This article was downloaded by: [Stech, Michael] On: 24 June 2010 Access details: Access Details: [subscription number 923331511] Publisher Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 3741 Mortimer Street, London W1T 3JH, UK Systematics and Biodiversity Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t913521959 Molecular, phytochemical and morphological characterization of the liverwort genus Radula in Portugal (mainland, Madeira, Azores) Michael Stecha; Manuela Sim-Simbc; M. Glória Esquíveld; Leena luísc; Susana Fontinhae; Carlos Lobof; César Garciac; Soraia Martinsc; Cristiana Vieirag; José Barrosoh; Luis G. Pedroh; A.Cristina S. Figueiredoh a Nationaal Herbarium Nederland, Universiteit Leiden branch, Leiden, the Netherlands b Faculdade de Ciências de Lisboa, DBV, Centro de Biologia Ambiental, Universidade de Lisboa, Lisboa, Portugal c Museu Nacional de História Natural, Lisboa, Portugal d Dep. Botânica e Engenharia Biológica, Centro de Botânica Aplicada à Agricultura, Lisboa, Portugal e Centro de Estudos da Macaronésia, Universidade da Madeira, Funchal, Madeira, Portugal f Jardim Botânico da Madeira, Funchal, Madeira, Portugal g Faculdade de Ciências do Porto, Departamento de Botânica, CIBIO/Departamento de Botânica - FCUP, Porto, Portugal h Faculdade de Ciências de Lisboa, DBV, Instituto de Biotecnologia e Bioengenharia, Centro de Biotecnologia Vegetal, Universidade de Lisboa, Lisboa, Portugal Online publication date: 23 June 2010 To cite this Article Stech, Michael , Sim-Sim, Manuela , Esquível, M. Glória , luís, Leena , Fontinha, Susana , Lobo, Carlos , Garcia, César , Martins, Soraia , Vieira, Cristiana , Barroso, José , Pedro, Luis G. and Figueiredo, A.Cristina S.(2010) 'Molecular, phytochemical and morphological characterization of the liverwort genus Radula in Portugal (mainland, Madeira, Azores)', Systematics and Biodiversity, 8: 2, 257 — 268 To link to this Article: DOI: 10.1080/14772001003723579 URL: http://dx.doi.org/10.1080/14772001003723579 PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. Systematics and Biodiversity (2010), 8(2): 257–268 Research Article Molecular, phytochemical and morphological characterization of the liverwort genus Radula in Portugal (mainland, Madeira, Azores) MICHAEL STECH1, MANUELA SIM-SIM2,3, M. GLÓRIA ESQUÍVEL4, LEENA LUÍS3, SUSANA FONTINHA5, CARLOS LOBO6, CÉSAR GARCIA3, SORAIA MARTINS3, CRISTIANA VIEIRA7, JOSÉ BARROSO8, LUIS G. PEDRO8 & A. CRISTINA S. FIGUEIREDO8 1 Nationaal Herbarium Nederland, Universiteit Leiden branch, PO Box 9514, 2300 RA Leiden, the Netherlands Universidade de Lisboa, Faculdade de Ciências de Lisboa, DBV, Centro de Biologia Ambiental, C2, Campo Grande, 1749-016 Lisboa, Portugal 3 Museu Nacional de História Natural, Jardim Botânico/CBA, Rua da Escola Politécnica, n◦ 58, 1250-102 Lisboa, Portugal 4 Centro de Botânica Aplicada à Agricultura, Dep. Botânica e Engenharia Biológica, ISA, UTL, 1399 Lisboa, Portugal 5 Centro de Estudos da Macaronésia, Universidade da Madeira, Campus da Penteada, 9000-390 Funchal, Madeira, Portugal 6 Jardim Botânico da Madeira, Caminho do Meio, Quinta do Bom Sucesso, 9050-251 Funchal, Madeira, Portugal 7 CIBIO/Departamento de Botânica – FCUP, Faculdade de Ciências do Porto, Departamento de Botânica, Edifı́cio FC4, Rua Do Campo Alegre, s/n◦ , 4169-007 Porto, Portugal 8 Universidade de Lisboa, Faculdade de Ciências de Lisboa, DBV, Instituto de Biotecnologia e Bioengenharia, Centro de Biotecnologia Vegetal, C2, Campo Grande, 1749-016 Lisboa, Portugal Downloaded By: [Stech, Michael] At: 06:32 24 June 2010 2 (Received 4 June 2009; revised 16 November 2009; accepted 23 February 2010) Relationships of the eight species of the liverwort genus Radula occurring in Portugal (mainland, the Madeira and Azores archipelagos), including the Macaronesian endemics R. jonesii and R. wichurae, were evaluated based on molecular, phytochemical and morphological–anatomical data. Maximum parsimony and Bayesian analyses were performed with sequences from three plastid DNA markers (trnS GGA -rps4 spacer, rps4 gene, trnLUAA intron), volatile oil compounds, as well as qualitative morphological–anatomical characters. In addition, the molecular data were subjected to maximum likelihood analysis. The eight taxa, R. aquilegia, R. carringtonii, R. complanata, R. holtii, R. jonesii, R. lindenbergiana, R. nudicaulis and R. wichurae, can be clearly distinguished from each other, either by molecular data alone or by combination of characters from all three data sets. Radula aquilegia is monophyletic according to the molecular data, but shows considerable, yet undescribed intraspecific morphological and phytochemical variability. Recognition of R. complanata and R. lindenbergiana as separate species, previously based solely on the paroecious vs. dioecious sexual condition, is moderately supported by the molecular phylogenetic analyses and strongly supported by the phytochemical data. The Radula species, narrowly distributed in Macaronesia and Atlantic Europe, probably have two different origins. For Radula holtii and R. nudicaulis, connections with Radula species from the Neotropics are indicated. The other species, among them the two Macaronesian endemics, are closely related with the R. complanata/R. lindenbergiana complex, which is widely distributed in the northern hemisphere. Key words: Macaronesia, Portugal, Radula, species characterization, trnS-rps4 region, trnLUAA intron, volatile oils Introduction Radula Dumort., the only genus of the family Radulaceae (Porellanae, Jungermanniidae, Marchantiophyta), comprises about 150 (–200) species (Frey & Stech, 2009). The genus is morphologically characterized by incubous, conduplicate-bilobed lateral leaves with the postical lobe (lobule) markedly smaller than the antical lobe. Radula has traditionally been postulated to repreCorrespondence to: Michael Stech. E-mail: stech@nhn. leidenuniv.nl ISSN 1477-2000 print / 1478-0933 online  C 2010 The Natural History Museum DOI: 10.1080/14772001003723579 sent an isolated lineage within leafy liverworts, based on morphological–anatomical (e.g. nearly always Radula-type branching, total absence of underleaves, lack of transfer cells in the gametophyte-sporophyte junction), cytological and phytochemical characters (Schuster, 1980a; Yamada, 1988). Radula’s species distribution is mainly tropical, but extends as far as southern Greenland in the North and subantarctic Campbell Island in the South (Renner & Braggins, 2004). In Europe, the highest diversity occurs in Atlantic Europe and the Macaronesian archipelagos. Eight Downloaded By: [Stech, Michael] At: 06:32 24 June 2010 258 M. Stech et al. Radula species are known to occur in Portugal (Bouman & Dirkse, 1990; Söderström et al., 2002; Luı́s et al., 2005): R. aquilegia (Hook.f. & Taylor) Gottsche, Lindenb. & Nees, R. carringtonii J.B. Jack, R. complanata (L.) Dumort., R. holtii Spruce, R. lindenbergiana Gottsche ex C. Hartm., R. nudicaulis Steph., and the Macaronesian endemics R. jonesii Bouman, Dirkse & Yamada and R. wichurae Steph. Of the latter two species, Radula wichurae is classified as ‘vulnerable’ (Sim-Sim et al., 2008) and R. jonesii as ‘endangered’ (IUCN, 2008). Three species are known to mainland Portugal (R. complanata, R. holtii, R. lindenbergiana), seven to Madeira (R. aquilegia, R. carringtonii, R. holtii, R. jonesii, R. lindenbergiana, R. nudicaulis, R. wichurae), and