DISSERTATION Titel der Dissertation Molecular studies in Bromeliaceae: Implications of plastid and nuclear DNA markers for phylogeny, biogeography, and character evolution with emphasis on a new classification of Tillandsioideae Verfasser Mag.rer.nat. Michael Harald Johannes Barfuss angestrebter akademischer Grad Doktor der Naturwissenschaften (Dr.rer.nat.) Wien, 2012 Studienkennzahl lt. Studienblatt: A 091 438 Dissertationsgebiet lt. Studienblatt: Botanik Betreuerin: Ao. Univ.-Prof. Dr. Rosabelle Samuel Acknowledgements Warm thanks go to my supervisor, Rosabelle Samuel, for her support and guidance through the years, and to Walter Till for his long-standing collaboration and sharing his taxonomic expertise on Bromeliaceae without which the research would not have been successful. Thanks go to Tod F. Stuessy, emeritus head of the department of Systematic and Evolutionary Botany at the University of Vienna, for making the laboratory facilities available. Above all, I would like to thank my colleagues from the molecular laboratory, Elfriede Grasserbauer, Verena Klejna and Gudrun Kohl, for their patience and support that enabled me to complete this work. I would like to extend my gratitude to people who provided plant material, especially Renate Ehlers, Liselotte and Helmut Hromadnik, Elton M.C. Leme, Jose M. Manzanares, and Juan P. Pinzón. Many thanks go to all co-authors, especially to Thomas J. Givnish, Georg Zizka and Katharina Schulte for helpful discussions, and also to Félix Forest and Mark W. Chase at the Royal Botanic Gardens, Kew, U.K., with whom I shared useful discussions on low-copy nuclear genes during my stay at the Jodrell Laboratory. Special thanks go to Ovidiu Paun, Anton Russell, and Hermann Voglmayr for valuable discussions, critical review of some parts of this thesis and technical support. Several members of the department of Systematic and Evolutionary Botany in Vienna have supported me intellectually and morally and this is gratefully acknowledged: AnneCaroline Cosendai, Kathere Emadzade, Patricio López Sepúlveda, Eva Temsch, Barbara Turner, and former students, Cordula Blöch, Sutee Duangjai, Eva Maria Mayr, Carolin A. Rebernig, Stefan Safer, Reinhold Stockenhuber. Finally I want to thank my family and all friends all over the world who have helped and supported me in various ways, and anybody else I may have forgotten to mention. Financial support provided by the University of Vienna, the Commission for Interdisciplinary Ecological Studies (KIÖS, grant no. 2007-02) at the Austrian Academy of Sciences (ÖAW), the SYNTHESYS (Synthesis of Systematic Resources) program funded by European Union, and the Austrian Science Foundation (FWF; grant no. P13690) is greatly acknowledged. 3 Danksagung Besonderer Dank geht an meine Betreuerin, Rosabelle Samuel, für ihre langjährige Unterstützung und Begleitung, und an Walter Till für seine fortwährende Zusammenarbeit und Einbringung seiner taxonomischen Bromelien-Expertise, ohne die diese Forschungsarbeit nicht erfolgreich gewesen wäre. Dank geht an Tod F. Stuessy, emeritierter Leiter des Departments für Botanische Systematik und Evolutionsforschung der Universität Wien, für die Bereitstellung der Laboreinrichtungen. Vor allem möchte ich meinen Kolleginnen aus dem Molekularlabor, Elfriede Grasserbauer, Verena Klejna und Gudrun Kohl, für ihre Geduld und Unterstützung danken, die es mir ermöglicht haben, diese Arbeit fertig zu stellen. Herzlicher Dank geht an mehrere Personen, die Pflanzenmaterial zur Verfügung gestellt haben, insbesondere Renate Ehlers, Helmut und Liselotte Hromadnik, Elton M.C. Leme, Jose M. Manzanares, und Juan P. Pinzón. Dank geht an alle Mitautoren, vor allem an Thomas J. Givnish, Georg Zizka und Katharina Schulte für hilfreiche Diskussionen und auch an Félix Forest und Mark W. Chase der „Royal Botanic Gardens, Kew“ (U.K.) für wertvolle Diskussionen über nukleäre Gene während meines Forschungsaufenthaltes im „Jodrell Laboratory“. Besonderer Dank geht an Ovidiu Paun, Anton Russell und Hermann Voglmayr für wertvolle Diskussionen, kritische Überprüfung einiger Teile dieser Arbeit und technische Unterstützung. Mehrere Kollegen des Departments für Botanische Systematik und Evolutionsforschung der Universität Wien haben mich intellektuell und moralisch unterstützt, wofür ich mich aufs Herzlichste bedanken möchte: Anne-Caroline Cosendai, Kathere Emadzade, Patricio López Sepúlveda, Eva Temsch, Barbara Turner; und ehemalige Studenten, Cordula Blöch, Sutee Duangjai, Eva Maria Mayr, Carolin A. Rebernig, Stefan Safer, Reinhold Stockenhuber. Abschließend möchte ich meiner Familie und allen Freunden auf der ganzen Welt, die mir geholfen und mich in verschiedenster Weise unterstützt haben, und all jene ich vergessen habe zu erwähnen, danken. Finanzielle Unterstützung wurde von der Universität Wien, der Kommission für Interdisziplinäre Ökologische Studien (KIÖS, Projektnr. 2007-02) an der Österreichischen Akademie der Wissenschaften (ÖAW), dem von der Europäischen Union geförderte SYNTHESYS-Programm („Synthesis of Systematic Resources“), und den Fonds zur Förderung der Wissenschaftlichen Forschung (FWF; Projektnr. P13690) bereit gestellt. 4 Contents Acknowledgements.............................................................................................................................. 3 Danksagung ....................................................................................................................................... 4 Abstract ....................................................................................................................................... 7 Kurzfassung................................................................................................................................. 9 Introduction .............................................................................................................................. 11 Molecular markers for plant phylogenetics ............................................................................. 11 Characteristics, taxonomic history, and previous molecular studies in Bromeliaceae ............ 12 Taxonomic history and previous molecular studies in Tillandsioideae .................................... 17 Aims of the present study......................................................................................................... 19 References ................................................................................................................................ 19 Part 1. Application of plastid DNA markers in phylogenetic reconstructions of Bromeliaceae............................................................. 27 Chapter 1. Phylogeny, adaptive radiation, and historical biogeography in Bromeliaceae: Insights from an eight-locus plastid phylogeny .................................................. 29 Part 2. Optimization of nuclear DNA markers and their application in phylogenetic reconstructions of subfamilies Bromelioideae and Tillandsioideae (Bromeliaceae) ....................................................... 55 Chapter 2. Phylogeny of Bromelioideae (Bromeliaceae) inferred from nuclear and plastid DNA loci reveals the evolution of the tank habit within the subfamily .............. 57 Chapter 3. Optimizing eight nuclear DNA markers for phylogenetic studies in recently diverged angiosperms: A case study in Bromeliaceae subfamily Tillandsioideae . 73 Part 3. Classification of Bromeliaceae subfamily Tillandsioideae............. 117 Chapter 4. Progress towards a new classification of Tillandsioideae ................................ 119 Chapter 5. A new classification of Bromeliaceae subfamily Tillandsioideae inferred from DNA sequences data of two genomes and morphology .................................. 129 Appendix ................................................................................................................................. 207 Conclusions ............................................................................................................................. 233 Curriculum Vitae ................................................................................................................... 235 5 Abstract The monocot family Bromeliaceae comprises approximately 3,140 species distributed in tropical and subtropical regions of the New World from the southern United States to southern Argentina. The family is subdivided into eight subfamilies; the most species-rich are Tillandsioideae and Bromelioideae. Taxonomic concepts within Bromeliaceae are highly problematic, since discriminating morphological characters have been shown to be homoplastic or plesiomorphic. The present study aims to provide a robust phylogenetic framework for Bromeliaceae, especially for the most diverse and complicated subfamilies Bromelioideae and Tillandsioideae. Resulting phylogenies provide a basis to estimate the usefulness of morphological characters and to propose or strengthen hypotheses concerning evolutionary traits, biogeography, age and origins of bromeliads. The main questions raised are: (1) Do additional sequence data from the plastid genome and a wider sampling within Bromeliaceae provide a better resolved, robust phylogenetic framework? What are the reasons for the low DNA sequence divergence observed up to now? (2) Can nuclear DNA sequences be successfully implemented for phylogenetic reconstruction? What are the challenges to optimize nuclear markers and do they perform better than plastid loci? (3) Can the resulting phylogeny based on plastid and nuclear DNA sequences together with the re-evaluated morphological characters provide a reasonable, stable classification? To provide a more robust phylogenetic hypothesis for the classification of Bromeliaceae, eight rapidly evolving plastid DNA markers (atpB-rbcL, matK, ndhF, psbA-trnH, rpl32-trnL, rps16, trnL intron, and trnL-trnF) and 90 bromeliad species were included in the current study and analysed using maximum-parsimony, maximum-likelihood, and Bayesian approaches. Results support the formerly proposed eight-subfamily classification based on the single plastid gene ndhF. Support values for five of the subfamilies increase, but that for Lindmanioideae, Puyoideae, and Bromelioideae decrease as a result of expanded taxon sampling, including several more divergent species. The initially proposed monophyletic origin of Puyoideae cannot be unambiguously confirmed. Calibration of the resulting phylogeny against time and biogeographic analysis reveals that Bromeliaceae originated in the Guayana Shield about 100 million years ago (Ma) and spread radially into adjacent areas ca. 16—13 Ma. Extant lineages arose between 20 and 5 Ma. Andean uplift facilitated diversification of core Tillandsioideae about 14.2 Ma and Bromelioideae 10.1 Ma, the latter having their greatest diversity in the Brazilian Shield due to dispersal from the Andes. The most species-rich genera did not appear before 8.7 My with a high diversification between 5 and 4 Ma, which is most likely the reason for the comparatively low sequence divergence. To test the usefulness of nuclear regions for phylogenetic reconstruction in Bromelioideae, DNA sequences of part of the low-copy nuclear gene phosphoribulokinase (PRK) and five 7 plastid loci (matK, 3'trnK intron, trnL intron, trnL-trnF, and atpB-rbcL) were investigated. Phylogenetic trees obtained from analyses of the PRK sequences do not contradict trees obtained from plastid markers. The PRK matrix shows a significantly higher number of potentially PICs (phylogenetically informative characters) than the plastid dataset (16.9% vs. 3.1%), which improves resolution and support in the resulting trees. Although PRK is not able to resolve relationships completely, the combined analysis with plastid markers yields good support for several uncertain relationships observed previously. The early diverging lineages can be identified (“basal bromelioids”) and the remainder of the subfamily clustered into a highly supported clade (“eu-bromelioids”). Results indicate that taxonomic circumscriptions within “core bromeliads” are still insufficient, and relationships complex and difficult to solve. Several genera appear polyphyletic, and Aechmea as well as Quesnelia remain the most complicated genera of the subfamily. Most-parsimonious character state reconstructions for two evolutionary traits (tank habit, sepal symmetry) indicate that both characters have undergone few transitions within the subfamily and thus are not as homoplasious as previously assumed. The comparative study of nuclear DNA sequences within Tillandsioideae shows that some nuclear markers are able to provide more information and a higher degree of resolution in phylogenetic trees than plastid markers. However, their utility does not depend only on sequence variability, but also on methodological challenges in using traditional Sangersequencing. The internal transcribed spacer of nuclear ribosomal DNA (ITS nrDNA) is not recommended as a suitable marker for phylogenetic investigations of Bromeliaceae due to the presence of strong secondary structures which create problems in amplification and sequencing as well as its low number of PICs for resolving deeper nodes. Amplified fragments of the genes malate synthase (MS) and RNA polymerase II, beta subunit (RPB2) are not helpful due to their small size and limited number of PICs. Glucose-6-phosphate isomerase (PGIC), nitrate reductase (NIA), and xanthine dehydrogenase (XDH) need to be further investigated. Phosphoribulokinase (PRK) and phytochrome C (PHYC) are useful nuclear markers and able to provide considerable resolution in phylogenetic trees, but some relationships are poorly supported. The combined analysis of nuclear DNA sequence data (PRK, PHYC) with the already existing plastid DNA sequence data (atpB-rbcL, matK, rbcL, partial rbcL-accD, rps16 intron, partial trnK intron, trnL intron and trnL-trnF) shows a significant increase of resolution within phylogenetic trees of Tillandsioideae. Nine accepted genera can be re-circumscribed and three new genera are described taxonomically based on morphology. For morphologically distinct species groups within Racinaea and Tillandsia, two new subgenera are erected. Viridantha has been downgraded to subgeneric rank. Poor sampling within the Cipuropsis-Mezobromelia clade and missing support for clades within Tillandsia prevent the recognition of further taxonomic groups. 8 Kurzfassung Die Ananasgewächse (Bromeliaceae) sind mit etwa 3.140 Arten eine der artenreichsten Familien der einkeimblättrigen Blütenpflanzen, die in tropischen und subtropischen Regionen der Neuen Welt vom Süden der Vereinigten Staaten bis in den Süden Argentiniens vorkommen. Die Familie ist in acht Unterfamilien unterteilt, deren artenreichste die Tillandsioideae und Bromelioideae sind. Taxonomische Konzepte innerhalb der Bromeliaceae sind sehr problematisch, da zur Unterscheidung herangezogene morphologische Merkmale homoplastisch oder plesiomorph sind. Die vorliegende Studie zielt darauf ab, eine gut gestützte Stammbaumrekonstruktion für die Bromeliaceae zu erstellen, insbesondere für die komplexen Unterfamilien Bromelioideae und Tillandsioideae. Die resultierenden Phylogenien sollen eine Grundlage bieten, die taxonomische Verwendbarkeit von morphologischen Merkmalen zu bewerten und Hypothesen zur Merkmalsevolution, Biogeographie, Alter und Herkunft der Bromelien aufzustellen oder zu untermauern. Die wichtigsten Fragenstellungen sind: (4) Zeigen eine breitere Probennahme innerhalb der Bromeliaceae sowie zusätzliche Sequenz-Daten aus dem Plastiden-Genom einen besser aufgelösten und gut gestützten Stammbaum? Was sind Gründe für die bisher beobachtete geringe DNS-SequenzDivergenz? (5) Können nukleäre DNS-Sequenzen erfolgreich zur Erstellung von Stammbaumrekonstruktion verwendet werden? Was sind die Herausforderungen um nukleäre Marker zu optimieren und sind sie besser geeignet als Plastiden-Marker? (6) Kann man mit Hilfe der auf Plastiden- und Kern-DNS-Sequenzen basierenden Phylogenie zusammen mit neu bewerteten morphologischen Merkmalen eine vernünftige, stabile Klassifikation erreichen? Um eine besser abgestützte phylogenetische Hypothese für die Klassifizierung der Bromeliaceae zu erreichen, wurden in der aktuellen Studie acht schnell evoluierende Plastiden-DNSMarker (atpB-rbcL, matK, ndhF, psbA-trnH, rpl32-trnL, rps16, trnL Intron und trnL-trnF) und 90 Bromelienarten ausgewählt und mit Maximum-Parsimony-, Maximum-Likelihood- und Bayesian-Methoden analysiert. Die Ergebnisse unterstützen die schon früher vorgeschlagene Einteilung in acht Unterfamilien, die auf einem einzigen Plastiden-Gen (ndhF) basiert. Unterstützungswerte für fünf der Unterfamilien haben sich erhöht, jedoch die für Lindmanioideae, Puyoideae und Bromelioideae sind infolge der erweiternden Taxon-Auswahl gesunken. Die ursprünglich vorgeschlagene Monophylie der Puyoideae kann nicht eindeutig bestätigt werden. Eine kalibrierte zeitliche Zuordnung der resultierenden Phylogenie und biogeographische Analysen zeigen, dass die Bromeliaceae vor rund 100 Millionen Jahren (My) am Guayana-Schild entstanden sind und sich radial in die angrenzenden Gebiete vor ca. 16-13 My ausgebreitet haben. Heute noch lebende Entwicklungslinien entstanden vor 20-5 My. Die Andenhebung begünstigte die Artbildung innerhalb der Kern-Tillandsioideae vor etwa 14,2 My und die der Bromelioideae vor 10,1 My, wobei letztere ihre größte Vielfalt am brasilianischen Schild aufgrund der Ausbreitung von den Anden her hat. Die heute artenreichsten Gattungen mit einer 9 erhöhten Artbildungsrate vor 5-4 Ma sind nicht vor 8,7 My erschienen, was höchstwahrscheinlich der Grund für die vergleichsweise niedrige Sequenzdivergenz ist. Um die Verwendbarkeit von nukleären Regionen für phylogenetische Untersuchungen in den Bromelioideae zu testen, wurden DNS-Sequenzen von einem Abschnitt des in geringer Kopienzahl vorliegenden nukleären Gens Phosphoribulokinase (PRK) und fünf PlastidenMarker (matK, 3'trnK Intron, trnL Intron, trnL-trnF und atpB-rbcL) untersucht. Stammbäume aus Analysen der PRK-Daten widersprechen nicht den Ergebnissen, die mit Hilfe von PlastidenMarkern erhalten werden. Die PRK-Matrix zeigt eine deutlich höhere Anzahl an potenziellen PICs (phylogenetisch informative Merkmale), die die Auflösung und die Absicherung der resultierenden Stammbäume verbessert, als der Plastiden-Datensatz (16,9% vs. 3,1%). Obwohl PRK nicht in der Lage ist die Verwandtschaftsbeziehungen vollständig aufzulösen, liefert die kombinierte Analyse mit den Plastiden-Markern eine gute Unterstützung für mehrere zuvor unsichere Verwandtschaftsverhältnisse. Die früh abzweigenden Entwicklungslinien konnten identifiziert werden („basal bromelioids“) und der Rest der Unterfamilie ist in einem hoch unterstützten Ast vereinigt („eu-bromelioids“). Die Ergebnisse zeigen, dass die taxonomischen Beziehungen von Gattungen innerhalb der „core bromelioids“ verflochten und schwierig zu lösen sind. Mehrere Gattungen erscheinen polyphyletisch und Aechmea sowie Quesnelia bleiben die komplexesten Gattungen der Unterfamilie. Most-Parsimonious-Rekonstruktionen von Merkmalszuständen für Trichterhabitus und Kelchblattsymmetrie zeigen, dass beide Merkmale innerhalb der Unterfamilie nur wenigen morphologischen Veränderungen unterworfen waren und daher nicht so stark homoplastisch sind wie bisher angenommen. Die vergleichende Studie von nukleären DNS-Sequenzen innerhalb der Tillandsioideae zeigt, dass einige nukleäre Marker in der Lage sind, mehr Informationen und eine höhere Auflösung in den Stammbäumen zu erzielen als Plastiden-Marker. Allerdings ist ihre Verwertbarkeit nicht nur von der Sequenzvariabilität, sondern auch von methodischen Herausforderungen der traditionellen Sanger-Sequenzierung abhängig. Der „internal transcribed spacer“ der nukleären ribosomalen DNS (ITS nrDNS) ist aufgrund von vorhandenen starken Sekundärstrukturen, die Probleme bei der Amplifikation und Sequenzierung bereiten, sowie seiner geringen Anzahl an PICs kein geeigneter Marker zum Auflösen tieferer Knoten in phylogenetischen Untersuchungen bei den Bromeliaceae. Amplifizierte Fragmente der Gene Malat-Synthase (MS) und RNA-Polymerase-II, Beta-Untereinheit (RPB2) sind aufgrund ihrer geringen Länge und der begrenzten Anzahl an PICs nicht sonderlich hilfreich. Glucose-6-Phosphat-Isomerase (PGIC), Nitrat-Reduktase (NIA) und Xanthin-Dehydrogenase (XDH) müssen weiter untersucht werden. Phosphoribulokinase (PRK) und Phytochrom C (PHYC) sind nützliche nukleäre Marker und bringen eine gute Auflösung innerhalb vieler Bereiche der Stammbäumen, jedoch sind einige Verwandtschaftsverhältnisse nicht aufgelöst oder nur schlecht unterstützt. Die kombinierte Analyse der Kern-DNS-Sequenzdaten (PRK, PHYC) mit den bereits vorhandenen Plastiden-DNS-Sequenzdaten (atpB-rbcL, matK, rbcL, einem Teilstück des rbcL-accD, rps16 intron, einem Teilstück des trnK intron, trnL intron und trnL-trnF) zeigt eine deutliche Steigerung der Auflösung innerhalb der Stammbäume der Tillandsioideae. Basierend auf morphologischen Merkmalen konnten neun bisher akzeptierte Gattungen neu definiert und drei Gattungen neu beschrieben werden. Für morphologisch abweichende Artengruppen innerhalb von Racinaea und Tillandsia werden zwei neue Untergattungen aufgestellt. Viridantha wird auf Untergattungsniveau herabgestuft. Geringe Probenahme im Cipuropsis-Mezobromelia-Ast und fehlende Unterstützung für Verwandtschaftsgruppen innerhalb von Tillandsia verhindern eine taxonomische Beschreibung weiterer Artengruppen. 10 Introduction Molecular markers for plant phylogenetics Since the first extensive publication of an angiosperm phylogeny based on plastid rbcL sequence data (Chase & al., 1993), molecular studies in plants using DNA sequences have become popular. Advancements in sequencing technologies in both hardware and software have facilitated the addition of vast amounts of sequence information to public databases (e.g., GenBank) and a great number of published phylogenetic studies over the last 20 years. Milestones in sequencing technology have included the development of the chain-termination method (Sanger-sequencing) by Sanger & al. (1977) and the release of affordable automated plate (ABI PRISM 377 Genetic Analyzer, Life Technologies) and, later, capillary sequencers (e.g., Applied Biosystems 3730 DNA analyser, Life Technologies; Amersham Bioscience MegaBACE 4000, GE Healthcare) for detection of fluorescently labelled fragments. Preparation, handling, and throughput of sequences improved greatly, and generating sequence data became a routine exercise. Since 2005, next generation sequencing (NGS) technologies have enabled highthroughput sequencing using fluorescence (454, Roche; Solexa, Illumina; SOLiD, Life Technologies) or pH change measurements (Ion Torrent, Life Technologies) for detection of incorporated nucleotides. Although different applications have recently been developed, high costs, handling and analysing the huge amounts of sequence data, and the rather uncontrolled sequencing of genomic fragments are still limiting factors for general application of NGS methodologies to plant phylogenetics. During the period 1993–2000 the most important sources of data for reconstructing plant phylogenies were a small number of plastid DNA loci and the multi-copy 18S-5.8S-26S nuclear ribosomal DNA (nrDNA) repeat unit, e.g., (1) plastid DNA markers: e.g., atpB (e.g., Hoot & al., 1995); atpB-rbcL (e.g., Manen & al., 1994); matK (e.g., Johnson & Solitis, 1994, 1995; Liang & Hilu, 1996); ndhF (e.g., Olmstead & Reeves, 1995; Olmstead & al., 2000); psbA-trnH (e.g., Sang & al., 1997); rbcL (e.g., Nickrent & Soltis, 1995); rps16 intron (e.g., Oxelman & al., 1997); trnT-trnLtrnF (e.g., Taberlet & al., 1991; Gielly & Taberlet, 1996; Sang & al., 1997); (2) nuclear ribosomal DNA markers: e.g., 18S nrDNA (e.g., Nickrent & Soltis, 1995); 26S nrDNA (e.g., Bult & al., 1995; Kuzoff & al., 1998); external transcribed spacer (ETS nrDNA) (e.g., Baldwin & Markos, 1999); internal transcribed spacer (ITS nrDNA) (e.g., White & al., 1990; Baldwin, 1992; Sun & al., 1994; Baldwin & al., 1995; Sang & al., 1995; Blattner, 1999; Douzery & al., 1999). More conserved gene regions (atpB, rbcL, 18S nrDNA, 26S nrDNA) were applied at higher taxonomic levels, faster evolving genes (matK, ndhF) at intermediate levels, and more variable introns and spacer regions (atpB-rbcL, psbA-trnH, rps16 intron, trnT-trnL-trnF, ETS nrDNA, ITS nrDNA) at interspecific levels. Although new primers for novel markers had been published (e.g., Demesure & al., 1995; Dumolin-Lapegue & al., 1997; Small & al., 1998), most studies 11 BARFUSS, M.H.J. MOLECULAR STUDIES continued to use these frequently sequenced plastid loci. Comparative studies on several plastid DNA markers including information on existing or new universal primers (Shaw & al., 2005, 2007) have increased the number of applicable plastid DNA markers and primers, but also promoted the further use of plastid DNA sequences solely. Additional published nrDNA primers (e.g., Gruenstaeudl & al, 2009) continue to facilitate the usage of the ITS nrDNA (e.g., Russell & al., 2010a; Hřibová & al., 2011). The first low-copy nuclear DNA genes investigated for their phylogenetic utility belonged to the phytochrome gene family (PHYA, PHYB, PHYC, PHYD, PHYE; e.g., Mathews & al., 1995; Mathews & Sharrock, 1996, 1997; Mathews & Donoghue, 1999; Mathews & al., 2000). Only a few additional studies had utilized other low-copy nuclear DNA loci (e.g., ADH: Morton & al., 1996; Sang & al., 1997; Sang & Zhang, 1999; ncpGS: Emshwiller & Doyle, 1999; NIA: Howarth & Baum, 2002; PEPC: Gehrig & al., 1998, 2001; RPB2: Denton & al., 1998). It has been shown that low-copy nuclear DNA sequences are more variable and phylogenetically informative markers than plastid DNA regions (e.g., Strand & al., 1997; Sang, 2002; Mort & Crawford, 2004; Small & al., 2004), but direct sequencing is unfeasible in many cases because of their biparental inheritance and high a frequency of heterozygous individuals carrying alleles of different lengths. In the case of multi-copy genes, differentiating between paralogous and orthologous sequences becomes problematic. The need for cloning, and difficulties in interpreting results due to evolutionary processes like hybridization (e.g., Sang, 2004; Ma & al., 2010) or incomplete lineage sorting (e.g., Piñeiro & al., 2009; Willyard & al., 2009) have limited their usage. However, lowcopy nuclear DNA sequences have been employed with success for some plant groups, e.g., angiosperms (RPB2: Oxelman & al., 2004; XDH: Morton & al., 2011), Arecaceae (MS: Lewis & Doyle, 2001, 2002; Thomas & al., 2006; PRK: Lewis & Doyle, 2002; RPB2: Thomas & al., 2006), Asteraceae (PGIC: Ford & al., 2006; Stuessy & al., 2011; GADPH: Vaezi & Brouillet, 2009), Orchidaceae (XDH: Górniak & al., 2010), Paeoniaceae (G3PAT: Tank & Sang, 2001), and Passifloraceae (ncpGS: Yockteng & Nadot, 2004; Clarkson & al., 2010). Nevertheless, most phylogenetic studies in flowering plants are still based on maternally inherited, single-copy plastid DNA regions and the ITS nrDNA, which is usually well homogenized due to the phenomenon of concerted evolution. The absence of universal primers for low-copy nuclear genes has limited their amplification in a wider range of taxonomic levels in angiosperms. In certain plant groups the study of few plastid DNA markers together with the ITS nrDNA has been shown to provide insufficient information for reconstructing phylogenies at lower taxonomic levels (e.g., Barfuss & al., 2005; Lukas, 2010; Safer, 2011). Two hypotheses can be formulated, either that in certain plant groups the plastid genome and nrDNA regions evolve at a much slower rate (e.g., Gaut & al., 1992, 1997), or that extant species have undergone a rapid diversification process in their recent history (e.g., Cronn & al., 2002; Hughes & Eastwood, 2006), which facilitated great morphological diversity, but comparatively little sequence divergence. In both cases a combined analysis of several plastid DNA markers (e.g., Chase & al., 2006; Russell & al., 2010b) or the study of faster-evolving low-copy nuclear loci (e.g., Álvarez & al., 2008; Russell & al., 2010a) are possible solutions. Characteristics, taxonomic history, and previous molecular studies in Bromeliaceae Bromeliaceae is among the most conspicuous and species-rich groups of monocot angiosperms, constituting an early-diverging lineage within the order Poales (Givnish & al., 2007, 2010, see Figure 1; APG III, 2009). They are distributed in various, ecologically diverse habitats in tropical and subtropical regions of the New World ranging from the southern United States 12 IN BROMELIACEAE INTRODUCTION Figure 1. Phylogram of the single most parsimonious tree resulting from analysis of the plastome data showing the phylogenetic position of Bromeliaceae as the earliest-diverging lineage of Poales. Branch lengths are proportional to the number of inferred substitutions along each branch. Bootstrap support for each node is shown above the corresponding branch. Monocots are highlighted with magenta branches; cyperids, xyrids, restiids, and graminids, with colored boxes. (Taken from Givnish & al., 2010). (Virginia) to Patagonia in southern Argentina (Smith & Downs, 1974; Givnish & al., 1997; Benzing 2000). Their greatest species diversity is found in mountainous regions at elevations between 1000 and 2500 m.a.s.l. throughout the entire distributional range. Only one species, Pitcairnia feliciana, is found outside this area in Western Africa, which has been shown to be a result of long-distance dispersal about 10 Ma (Givnish & al., 2007). This supports the hypothesis that extant bromeliads have developed much later than the Gondwanan continent breakup, similarly to other flowering plant families (e.g., Cactaceae). 13 BARFUSS, M.H.J. MOLECULAR STUDIES Bromeliads are morphologically very distinctive, hence their inclusion in a separate order Bromeliales. However, a relationship to Velloziaceae has been proposed by cladistic analyses of morphological characters (Gilmartin & Brown, 1987) and the early molecular studies supported genetic affinities to Rapateaceae or Mayacaceae (Clark & al., 1993; Duvall & al., 1993). The family is characterized by typical monocot flowers (K3, C3, A3+3, G(3)) arranged in often showy inflorescences, distinctive leaf rosettes that are often water-impounding (“tanks”), and leaves that bear absorptive trichomes for water and nutrient uptake. The latter two features have provided key innovations to favour colonization of habitats of alternating water and nutrient supply as epiphytes (e.g., Benzing, 2000; Crayn & al., 2004, see Figure 2). Additional adaptions to drought stress, like succulence and the evolution of the CAM photosynthetic pathway, have also enabled their establishment in less competitive, extreme and rocky environments as lithophytes (e.g., McWilliams, 1974; Givnish & al., 2007). Several species of the family are of important horticultural value and many botanical gardens, amateur botanists, and private collectors host a large number of species ex situ in living plant collections. In Bromeliaceae, ca. 3,140 species are classified into 58 genera, which have traditionally been assigned to three subfamilies (Smith & Downs, 1974, 1977, 1979; Smith & Till, 1998): (1) Bromelioideae: inferior ovarys, baccate fruits, and serrate or rarely entire leave margins; (2) Pitcairnioideae: superior to partly inferior ovarys, capsular fruits, bicaudate, winged or rarely naked seeds, and serrate or rarely entire leave margins; (3) Tillandsioideae: superior to semi- inferior ovarys, capsular fruits, plumose seed appendages, and exclusively entire leave margins. Recent molecular studies have revealed that both Bromelioideae and Tillandsioideae are monophyletic and well characterized by their traditional morphology (e.g., Terry & al., 1997a, b; Crayn & al., 2004; Barfuss & al., 2005; Givnish & al., 2004, 2007; Schulte & al., 2005). Pitcairnioideae is paraphyletic and morphological characters used to define this subfamily are plesiomorphic. Therefore Givnish & al. (2007, see Figure 3) established five additional subfamilies based on DNA sequence data and a re-evaluation of morphological characters to circumscribe natural, monophyletic lineages: Brocchinioideae, Lindmanioideae, Hechtioideae, Navioideae, and Puyoideae. The most species-rich and taxonomically complicated subfamilies of Bromeliaceae are Bromelioideae and Tillandsioideae. Subfamily Bromelioideae is the second largest with 32 genera and ca. 800 species (Smith & Till, 1998; Luther 2010). Although distributed over nearly the whole range of the family, the centre of diversity is found in the Atlantic rain forest of southeastern Brazil (Smith & Downs, 1979). The largest and taxonomically most complex genus of Bromelioideae is Aechmea (260 spp.). Subfamily Tillandsioideae is the largest subfamily with nine generally accepted genera and ca. 1,300 species (Smith & Downs, 1977; Smith & Till, 1998; Luther, 2010; Luther & Rabinowitz, 2010) found throughout the distribution range of Bromeliaceae, indicating the greatest potential to adapt to extreme environments. Centres of diversity are the northern Andes in Peru, Ecuador and Colombia, the Atlantic rain forest of south-eastern Brazil, and Central America (Smith & Downs, 1977). It is considered the most morphologically diverse subfamily, hosting the three large genera Guzmania (210 spp.), Vriesea (266 spp.), and the most species-rich genus of Bromeliaceae, Tillandsia (626 spp.). Generic and subgeneric delimitations within subfamilies based on morphology have been shown to be problematic and in need of urgent revision (e.g., Smith & Kress, 1989, 1990; 14 IN BROMELIACEAE INTRODUCTION Figure 2. Strict consensus tree and most-parsimonious reconstructions of the combined matK and rps16 intron data for 51 species of Bromeliaceae showing the evolution of life-form and photosynthetic pathway. Bootstrap values 13 are indicated above the corresponding branches. Carbon-isotope ratios (δ C values in ‰) are shown for the taxa analysed. The derived character-states “epiphytic” and “CAM” are highlighted in blue and red, respectively. (Taken from Crayn & al., 2004). Brown & Terry, 1992; Grant, 1993b, 1995a; Read & Baensch, 1994; Betancur & MirandaEsquivel, 1999; Brown & Leme, 2005; Betancur & Salinas, 2006; de Sousa & Wendt, 2008). Several studies investigating distinctive morphological features, which were not yet applied to the taxonomy, have provided evidence that many characters are plesiomorphic or homoplastic and hence in their current application fail to delimit natural groups within Bromeliaceae. Interand infrageneric relationships, especially of Bromelioideae and Tillandsioideae, still remained poorly understood and taxonomic circumscriptions inadequate, because studies have relied on a very restricted number of species in comparison to the species richness of the subfamilies and, often, morphological characters are missing from herbarium specimens. Repeated changes of taxonomic boundaries by transferring or excluding species, changes in taxonomic ranks, and newly proposed segregates since the last comprehensive monograph of the family (Smith 15 BARFUSS, M.H.J. MOLECULAR STUDIES Figure 3. Ultrametric tree for Bromeliaceae based on cross-verified penalized likelihood, showing inferred chronology of cladogenesis over the past 20 My. The new subfamilial classification is highlighted by brackets; membership of subfamilies in the three traditional subfamilies is indicated by shaded bars. Hollow bar = Bromelioideae; gray bar = Tillandsioideae; solid bars = Pitcairnioideae. (Taken from Givnish & al., 2007). & Downs, 1974, 1977, 1979), as well as the greatly increased number of newly described species (Luther, 2010) indicate an urgent need for modern taxonomic revision based on reliable morphological characters. The first molecular studies on Bromeliaceae used restriction site variation in plastid DNA (Ranker & al., 1990) to infer phylogenetic relationships within bromeliads. Early studies that included sequence data used the rbcL gene to clarify either subfamilial relationships or familial, ordinal and superordinal relationships of Bromeliaceae and related monocot families (Clark & al., 1993). Their results confirmed a monophyletic origin of Bromeliaceae, but relationships of the three subfamilies could not be clearly resolved, due to limited taxon sampling. Two studies on the ndhF gene from the small single-copy region of the plastid genome were published by Terry & al. (1997a, b). These studies focused on subfamilial relationships (1997a, 30 bromeliad species) and relationships below the subfamilial level of Tillandsioideae (1997b, 25 tillandsioid species). Since the resolution of trees obtained from analyses of ndhF sequences was rather 16 IN BROMELIACEAE INTRODUCTION low, subsequent publications used the more rapidly evolving trnL intron (Horres & al., 2000; 64 bromeliad species), matK, and rps16 intron sequences (Crayn & al., 2000, 2004; 51 bromeliad species) to construct phylogenies and infer evolution of certain traits of bromeliads (e.g., CAM photosynthesis, epiphytic habit). However, these results also showed low sequence divergence among taxa and it was initially speculated that the plastid genome of Bromeliaceae evolves at a much slower rate than in other angiosperm lineages (Gaut & al., 1992, 1997; Givnish & al., 2004). The greatest drawbacks of early bromeliad phylogenetic publications were their reliance on only a few sequences of the plastid genome and the inclusion of a very limited set of species. Therefore subsequent studies focused on the large subfamilies Tillandsioideae (Barfuss & al., 2005) and Bromelioideae (Schulte & Zizka, 2008) with a greater selection of taxa and four to seven plastid markers in concatenated data matrices for analyses. Although it was possible to resolve major phylogenetic units within the family and subfamilies, the results still showed low resolution and low support for some deeper nodes. Sequence divergence at generic and interspecific levels was still comparatively low. Nuclear DNA markers were only recently included in phylogenetic studies of bromeliads. Studies using limited sets of taxa were undertaken (1) for Bromelioideae using partial sequences of the gene phosphoribulokinase (PRK) (Schulte & al., 2009; see chapter 2); (2) for Tillandsia subg. Tillandsia using sequences of the internal transcribed spacer 2 (ITS2) and the ETS nrDNA (Chew & al., 2010); (3) for Aechmea sequencing parts of the genes RNA polymerase II, beta subunit (RPB2), glyceraldehyde-3-phosphate dehydrogenase (G3PDH), and ETS nrDNA (Sass & Specht, 2010); (4) for Puya using parts of exon 1 of phytochrome C (PHYC) (Jabaily & Sytsma, 2010); and (5) for Alcantarea investigating FLORICAULA/LEAFY (FLO/LFY) (Versieux & al., 2012). Despite the increased usage of nuclear regions, relationships within Bromelioideae and Tillandsioideae still remain inadequately resolved, particularly within complicated and species-rich genera such as Aechmea, Tillandsia, and Vriesea. Taxonomic history and previous molecular studies in Tillandsioideae In Tillandsioideae more than 1,300 species are classified into nine accepted genera: Alcantarea (28 spp.), Catopsis (18 spp.), Glomeropitcairnia (2 spp.), Guzmania (210 spp.), Mezobromelia (9 spp.), Racinaea (65 spp.), Tillandsia (626 spp.), Vriesea (266 spp.), and Werauhia (87 spp.). The large genera Tillandsia and Vriesea as well as Werauhia are further divided into subgenera or sections (Till, 2000a, b). Since the last monograph for Tillandsioideae by Smith & Downs (1977) there have been changes in taxonomic limits (Tillandsia, Vriesea), establishment of new genera (Alcantarea: Grant, 1995; Werauhia: Grant, 1995; Racinaea: Spencer & Smith, 1993; Viridantha: Espejo-Serna, 2002), and transfers of species to other genera (e.g., xerophytic Vriesea species: Grant, 1993b, 2005). These changes have demonstrated the unreliability of previous taxonomies based on traditional morphological characters (e.g., flower arrangement, corolla tube connations, stamen and style position, presence vs. absence of petal appendages). Additional studies on gross morphology of flowers have supported this view (e.g., Brown & Gilmartin, 1984, 1989; Böhme, 1988; Gross, 1988; Gortan, 1991; Halbritter, 1992) and shown that the three largest genera (Guzmania, Tillandsia, Vriesea) and Mezobromelia are paraphyletic. The new description and resurrection, respectively, of the genera Viridantha (Espejo-Serna, 2002) and Sodiroa (Betancur & Miranda-Esquivel, 1999) are not generally accepted (e.g., Barfuss & al., 2005; Luther 2010). The remaining five currently accepted genera seem to be well circumscribed, with Catopsis and Glomeropitcairnia being the most distinctive genera of the subfamily (Smith & Till, 1998; Till, 2000a). 17 BARFUSS, M.H.J. MOLECULAR STUDIES Figure 4. Phylogram of one most parsimonious tree of 110 Tillandsioideae and 12 outgroup taxa obtained from combined analysis of seven plastid markers. Branch length is given above the branches. Gray bars indicate the tribal classification of Tillandsioideae, black ones the other subfamilies and the super-outgroup (Rapateaceae). (Taken from Barfuss & al., 2005). The two earliest molecular studies focusing on Tillandsioideae were restricted to a single marker and a limited sample set (ndhF: Terry & al., 1997b; trnL intron: Horres & al., 2000). The outcomes were correspondingly poor: although they supported the monophyly of Tillandsioideae and the distinct positions of Catopsis and Glomeropitcairnia, relationships of the remaining genera were mostly unresolved. A more recent study with an increased taxon sampling and seven plastid DNA markers (Barfuss & al., 2005; see Figure 4) confirmed these results but was 18 IN BROMELIACEAE INTRODUCTION also able to resolve major taxonomic units and clarify relationships of additional genera (Alcantarea, Guzmania, Werauhia). The remaining genera turned out to be paraphyletic (Vriesea, Tillandsia) and/or poorly genetically differentiated (Racinaea, Tillandsia) as evident from short and weakly- or unsupported branches. This demonstrated the need for more rapidly evolving markers, preferably from the nuclear genome. Two recently published studies utilized either nrDNA regions (ITS2 and ETS nrDNA for Tillandsia subg. Tillandsia: Chew & al., 2010) or a single low-copy nuclear gene (FLO/LFY for Alcantarea: Versieux & al., 2012), but were restricted to small subgroups of Tillandsioideae and a poor sample selection. Aims of the present study The scope of the present investigation was defined by three main questions: (4) Do additional sequence data from the plastid genome and a wider sampling within Bromeliaceae provide a better resolved, robust phylogenetic framework? What are the reasons for the low DNA sequence divergence observed up to now? (5) Can nuclear DNA sequences be successfully implemented for phylogenetic reconstruction? What are the challenges to optimize nuclear markers, and do they perform better than plastid loci? (6) Can the resulting phylogenies based on plastid and nuclear DNA sequences together with the re-evaluated morphological characters provide a reasonable, stable classification? To answer these questions the work was divided into three parts, which correspond to the three chapters of the present thesis: (1) Application of additional plastid DNA sequence data to phylogenetic reconstruction in Bromeliaceae to test the eight-subfamily classification and to strengthen hypotheses concerning biogeography, origin, age, and key trait evolution of extant bromeliad lineages, pp. 27–54; (2) Optimization of nuclear DNA markers and their application to phylogenetic reconstructions within Bromeliaceae (in comparison to plastid DNA markers) to answer questions of relationships and character evolution within Bromelioideae and Tillandsioideae, pp. 55–116; (3) Implementation of phylogenetic results in the classification and taxonomy of Tillandsioideae, pp. 117–206. References Álvarez, I., Costa, A. & Feliner, G.N. 2008. Selecting single-copy nuclear genes for plant phylogenetics: A preliminary analysis for the Senecioneae (Asteraceae). J. Molec. Evol. 66: 276– 291. APG III (The angiosperm phylogeny group). 