tephrosia cinerea from T. chaquenha
SEM, Scanning Electron Microscopy

Galega cinerea L., Syst. Nat., 10th edn, 2: 1172 (1759). Cracca villosa var. cinerea (L.) Kuntze, Revis. Gen. Pl. 1: 173 (1891); Cracca cinerea (L.) Morong, Ann. New York Acad. Sci. 7: 79 (1892); Colinil cinereum (L.) C.H.Hitchc., Rep. (Annual) Missouri Bot. Gard. 4: 75 (1893).

Type: [no specimen data] (lecto: LINN HL-924.2, designated by W.Fawcett and A.B.Rendle, Fl. Jamaica 4: 20 (1920)).

Vicia littoralis Jacq., Enum. Syst. Pl. 27 (1760). Galega littoralis (Jacq.) L., Syst. Nat., 12th edn, 2: 497 (1767); Tephrosia littoralis (Jacq.) Pers., Syn. Pl. 2: 328 (1807); Tephrosia cinerea var. littoralis (Jacq.) Benth. in C.F.P. von Martius (ed.), Fl. Bras. 15(1A): 48 (1859); Cracca littoralis (Jacq.) Rydb., in N.L.Britton et al. (eds), N. Amer. Fl. 24(3): 178 (1923).

Type: [no specimen data] (lecto, here designated: LINN HL-924.3).

Tephrosia venustula Kunth., nov. gen. sp., 4th edn 6: 459 (1823).

Type: Guyana. Bordones, s. dat., F.Humboldt & A.Bonpland 420 (holo: P 660150).

Orobus domingensis Spreng., Syst. Veg., 16th edn, 3: 261 (1826).

Type: Republica Dominicana. Santo Domingo, En terreno abandonado, común, cerca del Jardim Botânico, Santo Domingo, 20 Sept 1973, A.H.Liogier 20188 (neo, here designated: NY 1652524).

Tephrosia procumbens Macfady., Fl. Jamaica 1: 256 (1837).

Type citation: ‘Galega No. 2, Browne, 289.’

Type: not seen.

Tephrosia gynothrix Miq., Linnaea 18: 29 (1844).

Type: Suriname. In plantatione de Zwarigheid, Vredenburger Kreek, Oct. 1842, H.C.Focke 666 (holo: U 3650).

Tephrosia scopulorum Brandegee, Univ. Calif. Publ. Bot. 6: 181 (1915).

Type: Mexico. San Geronimo, Oaxaca, July 1914, C.A.Purpus 7151 (holo: UC 175091; iso: F F0059920F, GH (not seen), MO 125154, NY 33816, US 3953).

Tephrosia cinerea var. typica Hassl., Repert. Spec. Nov. Regni Veg. 16: 166 (1919), nom. inval.

This species has a subshrubby habit, with decumbent and thin branches. In general, the branches, leaves and fruits are covered by grey indumentum. All material from Caatinga vegetation has axillary fascicles of flowers with long, slender pedicels. The length of the pedicel is a character that facilitates the recognition of this species. Furthermore, its fruits are usually very long and arched. Owing to the slenderness of the pedicel, the fruits are always orientated downward. The calyx lobes are strongly subulate, especially in fruit. The species is morphologically similar in appearance to T. chaquenha, but is easily distinguished by the above suite of characters

The protologue of Vicia littoralis Jacq. (1760) cited neither specimens nor illustrations. Because Jacquin’s specimens are known to be in numerous herbaria (D’Arcy 1970; Stafleu and Cowan 1979), searches for original material were undertaken at B, BM, CGE, H, JE, LINN, LIV, OXF, UPS and W. We located only one specimen (LINN HL-924.3) annotated ‘Vicia littoralis mea Jacquin’ in Jacquin’s handwriting (Burdet 1979), and, therefore, it is likely that it was seen by Jacquin. Accordingly, we designate this specimen as the lectotype of Vicia litoralis Jacq. An additional four specimens held at LINN (LINN HS-1214.10, LINN HS-1214.11, LINN HS-1214.12, and LINN HS-1214.14) lack locality or date information, and each is annotated as Galega cinerea L. These specimens are unlikely to represent the original material of Vicia littoralis. The other specimen located in one of the herbaria that may hold Jacquin material is W 29662, also lacking location or date, with the name Galega littoralis L. written on it.

Orobus domingensis Spreng. was described on the basis of a Carlo Bertero collection from Hispaniola in Santo Domingo. The protologue provides neither a description nor an illustration. Searches for original material of this name in B, BHU, BM, G, K, M, P and W were unsuccessful. Therefore, the specimen NY 1652524 is here proposed as a neotype for Orobus domingensis.

