Botanical Journal of the Linnean Society, 2010, 163, 305–359. With 16 figures Phylogenetics and classification of the pantropical fern family Lindsaeaceae SAMULI LEHTONEN1*, HANNA TUOMISTO1, GERMINAL ROUHAN2 and MAARTEN J. M. CHRISTENHUSZ3 1 Department of Biology, University of Turku, FI-20014, Finland Muséum national d’Histoire naturelle, UMR CNRS 7205 ‘Origine, Structure et Evolution de la Biodiversité’, 16 rue Buffon CP39, F-75005 Paris, France 3 Department of Botany, The Natural History Museum, Cromwell Road, London SW7 5DB, UK 2 Received 19 February 2010; revised 3 May 2010; accepted for publication 2 June 2010 The classification and generic definition in the tropical–subtropical fern family Lindsaeaceae have been uncertain and have so far been based on morphological characters only. We have now studied the evolutionary history of the Lindsaeaceae by simultaneously optimizing 55 morphological characters, two plastid genes (rpoC1 and rps4) and three non-coding plastid intergenic spacers (trnL-F, rps4-trnS and trnH-psbA). Our data set included all genera associated with Lindsaeaceae, except Xyropteris, and c. 73% of the currently accepted species. The phylogenetic relationships of the lindsaeoid ferns with two enigmatic genera that have recently been included in the Lindsaeaceae, Cystodium and Lonchitis, remain ambiguous. Within the monophyletic lindsaeoids, we found six wellsupported and diagnostic clades that can be recognized as genera: Sphenomeris, Odontosoria, Osmolindsaea, Nesolindsaea, Tapeinidium and Lindsaea. Sphenomeris was shown to be monotypic; most taxa formerly placed in that genus belong to the Odontosoria clade. Ormoloma is embedded within Lindsaea and therefore does not merit recognition as a genus. Tapeinidium is sister to a clade with some species formerly placed in Lindsaea that are morphologically distinct from that genus and are transferred to Osmolindsaea and Nesolindsaea, proposed here as two new genera. We do not maintain the current subgeneric classification of Lindsaea itself, because neither of the two generally accepted subgenera (Lindsaea and Odontoloma) is monophyletic, and most of the sections also appear unnatural. Nesolindsaea shows an ancient biogeographical link between Sri Lanka and the Seychelles and many of the main clades within Lindsaea have geographically disjunct distributions. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359. ADDITIONAL KEYWORDS: lindsaeoids – Nesolindsaea – Osmolindsaea – phylogeny – pteridophytes – sensitivity analysis. INTRODUCTION The fern family Lindsaeaceae comprise c. 200 species and are distributed widely in the tropics, with several species extending into the subtropics in temperate South America, East Asia and New Zealand (Kramer, 1957a). Early on, lindsaeoid ferns were considered a part of Davalliaceae (Presl, 1836; Fée, 1852; Christensen, 1938) and were later placed in Dennstaedtiaceae (Holttum, 1947; Tryon & Tryon, 1982; Kramer *Corresponding author. E-mail: samile@utu.fi & Green, 1990). The name Lindsaeaceae was proposed by Pichi-Sermolli (1970). Early molecular systematic studies placed Lindsaeaceae as sister to a clade uniting Dennstaedtiaceae and Pteridaceae (Hasebe et al., 1994; Wolf, Soltis & Soltis, 1994) and hence supported the recognition of Lindsaeaceae as a separate family. The six traditionally recognized lindsaeoid genera are Lindsaea Dryand. ex Sm. (c. 165 spp.), Odontosoria Fée (c. 12 spp.), Ormoloma Maxon (1 or 2 spp.), Sphenomeris Maxon (c. 11 spp.), Tapeinidium (C.Presl) C.Chr. (c. 18 spp.) and Xyropteris K.U.Kramer (1 sp.). Results of molecular studies have also led to the expansion of © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 305 306 S. LEHTONEN ET AL. Figure 1. Phylogenetic relationships within leptosporangiate ferns according to Schuettpelz et al. (2008). the family to include two enigmatic genera. These are Lonchitis L., which was earlier placed in Dennstaedtiaceae (e.g. Tryon & Tryon, 1982), and Cystodium J.Sm., previously considered to belong to the tree fern family Dicksoniaceae (Korall et al., 2006; Schuettpelz & Pryer, 2007). However, a separate family Lonchitidaceae has also been proposed for the genus Lonchitis on the basis of morphological differences with lindsaeoid ferns (Christenhusz, 2009). In a recent synopsis of fern phylogenetics, Schuettpelz & Pryer (2008) recognized Saccoloma Kaulf. (the sole genus of Saccolomataceae) as the sister lineage of Lindsaeaceae (Fig. 1). Earlier, however, Saccoloma has been placed as sister lineage to all other Polypodiales (Pryer et al., 2004; Korall et al., 2006; Perrie & Brownsey, 2007). Schneider et al. (2004) resolved Lonchitis as the sister of other Polypodiales and Saccoloma as the sister of Lindsaeaceae (which then excludes Lonchitis). Although the hypothesized phylogenetic relationships between these groups differ among studies, there is a consensus that Lindsaeaceae are one of the early-diverging lineages from the main stock of polypod ferns. The Early Cretaceous fossil remains with characteristic root anatomy (i.e. a sclerenchymatous outer cortex combined with an innermost cortical layer six cells wide) confirm a minimum age of c. 99 My for the lindsaeoid ferns (Schneider & Kenrick, 2001). Based on likelihood analyses, Pryer et al. (2004) estimated that lindsaeoids diverged c. 133 Mya. Generic delimitation on morphological grounds has been difficult within Lindsaeaceae. The lack of obvious morphological discontinuities led Tryon & Tryon (1982) and Kramer & Green (1990) to consider Sphenomeris congeneric with Odontosoria, but this classification has been criticized (Barcelona, 2000) and was not followed by Smith et al. (2006, 2008). Tapeinidium moorei (Hook.) Hieron. has been especially difficult to place and it has been included in Lindsaea, Odontosoria and Sphenomeris. In general terms, Lindsaea is characterized by sori that open towards the margin and have the indusium attached to the disc (Dryander, 1797). Ormoloma is characterized by the combination of simply pinnate lamina and strictly uninerval sori and Odontosoria and Sphenomeris by their cuneately–dichotomously divided lamina with apical sori and large spores. Tapeinidium has only a slight adaxial sulcation of the axes and pinnate leaf architecture, with the sori on lateral lobes, and Xyropteris differs in having continuous sori and basally auriculate pinnae. Lonchitis is morphologically clearly different from the other genera as it has sori in the sinus and succulent, dorsiventral, hairy rhizomes. Saccoloma has stout erect rhizomes and marginal cup-shaped sori, whereas Cystodium in many respects resembles a small tree fern, which explains its previous placement in Dicksoniaceae. The systematics of lindsaeoid genera has been extensively studied by Kramer (1957a, b, 1967a, b, c, 1970, 1971a, b, c, 1972a, b, 1988, 1989a, b, 1991; Kramer & Tindale, 1976). Kramer subdivided the largest genus, Lindsaea, into two subgenera based on rhizome anatomy (Kramer, 1967b). Subgenus Lindsaea included species with an essentially terrestrial, short to moderately long creeping rhizome with radially symmetric steles, and subgenus Odontoloma included species that are usually epiphytic and have long creeping rhizomes with strongly dorsiventral steles (Kramer, 1967b). The subgenera were further divided into 17 sections in subgenus © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE Lindsaea and six sections in subgenus Odontoloma (Kramer, 1957a, 1967b, 1972a). However, Kramer (1957a, 1967b, 1971a) clearly defined all these sections as paraphyletic entities, thus his classification was never meant to be phylogenetic in the sense of Hennig (1966). Lindsaeoid ferns have been poorly sampled in previous molecular studies. For example, Pryer et al. (2004) only included one lindsaeoid species and Schuettpelz & Pryer (2007) five. Here, we present the first comprehensive analysis of the phylogenetic relationships within Lindsaeaceae. Our analysis is based on simultaneous optimization of morphological and molecular evidence and it includes 165 of the 249 lindsaeoid taxa (species, subspecies and varieties) recognized by Kramer (1957a, b, 1967a, b, c, 1970, 1971a, 1972a, b, 1988, 1989a, b, 1991) and Kramer & Tindale (1976), with an additional five lindsaeoid taxa not recognized by Kramer and five outgroup taxa. On the basis of the results, we propose a new generic classification for the family. MATERIAL AND METHODS TAXON SAMPLING AND MORPHOLOGICAL DATA For this study, we aimed to sample as many taxa of Lindsaeaceae as possible. Our molecular sampling includes large numbers of silica-dried material collected by us (mostly from the Neotropics) or provided by colleagues (mostly from the Palaeotropics). In addition, many taxa were sampled using existing herbarium specimens deposited at AAU, BISH, BM, CAY, L, MO, P, TUR, U, US and Z (herbarium acronyms according to Thiers, 2010). We successfully sampled 195 specimens representing 175 taxa. All eight genera placed in the family by Smith et al. (2008) were included in our analyses, except the monotypic Xyropteris, for which DNA amplification failed. The sampling covered c. 71% of the generally accepted taxa of Lindsaea (76% of the species), 50% of the taxa of Tapeinidium (56% of the species), 56% of the taxa in the Sphenomeris-Odontosoria group (74% of the species), both Ormoloma spp. accepted by Kramer (1957a), one of the two Lonchitis spp. and the sole Cystodium sp. Overall, we obtained c. 69% taxon sampling and 73% species sampling within the family in the broad sense. In addition, two Saccoloma spp. representing the closely related Saccolomataceae were included. The uncertainty concerning the closest relatives of the lindsaeoid ferns may cause problems for the rooting of the tree. Depending on the study, either Saccoloma (Pryer et al., 2004; Korall et al., 2006; Perrie & Brownsey, 2007), Lonchitis (Schneider et al., 2004) or the Saccoloma–Lonchitis–Lindsaeaceae 307 (Schuettpelz & Pryer, 2007) clade is the sister lineage of all other polypod ferns. The analysis by Schuettpelz & Pryer (2007) was based on the largest number of data (rbcL, atpB and atpA assembled for 400 leptosporangiate ferns) and therefore can be considered the most reliable hypothesis presented to date. We based our outgroup selection on this hypothesis and rooted our tree with a dennstaedtioid polypod, Pteridium pinetorum C.N.Page & R.R.Mill. (northern bracken, Dennstaedtiaceae), a member of the proposed sister lineage to the Lindsaeaceae– Saccolomataceae clade. We identified the sequenced specimens by comparing them with c. 4000 herbarium specimens. Whenever type material or digital type images were available, they were used to verify the correct application of names. In other cases we named the specimens by following the specimen lists, keys and descriptions provided in the numerous publications of Kramer (e.g. Kramer, 1957a, 1967a, b, c, 1970, 1971a, b, c, 1972a, b, 1989a, 1991; Kramer & Tindale, 1976). Many of the sequenced specimens had been identified to species and annotated by K. U. Kramer himself, but even then we verified the consistency of identifications. We coded a total of 55 morphological characters for the studied taxa (Appendices 1 and 2). The characters were either novel or adapted from Pryer, Smith & Skog (1995), Stevenson & Loconte (1996) or Barcelona (2000). Spore characters were mostly coded according to information given in the literature (Kramer, 1957a, 1967a, b, c, 1970, 1971a, b, 1972a, b, 1989a, 1991; Kramer & Tindale, 1976; Kramer & Green, 1990), but most of the other characters were coded by direct observation of herbarium specimens. The collections consulted for our study are held at AAU, BISH, BM, CAY, K, KYO, L, MICH, MO, NSW, P, TUR, U, UC, US and Z. MOLECULAR DATA: SELECTION OF THE DNA REGIONS We obtained silica-dried material for 37 specimens and used herbarium material to broaden the molecular sampling by another 158 specimens. Because DNA fragmentation is a problem with herbarium material, we had to focus on relatively short sequences (Lehtonen & Christenhusz, 2010). This precluded using the conservative mitochondrial genome (Seberg & Petersen, 2006) and many of the plastid markers commonly used in fern systematics (e.g. atpA, atpB, rbcL). The nuclear genome presents particular problems because of paralogous sequences and common polyploidy among ferns (Schuettpelz et al., 2008). For amplification and sequencing we therefore selected the four relatively short plastid regions with varying rates of evolution, trnL-trnF, trnH-psbA, rpoC1 and © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 308 S. LEHTONEN ET AL. rps4 + rps4-trnS; the latter was further divided into two fragments for the analysis (see below), thus resulting in five DNA fragments in our analyses. Furthermore, we coded observed inversion events (see below) as binary characters in a separate data matrix (Appendices 3 and 4). trnL-trnF intergenic spacer This spacer region was among the first non-coding sequences used in plant systematics (Taberlet et al., 1991) and it has become one of the most widely used at lower taxonomic levels in fern systematics (Small et al., 2005). The spacer region amplifies almost universally in ferns with standard primers and it provides a high number of variable sites (Small et al., 2005). However, the region is typically short (c. 400 bp in our case) and reconstructing its evolution is complicated by numerous structural changes, including short inversions (Quandt & Stech, 2004; Kim & Lee, 2005; Lehtonen, Myllys & Huttunen, 2009). rpoC1 gene This gene was among the markers proposed and assessed for the bar coding of land plants (Chase et al., 2007; Hollingsworth et al., 2009), although it was not finally selected (CBOL Plant Working Group, 2009). It amplifies well in many plant groups, but the rate of evolution is not high (Chase et al., 2007). We sequenced a c. 730-bp long part of rpoC1. trnH-psbA intergenic spacer This region is now widely used in lower-level plant systematics, because of its high level of variation among closely related species and ease of amplification (Shaw et al., 2005). It had therefore been suggested as a potential DNA bar code for land plants (Chase et al., 2007). However, the region is typically short (c. 430 bp in our case) and usually does not provide sufficient information for phylogenetic reconstruction unless coupled with other regions (Shaw et al., 2005). Like the trnL-trnF spacer, the trnH-psbA spacer often contains many structural rearrangements, which complicates its use in phylogenetic inference (Bain & Jansen, 2006; Storchová & Olson, 2007; Borsch & Quandt, 2009; Lehtonen et al., 2009). rps4 gene The rps4 gene (c. 580 bp) was amplified in this study, together with the rps4-trnS intergenic spacer (c. 250 bp). For the phylogenetic analyses, the rps4 gene was separated from the spacer region and the gene and the spacer region were submitted for the analyses as separate fragments. This gene is often used in fern systematics, but because of its low substitution rate it is more useful at deeper phylogenetic levels (Small et al., 2005). rps4-trnS intergenic spacer The rps4-trnS spacer is the shortest of the fragments used in this study (c. 275 bp) and it has a relatively low substitution rate in comparison with many other non-coding plastid DNA regions (Shaw et al., 2005; Small et al., 2005). Nevertheless, the spacer is widely used in fern systematics (Small et al., 2005, and references therein). MOLECULAR DATA: LABORATORY PROCEDURES Total genomic DNA was extracted by using DNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA) or E.Z.N.A. SP Plant DNA Kit (Omega Bio-tek, Doraville, GA, USA). For silica-dried material, the manufacturer’s extraction protocols were followed, but for herbarium specimens a modified protocol (30 min incubation, 450 mL of cell lysis buffer, 50 mL of elution buffer, 10 min elution) as described by Drábková, Kirschner & Vlcek (2002) was used. Primers used for amplification and sequencing are listed in Table 1. DNA fragments were amplified using PureTaq RTG PCR beads (Amersham Biosciences, Piscataway, NJ, USA). Each reaction contained 5 mL of DNA template, Table 1. Primers used for PCR and sequencing Locus 5′–3′ Reference trnL-trnF e: GGTTCAAGTCCCTCTATCCC f: ATTTGAACTGGTGACACGAG Taberlet et al., 1991 Taberlet et al., 1991 trnH-psbA trnH: CGCGCATGGTGGATTCACAATCC psbA3′f: GTTATGCATGAACGTAATGCT C Tate & Simpson, 2003 Sang, Crawford & Stuessy, 1997 rpoC1 LP1: TATGAAACCAGAATGGATGG LP5: CAAGAAGCATATCTTGASTYGG Chase et al., 2007 Chase et al., 2007 rps4-trnS rps4.5′: ATGTCSCGTTAYCGAGGACCT trnSGGA: TTACCGAGGGTTCGAATCCCT C Small et al., 2005 Shaw et al., 2005 © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE 1 mL of each primer and 18 mL of double distilled water (ddH2O). The PCRs were carried out in a GeneAmp PCR System 9700. The amplification profiles were: initial denaturation (95 °C, 2 min) followed by 35 cycles of amplification, hybridization and extension (95 °C, 1 min; 57 °C, 1 min; 72 °C, 1 min), and 7 min of final extension at 72 °C for rps4-trnS, initial denaturation (95 °C, 2 min) followed by 35 cycles of amplification, hybridization and extension (95 °C, 1 min; 52 °C, 1 min; 72 °C, 1 min), and 7 min of final extension at 72 °C for trnL-trnF, initial denaturation (94 °C, 1 min) followed by 40 cycles of amplification and extension (94 °C, 30 s; 48 °C, 40 s; 72 °C, 40 s), and 5 min of final extension at 72 °C for trnH-psbA and rpoC1. PCR products were purified and sequenced in both directions under BigDye™ terminator cycling conditions by Macrogen Inc. (Seoul, South Korea; http:// www.macrogen.com). Sequences were assembled and incorrect bases corrected using 4Peaks (A. Griekspoor & T. Groothuis, http://mekentosj.com). The corrected sequences were initially aligned with ClustalX version 1.83.1 (Thompson et al., 1997) by using default alignment parameters. Putative inversions were identified by eye from the aligned sequence matrix. Before submitting the data to final POY analyses (see below), all the gaps inserted by ClustalX were removed. GenBank accession numbers and voucher specimen information are given in Table 2. Most of the sequences were newly generated, but some were taken from Lehtonen & Tuomisto (2007) and Lehtonen et al. (2009). CONCEPT OF HOMOLOGY, DATA OPTIMIZATION AND THE ROLE OF TOTAL EVIDENCE Characters are homologous when they share a common ancestry and then they tell us about phylogenetic history (Patterson, 1982). Therefore, homology statements and phylogenetic history are inseparable problems (Patterson, 1982; Löytynoja & Goldman, 2009). Homology should be assumed unless there is evidence against it, because otherwise building phylogenetic hypotheses becomes impossible (Hennig, 1966) and the evidence for refuting homology statements can only come from the character congruence test in phylogenetic analysis (Farris, 1983). From this, it follows that a phylogeneticist’s task is to maximize homology statements in the observed character data (e.g. De Laet, 2005; Phillips, 2006). Multiple sequence alignment is traditionally used to establish primary homology statements among sequence positions on the basis of phenetic similarity, and subsequent phylogenetic analysis is used to determine which of the primary homologies are considered real (‘secondary’) homologies (de Pinna, 1991). Although improved computing 309 capacity has made this two-step approach unnecessary (e.g. Lehtonen, 2008; Liu et al., 2008; Wheeler & Giribet, 2009), many authors still favour it (e.g. Simmons, 2004; Kjer, Gillespie & Ober, 2007; Ogden & Rosenberg, 2007). Direct optimization (DO) provides a means to determine homology correspondences and search for the optimal tree simultaneously. The homology correspondences are then viewed as the results of an analysis and the obtained topology maximizes the homology correspondences in the available data (Wheeler, 1996). This requires that all the relevant data are optimized together in a total-evidence character congruence test (Kluge, 1989; Phillips, 2006; Fitzhugh, 2006a). Consequently, if morphological characters are included in the analysis, they will affect the determination of homology correspondences in the molecular data (‘sequence alignment’), just as DNA characters may direct the optimization of morphological character-state transformations (Phillips, 2006; Agolin & D’Haese, 2009). We chose to use parsimony as the optimality criterion when selecting the preferred tree and homology correspondences (Farris, 1983). When no common model of evolution is assumed, maximum likelihood gives the same results (Tuffley & Steel, 1997; Steel & Penny, 2000), but requires parameterization of branch lengths in addition to topology (Goloboff, 2003). Our data include a large number of heterogeneously evolving characters (structural changes, e.g. minute inversions and repeats in non-coding sequences, morphological characters) for which realistic evolutionary models are not available (Borsch & Quandt, 2009). Lehtonen et al. (2009) demonstrated that parsimony outperformed maximum likelihood and Bayesian methods in inferring phylogenies from sequences containing small-scale structural rearrangements. We also prefer parsimony over Bayesian approaches because of the impossibility of flat priors (Steel & Pickett, 2006), the need to specify probability distributions for all parameters (Moyle et al., 2009) and the peculiar behaviour of Bayesian inference in the presence of missing data (Goloboff & Pol, 2005). We performed simultaneous analysis of all available evidence (total evidence; Kluge, 1989) instead of running partitioned analyses for different DNA fragments. Data partitioning can only be justified if the partitions are independent of each other (Fitzhugh, 2006a, 2006b), which is not the case with DNA sequences derived from the same plastid genome. Combined data partitions have been found to have higher nodal stability than separate partitions, indicating that the sequence alignment problem is less serious in the combined analysis (Aagesen, 2005). We did perform separate analyses for the different loci to inspect them visually for incongruence that might © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 310 S. LEHTONEN ET AL. Table 2. List of taxa used in this study with GenBank accession numbers and voucher information Taxon Voucher trnL-trnF trnH-psbA rpoC1 rps4 rps4-trnS Cystodium sorbifolium (Sm.) J.Sm. Lindsaea agatii (Brack.) Lehtonen & Tuomisto Lindsaea annamensis K.U.Kramer Lindsaea apoensis Copel. in Perkins Lindsaea arcuata Kunze Lindsaea austrosinica Ching Brass 23810 (BM) GU478726 n.a. n.a. n.a. n.a. Braithwaite 4519 (U) GU478854 GU478523 GU478605 n.a. n.a. Averyanov VH2540 (AAU) Elmer 11565 (U) GU478804 GU478501 GU478612 n.a. n.a. GU478865 GU478563 n.a. n.a. n.a. Lindsaea azurea Christ Lindsaea bifida (Kaulf.) Mett. ex Kuhn Lindsaea bifida (Kaulf.) Mett. ex Kuhn Lindsaea blanda Mett. ex Kuhn Lindsaea blotiana K.U.Kramer 1 Lindsaea blotiana K.U.Kramer 2 Lindsaea bolivarensis V.Marcano Lindsaea borneensis Hook. Lindsaea botrychioides A.St.