Abrodictyum parviflorum (Poir.) Bauret & Dubuisson, 2017

Dubuisson, Jean-Yves, Bauret, Lucie, Grall, Aurelie, Senterre, Bruno, Said, Ahamada H., Pynee, Kersley, Ebihara, Atsushi & Hennequin, Sabine, 2017, Taxonomic study of the fern genera Abrodictyum C. Presl and Trichomanes L. (Hymenophyllaceae, Polypodiidae) in the western Indian Ocean, and description of a new Abrodictyum species for Madagascar, Phytotaxa 327 (3), pp. 201-222 : 211-213

Abrodictyum parviflorum (Poir.) Bauret & Dubuisson , comb. nov.

Trichomanes parviflorum Poir. (1808: 83) . Type:— MADAGASCAR. Without locality, without date, L.-M.A. du Petit-Thouars s.n. (holotype P, P00482986!; isotype P, P00482987!).

= Trichomanes lanceolatum Poir. (1808: 83) . Type:— MADAGASCAR. Without locality, without date, L.-M.A. du Petit-Thouars s.n. (holotype P, P00482989!; isotype P, P00482990!).

= Trichomanes meifolium Bory ex Willd. (1810: 509, 510). ≡ Macroglena meifolia (Bory ex Willd.) Copel. (1938: 83) . ≡ Cephalomanes meifolium (Bory ex Willd.) K.Iwats. (1984: 177) . ≡ Abrodictyum meifolium (Bory ex. Willd) Ebihara & Dubuisson (2006: 244). Type:— LA REUNION, ‘Habitat in sylvis montium insulae Borboniae’, without date, J.B.G.M. Bory de St.-Vincent s.n. (holotype B, BW 20 201010!; isotype P, P00612303!).

= Trichomanes longisetum Bory ex Willd. (1810: 510, 511). Type:— LA REUNION. ‘Habitat in sylvis insulae Borboniae’, without date, without collector (holotype B, BW 20 211010!).

= Trichomanes foeniculaceum Hook. (1845: 135) .

Type:—‘ Mauritius and Bourbon’, without date, J.B.G.M. Bory de St.-Vincent 127 (isotype P, P00064992!).

Terrestrial ferns, occasionally lithophytic, rarely epiphytic on the base of trunks. Rhizomes erect to short-creeping, 0.8–2.0 cm in diameter, bearing long tufted erect red-brown trichomes, especially densely covering apex, and numerous robust roots. Fronds clustered, erect; stipes 5–10(–12) cm long, mostly canaliculate, wingless, bearing numerous scattered trichomes identical to those present on rhizomes; rachis slightly winged on its upper part and with trichomes similar to those on the stipes and rhizomes. Laminae (10–)15–25(–40) × 4–8(–10) cm, plane or with pinnae and pinnules oriented in three dimensions, the fronds having thus a brush-like appearance, variable, narrowly to widely lanceolate or elliptic, more rarely triangular, with attenuated apex and obtuse to decurrent and more rarely rounded to truncated base, tri-pinnate to more divided, translucent; pinnae 2.0–5.0 × 0.5–1.5 cm, with trichomes on their main axes similar to those on the rachis, sub-horizontal to oblique, basally sub-opposite then alternate, usually lanceolate and pinnate to bi-pinnate; pinnules dichotomous or pinnate; ultimate segments 0.1–0.3 mm wide, capillary with a thin lamina not exceeding 3 rows of cells on both sides of the veins, linear and single-veined with acute to rounded ends, the veins not reaching the margin when lamina developed, often heterogeneous with a variable lamina width on the same frond; main venation pinnate and anadromous. Sori paratactic, located mostly on acroscopic basal-most segments of proximal pinnules, 0.8–1.2 × 0.5–0.8 mm, usually numerous, sometimes more than 30 per pinna, free, obconic to slightly campanulate, truncated or with a very slightly dilated mouth; receptacle short to long-exerted.

