Unraveling the phylogeny of polygrammoid ferns (Polypodiaceae and Grammitidaceae): exploring aspects of the diversification of epiphytic plants

Abstract

We explore the phylogeny of the polygrammoid ferns using nucleotide sequences derived from three plastid loci for each of 98 selected species. Our analyses recovered four major monophyletic lineages: the loxogrammoids, two clades consisting of taxa restricted to the Old World, and a largely neotropical clade that also includes the pantropical Grammitidaceae. The loxogrammoid lineage diverges first and is sister to a large clade comprising the three remaining species-rich lineages. One paleotropical clade includes the drynarioid and selligueoid ferns, whereas the second paleotropical clade includes the platycerioids, lepisoroids, microsoroids, and their relatives. The grammitids nest within the neotropical clade, although the sister taxon of this circum-tropic, epiphytic group remains ambiguous. Microsorum and Polypodium, as traditionally defined, were recovered as polyphyletic. The relatively short branch lengths of the deepest clades contrast with the long branch lengths leading to the terminal groups. This suggests that the polygrammoid ferns arose through an old, rapid radiation. Our analysis also reveals that the rate of substitution in the grammitids is remarkably higher relative to other polygrammoids. Disparities in substitution rate may be correlated with one or more features characterizing grammitids, including species richness, chlorophyllous spores, and an extended gametophytic phase.

Introduction

Recent studies in fern systematics have improved not only our understanding of the deep relationships of ferns (Hasebe et al., 1995; Pryer et al., 1995, Pryer et al., 2001a) but also the phylogenetic relationships within major groups of derived ferns using either a single-gene approach (Crane et al., 1995; Gastony and Rollo, 1995; Gastony and Ungerer, 1997; Murakami et al., 1999) or a multi-gene approach (Schneider et al., 2002b, Schneider et al., in prep; Smith and Cranfill, 2002; Wolf, 1995). Unraveling the phylogeny of major derived fern groups, such as asplenioid, dryopteroid, and polygrammoid ferns, is of particular interest because these ferns are important components of current tropical vegetation. Comparisons of the phylogenetic patterns of ferns and angiosperms are a crucial component in our efforts to understand the evolution of modern floras. Polygrammoid ferns are not only one of the most species-rich groups of ferns but they also stand out among vascular plants as one of the most diverse and abundant groups of vascular epiphytes in tropical and subtropical forests, particularly in the paleotropics (Benzing, 1990). The lineage includes two traditionally recognized families: Grammitidaceae and Polypodiaceae (Hennipman et al., 1990; Parris, 1990), but recent studies suggest that Grammitidaceae nest within Polypodiaceae. These studies, however, are so far unable to determine the closest relatives of Grammitidaceae (Hasebe et al., 1995; Haufler et al., 2003; Pryer et al., 1995; Schneider et al., 2002b). Phylogenetic studies also support another group of epiphytic ferns, the davallioids (Davalliaceae), as being the sister clade to the polygrammoids (Hasebe et al., 1995; Pryer et al., 1995). Together, these fern groups comprise about 50% of extant epiphytic ferns (estimate based on species numbers given in Kubitzki, 1990).

Polypodiaceae were among the first groups of plants studied using a cladistic approach, but the utility of these early studies was limited because molecular data and associated analytical tools were unavailable. Nevertheless, the results of those studies (Bosman, 1991; Hennipman and Roos, 1982, Hennipman and Roos, 1983; Hovenkamp, 1986, Hovenkamp, 1996; Rödl-Linder, 1990; Roos, 1985) made substantial contributions to the most recent classification of Polypodiaceae (Hennipman et al., 1990). Unfortunately, high levels of homoplasy among morphological characters introduce considerable error in phylogenetic studies of this family (Hennipman and Roos, 1983; Hovenkamp, 1996). For example, several characters considered to be taxonomically important in the polygrammoid ferns have been demonstrated to be homoplastic in other fern groups. These characters include leaf dissection in asplenioid (Murakami et al., 1999) and blechnoid ferns (Cranfill, 2001), as well as spore shape in vittarioid ferns (Crane et al., 1995). Detailed studies of leaf venation (Mitsuta, 1984), paraphyses (Baayen and Hennipman, 1987), spore morphology (Hennipman, 1990; Van Uffelen, 1997), and stomatal patterns (Sen and Hennipman, 1981) have provided evidence that these characters are homoplastic in the polygrammoid ferns. Additional characters, such as DNA sequence data, can be used to resolve uncertainties of morphological character evolution. Molecular sequence data can be analyzed along with morphological data (Smith et al., 2000) or separately, as in the present study. The resulting phylogenetic framework could then be used to infer the evolution of morphological features, as has been done for the evolution of body plans of vascular plants (Schneider et al., 2002a). This kind of study can elucidate the evolution of morphological features in polygrammoid ferns and their taxonomic value (Schneider et al., 2002b).

The major objective of this study is to recover a global phylogeny of the polygrammoid ferns, utilizing three chloroplast genome loci (rbcL and rps4 genes; rps4-trnS intergenic spacer) sampled widely throughout the polygrammoid ferns. Because a more detailed study of the phylogeny of Grammitidaceae will be published separately (Ranker et al., in prep.), we include only a small selection of grammitid species in this study. Detailed phylogenetic studies of selected groups of Polypodiaceae, such as Gymnogrammitis (Schneider et al., 2002b) and Lecanopteris (Haufler et al., 2003), are already published. A further goal of our study is to evaluate currently accepted classifications of Polypodiaceae, in particular, the recent classification of Hennipman et al. (1990).

Our results have implications not only for the taxonomy of the group but also for studies focusing on the evolution of morphological, ecophysiological, and molecular features. The results of the present study provide insights into the evolution of epiphytic communities in the Cenozoic, suggesting rapid diversification that may be correlated with the establishment and expansion of angiosperm-dominated forests. We explore for the first time the heterogeneity of substitution rates in a highly derived and species-rich group of ferns for possible correlations with species richness, habitat change, and the evolution of chlorophyllous spores.