bs_bs_banner Botanical Journal of the Linnean Society, 2012, ••, ••–••. With 3 figures Molecular evolution and diversification of the moss family Daltoniaceae (Hookeriales, Bryophyta) with emphasis on the unravelling of the phylogeny of Distichophyllum and its allies BOON-CHUAN HO1,2*,†, LISA POKORNY2†, BENITO C. TAN3, JAN-PETER FRAHM1, A. JONATHAN SHAW2 and DIETMAR QUANDT1 1 Nees-Institut für Biodiversität der Pflanzen, Rheinische Friedrich-Wilhelms-Universität Bonn, Meckenheimer Allee 170, D-53115 Bonn, Germany 2 Department of Biology, Duke University, Durham, NC 27708, USA 3 Department of Biological Sciences, National University of Singapore, Singapore 119260 Received 19 January 2012; revised 4 May 2012; accepted for publication 31 May 2012 Phylogenetic relationships in Daltoniaceae (~200 species in 14 genera) are inferred from nucleotide sequences from five genes, representing all genomic compartments, using parsimony, likelihood and Bayesian methods. Alternative classifications for Daltoniaceae have favoured traits from either sporophytes or gametophytes; phylogenetic transitions in gametophytic leaf limbidia and sporophytic exostome ornamentation were evaluated using ancestral state reconstruction to assess the levels of conflict between these generations. Elimbate leaves and the cross-striate exostome are reconstructed as plesiomorphic states. Limbate leaves and papillose exostomes evolved at least two and six times, respectively, without reversals. The evolution of leaf limbidia is relatively conserved, but exostome ornamentation is highly homoplasious, indicating that superficial similarity in peristomes gives unreliable approximations of phylogenetic relatedness. Our phylogenetic analyses show that Achrophyllum and Calyptrochaeta are reciprocally monophyletic. Within core Daltoniaceae, relationships among taxa with elimbate leaves are generally well understood. However, taxa with limbate leaves form a monophyletic group, but resolved subclades correspond to biogeographical entities, rather than to traditional concepts of genera. Daltonia (~21 species), Distichophyllum (~100 species) and Leskeodon (~20 species) are polyphyletic. Seven nomenclatural changes are proposed here. As the current taxonomy of Daltoniaceae lacks phylogenetic consistency, critical generic revisions are needed. © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–••. ADDITIONAL KEYWORDS: Achrophyllum – ancestral states – Calyptrochaeta – Daltonia – exostome ornamentation – Leskeodon – limbidium. INTRODUCTION Insights into phylogenetic relationships enable a better understanding of morphology and morphological evolution. Because all land plants are characterized by an alternation of sporophytic and gametophytic generations, a phylogenetic context can help to clarify how evolutionary change has occurred in the two *Corresponding author. E-mail: calyptrochaeta@yahoo.com †These authors contributed equally in partial fulfilment of the requirements for their respective doctoral degrees. generations, in concert and independently. In mosses (Bryophyta), conflicting classifications with an emphasis on either sporophytes or gametophytes are common because both generations are relatively well developed and provide characters that may be useful for phylogenetic inference and classification. Daltoniaceae (Hookeriales: Bryophyta) exemplifies the systematic challenge that may result from focusing taxonomic inferences on traits from just one generation (i.e. gametophytes vs. sporophytes). Daltoniaceae comprises a prominent group of tropical and southtemperate mosses that prefer humid forest habitats. © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• 1 2 B. -C. Ho ET AL. Members of the family can be found in a wide range of habitats from terrestrial to epiphytic and occasionally epiphyllous (Fleischer, 1908; Miller, 1982; Buck, 1998). A few taxa are facultatively aquatic, being entirely submerged under water (de Winton & Beever, 2004), and there are reports of an epizoic species of Daltoniaceae living on the backs of weevils (Gressitt, Samuelson & Vitt, 1968; Gradstein, Vitt & Anderson, 1984). The family (~200 species in 14 genera) is characterized by: (1) plants sparingly branched and usually complanate foliate; (2) foliate shoots not well differentiated into primary and secondary axes; (3) leaves with one midrib or costa (unicostate), rarely none (ecostate only in Distichophyllidium M.Fleisch.); (4) marginal laminal cells mostly differentiated as a distinct border of elongate cells (limbate); (5) median laminal cells ± isodiametric; and (6) calyptra mitrate. About half the genera include only one or two species each (Adelothecium Mitt., Beeveria Fife, Benitotania H.Akiyama, T.Yamag. & Suleiman, Bryobrothera Thér., Crosbya Vitt., Ephemeropsis K.I.Goebel, Metadistochophyllum Nog. & Z.Iwats. and Leskeodontopsis Zanten), whereas the largest genus, Distichophyllum Dozy & Molk., contains about half the species in the family. The circumscription of Daltoniaceae and the relationships among genera are not entirely settled. In particular, the inclusion of Calyptrochaeta Desv. in the family still awaits confirmation. The relationships between Achrophyllum Vitt & Crosby, Calyptrochaeta and the rest of Daltoniaceae are uncertain. Furthermore, infrageneric relationships within Achrophyllum have never been assessed. A first approach exists for Calyptrochaeta (Pokorny, Oliván & Shaw, 2011). Within the well-supported core Daltoniaceae (i.e. Daltoniaceae s.l., excluding Calyptrochaeta and Achrophyllum), relationships among the genera, especially with regard to the species-rich Distichophyllum, are still in question. An understanding of morphological evolution in these mosses requires a better understanding of phylogenetic relationships (Buck, Cox & Shaw, 2005). HISTORY OF DALTONIACEAE The traditional widely adopted and broadly defined Hookeriaceae, which includes Daltoniaceae, was first proposed by Fleischer (1908). According to Fleischer, Daltonieae [with Daltonia and Crosbya (as Bellia Broth.)] and Distichophylleae [with Achrophyllum (as Pterygophyllum Broth.), Adelothecium, Calyptrochaeta (as Eriopus (Brid.) Brid.), Distichophyllidium M.Fleisch., Distichophyllum and Leskeodon Broth.] were recognized as two separate, but closely related, tribes with unicostate leaves. In general, Fleischer classified groups below the ordinal level primarily on gametophytic similarities and above that on sporophytic characters. Brotherus (1925) followed Fleischer’s classification, but recognized the tribes as subfamilies in Hookeriaceae. Subsequent systematic views differed in their emphasis on sporophytic vs. gametophytic traits for inferring relationships in the group. Crosby (1974) proposed a novel classification in which genera were grouped according to two basic peristome types (teeth surrounding the mouth of the sporophytic capsule or sporangium), regardless of their gametophytic similarities. The so-called hookeriaceous peristome has the exostome teeth (outer peristome) horizontally cross-striolate on the outer surfaces near their bases (Fig. 1C), endostomes (inner peristome) with high basal membranes and finely papillose segments and cilia absent to rudimentary. The daltoniaceous peristome, in contrast, has exostome teeth papillose throughout (Fig. 1D), endostomes with low or absent basal membranes and low papillose segments and cilia absent (Crosby, 1974). The genera Leskeodon and Crosbya were initially segregated from Distichophyllum and Daltonia, respectively, because of their dissimilar peristome (Brotherus, 1907). In Crosby’s classification, Daltonia, Distichophyllidium, Leskeodon, Leskeodontopsis and other ecostate and bicostate genera that share daltoniaceous peristomes were classified into one family, Daltoniaceae, whereas genera characterized by hookeriaceous peristomes were classified into Hookeriaceae. As a result, both families were highly heterogeneous in gametophytic structure. Buck (1987, 1988) took the opposite approach, emphasizing gametophytic rather than sporophytic features, and re-circumscribed Daltoniaceae to include the neotenic Ephemeropsis (gametophytes consisting of little more than protonemata, and a hookeriaceous peristome). As a consequence, the families distinguished in Hookeriales were heterogeneous with regard to hookeriaceous and daltoniaceous peristome morphologies. Hedenäs (1996) conducted a cladistic study on Hookeriales based on 75 morphological characters, about one-third of which came from the sporophyte and 14 came from the peristome. Not surprisingly, genera with similar peristomes were not grouped together, but few relationships received strong bootstrap support (BS). Buck et al. (2005) resolved the relationships within Hookeriales using nucleotide sequences from four genes. In general, this analysis supported the view that peristome characters can be homoplasious and should not, a priori, be relied upon to infer relationships. They found that Adelotheciaceae is phylogenetically nested in Daltoniaceae, but relationships among Achrophyllum, Calyptrochaeta and the rest of Daltoniaceae were ambiguous. Although only two Distichophyllum spp. were sampled, the genus appeared to be non-monophyletic. © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• EVOLUTION AND PHYLOGENY OF DALTONIACEAE 3 Figure 1. Characters for which character evolution was reconstructed using BayesTraits. A, Undifferentiated marginal laminal cells in mid-leaf in Benitotania elimbata (Suleiman 1901). B, Differentiated marginal laminal cells in mid-leaf in Daltonia marginata (Juri 10). C, Horizontally cross-striate ornamentation on the lower outside of the exostome in Hookeria acutifolia (Robbinson & Sharp 1955). D, Papillose ornamentation on the lower outside of the exostome in Daltonia angustifolia (Brass 20740). © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• 4 B. -C. Ho ET AL. EVOLUTION OF TAXONOMICALLY RELEVANT MORPHOLOGICAL CHARACTERS IN DALTONIACEAE Two characters have been considered to be especially important for