six to the Azores (R. aquilegia, R. carringtonii, R. holtii, R. lindenbergiana, R. nudicaulis, R. wichurae). In the study area, the genus Radula is abundant on tree trunks and branches (more rarely on leaves), periodically wet rocks or shaded vertical rock walls, soil slopes, and frequently growing on other bryophytes. Its diversity and abundance becomes higher in mountainous areas of the Macaronesian archipelagos (Luı́s et al., 2005). Resolving intrageneric relationships within Radula has long been problematic due to the uniformity and phylogenetic uninformativeness of many morphological– anatomical and karyological characters (e.g. Renner & Braggins, 2004; Zheng & Zhu, 2009). Earlier classifications such as the distinction of 11 sections (Castle, 1937, 1969; Grolle, 1970) were considered artificial by Yamada (1979) and Schuster (1980a, 1980b), who distinguished three subgenera (Cladoradula, Odontoradula, Radula) or four subgenera (Cladoradula, Metaradula, Odontoradula, Radula), respectively. Both systems were based on revisionary work on species from different geographic regions (Asian vs. mainly Laurasian taxa) and considered different morphological–anatomical characters (cf. the overview in Renner & Braggins, 2004). In contrast, Jones (1977), in his revision of the African Radula species, refrained from using intrageneric ranks and only indicated informal groups. Although revisionary work on the genus Radula is going on (e.g. Yamada, 1999; So, 2005a, 2005b, 2006) and new potentially informative characters are being employed (Renner & Braggins, 2005), the intrageneric classification remains problematic as long as no worldwide revision and no molecular phylogenetic studies have been carried out. Not only the intrageneric classification but also the delimitation of several species, among them European species that are subject of the present study, is problematic. For example, Radula lindenbergiana was considered a subspecies of R. complanata by Schuster (1980b), and several authors believed that sterile plants cannot be referred to either species based on morphology or habitat preferences (Bouman & Dirkse, 1990; Paton, 1999; Luı́s et al., 2005; Schumacker & Váňa, 2005). However, in European floras and checklists R. lindenbergiana is still often treated as a separate species (e.g. Schumacker & Váňa, 2005; Frey et al., 2006). The distinction between R. aquilegia and R. carringtonii is not completely clear either, which holds true especially for the Madeiran and some Azorean populations, as the diagnostic morphological characters used to identify these two species are not always reliable (e.g. Bouman & Dirkse, 1990; Paton, 1999). Although systematic work on the Macaronesian Radula species led to improved morphological species circumscriptions (Bouman & Dirkse, 1990), independent data sets are needed to evaluate the significance of the morphological characters and to assess the phylogenetic and biogeographic relationships of the respective taxa. As the genus Radula is well-known for its diversity of chemical compounds, we recently evaluated the chemotaxonomical value of volatile oil characters to distinguish the Portuguese Radula species by similarity analysis (Figueiredo et al., 2009). These data allow characterizing of all morphological species, except for R. aquilegia, which appeared as separated into two groups, and the closely related R. complanata and R. lindenbergiana. In the present paper we provide the first molecular analysis, based on plastid trnS GGA -rps4 spacer, rps4 gene, and trnLUAA intron sequences, of European Radula species, focusing on Portugal (mainland, Madeira and Azores archipelagos). Similar to recent molecular studies on the liverwort genus Plagiochila (Sim-Sim et al., 2005) and the moss genus Campylopus (Stech et al., 2007) in Madeira, we aim at obtaining a molecular species inventory of the genus Radula in Portugal. Additionally, we intend to provide improved species circumscription of European Radula species based on comparative phylogenetic analyses of molecular, phytochemical, and morphological–anatomical data as well as results from morphological revision of herbarium material. Materials and methods Plant material Material of Radula species from Portugal was collected during field trips by the authors to mainland Portugal (2001–2008), Madeira (2001–2008) and Azores (2007–2008). For R. complanata, material from Switzerland (collected by one of the authors in 2008) was used in the molecular and phytochemical analyses, as no suitable material from Portugal was available. In addition to the new collections, herbarium material from LISU, MADS, MADM, MADJ, PO, COI, S, L and B was used for morphological studies and partly for molecular analysis. In total, about 600 Radula specimens were determined and studied morphologically. Of these, 52 specimens were selected for detailed morphological–anatomical analysis, because they were recently collected, well-developed and/or presenting fruiting structures, and phytochemical and/or molecular data were available as well. Voucher information and geographic origins of the specimens included in the phylogenetic Characterization of the liverwort Radula in Portugal 259 Downloaded By: [Stech, Michael] At: 06:32 24 June 2010 Table 1. Geographic origin and voucher information of the Radula specimens and outgroup representatives analysed. Specimens included in the morphological analysis are indicated by an asterisk. Specimens included in the phytochemical and molecular analyses are identified by their abbreviations used in Figueiredo et al. (2009) or by their GenBank accession numbers, respectively. Taxon Abbr. Geographic origin Voucher no. (herb.) Abbr. in Figueiredo et al. Acc. no. (2009) trnS-rps4 R. antilleana Castle ∗ R. aquilegia (Hook.f. & Taylor) Gottsche, Lindenb. & Nees ∗ R. aquilegia ∗ R. aquilegia ∗ R. aquilegia ∗ R. aquilegia ∗ R. aquilegia ∗ R. aquilegia ∗ R. boryana (F. Weber) Nees ∗ R. carringtonii J.B.Jack ∗ R. carringtonii ∗ R. carringtonii ∗ R. carringtonii ∗ R. carringtonii ∗ R. carringtonii ∗ R. carringtonii ∗ R. carringtonii ∗ R. carringtonii ∗ R. carringtonii ∗ R. carringtonii ∗ R. carringtonii ∗ R. carringtonii ∗ R. complanata (L.) Dumort. ∗ R. complanata R. complanata R. episcia Spruce R. grandis Steph. R. gottscheana Taylor ∗ R. holtii Spruce ∗ R. holtii ∗ R. holtii ∗ R. holtii ∗ R. jonesii Bouman, Dirkse & Yamada ∗ R. lindenbergiana Gottsche ex C. Hartm. ∗ R. lindenbergiana ∗ R. lindenbergiana ∗ R. lindenbergiana ∗ R. lindenbergiana ∗ R. lindenbergiana ∗ R. lindenbergiana ∗ R. lindenbergiana ∗ R. nudicaulis Steph. ∗ R. nudicaulis ∗ R. nudicaulis ∗ R. nudicaulis ∗ R. nudicaulis ∗ R. wichurae Steph. ∗ R. wichurae ∗ R. wichurae ∗ R. wichurae ∗ R. wichurae ∗ R. wichurae ∗ R. wichurae ∗ R. wichurae ∗ R. wichurae ∗ R. wichurae ∗ R. wichurae Outgroup Leptoscyphus cleefii Fulford – a1 Peru Madeira 282919 (B) 232860 (LISU), 3276 (MADJ) – M07 GU191807 GU191821 GU189044 GU189055 a2 a3 a4 a5 a6 a7 – ca1 ca2 ca3 ca4 ca5 ca6 ca7 ca8 ca9 ca10 ca11 ca12 ca13 co1 co2 co3 – – – h1 h2 h3 h4 j1 l1 l2 l3 l4 l5 l6 l7 l8 n1 n2 n3 n4 n5 w1 w2 w3 w4 w5 w6 w7 w8 w9 w10 w11 Azores (Flores) Madeira Madeira Azores (Pico) Azores (Pico) Azores (Pico) S. Tomé e Prı́ncipe Madeira Madeira Madeira Azores (Flores) Madeira Madeira Azores (Pico) Azores (Pico) Azores (Pico) Azores (Pico) Azores (Flores) Azores (Flores) Azores (Flores) Switzerland Switzerland Germany Peru New Zealand Peru Portugal mainland Madeira Madeira Portugal mainland Madeira Portugal mainland Portugal mainland Madeira Madeira Madeira Madeira Portugal mainland Portugal mainland Madeira Madeira Madeira Madeira Madeira Madeira Madeira (Porto Santo) Azores (Flores) Azores (Flores) Madeira Azores (Graciosa) Azores (Terceira) Azores (Terceira) Azores (Pico) Azores (S. Jorge) Azores (S. Jorge) 232879 (LISU) 232855 (LISU), 3084 (MADJ) 232859 (LISU), 3085 (MADJ) 232878 (LISU) 232876 (LISU) 232877 (LISU) 233167 (LISU) 232858 (LISU) 233270 (LISU) 233269 (LISU) 232882 (LISU) 232856 (LISU), 3083 (MADJ) 232857 (LISU) 232887 (LISU) 232886 (LISU) 232884 (LISU) 232885 (LISU) 232880 (LISU) 232881 (LISU) 232883 (LISU) 232898 (LISU) 232899 (LISU) Frey 2–883 (herb. W. Frey) 256171 (B) Frahm 28–8 (herb. W. Frey) 256159 (B) 232895 (LISU) 232861 (LISU), 8450 (MADJ) 232862 (LISU) 232894 (LISU) 232863 (LISU), 3277 (MADJ) 232896 (LISU) 233267 (LISU) 232867 (LISU), 3268 (MADJ) 232864 (LISU) 232866 (LISU), 3086 (MADJ) 232865 (LISU), 3087 (MADJ) 232897 (LISU) 152346 (LISU) 232872 (LISU), 3280 (MADJ) 232871 (LISU), 3275 (MADJ) 232870 (LISU), 3081 (MADJ) 232869 (LISU), 3082 (MADJ) 232868 (LISU) 232874 (LISU), 3278 (MADJ) 232875 (LISU), 3279 (MADJ) 232893 (LISU) 232888 (LISU) 232873 (LISU), 3088 (MADJ) 232889 (LISU) 232890 (LISU) 232891 (LISU) 232892 (LISU) 232901 (LISU) 232900 (LISU) AF08 M06a M06b AP07a AP07c AP07b – M08 – – AF08c M06a M06b AP07a AP07b AP07c AP07d AF08a AF08b AF08d SW08a SW08b – – – – ML08b M06a M06b ML08a M07 ML08 – M06c M05 M06a M06b ML99 ML06 M08 M07 M06a M06b M06c M07 PS08 AF08 AF99a M06 AG99b AT07a AT07b AP07 ASJ08a ASJ08b GU191822 – – – – – GU191805 GU191816 GU191817 GU191819 GU191818 – – – – – – – – – GU191827 GU191826 GU191825 GU191808 GU191811 GU191806 – – – – GU191820 GU191823 GU191824 – – – – – – GU191810 GU191809 – – – GU191813 GU191812 GU191814 GU191815 – – – – – – – GU189056 – – – – – GU189042 GU189051 GU189052 GU189040 GU189053 – – – – – – – – – GU189059 GU189058 AY007636 GU189045 AY007635 GU189043 GU189060 – – – GU189054 GU189057 GU189041 – – – – – – GU189047 GU189046 – – – GU189049 GU189048 GU189050 GU189039 – – – – – – – – see Vanderpoorten & Long (2006) Germany see Vanderpoorten & Long (2006) Stech B950821.1 (L) – EU661835 DQ176709 – GU191804 AY007634 Ptilidium ciliare (L.) Hampe – Acc. no. trnL 260 M. Stech et al. analyses of molecular, phytochemical and morphological characters are summarized in Table 1. Detailed information on further morphologically studied material is available from the authors. DNA extraction, PCR and sequencing Downloaded By: [Stech, Michael] At: 06:32 24 June 2010 Plant material was thoroughly cleaned with distilled water and ultrasonic treatment. DNA extraction and PCR amplification of trnLUAA intron with primers CM and DM was performed as described in Frey et al. (1999), amplification of trnS GGA -rps4 spacer and rps4 gene with primers trnS-F and rps5′ as described in Hernández-Maqueda et al. (2008). R PCR products were cleaned with the SureClean solution  R (Bioline) or the Wizard DNA Clean-up kit (Promega). Manual or automated sequencing was performed as described in Frey et al. (1999) or by Macrogen Inc., South Korea (www.macrogen.com), respectively. GenBank accession numbers are given in Table 1. Alignment, sequence analysis and phylogenetic reconstructions Based on the criteria established in Kelchner (2000) and Quandt & Stech (2005), DNA sequences were manually R v0.995 (Müller et al., 2006). Sealigned using PhyDE lection of outgroup representatives was based on sequence availability, alignability of the included non-coding regions, and recent molecular phylogenetic reconstructions of liverworts. Ptilidium ciliare (L.) Hampe (Ptilidiaceae) was used as close outgroup based on the sister-group relationship of Ptilidium to the remaining Porellanae (e.g. He-Nygrén et al., 2006; see also Ahonen, 2004), and Leptoscyphus cleefii Fulford (Lophocoleaceae) as more remote outgroup from Jungermannianae, the sister-group of Porellanae (e.g. Forrest & Crandall-Stotler, 2004; He-Nygrén et al., 2006; Heinrichs et al., 2007). Sequences of the latter species were taken from Vanderpoorten & Long (2006). Phylogenetic reconstructions according to the maximum parsimony (MP) and maximum likelihood (ML) optimality criteria were performed using PAUP 4.0b10 (Swofford, 2002). Heuristic searches under parsimony were implemented using random sequence addition with 1000 replicates and employing the default settings otherwise. Gaps were treated as missing data. Heuristic bootstrap searches under parsimony were performed with 10 000 replicates and 100 random addition cycles per bootstrap replicate with the same options in effect. Maximum likelihood analyses were executed assuming a general time reversible (GTR) model and a rate variation among sites following a gamma distribution. GTR+Ŵ+I was chosen as the model that best fits the combined data according to the AIC criterion as evaluated by MrModeltest v2.3 (Nylander, 2004) employing MrMTgui (Nuin, 2005). The proposed settings by MrModeltest were exe- cuted in PAUP: Basefreq = (0.3765 0.1238 0.1321), Nst = 6, Rmat = (0.8027 3.1192 0.2388 0.9415 3.3011), Shape = 3.8729, Pinvar = 0.4675. Model testing for the three partitions (trnS-rps4 spacer, rps4 gene, trnL intron) separately indicated either GTR+Ŵ or GTR+I as best fitting models. Therefore, a Bayesian analysis (see below) with three separate partitions was carried out to test for incongruence with the analysis based on GTR+Ŵ+I as indicated by MrModeltest for the combined data set. Heuristic likelihood bootstrap searches were performed with 100 replicates. For further measurement of support, posterior probabilities were calculated using MrBayes v3.1 (Huelsenbeck & Ronquist, 2001). As in the maximum likelihood analysis, the GTR model of nucleotide substitution was employed, assuming site-specific rate categories following a gamma distribution. The a priori probabilities supplied were those specified in the default settings of the program. Posterior probability (PP) distributions of trees were created using the Metropolis-coupled Markov chain Monte Carlo (MCMCMC) method. Four runs with four chains (106 generations each) were run simultaneously, with the temperature of the single heated chain set to 0.2. Chains were sampled every 10 generations and the respective trees written to a tree file. Calculations of the consensus tree and the posterior probability of clades were performed based upon the trees sampled after the chains converged (we used 25% as default). Analysis of phytochemical characters The volatile oil compounds were isolated by distillationextraction and analysed by gas chromatography (GC) and gas chromatography-mass spectrometry (GC-MS) as reported by Figueiredo et al. (2009). All 48 specimens analysed in Figueiredo et al. (2009), plus R. boryana (F. Weber) Nees as outgroup representative, were included in the present study. In total, 101 chemical compounds were considered, including those 83 reported by Figueiredo et al. (2009), seven further distinct compounds identified based on mass spectra only (putative sesquisabinene A, isobazzanene, epi-β-santalene, γ -gurjunene, taynudol, cyclocolorenone and an