-Hil. Lindsaea bouillodii Christ Lindsaea brachypoda (Baker) Salomon Lindsaea brevipes Copel. Lindsaea cf. cambodgensis Christ Lindsaea carvifolia K.U.Kramer Lindsaea chienii Ching Lindsaea chrysolepis K.U.Kramer Lindsaea coarctata K.U.Kramer 1 Lindsaea coarctata K.U.Kramer 2 Lindsaea crispa Baker Lindsaea cubensis Underw. & Maxon Lindsaea cultrata (Willd.) Sw. Lindsaea cyclophylla K.U.Kramer Lindsaea digitata Lehtonen & Tuomisto Lindsaea diplosora Alderw. Lindsaea dissectiformis Ching Lindsaea divaricata Klotzsch Maas 5032 (U) Averyanov VH4816 (AAU) Cheesman 165 (U) Hatschbach 18248 (L) GU478788 GU478801 GU478506 GU478486 GU478615 n.a. GU478657 n.a. GU478359 n.a. GU478852 FJ360991 GU478536 FJ360901 n.a. FJ360946 n.a. n.a. n.a. n.a. Lehtonen 552 (TUR) GU478799 GU478485 GU478608 n.a. n.a. Palmer 747 (U) GU478806 GU478525 n.a. n.a. n.a. Rakotondrainibe 6350 (P) Rakotondrainibe 4327 (MO) Tuomisto 14478 (TUR) GU478813 GU478510 GU478633 GU478679 GU478381 GU478814 GU478511 n.a. n.a. n.a. FJ360992 FJ360902 FJ360947 GU478652 GU478354 Wong 2 (AAU) Windisch 4987 (AAU) GU478866 FJ360994 GU478538 FJ360904 n.a. FJ360949 n.a. GU478714 n.a. GU478416 Nielsen 592 (AAU) Balgooy 1529 (L) FJ360993 GU478848 FJ360903 GU478553 FJ360948 n.a. GU478716 n.a. GU478418 n.a. Braithwaite 4029 (MO) Iwatsuki 14505 (AAU) GU478809 GU478840 GU478541 GU478502 GU478627 n.a. n.a. n.a. n.a. n.a. Mjöberg 9 (BM) GU478860 GU478562 n.a. n.a. n.a. Kramer 7554 (U) Croft 216 (L) FJ360995 GU478850 FJ360905 GU478552 FJ360950 n.a. n.a. GU478709 n.a. GU478411 Tuomisto 14500 (TUR) FJ360996 FJ360906 FJ360951 GU478656 GU478358 Plowman 11667 (U) GU478786 GU478507 n.a. n.a. n.a. Parris 11469 (L) Morton 10035 (BM) GU478834 GU478795 GU478526 GU478494 n.a. n.a. n.a. GU478687 n.a. GU478389 Poulsen 128 (AAU) Liesner 23788 (MO) GU478859 GU478776 GU478554 GU478483 n.a. n.a. GU478668 n.a. GU478370 n.a. Tuomisto 14470 (TUR) EU146057 EU146041 EU146051 GU478698 GU478400 Nielsen 815 (AAU) Averyanov VH2557 (AAU) Tuomisto 14553 (TUR) GU478833 FJ360997 GU478535 FJ360907 GU478598 FJ360952 GU478722 n.a. GU478424 n.a. EU146052 EU146040 EU146042 GU478697 GU478399 © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE 311 Table 2. Continued Taxon Voucher trnL-trnF trnH-psbA rpoC1 rps4 rps4-trnS Lindsaea divergens Wall. ex Hook. & Grev. Lindsaea doryphora K.U.Kramer 1 Lindsaea doryphora K.U.Kramer 2 Lindsaea dubia Spreng. Abdullah 228 (U) GU478765 GU478465 n.a. n.a. n.a. Poulsen 341 (AAU) GU478868 GU478539 GU478636 GU478683 GU478385 Magintan 560 (AAU) GU478867 GU478540 n.a. n.a. n.a. Christenhusz 2312 (TUR) Larsen 42363 (AAU) Kessler 13597 (GOET) Tolentino HAL-049 (U) Tuomisto 14925 (TUR) GU478779 GU478493 GU478623 GU478701 GU478403 GU478853 n.a. GU478857 FJ361005 GU478522 GU478521 GU478557 FJ360914 n.a. GU478606 n.a. FJ360960 n.a. GU478678 n.a. GU478702 n.a. GU478380 n.a. GU478404 van der Werff 3418 (U) Streimann 8951 (L) Hotta 12856 (U) FJ361004 FJ360998 GU478820 FJ360913 FJ360908 GU478497 FJ360959 FJ360953 n.a. GU478689 GU478719 n.a. GU478391 GU478421 n.a. Deroin 112 (P) GU478818 n.a. n.a. n.a. n.a. Rakotondrainibe 4090 (MO) Stone 2506 (US) GU478810 GU478500 n.a. n.a. n.a. GU478766 GU478467 GU478594 n.a. n.a. Tuomisto 14927 (TUR) FJ360999 FJ360944 FJ360954 GU478700 GU478402 Wuzhishan fern survey 533 (MO) GU478807 GU478555 GU478604 GU478671 GU478373 Brownlie 978 (TUR) Tuomisto 12742 (TUR) FJ361000 FJ361001 FJ360909 FJ360910 FJ360955 FJ360956 GU478673 GU478660 GU478375 GU478362 Moran 3637 (AAU) GU478781 GU478487 n.a. GU478707 GU478409 Tuomisto 13049 (TUR) FJ374265 FJ374263 FJ374264 GU478717 GU478419 Charoenphol 5081 (AAU) Christenhusz 2826 (TUR) Rojas 2350 (BM) GU478855 n.a. GU478607 n.a. n.a. GU478772 GU478435 GU478592 GU478649 GU478351 n.a. GU478436 n.a. n.a. n.a. Blake 21210 (MO) Nielsen 542 (AAU) Poulsen 72 (AAU) Clemens 30733 (BM) Hennipman 3937 (U) Poulsen 1714 (TUR) Prance 15677 (U) GU478856 GU478822 GU478821 GU478764 FJ361002 FJ361003 GU478792 GU478520 GU478532 GU478531 GU478466 FJ360911 FJ360912 GU478490 n.a. n.a. n.a. n.a. FJ360957 FJ360958 n.a. GU478682 GU478669 GU478670 GU478711 n.a. n.a. n.a. GU478384 GU478371 GU478372 GU478413 n.a. n.a. n.a. Saiki 1093 (Z) Braithwaite 4304 (U) Tuomisto 14466 (TUR) GU478803 GU478831 FJ361006 GU478499 n.a. FJ360915 GU478596 GU478629 FJ360961 n.a. GU478681 GU478651 n.a. GU478383 GU478353 Lindsaea ensifolia Sw. 1 Lindsaea ensifolia Sw. 2 Lindsaea cf. fissa Copel. Lindsaea falcata (Dryand.) Rosenst. Lindsaea feei C.Chr. Lindsaea fraseri Hook. Lindsaea gomphophylla Baker Lindsaea goudotiana (Kunze) Mett. ex Kuhn Lindsaea grandiareolata (Bonap.) K.U.Kramer Lindsaea gueriniana (Gaudich.) Desv. Lindsaea guianensis (Aubl.) Dryand. Lindsaea hainaniana (K.U.Kramer) Lehtonen & Tuomisto Lindsaea harveyi Carrière Lindsaea hemiglossa K.U.Kramer 1 Lindsaea hemiglossa K.U.Kramer 2 Lindsaea hemiptera K.U.Kramer Lindsaea heterophylla Dryand. Lindsaea imrayana (Hook.) Perez 1 Lindsaea imrayana (Hook.) Perez 2 Lindsaea incisa Prent. Lindsaea integra Holttum 1 Lindsaea integra Holttum 2 Lindsaea jamesonioides Baker Lindsaea javanensis Blume 1 Lindsaea javanensis Blume 2 Lindsaea javitensis Humb. & Bonpl. ex Willd. Lindsaea kawabatae Kurata Lindsaea kingii Copel. Lindsaea lancea (L.) Bedd. var. lancea © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 312 S. LEHTONEN ET AL. Table 2. Continued Taxon Voucher trnL-trnF trnH-psbA rpoC1 rps4 rps4-trnS Lindsaea lancea (L.) Bedd. var. leprieurii (Hook.) K.U.Kramer 1 Lindsaea lancea (L.) Bedd. var. leprieurii (Hook.) K.U.Kramer 2 Lindsaea lancea (L.) Bedd. var. submontana Boudrie & Cremers Lindsaea lapeyrousei (Hook.) Baker 1 Lindsaea lapeyrousei (Hook.) Baker 2 Lindsaea leptophylla Baker Maas 2296 (U) GU478773 GU478481 n.a. GU478706 GU478408 Christenhusz 2552 (TUR) GU478780 GU478504 GU478617 GU478654 GU478356 Granville 14959 (CAY) GU478774 GU478480 GU478614 n.a. n.a. Hadley 201 (BM) n.a. GU478556 GU478626 n.a. n.a. Brownlie 1575 (U) FJ361007 FJ360916 FJ360962 GU478680 GU478382 GU478819 GU478509 n.a. n.a. n.a. GU478784 FJ361008 FJ361009 GU478837 GU478808 GU478505 FJ360917 FJ360918 GU478527 GU478542 n.a. FJ360963 FJ360964 n.a. GU478628 GU478718 GU478721 GU478674 n.a. n.a. GU478420 GU478423 GU478376 n.a. n.a. GU478815 GU478512 GU478634 n.a. n.a. malayensis Holttum Raharimalala 2017 (MO) Bernard 4775 (P) Braggins 589 (AAU) Alston 16699 (U) Elmer 16151 (U) Charoenphol 4999 (AAU) Rakotondrainibe 6349 (P) Tagawa 4781 (U) GU478836 GU478545 n.a. n.a. n.a. malayensis Holttum Schneider 222 (Z) GU478858 GU478530 GU478602 GU478672 GU478374 media R.Br. 1 media R.Br. 2 Gittins s.n. (AAU) van der Werff 11655 (MO) Liesner 7062 (MO) GU478842 GU478843 GU478515 GU478516 n.a. GU478597 n.a. n.a. n.a. n.a. GU478783 GU478478 n.a. n.a. n.a. Elmer 13488 (U) GU478864 GU478558 n.a. n.a. n.a. Kramer 8005 (U) GU478863 GU478559 n.a. n.a. n.a. Everist 8062 (AAU) Lan 5065 (L) FJ361010 GU478812 FJ360919 GU478513 FJ360965 n.a. n.a. n.a. n.a. n.a. Guillaumet 4206 (P) GU478811 n.a. n.a. n.a. n.a. Edwards 4414A (L) FJ361011 FJ360920 FJ360966 n.a. n.a. Cicuzza 104 (GOET) McKee 2617 (U) Larsen 42928 (U) GU478849 GU478829 FJ361012 GU478534 n.a. FJ360921 GU478631 n.a. FJ360967 GU478676 n.a. GU478685 GU478378 n.a. GU478387 Parris 11099 (AAU) GU478824 GU478548 GU478632 n.a. n.a. Kramer 8004 (U) GU478823 GU478547 GU478599 n.a. n.a. Averyanov VH4814 (AAU) FJ361013 FJ360922 FJ360968 n.a. n.a. Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea lherminieri Fée linearis Sw. lobata Poir. in Lam. longifolia Copel. lucida Blume Lindsaea Baker Lindsaea 1 Lindsaea 2 Lindsaea Lindsaea madagascariensis Lindsaea meifolia (Kunth) Mett. ex Kuhn Lindsaea merrillii Copel. ssp. merrillii Lindsaea merrillii Copel. ssp. yaeyamensis (Tagawa) K.U.Kramer Lindsaea microphylla Sw. Lindsaea millefolia K.U.Kramer 1 Lindsaea millefolia K.U.Kramer 2 Lindsaea monocarpa Rosenst. in C.Chr. Lindsaea multisora Alderw. Lindsaea nervosa Mett. Lindsaea oblanceolata Alderw. Lindsaea obtusa J.Sm. ex Hook. 1 Lindsaea obtusa J.Sm. ex Hook. 2 Lindsaea orbiculata (Lam.) Mett. ex Kuhn © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE 313 Table 2. Continued Taxon Voucher trnL-trnF trnH-psbA rpoC1 rps4 rps4-trnS Lindsaea ovoidea Fée Lindsaea oxyphylla Baker Lindsaea pacifica K.U.Kramer Lindsaea pallida Klotzsch Hatschbach 7083 (L) Gautier 2662 (P) Brownlie 979 (U) FJ361014 GU478816 FJ361015 FJ360923 GU478508 FJ360924 FJ360969 n.a. FJ360970 GU478713 n.a. GU478663 GU478415 n.a. GU478365 Christenhusz 2518 (TUR) Poulsen 342 (AAU) GU478778 GU478479 GU478624 GU478705 GU478407 GU478851 GU478533 n.a. GU478675 GU478377 Larsen 42927 (AAU) FJ361016 FJ360925 FJ360971 GU478684 GU478386 Boudrie 4254 (TUR) GU478775 GU478492 GU478622 GU478695 GU478397 Nielsen 816 (AAU) Kjellberg 3561 (BM) Kramer 2891 (U) Tuomisto 14940 (TUR) Brownlie 996 (U) FJ361021 GU478767 GU478782 FJ361017 FJ361018 FJ360930 GU478468 GU478477 FJ360926 FJ360927 FJ360976 n.a. n.a. FJ360972 FJ360973 GU478686 n.a. n.a. GU478659 GU478666 GU478388 n.a. n.a. GU478361 GU478368 Callmander 291 (P) Price 2789 (Z) GU478768 GU478838 GU478471 GU478544 GU478593 n.a. GU478661 n.a. GU478363 n.a. Christenhusz 3384 (TUR) Tuomisto 13045 (TUR) Croat 90913 (MO) Mackee 6609 (P) Wood 10051 (BISH) FJ361019 FJ360928 FJ360974 n.a. n.a. FJ361028 GU478771 GU478830 GU478825 FJ360937 GU478469 GU478519 GU478537 FJ360983 n.a. n.a. GU478601 GU478692 n.a. n.a. GU478662 GU478394 n.a. n.a. GU478364 Hou 245 (U) GU478861 GU478560 n.a. n.a. n.a. Alston 16692 (U) GU478805 GU478528 n.a. GU478723 GU478425 Brousilie 1049 (U) GU478826 GU478551 n.a. GU478665 GU478367 Christenhusz 3588 (TUR) FJ361020 FJ360929 FJ360975 GU478653 GU478355 Jaag 1730 (Z) Boudrie 4250 (TUR) Tolentino HAL-026 (U) Poole 2060 (Z) GU478847 GU478777 GU478832 GU478794 GU478564 GU478482 GU478549 GU478476 n.a. GU478618 GU478600 GU478621 n.a. GU478704 GU478710 GU478691 n.a. GU478406 GU478412 GU478393 Williams 14263 (US) GU478793 GU478475 GU478620 n.a. n.a. FJ361023 GU478828 FJ361024 FJ360932 GU478518 FJ360933 FJ360978 n.a. FJ360979 n.a. n.a. GU478688 n.a. n.a. GU478390 schizophylla (Baker) Hoogland 10860 (L) Munzinger 1254 (P) Christenhusz 2618 (TUR) Sledge 613 (Z) GU478802 GU478498 n.a. n.a. n.a. sessilis Copel. sp. 1 sp. 2 Koster 13846 (U) Cicuzza 891 (GOET) Kluge 7230 (GOET) GU478862 GU478846 FJ361022 GU478561 GU478546 FJ360931 n.a. GU478635 FJ360977 n.a. n.a. GU478664 n.a. n.a. GU478366 Lindsaea parallelogramma Alderw. Lindsaea parasitica (Roxb. ex Griffith) Hieron. Lindsaea parkeri (Hook.) Kuhn Lindsaea pectinata Blume Lindsaea pellaeiformis Christ Lindsaea pendula Klotzsch Lindsaea phassa K.U.Kramer Lindsaea pickeringii (Brack.) Mett. ex. Kuhn Lindsaea plicata Baker Lindsaea polyctena K.U.Kramer Lindsaea portoricensis Desv. 1 Lindsaea portoricensis Desv. 2 Lindsaea pratensis Maxon Lindsaea prolongata E.Fourn. Lindsaea propinqua Hook. in Night. Lindsaea pseudohemiptera (Alderw.) Lehtonen & Tuomisto Lindsaea pulchella (J.Sm.) Mett. ex Kuhn Lindsaea pulchra (Brack.) Carrière ex Seem. Lindsaea quadrangularis Raddi ssp. antillensis K.U.Kramer Lindsaea regularis Rosenst. Lindsaea reniformis Dryand. Lindsaea rigida J.Sm. Lindsaea rigidiuscula Lindm. 1 Lindsaea rigidiuscula Lindm. 2 Lindsaea rosenstockii Brause Lindsaea rufa K.U.Kramer Lindsaea sagittata Dryand. Lindsaea Christ Lindsaea Lindsaea Lindsaea © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 314 S. LEHTONEN ET AL. Table 2. Continued Taxon Voucher trnL-trnF trnH-psbA rpoC1 rps4 rps4-trnS Lindsaea sp. 3 Lindsaea schomburgkii Klotzsch Lindsaea seemannii J.Sm. in Seem. Lindsaea semilunata (C.Chr.) C.Chr. Lindsaea sphenomeridopsis K.U.Kramer Lindsaea spruceana Mett. ex Kuhn Lindsaea stolonifera Mett. ex Kuhn Lindsaea stricta (Sw.) Dryand. var. jamesoniiformis K.U.Kramer Lindsaea stricta (Sw.) Dryand. var. stricta Lindsaea stricta (Sw.) Dryand. var. parvula (Fée) K.U.Kramer Lindsaea subtilis K.U.Kramer Laegaard 13835 (AAU) Tuomisto 13044 (TUR) FJ361025 FJ361026 FJ360934 FJ360935 FJ360980 FJ360981 n.a. GU478694 n.a. GU478396 Hammel 3369 (AAU) GU478770 GU478470 n.a. n.a. n.a. Renz 14307 (U) GU478797 GU478489 GU478619 GU478693 GU478395 Renz 14184 (U) FJ361027 FJ360936 FJ360982 GU478708 GU478410 Quipuscoa 878 (AAU) GU478787 n.a. n.a. n.a. n.a. Braithwaite 4283 (U) GU478827 GU478524 n.a. n.a. n.a. Maas 6460 (U) GU478790 GU478473 GU478611 GU478696 GU478398 Boudrie 4246 (TUR) GU478791 GU478472 GU478610 GU478650 GU478352 Boudrie 4253 (TUR) GU478789 GU478474 GU478625 GU478690 GU478392 van der Werff 12855 (MO) Granville 15461 (CAY) Tuomisto 14170 (TUR) GU478817 GU478514 GU478595 n.a. n.a. GU478785 FJ361029 GU478495 FJ360938 GU478630 FJ360984 GU478703 GU478655 GU478405 GU478357 Hartley 12334 (AAU) Maguine 53824 (U) Kato 14212 (MO) FJ361030 GU478798 GU478839 FJ360939 GU478491 GU478543 FJ360985 n.a. GU478603 GU478677 GU478715 n.a. GU478379 GU478417 n.a. Tuomisto 13048 (TUR) FJ361031 FJ360940 FJ360986 GU478699 GU478401 Cameron s.n. (AAU) FJ361032 FJ360941 FJ360987 n.a. n.a. Tuomisto 13002 (TUR) GU478796 GU478488 GU478616 GU478658 GU478360 Bostock 638 (Z) Sledge 1377 (U) GU478841 GU478835 GU478517 GU478550 n.a. n.a. n.a. GU478720 n.a. GU478422 Alston 16860 (U) McKee 5304 (U) Lehtonen 621 (TUR) Crookes s.n. (U) Smith 5918 (U) FJ361034 GU478844 GU478800 GU478769 GU478845 FJ360943 GU478496 GU478484 GU478503 GU478529 FJ360989 n.a. GU478609 n.a. n.a. GU478667 n.a. GU478712 n.a. n.a. GU478369 n.a. GU478414 n.a. n.a. Christenhusz 3508 (TUR) Ballard 1390 (US) GU478725 GU478429 GU478567 GU478640 n.a. GU478757 GU478446 GU478570 n.a. n.a. Kramer 11073 (MO) GU478758 GU478447 n.a. n.a. n.a. Lindsaea surinamensis Posth. Lindsaea taeniata K.U.Kramer Lindsaea tenuifolia Blume Lindsaea tenuis Klotzch Lindsaea tetragona K.U.Kramer Lindsaea tetraptera K.U.Kramer Lindsaea trichomanoides Dryand. Lindsaea ulei Hieron. ex Christ Lindsaea walkerae Hook. Lindsaea venusta Kaulf. ex Kuhn Lindsaea werneri Rosenst. Lindsaea vieillardii Mett. Lindsaea virescens Sw. Lindsaea viridis Col. Lindsaea vitiensis K.U.Kramer Lonchitis hirsuta L. Nesolindsaea caudata (Hook.) Lehtonen & Christenh. Nesolindsaea kirkii (Hook. ex Baker) Lehtonen & Christenh. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE 315 Table 2. Continued Taxon Voucher trnL-trnF trnH-psbA rpoC1 rps4 rps4-trnS Odontosoria aculeata (L.) J.Sm. Odontosoria afra (K.U.Kramer) J.P.Roux Odontosoria africana F.Ballard Odontosoria angustifolia (Bernh.) C.Chr. Odontosoria biflora (Kaulf.) C.Chr. Odontosoria chinensis (L.) J.Sm. Odontosoria deltoidea (Copel.) Lehtonen & Tuomisto 1 Odontosoria deltoidea (Copel.) Lehtonen & Tuomisto 2 Odontosoria flexuosa Maxon Odontosoria fumarioides (Sw.) J.Sm. Odontosoria guatemalensis Christ Odontosoria jenmanii Maxon Odontosoria melleri (Hook.) C.Chr. Odontosoria retusa (Cav.) J.Sm. Odontosoria scandens (Desv.) C.Chr. Odontosoria schlechtendalii (C.Presl.) C.Chr. Odontosoria wrightiana Maxon Osmolindsaea japonica (Baker) Lehtonen & Christenh. Osmolindsaea sp. Osmolindsaea odorata (Roxb.) Lehtonen & Christenh. 1 Osmolindsaea odorata (Roxb.) Lehtonen & Christenh. 2 Osmolindsaea odorata (Roxb.) Lehtonen & Christenh. 3 Pteridium pinetorum C.N.Page & R.R.Mill Saccoloma elegans Kaulf. Saccoloma inaequale (Kunze) Mett. Sphenomeris clavata (L.) Maxon 1 Sphenomeris clavata (L.) Maxon 2 Tapeinidium amboynense (Hook.) C.Chr. Christenhusz 4242 (TUR) Hess-Wyss s.n. (Z) GU478740 GU478463 GU478580 n.a. n.a. GU478738 GU478454 GU478578 n.a. n.a. Lewate 5959 (P) GU478739 GU478455 n.a. n.a. n.a. Sarasin 65 (Z) GU478736 GU478451 n.a. n.a. n.a. Raulerson 5549 (BISH) GU478732 GU478453 GU478586 n.a. n.a. Mäkinen 98-239 (TUR) GU478731 GU478452 n.a. n.a. n.a. Mackee 19315 (P) GU478733 n.a. n.a. n.a. n.a. Buchhofz 1051 (P) GU478734 n.a. n.a. n.a. n.a. Ernst 1944 (BM) Christenhusz 3050 (TUR) Ventura 1090 (BM) GU478742 GU478744 GU478464 GU478461 GU478583 GU478579 GU478648 n.a. GU478350 n.a. GU478746 GU478458 GU478584 n.a. n.a. Zahner 10 (Z) Randriamanarivo 41 (MO) Fallen 620 (MO) GU478747 GU478737 GU478462 GU478456 n.a. n.a. n.a. n.a. n.a. n.a. GU478735 GU478450 n.a. n.a. n.a. Christenhusz 3460 (TUR) Nee 24879 (Z) GU478743 GU478459 GU478581 n.a. n.a. GU478745 GU478457 GU478585 n.a. n.a. Morton 4343 (BM) GU478741 GU478460 GU478582 n.a. n.a. Huang 329 (MO) GU478761 GU478433 GU478590 GU478644 GU478346 Du Puy 2397 (MO) Charoenphol 4708 (AAU) Averyanov VH3287 (AAU) Kramer 5993 (U) GU478763 GU478759 GU478434 GU478430 n.a. GU478588 n.a. GU478647 n.a. GU478349 GU478760 GU478431 GU478589 GU478646 GU478348 GU478762 GU478432 GU478591 GU478645 GU478347 Lehtonen 696 (TUR) GU478724 GU478426 GU478566 GU478637 n.a. Tuomisto 14948 (TUR) Tuomisto 14409 (TUR) GU478728 GU478727 GU478427 GU478428 GU478569 GU478568 GU478639 GU478638 n.a. n.a. Christenhusz 4218 (TUR) Axelrod 5550 (MO) GU478729 GU478448 GU478587 n.a. n.a. GU478730 GU478449 n.a. n.a. n.a. Dayton 26238 (BISH) GU478753 GU478441 GU478573 GU478642 GU478344 © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 316 S. LEHTONEN ET AL. Table 2. Continued Taxon Voucher trnL-trnF trnH-psbA rpoC1 rps4 rps4-trnS Tapeinidium calomelanos K.U.Kramer Tapeinidium denhamii (Hook.) C.Chr. Tapeinidium gracile Alderw. Tapeinidium longipinnulum (Ces.) C.Chr. Tapeinidium luzonicum (Hook.) K.U.Kramer Tapeinidium melanesicum K.U.Kramer Tapeinidium moorei (Hook.) Hieron. Tapeinidium novoguineense K.U.Kramer Tapeinidium pinnatum (Cav.) C.Chr. Price 1305 (Z) GU478755 GU478442 GU478575 n.a. n.a. Zogg 9266 (Z) GU478748 GU478440 GU478571 n.a. n.a. Kluge 7058 (GOET) Holttum s.n. (Z) GU478750 GU478754 GU478437 GU478439 GU478577 GU478574 n.a. GU478641 n.a. GU478343 Kessler 13610 (GOET) GU478751 GU478444 GU478576 GU478643 GU478345 Webster 14055 (BISH) GU478749 GU478443 GU478572 n.a. n.a. Mackee 196 (AAU) GU478756 GU478445 n.a. n.a. n.a. Brass 23048 (BM) GU478752 GU478438 n.a. n.a. n.a. Kessler 13613 (GOET) FJ360990 FJ360900 FJ360945 n.a. n.a. During the course of this study we also produced a trnL-trnF sequence (GenBank accession GU478565) of L. diplosora (voucher Wong 2439, L), but this specimen was not included in the analyses, as we failed to produce other sequences from it, and the trnL-trnF sequence was identical with the other L. diplosora sequence we used. n.a., not available. result from contamination or copy–paste errors during data manipulation. PHYLOGENETIC ANALYSES Our data set included three sequence anomalies which are most parsimoniously explained as short inversions. These were the 13-bp inversion in trnHpsbA spacer present in Sphenomeris clavata (L.) Maxon and Lindsaea linearis Sw., the 7-bp inversion frequently observed c. 200 bp upstream of the trnF gene in various Lindsaea and Odontosoria spp. and the 4-bp inversion located c. 350 bp upstream of the trnF gene in Lindsaea tenuifolia Blume and L. polyctena K.U.Kramer. The apparently inverted sequence fragments were replaced with their complement sequences in the sequence data and an additional binary character was coded for each inversion event to record the original orientation of the inverted sequence. It would be preferable to optimize the inversions directly under the framework of dynamic homology (Catalano, Saidman & Vilardi, 2009), but the current algorithms (Vinh, Varón & Wheeler, 2006; Wheeler, 2007) do not recognize this kind of minute inversions. The inversion events were mapped with delayed transformation (DELTRAN) optimization on the consensus trees to minimize the reversals and interpret character-state changes as convergences whenever possible (Swofford & Maddison, 1987; for detailed discussion, see Agnarsson & Miller, 2008). Character evolution was investigated with MacClade (Maddison & Maddison, 2000). We evaluated the robustness of the resulting phylogenetic hypotheses by calculating both nodal support and nodal stability measures. Nodal support indicates how well the data in question supports a given node, whereas nodal stability specifies how sensitive the node is to variation in analytical assumptions (Giribet, 2003). We obtained nodal support values by jackknifing the implied POY alignments (Farris et al., 1996; Giribet, 2005). Bremer support values (Bremer, 1988) were not used, because they would have been biased by the fact that we do not have 100% character sampling (DeBry, 2001). As the optimal topology is, by definition, the hypothesis that conforms with the data best (Fitzhugh, 2006b), we do not see any logical reason to collapse ‘poorly supported’ nodes to polytomies. Instead, the jackknife values are shown on the topologies to draw attention to areas where the results are most dependent on specific assumptions and can most easily be challenged. We evaluated the stability of our hypothesis by applying sensitivity analysis (Wheeler, 1995), i.e. by varying the assumptions made about transformation costs. The phylogenetic trees we present are based on equal transformation costs for all character-state changes, which maximizes the explanatory power of the hypothesis (Frost et al., 2001; Giribet, Edgecombe © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE Phylogenetic analyses were run in a parallel environment of 2 ¥ 2.26 GHz Quad-Core Intel Xeon Macintosh, using either POY (for molecular data, combined analysis of molecular and morphological data and sensitivity analyses; Varón, Vinh & Wheeler, 2010) or TNT (for morphological data analysis and jackknife calculations; Goloboff, Farris & Nixon, 2008). POY searches were conducted by a combination of 500 random addition starting trees with subtree pruning and reconnection (SPR) and tree bisection–reconnection (TBR) branch swapping, and evaluating all the suboptimal trees within 5% of the best cost, followed by 20 rounds of ratcheting (Nixon, 1999), 200 rounds of tree fusing (Goloboff, 1999) and final swapping with SPR and TBR. Sensitivity analyses were principally similar, but only 20 random addition starting trees were built, no suboptimal topologies were evaluated during TBR swapping and only 10 rounds of ratcheting were performed. Furthermore, in sensitivity analyses, the time allocated for each swapping round was limited to 3600 s for SPR, TBR and tree fusing and 20 000 s for ratcheting. The results of the sensitivity analyses were investigated with the aid of Cladescan (Sanders, 2010). In the morphological data analysis, we applied a search with 500 random addition starting trees swapped with ratcheting and holding 10 trees per replicate. Morphological characters were equally weighted and treated as non-additive (unordered) throughout all the analyses. For jackknife calculations, 100 pseudoreplicates with a deletion probability of e–1 were performed and, for each pseudoreplicate, 100 random addition starting trees were built and swapped with ratcheting. The command lines applied are given in Appendix 5. The numbers of potentially parsimony informative characters were calculated in PAUP* 4.0b10 (Swofford, 2002) using the implied alignments and treating a gap as a fifth character state. 