2009. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III. Bot. J. Linn. Soc. 161: 105–121. Baldwin, B.G. 1992. Phylogenetic utility of the internal transcribed spacers of nuclear ribosomal DNA in plants: an example from the Compositae. Molec. Phylog. Evol. 1: 3–16. Baldwin, B.G. & Markos, S. 1999. Phylogenetic utility of the external transcribed spacer (ETS) of 18S-26S rDNA: Congruence of ETS and ITS Trees of Calycadenia (Compositae). Molec. Phylog. Evol. 10: 449–463. Baldwin, B.G., Sanderson, M.J., Porter, J.M., Wojciechowski, M.F., Campbell, C.S. & Donoghue, M.J. 1995. The ITS region of nuclear ribosomal DNA: A valuable source of evidence on angiosperm phylogeny. Ann. Missouri Bot. Gard. 82: 247–277. 19 BARFUSS, M.H.J. MOLECULAR STUDIES Barfuss, M.H.J., Samuel, R., Till, W. & Stuessy, T.F. 2005. Phylogenetic relationships in subfamily Tillandsioideae (Bromeliaceae) based on DNA sequence data from seven plastid regions. Amer. J. Bot. 92: 337–351. Benzing, D.H. 2000. Bromeliaceae: profile of an adaptive radiation. Cambridge University Press, Cambridge, UK. Betancur, J. & Miranda-Esquivel, D.R. 1999. ¿Existe Sodiroa? Revista de la Academia Colombiana de Ciencias Exactas, Físicas y Naturales 23: 189–194. Betancur, J. & Salinas, N.R. 2006. El ocaso de Pseudaechmea (Bromeliaceae: Bromelioideae): the Pseudaechmea (Bromeliaceae: Bromelioideae) twilight. Caldasia 28: 157–164. Blattner, F.R. 1999. Direct amplification of the entire ITS region from poorly preserved plant material using recombinant PCR. BioTechniques 27: 1180–1186. Böhme, S. 1988. Bromelienstudien III. Vergleichende Untersuchungen zu Bau, Lage und systematischer Verwertbarkeit der Septalnektarien von Bromelien. Tropische und subtropische Pflanzenwelt 62: 125–274. Brown, G.K. & Gilmartin, A.J. 1984. Stigma structure and variation in Bromeliaceae - neglected taxonomic characters. Brittonia 36: 364–374. Brown, G.K. & Gilmartin, A.J. 1989. Stigma types in Bromeliaceae. A systematic survey. Syst. Bot. 14: 110–132. Brown, G.K. & Leme, E.M.C. 2005. The re-establishment of Andrea (Bromeliaceae: Bromelioideae), a monotypic genus from Southeastern Brazil threatened with extinction. Taxon 54: 63–70. Brown, G.K. & Terry, R.G. 1992. Petal appendages in Bromeliaceae. Amer. J. Bot. 79: 1051– 1071. Bult, C.J., Sweere, J.A. & Zimmer, E.A. 1995. Cryptic sequence simplicity, nucleotide composition bias, and molecular coevolution in the large subunit of ribosomal DNA in plants: implications for phylogenetic analyses. Ann. Missouri Bot. Gard. 82: 235–246. Chase, M.W. & al. 1993. Phylogenetics of seed plants: an analysis of nucleotide-sequences from the plastid gene rbcL. Ann. Missouri Bot. Gard. 80: 528–580. Chase, M.W., Fay, M.F., Devey, D.S., Maurin, O., Rønsted, N., Davies, T.J., Pillon, Y. & al. 2006. Multigene analyses of monocot relationships: A summary. Aliso 22: 63–75. Chew, T., De Luna, E. & González, D. 2010. Phylogenetic relationships of the pseudobulbous Tillandsia species (Bromeliaceae) inferred from cladistic analyses of ITS 2, 5.8S ribosomal RNA gene, and ETS sequences. Syst. Bot. 35: 86–95. Clark, W.D., Gaut, B.S., Duvall, M.R. & Clegg, M.T. 1993. Phylogenetic relationships of the Bromeliiflorae-Commeliniflorae-Zingiberiflorae complex of monocots based on rbcL sequence comparisons. Ann. Missouri Bot. Gard. 80: 987–998. Clarkson, J.J., Kelly, L.J., Leitch, A.R., Knapp, S. & Chase, M.W. 2010. Nuclear glutamine synthetase evolution in Nicotiana: Phylogenetics and the origins of allotetraploid and homoploid (diploid) hybrids. Molec. Phylog. Evol. 55: 99–112. Crayn, D.M., Terry, R.G., Smith, J.A.C. & Winter, K. 2000. Molecular systematic investigations in Pitcairnioideae (Bromeliaceae) as a basis for understanding the evolution of crassulacean acid metabolism (CAM). In K. L. Wilson and D. A. Morrison [eds.], Monocots: systematics and evolution, 569–579. CSIRO Publishing, Melbourne, Australia. Crayn, D.M., Winter, K. & Smith, J.A.C. 2004. Multiple origins of crassulacean acid metabolism and the epiphytic habit in the Neotropical family Bromeliaceae. Proc. Natl. Acad. Sci. U.S.A. 101: 3703–3708. 20 IN BROMELIACEAE INTRODUCTION Cronn, R.C., Small, R.L., Haselkorn, T. & Wendel, J.F. 2002. Rapid diversification of the cotton genus (Gossypium: Malvaceae) revealed by analysis of sixteen nuclear and chloroplast genes. Amer. J. Bot. 89: 707–725. Demesure, B., Sodzi, N. & Petit R.J. 1995. A set of universal primers for amplification of polymorphic non-coding regions of mitochondrial and chloroplast DNA in plants. Molec. Ecol. 4: 129–131. Denton, A.L., McConaughy, B.L. & Hall, B.D. 1998. Usefulness of RNA polymerase II coding sequences for estimation of green plant phylogeny. Molec. Biol. Evol. 15: 1082–1085. De Sousa, L.D.F. & Wendt, T. 2008. Taxonomy and conservation of the genus Lymania (Bromeliaceae) in the southern Bahian Atlantic Forest of Brazil. Bot. J. Linn. Soc. 157: 47–66. Douzery, E.J.P., Pridgeon, A.M., Kores, P., Linder, H.P., Kurzweil, H. & Chase, M.W. 1999. Molecular phylogenetics of Diseae (Orchidaceae): A contribution from nuclear ribosomal ITS sequences. Amer. J. Bot. 86: 887–899. Dumolin-Lapegue, S., Pemonge, M.H. & Petit, R.J. 1997. An enlarged set of consensus primers for the study of organelle DNA in plants. Molec. Ecol. 6: 393–397. Duvall, M.R., Clegg, M.T., Chase, M.W., Clark, W.D., Kress, W.J., Hills, H.G., Eguiarte, L.E., Smith, J.F., Gaut, B.S., Zimmer, E.A. & Learn Jr., G.H. 1993. Phylogenetic hypotheses for the monocotyledons constructed from rbcL sequence data. Ann. Missouri Bot. Gard. 80: 607–619. Emshwiller, E. & Doyle, J.J. 1999. Chloroplast-expressed glutamine synthetase (ncpGS): Potential utility for phylogenetic studies with an example from Oxalis (Oxalidaceae). Molec. Phylog. Evol. 12: 310–319. Espejo-Serna, A. 2002. Viridantha, un género nuevo de Bromeliaceae (Tillandsioideae) endémico de México. Acta Botánica Mexicana 60: 25–35. Ford, V.S., Lee, J., Baldwin, B.G. & Gottlieb, L.D. 2006. Species divergence and relationships in Stephanomeria (Compositae): PGIC phylogeny compared to prior biosystematic studies. Amer. J. Bot. 93: 480–490. Gaut, B.S., Clark, L.G., Wendel, J.F. & Muse, S.V. 1997. Comparisons of the molecular evolutionary process at rbcL and ndhF in the grass family (Poaceae). Molec. Biol. Evol. 14: 769– 777. Gaut, B.S., Muse, S.V., Clark, W.D. & Clegg, M.T. 1992. Relative rates of nucleotide substitution at the rbcL locus in monocotyledonous plants. J. Molec. Evol. 35: 292–303. Gehrig, H.H., Heute, V. & Kluge, M. 1998. Toward a better knowledge of the molecular evolution of phosphoenolpyruvate carboxylase by comparison of partial cDNA sequences. J. Mol. Evol. 46: 107–114. Gehrig, H.H., Heute, V. & Kluge, M. 2001. New partial sequences of phosphoenolpyruvate carboxylase as molecular phylogenetic markers. Molec. Phylog. Evol. 20: 262–274. Gielly, L. & Taberlet, P. 1996. A phylogeny of the European gentians inferred from chloroplast trnL (UAA) intron sequences. Bot. J. Linn. Soc. 120: 57–75. Gilmartin, A.J. & Brown, G.K. 1987. Bromeliales, related monocots, and resolution of relationships among Bromeliaceae subfamilies. Syst. Bot. 12: 493–500. Givnish, T.J., Ames, M., McNeal, J.R., McKain, M.R., Steele, P.R., dePamphilis, C.W., Graham, S.W. & al. 2010. Assembling the tree of the monocotyledons: Plastome sequence phylogeny and evolution of Poales. Ann. Missouri Bot. Gard. 97: 584–616. Givnish, T.J., Millam, K.C., Berry, P.E. & Sytsma, K.J. 2007. Phylogeny, adaptive radiation, and historical biogeography of Bromeliaceae inferred from ndhF sequence data. Pp. 3–26 In: 21 BARFUSS, M.H.J. MOLECULAR STUDIES Columbus, J.T., Friar, E.A., Porter, J.M., Prince, L.M. & Simpson M.G. (eds.), Monocots: Comparative Biology and Evolution – Poales. Claremont: Rancho Santa Ana Botanic Garden (RSABG). Givnish, T.J., Milliam, T.M., Evans, T.M., Hall, J.C., Berry, P.E. & Terry, R.G. 2004. Ancient vicariance or long-distance dispersal? Inferences about phylogeny and South AmericanAfrican disjunctions in Rapateaceae and Bromeliaceae based on ndhF sequence data. Int. J. Plant Sci. 165 (4, Suppl.), 35–54. Givnish, T.J., Sytsma, K.J., Smith, J.F., Hohn, W.J., Benzing, D.H. & Burkhardt, E.M. 1997. Molecular evolution and adaptive radiation in Brocchinia (Bromeliaceae: Pitcairnioideae) atop tepuis of the Guayana shield. Pp. 259–311. In: Givnish, T.J., Sytsma, K.J. (eds.), Molecular Evolution and Adaptive Radiation. Cambridge University Press, New York. Górniak, M., Paun, O. & Chase, M.W. 2010. Phylogenetic relationships within Orchidaceae based on a low-copy nuclear coding gene, XDH: Congruence with organellar and nuclear ribosomal DNA results. Molec. Phylog. Evol. 56: 784–795. Gortan, G. 1991. Narbenformen bei Bromeliaceen: Variationsmöglichkeiten und Überlegungen zu systematisch-taxonomischen Korrelationen. M.Sc. thesis, Universität Wien, Vienna, Austria. Grant, J.R. 1993b. True Tillandsias misplaced in Vriesea (Bromeliaceae: Tillandsioideae). Phytologia 75: 170–175. Grant, J.R. 1995a. Bromelienstudien. The resurrection of Alcantarea and Werauhia, a new genus. Tropische und subtropische Pflanzenwelt 91: 1–57. Grant, J.R. 2005 ("2004"). New Combinations and Names in Andean Pitcairnia, Tillandsia, and Werauhia (Bromeliaceae). Vidalia 2: 23–25. Gross, E. 1988. Bromelienstudien IV. Zur Morphologie der Bromeliaceen-Samen unter Berücksichtigung systematisch-taxonomischer Aspekte. Tropische und subtropische Pflanzenwelt 64: 415–625. Gruenstaeudl, M., Urtubey, E., Jansen, R.K., Samuel, R., Barfuss, M.H.J. & Stuessy, T.F. 2009. Phylogeny of Barnadesioideae (Asteraceae) inferred from DNA sequence data and morphology. Molec. Phylog. Evol. 51: 572–587. Hřibová, E., Čížková, J., Christelová, P., Taudien, S., de Langhe, E. & Doležel, J. 2011. The ITS15.8S-ITS2 sequence region in the Musaceae: Structure, diversity and use in molecular phylogeny. PLoS one 6: e17863 (1–11), doi: 10.1371/journal.pone.0017863. Halbritter, H. 1992. Morphologie und systematische Bedeutung des Pollens der Bromeliaceae. Grana 31: 197–212. Horres, R., Zizka, G., Kahl, G. & Weising, K. 2000. Molecular phylogenetics of Bromeliaceae: Evidence from trnL (UAA) intron sequences of the chloroplast genome. Pl. Biol. 2: 306– 315. Hoot, S.B. & Douglas, A.W. 1998. Phylogeny of the Proteaceae based on atpB and atpB-rbcL intergenic spacer region sequences. Australian Syst. Bot. 11: 301–320. Howarth, D.G. & Baum, D.A. 2002. Phylogenetic utility of a nuclear intron from nitrate reductase for the study of closely related plant species. Molec. Phylog. Evol. 23: 525–528. Hughes, Colin & Eastwood, R. 2006. Island radiation on a continental scale: Exceptional rates of plant diversification after uplift of the Andes. Proc. Natl. Acad. Sci. U.S.A. 93: 10334– 10339. Jabaily, R.S. & Sytsma, K.J. 2010. Phylogenetics of Puya (Bromeliaceae): Placement, major lineages, and evolution of Chilean species. Amer. J. Bot. 97: 337–356. 22 IN BROMELIACEAE INTRODUCTION Johnson, L.A. & Soltis, D.E. 1994. matK DNA sequence and phylogenetic reconstruction in Saxifragaceae s.s. Syst. Bot. 19: 143–156. Johnson, L.A. & Soltis, D.E. 1995. Phylogenetic inference in Saxifragaceae sensu stricto and Gilia (Polemoniaceae) using matK sequences. Ann. Missouri Bot. Gard. 82: 149–175. Kuzoff, R.K., Sweere, J.A., Soltis, D.E., Soltis, P.S. & Zimmer, E.A. 1998. The phylogenetic potential of entire 26S rDNA sequences in plants. Molec. Biol. Evol. 15: 251–263. Lewis, C.E. & Doyle, J.J. 2001. Phylogenetic Utility of the Nuclear Gene Malate Synthase in the Palm Family (Arecaceae). Molec. Phylogen. Evol. 19: 409–420. Lewis, C.E. & Doyle, J.J. 2002. A phylogenetic analysis of tribe Areceae (Arecaceae) using two low-copy nuclear genes. Pl. Syst. Evol. 236: 1–17. Liang, H. & Hilu, K.W. 1996. Application of the matK gene sequences to grass systematics. Can. J. Bot. 74: 125–134. Lukas, B. 2010. Molekulare Phylogenie und Phytochemie der Gattung Origanum. Dissertation, University of Vienna. Luther, H.E. 2010. An alphabetical list of bromeliad binomials. Pp. I-IV, 1–45 In: Rabinowitz, L. & Holst, B.K. (eds.), An alphabetical list of bromeliad binomials, 12th edition. Sarasota: The Sarasota Bromeliad Society & Marie Selby Botanical Gardens (MSBG). Luther, H.E. & Rabinowitz, L. 2010. De Rebus Bromeliacearum IV. Selbyana 30: 147–189. Ma, J.X., Li, Y.N., Vogl, C., Ehrendorfer, F. & Guo, Y.P. 2010. Allopolyploid speciation and ongoing backcrossing between diploid progenitor and tetraploid progeny lineages in the Achillea millefolium species complex: Analyses of single-copy nuclear genes and genomic AFLP. B.M.C. Evol. Biol. 10: 100 (1–11), doi:10.1186/1471-2148-10-100. Manen, J.-F., Natali, A. & Ehrendorfer, F. 1994. Phylogeny of the Rubiaceae-Rubieae inferred from the sequence of a cpDNA intergene region. Pl. Syst. Evol. 190: 195–211. Mathews, S. & Donoghue, M.J. 1999. The root of angiosperm phylogeny inferred from duplicate phytochrome genes. Science 286: 947–950. Mathews, S., Lavin, M. & Sharrock, R.A. 1995. Evolution of the phytochrome gene family and its utility for phylogenetic analyses of angiosperms. Ann. Missouri Bot. Gard. 82:.296–321. Mathews, S. & Sharrock, R.A. 1996. The phytochrome gene family in grasses (Poaceae): a phylogeny and evidence that grasses have a subset of the loci found in dicot angiosperms. Molec. Biol. Evol. 13: 1141–1150. Mathews, S. & Sharrock, R.A. 1997. Phytochrome gene diversity. Plant Cell Environ. 20: 666– 671. Mathews, S., Tsai, R.C. & Kellogg, E.A. 2000. Phylogenetic structure in the grass family (Poaceae): evidence from the nuclear gene phytochrome B. Amer. J. Bot. 87: 96–107. McWilliams, E.L. 1974. Evolutionary ecology. Pp 40–55 In L. B. Smith & R. J. Downs, Pitcairnioideae (Bromeliaceae). Flora Neotropica Monongraph 14, part 1. Hafner Press, New York, New York, USA. Mort, M.E. & Crawford, D.J. 2004. The continuing search: low-copy nuclear sequences for lower-level plant molecular phylogenetic studies. Taxon 53: 257–261. Morton, C.M. 2011. Newly sequenced nuclear gene (XDH) for inferring angiosperm phylogeny. Ann. Missouri Bot. Gard. 98: 63–89. Morton, B.R., Gaut, B. & Clegg, M.T. 1996. Evolution of alcohol dehydrogenase genes in the palm and grass families. Proc. Natl. Acad. Sci. U.S.A. 93: 11735–11739. Nickrent, D.L. & Soltis, D.E. 1995. A comparison of angiosperm phylogenies from nuclear 18S rDNA and rbcL sequences. Ann. Missouri Bot. Gard. 82: 208–234. 23 BARFUSS, M.H.J. MOLECULAR STUDIES Olmstead, R.G. & Reeves, P.A. 1995. Evidence for the polyphyly of the Scrophulariaceae based on chloroplast rbcL and ndhF sequences. Ann. Missouri Bot. Gard. 82: 176–193. Olmstead, R.G., Jansen R.K., Kim K.J. & Wagstaff, S.J. 2000. The phylogeny of the Asteridae s. l. based on chloroplast ndhF sequences. Molec. Phylog. Evol. 16: 96–112. Oxelman, B., Lidén, M. & Berglund, D. 1997. Chloroplast rps16 intron phylogeny of the tribe Sileneae (Caryophyllaceae). Pl. Syst. Evol. 206: 393–410. Oxelman, B., Yoshikawa, N., McConaughy, B.L., Luo, J., Denton, A.L. & Hall, B.D. 2004. RPB2 gene phylogeny in flowering plants, with particular emphasis on asterids. Molec. Phylog. Evol. 32: 462–479. Piñeiro, R., Costa, A., Aguilar, J.F. & Feliner, G.N. 2009. Overcoming paralogy and incomplete lineage sorting to detect a phylogeographic signal: A GapC study of Armeria pungens. Botany-Botanique 87: 164–177. Ranker, T.A., Soltis, D.E., Soltis, P.S. & Gilmartin, A.J. 1990. Subfamilial phylogenetic relationships of the Bromeliaceae: evidence from chloroplast DNA restriction site variation. Syst. Bot. 15: 425–434. Read, R.W. & Baensch, H.U. 1994. Ursulaea; a new genus of Mexican Bromeliads. J. Bromeliad Soc. 44: 205–211. Russell, A., Samuel, R., Klejna, V., Barfuss, M.H.J., Rupp, B. & Chase, M.W. 2010a. Reticulate evolution in diploid and tetraploid species of Polystachya (Orchidaceae) as shown by plastid DNA sequences and low-copy nuclear genes. Ann. Bot. (Oxford) 106: 37–56. Russell, A., Samuel, R., Rupp, B., Barfuss, M.H.J., Safran, M., Besendorfer, V. & Chase, M.W. 2010b. Phylogenetics and cytology of a pantropical orchid genus Polystachya (Polystachyinae; Vandeae; Orchidaceae): Evidence from plastid DNA sequence data. Taxon 59: 389– 404. Safer, S. (2011). Molecular and phytochemical investigations on the genus Leontopodium. Dissertation, University of Innsbruck. Small, R.L., Cronn, R.C. & Wendel, J.F. 2004. Use of nuclear genes for phylogeny reconstruction in plants. Aust. Syst. Bot. 17: 145–170. Samuel, R., Kathriarachchi, H., Hoffmann, P., Barfuss, M.H.J., Wurdack, K.J., Davis, C.C. & Chase, M.W. 2005. Molecular phylogenetics of Phyllanthaceae: Evidence from plastid matK and nuclear PHYC sequences. Amer. J. Bot. 92: 132–141. Sang, T. 2002. Utility of low-copy nuclear gene sequences in plant phylogenetics. Critical Reviews in Biochemistry and Molecular Biology 37: 121–147. Sang, T., Crawford, D.J. & Stuessy, T.F. 1995. Documentation of reticulate evolution in peonies (Paeonia) using sequences of internal transcribed spacer of nuclear ribosomal DNA: implications for biogeography and concerted evolution. Proc. Natl. Acad. Sci. U.S.A. 92: 6813– 6817. Sang, T., Crawford, D.J. & Stuessy, T.F. 1997. Chloroplast DNA phylogeny, reticulate evolution, and biogeography of Paeonia (Paeoniaceae). Amer. J. Bot. 84: 1120–1136. Sang, T., Donoghue, M.J. & Zhang, D. 1997. Evolution of alcohol dehydrogenase genes in peonies (Paeonia): phylogenetic relationships of putative non-hybrid species. Molec. Biol. Evol. 14: 994–1007. Sang, T., Pan, J., Zhang, D., Ferguson, D., Wang, C., Pan, K.-Y. & Hong, D.-Y. 2004. Origins of polyploids: an example from peonies (Paeonia) and a model for angiosperms. Biol. J. Linn. Soc.. 82: 561–571. 24 IN BROMELIACEAE INTRODUCTION Sang, T. & Zhang, D. 1999. Reconstructing hybrid speciation using sequences of low-copy nuclear genes: Hybrid origins of five Paeonia species based on Adh gene phylogenies. Syst. Bot. 24: 148–163. Sanger F, Nicklen, S & Coulson, A.R. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. U.S.A. 74: 5463–5467. Sass, C. & Specht, C.D. 2010. Phylogenetic estimation of the core bromelioids with an emphasis on the genus Aechmea (Bromeliaceae). Molec. Phylog. Evol. 55: 559–571. Schulte, K., Barfuss, M.H.J. & Zizka, G. 2009. Phylogeny of Bromelioideae (Bromeliaceae) inferred from nuclear and plastid DNA loci reveals the evolution of the tank habit within the subfamily. Molec. Phylogen. Evol. 51: 327–339. Schulte, K. & Zizka, G. 2008. Multi locus plastid phylogeny of Bromelioideae (Bromeliaceae) and the taxonomic utility of petal appendages and pollen characters. Candollea 63: 209– 225. Shaw, J., Lickey, E.B., Beck, J.T., Farmer, S.B., Liu, W.S., Miller, J., Siripun, K.C., Winder, C.T., Schilling, E.E. & Small, R.L. 2005. The tortoise and the hare II: Relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis. Amer. J. Bot. 92: 142– 166. Shaw, J., Lickey, E.B., Schilling, E.E. & Small, R.L. 2007. Comparison of whole chloroplast genome sequences to choose noncoding regions for phylogenetic studies in angiosperms: The tortoise and the hare III. Amer. J. Bot. 94: 275–288. Small, R.L., Ryburn, J.A., Cronn, R.C., Seelanan, T. & Wendel, J.F. 1998. The tortoise and the hare: choosing between noncoding plastome and nuclear Adh sequences for phylogenetic reconstruction in a recently diverged plant group. Amer. J. Bot. 85: 1301–1315. Smith, L.B. & Downs, R.J. 1974. Pitcairnioideae (Bromeliaceae). In Flora Neotropica. Monograph 14, vol. 1, 1–660. Hafner Press, New York, USA. Smith, L.B. & Downs, R.J. 1977. Tillandsioideae (Bromeliaceae). In Flora Neotropica. Monograph 14, vol. 2, 661–1492. Hafner Press, New York, USA. Smith, L.B. & Downs, R.J. 1979. Bromelioideae (Bromeliaceae). In Flora Neotropica. Monograph 14, vol. 3, 1493–2142. New York Botanical Garden, Bronx, New York, USA. Smith, L.B. & Kress, W.J. 1989. New or restored genera of Bromeliaceae. Phytologia 66, 70–79. Smith, L.B. & Kress, W.J. 1990. New genera of Bromeliaceae. Phytologia 69, 271–274. Smith, L.B. & Till, W. 1998. Bromeliaceae. In K. Kubitzki [ed.], The Families and Genera of Vascular Plants, vol. 4, 74–99. Spriger, Berlin, Germany. Spencer, M.A. & Smith, L.B. 1993. Racinaea, a new genus of Bromeliaceae (Tillandsioideae). Phytologia 74: 151–160. Strand, A.E., LeebensMack, J. & Milligan, B.G. 1997. Nuclear DNA-based markers for plant evolutionary biology. Molec. Ecol. 6: 113–118. Stuessy, T.F., Blöch, C., Villaseñor, J.L., Rebernig, C.A. & Weiss-Schneeweiss, H. 2011. Phylogenetic analyses of DNA sequences with chromosomal and morphological data confirm and refine sectional and series classification within Melampodium (Asteraceae, Millerieae). Taxon 60: 436–449. Sun, Y., Skinner, D.Z., Liang, G.H. & Hulbert, S.H. 1994. Phylogenetic analysis of Sorghum and related taxa using Internal Transcribed Spacers of nuclear ribosomal DNA. Theor. Appl. Genet. 89: 26–32. Taberlet, P., Gielly, L., Pautou G. & Bouvet, J. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Molecular Biology 17: 1105–1109. 25 Tank, D.C. & Sang, T. 2001. Phylogenetic utility of the glycerol-3-phosphate acyltransferase gene: Evolution and implications in Paeonia (Paeoniaceae). Molec. Phylog. Evol. 19: 421– 429. Terry, R.G., Brown, G.K. & Olmstead, R.G. 1997a. Examination of subfamilial phylogeny in Bromeliaceae using comparative sequencing of the plastid locus ndhF. Amer. J. Bot. 84: 664–670. Terry, R.G., Brown, G.K. & Olmstead, R.G. 1997b. Phylogenetic relationships in subfamily Tillandsioideae (Bromeliaceae) using ndhF sequences. Syst. Bot. 22: 333–345. Thomas, M.M., Garwood, N.C., Baker, W.J., Henderson, S.A., Russell, S.J., Hodel, D.R. & Bateman, R.M. 2006. Molecular phylogeny of the palm genus Chamaedorea, based on the low-copy nuclear genes PRK and RPB2. Molec. Phylogen. Evol. 38: 398–415. Till, W. 2000a. Tillandsioideae. In D. H. Benzing [ed.], Bromeliaceae: profile of an adaptive radiation, 555–572. Cambridge University Press, Cambridge, UK. Till, W. 2000b. Tillandsia and Racinaea. In D. H. Benzing [ed.], Bromeliaceae: profile of an adaptive radiation, 573–586. Cambridge University Press, Cambridge, UK. Vaezi, J. & Brouillet, L. 2009. Phylogenetic relationships among diploid species of Symphyotrichum (Asteraceae: Astereae) based on two nuclear markers, ITS and GAPDH. Molec. Phylog. Evol. 51: 540–553. Versieux, L.M., Barbará, T., Wanderley, M.d.G.L., Calvente, A., Fay, M.F. & Lexer, C. 2012. Molecular phylogenetics of the Brazilian giant bromeliads (Alcantarea, Bromeliaceae): implications for morphological evolution and biogeography. Molec. Phylog. Evol. 64: 177– 189. White, T.J., Bruns, T., Lee, S. & Taylor, J. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics, Pp. 315–322 in Innis, M.A., Gelfand, D.H., Sninsky, J.J. & White, T.J. (eds.), PCR protocols: A guide to methods and applications. San Diego: Academic Press. Willyard, A., Cronn. R. & Liston, A. 2009. Reticulate evolution and incomplete lineage sorting among the ponderosa pines. Molec. Phylog. Evol. 52: 498–511. Yockteng, R. & Nadot, S. 2004. Phylogenetic relationships among Passiflora species based on the glutamine synthetase nuclear gene expressed in chloroplast (ncpGS). Molec. Phylog. Evol. 31: 379–396. 26 Part 1 Application of plastid DNA markers in phylogenetic reconstructions of Bromeliaceae Chapter 1 Phylogeny, adaptive radiation, and historical biogeography in Bromeliaceae: Insights from an eight-locus plastid phylogeny Thomas J. Givnish, Michael H.J. Barfuss, Benjamin Van Ee, Ricarda Riina, Katharina Schulte, Ralf Horres, Philip A. Gonsiska, Rachel S. Jabaily, Darren M. Crayn, J. Andrew C. Smith, Klaus Winter, Gregory K. Brown, Timothy M. Evans, Bruce K. Holst, Harry Luther, Walter Till, Georg Zizka, Paul E. Berry & Kenneth J. Sytsma Status: published, American Journal of Botany 98 (5): 872–895. 2011. Contribution: data collection, preliminary data analyses, manuscript writing/editing. 31 BARFUSS, M.H.J. 32 BROMELIAD PHYLOGENY AND EVOLUTION PART 1, CHAPTER 1 33 BARFUSS, M.H.J. 34 BROMELIAD PHYLOGENY AND EVOLUTION PART 1, CHAPTER 1 35 BARFUSS, M.H.J. 36 BROMELIAD PHYLOGENY AND EVOLUTION PART 1, CHAPTER 1 37 BARFUSS, M.H.J. 38 BROMELIAD PHYLOGENY AND EVOLUTION PART 1, CHAPTER 1 39 BARFUSS, M.H.J. 40 BROMELIAD PHYLOGENY AND EVOLUTION PART 1, CHAPTER 1 41 BARFUSS, M.H.J. 42 BROMELIAD PHYLOGENY AND EVOLUTION PART 1, CHAPTER 1 43 BARFUSS, M.H.J. 44 BROMELIAD PHYLOGENY AND EVOLUTION PART 1, CHAPTER 1 45 BARFUSS, M.H.J. 46 BROMELIAD PHYLOGENY AND EVOLUTION PART 1, CHAPTER 1 47 BARFUSS, M.H.J. 48 BROMELIAD PHYLOGENY AND EVOLUTION PART 1, CHAPTER 1 49 BARFUSS, M.H.J. 50 BROMELIAD PHYLOGENY AND EVOLUTION PART 1, CHAPTER 1 51 BARFUSS, M.H.J. 52 BROMELIAD PHYLOGENY AND EVOLUTION PART 1, CHAPTER 1 53 54 Part 2 Optimization of nuclear DNA markers and their application in phylogenetic reconstructions of subfamilies Bromelioideae and Tillandsioideae (Bromeliaceae) Chapter 2 Phylogeny of Bromelioideae (Bromeliaceae) inferred from nuclear and plastid DNA loci reveals the evolution of the tank habit within the subfamily Katharina Schulte, Michael H.J. Barfuss & Georg Zizka Status: published, Molecular Phylogenetics and Evolution 51 (2): 327–339. 2009. Contribution: primer design, data collection, preliminary data analyses, manuscript writing/editing. 59 BARFUSS, M.H.J. 60 PHYLOGENY OF BROMELIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 2 61 BARFUSS, M.H.J. 62 PHYLOGENY OF BROMELIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 2 63 BARFUSS, M.H.J. 64 PHYLOGENY OF BROMELIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 2 65 BARFUSS, M.H.J. 66 PHYLOGENY OF BROMELIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 2 67 BARFUSS, M.H.J. 68 PHYLOGENY OF BROMELIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 2 69 BARFUSS, M.H.J. 70 PHYLOGENY OF BROMELIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 2 71 Chapter 3 Optimizing eight nuclear DNA markers for phylogenetic studies in recently diverged angiosperms: A case study in Bromeliaceae subfamily Tillandsioideae Michael H.J. Barfuss 1, Rosabelle Samuel 1 & Félix Forest 2 1 University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria 2 Royal Botanic Gardens, Kew, Jodrell Laboratory, Richmond, Surrey, TW9 3DS, U.K. Author for correspondence: Michael H.J. Barfuss, michael.h.j.barfuss@univie.ac.at Keywords: glucose-6-phosphate isomerase, cytosolic; malate synthase; low-copy nuclear DNA; nitrate reductase 1, [NADH]; ITS nuclear ribosomal DNA; phosphoribulokinase; phytochrome C; RNA polymerase II, beta subunit; xanthine dehydrogenase Running Title: Nuclear DNA markers in Tillandsioideae (Bromeliaceae) Status: in preparation, intended to be submitted to American Journal of Botany Contribution: data collection, data analyses, manuscript writing Abstract Comparative studies of nuclear DNA sequences within Tillandsioideae show that several nuclear markers are able to provide more information and a higher degree of resolution in phylogenetic trees than previously used plastid markers. However, their utility depends not only on sequence variability, but also on methodological challenges, especially when traditional Sanger-sequencing is used. The presence of strong secondary structures in ITS nrDNA creates problems in amplification and sequencing. Therefore it is not recommended as a suitable marker for phylogenetic investigations of Bromeliaceae, also because of low resolution at deeper nodes. However, ITS nrDNA regions might be useful for delimiting species. Amplified fragments of the genes malate synthase (MS) and RNA polymerase II, beta subunit (RPB2) are not helpful due to their relatively short length and restricted number of PICs. Glucose-6-phosphate isomerase (PGIC) needs to be further investigated, but sequence length and variation found in the two species sequenced suggests it will be suitable at generic and higher taxonomic levels. Likewise, nitrate reductase 1, [NADH] (NIA) and xanthine dehydrogenase (XDH) need further investigations; the first might be suitable for lower taxonomical levels, while the second would be better suited for higher levels. Phosphoribulokinase (PRK) seems to be one of the most promising markers, but length differences between bromeliad taxa and indels within alleles of heterozygous accessions made it the most difficult to edit and align and thus time-consuming. For phytochrome C (PHYC) generating sequences and the editing/alignment have been straight forward. PRK and PHYC are useful nuclear markers and able to provide considerable resolution in phylogenetic trees, though some relationships remain uncertain. 75 Introduction Despite rapid technical progress in molecular systematics, the goal of reconstructing fully resolved and well-supported phylogenetic trees in many species-rich and/or recently diverged angiosperm groups remains challenging (e.g., Lukas, 2010; Safer, 2011). Many molecular studies still analyze only plastid DNA markers or the multi-copy internal transcribed spacer nuclear ribosomal DNA (ITS nrDNA) of the 18S-5.8S-26S nrDNA repeat unit (e.g., Gruenstaudl & al., 2009; Russell & al., 2010). Availability of universal primers for recently characterized plastid markers (Shaw & al., 2005; 2007) and the ITS nrDNA region (White & al., 1990; Sun & al., 1994; Blattner, 1999; Gruenstaeudl & al., 2009) as well as the most advantageous single-copy nature of plastid DNA and the usually well-homogenized ITS nrDNA (concerted evolution) made them more suitable for phylogenetic studies. Usually these markers are simple to amplify, directly sequence, and align across lower taxonomic levels. Also, design of new primers became much easier due to the steadily increasing number of available sequences and complete plastid genomes publically available. However, a major challenge of commonly used plastid DNA and nrDNA markers is resolution of phylogenetic relationships at all taxonomic levels, even when many markers are combined. A common strategy to solve these problems is the use of lowcopy nuclear DNA (e.g., Strand & al., 1997; Sang, 2002; Chapman 2007; Álvarez & al., 2008), but these markers are often avoided for several reasons. Inheritance of plastid DNA is predominantly uniparental and maternal in plants, whereas nuclear DNA is biparentally inherited, although due to conversion nrDNA can become uniparental. Processes like incomplete lineage sorting, hybridization, coalescence times longer than species ages, and introgression may complicate interpretation of results (Piñeiro & al., 2009; Willyard & al., 2009). More difficulties arise when polyploid taxa or multi-copy nuclear genes and gene families are investigated (e.g., Sang & al., 2004). Analyzing and interpreting such data can be problematic because multiple alleles and paralogs are present. However, for particular evolutionary questions and certain plant groups with low variation in plastid DNA or nrDNA sequences, more rapidly evolving lowcopy nuclear DNA markers are a promising solution. An increasing number of studies published over the recent years have utilized such sequences to investigate phylogenetic relationships (e.g., Emshwiller & Doyle, 1999; Cronn & al., 2002; Yockteng & Nadot, 2004; Vaezi & Brouillet, 2009), reticulate evolution (e.g., Ma & al., 2010; Russell & al., 2010) and hybrid speciation (e.g., Sang & Zhang, 1999; Tank & Sang, 2001; Clarkson & al., 2010), or to uncover genome evolution (e.g., Wendel, 2000). The pineapple family (Bromeliaceae), with more than 3,100 species assigned to 58 genera, represents a large and ecologically important group of monocots for which infrafamilial classification and phylogenetic relationships have been problematic for many years. Several earlier studies have tried to create well-resolved trees using one or a few plastid markers with limited success (Terry & al., 1997a, b; Horres & al., 2000; Crayn & al., 2004). Accompanying studies focusing on the largest subfamilies Tillandsioideae (Barfuss & al., 2005) and Bromelioideae (Schulte & Zizka, 2008) as well as Bromeliaceae in general (Givnish & al., 2011), both including a wider selection of taxa and plastid markers, were able to resolve the main phylogenetic units within the family but still showed low support for some deeper nodes and comparatively little sequence divergence between genera and species. Efforts to sequence and combine many plastid spacer regions that are supposed to evolve more rapidly (Shaw & al., 2005, 2007) than previously sequenced markers for Bromeliaceae (Barfuss & al., 2005; Schulte & Zizka, 2008; 77 BARFUSS, M.H.J. NUCLEAR DNA MARKERS Givnish & al., 2011: atpB-rbcL, trnK-matK, ndhF, psbA-trnH, rpL32-trnL, rps16 intron, and 5'trnL-3'trnL-trnF) are of limited success in gathering sufficient phylogenetic information at lower taxonomic levels (M. H. J. Barfuss, unpublished data: atpI-atpH, psbD-trnT, psbJ-petA, rpoB-trnC, trnD-trnT, 3'trnV-ndhC,and trnQ-5'rps16). Low sequence divergence associated with plastid DNA sequences in phylogenetic studies of bromeliads were initially mainly attributed to generally low substitution rates in the bromeliad plastid genome (e.g., Gaut & al., 1992; Givnish & al., 2004, 2007). Insights from the recently published eight-locus plastid analysis of Bromeliaceae have shown that the family itself is old (100 My), but extant lineages of bromeliads arose less than 20 My ago (Givnish & al., 2011). The most species-rich clades are less than 5 My old. These results support the hypothesis that extant lineages have undergone recent and rapid radiations and subsequent diversification in newly formed ecological niches and not that substitution rates are lower than in other angiosperm lineages. The first molecular studies of bromeliads that include low-copy nuclear DNA sequences were recently published by Schulte & al. (2009) for Bromelioideae (phosphoribulokinase, PRK), Chew & al. (2010) for Tillandsia subgenus Tillandsia (internal transcribed spacer 2, ITS 2, external transcribed spacer, ETS nrDNA), Sass & Specht (2010) for Aechmea (RNA polymerase II, beta subunit, RPB2, glyceraldehyde-3-phosphate dehydrogenase, G3PDH, ETS nrDNA), Jabaily & Sytsma (2010) for Puya (phytochrome C, PHYC), and Versieux & al. (2012) for Alcantarea (FLORICAULA/LEAFY, FLO/LFY). However, relationships within Bromeliaceae still remain inadequately understood because these markers were applied only to restricted sets of taxa. In this study we aim (1) to investigate primers for several nuclear DNA markers of angiosperms with focus on monocots and Bromeliaceae especially; (2) to resolve lower taxonomic units of subfamily Tillandsioideae down to species groups; (3) to evaluate relationships of recently diverged species; and (4) to detect hybrids, possible hybrid speciation, and reticulate evolution. We also selected the ITS nrDNA since the complete region was never used in a phylogenetic study for a large set of samples above the subgeneric level. Materials and methods Selection of taxa For initial primer trials, three sets of few individuals of Tillandsioideae (2, 8, or 13) were selected. Larger sets of samples (64, 72, or 114) with a minimum of two accessions per major clade in other analyses were chosen based on analyses of seven plastid markers (Barfuss & al., 2005) to demonstrate variability of lineages. The complete set of samples (444) was finally analyzed for the most promising nuclear DNA markers. A sample list including botanical authors is given in Supplementary Data of chapter 5; accession details can be obtained from the first author upon request. Extraction of high-quality DNA Initially the CTAB procedure established by Doyle and Doyle (1987) modified for minipreps (1.5 or 2 mL tubes) was used. Due to the high content of polysaccharides, polyphenols, and other secondary metabolites in more than ⅔ of the Tillandsioideae species studied, this method did not yield sufficient amounts of high quality DNA, which is necessary to amplify suitable quantities of low-copy nuclear DNA. Therefore total genomic DNA was extracted from fresh or silicagel dried leaf tissue using the altered CTAB procedure of Tel-Zur & al. (1999) with further 78 IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 reagent modifications, and with adjustments within the protocol, mainly implemented to allow the samples to be processed as minipreps. Most important chemical changes from the original protocol include (1) the addition of 1% PVP-40 into the Sorbitol buffer, (2) the increase of the CTAB concentration to 3% and the NaCl concentration to 3 M in the high-salt CTAB buffer, and (3) the addition of PVP-40 into the extraction buffer. In most cases this method yielded superior quantities of high molecular weight DNA which could be used immediately for PCR. When good DNA was visualized on 1% TAE agarose gels, but initial PCR trials were not successful, DNA was further purified or newly extracted using a modified CTAB/DNeasy Plant Mini Kit (QIAGEN) procedure. Protocols have been developed either starting from DNA dissolved in TE buffer or after CTAB extraction before (and therefore not doing a) precipitation with isopropanol, to avoid precipitating of any unwanted contaminants (mainly polyphenolic compounds) which could inhibit Taq polymerase activity. This combined method yielded much better DNA extracts than using the QIAGEN kit alone. Detailed protocols of all modifications of the extraction procedures can be obtained from the first author upon request. Selection of nuclear markers The ITS nrDNA region (e.g., White & al., 1990; Sun & al., 1994; Gruenstaeudl & al., 2009; Chew & al., 2010) and seven low-copy nuclear DNA markers were selected and tested: malate synthase (MS; e.g., Lewis & Doyle, 2001, 2002), RNA polymerase II, beta subunit (RPB2; e.g., Denton & al., 1998; Oxelman & al., 2004; Thomas & al., 2006; Sass & Specht, 2010), glucose-6phosphate isomerase, cytosolic (G6PIC = phosphoglucose isomerase C, PGIC; Terauchi & al., 1997; Ford & al., 2006), nitrate reductase 1, [NADH] (NR1 = nitrate reductase apoenzyme, NIA : e.g., Howarth & Baum, 2002), xanthine dehydrogenase (XDH; Górniak & al., 2010; Morton, 2011), phosphoribulokinase (PRK; e.g., Lewis & Doyle, 2002; Thomas & al., 2006; Schulte & al., 2009), and phytochrome C (PHYC; e.g., Mathews & Donoghue, 1999; Samuel & al., 2005; Russell & al., 2010). Regions were selected based on the availability of primers and/or DNA sequences available in GenBank to be able to design primers for Bromeliaceae. The primary target was to obtain one or more genomic markers that are either homogenized (ITS nrDNA) or can be treated as effectively single-copy (other nuclear markers) to be sequenced directly. Other parameters for the selection of markers were (1) performance of primers and reagents for both amplification (e.g., primer annealing, primer-dimer formation) and sequencing (e.g., homogeneity), (2) occurrence of homopolymers and microsatellites, which can give problems in sequencing, (3) amplicon length, and (4) number of potential phylogenetically informative characters (PICs). The genetic structure of each low-copy nuclear DNA region was determined from the assembled RefSeq genome of Oryza sativa and is summarized in Table 1. Positions of primers selected finally are given in Figure 1, A–H. All sequences generated for this study will be deposited in GenBank (http://www.ncbi.nlm.nih.gov/genbank). Nomenclature of primers Positions of primers on chromosomes and within the coding sequence (CDS) of the genes were determined by BLAST searches, both within the entire database and the annotated genes of the assembled RefSeq genome of Oryza sativa (Tables 3–10). This was possible because all primers are located in exons at least for ¾ of their length. Names of previously published primers were taken from their primary publication, and those of newly generated or modified ones were derived from the CDS position of the last primer base at the 3' end. If a modified primer ended at the same position as the original one, a number after a hyphen was attached. 79 BARFUSS, M.H.J. NUCLEAR DNA MARKERS Letter extensions after a hyphen indicate primers specific to a particular plant group (an = angiosperms, mo = monocots, br = bromeliads). Table 1. Attributes of the low-copy nuclear genes based on GenBank information from the assembled RefSeq genome of Oryza sativa. Amplified regions are highlighted in red. Annotation of malate synthase and RNA polymerase II, beta subunit was incomplete or wrong and therefore done manually aided with reference sequences of Arabidopsis thaliana (RPB2) or Zea mays (MS). exon 1 intron 1 exon 2 intron 2 exon 3 intron 3 exon 4 intron 4 exon 5 intron 5 exon 6 intron 6 exon 7 intron 7 exon 8 intron 8 exon 9 intron 9 exon 10 intron 10 exon 11 intron 11 exon 12 intron 12 exon 13 intron 13 exon 14 intron 14 exon 15 intron 15 exon 16 intron 16 exon 17 intron 17 exon 18 intron 18 exon 19 intron 19 exon 20 intron 20 exon 21 intron 21 exon 22 intron 22 exon 23 intron 23 exon 24 intron 24 exon 25 80 malate synthase (MS) RNA polymerase II, beta subunit (RPB2) glucose-6-phosphate isomerase, cytosolic (G6PIC, PGIC) Chromosome: 4 (NC_008397) Gene: MS (Os04g0486950) mRNA: NM_001187064 Total range: 24,724,583-24,726,722 total length: 2131 (complement) Processed length: 1704 Protein product length: 567 Copy number: 1 Chromosome: 3 (NC_008396) Gene: RPB2 (Os03g0646800) mRNA: NM_001057305 Total range: 25,794,125-25,801,820 Total length: 7,696 Processed length: 3,555 Protein product length: 1184 Copy number: 1 Chromosome: 6 (NC_008399) Gene: PGIC-b (Os06g0256500) mRNA: NM_001063851 Total range: 8,151,838-8,158,107 Total length: 6,270 Processed length: 1,704 Protein product length: 567 Copy number: 2 Duplicate gene: PGIC-a (Os03g0776000) Incomplete annotation in GenBank! Incomplete annotation in GenBank! 1-375 376-459 460-785 786-1048 1049-1379 1380-1468 1469-2140 1-234 235-1792 1793-2014 2015-2085 2086-2376 2377-2582 2583-2780 2781-2907 2908-2972 2973-3054 3055-3141 3142-3254 3255-3375 3376-3466 3467-3511 3512-3587 3588-3725 3726-3837 3838-3910 3911-3989 3990-4066 4067-4148 4149-4256 4257-4374 4375-4491 4492-4790 4791-4946 4947-5046 5047-5148 5149-5302 5303-5428 5429-5503 5504-5581 5582-5656 5657-5752 5753-5858 5859-5974 5975-6047 6048-6117 6118-6196 6197-6250 6251-6328 6329-6472 6473-6563 6564-6863 6864-7046 7047-7469 7470-7582 7583-7696 375 84 326 263 331 89 672 234 1558 222 71 291 206 198 127 65 82 87 113 121 91 45 76 138 112 73 79 77 82 108 118 117 299 156 100 102 154 126 75 78 75 96 106 116 73 70 79 54 78 144 91 300 183 423 113 114 1-51 52-205 206-282 283-375 358-439 440-512 513-560 561-862 863-1018 1019-1093 1094-1190 1191-1288 1289-1331 1332-1721 1722-1789 1790-1879 1880-1964 1965-2044 2045-2093 2094-2314 2315-2387 2388-2465 2466-2526 2527-3460 3461-3527 3528-3596 3597-3653 3654-3746 3747-3800 3801-3883 3884-4012 4013-4538 4539-4602 4603-4712 4713-4786 4787-4869 4870-4923 4924-5042 5043-5111 5112-5740 5741-5830 5831-6002 6003-6110 6111-6222 6223-6270 51 154 77 75 82 73 48 302 156 75 97 98 43 390 68 90 85 80 49 221 73 78 61 934 67 69 57 93 54 83 129 526 64 110 74 83 54 119 69 629 90 172 108 112 48 IN TILLANDSIOIDEAE (BROMELIACEAE) exon 1 intron 1 exon 2 intron 2 exon 3 intron 3 exon 4 intron 4 exon 5 intron 5 exon 6 intron 6 exon 7 intron 7 exon 8 intron 8 exon 9 intron 9 exon 10 intron 10 exon 11 intron 11 exon 12 intron 12 exon 13 intron 13 exon 14 intron 14 exon 15 intron 15 exon 16 intron 16 exon 17 intron 17 exon 18 intron 18 exon 19 intron 19 exon 20 intron 20 exon 21 intron 21 exon 22 intron 22 exon 23 intron 23 exon 24 intron 24 exon 25 PART 2, CHAPTER 3 nitrate reductase 1, [NADH] (NR1, NIA) xanthine dehydrogenase (XDH) phosphoribulokinase (PRK) phytochrome C (PHYC) Chromosome: 8 (NC_008401) Gene: NIA-a (Os08g0468100) mRNA: NM_001068541 Total range: 23,121,355-23,126,236 Total length: 4,882 Processed length: 2,751 Protein product length: 916 Copy number: 3 Duplicate genes: NIA-b (Os02g0770800) NIA-c (Os08g0468700) Chromosome: 3 (NC_008396) Gene: XDH (Os03g0429800) mRNA: NM_001056955 Total range: 18,621,710-18,634,104 Total length: 12,395 Processed length: 4,110 Protein product length: 1,369 Copy number: 1 Chromosome: 2 (NC_008395) Gene: PRK-a (Os02g0698000) mRNA: NM_001054360 Total range: 29,588,876-29,590,869 Total length: 1,994 Processed length: 1,212 Protein product length: 403 Copy number: 2 Duplicate genes: PRK-b (Os04g0595700) complement: Os02g0697900 (partial) Chromosome: 3 (NC_008396) Gene: PHYC (Os03g0752100) mRNA: NM_001057831 Total range: 31,768,179-31,772,647 Total length: 4,469 Processed length: 3,414 Protein product length: 1,137 Copy number: 1 Similar genes: PHYA (Os03g0719800) PHYB (Os03g0309200) 1-1039 1040-1147 1148-1288 1289-3226 3227-3459 3460-3544 3545-4882 1-142 143-1269 1270-1369 1370-1846 1847-1955 1956-2595 2596-4098 4099-4938 4939-5181 5182-5291 5292-5606 5607-6859 6860-7099 8000-7435 7436-7720 7721-7795 7796-8017 8018-8565 8566-8727 8728-8871 8872-9006 9007-10912 10913-11149 11150-11247 11248-11421 11422-12152 12153-12395 1-545 546-838 839-923 924-1012 1013-1097 1098-1267 1268-1512 1513-1742 1743-1994 1-2065 2066-2169 2170-2986 2987-3379 3380-3673 3674-4231 4232-4469 1039 108 141 1938 233 84 1338 142 1127 100 477 109 640 1503 840 243 110 315 1253 240 336 285 75 222 548 162 144 135 1906 237 98 174 731 243 545 293 85 89 85 170 245 230 252 2065 104 817 393 294 558 238 81 BARFUSS, M.H.J. NUCLEAR DNA MARKERS Figure 1. Genetic structure of each nuclear DNA region based on GenBank information from the assembled RefSeq genome of Oryza sativa. Approximate positions of primers finally used are indicated by arrow heads. Annotation of malate synthase and RNA polymerase II, beta subunit was incomplete or wrong and therefore done manually aided with reference sequences of Arabidopsis thaliana (RPB2) or Zea mays (MS). 