unidentified component R), and 11 only present in the outgroup (α-terpinene, p-cymene, γ terpinene, 2,5-dimethyl styrene, terpinolene, p-cymen-8ol, thymol, carvacrol, α-gurjunene, and cis-calamenene). A data matrix, which is available on request, was generated based on the presence or absence of the respective compounds in the 49 samples (0 = absent, 1 = present). This matrix was analysed with PAUP 4.0b10 according to the maximum parsimony principle and subjected to a Bayesian analysis with MrBayes v3.1 (data type = restriction site for character states 0/1, coding = variable). Heuristic bootstrap searches under parsimony were performed with 1000 replicates and 10 random addition cycles per bootstrap replicate. Characterization of the liverwort Radula in Portugal Downloaded By: [Stech, Michael] At: 06:32 24 June 2010 Analysis of morphological-anatomical characters For each of 51 Radula specimens (plus R. boryana as outgroup representative) for which phytochemical data and/or DNA sequence data was generated, three stems from the main axis were dissected off and each stem analysed at 5–10 mm from the apex. Both quantitative and qualitative (discrete) characters were recorded, but in accordance with the literature a preliminary analysis indicated that the qualitative characters were more useful as diagnostic characters for distinguishing the Radula species than the quantitative ones. Therefore, we used 13 qualitative characters (Table 2) for phylogenetic analysis of the respective 52 specimens. Of these, 11 characters were coded as binary characters (0/1), whereas two characters were coded as multistate characters (0, 1 and 2). The respective character states for ‘perianth shape’ and ‘sexual condition’ could only be recorded from fertile specimens; in sterile specimens the missing data were indicated by ‘?’. The matrix was analysed with PAUP 4.0b10 according to the maximum parsimony principle and subjected to a Bayesian analysis with MrBayes v3.1 (data type = standard for character states 0/1 and multistate characters, coding = variable). Heuristic bootstrap searches under parsimony were performed with 1000 replicates and 10 random addition cycles per bootstrap replicate. Results Sequence characterization and alignment Lengths of the trnS-rps4 spacer vary between 31 nucleotides (nt) (R. antilleana Castle) and 125 nt (R. episcia Spruce) across the whole data set and within Radula. Lengths of the trnL intron range from 303 to 372 nt across the whole data set, within Radula from 341 (R. aquilegia) to 372 nt (R. boryana). The rps4 gene is 608 nt in all sequenced samples. For R. holtii only the trnL intron could be sequenced. The combined alignment comprises 1263 positions (trnS-rps4 positions 1–215, rps4 216–823, trnL 824–1263), of which 68 were excluded from phylogenetic analysis due to incomplete sequencing. Of the remaining 1195 included characters, 333 (27.9%) are variable, and 217 (18.2%, or 65.2% of the variable positions) are parsimony-informative (35 in trnS-rps4, 115 in rps4, 67 in trnL). 261 ML bootstrap values ≥ 60% and significant (≥ 94) posterior probabilities (PP) indicated above the branches. Bayesian analyses using the GTR+Ŵ+I model for the whole data set versus individual models (GTR+Ŵ or GTR+I, respectively) for the three partitions resulted in the same topology and (almost) equal PPs. In all analyses, the basal split between the afroneotropical species Radula boryana and R. gottscheana Taylor and a clade including all other Radula species receives maximum support. Within the latter clade, two of the Portuguese species, R. holtii and R. nudicaulis, form a maximally supported clade with the neotropical species R. antilleana and R. episcia. The remaining European Radula representatives form a clade, also with maximum support, of which the sister-group relationship to R. grandis Steph. from New Zealand is not well-supported. Radula aquilegia occupies a basal position within this second European Radula clade, followed by R. jonesii. The sister-group relationships of R. complanata and R. lindenbergiana as well as of R. carringtonii and R. wichurae are maximally supported. Separate analyses of trnS-rps4 spacer, rps4 gene and trnL intron, respectively, resulted in less well-resolved (consensus) trees, but did not reveal inconsistencies between the different markers as inferred from visual comparison of the respective tree topologies (data not shown). Analysis of phytochemical characters The maximum parsimony calculation of the phytochemical matrix (101 characters, of which 85 were parsimonyinformative) yielded two trees (lengths 167, CI excluding uninformative characters 0.559, RI 0.935, RC 0.560). Both trees differ in the branching order of the basal clades that are thus resolved as a polytomy in the strict consensus tree shown in Fig. 2. The specimens of R. carringtonii, R. complanata, R. holtii, R. nudicaulis and R. wichurae each form clades with significant bootstrap support (99−100% BS) and significant PP values of 96−100 (except R. wichurae). Radula holtii is resolved at a basal position without support. The Madeiran and Azorean specimens of R. aquilegia are separated on two clades with 99−100% BS and a PP of 98 for the clade of the Azorean specimens. Concerning species relationships, only the position of R. lindenbergiana as sister to R. wichurae receives significant BS (99%) and an almost significant PP of 93. Molecular phylogenetic analysis Analysis of morphological–anatomical characters The MP analysis retained a single most parsimonious tree (length 496, CI excluding uninformative characters 0.754, RI 0.889, RC 0.730). In the ML analysis a single optimal tree with the same topology as the MP tree was found (lnL = −3861.34269). The ML tree is shown in Fig. 1, with MP and The eight Radula species referred to Portugal can be distinguished morphologically by a set of diagnostic characters (Table 3), as inferred from the study of herbarium material. Characters diagnostic for single species are the presence of caducous leaves in Radula nudicaulis, the trumpet-shaped 262 M. Stech et al. Downloaded By: [Stech, Michael] At: 06:32 24 June 2010 Table 2. Matrix of 13 morphological–anatomical characters of the 52 Radula specimens subjected to phylogenetic analysis. Characters: 1. ramification: irregularly pinnate (0), regularly pinnate (1); 2. cell walls in the medullar cells: thick (0), thin (1); 3. leaf lobe extension: lobes not covering stem in dorsal view (0), lobes covering stem in dorsal view (1); 4. overlapping of leaves: not overlapping in dorsal view (0), overlapping in dorsal view (1); 5. apical margin of lobe and lobule: unbordered (0), bordered with small hyaline cells (1); 6. lobule shape: subquadrate (0), subrectangular (1), ovate (2); 7. lobule inflation: flat (0), inflated (1); 8. lobule apex: rounded (0), bluntly acute (1); 9. keel orientation: straight (0), convex (1), concave (2); 10. gemmae: absent (0), present (1); 11. caducous leaves: absent (0), present (1); 12. perianth shape: subrectangular (0), trumpet-shaped (1); 13. sexual condition: dioecious (0), monoecious (paroecious) (1). Taxon sample 1 2 3 4 5 6 7 8 9 10 11 12 13 R. boryana R. aquilegia R. aquilegia R. aquilegia R. aquilegia R. aquilegia R. aquilegia R. aquilegia R. carringtonii R. carringtonii R. carringtonii R. carringtonii R. carringtonii R. carringtonii