Plant Syst. Evol. 241: 1–12 (2003) DOI 10.1007/s00606-003-0032-z A taxonomic reassessment of the Vittiaceae (Hypnales, Bryopsida): evidence from phylogenetic analyses of combined chloroplast and nuclear sequence data A. Vanderpoorten1, B. Goffinet2, L. Hedenäs3, C. J. Cox4, and A. J. Shaw4 1 Université Catholique de Louvain, Unité d’Ecologie et Biogéographie, Louvain-la-Neuve, Belgium Department of Ecology and Evolutionary Biology, University of Connecticut, CT, USA 3 Swedish Museum of Natural History, Department of Cryptogamic Botany, Stockhalm, Sweden 4 Department of Biology, Duke University, NC, USA 2 Received November 29, 2001; accepted April 8, 2003 Published online: October 16, 2003
Springer-Verlag 2003 Abstract. The Vittiaceae are a small family of aquatic mosses that are defined based on gametophytic traits whose interpretation has led to conflicting taxonomic arrangements. Phylogenetic analyses of two cpDNA regions, trnL-trnF and atpB-rbcL, indicate that Vittia is nested within the Amblystegiaceae s. str., suggesting that the family Vittiaceae should not be recognized. Platylomella lescurii appears nested within the Thuidiaceae/Leskeaceae. This suggests that the series of character states shared by Vittia and Platylomella, including a differentiated leaf border, short laminal cells, stiff stems, and a thick costa, are convergent features that arose independently in unrelated lineages of aquatic Hypnales. Within the Amblystegiaceae, phylogenetic analyses of the two cpDNA regions combined with ITS sequence data show that Hypnobartlettia, Vittia elimbata spec. nov., V. pachyloma, and V. salina, despite their strong morphological similarity to aquatic Amblystegium species, form a clade that is sister to the Drepanocladus/Pseudocalliergon complex. This combined clade is unresolved at a polytomy that includes Amblystegium serpens and a clade including all the other Amblystegium species. The occurrence of A. serpens outside the strongly supported clade including other Amblystegium species suggests that A. serpens may be better accommodated in a distinct genus. Amblystegium serpens is the type species of Amblystegium and thus retains the name. The other species are accommodated in their own genus, Hygroamblystegium, including H. fluviatile, H. humile comb. nov., H. noterophyllum, H. tenax, and H. varium. Key words: Vittia, Platylomella, Hygroamblystegium, Amblystegiaceae, convergent evolution, aquatic mosses, ITS, atpB-rbcL spacer, trnF-trnL region. Unrelated plant species growing in similar habitats may often exhibit a similar appearance, a process known as convergent evolution. In fast flowing waters, plant species often have a suite of convergent morphological characters, including extensive attachment to their substrates, increased tensile strength, increased flexibility to dissipate energy through elasticity, and/or a streamline shape (Suren et al. 2000). Mosses, which are common in running waters, are rarely flexible but increase strength with a suite of morphological characters such as stiff 2 A. Vanderpoorten et al.: Taxonomic reassessment of the Vittiaceae wiry stems, small thick-walled leaf cells, a strong costa, and often pluri-stratose leaf borders or laminae (Vitt and Glime 1984). Due to an assumed greater exposure of the moss gametophyte to selective environmental pressures, much of the suprageneric classification of mosses has traditionally been based on sporophytic characters. However, the assumption that gametophytic characters are more labile than sporophytic ones has been challenged (e.g. Buck 1991; Buck et al. 2000; Hedenäs 1999, 2001, 2002). As a consequence, a number of gametophytic traits were given primacy at various taxonomic levels in recent moss classifications. For example, within the Hypnales, the most diverse of the two orders of pleurocarpous mosses, the thickness of leaf lamina has been interpreted as an important taxonomic character (Ochyra 1986a) to such an extent that new species [e.g. Palustriella pluristratosa Stech and Frahm (Stech and Frahm 2001b)], genera [e.g. Donrichardsia (Grout) H.A. Crum & L.E. Anderson (Crum and Anderson 1979), Gradsteinia Ochyra (Ochyra 1990, Ochyra et al. 1998), Ochyraea Vana (Vana 1986), Richardsiopsis Ochyra (Ochyra 1986b)] and, in some cases, families [e.g. the Donrichardsiaceae (Ochyra 1985) and the Vittiaceae (Ochyra 1987a)], have been erected to accommodate peculiar pleurocarpous aquatic mosses with variously multistratose leaf laminae. Recent analyses of morphological (Hedenäs 1995, 1997, 1998; Ochyra and Bednarek-Ochyra 1999; Ochyra and Vanderpoorten 1999) and molecular data (Stech and Frahm 1999, 2001a; Stech et al. 1999, Vanderpoorten et al. 2001, 2002a) suggest, conversely, that these taxa are often extreme expressions nested within larger genera, including Platyhypnidium Fleisch., Drepanocladus (Müll. Hal.) Roth, Hygrohypnum Lindb., and Amblystegium