82 IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 83 BARFUSS, M.H.J. 84 Table 2. Amplified markers, number of taxa studied, and PCR conditions of each marker. marker no. of taxa studied PCR conditions ITS nrDNA 114 1× 97°C for 2 min; 15× 99°C for 20 s, 64°C for 15 s, 72°C for 2 min; 20× 99°C for 15 s, 64°C for 10 s, 72°C for 2 min + 5 s/cycle; 1× 72°C for 7 min; 1× 4°C for ∞. MS 13 1× 95°C for 2 min; 5× 95°C for 30 s, 60°C for 30 s, 72°C for 2 min; 30× 95°C for 30 s, 54°C for 30 s, 72°C for 2 min; 1× 72°C for 7 min; 1× 4°C for ∞. RPB2 8 1× 95°C for 2 min; 40× 95°C for 30 s, 55°C for 30 s, 72°C for 1 min; 1× 72°C for 7 min; 1× 4°C for ∞. PGIC 2 1× 95°C for 2 min; 35× 95°C for 30 s, 50°C for 30 s, 70°C for 2 min; 1× 70°C for 7 min; 1× 4°C for ∞. NIA 71 1× 95°C for 2 min; 10× 95°C for 30 s, 65°C - 1°C/cycle for 30 s, 70°C for 2 min; 30× 95°C for 30 s, 58°C for 30 s, 70°C for 2 min; 1× 70°C for 7 min; 1× 4°C for ∞. XDH 65 1× 95°C for 2 min; 10× 95°C for 30 s, 65°C - 1°C/cycle for 30 s, 70°C for 2 min; 30× 95°C for 30 s, 58°C for 30 s, 70°C for 2 min; 1× 70°C for 7 min; 1× 4°C for ∞. PRK 444 1× 95°C for 2 min; 15× 95°C for 30 s, 62°C for 30 s, 70°C for 1 min 30 s; 20× 95°C for 30 s, 62°C for 30 s, 70°C for 1 min 30 s + 5 s/cycle; 1× 70°C for 7 min; 1× 4°C for ∞. PHYC 444 1× 95°C for 2 min; 20× 95°C for 30 s, 59°C for 30 s, 70°C for 1 min 30 s; 20× 95°C for 30 s, 59°C for 30 s, 70°C for 1 min 30 s + 5 s/cycle; 1× 70°C for 7 min; 1× 4°C for ∞. NUCLEAR DNA MARKERS name sequence (5'–3') position length direction use/success specificity reference primers initially aivailable 17SE (AB101) ACGAATTCATGGTCCGGTGAAGTGTTCG 18S nrDNA, 3' end 28 bp forward PCR unsuccessful angiosperms Sun & al., 1994 ITS-A GGAAGGAGAAGTCGTAACAAGG 18S nrDNA, 3' end 22 bp forward primer design angiosperms Blattner, 1999 ITS5 GGAAGTAAAAGTCGTAACAAGG 18S nrDNA, 3' end 22 bp forward PCR unsuccessful universal White & al., 1990 ITS1 TCCGTAGGTGAACCTGCGG 18S nrDNA, 3' end 19 bp forward not used universal White & al., 1990 ITS-D CTCTCGGCAACGGATATCTCG 5.8S nrDNA, 5' end 21 bp forward primer design angiosperms Blattner, 1999 ITS3 GCATCGATGAAGAACGCAGC 5.8S nrDNA 20 bp forward not used universal White & al., 1990 ITS2 GCTGCGTTCTTCATCGATGC 5.8S nrDNA 20 bp reverse PCR unsuccessful universal White & al., 1990 ITS-C GCAATTCACACCAAGTATCGC 5.8S nrDNA 21 bp reverse PCR unsuccessful angiosperms Blattner, 1999 ITS-B CTTTTCCTCCGCTTATTGATATG 26S nrDNA, 5' end 23 bp reverse not used angiosperms Blattner, 1999 ITS4 TCCTCCGCTTATTGATATGC 26S nrDNA, 5' end 20 bp reverse PCR unsuccessful universal White & al., 1990 26SE (AB102) TAGAATTCCCCGGTTCGCTCGCCGTTAC 26S nrDNA, 5' end 28 bp reverse PCR unsuccessful angiosperms Sun & al., 1994 IN TILLANDSIOIDEAE (BROMELIACEAE) Table 3. Primers tested for the internal transcribed spacer of nuclear ribosomal DNA (ITS nrDNA). Highlighted primers performed best and were selected for PCR or sequencing. bp = base pairs. primers subsequently tested ITS18Sf ACCGATTGAATGGTCCGGTGAAGTGTTCG 18S nrDNA, 3' end 29 bp forward PCR angiosperms Gruenstaeudl & al., 2009 ITS18Scsf GAATGGTCCGGTGAAGTGTTCG 18S nrDNA, 3' end 22 bp forward sequencing angiosperms this study ITS5c AGAGGAAGGAGAAGTCGTAACAAGGT 18S nrDNA, 3' end 26 bp forward sequencing angiosperms this study; modified from “ITS5” & “ITS-A” ITS5.8Sf ACTCTCGGCAACGGATATCTCGGCTC 5.8S nrDNA, 5' end 26 bp forward PCR angiosperms Gruenstaeudl & al., 2009 ITS5.8Scsf GACTCTCGGCAACGGATATCTCG 5.8S nrDNA, 5' end 23 bp forward sequencing angiosperms this study; modified from “ITS-D” ITS-E CGGCAACGGATATCTCGGCTC 5.8S nrDNA, 5' end 21 bp forward sequencing angiosperms Blattner 1999 ITS5.8Scsr GATGCGTGACGCCCAGGCAG 5.8S nrDNA, 3' end 20 bp reverse sequencing angiosperms this study ITS5.8Sr ATGCGTGACGCCCAGGCAGACGTG 5.8S nrDNA, 3' end 24 bp reverse PCR angiosperms Gruenstaeudl & al., 2009 ATGCGTGACGCCCAGGCAGRCGTG 5.8S nrDNA, 3' end 24 bp reverse PCR angiosperms this study; modified from “ITS5.8Sr” TGACGCCCAGGCAGRCGTGC 5.8S nrDNA, 3' end 20 bp reverse sequencing angiosperms this study 26SE-2 CGGTTCGCTCGCCGTTACTA 26S nrDNA, 5' end 20 bp reverse sequencing angiosperms this study ITS26Scsr GGACGCTTCTCCAGACTACAATTCG 26S nrDNA, 5' end 25 bp reverse sequencing angiosperms this study ITS26Sr CTGAGGACGCTTCTCCAGACTACAATTCG 26S nrDNA, 5' end 29 bp reverse PCR angiosperms Gruenstaeudl & al., 2009 85 PART 2, CHAPTER 3 ITS5.8Sr-an ITS5.8Scsr-2 BARFUSS, M.H.J. 86 Table 4. Primers tested for malate synthase (MS). Only highlighted primers worked and were selected for PCR and sequencing. Position of primers was determined from the assembled RefSeq genome of Oryza sativa. bp = base pairs. CDS = coding sequence. Chr = chromosome. name sequence (5'–3') CDS position (bp) Chr position (bp) length direction use/success specificity reference primers initially tested ms400f GGAAGATGRTCATCAAYGCNCTYAAYTC exon 1, 329–356 24,726,367–24,726,394 28 bp forward unsuccessful angiosperms Lewis & Doyle 2001 ms356f GGAAGATGRTMATCAAYGCRCTKAAYTC exon 1, 329–356 24,726,367–24,726,394 28 bp forward unsuccessful angiosperms D. Springate, RBG Kew, unpublished; modified from “ms400f” ms526f GGACTATAAGCTTCCATGACCTC exon 2, 455–477 24,726,162–24,726,184 23 bp forward unsuccessful palms, some other monocots Lewis & Doyle 2001 ms943r GTCTTNACRTAGCTGAADATRTARTCCC exon 3, 872–899 24,725,504–24,725,477 28 bp reverse unsuccessful angiosperms Lewis & Doyle 2001 ms1408r CCARTTCTGVACBCKGCTGATCTCCGC exon 4, 1,408–1,434 24,724,853–24,724,879 27 bp reverse unsuccessful angiosperms D. Springate, RBG Kew, unpublished ms1488r TTCRTAYYTNAKCCAYTGCCAGTTYTG exon 4, 1,426–1,452 24,724,835–24,724,861 27 bp reverse unsuccessful angiosperms Lewis & Doyle 2001 primers subsequently tested GGAAGATGRTCATCAAYGCNCTBAAYTC exon 1, 329–356 24,726,367–24,726,394 28 bp forward unsuccessful monocots this study; modified from “ms400f” ms428f (MSe2f) GCTGGGARAACCTGATGARVGGCCA exon 2, 404–428 24,726,211–24,726,235 25 bp forward PCR/sequencing monocots this study ms553r (MSe2r) CCGTCGAYGAGRATRTGGGCCTC exon 2, 553–575 24,726,064–24,726,086 23 bp reverse unsuccessful monocots this study ms837f (MSe3f) CARATGRAYGAGATVCTNTACGAGCTG exon 3, 811–837 24,725,539–24,725,565 27 bp forward unsuccessful monocots this study ms960r (MSe3r) GGARRTCGGAGTAGCTSCKCATGAAG exon 3, 960–985 24,725,391–24,725,416 26 bp reverse PCR/sequencing monocots this study ms1411r (MSe4r) TGCCAGTTCTGVACBCKGCTGATCTC exon 4, 1,411–1,436 24,724,851–24,724,876 26 bp reverse unsuccessful monocots this study; modified from “ms1488r” NUCLEAR DNA MARKERS ms356f-2 name sequence (5'–3') CDS position (bp) Chr position (bp) length direction use/success specificity reference P6F TGGGGAATGATGTGTCCTGC exon 11, 1,513–1,532 25,798,152–25,798,171 20 bp forward unsuccessful angiosperms Denton & al., 1998 P7F CCYCGTAATACWTAYCARTCWGC exon 17, 2,164–2,186 25,799,631–25,799,653 23 bp forward unsuccessful angiosperms Denton & al., 1998 P7R CCCATGGCTTGCTTCCCCAT exon 17 2,188–2,207 25,799,655–25,799,674 20 bp reverse unsuccessful angiosperms Denton & al., 1998 P10F CARGARGATATGCCATGGAC exon 23, 2,851–2,870 25,800,820–25,800,839 20 bp forward PCR/sequencing angiosperms Denton & al., 1998 P10R CCCATRATACACTCAATGAGYTG exon 23, 2,938–2,960 25,800,907–25,800,929 23 bp reverse unsuccessful angiosperms Denton & al., 1998 P11aR GTGAATCTTGTCATCMACCATATGC exon 24, 3,153–3,177 25,801,305–25,801,329 25 bp reverse PCR/sequencing angiosperms Denton & al., 1998 IN TILLANDSIOIDEAE (BROMELIACEAE) Table 5. Primers tested for RNA polymerase II, beta subunit (RPB2). Only highlighted primers worked and were selected for PCR and sequencing. Position of primers was determined from the assembled RefSeq genome of Oryza sativa. bp = base pairs. CDS = coding sequence. Chr = chromosome. Table 6. Primers tested for glucose-6-phosphate isomerase, cytosolic (PGIC). Only highlighted primers worked and were used for PCR and sequencing. Position of primers was determined from the assembled RefSeq genome of Oryza sativa. bp = base pairs. CDS = coding sequence. Chr = chromosome. name sequence (5'–3') CDS position (bp) Chr position (bp) length direction use/success specificity reference AA11F (93.4) TTYGCNTTYTGGGAYTGGGT exon 11, 793–812 8,154,188–8,154,207 20 bp forward PCR/sequencing universal Ford & al., 2006 (Terauchi & al., 1997) AA16F ATGGARAGYAAYGGNAARGG exon 16, 1,075–1,094 8,155,727–8,155,746 20 bp forward unsuccessful angiosperms Ford & al., 2006 AA16R CCYTTNCCRTTRCTYTCCAT exon 16, 1,075–1,094 8,155,727–8,155,746 20 bp reverse PCR/sequencing angiosperms Ford & al., 2006 CCCCAYTGRTCRAAIGARTTDATICCCCA exon 21, 1,504–1,532 8,157,623–8,157,651 29 bp reverse unsuccessful angiosperms Ford & al., 2006 CCCCAYTGRTCRAANGARTTDATNCCCCA exon 21, 1,504–1,532 8,157,623–8,157,651 29 bp reverse unsuccessful angiosperms this study; modified from “AA21RM” 93.9H TCIACICCCCAITGRTCTAAIGARTTIAT exon 21, 1,510–1,538 8,157,629–8,157,657 29 bp reverse unsuccessful universal Terauchi & al., 1997 yamv TCIACICCCCAITGRTCAAAIGARTTIAT exon 21, 1,510–1,538 8,157,629–8,157,657 29 bp reverse unsuccessful angiosperms Ford & al., 2006; modified from “93.9H” pgic1510r TCNACNCCCCANTGRTCAAANGARTTNAT exon 21, 1,510–1,538 8,157,629–8,157,657 29 bp reverse unsuccessful angiosperms this study; modified from “yamv” 87 PART 2, CHAPTER 3 AA21RM pgic1504r BARFUSS, M.H.J. 88 Table 7. Primers tested for nitrate reductase 1, [NADH] (NIA). Highlighted primers performed best and were selected for PCR and/or sequencing. Position of primers was determined from the assembled RefSeq genome of Oryza sativa. bp = base pairs. CDS = coding sequence. Chr = chromosome. name sequence (5'–3') CDS position (bp) Chr position (bp) length direction use/success specificity reference primers initially tested NIA2F TCBGTGATTACGACGCCGTGTCATGA exon 2, 1,087–1,112 23,122,549–23,122,574 26 bp forward unsuccessful angiosperms Howarth & Baum 2002 NIA2R GACCARAARCACCARCACCARTAYT exon 3, 1,283–1,307 23,124,683–23,124,707 25 bp reverse unsuccessful angiosperms Howarth & Baum 2002 NIA3F AARTAYTGGTGYTGGTGYTTYTGGTC exon 3, 1,282–1,307 23,124,682–23,124,707 26 bp forward PCR/sequencing angiosperms Howarth & Baum 2002 NIA3R GAACCARCARTTGTTCATCATDCC exon 4, 1,414–1,437 23,124,899–23,124,922 24 bp reverse PCR/sequencing angiosperms Howarth & Baum 2002 primers subsequently tested nia410f (nia1f) CTNATGCACCACGGNTTCATCAC exon 1, 388–410 23,121,742–23,121,764 23 bp forward PCR/sequencing angiosperms this study nia413f (nia1f-2) CTNATGCACCACGGNTTYATCACCCC exon 1, 388–413 23,121,742–23,121,767 26 bp forward PCR angiosperms this study nia1042r (nia2r-3) TGATKATRTAYTCYGGCTTRTACCACCA exon 2, 1,042–1,069 23,122,504–23,122,531 28 bp reverse PCR/sequencing angiosperms this study nia1073f (nia2f-2) GGTACAAGCCRGARTAYATMATYAACGA exon 2, 1,046–1,073 23,122,508–23,122,535 28 bp forward unsuccessful angiosperms this study nia1108r (nia2r-2) GTTGATGGGYARRATCTCSTCGTG exon 2, 1,108–1,131 23,122,570–23,122,593 24 bp reverse unsuccessful angiosperms this study nia1390r 1 (nia3r-2) ACCATGACGTTCCAGATGAGCTTCTC exon 3, 1,390–1,413 (-1,415) 23,124,790–23,124,815 26 bp reverse unsuccessful angiosperms this study 1 italicized numbers indicate base positions that partly lie within introns (GT motive of intron start). NUCLEAR DNA MARKERS note: name sequence (5'–3') CDS position (bp) Chr position (bp) length direction use/success specificity reference xdh422f (xdh4f-1) CCYGGTTTYRTBATGTCVATGTATGC exon 4, 397–422 18,624,350–18,624,375 26 bp forward PCR monocots this study xdh437f (xdh4f-2) GTCVATGTATGCVTTRYTRMGRTCAAG exon 4, 411–437 18,624,364–18,624,390 27 bp forward PCR monocots this study xdh479f (xdh4f-3) TGARGARCARATYGAAGAWWGCCTTGC exon 4, 453–479 18,624,406–18,624,432 27 bp forward PCR/sequencing monocots this study xdh974f (xdh4intf-2) GTTACCCATGTGGCGGAGCTTAATGC exon 4, 949–974 18,624,902–18,624,927 26 bp forward sequencing bromeliads this study xdh1033f-br (xdh4intf-1) GTGCTTCTGTGAGACTGACACAGCTCC exon 4, 1,007–1,033 18,624,960–18,624,986 27 bp forward sequencing bromeliads this study xdh1165r-br (xdh4intr-1) TAGGTCAGATATTGGACTAGCAGTACA exon 4, 1,165–1,191 18,625,118–18,625,144 27 bp reverse sequencing bromeliads this study xdh1169r-br (xdh4intr-2) GGGTTTAGGTCAGATATTGGACTAGCAG exon 4, 1,169–1,196 18,625,122–18,625,149 28 bp reverse sequencing bromeliads this study xdh1611r (xdh4r-1) CGRAAYTCHACCATYCCMCCWGGWGCA exon 4, 1,611–1,637 18,625,564–18,625,590 27 bp reverse PCR/sequencing monocots this study xdh1612r (xdh4r-2) CGRAAYTCHACCATYCCMCCWGGWGC exon 4, 1,612–1,637 18,625,565–18,625,590 26 bp reverse PCR monocots this study xdh1880f (xdh5f-1) GGTYACWGGDGARGCDGAATATRCTGA exon 5, 1,854–1,880 18,626,647–18,626,673 27 bp forward unsuccessful monocots this study xdh1883f (xdh5f-2) ACWGGDGARGCDGAATATRCTGAYGA exon 5, 1,858–1,883 18,626,651–18,626,676 26 bp forward unsuccessful monocots this study ACYTGDGTRGATGAWATCATRTGAAYTTC exon 6, 2,386–2,412 (–2,414) 18,627,289–18,627,317 29 bp reverse unsuccessful monocots this study ACYTGDGTRGATGAWATCATRTGAA exon 6, 2,390–2,412 (–2,414) 18,627,293–18,627,317 25 bp reverse unsuccessful monocots this study xdh2386r1 (xdh6r-1) 1 xdh2390r (xdh6r-2) note: 1 IN TILLANDSIOIDEAE (BROMELIACEAE) Table 8. Primers tested for xanthine dehydrogenase (XDH). Highlighted primers performed best and were selected for PCR and/or sequencing. Position of primers was determined from the assembled RefSeq genome of Oryza sativa. bp = base pairs. CDS = coding sequence. Chr = chromosome. italicized numbers indicate base positions that partly lie within introns (GT motive of intron start). PART 2, CHAPTER 3 89 name sequence (5'–3') CDS position (bp) Chr position (bp) length direction use/success specificity AAYGAYTTTGAYCTYATGTATGARCARGT exon 1, 394–422 29,589,269–29,589,297 29 bp forward unsuccessful angiosperms Lewis & Doyle 2002 prk596f (prk663f) GAYTTCAGYATYTAYTTRGACAT exon 2, 574–596 29,589,742–29,589,764 23 bp forward PCR/sequencing angiosperms D. Springate, RBG Kew, unpublished prK973r (prk1040r) TCTRTCAAATTGYCCRTCCATYTC exon 5, 973–996 29,590,630–29,590,653 24 bp reverse PCR/sequencing angiosperms D. Springate, RBG Kew, unpublished prk1167r ATGGTYTGRAANARACCNGTNCCRTTGTTGC exon 5, 1,100–1,130 29,590,757–29,590,787 31 bp reverse PCR/sequencing angiosperms Lewis & Doyle 2002 26 bp forward PCR bromeliads this study; modified from “prk622f“ BARFUSS, M.H.J. 90 Table 9. Primers tested for phosphoribulokinase (PRK). Highlighted primers performed best and were selected for PCR or sequencing. Position of primers was determined from the assembled RefSeq genome of Oryza sativa. bp = base pairs. CDS = coding sequence. Chr = chromosome. reference primers initially tested prk488f 1 primers subsequently tested TCAGCAATGAGGTTAAATTTGCATGG exon 2, 596–621 29,589,764–29,589,789 prk622f CAGCAATGAGGTTAAATTTGCATGGA exon 2, 597–622 29,589,765–29,589,790 26 bp forward PCR/sequencing bromeliads Schulte, Barfuss & Zizka 2009 prk630f2 AAATTTGCATGGAAAATTCAGGTC exon 2, 610-630 (–633) 29,589,778–29,589,801 24 bp forward sequencing bromeliads this study CTGCAGATCCGCAGAAGAAATATGC exon 4, (710–) 716–734 29,590,137–29,590,161 25 bp forward sequencing bromeliads Schulte, Barfuss & Zizka 2009 CCGCAGATCCGCAGAAGAAATTTTC exon 4, (710–) 716–734 29,590,137–29,590,161 25 bp forward sequencing Brocchinia this study; modified from “prk734f” prk889r GGGTATGAGCATGTCAATTTCCTCCC exon 4, 889–914 29,590,316–29,590,341 26 bp reverse sequencing bromeliads Schulte, Barfuss & Zizka 2009 prk890r GGGTATGAGCATGTCAATTTCCTCC exon 4, 890–914 29,590,317–29,590,341 25 bp reverse sequencing bromeliads this study; modified from “prk889r” prk1057r CTTCAGCATTTGTTGTGTCACCTC exon 5, 1,057–1,080 29,590,714–29,590,737 24 bp reverse sequencing bromeliads this study prk1069r GAAAATCTGCRTGCTTCAGCATTTG exon 5, 1,069–1,093 29,590,726–29,590,750 25 bp reverse PCR/sequencing bromeliads Schulte, Barfuss & Zizka 2009 prk1069r-2 GGAAAATCTGCRTGCTTCAGCATTTG exon 5, 1,069–1,094 29,590,726–29,590,750 26 bp reverse PCR bromeliads this study; modified from “prk1069r” prk734f (prk735f) prk734f-2 notes: 2,4 1 2,3 unlike given in Lewis & Doyle (2002) primer position of “prk488f” is in exon 1 but not in exon 3. italicized numbers indicate base positions that partly lie within introns. 3 position of “prk735f” was initially wrongly determined (Schulte & al, 2009) and therefor corrected from “prk734f”. 4 for the genus Brocchinia primer “prk734f” had to be modified because of mismatches that give problems in sequencing. 2 NUCLEAR DNA MARKERS prk621f name sequence (5'–3') CDS position (bp) Chr position (bp) length direction use specificity reference primers used initially for primer design PHYupstream1 TCWGGNAARCCNTTYTAYGC exon 1, 484–503 31,768,662–31,768,681 20 bp forward primer design angiosperms Mathews & Donoghue 1999 PHYupstream2 CCITTYTAYGSIATHYTICAYMG exon 1, 493–515 31,768,671–31,768,693 23 bp forward primer design angiosperms Mathews & Donoghue 1999 PhyCdownstream GRATKGCATCCATYTCMAYRTC exon 1, 1,705–1,726 31,769,883–31,769,904 22 bp reverse primer design angiosperms Mathews & Donoghue 1999 primers subsequently tested Russell & al., 2010b, modified from “PHYupstream1” this study, modified from “PHYupstream2” this study, modified from “PHYupstream2” phyc503f-mo TCVGGGAAGCCSTTYTAYGC exon 1, 484–503 31,768,662–31,768,681 20 bp forward PCR/sequencing monocots phyc515f-mo AAGCCCTTYTACGCVATMATGCACCG exon 1, 490–515 31,768,668–31,768,693 26 bp forward PCR/sequencing monocots phyc515f-br AAGCCCTTYTACGCTATCCTGCACCG exon 1, 490–515 31,768,668–31,768,693 26 bp forward PCR bromeliads phyc524f-br GCTATCCTGCACCGGATCGAYGT exon 1, 502–524 31,768,680–31,768,702 23 bp forward sequencing bromeliads this study this study this study phyc974f-mo GCTCCTCAYGGCTGYCAYGCTCA exon 1, 952–974 31,769,130–31,769,152 23 bp forward sequencing some monocots: bromeliads, orchids phyc974f-br GCTCCTCACGGCTGCCACGCTCA exon 1, 952–974 31,769,130–31,769,152 23 bp forward sequencing bromeliads some monocots: bromeliads, orchids some monocots: bromeliads, orchids some monocots: bromeliads, orchids some monocots: bromeliads, orchids IN TILLANDSIOIDEAE (BROMELIACEAE) Table 10. Primers used/tested for phytochrome C (PHYC). Highlighted primers performed best and were selected for PCR or sequencing. Position of primers was determined from the assembled RefSeq genome of Oryza sativa. bp = base pairs. CDS = coding sequence. Chr = chromosome. this study phyc991f-mo CCAYGCTCARTAYATGGCTAATATGG exon 1, 966–991 31,769,144–31,769,169 26 bp forward sequencing phyc1145r-mo CCTGMARCARGAACTCACAAGCATATC exon 1, 1,145–1,171 31,769,323–31,769,349 27 bp reverse sequencing phyc1145r2-mo CAACAGGAACTCACAAGCATATC exon 1, 1,145–1,167 31,769,323–31,769,345 23 bp reverse sequencing phyc1210r-mo GGATATGCTTCTCCTTTGYTTGAGC exon 1, 1,210–1,234 31,769,388–31,769,412 25 bp reverse sequencing phyc1690r-br TCAACATCTTCCCAYGGGAGGCT exon 1, 1,690–1,712 31,769,868–31,769,890 23 bp reverse sequencing bromeliads this study phyc1699r-mo ATWGCATCCATTTCAACATCYTCCCA exon 1, 1,699–1,724 31,769,877–31,769,902 26 bp reverse PCR/sequencing monocots this study phyc1699r-br ATWGCATCCATTTCAACATCTTCCCA exon 1, 1,699–1,724 31,769,877–31,769,902 26 bp reverse PCR bromeliads this study phyc1705r-mo GRATWGCATCCATYTCAACATC exon 1, 1,705–1,726 31,769,883–31,769,904 22 bp reverse PCR/sequencing monocots Russell & al., 2010b, modified from “PhyCdownstream” Russell & al., 2010b this study; modified from “phyc1145r-mo” this study PART 2, CHAPTER 3 91 BARFUSS, M.H.J. 92 Table 11. Attributes of analyzed matrices and parsimony scores of equally most parsimonious trees after analysis using PAUP*. ITS nrDNA was analysed with (incl.) and without (excl.) the problematic region of ITS1. Data of plastid DNA was taken from Barfuss & al., (2005). The sequence length is given for individual markers only. bp = base pairs, var. char. = variable characters, PUIC = parsimony uninformative characters, PIC = parsimony informative characters, CI = consistency index, RI = retention index, RC = rescaled consistency index, HI = homoplasy index. marker no. of taxa studied amplicon lenght [bp] no. of ch. no. of constant ch. no. of variable ch. no. of PUIC. no. of PICs no. of trees tree length CI RI RC HI ITS nrDNA (incl.) 111 884–1,017 1,291 899 (69.6%) 392 (30.4%) 173 (13.4%) 219 (17.0%) >10,000 936 0.590 0.748 0.441 0.410 ITS nrDNA (excl.) 111 884–1,017 1,057 788 (74.5%) 269 (25.5%) 120 (11.4%) 149 (14.1%) >10,000 634 0.588 0.779 0.458 0.412 MS 13 659–707 710 n. d. n. d. n. d. n. d. n. d. n. d. n. d. n. d. n. d. n. d. RPB2 8 520–522 523 n. d. n. d. n. d. n. d. n. d. n. d. n. d. n. d. n. d. n. d. PGIC 2 983 983 n. d. n. d. n. d. n. d. n. d. n. d. n. d. n. d. n. d. n. d. NIA 72 649–800 1,022 734 (71.8%) 288 (28.2%) 164 (16.1%) 124 (12.1%) >10,000 527 0.691 0.784 0.542 0.309 XDH 64 1,110 1,110 860 (77.5%) 250 (22.5%) 138 (12.4%) 112 (10.1%) >10,000 369 0.743 0.740 0.549 0.257 PRK 444 831–1,692 3291 2,554 (77.6%) 737 (22.4 %) 203 (6.2%) 534 (16.2%) >10,000 2,232 0.478 0.870 0.416 0.522 PHYC 444 1159–1,192 1,228 743 (60.5%) 485 (39.5%) 116 (9.4%) 369 (30.1%) >10,000 1,398 0.473 0.867 0.410 0.527 plastid DNA 120 ∑ of 7 markers 6,277 5,294 (84,3 %) 983 (15,7 %) 440 (7.0%) 543 (8.7%) >10,000 1,713 0.654 0.830 0.543 0.346 NUCLEAR DNA MARKERS IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 Amplification, sequencing, and cloning of low-copy nuclear DNA markers Gradient PCR conditions were optimized with either a Mastercycler gradient (Eppendorf) or a Veriti 96-Well Thermal Cycler (Applied Biosystems, Life Technologies) using their gradient function or independent temperature blocks, respectively. Gradient PCR was usually done in two successive steps, (1) running gradients for the annealing temperature, and (2) running gradients for the extension temperature. The first step is necessary to find the best annealing temperature for a given primer pair, which can differ greatly from calculated melting temperatures using different programs for oligonucleotides (e.g., NetPrimer; http://www.premierbiosoft.com/netprimer/index.html). General PCR conditions for the first step were: 1× 95°C for 2 min; 35× 95°C for 30 s, X°C for 30 s, 72°C for Y min; 1× 72°C for 7 min; and 4°C for ∞; where X stands for the annealing temperature variation and Y for the extension time used. The latter was dependent on the expected fragment length and ranged between one and three minutes according to the general rule that Taq amplifies at a rate of 1 kb per minute. Usually two rounds of annealing temperature gradients were performed, initially with a wide temperature variation of 48–68°C and later a narrower range determined based on the results of the first round. The second step was done because the extension temperature depends not only on the working optimum of Taq polymerase but also on the base composition (GC content) in certain areas of the amplified region (Su & al., 1996). AT-rich sections might need a reduced extension temperature, since these segments start to denature sometimes even below 72°C and Taq polymerase is not able to function. The online program POLAND (Steger, 1994; http://www.biophys.uni-duesseldorf.de/html/local/POLAND/poland.html) can be used to check for 50% temperature probabilities of each base in a given sequence, which is a theoretical calculation using different formulas for how strongly bases are bound to each other. This presents a problem with this approach because at least one sequence of the targeted region for a given study group is required a priori to conduct this analysis. The general PCR conditions for the second step were: 1× 95°C for 2 min; 35× 95°C for 30 s, X°C for 30 s, Z°C for Y min; 1× Z°C for 7 min; and 4°C for ∞; where Z refers to a temperature variation of 60–72°C. Usually one gradient run was sufficient to verify whether a denaturing temperature of 72°C is too high. A third gradient step to determine the denaturing temperature was necessary only for ITS nrDNA, because Taq is not able to pass through GC-rich regions, which are still bound at 72°C because they were not denatured at 95°C completely. If necessary, further PCR optimization steps were undertaken and/or specific PCR conditions established for each marker (Table 2). Occasionally other parameters for markers such as two-step PCR, inclusion of a touchdown cycle, or an increase of the extension time by 5 s at later cycles has been shown to produce greater amounts of PCR product. Each gradient PCR was done in 0.2 mL 12-strip PCR tubes (or 96-well PCR plates) using 10 µL reactions including 9 µL 1.1× ReddyMix PCR Master Mix containing 2.5 mM MgCl2 (AB-0619; Thermofisher, ABgene), 0.4 µM of each primer (0.2 µL at 20 µM = 20 pmol/µL), 4% (0.4 µL) dimethysulfoxide (DMSO), and 0.2 µL (approximately 25 ng/µL) template DNA. Amplifications of each marker using optimized conditions were usually carried out one a Mastercycler gradient in 25 µL reactions using 22.5 µL 1.1× ReddyMix PCR Master Mix containing 2.5 mM MgCl2, 0.5 µL (0.4 µM) of each primer at 20 µM, 1 µL (4%) DMSO, and 0.5 µL template DNA. Volumes were scaled up proportionally when more PCR product was necessary to perform cycle sequencing with additional primers. When single bands were present on agarose 93 BARFUSS, M.H.J. NUCLEAR DNA MARKERS gels, PCR products were purified either with a 1:2 mixture of exonuclease I (20 units/µL; Fermentas) and alkaline phosphatase (either shrimp or FastAP thermosensitive alkaline phosphatase, 1 unit/µL; Fermentas) or ExoSAP-IT (Amersham, GE Healthcare) to degrade single stranded DNA fragments and dNTPs (Werle & al., 1994); 2.5 µL of the enzyme mixture (scaled up proportionally for higher volumes) was added to each 25 µL PCR reaction and incubated at 37°C for 45 min, followed by deactivating the enzymes at 85°C for 15 min. Cycle sequencing reactions were usually performed on a 96-Well GeneAmp PCR System 9700 (Applied Biosystems, Life Technologies) using a modified reaction protocol: 0.5 µL BigDye Terminator v3.1 (Applied Biosystems, Life Technologies), 1 µL (0.4 µM) primer at 4 µM, 1.75 µL 5× sequencing buffer, 3–6.75 µL purified PCR product, and 0–3.75 µL PCR-grade water. The general temperature profile (except for ITS nrDNA) was: 1× 96°C for 1 min; 35× 96°C for 10 s, 50°C for 5 s; 60°C for 3 min; 4°C for ∞. Modifications to the original profile are increased number of cycles (orig. 25 cycles) and reduced extension time (orig. 4 min). Cycle sequencing products were purified by gel filtration using MultiScreen filter plates (MAHVN4550; Milipore) and the cross-linked dextran gel Sephadex G-50 Superfine or Fine (GE Healthcare) according to the Millipore protocol (Tech Note TN053) with few modifications (more details provided upon request from the first author). Sequences were run on either a 16-capillary sequencer (3130xl Genetic Analyzer, Applied Biosystems, Life Technologies) or a 48-capillary sequencer (3730 DNA Analyzer; Applied Biosystems, Life Technologies) following manufacturer’s instructions. When cloning was required, the proof-reading Phusion High-Fidelity DNA polymerase (F530S; Finnzymes) was used to reduce the accumulation of polymerase errors in the cloned sequences and the formation of artificial chimeric PCR products. PCR reactions were 50 µL, including 34.7 µL purified water, 10 µL 5× Phusion High Fidelity PCR buffer, 0.8 µL MgCl2 at a concentration of 50 mM, 1 µL dNTPs at 10 mM each, 0.5 µL Phusion DNA polymerase (2U/µL), 1 µL of each primer at 20 mM, and 1 µL template DNA. Thermocycling was performed according to manufacturer’s instructions with slight modifications to fit the optimize PCR conditions for each marker. Amplified DNA was verified and purified on 1% TAE agarose gels using the Invisorb Spin DNA Extraction Kit (Invitek) following the manufacturer’s instructions. Since the proof reading DNA polymerase is not able to create 3' A-overhangs, which is necessary to proceed with standard TA cloning, purified PCR products were incubated at 72°C for 20 min using a mixture of 0.5 µL dATPs at 10 mM, 2.5 µL 10× Taq PCR buffer containing (NH4)2SO4, 0.1 µL Taq polymerase (5U/µL), and 1.5 µL MgCl2 at 25 mM per 25 µL purified PCR product (all products from Fermentas). Cloning was performed using the pGEM-T Easy Vector System (Promega) following manufacturer’s instructions. Successfully transformed colonies were suspended in 1× TE buffer, denatured for 10 min at 95°C, placed on ice, centrifuged to collect cell debris, and used as a template for PCR. Subsequent amplification and sequencing were performed as described above. Amplification and sequencing of ITS nrDNA PCR of the ITS nrDNA (Figure 1, A) was initially attempted using the universal primers “ITS2”, “ITS3”, “ITS4” and “ITS5” (White & al., 1990; Baldwin & al., 1995) and the angiosperm-specific primers “17SE” and “26SE” (Sun & al., 1994; Douzery & al., 1999). No useful PCR products were obtained at the first step of gradient PCR using standard reactions. Therefore the new angiosperm-specific PCR primers “ITS18Sf”, “ITS5.8Sf”, “ITS5.8Sr”, and “ITS26Sr” with a significantly higher annealing temperature were employed (Gruenstaeudl & al., 2009). These primers have been developed using large-unit ribosomal sequences of angiosperms from GenBank plus a 94 IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 sequences of Bromeliaceae. The new internal 5.8S nrDNA primers, “ITS5.8Sf” and “ITS5.8Sr”, were used because “ITS3” and “ITS2” of White & al. (1990) do not produce overlapping reads and have a low annealing temperature. For further PCR trials, standard reactions were altered by adding DMSO in final concentrations of up to 10% and/or betaine up to 1 M; these additives are thought to aid in amplification by reducing GC binding during extension. Betaine had no obvious effect, but DMSO at a final concentration of 3% produced a PCR product of the expected length in at least some samples. Therefore, in a final optimization step a gradient run for the denaturing temperature was performed using a temperature range of 94–99°C. Already at 96°C more samples showed the expected PCR product and at a denaturing temperature of 99°C consistently all samples had comparatively strong bands. Additionally an extension time twice as high as the general rule (1 kb/m) increased the amount significantly. The optimized PCR conditions are given in Table 2. For the initial phase of cycle sequencing for ITS nrDNA, all four newly generated primers and the standard reaction protocol and PCR conditions were used. Two problems in these sequences were detected. Firstly electropherograms had n-1 patterns because of high primer melting temperatures (around 70 °C), which was solved by using the primers from Blattner (1999) and/or several newly designed sequencing primers that have annealing temperatures around 65°C. The second problem was that only internal primers sequencing towards the 26S nrDNA gave complete reads. In all other cases signal intensity dropped and was mostly completely lost. In such cases, the dGTP BigDye Terminator v3.0 Cycle Sequencing Kit (Applied Biosystems, Life Technologies) procedure has been modified. To be able to run this chemistry on the capillary sequencers it was mixed in a ratio of 1:4 (dGTP v3.0:v3.1) with the BigDye Terminator v3.1 to avoid strong banding compression patterns that are visible in capillary electrophoresis by using the dGTP BigDye Terminator v3.0 alone. Also the effect of DMSO and betaine was tested. The general sequencing protocol was modified by using 1 µL per reaction of the BigDye Terminator mix, 1 µL (0.4 µM) primer at 4 µM, 1 µL 5× sequencing buffer, 1.6 µL (0.8 M) betaine at 5 M, 0.4 µL (4%) DMSO, 4 µL purified PCR product, and 1 µL PCR-grade water. Cycling conditions were changed to: 1× 96°C for 1 min, 99°C for 30 s, and 60°C for 3 min; 35× 96°C for 20 s, 60°C for 3 min; and 4°C for ∞. No annealing step was necessary since all primers used for sequencing have annealing temperatures higher than 60°C. Details of all PCR and sequencing primers tested and finally used for ITS nrDNA are given in Table 3. Analyses of nuclear DNA sequences Raw sequences were initially analyzed and edited using the Sequencing Analysis Software v5.3 (Applied Biosystems, Life Technologies). Forward and reverse DNA strands from both PCR and/or internal sequencing primers were assembled with the SeqMan Pro module of the Lasergene v8.1 software package (DNASTAR). Contigs were edited and allelic consensus sequences were exported as text file in fasta format. Fasta files were pooled and aligned with MUSCLE v3.8 (Edgar, 2004a, b) leaving default parameters unchanged and then adjusted by eye with BioEdit v7.0.5 (Hall, 1999) following the guidelines of Kelchner (2000) and Borsch & Quandt (2009). In polymorphic DNA regions containing numerous insertion/deletion (indel) events related taxa were aligned first. Later these blocks of aligned taxa were aligned using the profile alignment function of MUSCLE and adjusted by eye using BioEdit. 95 BARFUSS, M.H.J. NUCLEAR DNA MARKERS Figure 2. Secondary structure (hairpins) for eight selected taxa of Tillandsioideae of the same problematic region of the ITS1 obtained from RNA folding on the mfold web sever. 96 IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 97 Figure 3. Alignment of a section of the ITS1 region showing the problematic region that has an extremely stable secondary structure, which gives problems in PCR and sequencing. BARFUSS, M.H.J. NUCLEAR DNA MARKERS The Alignment of ITS nrDNA was guided by their secondary structures (Figures 2, 3). Stem-loop regions (hairpins) were identified either visually or on the mfold web server (Zuker, 2003). A certain GC rich region within ITS1 could not be unambiguously aligned (Figure 3). Aligned ITS nrDNA sequences were checked for the presence of the conserved angiosperm motifs -GGCRY[4–7 N]-GYGYCAAGGAA- in ITS1 (Liu & Schardl, 1994), -GAATTGCAGAATCC- in the 5.8S nrDNA region (Jobes & Thien, 1997), and the conserved (C1–C6) and variable (V1–V6) domains in ITS2 (Hershkovitz & Zimmer, 1996) to exclude pseudogenes. Alignments of low-copy nuclear DNA markers were adjusted by checking the reading frame, by determining the angiosperm-specific intron-flanking GT-AG motive in intron regions, and by using annotated genes of the assembled RefSeq genome of Oryza sativa and other annotated sequences from GenBank. Aligned matrices can be obtained from the first author upon request. Since PCR products can contain two or more alleles and were mostly sequenced directly in this study, heterozygous individuals contain two types of polymorphisms: (1) single nucleotide polymorphisms (SNPs) that are visible as double peaks at certain positions, and (2) length polymorphisms caused by insertion/deletion (indels) events, which result in polymorphisms at every subsequent position in the electropherogram. Unambiguous SNPs were coded using the IUPAC-IUB symbols for nucleotide nomenclature. Indels made assembly of contigs difficult. To be able to assemble such contigs, the type of indel was identified by visual inspection and usually the shorter allele was ignored and manually overwritten by the sequence of the longer. Assembled contigs were edited with special care to correct for handling mistakes and the initially ignored sequence of the shorter allele was checked not to miss any mutations. If two or more indels between priming sites were observed, editing was highly problematic and errorprone; in such cases, these individuals were excluded or the PCR products cloned and sequenced. Nuclear DNA alignments were inspected prior to phylogenetic analyses for the impact of SNPs and indels within allelic consensus sequences of a given individual (Jabaily & Sytsma, 2010). The presence of similar alleles in more distantly related taxa may indicate possible hybridization events including reticulations and introgressions, polyploidization, and occurrence of incomplete lineage sorting. When no significant conflict was detected (homoplasious SNPs within accessions of distantly related taxa were ignored), sequences were included and differences were assumed to be most likely due to allelic variation within a single species or within a species complex. Otherwise individuals were excluded or sequenced after cloning. Sequences resulting from PCR-mediated, artificial recombinations of alleles were sometimes observed and excluded. PAUP* version 4.0b10 (Swofford, 2003) has been used for all analyses using maximumparsimony with unordered and equally weighted characters. A two-step heuristic search method was used: (1) 1,000 random sequence additions, TBR branch swapping holding 10 trees each step, and saving 10 shortest trees per replicate; (2) trees obtained from the first analyses were taken as starting trees and heuristic searches were conducted until all or a maximum of 10,000 shortest trees were saved, but the tree search was allowed to continue to check all input trees for possible shorter ones (swapping to completion). Gaps were treated as missing characters, and no gap coding procedure was applied because of the many indels within accessions that were ignored within allelic consensus sequences. The matrices including all characters were analyzed for five nuclear DNA markers (ITS nrDNA, NIA, XDH, PRK, PHYC) and for ITS nrDNA alone in a matrix excluding the difficult region of ITS1 (Figures 2,3), since the alignment was ambiguous for that region and many gaps had to be introduced. For compari98 IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 son, the plastid matrix of Barfuss et al. (2005) was reanalyzed with the same phylogenetic methods, but the super-outgroup Stegolepis (Rapateaceae) has been excluded to be comparable with results of the other loci. Statistics for each marker are given in Table 11. Support was estimated with the bootstrap (Felsenstein, 1985). Bootstrap percentages (BP) were calculated using 1,000 replicates, TBR branch swapping, and simple sequence addition, holding 10 trees each step and saving 100 shortest trees per bootstrap replicate. Results Direct sequencing of the eight nuclear DNA markers was possible in nearly all accessions. Comparisons of amplified fractions of nuclear DNA markers to the assembled RefSeq genome of Oryza sativa showed that MS, RPB2, XDH, and PHYC are single-copy in rice (Table 1). Although present in multiple copies in Oryza sativa, blast of PGIC, NIA, and PRK fragments produced in this study showed greatest similarity to the same single paralog. Nuclear ribosomal internal transcribed spacer (ITS nrDNA) Previously published protocols and PCR primers did not consistently produce PCR products, indicating that either the primers did not match the target or the protocol was inappropriate. The newly developed primers with a significantly higher annealing temperature and higher DMSO concentrations produced PCR products in few samples if the standard denaturing temperature was used. Therefore high strong secondary structure are likely present in bromeliad ITS nrDNA, which is insufficiently denatured at 95°C to allow successful amplification at 72°C (Figure 2). At higher denaturing temperatures PCR bands of the expected length began to appear but yields were variable; at 99°C, all samples amplified consistently. As would be suspected from high GC content (Figure 3), cycle sequencing was also problematic. By altering the sequencing chemistry it was possible to get full ITS nrDNA sequences for all included accessions, although a three-step loss of signal still occurred. Some taxa gave problems even with this modified protocol and were finally excluded. With the primers finally used in this study, a standard ITS nrDNA sequence was composed of 123 bp of 18S nrDNA, ITS1 (of variable length), 164 bp for 5.8S nrDNA, ITS2 (of variable length), and 139 bp of 26S nrDNA. Fragment size ranged from 884–1,017 bp (in Tillandsia biflora and Vriesea scalaris, respectively). No pseudogenes were found, although slight changes to the published motifs of the conserved and variable domains described previously for ITS2 were found. The final matrices included 1,291 (including all characters) and 1,057 (excluding the problematic ITS1 region) positions, respectively. Other statistics are given in Table 11. For ITS nrDNA strict consensus trees (Figure 4), bootstrap consensus trees (Figure 5) and phylograms (Figure 6) are shown, resulting from analyses of the total matrix as well as of a matrix excluding the problematic ITS1 region. Although several clades of related species received bootstrap support, backbone relationships are insufficiently resolved. Each accession is different from the rest (i.e. long terminal branches) but synapomorphies are few (i.e. short internal branches), as can be seen in one of the individual parsimony trees (Figure 6, Table 11). 