Fridericia triplinervia (Mart. ex DC.) L. G. Lohmann

11. Fridericia triplinervia (Mart. ex DC.) L. G. Lohmann in Ann. Missouri Bot. Gard. 99: 446. 2014 ≡ Bignonia triplinervia Mart. ex DC., Prodr. 9: 153. 1845 ≡ Petastoma triplinervia (Mart. ex DC.) Miers View in CoL in Proc. Roy. Hort. Soc. London 3: 195. 1863 ≡ Arrabidaea triplinervia (Mart. ex DC.) Baill. ex Bureau View in CoL in Vidensk. Meddel. Dansk Foren. Kjøbenhavn 1893: 99. 1894 ≡ Saritaea triplinervia (Mart. ex DC.) Dugand View in CoL in Caldasia 3: 266. 1945. – Lectotype (designated by Arbo 2018: 112): Brazil, São Paulo, Dec 1817, C. F. P. von Martius s.n. (M accession code M-86356! [photo in NY!]; isolectotypes: G-DC barcode G00133252 [image!], K n.v., M accession code M-86357! [between Paranangaba and Mineiros in label of specimen]). – Fig. 9B–E, 10.

= Arrabidaea ateramnantha Bureau & K. Schum. in Fl. Bras. 8(2): 68. 1896. – Lectotype (designated here): Brazil, EspÍrito Santo, Itapemirim [in Brasiliae australiore loco haud accuratius indicato], 12 Feb 1876, A. Glaziou 11228 (P barcode P00587292!; isolectotype: P barcode P00587291!).

Note — In the protologue of Arrabidaea ateramnantha, Bureau & Schumann (1896) did not indicate where the type was deposited. We found two gatherings of Glaziou 11228 in P. We chose the specimen P00587292 as lectotype because it is the most complete and well-preserved specimen.

= Arrabidaea triplinervia var. brachycalyx Hassl. in Bull. Herb. Boissier 6(App. 1): 26. 1898. – Lectotype (designated by Arbo 2018: 112): Paraguay, Cordillera de Altos, Jan, E. Hassler 31 (P barcode P03586448!; isolectotypes: G barcode G00085611 n.v., K barcode K000449155!, NY barcode NY00313123!, P barcode P03586447!).

– “ Fridericia triplinervia ” (Mart. ex DC.) L. G. Lohmann, Cat. Pl. Fung. Brasil 1: 766. 2010. Designation not validly published (Art. 41.1).

Morphological description — Lianas or scandent shrubs, evergreen, up to 12 m high. Branches terete, striate and lenticellate, brown or grey when dry, glabrous and lepidote; with inconspicuous interpetiolar ridge, with (sometimes inconspicuous) interpetiolar glandular fields; prophylls of axillary buds broadly triangular, not apiculate, 0.6–0.8 mm long. Leaves 2-foliolate; petiole 1–3.1 cm long, lepidote with glandular peltate trichomes to pubescent, with simple trichomes; petiolules 0.7–2 cm long, lepidote with peltate glandular trichomes to pubescent, with simple trichomes; blade chartaceous or coriaceous, smooth, margin entire to revolute, leaflets ovate to elliptic, 5–13 × 2.5–7.5 cm, base cuneate, rounded or subcordate, apex acuminate, adaxially not vernicose, glabrous throughout, abaxially glabrous to lepidote throughout (sometimes pubescent, with simple trichomes at domatia), venation palmate actinodromous basal or suprabasal, secondary veins raised, tertiary veins raised, with evident pocket ( 2–4 mm long or 1–2 mm in C Amazonia) and trichome tuft domatia. Inflorescences terminal and axillary, thyrsoid with abortion of buds, with 1 or 2(or 3) orders, first-order peduncles 1–5.5(–8.6) cm long, second-order 0.4–2.3 cm, third-order c. 0.5 cm, glabrous; bracts linear, (0.5–) 2–3 mm long, caducous; bracteoles linear, 0.8–1.5 mm long, caducous; pedicels 0.4–1.2 cm long. Calyx tubular to narrowly obconic, not costate, truncate (sometimes irregularly split), 0.9–1.6 × 0.4–0.7 cm, lepidote with glandular peltate trichomes, chartaceous, green to yellowish green, with patelliform glands and glandular areas evenly distributed. Corolla infundibular with abrupt widening above to calyx, strongly zygomorphic, furrowed, 4.2–6.6 cm long, 1.4–2.1 cm wide at tube mouth, externally puberulous, without glands, lobes 1.1–1.6 × 1–1.5 cm, margin rounded or acuminate and undulate, dark pink, purple or white. Androecium with all stamens included; longer filaments 1.8–2 cm long, short- er filaments 1.2–1.8 mm; staminode not seen; anthers 3.6–4 mm long, connectives protruding 0.8–1.5 mm. Gynoecium with ovary cylindric, 3.5–4.2 × c. 0.8 mm, furrowed, glabrous, style 2.9–3.4 cm long; stigma lanceolate; nectar disk annular and pulviniform under ovary, 0.8–1.3 × c. 2.5 mm. Fruit linear, flat, margins slightly raised, central ridge slightly raised, valves coriaceous, rough, 18–40 × 1.2–1.7 cm, glabrous; septum woody. Seeds elliptic, body elliptic, 0.9–1.2 × 2.2–2.3 cm, wings hyaline, 0.1–0.2 cm wide, margins entire.