R. carringtonii R. carringtonii R. carringtonii R. carringtonii R. carringtonii R. carringtonii R. carringtonii R. complanata R. complanata R. holtii R. holtii R. holtii R. holtii R. jonesii R. lindenbergiana R. lindenbergiana R. lindenbergiana R. lindenbergiana R. lindenbergiana R. lindenbergiana R. lindenbergiana R. lindenbergiana R. nudicaulis R. nudicaulis R. nudicaulis R. nudicaulis R. nudicaulis R. wichurae R. wichurae R. wichurae R. wichurae R. wichurae R. wichurae R. wichurae R. wichurae R. wichurae R. wichurae R. wichurae a1 a2 a3 a4 a5 a6 a7 ca1 ca2 ca3 ca4 ca5 ca6 ca7 ca8 ca9 ca10 ca11 ca12 ca13 co1 co2 h1 h2 h3 h4 j1 l1 l2 l3 l4 l5 l6 l7 l8 n1 n2 n3 n4 n5 w1 w2 w3 w4 w5 w6 w7 w8 w9 w10 w11 1 0 0 0 1 1 1 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 1 1 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 1 1 1 1 1 1 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0 0 0 1 0 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1 1 0 0 1 1 0 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 2 1 0 1 1 1 1 1 0 0 0 1 0 1 0 0 0 0 0 0 (01) 0 0 0 0 0 0 0 1 (01) 1 1 (01) 1 1 (01) 0 0 0 0 0 2 2 (02) 2 (02) 2 (02) 2 2 (02) (02) 0 1 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 (01) (01) (01) (01) (01) (01) (01) 0 0 (01) (01) (01) 0 0 0 0 0 (01) 0 (01) 1 1 0 0 0 0 1 1 (01) 1 1 1 (01) 1 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 1 1 1 1 1 0 0 (01) 0 0 1 0 0 0 0 0 0 0 0 (01) 0 (01) (01) (01) (02) (01) 2 0 (01) 0 1 1 (01) (01) (01) (01) (01) (01) (02) (02) (02) (02) (02) (02) (02) (02) (02) (02) (02) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 ? ? ? ? ? ? 0 ? ? ? ? ? ? 0 ? ? ? ? 0 ? ? 0 0 1 ? 1 ? 0 ? ? ? 0 0 0 0 0 ? ? ? ? ? ? ? ? 0 0 0 ? ? ? ? ? ? ? ? ? ? ? 0 ? ? ? ? ? ? 0 ? ? ? ? 0 ? ? 1 1 1 ? 1 ? 1 ? ? ? 0 0 0 0 0 ? ? ? ? ? ? ? ? 0 0 0 ? ? ? ? ? 1 0 0 1 1 1 1 1 1 1 1 1 1 1 Characterization of the liverwort Radula in Portugal 263 R. wichurae w1 Madeira 95/97/100 R. wichurae w2 Madeira (Porto Santo) R. wichurae w3 Azores (Flores) 100 R. wichurae w4 Azores (Flores) R. carringtonii ca1 Madeira R. carringtonii ca2 Madeira 95/93 /100 84/95/100 R. carringtonii ca3 Madeira R. carringtonii ca4 Azores (Flores) R. complanata co1 Switzerland 87/87 /100 R. complanata co2 Switzerland Downloaded By: [Stech, Michael] At: 06:32 24 June 2010 79/86/100 100 R. complanata co3 Germany R. lindenbergiana l1 Portugal 63/62 R. lindenbergiana l2 Portugal /-- 100 R. jonesii j1 Madeira 82/--/-- R. aquilegia a1 Madeira 86/79/96 R. aquilegia a2 Azores (Flores) R. grandis New Zealand 100 75/91/-66/60/-- R. episcia Peru R. holtii h1 Portugal R. nudicaulis n1 Madeira 100 100 R. nudicaulis n2 Madeira 97/98/100 R. antilleana Peru 100 R. boryana São Tomé R. gottscheana Peru Leptoscyphus cleefii Ptilidium ciliare 0.1 Fig. 1. Maximum likelihood tree based on plastid trnS-rps4 spacer, rps4 gene, and trnL intron sequences of 24 Radula specimens as well as Leptoscyphus cleefii and Ptilidium ciliare as outgroup representatives. Bootstrap support values ≥ 60% from respective maximum parsimony and maximum likelihood analyses as well as significant (≥ 94) Bayesian posterior probabilities are given at the branches (maximum support in all analyses is indicated by ‘100’). perianth in R. holtii and the lobe and lobule margins being bordered with small hyaline cells in R. wichurae. Discoid marginal gemmae on lobes and lobules are only found in R. complanata and R. lindenbergiana, differentiation be- tween thick-walled cortical and thin-walled medullary cells in the stem occurs in R. holtii and R. nudicaulis. However, species delimitations are not always sufficiently clear when based solely on morphological–anatomical characters. For 264 M. Stech et al. 100/-- 99/(93) 100/96 Downloaded By: [Stech, Michael] At: 06:32 24 June 2010 100/100 99/-- 100/98 100/100 100/100 100/96 R. wichurae w1 Madeira R. wichurae w2 Madeira (Porto Santo) R. wichurae w3 Azores (Flores) R. wichurae w4 Azores (Flores) R. wichurae w5 Madeira R. wichurae w6 Azores (Graciosa) R. wichurae w7 Azores (Terceira) R. wichurae w8 Azores (Terceira) R. wichurae w9 Azores (Pico) R. wichurae w10 Azores (São Jorge) R. wichurae w11 Azores (São Jorge) R. lindenbergiana l1 Portugal R. lindenbergiana l3 Madeira R. lindenbergiana l4 Madeira R. lindenbergiana l5 Madeira R. lindenbergiana l6 Madeira R. lindenbergiana l7 Portugal R. lindenbergiana l8 Portugal R. carringtonii ca1 Madeira R. carringtonii ca4 Azores (Flores) R. carringtonii ca5 Madeira R. carringtonii ca6 Madeira R. carringtonii ca7 Azores (Pico) R. carringtonii ca8 Azores (Pico) R. carringtonii ca9 Azores (Pico) R. carringtonii ca10 Azores (Pico) R. carringtonii ca11 Azores (Flores) R. carringtonii ca12 Azores (Flores) R. carringtonii ca13 Azores (Flores) R. aquilegia a1 Madeira R. aquilegia a3 Madeira R. aquilegia a4 Madeira R. aquilegia a2 Azores (Flores) R. aquilegia a5 Azores (Pico) R. aquilegia a6 Azores (Pico) R. aquilegia a7 Azores (Pico) R. nudicaulis n1 Madeira R. nudicaulis n2 Madeira R. nudicaulis n3 Madeira R. nudicaulis n4 Madeira R. nudicaulis n5 Madeira R. complanata co1 Switzerland R. complanata co2 Switzerland R. jonesii j1 Madeira R. holtii h1 Portugal R. holtii h4 Portugal R. holtii h2 Madeira R. holtii h3 Madeira R. boryana São Tomé Fig. 2. Strict consensus of two most parsimonious trees resulting from maximum parsimony analysis of volatile oil characters of 48 samples from nine Radula species, using R. boryana as outgroup representative. Bootstrap support values ≥ 60% and significant (≥ 94) Bayesian posterior probabilities are given at the branches. example, Radula aquilegia generally differs from R. carringtonii by inflated subrectangular rather than flat quadrate lobules. Some populations of R. aquilegia, however, show subquadrate and only slightly inflated lobules that may resemble those of R. carringtonii. Moreover, lobule morphology varies between populations of R. aquilegia collected in distinct islands of the Azores archipelago. The populations collected in the islands of the central group (Pico I.) present typical inflated subrectangular lobules, which contrasts with the populations from the western group (Flores I.) whose lobules are less inflated and subquadrate. The R. aquilegia populations from Madeira Island are less variable and usually have scarcely inflated subquadrate lobules, which sometimes make their distinction from R. carringtonii difficult. In addition to the morphological differences, such plants from Flores I. are more fragile than the typical ones and always found in shady, humid rock crevices and fissures and walls in the forest. Radula complanata and R. lindenbergiana are highly similar in all vegetative characters and can only be distinguished in the presence of sexual structures, since R. complanata is paroecious and R. lindenbergiana dioecious. Radula jonesii could be confused with R. holtii when sterile, except for the absence of thin-walled medullary cells of the stem. In the MP analysis of the 13 diagnostic morphological– anatomical characters 8000 most parsimonious trees were recovered (length 38, CI 0.395, RI 0.807, RC 0.318). In the strict consensus tree (Fig. 3) the samples of R. nudicaulis, all but one R. aquilegia, and one R. carringtonii form a basal polytomy together with the outgroup R. boryana. The specimens of R. holtii and R. wichurae each form one clade. Their relationships with