99 BARFUSS, M.H.J. 100 NUCLEAR DNA MARKERS Figure 4. Strict consensus trees of the ITS nrDNA: analysis of the complete DNA sequence matrix (A), analysis excluding the problematic region of ITS 1 (B). IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 101 Figure 5. Bootstrap consensus trees of the ITS nrDNA: analysis of the complete DNA sequence matrix (A), analysis excluding the problematic region of ITS 1 (B). BARFUSS, M.H.J. 102 NUCLEAR DNA MARKERS Figure 6. Phylograms selected randomly from 10,000 trees saved during maximum parsimony analyses of the ITS nrDNA: analysis of the complete DNA sequence matrix (A), analysis excluding the problematic region of ITS 1 (B). IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 Malate synthase (MS) and RNA polymerase II, beta subunit (RPB2) MS (Figure 1, B) was initially attempted using the primers “ms400f”, “ms526f”, “ms943r” and “ms1488r” of Lewis & Doyle (2001) and the modified or new primers “ms356f” and “ms1408r” (D. Springate, RBG Kew, London, unpublished) in all possible combinations. Since no amplifications were obtained, new primers were designed based on MS sequences obtained from GenBank. Three forward and three reverse primers were developed covering a region from partial exon 1 to partial exon 4. The position of internal primers was designed to amplify three overlapping fragments that might be assembled as one piece. New primers were tested in all possible combinations, but only amplification of the fragment containing intron 2 using the primers “ms428f” and “ms960r” was successful and further optimized (Table 2). PCR primers were used as sequencing primers. Details of all primers tested for MS are given in Table 4. RPB2 (Figure 1, C) was tested using primers “P6F”, “P7R”, “P7F”, “10R”, “P10F”, and “11aR” (Table 5) of Denton & al. (1998) in all possible combinations for a region covering partial exon 11 to partial exon 24. Only the primer combination “P10F”/”P11aR” successfully amplified a fragment that contains intron 23. PCR conditions were further optimized (Table 2) and the product finally sequenced using PCR primers. Both markers were tested for two small sets from Tillandsia species complexes (T. tectorum, T. plumosa) with Mezobromelia hutchisonii as outgroup. The amplified fragment of MS (659–707 bp) covered a region from partial exon 2 (279–282 bp) to partial exon 3 (261 bp) (Figure 1, B), that of RPB2 (520–522 bp) a region from partial exon 23 (148 bp), to partial exon 24 (134 bp) (Figure 1, C). Both markers displayed indel variation, but only few substitutions. No phylogenetic analysis was conducted due to the limited number of taxa and sequence variability. Glucose-6-phosphate isomerase, cytosolic (PGIC) Amplification of PGIC (Figure 1, D) was attempted using primers of Ford & al. (2006) in different combinations. Primer combination “AA11F”/”AA16R” for a fragment covering parts of exon 11 to exon 16, respectively, was the only successful one. PCR primers were used as sequencing primers. Optimized PCR conditions and primer details are given in Table 2 and Table 6, respectively. PGIC was currently tested only for two species, one each from Catopsis and Alcantarea. The amplified fragment (983 bp) covered a region from partial exon 11 (17 bp) to partial exon 16 (6 bp) (Figure 1, D). Two 1-bp insertions and 53 substitutions were observed. No phylogenetic analysis was performed, due to limited sampling. Nitrate reductase 1, [NADH] (NIA) NIA (Figure 1, E) was initially tested using the primers combinations “NIA2F”/”NIA2R” (targeting intron 2) and “NIA3F”/”NIA3R” (targeting intron 3) of Howarth & Baum (2002). Amplification of intron 2 failed and the primers for intron 3 only amplified a ca. 200 bp fragment. Therefore new primers were designed to cover a region from partial exon 1 to partial exon 3. Several primer combinations have been tested, but only the combination “nia410f”/”nia1042r” covering a part of exon 1 and the complete intron 1 gave satisfactory results and was therefore optimized further (Table 2). PCR primers were used as sequencing primers. Primer details are given in Table 7. 103 BARFUSS, M.H.J. 104 NUCLEAR DNA MARKERS Figure 7. Strict consensus tree (A) and bootstrap consensus tree (B) of NIA. IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 105 Figure 8. Strict consensus tree (A) and bootstrap consensus tree (B) of XDH. BARFUSS, M.H.J. NUCLEAR DNA MARKERS NIA was investigated for an enlarged sample set (72 accessions). The amplified fragment (729– 798 bp) covered a region from parts of exon 1 (629–638 bp), intron 1 (of variable length), to partial exon 2 (15 bp) (Figure 1, E). Fragment length ranged from 649–800 bp (Brocchinia micrantha, Tillandsia duratii, respectively). Some accessions displayed a microsatellite at the 3' end of intron1. A strict consensus and a bootstrap consensus tree are shown in Figure 7. Backbone relationships are not resolved, but several terminal clades receive support. Xanthine dehydrogenase (XDH) Amplification of XDH (Figure 1, F) was attempted using primers that were newly designed for partial exon 4 and complete intron 5 using sequences downloaded from GenBank. Primers of the phylogenetic studies of Górniak & al. (2010) and Morton (2011) became available only after the current study was completed. Amplification of intron 5 failed so only the successfully amplified part of exon 4 was further optimized (Table 2). Best amplicons were achieved using the primer combination “xdh479f”/”xdh1611r”. These were initially also used as sequencing primers, but in addition internal sequencing primers were developed to produce acceptable traces for the whole fragment. All primers are listed in Table 8. XDH was investigated for an enlarged sample set (64 accessions). PCR primers produced fragments with part of exon 4 (1,107–1,110 bp). No indels within accessions were found, and only a 3-bp indel for Brocchinia micrantha was observed in the entire DNA matrix. Resolution is restricted to some terminal clades (Figure 8). Phosphoribulokinase (PRK) For PRK (Figure 1, G) initial gradient PCR runs were performed using degenerate primers originally designed for angiosperms, monocots, and/or palms (Lewis & Doyle 2002; D. Springate, RGB Kew, London, unpublished). Primer combinations “prk663f”/”prk1040r” and “prk663f”/”prk1167r” spanning a region of partial exon 2 to partial exon 5 amplified single, weak PCR products only in samples having high-quality DNA extracts, whereas any combination using “prk488f” located in exon 1 yielded no amplification. After successful sequencing of PCR products, a second set of bromeliad-specific PCR primers (“prk622f”, “prk1069r”) and additional internal primers (“prk734f”, “prk889r”) sitting in exon 4 were designed, which were initially used for Bromelioideae (Bromeliaceae) by Schulte & al. (2009). However, these primers often showed primer-dimer formation. Therefore in a last phase of optimization, three primers were slightly modified (“prk621f”, “prk890r”, “prk1069r-2”) to allow higher annealing temperatures and two additional nested sequencing primers (“prk630f”, “prk1057r”) were developed to avoid sequencing problems caused by incorporation of dimers into the products. For the earliest diverging genus of Bromeliaceae, Brocchinia, “prk734f” had to be adjusted because of mismatches present at the 3' end (prk734f-2). The optimized PCR conditions and all primers tested are given in Table 3 and Table 9, respectively. Final PCR primers (“prk621f”, “prk1069r-2”) amplified fragments ranging from 831–1,692 bp in Tillandsia ionochroma and Guzmania patula, respectively (see Supplementary Data of chapter 5). A standard PRK sequence was composed of parts of exon 2 (9 bp), intron 2 (of variable length), exon 3 (85 bp), intron 3 (of variable length) exon 4 (245 bp), intron 4 (of variable length) and parts of exon 5 (108 bp) (Figure 1, G). Primers “prk621f” and “prk1069r-2” were designed to be highly specific for PRK of Bromeliaceae, yielding clean, single PCR bands even for accessions for which initial and second primer sets produced faint or no bands. 106 IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 107 Figure 9. Simplified strict consensus trees of equally-most-parsimonious trees found in parsimony analyses of individual datasets supplemented with bootstrap support from parsimony bootstrap analysis either above or below branches; triangles sitting in terminal position on a branch indicate monophyletic groups, whereas triangles emerging from within a node indicate several branches of that group on a polytomy. The tree for plastid DNA was modified according to results of Barfuss & al., (2005) and is shown for comparison. BARFUSS, M.H.J. NUCLEAR DNA MARKERS Direct sequencing was successful in about ⅔ (67.2%) of the accessions with few SNPs (23.6%). Intra-accession indels were detected in 146 (32.8%) accessions. The number of indels within a given accession ranged from 1–5. Indel variation was only found in the less conserved introns and never in exons. Three taxa had to be cloned, since the number of indels was too great to be unambiguously edited: Vriesea psittacina, Guzmania graminifolia, and Tillandsia platyrhachis. Occasionally others were cloned to unambiguously verify the length of both alleles [see Appendix]. SNPs were found in 241 (54.3%) out of 444 sampled accessions. Only 193 (43.5%) samples were found to be homozygous for the region sequenced; inter- and intraindividual variation in introns was high. The number of SNPs in a given accession ranged from 1–18 (highest number in Tillandsia fasciculate, B0076) were primarily found in introns and less frequently in exons, which primarily had synonymous substitutions. Homopolymers were found in several accessions in intron 2 and 4, respectively (see Supplementary Data, chapter 5). Statistics for the PRK matrix and trees are given in Table 11. Indels are common at all hierarchical levels, yielding a final matrix of 3,291 characters. Evaluation of SNPs and indels within accessions showed no significant impact on subsequent data analysis (signal conflict), although homoplasy was present. Therefore coded SNPs were used except for the three cloned taxa. Phylogenetic relationships based on PRK variation are shown in Figure 9, which is the strict consensus tree with bootstraps indicated. Strongly supported are the sister group relationship (BP 100) of Glomeropitcairnia (BP 100) and Catopsis (BP 100) and core Tillandsioideae (BP 100). Tribes Vrieseeae and Tillandsieae as well as core Tillandsieae are not supported. The Cipuropsis and Vriesea groups are strongly supported (BP 91 and 100, respectively). Werauhia is strongly supported (BP 100), whereas the Cipuropsis-Mezobromelia group has only weak BP (58). The same picture is seen for Alcantarea (BP 100) and Vriesea (BP 61). Vriesea is monophyletic only when Andean members located in the Cipuropsis-Mezobromelia group, Tillandsia malzinei and former xerophytic gray-leaved members, the latter two now placed within Tillandsia s. str., are excluded (data not shown). The sister group position of the Vriesea group and a clade containing Tillandsia s. str. and the T. dodsonii group was surprising but only weakly supported (BP 63). Monophyly of Guzmania is not supported, and three clades of Guzmania species (data not shown) occur in a polytomy with several others. The T. wagneriana (BP 64) and T. lindenii (BP 100) groups are resolved as sisters (BP 65). Racinaea is strongly supported (BP 96), but its relationships to other groups remain unresolved. Tillandsia s. str. and T. dodsonii group are resolved as sister groups with a moderate support (BP 81) with the T. dodsonii group being strongly supported (BP 96), but unresolved with several other clades of Tillandsia species. Phytochrome C (PHYC) For partial exon 1 of PHYC (Figure 1, H), initial PCRs were performed using the four degenerate primers “phyc0503f-mo”, “phyc0515f-mo”, “phyc1699r-mo”, and “phyc1705r-mo”, which were obtained from primer sequences originally published by Mathews & Donoghue (1999) and modified according to monocot sequences downloaded from GenBank. All primer combination worked, but showed strong primer-dimer formation. Best amplifications were achieved with “phyc0515f-mo” and “phyc1699r-mo”. However, all combinations were successfully sequenced, but showed strong signals of dimers at the start of the traces. Therefore bromeliadspecific PCR primers (“phyc0515f-br”, “phyc1699r-br”) were produced and optimized (Table 2) based on sequence information obtained from “phyc0503f-mo” and “phyc1705r-mo” since 108 IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 their annealing position is outside the annealing positions of these primers. In addition, nested sequencing primers (“phyc0524f-br”, “phyc1690r-br”) were designed to avoid sequencing problems as much as possible. Several internal primers creating overlapping reads towards the end of the fragment were also developed to sequence the whole PCR product. Monocotspecific ones were also tested successfully for Polystachya (Orchidaceae) by Russell & al. (2010). Four internal primers giving the best reads were finally selected for sequencing. Details of all primers are given in Table 10. Only fragments of partial exon 1 (Figure 1, H) were amplified ranging from 1,159–1,192 bp in Werauhia insignis and Guzmania acorifolia/G. condensata, respectively (see Supplementary Data of chapter 5), but mostly 1,177 bp. Direct sequencing was straightforward in nearly all cases. In contrast to PRK, no length differences between the two alleles of an accession were detected; this is likely due to the fact that only conserved parts of the exon 1 were amplified in PHYC. Only two obvious hybrids of distantly related taxonomic groups, where one of the possible parental taxa displayed two 3 bp insertions, were observed; these were finally excluded. Sequences of PHYC generally showed clean raw data with no or only few polymorphic sites in the electropherograms. Most individuals are either homozygous or show only little intraspecific allelic variation, as can be expected for relatively conserved exon. SNPs were found in 220 taxa out of 444 sampled individuals. About 224 samples were found to be homozygous for the amplified region. The number of SNPs in a given heterozygous taxon ranged from 1–15 (the upper limit seen in T. aff. cucaensis Wittm. B0735, see Supplementary Data of chapter 5). Distribution of polymorphic sites was even across the amplified portion of PHYC with no particular preference for a specific region, but significantly higher at the 3rd codon position (i.e. synonymous substitutions). Information on the analyzed PHYC matrix and trees are given in Table 9. Indels are rare, and present either in single species or few closely related species, leading to a final matrix size of 1,228 characters. Like in PRK, all polymorphic sites have been checked for their impact on subsequent data analysis, which did not show significant conflicts. Therefore the polymorphism-coded PHYC dataset was used for all analyses, with the exception of three taxa, which were cloned for PRK and therefore also for PHYC, to allow the combination of corresponding allelic sequences. Figure 9 shows the strict consensus tree of a MP analysis with BP, which summarizes phylogenetic relationships based on PHYC sequence data. As in PRK tree, both Glomeropitcairnia and Catopsis are strongly supported (BP 100) as sisters (BP 100), with the latter clade sister to a strongly supported core Tillandsioideae (BP 100). Tribe Vrieseeae is paraphyletic, but the resulting grade has only a weak support (BP 58). The Vriesea group (BP 100 ), Alcantarea (BP 97) and Vriesea (BP 89) are monophyletic, the last as found in PRK, but only when the Andean members, Tillandsia malzinei and former xerophytic gray-leaved species are excluded (data not shown). The Cipuropsis group splits into two lineages, Werauhia (BP 100) and the Cipuropsis-Mezobromelia group (BP 67). Mezobromelia is weakly supported (BP 57) as nested in the Cipuropsis-Mezobromelia group (data not shown), which in addition contains taxa previously assigned to Vriesea and Tillandsia (data not shown). Tribe Tillandsieae are strongly supported (BP 92). The first lineage that splits from core Tillandsieae is Guzmania (BP 85), with Mezobromelia hutchisonii nested inside (data not shown). Core Tillandsieae is only weakly supported (BP 58) and relationships of genera and groups are unresolved. The T. lindenii group (89 BP), Racinaea (no BP) and the T. dodsonii group (BP 86) are monophyletic but all three are placed in a polytomy with different clades of Tillandsia s. str. and the T. wagneriana group. 109 BARFUSS, M.H.J. NUCLEAR DNA MARKERS Discussion Challenges of nuclear DNA markers The fact that nuclear DNA sequences are much less used than plastid ones became evident in initial phases of the current study. Several problems highlighted earlier were also encountered during the optimization of nuclear markers for Tillandsioideae. The most time-consuming steps were primer design, extraction of high-quality DNA, PCR optimization, improving sequencing results of ITS nrDNA, and editing as well as aligning intron-containing markers and ITS nrDNA. The only advantage of Tillandsioideae is that it presumably consists of more than 98% diploid species, which clearly helps in data collection, evaluation and interpretation. The major challenges of nuclear DNA sequences for phylogenetic studies in Tillandsioideae are: (1) limited availability of effective primers; (2) occurrence of strong secondary structure (ITS nrDNA, NIA) that make PCR and sequencing difficult; (3) a short fragment length and consequentially a small number of PICs (e.g., amplified fragments of MS, RPB2); (4) heterozygous individuals with SNPs and indel mutations, the latter mainly in markers containing polymorphic introns (e.g., PRK), making sequence editing time-consuming and increased laboratory costs if cloning is required; and (5) alignment of markers containing polymorphic introns with indel variation. A proper pre-evaluation of nuclear DNA markers and the selection of more conserved regions (e.g., PHYC, PGIC, XDH) would have minimized these complications. A review article on nuclear DNA markers used in the Department of Systematic and Evolutionary Botany at the University of Vienna for different angiosperm lineages is in progress (Barfuss & al., in prep.). Nuclear versus plastid DNA markers The current results support the hypothesis that several nuclear DNA markers evolve much more rapidly than do plastid markers. The number of potential PICs per sequenced base pair in PHYC is 2-fold, in PRK 3-fold higher than in plastid DNA markers, which means sequencing these more variable nuclear DNA markers requires much lesser sequencing effort to have the same level of resolution as combined plastid DNA markers. However, homoplasy in three out of five nuclear data sets examined here was also higher (Table 11). Conserved regions like XDH evolve at about the same rate as plastid regions. Barfuss & al. (2005) and Givnish & al. (2011) have shown that the combination of up to eight plastid markers is insufficient to resolve phylogenetic relationships of Bromeliaceae down to the subgeneric rank (Figure 9). Inclusion of properly chosen plastid markers will definitely improve the already published plastid analyses, but this is costly and time-consuming. Whether sequencing more plastid or additional nuclear markers is more efficient and useful for the scientific questions being investigated needs to be evaluated before large datasets are sequenced. Phylogenetic utility of nuclear markers ITS nrDNA is not suitable for well resolved backbone relationships, although a lot of intraindividual variation was present (Figure 6). This is remarkable, since in many other angiosperm lineages (e.g., Barnadesioidae (Asteraceae): Gruenstaeudl & al., 2009) and especially in monocots (e.g., Musaceae: Hřibová, 2011; Polystachya (Orchidaceae): Russell & al., 2010;), ITS nrDNA has been applied successfully below family level. The strong secondary structure could have ecological significance for adaptation to hot and dry environments (Hurst & Merchant, 2001), as epi- and lithophytes are often exposed to changing microclimatic conditions. Howev110 IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 er, orchid twig epiphytes in subtribe Oncidiinae that occupy the same micro-environments as many species of Tillandsia exhibit much higher levels of ITS nrDNA variation than do those species occupying more mesic epiphytic habitats (Chase, 2009). In this context, it is interesting to note that mesophytic bromeliad taxa (e.g., Vriesea, Werauhia) display a shorter problematic ITS1 region than xerophytic species (Tillandsia). Because of low resolution and great difficulties in amplification, sequencing and alignment, ITS nrDNA cannot be recommended as marker generally applied in phylogenetic studies of Bromeliaceae, but might be currently the only marker to discriminate species for barcoding purpose (Michael H. J. Barfuss, unpublished data). Amplified fragments of MS and RPB2 are relatively short and information content is limited compared to other markers studied. Therefore these genomic regions were not further investigated for an enlarged set of samples. Although PGIC seems promising regarding ease of sequencing and alignment at various taxonomic levels within Bromeliaceae, it was not further tested for other samples of Tillandsioideae. Trees of NIA and XDH show limited resolution of deeper nodes (Figures 7, 8, respectively). However, several terminal clades receive bootstrap support, and both markers might be useful in adding more phylogenetic information in a combined analysis of nuclear DNA markers. Results of XDH show that this marker is better suited to higher taxonomic levels (as in Orchidaceae; Gorniák et al., 2010), whereas NIA might help better discriminate closely related taxa. Preliminary results indicated that PRK and PHYC were the two most promising nuclear DNA markers tested in terms of amplification and sequencing ease as well as in providing high numbers of PICs. Therefore these regions were sequenced for all available Tillandsioideae accessions. As for combined plastid data, individual nuclear DNA markers are unable to resolve relationships completely (Figure 9). Although several clades are retrieved and highly supported by most markers, they still provides different degrees of resolution and levels of support below the subfamily rank. PHYC is easy to align across the analyzed taxa, but it fails to provide resolution in more terminal nodes, especially in core Tillandsieae. PRK is difficult and not suitable, especially at higher levels, in the light of the numerous indels that have to be introduced into the alignment, but it helps to provide more resolution at intermediate levels. Recently diverged species To address questions concerning relationships of recently diverged species within genera of Tillandsioideae, our sampling was far too limited and turned out to be problematic in taxonomically complicated species complexes, especially when consensus allelic sequences are used. However, in the complete data sets of PRK and PHYC we sometimes included two or sometimes more accessions of the same taxon, which mostly cluster together or into clades of closely related species. We also found some cases where accessions of the same species but of a different subspecific rank did not cluster, although these accessions were still not far from each other. Whether these patterns are results of missing PICs, homoplasy, allelic consensus sequences, or questionable species concepts would need to be investigated further. It is clearly necessary to include multiple accessions of the same species/taxon, if possible over their whole distribution, and preferably complete gene sequences including all exons and introns to gain more linked information. In addition more cloning for divergent alleles of an individual, especially in complicated species complexes, is required to have pure allelic data. 111 BARFUSS, M.H.J. NUCLEAR DNA MARKERS Hybrids, hybrid speciation and reticulate evolution The two species suspected to be of hybrid origin, T. dorotheae and T. marconae, could be clearly confirmed by sequencing nuclear DNA. Detection was possible since parental taxa belong to different lineages within Tillandsia with relatively high sequence divergence. The two alleles of each sample cluster with high support with the clade of their parents. Parents of T. dorotheae are T. albertiana and a species of the T. argentina complex, that of T. marconae, T. landbeckii and a species of the T. purpurea complex. Most living plants in the botanical gardens or private collections are of vegetatively propagated origin from one or a few wild collections. Whether these species are allopolyploids, homoploid hybrids, or just F1 individuals of an occasional recent hybridization event would need to be investigated by more documented material. Except for these two cases, we currently do not have well-supported instances of hybridization, reticulate evolution or incomplete lineage sorting in other species. Acknowledgements The authors thank Tod F. Stuessy for providing lab facilities; many botanical gardens and individual people for providing plant material; Walter Till for taxonomic expertise, Mark W. Chase for a critically review of the manuscript. Financial support was provided by the University of Vienna, the Commission for Interdisciplinary Ecological Studies (KIÖS) at the Austrian Academy of Sciences (ÖAW) to W. Till and M.H.J. Barfuss (2007-02), and the SYNTHESYS (Synthesis of Systematic Resources) program funded by European Union. References Álvarez, I., Costa, A. & Feliner, G.N. 2008. Selecting single-copy nuclear genes for plant phylogenetics: A preliminary analysis for the Senecioneae (Asteraceae). J. Molec. Evol. 66: 276– 291. Baldwin, B.G., Sanderson, M.J., Porter, J.M., Wojciechowski, M.F., Campbell, C.S. & Donoghue, M.J. 1995. The ITS region of nuclear ribosomal DNA: A valuable source of evidence on angiosperm phylogeny. Ann. Missouri Bot. Gard. 82: 247–277. Barfuss, M.H.J., Samuel, R., Till, W. & Stuessy, T.F. 2005. Phylogenetic relationships in subfamily Tillandsioideae (Bromeliaceae) based on DNA sequence data from seven plastid regions. Amer. J. Bot. 92: 337–351. Blattner F.R. 1999. Direct amplification of the entire ITS region from poorly preserved plant material using recombinant PCR. BioTechniques 27: 1180–1186. Borsch T. & Quant, D. 2009. Mutational dynamics and phylogenetic utility of noncoding chloroplast DNA. Pl. Syst. Evol. 282: 169–199. Chapman, M.A., Chang, J., Weisman, D., Kesseli, R.V. & Burke, J.M. 2007. Universal markers for comparative mapping and phylogenetic analysis in the Asteraceae (Compositae). Theor. Appl. Gen. 115: 747–755. Chase, M.W., Williams, N.H., Donisete de Faria, A., Neubig, K.M., do Carmo E. Amaral, M. & Whitten, W.M. 2009. Floral convergence in Oncidiinae (Cymbidieae; Orchidaceae): an expanded concept of Gomesa and a new genus Nohawilliamsia. Ann. Bot. (Oxford) 104: 387–402. 112 IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 Chew, T. ,De Luna, E. & González, D. 2010. Phylogenetic relationships of the pseudobulbous Tillandsia species (Bromeliaceae) inferred from cladistic analyses of ITS 2, 5.8S ribosomal RNA gene, and ETS sequences. Syst. Bot. 35: 86–95. Clarkson, J.J., Kelly, L.J., Leitch, A.R., Knapp, S. & Chase, M.W. 2010. Nuclear glutamine synthetase evolution in Nicotiana: Phylogenetics and the origins of allotetraploid and homoploid (diploid) hybrids. Molec. Phylog. Evol. 55: 99–112. Crayn, D.M., Winter, K. & Smith, J.A.C. 2004. Multiple origins of crassulacean acid metabolism and the epiphytic habit in the Neotropical family Bromeliaceae. Proc. Natl. Acad. Sci. U.S.A. 101: 3703–3708. Cronn, R.C., Small, R.L., Haselkorn, T. & Wendel, J.F. 2002. Rapid diversification of the cotton genus (Gossypium: Malvaceae) revealed by analysis of sixteen nuclear and chloroplast genes. Amer. J. Bot. 89: 707–725. Denton, A.L., McConaughy, B.L. & Hall, B.D. 1998. Usefulness of RNA polymerase II coding sequences for estimation of green plant phylogeny. Molec. Biol. Evol. 15: 1082–1085. Douzery, E.J.P., Pridgeon, A.M., Kores, P., Linder, H.P., Kurzweil, H. & Chase, M.W. 1999. Molecular phylogenetics of Diseae (Orchidaceae): A contribution from nuclear ribosomal ITS sequences. Amer. J. Bot. 86: 887–899. Doyle, J.J. & Doyle, J.L. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem. Bull. Bot. Soc. Amer. 19: 11–15. Edgar, R.C. 2004a. MUSCLE: A multiple sequence alignment method with high accuracy and high throughput. Nucl. Acids Res. 32: 1792–1797. http://www.drive5.com/muscle/ Edgar, R.C. 2004b. MUSCLE: A multiple sequence alignment method with reduced time and space complexity. B.M.C. Bioinf. 5: 113. Emshwiller, E. & Doyle, J.J. 1999. Chloroplast-expressed glutamine synthetase (ncpGS): Potential utility for phylogenetic studies with an example from Oxalis (Oxalidaceae). Molec. Phylog. Evol. 12: 310–319. Felsenstein, J. 1985. Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39: 783–791. Ford, V.S., Lee, J., Baldwin, B.G. & Gottlieb, L.D. 2006. Species divergence and relationships in Stephanomeria (Compositae): PGIC phylogeny compared to prior biosystematic studies. Amer. J. Bot. 93: 480–490. Gaut, B.S., Muse, S.V., Clark, W.D. & Clegg, M.T. 1992. Relative rates of nucleotide substitution at the rbcL locus in monocotyledonous plants. J. Molec. Evol. 35: 292–303. Givnish, T.J., Barfuss, M.H.J., Van Ee, B., Riina, R., Schulte, K., Horres, R., Gonsiska, P.A., Jabaily, R.S., Crayn, D.M., Smith, J.A.C., Winter, K., Brown, G.K., Evans, T.M., Holst, B.K., Luther, H.E., Till, W., Zizka, G., Berry, P.E. & Sytsma, K.J. 2011. Phylogeny, adaptive radiation, and historical biogeography in Bromeliaceae: Insights from an eight-locus plastid phylogeny. Amer. J. Bot. 98: 872–895. Givnish, T.J., Millam, K.C., Berry, P.E. & Sytsma, K.J. 2007. Phylogeny, adaptive radiation, and historical biogeography of Bromeliaceae inferred from ndhF sequence data. Pp. 3–26 In: Columbus, J.T., Friar, E.A., Porter, J.M., Prince, L.M. & Simpson M.G. (eds.), Monocots: Comparative Biology and Evolution – Poales. Claremont: Rancho Santa Ana Botanic Garden (RSABG). 113 BARFUSS, M.H.J. NUCLEAR DNA MARKERS Givnish, T.J., Milliam, T.M., Evans, T.M., Hall, J.C., Berry, P.E. & Terry, R.G. 2004. Ancient vicariance or long-distance dispersal? Inferences about phylogeny and South AmericanAfrican disjunctions in Rapateaceae and Bromeliaceae based on ndhF sequence data. Int. J. Plant Sci. 165 (4, Suppl.), 35–54. Górniak, M., Paun, O. & Chase, M.W. 2010. Phylogenetic relationships within Orchidaceae based on a low-copy nuclear coding gene, XDH: Congruence with organellar and nuclear ribosomal DNA results. Molec. Phylog. Evol. 56: 784–795. Gruenstaeudl, M., Urtubey, E., Jansen, R.K., Samuel, R., Barfuss, M.H.J. & Stuessy, T.F. 2009. Phylogeny of Barnadesioideae (Asteraceae) inferred from DNA sequence data and morphology. Molec. Phylog. Evol. 51: 572–587. Hall, T.A. 1999. BioEdit: A user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl. Acids Symp. Ser. 41: 95–98. Hershkovitz, M.A. & Zimmer E.A. 1996. Conservation patterns in angiosperm ITS2 sequences. Nucleic Acids Res. 24: 2857–2867. Hřibová, E.,Čížková, J., Christelová, P., Taudien, S., de Langhe, E. & Doležel, J. 2011. The ITS15.8S-ITS2 sequence region in the Musaceae: Structure, diversity and use in molecular phylogeny. PLoS one 6: e17863 (1–11), doi: 10.1371/journal.pone.0017863. Horres, R., Zizka, G., Kahl, G. & Weising, K. 2000. Molecular phylogenetics of Bromeliaceae: Evidence from trnL (UAA) intron sequences of the chloroplast genome. Pl. Biol. 2: 306– 315. Hurst, L.D. & Merchant, A.R. 2001. High guanine-cytosine content is not an adaptation to high temperature: a comparative analysis amongst prokaryotes. Proc. Biol. Sci. 268: 493–497. Howarth, D.G. & Baum, D.A. 2002. Phylogenetic utility of a nuclear intron from nitrate reductase for the study of closely related plant species. Molec. Phylog. Evol. 23: 525–528. Jabaily, R.S. & Sytsma, K.J. 2010. Phylogenetics of Puya (Bromeliaceae): Placement, major lineages, and evolution of Chilean species. Amer. J. Bot. 97: 337–356. Jobes, D.V. & Thien, L.B. 1997. A conserved motif in the 5.8S ribosomal RNA (rRNA) gene is a useful diagnostic marker for plant internal transcribed spacer (ITS) sequence. Plant Mol Biol Rep 15: 326–334. Kelchner, S.A. 2000. The evolution of noncoding chloroplast DNA and its application in plant systematics. Ann. Missouri Bot. Gard. 87: 482–498. Lewis, C.E. & Doyle, J.J. 2001. Phylogenetic Utility of the Nuclear Gene Malate Synthase in the Palm Family (Arecaceae). Molec. Phylogen. Evol. 19: 409–420. Lewis, C.E. & Doyle, J.J. 2002. A phylogenetic analysis of tribe Areceae (Arecaceae) using two low-copy nuclear genes. Pl. Syst. Evol. 236: 1–17. Liu, J.-S. & Schardl, C.L. 1994. A conserved sequence in internal transcribed spacer 1 of plant nuclear rRNA genes. Plant Mol. Biol. 26: 775–778. Lukas, B. 2010. Molekulare Phylogenie und Phytochemie der Gattung Origanum. Dissertation, University of Vienna. Ma, J.X., Li, Y.N., Vogl, C., Ehrendorfer, F. & Guo, Y.P. 2010. Allopolyploid speciation and ongoing backcrossing between diploid progenitor and tetraploid progeny lineages in the Achillea millefolium species complex: Analyses of single-copy nuclear genes and genomic AFLP. B.M.C. Evol. Biol.10: 100 (1–11), doi:10.1186/1471-2148-10-100. Mathews, S. & Donoghue, M.J. 1999. The root of angiosperm phylogeny inferred from duplicate phytochrome genes. Science 286: 947–950. 114 IN TILLANDSIOIDEAE (BROMELIACEAE) PART 2, CHAPTER 3 Morton, C.M. 2011. Newly sequenced nuclear gene (XDH) for inferring angiosperm phylogeny. Ann. Missouri Bot. Gard. 98: 63–89. Oxelman, B., Yoshikawa, N., McConaughy, B.L., Luo, J., Denton, A.L. & Hall, B.D. 2004. RPB2 gene phylogeny in flowering plants, with particular emphasis on asterids. Molec. Phylog. Evol. 32: 462–479. Piñeiro, R., Costa, A., Aguilar, J.F. & Feliner, G.N. 2009. Overcoming paralogy and incomplete lineage sorting to detect a phylogeographic signal: A GapC study of Armeria pungens. Botany-Botanique 87: 164–177. Russell, A., Samuel, R., Klejna, V., Barfuss, M.H.J., Rupp, B. & Chase, M.W. 2010. Reticulate evolution in diploid and tetraploid species of Polystachya (Orchidaceae) as shown by plastid DNA sequences and low-copy nuclear genes. Ann. Bot. (Oxford) 106: 37–56. Safer, S. (2011). Molecular and phytochemical investigations on the genus Leontopodium. Dissertation, University of Innsbruck. Samuel, R., Kathriarachchi, H., Hoffmann, P., Barfuss, M.H.J., Wurdack, K.J., Davis, C.C. & Chase, M.W. 2005. Molecular phylogenetics of Phyllanthaceae: Evidence from plastid matK and nuclear PHYC sequences. Amer. J. Bot. 92: 132–141. Sang, T. 2002. Utility of low-copy nuclear gene sequences in plant phylogenetics. Critical Reviews in Biochemistry and Molecular Biology 37: 121–147. Sang, T., Pan, J., Zhang, D., Ferguson, D., Wang, C., Pan, K.-Y. & Hong, D.-Y. 2004. Origins of polyploids: an example from peonies (Paeonia)and a model for angiosperms. Biological Journal of the Linnean Society. 82: 561–571. Sang, T. & Zhang, D. 1999. Reconstructing hybrid speciation using sequences of low-copy nuclear genes: Hybrid origins of five Paeonia species based on Adh gene phylogenies. Syst. Bot. 24: 148–163. Sass, C. & Specht, C.D. 2010. Phylogenetic estimation of the core bromelioids with an emphasis on the genus Aechmea (Bromeliaceae). Mol. Phyl. Evol. 55: 559–571. Schulte, K., Barfuss, M.H.J. & Zizka, G. 2009. Phylogeny of Bromelioideae (Bromeliaceae) inferred from nuclear and plastid DNA loci reveals the evolution of the tank habit within the subfamily. Molec. Phylogen. Evol. 51: 327–339. Schulte, K. & Zizka, G. 2008. Multi locus plastid phylogeny of Bromelioideae (Bromeliaceae) and the taxonomic utility of petal appendages and pollen characters. Candollea 63: 209– 225. Shaw, J., Lickey, E.B., Beck, J.T., Farmer, S.B., Liu, W.S., Miller, J., Siripun, K.C., Winder, C.T., Schilling, E.E. & Small, R.L. 2005. The tortoise and the hare II: Relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis. Amer. J. Bot. 92: 142– 166. Shaw, J., Lickey, E.B., Schilling, E.E. & Small, R.L. 2007. Comparison of whole chloroplast genome sequences to choose noncoding regions for phylogenetic studies in angiosperms: The tortoise and the hare III. Amer. J. Bot. 94: 275–288. Steger, G. 1994. Thermal denaturation of double-stranded nucleic acids: prediction of temperatures critical for gradient gel electrophoresis and polymerase chain reaction. Nucleic Acids Res. 22: 2760–2768. Strand, A.E., LeebensMack, J. & Milligan, B.G. 1997. Nuclear DNA-based markers for plant evolutionary biology. Molecular Ecology 6: 113–118. Su, X.-Z., Wu, Y., Sifri, C.D. & Wellems, T.E. 1996. Reduced extension temperatures required for PCR amplification of extremely A+T-rich DNA. Nucleic Acids Research 24: 1574–1575. 115 Sun, Y., Skinner, D.Z., Liang, G.H. & Hulbert, S.H. 1994. Phylogenetic analysis of Sorghum and related taxa using Internal Transcribed Spacers of nuclear ribosomal DNA. Theor. Appl. Genet. 89: 26–32. Swofford, D.L. 2003. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4.0b10. Sunderland: Sinauer Associates. Tank, D.C. & Sang, T. 2001. Phylogenetic utility of the glycerol-3-phosphate acyltransferase gene: Evolution and implications in Paeonia (Paeoniaceae). Molec. Phylog. Evol. 19: 421– 429. Tel-Zur, N., Abbo, S., Myslabodski, D. & Mizrahi, Y. 1999. Modified CTAB procedure for DNA isolation from epiphytic cacti of the genera Hylocereus and Selenicereus (Cactaceae). Pl. Molec. Biol. Reporter 17: 249–254. Terauchi, R., Terachi, T. & Miyashita, N.T. 1997. DNA polymorphism at the Pgi locus of a Wild Yam, Dioscorea tokoro. Genetics 147: 1899–1914. Terry, R.G., Brown, G.K. & Olmstead, R.G. 1997a. Examination of subfamilial phylogeny in Bromeliaceae using comparative sequencing of the plastid locus ndhF. Amer. J. Bot. 84: 664–670. Terry, R.G., Brown, G.K. & Olmstead, R.G. 1997b. Phylogenetic relationships in subfamily Tillandsioideae (Bromeliaceae) using ndhF sequences. Syst. Bot. 22: 333–345. Thomas, M.M., Garwood, N.C., Baker, W.J., Henderson, S.A., Russell, S.J., Hodel, D.R. & Bateman, R.M. 2006. Molecular phylogeny of the palm genus Chamaedorea, based on the low-copy nuclear genes PRK and RPB2. Molec. Phylogen. Evol. 38: 398–415. Vaezi, J. & Brouillet, L. 2009. Phylogenetic relationships among diploid species of Symphyotrichum (Asteraceae: Astereae) based on two nuclear markers, ITS and GAPDH. Molec. Phylog. Evol. 51: 540–553. Versieux, L.M., Barbará, T., Wanderley, M.d.G.L., Calvente, A., Fay, M.F. & Lexer, C. 2012. Molecular phylogenetics of the Brazilian giant bromeliads (Alcantarea, Bromeliaceae): implications for morphological evolution and biogeography. Molec. Phylog. Evol. 64: 177– 189. Wendel, J.F. 2000. Genome evolution in polyploids. Plant Molecular Biology 42: 225–249. Werle, E., Schneider, C., Renner, M., Volker, M. & Fiehn, W. 1994. Convenient single-step, one tube purification of PCR products for direct sequencing. Nucleic Acids Research 22: 4354–4355. White, T.J., Bruns, T., Lee, S. & Taylor, J. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics, Pp. 315–322 in Innis, M.A., Gelfand, D.H., Sninsky, J.J. & White, T.J. (eds.), PCR protocols: A guide to methods and applications. San Diego: Academic Press. Willyard, A., Cronn. R. & Liston, A. 2009. Reticulate evolution and incomplete lineage sorting among the ponderosa pines. Molec. Phylog. Evol. 52: 498–511. Yockteng, R. & Nadot, S. 2004. Phylogenetic relationships among Passiflora species based on the glutamine synthetase nuclear gene expressed in chloroplast (ncpGS). Molec. Phylog. Evol. 31: 379–396. Zuker, M. 2003. Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res 31: 3406–3415. 116 Part 3 Classification of Bromeliaceae subfamily Tillandsioideae Chapter 4 Progress towards a new classification of Tillandsioideae Walter Till & Michael H.J. Barfuss Status: published, Journal of the Bromeliad Society 56 (6): 253–259. 2006. Contribution: data collection, data analysis, manuscript writing/editing 121 BARFUSS, M.H.J. 122 PROGRESS TOWARDS A NEW CLASSIFICATION OF TILLANDSIOIDEAE PART 3, CHAPTER 4 123 BARFUSS, M.H.J. 124 PROGRESS TOWARDS A NEW CLASSIFICATION OF TILLANDSIOIDEAE PART 3, CHAPTER 4 125 BARFUSS, M.H.J. 126 PROGRESS TOWARDS A NEW CLASSIFICATION OF TILLANDSIOIDEAE PART 3, CHAPTER 4 127 Chapter 5 A new classification of Bromeliaceae subfamily Tillandsioideae inferred from DNA sequences data of two genomes and morphology Michael H.J. Barfuss, Walter Till & Rosabelle Samuel University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria Author for correspondence: Michael H.J. Barfuss, michael.h.j.barfuss@univie.ac.at Keywords: Alcantarea; Catopsis; Cipuropsis; plastid DNA; Glomeropitcairnia; Guzmania; Josemania; Lectotypifications; Lemeltonia; Mezobromelia; PHYC; PRK; Racinaea; R. subg. Pseudophytarrhiza; Rothowia; low-copy nuclear DNA; Tillandsia; T. subg. Pseudovriesea; T. subg. Viridantha; Vriesea, Werauhia Running Title: New classification of Tillandsioideae (Bromeliaceae) Status: in preparation, intended to be submitted to Taxon Contribution: data collection, data analyses, manuscript writing Abstract In order to establish a natural and stable classification system for Tillandsioideae we conducted phylogenetic analyses of newly generated low-copy nuclear DNA sequence data (PHYC, PRK) and previously published plastid DNA markers (atpB-rbcL, matK, rbcL, partial rbcL-accD, rps16 intron, partial trnK intron, trnL intron, trnL-trnF), together with a re-evaluation of morphological characters. Parsimony analysis revealed the following, mostly well-supported general branching pattern for Tillandsioideae: ((Glomeropitcairnia, Catopsis) (((Alcantarea, Vriesea) (Werauhia, Cipuropsis-Mezobromelia clade)) (Guzmania, ((Josemania, Racinaea) (Rothowia, (Lemeltonia, Tillandsia)))))). Core Tillandsioideae are well supported by all analyses, but Bayesian analysis yielded a slightly different branching pattern for core Tillandsieae: (Guzmania, ((Josemania, Rothowia) (Racinaea (Lemeltonia, Tillandsia))). Based on the results a new classification of Tillansioideae is presented. Two new subtribes (Cipuropsidinae, Vrieseinae), three new genera (Josemania, Lemeltonia, Rothowia), and three new subgenera (Racinaea subg. Pseudophytarrhiza, Tillandsia subg. Viridantha, T. subg. Pseudovriesea) are described and several species are reclassified. A key to the genera of Tillandsioideae is also provided. Lectotypes are selected for Catopsis subg. Tridynandra, Tillandsia sect. Conostachys, and Tillandsia sect. Eriophyllum. Classification of some unresolved phylogenetic units remains informal and needs further attention, especially within the Cipuropsis-Mezobromelia clade and the genus Tillandsia. 131 Introduction Tillandsioideae Burnett is the largest and morphologically most diverse subfamily of Bromeliaceae Juss. with more than 1300 species in nine generally accepted genera (Smith & Downs, 1977; Smith & Till, 1998; Till, 2000a; Grant & Zijlstra, 1998; Luther & Rabinowitz, 2010; species numbers according to Luther, 2010): Alcantarea (E. Morren ex Mez) Harms (28 spp.), Catopsis Griseb. (18 spp.), Glomeropitcairnia (Mez) Mez (2 spp.), Guzmania Ruiz & Pav. (210 spp.), Mezobromelia L.B. Sm. (9 spp.), Racinaea M.A. Spencer & L.B. Sm. (65 spp.), Tillandsia L. (626 spp., in 6 subgenera), Vriesea Lindl. (266 spp., in two sections), and Werauhia J.R. Grant (87 spp., in 2 sections). The genus Viridantha Espejo (8 spp.; Espejo-Serna, 2002; Espejo-Serna & al., 2007; López-Ferrari & Espejo-Serna, 2009) is included in Tillandsia, following most recent authors, and is subdivided into two sections, sect. Viridantha (5 spp.) and sect. Caulescens Espejo (3 spp.). A well-established taxonomic concept of the whole subfamily is of much interest, because tillandsioids are an important component of neotropical ecosystems accounting for a high portion of recorded (epiphytic) biodiversity, are of special importance for conservationists since many species are endemic to certain regions, have the greatest number of horticulturally important Bromeliaceae species, and are popular amongst bromeliad collectors and amateur botanists. Despite its importance, a well-supported and broadly accepted classification scheme has not been achieved to date. The main reasons are the lack of definite morphological characters to delimit genera, subgenera and sections, and the low variability of DNA markers studied up to now, weakening the possibility to reconstruct a solid phylogenetic framework with wellsupported, monophyletic entities (see Hörandl & Stuessy (2010) for the definitions of evolutionary terms used in this study). Traditional morphological classifications A large number of species in the subfamily exhibit considerable interspecific morphological variation especially in reproductive organs (e.g., inflorescences and flowers) and in general habit, with life forms ranging from entirely mesophytic and water-impounding to strongly xerophytic, rarely terrestrial to commonly epiphytic or lithophytic (e.g., Gilmartin, 1983; Gilmartin & Brown, 1986; Benzing, 2000; Stefano & al., 2008). Early bromeliad taxonomists considered floral morphological characters to be most important for classification, since there is sometimes no diagnostic variation in the vegetative structures of distantly related species (Baker, 1889; Mez, 1896, 1934–35; Smith, 1951; Smith & Downs, 1977). Smith & Downs, authors of the last monograph of Bromeliaceae (Pitcairnioideae Harms: 1974; Tillandsioideae: 1977; Bromelioideae: 1979) maintained the use of these floral characteristics in their taxonomic treatments: the four main characters used to differentiate genera of Tillandsioideae are the position of the ovary, the structure of seeds, the presence vs. absence of petal appendages, and free vs. conglutinate petals. However, in their "Preface" Smith & Downs (1977: 663) state about the difficulty of identifying Tillandsioideae genera "once the crucial petals and stamens are lost" and the sometimes inconsistent placement of species by Mez (1934–35; cf. the different treatment of several xerophytic, grey-leaved Vriesea species by both authors). Therefore, many species are misplaced in both monographs according to these generic definitions, since their classification was often based on herbarium specimens without perfect flowers (e.g., Utley, 1978; Weber & Smith, 1983; Utley & Luther, 1991; Grant, 1993a). 