4 RESULTS 1 2 SAMPLING 0.5 transversion:transition & Wheeler, 2001; Terry & Whiting, 2005a, 2005b; Espinasa, Flick & Giribet, 2007; Whiting et al., 2008). However, equal costs may lead to illogical, non-metric results and actually cannot be applied when ‘true’ indel lengths are variable (Giribet & Wheeler, 2007), and some studies have preferred non-equal costs (e.g. Laamanen et al., 2005; Sørensen, Sterrer & Giribet, 2006; Boyer & Giribet, 2007; Giribet & Wheeler, 2007; Kutty et al., 2007; Lindgren & Daly, 2007). To test the robustness of the equal-cost hypothesis, we also applied another 15 transformation cost regimes. Theoretically, the available parameter space is infinite, but in practice this space can be narrowed down to a reasonable set of parameters. The lower limit for a transversion–transition cost ratio must be 0.5 (otherwise transformation A → C → G would be cheaper than A → G) and the gap cost must be at least one half of the cost of transformations (Wheeler, 1995). If the gap cost reaches a high enough level, the resulting tree starts to reflect the history of gaps more than of substitutions, until substitution information becomes completely overruled by the gaps (Spagna & Álvarez-Padilla, 2008). Simulations and empirical data suggested that the upper limit for a gap cost should not be more than four times the highest nucleotide transformation cost (Spagna & Álvarez-Padilla, 2008). Based on these constraints, we varied the transversion–transition and indel–transversion cost ratio in four increments (0.5, 1, 2, 4) in order to equally sample across the reasonable parameter space and hence analysed 16 different transformation cost regimes (Fig. 2). The separately coded inversion events and, in the total-evidence analysis, the morphological characters, were given the same weight as indels (Wheeler & Hayashi, 1998; Giribet et al., 2002; Schulmeister, Wheeler & Carpenter, 2002; Grazia, Schuh & Wheeler, 2008). 317 0.5 1 2 4 substitution:indel Figure 2. Transformation measure nodal stability. cost regimes applied to SUCCESS Although some morphological characters could not be recorded for all taxa, in practice our taxon sampling was limited by the availability of specimens suitable for DNA extraction. Our final sampling covered 195 terminals representing 175 taxa (as accepted here) or approximately 69% of the taxa (73% of the species) generally accepted in Lindsaeaceae. All genera of the family were included in our analyses, except the monotypic Xyropteris, which did not yield DNA sequences of adequate quality. Of the other genera, the most complete sampling was achieved for Lindsaea (71% of accepted taxa) and the least complete for Tapeinidium (50% of accepted taxa). Some taxa were represented by more than a single specimen. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 318 S. LEHTONEN ET AL. Amplification and sequencing success varied widely among the sampled loci, such that trnL-trnF sequences were obtained for 98% of the sampled taxa but rps4-trnS sequences for only 43% (Table 3). The proportion of missing data at the data set level (195 taxa ¥ 7 sequences) was approximately 33% and, at the character level (195 taxa ¥ 3636 aligned characters), c. 25%. Sequence lengths were 567–580 bp for rps4, 231– 295 bp for the rps4-trnS spacer, 378–534 bp for the trnL-trnF spacer and 426–495 bp for the trnH-psbA spacer. The sequenced rpoC1 region was the same length in all species (732 bp), but the length of the POY implied alignment was 734 bp. The number of potentially parsimony informative characters in the implied alignments varied from 186 in rpoC1 to 589 in the trnL-trnF spacer. PHYLOGENETIC HYPOTHESES Molecular data When analysed under equal transformation costs, the molecular data resulted in a single most parsimonious tree of 5483 steps (Figs 3–7). This tree resolved Lonchitis, Saccoloma and Cystodium as the first branching lineages of the ingroup, with Cystodium being sister to all traditional lindsaeoid genera. Within the lindsaeoids, Sphenomeris clavata formed the first diverging lineage. All other species previously placed in Sphenomeris were mixed with Odontosoria in the next diverging lineage. The next clade brought together species that have previously not been associated with each other. Tapeinidium is a well-established genus, but it was found to be a sister to a group of species previously placed in Lindsaea; here, these are treated as two new genera, Nesolindsaea and Osmolindsaea. All other Lindsaea spp. formed a single clade. Core Lindsaea is discussed here by using informal clades numbered from I to XIII. The first diverging lineage within core Lindsaea (clade I) is composed of two morphologically unusual species from Central America and northern South America. The six species of clade II are morphologically diverse, have great variation in the studied sequences and are distributed from Madagascar through Melanesia to New Zealand. Clade III consists of a single species from Australia, L. incisa Prent., which is rather distinct from other species based on the studied sequences. Clade IV consists of species distributed from Australia through Melanesia to South-East Asia. The species in clade V are true epiphytes, an unusual habit within Lindsaea. Clade VI consists of a large number of species from the Melanesian and Pacific islands. The phylogenetic relationships among clades VII– XIII were resolved only partly, such that clades VII (VIII–IX), X and (XI–XIII) formed a polytomy. Clade VII consists of only two Australasian species. Clades VIII and IX were resolved as sisters. Clade VIII consists of mostly Asian species, but includes L. grandiareolata (Bonap.) K.U.Kramer from Madagascar and L. vieillardii Mett. from New Caledonia. Lindsaea vieillardii was placed as the first branching species in clade VIII with rather low stability and no jackknife support. Clade IX includes many species endemic to Madagascar, but also species with wider distribution in the Palaeotropics. Clade X consists of species endemic to New Caledonia or New Zealand. Clades XI–XIII are strictly Neotropical, and indeed all Neotropical Lindsaea spp. were placed in these clades, except the two species that form clade I. However, we failed to sample L. macrophylla Kaulf., a Neotropical species that morphologically closely resembles some species in clade IX. All species in clade XI are endemic to the Atlantic coastal rain forests of south-eastern Brazil. Clade XII consists of L. imrayana (Hook.) Perez [Ormoloma imrayanum Hook.) Maxon], which differs from the other species by its distinct morphology and marked differences in the studied sequences. Although the position of this lineage was unstable in the sensitivity analysis, it was always placed well within core Lindsaea. Most Neotropical species were placed in clade XIII. Although large, this clade cannot be usefully subdivided because the sequences we used were almost invariable among the species. Consequently, most of the nodes within clade XIII are highly unstable and lack jackknife support. Table 3. The number of taxa sampled for each of the included data sets Sequence range (bp) Implied alignment (bp) PI characters* No. of terminal taxa % missing taxa Morphology Inversion rps4 rps4-trnS trnL-trnF trnH-psbA rpoC1 55 55 55 195 0 3 3 3 192 2 567–580 605 208 87 55 231–295 411 225 83 57 378–534 1056 589 192 2 426–495 772 411 186 5 732 734 186 117 40 *Potentially parsimony-informative characters. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE 319 Sphenomeris Odontosoria Nesolindsaea Osmolindsaea Tapeinidium Lindsaea 10 changes Figure 3. Phylogram showing the topology and branch lengths of the single most parsimonious tree obtained in a POY analysis of molecular data with equal transformation costs. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 320 S. LEHTONEN ET AL. Figure 4. Phylogenetic relationships among Lindsaeaceae based on POY analysis of molecular data with equal transformation costs. DELTRAN optimization of the inversion events is shown on the branches. Filled diamonds: 13-bp inversions in the trnH-psbA spacer from A-rich type to T-rich type. Filled circles: 4-bp inversions in the trnL-trnF spacer from GT-type to CA-type. Squares: 7-bp inversions in the trnL-trnF spacer, open squares indicate inversion from CT-type to GA-type and filled squares the reverse. Jackknife support values are shown above the branches, with values below 50 omitted. The results of sensitivity analysis are shown below the branches. Transformation cost regimes (see Fig. 2) under which a node is resolved as monophyletic is indicated by a black box and unresolved or unsupported relationships are indicated by white box. In the cases where we propose a change in nomenclature, the old taxon names (Kramer, 1957a, 1967a, b, 1970, 1971a, 1972a, b) are shown in parentheses. Continues in Figure 5. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE Figure 5. Continued from Figure 4, continues in Figure 6. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 321 322 S. LEHTONEN ET AL. Figure 6. Continued from Figure 5, continues in Figure 7. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE Figure 7. Continued from Figure 6. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 323 324 S. LEHTONEN ET AL. On average, the nodes in the molecular tree received 56% jackknife support and were resolved as monophyletic groups with nine transformation cost regimes out of the 16 used in the sensitivity analysis. The inversion events were found to be highly homoplastic. According to DELTRAN optimization, it must be assumed that the 4-bp inversion in the trnLtrnF spacer has occurred twice and that the 7-bp inversion has changed twice from the original orientation, and 20 times back to it. Morphological data When analysed under equal transformation costs, the morphological data resulted in three equally parsimonious trees of 466 steps. The strict consensus tree had 116 resolved nodes (Figs 8–11), 23 of which were shared with the molecular tree. The morphological data agreed with the molecular data in placing the traditional lindsaeoid genera in a monophyletic clade that excludes Lonchitis and Cystodium (77% jackknife support) and in recognizing a monophyletic Lindsaea (clades I–XIII of the molecular tree; without jackknife support) and a Neotropical clade within Odontosoria (with 96% jackknife support). Morphological characters were highly homoplastic (consistency index = 0.16, retention index = 0.76), resulting in relatively poor resolution and low average jackknife support (13%). Combined data The simultaneous analysis of molecular and morphological data under equal transformation costs resulted in a single most parsimonious tree with 6117 steps (Figs 12–15). The strict consensus tree had, in total, 184 resolved nodes, and 126 of these were shared with the molecular tree. In broad terms, the topologies obtained with the molecular and combined data sets were similar and the differences were in some species-level groupings within the 13 major clades. For example, the two analyses resolved some sister species relationships in the Odontosoria–Sphenomeris clade differently. Within core Lindsaea, the main clades I–XIII were identical in the two analyses, with the single exception of L. vieillardii which was resolved as sister to the reminder of clade VIII in the molecular analysis, but as sister to clade IX in the combined analysis. The molecular analysis was unable to resolve the divergence order among clades VII, VIII (IX–X) and (XI– XIII). In the combined analysis, this polytomy was resolved, although with no jackknife support and low stability (5/16). The sister species relationships resolved within clades II, V and VI differed markedly between the molecular and combined tree. The nodes of the combined tree were better supported and more stable than those of the molecular tree in the case of clade II. In contrast, the addition of morphological data reduced the overall support and stability of the nodes in clade V and the nodes in clade VI were generally unstable in both trees. The single-species clade III (Lindsaea incisa) was in the combined analysis resolved as sister to clades IV–XIII, whereas in the molecular tree it had a more derived position as sister to clades IV–VI. Although clade XIII contains the same taxa in the molecular and combined tree, its internal topology was resolved quite differently by the two analyses. Both analyses recognized the same monophyletic group, with L. reniformis Dryand. as sister to the remaining species, but otherwise both topologies were highly unstable and poorly supported. The combined tree explained the inversions more parsimoniously than the molecular tree. The distribution of the 4-bp inversions in the trnL-trnF spacer can be explained in the combined tree by a single change. This inversion was only observed in L. tenuifolia and L. polyctena of clade VI, which were placed as sisters in the combined tree. The 7-bp inversion can be explained by assuming two changes from the original orientation and 19 changes back to it. Despite the high levels of homoplasy and low average jackknife support in the morphological analysis, the combined analysis resulted in a much higher average jackknife support (65%) than the molecular analysis. Average nodal stability (8/16) was slightly lower, however. DISCUSSION RELIABILITY OF THE PHYLOGENETIC RESULTS In general, our taxon sampling (c. 73% of the species) can be considered sufficient for reliable phylogeny estimation (e.g. Zwickl & Hillis, 2002). Nevertheless, in some groups (e.g. Tapeinidium) our sampling remained rather poor and our inability to sequence all loci from all taxa may compromise the reliability of some groupings. The DNA sequences of some closely related taxa were so similar that the corresponding phylogenetic hypotheses are tentative at best. However, the comparison of molecular and morphological evidence in the light of nodal support and stability measures strengthens the discussion on evolutionary relationships. Although Lonchitis and Cystodium have traditionally not been considered lindsaeoid genera, nowadays they are usually included in Lindsaeaceae (Schuettpelz, Korall & Pryer, 2006; Smith et al., 2006; Schuettpelz & Pryer, 2007). Our analyses placed Cystodium close to the lindsaeoids, but Lonchitis branched off from the lindsaeoid lineage even before Saccoloma did. The phylogenetic position of Lonchitis was well supported and stable in our analyses, but that of Cystodium was unstable, and © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE 325 Figure 8. Strict consensus of the three equally parsimonious trees obtained in the analysis of the morphological data. The thick branches represent nodes that were also supported by the molecular data (Figs 4–7). Jackknife support values are shown above the branches, with values below 50 omitted. Old taxon names are shown in parentheses. Continues in Figure 9. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 326 S. LEHTONEN ET AL. Figure 9. Continued from Figure 8, continues in Figure 10. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE Figure 10. Continued from Figure 9, continues in Figure 11. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 327 328 S. LEHTONEN ET AL. Figure 11. Continued from Figure 10. several transformation cost regimes in the molecular and combined analyses resolved it as sister to Saccoloma. Unfortunately, the only DNA sequence that we managed to obtain from Cystodium was the trnL-trnF intergenic spacer. This, together with the fact that our sampling included only two species of Saccoloma and one dennstaedtioid, renders our results on the phylogenetic position of Cystodium and Lonchitis only tentative. Our analyses strongly support their monophyly of the traditional lindsaeoid genera. The lindsaeoid clade obtained 77% jackknife support in the morphological analysis and 100% jackknife support in the DNA and total-evidence analyses. In the DNA and total-evidence analyses, the clade was also present under all transformation cost regimes. High branch support or stability values as such, however, cannot guarantee that the correct phylogeny has been inferred. Our molecular sampling was based on plastid sequences and the obtained tree can be considered a good estimate of plastid evolution, but plastid phylogenies may in some cases drastically differ from those inferred from nuclear markers (e.g. Jakob & Blattner, 2006). Incongruence can result from hybridization as well as from incomplete lineage sorting of ancient polymorphisms (Funk & Omland, 2003). A wide sampling of independent loci and multiple individuals per species is required to detect such patterns (Maddison & Knowles, 2006). It is generally assumed that incomplete lineage sorting complicates the inference of recently diverged lineages only, but ancient plastid © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE 329 Figure 12. Phylogenetic relationships among Lindsaeaceae based on combined POY analysis of morphological and molecular data under equal transformation costs. DELTRAN optimization of the inversion events is shown on the branches. Filled diamonds: 13-bp inversions in the trnH-psbA spacer from A-rich type to T-rich type. Filled circles: 4-bp inversions in the trnL-trnF spacer from GT-type to CA-type. Squares: 7-bp inversions in the trnL-trnF spacer, open squares indicate inversion from CT-type to GA-type and filled squares the reverse. Jackknife support values are shown above the branches, with values below 50 omitted. The results of sensitivity analysis are shown below the branches. Transformation cost regimes (see Fig. 2) under which a node is resolved as monophyletic is indicated by a black box, unresolved or unsupported relationships are indicated by a white box. Old taxon names are shown in parentheses. Continues in Figure 13. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 330 S. LEHTONEN ET AL. Figure 13. Continued from Figure 12, continues in Figure 14. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE Figure 14. Continued from Figure 13, continues in Figure 15. