Martiodendron elatum (Ducke) Gleason var. occidentale (Ducke) R.C. Koeppen . Brittonia 14(2): 207 (1962).

≡ Martiusia elata Ducke var. occidentalis Ducke . Arquivos do Instituto de Biologia Vegetal 2: 40 (1935). ≡ Martiusia elata Ducke forma occidentale Ducke. Arquivos do instituto de Biologia Vegetal 2: 40 (1935).

Type:— Brazil: Amazonas: rio Purus, super ostium fluminis Acre, silva riparia rarius innnundabilis, arbor magna radicibus tabularibus magnis, flos aureis, 11-III-1933. Ducke, A., s.n. RB24187. Barcode 00146817 (Lectotype here designated: RB!; Isolectotypes: S!; K!; U!; US!; WIS!; Y) .

= Martiodendron macrocarpon Gleason. Phytologia 1:142 (1935). Type:— Brazil: Amazonas: near mouth of Rio Embira (tributary of Rio Tarauaca), 7°30’S 70°15’W. Tree, 110 feet high, with yellow flowers, very hard wood, on varzea land. 26-VI-1933, Krukoff, B.A. 5015 NY00004509 (Lectotype here designated: NY!; Isolectotypes: A!; BRI!; CAS!; DS; F!; G; LE!; M!; MAD; MICH!; MO!; RB!; S!; U!; UC; US!) ≡ Martiusia elata Ducke forma angustifolia Ducke. Leguminosas da Amazônia Brasileira. 87 (1939).

Diagnosis: — Martiodendron elatum var. occidentale differs from Martiodendron elatum var. elatum for having (9–)11(–13) leaflets generally smaller and narrower, the terminal ones (6–)7–11(–13) × (1.4–)2–3.2(–4.8) cm, (2.5–)3–4.6 times longer than wide; a stable number of stamens, usually 5, rarely 4, almost always 0–1 staminode; fruits narrower, 4.6–6(–7) cm wide, (2.4–)2.7–3.8 times longer than wide, with both relatively straight (Figures 3 and 5).

Distribution and Habitat: —The variety is found in the far north of Bolivia and the southern Brazilian Amazon, in the states of Acre, Rondônia, Mato Grosso, and southern and western Amazonas (Figure 15), as well as possibly in the far east of Peru close to the border with Brazil, inhabiting the basins of the Purus, Acre, Madeira, and Juruá Rivers, generally in upland forests, more rarely in floodplains and igapós.

Etymology: —The variety name comes from the Latin occidentalis, referring to the western distribution (Bolivia, Acre, Rondônia, Mato Grosso, and southern Amazonas) compared to the easternmost distribution of the type variety known at the time (southwest of Pará and southeast of Amazonas).

Nomenclatural Comments: —In the protologue of the variety, Ducke, (1935) cited an holotype: Ducke, A., s.n. H.J.B.R. 24187 and Koeppen & Iltis (1962) cited RB 24187. However, like in M. elatum, this herbarium number is not a unique identifier of a single specimen, but an identifier of the gathering. Thus, we deemed it necessary to make a lectotypification to establish the duplicate in RB as the lectotype. M. macrocarpon has one single material cited by Gleason (1935), but the author did not indicate in which herbarium the material was deposited. Koeppen & Iltis (1962) also didn’t specially cite any herbarium, so a lectotypification was also necessary.

Taxonomic Comments: —The continuous distribution and the small level of overlap for the diagnostic characters between this taxon and the typical form of the species led us to consider it as a variety, corroborating the ideas of Koeppen & Iltis (1962). Practically none of the characters proposed by Ducke (1922) in the description of M. elatum is stable and/or distinctive for this taxon, such as pyramidal inflorescences, smaller flowers, and sericeous fruits. The only characteristic of Ducke maintained here is that of larger fruits, so the definition of M. elatum closest to the one proposed here is that of Koeppen & Iltis (1962). In the description of M. macrocarpon (Gleason, 1935), which is maintained here as a synonym for M. elatum var. occidentale, as proposed by Koeppen & Iltis (1962), Gleason points out some differences that were also not corroborated in the present work, such as the impression of the veins on the abaxial face of the leaflets, more pubescent carpels or shape of the veins in fruits. We found that none of these characters can differentiate this taxon.

The difference in axillary bud size between the two varieties of M. elatum described by Koeppen & Iltis (1962) seem to be highly variable and, therefore, not enough to distinguish any taxon, even if infraspecific. Koeppen & Iltis points out that the presence of two carpels per flower is reasonably common for variety occidentale . On the other hand, Silva et al. (2005) indicates that the variety occidentale would always have bicarpellary gynoecium, which was not observed here. No flower with two carpels was observed in this work, despite the large number of specimens and flowers analyzed, including all specimens cited in the two previous works. The variation in leaflet number and size and fruit width and shape seems consistent with Koeppen’s division, with more leaflets and longer fruits in variety occidentale and fewer leaflets and wider and generally more curved fruits in variety elatum .