133 BARFUSS, M.H.J. NEW CLASSIFICATION Revisions of floral morphological characters used and neglected in Smith & Down's monographs pointed out the potential and limits of these features to classify Bromeliaceae species (petal appendages: Brown & Terry, 1992; stigmas and papillae: Brown & Gilmartin, 1984, 1989b; Schill & al., 1988; Gortan, 1991; septal nectaries: Böhme, 1988; pollen: Halbritter, 1988, 1992; and seeds: Gross, 1988). No taxonomic change was proposed in any of these studies, because species sampling was too limited to generally apply recognized characteristics to higher taxonomic units. Resurrections and segregations of subgenera and closely related species complexes at generic status were done later based on the consideration of these additional and reevaluated morphological characters (Alcantarea: Grant, 1995a; Racinaea: Spencer & Smith, 1993; Werauhia: Grant, 1995a; and Viridantha: Espejo-Serna, 2002). Some of these new genera are not always accepted, because remaining species of Tillandsia and Vriesea are left in their old generic circumscriptions without alteration of their morphological definitions. Even more surprising and hardly acceptable for some bromeliad specialists are rearrangements of former xerophytic, grey-leaved Vriesea species done by Grant (1993b, 1994b, 1995b, 2005), since the new placement clearly contradicts the definition of the genus Tillandsia established by earlier botanists, although for example Mez (1934–35) did not always follow strictly this characterization when new morphological evidence arose. The controversies about the correct generic placement of certain species have led to the situation that modern bromeliad taxonomists follow different opinions regarding a generally accepted classification scheme (Grant & Zijlstra, 1998; Smith & Till, 1998; Till, 2000a, b; Barfuss & al., 2005; Luther & Sieff, 1994, 1997; Luther, 2001, 2008, 2010; Luther & Rabinowitz, 2010), since some of these modifications were based on morphological dissimilarity and not on a shared character or character combinations (see Tables 4 and 5 for a comparison of traditional and most recent classification systems). The most important morphological character at the generic level within Tillandsioideae (and Bromeliaceae in general) is the presence vs. absence of petal appendages (Smith & Downs, 1974, 1977, 1979; Brown & Terry, 1992), which delimits the genus pairs Tillandsia/Vriesea and Guzmania/Mezobromelia. The taxonomic value of this character has been questioned (e.g., Read, 1968; Utley, 1978; Gardner, 1982; Gilmartin, 1983; Beaman, 1989) although all bromeliad monographers used it in their taxonomic treatments (Wittmack, 1888; Baker, 1889; Mez, 1896; Harms, 1930; Mez, 1934–35; Smith & Downs, 1977). Field botanists discovered that within populations of a single species, individuals with and without petal appendages can coexist (Read, 1968; W. Till, pers. com.), which would mean, following the strict generic concept of Smith and Downs (1977), these individuals should be classified in different genera. Despite these problems with using petal appendages as a strict and definite character, Benzing & al. (2000) countered that "...the frequently disparaged petal appendage would regain some lost currency as one of the more useful among the traditional characters... [if] occasional exceptions" are accepted. Defining infrageneric units morphologically (e.g., within Tillandsia: Till, 2000b) is an analogous difficulty. Only a few classical taxonomic studies have treated whole subgenera or species complexes of Tillandsia: T. subg. Anoplophytum (Tardivo, 2002), T. subg. Diaphoranthema (Till, 1984, 1992), T. subg. Phytarrhiza (Gilmartin, 1983; Gilmartin & Brown, 1986), T. subg. Pseudalcantarea (Beaman, 1989; Beaman & Judd, 1996), T. subg. Tillandsia (Gardner, 1982, 1986a, b; including some taxa from T. subg. Allardtia sensu Smith & Downs, 1977), the Tillandsia gardneri complex (Ehlers, 1997), Tillandsia macdougallii complex (Granados Mendoza, 2008), the Tillandsia plumosa complex (= Viridantha) (Espejo-Serna, 2002; Ehlers, 2009), and the Tillandsia tectorum complex (Hromadnik, 2005). Tillandsia subg. Allardtia was lastly only treated 134 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 within the "Tillandsia and Racinaea" chapter by Till (2000b). The separation of T. subg. Allardtia and T. subg. Anoplophytum is based on the character "straight vs. plicate filaments" and T. subg. Tillandsia is distinguished from both by its "exserted stamens and style". Till (2000b) stated that "the distinction between subgenera Allardtia and Anoplophytum is weak, and their separation may not be justified, at least not as proposed by Smith & Downs" (1977). The existence of closely related species (Till, 2000b) showing either of these subgeneric characteristics (straight vs. plicate filaments, respectively) and previous ontogenetic studies on filament plication (Evans & Brown, 1989) demonstrate that this distinction is inappropriate. This is reflected by the opinion of Till (2000b), where he moved the Andean species with plicate filaments from T. subg. Anoplophytum sensu Smith & Downs (1977) into T. subg. Allardtia. Several species of other subgenera of Tillandsia, especially of T. subg. Pseudalcantarea and T. subg. Phytarrhiza sensu Smith & Downs (1977), show similar patterns of morphological convergence or parallelism. Since traditional generic, subgeneric and sectional treatments are exclusively based on flower morphology, the most plausible explanation for their evolution is that pollination syndromes can adapt relatively fast to new environmental situations, leading to floral characteristics which could have evolved in parallel or convergently in different phylogenetic lineages. This might be caused by a high capability of bromeliads in general to adapt easily to different ecological niches and modified environmental conditions (e.g., Gardner, 1986a; Benzing, 2000; Kessler, 2002; Schmidt-Lebuhn & al., 2007; Krömer & al., 2008). Despite known problematic issues of previously-used floral morphological characters, their usage still persists, since no other diagnostic characters or character combinations have yet been identified or are sufficiently known to unambiguously circumscribe natural taxonomic units. Previous molecular phylogenetic studies Molecular phylogenetic studies which included members of Tillandsioideae revealed a holophyletic (= monophyletic s.str.; see Hörandl & Stuessy, 2010) subfam. Tillandsioideae (e.g., Terry & Brown, 1996, Terry & al., 1997a, b; Horres & al. 2000; Crayn & al., 2004; Givnish & al., 2004; Givnish & al., 2007; Givnish & al., 2011). Also, the early diverging clades comprising Catopsis and Glomeropitcairnia were well supported in most studies. However, resolution in core Tillandsioideae genera (Alcantarea, Guzmania, Mezobromelia, Racinaea, Tillandsia, Viridantha, Vriesea, and Werauhia; = Tillandsioideae s.str., Terry & al., 1997b) was greatly lacking, because phylogenetic studies only relied on a single or few plastid DNA regions. First reliable insights into relationships of core Tillandsioideae were brought by the seven marker analyses of Barfuss & al. (2005). The main results of this study were the split of core Tillandsioideae into two main lineages and the para- and/or polyphyly of the largest tillandsioid genera Tillandsia and Vriesea based on the information content of these plastid DNA markers. All other lastly segregated genera were found to be holophyletic, but they were clearly nested either within Vriesea (Alcantarea, Werauhia) or within Tillandsia (Racinaea, Viridantha) according to their traditional generic circumscriptions. Taxonomic changes proposed since the monograph of Smith & Downs's (1977) were only partly supported and clearly showed, that morphology based taxonomy is in conflict with molecular results. However, the taxon sampling was too limited for making a conclusive decision regarding holophyly/paraphyly of these genera. The study also showed that none of the previously applied infrageneric concepts of Tillandsia, Vriesea, and Werauhia using traditional morphological characters were fully supported by plastid DNA sequence data indicating convergent or parallel evolution of certain features. A 135 BARFUSS, M.H.J. NEW CLASSIFICATION preliminary report of Till & Barfuss (2006) including results of one single-copy nuclear gene (Phosphoribulokinase = PRK; not explicitly stated in the text but used to create Figures 1 and 2 of Till & Barfuss, 2006; 254–255) further corroborates these findings. However, the nested position within Tillandsia of meso- and semi-mesophytic taxa of T. subg. Phytarrhiza and Racinaea is no longer supported whereas Viridantha still remains nested within Tillandsia. Nevertheless, the shortcomings of all previous DNA studies were that they suffered from an extremely low number of parsimony-informative nucleotide characters (PICs), a serious undersampling of taxa in relation to the actually accepted species and the dependence on few analyzed plastid DNA regions. Despite the split of Tillandsioideae into 4 main lineages, which were considered as separate tribes (Pogospermeae Brongn. (1864) = Catopsideae Harms (1930), Glomeropitcairnieae Harms, Tillandsieae Rchb., Vrieseeae W. Till & Barfuss) in Barfuss & al. (2005) and the transfer of W. insignis (Mez) W. Till, Barfuss & Samuel in Barfuss & al. (2004) no further taxonomic conclusions were drawn in any previous phylogenetic study, leaving the most interesting questions unanswered: what are the generic and infrageneric units within Tillandsioideae and how can these be defined genetically and morphologically? Aiming at answering these question we conducted phylogenetic analyses of previouslyused plastid and newly-generated nuclear DNA sequences, i.e., (1) plastid atpB-rbcL, matK, rbcL, partial rbcL-accD, rps16 intron, partial trnK intron, trnL intron, and trnL-trnF (taken from Barfuss & al., 2005) and (2) nuclear PHYC (phytochrome C) and PRK (phosphoribulokinase) together with a re-evaluation of potentially useful morphological characters of the habit and various floral elements. Materials and Methods Taxon selection Plant material was primarily selected based on results of Barfuss & al. (2005) and the taxonomic treatment of Smith & Downs (1977). Ninety-seven accessions out of 122 from the first study together with 347 new accessions were included in the present investigation, making a total of 444 individuals of tillandsioid species. A complete list of all plant material examined including nomenclatoric authors can be found in the supplementary data (Appendix). Sampling within recognized genera, subgenera and sections was significantly increased: Alcantarea (23 acc.), Catopsis (10 acc.); Glomeropitcairnia (3 acc.), Guzmania (52 acc.), Mezobromelia (7 acc.), Racinaea (23 acc.), Vriesea sect. Vriesea (29 acc.), V. sect. Xiphion p.p., typo excluso (23 acc.), Werauhia sect. W. (9 acc.), and W. sect. Jutleya (7 acc.); the main focus has been within the species-richest genus Tillandsia (258 acc.): T. subg. Allardtia (100 acc.), T. subg. Anoplophytum (44 acc.), T. subg. Phytarrhiza (meso- and semi-mesophytic: 31 acc., xerophytic: 10 acc.), T. subg. Pseudalcantarea (7 acc.), T. subg. Diaphoranthema (12 acc.), and T. subg. Tillandsia (54 acc.). Members of T. subg. Pseudalcantarea and meso- and semi-mesophytic taxa of T. subg. Phytarrhiza were nearly completely sampled (missing only T. narthecioides Presl), to clarify issues of high levels of polyphyly within these subgenera (Gilmartin & Brown, 1986; Beaman & Judd, 1996; Till, 2000b), and to test for either an early split of species from the rest of Tillandsia (T. subg. Pseudalcantarea) or affinities to the genus Racinaea (meso- and semimesophytic T. subg. Phytarrhiza) based on earlier plastid DNA results (Barfuss & al., 2005). Additional tillandsioid taxa were chosen to include species with a distinct morphology, which 136 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 indicates an inadequate generic, subgeneric or sectional placement by Smith & Downs (1977), an ancestral position within recognized taxa or probable separate phylogenetic units within Tillandsioideae, e.g., the large, broad-leaved and water-impounding or three-pinnate species of T. subg. Tillandsia occurring in north-western South America (e.g., Till, 2000b; León & Sagástegui, 2008), the former xerophytic, grey-leaved Vriesea species (e.g., Grant, 1993b), or the Tillandsia tectorum complex (e.g., Hromadnik, 2005) which is associated with the Tillandsia plumosa complex (= Viridantha) in an earlier study (Barfuss & al., 2005). As an outgroup accessions of Glomeropitcairnia and Catopsis were used, since in all previous investigations these genera came out as two early diverging lineages, both being sister to each other and together being sister to the core Tillandsioideae. Relevant type species for genera and infrageneric units were investigated for nearly all taxa except Racinaea and Cipuropsis, i.e., R. cuspidata (L.B. Sm.) M.A. Spencer & L.B. Sm. and Ci. subandina Ule (syn. V. subandina (Ule) L.B. Sm. & Pittendr.), respectively. Most material came from the botanical garden of the University of Vienna and the private collection of H. & L. Hromadnik. Additional samples were donated by the botanical gardens of Berlin-Dahlem, Göttingen, Heidelberg, Kew, München, Linz, and Sarasota and the private collections of W. Adlassnig, R. Ehlers, P. Lechner, E.M.C. Leme, J.M. Manzanares, J.P. Pinzon, M. Speckmaier and M. Winkler. Some accessions were also collected by the first and/or second author during scientific fieldtrips and/or student expeditions to Venezuela (2000), the Dominican Republic (2001, 2002), and Ecuador (2003, 2004, 2006). Species nomenclature and infrageneric classification basically follows Smith & Downs (1977) with alterations by Luther & Sieff (1994, 1997), Luther (2001, 2008, 2010), Luther & Rabinowitz (2010) and Grant (1993b, 1994b, 1995b, 2005). Additional taxa accepted, but classified differently or missing in the taxonomic literature followed are Guzmania fusispica Mez & Sodiro [≠ G. osyana (E. Morren) Mez], Tillandsia buseri var. nubicola Gilmartin [≠ T. buseri Mez], T. fosteri Gilmartin [≠ T. demissa L.B. Sm.], T. macropetala Wawra [≠ T. viridiflora (Beer) Baker], T. malzinei (E. Morren) Baker [= Vriesea malzinei E. Morren], T. recurvifolia var. subsecundifolia (W. Weber & Ehlers) W. Till [≠ T. leonamiana E. Pereira], and the recently described new species Racinaea tillii Manzan. & Gouda (Manzanares & Gouda, 2010). However, from the results chapter onwards the new nomenclature is used, which is presented in the chapter "Classification of Tillandsioideae" and can be visualized in Figure 2 (see Appendix) and Tables 4 and 5. Gene selection Data from seven plastid DNA regions were taken from a preceding study (Barfuss & al., 2005). No additional plastid sequences were generated, because the original sampling already covered most of the accepted taxonomic lineages and preference was given to new nuclear DNA sequences. Eight nuclear gene regions were initially screened for a subset of Tillandsioideae accessions used in Barfuss & al. (2005) to select appropriate genomic regions for sequencing the more exhaustive sample set. DNA regions surveyed included the multi-copy internal transcribed spacer of the nuclear ribosomal DNA (ITS nrDNA; e.g., Chew & al., 2010), and the lowcopy nuclear genes glucose-6-phosphate isomerase, cytosolic (PGIC; e.g., Ford & al., 2006), malate synthase (MS; e.g., Lewis & Doyle, 2001, 2002), nitrate reductase (NIA: e.g., Howarth & Baum, 2002), phosphoribulokinase (PRK; e.g., Lewis & Doyle, 2002; Thomas & al., 2006), phytochrome C (PHYC; e.g., Mathews & Donoghue, 1999; Samuel & al., 2005), RPB2 (RNA polymerase II, beta subunit; e.g., Denton & al., 1998; Oxelman & al., 2004), and xanthine dehydrogenase (XDH; e.g., Górniak & al., 2010). Parameters for the selection of effective DNA regions 137 BARFUSS, M.H.J. NEW CLASSIFICATION were (1) the number of loci obtained, (2) the performance of primers and PCR conditions for both amplification and sequencing, (3) the occurrence of homopolymers and microsatellites, which are often hard to sequence, (4) the length of the PCR products and (5) the number of potentially phylogenetically informative characters. The primary target was to obtain one or more genomic regions which are either well homogenized (ITS nrDNA) or effectively singlecopy (other nuclear regions) to be mostly sequenced directly. PHYC and PRK were finally selected for phylogenetic analyses (see "Results" for the reasons and chapter 3). All nuclear sequences where generated newly for this study and have been archived in NCBI's GenBank (http://www.ncbi.nlm.nih.gov/genbank). Molecular data and analyses Total DNA extractions were done from fresh or silicagel-dried material mainly following a sorbitol/CTAB-based method for difficult plant tissue (Tel-Zur & al., 1999) with modifications according to Russell & al. (2010a). In some cases the 2x CTAB procedure described by Doyle & Doyle (1987) adopted for mini columns, the DNeasy® Plant Mini Kit (QIAGEN®) following the manufacturers protocol, or a combination of both were used; details for the latter can be requested from the first author. Initial primer sequences and PCR conditions for PRK and PHYC were described by Schulte & al. (2009) and Russell & al. (2010b), respectively. Details for general PCR and sequencing reactions, occasional cloning and other laboratory materials and methods followed chapter 3 of this thesis. Both the PHYC and PRK genes were amplified and sequenced using sometimes slightly modified or additional primers. Details of all primers used are given in Tables 1 (PHYC) and 2 (PRK) (see also chapter 3 for more primer details). Raw sequences were initially analysed and edited using the Sequencing Analysis Software v5.3 (Applied Biosystems®, Life Technologies™ Corporation). A contig of forward and reverse DNA strands from both PCR and/or internal sequencing primers was generated with the SeqMan Pro module of the Lasergene® v8.1 software package (DNASTAR, Inc.). Sequences were edited and a consensus was exported as text file in fasta format. Usually fasta files were pooled together and aligned initially with MUSCLE v3.8 (Edgar, 2004a, b) and then adjusted by eye with the BioEdit v7.0.5 (Hall, 1999) following the guidelines of Kelchner (2000) and Borsch & Quandt (2009). In very polymorphic DNA regions containing numerous indels related taxa were aligned first. Relationships were pre-estimated from previous results or preliminary phylogenetic analyses. Only then were individual alignments aligned together using the profile alignment function of MUSCLE and adjusted by eye using BioEdit. Aligned matrices can be obtained from the first author upon request. When allelic length variation between priming sites was observed, editing of sequences coming from both directions was not straightforward and done in such a way that the missing DNA sequence parts of one allele (largely small inserted repeats or deletions) were ignored and allelic consensus sequences were used further. This was accomplished by identifying the type of indel and extracting the longer allele from shifted electropherograms, which are a result of non-homologous parts of the two different alleles to be overlaid. If the number of indels was too high with two or more indels between priming sites in both alleles, editing of sequences became difficult and doubtful, so that in such cases PCR products were reinvestigated with a cloning step included. Edited indel containing consensus sequences were only included for analyses, when prior evaluation steps on polymorphic sites (see below) indicated that allelic indels are also most likely due to variation within a single species or within a closely related species complex. 138 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 Table 1. Primers used for PHYC with their location given according to base positions in exon 1 of the PHYC reference sequence of Oryza sativa (AF141942). name sequence (5'–3') position (direction) comments reference phyc515f-br AAGCCCTTYTACGCTATCCTGCACCG 490–515 (up) Bromeliad-specific PCR primer This study phyc524f-br GCTATCCTGCACCGGATCGAYGT 502–524 (up) Bromeliad-specific internal sequencing primer This study phyc974f-br GCTCCTCACGGCTGCCACGCTCA 952–974 (up) Bromeliad-specific internal sequencing primer This study phyc1145r2-mo CAACAGGAACTCACAAGCATATC 1,167–1,145 (down) monocot-specific internal sequencing primer This study phyc1690r-br TCAACATCTTCCCAYGGGAGGCT 1,712–1,690 (down) Bromeliad-specific internal sequencing primer This study phyc1699r-br ATWGCATCCATTTCAACATCTTCCCA 1,724–1,699 (down) Bromeliad-specific PCR primer This study Table 2. Primers used for PRK with their location given according to base positions in exon regions of the PRK reference sequence of Oryza sativa (NM_001054360); exon 1 (1–545), exon 2 (546–630), exon 3 (631–715), exon 4 (716– 960), exon 5 (961–1,212).Bold italicized numbers denote positions where primers are located partly within intron parts of PRK. name sequence (5’–3’) position (direction) comment reference prk621f TCAGCAATGAGGTTAAATTTGCATGG exon 2, 596–621 (up) Bromeliad-specific PCR primer This study prk630f AAATTTGCATGGAAAATTCAGGTC exon 2, intron 2, 610–633 (up) Bromeliad-specific internal sequencing primer This study prk734f CTGCAGATCCGCAGAAGAAATATGC intron 3, exon 4, 710–735 (up) Bromeliad-specific internal sequencing primer Schulte & al., 2009 prk890r GGGTATGAGCATGTCAATTTCCTCC exon 4, 914–890 (down) Bromeliad-specific internal sequencing primer This study prk1057r CTTCAGCATTTGTTGTGTCACCTC exon 5, 1,080–1,057 (down) Bromeliad-specific internal sequencing primer This study prk1069r2 GGAAAATCTGCRTGCTTCAGCATTTG exon 5, 1,094–1,069 (down) Bromeliad-specific PCR primer This study Table 3. Attributes of analyzed matrices of the five datasets D1–5 and parsimony scores of equally most parsimonious trees after analysis using PAUP*. The sequence length is given for individual markers only. bp = base pairs, var. char. = variable characters, PIC = parsimony informative characters, CI = consistency index, RI = retention index. matrix lenght [bp] plastid DNA (D1) no. of char. no. of var. char. no. of PICs no. of trees tree length CI RI 6,115 744 (12%) 378 (6%) >100,000 1,210 0.685 0.831 PHYC (D2) 1,159–1,192 1,228 485 (39%) 369 (30%) >100,000 1,398 0.473 0.867 PRK (D3) 831–1,692 3,291 737 (22%) 534 (16%) >100,000 2,232 0.478 0.870 PHYC + PRK (D4) 4,519 1,222 (27%) 903 (20%) >100,000 3,775 0.458 0.859 PHYC + PRK + plastid DNA (D5) 10 634 1,966 (18%) 1,297 (12%) >100,000 5,043 0.507 0.853 139 BARFUSS, M.H.J. NEW CLASSIFICATION Table 4. Accepted tribal, subtribal and generic concepts and informal clades of Tillandsioideae (Bromeliaceae) in comparison to earlier classification systems (based on the placement of the nomenclatoric types). i.s. = incertae sedis. Smith & Till (1998) Barfuss & al. (2005) genus genus genus tribe genus/subtribe/clade tribe/clade Glomeropitcairnia Glomeropitcairnia Glomeropitcairnia Glomeropitcairnieae Glomeropitcairnia Glomeropitcairnieae Catopsis Catopsis Catopsis Catopsideae Catopsis Pogospermeae Alcantarea Alcantarea Alcantarea Vriesea Vriesea Vriesea Werauhia Werauhia Werauhia Vriesea Chrysostachys clade Vriesea tuerckheimii (i.s.) Mezobromelia Mezobromelia Mezobromelia Mezobromelia Vrieseeae Core Tillandsioideae Splendens clade Vriesea Cipuropsidinae Cipuropsis Cipuropsis-Mezobromelia clade Vrieseeae Vriesea Barfuss, Till & Samuel Vrieseinae Smith & Downs (1977) Singularis clade Josemania Tillandsia Tillandsieae Tillandsia Rothowia Tillandsia Viridantha Guzmania 140 Racinaea Racinaea Racinaea Guzmania Guzmania Guzmania Core Tillandsieae Lemeltonia Tillandsia Tillandsieae OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 Table 5. Accepted subgeneric concept, informal clades, and unclassified species of Tillandsia (Bromeliaceae) and recent generic segregates in comparison to earlier classification systems (based on the placement of the nomenclatoric type). A dashed line indicates further informal subdivision of a given taxon. subg. = subgenus, p.p. = pro parte, i.s. = incertae sedis. Smith & Downs (1977) Till (2000) Barfuss & al. (2005) Barfuss, Till & Samuel Tillandsia subg. Pseudalcantarea Tillandsia subg. Pseudalcantarea Tillandsia subg. Pseudalcantarea Tillandsia subg. Pseudalcantarea Vriesea p.p., typo excluso (xerophytic, grey-leaved) Tillandsia subg. Pseudovriesea Tillandsia subg. Tillandsia Rauhii clade Tillandsia subg. Tillandsia Tillandsia subg. Tillandsia Tillandsia subg. Tillandsia (syn. Tillandsia subg. Allardtia) Viridantha Tillandsia subg. Viridantha Biflora clade Tillandsia australis (i.s.) Tillandsia subg. Allardtia Tillandsia edithae (i.s.) Tillandsia subg. Allardtia Tillandsia subg. Allardtia p.p. majore, typo excluso Tillandsia disticha (i.s.) Tillandsia pseudomicans (i.s.) Tillandsia sphaerocephala (i.s.) Xiphioides clade Tillandsia esseriana (i.s.) Tillandsia subg. Anoplophytum Tillandsia subg. Anoplophytum Tillandsia subg. Anoplophytum Gardneri clade Tillandsia subg. Anoplophytum Tillandsia albertiana (i.s.) Tillandsia nana (i.s.) Tillandsia subg. Diaphoranthema Tillandsia subg. Diaphoranthema Tillandsia subg. Diaphoranthema Tillandsia subg. Phytarrhiza (xerophytic) Tillandsia subg. Phytarrhiza (xerophytic) Tillandsia subg. Diaphoranthema Tillandsia subg. Phytarrhiza Purpurea clade Josemania Tillandsia subg. Phytarrhiza Tillandsia subg. Phytarrhiza (meso-/semi-mesophytic) Tillandsia subg. Phytarrhiza (meso-/semi-mesophytic) Lemeltonia Rothowia Racinaea subg. Pseudophytarrhiza Tillandsia subg. Pseudo-Catopsis Racinaea Racinaea Racinaea subg. Racinaea 141 BARFUSS, M.H.J. 142 NEW CLASSIFICATION Figure 1. Simplified strict consensus trees of equally most-parsimonious trees found in parsimony analyses of individual datasets (D1–D5) supplemented with BP from parsimony bootstrap analysis either above or below branches; triangles sitting in terminal position on a branch indicate holophyletic units, whereas triangles emerging from within a node indicate several branches of that unit on a polytomy. Only relationships of all except two recognized genera and one informal clade are displayed. Cipuropsis and Mezobromelia are included in the Cipuropsis-Mezobromelia clade including informal clades. D1 = plastid DNA, D2 = PHYC, D3 = PRK, D4 = PHYC + PRK, D5 = plastid DNA ( + PHYC + PRK. OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 Nuclear alignments were evaluated prior to phylogenetic analyses for the impact of polymorphic sites (Single Nucleotide Polymorphism = SNP). The need for these additional steps before phylogenetic analyses has been pointed out by Jabaily & Sytsma (2010) and Russell & al. (2010b) in detail. The occurrence of SNPs in resulting electropherograms of nuclear allelic consensus sequences was inferred as allelic variation within the given sample. In the final alignments these positions were carefully evaluated to see if their patterns suggest the presence of similar alleles in more distantly related taxa indicating polyploidization, hybridization, reticulation, introgression events or the occurrence of ancient alleles (incomplete lineage sorting). The aligned plastid DNA matrix with some taxa deleted (because of missing nuclear DNA sequence information) was taken from Barfuss & al. (2005). A 25 bp region from atpB-rbcL (a partly inverted composed poly-A/poly-T region) where homology could not be unambiguously assessed, was excluded here from the analyses (positions 4,321–4,345 of the plastid DNA dataset and positions 8,840–8,864 of the combined dataset, respectively). Five datasets were analyzed, i.e., (D1) plastid DNA, (D2) PHYC, (D3) PRK, (D4) PHYC + PRK, and (D5) PHYC + PRK + plastid DNA, and compared with results from Barfuss & al. (2005). In four cases sequences of the plastid DNA dataset had to be combined with nuclear sequences from a different source of the same species (G. variegata L.B. Sm., T. barclayana Baker, T. lindenii Regel, T. stricta Sol. ex Sims), since the original sample of Barfuss & al. (2005) was not available anymore. These species could not be excluded because they represent genetic entities that would otherwise be either underrepresented or non-existent in the combined dataset. Alleles resulting from cloned nuclear sequences were combined with duplicated plastid DNA sequences only when their position in individual nuclear trees did not significantly contradict the plastid DNA tree topologies. Analyses of the aligned and evaluated individual and combined datasets included maximum parsimony (MP) and Bayesian metropolis-coupled MCMC inference (BI). MP analyses were implemented in the program PAUP* 4.0b10 (Swofford, 2003) using a two-step heuristic search strategy due to the complexity of DNA sequence data. The first step of heuristic searches was done with 1,000 random sequence additions, TBR branch swapping and holding 10 trees each step to reduce time spent swapping on non-optimal trees, and saving always 100 shortest trees per replicate, even if the trees were not the shortest over all replicates (to check for multiple island in the resulting tree space). In the second round trees obtained from the first analyses were taken as starting trees and heuristic searches were conducted until all or a maximum of 100,000 shortest trees were saved, but the tree search was allowed to continue to swap on all input trees for shorter ones (swapping to completion). Gaps were always treated as missing characters and no gap coding procedure was applied, mainly because gaps resulting from allelic length variation in the very indel-rich PRK dataset (D3) were ignored. BI analyses were done with the program MrBayes 3.1 (Huelsenbeck & Ronquist, 2001; Ronquist & Huelsenbeck, 2003) using the MPI version (Altekar & al., 2004) at the University of Oslo Bioportal (http://www.bioportal.uio.no/). Best-fitting nucleotide substitution models were determined in advance using MrModeltest 2.3 (Nylander, 2004) following the Akaike information criterion. Although the best model for each dataset was determined as GTR + I + G, analyses of the two combined datasets D4 and D5 were still partitioned (PHYC, PRK, plastid DNA) to allow different parameters for each individual region to be estimated separately. Four independent runs of four MCMC chains were done for ten million generations, sampling every 1,000 generations and with a burn-in of 8 to 25 %, which was evaluated for each run individually. Other parameters and priors were left as default. 143 Genus/clade Habit Petals Petal appendages Filaments Petalfilament Anthers Aperture type Glomeropitcairnia m f p f f f diffuse Catopsis m f a f f f Catopsis type Guzmania m c a / (p) f c / (f) f / (c) diffuse / inaperturate Josemania m / sm f a f f f insulae Lemeltonia sm / (m) f a c f f insulae Ovary position ½–⅔ inferior superior– ⅛ inferior max. ⅓ inferior max. ⅓ inferior max. ⅓ inferior Ovule appendage (chalazal) Setal nectary ducts Stigma type Pseudopappus Endostome type Embryo type p: ap > ov ? cb micropylar ? ? pp ascending se chalazal d g a horizontal se / cb micropylar a / b / (e) a / (b) a ? pi micropylar ? ? a horizontal cf micropylar d b micropylar c / d / f (?) b m f a f f f insulae to diffuse max. ⅓ inferior a horizontal Rothowia m f a f f f insulae to diffuse max. ⅓ inferior a ? cb micropylar e b / (–a) max. ⅓ inferior p: ap ≤ ov or (ap > ov) / a horizontal se/ cs / (cb) / (cp) micropylar e/f/g/h/i /k/l/m/n b/c/d/e/ f / (–g) a ? se micropylar c a a ? se micropylar ? ? a descending cup micropylar d a / (b) a ? se micropylar ? ? a ? se micropylar ? ? a ? cs micropylar c/d a p: ap ≤ ov ? cupp micropylar ? ? p: ap > ov ? cp / (ce) micropylar ? ? p: ap ≤ ov / (a) descending cb / tl micropylar e / (–c) b / (a) Tillandsia m / (sm) / (sx) / x f a / (p) f f f operculum / diffuse, insulae, Alcantarea type Cipuropsis m c (ca. ¼) p f c (ca. ¼) f diffuse to insulae Mezobromelia m c (ca. ½) p f c (ca. ½) c diffuse / inaperturate Werauhia m f p / (a) f f f insulae Chrysostachys clade m c (ca. ¼) p f c (ca. ¼) ? ? Singularis clade m c (ca. ¼) a f c (ca. ¼) f fine insulae Splendens clade m f p f f f ? V. tuerckheimii m f p f c (ca. ¼) f ? Alcantarea m / (sm) f Vriesea m / (sm) / (sx) f / (c) short p p f f f f / (c) short f f Alcantarea type insulae max. ⅓ inferior max. ⅓ inferior max. ⅓ inferior max. ⅓ inferior max. ⅓ inferior max. ⅓ inferior max. ⅓ inferior max. ⅓ inferior max. ⅓ inferior NEW CLASSIFICATION Racinaea se / (slightly cs) / (cf) BARFUSS, M.H.J. 144 Table 6. Morphological characters used to differentiate among genera and informal clades of Tillandsioideae. Taxa are arranged according to the "Classification of Tillandsioideae" chapter. Values for the septal nectary ducts, endostome type, and embryo type were exclusively taken from literature. m = mesophytic, sm = semi-mesophytic, sx = semi-xerophytic, x = xerophytic; f = free, c = connate/conglutinate; p = present, a = absent; ap = appendage, ov = ovule, pp = pseudopappus; cb = convolute-blade, ce = conduplicate-erect, cf = coralliform, cp = conduplicate-patent, cs = conduplicate-spiral, cup = cupulate, cupp = cupulate-papillate, se = simple-erect, tl = tubolaciniate, pi = pinnatisect; ? = missing data; (?) = information questionable in original source; data in () = rarely. Habit Petal shape Petal appendages Ovule appendage (chalazal) Stigma type Endostom type Embryo type Tillandsia subg. Anoplophytum (sx) / x lingulate a a / p: ap ≤ ov se / cs m (e–f) / f Tillandsia subg. Diaphoranthema x lingulate / unguiculate a p: ap ≤ ov se m/n e Tillandsia subg. Phytarrhiza x unguiculate a p: ap ≤ ov se m e/f Tillandsia subg. Pseudalcantarea m lingulate a a (cs) / cp f b Tillandsia subg. Pseudovriesea sx / (x) lingulate p / (a) p: ap ≤ ov cs / ce f (b–) / ± f Tillandsia subg. Tillandsia m / sm / sx / x lingulate a / (p) p: ap ≤ ov / ap > ov cs / (se) e/g/i/k b / c / f (–g) Tillandsia subg. Viridantha x lingulate a a / p: ap ≤ ov se / cb h/l c/f Biflora clade m / sm / (sx) lingulate a a / p: ap ≤ ov cs / (se) e/g/h b/f Gardneri clade x lingulate a a / (p: ap ≤ ov) se m (e–f) / f Purpurea clade x unguiculate a a / (p: ap ≤ ov) cs h/l c–f Rauhii clade m lingulate a p: a ≤ o/( ap > ov) cs / (cp) h d Xiphioides clade x unguiculate a p: ap ≤ ov se / (cs) m f 145 PART 3, CHAPTER 5 Subgenus/clade OF TILLANDSIOIDEAE (BROMELIACEAE) Table 7. Morphological characters used to differentiate among subgenera and informal clades of Tillandsia. Taxa are arranged according to the “Classification of Tillandsioideae” chapter. m = mesophytic, sm = semi-mesophytic, sx = semi-xerophytic, x = xerophytic; p = present, a = absent; ap = appendage, ov = ovule, cb = convolute-blade, ce = conduplicate-erect, cp = conduplicate-patent, cs = conduplicate-spiral, se = simple-erect; data in () = rarely. BARFUSS, M.H.J. NEW CLASSIFICATION Support was obtained from 50 % Bayesian majority rule consensus trees (Posterior Probabilities, PP) and from parsimony Bootstrap analyses done in PAUP* 4.0b10. Bootstrap percentages (BPs) were calculated using 1,000 replicates, TBR branch swapping, simple sequence addition, holding 10 trees each step and saving 100 shortest trees per bootstrap replicate. Combinability of plastid DNA, PHYC and PRK datasets was determined by visual inspection of individual bootstrap consensus trees (Whitten & al., 2000). Measures of incongruence like the incongruence length difference test (ILD; Farris & al, 1994, 1995; Lee, 2001) have been demonstrated not always to be useful indicators of data combinability, specifically in the light of relatively recently divergent plant groups with a lower sequence divergence in plastid DNA compared to nuclear DNA (Yoder & al., 2001; Reeves & al., 2001). We use the following descriptions for categories of BP: unsupported, <50%; weak, 50–79%; moderate, 80–89%; strong, 90–100%. In the case of only weakly- to moderately-supported incongruent topologies between the individual trees, direct combination was regarded as appropriate. The artifact of long-branch attraction (LBA) can sometimes yield high moderate or strong BP for a phylogenetically wrong clade, therefore the number of autapomorphies for and the taxon sampling within the specific clades are also considered (Bergsten, 2005). Morphological data and analyses Morphological characters were collected and documented over several years from herbarium sheets, liquid flower preservations, fresh material cultivated in the botanical garden of Vienna or from several donors (see "Taxon selection") using a stereo microscope (Olympus SZ60 Stereo Zoom Microscope) or SEM technology (either with a JEOL T-300 or JEOL JSM-6390 Scanning Electron Microscope). Characters of some taxa were extracted and carefully evaluated from the original description of the respective taxon, the last monograph (Smith & Downs, 1977) or selected literature (e.g., Brown & Gilmartin 1984, 1989b; Böhme, 1988; Groß, 1988; Brown & Terry, 1992; Halbritter, 1992). Except for the general habit, the gathered morphological information mainly comprised characters of diverse floral elements like sepals, petals (incl. petal appendages), stamens (incl. filaments, anthers and pollen), carpels (incl. stigma and ovules) as well as ovary position and seed morphology (Tables 6, 7). Characters were manually mapped on resulting phylogenetic trees and empirically useful morphological traits taken for the circumscription of obtained clades. No phylogenetic analysis of a morphological data matrix was performed, because of missing data in many of accessions studied. Although several species could not be investigated for all used morphological traits, it was assumed that characteristics can be applied to higher taxonomic ranks. Results Table 3 provides the attributes of the analyzed matrices, together with the details of the equally-most-parsimonious trees found in parsimony analyses of each of the five datasets. Tables 6 and 7 summarize morphological traits found useful so far and used to differentiate between genera, subgenera and informal clades. Simplified strict consensus trees of shortest trees from parsimony analyses supplemented with BP are shown in Figure 1, D1–5 for comparing main differences in tree topologies between each dataset. One most parsimonious phylogram (Figure 2, see Appendix) from total combined analysis that reflects our understanding on phylogenetic relationships best was chosen and supplemented with BP and PP. Alternative topologies of terminal branches and backbone relationships from Bayesian analysis are displayed in boxes 146 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 mostly in front of the corresponding clades (terminals) or on the upper-left side (backbones) of Figure 2. Their positions within the parsimony tree are highlighted by capital letters (Figure 2, A–P: terminals, Q–S: backbones). Other trees from different analyses (parsimony and Bayesian) of each individual dataset are not shown and can be requested from the first author. Only relationships of genera accepted in the present investigation (except Cipuropsis and Mezobromelia) and one informal infrasubtribal clade (Cipuropsis-Mezobromelia clade) are shown and their results presented for each individual dataset (Figure 1, D1–5). Results including all tribal, subtribal, generic, and infrageneric units and informal infrageneric clades are displayed for the total combined dataset only (Figure 2). The nomenclature, taxonomy and morphological characterization of clades, and the taxonomic position of each species studied can be obtained from the chapter "Classification of Tillandsioideae", Figure 2 or Tables 4 and 5, respectively. Main geographic distributions are given in the key to the genera. Gene selection From eight genomic regions screened initially, only PRK and PHYC were selected because (1) they were easy to amplify and sequence in most cases, (2) had the most variation in relation to their average length, and (3), most importantly, amplifications resulted in single, clear bands. This indicates an effectively single-copy nature of both nuclear regions and therefore allows PCR products to be sequenced directly without cloning, at least in individuals without allelic indel variation or hybrid origin. This is in agreement with previous studies of these genes in other monocot genera like Puya Molina (Bromeliaceae: Jabaily & Sytsma, 2010), Oryza sativa L. (Poaceae: Chen & al., 2004), or Chamaedorea Willd. (Arecaceae: Thomas & al., 2006). Reports of possible multiple loci (paralogs) of PRK in subfamily Bromelioideae (Schulte & al., 2009) cannot be confirmed for Tillandsioideae and are possibly due to a gene duplication event within Bromelioideae (D. Silvestro & K. Schulte, unpubl. data). Within PRK introns few polynucleotide runs (SSRs) were observed and could not be avoided (see supplementary data): (1) in intron four a more frequent SSR was detected in most tillandsioid accessions (except for taxa were the SSR containing region has been lost), (2) in intron two few Guzmania species exhibit a large insertion containing a second SSR, and (3) in intron three of G. multiflora (André) André ex Mez, which also exhibits a larger insertion, a third SSR was observed. The other regions were discarded either because they made problems in amplifications and sequencing (ITS nrDNA), primer combinations only amplified a comparatively short DNA fragment (RPB2, MS), or sequence variation and taxon sampling was to low (NIA, PGIC, XDH). Nuclear data evaluation Details of allelic length variation (heterozygous indels) and polymorphic sites (SNPs) are given separately for each nuclear marker (see below). Evaluation of both PHYC and PRK data revealed that none of the investigated samples showed evidence for polyploidization, hybridization, reticulation, introgression or incomplete lineage sorting involving more distantly related species. Although polyploidy (either auto- or allopolyploidy) can have an major impact on the allele distribution and frequency within an/all affected species, it can be rejected for most species analyzed here since in most cases allelic variation can be traced back to two alleles (see supplementary data). This is also in agreement with previous findings that most Tillandsioideae species (and Bromeliaceae species in general) show a diploid chromosome number of 2n = 2x = 50 (e.g., Brown & Gilmartin, 1989a; Palma-Silva & al., 2004). In two species within T. subg. Diaphoranthema (T. capillaris Ruiz & Pav., T. virescens Ruiz & Pav.), allelic variation of SNPs 147 BARFUSS, M.H.J. NEW CLASSIFICATION supports already documented cases of polyploidy with 2n = 4x = 100 (Till, 1992), either autopolyploidization or allopolyploidization involving closely related species. However, these were ignored since they have no impact on the current results (incl. branching patterns and statistical support) and are insignificant for the current study. A possible case of a hybrid origin involving sister taxa was found in the sample determined as Ro. platyrhachis. One allele of this accession is strongly associated with Ro. wagneriana. Sequence variation between or within other closely related species in terminal branches was too low to either reject or confirm any of the above mentioned hypotheses. Therefore, these were ignored, since the scope of the present study is to resolve deeper nodes primarily and not individual relationships of recently diverged species. Otherwise it would also be necessary to question and investigate the species concept used involving a very broad sampling and other molecular markers. However, polytomies and less resolved or supported clades at any hierarchical level in resulting trees were primarily caused by a lack of phylogenetic signal rather than conflicting polymorphic sites. Plastid DNA data and analyses (dataset D1) Barfuss & al. (2005) already explored attributes and detailed results from the plastid DNA data analyses. A summary of generic relationships based on a strict consensus tree of a maximum parsimony analysis supplied with BP is given in Figure 1, D1. Differences in the attributes between the first plastid DNA study and the present investigation (Table 3) are due to the deletion of several taxa from the plastid DNA matrix, where no nuclear DNA data was available, and the exclusion of a 25 bp region from atpB-rbcL. Deleting these taxa and base pairs had no significant effect on the topology of obtained trees, identical to those of Barfuss & al. (2005). PHYC data and analyses (dataset D2) Characteristics of the analyzed PHYC matrix are given in Table 3. PCR primers amplified fragments of partial exon one ranging from 1,159 (Werauhia insignis) to 1,192 bp (Guzmania acorifolia (Griseb.) Mez, G. condensata Mez & Wercklé), but mostly a fragment of 1,177 bp. According to the tree topology, the latter size can be inferred as the ancestral state. Indel events are rare and derived, and present either in single species or few closely related species in terminal branches. Indication of allelic length variation within individuals caused by heterozygous indels was not found in any of the sampled taxa, since in PHYC only conserved parts of the exon 1 were amplified. Allelic nucleotide variation as indicated by obvious double peaks at a few positions was found in 220 taxa out of 444 sampled individuals. 