Schimp. Hence, the taxonomic position of many aquatic mosses has been an area of controversy. For example, the peculiar gametophytic features of one such aquatic moss, Vittia pachyloma (Mont.) Ochyra (Vittiaceae), including multistratose leaf laminae, short laminal cells, a strong, percurrent to excurrent costa, and a leaf margin composed of many layers of linear, thick-walled cells (Ochyra 1987a), led to a large number of alternative taxonomic placements. Montagne (1838) first interpreted the taxon, with hesitation, as a member of Gymnostomum Nees and Hornsch. (Pottiaceae), indicating its similarity, especially the possession of a thickened leaf border, with Cinclidotus P. Beauv. Subsequently, Hooker and Wilson (1844) transferred the taxon into Sciaromium (Mitt.) Mitt. within the Amblystegiaceae. This treatment was followed in subsequent classification systems (Vitt 1984) until Ochyra (1987a) erected the monospecific genus Vittia and family Vittiaceae to accommodate the taxon. This taxonomic position, however, remained debatable, and was discussed by Hedenäs (1995) who interpreted the taxon as a specialized Amblystegium with a few autapomorphies presumably connected with the adaptation to growth in running water. Most recently, Buck and Goffinet (2000) and Ochyra and Matteri (2001) retained Vittia as a distinct genus within the Amblystegiaceae. Other taxa with at least partly polystratose leaf laminae and a well-defined leaf border, such as Platylomella A. L. Andrews, were also included in the family (Buck and Goffinet 2000). In this paper, we use variation in sequences from chloroplast and nuclear loci to clarify the phylogenetic relationships of Vittia pachyloma and the recently described species, V. salina Hedenäs and Munoz (Hedenäs and Munoz 2002). The analysis also included other aquatic Hypnales to determine whether their morphological similarities reflect common ancestry or convergent responses to growth in running waters. Material and methods Taxon sampling. Two sampling strategies were employed, the first to confirm the placement of Vittia within the Amblystegiaceae and to assess familial relationships of morphologically similar taxa. The second dataset was constructed to assess sister group relationships within the Amblystegiaceae. Vittia and related taxa included six accessions A. Vanderpoorten et al.: Taxonomic reassessment of the Vittiaceae of V. pachyloma from different Chilean provinces; two accessions of V. salina; one accession of a puzzling aquatic moss from Bolivia previously interpreted as an extreme Amblystegium expression; and three accessions of Platylomella lescurii (Sull.) A. L. Andrews, a rheophilous moss endemic to the Appalachian mountains (Table 1). The inclusion of these taxa within the Amblystegiaceae was tested by sampling members of the Amblystegiaceae s. str., plus representative members of Hypnalean families whose potential relationships with the Amblystegiaceae have been suggested by morphological (Hedenäs 1995, 1998; Ochyra and Vanderpoorten 1999) and molecular data (Vanderpoorten et al. 2002b) (dataset 1, appendix 1). These Hypnalean families include the Anomodontaceae (Anomodon Hook. and Tayl.), Thuidiaceae (Thuidium Bruch and Schimp, Haplocladium Müll. Hal., Abietinella Müll. Hal., Helodium Warnst.), Hypnaceae (Hypnum Hedw., Caribaeohypnum Ando and Higuchi, Ptilium De Not.), Rhytidiaceae (Rhytidium (Sull.) Kindb), Brachytheciaceae (Platyhypnidium), and Leskeaceae (Leskea Hedw.). Neckera pennata Hedw. and N. douglasii Hook., two members of the suborder Neckerineae (Brotherus 1925, De Luna et al. 2000), were used as outgroups. Generic relationships within the Amblystegiaceae were investigated using a taxon sampling including members of all the genera currently included within the family sensu stricto (dataset 2, appendix 1). The tree was rooted with Palustriella and Cratoneuron, which are sister to the other members of the family (Vanderpoorten et al. 2002b). Phylogenetic analyses. Familial relationships within the Hypnales were investigated using the two cpDNA regions trnL-trnF and atpB-rbcL that were sequenced according to the protocols described in Vanderpoorten et al. (2002b), and which could be unambiguously aligned across the Hypnales (dataset 1). The cpDNA dataset was analyzed under maximum likelihood after selection of the substitution model best fitting the data. The model was selected on the basis of the likelihood of a neighbor-joining tree under 56 different nested models as implemented by Modeltest 3.04 (Posada and Crandall 1998). The model parameters were subsequently fixed in heuristic searches with 300 random addition replicates with TBR branch swapping. Support for clades was assessed using bootstrap analyses which were conducted with 100 3 replicates, using simple taxon addition, and saving no more than 20,000 trees per replicate. Within the Amblystegiaceae, the sample of closely related