generic separation in this group, i.e. gametophytic leaf limbidia (elimbate vs. limbate, see Fig. 1A,B) and sporophytic exostome ornamentation (cross-striate vs. papillose, see Fig. 1C,D). Gametophytic leaf limbidia, differentiated leaf borders consisting of narrower elongated, thick-walled cells (Fig. 1B), have long been associated with Hookeriales (Hedenäs, 1999). The character is commonly used to separate the genera of Hookeriaceae s.l. (e.g. Brotherus, 1907, 1925), although the functional importance of limbidia is obscure. There is now general agreement that sporophytic features are not as stable, reliable or informative for inferring phylogenetic relationships as formerly supposed (e.g. Hedenäs, 2001, 2002; Vanderpoorten et al., 2002; Huttunen et al., 2004; Olsson et al., 2009a; Quandt et al., 2009; Liu, Budke & Goffinet, 2012; Pokorny et al., 2012). Nonetheless, Buck (1991, 2007) considered that peristome features could still be valuable for distinguishing genera in some families. In Hookeriales, ancestral state reconstructions using phylogenetic trees inferred from molecular data indicate that exostome ornamentation is homoplasious (Pokorny et al., 2012). This result corroborates earlier suggestions by Buck (1987), Whittemore & Allen (1989) and Tan & Robinson (1990). In Daltoniaceae, there are at least two generic pairs that have similar gametophytic traits, but different peristome structure and ornamentation (Crosbya – hookeriaceous vs. Daltonia – daltoniaceous; Distichophyllum – hookeriaceous vs. Leskeodon – daltoniaceous). The resolution of phylogenetic relationships in the family will facilitate an understanding of the patterns of evolution in both leaf limbidia and peristome structure, especially exostome ornamentation (Fig. 1). Consequently, their value in generic delimitation can be assessed. This study was conducted with the following aims: (1) to assess the evolution of two taxonomically important morphological characters (leaf limbidia and exostome ornamentation); (2) to resolve relationships among the genera of Daltoniaceae; (3) to infer infrageneric relationships in Calyptrochaeta and Achrophyllum; and (4) to assess interspecific relationships in the large and widespread (and apparently polyphyletic) genus Distichophyllum. MATERIAL AND METHODS Plant names used for taxa in this study follow the Tropicos database (http://www.tropicos.org), except when otherwise indicated. Authorities of species and variety names are indicated in Supporting Information Table S1. Each voucher is annotated with the taxon name followed by a two-letter country code in parentheses according to the ‘ISO 3166-1 alpha-2’ (http://www.iso.org/iso/english_country_names_and_ code_elements). In some cases, a single letter suffix was added to indicate collection from different regions within a large country (see Table S1). TAXON SAMPLING AND MOLECULAR PROTOCOLS One hundred and twenty-six accessions were sampled for DNA, including 18 exemplars from the other seven hookerialean families sensu Buck et al. (2005) as outgroups. The ingroup consisted of 94 species from 12 of the 14 genera in Daltoniaceae s.l. As phylogenetic relationships in Daltonia have recently been assessed (Yu et al., 2010), only a selection of representatives of Daltonia to cover the genetic diversity of the genus was included in this study (10 of 21). However, as many species as possible were sampled from other larger genera, such as Achrophyllum (six of eight), Calyptrochaeta (11 of 29) and Leskeodon (seven of 20). Sampling of Distichophyllum (37 species and eight varieties) represented about one-third of the ~100 accepted species (Crosby et al., 1999). In Distichophyllum, the sampling included two or more accessions for species that showed large morphological variability and/or formed species complexes that were difficult to resolve morphologically. Nucleotide sequences (593 of 630 available) were obtained from five regions representing all three genomes, including 320 newly generated nucleotide sequences for this study (Table S1). Sequenced regions comprised: (1) the plastid trnS-rps4 region [i.e. rps4 plus the trnS–rps4 intergenic spacer (IGS), hereafter rps4]; (2) the plastid trnL–F region, including the trnLUAA group I intron and the trnL–F IGS (hereafter trnLF); (3) the mitochondrial nad5 group I intron (hereafter nad5); (4) the nuclear ribosomal ITS1–5.8S–ITS2 (hereafter ITS) region; and (5) partial nuclear large ribosomal RNA subunit (hereafter 26S). Voucher information and GenBank accession numbers are summarized in Table S1. All 73 new vouchers used in this study were identified or re-confirmed by the first author, as many species have not been critically evaluated in a taxonomic revision. Total genomic DNA extractions were performed from dried herbarium vouchers using a modified cetyltrimethylammonium bromide (CTAB) protocol (Doyle & Doyle, 1990), as in Shaw (2000). Amplifications of the selected DNA regions were carried out following standard protocols and primers, as outlined in Olsson et al. (2009b) and Shaw, Cox & Boles (2003). © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• EVOLUTION AND PHYLOGENY OF DALTONIACEAE 5 Table 1. List of hotspots (ambiguous alignments) and inversions with their corresponding positions in the final concatenated data matrix (Supporting Information) Nr. Position Gene Nr. Position Gene Hs1 Iv1 Hs2 Iv2 Hs3 Hs4 Hs5 Hs6 Iv3 Hs7 957–960 961–965 970–973 1131–1137 1212–1215 1281–1289 1339–1347 1553–1558 1604–1610 1650–1655 trnL-F trnL-F trnL-F trnL-F trnL-F trnL-F trnL-F trnL-F trnL-F trnL-F Hs8 Hs9 Hs10 Hs11 Hs12 Hs13 Hs14 Hs15 Hs16 Hs17 Hs18 2526 2530 2993–3042 3109–3111 3131–3135 3432–3437 3601–3603 3763–3766 4192–197 4280–4290 5016–5018 5042–5043 nad5 nad5 ITS ITS ITS ITS ITS ITS ITS 26S 26S Genomic region sequences in the concatenated data matrix were arranged in the order: rps4, 1–837; trnL-F, 838–1695; nad5, 1696–3042; ITS, 3043–4501; 26S, 4502–5525. Hs, hotspot; Iv, inversion. DNA SEQUENCE EDITING AND ALIGNMENT Forward and reverse sequences were assembled and edited for inaccuracy using either PhyDE 0.995 (http://www.phyde.de) or Sequencher v4.1 (Gene Codes Corp.). Consensus sequences were aligned manually in PhyDE 0.995 applying the guidelines outlined in Kelchner (2000), Borsch et al. (2003), Quandt & Stech (2005) and Morrison (2006). Simple sequence repeats were positionally isolated on the basis of strict motif recognition, as advocated by Kelchner (2000) and Quandt & Stech (2005). Regions of ambiguous alignment (hotspots) in the data matrix were defined as outlined in Olsson et al. (2009b) and excluded from phylogenetic analyses (Table 1). Hairpin-associated inversions, which were visually identified, were positioned separately in the alignment (see Table 1). Instead of coding for the presence or absence of inversions for the phylogenetic analyses, they were reversed and complemented in a second alignment file to retrieve the information within the detected inversion (cf. Quandt, Müller & Huttunen, 2003; Borsch & Quandt, 2009). Both alignments can be found as online supporting information (Data files S1 & S2) and are also deposited in TreeBASE (http://www.treebase.org) under http://purl.org/phylo/ treebase/phylows/study/TB2:S11127. DNA SEQUENCE DATA ANALYSES Analyses using maximum parsimony (MP), maximum likelihood (ML) and Bayesian Inference (BI) were performed with or without additional information from the simple indel coding (abbreviated as sic) approach of Simmons & Ochoterena (2000). Preliminary analyses on the concatenated nuclear and organellar datasets were first carried out to check for conflicts [i.e. compare nodes with at least 70% BS or 0.95 posterior probability (PP); see method in MasonGamer & Kellogg, 1996) before final analyses on the total combined data matrix. The concatenated combined data matrix was analysed without indel coding, with indel coding in the organellar dataset only and with indel coding for the complete dataset (written as subscript w/o, sic-org and sic, respectively; see Table S2 for detailed definition). The computer program SeqState (Müller, 2005) was used to generate a ready-to-use NEXUS file containing the sequence alignment with an automatically generated binary indel matrix appended. Command files for the parsimony ratchet (Nixon, 1999) and likelihood ratchet (Morrison, 2007) under the GTR + G + I model were both generated using the program PRAP2 (Müller, 2007), applying the default configuration, and executed in PAUP 4.0b10 (Swofford, 2002). Heuristic bootstrap searches under MP were performed with 10 000 replicates (Hillis & Bull, 1993; cf. Müller, 2005) in PAUP 4.0b10, whereas, under ML, 400 replicates (cf. Pattengale et al., 2009) were performed in GARLI 0.96b8 (Zwickl, 2006) with default settings. Bayesian analyses were performed with MrBayes v3.1.2 (Huelsenbeck & Ronquist, 2001; Ronquist & Huelsenbeck, 2003), applying the GTR + G + I substitution model and the restriction site model (an F81like model; MrBayes manual) for the sequence data and the binary indel partitions, respectively. To allow for heterogeneous DNA substitution patterns, the dataset was divided into four sequence data partitions including partition 1 [plastid (rps4 + trnLF)], 2 [mitochondrial (nad5)], 3 [nuclear (ITS + 26S)] and 4 © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• 6 B. -C. Ho ET AL. [coded indel] scores. Model parameters for each partition were optimized independently. We performed an analysis with partitioning of the dataset, but without including the indels, to evaluate the effect of indel inclusion. Another analysis of the dataset without sequence-based partitioning (and without indel coding) was carried out to evaluate the effects of data partitioning. The a priori probabilities supplied were those specified in the default settings of the program. PP distributions of trees were estimated using the Metropolis-coupled Markov chain Monte Carlo (MCMCMC) method. Eight runs with four chains of 6 ¥ 106 generations each were run simultaneously. Chains were