Galega domingensis Willd., Sp. Pl., 4th edn, 3(2): 1249 (1802); Cracca domingensis (Willd.) Rydb., in N.L.Britton et al. (eds), N. Amer. Fl. 24(3): 181 (1923).

Type: Dominica. … St. Domingo…, s. dat., F.Bredemeyer s.n. (holo: B-W 13942-010).

Tephrosia ascendens Macfady., Fl. Jamaica, 1: 257 (1837).

Type: Jamaica, s. dat., J.Macfadyen s.n. (holo: K 502616).

Tephrosia tenella A.Gray, Smithsonian Contr. Knowl. 5(6):36 (1853). Cracca tenella (A.Gray) Rose, Contr. USA Natl. Herb. 12(7): 271 (1909).

Type: United States. San Pedro: Sonora, 8 Sep. 1851, C.Wright 966, (lecto, here designated: GH 107070; isolecto: GH 63573, GH 63572, K 848919, K 848920, MO 217373, MO 277489, NY 33799, PH 30249, US 3979).

Tephrosia leptostachya var. leptophylla Benth. in C.F.P. von Martius (ed.), Fl. Bras. 15(1A): 49 (1862), syn. nov.

Type: Brazil. Goiás: Alegre, s. dat., Pohl s.n. (holo: K 921500).

Cracca rusbyi Rydb. in N.L.Britton et al. (eds), N. Amer. Fl. 24(3): 181 (1923).

Type: Mexico. Santa Catarina, Oaxaca, 14 July 1910 H.H.Rusby 69 (holo: NY 6596).

The name Tephrosia domingensis dropped into disuse after Bentham (1862) treated it as a synonym of T. leptostachya DC. and published T. leptostachya var. leptophylla Benth., on the basis of material of T. domingensis. After that, all South American collections were treated as T. leptostachya. After reviewing the type specimen of T. domingensis, we consider that the South American material treated as T. leptostachya is better considered as a separate species, T. domingensis.

Tephrosia domingensis is recognised by its decumbent habit, but in some situations it may become ascendant. Its branches are slender, strongly lignified, sometimes wine red in colour, and glabrescent. The leaves have very narrow and glabrescent leaflets with the apex usually acute, but occasionally round. The pseudoracemes are long with few flowers. Their legumes are plane, with the edges of the valves prominent and with a persistent calyx; the calyx lobes are very delicate. The seeds are usually flattened, and become dark and grey-speckled when old.

In the protologue of Tephrosia tenella, Gray (1853) stated ‘…near the San Pedro, Sonora; Sept. (966)’. Stafleu and Cowan (1976) stated that the herbarium, types, original manuscripts, archives and correspondence of Asa Gray are at GH. Three specimens of C.Wright 966 were located at GH (GH 63572, GH 63573, and GH 107070). The protologue describes flowers and fruits, and all the three specimens fit the protologue. The specimen GH 107070 has a label on it (certainly not prepared by Gray), identifying it as the holotype. However, on the same sheet, there are two different collections, namely, GH 107070 (two plants on right half of sheet) and GH 263484 (two plants on left half of sheet, with the collection number 929 on the label). Because GH 107070 is the most informative specimen, with both flowers and well-developed fruit, it is here formally designated as the lectotype of T. tenella A.Gray.

Type: Paraguay. In valle fluminis Y – acá. in campo glareoso, pr. Pirihebuy, Dec. 1900, E.Hassler 6721 (holo: G 400037).

Tephrosia nervosa Chodat & Hassl., Bull. Herb. Boisser, sér. 2, 4: 839 (1904), nom. illeg., non Pers. (1807).Tephrosia adunca var. acutifolia Chodat & Hassl., Bull. Herb. Boissier, sér. 2, 4: 879 (1904), syn. nov.

Type: Paraguay, Iter ad Yerabales montium Sierra de Maracayu, in regione flumis Capibari, 1900, E.Hassler 4381 (lecto, here designated: G 400034; isolecto: G 400033, G 400035, G 448509, G [no barcode accession number].

Tephrosia marginata var. pseudorufescens Hassl., Repert. Spec. Nov. Regni Veg. 16: 163 (1919).

Type: Paraguay. In campo Nandurucay Sierra de Maracayu, 1898–1899, E.Hassler 4346 (lecto, here designated: G 400038; isolecto: G 400036, G 400044, NY 33830). Remaining syn: Paraguay. In regione Yerbalium de Maracayú, E.Hassler 4381, BM 538077, G 400033, G 400034, G 400035, K 502595, NY 33831; Paraguay, Prope Caaguazú in campis combustis, E.Hassler 9126, G 400026).

Tephrosia marginata var. cinerascens Hassl., Repert. Spec. Nov. Regni Veg. 16: 163 (1919), nom. inval.