species allowed the use of the internal transcribed spacers (ITS) of the 18S-26S nrDNA, in combination with the two cpDNA regions. The decision to combine the partitions was made after comparing both the topologies and the support for the branches as assessed by nonparametric bootstrapping. If a taxon was resolved as part of two distinct monophyletic clades supported each by bootstrap proportions of 70% or higher in bootstrap analyses of individual partitions, these partitions would be considered incongruent (Mason-Gamer and Kellog 1996). In such cases, taxa causing the incongruence were removed from the analysis. The ML analyses were conducted as described above. Results The substitution model maximizing the likelihood of the neighbor-joining tree derived from the first cpDNA dataset ()lnL ¼ 5080.6074) was a general time-reversible model (Rodriguez et al. 1990). The rate heterogeneity among sites was assumed to follow a gamma distribution with a shape parameter of 0.8564. The model was implemented with the following settings: rate matrix R (A-C) ¼ 1.0958; R(A-G) ¼ 2.4858; R(A-T) ¼ 0.3164; R(C-G) ¼ 2.2292; R(C-T) ¼ 2.4858; R(G-T) ¼ 0.3164; base frequencies A ¼ 0.4074; C ¼ 0.1012; G ¼ 0.1252; T ¼ 0.3662; proportion of invariable sites ¼ 0.3762. Heuristic searches under ML resulted in a unique most likely tree ()lnL ¼ 4573.2927) (Fig. 1). Taxa with multistratose laminae, including Platylomella, Vittia, Palustriella, Platyhypnidium, and Donrichardsia, were scattered throughout the phylogeny. The genus Vittia was nested within the Amblystegiaceae s. str., supported with a bootstrap value of 90%. Platylomella lescurii was not included within the family and its relationships with the Leskeaceae/Thuidiaceae complex was supported at 77%. Donrichardsia formed a clade with Platyhypnidium and Hygrohypnum montanum that was supported at 84%. Separate analysis of cpDNA and ITS sequence datasets within the Amblystegiaceae Locality Drepanocladus longifolius (Mitt.) Broth. ex Par. Hypnobartlettia fontana Ochyra Vittia elimbata Hedenäs, Vanderpoorten & Goffinet Vittia pachyloma (Mont.) Ochyra Chile, Prov. of Cordillera Vittia salina Hedenäs & Munoz Platylomella lescurii (Sull.) Andr. Voucher GenBank accession number Habitat (ITS/atpB-rbcL/trnL-trnF) Goffinet 5603 (DUKE, BG, AV) New Zealand, South Island, Frahm X12-1 Nelson District (BONN, AV, BG) Bolivia, La Paz Lewis 87816 (MO, DUKE, S, NY) Chile, Prov. of Cordillera Goffinet 5605 (DUKE, BG) Chile, Prov. of Bio-Bio Goffinet 5582 (DUKE, BG) Chile, Prov. La Araucania Goffinet 5534 (DUKE, BG) Chile, Magallanes Goffinet 6679 (BG) Chile, Prov. of Cordillera Goffinet 5610 (DUKE, BG) Argentina, Catamarca Biasuso et al. 99-60 (S) AY062888/AY062884/ AY062890 AY242368/AY242359/ AY242376 AF464999/AF464976/ AF465017 AY062886/AY062883/ AY062889 AY242365/AY242356/ AY242384 AY242366/AY242357/ AY242383 AY242367/AY242358/ AY242375 AY062887/AY062885/ AY062891 AY242369/AY242360/ AY242377 Argentina, Catamarca Biasuso et al. 99-56a (S) USA, North Carolina Vanderpoorten s.n. (AV, BG) Vanderpoorten s.n. (AV, BG) Vanderpoorten 4889 (AV, BG) AY242370/AY242361/ AY242378 AY242371/AY242362/ AY242379 AY242372/AY242363/ AY242380 AY242373/AY242364/ AY242381 USA, South Carolina USA, North Carolina Mountain stream Limestone dwell, submerged in water In running water at 3810 m Mountain stream Mountain stream Mountain stream Mountain stream Mountain stream Rocks and sandy banks of salt water rivulet at 3700 m Fast running salt water stream at 3000 m Fast-flowing mountain stream Fast-flowing mountain stream Fast-flowing mountain stream A. Vanderpoorten et al.: Taxonomic reassessment of the Vittiaceae Taxon 4 Table 1. Voucher information and GenBank accession numbers of the taxa for which the ITS (nrDNA repeat), atpB-rbcL intergenic spacer, and trnL-trnF region have been obtained and combined with a previous dataset for a phylogenetic reassessment of the Vittiaceae. BG and AV represent the private herbaria of Bernard Goffinet and Alain Vanderpoorten, respectively A. Vanderpoorten et al.: Taxonomic reassessment of the Vittiaceae 5 Hypnum pallescens Ptilium crista-castrensis Rhytidium rugosum Abietinella abietina Haplocladium virginianum Leskea gracilescens 52 Helodium blandowii 77 Thuidium delicatulum Campylophyllum halleri Hygrohypnum smithii 72 Scorpidium revolvens Tomentypnum nitens Hygrohypnum luridum Serpoleskea confervoides Campyliadelphus chrysophyllus 85 59 Amblystegium serpens Drepanocladus aduncus 90 Pseudo-calliergon trifarium 90 Vittia pachyloma Vittia salina Vittia elimbata Palustriella falcata 69 Hygrohypnum ochraceum Sanionia uncinata Hamatocaulis vernicosus Calliergonella cuspidata Loeskypnum badium 100 89 Straminergon stramineum Warnstorfia fluitans Calliergon cordifolium Caribaeohypnum polypterum Platydictya jungermannioides 51 Conardia compacta Platygyrium repens 100 98 84 Platyhypnidium riparioides Hygrohypnum montanum