Selected List of Additional Specimens Examined: (20 of 43 analyzed specimens): — BOLIVIA: Pando: Federico Román, 09°41’S 65°24’W. 24-VII-2002, Urrelo F.633 (F; LPB); BRAZIL: Acre: Acrelândia: basin of rio Madeira, rio Abun ã, Porto Dias, 9°59’18”S 66° 46’40”W. Tree, 25 m × 48 cm, 17-V-2009, Daly, D. C. 13736 (RB; NY); Bujari: riozinho do Andirá, ramal do Espinhara, 9°42’31”S 68° 8’ 16”W. Árvore, 7 m, 7-IX-2013, Costa, D. S. 297 (RB); Rio Branco: Br-364, ramal Bujari, km8. Árvore, 20 m, 23-VI-1989, Saraiva, R. S. 22 (INPA); Sena Madureira: riozinho do Andirá, 9°43’45”S 68°8’53”W. Árvore, 35 m, 15-VI-1995, Oliveira, A. R.S. 631 (NY); Xapuri: Reserva Extrativista Chico Mendes, 1°33’52”S 68° 18’11”W, terra firme. Árvore, 30 m × 60 cm, 3-VI-1999, Ehringhaus, C. 1026 (NY); Amazonas: Coari: Base de operaçıes geólogo Pedro de Moura, 4°57’57”S 65°5’56”W. Árvore, 20 m × DAP 35 cm, 14-VI-2009, Melo, M. G. G. 55 (INPA); Envira: near mouth of rio Embira, tributary of rio Tarauaca, 7°30’S 70°15’W. Tree, 110 feet high, 21-VI-1923, Krukoff, B. A. 4950 (F; K; NY; US); Humaitá: rio Madeira, 29-VI-1936, Ducke, A. s.n. RB35074 (RB); Tefé: rio Urucu, base da Petrobras, 4°0’S 65°0’W. Árvore, 30-III-2004, Amaral, I. L. 2814 (INPA); Mato Grosso: Aripuan ã, gleba Aripuan ã a 25 km da cidade em direç ã o sudoeste, 10°16’35”S 59° 36’ 30”W. Árvore, 6 m, 5-VII-1997, Árbocz, G. F. 4175 (ESA; RB); Rondônia: Itapu ã do Oeste, Flona do Jamari, Concess ã o AMATA. Árvore, 25 m, 23-III-2010, Amoêdo, S. L. 16 (INPA; RB; RON); Jaci Paraná: margem esquerda do rio Madeira, 9°33’31”S 64°53’27”W, 24-III-2010, Pereira-Silva, G. 15173 (CEN; INPA; RB); Ji Paraná: Gleba G, km20. Árvore, 30 m, 9-IV-1983, Paulino Filho, H. F. 98 (INPA); Machadinho do Oeste: Tabajara, rio Machado, 8°57’80”S 61° 56’14”W. Árvore, 15 m, 1-VI-2015, Bigio, N. C. 1684 (RB); Porto Velho: margem esquerda do rio Madeira, 9°23’20”S 64°50’49”W. Árvore, 23-VI-2010, Pereira Silva, G. 15470 (CEN; INPA; RB); Guajará Mirim. 28-IX-1996, Araujo, J. 2609-076-1996 (RON); Abun ã, próximo a antiga estrada de ferro Abun ã -Mutumparaná a 2 km de Abun ã, 9°40’S 65° 20’W. Árvore, 28 m × 65 cm, 19-IV-1987, Cid Ferreira, C. A. 8872 (INPA; NY); Presidente Médici: estrada para Alvorada, km 35. Árvore, 25 m × 60 cm, 23-IV-1983, Paulino-Filho, H. F. 158 (INPA); rio Jarú, Colocaç ã o Primavera, seringal 71. Árvore, 45-50 m, 1.71 m, 14-VI-1972, Oliveira, E. 5992 (IAN).

List With Summary Data of Additional Specimens Examined: (23 of 43 analyzed specimens): — BRAZIL: Acre: Pardo, M. 85 (NY); Lima, L. 632 (NY); Daly, D.C. 8131 (INPA; NY); 9438 (NY); Santos, J.U. 164 (MG); Figueiredo, C. 762 (NY); Krukoff, B. 5401 (F; NY; RB); Amazonas: Clóvis s.n. INPA217423 (INPA); Souza, L.A.G. 30_10 (INPA); Krukoff, B.A. 6619 (RB); Rondônia: Silva, M.G. 6127 (INPA; MG); Paulino Filho, H.F. 51 (INPA); Bigio, N.C. 1815 (RB); Silva, M.G. 6080 (MG); Pereira Silva, G. 16329 (CEN; IAN; INPA; NY; RB); Ulisses 12980 (INPA); Santos, A.A. 3613 (CEN; IAN; INPA; NY); Araújo, J. 0809-082-1997 (RON); Maciel, U.N. 1600 (K; MG); Vinha, E.; 3101-156-1997 (RON); Silva, M.G. 6201 (INPA; MG; UEC); Antônio, M. 2407-136-2997 (RON); Araújo, J. 2807-135-1997 (RON)