Chodat and Hassler (1904) described this species as Tephrosia nervosa on the basis of material collected in Paraguay. However, this name was an illegitimate later homonym, and Hassler (1919) subsequently published the replacement name T. marginata and used characters of the indumentum to describe two varieties, namely, Tephrosia marginata var. cinerascens Hassl., nom. inval. and T. marginata var. pseudorufescens Hassl. However, the diagnostic characters used to separate the varieties are tenuous, and these taxa are not accepted here. Tephrosia marginata var. cinerascens is invalid under ICN Art. 26.2 (Turland et al. 2018; known informally as ‘The Shenzhen Code’ or simply ‘the Code’, see http://www.iapt-taxon.org/nomen/main.php) because one of the cited syntypes (Hassler 6721) is the type of T. nervosa (= T. marginata).

This species morphologically resembles Tephrosia chaquenha regarding the shape of the leaflets. However, T. marginata is the only species with linear leaflets with a prominent marginal rib. Besides this, T. marginata has a ferruginous indumentum, long pseudoraceme (8–20 cm long), and wine red-coloured flowers, whereas T. chaquenha presents a grey indumentum, short pseudoraceme (2.5–13 cm long) and pink flowers.

In the protologue of Tephrosia adunca var. acutifolia, Chodat and Hassler (1904) did not cite any collection. Nevertheless, Chodat and Hassler (1907, p. 826) published a correction to their original protologue citing E.Hassler 4381 as the type collection. We are confident, therefore, that this is the original collection analysed by the authors when describing their new taxon. Four specimens under this number were located at G (G 400033, G 400033a, G 400034, G 400035, and G 448509). We are designating G 400034 as the lectotype of T. adunca var. acutifolia Chodat & Hassl., because this is the most informative specimen, having flowers and young fruits (other duplicates have only flowers). Moreover, the original description and the type collection present no morphological characters that differentiate the taxon from T. marginata; therefore, we are synonymising T. adunca var. acutifolia under T. marginata.

In the protologue of Tephrosia marginata var. pseudorufescens, Hassler (1919, p. 163) cited the syntypes ‘Hassler 4346, 4381 id. in campis siccis Caaguazú Hassler 9126’. Several specimens were located at BM, G, K, and NY. Among the specimens at G (where the Hassler types are housed), G 400038 has only fruits, whereas the others have both flowers and fruits. Because the protologue describes fruits but does not mention flowers, we are here designating G 400038 as the lectotype of Tephrosia marginata var. pseudo-rufescens Hassl.

Cracca vogelii (Hook. f.) Kuntze, Revis. Gen. Pl. 1: 175 (1891).

Type: Nigeria. Niger, s. dat., Vogel s.n. (holo: K 262618).

Tephrosia megalantha Micheli, Bull. Soc. Roy. Bot. Belgique 36: 57 (1897), syn. nov.

Type: Democratic Republic of the Congo. Lusambo, s. dat., E.Laurent s.n. (lecto, here designated: BR 8947048).

Tephrosia periculosa Baker, Bull. Misc. Inform. Kew. 1897 128–129: 258 (1897), syn. nov.

Type: Malawi. Kondowe to Karonga, s. dat., A.Whyte 324 (holo: K 262825).

Tephrosia vogelii is a well-circumscribed species. In the vegetative stage, its branches are strongly grooved. Both the branches and the leaves are densely covered by strigose and rufous trichomes. The species can be recognised more easily in the reproductive state by its caducous bracts, a character that is absent in other South American species. The flowers have a tomentose indumentum on the calyx lobes, and the standard has indumentum on its dorsal surface; these characters are present only in this species. The indumentum of the fruit is woolly, which is another exclusive character of this introduced species in South America.

Tephrosia vogelii was introduced to South America probably because of its ornamental potential and ichthyotoxic properties. Most likely, the species originated in Central Africa. It is grown in Africa for the extraction of insecticidal substances. In South America, T. vogelii is infrequently encountered and only a small number of collections exist in herbaria.

In the protologue of Tephrosia megalantha, M. Micheli (Leguminosaceae, cited in Chalon et al. 1897, pp. 53–77) cited ‘Congo (Dewèvre n. 520)’ and ‘Lusambo (Laurent).’ According to Stafleu and Cowan (1981), the herbarium and types of Marc Micheli are housed at G, but no original material was located in this herbarium. One specimen of E.Laurent s.n. was located at BR. Therefore, we are designating BR-8947048 as the lectotype of Tephrosia megalantha Micheli.

The authors declare that they have no conflict of interest.

We thank CAPES and CNPq (processes 563550/2010-4 and 201550/2010-5), and FAPESP (process 2010/52488-9) for financial and logistical support to the first author.