224 samples were found to be homozygous for the region sequenced. The number of SNPs in a given heterozygous taxon ranged from 1–15 (the latter seen in T. aff. cucaensis Wittm. B0735). Distribution of polymorphic sites was evenly across the whole range of PHYC with no particular preference for a specific gene region, but significantly higher at the 3rd codon position caused mostly by synonymous substitutions. Evaluation of the effects of SNPs in PHYC data indicate that there is not a strong phylogenetic bias in not having allelic data and allelic consensus sequences can be used. We therefore only used the polymorphism-coded PHYC dataset for all analyses, with the exception of three taxa, which were cloned for PRK and therefore also for PHYC, to be able to combine corresponding allelic sequences. Figure 1, D2 is the strict consensus tree of a MP analysis with BP, which summarizes the generic relationships based on PHYC sequence data. As in plastid DNA tree, both Glomeropitcairnia (tribe Glomeropitcairnieae) and Catopsis (tribe Pogospermeae) are strongly supported (Bootstrap percentage = BP 100) and sisters (BP 100), with the latter clade sister to a strongly148 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 supported core Tillandsioideae (BP 100). Tribe Vrieseeae is paraphyletic in PHYC results, but the resulting grade of subtribes Vrieseinae and Cipuropsidinae has only a weak BP (58). Vriesea is holophyletic (BP 89) only when Andean members located in the Cipuropsis-Mezobromelia clade, T. malzinei and former xerophytic gray-leaved members, the latter two now placed within Tillandsia, are excluded. The sister group relationship of Vriesea and Alcantarea (= subtribe Vrieseinae) is strongly supported (BP 100), as is the holophyletic origin of Alcantarea (BP 97). Subtribe Cipuropsidinae is holophyletic (BP 90) and splits into two lineages, Werauhia (BP 100) and the Cipuropsis-Mezobromelia clade (BP 67). Mezobromelia is a weakly-supported holophyletic lineage (BP 57) nested within the Cipuropsis-Mezobromelia clade (not shown), which in addition contains taxa previously assigned to Vriesea and Tillandsia. Tribe Tillandsieae is strongly supported (BP 92). The first lineage which splits from core Tillandsieae is Guzmania (BP 85), with G. hutchisonii (syn. M. hutchisonii) nested inside (not shown). Core Tillandsieae is only weakly supported (BP 58) and relationships of genera are unresolved. Holophyletic taxa are Josemania (BP 89), Racinaea (no BP) and Lemeltonia (BP 86), but all three are placed on a polytomy with different clades of Tillandsia and Rothowia. PRK data and analyses (dataset D3) Attributes of the analyzed PRK matrix are given in Table 3. PCR primers amplified fragments ranging from 831 (G. patula Mez & Wercklé) to 1,692 bp (T. ionochroma André ex Mez), with a mean of 1,086 bp. Indel events are common at all hierarchical levels of obtained trees, both between individuals of a single species and different higher clades. Unlike PHYC, evidence of allelic length variation within individuals was detected in 146 of sampled taxa, since the amplified region covered three less conserved introns and four exons (exon 2 to exon 5), with the second and the fifth exon being sequenced only partially. The number of heterozygous indels ranged from 1 to 5 in a given taxon. Indels were only found in the less conserved intron regions and never in the PRK exons. Three taxa had to be cloned, since indel variation between alleles was too high to perform unambiguous editing, i.e., V. psittacina (Hook.) Lindl., G. graminifolia (André ex Baker) L.B. Sm., and Ro. platyrhachis. Some others were occasionally cloned to verify allele length (see supplementary data). Allelic nucleotide variation was found in 241 taxa out of 444 sampled individuals. Only 193 samples were found to be homozygous for the whole region sequenced concerning indels and SNPs, which seems to be due to the relatively high rate of evolution in intron sequences. The number of SNPs in a given heterozygous taxon ranged from 1 to 18 (the latter seen in T. fasciculata Sw. B0076). Polymorphic sites were primarily found in the less conserved intron parts of the gene and less frequent in exons with mostly synonymous substitutions. Like in PHYC, evaluation of the effects of SNPs in PRK data indicate that there is not a strong phylogenetic bias in not having allelic data and allelic consensus sequences can be used. We therefore also only use the polymorphism-coded PRK dataset for all analyses, with the exception of the three cloned taxa. Generic relationships based on PRK sequence data are shown in Figure 1, D3, which is the strict consensus tree of a MP analysis supplemented with BP. Backbone relationships are much less resolved than in the PHYC tree. Topologically identical are the sister group relationship (BP 100) of Glomeropitcairnia (BP 100) and Catopsis (BP 100) and core Tillandsioideae (BP 100). Subtribes Cipuropsidinae and Vrieseinae are strongly supported (BP 91 and 100, respectively), whereas tribes Vrieseeae and Tillandsieae as well as core Tillandsieae are not supported. Werauhia is strongly supported by PRK sequence data (BP 100), whereas the CipuropsisMezobromelia clade has only weak BP (BP 58). The same picture is seen for Alcantarea (BP 149 BARFUSS, M.H.J. NEW CLASSIFICATION 100) and Vriesea (BP 61), respectively. The relationship of Vriesea and Alcantarea to a clade containing Tillandsia and Lemeltonia is surprising, but this is only weakly supported (BP 63). Guzmania is not supported and three different clades of Guzmania species (not shown) occur in a polytomy with several other clades. Rothowia and Josemania are resolved as sisters (BP 65), both also being holophyletic (BP 64 and 100, respectively). Racinaea is strongly supported (BP 96), but relationships to other genera remain unresolved. Tillandsia and Lemeltonia are resolved as sister taxa with a moderate BP (BP 81) with Lemeltonia being strongly supported (BP 96), but in a polytomy with several other clades of Tillandsia species. Combined nuclear data (dataset D4) Analyses of concatenated nuclear datasets D2 and D3 (Figure 1, D4) combine tree topologies obtained from individual analyses of PHYC and PRK (Figure 1, D2–3). Most accepted infrasubfamilial clades and genera are holophyletic with mostly moderate or high BP, i.e., Glomeropitcairnia (BP 100), Catopsis (BP 100), core Tillandsioideae (BP 100), Vrieseinae (BP 100), Vriesea (BP 89), Alcantarea (BP 100), Cipuropsidinae (BP 99), Werauhia (BP 100), CipuropsisMezobromelia clade (BP 81), Tillandsieae (BP 81), core Tillandsieae (BP 77), a sister relationship of Josemania and Rothowia (BP 79), Josemania (BP 100), Racinaea (BP 99), Lemeltonia (BP 100), and the sister relationship of Lemeltonia and Tillandsia (BP 95). The grade of paraphyletic Vrieseeae taxa (BP 60), Guzmania (BP 61), and Rothowia (BP 70) are weakly supported. No BP is given for the holophyletic Tillandsia and the sister group position of Racinaea to the clade containing Josemania and Rothowia. Combined plastid and nuclear data (dataset D5) Results of combined analyses of all markers used are shown in two figures. Figure 1, D5 presents a strict consensus tree summarizing mostly generic relationships. Figure 2 is more inclusive and displays relationships of all accessions investigated including subgeneric units and informal infrasubfamilial and -generic clades. These are either displayed as a phylogram from parsimony analysis with BP and PP (Figure 2, complete tree) or parts of a phylogram from Bayesian analysis showing alternative topologies (Figure 2, A–S). As in the combined nuclear dataset (Figure 1, D4), all generic entities and accepted infrasubfamilial units are holophyletic, but are here supported by mostly high or moderate values: Glomeropitcairnia (tribe Glomeropitcairnieae; BP 100/PP 100), Catopsis (tribe Pogospermeae; BP 100/PP 100), Werauhia (BP 100/PP 100), Splendens clade (BP 100/PP 100), Chrysostachys clade (BP 85/PP 100), Cipuropsis (BP 82/PP 98), Alcantarea (BP 100/PP 100), Vriesea (BP 89/PP 100), Josemania (BP 100/PP 100), Racinaea (BP 100/PP 100), Rothowia (BP 81/PP 100), and Lemeltonia (BP 100/PP 100); weak BP is currently given only for Mezobromelia (BP 66/PP 100), Guzmania (BP 68/PP 100), and Tillandsia (BP 62/PP 96). Tribes and subtribes are also strongly supported: Vrieseeae (BP 93/PP 100), Cipuropsidinae (BP 100/PP 100), Vrieseinae (BP 100/PP 100), and Tillandsieae (BP 100/PP 100). Unclassified units that were also considered in this study, but not treated taxonomically are: core Tillandsioideae (BP 100/PP 100), and core Tillandsieae (BP 93/PP 100). Relationships of the latter including genera Josemania, Lemeltonia, Racinaea, Rothowia and Tillandsia are resolved differently in parsimony and Bayesian analysis (Fig 2) and show currently either no BP or only PP. Infrageneric units that are well characterized and strongly to moderately supported (referring to BP only) are (1) within Racinaea: R. subg. Pseudophytarrhiza (BP 100/PP 100) and R. subg. Racinaea (BP 100/PP 100); (2) within Tillandsia: T. subg. Pseudalcantarea (BP 99/PP 100), T. subg. Viridantha (BP 95/PP 100), the Purpurea clade (BP 100/PP 100), 150 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 the Gardneri clade (BP 100/PP 100), T. subg. Anoplophytum (BP 99/PP 100), T. subg. Phytarrhiza (BP 100/PP 100), and the Xiphioides clade (BP 91/PP 100). Weakly supported are T. subg. Pseudovriesea (BP 68/PP 99), T. subg. Tillandsia, the Rauhii clade (BP 57/no PP), and the Biflora clade. Tillandsia subg. Diaphoranthema is paraphyletic, although internal clades receive strong statistical support (Figure 2, Q). Additional clades which are neither used formally nor named, but present in most markers studied are (1) a clade containing the Gardneri clade, T. subg. Anoplophytum, T. subg. Diaphoranthema, T. subg. Phytarrhiza, the Xiphioides clade and the unclassified species T. albertiana, T. edithae, and T. esseriana (BP 84/PP 100), and (2) the same but without the earlier branching Gardneri clade (BP 63/PP 100); and a terminal clade comprising T. subg. Diaphoranthema, T. subg. Phytarrhiza, and the Xiphioides clade (BP 91/PP 100). Morphological data and analysis Mapping morphological characters onto phylogenetic trees (data not shown) resulted in the following diagnostic traits to be used for delimiting generic clades in the present study (Table 6): ovary position, seed morphology, stigma structure, corolla structure, stamen structure, the presence vs. absence of petal appendages, and the mesophytic vs. xerophytic habit. Other characters considered to be important, but currently not used as diagnostic characters here are ovule and pollen morphology (Table 6 and 7). The main reason for not using these characters for defining genera is the high amount of missing data for many accessions, since pollen and ovules have been rarely available for detailed study. However, these characters are used to characterize new generic and subgeneric units, where data was available and most taxa could be examined. Despite of the limited sampling, the congruence between morphological characters and the molecular phylogeny is convincing. The key characters for delimiting genera in the core Tillandsioideae (all genera except Catopsis and Glomeropitcairnia) are the stigma morphology, followed by the corolla and stamen structure. Three new stigma types are recognized (conduplicate-erect, conduplicate-patent, and pinnatisect) in addition to the already described ones (Brown & Gilmartin, 1989b). An initial trial to completely avoid the use of petal appendages as diagnostic characters failed, since some genetic entities (Guzmania, Tillandsia, internal clades of the CipuropsisMezobromelia clade) share similar character states in their morphological traits currently used for differentiation. Therefore petal appendages are also used in the present taxonomic concept, but with a reduced taxonomic value, since no other morphological characters have yet been identified or are satisfactorily known. Petal appendages still have to be utilized for the differentiation of Mezobromelia from Guzmania and Cipuropsis from Tillandsia, with the following exceptions: (1) T. subg. Pseudovriesea and T. malzinei within Tillandsia (petal appendages present in these taxa, whereas absent from the rest of Tillandsia species), (2) G. hutchisonii within Guzmania (petal appendages present in this taxon, whereas absent from the rest of Guzmania species), (3) W. insignis within Werauhia (petal appendages absent from this taxon, whereas present in the rest of Werauhia species), and (4) T. asplundii and T. singularis within the Cipuropsis-Mezobromelia clade (petal appendages absent from these taxa, whereas present in all other Cipuropsis-Mezobromelia clade species). The species Ci. amicorum (syn. T. amicorum) and M. schimperiana (syn. T. schimperiana) are clearly misplaced in the genus Tillandsia, not only from the genetic side: all investigated material of these species exhibits petal appendages. The placement of Ci. amicorum into Tillandsia was most probably based on herbarium-preserved flowers, where petal appendages can be easily overlooked, and the 151 BARFUSS, M.H.J. NEW CLASSIFICATION placement of M. schimperiana was based on a post-anthetic herbarium specimen, assuming that according to the overall morphology it would best fit into Tillandsia. DISCUSSION In contrast to "Classification system of Tillandsioideae", arrangement of discussed taxonomic units and clades is not strictly hierarchical and alphabetical and more orientated towards the tree topology (Figure 2, see Appendix). This is necessary, since additional clades that were not formally treated (e.g., core Tillandsioideae and core Tillandsieae and the CipuropsisMezobromelia clade), and not accepted units (e.g., T. subg. Allardtia) are included as well. However, within higher units taxa and clades are arranged alphabetically to maintain clarity. DNA data and analysis Individual datasets D1–D3 exhibit big differences in their attributes (Table 3) and in their usefulness in resolving phylogenetic relationships, also evident from individual bootstrap consensus trees (Figure 1, D1–D3). No marker alone, neither plastid DNA (Figure 1, D1) nor individual nuclear DNA (Figure 1, D2–D3), is able to completely resolve branching patterns, but nuclear sequences in combination (on an average of about 2,300 bp) are yielding even better resolution within Tillandsioideae than combined plastid sequences (on an average of about 5,000 bp). Although the information content of nuclear datasets D2 and D3 is much higher, the amount of homoplasy also increased compared to plastid DNA data. The easiest markers to work with so far are the matK gene within the plastid DNA and PHYC within nuclear DNA markers. PRK is difficult and not very suitable, especially at higher levels, in the light of the numerous indels that have to be introduced into the alignment. Several species that were genetically investigated by multiple accessions are apparently of non-holophyletic origin (Figure 2); for some this can be explained as a result of missing DNA sequence information, evident from very short branches in respective terminal clades: e.g., L. scaligera, M. schimperiana, and M. bicolor; some taxa appear paraphyletic: e.g., A. nevaresii Leme and R. fraseri (Baker) M.A. Spencer & L.B. Sm.; and for others the results support a polyphyletic origin: e.g., varieties of R. tetrantha (Ruiz & Pav.) M.A. Spencer & L.B. Sm., T. fendleri Griseb., T. guatemalensis, some varieties and forms of T. tectorum E. Morren, and subspecies of T. xiphioides. Although this suggests the need of a critical revision of species boundaries within Tillandsioideae, sampling needs to be increased to really address this question. Therefore species circumscription is not treated within our study as it has no relevance for the outcome. Morphological data and analysis As already shown by previous phylogenetic and morphological studies, the four diagnostic characters used by Smith& Downs (1977) to differentiate genera within Tillandsioideae are insufficient for achieving an evolutionary based, phylogenetically accurate classification system; these morphological characters are the ovary position, apical and/or basal seed appendages, free vs. conglutinate/connate petals, and present vs. absent petal appendages. The first two characters are applied in this study in the same way for separating Glomeropitcairnia and Catopsis from core Tillandsioideae. The other two characters cannot be applied as proposed by Smith & Downs (1977), since they involve either convergent/parallel evolution or symplesiomorphic character states, evident in the evolution of a fused corolla tube in both tribes Til152 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 landsieae (Guzmania) and Vrieseeae (Mezobromelia) or the occurrence of petal appendages in different phylogenetic lineages. Petal appendages are primarily found in tribe Vrieseeae with only a few exceptions (see "Results"); in contrast, petal appendages are mostly absent in Tillandsieae, but also here a few exceptions are present. Most morphological characters used in this study need a future, much detailed investigation of an increased sample of Tillandsioideae, e.g., stigma or septal nectary morphology. Glomeropitcairnia Glomeropitcairnia Harms. is well supported, isolated, genetically and morphologically well differentiated and comprises only two species, i.e., Gl. erectiflora Mez and the type species Gl. penduliflora (Griseb.) Mez. Till & al. (1997) illustrated and portrayed both species side by side for the first time showing important, previously undocumented morphological characteristics, i.e., a convolute-blade like stigma with unique papillae and a diffuse aperture of the pollen. The genus was originally described by Mez (1896) as a subgenus of Pitcairnia L'Hér. based on the character state of the position of the ½–⅔ inferior ovary. Initial controversies over the correct subfamilial placement have mostly been solved since Harms (1930) established the separate, monotypic tribe Glomeropitcairnieae under subfam. Tillandsioideae based on the common character of plumose seed appendages. Gilmartin & al. (1989) confirmed the placement within Tillandsioideae based on cladistic analyses of morphological characters and concluded the genus to be the sister taxon to Guzmania and Mezobromelia. Ranker & al. (1990) on the other hand analyzed restriction site variation of the plastid genome and suggested to treat Glomeropitcairnia as a monotypic subfamily within Bromeliaceae because of its genetic divergence from the other tillandsioid genera. Whereas the placement within Tillandsioideae is well supported by all published DNA sequence analyses (originally using the ndhF gene by Terry & Brown, 1996) a close relationship to any of the other genera and the exclusion from Tillandsioideae is not supported. Catopsis Other than Glomeropitcairnia, Catopsis (type species C. nitida (Hook.) Griseb.) was always placed within Tillandsioideae. Harms (1930) emphasized the recognition of a monotypic tribe for the genus based on its unique seed characteristics. Catopsis is very isolated within Tillandsioideae showing apparent pollen, septal nectary, and seed morphological characters that differentiate it from the rest of the genera within the subfamily (e.g., Böhme, 1988; Halbritter, 1992; Palací, 1997). Because of its unique position, which is also supported by a larger number of autapomorphies for the genus in the phylogenetic trees, a separate, monotypic tribe is also justified by DNA sequence data. Not Harms (1930), who established Catopsideae for Catopsis within Tillandsioideae as originally indicated in Barfuss & al (2005), but Brongniart (1864) was the first who proposed Pogospermeae as a monotypic tribe for Pogospermum, which later became a synonym of Catopsis. Therefore Pogospermeae is the oldest available and correct name at tribal level which has to be used. Core Tillandsioideae = Tillandsioideae s.str. This term for a clade found in phylogenetic investigations of Bromeliaceae was first introduced by Terry & al. (1997b) and comprised the genera Guzmania, Mezobromelia, Tillandsia (incl. Racinaea), and Vriesea (incl. Werauhia). They did not investigate any Alcantarea species, but Horres & al. (2000) showed that Alcantarea is also part of the core Tillandsioideae. Crayn & al. 153 BARFUSS, M.H.J. NEW CLASSIFICATION (2004) had a different view in circumscribing this group and excluded Werauhia and Mezobromelia, maybe because they did not include any Alcantarea or true Vriesea species. Their only "Vriesea" species was T. malzinei, which is actually not part of the true Vriesea alliance (see below) but belongs to Tillandsia and is in sister position to T. funckiana Baker (Terry & al., 1997b). We are following the original idea of core Tillandsioideae and restrict the term to all currently accepted genera and informal clades except Catopsis and Glomeropitcairnia. In all phylogenetic trees obtained from different datasets (Figure 1, D1–5), regardless which genome (plastid, nuclear) was studied, core Tillandsioideae always resolve as a holophyletic unit with maximal BP. Vrieseeae Tribe Vrieseeae corresponds roughly to Vriesea s.l. (= Vriesea sensu Smith & Downs, 1977), with the exception of some taxa, which are now either excluded (T. malzinei and former xerophytic, gray-leaved Vriesea species) or included (Mezobromelia sensu typo and elements of former T. subg. Allardtia, i.e., Ci. amicorum, M. schimperiana, T. asplundii, and T. singularis). Mezobromelia sensu typo was only excluded in Barfuss & al. (2005), because conclusions were based on a misidentified Guzmania species. In the current study the tribe is subdivided into two main lineages, i.e., subtribe Cipuropsidinae with Cipuropsis, Mezobromelia, Werauhia and three informal clades (see below), and subtribe Vrieseinae with Alcantarea and Vriesea. Both subtribes can be distinguished by morphological characters of stigmas and ovules (see below). Vrieseeae is holophyletic in the plastid DNA tree (Figure 1, D1, BP 100), but not in the case of individual and combined nuclear markers (Figure 1, D2–D4). Although topologies of these are conflicting, supports for these alternative branching patterns involving taxa of Vrieseeae are generally weak. The reason seems to be the lack of characters to resolve backbone relationships in this part of the tree with better confidence. We still trust in the resolution of the plastid DNA tree and are fairly confident that more nuclear DNA sequences will provide the same result as plastid DNA data, since BP of the alternative topology from combined nuclear data (Figure 1, D4) still remains weak (BP 60). Cipuropsidinae Subtribe Cipuropsidinae is mainly composed of the Andean, Caribbean, and Mesoamerican species of Mezobromelia and Vriesea sensu Smith & Downs (1977), which do not show Vrieseinae-type stigmas. In contrast to Vrieseinae, which usually have very long appendages on the ovules, taxa of this clade generally display no or only very short appendages, except for Vriesea tuerckheimii. The subtribal name is based on the oldest generic name available for a species falling within this clade (Ci. subandina). It is divided into two main lineages, one corresponding to taxa assigned to the Cipuropsis-Mezobromelia clade, the other one to Werauhia. 154 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 Cipuropsis-Mezobromelia clade Presently, the Cipuropsis-Mezobromelia clade is an assemblage of previously recognized species of Mezobromelia sensu typo, Tillandsia, and Vriesea (incl. Cipuropsis). Besides Tillandsia s.str, this is the second clade where a fully resolved classification system cannot yet be applied. Species and clades found in this phylogenetic unit are (1) V. glutinosa and V. splendens (Splendens clade), (2) V. chrysostachys and V. ospinae (Chrysostachys clade), (3) V. tuerckheimii, (4) T. asplundii and T. singularis (Singularis clade), (5) Ci. amicorum and Ci. zamorensis (syn. V. zamorensis) (Cipuropsis), (6) M. bicolor, M. pleiosticha, and M. schimperiana (Mezobromelia). Vriesea species found here differ from Vriesea s.str. in not having the convolute-blade or the tubolaciniate stigma morphology but either simple-erect or conduplicate-spiral, respectively. Cipuropsis amicorum and M. schimperiana have been misplaced in Tillandsia (petal appendages present in both species), and the Singularis clade lacks petal appendages and was therefore previously treated within T. subg. Allardtia, but differs clearly from that subgenus in other morphological features, e.g., pollen with fine insulae on the aperture, leaf sheaths drying silvery-grey, and tubular yellow corolla. V. tuerckheimii remains unclassified and without assignment to a group. We are presently proposing generic rearrangements only for three species within the Cipuropsis and Mezobromelia clades (Ci. amicorum, Ci. zamorensis, and M. schimperiana) mainly because of limited access to plant material for additional species that are supposed to group within the Cipuropsis-Mezobromelia clade. Although investigated species of this clade form distinct, moderately or strongly-supported phylogenetic units, the application of morphological characters is currently too vague to propose new genera. Cipuropsis Cipuropsis was proposed as a monotypic genus for Ci. subandina (syn. V. subandina). Its generic status was later accepted by Harms (1930), but not by Mez (1934–35) and Smith & Downs (1977), who treated the genus as a synonym of Vriesea. The only diagnostic character believed to be natural and discriminative for species of Tillandsia and Vriesea by the latter two have been petal appendages while other morphological differences between species complexes were considered to be within the morphological range of these two genera. Unfortunately, it has not been possible for us to collect or receive fresh or well-preserved plant material for genetic analyses of Ci. subandina, but because of its strong morphological similarity with and obvious relationships to Ci. zamorensis based on herbarium specimens (incl. type material) of both species (W. Till & M.H.J. Barfuss, unpubl. data), it is most likely that Ci. subandina species groups within this clade. As Cipuropsis is the oldest generic name available for a group of species within the Cipuropsis-Mezobromelia clade, it is restored from the synonymy of Vriesea and treated as a separate genus within subtribe Cipuropsidinae. Two species previously treated either under Tillandsia or Vriesea are transferred here, i.e., Ci. amicorum and Ci. zamorensis, respectively. Further taxa not investigated but supposed to belong to Cipuropsis are V. dubia (L.B Sm.) L.B. Sm., V. duidae (L.B Sm.) Gouda, V. elata (Baker) L.B. Sm., V. lutheriana J.R. Grant, and V. rubra (Ruiz & Pav.) Beer. This taxonomic change is justified, since Ci. subandina will always be part of the Cipuropsis-Mezobromelia clade. Even if the taxonomic circumscription of Cipuropsis will change, other species of the Cipuropsis-Mezobromelia clade would have to be assigned to Cipuropsis for reasons of priority. 155 BARFUSS, M.H.J. NEW CLASSIFICATION Mezobromelia Species previously classified within Mezobromelia s.l. belong to two phylogenetically unrelated units. The first lineage, of which we have only investigated G. hutchisonii, is placed within Guzmania; the second one, in which the type species M. bicolor is located (= Mezobromelia s.str.), is nested within the Cipuropsis-Mezobromelia clade as sister to Cipuropsis. Mezobromelia schimperiana is genetically definitely not a Tillandsia species, as evident also from flower morphology, because petals are connate and appendaged in all investigated specimens. It shares great morphological similarity with species assigned to Mezobromelia s.str. (J.M. Manzanares, unpubl. data), which is also evident genetically; the species is therefore assigned to this genus. Chrysostachys clade This clade is composed of two Vriesea species (the name-giving V. chrysostachys, and V. ospinae) which were already considered to be a distinct species complex under Vriesea (W. Till, unpubl. data). Both species show simple-erect stigma morphology. Singularis clade Tillandsia asplundii and the name-giving T. singularis are always nested within the CipuropsisMezobromelia clade, despite having no petal appendages. Both species are sister taxa and together they are in sister position to the clade containing both genera Cipuropsis and Mezobromelia. Some similar Tillandsia species like T. delicatula L.B. Sm., T. pinnata Mez & Sodiro, and T. truncata L.B. Sm. might also cluster with these two species, but due to the lack of material, these species were not investigated genetically. This species complex was previously treated under T. subg. Allardtia and forms a distinct clade within Cipuropsis-Mezobromelia. Generic status might be justified, when a broader sampling has been studied. Splendens clade Presently it is composed of two species in northern South America. While V. glutinosa had been placed in V. sect. Xiphion by Smith & Downs (1977), V. splendens (the name-giving taxon) was classified as a member of V. sect Vriesea. The recognition of a formal status of this group, to which also V. soderstromii L.B. Sm. may belong, remains to be studied further. Werauhia Werauhia (type species W. gladioliflora (H. Wendl.) J.R. Grant) is holophyletic and well supported by all DNA regions studied. It is also supported by morphological characters (e.g., cupulate stigma type) that were already recognized by J.R. Grant before molecular data was available. He therefore correctly excluded these taxa from Vriesea and erected the genus Werauhia with two sections. Werauhia sect. Jutleya (type species W. pedicellata (Mez & Wercklé) J.R. Grant) is holophyletic, but in association with taxa of W. sect. Werauhia in two grades, making the latter paraphyletic. To retain the current classification, W. sect. Werauhia would have to be split into two groups based on the current taxon sampling. However, such actions awaits significantly increased species numbers in future studies. 156 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 Vrieseinae Subtribe Vrieseinae is composed of the two well supported genera Alcantarea and Vriesea as they are defined in this study. Characteristics for this subtribe are ovules with usually long appendages and geographic distributions with centers of diversity in eastern Brazil. Stigma types found in this clade are convolute-blade (Brown & Gilmartin, 1989b), tubolaciniate (Leme & Brown, 2004), conduplicate-patent (Versieux & Wanderley, 2007), conduplicate-erect. Alcantarea Contrary to V. subg. Vriesea, V. subg. Alcantarea (Mez, 1891–94) sensu Smith & Downs (1977) (type species A. regina (Vell.) Harms, lectotypified by Grant & Zijlstra, 1998) is a holophyletic group if a few misplaced species are excluded (Grant, 1995a). J.R. Grant reestablished the generic status proposed by Harms (1929) based on additional morphological evidence (Grant, 1995a). Morphological characters of this genus have been recently revised by Leme (2007, 2009) and the whole Alcantarea by Versieux (2009) and Versieux & al. (2010, 2012). Brown & Gilmartin (1989b) and Grant (1995a) originally indicated the stigma type of Alcantarea to be the same as Vriesea (convolute-blade). Figure 23 of Brown & Gilmartin (1989b) is wrongly assigned to V. geniculata (Wawra) Wawra (= A. geniculata (Wawra) J.R. Grant) and displays a true Vriesea species; indications of a convolute-blade stigma morphology for other Alcantarea species (as Vriesea) in their appendix are also incorrect. Leme (2007) demonstrated that the stigma of Alcantarea is of a different type than that of Vriesea. In his opinion, the stigma is conduplicate-spiral, but this view is largely not in concordance with the original idea of this stigma type by Brown & Gilmartin (1989b). Therefore two new stigma terms are used (conduplicate-patent and conduplicate-erect, respectively) as additional morphological characters to define Alcantarea. These new stigma types are illustrated either in Versieux & Wanderley (2007) and Leme (2007), or Leme (2009). Vriesea Vriesea subg. Vriesea sensu Smith & Downs (1977; type species V. psittacina) is an assemblage of different phylogenetic units including Werauhia, species now assigned to the CipuropsisMezobromelia clade, the former xerophytic, gray-leaved Vriesea species and T. malzinei, the latter two in fact belonging to Tillandsia (Terry & al., 1997b). Utley (1978, 1983) and Grant (1995) already recognized that V. subg. Vriesea was an assemblage of different evolutionary units, and Grant (1995) finally separated Werauhia (including the type species) from the rest of V. subg. Vriesea sect. Xiphion. Former xerophytic, gray-leaved Vriesea species were excluded in a series of four publications (Grant, 1993b, 1994b, 1995b, 2005). First evidence that T. malzinei has to be assigned to Tillandsia s.str. was provided by Terry & al. (1997b) and this is confirmed in the present study. Former Vriesea species of the Cipuropsis-Mezobromelia clade have never been excluded from V. subg. Vriesea, but differ clearly in not having the convolute-blade or tubolaciniate stigma type. These two stigma types are therefore the main characteristics of the predominantly Brazilian Vriesea species and are used to differentiate Vriesea s.str. in conjunction with habit or petal appendages from the remaining tillandsioid genera. Sections Vriesea and Xiphion p.p., typo excluso of Vriesea s.str. are para- and/or polyphyletic, but placement of species assigned to different sections is not by random in the phylogenetic context. Xiphion characters (included stamens) seem to be plesiomorphic within Vriesea, evident from clades in a grade at the base of Vriesea. Vriesea sect. Vriesea is derived (exserted stamens), but pa157 BARFUSS, M.H.J. NEW CLASSIFICATION raphyletic, because one lineage within this clade shows Xiphion characters, a possible reversal to the ancestral state. This hypothesis is supported by both the parsimony and Bayesian trees, but has no statistical support, which demonstrates the need of further molecular investigations of Vriesea for achieving a proper infrageneric classification. Tillandsieae Contrary to Vrieseeae, which is currently resolved as a holophyletic unit only by plastid DNA data, Tillandsieae shows a holophyletic origin with moderate or high BP in nearly all DNA markers studied except in the PRK tree, where core Tillandsioideae clades display a polytomy (Figure 1, D3). This tribe contains Guzmania, as well as the core Tillandsieae genera Josemania, Lemeltonia, Racinaea, Rothowia, and Tillandsia. Since inclusion of Mezobromelia in Barfuss & al. (2005) was based on a misidentified Guzmania species and the actual type of Mezobromelia is placed in Vrieseeae, it is excluded here from Tillandsieae. The clade corresponds to what Crayn & al. (2004) considered as the core Tillandsioideae in their study (see above). Guzmania Circumscription of Guzmania (type species G. monostachia) sensu Smith & Downs (1977) has not changed over more than three decades. Some taxa were transferred from other genera to Guzmania after publishing of the monograph only because complete floral characteristics were available for some species and clearly indicated that they were misplaced according to the generic definition (e.g., Utley, 1978; Luther, 1998). Current DNA results mostly confirm the old circumscription with the exception of the inclusion of one morphologically distinct group of former Mezobromelia with laxly-branched inflorescences and exposed stamens. Out of this group only G. hutchisonii was investigated. This taxon was initially treated as Tillandsia species by Smith & Downs (1977), but was transferred to Mezobromelia based on the floral character of conglutinate/connate petals bearing appendages (Weber & Smith, 1983). It differs from Mezobromelia sensu typo (M. bicolor) in the inflorescence architecture (candelabra-like with compact lateral branches in Mezobromelia s.str. vs. laxly-branched in G. hutchisonii) and in the stamen structure (conglutinate anthers and included stamens in Mezobromelia s.str. vs. free anthers and exserted stamens in G. hutchisonii). Therefore, the exclusion from Mezobromelia s.str. and the transfer of G. hutchisonii to Guzmania is justified genetically and morphologically. In addition, evidence from plastid DNA data (M.H.J. Barfuss, unpubl. data) suggests close relationships to G. diffusa L.B. Sm. and G. bakeri (Wittmack) Mez. A number of Guzmania species which have not been analyzed as well (e.g., G. candelabrum (André) André ex Mez, G. dalstroemii H. Luther, G. hirtzii H. Luther, and G. lemeana Manzan.) are showing similar inflorescences architecture and therefore might also be related to G. hutchisonii (Smith & Downs, 1977). Mezobromelia fulgens L.B. Sm. is also closely related genetically (M.H.J. Barfuss, unpubl. data). Further Mezobromelia species with free and exposed anthers, which are supposed to be closely related to G. hutchisonii and therefore believed to belong to the same clade but have not been investigated, are M. brownii H. Luther and M. lyman-smithii Rauh & Barthlott. Apart from these hypotheses concerning relationships, however, a distinction between Mezobromelia and Guzmania based just on the character of petal appendages is clearly inappropriate and does not reflect true phylogenetic relationships. Whether the way how the corolla tube is developed (either early or late sympetalous; Leins & Erbar, 2010) could serve as a discriminative character (most probably truly connate in Mezobromelia vs. only conglutinate in Guzmania) needs a further, detailed investigation. 