Donrichardsia macroneuron Anomodon attenuatus Neckera douglasii 100 Neckera pennata 0.005 substitutions/site (dataset 2) gave no evidence of incongruence between relationships implied by cpDNA vs. nrDNA. The datasets were consequently combined. The substitution model maximizing the likelihood of the neighbor-joining tree derived Fig. 1. Phylogram of the unique most likely tree from a maximum likelihood analysis of trnL-trnF and atpB-rbcL sequence data for a taxon sampling including the Amblystegiaceae (highlighted in gray) and related Hypnalean families (dataset 1). Branches of maximum length=0 were collapsed. Bootstrap values higher than 50% are indicated below the branches. Taxa in boldface have (or can have in the case of Palustriella and Platyhypnidium) multistratose leaf laminae from the combined data set ()lnL ¼ 4130.4741) was a general time-reversible model (Rodriguez et al. 1990) with the rate heterogeneity among sites following a gamma distribution (shape parameter ¼ 0.1645). The model 6 A. Vanderpoorten et al.: Taxonomic reassessment of the Vittiaceae was implemented with the following settings: rate matrix R (A-C) ¼ 1.0529; R(A-G) ¼ 1.6299; R(A-T) ¼ )3636; R(C-G) ¼ 1.0725;1.0725; R(C-T) ¼ 1.6299; R(G-T) ¼ 1.0000; base frequencies A ¼ 0.3104; C ¼ 0.1892; G ¼ 0.1935; T ¼ 0.3070; proportion of invariable sites ¼ 0. Heuristic searches under ML resulted in a unique most likely tree ()lnL ¼ 4129.1662) (Fig. 2). Vittia constituted with Hypnobartlettia a clade that was sister to the Drepanocladus/Pseudo-calliergon complex (Fig. 2). This combined clade was unresolved at a polytomy that included Amblystegium serpens (Hedw.) Schimp.and a clade, supported at 100%, including all the other Amblystegium species. Within the Vittia clade, accessions of Vittia pachyloma formed a clade supported at 77%. The two accessions of V. salina composed a monophyletic group with 76% support. Vittia elimbata, described as a new species here (Fig. 3, Table 2), appeared to be sister to V. salina. Vittia elimbata shares with all these taxa, except V. salina, a partly bistratose upper leaf lamina but differs from all these species in having unistratose leaf margins. Furthermore, V. elimbata differs from V. pachyloma in having much narrower pseudoparaphyllia. Based on leaf morphology and anatomy, V. elimbata is similar to Hypnobartlettia fontana, but the latter has paraphyllia, its lamina is bistratose almost throughout rather than in the upper portion only, and its pseudoparaphyllia are broader. Discussion Phylogenetic relationships and circumscription of Vittia. The results presented here strongly support recent interpretations based on morphology (Hedenäs 1995, Buck and Goffinet 2000, Ochyra and Matteri 2001) that Vittia is a member of the Amblystegiaceae. Together with other families erected to accommodate other pleurocarpous mosses with multistratose leaf laminae such as the Hypnobartlettiaceae and the Donrichardsiaceae (Hedenäs 1995; Stech and Frahm 1999, 2000, 2001a; Stech et al. 1999), the Vittiaceae should be abandoned. Within the Amblystegiaceae, Vittia forms with Hypnobartlettia a clade that is sister to the Drepanocladus/Pseudo-calliergon complex. This combined clade is unresolved at a polytomy that included Amblystegium serpens and a clade including all the other Amblystegium species. Although the apparent non-monophyly of the genus Amblystegium may reflect insufficient phylogenetic signal in our datasets, the occurrence of A. serpens outside the strongly supported clade including other Amblystegium species suggests that A. serpens may be better accommodated in a distinct genus. Amblystegium serpens is the type species of Amblystegium and thus retains the name, probably together with A. subtile (Hedw.) Schimp. (not included in this study). The other species [A. fluviatile (Hedw.) Schimp., A. humile (P. Beauv.) Crundw., A. tenax (Hedw.) C.E.O. Jensen, A. varium (Hedw.) Lindb., and probably A. noterophyllum (Sull.) Holz. (not included in this study)] should be accommodated in Hygroamblystegium Hedw., a genus that we previously hesitated to recognize based on phylogenetic analyses of a smaller dataset (Vanderpoorten et al. 2002a). Amblystegium tenax, A. fluviatile, and A. noterophyllum, were already tentatively included by some authors within Hygroamblystegium on the basis of a larger size, stronger and longer costa, and a tendency to occur in aquatic habitats (e.g. Buck 1998). Our results support the inclusion of A. varium within Hygroamblystegium (Buck 1998) and we see no reason to recognize the genus Orthotheciella Ochyra for this species (Ochyra 1998, Ochyra and Matteri 2001). Hence, Hypnum sect. Orthotheciella Müll. Hal. [Forschungsreise Gazelle, 4 (Bot.): 36. 1889] and Orthotheciella (Müll. Hal.) Ochyra [The Moss Flora of King George Island Antarctica: 224. 