the specimens of R. complanata, R. jonesii, R. lindenbergiana, all but one R. carringtonii, and one R. aquilegia are unresolved. No Characterization of the liverwort Radula in Portugal 265 Table 3. Morphological–anatomical diagnostic characters of the eight Radula species occurring in Portugal. Cauloid (stem) Leaf lobes Leaf lobules R. aquilegia Irregularly to regularly pinnately branched; medullary cells thick-walled R. carringtonii Irregularly to (rarely) regularly pinnately branched; medullary cells thick-walled Irregularly pinnately branched; medullary cells thick-walled Imbricate or rarely contiguous; generally covering the stem in dorsal view; margins unbordered Imbricate or rarely contiguous; generally covering the stem; margins unbordered Imbricate; generally covering the stem; margins unbordered Subrectangular or rarely subquadrate; inflated; apex rounded; keel convex; margins unbordered Subquadrate or rarely subrectangular; flat; apex rounded; keel straight; margins unbordered Subquadrate to subrectangular; flat; apex rounded; keel straight to slightly convex; margins unbordered Subquadrate; flat; apex rounded; keel straight to convex; margins unbordered Subquadrate; flat; apex from strictly rounded to bluntly acute; keel straight to concave; margins unbordered Subrectangular to subquadrate; flat; apex rounded; keel straight to slightly convex, rarely concave; margins unbordered Subquadrate; inflated; apex bluntly acute to rounded; keel straight to convex; margins unbordered Ovate, sometimes almost quadrate; flat; apex rounded; keel straight to slightly concave; margins bordered with small hyaline cells R. complanata Irregularly pinnately branched; medullary cells thin-walled and hyaline Regularly pinnately branched; medullary cells thick-walled Contiguous; not covering the stem; margins unbordered R. lindenbergiana Irregularly to regularly pinnately branched; medullary cells thick-walled Imbricate to contiguous; generally covering the stem; margins unbordered R. nudicaulis Irregularly to regularly pinnately branched; medullary cells thin-walled and hyaline Irregularly to regularly pinnately branched; medullary cells thick-walled Imbricate to contiguous; generally covering the stem; margins unbordered Imbricate; generally covering the stem; margins bordered with small hyaline cells R. holtii Downloaded By: [Stech, Michael] At: 06:32 24 June 2010 R. jonesii R. wichurae Imbricate to contiguous; not covering the stem; margins unbordered significant bootstrap support and no posterior probabilities in the respective Bayesian analysis were achieved. Discussion All three data sets employed in the present study (DNA sequence data, volatile oil composition, and morphological–anatomical characters) contribute to inferring species circumscriptions and phylogenetic relationships of the Radula species referred to Portugal, albeit to different degrees. The molecular phylogenetic reconstruction (Fig. 1) provides the highest resolution in terms of species relationships, although this may partly be due to the different taxon sampling with further Radula species included from other geographic regions. Radula aquilegia, R. carringtonii, R. holtii, R. jonesii, R. nudicaulis, R. wichurae and the R. complanata/R. lindenbergiana complex can be clearly distinguished from each other based on the molecular data. The phytochemical data (Fig. Sexual reproduction Asexual reproduction Dioecious; perianth subrectangular Gemmae and caducous leaves absent Dioecious; perianth subrectangular Gemmae and caducous leaves absent Monoecious (paroecious); perianth subrectangular Discoid marginal gemmae present at lobe and lobule margins; caducous leaves absent Gemmae and caducous leaves absent Monoecious (paroecious); perianth trumpet-shaped Monoecious (paroecious); perianth subrectangular Gemmae and caducous leaves absent Dioecious; perianth subrectangular Discoid marginal gemmae present at lobe and lobule margins; caducous leaves absent Reproductive structures not observed Caducous leaves present, especially in adult plants; gemmae absent Gemmae and caducous leaves absent Dioecious; perianth subrectangular 2) support these species circumscriptions except for R. aquilegia, of which the Azorean and Madeiran samples form separate clades (see discussion below). Support for the clades of R. aquilegia, R. carringtonii, R. complanata and R. wichurae is comparable to that obtained in the molecular analyses, and higher for R. lindenbergiana. The latter is clearly separated from R. complanata when the phytochemical data are analysed in a phylogenetic framework, in contrast to the similarity analysis performed in Figueiredo et al. (2009). The morphological–anatomical characters are less informative in the phylogenetic approach (Fig. 3), with only the specimens of R. holtii and R. wichurae being resolved as distinct clades, although several characters are diagnostic for single species or species pairs (cf. results). The low resolution and absence of statistical support may result from the low number of characters, but also from a considerably higher homoplasy and lower synapomorphy content in the morphological data set, as inferred from the low rescaled consistency (RC) value of the resulting trees Downloaded By: [Stech, Michael] At: 06:32 24 June 2010 266 M. Stech et al. R. wichurae w1 Madeira R. wichurae w2 Madeira (Porto Santo) R. wichurae w3 Azores (Flores) R. wichurae w4 Azores (Flores) R. wichurae w5 Madeira R. wichurae w6 Azores (Graciosa) R. wichurae w7 Azores (Terceira) R. wichurae w8 Azores (Terceira) R. wichurae w9 Azores (Pico) R. wichurae w10 Azores (São Jorge) R. wichurae w11 Azores (São Jorge) R. holtii h1 Portugal R. holtii h4 Portugal R. holtii h2 Madeira R. holtii h3 Madeira R. lindenbergiana l1 Portugal R. lindenbergiana l3 Madeira R. lindenbergiana l2 Portugal R. lindenbergiana l4 Madeira R. lindenbergiana l5 Madeira R. lindenbergiana l6 Madeira R. lindenbergiana l7 Portugal R. lindenbergiana l8 Portugal R. complanata co1 Switzerland R. complanata co2 Switzerland R. jonesii j1 Madeira R. carringtonii ca1 Madeira R. carringtonii ca2 Madeira R. carringtonii ca3 Madeira R. carringtonii ca4 Azores (Flores) R. carringtonii ca5 Madeira R. carringtonii ca7 Azores (Pico) R. carringtonii ca8 Azores (Pico) R. carringtonii ca9 Azores (Pico) R. carringtonii ca10 Azores (Pico) R. carringtonii ca11 Azores (Flores) R. carringtonii ca12 Azores (Flores) R. carringtonii ca13 Azores (Flores) R. aquilegia a2 Azores (Flores) R. nudicaulis n3 Madeira R. nudicaulis n4 Madeira R. nudicaulis n5 Madeira R. nudicaulis n1 Madeira R. nudicaulis n2 Madeira R. carringtonii ca6 Madeira R. aquilegia a1 Madeira R. aquilegia a3 Madeira R. aquilegia a4 Madeira R. aquilegia a5 Azores (Pico) R. aquilegia a6 Azores (Pico) R. aquilegia a7 Azores (Pico) R. boryana São Tomé Fig. 3. Strict consensus of 8000 most parsimonious trees resulting from maximum parsimony analysis of morphological-anatomical characters of 52 samples from nine Radula species, using R. boryana as outgroup representative. (0.318 versus 0.560 in the phytochemical and 0.730 in the molecular analysis). In fact, only three out of the 13 characters show individual RC values of 1.00 (characters 5, 12 and 13, cf. Table 2), while the RC values of the remaining characters range between 0.08–0.59. Despite the low resolution of the phylogenetic