sampled every 1000 generations with trees written to a tree file. The program Tracer v1.5 (Rambaut & Drummond, 2009) was used to evaluate the burn-in and to examine the log likelihoods, ensuring that all parameters converged to a stationary phase with sufficient effective sample size (ESS). Calculations of the consensus tree and PP of the clades were performed based on the trees sampled after the chains converged at generation 1 000 000 for the dataset with partitioning (with or without simple indel coding) and 600 000 without partitioning. Consensus topologies and support values from the different methodological approaches were compiled and drawn using TreeGraph 2.0.45-197 beta (Stöver & Müller, 2010). MORPHOLOGICAL was pruned from these 1000 rooted trees with PAUP 4.0b10 (Swofford, 2002) prior to the analyses. A reversible-jump MCMC approached was utilized, as advocated by Pagel & Meade (2006), because it can simultaneously test the five models of character state transformation available in the program and reconstruct ancestral states. A reversible-jump hyperprior with a uniform distribution on the values of 0–30 was used to seed the mean of the prior exponential distribution of the rates of state transition. Ancestral states were only reconstructed for deeper nodes of the backbone and other selected nodes of interest. Analyses were carried out using the ‘addMRCA’ (most recent common ancestor) command so that the reconstruction will consider the node in any sampled tree that minimally contains all the specified taxa (i.e. the node might include a number of other taxa). Rate deviation (rd) was adjusted (rd = 70 for C1, rd = 60 for C2) to yield an acceptance rate of about 15–40%, as recommended in the program manual. The analyses were performed for 100 ¥ 106 iterations with a sampling frequency of 5000. Tracer v1.5 was used to ascertain that the chains in both analyses had reached convergence after the default burn-in (50 000 iterations). Ancestral states (i.e. state 0 and 1) for each reconstructed node were evaluated by taking the arithmetic means of the sampled PPs for each character state. Mean PPs, shown as proportions on piecharts plotted on a cladogram, were drawn with TreeGraph 2.0.45-197 beta. DATA AND ANCESTRAL STATE RECONSTRUCTION As leaf limbidia (character C1: elimbate, 0; limbate, 1) and exostome ornamentation (character C2: crossstriate, 0; papillose, 1) are key characters used to distinguish genera in Daltoniaceae (Fig. 1), their evolution was reconstructed. Character states were scored from the voucher specimens, supplemented by published descriptions where necessary. Literaturebased information was necessary, especially for exostome structure, as sporophytes at the correct developmental stage are uncommon and molecular vouchers often lacked them. Ancestral states were evaluated analytically using an MCMC approach implemented in the BayesMultiState module in the BayesTraits v1.0 package (Pagel, Meade & Barker, 2004; Pagel & Meade, 2006). The MCMC method has the advantage of accounting for uncertainty in both phylogenetic topology and character mapping (Pagel et al., 2004; Ronquist, 2004). A total of 1000 BI trees (i.e. 125 random trees after burn-in from each of the eight BIhom runs (i.e. BI with a single homogeneous model, see Table S2) were used for character evolution analyses. To enable bifurcating branching at the root node, a prerequisite for BayesTraits to execute, Hypopterygiaceae (the most distant outgroup) RESULTS ALIGNMENT AND SEQUENCE ANALYSES Ninety-four per cent of the sequences were successfully obtained: all rps4 and trnL-F, 87% nad5, 95% ITS and 88% 26S accessions (Table S1), with unaligned amplicon lengths of 623–729, 406–479, 920– 1149, 655–797 and 985–999 nucleotide bases, respectively. Excluding the hotspots, the concatenated and aligned data matrix consists of 5365 characters in total: 1634 plastid, 1291 mitochondrial and 2440 nuclear positions. In total, 18 hotspots were identified, occurring mostly in trnL-F and ITS (see Table 1). A total of 960 indels was coded, two-thirds belonging to ITS sequences. Simple sequence repeats contributed to most of the length variation in trnL-F. Within rps4, the rps4-trnS IGS accounted for most of the sequence length variability. However, a 90-base nucleotide repeat within rps4, characterizing Ephemeropsis trentepohlioides (Renner) Sainsbury, is noteworthy. Length mutations in the nad5 and 26S sequences were limited. The coded indels also increased the number of potentially parsimony informative (PI) characters from 1120 to 1630 © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• EVOLUTION AND PHYLOGENY OF DALTONIACEAE (Table S2). The PI characters of the nuclear ribosomal genome nearly doubled with the inclusion of coded indels as characters (from 483 to 870). Among the five markers, ITS contributed the highest number of PI characters (380) in the combined data matrix, similar in number to the combined plastid genome sequences (i.e. rps4 + trnL-F). PHYLOGENETIC ANALYSES Phylogenetic trees were rooted with Hypopterygiaceae (Cyathophorum P.Beauv., Hypopterygium Brid. and Lopidium Hook.f. & Wilson) based on the results of Buck et al. (2005). Both MP and BI analyses of the combined nuclear (ITS + 26S) and organellar (nad5 + rps4 + trnL-F) datasets revealed no significant conflicts (i.e. nodes with at least 70% BS or 0.95 PP; Supporting Information Figs S1 and S2). The two datasets were hence combined for the final analyses. Clade support was assessed by BS for parsimony (three datasets: BSw/o, BSsic, BSsic-org, see Table S2 for abbreviations of the subscripts) and likelihood (BSML) analyses, and by PP for BIs (four datasets: PPhom, PPw/o, PPsic, PPsic-org, see Table S2 for abbreviations of the subscripts). Support values were considered to be ‘adequate’ when BS ⱖ 70% (Hillis & Bull, 1993) (100%, ‘maximum support’; ⱖ 80%, ‘well supported’; 80–70%, ‘moderately supported’; < 70%, ‘poorly supported’) or when PP ⱖ 0.95 (1.00, ‘maximum support’; ⱖ 0.99, ‘well supported’; 0.95–0.99, ‘moderately supported’; < 0.95, ‘poorly supported’). All MP analyses of the concatenated datasets with various simple indel coding schemes gave almost identical results. However, MP analyses of the datasets with simple indel coding (i.e. MPsic and MPsic-org) generally provided stronger BS than analyses without coded indels (MPw/o) (Fig. S1). Results from BIs of all four datasets (i.e. BIhom, BIw/o, BIsic, BIsic-org) showed no conflict with the MP consensus trees and had higher resolution with generally higher support. The tree topologies from the three datasets with sequence partitioning were almost identical, except for a few distal branches with poor PP values (see Fig. S2). For all analyses of the partitioned datasets, individual runs did not have sufficient ESS (<200) for the ‘Tree Length’ (TL{all}) parameter. Nevertheless, BIhom gave the best overall scores for each parameter among the four BI analyses (Table S2). ML analyses were carried out only on the dataset without indel coding. Likelihood ratchet analysis resulted in 12 best trees, with two main topologies that differed in the placement of the Calyptrochaeta clade. One-third of the ML trees (ln L = -36831.66) resolved Calyptrochaeta as the sister group of the rest of Daltoniaceae, and this relationship was 7 also supported maximally in BIw/o. The remaining nine trees (ln L = -36831.15) had Calyptrochaeta forming a sister-group relationship with the Schimperobryaceae–Hookeriaceae–Leucomiaceae– Pilotrichaceae clade. One of the ML trees with the latter topology is shown in Figure 2A and B, with support values BSsic, BSML, PPsic and PPw/o. Alternative topologies relating to the placement of Calyptrochaeta are shown by dotted lines. The same tree is also presented as a phylogram to illustrate the branch lengths within and among clades (Fig. S3). The precise phylogenetic position of Calyptrochaeta remains uncertain (Fig. 2A). Similarly, the placement of Achrophyllum, emerging as the sister group of the remaining Daltoniaceae (i.e. core Daltoniaceae), does not have adequate support in all analyses, except BIsic and BIsic-org. Monophyly of both Calyptrochaeta and Achrophyllum is maximally supported. In Calyptrochaeta, the position of C. asplenioides (Brid.) Crosby is ambiguous and different sister-group relationships are resolved under ML and BI (Fig. 2A). With MP (Fig. S1A), C. asplenioides, C. cristata (Hedw.) Desv. and the other nine Calyptrochaeta spp. form a trichotomy. The species-rich clade sits on a long branch that bifurcates into an Australasian–Patagonian subclade and an Asian subclade (Fig. S3). The relationships of the two Australasian species C. brownii (Dixon) J.K.Barlett and C. otwayensis Streimann are unresolved in all analyses. All six Asian species sampled form a monophyletic group, although support values are inadequate (BSML = 67, PPhom = 0.85). In Achrophyllum, A. haesselianum (Matteri) Matteri is the sister group of all other species in the genus. The type of the genus, A. quadrifarium (Sm.) Vitt & Crosby, forms a sister-group relationship with the remaining four species, which are largely unresolved (Fig. 2A). Within the maximally supported core Daltoniaceae, Beeveria distichophylloides (Broth & Dixon) Fife and Distichophyllum microcarpon (Hedw.) Mitt. form the sister group of the remaining family. These two species, both on long branches (Fig. S3), are moderately supported as sister groups by the dataset with indels coded (Fig. 2A). Support values are lower without indels coded. Remaining exemplars of core Daltoniaceae comprise two sister clades of unequal size, the smaller including Ephemeropsis, Bryobrothera, Adelothecium and Benitotania. Adelothecium and Benitotania form the sister group of Bryobrothera and these three monotypic genera constitute, in turn, the sister group of Ephemeropsis. The well-supported and larger clade consists of various assemblages of different species in Daltonia, Distichophyllum, Leskeodon and a few other small genera. Within this clade, a