1998] are synomized with Hygroamblystegium. Amblystegium humile, conversely, has been either included within Amblystegium (Crundwell 1981), or even in Leptodictyum because of populations sharing with the latter a distinctly complanate habit (Ochyra 1981). The new combination, Hygroamblystegium humile (P. Beauv.) A. Vanderpoorten et al.: Taxonomic reassessment of the Vittiaceae 7 Fig. 2. Phylogram of the unique most likely tree from a maximum likelihood analysis of trnL-trnF, atpBrbcL and ITS sequence data for the Amblystegiaceae (dataset 2). Branches of maximum length=0 were collapsed. Bootstrap values higher than 50% are indicated below the branches Vanderpoorten, Hedenäs & Goffinet, comb. nov. (Basionym: Hypnum humile P. Beauv., Prodrome des Cinquième et Sixième Familles de l’Aethéogamie 65. 1805), is therefore introduced. As circumscribed here, Hygroamblystegium differs morphologically from both Amblystegium s. str. and Leptodictyum by the occasional presence of foliose paraphyllia. Morphologically, Hygroamblystegium is similar to the Hypnobartlettia/Vittia complex. Both complexes share a series of characters assumed to correlate with the occurrence in fast-flowing water (Vitt and Glime 1984), such as a thick costa and stiff stems. Vittia elimbata especially resembles an extreme expression of Hygroamblystegium in its general appearance 8 A. Vanderpoorten et al.: Taxonomic reassessment of the Vittiaceae Fig. 3. Vittia elimbata spec. nov. A Habit; B Stem leaves; C Branch leaf; D Pseudoparaphyllia; E Axillary hairs; F Median cells of stem leaf lamina; G Alar cells of stem leaf; H Transverse sections of stem leaf, 65–75% way up leaf. Scales: a: F, G; b: E, H; c: D; d: B, C; e: A and microscopical features, including the lack of a defined leaf border and undifferentiated alar cells. The taxa of the Vittia/Hypnobartlettia complex differ from Hygroamblystegium in having a variously multistratose leaf lamina and thickened leaf borders. These features, however, are not strongly expressed in all the species of the complex. In Vittia pachyloma, the leaf margins are (3–)4–8 stratose, and the lamina varies from unistratose to partly or entirely 2–4-stratose, whereas in both V. salina and V. elimbata the bistratose portions of the lamina are less strongly developed. Vittia salina has a partly or entirely bistratose leaf border sometimes becoming 3(-4)-stratose near the base but has an otherwise unistratose lamina. Vittia elimbata has small bistratose portions in its upper lamina, especially close to the costa. The same kinds of peculiar gametophytic features are encountered in the Haitian Limbella bartlettii (H. A. Crum & Steere) W. R. Buck but are combined in a different way. Limbella bartlettii possesses a 2(-3)-stratose leaf margin and a leaf lamina that is bistratose A. Vanderpoorten et al.: Taxonomic reassessment of the Vittiaceae 9 Table 2. Description of Vittia elimbata Hedenäs, Vanderpoorten & Goffinet, sp. nov. Diagnosis: V. salinae affinis, sed limbidio absenti et lamina folii superne ex parte bistratosa. Type. Bolivia. Dept. La Paz, Inquisivi, Laguna Ventanani, along the Arco Pongo trail between JapoCargadero, SW slope of Cerro Ventanani c. 3 km N of Japo and 28 km NE of Inquisivi, 1645¢ S, 6656¢ W, 3810 m a.s.l., in running water on face of small waterfall, 6 August 1987, M. Lewis 87816 (DUKE, holotype; isotypes in MO, NY, S). Plants medium-sized, yellow-green. Stem stiff, irregularly or pinnately branched in one plane, with a central strand, and a cortex of 2(-3) layers of small and incrassate cells; pseudoparaphyllia triangular or narrowly so; paraphyllia absent; axillary hairs with 1-celled upper part, 8.5–10.5 lm wide, hyaline when young, basal 1–2 cells quadrate or rectangular, brown; rhizoids not seen, but rhizoid initials present on stem just below leaf costa insertion. Stem leaves 1.6–1.7 mm long, 0.6–0.8 mm wide, erect-patent to almost patent, straight or slightly homomallous, rounded-triangular or ovate-triangular, gradually narrowed to acuminate or sometimes blunt apex, slightly concave; margin finely and obtusely denticulate; costa single, percurrent or shortly excurrent, 105.0–136.5 lm wide at base, diffusely delimited above, 4–5-stratose below, dorsal layers of smaller cells than ventral layers, in upper leaf biconvex, (4-)5-stratose; median laminal cells 21.0– 61.0 · 6.5–10.0(-10.5) lm, slightly incrassate, eporose, in upper part near costa bistratose, close to leaf apex frequently bistratose in larger portions; alar cells rectangular, shortly rectangular or especially in upper part of group often quadrate, not or sometimes in basalmost portion slightly widened, forming a relatively large, rather indistinct, transverse or triangular group that extends from margin 1/2–2/3 of distance to costa, shortly decurrent. Branch leaves smaller and narrower than stem leaves, proximal branch leaves ovate to orbicular, with acute or rounded apiculate apex. [Sexual branches and sporophyte unknown]. in streaks. This species