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 331 332 S. LEHTONEN ET AL. Figure 15. Continued from Figure 14. haplotypes may persist for millions of years, potentially also affecting deeper phylogenetic levels (Jakob & Blattner, 2006). In the present phylogenetic analysis of Lindsaeaceae, these problems may be responsible for the observed polyphyly of species and lack of support and stability, especially within the possibly rapidly radiated South American clade XIII. Our revised taxonomy, however, is supported by both molecular evidence and morphological studies of herbarium specimens. REVISED GENERIC CLASSIFICATION OF LINDSAEACEAE The phylogenetic evidence presented above indicates that the prevailing generic limits among the traditional lindsaeoid genera are artificial. Three of the genera as currently circumscribed are either para- or polyphyletic: Sphenomeris, Odontosoria and Lindsaea itself. Tapeinidium and Ormoloma are monophyletic, but Ormoloma is so deeply embedded within Lind- © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE 333 Table 4. Morphological characters supporting the six lindsaeoid genera in the total evidence analysis Sphenomeris Odontosoria Osmolindsaea Nesolindsaea Tapeinidium Lindsaea 8. INDUMENT, NUMBER OF CELL SERIES AT THE BASE: 2–3 20. ARCHITECTURE: flexuose 50. NUMBER OF SPORANGIA/ SORUS: few* 15. LENGTH: short 25. VEIN ORDERS: three 23. TEXTURE: coriaceous* 33. SHAPE: equal sided 36. SEGMENTS: connected by wing* 15. LENGTH: short* 16. RACHIS ABAXIALLY: keeled* 21. NATURE OF SULCUS: broad* 25. VEIN ORDERS: two* Numbers refer to character numbers in Appendix 1. Characters supporting a genus but variable within it are marked with *. saea that maintaining it as a separate genus is not justified. Several species traditionally placed in Lindsaea are closely related to Tapeinidium. One of these is sister to the Tapeinidium clade and also fits morphologically in this rather uniform genus. The other species are morphologically so unlike Tapeinidium that they are more practically separated into two new genera, which are also highly differentiated based on the studied DNA sequences. Here we describe the new genera Osmolindsaea and Nesolindsaea and provide revised circumscriptions for the old ones. SPHENOMERIS MAXON, NOM. CONS. Synonym: Stenoloma Fée, nom. rej. Typus generis: Sphenomeris clavata (L.) Maxon Taxonomic history: The typification of Sphenomeris has caused some confusion. Originally Fée (1852) described two genera, Odontosoria with one species (O. uncinella (Kunze) Fée) and Stenoloma with ten species. Farwell (1931) designated Stenoloma clavata (L.) Fée as the type of the genus. Probably not aware of this earlier typification, Pichi-Sermolli (1953) typified Stenoloma with S. dumosa Fée (= Odontosoria aculeata (L.) J.Sm.). The generic name Sphenomeris of Maxon (1913) was later conserved against the older Stenoloma (Lanjouw et al., 1956). Following Morton’s (1959) proposal, the Committee for Pteridophyta (1959) changed the typification from S. dumosa to S. clavata, in order to keep Sphenomeris in the list of conserved names and to follow the original generic description more closely. Kramer (1971a) criticized the re-typification and argued that S. clavata does not fit in the original generic description. However, the generic typification by Farwell (1931) makes the later typification and re-typification superfluous. Morphological characters used to circumscribe Sphenomeris and Odontosoria were found to be unsatisfactory by Kramer (1972a) and Sphenomeris has sometimes been treated as a synonym of Odontosoria (Tryon & Tryon, 1982; Kramer & Green, 1990). However, both generic names are still in use (e.g. Smith et al., 2006; Schuettpelz & Pryer, 2008). Barcelona (2000) suggested that three new genera should be recognized in the Sphenomeris–Odontosoria group, but did not formally describe them. In our analyses, most species of the Sphenomeris–Odontosoria group formed a monophyletic clade, but the type species of Sphenomeris (S. clavata) was placed in a clade of its own as sister to all other lindsaeoids. Therefore, we recognize Sphenomeris in a narrow sense and transfer most species to a broadly defined Odontosoria. According to this circumscription, Sphenomeris is endemic to the Neotropics. Sphenomeris clavata is distributed from southern Florida to the Bahamas and the Greater Antilles (Kramer, 1957a). Two similar species (S. spathulata (Maxon) K.U.Kramer and S. killipii (Maxon) K.U.Kramer) have been described from Colombia (Maxon, 1947; Kramer, 1957a), but both are apparently only known from the type material. We tentatively accept these species as members of Sphenomeris, but their status needs to be verified. In the total evidence tree, Sphenomeris has two morphological apomorphies: less than three scale cells in the rhizome scales and flexuose rachises (Table 4). These characters are not unique to Sphenomeris, so it is nested within Odontosoria in our morphological tree. However, Sphenomeris is quite distinct based on the studied DNA sequences and terminates a long lineage. The separation of Sphenomeris and other lindsaeoid genera is well supported, although not fully stable. We failed to amplify the most conservative locus (rps4) for Sphenomeris, so we cannot rule out the possibility of long-branch attraction towards © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 334 S. LEHTONEN ET AL. the long root node. Sphenomeris clavata is reported to have a chromosome number N = 38 (Walker, 1966). This differs from other chromosome counts in Lindsaeaceae and provides additional support for the evolutionary distinctness of Sphenomeris. ODONTOSORIA FÉE Synonym: Lindsayopsis Kuhn Typus generis: Odontosoria uncinella Fée = O. scandens (Desv.) C.Chr. (Kunze) Taxonomic history: The name Odontosoria was first used for a section of Davallia by Presl (1836). Fée (1852) used Odontosoria at the generic rank, basing it on the single species O. uncinella, which was not published until 1850, and was therefore not known to Presl. Kuhn (1882) wrongly credited the genus Odontosoria to Presl and described a new genus Lindsayopsis for O. scandens, hence providing a synonym for Odontosoria. Diels (1899) also accredited Odontosoria to Presl, but recognized two sections in the genus, ‘Eu-Odontosoria’ for rather small plants with erect habit and determinate growth and ‘Stenoloma’ for scandent plants with indefinite growth. This synonymized the two genera (Odontosoria and Stenoloma) proposed by Fée (1852). Maxon (1913) decided that both sections need to be recognized at the generic level, but stated that the name Odontosoria is to be applied to the scandent group based on the type species and description by Fée. He clearly stated that O. chinensis (L.) J.Sm. and S. clavata have no place in the genera Lindsaea, Schizoloma or Odontosoria, and he therefore proposed the genus Sphenomeris, with S. clavata as the type. Tryon & Tryon (1982) and Kramer & Green (1990) considered Sphenomeris as congeneric with Odontosoria. The thesis of Barcelona (2000) provided a good overview of the genus Odontosoria, but proposed a generic classification that is not well founded or effectively published (article 30.5 of the ICBN; McNeill et al., 2006). Most of the species that we include in Odontosoria have already been placed in this genus by some earlier author. Here we propose just one new combination. ODONTOSORIA DELTOIDEA (C.CHR.) LEHTONEN & TUOMISTO, COMB. NOV. Basionym: Lindsaea deltoidea C.Chr., Index Filic. 393, 1906; based on Lindsaea elongata Labillardière, Sert. Austro-Caled. 6, t. 9, 1824, non Cavanilles, 1802. Type: New Caledonia, Labillardière s.n. (holotype P, P00633680!, isotype P, P00633679!). Notes: Odontosoria as circumscribed here is a pantropical genus. All Neotropical Odontosoria included in our analyses were resolved into a single wellsupported and highly stable clade. The first division within this clade separates the spineless O. schlechtendalii (C.Presl) C.Chr. and O. guatemalensis Christ from the rest of the Neotropical species, which are all spiny. This sister clade relationship is supported by all analyses, including the one based on morphology only. Tryon & Tryon (1982) considered O. wrightiana Maxon as a narrow-segmented form of O. aculeata (L.) J.Sm., but our analyses do not support this view. We were unable to obtain molecular data from three Neotropical species, which all share morphological synapomorphies with the Neotropical clade and are assumed to belong to it (O. colombiana Maxon, O. gymnogrammoides Christ and the recently described O. reyesii Caluff). The Neotropical clade is sister to a clade that includes O. chinensis, which is distributed throughout the Palaeotropics (except continental Africa) and is a well-known example of a species complex with various levels of polyploidy and hybridization and probably several cryptic species (Lin, Kato & Iwatsuki, 1994, 1996). The two African species (O. afra (K.U.Kramer) J.P.Roux and O. africana F.Ballard) are morphologically highly similar, but differ in the angle between primary and secondary divisions, which is straight in O. africana and acute in O. afra (Kramer, 1971b). The two had identical DNA sequences, however, and they may be conspecific. The third main clade within Odontosoria includes species from Madagascar (O. melleri (Hook. ex Baker) C.Chr.), Malesia (O. retusa (Cav.) J.Sm.) and New Caledonia [O. angustifolia (Bernh.) and O. deltoidea]. The New Caledonian species are resolved as sisters with high support and stability. The DNA sequencing failed for another New Caledonian endemic, Sphenomeris alutacea (Mett.) Copel., but based on morphological characters it most likely belongs to this clade as well. The combined analysis resolved O. retusa and O. melleri as sisters with high support and stability, unlike the molecular analysis in which they formed successive lineages in a poorly supported and unstable grouping. Odontosoria is rather difficult to define morphologically. In the total evidence analysis, the genus was supported by the number of sporangia per sorus being small, but this character is not shared by all species (Table 4). Cytological data have played an important role in the taxonomy of the O. chinensis complex (Lin, Kato & Iwatsuki, 1990; Lin et al., 1994, 1996). Odontosoria biflora has N = 48, whereas in O. chinensis both diploid (N = 48) and tetraploid (N = 96) populations are known (Lin et al., 1990, 1994). Small deviations from these basic © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE 335 Figure 16. Cross sections of lindsaeoid rhizomes: A, solenostele with an internal sclerified pith (Osmolindsaea odorata, Ohba 3374, U); B, solenostele (Nesolindsaea caudata, Comanor 1077, MO); C, ‘Lindsaea-type’ protostele (Lindsaea ovoidea, Lehtonen 558, TUR). numbers also occur and sometimes segregate species have been recognized as a result (Lin et al., 1994, 1996). Both Neotropical species with available chromosome counts (O. fumarioides (Sw.) J.Sm. and O. jenmanii Maxon) are clearly tetraploid (N ª 96; Walker, 1966), but N ª 88 has been reported for O. deltoidea and O. retusa of the third Odontosoria clade (Brownlie, 1965; Walker in Kramer, 1971a). OSMOLINDSAEA (K.U.KRAMER) LEHTONEN & CHRISTENH., GEN. & STAT. NOV. Basionym: Lindsaea Dryand. ex Sm. section Osmolindsaea K.U.Kramer, Blumea 15: 560 (1967). Typus generis: Osmolindsaea odorata (Roxburg) Lehtonen & Christenh. Distinction: The spores of Osmolindsaea are monolete and the rhizomes have a solenostele with an internal sclerified pith (Fig. 16A). In the total evidence analysis, the genus is supported by having short petioles (Table 4). Taxonomic history: Kramer (1967b, 1971a) treated Osmolindsaea as a section of Lindsaea. However, the clade including Osmolindsaea is sister to Tapeinidium and synonymizing Tapeinidium with Lindsaea would result in a genus that is morphologically difficult to define. Therefore, we describe Osmolindsaea here as a new genus of Lindsaeaceae. The name refers to the scent of the fresh leaves, ‘osme’ meaning fragrance in Greek. The following new combinations are proposed here: Osmolindsaea odorata (Roxb.) Lehtonen & Christenh., comb. nov. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 336 S. LEHTONEN ET AL. Basionym: Lindsaea odorata Roxb., Calcutta Journal of Natural History 4: 511 (1846). Type: A plant from the Garrow Hills, India; no specimen extant. Lectotype (designated by Kramer, 1967b): in W. Roxburgh, Icones Roxburghianae (K!). Osmolindsaea japonica (Baker) Lehtonen & Christenh., comb. nov. Basionym: Lindsaea cultrata (Willd.) Sw. var. japonica Baker, Synopsis Filicum 1st ed. 105 (1868). Synonym: Lindsaea japonica (Baker) Diels, Die Natürlichen Pflanzenfamilien, I, 4: 221 (1899). Lectotype (designated here): Japan, Nagasaki, growing on large boulder in streams, Oldham 477 (upper plant), ex Herb. Hookerianum (K!, duplicates K!, GH). Note: Osmolindsaea japonica has usually been considered as a variety of L. odorata (Kramer, 1972b), but it is clearly distinct both in morphological characters and based on the studied DNA sequences. Osmolindsaea japonica has continuous sori, entire pinnules, a large terminal segment and sparsely distributed, minute rhizome scales. In contrast, O. odorata has interrupted sori, incised pinnules, gradually reduced upper pinnules and densely packed long rhizome scales. Distribution: Osmolindsaea is distributed from East Africa and Madagascar through India and the Malay Peninsula, north to Japan and Korea and east to the Solomon Islands. Notes: Lindsaea odorata Roxb. var. darjeelingensis T.Sen & U.Sen is similar to O. odorata, but may prove to be a different taxon. The Madagascan specimen included in our study clearly represents an undescribed species, which differs from the other Osmolindsaea spp. in having broad rhizome scales. This species will be described in a more detailed taxonomic revision of Osmolindsaea (S. Lehtonen, H. Tuomisto, G. Rouhan, M. J. M. Christenhusz, unpubl. data). Osmolindsaea spp. have a solenostele with an internal sclerified pith, in contrast to Lindsaea spp., which have a protostele. Osmolindsaea japonica is reported to have N ª 75 (Lin et al., 1990) and O. odorata N = 150 (as L. cultrata, Manton & Sledge, 1954; Mehra & Khanna, 1959). NESOLINDSAEA LEHTONEN & CHRISTENH., GEN. NOV. Diagnosis: Genus novum Lindsaeae simile, a qua rhizomate solenostelam ferente differt; sporae triletae. Synonym: Lindsaea Dryand. ex Sm. section Aulacorhachis K.U.Kramer Typus generis: Nesolindsaea caudata (Hook.) Lehtonen & Christenh. Distinction: Members of Nesolindsaea superficially resemble typical Lindsaea. However, all Lindsaea spp. have protosteles, whereas Nesolindsaea spp. have solenosteles, which is a character they share with Osmolindsaea (Fig. 16B). The spores of Nesolindsaea are trilete, in contrast to the monolete spores in Osmolindsaea, Xyropteris and Tapeinidium. In the total evidence analysis, the genus is supported by three orders of lamina veins (Table 4). The only reported chromosome count is N = 82 for N. caudata (Manton & Sledge, 1954). Taxonomic history: All previous taxonomic works have placed the two species of Nesolindsaea in the genus Lindsaea. Kramer (1972a) discussed the affinity between these two species, but he placed them in different sections on the basis of several distinguishing characters (interrupted vs. continuous sori, pubescent vs. glabrous axes and different insertion of rhizome scales). Kramer (1972b) placed Lindsaea caudata Hook. in the monotypic section Aulacorhachis. Kramer (1972a) placed Lindsaea kirkii Hook. in section Temnolindsaea, but this section is based on L. klotzschiana Moritz (= L. feei C.Chr.), which is a true Lindsaea. The section name Aulacorhachis is not used here at the generic level, because it refers to the furrowed midrib of N. caudata, which is absent in N. kirkii. Because both species are island endemics, we chose the Greek name component ‘neso-’, meaning island. The following new combinations are proposed here: Nesolindsaea caudata (Hook.) Lehtonen & Christenh., comb. nov. Basionym: Lindsaea caudata Hook., Species Filicum, ed. 1: 215 (1846). Type: Ceylon, Adam’s Peak, Walker s.n. (holotype K!). Nesolindsaea kirkii (Hook.) Lehtonen & Christenh., comb. nov. Basionym: Lindsaea kirkii Hook., Synopsis Filicum, ed. 1: 108 (1868). Lectotype (designated here): Seychelles, 1862, Kirk s.n. (K!, duplicate GH). Distribution and biogeography: Nesolindsaea has a highly restricted distribution, with N. caudata being endemic to Sri Lanka and N. kirkii to the granitic islands in the Seychelles. The long branches indicate a great age for the separation of the two species. The Seychelles are continental in origin and separated from mainland India and Sri Lanka during the early Tertiary (Chaubey et al., 1998). ‘Lemurian stepping stones’ during the Eocene–Oligocene may explain the biogeographical relationships between India, Sri © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE 337 Lanka and Madagascar via the Seychelles (Schatz, 1996). Biogeographical links between India, Sri Lanka and the Seychelles have been shown before in various groups, including frogs (Biju & Bossuyt, 2003), Dillenia L. (Hoogland, 1952) and Nepenthes L. (Meimberg et al., 2001), and we hereby provide a novel example from the ferns. The illustration of Nesolindsaea kirkii (as Lindsaya kirkii) from Baker (1871–1875; Plate 28) is reproduced here as Fig. 17. Taxonomic history: This monotypic genus has been found in Borneo and Sumatra and is known from only a few collections. The affinities of X. stortii have been difficult to establish. It was first described as Schizoloma (Alderwerelt van Rosenburgh, 1914), which is now a synonym of Lindsaea, but has also been included in Tapeinidium (Copeland, 1929) and was placed in its own genus by Kramer (1957b) on the basis of a unique combination of characters. TAPEINIDIUM (C.PRESL) C.CHR. Notes: Unfortunately we failed in our attempts to sequence this highly interesting species. Xyropteris is simply pinnate and differs from Tapeinidium in having continuous sori and large auricles at the bases of pinnae. Kramer (1971a) considered the two genera to be closely related. No chromosome counts are available. Basionym: Microlepia subgenus Tapeinidium C.Presl. Synonym: Protolindsaya Copel. Typus generis: Tapeinidium pinnatum (Cav.) C.Chr. Taxonomic history: Tapeinidium was originally described as an infrageneric taxon within Microlepia (Presl, 1851), but was already treated as a genus by Fée (1852), albeit with the incorrect name Wibelia. Kramer (1967c) provided a taxonomic revision of the genus. Notes: Tapeinidium is a genus of approximately 18 species distributed from southern India to Samoa. Our phylogenetic evidence strongly supports the inclusion of Tapeinidium in Lindsaeaceae. This contrasts with the early molecular hypothesis of Wolf (1995), which placed it in the dryopteroid clade. Our circumscription of Tapeinidium includes the New Caledonian endemic T. moorei. This species was considered to belong to Tapeinidium by Hieronymus (1920), but Kramer (1967a) placed it in Lindsaea. Stelar anatomy is variable in Tapeinidium: the large species have a true solenostele (Kramer, 1971a), but the smaller ones, for example T. moorei, have a protostele with an internal sclerotic strand near the leaf bases (Kramer, 1967a). Tapeinidium species are characterized by uninerval (rarely binerval) sori and laterally quite adnate rigid indusia. In our total evidence analysis, the genus was supported by equally sided ultimate segments, connection of the segments and coriaceous lamina (Table 4). Not all species, however, have all these characters. More comprehensive sampling of Tapeinidium is desirable in the future, as we were able to obtain molecular data from only ten species. Cytology of the genus is not well known, but N ª 150 has been reported for T. pinnatum (Lin et al., 1990). XYROPTERIS K.U.KRAMER Typus generis: K.U.Kramer Xyropteris stortii (Alderw.) LINDSAEA DRYAND. Synonyms: Guerinia J.Sm., Humblotiella Tardieu, Isoloma J.Sm, Lindsaenium Fée, Odontoloma J.Sm., Ormoloma Maxon, Sambirania Tardieu, Schizolegnia Alston, Schizoloma Gaudich., Synaphlebium J.Sm. Typus generis: Lindsaea trapeziformis Dryand. = Lindsaea lancea (L.) Bedd. Taxonomic history: The genus Lindsaea was described by Dryander (1797), who included 10 species in the genus. More species were added by several workers (e.g. Klotzsch, 1844; Fée, 1852). The genus name was often incorrectly spelled Lindsaya until Copeland (1947) restored the original spelling. Kramer (1957a, 1967b, 1971a, 1972a) was the first to provide a proper revision of the genus and he broadened the circumscription of Lindsaea considerably. For example, he (Kramer, 1967b, 1972a) placed Schizoloma, Isoloma, Humblotiella and Sambirania as sections under Lindsaea. Kramer (1957a, 1967b, 1971a, 1972a) proposed the subdivision of Lindsaea into two subgenera and 23 sections and this has been followed by all later workers (e.g. Walker, 1973; Tryon & Tryon, 1982). The present paper is the first attempt to investigate phylogenetic relationships among Lindsaea spp. in detail. In the process, it has become obvious that the varieties of some previously recognized species do not form a monophyletic group. The following new combinations are therefore proposed: Lindsaea pseudohemiptera (Alderw.) Lehtonen & Tuomisto, comb. nov. Basionym: Lindsaea repens var. pseudohemiptera Alderw. Bulletin du Jardin Botanique de Buitenzorg, III, 2: 157 (1920). © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 338 S. LEHTONEN ET AL. Figure 17. Nesolindsaea kirkii, reproduced from Baker (1871–1875). © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE Type: Sumatra, Mt. Merapi, Bünnemeijer 5419 (lectotype BO, isolectotypes L!, SING, U!, US!). Lindsaea hainaniana (K.U.Kramer) Lehtonen & Tuomisto, comb. nov. Basionym: Lindsaea lobata Poiret var. hainaniana K.U.Kramer, Gardens’ Bulletin, Singapore 26: 37 (1972). Type: Hainan, Liang 63789 (holotype US!, isotype K!). Lindsaea agatii (Brack.) Lehtonen & Tuomisto, comb. nov. Basionym: Schizoloma agatii Brack. in United States Exploring Expedition. Botany. Cryptogamia. Filices 16: 216 (1854). Type: Fiji, U.S. Expl. Exped. s.n. (holotype US!, isotypes K!, P, P00633682!). Synonym: Lindsaea ensifolia Sw. ssp. agatii (Brack.) K.U.Kramer Kramer (1967b) described a new species, L. sarawakensis K.U.Kramer and considered the earlier names L. diplosora Alderw. and L. diplosora Alderw. var. acrosora C.Chr. as synonyms of L. rigida J.Sm. (Kramer, 1971a). In our opinion L. rigida and L. diplosora are distinct taxa, but L. sarawakensis is conspecific with L. diplosora. Furthermore, we do not give L. diplosora var. acrosora a distinct taxonomic status. Consequently, we accept Lindsaea diplosora Alderw. (holotype: Matthew 523 BO, photograph in AAU!) as a species with the heterotypic synonyms Lindsaea sarawakensis K.U.Kramer (holotype: Mjöberg 9, P!) and Lindsaea diplosora Alderw. var. acrosora C.Chr. (lectotype (designated here): Mjöberg 94, S!, duplicate BM!). Lindsaea lapeyrousei (Hook.) Baker was originally described based on two samples: one from Vanikoro in the Solomon Islands (Moore s.n., K) and one from Fiji (Milne s.n., K). Kramer labelled the Moore specimen as ‘holotype’ and the Milne specimen as ‘paratype’, but they are both syntypes. The latter was then placed into Kramer’s L. lapeyrousei ssp. fijiensis K.U.Kramer (type: Fiji, Degener & Ordonez 13734a, holotype L, isotypes MICH, BISH, GH, K). The lectotype of L. lapeyrousei (designated here) is ‘Vaniholla’, C. Moore s.n. (K!), and we place all samples from the Solomon islands and Fiji under this (including L. lapeyrousei ssp. fijiensis and L. kajewskii Copel.). The name honours French explorer Jean-François de La Pérouse, who was shipwrecked in Vanikoro in 1788, and is correctly cited as L. lapeyrousei, instead of L. lapeyrousii (article 60.7 of the ICBN; McNeill et al., 2006). Notes: Lindsaea is a pantropical genus of approximately 150 species, some of which extend to the subtropics in Japan, Australia and New Zealand. The monophyly of Lindsaea (as circumscribed here) was supported in all our analyses. Kramer (1957a) defined Lindsaea as having laterally free indusia, but this character is variable. Several Lindsaea spp. have 339 laterally adnate indusia and laterally free indusia are not restricted to Lindsaea either. In the total evidence analysis, Lindsaea was supported by short petioles, abaxially keeled rachises, broad rachis sulcation and twice-ordered laminar venation (Table 4). However, none of these characters is present in all Lindsaea spp. and some are not even the prevailing characters in the genus. Neither of the subgenera defined by Kramer (1967b) is monophyletic in our analyses. Species placed in subgenus Odontoloma were resolved in clades V, VI and IX, thus also rendering subgenus Lindsaea paraphyletic. Consequently, we do not recognize the subgeneric classification. Most of the proposed sections within Lindsaea are not monophyletic and those which are monophyletic are embedded within other sections. Therefore, we do not recognize the sections either. Instead of presenting a new formal subgeneric classification here, we will discuss the phylogenetic lineages within Lindsaea by using informal clades numbered from I to XIII. SUBDIVISION OF LINDSAEA CLADE I This clade consists of the Central American L. seemannii J.Sm. and L. pratensis Maxon and is well supported and highly stable in the molecular, morphological and combined analyses. Both species belong to Kramer’s (1957a) section Tropidolindsaea, but we failed to sample another two species of the section. The Caribbean L. protensa C.Chr. is probably a member of clade I, but L. adiantoides J.Sm. from the Philippines is more likely to belong to our clade II. Kramer (1957a, 1971a) correctly suggested that these species are isolated within the genus. Members of clade I resemble the species in clade II in having a keeled rachis and clustered fronds. Spores are monolete in clade I, whereas most species of clade II have trilete spores. There are no chromosome counts available for clade I. CLADE II This well-supported clade includes all species of Kramer’s (1967b) section Isoloma (sometimes treated as a distinct genus), but also L. viridis Col. (of the monotypic section Chlorolindsaea; Kramer & Tindale, 1976) and L. plicata Baker (the only member of section Sambirania that we were able to sample; Kramer, 1972a). The clade is found in Madagascar and New Zealand and from Malesia to western Polynesia. Common morphological characters in this clade are keeled rachises, clustered fronds and dark axes. Most of the species have non-dimidiate and basally auriculate pinnules, but L. viridis and L. plicata have deeply dissected pinnules. Spores are trilete, except in L. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 340 S. LEHTONEN ET AL. viridis, in which they are monolete. The studied sequences are highly divergent within the clade, suggesting a great age for the divergences. Lindsaea gueriniana (Gaudich.) Desv. and L. pellaeiformis Christ were not resolved as sister species despite their morphological similarity. We were unable to sequence two species of section Isoloma that most likely belong to this clade: L. ovata J.Sm. and L. philippinensis K.U.Kramer. The latter, however, seems morphologically indistinguishable from L. jamesonioides Baker and probably should be considered synonymous with that species. The only available chromosome count for clade II is N = 88 for L. viridis (Brownlie, 1961). Kramer (1970) divided L. lapeyrousei into two subspecies, with the specimens from Fiji placed in ssp. fijiensis. The subspecies are recognized by slight differences in pinnule segmentation and in the distance between the indusium and the pinnule margin. However, the morphological variation is overlapping and we found no differences in the DNA sequences of the specimens sampled. Hence the recognition of two subspecies seems unnecessary. The available chromosome counts for clade IV are N = 47 for both L. brachypoda (Baker) Salomon (as L. concinna, Kramer, 1957a) and L. lucida (Mitui, 1976). CLADE V CLADE III This clade only includes L. incisa Prent., an endemic species from eastern Australia. Kramer & Tindale (1976) suggested that L. microphylla Sw., L. linearis Sw. and L. dimorpha F.M.Bailey (not included in our study) could be the closest relatives of L. incisa. Indeed, L. linearis and L. microphylla are resolved as a clade, but without close affinity with L. incisa. In the molecular tree, L. incisa is resolved as sister to the clade [IV (V VI)], but the combined analysis placed it as sister to clades IV–XIII. The position in the molecular tree is quite unstable and lacks support, but the more isolated phylogenetic position in the combined analysis is rather stable and well supported. Lindsaea incisa has trilete spores. No chromosome counts are available. CLADE IV This clade and the phylogenetic relationships within it are well supported and mostly stable in the molecular and the combined analyses. With the exclusion of L. kingii Copel., this clade was also resolved in the morphological analysis. Kramer (1971a) treated L. brevipes Copel. as a subspecies of L. lucida Blume and Kramer (1970) suggested that both, especially ssp. brevipes, are closely related with L. lapeyrousei (Hook.) Baker. This hypothesis is supported by our analyses, as L. brevipes was resolved as sister to L. lapeyrousei, and these two together as sister to L. lucida. As the two proposed subspecies of L. lucida did not form a monophyletic group in any of our analyses, we recognize them as separate species. Lindsaea brevipes usually has short petioles and more deeply incised and smaller pinnules than L. lucida, but the exact boundary between the two is fuzzy. According to the literature, L. brevipes occurs only in eastern Malesia but L. lucida is distributed from Malesia to the Himalayas; L. lapeyrousei is confined to the Bismarck Archipelago, the Santa Cruz (Solomon) Islands and Fiji. The species in clade V are epiphytes and thus form an unusual group within the predominantly terrestrial Lindsaea. The monophyly of the group is well supported and highly stable in the molecular analysis, although less so in the combined analysis. Lindsaea repens (Bory) Thwaites sensu Kramer (1971a) was not supported as a monophyletic entity in our analyses. Out of the 11 varieties (Kramer, 1967b, 1970, 1971a), we were able to sequence three: var. pectinata (Blume) Mett. ex Kuhn, var. sessilis (Copel.) K.U.Kramer and var. pseudohemiptera Alderw. We accept these varieties as distinct species and we assume that many of the unsampled varieties should also have a status at the species level. In the combined analysis, L. pectinata Blume was resolved as the sister of L. oblanceolata Alderw. with high support and low stability, but in the molecular analysis they were poorly supported and of unstable successive lineages. Nevertheless, these two species are morphologically highly similar with somewhat rounded and only shallowly incised pinnules and typically plurinerval sori. Lindsaea pectinata and L. oblanceolata have widely overlapping distribution in Malesia and Thailand. They can be distinguished by L. pectinata having much shorter sori and pinnules with numerous shallow incisions, in contrast to L. oblanceolata, which has only a few shallow incisions and irregularly interrupted sori. Lindsaea sessilis Copel. was resolved as sister to Lindsaea sp. 1, both in the molecular and the combined analyses. Lindsaea sp. 1 may represent an undescribed species, as we were unable to mach it with any species description. It has pinnules quite similar to those of L. sessilis, but differs from that species in having long petioles and bipinnate fronds. The specimen Cicuzza 891 (Z!) was originally identified as L. microstegia Copel., but it clearly differs from that species in pinnule shape. Lindsaea sessilis differs from the other species by having relatively narrow incisions in pinnules and reniform indusia. Lindsaea pseudohemiptera has irregularly incised pinnules, © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE with the incisions being quite deep, especially towards the pinnule apices. In this character, it resembles its closest relative, L. carvifolia K.U.Kramer, which also has deeply incised pinnules. Apparently juvenile plants of this clade often have deeply incised pinnules and the repeated evolution of deeply incised adult forms can be explained as convergent retention of juvenile traits (paedomorphosis). Another species with deeply incised pinnules is L. fissa Copel. The specimen we sampled for molecular analysis was, however, sterile and had juvenile fronds only. Its species identification is thus uncertain, but the specimen is morphologically identical to juvenile leaves of typical L. fissa. In the molecular analysis, L. cf. fissa was resolved as sister to L. merrillii Copel. ssp. yaeyamensis (Tagawa) K.U.Kramer, but the combined analysis resolved L. merrillii ssp. yaeyamensis as sister to L. merrillii ssp. merrillii. Because of these conflicting results and uncertain identification of the specimen representing L. fissa, we accept Kramer’s (1972b) subspecific ranks for L. merrillii for the time being. The chromosome numbers N ª 44 (L. sessilis; Walker in Kramer, 1971a) and N ª 47 (L. doryphora and L. pectinata; Kramer, 1957a) have been reported for members of clade V. CLADE VI This clade is composed of both terrestrial and epiphytic groups and was resolved as well supported, but rather unstable sister to the purely epiphytic clade V. Also the phylogenetic relationships within the clade are mostly unstable and some groupings were differently resolved in the molecular and the combined analyses. The first diverging lineage within this clade is, however, well supported and highly stable. This lineage contains L. diplosora Alderw. (= L. sarawakensis) from Borneo, L. sp. 2 from Sulawesi, L. rigida J.Sm. from Malesia and L. monocarpa Rosenst. from New Guinea. All these species have apically located sori. Lindsaea rigida and L. monocarpa have coriaceous pinnules, but can be separated because L. monocarpa has almost entire pinnules, hidden veins and a single apical sorus, whereas L. rigida has dentate–crenate pinnules, elevated veins and 1–4 sori per pinnule. Lindsaea diplosora and L. sp. 2 have herbaceous pinnules. Pinnules are less incised, more dense and inserted to the rachis at a more acute angle in L. diplosora than in L. sp. 2. Although the node separating L. diplosora and L. sp. 2 is well supported and highly stable, it may represent an incongruence between gene and species trees. Hence, further sampling is desirable before formal taxonomic actions are taken. In addition to the voucher specimens 341 mentioned in Table 2, L. diplosora is represented by Chew 1811 (K!) and Parris 11612 (L!), L. rigida by Anderson 2815 (MO!), Fallen 439 (MO!), Parris 6026 (K!) and Richards 1703 (K!) and L. sp. 2 by Kluge 7248 (GOET!). The next diverging lineage is composed of three species: L. propinqua Hook. and L. pacifica K.U.Kramer from the Pacific islands and L. azurea Christ from Malesia. Of these species, L. propinqua is morphologically highly variable across its wide range and may actually include several species. The next diverging lineage corresponds with the section Penna-arborea, one of the few sections supported as a monophyletic entity. All the species of this clade typically grow as epiphytes. We found L. pulchra (Brack.) Carruthers ex Seem. sensu Kramer (1970) to be polyphyletic. Kramer (1970) himself noted that specimens from the Bismarck Archipelago and Solomon Islands have unusually large pinnules and we recognize them as L. stolonifera Mett. ex Kuhn. In the molecular analysis, this species was resolved as sister to L. pulchella (J.Sm.) Mett. ex Kuhn, but in the combined analysis as sister to L. chrysolepis K.U.Kramer. Lindsaea blanda Mett. ex Kuhn was treated as a variety of L. pulchella by Kramer (1971a), but is recognized here as a separate species. The remaining species are terrestrial and form a clade with L. parallelogramma Alderw. as the first branching species. Lindsaea tenuifolia Blume, L. polyctena K.U.Kramer, L. multisora Alderw., L. tetragona K.U.Kramer and L. longifolia Copel. form a clade with slightly different internal relationships between the molecular and the combined trees. In the combined analysis, L. tenuifolia and L. polyctena are resolved as sisters, with high support and stability, in contrast to their highly unstable position as successive lineages in the molecular analysis. A sister group relationship is supported by the fact that these two species share a unique 4-bp inversion in the trnL-trnF intergenic spacer region. We found L. lobata Poiret sensu Kramer (1972b) to be polyphyletic and correspondingly recognize his var. hainaniana at the species level as L. hainaniana. Lindsaea cultrata (Willd.) Sw. and L. integra Holtt. are often confused, but L. cultrata is generally larger, often bipinnate, has some incisions in the upper margin of the pinnules and has pinnules more than twice as long as broad. Lindsaea integra, however, is generally smaller, simply pinnate and has almost entire pinnules that are relatively broader than in L. cultrata and, at most, twice as long as broad. Lindsaea cultrata has also often been confused with Osmolindsaea odorata (Kramer, 1967b). This explains the observations of a solenostelic rhizome in L. cultrata (GwynneVaughan, 1903; Barcelona, 2000), which actually has © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 342 S. LEHTONEN ET AL. a protostele in contrast to a solenostele in Osmolindsaea. According to the available cytological studies, L. parallelogramma and L. integra (‘L. nitida’) have N = 47 (Kramer, 1957a), L. blanda has N = 44–45 (Walker in Kramer, 1971a) and L. obtusa J.Sm. ex Hook. has N ª 50 or 100 (Holttum, 1954). CLADE VII Lindsaea microphylla and L. linearis formed a wellsupported clade that was also stable in the molecular analysis. This clade probably also includes L. dimorpha, although we were unable to obtain any sequences from this species. All the species occur in Australia, but L. linearis extends to New Caledonia, Norfolk Island and New Zealand (Kramer & Tindale, 1976) and L. dimorpha extends to New Caledonia (Kramer, 1967a). Both L. linearis and L. dimorpha are dimorphic with dissimilar sterile and fertile fronds, which is unusual in Lindsaea. Lindsaea microphylla is not as clearly dimorphic, but its sterile fronds typically have larger segments than those in fertile fronds and sterile fronds usually persist in fertile plants (Kramer & Tindale, 1976). This clade appears to be distinct also cytologically, as the available chromosome count N = 34 for L. linearis (Brownlie, 1958) is unique in the genus. CLADE VIII Clades VIII and IX together roughly correspond to section Schizoloma. Lindsaea vieillardii Mett., a New Caledonian endemic, was resolved as the first diverging lineage of clade VIII in the molecular analysis, but as the first diverging lineage of clade IX in the combined analysis. In both analyses the position of this species was unstable and lacked jackknife support. The rest of clade VIII is more stable and more strongly supported. Lindsaea schizophylla (Baker) Christ is endemic to Sri Lanka. We sequenced two specimens originally identified as this species and they were resolved as sisters. However, only one of them has the deeply dissected pinnules and rather short rhizomes typical of L. schizophylla. The other specimen (Laegaard 13835 AAU!) has long rhizomes and only slightly incised, flabellate pinnules and apparently represents an unknown species (L. sp. 3). Lindsaea bouillodii Christ was resolved as sister to L. cf. cambodgensis Christ. However, the latter specimen has much larger pinnules than typical L. cambodgensis and may actually represent another species. Molecular data alone were insufficient to resolve the relationships of six apparently closely related species: L. grandiareolata (Bonap.) K.U.Kramer from Madagascar, L. dissectiformis Ching from Vietnam and Hainan, L. chienii Ching from Japan to Thailand, L. annamensis K.U.Kramer from Vietnam, L. kawabatae Kurata from Japan and L. austrosinica Ching from South China and Indo-China. Lindsaea grandiareolata is morphologically the most distinct. It has lanceolate pinnules and anastomosing veins, unlike any other species in the group. The other species are more homogeneous and their status as separate species has been questioned. The main distinctive characters are in the dissection of the lamina. Kramer (1972b) suggested that L. dissectiformis and L. kawabatae may be conspecific, and Dong & Zhang (2006) considered L. annamensis as a synonym of L. chienii but accepted L. dissectiformis as a distinct species. In our opinion, L. annamensis and L. chienii are distinct and for the moment we accept all the species accepted by Kramer (1972b). Dong & Zhang (2006) used the name L. eberhardtii (Christ) K.U.Kramer, but, as explained by Kramer (1972b), the basionym Stenoloma eberhardtii is a nomen nudum and the correct name for this species is L. dissectiformis. The available chromosome counts are N ª 47 for both L. chienii (Kurita & Nishida, 1963) and L. vieillardii (Brownlie, 1965), although the latter species may belong to clade IX. CLADE IX This clade received poor support and was unstable and the position of L. vieillardii as the first diverging lineage in the combined analysis is questionable. The clade as a whole is distributed widely around the Indian Ocean and the Pacific. The Madagascan species were grouped together in the combined analysis, but L. oxyphylla Baker was excluded from this group in the molecular analysis. The species of clade IX usually have equilateral pinnae or undivided pinna apices. Lindsaea ensifolia Sw. and L. heterophylla Dryand. have presented taxonomic problems because of their great morphological variation (Dixit & Ghosh, 1980, 1983). High chromosome numbers have been reported for L. ensifolia (N = 88; Manton & Sledge, 1954), indicating that polyploidy may have been involved in the evolution of this complex. Hybrid origin of L. heterophylla has also been suggested; the species is reported to often have irregularly shaped and abortive spores (Kramer, 1971a). The specimen included in our analyses had both GA- and CT-type (Lehtonen et al., 2009) loop regions of the trnL-trnF stem-loop structure, which must mean that either the specimen has two types of plastid genomes, or the inversion has occurred in a somatic cell. In comparison, the CT-type inversion was found in L. ensifolia and the GA-type in L. agatii. This species complex also includes several taxa that we were not able to include in our analyses: L. ensifolia ssp. coriacea (Alderw.) K.U.Kramer, L. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE rutlandica R.D.Dixit & B.Ghosh and L. beddomei R.D.Dixit & B.Ghosh. The latter two taxa were not accepted by Fraser-Jenkins (2008), who considered L. beddomei conspecific with L. heterophylla and L. rutlandica conspecific with L. walkerae Hook. We have not been able to study the type specimens of these taxa. Lindsaea fraseri Hook. and L. media R.Br. are also problematic species that require further study. Kramer & Tindale (1976) found only abortive spores in L. fraseri and speculated that this species may be a hybrid between L. media and L. agatii. We found no direct evidence of hybrid origin, but almost identical sequences of L. media and L. fraseri suggest a close affinity between these two species. The Madagascan species formed a monophyletic clade both in the molecular and combined analyses, with the latter receiving higher jackknife support and stability. In both analyses, L. madagascariensis Baker and L. leptophylla Baker were resolved as sister species, as were L. goudotiana (Kunze) Mett. ex Kuhn and L. subtilis K.U.Kramer. Although L. blotiana K.U.Kramer and L. millefolia K.U.Kramer are morphologically distinct, with the pinnules of L. blotiana being clearly less divided, the two had no variation in the studied sequences. The finely dissected pinnules of L. millefolia give the species an appearance similar to the unrelated L. dissectiformis of clade VIII, L. bifida of clade XI and L. sphenomeridopsis of clade XIII. CLADE X This clade was rather unstable and not well supported in our molecular analysis, but in the combined analysis both support and stability measures were higher. The close relationship among these species had already been suggested by Kramer (1967a) and Kramer & Tindale (1976). Clade X is biogeographically well defined, with three species endemic to New Caledonia (L. rufa K.U.Kramer, L. nervosa Mett. and L. prolongata E.Fourn.) and one distributed in New Zealand, Australia and Tasmania (L. trichomanoides Dryand.). We found no differences in the studied sequences between L. rufa and L. trichomanoides, which suggests a recent divergence of these species (for a recent discussion of New Caledonian biogeography, see Grandcolas et al., 2008; Heads, 2008). Lindsaea rufa and L. trichomanoides are morphologically very similar, but the former is more robust. The molecular analysis was unable to resolve the branching order of clades VII (VIII IX), X and [XI (XII XIII)]. In the combined analysis, clade X was resolved as the sister to the South American clade [XI (XII XIII)], although with low stability and no support. Available chromosome numbers in this clade are N = 42 for L. trichomanoides (as L. cuneata, Brownlie, 343 1957) and N ª 40 for L. prolongata (Brownlie, 1965). Similar chromosome numbers have only been reported from Neotropical Lindsaea, supporting the sister relationship as recognized in the combined analysis. CLADE XI This well-supported but relatively unstable clade is composed of species endemic to the Atlantic coastal rain forests of Brazil (Mata Atlântica). The studied sequences of Lindsaea bifida (Kaulf.) Mett. ex Kuhn were barely different from those of L. virescens Sw. and a second well-supported and stable sister species pair was L. ovoidea Fée with L. botrychioides A.St.Hil. Both Lindsaea virescens and L. bifida have finely dissected laminas, whereas L. ovoidea and L. botrychioides have more or less flabellate pinnules. Lindsaea bifida is morphologically almost inseparable from L. sphenomeridopsis K.U.Kramer from the Guyana shield and Kramer (1957a) considered the two closely allied. However, our analyses resolved L. sphenomeridopsis to clade XIII. No chromosome counts are available for clade XI. CLADE XII Ormoloma has been considered a separate genus but a close relative of Lindsaea (Kramer, 1957a). All of our analyses resolve Ormoloma to be deeply nested within Lindsaea, although its position within the genus is unstable. Ormoloma is divergent both morphologically and based on the sequence data, and low stability suggests possible issues caused by these long branches. However, cytological studies of Lindsaeaceae have reported N = 42 only from Ormoloma (Wagner, 1980) and some members of clades X and XIII. The phylogenetic position of the Caribbean Ormoloma within the Neotropical Lindsaea clade is also biogeographically reasonable. However, additional sampling of more conservative loci is needed to confirm the placement of Ormoloma. Two species have been recognized, but their status as distinct species was already rejected by Tryon & Tryon (1982). We were unable to find any constant morphological or molecular differences between the two proposed species and