Distribution and habitat: —Western Indian Ocean, including Seychelles, expected from the Comoros. Usually terrestrial in the understory of rainforests, occasionally lithophytic, more rarely epiphytic on the base of trunks, often near streams, from low to high elevations (90–2,200 m) ( Table 1).

Representative specimens examined: — MADAGASCAR. Antsiranana: Doany, PN de Marojejy, October 2001, H. Rasolohery 567 (P00248515) ; Andapa, Forêt de Betaolana 16 October 1999, F. Rakotondrainibe et al. 4950 (P00179565), October 1999, F. Rakotondrainibe et al. 4965 (P00179589), F. Rakotondrainibe et al. 4990 (P00179614), F. Rakotondrainibe et al. 4991 (P00179615) ; Andapa, Forêt d’Analabe (massif d’Anjanaharibe-Sud , vers W), November 1999, F. Rakotondrainibe et al. 5165 (P00181155) ; Andapa , RNI 12 du Marojejy, October 1996, F. Rakotondrainibe 3503 (P00085045) ; Parc National de la Montagne d’Ambre , March 1992, F. Rakotondrainibe 1538 (P00085767) ; Montagne d’Ambre, Grand Lac , October 2004, T. Janssen et al. 2447 (P00590684) ; Andapa, Parc national de Marojejy, Abords de la piste entre camp 2 et camp 3, September 2015, G. Rouhan et al. 1576 (P02434064) .

Fianarantsoa: Andrambovato, Forêt d’Andrambovato , October 2000, F. Rakotondrainibe et al. 5986 (P00212491) ; Ambalavao, Ambatomboay, May 1995, F. Rakotondrainibe 2694 (P00059893). Tamatave: Massif de l’Andrangovalo au sud-est du Lac Alaotra, October 1937, H. Humbert 17766 (P01526654) .

— MAURITIUS. Mt Cocotte, October 2010, S. Hennequin et al. 318 (MAU 0008184, P02432374, P02432375).

— LA REUNION. Without locality, without date, L.-M.A. du Petit-Thouars s.n. (P00064996) ; Hauts de St Denis , 1833, Goudot s.n. (P00612302) ; without locality, 1834, Bernier 33 (P00064995, on the same sheet as the A. meifolium isotype); Brûlé de St Denis , 1886, C. Keller s.n. (P01526528) ; Forêt des Makes , May 1947, J. Bosser 11152 (P01526536) ; Saint-Philippe , Forêt de Mare Longue, 1 December 1967, T. Cadet 1183 (P01526521) ; Forêt des Cabris , Plaine des Palmistes, November 1985, K. Kramer et al. 9363 (P01526512) ; Bébour , April 1999, J.-Y. Dubuisson HR-1999-21 (P), HR-1999-22 (P) ; Saint-Philippe , Forêt de Mare Longue, April 2002, J.-Y. Dubuisson HR-2002-13 (P02433295, mix of plane [left] and brush-like [right] forms) ; Le Tremblet , April 2003, J.-Y. Dubuisson HR 2003-2 (P, brush-like form), HR 2003-2’ (P, plane form) .

— SEYCHELLES. Curieuse: Ouest de Mont Curieuse , September 2009, B. Senterre 5675 (SEY) . La Digue: Belle Vue , March 2009, B. Senterre et al. 5487 (SEY) . Praslin: Fond Azore, sommet Ouest , August 2008, B. Senterre 5402 (SEY) ; Rivière Salazie , December 2012, B. Senterre et al. 6400 (K, P, SEY) .

Note: —The species traditionally named and identified A. meifolium is widespread in rainforests of Madagascar and Mascarenes (especially La Réunion Island) from low to high elevations ( Tardieu-Blot 1951, 2008; Grangaud 2010; pers. obs.). It is usually described as expressing large erect highly divided fronds with thin capillary segments and with the different frond axes (from pinna to ultimate segments) mostly oriented in the three space dimensions, hence giving the bootle-brush-like shape ( Fig. 1C ). By contrast, all the other currently validated Abrodictyum species in the region (except A. cylindratum sp. nov. described above) show clearly plane fronds in the wild. Furthermore, the typical variety of A. meifolium usually displays narrowly elliptic/lanceolate fronds as already introduced and illustrated by the type from La Réunion (see also Fig. 1C ).