most likely also belongs to Vittia, but the lack of recent collections precluded the inclusion of the species in the present molecular study. Morphological evolution of Vittia. A multistratose leaf lamina is a synapomorphy for the Vittia/Hypnobartlettia complex, which shares this peculiar leaf anatomy with many unrelated aquatic Hypnales such as Donrichardsia macroneuron, Platyhypnidium torrenticola (Ochyra, C. Schmidt & Bültmann) Ochyra & BednarekOchyra, P. mutatum Ochyra & Vanderpoorten, and Ochyraea tatrensis Vana. Thus, polystratification of leaf lamina appears as a convergent feature associated with strong water flow that independently arose in a number of unrelated aquatic Hypnalean taxa. The results thus support previous interpretations that morphological features associated with physical reinforcement of leaves are homoplastic and unstable (Vitt and Glime 1984). However, the present analyses do not support the interpretation that all the members of the families formerly defined by such features, including the Donrichardsiaceae, Hypnobartlettiaceae, and Vittiaceae, are extreme forms of Amblystegiaceae (Crosby et al. 1999, Buck and Goffinet 2000). In fact, among the aquatic pleurocarpous mosses with multistratose leaf laminae, molecular evidence only confirms the placement of Hypnobartlettia, Vittia, and Gradsteinia andicola within the Amblystegiaceae. Richardsiopsis lacustris (Herz. & Rich.) Ochyra, a rare endemic moss occurring in Brazil, Chile, Ecuador, and Peru, is synonym to Drepanocladus perplicatus (Dus.) Roth and also belongs to the family (Hedenäs 1997). Conversely, the data presented here show that Platylomella lescurii, which has traditionally been interpreted as a specialized rheophilous Amblystegium solely differing by a differentiated leaf border (Crum and Anderson 1981), is related to the Thuidiaceae/Leskeaceae. Similarly, Ochyraea tatrensis was shown to be closely related to Hygrohypnum smithii (Stech and Frahm 2001a), a taxon that formed with Campylophyllum a clade sister to the Leskeaceae/ 10 A. Vanderpoorten et al.: Taxonomic reassessment of the Vittiaceae Thuidiaceae and the Hypnaceae in the present analyses. Limbella tricostata (Sull.) Müll. Hal. ex E.B. Bartram was shown to be unrelated to the Amblystegiaceae and transferred to the Thamnobryaceae (Ochyra 1987b), a hypothesis supported by the results of a phylogenetic survey of the pleurocarps based on rbcL sequence data (Arikawa and Higuchi 1999). These results parallel previous analyses of morphological evolution in the Hypnales, suggesting species associations that were not expected from analyses of morphological variation (Vanderpoorten et al. 2002b). Phylogenetic analyses of pleurocarpous mosses including molecular evidence indicate that convergent evolution and reversals in morphological characters have been common, and underline that in many cases molecular data are necessary to resolve relationships at the familial level. This study was made possible through financial support by the Belgian Funds for Scientific Research (FNRS) to AV and by NSF Grant DEB-0089633 to BG and DEB-0089131 to AJS. The authors thank N. Hax for her assistance with laboratory work, S. McDaniel for discussing with us of a first draft of this paper, R. Ochyra who confirmed the identification of Drepanocladus longifolius, and J.-P. Frahm for providing material of Hypnobartlettia fontana. Appendix 1. Taxon sampling and Genbank accession numbers (trnL-trnF, atpB-rbcL, ITS) from Vanderpoorten et al. (2002b) Abietinella abietina (Hedw.) Fleisch., AY009850, AF322308, AY009802-Amblystegium fluviatile (Hedw.) B, S. & G., AY009822, AF322324, AF168154-A. humile (P. Beauv.) Crundw., AY009823, AF322359, AF168165- A. serpens (Hedw.) B., S. & G., AY009827, AF322326, AF168152- A. tenax (Hedw.) C. Jens. #I, AY980024, AF322360, AF168164- A. tenax (Hedw.) C. Jens. #III, AY009821, AF322327, AF168157- A. varium (Hedw.) Lindb., AY009825, AF322328, AF168159- Anacamptodon splachnoides (Brid.) Brid., AY009816, AF322336, AY009810- Anomodon attenuatus (Hedw.) Hueb., AY009851, AF322309, AF168133Calliergon cordifolium (Hedw.) Kindb, AY009836, AF322341, AF168146- Calliergonella cuspidata (Hedw.) Loeske, AY009859, AF322310, AF168145- Campyliadelphus chrysophyllus (Brid.) Kanda, AY009831, AF322355, AF168150- Campylium stellatum (Hedw.) C. Jens., AY009832, AF322354, AF168151- Campylophyllum halleri (Sw. ex Hedw.) Fleisch., AY009853, AF322311, AF168134- Caribaeohypnum polypterum (Mitt.) Ando & Higuchi, AY009846, AF322353, AY009799- Conardia compacta (Muell. Hal.) Robins., AY009865, AF322312, AY009806- Cratoneuron filicinum (Hedw.) Spruce, AY009817, AF322332, AY009812- C. filicinum#2 (Hedw.) Spruce, AY009826, AF322362, AF168155- Cratoneuropsis relaxa (Hook. & Wils.) Fleisch., AF098525, AF322337, AF152388-AF152391- Donrichardsia macroneuron (Grout) Crum & Anderson, AY009848, AF322323, AF167350- Drepanocladus