group of Neotropical © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• 8 B. -C. Ho ET AL. Figure 2. See caption on next page. © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• EVOLUTION AND PHYLOGENY OF DALTONIACEAE 9 Figure 2. Cladogram of one of 12 maximum likelihood trees found using a ratchet approach (see text). Support values above the branches are posterior probabilities based on a nonpartitioned model (PPhom), followed by likelihood bootstrap support values (BSML); both excluded indels in the analysis. Values below the branches are posterior probabilities based on a partitioned model (PPsic), followed by parsimony boostrap support values (BSsic), both with indels coded in the matrix. Alternative topologies of parts of branches are shown as dotted lines. ‘#’ denotes maximum support (BS = 100%, PP = 1.00); ‘-’ denotes conflicting or unresolved topology of the corresponding analysis; Dalt 1–2, Dist 1–8 and Lesk 1–3 are abbreviations for Daltonia clade 1–2, Distichophyllum clade 1–8 and Leskeodon clade 1–3, respectively. Leskeodon (Lesk1), together with six Distichophyllum spp. from South America, Australasia and Asia (Dist1), is sister group to the rest of the sampled taxa. The Lesk1–Dist1 clade is on a long branch (Fig. S3) and receives maximum support in nearly all analyses (Fig. 2A). The placement of Distichophyllum ellipticum Herzog within this clade is uncertain, with some analyses favouring a topology of D. ellipticum forming © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• 10 B. -C. Ho ET AL. a sister-group relationship with D. fernandezianum Broth., which is in turn sister group to the rest of Dist1 (Fig. 2A, shown as dotted lines). Next, a well-supported group of Distichophyllum spp. confined to southern Australasia and Patagonia (Dist2) is sister to the rest of the taxa (bottom of Fig. 2A plus 2B). Crosbya straminea (Mitt. ex Beckett) Vitt and Distichophyllum mniifolium (Hornsch.) Sim (clade Dist3), New Zealand and South African endemics, respectively, are the next two successive splits, forming a grade. Following these, there are three successive clades that form three short branches lacking support (Figs 2B and S3). These include a small clade of Hawaiian endemics, Distichophyllum freycineti (Schwägr.) Mitt. and D. paradoxum (Mont.) Mitt. (Dist4), a clade consisting largely of Daltonia spp. plus a few morphologically atypical (see discussion for details) Distichophyllum spp. (Dist5 + Dalt1), and a large clade of exclusively Old-World-Pacific taxa. Most internal nodes within the well-supported Dist5–Dalt1 clade receive maximum PP and at least 85% support from BSML and BSw/o. The Old-World-Pacific clade is well supported in Bayesian analysis, but lacks BS support in the MP and ML analyses. Within this clade, Distichophyllidium nymanianum M.Fleisch., Leskeodon seramensis H.Akiyama (Lesk2) and the rest of the taxa (Dist7 + Dist8) form another three short branches with almost zero branch lengths (Fig. S3). Nevertheless, topologies in almost all analyses (except MPsic) indicate that at least Lesk2 is sister group to the Dist7–Dist8 clade (Fig. 2B). The latter clade is poorly supported (except BSML = 78), but its two main subclades (Dist7 and Dist8) are maximally supported. The Dist7 clade (within which Lesk3 is nested) consists of epiphytes and two species complexes surrounding Distichophyllum nigricaule Mitt. ex Bosch. & Sande Lac. and D. collenchymatosum Cardot. Both species complexes receive good support, but it is unclear which is the sister group of D. cuspidatum (Dozy & Molk.) Dozy & Molk. (Fig. 2B). This group is the sister group of Dist8, within which Dalt2 is nested. Most of the internodes within the Dist8 clade are well supported, although the position of Daltonia armata E.B.Bartram is ambiguous. Species with multiple accessions from different islands, such as Distichophyllum leiopogon Dixon (better known as D. cucullatum E.B.Bartram, see Ho, Tan & Nathi, 2010), D. spathulatum Broth. and D. tortile Dozy & Molk. ex Bosch & Sande Lac., are resolved as monophyletic with maximum support. The sister group of D. tortile is a wellsupported clade consisting of D. schmidtii Broth. plus all species sampled from southeast Africa and adjacent islands in the western Indian Ocean. One of the 12 ML trees has D. schmidtii resolved as sister group of D. mascarenicum Besch. and D. rakotomariae Crosby (not shown), which also has some poor support from all BI analyses (Fig. S2B). ANCESTRAL STATE RECONSTRUCTION Reconstructed evolutionary transitions in leaf limbidia (C1) and exostome ornamentation (C2) are plotted on the tree shown in Figure 3. The reversible jump MCMC chains visited the one-rate ‘0Z’ model (i.e. q01 > 0, q10 = 0, where q01 is the transition rate from state 0 to 1 and vice versa) ~96% and 68% of all post-burnin iterations for characters C1 and C2, and, for the ‘00’ model (i.e. q01 = q10), ~4% and 31%, respectively (visited other models occasionally). No state reversals were detected for either character. Reconstruction of leaf limbidia indicates that the ancestral state at the root node is elimbate (mean PP of 0.9999) and that leaf limbidia evolved twice (Fig. 3, circles above the branches). A transition from elimbate to limbate leaves occurs at the root of the Calyptrochaeta clade and the clade corresponding to the Daltonia– Distichophyllum–Leskeodon complex (Lesk1, Dist1 in Fig. 2A). For exostome ornamentation, the ancestral state at the root is reconstructed as striate (PP = 0.9961). At least six independent transitions to papillose from striate exostome teeth were detected (Fig. 3, circles below branches). These transitions correspond to the Neotropical Leskeodon (Lesk1), the main Daltonia clade (i.e. Dalt1 + Dist5) and four scattered individual species (Distichophyllidium nymanianum, Leskeodon seramensis, L. acuminatus and Daltonia armata). The last two taxa clearly represent cases of papillose exostomes arising within clades with striate exostome. Although character reconstructions of the ancestral nodes immediately before Distichophyllidium nymanianum and L. seramensis have a higher mean PP of being cross-striate (0.7887 and 0.9138, respectively) than papillose, the precise position of these two taxa is uncertain (Fig. 2B). This may play a major role in obscuring a definitive tally for the number of times papillose exostome teeth arose. DISCUSSION ASSESSING THE EVOLUTION OF TWO TAXONOMICALLY IMPORTANT MORPHOLOGICAL CHARACTERS Ancestral state reconstructions show that the elimbate leaves and cross-striate exostome are plesiomorphic states in Daltoniaceae s.l. Limbate leaves evolved twice and represent a synapomorphy for Calyptrochaeta and for Distichophyllum plus its allied genera. The majority of species in core Daltoniaceae have limbate leaves (without known © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• EVOLUTION AND PHYLOGENY OF DALTONIACEAE © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• 11 Figure 3. Majority consensus tree of 1000 resampled trees used for character evolution analysis in BayesTraits (see text). Circles plotted beside the nodes represent the proportions of the mean posterior probability (PP) of each state in Character C1 (above branches) and C2 (below branches). Numbers following the sample names correspond to character codings: C1 (leaf limbidia): elimbate (white in circle plot, coded as 0), limbate (grey, coded as 1); C2 (exostome ornamentation): cross-striate (white in circle plot, coded as 0), papillose (grey, coded as 1). Unknown or inapplicable states are denoted by ‘–’. 12 B. -C. Ho ET AL. functionality) and are all resolved in one clade, corresponding to the traditional Daltonia, Distichophyllidium, Distichophyllum, Crosbya and Leskeodon. All species in these five genera (except Distichophyllum microcarpon, discussed below), including those not sampled here, possess at least some traces of a limbidium. Thus, states of this character correspond well with phylogenetic relationships. The only exception, Calyptrochaeta, has limbate leaves, but with short and unequally forked costae near the base. Costae in other genera in the family are single, except for Distichophyllidium, which is ecostate, and Achrophyllum, which has elimbate leaves with a single costa that is forked above the base. Papillose exostome teeth, characteristic of daltoniaceous peristomes, evolved independently from striate teeth multiple times in Daltoniaceae. Species with papillose exostomes were traditionally placed in the largely epiphytic Daltonia or Leskeodon. Although papillose exostomes are consistent in clades Lesk1 and Dalt1, some species in the traditional Daltonia and Leskeodon are intermixed with species of Distichophyllum, a genus traditionally circumscribed as having striate exostomes. Daltonia armata was described in Daltonia based mainly on leaf morphology without description of the peristome (Bartram, 1944). Its papillose exostome is confirmed here from the type specimens (A. Lynn Zwickey 638 FH!, studied by the first author). Leskeodon acuminatus was first described as a Distichophyllum, but was transferred to Leskeodon because of its peristome type (Fleischer, 1908). Daltonia armata and Leskeodon acuminatus are nested deeply within a clade of taxa with striate peristomes. Traditional segregation of Leskeodon from Distichophyllum merely by its daltoniaceous peristome (papillose exostome) is rejected here. For the other generic pair, the separation of the hookeriaceous Crosbya, with only two accepted close species, from Daltonia is supported. Thus, contrary to Buck’s (1991, 2007) suggestion, exostome ornamentation is not always a reliable character for distinguishing genera in Daltoniaceae. Whittemore & Allen (1989) observed that the two types of perisome exhibit different and opposite hydroscopic movements in response to changing humidity. Shifts to a daltoniaceous peristome could be associated with the change to an epiphytic lifestyle during evolution. Our observations may have implications for generic delimitation in