consequently accept only Lindsaea imrayana (Hook.) Perez. CLADE XIII This large clade is well supported but rather unstable and includes only Neotropical species. Although some groups within the clade are morphologically clearly distinct, we found practically no variation among the species at the sequence level. Hence, the topology © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 344 S. LEHTONEN ET AL. within the clade is poorly supported and highly unstable and only a few observations are worth making. Both the molecular and combined analyses resolved Lindsaea feei C.Chr. from northern South America as sister to the rest of clade XIII (this species was called L. klotzschiana Moritz by Kramer, 1957a; for a discussion on the nomenclature, see Morton, 1967). Lindsaea pendula Klotzsch and L. meifolia (Kunth) Mett. are morphologically unique within the genus in having pendulous segments and they are resolved as sister species in all analyses, with full stability. One larger monophyletic group, which includes species from L. cyclophylla K.U.Kramer to L. arcuata Kunze, is recognized in both the molecular and the combined analysis with high support and, in the case of the molecular analysis, with high stability as well. However, this clade includes L. cyclophylla but excludes the two other species of Kramer’s (1957a) section Haplolindsaea, namely L. sagittata (Aublet) Dryand. and L. reniformis Dryand. All three species are morphologically similar with a simple lamina. The polyphyletic origin of Haplolindsaea is unexpected, although the different spore architecture of L. cyclophylla (monolete spores in contrast with trilete spores in L. sagittata and L. reniformis) may be a morphological indication of this. Our sampling included all three varieties of L. stricta accepted by Kramer (1957a) and we failed to resolve them as a monophyletic group. Several other species were also not supported as monophyletic entities; these include L. portoricensis Desv., L. rigidiuscula Lindm. and L. lancea (L.) Bedd. In the case of L. lancea, the non-monophyletic origin was expected on the basis of the great morphological differences between the varieties. We follow our earlier justification (Tuomisto, 1998) and treat L. falcata Dryand. as a full species instead of a variety under L. lancea, contrary to Rosenstock (1906) and Kramer (1957a). Species of clade XIII have apparently diverged recently. As we only used plastid sequences, hybridization and incomplete lineage sorting because of retained ancient polymorphisms may have confused the phylogenetic inference (Maddison & Knowles, 2006). The presence of polyploidy within the clade also suggests that hybridization may have taken place. Available chromosome counts are N = 42 for L. divaricata Klotzsch and L. falcata (Tryon, Bautista & da Silva Araújo, 1975), N = 84 for L. schomburgkii Klotzsch (Tryon et al., 1975), N = 88 for L. portoricensis (Walker, 1966) and either N ª 84 or N ª 88 for L. arcuata (Mickel et al., 1966; Smith & Mickel, 1977) and L. quadrangularis Raddi (Tryon et al., 1975). More detailed studies using multiple rapidly evolving nuclear markers are needed to solve the taxonomic problems within the clade. EVOLUTIONARY RELATIONSHIPS WITHIN THE LINDSAEOIDS Based on our phylogenetic analyses, the first lineage to diverge from the lindsaeoid stock is Sphenomeris clavata. Its branch is long and located near the root, thus raising concerns of long-branch attraction (Bergsten, 2005). The node combines 100% jackknife support with low stability, which can be an indication of long-branch problems (Giribet, 2003). Further sampling of more conservative genes is needed to verify the phylogenetic position of Sphenomeris. The next diverging lineage is Odontosoria, which is here circumscribed such that it also includes all the sampled species of Sphenomeris, except S. clavata. The clade is pantropical, but the Neotropical species form a monophyletic and strongly supported group nested within it (96% jackknife support in morphological analysis, 100% support and high stability in molecular and combined analyses). All the Neotropical species are climbers and they share several morphological synapomorphies (scrambling fronds with indeterminate growth, flexuose rachises, opposite and reflexed pinnae and large spores; Table 4). An early study based on rbcL sequences (Wolf, 1995) suggested that Tapeinidium should be excluded from Lindsaeaceae, but this result must have been based on erroneous identification or sample contamination; the position of Saccoloma in that study was also problematic. In our study, Tapeinidium forms a clade with several curious species that have traditionally been placed in Lindsaea, but which we place in the new genera Osmolindsaea and Nesolindsaea. This clade is Palaeotropical, with a centre of diversity in Melanesia. It is separated from Lindsaea by a long, well-supported and highly stable branch. The sister relationship between Lindsaea and (Tapeinidium (Nesolindsaea Osmolindsaea)) clade obtains high jackknife support (99%), but only 3/16 stability in the combined analysis. The sister relationship (Odontosoria Lindsaea) actually obtains higher stability (5/16), but did not emerge in the equal-cost regime favoured by us. The monophyly of a strictly defined Lindsaea is well established with 100% jackknife support in the molecular and combined analyses. The morphological analysis resolved the clade as well, but without jackknife support. The clade is stable in the combined analysis (13/16), but less so in the molecular analysis (8/16). The monophyletic Lindsaea also includes Ormoloma, an enigmatic Neotropical group previously recognized as a separate genus. EVOLUTION OF MORPHOLOGICAL AND ANATOMICAL CHARACTERS The ‘Lindsaea-type’ protostele (Fig. 16C) is present in all Lindsaea and Sphenomeris, and in some Tapein- © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE idium and Odontosoria, but is unknown among adult members of other ferns (Kramer, 1957a). Nesolindsaea and most Odontosoria have solenosteles. The stelar anatomy of Osmolindsaea and most Tapeinidium may be considered intermediate, as internal sclerified pith is typically present at least near the frond bases. Kramer & Green (1990) considered solenostele as the primitive state in lindsaeoids, from which protostele was derived by simplification. The protostele is clearly a derived state within Polypodiales (Wolf, 1995), but as it is present in both Lindsaea and Sphenomeris, character optimization of stelar anatomy remains ambiguous for the common ancestor of lindsaeoids. Some Lindsaea spp. (e.g. L. bifida, L. sphenomeridopsis, L. millefolia, L. dissectiformis) have highly dissected leaves similar to those of species in several other lindsaeoid genera. Kramer (1957a) therefore considered high degree of leaf dissection to be a primitive character. However, our analyses suggest that the common ancestor of modern Lindsaea had erose–crenate pinnule margins and that both the entire and the highly dissected pinnule forms have evolved independently several times. Early Lindsaea probably had once pinnate leaves, from which bipinnate leaves have evolved several times with many reversals back to the once pinnate form. Large terminal pinnule appears to have evolved multiple times from the more primitive state of reduced distal pinnules. Most species of Lindsaeaceae have free veins, but anastomosing veins prevail in clade VI and may be a synapomorphy of this lineage, although with some reversals to free venation. Anastomosing veins also occur in several species of clades VIII and IX and several independent origins of vein fusion must be assumed. The coenosorus appears to be derived from the uni-binerval sorus, in accordance with Kramer’s (1957a) hypothesis. However, continuous sori have evolved separately in Nesolindsaea, Osmolindsaea, Lindsaea and, possibly, Xyropteris (the phylogenetic position of which remains unknown). The coenosorus is present among the earliest lineages of Lindsaea, although it is most common in the derived clade XIII. A narrow indusium is a derived character, as already assumed by Kramer (1957a), with multiple independent origins. Just like the evolution of laminar architecture, the evolution of sorus structure is also confounded by convergence and reversals. Monolete spores have been derived from trilete ancestors more than once within Lindsaeaeae. However, it is not clear whether the common ancestor of modern Lindsaea had monolete or trilete spores. Tapeinidium and Osmolindsaea have monolete spores, but Nesolindsaea, Sphenomeris and most 345 Odontosoria have trilete spores. One transformation from trilete to monolete type must be assumed in Odontosoria. All species of Lindsaea clade I have monolete spores, as does L. viridis of clade II. Trilete spores are optimized for the common ancestor of Lindsaea clades II–XIII. Therefore, monolete and trilete spores can be equally parsimoniously assumed to represent the ancestral state for Lindsaea. In the combined analysis, L. viridis was resolved as the first branching species of clade II, but in the molecular analysis it was more derived. If the former is correct, the monolete spores of clade I and L. viridis can be considered homologous (although it would still be equally parsimonious to assume independent evolution from a trilete ancestor). Apart from these clades, monolete spores in Lindsaea are restricted to a few species of clade XIII. These species do not form a monophyletic group in our analyses, thus suggesting at least two independent origins of monolete spores within clade XIII. Species of Lindsaeaceae are typically terrestrial, but epiphytism occurs in two groups within Lindsaea. Clade V is predominantly epiphytic and several epiphytic species form a monophyletic group within clade VI. These two occurrences are most parsimoniously explained as independent origins of epiphytism, but the origin in closely related groups may indicate that some preadaptations to epiphytic habit were already present in their common ancestor. We failed to sample L. odontolabia (Baker) K.U.Kramer, an epiphytic species endemic to Madagascar, which may be a member of clade IX. In any case, epiphytism is a derived character in the family, as suggested by previous authors (Copeland, 1947; Kramer, 1971a), although the fossilized roots of a lindsaeoid fern dating to the Early Cretaceous were growing epiphytically on a tree fern trunk (Schneider & Kenrick, 2001). Pryer et al. (2004) used this fossil to constrain the minimum age of Lindsaeaceae in their estimation of leptosporangiate divergence times. On the one hand, if the fossil taxon is a member of an extant epiphytic Lindsaea clade, rather than some now extinct epiphytic lineage, this may have led to serious underestimation of the family age. On the other hand, several terrestrial Lindsaea spp. can occasionally grow as epiphytes on the lower parts of tree fern trunks. Another possible calibration point for the molecular clock is provided by fossil Lindsaea spores from Upper Miocene deposits in New Zealand (Mildenhall, 1980). There are only three extant Lindsaea spp. in New Zealand (L. viridis, L. trichomanoides, L. linearis; Kramer & Tindale, 1976) and each of them belongs to a different clade. If the fossil spores could be associated with a modern species, a more exact dating of the Lindsaea clades might be possible. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 346 S. LEHTONEN ET AL. ACKNOWLEDGEMENTS We thank Michael Kessler and Michel Boudrie for the plant material they kindly provided for our project, Veikko Rinne for taking the cross-section images of rhizomes and anonymous reviewers for their constructive comments on the earlier version of the manuscript. We are grateful to the curators of AAU, BISH, BM, CAY, KYO, L, MO, P, TUR, U, US and Z for allowing their specimens to be destructively sampled for the molecular study and to the curators of K, MICH, NSW and UC for providing digital images or material on loan. The Willi Hennig Society is acknowledged for making TNT publicly available. This study was funded by a grant from the Academy of Finland to H.T. and by grants from the Kone Foundation and the Oskar Öflund Foundation to S.L. REFERENCES Aagesen L. 2005. Direct optimization, affine gap costs, and node stability. Molecular Phylogenetics and Evolution 36: 641–653. Agnarsson I, Miller JA. 2008. Is ACCTRAN better than DELTRAN? Cladistics 24: 1–7. Agolin M, D’Haese CA. 2009. An application of dynamic homology to morphological characters, direct optimization of setae sequences and phylogeny of the family Odontellidae (Poduromorpha, Collembola). Cladistics 25: 353–385. Alderwerelt van Rosenburgh CRWK van. 1914. New or interesting Malayan ferns 6. Bulletin du Jardin Botanique de Buitenzorg, sér. 2 1–60. Bain JF, Jansen RK. 2006. A chloroplast DNA hairpin structure provides useful phylogenetic data within tribe Senecioneae (Asteraceae). Canadian Journal of Botany 84: 862–868. Baker JG. 1871–1875. On the Seychelles fern flora. Transactions of the Royal Irish Academy, Science 25: 509–518. Barcelona JF. 2000. Systematics of the fern genus Odontosoria sensu lato (Lindsaeaceae). Unpublished D. Phil. Thesis, Miami University. Bergsten J. 2005. A review of long-branch attraction. Cladistics 21: 163–193. Biju SD, Bossuyt F. 2003. New frog family from India reveals and ancient biogeographical link with the Seychelles. Nature 425: 711–714. Borsch T, Quandt D. 2009. Mutational dynamics and phylogenetic utility of noncoding chloroplast DNA. Plant Systematics and Evolution 282: 169–199. Boyer SK, Giribet G. 2007. A new model Gondwanan taxon, systematics and biogeography of the harvestman family Pettalidae (Arachnida, Opiliones, Cyphophthalmi), with a taxonomic revision of genera from Australia and New Zealand. Cladistics 23: 337–361. Bremer K. 1988. The limits of amino-acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42: 795– 803. Brownlie G. 1957. Cyto-taxonomic studies on New Zealand Pteridaceae. New Phytologist 56: 207–209. Brownlie G. 1958. Chromosome numbers in New Zealand ferns. Transactions of the Royal Society of New Zealand 85: 213–216. Brownlie G. 1961. Additional chromosome numbers–New Zealand ferns. Transactions of the Royal Society of New Zealand, Botany 1: 1–4. Brownlie G. 1965. Chromosome numbers in some Pacific Pteridophyta. Pacific Science 19: 493–497. Catalano SA, Saidman BO, Vilardi JC. 2009. Evolution of small inversions in chloroplast genome, a case study from a recurrent inversion in angiosperms. Cladistics 25: 93–104. CBOL Plant Working Group. 2009. A DNA barcode for land plants. Proceedings of the National Academy of Sciences of the United States of America 106: 12794–12797. Chase MW, Cowan RS, Hollingsworth PM, van den Berg C, Madriñan S, Petersen G, Seberg O, Jørgensen T, Cameron KM, Carine M, Pedersen N, Hedderson TAJ, Conrad F, Salazar GA, Richardson JE, Hollingsworth ML, Barraclough TG, Kelly L, Wilkinson M. 2007. A proposal for a standardised protocol to barcode all land plants. Taxon 56: 295–299. Chaubey AK, Bhattacharya GE, Murty GPS, Srinivas K, Ramprasad T, Gopala Rao D. 1998. Early Tertiary seafloor spreading magnetic anomalies and paleo-propagators in the northern Arabian Sea. Earth and Planetary Science Letters 154: 41–52. Christenhusz MJM. 2009. Index Pteridophytorum Guadalupensium or a revised checklist to the ferns and club mosses of Guadeloupe (French West Indies). Botanical Journal of the Linnean Society 161: 213–277. Christensen C. 1938. Filicinae. In: Verdoorn F, ed. Manual of pteridology. The Hague: Martinus Nijhoff, 522–550. Committee for Pteridophyta. 1959. Nomina generica conservanda. Taxon 8: 205–207. Copeland EB. 1929. The oriental genera of the Polypodiaceae. University of California Publications in Botany 16: 45–128. Copeland EB. 1947. Genera filicum. Waltham, MA: Chronica Botanica. De Laet JE. 2005. Parsimony and the problem of inapplicables in sequence data. In: Albert VA, ed. Parsimony phylogenetics and genomics. New York: Oxford University Press, 148–159. DeBry RW. 2001. Improving interpretation of the decay index for DNA sequence data. Systematic Biology 50: 742–752. Diels L. 1898–1899. Polypodiaceae. In: Engler A, Prantl KAE, eds. Die natürlichen Pflanzenfamilien. Teil I, abt. 4., Leipzig: Verlag von Wilhelm Engelmann, 139–339, published in two parts. Dixit RD, Ghosh B. 1980. On the identity and nomenclature of Lindsaea heterophylla Dryand. and Lindsaea heterophylla Bedd. Proceeding of the Indian Academy of Sciences (Plant Science) 89: 179–183. Dixit RD, Ghosh B. 1983. The genus Lindsaea Dryand ex Smith in India. Proceeding of the Indian Academy of Sciences (Plant Science) 92: 233–258. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE Dong S-Y, Zhang X-C. 2006. A taxonomic revision of the fern genus Lindsaea (Lindsaeaceae) from Hainan, China. Acta Phytotaxonomica Sinica 44: 258–271. Drábková L, Kirschner J, Vlcek C. 2002. Comparison of seven DNA extraction and amplification protocols in historical herbarium specimens of Juncaceae. Plant Molecular Biology Reporter 20: 161–175. Dryander J. 1797. Lindsaea, a new genus of ferns. Transactions of the Linnean Society of London 3: 39–43. Espinasa L, Flick C, Giribet G. 2007. Phylogeny of the American silverfish Cubacubaninae (Hexapoda, Zygentoma, Nicoletiidae), a combined approach using morphology and five molecular loci. Cladistics 23: 22–40. Farris JS. 1983. The logical basis of phylogenetic analysis. In: Platnick NI, Punk VA, eds. Advances in cladistics, Proceedings of the third meeting of the Willi Hennig Society, Vol. 2. New York: Columbia University Press, 7–36. Farris JS, Albert VA, Källersjö M, Lipscomb D, Kluge AG. 1996. Parsimony jackknifing outperforms neighborjoining. Cladistics 12: 99–124. Farwell OA. 1931. Fern notes II. Ferns in the herbarium of Parke, Davis & Co. American Midland Naturalist 12: 233– 311. Fée ALA. 1852. Genera filicum. Exposition des genres de la famille des Polypodiacées (Classe des Fougères). Mémoires sur les familles des fougères vol. 5. Strasburg: BergerLevrault. Fitzhugh K. 2006a. The ‘requirement of total evidence’ and its role in phylogenetic systematics. Biology and Philosophy 21: 309–351. Fitzhugh K. 2006b. The abduction of phylogenetic hypotheses. Zootaxa 1145: 1–110. Fraser-Jenkins CR. 2008. Endemics and pseudo-endemics in relation to the distribution patterns of Indian Pteridophytes. Taiwania 53: 264–292. Frost DR, Rodrigues MT, Grant T, Titus TA. 2001. Phylogenetics of the lizard genus Tropidurus (Squamata, Tropiduridae, Tropidurinae), direct optimization, descriptive efficiency, and sensitivity analysis of congruence between molecular data and morphology. Molecular Phylogenetics and Evolution 21: 352–371. Funk DJ, Omland KE. 2003. Species-level paraphyly and polyphyly: frequency, causes, and consequences, with insights from animal mitochondrial DNA. Annual Review of Ecology, Evolution and Systematics 34: 397–423. Giribet G. 2003. Stability in phylogenetic formulations and its relationship to nodal support. Systematic Biology 52: 554–564. Giribet G. 2005. Generating implied alignment under direct optimization using POY. Cladistics 21: 396–402. Giribet G, Edgecombe GD, Wheeler WC. 2001. Arthropod phylogeny based on eight molecular loci and morphology. Nature 413: 157–161. Giribet G, Edgecombe GD, Wheeler WC, Babbitt C. 2002. Phylogeny and systematic position of Opiliones, a combined analysis of Chelicerata relationships using morphological and molecular data. Cladistics 18: 5–70. 