In Madagascar and the Mascarenes, many specimens with thin to capillary segments have been considered to belong to distinct species. These species include Trichomanes parviflorum (type from Madagascar: L.-M.A. du Petit-Thouars s.n., P00482986) with large, plane and widely lanceolate fronds, T. lanceolatum (type from Madagascar: L.-M.A. du Petit-Thouars s.n., P00482989) with typical lanceolate fronds (and not as wide as in T. parviflorum ), T. longisetum Bory ex Willdenow (1810: 510 , 511) (type from La Réunion: sin. coll., BW 20 211010) with small brush-like fronds, and T. foeniculaceum Hooker (1845: 135) (type from La Réunion or Mauritius: J.B.G.M. Bory de St.-Vincent 127, P00064992) which does not seem significantly different from T. parviflorum . Tardieu-Blot (1951) treated first T. longisetum and T. foeniculaceum as synonyms of T. meifolium , and then considered T. parviflorum a distinct species making T. foeniculaceum a synonym ( Tardieu-Blot 2008). For Roux (2009), T. longisetum and T. foeniculaceum are synonyms of A. meifolium while T. parviflorum and T. lanceolatum are synonyms of A. rigidum , at least the palaeotropical populations which are now referred to A. pseudorigidum , a species restricted to the Afro-Malagasy region ( Dubuisson et al. 2016).

Many old taxonomic revisions from 19th and 20th centuries were conducted in European institutions based on old material that had been sent from tropical areas. These old collections often lacked ecological data, and the botanists who studied them generally had not observed those ferns in the wild. This is true for the majority of the impressive Tardieu-Blot’s work, which was produced in the P Herbarium only. This fact can explain why different forms of a single polymorphic species have been considered to be distinct taxa. Our observations in the field in the western Indian Ocean, combined with the inferred phylogeny, give us enough evidence to say that all the forms named A. meifolium , T. parviflorum , T. lanceolatum , T. longisetum and T. foeniculaceum belong to a single highly polymorphic morphospecies, here referred to as A. meifolium (see Fig. 1 ). It is actually frequent to find all the morphological intermediates in one locality (e.g., specimens illustrated in Fig. 1C and 1D were collected at the same place; that is also the case for J.Y. Dubuisson HR-2002-13 - P02433295 that includes a mix of plane and brush-like forms). In addition, not all A. meifolium specimens display capillary segments with lamina reduced to a single row of cells (see Fig. 2A, B ). Many specimens have more laminar tissue, but the lamina does not exceed 3 cells on each side of the veins (see Fig. 2C, D ), and the development of the lamina is often heterogeneous, with specimens exhibiting both capillary segments and segments with a developed lamina on the same frond. Furthermore, most brush-like forms have capillary segments, whereas segments with lamina 2–3 cells wide are mostly observed in plane forms (e.g., T. parviflorum ). We also observed that brush-like forms are more often found in more or less open places that could be fully exposed to the sun and present potentially significant variations in hygrometry, while the plane forms mostly occur in the shadiest and wettest places, on banks near streams, with potentially limited variation in light and hygrometry. This suggests that the high polymorphism in shape, and maybe in lamina development, could be conditioned by micro-local environmental conditions such as the light and the variations of hygrometry. This polymorphism and/or plasticity would therefore allow the species to be widespread in all the rainforests and at all elevations in Madagascar and the Mascarenes. For Roux (2009), A. meifolium is restricted to Madagascar and La Réunion. However, we also observed, collected and identified the species in Mauritius and Seychelles. Its presence in Comoros awaits confirmation ( Saïd et al. 2017). The molecular phylogeny confirms the monophyly of typical A. meifolium and the specimens resembling T. parviflorum and T. lanceolatum , and by excluding A. cylindratum sp. nov. ( Fig. 3 ). If we recognize here all five species presented above as a single one ( A. meifolium ), the anteriority rule designates Trichomanes parviflorum as the valid name. Trichomanes parviflorum belongs to the Abrodictyum clade and must therefore be combined under Abrodictyum . Trichomanes lanceolatum was described by the same author, on the same year and on the same page as T. parviflorum ( Poiret 1808, p. 83) , but the latter species is cited first in the page, justifying that T. parviflorum is here proposed as the basionym.