the large family Pilotrichaceae, in which genera are similarly distinguished by peristome types regardless of gametophytic similarities (e.g. Lepidopilum vs. Lepidopilidium, Stenodictyon vs. Stenodesmus) (Buck, 1998). Although molecular data are available for some of these taxa, a study focus in Pilotrichaceae, with better sampling, is necessary to ascertain our hypothesis (see also Pokorny et al., 2012). Basic peristome types that differ in development correspond to deep lineages in the mosses (Shaw, Anderson & Mishler, 1987, 1989a, b; Shaw & Anderson, 1988; Goffinet et al., 1999; Shaw, Szövényi & Shaw, 2011). The differences, however, relate to patterns of cell division in the apical region of young sporophytes, long before the teeth develop, whereas the sorts of differences that have historically been used to distinguish genera and families in Hookeriales and other groups of pleurocarpous mosses pertain to superficial ornamentation of the teeth. Differences in ornamentation develop just before maturation and the release of spores. Our results add to a growing body of evidence that peristomal differences that occur late in development are phylogenetically mutable (e.g. Hedenäs, 2001, 2002; Vanderpoorten et al., 2002; Huttunen et al., 2004; Olsson et al., 2009a; Quandt et al., 2009; Liu et al., 2012; Pokorny et al., 2012). Nevertheless, the systematic significance of peristome variation needs to be assessed on a group-by-group basis, ideally using independent evidence to resolve phylogenetic relationships. RESOLVING RELATIONSHIPS AMONG THE GENERA IN DALTONIACEAE The topology of the Hookeriales backbone closely resembles results from Buck et al. (2005). Buck et al. (2005) were unable to unambiguously resolve the relationships of Calyptrochaeta and Achrophyllum relative to Hookeriaceae and Daltoniaceae. Despite the increased taxon sampling in our study, resolution of their relationships remains inconclusive. Nevertheless, both genera are resolved as monophyletic with maximum support. Other approaches to resolving their relationships to other genera could be to increase genomic sampling with a reduced taxon dataset or to try other analysis methods, such as network-based techniques. The topology of the elimbate taxa of core Daltoniaceae differs slightly from that reconstructed by Buck et al. (2005) and Gradstein & Wilson (2008). In their studies, Beeveria and Ephemeropsis formed a clade (PP < 0.95, BS = 52) that is the sister group to the rest of core Daltoniaceae. In our study, Beeveria plus Distichophyllum microcarpon (not sampled in their studies) form a sister-group relationship to the rest of core Daltoniaceae, and Ephemeropsis is the sister group of the remaining elimbate taxa (Fig. 2A). Leaves of D. microcarpon, unlike any other species of Distichophyllum, have no trace of differentiated leaf border or limbidium, and hence the taxon is unlikely to belong to Distichophyllum or any of its allied genera. Distichophyllum microcarpon could be © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• EVOLUTION AND PHYLOGENY OF DALTONIACEAE transferred to Beeveria to reflect its closer affinity to that genus than to other limbate groups, although the relationship is not adequately supported in all analyses here. Allan Fife (unpubl. data) intends to create a new genus to accommodate this taxon based on morphology; our findings would lend additional support to this taxonomic change. The monophyly of, and relationships among, Adelothecium, Benitotania and Bryobrothera have been assessed without disagreement (Akiyama et al., 2003; Buck et al., 2005; Gradstein & Wilson, 2008) and are re-confirmed here. However, the sistergroup relationship of this clade with the genus Ephemeropsis, found here, does not agree with the findings of Buck et al. (2005) or Gradstein & Wilson (2008). The relationships of Ephemeropsis trentepohlioides and Beeveria distichophylloides in their studies could be attributed to long-branch attraction. Our results indicate an adequately supported clade consisting of Adelothecium, Benitotania, Bryobrothera and Ephemeropsis, which are all epiphytic/ epiphyllous taxa. The sister clade of Adelothecium, Benitotania, Bryobrothera and Ephemeropsis, which consists of taxa with limbate leaves, is well supported as monophyletic in our study. Nevertheless, phylogenetic relationships within this clade contradict classical concepts of Daltonia, Distichophyllum and Leskeodon. Specifically, the small genera Crosbya and Distichophyllidium are positioned within these genera. Adequately supported (i.e. BS ⱖ 70%, PP ⱖ 0.95) clades show various combinations of species currently classified in different genera. The phylogenetic evaluation of the limbate taxa is discussed below with regard to the reorganization of Distichophyllum. INFRAGENERIC RELATIONSHIPS WITHIN CALYPTROCHAETA AND ACHROPHYLLUM In terms of morphology, the type species C. cristata deviates from the most common appearance of the species in the genus, with laminal cells evenly thinwalled and leaf marginal teeth consisting of a variable number of cells. Typical Calyptrochaeta spp. often have thick-walled laminal cells, at least at the cell corners, and the majority of the leaf marginal teeth consist of parts of two adjacent border cells. These morphological differences seem to support the topology in BI (Fig. S2A), in which C. cristata forms a sister-group relationship to the rest of the sampled species. The Australasian species, C. brownii and C. otwayensis, have almost identical sequences, which may explain the unresolved topology. In our opinion, the morphological variation between the two species is sufficiently low to be considered conspecific. 13 This is the first attempt to evaluate species relationships in the genus Achrophyllum using molecular data. Within Achrophyllum, A. haesselianum and A. quadrifarium are morphologically distinct, as the plants have pale green coloration and scarcely toothed to subentire leaf margins (Sainsbury, 1955; Matteri, 1972). The four species reconstructed as the sister group of A. haesselianum are dark green plants with erose-dentate leaf margins. Species in the latter group are difficult to distinguish morphologically and largely unresolved in the phylogenetic tree. Matteri (1972) and Robinson (1975) had different concepts for taxa in this species complex, evident from contrasting morphological features used in their identification keys. Moreover, features used by both authors for species identification, such as size of marginal teeth, length of costa, laminal cell size, degree of wall thickening at cell corners etc., are quite variable. Notably, Robinson (1975) proposed synonymy of A. crassirete (Matteri) Matteri and A. magellanicum (Besch.) Matteri under A. anomalum (Schwägr.) H.Rob. and A. dentatum Hook.f. & Wilson) Vitt & Crosby, respectively. REORGANIZATION OF DISTICHOPHYLLUM AND ITS ALLIES This study has confirmed the heterogeneity of Distichophyllum and shows the complexity of its relationships to Crosbya, Daltonia, Distichophyllidium and Leskeodon. Considering the difficulty of identifying morphological synapomorphies for internal clades, one option is to consider the entire clade as a single genus. This would require the generic names Distichophyllum and Leskeodon, together with a few others, to be synonymized with the oldest name, Daltonia. This approach would require numerous new combinations and would tend to disrupt nomenclatural stability. We propose that it is a more reasonable solution to reorganize and adjust the traditional concepts of the genera, which will require additional morphological study. Resolved clades within Distichophyllum and its allies correspond more to biogeographical entities than to traditional concepts of genera. For instance, Lesk1 consists of species limited to the Neotropics, species in the Dist2 clade appear to have a so-called ‘Nothofagustype’ distribution (Seki, 1973), the two Hawaiian endemics are closely related (Dist4) and all species in Dist7 and Dist8 are restricted to the Old World. The only exception is found in the Dist1 clade, where Asian species are nested within groups of southern South American and Australasian species. Our phylogenetic analyses also show that Daltonia species occurring almost exclusively in the Himalayan region belong to species-poor clades that constitute a paraphyletic © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• 14 B. -C. Ho ET AL. assemblage, whereas species with a transcontinental distribution form a monophyletic group, thus agreeing with Yu et al. (2010). Although the genus Leskeodon is phylogenetically heterogeneous, excluding the Old World species would make the remaining members monophyletic and exclusively Neotropical (Lesk1 in Fig. 2A). However, some species currently in Distichophyllum (Dist1) are the sister group of Lesk1. These two clades are separated by short branches, but together are subtended by a longer branch (Fig. S3). Morphologically, all species in this Lesk1–Dist1 clade have small isodiametric laminal cells that are more or less homogeneous in size, except along the costa near the base. Moreover, plants of D. maibarae Besch. and D. montagneanum from Asia have remarkably similar leaf morphology to those of L. andicola from the New World. Although W. R. Buck (pers. comm., February 2010) believes that members of the two clades are sufficiently distinct to be different genera, we consider it better to treat both Dist1 and Lesk1 as a single genus. As an exemplar of L. auratus (type species of Leskeodon; Welch, 1966) belongs in the Lesk1–Dist1 clade, we thus refer to the Lesk1–Dist1 clade as Leskeodon. However, the formal transfers of the names D. crispulum, D. ellipticum, D. fernandezianum, D. montagneanum and D. rotundifolium to Leskeodon are postponed because the types of these names have not been examined. Within this clade, D. maibarae and D. montagneanum are distinguished exclusively by the presence or absence of some long erect hairs on the calyptrae (Mohamed & Robinson, 1991). The taxonomic