347 Giribet G, Wheeler WC. 2007. The case for sensitivity, a response to Grant and Kluge. Cladistics 23: 294–296. Goloboff PA. 1999. Analyzing large data sets in reasonable times, solutions for composite optima. Cladistics 15: 415– 428. Goloboff PA. 2003. Parsimony, likelihood, and simplicity. Cladistics 19: 91–103. Goloboff PA, Farris JS, Nixon KC. 2008. TNT, a free program for phylogenetic analysis. Cladistics 24: 774– 786. Goloboff PA, Pol D. 2005. Parsimony and bayesian phylogenetics. In: Albert VA, ed. Parsimony phylogenetics and genomics. New York: Oxford University Press, 148– 159. Grandcolas P, Murienne J, Robillard T, DeSutterGrandcolas L, Jourdan H, Guilbert E. 2008. New Caledonia, a very old Darwinian island? Philosphical Transactions of the Royal Society of London. Series B, Biological Sciences 363: 3309–3317. Grazia J, Schuh RT, Wheeler WC. 2008. Phylogenetic relationships of family groups in Pentatomoidea based on morphology and DNA sequences (Insecta, Heteroptera). Cladistics 24: 932–976. Gwynne-Vaughan DT. 1903. Observations on the anatomy of solenostelic ferns. II. Annals of Botany 17: 689–742. Hasebe M, Omori T, Nakazawa M, Sano T, Kato M, Iwatsuki K. 1994. rbcL gene sequences provide evidence for the evolutionary lineages of leptosporangiate ferns. Proceedings of the National Academy of Sciences of the United States of America 91: 5730–5734. Heads M. 2008. Panbiogeography of New Caledonia, southwest Pacific, basal angiosperms on basement terranes, ultramafic endemics inherited from volcanic island arcs and old taxa endemic to young islands. Journal of Biogeography 35: 2153–2175. Hennig W. 1966. Phylogenetic systematics. Urbana, IL: University of Illinois Press. Hieronymus G. 1920. Kleine Mitteilungen über Pteridophyten III. Hedwigia 62: 12. Hollingsworth ML, Clark A, Forrest LL, Richardson J, Pennington RT, Long DG, Cowan R, Chase MW, Gaudeul M, Hollingsworth PM. 2009. Selecting barcoding loci for plants, evaluation of seven candidate loci with species level sampling in three divergent groups of land plants. Molecular Ecology Resources 9: 439–457. Holttum RE. 1947. A revised classification of leptosporangiate ferns. Journal of the Linnean Society (Botany) 53: 123–159. Holttum RE. 1954. A revised flora of Malaya, Vol. II. Singapore: Ferns of Malaya. Hoogland RD. 1952. A revision of the genus Dillenia. Blumea 7: 1–145. Jakob SS, Blattner FR. 2006. A chloroplast genealogy of Hordeum (Poaceae): long-term persisting haplotypes, incomplete lineage sorting, regional extinction, and the consequences for phylogenetic inference. Molecular Biology and Evolution 23: 1602–1612. Kim K-J, Lee H-L. 2005. Widespread occurrence of small © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 348 S. LEHTONEN ET AL. inversions in the chloroplast genomes of land plants. Molecules and Cells 19: 104–113. Kjer KM, Gillespie JJ, Ober KA. 2007. Opinions on multiple sequence alignment, and an empirical comparison of repeatability and accuracy between POY and structural alignment. Systematic Biology 56: 133–146. Klotzsch JF. 1844. Beiträge zu einer Flora der AequinoctialGegenden der neuen Welt. Filices. Linnaea 18: 525–556. Kluge AG. 1989. A concern for evidence and a phylogenetic hypothesis of relationships among Epicrates (Boidae, Serpentes). Systematic Zoology 38: 7–25. Korall P, Conant DS, Schneider H, Ueda K, Nishida H, Pryer KM. 2006. On the phylogenetic position of Cystodium, it’s not a tree fern – it’s a polypod! American Fern Journal 96: 45–53. Kramer KU. 1957a. A revision of the genus Lindsaea in the new world with notes on allied genera. Acta Botanica Neerlandica 6: 97–290. Kramer KU. 1957b. A new genus of lindsaeoid ferns. Acta Botanica Neerlandica 6: 599–601. Kramer KU. 1967a. The lindsaeoid ferns of the Old World. I. New Caledonia. Acta Botanica Neerlandica 15: 562– 584. Kramer KU. 1967b. The lindsaeoid ferns of the Old World III. Notes on Lindsaea and Sphenomeris in the Flora Malesiana area. Blumea 15: 557–576. Kramer KU. 1967c. The lindsaeoid ferns of the Old World II. A revision of Tapeinidium. Blumea 15: 545–556. Kramer KU. 1970. The lindsaeoid ferns of the Old World V. The smaller Pacific islands. Blumea 18: 157–194. Kramer KU. 1971a. Lindsaea-group. Flora Malesiana 1: 177–254. Kramer KU. 1971b. Lindsaeaceae. In: Bamps P, ed. Flore du Congo du Rwanda et du Burundi. Brussels: Jardin Botanique National de Belgique, 1–7. Kramer KU. 1971c. The lindsaeoid ferns of the Old World – VIII Sphenomeris in continental Africa. Bulletin du Jardin Botanique National de Belgique 41: 353–355. Kramer KU. 1972a. The lindsaeoid ferns of the Old World – IX Africa and its islands. Bulletin du Jardin Botanique National de Belgique 42: 305–345. Kramer KU. 1972b. The lindsaeoid ferns of the Old World VI. Continental Asia, Japan and Taiwan. The Gardens’ Bulletin Singapore 24: 1–48. Kramer KU. 1988. Lindsaea terrae-reginae, a new fern species from Queensland. Telopea 3: 287–289. Kramer KU. 1989a. The lindsaeoid ferns of the Old World X. Further notes on taxa described from China. Acta Phytotaxonomica et Geobotanica 40: 75–76. Kramer KU. 1989b. Lindsaea mesarum, a new fern species from the Roraima sandstone. Annals of the Missouri Botanical Garden 76: 605–607. Kramer KU. 1991. Two remarkable new lindsaeas from South America. Botanica Helvetica 101: 227–233. Kramer KU, Green PS. 1990. Pteridophytes and gymnosperms. In: Kubitzki K, ed. The families and genera of vascular plants. Berlin: Springer-Verlag, 1–404. Kramer KU, Tindale MD. 1976. The Lindsaeoid ferns of the Old World VII. Australia and New Zealand. Telopea 1: 91–128. Kuhn M. 1882. Die Gruppe der Chaetopterides unter den Polypodiaceen. Festschrift zu dem fünfzigjährigen Jubilaum der Königstädtischen Realschule zu Berlin 323–348. Kurita S, Nishida M. 1963. Cytotaxonomy of Japanese Sphenomeris. Journal of Japanese Botany 38: 4–7. Kutty SN, Bernasconi MV, Sifner F, Meier R. 2007. Sensitivity analysis, molecular systematics and natural history evolution of Scathophagidae (Diptera, Cyclorrhapha, Calyptratae). Cladistics 23: 64–83. Laamanen TR, Meier R, Miller MA, Hille A, Wiegmann BM. 2005. Phylogenetic analysis of Themira (Sepsidae, Diptera), sensitivity analysis, alignment, and indel treatment in a multigene study. Cladistics 21: 258–271. Lanjouw J, Baehni C, Robyns W, Rollins RC, Ross R, Rousseau J, Schulze GM, Smith AC, de Vilmorin R, Stafleu FA. 1956. International code of botanical nomenclature, adopted by the Eighth International Botanical Congress, Paris, July 1954. Regnum Vegetabile vol. 8. Lehtonen S. 2008. Phylogeny estimation and alignment via POY versus Clustal + PAUP*, a response to Ogden and Rosenberg (2007). Systematic Biology 57: 653–657. Lehtonen S, Christenhusz MJM. 2010. Historical herbarium specimens in plant molecular systematics – an example from the fern genus Lindsaea (Lindsaeaceae). Biologia 65: 204–208. Lehtonen S, Myllys L, Huttunen S. 2009. Phylogenetic analysis of non-coding plastid DNA in the presence of short inversions. Phytotaxa 1: 3–20. Lehtonen S, Tuomisto H. 2007. Ecological and genetic notes on Lindsaea digitata (Lindsaeaceae), a new fern species from western Amazonia. Nordic Journal of Botany 25: 145– 151. Lin S-J, Kato M, Iwatsuki K. 1990. Sporogenesis, reproductive mode, and cytotaxonomy of some species of Sphenomeris, Lindsaea, and Tapeinidium (Lindsaeaceae). American Fern Journal 80: 97–109. Lin S-J, Kato M, Iwatsuki K. 1994. A taxonomic study of the fern genus Sphenomeris (Lindsaeaceae) in Japan. Journal of Japanese Botany 69: 127–141. Lin S-J, Kato M, Iwatsuki K. 1996. Electrophoretic evidence for the origins of tetraploids and hybrids of Sphenomeris chinensis sensu lato (Lindsaeaceae) in Japan. Journal of Plant Research 109: 201–209. Lindgren AR, Daly M. 2007. The impact of length-variable data and alignment criterion on the phylogeny of Decapodiformes (Mollusca, Cephalopoda). Cladistics 23: 464–476. Liu K, Nelesen S, Raghavan S, Linder CR, Warnow T. 2008. Barking up the wrong treelength, the impact of gap penalty on alignment and tree accuracy. IEEE/ACM Transactions on Computational Biology and Bioinformatics. Available at http://doi.ieeecomputersociety.org/10.1109/ TCBB.2008.63 Löytynoja A, Goldman N. 2009. Uniting alignments and trees. Science 324: 1528–1529. McNeill J, Barrie FR, Burdet HM, Demoulin V, Hawksworth DL, Marhold K, Nicolson DH, Prado J, Silva PC, © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE Skog JE, Wiersema JH, Turland NJ. 2006. International Code of Botanical Nomenclature (Vienna Code). Regnum Vegetabile 146. Lichtenstein: A.R.G. Gantner Verlag. Maddison DR, Maddison WP. 2000. Macclade version 4, analysis of phylogeny and character evolution. Sunderland, MA: Sinauer Associates. Maddison WP, Knowles LL. 2006. Inferring phylogenies despite incomplete lineage sorting. Systematic Biology 55: 21–30. Manton I, Sledge WA. 1954. Observations on the cytology and taxonomy of the pteridophyte flora of Ceylon. Philosphical Transactions of the Royal Society of London. Series B, Biological Sciences 238: 127–185. Maxon WR. 1913. Studies of tropical American ferns 4. Contributions of the United States National Herbarium 17: 133–179. Maxon WR. 1947. New ferns from the northern Andes. Contributions from the Gray Herbarium of Harvard University 165: 69–76. Mehra PN, Khanna KR. 1959. Cytology of some Himalayan ferns. Journal of Genetics 56: 296–314. Meimberg H, Wistuba A, Dittrich P, Heubl G. 2001. Molecular phylogeny of Nepenthaceae based on cladistic analysis of plastid trnK intron sequence data. Plant Biology 3: 164–175. Mickel JT, Wagner WH, Lim Chen K. 1966. Chromosome observations on the ferns of Mexico. Caryologia 19: 95–102. Mildenhall DC. 1980. New Zealand late cretaceous and Cenozoic plant biogeography, a contribution. Palaeogeography, Palaeoclimatology, Palaeoecology 31: 197–233. Mitui K. 1976. Chromosome numbers of some ferns in the Ryukyu Islands. Journal of Japanese Botany 51: 33–41. Morton CV. 1959. A proposal regarding the conserved name Sphenomeris. Taxon 8: 29. Morton CV. 1967. On the publication of names by means of illustrations with analyses. Taxon 16: 119–121. Moyle RG, Chesser RT, Brumfield RT, Tello JG, Marchese DJ, Cracraft J. 2009. Phylogeny and phylogenetic classification of the antbirds, ovenbirds, woodcreepers, and allies (Aves, Passeriformes, infraorder Furnariides). Cladistics 25: 386–405. Nixon KC. 1999. The parsimony ratchet, a new method for rapid parsimony analysis. Cladistics 15: 407–414. Ogden TH, Rosenberg MS. 2007. Alignment and topological accuracy of the direct optimization approach via POY and traditional phylogenetics via ClustalW + PAUP*. Systematic Biology 56: 182–193. Patterson C. 1982. Morphological characters and homology. In: Hawksworth DL, ed. Problems of phylogenetic reconstruction. London: Academic Press, 21–74. Perrie L, Brownsey P. 2007. Molecular evidence for longdistance dispersal in the New Zealand pteridophyte flora. Journal of Biogeography 34: 2028–2038. Phillips AJ. 2006. Homology assessment and molecular sequence alignment. Journal of Biomedical Informatics 39: 18–33. Pichi-Sermolli REG. 1953. The nomenclature of some ferngenera. Webbia 9: 387–454. 349 Pichi-Sermolli REG. 1970. A provisional catalogue of the family names of pteridophytes. Webbia 25: 219–297. Pinna MCC. 1991. Concepts and tests of homology in the cladistic paradigm. Cladistics 7: 367–394. Presl CB. 1836. Tentamen pteridographiae, seu, genera filicacearum praesetim juxta venarum decursum et distributionem exposita. Prague: A. Haase. Presl CB. 1851. Epimeliae botanicae. Prague: A. Haase. Pryer KM, Schuettpelz E, Wolf PG, Scheider H, Smith AR, Cranfill R. 2004. Phylogeny and evolution of ferns (monilophytes) with a focus on the early leptosporangiate divergences. American Journal of Botany 88: 1118– 1130. Pryer KM, Smith AR, Skog JE. 1995. Phylogenetic relationships of extant ferns based on evidence from morphology and rbcL sequences. American Fern Journal 85 (Special Issue): 205–282. Quandt D, Stech M. 2004. Molecular evolution of the trnTUGU-trnFGAA region in bryophytes. Plant Biology 6: 545– 554. Rosenstock E. 1906. Beiträge zur Pteridophytenflora Südbrasiliens. Hedwigia 46: 57–144. Sanders JG. 2010. Program note, Cladescan, a program for automated phylogenetic sensitivity analysis. Cladistics 26: 114–116. Sang T, Crawford DJ, Stuessy TF. 1997. Chloroplast DNA phylogeny, reticulate evolution, and biogeography of Paeonia (Paeoniaceae). American Journal of Botany 84: 1120–1136. Schatz GE. 1996. Malagasy/Indo-Australo-Malesian phytogeographic connections. In: Lourenço WR, ed. Biogeography of Madagascar. Paris: Editions ORSTOM, 73–84. Schneider H. 1996. The root anatomy of ferns, a comparative study. In: Camus JM, Gibby M, Johns RJ, eds. Pteridology in perspective: pteridophyte symposium ‘95. Proceedings of the Holttum memorial pteridophyte symposium, Kew 1995. Kew: Royal Botanic Gardens, 271–283. Schneider H, Kenrick P. 2001. An early Cretaceous root-climbing epiphyte (Lindsaeaceae) and its significance for calibrating the diversification of polypodiaceous ferns. Review of Paleobotany and Palynology 115: 33– 41. Schneider H, Schuettpelz E, Pryer KM, Cranfill R, Magallón S, Lupia R. 2004. Ferns diversified in the shadow of angiosperms. Nature 428: 553–557. Schuettpelz E, Crusz AL, Windham MD, Pryer KM. 2008. The utility of nuclear gapCp in resolving polyploid fern origins. Systematic Botany 33: 621–629. Schuettpelz E, Korall P, Pryer KM. 2006. Plastid atpA data provide improved support for deep relationships among ferns. Taxon 55: 897–906. Schuettpelz E, Pryer KM. 2007. Fern phylogeny inferred from 400 leptosporangiate species and three plastid genes. Taxon 56: 1037–1050. Schuettpelz E, Pryer KM. 2008. Fern phylogeny. In: Ranker TA, Haufler CH, eds. The biology and evolution of ferns and lycophytes. Cambridge: Cambridge University Press, 395–416. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 350 S. LEHTONEN ET AL. Schulmeister S, Wheeler WC, Carpenter JM. 2002. Simultaneous analysis of the basal lineages of Hymenoptera (Insecta) using sensitivity analyses. Cladistics 18: 455–484. Seberg O, Petersen G. 2006. Mitochondrial DNA sequences in plant phylogenetics and evolution-symposium at XVII IBC, Vienna, Austria. Taxon 55: 833–835. Shaw J, Lickey EB, Beck JT, Farmer SB, Liu W, Miller J, Siripun KC, Winder CT, Schilling EE, Small RL. 2005. The tortoise and the hare II, relatively utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis. American Journal of Botany 92: 142–166. Simmons MP. 2004. Independence of alignment and tree search. Molecular Phylogenetics and Evolution 31: 874– 879. Small RL, Lickey EB, Shaw J, Hauk WD. 2005. Amplification of noncoding chloroplast DNA for phylogenetic studies in lycophytes and monilophytes with a comparative example of relative phylogenetic utility from Ophioglossaceae. Molecular Phylogenetics and Evolution 36: 509–522. Smith AR, Mickel JT. 1977. Chromosome counts for Mexican ferns. Brittonia 29: 391–398. Smith AR, Pryer KM, Schuettpelz E, Korall P, Schneider H, Wolf PG. 2006. A classification for extant ferns. Taxon 55: 705–431. Smith AR, Pryer KM, Schuettpelz E, Korall P, Schneider H, Wolf PG. 2008. Fern classification. In: Ranker TA, Haufler CH, eds. The biology and evolution of ferns and lycophytes. Cambridge: Cambridge University Press, 417–467. Sørensen MV, Sterrer W, Giribet G. 2006. Gnathostomulid phylogeny inferred from a combined approach of four molecular loci and morphology. Cladistics 22: 32–58. Spagna JC, Álvarez-Padilla F. 2008. Finding an upper limit for gap costs in direct optimization parsimony. Cladistics 24: 787–801. Steel M, Penny D. 2000. Parsimony, likelihood, and the role of models in molecular phylogenetics. Molecular Biology and Evolution 17: 839–850. Steel M, Pickett KM. 2006. On the impossibility of uniform priors on clades. Molecular Phylogenetics and Evolution 39: 585–586. Stevenson DW, Loconte H. 1996. Ordinal and familial relationships of pteridophyte genera. In: Camus JM, Gibby M, Johns RJ, eds. Pteridophyte symposium ‘95. Proceedings of the Holttum memorial pteridophyte symposium, Kew 1995. Kew: Royal Botanic Gardens, 435–467. Storchová H, Olson MS. 2007. The architecture of the chloroplast psbA-trnH non-coding region in angiosperms. Plant Systematics and Evolution 268: 235–256. Swofford DL. 2002. PAUP*, phylogenetic analysis using parsimony and other methods. Ver. 4.0b10. Sunderland, MA: Sinauer Associates. Swofford DL, Maddison WP. 1987. Reconstructing ancestral character states under Wagner parsimony. Mathematical Biosciences 87: 199–229. Taberlet P, Gielly L, Pautou G, Bouvet J. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Molecular Biology 17: 1105–1109. Tate JA, Simpson BB. 2003. Paraphyly of Tarasa (Malvaceae) and diverse origins of the polyploid species. Systematic Botany 28: 723–737. Terry MD, Whiting MF. 2005a. Mantophasmatodea and phylogeny of the lower neopterous insects. Cladistics 21: 240–257. Terry MD, Whiting MF. 2005b. Comparison of two alignment techniques within a single complex data set, POY versus Clustal. Cladistics 21: 272–281. Thiers B. 2010. [continuously updated]. Index herbariorum, a global directory of public herbaria and associated staff. New York: Botanical Garden’s Virtual Herbarium. Available at http://sweetgum.nybg.org/ih/ (accessed 8 July 2009). Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG. 1997. The ClustalX windows interface, flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 24: 4876–4882. Tryon AF, Bautista HP, da Silva Araújo I. 1975. Chromosome studies of Brazilian ferns. Acta Amazonica 5: 35–43. Tryon R, Tryon A. 1982. Ferns and allied plants with special reference to tropical America. New York: Springer-Verlag. Tuffley C, Steel M. 1997. Links between maximum likelihood and maximum parsimony under a simple model of site substitution. Bulletin of Mathematical Biology 59: 581–607. Tuomisto H. 1998. What satellite imagery and large-scale field inventories can tell about biodiversity patterns in Amazonian forests. Annals of the Missouri Botanical Garden 85: 48–62. Varón A, Vinh LS, Wheeler WC. 2010. POY version 4, phylogenetic analysis using dynamic homologies. Cladistics 26: 72–85. Vinh LS, Varón A, Wheeler WC. 2006. Pairwise alignment with rearrangements. Genome Informatics 17: 141–151. Wagner FS. 1980. New basic chromosome numbers for genera of neotropical ferns. American Journal of Botany 67: 733–738. Walker TG. 1966. A cytotaxonomic survey of the pteridophytes of Jamaica. Transactions of the Royal Society of Edinburgh 66: 169–237. Walker TG. 1973. Evidence from cytology in the classification of ferns. In: Jermy AC, Crabbe JA, Thomas BA, eds. The phylogeny and classification of the ferns. Botanical Journal of the Linnean Society 67 (Suppl. 1): 91–110. Wheeler WC. 1995. Sequence alignment, parameter sensitivity, and the phylogenetic analysis of molecular data. Systematic Biology 44: 321–331. Wheeler WC. 1996. Optimization alignment, the end of multiple sequence alignment in phylogenetics? Cladistics 12: 1–9. Wheeler WC. 2007. Chromosomal character optimization. Molecular Phylogenetics and Evolution 44: 1130–1140. Wheeler WC, Giribet G. 2009. Phylogenetic hypotheses and the utility of multiple sequence alignment. In: Rosenberg M, ed. Sequence alignment, methods, models, concepts, and strategies. Berkeley, CA: University of California Press, 95–104. Wheeler WC, Hayashi CY. 1998. The phylogeny of the extant chelicerate orders. Cladistics 14: 173–192. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE Whiting MF, Whiting AS, Hastriter MW, Dittmar K. 2008. A molecular phylogeny of fleas (Insecta, Siphonaptera), origins and host associations. Cladistics 24: 677–707. Wolf PG. 1995. Phylogenetic analyses of rbcL and nuclear ribosomal RNA gene sequences in Dennstaedtiaceae. American Fern Journal 85: 306–327. Wolf PG, Soltis PS, Soltis DE. 1994. Phylogenetic relationships of dennstaedtioid ferns, evidence from rbcL sequences. Molecular Phylogenetics and Evolution 3: 383–392. Zwickl DJ, Hillis DM. 2002. Increased taxon sampling greatly reduces phylogenetic error. Systematic Biology 51: 588–598. APPENDIX 1 Morphological characters. ROOTS 0. ANATOMY: (0) Dennstaedtia-type; (1) Lonchitistype; (2) Osmunda-type; (3) Lindsaea-type. Character state descriptions follow Schneider (1996). RHIZOMES 1. HABIT: (0) short; (1) long. This character is modified from character 1 of Barcelona (2000). 2. RHIZOME SYMMETRY: (0) radial; (1) dorsiventral. This character corresponds to character 26 of Pryer et al. (1995). 3. MATURE RHIZOME STELE TYPE: (0) protostele; (1) solenostele; (2) dictyostele. This character corresponds to character 27 of Pryer et al. (1995). 4. XYLEM WIDTH: (0) < 1/3 of rhizome diameter; (1) ! 1/2 of rhizome diameter. 5. LEAVES ISSUING AT: (0) right angle; (1) acute angle. 6. DIAMETER: (0) < 5 mm; (1) > 1 cm. This character is modified from character 4 of Barcelona (2000). INDUMENT 7. INDUMENT, ATTACHMENT: (0) normal; (1) inserted on mounds. 8. INDUMENT, NUMBER OF CELL SERIES AT THE BASE: (0) one; (1) 2 or 3; (2) 4–10; (3) > 10. This character is modified from character 18 of Barcelona (2000). 9. INDUMENT, LENGTH: (0) < 2.5 mm; (1) > 2.5 mm. 11. ARCHITECTURE: (0) determinate; (1) indeterminate. This character corresponds to character 20 of Barcelona (2000). STIPES 12. SHAPE IN CROSS SECTION: (0) subterete; (1) triangular; (2) quadrangular. This character corresponds to character 23 of Barcelona (2000). 13. NUMBER OF VASCULAR BUNDLES: (0) two– many; (1) single. This character is modified from character 28 of Barcelona (2000). 14. WINGS: (0) absent; (1) present. 15. LENGTH: (0) short, < 1/4 of the lamina; (1) long, ! 1/2 of the lamina. This character is modified from character 29 of Barcelona (2000). RACHISES 10. HABIT: (0) erect; (1) scrambling. This character corresponds to character 19 of Barcelona (2000). AND MINOR AXES 16. RACHIS ABAXIALLY: (0) terete; (1) bi-angular; (2) keeled. 17. SURFACE: (0) unarmed; (1) spinules/spiny. This character corresponds to character 45 of Barcelona (2000). 18. SPINE LENGTH: (0) < 1.5 mm; (1) 2–3 mm. This character was coded as inapplicable for the taxa without spines. 19. INDUMENT: (0) macroscopically hairy; (1) macroscopically glabrous. 20. ARCHITECTURE: (0) straight; (1) flexuose (zigzagging). This character corresponds to character 48 of Barcelona (2000). 21. NATURE OF SULCUS: (0) narrow; (1) broad; (2) absent. LAMINAE 22. DIMORPHISM: (0) monomorphic; (1) dimorphic. This character is modified from character 32 of Barcelona (2000). 23. TEXTURE: (0) herbaceous; (1) coriaceous. This character is modified from character 36 of Barcelona (2000). 24. DROMY AT THE BASE OF BLADE (LOWERMOST PINNAE): (0) catadromous; (1) anadromous. This character corresponds to character 14 of Pryer et al. (1995). 25. VEIN ORDERS: (0) one; (1) two; (2) three; (3) four or more. This character corresponds to character 5 of Pryer et al. (1995). PINNAE FRONDS 351 AND PINNULES 26. PINNAE ATTATCHMENT IN ADULTS: (0) opposite; (1) alternate. This character is modified from character 51 of Barcelona (2000). © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 352 S. LEHTONEN ET AL. 27. DISPOSITION: (0) ascending; (1) reflexed. This character is modified from character 50 of Barcelona (2000). 28. TERMINAL PINNA ARCHITECTURE: (0) conform; (1) confluent. This character corresponds to character 55 of Barcelona (2000). 29. PINNULE INSERTION: (0) at ridges; (1) below ridges. 30. LOWER PINNULES: (0) not much reduced; (1) strongly reduced; (2) aphlebioid. 31. PINNULES PENDULOUS: (0) no; (1) yes. 32. PINNAE SURFACE: (0) macroscopically hairy; (1) macroscopically glabrous. ULTIMATE SEGMENTS 33. SHAPE: (0) equal sided; (1) cuneate-rhombic; (2) dimidiate. This character is modified from character 75 of Barcelona (2000). 34. SEGMENT BASE: (0) non-auriculate (1) auriculate. 35. SEGMENT MARGIN: (0) entire; (1) erose– crenate; (2) incised. This character is modified from character 78 of Barcelona (2000). 36. SEGMENTS: (0) stalked; (1) sessile; (2) connected by wing. VENATION 37. VEINS: (0) running out to the margin; (1) stop short of the margin. 38. VEIN FUSION: (0) non-anastomosing; (1) anastomosing. This character is modified from character 7 of Pryer et al. (1995). Here, even nonconsistent formation of areoles is coded as anastomosing, as this happens only in certain species, while most of the species constantly have free veins. INDUSIA 39. ORIGIN OF INDUSIA OR INDUSIAL COMPONENTS: (0) leaf margin; (1) abaxial leaf surface. This character corresponds to character 53 of Pryer et al. (1995). 40. FALSE INDUSIUM: (0) present; (1) absent. This character corresponds to character 57 of Stevenson & Loconte (1996). 41. INDUSIAL OPENING: (0) introrse; (1) extrorse; (2) suprasoral. This character corresponds to character 55 of Pryer et al. (1995). 42. EXTENT OF LATERAL ADNATION: (0) free (1) adnate. This character is modified from character 84 of Barcelona (2000). 43. SORI LOCATION: (0) inner to outer part of the segment margin; (1) only in the outer part of the upper segment margin. 44. INDUSIUM: (0) reaching margin; (1) intramarginal; (2) longer than and protruding beyon lamina. This character is inapplicable for taxa with leaf marginal origin of indusium. 