Furthermore, the Asiatic species Abrodictyum pluma ( Hooker 1854: t997) Ebihara & Iwatsuki (2006: 243) also exhibits brush-like fronds with capillary segments and is often identified as A. meifolium in collections, but the species also shows creeping rhizome (vs. erect in A. parviflorum comb. nov.) and is molecularly clearly distinct ( Fig. 3 ).

Figures

FIGURE 1. Polymorphism in lamina shape of some similar Abrodictyum species in the western Indian Ocean. A. Narrowly cylindrical brush-like Malagasy A. cylindratum Dubuisson sp. nov. (G. Rouhan 1219; photograph by G. Rouhan). B. Habitus of a narrowly cylindrical brush-like Malagasy A. cylindratum Dubuisson sp. nov.(G. Rouhan et al. 1595, P02434066). C. Habitus of La Réunion A. meifolium (Bory ex.Willd) Ebihara & K.Iwats. exhibiting the shape of the species type with elliptic-lanceolate brush-like fronds (J.-Y. Dubuisson HR-2003- 2, P). D. Habitus of La Réunion A. meifolium with lanceolate plane fronds, specimen with such frond shape would correspond to distinct Trichomanes parviflorum Poir. according to Tardieu-Blot (2008) (J.-Y. Dubuisson HR-2003-2, P). E. Plane form of T. parviflorum in the wild, La Réunion (photograph by L. Bauret).

FIGURE 2. Anatomical details of pinnae, pinnules and segments for Abrodictyum and Trichomanes species of the western Indian Ocean. A. Capillary segments with a thin lamina of one row of cells of brush-like A. meifolium (Bory ex Willd.) Ebihara & K.Iwats (J.-Y. Dubuisson HR-2003-2, P). B. Details of capillary segments of brush-like A. meifolium (J.-Y. Dubuisson HR-2003-21, P). C. Segments with thin lamina of A. meifolium (H. Humbert 17766, P01526554), such similar segments are also observed in plane T. parviflorum Poir. D. Details of segments with thin lamina of A. meifolium (H. Humbert 17766, P01526554), such similar segments are also observed in plane T. parviflorum. E. Segments with homogeneous thin lamina of A. angustimarginatum (Bonap.) J.P. Roux (F. Rakotondrainibe 6392, P00244962, first identified as T. aff. meifolium). F. Pinnules of A. pachyphlebium (C.Chr.) Bauret & Dubuisson (G. Rouhan et al. 331, P00749278). G. Segments of A. tamarisciforme (Jacq.) Ebihara & Dubuisson with the particular distinct marginal row of cells (as detailed in inset) (J.-Y. Dubuisson HR-2002-15, P02433297). H. Pinnules of T. boivinii Bosch (F. Rakotondrainibe 1788, P00100302). I. Pinnae of T. madagascariense (Bosch) T. Moore (G. Rouhan et al. 367, P00749322).

FIGURE 3. Bayesian phylogeny inferred fromrbcL sequences showing relationshipswithin genus Abrodictyum (notedA) andTrichomanes (noted T). Support values on nodes are posterior probabilities. The new taxonomic treatment discussed here is indicated on the right. Taxa in coloured boxes represent Afro-western Indian Ocean clades here noted α and β as discussed in the text.