158 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 A holophyletic origin of Guzmania is supported by three datasets (PHYC D2, nuclear combined D4, and total combined D5), whereas in dataset D3 (PRK) Guzmania clades display a polytomy with other core Tillandsioideae clades (Figure 1). Present in all phylogenetic trees are three main clades of Guzmania taxa (shown only in Figure 2), i.e., (1) G. hutchisonii, (2) Guzmania s.str., and (3) the remaining Guzmania species (incl. Sodiroa). Guzmania appears paraphyletic in the plastid DNA study of Barfuss & al. (2005), but the sample investigated as M. pleiosticha (Griseb.) Utley & H. Luther (acc. no. MB-15 = B0015) turned out to be a misidentified G. variegata. Guzmania hutchisonii appeared as the earliest diverging lineage within Tillandsieae in Barfuss & al. (2005). This is most probably due to numerous autapomorphies for the species responsible for LBA, which is further enforced by the low number of parsimony-informative plastid DNA characters and missing samples. The genetic divergence of clades 1 and 3 from Guzmania s.str. (clade 2) in all markers investigated is also reflected by morphology and indicates derived status within the genus. The hypothesis that in clade 3 a recent radiation has taken place is supported by short branches relative to taxa of Guzmania s.str. and G. hutchisonii. Nevertheless, PHYC and combined nuclear DNA sequences data consistently place all three clades into one linage (BP 85 vs. 61, respectively) and total-combined data give good resolution, and again increasing statistical support (BP 68/PP 100) indicating the whole group to be a natural entity. Support for the separation of a genus Sodiroa (e.g., Baker, 1889; Harms, 1930; Mez, 1934–35; not accepted by but mentioned in the key of Smith & Downs, 1977: 1284) that is defined by connate sepals is currently not provided by DNA data, and this character probably evolved convergently or in parallel, or reversals to the ancestral state (free sepals) have occurred. This is in contrast to the acceptance by Betancur & Miranda-Esquivel (1999), who reevaluated morphological characters phylogenetically without the application of molecular tools. Further generic segregations or infrageneric divisions of clades (e.g., the three main clades) need a careful evaluation of morphological characters, more relatives of G. hutchisonii, and more genetic evidence to obtain well resolved relationships within Guzmania. Core Tillandsieae = Tillandsieae s.str. This clade contains all segregates of Tillandsia since the treatment of Smith & Downs (1977) and can be referred to Tillandsia s.l. (= Tillandsia sensu Smith & Downs, 1977) if wrongly assigned taxa are excluded or included (see below). The nested placement of the here segregated genera Josemania, Lemeltonia, Racinaea, and Rothowia within Tillandsia in the earlier study of Barfuss & al. (2005) obviously was an artifact caused by the low number of parsimonyinformative characters. The current study clearly shows that all four segregates constitute holophyletic lineages not nested within Tillandsia s.str. Since Tillandsia sensu s.l. is morphologically very heterogeneous and the already earlier segregated genus Racinaea is widely accepted, recognition of Josemania, Lemeltonia, and Rothowia is necessary for these morphologically divergent holophyletic lineages. Josemania This new genus (type species J. lindenii) contains taxa previously ascribed to T. subg. Phytarrhiza based on the character of broad and conspicuous petal blades. In 1869 Regel proposed T. sect. Wallisia for the single species T. lindenii. E. Morren (1870) changed the status and erected the genus Wallisia which was not followed by bromeliad taxonomists. According to Art. 52.1 of the international code of botanical nomenclature (McNeill & al., 2006), this name is illegitimate at generic level and therefore the new name Josemania is proposed. All species 159 BARFUSS, M.H.J. NEW CLASSIFICATION classified here, i.e., J. anceps, J. cyanea, J. lindenii, J. pretiosa and J. umbellata, are easily recognized by their inflorescence structure, characteristic flowers and the pinnatisect stigma morphology (already illustrated in Morren, 1871). Lemeltonia Species of this new genus (type species L. dodsonii) have also been part of former, polyphyletic T. subg. Phytarrhiza sensu Smith & Downs (1977) and appear to be the sister genus to Tillandsia s.str. according to the current generic circumscriptions. The genus seems to be somehow intermediate between Racinaea, Rothowia, Josemania on the one hand and Tillandsia s.str. on the other. Species assigned to this genus are L. acosta-solisii, L. cornuta, L. dodsonii, L. monadelpha, L. narthecioides, L. scaligera, and L. triglochinoides. Lemeltonia narthecioides has not been included into the recent phylogenetic analyses, but high support for the inclusion into this genus comes from previous (Barfuss & al.,2005) and unpublished DNA data of the first author as well as from morphology. Racinaea Racinaea (Grant, 1994a; type species R. cuspidata) constitutes a separate entity within tribe Tillandsieae, which remained questionable after investigation of plastid DNA markers (Barfuss & al., 2005). Racinaea subg. Racinaea consists of all species formerly assigned to T. subg. Pseudo-Catopsis plus R. dyeriana (syn. T. dyeriana), a species previously classified within T. subg. Phytarrhiza but showing similar morphological features except those of strongly asymmetrical sepals and inconspicuous petal blades. However, subsymmetric petals and conspicuous petal blades are also found in other Racinaea species (e.g., R. elegans (L.B. Sm.) M.A. Spencer & L.B. Sm., R. riocreuxii (André) M.A. Spencer & L.B. Sm.), which are without doubt part of this genus. Racinaea subg. Pseudophytarrhiza has been established for the two species R. venusta (type species) and R. hamaleana, which have been transferred here from T. subg. Phytarrhiza too. These two species are genetically and morphologically distinct from the rest of Racinaea, but phylogenetically always in sister position to R. subg. Racinaea with strong statistical support. Rothowia This is a very isolated linage composed of three described species, i.e., Ro. laxissima, Ro. platyrhachis, and the type species Ro. wagneriana, which previously have been treated under T. subg. Phytarrhiza. The genus is entirely mesophytic and displays well-established phytotelmata (water-impounding rosettes). The elate inflorescence rachises and the obconic stigma are distinctive. Tillandsia Tillandsia s.l. (type species T. utriculata, incl. T. subg. Pseudo-Catopsis = Racinaea) is a morphologically very heterogeneous, but holophyletic group with respect to few exceptions, which are now either found in Vrieseeae or have been transferred from there (see below). Its circumscription corresponds roughly to the clade indicated as core Tillandsieae (Figure 2). Subsequent to the monograph, parts of the genus were studied to enable better recognition of individual species groups. The segregation of Racinaea and Viridantha has caused dispute among bromeliad taxonomists, among whom the first genus is broadly accepted and the latter not. 160 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 Only on the base of our molecular data we can propose a reliable classification for several segregates (see above). However, Tillandsia s.str. remains in a provisional state. Currently Tillandsia s.str. is circumscribed as the terminal supported backbone clade of our phylogenetic reconstruction (Figure 2) which includes either T. subg. Pseudalcantarea (parsimony analysis) or T. subg. Pseudovriesea (Bayesian analysis) as the most ancient lineage. Excluded are morphologically and genetically clearly distinguishable taxa (Josemania, Lemeltonia, Racinaea, Rothowia), which were previously treated as or integrated into separate subgenera (T. subg. Pseudo-Catopsis and meso-/semi-mesophytic T. subg. Phytarrhiza, respectively). The genus is divided into several subgenera and informal infrageneric clades, with their formal status depending mainly on phylogenetic support values (BP and PP), the number of investigated species, morphological characters to define them, and already well-established taxa in literature. As indicated under individual subgenera or infrageneric clades, their species circumscriptions have sometimes significantly changed since the monograph of Smith & Downs (1977). Tillandsia subg. Allardtia Tillandsia subg. Allardtia is not accepted under Tillandsia. The subgenus in the sense of Smith & Downs (1977) is an artificial assemblage of at least 14 phylogenetic units according to the current species sampling: (1) the Singularis clade is placed within Cipuropsidinae, (2) T. guatemalensis (type species of T. subg. Allardtia), T. selleana Harms, and T. leiboldiana Schltdl. are forming one group and are nested within T. subg. Tillandsia together with (3) T. adpressiflora Mez, (4) T. secunda Knuth, and (5) T. remota Wittm. in different positions; (6) T. barthlottii and T. myriantha are grouping with T. subg. Pseudovriesea; (7) the Tillandsia tectorum complex and the Tillandsia plumosa complex (= Viridantha) are closely related and form a distinct unit (T. subg. Viridantha); (8) xerophytic Andean members of T. subg. Allardtia are located in the Xiphioides clade together with Andean species of former T. subg. Anoplophytum; (9) T. disticha (i.s. = incertae sedis) is a taxon genetically very isolated from all other Tillandsia species; (10) T. edithae (i.s.) and (11) T. australis (i.s.) are currently unclassified, but show affinities to T. subg. Anoplophytum s.str; (12) T. pseudomicans (i.s.) and (13) T. sphaerocephala (i.s.) are also unclassified because of unstable positions in different types of analysis; (14) and the rest of T. subg. Allardtia sensu Smith & Downs (1977) containing most species seems to be a distinct holophyletic group, but presently not well supported because of low variability in both plastid and nuclear DNA markers. Therefore the Singularis clade (phylogenetic unit 1) is shifted to Cipuropsidinae; T. subg. Allardtia sensu typo has to be considered as a synonym of T. subg. Tillandsia with the six taxa mentioned earlier transferred there (2–5); T. barthlottii and T. myriantha are classified within T. subg. Pseudovriesea (6); the Tillandsia tectorum complex is transferred to T. subg. Viridantha (7); xerophytic Andean members are transferred to the Xiphioides clade (8); T. australis, T. disticha, T. edithae, T. pseudomicans, and T. sphaerocephala are unclassified (9–13); and the rest of T. subg. Allardtia (14) is provisionally treated as a distinct clade without formal nomenclatoric adjustments (Biflora clade). Tillandsia subg. Anoplophytum Current and previous results show that T. subg. Anoplophytum sensu Smith & Downs (1977; type species T. stricta) is (1) paraphyletic (regarding the clade that is most inclusive for species previously classified under T. subg. Anoplophytum; see Figure 2, S (BP 84/PP 100), including 161 BARFUSS, M.H.J. NEW CLASSIFICATION Gardneri clade, excluding T. australis), since (a) both T. subg. Diaphoranthema and T. subg. Phytarrhiza s.str. are nested within this clade, (b) the Xiphioides clade is in terminal position containing also species formerly assigned to T. subg. Allardtia, and (c) the currently unclassified T. edithae (T. subg. Allardtia sensu Smith & Downs, 1977) is also included; and (2) polyphyletic because T. macbrideana and T. pseudomacbrideana, elements of former T. subg. Anoplophytum, are now part of the Biflora clade. A postulated holophyletic origin of T. subg. Anoplophytum by Tardivo (2002) based on a phylogenetic analysis of selected morphological characters is not supported, mainly because subgeneric circumscription was according to Smith & Downs' vague morphological definition. The distinction between T. subg. Anoplophytum and T. subg. Allardtia based on the character of "plicate vs. straight" filaments is obviously inappropriate and does not reflect true relationships. Support for this view comes not only from the phylogenetic results but also from ontogenetic studies on species from other subgenera (Evans & Brown, 1989), where plicate filaments are also reported for T. subg. Allardtia and T. subg. Tillandsia. In its current circumscription T. subg. Anoplophytum is restricted to T. stricta and its close relatives, which have their distribution center in south-eastern Brazil. The taxon is characterized by simple inflorescences with polystichously arranged flowers. Tillandsia subg. Diaphoranthema In the current analyses T. subg. Diaphoranthema (type species T. recurvata (L.) L.) is a paraphyletic lineage within Tillandsia, but this is not well supported. The investigated species are forming a clade in the parsimony analysis in sister position to the Xiphioides clade, but with T. subg. Phytarrhiza s.str. nested inside. In the Bayesian analysis the three clades composed of species of T. subg. Diaphoranthema (Figure 2, S), i.e., Capillaris clade (T. capillaris, T. funebris A. Cast., T. kuehasii W. Till, and T. virescens) Usneoides clade (T. landbeckii ssp. andina W. Till, T. mollis H. Hrom. & W. Till, T. usneoides (L.) L., and T. cf. usneoides), and T. recurvata, are associated in different degrees with both T. subg. Phytarrhiza and the Xiphioides clade. To test for a holophyletic origin of T. subg. Diaphoranthema in the current circumscription additional DNA markers and taxa are required. The stigma type of all species in this subgenus is simple-erect according to Brown & Gilmartin (1989b), lobes are short rendering a capitate stigma resembling that of T. subg. Phytarrhiza s.str. This morphological similarity points towards a close relationship of both subgenera in their actual circumscription (Till, 1992). Both taxa share an orange pollen color which is absent in remaining Tillandsia. Tillandsia subg. Phytarrhiza Tillandsia subg. Phytarrhiza sensu Smith & Downs (1977; type species T. duratii) is clearly artificial and polyphyletic. The division into mesophytic, semi-mesophytic and xerophytic members was a first attempt of grouping these species into natural units (Gilmartin, 1983; Gilmartin & Brown, 1986; Till, 2000b). This view is only partly supported by DNA sequence data. Mesophytic and semi-mesophytic members belong to four different evolutionary units, three of them representing separate phylogenetic entities, which deserve generic status (Josemania, Rothowia, Lemeltonia). The fourth, containing R. dyeriana, R. hamaleana and R. venusta, is strongly associated with Racinaea and therefore these species are transferred to this genus. Xerophytic members likewise are a polyphyletic group within Tillandsia. They also display at least two phylogenetically distinct units, (1) species associated with T. purpurea Ruiz & Pav. (Purpurea clade), and (2) species around T. duratii, showing strong affinities to T. subg. Diaphoranthema. In the current treatment T. subg. Phytarrhiza is restricted to T. duratii and its close 162 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 relatives and is accepted as a separate subgenus, despite of its nested position (see "Classification of Tillandsioideae" for reasons). Other taxa examined so far are T. kirschnekii Rauh & W. Till, T. paleacea C. Presl, and T. aff. streptocarpa Baker. Whether its subgeneric status can be kept or T. subg. Diaphoranthema must be merged as suggested by Till (1992) needs further investigation of more species and additional DNA markers to obtain well-resolved relationships. Polyphyletic Tillandsia subg. Phytarrhiza in the sense of Smith & Downs (1977) has been studied by Gilmartin (1983) and Gilmartin & Brown (1986) in more detail, and by Till (1992) for a morphological comparison to T. subg. Diaphoranthema (see above). The first two studies used cladistic analyses of morphological characters (mainly obtained from Smith & Downs, 1977) primarily to test hypotheses concerning evolution of xerophytes and mesophytes within the subgenus. The main problem of both studies is the assumption that T. subg. Phytarrhiza is holophyletic, which does not match the molecular results. Their postulation for monophyly is based on the character of broad and conspicuous petal blades, which is obviously a character that evolved convergently or in parallel in different lineages. Identification of species groups within T. subg. Phytarrhiza (and T. subg. Diaphoranthema as sister or outgroup) of Gilmartin & Brown (1986) prior to the computer analyses for phylogenetic reconstruction is very critical and resulting trees should be taken with caution. Initial analyses of Gilmartin (1983) are more interesting, since analyses are based on single species and not on species groups. Three important clades from her phenetic analysis correspond to the newly erected genera Josemania, Lemeltonia, and Rothowia. The morphological study of T. subg. Diaphoranthema and T. subg. Phytarrhiza by Till (1992) support close relationships of both subgenera in their current circumscriptions. Tillandsia subg. Pseudalcantarea Tillandsia subg. Pseudalcantarea currently consist of three species: T. grandis Schltdl., T. macropetala , and T. viridiflora (type species). Further species that were previously assigned to this subgenus belong to different phylogenetic units. T. baliophylla lies within the Biflora clade and T. paniculata within T. subg. Tillandsia; therefore these two species are excluded from T. subg. Pseudalcantarea. Tillandsia heterophylla E. Morren, a simulator of morphological characters fitting the description of T. subg. Pseudalcantarea (Till, 2000b), also appears not closely related to T. subg. Pseudalcantarea. Instead, this species is part of group III species of Gardner's classification of T. subg. Tillandsia (Gardner, 1982, 1986b), which are, however, located in the Biflora clade and not part of T. subg. Tillandsia. The view of Beaman & Judd (1996) that T. subg. Pseudalcantarea should be restricted only to the species T. viridiflora and T. baliophylla, and the transfer of T. grandis to T. subg. Tillandsia is not supported by neither plastid DNA nor nuclear DNA sequence data. Tillandsia subg. Pseudalcantarea with its three here accepted species forms a distinct group being sister to the remaining clades of Tillandsia in the parsimony analysis (Figure 2), but this position is only weakly supported and not present in Bayesian analysis (Figure 2, R, here T. subg. Pseudovriesea is the earliest branch within Tillandsia s.str.). Future studies including more DNA markers will help clarifying the correct phylogenetic position of this subgenus which might merit generic status. Tillandsia subg. Pseudovriesea The new T. subg. Pseudovriesea (type species T. tequendamae) is composed of all xerophytic, grey-leaved species previously assigned to Vriesea and some Tillandsia species which were 163 BARFUSS, M.H.J. NEW CLASSIFICATION either never formally transferred to Vriesea, despite having petal appendages (i.e., T. myriantha, formerly part of T. subg. Allardtia), or because the investigated material did not show appendages on the petals so far (T. barthlottii, T. spathacea; the former being previously part of T. subg. Allardtia, the latter of T. subg. Tillandsia). J.R. Grant already recognized the close relationship to Tillandsia based on overall morphological similarity (habit) and differences in stigma morphology to Vriesea s.str. (convolute-blade or tubolaciniate vs. conduplicate-spiral). He transferred all xerophytic, grey-leaved Vriesea species to Tillandsia, at first those with exserted stamens and styles (Grant, 1993b, 1994b, 1995b), later also species with included stamens and styles (Grant 2005). The main difference to T. subg. Tillandsia is that most of these species have retained (or regained) their petal appendages, a character which is rarely found in other species of Tillandsia. Unfortunately, petal appendages are not synapomorphic for all taxa within T. subg. Pseudovriesea, since few members do not show petal appendages so far in the investigated flowers. To really address this issue populational, investigations on petal appendages would be necessary. The long standing debate on whether xerophytic, grey-leaved Vriesea species are members of Vriesea (Smith & Downs, 1977), of T. subg. Tillandsia (Grant, 1993b, 1994b, 1995b), or of a separate phylogenetic unit within Tillandsia (Barfuss & al., 2005), is now resolved. They fall neither into Vriesea s.str. nor into T. subg. Tillandsia (as Grant originally thought when he did the first rearrangements of taxa with exserted stamens and styles), but they are clearly members of Tillandsia deserving a subgenus of their own. Evidence from all plastid and nuclear DNA markers are supporting this view and earlier speculations of a possible ancient hybrid origin of this group between any Vriesea and Tillandsia species can be rejected. Tillandsia subg. Tillandsia Species of T. subg. Tillandsia were consistently separated from the other subgenera since Mez (1896) based on the floral characters of exserted stamens and styles, which was followed by Smith & Downs (1977). Gardner (1982, 1986b) established 5 groups within T. subg. Tillandsia sensu Smith & Downs (1977), but also included some possibly wrongly classified Allardtia species. Based on results of the current phylogeny, this subgenus should be restricted (1) to group I and II species of her classification, (2) to T. filifolia Schltdl. & Cham. of group IV, which in her investigations shows also affinities to group II, as well as (3) to some taxa previously assigned to T. subg. Allardtia (T. adpressiflora, T. guatemalensis, T. leiboldiana, T. remota, T. secunda, T. selleana). As the type species of T. subg. Allardtia, T. guatemalensis, is nested within T. subg. Tillandsia, Allardtia necessarily becomes a synonym of Tillandsia subg. Tillandsia. Group V species are actually species of T. subg. Viridantha, group III species are mainly nested within the Biflora clade and the second species of her group IV, T. disticha, shows currently no supported affinity to any infrageneric unit within Tillandsia and is therefore unclassified. Wrongly assigned is T. rauhii (group III, now Rauhii clade). Some species with undetermined affiliation in Gardner's treatment can now be assigned to clades within Tillandsia: T. cryptopoda L.B. Sm. and T. plagiotropica Rohweder clearly belong to T. subg. Tillandsia, group I; T. extensa Mez is part of group II, T. ecarinata, T. ferreyrae and T. teres are members of the Rauhii clade and not supported to belong to T. subg. Tillandsia. Support for subgroups of Gardner's group I (1986b) is currently also not given (see also Granados Mendoza, 2008). Although the subgenus has currently no statistical support, validation for the inclusion of the investigated taxa comes from previous studies and unpublished data (Terry & al., 1997b; Barfuss & al., 2005; M.H.J. Barfuss, unpubl. data). Four main clades can be recognized: (1) a 164 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 clade containing T. secunda (T. subg. Allardtia sensu Smith & Downs, 1977), T. malzinei (V. sect. Vriesea sensu Smith & Downs, 1977), and T. paniculata (T. subg. Pseudalcantarea sensu Smith & Downs, 1977); all these taxa have not been included into T. subg. Tillandsia before; (2) a clade that W. Rauh considered as the Tillandsia mima complex (Rauh, 1973; group II species of Gardner, 1986b) including T. adpressiflora (T. subg. Allardtia sensu Smith & Downs, 1977) and T. hildae Rauh (group II, Gardner, 1986b); (3) a clade which contains T. utriculata and therefore could be recognized as T. subg. Tillandsia s.str. (also group II species plus T. filifolia of group IV, Gardner, 1986b); and (4) a clade containing the majority of species (group I species, Gardner, 1986b, plus T. subg. Allardtia s.str.). Interesting are the geographical distributions of these clades. Whereas the early diverging clades 1 and 2 are outside the diversity center of T. subg. Tillandsia and mainly found in northern South America and the Caribbean, the two core clades have their main distribution in Mexico and Central America. This pattern gives important hints for future biogeographical studies in Tillandsioideae. Pseudobulbous taxa of T. subg. Tillandsia were studied by Chew & al. (2010). Results based on ITS 2 and ETS data support the polyphyletic origin of this habit within T. subg. Tillandsia and even within the whole genus Tillandsia, although statistical support unfortunately is mostly weak. An enriched taxon sampling and additional molecular markers would help clarifying the interesting question, how often the pseudobulbous habit arose within Tillandsia. Tillandsia malzinei is a special case within T. subg. Tillandsia. Genetically, from both nuclear and plastid DNA data, this species is doubtlessly nested in the Tillandsia clade, despite bearing appendages at the base of the petals. It is genetically very distinct from other Tillandsia species, but shows some genetic affinities to T. funckiana (Terry & al., 1997b) and other members of T. subg. Tillandsia, but not to T. subg. Pseudovriesea. Therefore we are currently treating this species as a strongly isolated taxon within T. subg. Tillandsia, due to its high number of molecular autapomorphies and the distinct morphology. Tillandsia subg. Viridantha Viridantha (Espejo-Serna, 2002) with T. plumosa as type species does not deserve generic rank under subfam. Tillandsioideae based on the current generic circumscriptions. Taxa of Viridantha clearly are member of Tillandsia s.str., but not related to species of T. subg. Tillandsia (group V: Gardner, 1982, 1986b) or associated with the Biflora clade (former T. subg. Allardtia sensu Smith & Downs, 1977). This is also supported by flower morphology (Brown & Gilmartin, 1989b) since some Viridantha species resemble the convolute-blade stigma type (W. Till & M.H.J. Barfuss, unpubl. data; but also illustrated in Gardner (1982) and Espejo-Serna (2002)), a character not found in any other subgroup of Tillandsia. Instead, these taxa are closely related to the Tillandsia tectorum complex (Hromadnik, 2005), which was previously included in T. subg. Allardtia and T. subg. Anoplophytum sensu Smith & Downs (1977). We therefore consider species of Viridantha and the Tillandsia tectorum complex as members of a separate subgenus. An internal classification of this subgenus needs a deeper and more comprehensive study of all species assigned to T. subg. Viridantha. The sectional treatment of Espejo-Serna (2002: Vi. sect. Viridantha and Vi. sect. Caulescens) is currently not supported. Descriptions of new taxa during the last few years and monographic treatments of both species complexes (e.g., Espejo-Serna, 2002; Hromadnik, 2005; Ehlers, 2009) using traditional, morphology based taxonomy show the need of a taxonomic revision also on a (phylo-)genetic basis. The placement of T. sphaerocephala as sister to T. subg. Viridantha in the parsimony analysis seems to be an artifact with no statistical support caused by low sequence variability and 165 BARFUSS, M.H.J. NEW CLASSIFICATION the lack of other species related to this taxon. Bayesian analysis places T. sphaerocephala in sister position to the Purpurea clade (Figure 2, L), which in the parsimony analysis groups with unclassified T. nana and T. pseudomicans, a possible artifact as well (PP 60). Biflora clade This is one of the most critical species complexes within Tillandsia (name-giving taxon T. biflora) and comprises mostly taxa previously assigned to T. subg. Allardtia. Some members of this clade have also been classified in other subgenera, but clearly belong into this phylogenetic unit: (1) most group III species of Gardner's treatment of T. subg. Tillandsia (Gardner 1982, 1986b), i.e., T. deppeana; T. heterophylla, T. imperialis, and T. multicaulis, (2) T. baliophylla of T. subg. Pseudalcantarea sensu Smith & Downs (1977), and (3) elements of T. subg. Anoplophytum sensu Smith & Downs (1977), i.e., T. macbrideana and T. pseudomacbrideana. Because the type species of T. subg. Allardtia, T. guatemalensis, is now part of T. subg. Tillandsia, a new subgenus would have to be created for these taxa. DNA data weakly links the species to a holophyletic group with no statistical support and definite morphological characters are currently missing. Therefore, we keep this complex as an informal clade under Tillandsia. Although the huge majority of species was previously classified within T. subg. Allardtia, included style and stamens with no filament plication can obviously not serve as an autapomorphic character for this complex. Gardneri clade Earlier, this group of species (name-giving taxon T. gardneri) was classified within T. subg. Anoplophytum. It is composed only of rose-flowered species distributed in eastern Brazil. The DNA results suggest that this species complex has had a relatively long time of independent evolution. Ehlers (1997) monographed them as "red-flowered" species and included 14 Brazilian members plus T. paraensis Mez. The latter belongs to T. subg. Tillandsia but was not investigated. We investigated only 3 species (T. brachyphylla, T. gardneri, T. globosa) and are not confident enough to propose a separate formal unit for these taxa. Therefore, this complex is recognized as an informal clade within Tillandsia. Purpurea clade As a part of former T. subg. Phytarrhiza, this clade contains xerophytic members (name-giving taxon T. purpurea) and is in an isolated position within Tillandsia and not closely related to any other subgroup. It comprises the following taxa investigated so far: T. aurea, T. cacticola, T. purpurea, and T. straminea. Species of the Purpurea clade are distinguished from T. subg. Phytarrhiza s.str. (simple-erect stigma type) by its conduplicate-spiral stigmas (Brown & Gilmartin, 1989b) and the often bicolored petals (crème-colored with violet tips), but for establishing a separate subgenus more taxa and accessions have to be investigated. Tillandsia nana and T. pseudomicans are sister to this clade in the parsimony analysis, but with weak BP (57). In Bayesian analysis T. sphaerocephala is sister to the Purpurea clade, but also only weakly supported. Both scenarios are probably an artifact caused by missing Tillandsia taxa and sequence information. Rauhii clade Species of the Rauhii clade (T. ecarinata, T. ferreyrae, the name-giving T. rauhii, and T. teres) have been previously classified within T. subg. Tillandsia (Smith & Downs, 1977). Gardner 166 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 (1982, 1986b) treated T. rauhii within group III of T. subg. Tillandsia, but this taxon is neither related to this group of species, nor currently placed within T. subg. Tillandsia. Tillandsia ecarinata, T. ferreyrae and T. teres were not classified into any subunit of T. subg. Tillandsia and treated as "incertae sedis" by Gardner (1982, 1986b), since floral characteristics could not be studied at that time. Despite of missing plastid DNA sequence data for all taxa except T. rauhii, placement of all four species within the Rauhii clade is still supported in parsimony analyses with a weak BP (57). The Rauhii clade shows an isolated position within Tillandsia, but backbone relationships between this and other phylogenetic units are unsupported. Placement is very likely to change if more DNA sequence data is gathered. Thus no formal unit is proposed for species of the Rauhii clade until placement of these taxa is clarified. Xiphioides clade Taxa falling into this phylogenetic unit (name-giving taxon T. xiphioides) have been mainly placed within T. subg. Anoplophytum based on the character of plicate filaments. According to the tree topologies, this group constantly is closely related to subgenera Anoplophytum s.str., Diaphoranthema, Phytarrhiza s.str., the Gardneri clade and some unclassified taxa, i.e., T. albertiana, T. edithae and T. esseriana (Figure 2). Till (2000b) recognized close relationships of taxa under both T. subg. Allardtia and T. subg. Anoplophytum sensu Smith & Downs (1977), but shifted the xerophytic Andean members of T. subg. Anoplophytum with plicate filaments to Allardtia. Genetically they are neither related to T. subg. Allardtia s.str. (which is now a synonym of T. subg. Tillandsia) nor to the rest of former Allardtia (= Biflora clade). From an evolutionary point of view it would have been better to shift xerophytic Andean members of T. subg. Allardtia with straight filaments into T. subg. Anoplophytum. As T. subg. Anoplophytum is currently circumscribed, the Xiphioides clade would need a separate infrageneric unit. However, due to the unresolved relationships of T. subg. Phytarrhiza and T. subg. Diaphoranthema and the unclassified taxa mentioned earlier, as well as currently missing morphological characteristics, the Xiphioides clade is treated here as an informal unit within Tillandsia. Incertae sedis—Unclassified Tillandsia species Taxonomists always try to put taxa into a hierarchical scheme. Unfortunately there are a number of Tillandsia species that show either no or only unsupported affinities to certain groups, but appear as separate, maybe relict ancient lineages, where their ancestors either did not diversify or related species became extinct or have yet not been discovered. Unclassified investigated taxa so far are T. albertiana, T. australis, T. disticha, T. edithae, T. esseriana, T. nana, T. pseudomicans and T. sphaerocephala. Morphological characteristics of these species seem to be either plesiomorphic or derived involving convergence or parallelism. Currently we cannot assign these species with good phylogenetic support values to any infrageneric unit. Additional taxa and new molecular markers may help to place these taxa with better confidence in the phylogenetic framework. 167 BARFUSS, M.H.J. NEW CLASSIFICATION Classification of Tillandsioideae In classifying Tillandsioideae we considered aspects of biological (evolutionary) classification of Hörandl & Stuessy (2010) and agree in principle with statements of Nickrent & al. (2010), who gave a good overview of their philosophy in reclassifying Santalales into different families. Although their classification was done on a higher level, their principal points apply in our view to any taxonomic level. Classifications are most useful when they simultaneously serve many functions as stressed by both studies mentioned above. Nickrent & al. (2010) list several criteria used to decide how to circumscribe taxonomic units. For the taxonomy of Tillandsioideae we are currently emphasizing the following (Nickrent & al., 2010, slightly modified), arranged according to their importance : (1) rejection of polyphyly, support for holophyly (= monophyly s.str.) and paraphyly, if well supported, (2) phylogenetic information (= minimizing redundancy), (3) stability (= minimizing nomenclatural changes), (4) ease of identification (= recognizability), (5) recognition of groups that are well-established in the literature, but where supporting information is currently partly missing. Hörandl & Stuessy (2010: 1650) list also four main criteria (theoretical foundation based on natural processes, predictivity, information content, and practicability), which are similar to those mentioned above, and they recommend to adhere to these principles of biological classification by following a protocol of five procedural steps. They emphasized the recognition of paraphyletic groups as natural units of biological classification. Contrary to Nickrent & al. (2010) we agree with their opinion that paraphyly is a natural stage in the evolution of taxa, and that it occurs regularly along with holophyly. However, in our view it should only be accepted for a final classification if all possible biological sources are considered and well-supported conclusions can be drawn. For an acceptance of conclusive, well-supported paraphyletic groups within Tillandsioideae, information from biological sources (in this study from DNA sequence data and morphology) is currently insufficient. Statistically unsupported paraphyly is preliminarily accepted for one already wellestablished taxon (T. subg. Diaphoranthema), where additional information could also lead to holophyly; unsupported paraphyly may also be tentatively accepted in a not yet fully resolved classification for a group, in which additional information is currently missing; it may be most relevant for most-recently divergent plant groups and at the species level. Therefore, until conclusive and well-supported clades in the complete phylogeny of all Tillandsioideae are presented, we are not describing or accepting new taxonomic units based on paraphyly, but it could be of major importance in a future, more detailed classification than presented here. Our main attempts are to adhere to the above emphasized five principles. Units that are well characterized are formally recognized, whereas others, for which significant supporting information is missing, are treated as informal clades or unclassified taxa (= incertae sedis) in our classification system. The subfamilial classification including accepted tribal, subtribal, generic and subgeneric ranks and informal clades is listed below in hierarchical, followed by alphabetical order. Two new subtribes (Cipuropsidinae, Vrieseinae), three new genera (Josemania, Lemeltonia, Rothowia), and three new subgenera (Racinaea subg. Pseudophytarrhiza, T. subg. Viridantha, and T. subg. Pseudovriesea) are described. Several species are therefore reclassified. Lectotypes are selected for Catopsis subg. Tridynandra Mez, Tillandsia sect. Conostachys Griseb., Tillandsia sect. Eriophyllum K. Koch. A comparison of accepted genera of Tillandsioideae and subgenera of Tillandsia to previous classification systems of either Smith & Downs (1977), Smith & Till (1998), Till (2000a, b), or Barfuss & al. (2005) can be found in Tables 4 and 5, re168 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 spectively. Phylogenetic relationships of all investigated species and their current classification are summarized in Figure 2 (see Appendix). A complete synonymy (homotypic and heterotypic) is presented for all accepted taxa above the species level and for all reclassified species, taxonomic synonyms with the monograph of Smith & Downs (1977) being the primary reference are given with investigated species mentioned to allow a comparison of the new vs. old taxonomic circumscriptions in the most critical groups. Morphological characteristics are listed for all taxa above the subgeneric level and for newly described infrageneric units and the geographical distributions of well-characterized clades can be obtained from the key to the genera. Bromeliaceae Juss. subfamily Tillandsioideae Burnett, Outl. Bot.: 442 (Jun 1835), as "Tillandsidae".—Type: Tillandsia L. Fruit a septicidal capsule, seeds with hairs or hairlike appendages, wind dispersed. Tribe Glomeropitcairnieae Harms, in Engl. & Prantl, Nat. Pflanzenfam. ed. 2, 15a: 100, 115 (1930).—Type: Glomeropitcairnia (Mez) Mez. Ovary about ½–⅔ inferior, capsule only partly septicidal; petal appendages present; stigma of the convolute-blade type; seeds with a long filiform, undivided chalazal appendage and a plumose micropylar flight apparatus; pollen with a diffuse aperture, exine reticulate. 1. Glomeropitcairnia (Mez) Mez, Bull. Herb. Boissier, sér. 2, 5: 232 (1905) ≡ Pitcairnia subg. Glomeropitcairnia Mez, in C.DC., Monogr. phan 9: 463 (1896).—Type: Tillandsia penduliflora Griseb. Tribe Pogospermeae Brongn., Ann. Sci. Nat. Bot., sér. 5, 1: 327 (1864).—Type: Pogospermum Brongn. (= Catopsis Griseb.). = Bromeliaceae tribe Catopsideae Harms, in Engl. & Prantl, Nat. Pflanzenfam. ed. 2, 15a: 100, 130 (1930).—Type: Catopsis Griseb. Ovary max. ⅛ inferior to completely superior, capsule septicidal; petal appendages lacking; stigma of the simple-erect type; flight apparatus of the seeds of multicellular chalazal hairs; pollen with sharply cut aperture margins, exine reticulate (Catopsis type). 2. Catopsis Griseb., Nachr. Königl. Ges. Wiss. Georg-August-Univ. [1]: 10, 12 (13 Jan 1864) ≡ Catopsis subg. Eucatopsis Mez, in C.DC., Monogr. phan. 9: 619 (1896).—Type: Tillandsia nitida Hook. = Tussacia Willd. ex Beer, Fam. Bromel.: 99 (1856), non Rchb., Iconogr. bot. exot. 1: x, 28 (1827: Gesneriaceae) nec Willd. ex Schult. & Schult. f., in Roem. & Schult., Syst. veg. 7(1): x, 57 (1829: Incertae sedis), nom illeg. (Art. 53.1) ≡ Tussaria Griseb., Fl. Brit. W. I.: 599 (1864), orth. var.—Type: Tillandsia vitellina Klotzsch = Catopsis nutans (Sw.) Griseb. = Pogospermum Brongn., Ann. Sci. Nat. Bot. Sér. 5, 1: 327 (Jun 1864).—Type: Pogospermum flavum Brongn. = Catopsis nutans (Sw.) Griseb. = Catopsis subg. Tridynandra Mez, in C.DC., Monogr. phan. 9: 620 (1896).—Lectotype (proposed): Catopsis morreniana Mez. 169 BARFUSS, M.H.J. NEW CLASSIFICATION Tribe Tillandsieae Rchb., Consp. regn. veg.: 62 (Dec 1828–Apr 1829) ≡ Bromeliaceae tribe Tillandsieae Dumort., Anal. Fam. Pl.: 55 (Oct 1829) ≡ Bromeliaceae tribe Tillandzieae Rchb., in Uphof, Pflanzengattungen: 205 (1910), orth. var.—Type: Tillandsia L. Ovary for max. ⅓ inferior, capsule septicidal; petal appendages usually lacking; stigma mainly of the conduplicate-spiral or simple-erect type, rarely of the convolute-blade type (Guzmania p.p., Rothowia, Tillandsia subg. Viridantha p.p.), occasionally of the coralliform type (Lemeltonia, Racinaea p.p.) or with pinnatisect margins (Josemania); seeds with a flight apparatus of pseudohairs at the micropylar end, endostome of type d (Groß, 1988) and embryo of type g (Groß, 1988); pollen mainly with diffuse apertures, occasionally of the insulae type, rarely of the Alcantarea type (former Vriesea imperialis type; Halbritter, 1988, 1992), or inaperturate. 3. Guzmania Ruiz & Pav., Fl. peruv. 3: 37, pl. 261 (1802).—Type: Guzmania tricolor Ruiz & Pav. = Guzmania monostachia (L.) Rusby ex Mez. = Caraguata Plum. ex Lindl., Bot. Reg. 13: sub pl. 1068 (1827), nom. illeg. (Art. 6.4), non Plum. ex Adans. (1763).—Type: Tillandsia lingulata L. = Devillea Bertero ex Schult. & Schult. f., in Roem. & Schult., Syst. veg. 7(2): 1229 (1830).— Type: Devillea speciosa, nom. nud. = Tillandsia sect. Conostachys Griseb., Nachr. Königl. Ges. Wiss. Georg-August-Univ. [1]: 19 (13 Jan 1864) ≡ Tillandsia subg. Conostachys (Griseb.) Baker, J. Bot. 26: 167 (1888) ≡ Vriesea subg. Conostachys (Griseb.) Mez, in Mart., Eichler & Urban, Fl. bras. 3(3): 516 (1894).—Lectotype (proposed): Tillandsia acorifolia Griseb. = Massangea E. Morren, Belgique Hort. 27: 59, 199, pl. 8, 9 (1877) ≡ Caraguata [Plum. ex Lindl.] subg. Massangea (E. Morren) Baker, Handb. Bromel.: 149 (1889).—Type: Tillandsia musaica Linden & André. = Sodiroa André, Bull. Soc. Bot. France 24: 167 "1877" (1878).—Lectotype (Smith & Downs, 1977: 1275): Sodiroa graminifolia André ex Baker. = Schlumbergeria E. Morren, Belgique Hort. 28: 311 (1878) ≡ Schlumbergera E. Morren, Belgique Hort. 33: 46 (1883), orth. var., non Lem. (1858: Cactaceae) ≡ Caraguata [Plum. ex Lindl.] subg. Schlumbergeria (E. Morren) Baker, Handb. Bromel.: 149 (1889).—Type: Schlumbergeria roezlii E. Morren. = Thecophyllum André, Bromel. Andr.: 107 (1889).—Type: Thecophyllum wittmackii André. = Chirripoa Suess., Bot. Jahrb. Syst. 72: 293, pl. 4, Fig. 11 (1942).—Type: Chirripoa solitaria Suess. = Guzmania polycephala Mez & Wercklé ex Mez – Mezobromelia L.B. Sm. p.p., typo excluso. Mesophytic, usually acaulescent; flowers mostly polystichously arranged; petals connate/conglutinate and mostly unappendaged; filaments free, anthers free or connate; pollen inaperturate or with diffuse aperture; stigma of the simple-erect or convolute-blade type; seeds with an endostome of type a and b, rarely e (Groß, 1988) and embryo of type a, rarely b (Groß, 1988). 170 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 Guzmania hutchisonii (L.B. Sm.) Barfuss & W. Till, nom. prov. ≡ Tillandsia hutchisonii L.B. Sm., Phytologia 13: 145, pl. 7, figs. 23, 24 (1966) ≡ Mezobromelia hutchisonii (L.B. Sm.) W. Weber & L.B. Sm., J. Bromeliad Soc. 33(3): 121 (1983).—Type: Hutchison & Wright 6801: US (holo), UC, USC (iso). = Mezobromelia trollii Rauh, Trop. Subtrop. Pflanzenwelt 21: 5, figs. 1–3 (1977).—Type: Rauh 40104: HEID (holo), US (iso). 4. Josemania Barfuss & W. Till, nom. prov. pro Wallisia (Regel) E. Morren ≡ Tillandsia sect. Wallisia Regel, Index Seminum Hort. Bot. Petrop. "1868": 92 (Mar 1869) ≡ Wallisia (Regel) E. Morren, Belgique Hort. 20: 97 (1870), nom. illeg. (Art. 52.1) ≡ Vallesia C. Wright ex Sauvalle, Anales Acad. Ci. Méd. Fís. Nat. Habana 8: 53 (1871), orth. var., non Vallesia Ruiz & Pav. (1794: Apocynaceae) ≡ Tillandsia subg. Wallisia (Regel) Baker, J. Bot. 26: 46 (1888).— Type: Tillandsia lindenii Regel. – Tillandsia subg. Phytarrhiza (Vis.) Baker p.p., typo excluso. Leaves narrowly triangular, usually longitudinally reddish striped near the base, inflorescence distichous, usually ovate-laceolate in outline (except Josemania umbellata), petals blue-violet with a strongly widened blade and constricted basally into a claw, stamens and style deeply included within the corolla, anthers subbasifixed, stigma lobes with pinnatisect and papillate margins, ovules slenderly cylindric, obtuse. Named after José Manuel Manzanares (1957–) from Quito, Ecuador, leading authority of Ecuadorian Bromeliaceae. Josemania anceps (Lodd.) Barfuss & W. Till, nom. prov. ≡ Tillandsia anceps Lodd., Bot. Cab. 8: pl. 771 (1823) ≡ Platystachys anceps (Lodd.) Beer, Fam. Bromel.: 80 (1856) ≡ Vriesea anceps (Lodd.) Lem., Ill. Hort. 6 (Misc.): 15 (1859) ≡ Phytarrhiza anceps (Lodd.) E. Morren, Belgique Hort. 29: 368, pls. 20, 21 (1879).—Type: Adam in Loddiges Hort. s.n.: CGE?, K? (holo), in absence: Original illustration. = Tillandsia lineatifolia Mez, in C.DC., Monogr. phan. 9: 686 (1896).—Type: Fendler 2447: GOET (holo). Josemania cyanea (Linden ex K. Koch) Barfuss & W. Till, nom. prov. ≡ Tillandsia cyanea Linden ex K. Koch, Wochenschr. Vereines Beförd. Gartenbaues Königl. Preuss. Staaten 10: 140 (1867).—Type: Linden Hort. s.n.: B? (holo). = Tillandsia lindenii E. Morren, Belgique Hort. 19: 321, pl. 18 (Nov 1869), as „lindeni", nom. illeg., non Regel (Mar 1869) ≡ Vriesea lindenii (E. Morren) Lem., Ill. Hort. 16: pl. 610 (1869) , nom. illeg. ≡ Tillandsia morreniana Regel, Gartenflora 19: 41 (1870) ≡ Wallisia lindenii (E. Morren) E. Morren, Belgique Hort. 20: 102 (1870), nom. illeg. ≡ Phytarrhiza lindenii (E. Morren) E. Morren, Belgique Hort. 29: 297 (1879), nom. illeg. ≡ Tillandsia lindenii [E. Morren] var. genuina E. Morren, Gard. Chron., ser. 2, 12: 460 (1879) ≡ Phytarrhiza lindenii var. genuina E. Morren, Belgique Hort. 29: 297 (1879).—Type: Wallis in Linden Hort. s.n.: LG? (holo). = Tillandsia coerulea Linden ex K. Koch, Wochenschr. Vereines Beförd. Gartenbaues Königl. Preuss. Staaten 13: 197 (1870), nom. nud. = Tillandsia lindenii [E. Morren] vera Dombrain, Floral Mag. 11: pl. 44 (1872).—Type: Original illustration. ?= Tillandsia lindenii [E. Morren] var. violacea hort. ex André, Rev. Hort. 58: 61 (1886).— Type: not indicated. 171 BARFUSS, M.H.J. NEW CLASSIFICATION = Tillandsia lindenii [E. Morren] superba rosea Dauthenay, Rev. Hort. 70: 539 (1898), nom. illeg.—Type: not indicated. = Tillandsia lindenii [E. Morren] vera superba Duval, Gartenwelt 5: 164, Fig. (1901), nom. illeg.—Type: not indicated. Josemania lindenii (Regel) Barfuss & W. Till, nom. prov. ≡ Tillandsia lindenii Regel, Index Seminum Hort. Bot. Petrop. "1868": 92 (Mar 1869) ≡ Wallisia lindenii (Regel) E. Morren, Belgique Hort. 20: 97 (1870), nom. illeg.—Type: Wallis in St. Petersburg Hort. s.n.: LE (holo). ?= Tillandsia lindeniana Regel, Gartenflora 18: 193, pl. 619 (Jul 1869).—Type: Linden Hort. s.n.: LE? (holo). = Tillandsia lindenii [E. Morren] var. luxurians E. Morren, Belgique Hort. 21: 289, pls. 20, 21 (1871) ≡ Phytarrhiza lindenii var. luxurians E. Morren (E. Morren), Belgique Hort. 29: 299 (1879) ≡ Tillandsia lindenii [E. Morren] var. luxurians (E. Morren) L.B. Sm., Contr. U. S. Natl. Herb. 29: 494 (1951), comb. illeg. superfl. ≡ Tillandsia lindenii [E. Morren] var. abundans L.B. Sm., Phytologia 20: 166 (1970), nom. nov. superfl.