Phenology — Produces flowers all year round. Fruits were collected from February to October.

Distribution and habitat — Fridericia triplinervia is distributed in C South America in dry and humid forests (semideciduous Atlantic and lowland Amazonian rainforests), as well as in dry and humid savannahs (Cerrado, Chaco and Pantanal). It is distributed in Bolivia ( Santa Cruz), Brazil ( Alagoas, Amazonas, Bahia, Ceará, Distrito Federal, EspÍrito Santo, Goiás, Mato Grosso, Mato Grosso do Sul, Minas Gerais, Paraná, Pernambuco, Rio de Janeiro, Rondônia and São Paulo) and in Paraguay ( Caazapá, Canendiyú, Central, Cordillera, Guairá, ParaguarÍ and Presidente Hayes).

Conservation status — Fridericia triplinervia is categorized as Least Concern (LC) based on its Extent of Occurrence ( 6,586,053 km 2) and Area of Occupancy ( 240,000 km 2).

Remarks — Fridericia triplinervia from dry areas is easily recognized by the conspicuous pocket domatia ( 2–4 mm long) found on the axils of the leaflet veins. This feature is shared with Lundia damazioi C. DC. , a species that is frequently confused with F. triplinervia in the Brazilian Atlantic rainforest and the W borders of the Brazilian Cerrado. However, L. damazioi can be easily distinguished by a series of features that are lacking in F. triplinervia , such as a prominent interpetiolar ridge, cordate leaflets, villous anthers and ciliate stigma ( Kaehler & Lohmann 2021b). In turn, F. triplinervia bears a conspicuous nectar disk, a feature that is lacking in L. damazioi . The specimens of F. triplinervia from C Amazonia bear smaller pocket domatia ( 1–2 mm long) and are similar to F. japurensis and F. schumanniana due to the ovate to elliptic leaflets with actinodromous basal or suprabasal venation. Fridericia triplinervia has tubular to narrowly obconic calyces with truncate rims (vs broadly obconic with a 5-lobed or irregularly split rim in F. schumanniana ) and has an infundibular corolla with an abrupt widening above the calyx (vs infundibular lacking an abrupt widening above the calyx in F. japurensis and F. schumanniana ). Fridericia triplinervia has coriaceous to woody fruits with rough valves, whereas F. japurensis and F. schumanniana have woody fruits, the former with hispid and sticky valves, the latter with verrucose valves.

Species exclusa

Tecoma moritziana Kraenzl. View in CoL in Repert. Spec. Nov. Regni Veg. 17: 218. 1921 ≡ Tabebuia moritziana (Kraenzl.) Schnee View in CoL in Pittier & al., Cat. Fl. Venez. 2: 409. 1947. – Syntype: Venezuela, El Palmar , Moritz 238 (B probably destroyed).

Remarks — Gentry (1982a) synonymized Tecoma moritziana in Fridericia grosourdyana based on the protologue description because the type of the former was probably destroyed. Gentry based the synonymization on the noticeable capitate glandular trichomes found on the inflorescence branches and calyx. Gentry (1982a) also highlighted differences in calyx length, with the calyx of T. moritziana smaller than that of F. grosourdyana . Moreover, F. grosourdyana can be distinguished by the shorter petiole, up to 8 cm (vs c. 10 cm in T. moritziana ), the elliptic to obovate central leaflet (vs broadly ovate in T. moritziana ), the tubular calyx (vs broadly campanulate in T. moritziana ), narrower calyx 0.25–0.4 cm wide (vs c. 0.7 cm wide in T. moritziana ) and 2 exserted stamens (vs all stamens included in T. moritziana ). Given all the morphological differences between these taxa, we do not recognize T. moritziana as a synonym of F. grosourdyana . Furthermore, T. moritziana lacks several diagnostic features of Fridericia , such as the triangular and minute prophylls, simple tendrils and lepidote ovaries with a single series of ovules on each placenta. We therefore exclude T. moritziana from Fridericia , although further studies are needed to verify its best generic placement.