value of this character has been questioned (Ho et al., 2010). The sampled Chinese accession of this species complex has naked calyptrae and should be named D. montagneanum, a new country record. However, the tree topology suggests that this plant is closer to the Japanese plants, where only D. maibarae, with hairy calyptrae, is known. The two species are gametophytically inseparable and polymorphic in terms of size and colour. No molecular or morphological evidence currently supports the separation of the two species, but a more detailed study with populations sampled throughout the geographical range of occurrence may resolve this taxonomic issue. The position of the New Zealand endemic genus Crosbya is almost identical to that in earlier published results (Buck et al., 2005). Gametophytes of Crosbya closely resemble those of Daltonia, except for the excurrent costa. Unlike Daltonia spp., the two species of Crosbya are both dioecious and have a hookeriaceous peristome (Vitt, 1977). Gametophytic similarity of these two genera may reflect convergence, as both are mostly epiphytes (Vitt, 1977). However, Crosbya spp. appear to be limited to tree trunks and branches, sometimes on boulders, whereas Daltonia spp. most commonly grow on twigs and leaves. All Daltonia spp., except D. armata, fall within the well-supported Dist5–Dalt1 clade. A few peculiar Asian (Himalayan) Distichophyllum spp. with more or less carinate leaves also belong to this clade. Daltonia cf. apiculata, Distichophyllum heterophyllum, D. meizhiae B.C.Tan & P.J.Lin and D. wanianum B.C.Tan & P.J.Lin produce gemmae on the dorsal side of the leaf costa (Ho et al., 2010). These Distichophyllum spp. also have ± rectangular basal laminal cells, which is a typical trait of some Daltonia spp. (Yu et al., 2010). Hence, the nomenclatural transfer of these species of Distichophyllum into Daltonia can be justified and proposed in this article. The limited sampling of the International Union for the Conservation of Nature (IUCN) red-listed Distichophyllum carinatum reveals that the Asian exemplars are the sister group of the European (German) exemplar, but this should not be interpreted as meaning that the species originated in Europe (Fig. 2B). The two available sequences from the Japanese voucher are identical to those of the Chinese exemplar. Comparing all gene sequences from German and Chinese vouchers (Table S1), only five nucleotide differences are detected in ~4500 nucleotides. It is most likely that the species originated in continental Asia (see discussion in Ho et al., 2010) where the majority of the closely related Daltonia spp. occur. The small genus Distichophyllidium is represented in the present study by only the type species. In the absence of the other four species in the genus, its monophyly and generic relationships cannot be determined. Buck et al. (2005), with the sampling of only two species each of Distichophyllum and Daltonia, showed a well-supported sister-group relationship between Distichophyllidium and Daltonia, a topology not resolved in our study with better sampling. The Dist7–Dist8 clade is considered to be the core of Distichophyllum, because an exemplar of the type species, D. spathulatum (Buck et al., 2005), is included here. The core Distichophyllum is a poorly supported clade consisting of two well-supported sister clades. We provisionally recognize the whole clade as Distichophyllum in the strict sense. At best, the two subclades Dist7 and Dist8 should be recognized at the infrageneric level in Distichophyllum, as there appears to be no morphological characters distinguishing them. Within Dist7, the nesting of Distichophyllum succulentum within D. collenchymatosum corroborates the suggestion in Ho et al. (2010) that the two names might be conspecific. However, without examination of type specimens, particularly those of the little known D. succulentum, it is better to postpone the synonymization of these names. © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• EVOLUTION AND PHYLOGENY OF DALTONIACEAE Two clades of epiphytic species occur within Dist7. The first consists only of Distichophyllum cuspidatum. In the other epiphytic clade, D. jungermannioides (Müll.Hal.) Bosch & Sande Lac., a species commonly found at the base of trees and shrubs, is the sister group of two other true epiphytes including Leskeodon acuminatus. Consequently, revival of the original basionym Distichophyllum acuminatum Bosch & Sande Lac. is proposed here for the latter species. All exemplars of the D. nigricaule complex, including plants that vary in size and degree of laminal cell size differentiation, are grouped together, but without resolution. The morphological variability within this species appears to be unrelated to genealogy. Thus, synonymy of the two varietal names, as proposed previously by some authors (e.g. Bartram, 1939; Gangulee, 1977), is supported here. Yu et al. (2010) speculated that Daltonia armata (not sampled in that study) may belong to the paraphyletic assemblage of Daltonia with limited geographical distribution, and not within the monophyletic transcontinentally distributed clade of Daltonia. Current analyses support the nesting of this species (Dalt2) within the Dist8 clade of the core Distichophyllum. The removal of this species from Daltonia has been suggested (Ho et al., 2010), but its placement was uncertain. The current phylogenetic assessment justifies its transfer into Distichophyllum s.s. The long branch length leading to this species (Fig. S3) indicates that rapid evolution could explain its highly modified morphology, but this suggestion remains to be tested. Taxonomic assessment of other species is not attempted in this study because additional morphological evaluation is required. These species belong to clades that either show unresolved relationships or have no known morphological synapomorphies. Hence, the names are tentatively retained in their currently accepted genera. These include the clades Dist2, Dist3 (with only Distichophyllum mniifolium), Dist4, (Hawaiian endemics) and Lesk2 (with only Leskeodon seramensis). Distichophyllum subnigricaule is heterogeneous, with its two established varieties appearing in different clades. Morphological similarities of the two varieties may be convergent. Thus, it is best to raise D. subnigricaule var. hainanense to species level and to treat the two taxa as separate species. For other species that are demonstrably nonmonophyletic, including Daltonia apiculata, Distichophyllum osterwaldii and D. pulchellum, the status of possible new taxa cannot be decisively assessed without further taxonomic study. Several relationships within the family could not be evaluated because sequences of some critical taxa were not obtained. Future phylogenetic studies need to include: (1) Metadistichophyllum rhizophorum 15 (M.Fleisch.) Nog. & Z.Iwats., which is a monotypic South-East Asian genus sometimes considered to be synonymous with Distichophyllum (Crosby, 1974; Akiyama, 1990); (2) Leskeodontopsis pustulata Zanten, a monotypic genus of rare occurrence in New Guinea; (3) Distichophyllum flavescens (Mitt.) Paris, the type species of the genus Discophyllum Mitt. from the Pacific Islands; (4) Distichophyllum noguchianum B.C.Tan, the type species of section Platyovatophyllum B.C.Tan known only from the Philippines; (5) Leskeodon palmarum (Mitt.) Broth., morphologically distinctive (Buck, 1998) and the only species in Leskeodon section Longiseti Broth. from the Neotropics. In addition, species, such as Achrophyllum javense (J. Froehl.) Z.Iwats. from South-East Asia, Calyptrochaeta setigera (Mitt.) W.R.Buck endemic to Brazil, Distichophyllidium jungermanniaceum M.Fleisch., also from South-East Asia, and Distichophyllum santosii E.B.Bartram from Borneo and the Philippines, could be important for the clarification of infrageneric relationships. CONCLUSION Being lost in a sea of similar gametophytic characters among genera and under the influence of Philibert’s principles of peristome conservatism, it is no surprise that workers in the past used easily observed differences in exostome ornamentation as key characters to delimit genera and families. The heterogeneity of papillose exostomes in the limbate Daltoniaceae means that traditional concepts of several genera require taxonomic recircumscription to reflect new insights about phylogenetic relationships. However, finding a set of ‘good’ morphological features to delimit the newly recognized clades is still challenging. Our study has revealed new information about relationships among genera within the Daltoniaceae. However, precise relationships of certain species, genera and clades still remain obscure. Traditional genera in the limbate Daltoniaceae are, in many cases, not supported by our molecular data and suggest convergent evolution. The abundance of homoplasy in morphological traits has hampered an accurate circumscription of genera to reflect natural groupings. Critical generic revisions of Daltonia, Distichophyllum and Leskeodon are essential for the construction of a new taxonomic system and the identification of morphological synapomorphies for resolved clades, if such synapomorphies exist. PROPOSED NEW NOMENCLATURAL COMBINATIONS AND NEW SYNONYMIES Daltonia carinata (Dixon & W.E.Nicholson) B.C.Ho & L.Pokorny, comb. nov. – Basionym: Distichophyl- © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• 16 B. -C. Ho ET AL. lum carinatum Dixon & W.E.Nicholson in Dixon Rev. Bryol. 36: 24. f. 1–7. 1909. – Type: Austria, Salzburg, [Salzkammergut], St. Wolfgang See, Zinkenbach, alt. 700 m; creeping on other mosses upon dripping rocks in ravine. H.N. Dixon & W.E. Nicholson s.n., 3.viii.1908 (holotype: BM!; isotypes DUKE!, E!, H!, NY!, MO n.v., S n.v.). Daltonia heterophylla (Wilson ex Mitt.) B.C.Ho & L.Pokorny, comb. nov. – Basionym: Mniadelphus heterophyllus Wilson ex Mitt. J. Proc. Linn. Soc., Bot. Suppl. 2: 144. 1859. – Distichophyllum heterophyllum (Wilson ex Mitt.) Paris, Index Bryol. 389. 