Hymenophyllaceae

Dubuisson, Jean-Yves, Pechon, Timothée Le, Bauret, Lucie, Rouhan, Germinal, Reeb, Catherine, Boucheron-Dubuisson, Elodie, Selosse, Marc-André, Chaussidon, Cécile, Dajoz, Isabelle, Pynee, Kersley, Grangaud, Edmond, Robert, Yannis, Tamon, Jean- Maurice & Hennequin, Sabine, 2018, Disentangling the diversity and taxonomy of Hymenophyllaceae (Hymenophyllales, Polypodiidae) in the Mascarene archipelago, with ecological implications, Phytotaxa 375 (1), pp. 1-58 : 40-41

Hymenophyllaceae endemicity and ecology on the Mascarenes

Before this study, a single species was considered to be endemic to the archipelago ( Didymoglossum barklyanum ). We demonstrate here that we can add two further species: H. hygrometricum and H. fumarioides . The presence of these two species outside of the Mascarenes, and especially in Madagascar, as reported in floras and/or herbaria, appears not to be verified and thus likely erroneous. The confirmation of endemicity for the variety C. inopinatum var. tamonii and for the forma H. capillare f. nanum , needs additional investigations in the wild, especially in Madagascar. Crepidomanes minutum var. mascarenense , initially considered as endemic to the Mascarenes, is actually present on all the islands of the western Indian Ocean hosting Hymenophyllaceae (except Rodrigues). In addition, similar (but not related) dwarf forms with rare frond proliferation (and named “ saxifragoides ”) are frequently observed in Asia and the Pacific ( Nitta et al. 2011).

At the specific level, the endemicity for Hymenophyllaceae remains low (13%). This contrasts with the endemicity of the family in Madagascar, the closest continental island, with at least 16 endemic species out of 43 (37.2%) according to Tardieu-Blot’s data (1951), or with a corrected estimate of 19 out of 48 (39.6%) according to the recent revisions (F. Rakotondrainibe, pers. comm.; Roux, 2009; Dubuisson et al. 2013, 2016, 2017; Bauret et al. 2015; Saïd et al. 2017). This is not unexpected if we compare the estimated ages of the isolation of Madagascar and of the appearance of the Mascarene Plateau (85 Myrs vs. 35 Myrs) and if we assume that endemicity potentially increases with the age of an island.

This relatively low endemicity is nevertheless more or less equivalent to that found in the archipelago for the fern families Aspleniaceae (13%), Blechnaceae (20%), Polypodiaceae (17.4%), Pteridaceae (10%), and Thelypteridaceae (14.3%), but contrasts to that of Dryopteridaceae (and principally genus Ctenitis ; 52.2%), Lomariopsidaceae (3 out of 4 species; 75%), and Cyatheaceae (all 4 Alsophila species; 100%), according to Hennequin et al. (2014). The high diversity (at least 23 species) combined with the low endemicity and the fact that all the three endemic species are not phylogenetically closely related (i.e., they do not cluster into a single distinct clade), suggest that the diversity of the Hymenophyllaceae in the archipelago would have originated from multiple colonization events rather than local diversification following few colonization. All non-endemic Hymenophyllaceae are present in Madagascar and the three endemic Mascarenan species also are closely related to species present in the western Indian Ocean ( Ebihara et al. 2007; Hennequin et al. 2003, 2010; Dubuisson et al. 2013, 2016, 2017; pers. unpublished molecular data). This strongly suggests a Malagasy origin for the Mascarenan Hymenophyllaceae flora, in addition to a probable local anagenetic origin for each of the three endemic species.