reconstruction based on morphology, the present study shows that all eight Radula taxa referred to Portugal can be clearly distinguished from each other, either by molecular data alone or by combination of molecular, phytochemical, and morphological–anatomical characters. Radula carringtonii is clearly separated from the morphologically similar R. aquilegia according to the molecular and phytochemical data, and is resolved with maximum support as sister to the Macaronesian endemic R. wichurae in the molecular tree (Fig. 1). Although phylogenetic relationships of the second Macaronesian endemic R. jonesii are less well-supported (Fig. 1), this species can be clearly distinguished from the other Radula species, especially from the most similar R. holtii, by DNA sequences, volatile oil composition (Fig. 2; Figueiredo et al., 2009), and the different stem anatomy. The more complicated cases concern the incongruence of the three data sets in R. aquilegia and the differentiation between R. complanata and R. lindenbergiana. The separation of the Azorean and Madeiran samples of R. aquilegia in the volatile oil analyses (Fig. 2; Figueiredo et al., 2009) may indicate that the morphological species R. aquilegia comprises different cryptic taxa, which may have originated from independent evolution on both archipelagos. Our morphological studies revealed a larger intraspecific variability in R. aquilegia than previously known, which, however, shows a different geographic pattern than the phytochemical variation. The Azorean flora comprises both the typical R. aquilegia and a morphological–ecological form closer to R. carringtonii, the latter occurring in the western group of Azores islands but also in Madeira. As the molecular data clearly support the morphological species R. aquilegia as a monophyletic group, the observed morphological and phytochemical variation may thus be influenced by different ecological conditions on the archipelagos, but does not justify recognizing formal (intra-)specific taxa. Radula complanata and R. lindenbergiana have always been considered closely related, or even conspecific (Schuster, 1980a). Several authors emphasize that separation is only possible in the presence of reproductive structures (Schuster, 1980a; Paton, 1999; Luı́s et al., 2005; Schumacker & Váňa, 2005). The molecular analyses support a close relationship of both species (Fig. 1), but statistical support is rather low especially for the clade of R. lindenbergiana, since the sequences of the included specimens of both taxa differ by only three substitutions (two in rps4 gene, one in trnL intron). In addition, different types of loops of a putative hairpin secondary structure in the most variable stem-loop region P8 of the trnL intron (cf. Quandt & Stech, 2005), an A-rich type in R. complanata and a T-rich type in R. lindenbergiana (Fig. 4), may serve as a molecular character. Both loop types originate from inversion of the loop sequence. However, such inversions in non-coding plastid DNA regions of bryophytes have been shown to be homoplastic at different taxonomic levels (e.g. Quandt et al., 2003; Stech, 2004; Quandt & Stech, 2005) and can also occur within one species (own unpublished results). In contrast, partial sequences of the more variable internal transcribed spacers (ITS) of the nuclear ribosomal DNA clearly allow distinguishing R. complanata and R. lindenbergiana (own unpublished results). Unfortunately, tree construction was not possible as no reliable ITS sequences of the other Radula species could be obtained, mainly due to fungal contamination. Taking the molecular data together with the difference in volatile oil composition and sexual condition, we consider a separation of R. complanata and R. lindenbergiana at species level reasonable. However, Characterization of the liverwort Radula in Portugal A A T T T G A C G A A T A A A A A A C T G C T T A T T T T T G A C G A A T T T A A A G C T G C T T A Downloaded By: [Stech, Michael] At: 06:32 24 June 2010 Fig. 4. Putative hairpin secondary structure in stem-loop region P8 of the trnL intron. A. A-rich type of loop in Radula complanata (G = −10.8). B. T-rich type of loop (or U in the respective RNA) in R. lindenbergiana (G = −10.6). an extended sampling, including Portuguese material of R. complanata, and analysis of further molecular markers is necessary to arrive at a final conclusion. Despite the limited taxon sampling across the genus Radula, the molecular phylogenetic reconstructions allow first conclusions concerning intrageneric and biogeographic relationships of the species under study. The basal split between R. boryana/R. gottscheana and the remaining species corresponds to the distinction of subgenera Cladoradula and Radula, respectively, by Schuster (1980a). The inclusion of R. aquilegia and R. carringtonii in subg. Cladoradula by Castle (1969) is rejected. Within subgenus Radula, relationships are difficult to analyse as not all European species have been decidedly assigned to intrasubgeneric taxa. At this taxonomic level, at least the hitherto recognised taxa of Castle (1969) may not represent monophyletic groups, as indicated by the positions of R. wichurae (section Marginata), close to R. complanata/R. lindenbergiana and R. holtii (section Radula), and of R. nudicaulis (also section Radula) being separated from this clade. The Radula species narrowly distributed in Macaronesia and Atlantic Europe probably have two different origins. For Radula holtii and R. nudicaulis connections with the Neotropics are indicated by the molecular data, which corresponds to the pronounced neotropical influence on the Macaronesian bryoflora recently indicated by a number of molecular studies of mosses (e.g. Stech et al., 2007) and especially liverworts (e.g. Feldberg & Heinrichs, 2005; SimSim et al., 2005; Vanderpoorten & Long, 2006). The other species, among them the Macaronesian endemics R. jonesii and R. wichurae, are closely related with, and may be derived from, the R. complanata/R. lindenbergiana complex, which is widely distributed in the northern hemisphere. Relationships with northern-hemisphere species have also been revealed in previous studies of Macaronesian mosses (e.g. Stech et al., 2007, 2008) and liverworts (e.g. Sim-Sim et al., 2005; Hentschel et al., 2007). Biogeographic patterns of the European Radula species thus seem to be rather com- 267 plex, similar to other diverse and species-rich genera such as Plagiochila (Sim-Sim et al., 2005) and Campylopus (Stech et al., 2007). Acknowledgements This study was partially funded by Fundação para a Ciência e a Tecnologia (Lisbon) under research contract POCI/AGR/57487/2004. Sincere thanks are due to B. Giesicke (Berlin) and M.C.M. Eurlings (Leiden) for technical assistance. References AHONEN, I. 2004. Molecular phylogeny of the liverwort order Porellales (Marchantiophyta, Jungermanniopsida). In: GOFFINET, B., HOLLOWELL, V. & MAGILL, R., Eds., Molecular systematics of bryophytes. Monographs in Systematic Botany from the Missouri Botanical Garden 98, 168–188. BOUMAN, A.C. & DIRKSE, G.M. 1990. The genus Radula in Macaronesia. Lindbergia 16, 119–127. CASTLE, H. 1937. A revision of the genus Radula. Introduction and Part I. Subgenus Cladoradula. Annales Bryologici 9, 13–56. CASTLE, H. 1969. Radula (L.) Dumortier – A synopsis of the taxonomic revision of the genus. Revue Bryologie et Lichénologie 36, 5–44. FELDBERG, K. & HEINRICHS, J. 2005. On the identity of Herbertus borealis (Jungermanniopsida: Herbertaceae), with notes on the possible origin of H. sendtneri. Journal of Bryology 27, 343–350. FIGUEIREDO, A.C., SIM-SIM, M., BARROSO, J.G., PEDRO, L.G., ESQUÍVEL, M.G., FONTINHA, S., LUÍS, L., MARTINS, S. & STECH, M. 2009. Liverwort Radula species from Portugal: chemotaxonomical evaluation of volatiles composition. Flavour and Fragrance Journal 24, 316–325. FORREST, L.L. & CRANDALL-STOTLER, B.J. 2004. A phylogeny of the simple thalloid liverworts (Jungermanniopsida, Metzgeriidae) as inferred from five chloroplast genes. In: GOFFINET, B., HOLLOWELL, V. & MAGILL, R., Eds., Molecular Systematics of Bryophytes. Monographs in Systematic Botany from the Missouri Botanical Garden 98, 119–140. FREY, W., FRAHM, J.