1896. – Type: [Nepal,] Nangli [‘Sikkim’], 10 000 ft; J.D. Hooker ‘690’ (holotype: NY-Mitten!; isotype: L!). The protologue of the name in Mitten (1859) states: ‘In Himalayae orient. Reg. temp., Sikkim, J.D. Hooker no. 690’. The holotype in NY-Mitten is labeled: ‘690 (in pencil), Nangkli, 10 000’, and it has been confirmed by D. G. Long (Royal Botanic Garden Edinburgh, pers. comm. January 2010) that the type locality of this name ‘Nangkli’ should be in Nepal, but has been erroneously regarded as Sikkim by Mitten. A supposed isotype in BM bears the label ‘690 Herb Ind. Or. Hook. fil. & Thomson (no. 458 in pencil). Hab. Tonglo, Reg. temp. Sikkim Himalaya, alt 10 000 ft. Coll. J.D.H.’. However, on close examination of this BM specimen, it turns out to be a different species, identifiable to Distichophyllum succulentum (Mitt.) Broth. It is known that the ‘Herb. Ind. Or. Hook. Fil. & Thomson’ numbers were issued by the herbarium K for distribution of presumed identical species, not duplicates of the same collection (fide D.G. Long); thus, the BM specimen is considered here as not a type. Although only the above two specimens of this seemingly rare species have been reported in the literature (Gangulee, 1977), one other historical specimen of this species has been located in L (Herb. v.d. Sande Lacoste) labelled: ‘Mniadelphus heterophyllus M., Himalaya orient. Hb Mitten’. As with the holotype in NY, this specimen also has a few stems of Cyathophorum hookerianum (Griff.) Mitt. mixed in, suggesting that they originated from the same collection. Mitten exchanged specimens with Dutch bryologists (see Touw, 2007), and thus the specimen at L is interpreted here as an isotype. Daltonia meizhiae (B.C.Tan & P.J.Lin) B.C.Ho & L.Pokorny, comb. nov. – Basionym: Distichophyllum meizhiae B.C.Tan & P.J.Lin Trop. Bryol. 10: 55. f. 2, 8–12. 1995, ‘meizhii’. – Type: China. Yunnan Province, Gongshan-xian (county), Du-long-jiang Commune, on boulder by the Ching-lang-tang river bank, about 1300 m elev. Mei-zhi Wang 10040, viii.1982 (holotype: PE!). Daltonia waniana (B.C.Tan & P.J.Lin) B.C.Ho & L.Pokorny, comb. nov. – Basionym: Distichophyllum wanianum B.C.Tan & P.J.Lin, Trop. Bryol. 10: 57. f. 1, 13–18. 1995. – Type: China. Yunnan Province, Luchun, on branches. M. Zhang 550 (holotype: IBSC n.v.; isotypes: KUN n.v., FH!). Distichophyllum armatum (E.B.Bartram) B.C.Ho & L.Pokorny, comb. nov. – Basionym: Daltonia armata E.B.Bartram, Farlowia 1: 508, f. 21–24. 1944. – Type: Philippines, Mindanao, Lanao Prov., vicinity of Dansalan [ = Marawi], Sacred Mountain, alt. 700–800 m, on culm of climbing bamboo, A. Lynn Zwickey 638, 3.xi.1938. (holotype: FH!; isotype: FH! MICH n.v.). Distichophyllum hainanense (P.J.Lin & B.C.Tan) B.C.Ho & L.Pokorny, stat. nov. – Basionym: Distichophyllum subnigricaule var. hainanense P.J.Lin & B.C.Tan Harvard Pap. Bot. 7: 43. f. 33: E–I. 1995. – Type: China. ‘Hainan, Mt. Diao-luo, on root of tree, c. 1050 m.’ P.-J. Lin et al. 945A, iii.1990 (holotype: IBSC n.v.; isotype: FH!). Leskeodon maibarae (Besch.) B.C.Ho & L.Pokorny, comb. nov. – Basionym: Distichophyllum maibarae Besch., J. Bot. (Morot) 13: 40. 1899 – Type: Japon, Nippon central [Honshu], Maibara, associé au Symphyogyna sublobata, Faurie 11130, 7.xi.1893 (holotype: BM!; isotypes: FH, H-Br!). ACKNOWLEDGEMENTS This research was supported jointly by the Deutscher Akademischer Austausch Dienst (DAAD) budget 331-4-04-002 to BCH, Fulbright Commission grant no. 1582431 and the Ramón Areces Foundation to LP, National Science Foundation (NSF) grant no. DEB-0529593-002 to AJS, and Deutsche Forschungsgemeinschaft (DFG) grant no. QU 153/3-1 to DQ. The authors thank the curators of the following herbaria and A. Schäfer-Verwimp for specimen loans: BM, BORH, CHR, CONC, DR, E, EGR, H, KLU, L, MO, NICH, NY, PSU, S, SING, STU, SZG. The Tan Chin Kee Foundation in the Philippines provided travel support to BCH to present this study at the 8th Flora Malesiana Symposium, and Mr K. T. Yong from Universiti Malaya helped to arrange fieldtrips for BCH. We are grateful to two anonymous reviewers for comments on the manuscript. Purified PCR products were sequenced by Macrogen Inc., South Korea (http://www.macrogen.com) or at the Duke IGSP Genome Sequencing and Analysis Core Facility (http://www.genome.duke.edu/cores/sequencing). Part of the Bayesian analyses was carried out using resources of the Computational Biology Service Unit from Cornell University (http://cbsuapps.tc.cornell. edu/acknowledgement.htm). REFERENCES Akiyama H. 1990. Taxonomic studies of mosses of Seram and Ambon (Moluccas, East Indonesia) collected by Indonesian– © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• EVOLUTION AND PHYLOGENY OF DALTONIACEAE Japanese botanical expeditions II. Journal of the Faculty of Science, University of Tokyo, Section III. Botany 14: 385– 413. Akiyama H, Tsubota H, Yamaguchi T, Suleiman M. 2003. The new genus Benitotania (Daltoniaceae, Bryopsida) from Mt. Kinabalu. The Bryologist 106: 454–459. Bartram EB. 1939. Mosses of the Philippines. Philippine Journal of Science 68: 1–437. Bartram EB. 1944. New and noteworthy Philippine mosses. Farlowia 1: 503–513. Borsch T, Hilu KW, Quandt D, Wilde V, Neinhuis C, Barthlott W. 2003. Non-coding plastid trnT-trnF sequences reveal a well resolved phylogeny of basal angiosperms. Journal of Evolutionary Biology 16: 558–576. Borsch T, Quandt D. 2009. Mutational dynamics and phylogenetic utility of non-coding plastid DNA. Plant Systematics and Evolution 282: 169–199. Brotherus VF. 1907. Pilotrichaceae, Nemataceae, Hookeriaceae and Hypopterygiaceae. In: Engler A, Prantl K, eds. Die natürlichen Pflanzenfamilien [nebst ihren Gattungen und wichtigeren Arten, insbesondere den Nutzpflanzen, bearbeitet unter Mitwirkung zahlreicher hervorragender Fachgelehrten] Ed. 1, T. 1, Abt. 3, H. 2. Leipzig: Wilhelm Engelmann, 916–972. Brotherus VF. 1925. Hookeriales. In: Engler A, Prantl K, eds. Die natürlichen Pflanzenfamilien [nebst ihren Gattungen und wichtigeren Arten, insbesondere den Nutzpflanzen, unter Mitwirkung zahlreicher hervorragender Fachgelehrten] Ed. 2, Bd. 11. Leipzig: Wilhelm Engelmann, 214–278. Buck WR. 1987. Taxonomical and nomenclatural arrangement in the Hookeriales with notes on West Indian taxa. Brittonia 39: 210–224. Buck WR. 1988. Another view of familial delimitation in the Hookeriales. Journal of the Hattori Botanical Laboratory 64: 29–36. Buck WR. 1991. The basis for familial classification of pleurocarpous mosses. In: Miller NG, ed. Advances in bryology, vol. 4, Bryophyte systematics. Berlin/Stuttgart: J. Cramer, 169–185. Buck WR. 1998. Pleurocarpous mosses of the West Indies. Memoirs of the New York Botanical Garden 82: 1–400. Buck WR. 2007. The history of pleurocarp classification: two steps forward, one step back. In: Newton AE, Tangney RS, eds. Systematics association special volume 71, Pleurocarpous mosses, systematics and evolution. Boca Raton, FL: CRC Press, 1–18. Buck WR, Cox CJ, Shaw AJ. 2005. [2004]. Ordinal relationships of pleurocarpous mosses, with special emphasis on the Hookeriales. Systematics and Biodiversity 2: 121–145. Crosby MR. 1974. Toward a revised classification of the Hookeriaceae (Musci). Journal of the Hattori Botanical Laboratory 38: 129–141. Crosby MR, Magill RE, Allen B, He S. 1999. A checklist of the mosses (updated February 2000). St. Louis, MO: Missouri Botanical Garden. Doyle JJ, Doyle JL. 1990. Isolation of plant DNA from fresh tissue. Focus 12: 13–15. 17 Fleischer M. 1908. Die Musci der Flora von Buitenzorg (zugleich Laubmoosflora von Java). Flore de Buitenzorg V, Bd. 3. Leiden: E. J. Brill. Gangulee HC. 1977. Mosses of eastern India and adjacent regions 6. Hookeriales. Calcutta: H. C. Gangulee. Goffinet B, Shaw J, Anderson LE, Mishler BD. 1999. Peristome development in mosses in relation to systematics and evolution. V. Diplolepideae: Orthotrichaceae. The Bryologist 102: 581–594. Gradstein SR, Vitt DH, Anderson RS. 1984. The epizoic occurrence of Daltonia angustifolia (Musci) in Papua New Guinea. Cryptogamie: Bryologie, Lichénologie 5: 47–50. Gradstein SR, Wilson R. 2008. Protonemal neoteny in bryophytes. In: Mohamed H, Baki BB, Nasrulhaq-Boyce A, Lee PKL, eds. Bryology in the new millennium. Kuala Lumpur: University of Malaya, 1–11. Gressitt JL, Samuelson GA, Vitt DH. 1968. Moss growing on living Papuan moss-forest weevils. Nature 217: 765–767. Hedenäs L. 1996. A cladistic overview of the ‘Hookeriales’. Lindbergia 21: 107–143. Hedenäs L. 1999. How important is phylogenetic history in explaining character states in pleurocarpous mosses? Canadian Journal of Botany 77: 1723–1743. Hedenäs L. 2001. Environmental factors potentially affecting character states in pleurocarpous mosses. The Bryologist 104: 72–91. Hedenäs L. 2002. Important complexes of intercorrelated character states in pleurocarpous mosses. Lindbergia 27: 104–121. Hillis DM, Bull JJ. 1993. An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Systematic Biology 42: 182–192. Ho B-C, Tan BC, Nathi Y. 2010. New and noteworthy records of Distichophyllum (Daltoniaceae, Bryopsida) and allied genera in Asia and Australasia. Tropical Bryology 31: 106–122. Huelsenbeck JP, Ronquist F. 2001. MRBAYES. Bayesian inference of phylogeny. Bioinformatics 17: 754–755. Huttunen S, Ignatov MS, Müller K, Quandt D. 2004. Phylogeny and evolution of epiphytism in the three moss families Meteoriaceae, Brachytheciaceae, and Lembophyllaceae. Monographs in Systematic Botany from the Missouri Botanical Garden 98: 328–361. Kelchner SA. 2000. The evolution of non-coding chloroplast DNA and its application in plant systematics. Annals of the Missouri Botanical Garden 87: 482–498. Liu Y, Budke JM, Goffinet B. 2012. Phylogenetic inference rejects sporophyte based classification of the Funariaceae (Bryophyta): rapid radiation suggests rampant homoplasy in sporophyte evolution. Molecular Phylogenetics and Evolution 62: 130–145. Mason-Gamer RJ, Kellogg EA. 1996. Testing for phylogenetic conflict among molecular data sets in the tribe Triticeae (Gramineae). Systematic Biology 45: 