Most Hymenophyllaceae taxa are colonial epiphytes and the gametophytes in the family are usually colonial and perennial. Dassler & Farrar (1997, 2001) proposed that colonial-epiphytic taxa with colonial long-lived gametophytes would have more success in colonizing distant oceanic islands than terrestrial taxa which have less strongly colonial and shorter-lived gametophytes ( Fig. 19 ). The colonial and long-lived status of epiphytic gametophytes would actually increase the probability of fertilization between two distinct gametophytes, allowing relatively fast colonization and installation of epiphytic sporophytes.According to Dassler & Farrar’s proposition, we could hypothesize that on distant oceanic islands (1) there are more epiphytic than terrestrial taxa, and (2) under competition, numerous species which would have succeeded in colonizing the same habitat at more or less the same time, would have fewer possibilities to diversify but could coexist if niches are not fully overlapping or if the available area is large enough to host them. On the Mascarenes, the first hypothesis seems probable: of the 23 species currently present on the archipelago, 20 are colonial epiphytes, one is hemi-epiphytic, one is an individual epiphyte, and a single is terrestrial ( A. parviflorum ) (4.3%), whereas Madagascar hosts at least 8 terrestrial species out of 48 species (16.7%) including A. parviflorum . The Mascarenan rainforests host many colonial epiphytic Hymenophyllaceae species that would have originated from numerous colonization events, resulting in a maximum of 17 colonial epiphytic taxa in the lowland rainforest and up to 13 taxa in the montane rainforest (see Table 3 and Fig. 20 ). This does not contradict the second hypothesis. By contrast, the twelve Mascarenan endemic and terrestrial Ctenitis species ( Dryopteridaceae ) have derived from a single common ancestor, hence illustrating a local cladogenetic radiation ( Hennequin et al. 2017). The single Mascarenan terrestrial Hymenophyllaceae species, A. parviflorum , does not show an equivalent local radiation. With its gametophytes that are more long-lived than those of Dryopteridaceae and its ecological opportunism allowing a widespread distribution in all rainforest types on almost the entire elevational gradient in the western Indian Ocean (except the Comores), A. parviflorum would have quickly succeeded in colonizing the wet forests of the archipelago. This rapid colonization appears to have prevented a diversification process as well as, perhaps, the colonization by other terrestrial species of Hymenophyllaceae . By comparison, on the Comores archipelago, which is equidistant between Madagascar and southern Africa (see Fig. 1 ) and has more or less the same age as the current Mascarene islands, A. parviflorum is lacking and replaced by the terrestrial and closely related A. pseudorigidum Bauret & Dubuisson ( Dubuisson et al. 2016: 161), also widespread in Madagascar, which occupies the same niches along the entire local elevational gradient ( Dubuisson et al. 2016, 2017; Saïd et al. 2017). The latter species appears to have succeeded in colonizing the Comores archipelago instead of A. parviflorum , by competition or stochastically, and reciprocally on the Mascarenes for A. parviflorum . Both examples of species of Hymenophyllaceae and genus Ctenitis would therefore illustrate the importance of history, ecological preferences, and especially the gametophyte biology, too often neglected in ecological studies on ferns (as reviewed Walker et al. 2010), for explaining the local assemblage of the fern flora and contrasted patterns between lineages and localities.

Figures

FIGURE 1. Location of the Mascarene Archipelago (or Mascarenes) in the western Indian Ocean (from Pynee et al. 2012).

FIGURE 19. Importance of gametophyte ecology and biology for the sporophyte settlement on distant islands according to Dassler & Farrar’s (2001) hypothesis. The long-lived and colonial epiphytic gametophytes (above) would have the possibilities to realize fertilization many seasons/years (T) after the spore germination, whereas short-lived and non-colonial terrestrial gametophytes (below) would fail in realizing the fertilization if they are not close enough, and die at the end of the season (red crosses). Hence, the epiphytic taxa with longlived and colonial gametophytes would have a higher probability of establishment than terrestrial taxa. Hymenophyllaceae and the genus Ctenitis (Dryopteridaceae) illustrate such contrasts in the Mascarene archipelago, explaining the difference in the level of endemicity (see text for an in-depth discussion).

FIGURE 20. Elevational distribution on each Hymenophyllaceae taxon on the windward (red plots) and leeward (blue plots) sides of La Réunion, and correspondence with thermal zones or ranges according to Cadet (1980) and provided here for the windward side, with ranges of annual average temperatures. The dotted horizontal lines correspond to the limits of leeward ecosystems. The plots indicate the average elevation. Crepidomanes trinerve is here lacking because it was only observed at a single locality on the leeward side (‘Les Canots’ at ~800 m). The range of each taxon, as illustrated by the vertical lines, is assumed to correspond to the range between the minimum and the maximum elevation at which they have been observed, even if the taxon has not been observed in all the intermediate localities between such limits.