aduncus (Hedw.) Warnst., AY009828, AF322331, AF180949- D. sendtneri (Schimp. ex Muell. Hal.) Warnst., AY009818, AF322338, AY009811- D. sordidus (Muell. Hal.) Hedenäs, AY009868, AF322333, AY009792- Hamatocaulis vernicosus (Mitt.) Hedenäs, AY009819, AF322343, AF315073- Haplocladium virginianum (Brid.) Broth., AF161133, AF322305, AF168160- Helodium blandowii (Web. & Mohr) Warnst., AY009852, AF322313, AY009803- Hygrohypnum luridum (Hedw.) Jenn., AY009862, AF322339, AF168137- H. montanum (Lindb.) Broth., AY009863, AF322320, AY009804- H. ochraceum (Wils.) Loeske, AY009861, AF322345, AF168138- H. smithii (Sw.) Broth., AY009856, AF322306, AF168139- Hypnum pallescens (Hedw.) P. Beauv., AY009844, AF322314, AY009796- Leptodictyum riparium (Schimp.) Warnst., AY009830, AF322325, AF168163- Leskea gracilescens Hedw., AF161135, AF322356, AF176277- Neckera douglasii Hook., AF315070, AF322358, AY009808- N. pennata Hedw., AF315072, AF322357, AY009809- Palustriella falcata (Brid.) Hedenäs, AY009829, AF322330, AF168158- Platydictya jungermannioides (Brid.) Crum, AY009857, AF322307, AF168162- Platygyrium repens (Brid.) Schimp., AF161131, AF322317, AY009798- Platyhypnidium riparioides (Hedw.) Dix., AF260908, AF230981-AF230996-Pseudo-calliergon trifarium (Web. & Mohr) Loeske, AY009835, AF322329, AY009793- Pseudo-calliergon turgescens (Jens.) Loeske, AY009843, AF322335, AY009794- Ptilium crista-castrensis (Hedw.) De Not., AY009847, AF322316, AY009800Rhytidium rugosum (Hedw.) A. Vanderpoorten et al.: Taxonomic reassessment of the Vittiaceae 11 Appendix 1 (Continued) Kindb., AY009849, AF322318, AY009801- Sanionia uncinata (Hedw.) Loeske, AY009860, AF322321, AF168148- Serpoleskea confervoides (Brid.) Kartt. AY009858, AF322334, AF168142- Straminergon stramineum (Kindb.) Hedenäs, AY009833, AF322349, AF168143- Thuidium delicatulum (Hedw.) B., S. & G., AF161132, AF322322, AF176278- Tomentypnum nitens (Hedw.) Loeske, AY009854, AF322352, AF168161- Warnstorfia fluitans (Hedw.) Warnst., AY009838, AF322350, AF168149. References Arikawa T., Higuchi M. (1999) Phylogenetic analysis of the Plagiotheciaceae (Musci) and its relatives based on rbcL sequences. Cryptogamie, Bryol. 20: 231–245 Brotherus V. F. (1925) Musci (Laubmoose). In: Engler A. (ed.) Die natürlichen Pflanzenfamilien. W. Engelmann, Leipzig, pp. 32–349. Buck W. R. (1991) The basis for familial classification of pleurocarpous mosses. Adv. Bryol. 4: 169–185. Buck W. R. (1998) Pleurocarpous mosses of the west Indies. Mem. New York Bot. Gard. 82: 1–400. Buck W. R., Goffinet B. (2000) Morphology and classification of mosses (Bryopsida). In: Shaw A. J., Goffinet B. (eds.) Bryophyte biology. Cambridge University Press, Cambridge, pp. 71–123. Buck W. R., Goffinet B., Shaw A. J. (2000) Testing morphological concepts of orders of pleurocarpous mosses (Bryophyta) using phylogenetic reconstruction based on trnL-trnF and rps4 sequences. Mol. Phylogenet. Evol. 16: 180–198. Crosby M. R., Magill R. E., Allen B., He S. (1999) A checklist of the mosses. Missouri Botanical Garden, St Louis. Crum H. A., Anderson L. E. (1979) Donrichardsia, a new moss genus of Amblystegiaceae (Musci). Fieldiana, Botany N.S. 1: 1–8. Crum H. A., Anderson L. E. (1981) Mosses of Eastern North America. Columbia University Press, NY. Crundwell A. C. (1981) New combinations in European mosses. II. Gymnostomum and Amblystegium. J. Bryol. 11: 603–604. De Luna E., Buck W. R., Akiyama H., Arikawa T., Tsubota H., Gonzalez D., Newton A. E., Shaw A. J. (2000) Ordinal phylogeny within the Hypnobryalean pleurocarpous mosses inferred from cladistic analyses of three chloroplast DNA sequence datasets: trnL-F, rps4, and rbcL. Bryologist 103: 242–256. Hedenäs L. (1995) Higher taxonomic level relationships among diplolepidous pleurocarpous mosses – a cladistic overview. J. Bryol. 18: 723–781. Hedenäs L. (1997) The Drepanocladus s. str. species with excurrent costae (Musci: Amblystegiaceae). Nov. Hedw. 64: 535–547. Hedenäs L. (1998) An evaluation of phylogenetic relationships among the Thuidiaceae, the Amblystegiaceae, and the temperate members of the Hypnaceae. Lindbergia 22: 101–133. Hedenäs L. (1999) How important is phylogenetic history in explaining character states in pleurocarpous mosses? Canad. J. Bot. 77: 1723–1743. Hedenäs L. (2001) Environmental factors potentially affecting character states in pleurocarpous mosses. Bryologist 104: 72–91. Hedenäs L. (2002) Important complexes of intercorrelated character states in pleurocarpous mosses. Lindbergia 27: 104–121. Hedenäs L., Munoz J. (2002) Vittia salina L. Hedenäs & J. Munoz, sp. nov., a new moss from Argentina. J. Bryol. 