45. INDUSIUM SHAPE: (0) obovate to ovate; (1) elliptic; (2) horseshoe-shaped; (3) flabellate. This character is modified from character 85 of Barcelona (2000). 46. INDUSIUM REFLEXED: (0) no; (1) yes. 47. INDUSIUM MARGIN: (0) erose; (1) entire. This character is modified from character 92 of Barcelona (2000). 48. INDUSIUM WIDTH: (0) less than 0.3 mm; (1) 0.3–0.6 mm; (2) 0.7–1 mm. This character is modified from character 94 of Barcelona (2000). SPORANGIA 49. SIZE: (0) < 160 mm; (1) 200–250 mm; (2) > 250 mm. 50. NUMBER OF SPORANGIA/SORUS: (0) few (< 12); (1) many (usually > 20). This character corresponds to character 51 of Pryer et al. (1995). 51. ANNULUS CELLS: (0) 9–12; (1) 15–23. SPORES 52. TYPE: (0) monolete; (1) trilete. This character corresponds to character 107 of Barcelona (2000). 53. SIZE: (0) 18–35 mm; (1) 35–43 mm; (2) 50–60 mm. This character is modified from character 109 of Barcelona, (2000). 54. PERISPORE (EPISPORE) SURFACE: (0) (nearly) smooth or plain; (1) obviously patterned or sculptured. This character corresponds to character 64 of Pryer et al. (1995). © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE 353 APPENDIX 2 Morphological data matrix. Polymorphic entries: a = [01], b = [12], c = [13]. Inapplicable data = ‘–’. Missing data = ‘?’. 111111111122222222223333333333444444444455555 0123456789012345678901234567890123456789012345678901345 Nesolindsaea caudata Nesolindsaea kirkii Cystodium sorbifolium Lindsaea agatii Lindsaea annamensis Lindsaea apoensis Lindsaea arcuata Lindsaea austrosinica Lindsaea azurea Lindsaea bifida 1 Lindsaea bifida 2 Lindsaea blanda Lindsaea blotiana 2 Lindsaea blotiana 2 Lindsaea bolivarensis Lindsaea borneensis Lindsaea botrychioides Lindsaea bouillodii Lindsaea brachypoda Lindsaea brevipes Lindsaea cf. cambodgensis Lindsaea carvifolia Lindsaea chienii Lindsaea chrysolepis Lindsaea coarctata 1 Lindsaea coarctata 2 Lindsaea crispa Lindsaea cubensis Lindsaea cultrata Lindsaea cyclophylla Lindsaea digitata Lindsaea diplosora Lindsaea dissectiformis Lindsaea divaricata Lindsaea divergens Lindsaea doryphora 1 Lindsaea doryphora 2 Lindsaea dubia Lindsaea ensifolia 1 Lindsaea ensifolia 2 Lindsaea falcata Lindsaea feei Lindsaea cf. fissa Lindsaea fraseri Lindsaea gomphophylla Lindsaea goudotiana Lindsaea grandiareolata Lindsaea gueriniana Lindsaea guianensis 301100011100010100-001001210100002000101111003010111110 301100001100010100-100001210100002010101110021011111110 ?00201100100000100-00000031010001001200a021020002211111 301000001000210110-101001210000000000111110003011010100 301000001000210110-100001210100001020101110001001010100 311011003100010010-102001110110002020101110001111010100 301000001000210110-100001210000002000101110003110010100 301000001000211110-101001210000001010101110013001010100 301000001000010110-101001210100002010111111013110010100 301000001000210110-101001210100001020101110001110010100 301000001000210110-101001210100001020101110001110010100 311000001000010010-100001110100002010101110011111010100 311000002000011110-1010012100000010101011100020010?0100 311000002000010110-101001210100001010101110011101010?00 301000001000211110-101001210000002000101110013110010100 311011001100010100-100001210100002000101110003110010100 301000001000210a10-100001110000001010101110003101010100 301000001000010110-101001210100001010101110001001010100 301000002000210010-100101b10100002000101110001011010100 301000001000210010-101001110101002010101110001011010100 301000001000010110-100001210100001010101110011001010100 311011003100010010-102001110111002020101110001110010100 301000002000010110-101001210100001010101110011001010100 311000003100010010-101001110100002010111111001011010100 301000001000010110-101001210000002000101110013110010100 301000001000010110-101001210000002000101110013110010100 301000001000010110-100001210100002010111110003011010100 301000001000210110-100101210000001000101110013001010000 301000001000111110-101001210000002000111110001110010100 3010000010000111-0-1010011-00---0--0-101110013111010000 301000001000011110-101001210000002000101110013110010100 311000003100010100-101001210100002010101111101011010100 301000001000010110-100001210100001020101110011001010100 301000001000011110-101001210000002000101110013110010100 301000002000010020-101001110001000101101110003010010100 301001003000010a00-100001210000002000101110013110010100 301001003000010a00-100001210000002000101110013110010100 301000001000211110-101001110000002010101110003110010100 301000001000210110-100001210000000000111110003011010100 301000001000210110-100001210000000000111110003011010100 301000001000211110-101001110000002000101110013110010100 301000001000010110-100001210100002010101110011101010100 311011003100010010-102001110101002020101110001011100100 301000001000010110-100001210000000010111110003011?1???? 301000001000011100-101001110100001010101110003001010100 311000001000010010-101001210100001020101110022011010100 301000001000011110-101001210000000000111110003011011100 301000002000010120-101011110100000100101110003011111100 301000001000010100-100001210100002000101110013110010100 © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 354 S. LEHTONEN ET AL. APPENDIX 2 Continued 111111111122222222223333333333444444444455555 0123456789012345678901234567890123456789012345678901345 Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea hainaniana harveyi hemiglossa 1 hemiglossa 2 hemiptera heterophylla imrayana 1 imrayana 2 incisa integra 1 integra 2 jamesonioides javanensis 1 javanensis 2 cf. javitensis kawabatae kingii lancea var. lancea lancea var. leprieurii 1 lancea var. leprieurii 2 lancea var. submontana lapeyrousei 1 lapeyrousei 2 leptophylla lherminieri linearis lobata longifolia lucida madagascariensis malayensis 1 malayensis 2 media 1 media 2 meifolia merrillii ssp. merrillii merrillii ssp. yaeyamensis microphylla millefolia 1 millefolia 2 monocarpa multisora nervosa oblanceolata obtusa 1 obtusa 2 orbiculata ovoidea oxyphylla pacifica pallida parallelogramma 301000001000010110-100001210100002010111110001111010100 301000001000010110-100001210000002010111110011011010100 301000001000010110-101001210000002000101110013110010100 301000001000010110-101001210000002000101110013110010100 301000001000011100-101001210100002000101110003010010100 301000001000010110-100001210000000010111110003011010100 311000003100010110-1000012-0000000010101110011111010100 311000003100010110-1000012-0000000010101110011111010100 301000002010210010-101001110100001000101110001002111100 301000001000210010-101001b10000002000111110003111010100 301000001000210010-101001b10000002000111110003111010100 301000001100110020-101011110000000101101110003012111100 301000001000010110-100001210000001010101110003011010100 301000001000010110-100001210000001010101110003011010100 301000001000010110-101011210100002000101110003000110100 301000001000010110-101001210100001010101110011001010100 301000002100210110-100001210101002010101110001001010100 301000001000210110-101001210000002000101110013110010100 301000001000011110-101001110000002000101110013110010100 301000001000011110-101001110000002000101110013110010100 301000001000211110-101001110000002000101110003110010100 301000001000210010-101001110101002020101111001001010100 301000001000210010-101001110101002020101111001001010100 311000001000011000-101001210000001010101110021011010100 301000001000210110-101001210000002000101110013110010100 301000001000010a00-101101110000001000101110003002011110 301000001000010110-100001210100002010111110001111010100 301000001000010110-101001210100002010111110001011010100 301000001000210010-100001110100002010101110001011010100 311000002000010110-101001210100001010101110011101010?00 301000001000010110-100001210100002010111110011110010100 301000001000010110-101001210100002010111110001010010100 301000001000010110-100001210000000000111110003011?1???? 301000001000010110-1000012101000010101a1110001001010100 301000001000010100-101011210100101000101110000011011110 311011003100010010-101001110101002010101110011010000100 311011003100010010-101001110101002010101110001010000100 301000002100010000-101001210100001010101110001001011100 301000002000011110-101001210100001020101110011110010100 301000002000011110-101001210100001020101110011110010100 311000001100010110-101011210100002010111111121001010100 301000001000210110-1010012101000020101011100011100?0100 301000002000010110-100001210000001010101110001100010100 311011003100010010-102001110a11002000101110003111010100 301000001000010110-101001210100002010111110001111010100 301000001000010110-101001210100002010111110001111010100 301000001000010a10-100001210100001010101110003001010100 301000001000010110-100001210100001010101110001101010100 301000001000010100-101001110100000010101110023000010??? 301000001000010110-1010012101000020101111100010110?0100 301000001000010110-100001210000002000101110001001010010 301000001000211110-101001210100002010111110001010010100 © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE 355 APPENDIX 2 Continued 111111111122222222223333333333444444444455555 0123456789012345678901234567890123456789012345678901345 Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea parasitica parkeri pectinata pellaeiformis pendula phassa pickeringii plicata polyctena portoricensis 1 portoricensis 2 pratensis prolongata propinqua pseudohemiptera pulchella pulchra quadrangularis ssp. antillensis regularis reniformis rigida rigidiuscula 1 rigidiuscula 2 rosenstockii rufa sagittata schizophylla schomburgkii seemannii semilunata sessilis sp. 1 sp. 2 sp. 3 sphenomeridopsis spruceana stolonifera stricta var. jamesoniiformis stricta var. parvula stricta var. stricta subtilis surinamensis taeniata tenuifolia tenuis tetragona tetraptera trichomanoides ulei venusta vieillardii virescens 311011003100010a00-101001210000002000101110003111010100 301000001000011110-101001210000002000101110003111010100 311011003100010010-102001110111002010101110011010010100 301000002100010a20-101011110100000100101110003011111100 301000001000010100-101011210100101000101110001001110110 301000001000210110-101001210000002000101110013110010100 311010003000010010-1010011101010020101a1110011111010100 30100000?000010020-101001110100000010101110001012111120 301000001000210110-101001210100002020101110011110010100 301000001000010a10-101011210100002000101110013101010100 301000001000010a10-101011210100002000101110013101010100 301000003100010020-101011110101002000101111003001111020 301000001000010110-100001210000001010101110001010010100 311000001100010110-100001210100002010111110001011010100 311011003100010010-102001110111002010101110011110010100 311000001000010010-1010011101000020101a1110011011010100 311000001000210010-101001110100002010111110001111010100 301000001000210110-100001210000002000101110003110010100 311011003100010a10-101001b10100002010101110001111010100 3010000010000111-0-1010011-00---0--0-101110013111010100 311000001100010110-1010112101000020101a1111101001110100 301000001000211110-101011210000002000101110003110110100 301000001000211110-101011210000002000101110003110110100 311010003100010110-100001210100002020101110011110010100 301000001000010110-100001210000001010101110001101010100 3010000010000111-0-1010011-00---0--0-101110013111010100 311000001000010110-101001210100001010101110001001010100 301000001000211110-101011110000002000101110003110010100 301000003100010020-101001110101002010101111003011111020 301000001000211110-101001110000002000101110003110010100 311011003100010010-102001110111002010101110011110010100 311010003100010110-1010012101000020101011100111100?0100 311000001100010110-1010112101000020101a1111101001110100 311000001000010100-100001210100001010101110001001010100 301000001000011110-101001210100001020101110011001010100 301000001000010100-101001210100002000101110003110010100 311000001000210010-101001110100002010111110001111010100 301000001000010a00-101011210100002000101110003101110110 301000001000010a00-101011210100002000101110003101110110 301000001000010100-101011210100002000101110003101110110 311000002000010010-100001110100001010101110022011010100 301000001000111110-101001110000002000101110013010010100 301000001000210110-100001210000002000101110013010010100 301000001000110120-101001210100002020101110011110010100 301000001010010010-101001210100002010101110011101010100 301000001000210110-101001210100002020101110001110010100 301000001000011110-101001210000002000101110003010010100 301000003000010110-100001210100001010101110001001010100 301000001000011100-101001110000002000101110013110010100 301000001000210110-101001210100002010111110001011010100 301000002000210110-101001210000000010111110003011010100 301000001000210110-101001210100001010101110001100010100 © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 356 S. LEHTONEN ET AL. APPENDIX 2 Continued 111111111122222222223333333333444444444455555 0123456789012345678901234567890123456789012345678901345 Lindsaea viridis Lindsaea vitiensis Lindsaea walkerae Lindsaea werneri Lonchitis hirsuta Odontosoria aculeata Odontosoria afra Odontosoria africana Odontosoria angustifolia Odontosoria biflora Odontosoria chinensis Odontosoria deltoidea 1 Odontosoria deltoidea 2 Odontosoria flexuosa Odontosoria fumarioides Odontosoria guatemalensis Odontosoria jenmanii Odontosoria melleri Odontosoria retusa Odontosoria scandens Odontosoria schlechtendalii Odontosoria wrightiana Osmolindsaea japonica Osmolindsaea odorata 1 Osmolindsaea odorata 2 Osmolindsaea odorata 3 Osmolindsaea sp. Pteridium pinetorum Saccoloma elegans Saccoloma inaequale Sphenomeris clavata 1 Sphenomeris clavata 2 Tapeinidium amboynense Tapeinidium calomelanos Tapeinidium denhamii Tapeinidium gracile Tapeinidium longipinnulum Tapeinidium luzonicum Tapeinidium melanesicum Tapeinidium moorei Tapeinidium novoguineense Tapeinidium pinnatum 301000001100010020-100001110100001020101110001002111020 311000001000210010-101001110100002010101110011111010100 301000001000010100-101011210000000000101110003011111100 311000001000210110-101001110100002010111110011111010100 101100100000000100-00100031010001000210010-0-----211121 3111000011110100011111011301102001010101111020011101120 301000002100010000-1010013101000010201011110210012011b0 301000002110010000-110011311100001020101111020001101110 301?00103000010100-100011310100001020101111001001101110 301000001100010100-101011310100001020101111000012101010 301000002100010100-10101131010000102010111100a001201010 311100102100010100-100011310100001010101111021001111110 311100102100010100-100011310100001010101111021001111110 3111000011110100010111001301102001010101111000001201120 3111000011110100011111011301102001020101111000012201120 311100001111010000-111001301102001020101111020012201120 3111000031110100011111011301102001010101111021012101120 301100101100010100-110011310100001010101110021011201110 301100103100010100-10001131010000101010111000c001201110 3111000011110100010111011301102001010101111001012101120 31110000?111010000-111001301102001020101111021012101120 3111000011110100011110011301102001010101111020011101120 301100001000010000-100001110000002000101111003012111010 301100001000010000-100001110100002010101111001001111010 301100001000010000-100001110100002010101111001001111010 301100001000010000-100001110101002010101111001002111010 301100001000010000-100001110101002010101111001001111010 011100100000000100-00000031000001000200000-0-----111101 200201103100010100-1000012-0000000000001110001101211101 200201103100010100-100001310100000012001111000002201101 301000002100010100-110001310100001020101111021002111120 301000002100010100-110001310100001020101111021002111120 301100001100010110-101011310100000012101111010011101000 301000001000010100-101011310100000012101111001011111000 301000001000110110-101001310100000012101110020001101000 301100001100110110-1010112101000000101011110100011?1000 301100003100210110-101011210000000010101111001001111000 301100001100110110-1010113101000000121011110100111?1010 301100001000010100-101011210000000010101111011001111000 301000011000010100-100011310100000012101111001001111020 311100003100010100-101011310100000012101111011011101000 301100001100110110-1010112101000000101011110100011?1000 APPENDIX 3 DNA-structure characters. 0. 4-bp INVERSION IN trnL-trnF: (0) GT-type; (1) CA-type. 1. 7-bp INVERSION IN trnL-trnF: (0) CT-type; (1) GA-type. 2. 13-bp INVERSION IN trnH-psbA: (0) T-type; (1) A-type. © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE 357 APPENDIX 4 DNA-structure data matrix. Polymorphic entries: a = [01]. Missing data = ‘?’. Nesolindsaea caudata Nesolindsaea kirkii Cystodium sorbifolium Lindsaea agatii Lindsaea annamensis Lindsaea apoensis Lindsaea arcuata Lindsaea austrosinica Lindsaea azurea Lindsaea bifida 1 Lindsaea bifida 2 Lindsaea blanda Lindsaea blotiana 1 Lindsaea blotiana 2 Lindsaea bolivarensis Lindsaea borneensis Lindsaea botrychioides Lindsaea bouillodii Lindsaea brachypoda Lindsaea brevipes Lindsaea cf. cambodgensis Lindsaea carvifolia Lindsaea chienii Lindsaea chrysolepis Lindsaea coarctata 1 Lindsaea coarctata 2 Lindsaea crispa Lindsaea cubensis Lindsaea cultrata Lindsaea cyclophylla Lindsaea digitata Lindsaea diplosora Lindsaea dissectiformis Lindsaea divaricata Lindsaea divergens Lindsaea doryphora 1 Lindsaea doryphora 2 Lindsaea dubia Lindsaea ensifolia 1 Lindsaea ensifolia 2 Lindsaea falcata Lindsaea feei Lindsaea cf. fissa Lindsaea fraseri Lindsaea gomphophylla Lindsaea goudotiana Lindsaea grandiareolata Lindsaea gueriniana Lindsaea guianensis Lindsaea hainaniana Lindsaea harveyi Lindsaea hemiglossa 1 Lindsaea hemiglossa 2 Lindsaea hemiptera 000 000 00? 010 000 010 010 000 010 000 000 010 000 000 010 010 010 000 000 010 010 010 000 010 010 010 010 000 000 010 010 010 000 000 010 010 010 010 000 ??? 010 010 010 010 010 00? 000 000 010 010 010 010 010 000 Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea heterophylla imrayana 1 imrayana 2 incisa integra 1 integra 2 jamesonioides javanensis 1 javanensis 2 cf. javitensis kawabatea kingii lancea var. lancea lancea var. leprieurii 1 lancea var. leprieurii 2 lancea var. submontana lapeyrousei 1 lapeyrousei 2 leptophylla lherminieri linearis lobata longifolia lucida madagascariensis malayensis 1 malayensis 2 media 1 media 2 meifolia merrillii ssp. merrillii merrillii ssp. yaeyamensis microphylla millefolium 1 millefolium 2 monocarpa multisora nervosa oblanceolata obtusa 1 obtusa 2 orbiculata ovoidea oxyphylla pacifica pallida parallelogramma parasitica parkeri pectinata pellaeiformis pendula phassa pickeringii © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 0a? 010 ??? 010 000 000 010 010 010 000 000 01? 010 010 010 010 010 ??? 010 010 001 010 010 010 010 010 010 010 010 010 010 010 010 000 00? 010 010 00? 010 010 010 000 010 010 010 010 000 010 010 010 010 010 000 010 358 S. LEHTONEN ET AL. APPENDIX 4 Continued Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea Lindsaea plicata polyctena portoricensis 1 portoricensis 2 pratensis prolongata propinqua pseudohemiptera pulchella pulchra quadrangularis ssp. antillensis regularis reniformis rigida rigidiuscula 1 rigidiuscula 2 rosenstockii rufa sagittata schizophylla schomburgkii seemannii semilunata sessilis sp. 1 sp. 2 sp. 3 sphenomeridopsis spruceana stolonifera stricta var. jamesoniiformis stricta var. parvula stricta var. stricta subtilis surinamensis taeniata tenuifolia tenuis tetragona tetraptera trichomanoides ulei venusta vieillardii 000 110 000 010 010 000 010 010 010 010 010 010 010 0?0 010 010 010 000 010 010 010 010 010 010 010 010 010 010 01? 010 010 000 010 000 010 010 110 010 010 000 000 010 010 010 Lindsaea virescens Lindsaea viridis Lindsaea vitiensis Lindsaea walkerae Lindsaea werneri Lonchitis hirsuta Odontosoria aculeata Odontosoria afra Odontosoria africana Odontosoria angustifolia Odontosoria biflora Odontosoria chinensis Odontosoria deltoidea 1 Odontosoria deltoidea 2 Odontosoria flexuosa Odontosoria fumarioides Odontosoria guatemalensis Odontosoria jenmanii Odontosoria melleri Odontosoria retusa Odontosoria scandens Odontosoria schlechtendalii Odontosoria wrightiana Osmolindsaea japonica Osmolindsaea odorata 1 Osmolindsaea odorata 2 Osmolindsaea odorata 3 Osmolindsaea sp. Pteridium pinetorum Saccoloma elegans Saccoloma inaequale Sphenomeris clavata 1 Sphenomeris clavata 2 Tapeinidium amboynense Tapeinidium calomelanos Tapeinidium denhamii Tapeinidium gracile Tapeinidium longipinnulum Tapeinidium luzonicum Tapeinidium melanesicum Tapeinidium moorei Tapeinidium novoguineense Tapeinidium pinnatum 000 01? 010 000 010 00010 010 010 010 010 010 01? 01? 010 000 000 000 010 0?0 000 000 010 00? 00? 000 000 000 000 0000001 001 000 000 000 000 000 000 000 000 000 000 © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359 PHYLOGENETICS OF LINDSAEACEAE APPENDIX 5 Command lines applied. Morphological data analysis using TNT: log filename; mxram 256; p datafile; taxname=; out Pteridium_pinetorum; hold 100000; mult=ratchet replic 500 hold 10; length; tsave *filename; save; tsave /; nelsen; tsave *filename; save; tsave /; quit; Jackknife analyses using TNT: log filename; mxram 256; p datafile; taxname=; out Pteridium_pinetorum; hold 100000; resample jak replications 100 [mult=ratchet replic 100 hold 1]; quit; Molecular and combined analyses using POY: read(“datafiles”) report(“filename”,timer:“search start”) transform(tcm:“111.txt”) set(seed:1,log:“filename”,root:“Pteridium_ pinetorum”) build(500) 359 swap(threshold:5.0) select() perturb(transform(static_approx),iterations:20, ratchet:(0.25,3)) select() fuse(iterations:200,swap()) select() report(“filename”,treestats) report(“filename”,trees) report(“filename”,consensus) report(“filename”,graphconsensus) report(“filename”,diagnosis) report(“filename”,ia) report(“filename”,cross_references:all) report(“filename”,timer:“search_end”) set(nolog) exit() Sensitivity analyses using POY: read(“datafiles”) set(seed:1,log:“filename”,root:“Pteridium_ pinetorum”) store(“original_data”) transform(tcm:“filename”) transform((names:(“filename”),weightfactor:N)) build(20) swap(timeout:3600) select() perturb(transform(static_approx),iterations:10, ratchet:(0.25,3), swap(timeout:20000)) select() fuse(iterations:200,swap()) select() report(“filename”,trees:(total),“filename”,consensus,“filename”,graphconsensus) use(“original_data”) . . . © 2010 The Linnean Society of London, Botanical Journal of the Linnean Society, 2010, 163, 305–359