524–545. Matteri CM. 1972. Las Hookeriaceae (Musci) andinopatagóniacas. II. Revista del Museo Argentino de Ciencias Naturales ‘Bernardino Rivadavia’ e Instituto Nacional © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• 18 B. -C. Ho ET AL. de Investigación de las Ciencias Naturales, Botánica 4: 243–280. Miller HA. 1982. Bryophyte evolution and geography. Biological Journal of the Linnean Society 18: 145–196. Mitten W. 1859. Musci Indiae Orientalis; an enumeration of the mosses of the East Indies. Journal of the Proceedings of the Linnean Society, Botany, Supplement 1: 1–96; 2: 97–171. Mohamed H, Robinson H. 1991. A taxonomic revision of the moss families Hookeriaceae and Hypopterygiaceae in Malaya. Smithsonian Contributions to Botany 80: i–iv, 1–44. Morrison DA. 2006. Multiple sequence alignment for phylogenetic purposes. Australian Systematic Botany 19: 479– 539. Morrison DA. 2007. Increasing the efficiency of searches for the maximum likelihood tree in a phylogenetic analysis of up to 150 nucleotide sequences. Systematic Biology 56: 988–1010. Müller K. 2005. SeqState – primer design and sequence statistics for phylogenetic DNA data sets. Applied Bioinformatics 4: 65–69. Available at: http://bioinfweb.info/Software/ SeqState [accessed May 2009]. Müller K. 2007. PRAP2 – parsimony and likelihood ratchet analyses with PAUP, v.0.9. Available at: http:// systevol.nees.uni-bonn.de/software/PRAP2 [accessed May 2009]. Nixon KC. 1999. The parsimony ratchet, a new method for rapid parsimony analysis. Cladistics 15: 407–414. Olsson S, Buchbender V, Enroth J, Hedenäs L, Huttunen S, Quandt D. 2009a. Evolution of the Neckeraceae (Bryophyta): resolving the backbone phylogeny. Systematics and Biodiversity 7: 419–432. Olsson S, Buchbender V, Enroth J, Huttunen S, Hedenäs L, Quandt D. 2009b. Phylogenetic analyses reveal high levels of polyphyly among pleurocarpous lineages as well as novel clades. The Bryologist 112: 447–466. Pagel M, Meade A. 2006. Bayesian analysis of correlated evolution of discrete characters by reversible-jump Markov chain Monte Carlo. American Naturalist 167: 808–825. Pagel M, Meade A, Barker D. 2004. Bayesian estimation of ancestral character states on phylogenies. Systematic Biology 53: 673–684. Pattengale ND, Alipour M, Bininda-Emonds ORP, Moret BME, Stamatakis A. 2009. How many bootstrap replicates are necessary? Research in Computational Molecular Biology 5541: 184–200. Pokorny L, Ho B-C, Frahm J-P, Quandt D, Shaw AJ. 2012. Phylogenetic analyses of morphological evolution in the gametophyte and sporophyte generations of the moss order Hookeriales (Bryopsida). Molecular Phylogenetics and Evolution 63: 351–364. Pokorny L, Oliván G, Shaw AJ. 2011. Phylogeographic patterns in two southern Hemisphere species of Calyptrochaeta (Daltoniaceae, Bryophyta). Systematic Botany 36: 542–553. Quandt D, Huttunen S, Tangney R, Stech M. 2009. Back to the future? Molecules take us back to the 1925 classification of the Lembophyllaceae (Bryopsida). Systematic Botany 34: 443–454. Quandt D, Müller K, Huttunen S. 2003. Characterisation of the chloroplast DNA psbt-H region and the influence of dyad symmetrical elements on phylogenetic reconstructions. Plant Biology 5: 400–410. Quandt D, Stech M. 2005. Molecular evolution of the chloroplast trnLUAA intron in bryophytes. Molecular Phylogenetics and Evolution 36: 429–443. Rambaut A, Drummond AJ. 2009. Tracer v1.5. Available at: http://tree.bio.ed.ac.uk/software/tracer/ [accessed December 2009]. Robinson H. 1975. The mosses of Juan Fernandez Islands. Smithsonian Contributions to Botany 27: i–iv, 1–88. Ronquist F. 2004. Bayesian inference of character evolution. Trends in Ecology and Evolution 19: 475–481. Ronquist F, Huelsenbeck JP. 2003. MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574. Sainsbury GOK. 1955. A handbook of the New Zealand mosses. Bulletin of the Royal Society of New Zealand 5: 1–490. Seki T. 1973. Distributional patterns of the Patagonian mosses. Proceedings of the Japan Society of Plant Taxonomists 3: 13–15. Shaw AJ. 2000. Molecular phylogeography and cryptic speciation in the mosses, Mielichhoferia elongata and M. mielichhoferiana (Bryaceae). Molecular Ecology 9: 595–608. Shaw AJ, Cox CJ, Boles SB. 2003. Polarity of peatmoss (Sphagnum) evolution: who says bryophytes have no roots? American Journal of Botany 90: 1777–1787. Shaw AJ, Szövényi P, Shaw B. 2011. Bryophyte diversity and evolution: windows into the early evolution of land plants. American Journal of Botany 98: 1–18. Shaw J, Anderson LE. 1988. Peristome development in mosses in relation to systematics and evolution. II. Tetraphis pellucida (Tetraphidaceae). American Journal of Botany 75: 1019–1032. Shaw J, Anderson LE, Mishler BD. 1987. Peristome development in mosses in relation to systematics and evolution. I. Diphyscium foliosum (Buxbaumiaceae). Memoirs of the New York Botanical Garden 45: 55–70. Shaw J, Anderson LE, Mishler BD. 1989a. Peristome development in mosses in relation to systematics and evolution. III. Funaria hygrometrica, Bryum pseudocapillare, and B. bicolor. Systematic Botany 14: 24–36. Shaw J, Anderson LE, Mishler BD. 1989b. Peristome development in mosses in relation to systematics and evolution. IV. Haplolepideae: Ditrichaceae and Dicranaceae. The Bryologist 92: 314–325. Simmons MP, Ochoterena H. 2000. Gaps as characters in sequence-based phylogenetic analyses. Systematic Biology 49: 369–381. Stöver BC, Müller KF. 2010. TreeGraph 2: combining and visualizing evidence from different phylogenetic analyses. BioMed Central Bioinformatics 11: 7. Available at: http:// treegraph.bioinfweb.info/ [accessed September 2010]. Swofford DL. 2002. PAUP*4b10 – Phylogenetic analysis using parsimony (*and other methods). Windows Version 4. Sunderland, MA: Sinauer Associates. © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–•• EVOLUTION AND PHYLOGENY OF DALTONIACEAE Tan BC, Robinson H. 1990. A review of Philippine hookeriaceous taxa (Musci). Smithsonian Contributions to Botany 75: i–iv, 1–41. Touw A. 2007. Original specimens kept in the Nationaal Herbarium Nederland (Universiteit Leiden Branch) of Asian and South American moss taxa published by F. Dozy, J. H. Molkenboer, R. B. van den Bosch and C. M. van der Sande Lacoste. NHN-Electronic Publication 1: 1–104. Vanderpoorten A, Hedenäs L, Cox CJ, Shaw AJ. 2002. Circumscription, classification, and taxonomy of Amblystegiaceae (Bryopsida) inferred from nuclear and chloroplast DNA sequence data and morphology. Taxon 51: 115–122. Vitt DH. 1977. A taxonomic study of the genus Crosbya (= Bellia; Musci). Canadian Journal of Botany 55: 2079– 2089. 19 Welch WH. 1966. The Hookeriaceae of Mexico. The Bryologist 69: 1–68. Whittemore A, Allen B. 1989. The systematic position of Adelothecium Mitt. and the familial classification of the Hookeriales (Musci). The Bryologist 92: 261–272. de Winton MD, Beever JE. 2004. Deep-water bryophyte records from New Zealand lakes. New Zealand Journal of Marine and Freshwater Research 38: 329–340. Yu J, Devos N, Majestyk P, Shaw AJ. 2010. Intercontinentally disjunct species are derived rather than relictual in the moss genus Daltonia (Bryophyta). Taxon 59: 459–465. Zwickl DJ. 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. Unpublished PhD dissertation, The University of Texas, Austin, TX. Available at: http://garli.nescent.org SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article: Figure S1. Strict consensus cladogram obtained from parsimony ratchet analysis of the concatenated dataset without indels coded, with bootstrap support values noted beside the branches. Values above the branches are analyses that exclude indels (BSw/o); those below the branches are with indels coded for the entire sequence data (BSsic) or coded for only the organellar sequence data (BSsic-org). Figure S2. Majority consensus cladogram for the Bayesian analysis of the concatenated dataset in which indels were coded (BIw/o), with posterior probabilities (PPs) obtained from various modified dataset and model implementation. Support values above the branches are for analyses with indels excluded and use a homogeneous DNA substitution model for the matrix (PPhom), followed by values for an analysis using partitioned models (PPw/o). Values below the branches include indels and consider the entire sequence data (PPsic) or only the organellar sequence data (PPsic-org); in both cases, the matrix was partitioned into different regions. Figure S3. Phylogram of one of the 12 trees from a maximum likelihood analysis (this is the same tree as in Fig. 2, but shows branch lengths). Data file S1. Dalt_final_alignm.nex. Final alignment of the concatenated five-gene data matrix in the order rps4, trnLF, nad5, ITS and 26S. See Table 1 for the nucleotide positions of the genes, hotspots (ambiguous alignments) and inversions. Data file S2. Dalt_comb_hx_ir_ML_trees.nex. Alignment of the five-gene data matrix (without indel coding) used in various phylogenetic analyses. The ambiguous aligned segments (hotspots) have been trimmed. Detected inversions have been reversed and complimented. Twelve resulting maximum likelihood trees were embedded. Table S1. Species sampled, voucher information and GenBank accessions for rps4, trnLF, nad5, ITS and 26S of 126 samples (593 of 630 available). An asterisk (*) indicates the type species of a genus. Country codes follow those of ISO 3166-1 alpha-2; additional regional abbreviations: ID-C, Celebes; ID-J, Java; ID-M, Moluccas; ID-S, Sumatra; MY-E, East Malaysia (Sarawak and Sabah); MY-W, West Malaysia (Peninsula). Sequences generated for this study are written in boldface, those not available are denoted by ‘n.a.’. Herbarium acronyms follow those of the Index Herbariorum. Table S2. Comparsion of likelihood scores and effective sample sizes (ESS) for Bayesian inference under the GTR + G + I substitution model and the restriction site model on various modified datasets. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. © 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2012, ••, ••–••