24: 163–165. Hooker J. D., Wilson W. (1844) Musci antarctici; being characters with brief descriptions of the new species of mosses discovered during the voyage of H. M. Discovery ships, Erebus and Terror, in the southern circumpolar regions, together with those of Tasmania and New Zealand. J. Bot. (London) 3: 533–566. Mason-Gamer R., Kellogg E. (1996) Testing for phylogenetic conflict among molecular data sets in the tribe Triticeae (Gramineae). Syst. Biol. 45: 524–545. Montagne C. (1838) Centurie de plantes cellulaires exotiques nouvelles. Musci Ann. Sci. Nat. Bot. sér. 2, 9: 49–57. Ochyra R. (1981) Musci Poloniae exsicatti, ab Instituto Botanico Academiae Scientarum Polonae editi, centuria V. Fragm. Flor. Geobot. 26: 367–418. Ochyra R. (1985) Hypnobartlettia fontana gen. et sp. nov. (Musci: Hypnobartletticaeae fam. nov.), a unique moss from New Zealand. Lindbergia 11: 2–8. 12 A. Vanderpoorten et al.: Taxonomic reassessment of the Vittiaceae Ochyra R. (1986a) Sciaromiadelphus A. Abr. and I. Abr.- the relationship between extant and fossil moss specimen. J. Hatt. Bot. Lab. 61: 65–74. Ochyra R. (1986b) On the taxonomic position of Sciaromium lacustre Herz. and Rich. in Rich. J. Bryol. 14: 109–115. Ochyra R. (1987a) A revision of the moss genus Sciaromium (Mitt.) Mitt. II. The section Limbidium Dusen, with a description of Vittia gen. nov. (Vittiaceae, fam. nov.). J. Hatt. Bot. Lab. 62: 387–415. Ochyra R. (1987b) On the taxonomy and family placement of the moss genus Limbella (C. Müll.) Broth. J. Bryol. 14: 465–485. Ochyra R. (1990) Gradsteinia andicola, a remarkable aquatic moss from South America. Trop. Bryol. 3: 19–28. Ochyra R. (1998) The Moss Flora of King George Island Antarctica. Polish Academy of Sciences, W. Szafer Institute of Botany, Cracow. Ochyra R., Bednarek-Ochyra H. (1999) Platyhypnidium grolleanum, a new species from the Sudetes (Central Europe). Haussknechtia Beih. 9: 256–264. Ochyra R., Matteri C. M. (2001) Bryophyta, Musci, Amblystegiaceae. Flora criptogámica de Tierra del Fuego 14 (10): 1–95. Ochyra R., Vanderpoorten A. (1999) Platyhypnidium mutatum, a mysterious new moss from Germany. J. Bryol. 21: 183–189. Ochyra R., Schmidt C., Bültmann H. (1998) Gradsteinia torrenticola, a new aquatic moss species from Tenerife. J. Bryol. 20: 403–409. Posada D., Crandall K. A. (1998) Modeltest: Testing the model of DNA substitution. Bioinformatics 14: 817–818. Rodriguez F., Oliver J. F., Marin A., Medina J. R. (1990) The general stochastic model of nucleotide substitution. J. Theor. Biol. 142: 485–501. Stech M., Frahm J. P. (1999) The status and systematic position of Platyhypnidium mutatum and the Donrichardsiaceae based on molecular data. J. Bryol. 21: 191–195. Stech M., Frahm J. P. (2000) The systematic position of Gradsteinia andicola Ochyra (Donrichardsiaceae, Bryopsida): Evidence from nrDNA internal transcribed spacer sequences. Trop. Bryol. 18: 75–85. Stech M., Frahm J. P. (2001a) The systematic position of Ochyraea tatrensis (Hypnobartlettiaceae, Bryopsida) based on molecular data. Bryologist 104: 199–203. Stech M., Frahm J. P. (2001b) Palustriella pluristratosa spec. nov (Amblystegiaceae, Bryopsida), a new aquatic moss species with pluristratose lamina from Switzerland. Bot. Helvet. 111: 139–150. Stech M., Frey W., Frahm J. P. (1999) The status and systematic position of Hypnobartlettia fontana Ochyra and the Hypnobartlettiaceae based on molecular data. Studies in austral temperate rain forest bryophytes 4. Lindbergia 24: 97–102. Suren A. M., Smart G. M., Smith R. A., Brown S. L. R. (2000) Drag coefficients of stream bryophytes: experimental determination and ecological significance. Freshw. Biol. 45: 309–317. Vana J. (1986) Ochyraea tatrensis gen. et spec. nov., a remarkable pleurocarpous moss from Czchoslovakia. J. Bryol. 14: 261–267. Vanderpoorten A., Shaw A. J., Goffinet B. (2001) Testing controversial alignments in Amblystegium and related genera (Amblystegiaceae: Bryopsida). Evidence from rDNA ITS sequences. Syst. Bot. 28: 470–479. Vanderpoorten A., Hedenäs L., Cox C. J., Shaw A. J. (2002a) Circumscription, classification, and taxonomy of Amblystegiaceae inferred from nuclear and chloroplast DNA sequence data and morphology. Taxon 51: 115–122. Vanderpoorten A., Hedenäs L., Cox C. J., Shaw A. J. (2002b) Phylogeny and morphological evolution of the Amblystegiaceae (Bryopsida). Mol. Phylogenet. Evol. 23: 1–21. Vitt D. H. (1984) Classification of the Bryopsida. In: Schuster R. M. (ed.) New manual of bryology, vol 2. Hattori Botanical Laboratory, Nichinan, pp. 697–759. Vitt D. H., Glime J. M. (1984) The structural adaptations of aquatic Musci. Lindbergia 10: 95–110. Addresses of the authors: A. Vanderpoorten, Université Catholique de Louvain, Unité d’Ecologie et Biogéographie, 4–5 Place Croix du Sud, B-1348 Louvain-la-Neuve, Belgium (e-mail: vanderpoorten@ecol.ucl.ac.be). B.Goffinet, Department of Ecology and Evolutionary Biology, 75 North Eagleville Road, University of Connecticut, Storrs, CT 06268–3043 USA. L. Hedenäs, Swedish Museum of Natural History, Department of Cryptogamic Botany, Box 50007, S-104 05 Stockholm, Sweden. C. J. Cox and A. J. Shaw, Duke University, Department of Biology, Durham, NC 27708, USA.