-P., FISCHER, E. & LOBIN, W. 2006. The Liverworts, Mosses and Ferns of Europe (English edition revised and edited by BLOCKEEL, T.L.). Harley Books, Colchester. FREY, W. & STECH, M. 2009. Marchantiophyta, Bryophyta, Anthocerotophyta. In: FREY, W., Ed., Syllabus of Plant Families. A. Engler′ s Syllabus der Pflanzenfamilien, 13th edition, Part 3. Bryophytes and Seedless Vascular Plants. Gebr. Borntraeger Verlagsbuchhandlung, Stuttgart, pp. 13–263. FREY, W., STECH, M. & MEISSNER, K. 1999. Chloroplast DNArelationship in palaeoaustral Lopidium concinnum (Hypopterygiaceae, Musci). An example of stenoevolution in mosses. Studies in austral temperate rain forest bryophytes 2. Plant Systematics and Evolution 218, 67–75. GROLLE, R. 1970. Radula casteli sp. nov. und Anmerkungen zur Gattung Radula. The Bryologist 73, 662–668. HEINRICHS, J., HENTSCHEL, J., WILSON, R., FELDBERG, K. & GRADSTEIN, S.R. 2007. Evolution of leafy liverworts (Jungermanniidae, Marchantiophyta): estimating divergence times from chloroplast DNA sequences using penalized likelihood with integrated fossil evidence. Taxon 56, 31–44. HENTSCHEL, J., ZHU, R.-L., LONG, D.G., DAVISON, P.G., SCHNEIDER, H., GRADSTEIN, S.R. & HEINRICHS, J. 2007. A phylogeny of Porella (Porellaceae, Jungermanniopsida) based on nuclear Downloaded By: [Stech, Michael] At: 06:32 24 June 2010 268 M. Stech et al. and chloroplast DNA sequences. Molecular Phylogenetics and Evolution 45, 693–705. HE-NYGRÉN, X., JUSLÉN, A., AHONEN, I., GLENNY, D. & PIIPPO, S. 2006. Illuminating the evolutionary history of liverworts (Marchantiophyta)—towards a natural classification. Cladistics 22, 1–31. HERNÁNDEZ-MAQUEDA, R., QUANDT, D., WERNER, O. & MUÑOZ, J. 2008. Phylogeny and classification of the Grimmiaceae/Ptychomitriaceae complex (Bryophyta) inferred from cpDNA. Molecular Phylogenetics and Evolution 46, 863–877. HUELSENBECK, J.P. & RONQUIST, F. 2001. MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics 17, 754–755. IUCN. 2008. Bryophyte Specialist Group 2000. Radula jonesii. In: 2008 IUCN Red List of Threatened Species. http://www.iucnredlist.org (accessed 5 May 2009). JONES, E.W. 1977. African Hepatics. XXX. The genus Radula Dumortier. Journal of Bryology 9, 461–504. KELCHNER, S.A. 2000. The evolution of non-coding chloroplast DNA and its application in plant systematics. Annals of the Missouri Botanical Garden 87, 482–498. LUÍS, L., DRAPER, D. & SIM-SIM, M. 2005. The distribution of the genus Radula in mainland Portugal and the Madeira Archipelago. Lindbergia 30, 3–10. MÜLLER, K., MÜLLER, J., NEINHUIS, C. & QUANDT, D. 2006. PhyDE – Phylogenetic Data Editor, v0.995. Program distributed by the authors. http://www.phyde.de (accessed 23 February 2010). NUIN, P.A.S. 2005. MTgui—a simple interface to ModelTest. Program distributed by the author. University of Toronto, http://www.genedrift.org/mtgui.php (accessed 23 February 2010). NYLANDER, J.A.A. 2004. MrModeltest v2. Program distributed by the authors, Evolutionary Biology Centre, Uppsala University. PATON, J.A. 1999. The Liverwort Flora of the British Isles. Harley Books, Colchester. 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 and secondary structure of the chloroplast trnL intron in bryophytes. Molecular Phylogenetics and Evolution 36, 429–443. RENNER, M.A.M. & BRAGGINS, J.E. 2004. The sterile gametophyte as a source of informative characters in the genus Radula (Radulaceae: Hepaticae). Nova Hedwigia 78, 243–268. RENNER, M.A.M. & BRAGGINS, J.E. 2005. Systematically relevant characters of the Radula sporophyte (Radulaceae: Hepaticae). Nova Hedwigia 81, 271–300. SCHUMACKER, R. & VÁŇA, J. 2005. Identification Keys to the Liverworts and Hornworts of Europe and Macaronesia (Distribution and Status). 2nd edition. Sorus, Poznań. SCHUSTER, R.M. 1980a. Phylogenetic studies on Jungermanniidae II. Radulineae (Part I). Nova Hedwigia 32, 637–693. SCHUSTER, R.M. 1980b. The Hepaticae and Anthocerotae of North America, East of the Hundredth Meridian, Vol. 3. Columbia University Press, New York. SIM-SIM, M., ESQUÍVEL, M.G., FONTINHA, S. & STECH, M. 2005. The genus Plagiochila (Plagiochilaceae, Hepaticophytina) in Madeira Island—Molecular relationships, ecology, and biogeographic affinities. Nova Hedwigia 81, 449– 462. SIM-SIM, M., LUÍS, L., GARCIA, C., FONTINHA, S., LOBO, C., MARTINS, S. & STECH, M. 2008. New data on the status of bryophytes from Madeira Island with reference to new species. Journal of Bryology 30, 226–228. SO, M.L. 2005a. A synopsis of Radula (Radulaceae, Marchantiophyta) in New Zealand and Tasmania. Journal of the Hattori Botanical Laboratory 98, 149–174. SO, M.L. 2005b. Radula (Radulaceae, Marchantiophyta) in Hawaii. Journal of the Hattori Botanical Laboratory 98, 175–191. SO, M.L. 2006. Radula (Radulaceae, Marchantiophyta) in the South Pacific. Journal of the Hattori Botanical Laboratory 99, 207–232. SÖDERSTRÖM, L, URMI, E. & VÁŇA, J. 2002. Distribution of Hepaticae and Anthocerotae in Europe and Macaronesia. Lindbergia 27, 21–36. STECH, M. 2004. Supraspecific circumscription and classification of Campylopus (Dicranaceae, Bryopsida) based on inferences from sequence data. Systematic Botany 29, 817–824. STECH, M., SIM-SIM, M., ESQUÍVEL, M.G., FONTINHA, S., TANGNEY, R., LOBO, C., GABRIEL, R. & QUANDT, D. 2008. Explaining the ‘anomalous’ distribution of Echinodium Jur. (Bryopsida): independent evolution in Macaronesia and Australasia. Organisms Diversity and Evolution 8, 282–292. STECH, M., SIM-SIM, M. & FRAHM, J.-P. 2007. Campylopus (Leucobryaceae, Bryopsida) on Madeira Island—molecular relationships and biogeographic affinities. Nova Hedwigia Beiheft 131, 91–100. ∗ SWOFFORD, D.L. 2002. PAUP : Phylogenetic Analysis Using Par∗ simony ( and Other Methods) Version 4.0b10. Sinauer Associates Inc., Sunderland, MA. VANDERPOORTEN, A. & LONG, D.G. 2006. Budding speciation and neotropical origin of the Azorean endemic liverwort, Leptoscyphus azoricus. Molecular Phylogenetics and Evolution 40, 73–83. YAMADA, K. 1979. A revision of Asian taxa of Radula, Hepaticae. Journal of the Hattori Botanical Laboratory 45, 201–322. YAMADA, K. 1988. Studies on Taiwan Hepaticae VII. Radulaceae. Bulletin of the National Science Museum, Tokyo; Série B 14, 41–51. YAMADA, K. 1999. Taxonomy of the section Amentulosae of the genus Radula (Radulaceae, Hepaticae). Haussknechtia Beiheft 9, 391–397. ZHENG, M. & ZHU, R.-L. 2009. Karyological studies on some species of Radula (Radulaceae, Jungermanniopsida, Marchantiophyta). Nova Hedwigia 88, 229–244.