—Type: Belgique Hort. 21: pls. 20, 21 (1871). = Phytarrhiza lindenii var. koutsinskyana E. Morren, Belgique Hort. 30: 80 (1880) ≡ Tillandsia lindenii [E. Morren] var. koutsinskyana (E. Morren) L.B. Sm., Contr. U. S. Natl. Herb. 29: 494 (1951).—Type: Warsaw Hort. s.n.: LG?, WA? (holo). ?= Tillandsia lindenii [E. Morren var.] latispatha Van Houtte ex André, Rev. Hort. 60: 201 (1888).—Type: Van Houtte s.n.: K? (holo). = Tillandsia lindenii [E. Morren] var. duvalii Duval ex André, Rev. Hort. 71: 516 (1899) ≡ Tillandsia lindenii [Regel] var. × duvalii (Duval ex André) L.B. Sm., Contr. U. S. Natl. Herb. 29: 493 (1951) ≡ Tillandsia lindenii [Regel] var. duvaliana L.B. Sm., Phytologia 20: 166 (1970), nom. nov. superfl.—Type: Duval Hort. s.n.: ? = Tillandsia lindenii Hasack, Möllers Deutsche Garten-Z. 15: 93, Fig. (1900) non Regel (1869), nom. illeg.—Type: Original illustration. = Tillandsia × duvali Duval, Gartenwelt 5: 164, Fig. (1901).—Type: Duval Hort. s.n.: ? (holo), in absence: Original illustration. = Tillandsia lindenii [Regel] var. caeca D. Barry, Bromeliad Soc. Bull. 12: 5 (1962) ≡ Tillandsia lindenii [Regel] var. × caeca D. Barry, in Smith & Downs, Fl. Neotrop. 14(2): 846 (1977).—Type: Barry Hort. s.n.: US (holo). Josemania pretiosa (Mez) Barfuss & W. Till, nom. prov. ≡ Tillandsia pretiosa Mez, Repert. Spec. Nov. Regni Veg. 16: 78 (1919).—Type: Sodiro 171/39: B (holo). = Tillandsia cyanea var. elatior L.B. Sm., Phytologia 5: 181 (1955).—Type: Fagerlind & Wibom 1947: S (holo). Josemania umbellata (André) Barfuss & W. Till, nom. prov. ≡ Tillandsia umbellata André, Rev. Hort. 58: 60, pl. (1886).—Type: André K-317: K (holo). = Tillandsia lindenii [E. Morren] var. regeliana E. Morren, Belgique Hort. 20: 225, pl. 12 (1870) ≡ Phytarrhiza lindenii var. regeliana (E. Morren) E. Morren, Belgique Hort. 29: 298 (1879).—Type: Wallis s.n.: LG? (holo). = Tillandsia lindenii [E. Morren] var. major Dombrain, Floral Mag. 10: pl. 529 (1871) ≡ Tillandsia lindenii [E. Morren] var. intermedia E. Morren, Rev. Hort. 50: 390 (1878), nom. nov. superfl. ≡ Phytarrhiza lindenii var. intermedia E. Morren, Belgique Hort. 29: 298 (1879), nom. illeg.—Type: Veitch Hort. s.n.: ? (holo), in absence: Original illustration. 172 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 ?= Vriesea violacea hort. ex Houllet, Rev. Hort. 44: 230 (1872), nom. nud. = Tillandsia lindenii [E. Morren] var. rutilans Linden ex Houllet, Rev. Hort. 44: 230 (1872), nom. nud. = Tillandsia lindenii [E. Morren] var. tricolor André, Ill. Hort. 24: 190 (1877) ≡ Tillandsia cyanea var. tricolor (André) L.B. Sm., Contr. U. S. Natl. Herb. 29: 491 (1951).—Type: André 4040: K (holo). = Tillandsia lindenii [E. Morren var.] splendida Carrière, Rev. Hort. 54: 12, pl. (1882).— Type: Thibaut & Keteleer Hort s.n.: ? (holo). ?= Tillandsia lindenii [E. Morren] var. violacea hort. ex André, Rev. Hort. 58: 61 (1886), nom. nud. = Tillandsia lindenii [E. Morren] vera major Duval, Gartenwelt 5: 164, Fig. (1901), nom. illeg.—Type: Original illustration. 5. Lemeltonia Barfuss & W. Till, nom. prov.—Type: Tillandsia dodsonii L.B. Sm. A genere Tillandsia L. s.str. filamentis basi connatis, stigmate coralliforme et ovulis obtusis claviformibusque, seminibus cum endostomio breviter cylindrico et embryone cum radicula plusminusve distincta differt. Petala alba vel rariter flavescentia. – Tillandsia subg. Phytarrhiza (Vis.) Baker p.p., typo excluso. Plants acaulescent or rarely caulescent; leaves very narrowly triangular, not forming a tank rosette; inflorescence green or brown, lax with the often fragrant flowers mostly spreading; petals white or rarely yellowish; stamens and style deeply included within the corolla, filaments connate among themselves at least at the base; stigma coralliform (Brown & Gilmartin 1989b); ovules clavate, obtuse; seeds with a short-cylindric endostome (type d, Groß 1988) and embryo with the radicule ± separated (type b, Groß 1988). Named after Elton Martinez Carvalho Leme (1960–) from Rio de Janeiro, Brazil, leading authority of Brazilian Bromeliaceae. Lemeltonia acosta-solisii (Gilmartin) Barfuss & W. Till, nom. prov. ≡ Tillandsia acosta-solisii Gilmartin, Phytologia 16: 160 (1968).—Type: Teuscher 2275-56: US (holo). Lemeltonia cornuta (Mez & Sodiro) Barfuss & W. Till, nom. prov. ≡ Tillandsia cornuta Mez & Sodiro, Bull. Herb. Boissier, sér. 2, 5: 106 (1905).—Type: Sodiro 171/42: B (holo). Lemeltonia dodsonii (L.B. Sm.) Barfuss & W. Till, nom. prov. ≡ Tillandsia dodsonii L.B. Sm., Phytologia 28: 32, pl. 2, figs. f, g (1974).—Type: Dodson 5225: US (holo). Lemeltonia monadelpha (E. Morren) Barfuss & W. Till, nom. prov. ≡ Phytarrhiza monadelpha E. Morren, Belgique Hort. 32: 168, pl. 7 (1882) ≡ Tillandsia monadelpha (E. Morren) Baker, J. Bot. 25: 281 (1887).—Type: Linden Hort. s.n.: LG (holo). = Tillandsia graminifolia Baker, J. Bot. 25: 281 (1887).—Type: Martin s.n.: BM (lecto); Poiteau s.n.: P? (syn); Sagot 859: P (syn); Fendler 828: K (syn); Parker s.n.: K (syn). = Catopsis (Andrea) alba E. Morren ex Baker, Handb. Bromel.: 192 (1889), nom. nud. = Tillandsia monobotrya Mez, Repert. Spec. Nov. Regni Veg. 16: 77 (1919).—Type: Wercklé 150 = Inst. Costar. 17444: B (holo). Lemeltonia narthecioides (C. Presl) Barfuss & W. Till, nom. prov. ≡ Tillandsia narthecioides C. Presl, Reliq. haenk. 1: 125 (1827).—Type: Haenke s.n.: PR (holo). 173 BARFUSS, M.H.J. NEW CLASSIFICATION Lemeltonia scaligera (Mez & Sodiro) Barfuss & W. Till, nom. prov. ≡ Tillandsia scaligera Mez & Sodiro, Bull. Herb. Boissier, sér. 2, 5: 107 (1905).—Type: Sodiro 171/4: B (holo). Lemeltonia triglochinoides (C. Presl) Barfuss & W. Till, nom. prov. ≡ Tillandsia triglochinoides C. Presl, Reliq. haenk. 1: 125 (1827).—Type: Haenke s.n.: PR (holo). = Tillandsia hartwegiana Brongn. ex Baker, Handb. Bromel.: 171 (1889), nom. nud. 6. Racinaea M.A. Spencer & L.B. Sm., Phytologia 74: 152 (1993).—Type: Tillandsia cuspidata L.B. Sm. 6.1. Racinaea subg. Pseudophytarrhiza Barfuss & W. Till, nom. prov.—Type: Tillandsia venusta Mez & Wercklé. A subgenere typica petalis latioribus unguiculatibusque et stigmate coralliforme differt. Petals with broader blade and distinct claw; filaments free; stigma coralliform. – Tillandsia subg. Phytarrhiza (Vis.) Baker p.p., typo excluso. Racinaea hamaleana (E. Morren) Barfuss & W. Till, nom. prov. ≡ Tillandsia hamaleana E. Morren, Gard. Chron. "1869"(2): 460 (1869) ≡ Wallisia hamaleana (E. Morren) E. Morren, Belgique Hort. 20: 97, pl. 5 (1870), nom. illeg. ≡ Phytarrhiza hamaleana (E. Morren) E. Morren, Belgique Hort. 29: 297 (1879).—Type: Wallis s.n. in Hort. E. Morren: LG? (holo). = Tillandsia commelyna E. Morren, Belgique Hort. 20: 97 (1870), nom. nud. = Tillandsia platypetala Baker, J. Bot. 26: 46 (1888).—Type: Hartweg s.n.: K (holo). = Tillandsia nubis Gilmartin, Phytologia 16: 161 (1968).—Type: Naundorff s.n.: US (holo). Racinaea venusta (Mez & Wercklé) Barfuss & W. Till, nom. prov. ≡ Tillandsia venusta Mez & Wercklé, Bull. Herb. Boissier, sér. 2, 5: 108 (1905).—Type: Wercklé Bromel. Costar. 95: B (holo). 6.2. Racinaea subg. Racinaea. = Tillandsia subg. Pseudo-Catopsis Baker, Handb. Bromel.: 157, 192 (Aug–Oct 1889).— Type: Tillandsia parviflora Ruiz & Pav.! (Baker 1889: 192). = Tillandsia sect. Pseudo-Catopsis André, Bromel. Andr.: 62, 66 (Sept–Dec 1889), nom. illeg. (Art. 53.4).—Lectotype (Smith & Downs, 1977: 670): Tillandsia ropalocarpa André. – Tillandsia subg. Phytarrhiza (Vis.) Baker p.p., typo excluso. Mesophytic, usually acaulescent; petals free and unappendaged; filaments and anthers free; pollen aperture diffuse or with insulae; stigma simple-erect or slightly conduplicate-spiral, rarely coralliform. Racinaea dyeriana (André) Barfuss & W. Till, nom. prov. ≡ Tillandsia dyeriana André, Énum. Bromél. 1888: 8 (13 Dec 1888).—Type: André 4256: K (holo), NY (iso). = Tillandsia rutschmannii Rauh, Trop. Subtrop. Pflanzenwelt 12: 5, figs. 1–3, [4a], 5 (1974).—Type: Naundorff s.n. = Hort. Bot. Heidelberg 31701: HEID (holo), US (iso). 7. Rothowia Barfuss & W. Till, nom. prov.—Type: Tillandsia wagneriana L.B. Sm. A genere Tillandsia L. s.str. stigmate obconico typo laminis-convolutis simili, ovulis obtusis vel subobtusis anguste subcylindrico-claviformibus, seminibus cum endostomio conico vel cylindrico et embryone cum radicula plusminusve distincta differt. Petala caerulea. – Tillandsia subg. Phytarrhiza (Vis.) Baker p.p., typo excluso. 174 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 Plants acaulescent; leaves lingulate, acute, with a distinct leaf-sheath, forming a tank-rosette; inflorescence pink, once to twice branched, rhachis ± alate, floral bracts glabrous; petals blue; stamens and style deeply included within the corolla; filaments lingulate, narrowed at the apex, anthers basifixed; stigma obconic, resembling the convolute-blade type; ovules slenderly cylindric, obtuse or subobtuse; seeds with a conical or cylindrical endostome (type e, Groß 1988) and embryo with a ± distinct radicule (type b, Groß 1988). Named after Ronald Thomas Wagner (1939–1962). Rothowia laxissima (Mez) Barfuss & W. Till, nom. prov. ≡ Tillandsia laxissima Mez, Bull. Herb. Boissier, sér. 2, 5: 108 (1905).—Type: Bang 2301 p.p.: B (holo), G, GH, LE, M, MO, NY, PH, US, W, WU (iso). Rothowia laxissima var. moorei (H. Luther) Barfuss & W. Till, nom. prov. ≡ Tillandsia laxissima var. moorei H. Luther, Selbyana 20(1): 13 (1999).—Type: Moore s.n.: SEL (holo). Rothowia platyrhachis (Mez) Barfuss & W. Till, nom. prov. ≡ Tillandsia platyrhachis Mez, in C.DC., Monogr. phan. 9: 848 (1896).—Type: Kalbreyer 1328: K (holo). = Tillandsia platyrhachis var. alba Rauh & Hirtz, Trop. Subtrop. Pflanzenwelt 53: 11, figs. 5–7 (1985).—Type: Rauh & Hirtz 37571: HEID (holo). = Tillandsia platyrhachis var. magnifica Rauh & von Bismarck, Trop. Subtrop. Pflanzenwelt 53: 6, figs. 2–5 (1985).—Type: Rauh & von Bismarck 66107: HEID (holo). Rothowia wagneriana (L.B. Sm.) Barfuss & W. Till, nom. prov. ≡ Tillandsia wagneriana L.B. Sm., Phytologia 9: 254, pl. 4, figs. 1–3 (1963).—Type: Moore 310: US (holo). 8. Tillandsia L., Sp. pl.: 286 (1753).—Type: Tillandsia utriculata L. Meso- to xerophytic, acaulescent to caulescent; petals free, usually not appendaged; filaments and anthers free; pollen aperture diffuse, with insulae or operculum, or of the Alcantarea type; stigma simple-erect or conduplicate-spiral, rarely of the convolute-blade type, seeds of various endostome types (Groß, 1988: e–n) and embryo types (Groß, 1988: b–g). 8.1. Tillandsia subg. Anoplophytum (Beer) Baker, J. Bot. 25: 212 (1887) ≡ Anoplophytum Beer, Flora 37: 346 (1854) ≡ Tillandsia sect. Anoplophytum (Beer) Griseb., Fl. Brit. W. I.: 597 (1864).—Type: Tillandsia stricta Sol. ex Sims. = Tillandsia sect. Eriophyllum K. Koch, Index Seminum Hort. Bot. Berol. "1873", App. 4: 1 (1874), as “Eriophorum”.—Lectotype (proposed): Tillandsia selloa K. Koch. (= Tillandsia linearis Vell.). 8.2. Tillandsia subg. Diaphoranthema (Beer) Baker, J. Bot. 16: 236 (1878) ≡ Diaphoranthema Beer, Flora 37: 349 (1854) ≡ Tillandsia sect. Diaphoranthema (Beer) K. Koch, Index Seminum Hort. Bot. Berol. "1873", App. 4: 1 (1874).—Type: Renealmia recurvata L. = Dendropogon Raf., Neogenyton: 3 (1825) ≡ Tillandsia sect. Strepsia Nutt., Gen. N. Amer. pl. 1: 208 (1818) ≡ Strepsia (Nutt.) Steud., Nomencl. Bot. ed. 2.2: 645 (1841), nom. illeg.— Type: Renealmia usneoides L. 8.3. Tillandsia subg. Phytarrhiza (Vis.) Baker, J. Bot. 25: 212, 214 (1887) ≡ Phytarrhiza Vis., Mem. R. Ist. Venet. Sci. 5: 340, pl. (1855) ≡ Tillandsia sect. Phytarrhiza (Vis.) K. Koch, Index Seminum Hort. Bot. Berol. "1873", App. 4: 1 (1874).—Type: Tillandsia duratii Vis. 175 BARFUSS, M.H.J. NEW CLASSIFICATION 8.4. Tillandsia subg. Pseudalcantarea Mez, in Engl., Pflanzenr. 4.32: 437, 455 (1935).—Type: Platystachys viridiflora Beer. 8.5. Tillandsia subg. Pseudovriesea Barfuss & W. Till, nom. prov.—Type: Tillandsia tequendamae André. Folia anguste triangularia, xeromorphica, dense lepidota, petala plerumque biappendiculata, saepe bicoloria, rariter marginibus crenulatis, stigma laminis conduplicato-spiralibus. = Tillandsia sect. Trianisandra André, Bromel. Andr.: 63 (1889).—Type: Tillandsia heterandra André. − Tillandsia subg. Allardtia A. Dietr. p.p., typo excluso.—studied taxa: T. myriantha Baker, T. barthlottii Rauh. − Tillandsia subg. Tillandsia p.p., typo excluso.—studied taxa: T. spathacea Mez & Sodiro. − Vriesea Lindl. p.p., typo excluso.—studied taxa: xerophytic, grey-leaved Vriesea spp. (see Appendix for Figure 2 and supplementary data for all spp. included). Leaves narrowly triangular, xeromorphic, densely lepidote, petals usually with two basal appendages, often two-colored, sometimes with crenulated margins, stigma usually conduplicate-spiral, rarely conduplicate-erect, pollen with diffuse sulcus area; seeds with an elongatecylindric endostome (type f, Groß, 1988), embryo of type (b–) or ± f (Groß, 1988). 8.6. Tillandsia subg. Tillandsia ≡ Caraguata Plum. ex Adans., Fam. pl. 2: 67, 532 (Jul–Aug 1763), nom. illeg. (Art. 52.1).—Type: Tillandsia utriculata L. = Renealmia L., Sp. pl.: 286 (1753), nom. rejic. vs. Renealmia L. f. (1782).—Type: Renealmia paniculata L. = Bonapartea Ruiz & Pav., Fl. peruv. 3: 38, pl. 262 (1802) ≡ Misandra F. Dietr., Nachtr. vollst. Lex. Gärtn. 5: 102 (1819) non Comm. ex Juss. (1789: Haloragaceae), nom. illeg. ≡ Acanthospora Spreng., Syst. veg. 2: 25 (1825), nom. illeg. ≡ Buonapartea Sweet, Hort. brit. ed. 3: 706 (1839), orth. var.—Type: Bonapartea juncea Ruiz & Pav. = Allardtia A. Dietr., Berliner Allg. Gartenzeitung 20: 241 (1852) ≡ Platystachys K. Koch, Index Seminum Hort. Bot. Berol. "1854", App.: 11 (1855), nom. illeg. ≡ Tillandsia sect. Allardtia (A. Dietr.) E. Morren, Belgique Hort. 27: 272 (1877) ≡ Tillandsia subg. Allardtia (A. Dietr.) Baker, J. Bot. 26: 40 (1888).—Type: Allardtia cyanea A. Dietr. (≡ Tillandsia guatemalensis L.B. Sm.). = Platystachys Beer, Fam. Bromel.: 18, 80 (1856), nom. illeg. (Art. 53, Note 1) ≡ Tillandsia sect. Platystachys (K. Koch) Benth. & Hook. f., Gen. pl. 3(2): 670 (1883) ≡ Tillandsia subg. Platystachys (K. Koch) Baker, J. Bot. 25: 212, 236 (1887).—Lectotype (Smith & Downs, 1977: 668): Tillandsia setacea Sw. = Pityrophyllum Beer, Fam. Bromel.: 17, 79 (1856) ≡ Tillandsia sect. Pityrophyllum (Beer) K. Koch, Index Seminum Hort. Bot. Berol. "1873", App. 4: 1 (1874) ≡ Tillandsia subg. Pityrophyllum (Beer) Baker, J. Bot. 26: 39 (1888).—Type: Tillandsia ionantha Planch. = Vriesea sect. Cylindrostachys Wittm., in Engl. & Prantl, Nat. Pflanzenfam. II.4: 59 (1888) ≡ Vriesea subg. Cylindrostachys (Wittm.) Harms, in Engl., Nat. Pflanzenfam. ed. 2, 15a: 124 (1930).—Type: Vriesea malzinei E. Morren. − Tillandsia subg. Pseudalcantarea Mez p.p., typo excluso.—studied taxa: T. paniculata (L.) L. − Vriesea sect. Xiphion (E. Morren) Wawra ex Wittm. p.p., typo excluso.—studied taxa: V. malzinei E. Morren. 176 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 8.7. Tillandsia subg. Viridantha (Espejo) Barfuss & W. Till, nom. prov. ≡ Viridantha Espejo, Acta Bot. Mex. 60: 27 (2002).—Type: Tillandsia plumosa Baker. = Viridantha sect. Caulescens Espejo, Acta Bot. Mex. 60: 31 (2002).—Type: Tillandsia tortilis Klotzsch ex Baker. − Tillandsia subg. Allardtia A. Dietr. p.p., typo excluso.—studied taxa: Tillandsia plumosa complex, Tillandsia tectorum complex (see Appendix for Figure 2 and supplementary data for all spp. included). − Tillandsia subg. Anoplophytum p.p., typo excluso.—studied taxa: T. heteromorpha Mez. − Tillandsia subg. Tillandsia p.p., typo excluso.—studied taxa: T. lepidosepala L.B Sm. 8.8. Biflora clade. Name giving taxon: Tillandsia biflora Ruiz & Pav. − Tillandsia subg. Allardtia A. Dietr. p.p. majore, typo excluso (see Appendix for Figure 2 and supplementary data for all spp. included). − Tillandsia subg. Pseudalcantarea Mez p.p., typo excluso.—studied taxa: T. baliophylla Harms. – Tillandsia subg. Anoplophytum (Beer) Baker p.p., typo excluso.—studied taxa: T. macbrideana L.B. Sm., T. pseudomacbrideana Rauh. − Tillandsia L. subg. Tillandsia p.p., typo excluso.—studied taxa: T. deppeana Steud., T. imperialis E. Morren ex Roezl, T. multicaulis Steud. 8.9. Gardneri clade. Name giving taxon: Tillandsia gardneri Lindl. – Tillandsia subg. Anoplophytum (Beer) Baker p.p., typo excluso.—studied taxa: T. brachyphylla Baker, T. gardneri Lindl., T. globosa Wawra. 8.10. Purpurea clade. Name giving taxon: Tillandsia purpurea Ruiz & Pav. – Tillandsia subg. Phytarrhiza (Vis.) Baker p.p., typo excluso.—studied taxa: T. aurea Mez, T. cacticola L.B. Sm., T. purpurea Ruiz & Pav., T. straminea Kunth. 8.11. Rauhii clade. Name giving taxon: Tillandsia rauhii L.B. Sm. − Tillandsia L. subg. Tillandsia p.p., typo excluso.—studied taxa: T. ecarinata L.B. Sm., T. ferreyrae L.B. Sm., T. rauhii L.B. Sm., T. teres L.B. Sm.. 8.12. Xiphioides clade. Name giving taxon: Tillandsia xiphioides Ker Gawl. – Tillandsia subg. Anoplophytum (Beer) Baker p.p., typo excluso.—studied taxa: xerophytic, Andean spp. (see Appendix for Figure 2 and supplementary data for all spp. included). − Tillandsia subg. Allardtia A. Dietr. p.p., typo excluso.—studied taxa: xerophytic, Andean spp. (see Appendix for Figure 2 and supplementary data for all spp. included). 8.13. Incertae sedis. The following taxa are currently unclassified within Tillandsia L.: T. albertiana Verv., T. australis Mez, T. disticha Kunth, T. edithae Rauh, T. esseriana Rauh & L.B. Sm., T. nana Baker, T. pseudomicans Rauh, T. sphaerocephala Baker. 177 BARFUSS, M.H.J. NEW CLASSIFICATION Tribe Vrieseeae W. Till & Barfuss, Amer. J. Bot. 92: 348 (2005).—Type: Vriesea Lindl. Ovary for a small part inferior, capsule septicidal; petal appendages usually present; stigma mainly of the convolute-blade and cupulate types, less often of the tubolaciniate, conduplicate-erect, conduplicate-patent, conduplicate-spiral, or simple-erect types; seeds with a flight apparatus of pseudohairs at the micropylar end; pollen mainly with exine fragments at the aperture (insulae type) or with solid aperture margins (Alcantarea type), less often with a diffuse aperture. Subtribe Cipuropsidinae Barfuss & W. Till, nom. prov.—Type: Cipuropsis Ule. A subtribu typica ovulis obtusis vel subobtusis et stigmatibus simplicibus-erectis, condplicatospialibus vel cupulatis papillosisque differt. Ovules obtuse or subobtuse, stigma of the simple-erect, conduplicate-spiral, or cupulate types with papillae (Brown & Gilmartin 1989b). 9. Cipuropsis Ule, Verh. Bot. Vereins Prov. Brandenburg 48: 148 (1907).—Type: Cipuropsis subandina Ule. – Tillandsia subg. Allardtia A. Dietr. p.p., typo excluso.—studied taxa: T. amicorum I. Ramírez & Bevil. − Vriesea sect. Xiphion (E. Morren) Wawra ex Wittm. p.p., typo excluso.—studied taxa: V. zamorensis (L.B. Sm.) L.B. Sm. Petals connate for 25% into a common tube, filaments connate for 5 mm with the petals like the petals themselves, petal appendages present; seeds with an endostome of type c (Groß, 1988) and embryo of type a (Groß, 1988). Cipuropsis amicorum (I. Ramírez & Bevil.) Barfuss & W. Till, nom. prov. ≡ Tillandsia amicorum I. Ramírez & Bevil., Acta Bot. Venez. 15: 149, pl. 1 (1989).—Type: Rutkis 452: VEN (holo); Steyermark, Bu[n]ting & Wessels-Boer 100258: MO, VEN (para); Carreño s.n.: VEN (para); Steyermark & Wessels-Boer 100448: MO, VEN (para). Cipuropsis zamorensis (L.B. Sm.) Barfuss & W. Till, nom. prov. ≡ Tillandsia zamorensis L.B. Sm., Phytologia 4: 213, pl. 1, figs. 3–5 (1953) ≡ Vriesea zamorensis (L.B. Sm.) L.B. Sm., Phytologia 20: 174 (1970).—Type: Scolnik 1500: NY (holo). 10. Mezobromelia L.B. Sm., Proc. Amer. Acad. Arts 70: 151 (1935).—Type: Mezobromelia bicolor L.B. Sm. – Tillandsia subg. Allardtia A. Dietr. p.p., typo excluso.—studied taxa: ?T. buseri Mez, T. schimperiana Wittm. – Vriesea sect. Xiphion (E. Morren) Wawra ex Wittm. p.p., typo excluso.—studied taxa: V. rubrobracteata Rauh. Mesophytic, acaulescent; flower arrangement distichous or polystichous; petals connate/conglutinate and appendaged; filaments free, anthers connate; pollen with diffuse aperture or inaperturate; stigma of the simple-erect type. 178 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 Mezobromelia schimperiana (Wittm.) Barfuss & Manzan, nom. prov. ≡ Tillandsia schimperiana Wittm., Bot. Jahrb. Syst. 11: 67 (1889).—Type: Lehmann XXVI: G (holo). = Vriesea rubrobracteata Rauh, Trop. Subtrop. Pflanzenwelt 27: 13, Fig. 5 (1979).—Type: Rauh 37412: HEID (holo). ?= Tillandsia buseri Mez, Bull. Herb. Boissier, sér. 2, 3: 145 (1903).—Type: Langlassé 102a: G (holo), B, K (iso). 11. Werauhia J.R. Grant, Trop. Subtrop. Pflanzenwelt 91: 28 (1995) ≡ Tillandsia sect. Xiphion E. Morren, Belgique Hort. 24: 292 (1874), Lectotype (Smith & Downs, 1977: 1069) ≡ Vriesea sect. Xiphion (E. Morren) Wawra ex Wittm., in Engl. & Prantl, Nat. Pflanzenfam. II.4: 59 (1888).—Type: Tillandsia gladioliflora H. Wendl. = Werauhia sect. Jutleya J.R. Grant, Trop. Subtrop. Pflanzenwelt 91: 28, 39.—Type: Thecophyllum pedicellatum Mez & Wercklé. – Tillandsia subg. Allardtia A. Dietr. p.p., typo excluso.—studied taxa: T. insignis Mez. Mesophytic, usually acaulescent; petals free, mostly appendaged; filaments and anthers free; pollen of the insulae type; stigma cupulate without papillae; seeds with an endostome of type d (Groß, 1988) and embryo of type a, rarely b (Groß, 1988). 12. Chrysostachys clade. Name giving taxon: Vriesea chrystostachys E. Morren. – Vriesea sect. Xiphion (E. Morren) Wawra ex Wittm. p.p., typo excluso.—studied taxa: V. chrystostachys E. Morren, V. ospinae H. Luther. 13. Singularis clade. Name giving taxon: Tillandsia singularis Mez & Wercklé. – Tillandsia subg. Allardtia p.p., typo excluso.—studied taxa: T. asplundii L.B. Sm., T. singularis Mez & Wercklé. 14. Splendens clade. Name giving taxon: Vriesea splendens (Brongn.) Lem. – Vriesea sect. Vriesea p.p., typo excluso.—studied taxa: V. splendens (Brongn.) Lem.. – Vriesea sect. Xiphion (E. Morren) Wawra ex Wittm. p.p., typo excluso.—studied taxa: V. glutinosa Lindl. 15. Incertae sedis. The following taxon is currently unclassified within the Cipuropsis-Mezobromelia clade: Vriesea tuerckheimii (Mez) L.B. Sm. Subtribe Vrieseinae Barfuss & W. Till, nom. prov.—Type: Vriesea Lindl. Ovula distincte appendiculata in chalaza, stigmata laminis convolutis, conduplicatis vel tubolaciniatis. Ovules distinctly appendaged at the chalaza, stigma of the convolute-blade (Brown & Gilmartin 1989b) or tubolaciniate type (Leme & Brown, 2004: Vriesea plurifolia) in Vriesea, and of the conduplicate-patent (Versieux & Wanderley, 2007; Leme 2007) or rarely of the conduplicateerect (Leme, 2009: Alcanatrea roberto-kautskyi; Leme, unpubl. data: Alcantarea cerosa) in Alcantarea. 179 BARFUSS, M.H.J. NEW CLASSIFICATION 16. Alcantarea (E. Morren ex Mez) Harms, Notizbl. Bot. Gart. Berlin-Dahlem 10: 802 (1929) ≡ Vriesea subg. Alcantarea E. Morren ex Mez, in Mart., Eichler & Urban, Fl. bras. 3(3): 516 (1894), non Alcantara Glaz. (1909: Asteraceae), nom nud., nec Alcantara Glaz. ex Baroso (1969: Asteraceae), Lectotype (Grant & Zijlstra, 1998: 93) ≡ Vriesea Gruppe „Reginae" Wawra, Itin. Princip. S. Coburgi 1: 158 (1883) ≡ Vriesea sect. Reginae (Wawra) Wittm., in Engl. & Prantl, Nat. Pflanzenfam. II.4: 58 (1888).—Type: Tillandsia regina Vell. = Tillandsia sect. Macrocyathus K. Koch, Index Seminum Hort. Bot. Berol. „1873",App. 4: 1, 6 (1874).—Type: Tillandsia gigantea sensu K. Koch (1874) (= Vriesea glazioviana Lem.). Mesophytic, acaulescent; petals free and appendaged; filaments and anthers free; pollen of the Alcantarea type; stigma conduplicate-patent, rarely conduplicate-erect. 17. Vriesea Lindl., Edwards’s Bot. Reg. 29: sub pl. 10 (1843), as "Vriesia", nom. conserv. vs. Hexalepis Raf., Fl. tellur. 4: 24 (1838) as well as vs. Vriesea Hassk., Flora 25 (Beibl.): 27 (1842: Scrophulariaceae) ≡ Hexalepis Raf., Fl. tellur. 4: 24 (1838), nom. rejic. vs. Vriesea Lindl. ≡ Tillandsia sect. Vriesea (Lindl.) Griseb., Nachr. Königl. Ges. Wiss. Georg-AugustUniv. [1]: 17 (13 Jan 1864) ≡ Vriesea ?sect. Psittacinae Wawra, Itin. princ. S. Coburgi 1: 158 (1883), nom. illeg. ≡ Tillandsia subg. Vriesea (Lindl.) Baker, J. Bot. 26: 47 (1888) ≡ Vriesea subg. Euvriesea Mez, in Mart., Eichler & Urban, Fl. bras. 3(3): 513 (1894) ≡ Vriesea sect. Genuinae Mez, in Mart., Eichler & Urban, Fl. bras. 3(3): 513 (1894) ≡ Neovriesia Britton ex Britton & P. Wilson, Bot. Porto Rico 5: 141 (1923), nom. illeg.—Type: Tillandsia psittacina Hook. = Tillandsia sect. Synandra K. Koch, Index Seminum Hort. Bot. Berol. "1873", App. 4: 1 (1874).—Type: Vriesea corallina Regel. = Tillandsia sect. Cyathophora K. Koch, Index Seminum Hort. Bot. Berol. "1873", App. 4: 1 (1874).—Type: Encholirion saundersii Carrière. = Vriesea sect. Platzmanniae E. Morren, Belgique Hort. 25: 349 (1875).—Type: Vriesea platzmannii E. Morren. = Vriesea sect. Brachystachyae Wawra, Itin. princ. S. Coburgi 1: 158 (1883).—Type: Vriesea carinata Wawra. = Vriesea sect. Macrostachyae Wawra, Itin. princ. S. Coburgi 1: 158 (1883).—Type: Vriesea conferta Gaudich. – Vriesea subg. Conostachys (Griseb.) Mez, in Mart., Eichler & Urban, Fl. bras. 3(3): 516 (1894), typo excluso.—studied taxa: V. poenulata (Baker) E. Morren ex Mez, V. rubida E. Morren ex Mez (= V. ventricosa (Wawra) Mez). – Vriesea sect. Xiphion (E. Morren) Wawra ex Wittm. p.p. majore, typo excluso (see Appendix for Figure 2 and supplementary data for all spp. included). Mesophytic to semi-mesophytic, usually acaulescent; petals free and appendaged; filaments and anthers free; pollen of the insulae type; stigma of the convolute-blade type or tubolaciniate. 180 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 Key to the genera of Tillandsioideae This key must remain provisional as the resolution in the molecular trees is not sufficient in some groups and morphological features are lacking for numerous taxa. However, it demonstrates that the molecular phylogeny can be supported in most cases using ovary, seed, stigma, anther, and petal morphology. The Cipuropsis-Mezobromelia clade and Tillandsia continue to be critical groups, either because of missing taxa, low phylogenetic resolution, or uncertain application of morphological characters. Therefore, informal clades of the CipuropsisMezobromelia clade and subgenera and informal clades of Tillandsia are usually not apportioned, except for cases, where differentiating characters have to be utilized for the discrimination of a genus from a subgeneric unit, informal clade or species. 1 1* 2 2* 3 3* 4 4* Ovary ½–⅔ inferior.—Seeds long appendaged on both ends, chalazal appendage very long, filiform, undivided, micropylar appendage plumose. Flowers polystichous. Pollen with diffuse aperture. Stigma lobes with undulate margins (convolute-blade type).— Lesser Antilles extending to northeastern Venezuela............................... Glomeropitcairnia Ovary max. ⅓ inferior to completely superior.—Seeds usually long appendaged only on one end, but chalazal appendage sometimes well developed (e.g., Alcantarea, Tillandsia p.p., Vriesea p.p., Vriesea tuerckheimii). Flowers polystichous or distichous .......................2 Seeds with an undivided micropylar appendage, chalazal multicellular hairs folded at maturity and forming the flight apparatus.—Ovary max. ⅛ inferior to completely superior. Pollen with sharp cut aperture margins (Catopsis type). Stigma with erect lobes (simpleerect type), rarely somewhat twisted (tending towards conduplicate-spiral type). Petals usually white and forming a campanulate corolla, rarely yellow and spreading.—Central America and Antilles extending to northern South America and Southeastern Brazil ............ .................................................................................................................................... Catopsis Seeds with a divided micropylar appendage forming a coma (flight apparatus), chalazal appendage undivided (very rarely divided) and short, or lacking, rarely long and occasionally somewhat divided (Alcantarea), not folded at maturity .................................................3 Micropylar appendage rather short, about equaling the seed proper, chalazal appendage distinctly larger than the seed proper, sometimes somewhat divided. Petals long, recurved, exposing the long exserted stamens and style.—Pollen with solid aperture margins. Stigma lobes conduplicate and spreading (conduplicate-patent type), rarely not spreading (conduplicate-erect type: Alcantarea cerosa, Alcantarea roberto-kautskyi), often somewhat twisted around their longitudinal axis.—Endemic to Southeastern Brazil ...... ............................................................................................................................... Alcantarea Micropylar appendage distinctly longer than the seed proper, chalazal appendage lacking to about half as long as the seed proper, nearly always undivided. Petals forming a tubular or campanulate corolla, at least in the proximal half ............................................................4 Petals conglutinated/connate into a distinct tube (distinctly longer than 25%).—Petals white, yellow, or green. Seeds without a distinct chalazal appendage .................................5 Petals not conglutinated/connate into a distinct tube, completely or at least ¾ free from each other but sometimes connate at the base (max. 25 % of entire length).—Petals violet, pink, red, orange, yellow, green, white, and rarely bicolored. Seeds sometimes with a distinct chalazal appendage up to the length of the seed proper, rarely longer ...................6 181 BARFUSS, M.H.J. 5 5* 6 6* 7 7* 8 8* 9 9* 10 10* 182 NEW CLASSIFICATION Petals with basal appendages and anthers forming a tube around the stigma.—Stigma with erect lobes (simple-erect type). Pollen with a diffuse aperture or nearly inaperturate.—Northern Andes extending to the Antilles, the Guianas, and Bolivia .... Mezobromelia Petals without basal appendages or anthers not forming a tube around the stigma.— Stigma with erect lobes (simple-erect type) or infundibuliform with undulate margins (convolute-blade type). Pollen with a diffuse aperture or inaperturate.—Andean and Central America, extending to the Antilles, the Guianas, and Eastern Brazil ................Guzmania Stigma lobes fused and forming a cup, margins entire (cupulate type) and without papillae.—Plants mostly forming impounding rosettes. Corolla tubular or campanulate, sometimes zygomorph (stamens then asymmetrically arranged. Pollen with irregular aperture margins and exine fragments in the aperture (insulae type). Seeds with no or only very short chalazal appendage.—Central America extending to Ecuador (and Bolivia) . Werauhia Stigma lobes not of the cupulate type without papillae, but rarely resembling cupulate with crenulate (tubolaciniate type) (Vriesea p.p.) or papillate (Vriesea tuerckheimii) margins ......................................................................................................................................... 7 Stigma with undulate and papillate margins (convolute-blade type), rarely with crenulate margins (tubolaciniate type), plants meso- or semi-mesophytic, petal appendages present.—Corolla tubular or campanulate, stamens and style exserted or included. Petals yellow (often with green tips), cream, or brownish (–red), rarely white, free or rarely short connate into a common tube with the filament bases. Pollen with irregular aperture margins and exine fragments in the aperture (insulae type). Seeds with a distinct chalazal appendage.—Eastern Brazil extending to Venezuela and Peru ...................................... Vriesea Stigma lobes usually not of the convolute-blade or tubolaciniate types, if rarely resembling convolute-blade then plants xerophytic and petal appendages absent ...................... 8 Filaments connate among each other at least at the base but sometimes for nearly their whole length, free from the petals. Stigma coralliform.—Leaves narrowly triangular. Petals white ore yellowish. Ovules clavate, obtuse. Seeds without chalazal appendage.—Central America to Peru extending to eastern Venezuela and the Guianas ................... Lemeltonia Filaments free from each other, but sometimes connate/agglutinated to the petals (e.g., Cipuropsis and Racinaea). Stigma usually not of the coralliform type, if rarely resembling the coralliform type then filaments free from each other (Racinaea subg. Pseudophytarrhiza and Racinaea dyeriana) ............................................................................................ 9 Stigma pinnatisect.—Inflorescence undivided, distichous, flat, lanceolate-elliptic in outline. Petals violet, their blades usually strongly enlarged. Leaf bases mostly longitudinally red striped. Ovules slenderly cylindric, obtuse. Seeds without chalazal appendage.— Ecuador and Peru, Josemania anceps extending to Guatemala, eastern Venezuela, the Guianas, and northern Brazil .................................................................................. Josemania Stigma not pinnatisect ......................................................................................................... 10 Stigma obconical and resembling the convolute-blade type, mesophytic.—Leaves lingulate, forming impounding rosettes. Inflorescence compound, pink, rhachis usually alate. Flowers distichous, petals with enlarged blade, violet. Seeds with a minute chalazal appendage only.—Andean, Ecuador to Bolivia ........................................................... Rothowia Stigma not obconical and not resembling the convolute-blade type, if rarely of the convolute-blade type, then plants xerophytic and not forming impounding rosettes (Tillandsia plumosa complex)................................................................................................................ 11 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 11 Sepals distinctly asymmetric, stigma simple-erect or slightly conduplicate-spiral (both reduction forms of the coralliform type), if subsymmetric then stigma resembling the coralliform type and leaves lingulate (Racinaea subg. Pseudophytarrhiza and Racinaea dyeriana). Seeds without chalazal appendage.—Flowers small and inconspicuous in most species, petals white to yellow, rarely blue (Racinaea hamaleana).—Andean, mainly Ecuador, extending to Central America, Bolivia, the Greater Antilles and the Guianas, and Southeastern Brazil .................................................................................................. Racinaea 11* Sepals symmetric, stigma usually simple-erect or conduplicate-spiral, rarely convolutebladed (Tillandsia subg. Viridantha p.p.) or resembling cupulate with papillate margins (Vriesea tuerckheimii). Seeds usually with a distinct chalazal appendage.......................... 12 12 Petals without basal appendages ........................................................................................ 13 12* Petals with basal appendages ............................................................................................. 14 13 Petals and filaments basally connate or leaf sheaths drying silver-gray.—Petals white or yellow ............................................................................................................. Singularis clade 13* Petals and filaments basally not connate and leaf sheaths not drying silver-gray.—From the southern United Stated to the Antilles, central Argentina and Uruguay with centers of diversity in northern Central America and the northern and central Andes (not further keyed out except for taxa of Tillandsia subg. Pseudovriesea bearing petal appendages)....... .................................................................................................................................. Tillandsia 14 Leaves narrowly triangular, xeromorphic, densely lepidote.—Petals usually violet, often bicolored with green, sometimes with crenulated margins.—Northern and central Andes extending to Bolivia, Mesoamerica, the Antilles, and eastern Venezuela ............................... ........................................................................................ Tillandsia subg. Pseudovriesea p.p. 14* Leaves lingulate, usually mesomorphic, not densely lepidote.—Corolla tubular, stamens and style usually included, rarely exserted. Petals white or yellow, rarely red. Pollen with a diffuse aperture ................................................................................................................... 15 15 Plants Andean extending to Venezuela, Trinidad, and the Guianas ........................................ .............................. Cipuropsis, Chrysostachys clade, Splendens clade, Vriesea tuerckheimii 15* Plant from Mexico ..................................................................................... Tillandsia malzinei 183 BARFUSS, M.H.J. NEW CLASSIFICATION Conclusions Much progress has been achieved in exploring phylogenetic relationships of Tillandsioideae by adding nuclear DNA information to the already existing and published plastid DNA data (Barfuss & al., 2005). Most accepted genera are well differentiated by both DNA and morphological data in their actual circumscriptions (Alcantarea, Catopsis, Glomeropitcairnia, Josemania, Lemeltonia, Racinaea, Rothowia, Vriesea, Werauhia). Guzmania in its present circumscription is holophyletic, but whether the corolla tube formation (Leins & Erbar, 2010) is a synapomorphic character for all Guzmania and different to that of Mezobromelia needs to be further tested. Generic definitions of the Cipuropsis-Mezobromelia clade need further attention, mostly in adding missing taxa and more DNA data as well as in evaluating morphological characters for defining clades and assessing their phylogenetic position relative to the preliminarily accepted genera Cipuropsis and Mezobromelia. Tillandsia as presently defined is holophyletic according to DNA data, but morphologically still very diverse, which is also demonstrated by the numerous subgenera and informal clades. Although several questions concerning relationships of species of former Tillandsia subg. Phytarrhiza and T. subg. Pseudalcantarea are solved, in other subgenera statistical support for their holophyletic origin is still missing. Relationships between clades and subgenera are also mostly unclear and lack statistical support. A solution has to be found for the unclassified species of Tillandsia. To get a fully resolved picture of relationships within Tillandsioideae, more nuclear DNA markers and a careful revision of morphological characters already identified as useful for discriminating taxa are still needed. Acknowledgments The authors thank Tod F. Stuessy for providing lab space; many botanical gardens and individual people for providing plant material (see "Taxon selection"); the "Botanischer Garten der Universität Wien" (HBV) for cultivating the majority of samples used; Jason Grant for valuable discussions; Eva Maria Mayr for helping with plant collections in Ecuador (2006); and Anton Russell for critically reviewing the manuscript. Financial support was partly provided by the Austrian Science Foundation (FWF) to W. Till (grant no. P13690-BIO), by the Commission for Interdisciplinary Ecological Studies (KIÖS) at the Austrian Academy of Sciences (ÖAW) to W. Till and M.H.J. Barfuss (2007-02), and by the Dept. of Systematic and Evolutionary Botany (Universität Wien). 184 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 References Altekar, G., Dwarkadas, S., Huelsenbeck, J.P. & Ronquist, F. 2004. Parallel Metropolis-coupled Markov chain Monte Carlo for Bayesian phylogenetic inference. Bioinformatics 20: 407– 415. Baker, J.G. 1878. A synopsis of the species of Diaphoranthema. J. Bot. 16: 236–241. Baker, J.G. 1887. Synopsis of Tillandsieae. J. Bot. 25: 52–55, 115–118, 171–177, 211–215, 234– 246, 277–281, 303–306, 344–347. Baker, J.G. 1888. A synopsis of Tillandsieae. J. Bot. 26: 12–17, 39–50, 79–82, 104–111, 137– 144, 167–172. Baker, J.G. 1889. Handbook of the Bromeliaceae. London: G. Bell & Sons. Barfuss, M., Samuel, M.R. & Till, W. 2004. Molecular phylogeny in subfamily Tillandsioideae (Bromeliaceae) based on six cpDNA markers: an update. J. Bromeliad Soc. 54: 9–17, 48. Barfuss, M.H.J., Samuel, R., Till, W. & Stuessy, T.F. 2005. Phylogenetic relationships in subfamily Tillandsioideae (Bromeliaceae) based on DNA sequence data from seven plastid regions. Amer. J. Bot. 92: 337–351. Beaman, R.S. 1989. Systematics of Tillandsia subgenus Pseudalcantarea (Bromeliaceae) and cladistic relationships of the Bromelioideae, Pitcairnioideae, and Tillandsioideae. Dissertation, University of Florida. Beaman, R.S. & Judd, W.S. 1996. Systematics of Tillandsia subgenus Pseudalcantarea (Bromeliaceae). Brittonia 48: 1–19. Benzing, D.H. 2000. Bromeliaceae: Profile of an adaptive radiation. Cambridge: Cambridge University Press. Bergsten, J. 2005. A review of long-branch attraction. Cladistics 21: 163–193. Betancur, J. & Miranda-Esquivel, D.R. 1999. ¿Existe Sodiroa? Revista Acad. Colomb. Ci. Exact. 23: 189–194. Böhme, S. 1988. Bromelienstudien III. Vergleichende Untersuchungen zu Bau, Lage und systematischer Verwertbarkeit der Septalnektarien von Bromelien. Trop. Subtrop. Pflanzenwelt 62: 125–274. Borsch, T. & Quant, D. 2009. Mutational dynamics and phylogenetic utility of noncoding chloroplast DNA. Pl. Syst. Evol. 282: 169–199. Brongniart, A. 1864. Note sur un nouveau genre de la famille des Broméliacées. Ann. Sci. Nat., Bot. 1: 325–329. Brown, G.K. & Gilmartin, A.J. 1984. Stigma structure and variation in Bromeliaceae— neglected taxonomic characters. Brittonia 36: 364–374. Brown, G.K. & Gilmartin, A.J. 1989a. Chromosome numbers in Bromeliaceae. Amer. J. Bot. 76: 657–665. Brown, G.K. & Gilmartin, A.J. 1989b. Stigma types in Bromeliaceae—a systematic survey. Syst. Bot. 14: 110–132. Brown, G.K. & Terry, R.G. 1992. Petal appendages in Bromeliaceae. Amer. J. Bot. 79: 1051– 1071. Chen, X., Yu, T., Xiong, J., Zhang, Y., Hua, Y., Li, Y. & Zhu, Y. 2004. Molecular cloning and expression analysis of rice phosphoribulokinase gene that is regulated by environmental stresses. Molec. Biol. Rep. 31: 249–255. 185 BARFUSS, M.H.J. NEW CLASSIFICATION Chew, T., De Luna, E. & González, D. 2010. Phylogenetic relationships of the pseudobulbous Tillandsia species (Bromeliaceae) inferred from cladistic analyses of ITS 2, 5.8S ribosomal RNA gene, and ETS sequences. Syst. Bot. 35: 86–95. Crayn, D.M., Winter, K. & Smith, J.A.C. 2004. Multiple origins of crassulacean acid metabolism and the epiphytic habit in the Neotropical family Bromeliaceae. Proc. Natl. Acad. Sci. U.S.A. 101: 3703–3708. Denton, A.L., McConaughy, B.L. & Hall, B.D. 1998. Usefulness of RNA Polymerase II coding sequences for estimation of green plant phylogeny. Molec. Biol. Evol. 15: 1082–1085. Doyle, J.J. & Doyle, J.L. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem. Bull. Bot. Soc. Amer. 19: 11–15. Edgar, R.C. 2004a. MUSCLE: A multiple sequence alignment method with high accuracy and high throughput. Nucl. Acids Res. 32: 1792–1797. http://www.drive5.com/muscle/ Edgar, R.C. 2004b. MUSCLE: A multiple sequence alignment method with reduced time and space complexity. B.M.C. Bioinf. 5: 113. Ehlers, R. 1997 ("1996"). Die rotblühenden brasilianischen Tillandsien/The red-flowered tillandsias from Brazil. Bromelie Sonderheft 3: 1–68. Ehlers, R. 2009. Die grünblühenden, kleinen, grauen Tillandsien Mexikos/The green-blooming small, grey tillandsias from Mexico. Bromelie Sonderheft 6: 1–144. Espejo-Serna, A. 2002. Viridantha, un género nuevo de Bromeliaceae (Tillandsioideae) endémico de México. Acta Bot. Mex. 60: 25–35. Espejo-Serna, A., López-Ferrari, A.R., Martínez-Correa, N. & Pulido-Esparza, V.A. 2007. Bromeliad flora of Oaxaca, Mexico: richness and distribution. Acta Bot. Mex. 81: 71–147. Evans, T.M. & Brown, G.K. 1989. Plicate staminal filaments in Tillandsia subgenus Anoplophytum (Bromeliaceae). Amer. J. Bot. 76: 1478–1485. Farris, J.S., Källersjö, M., Kluge, A.G. & Bult, C. 1994. Testing significance of congruence. Cladistics 10: 315–319. Farris, J.S., Källersjö, M., Kluge, A.G. & Bult, C. 1995. Constructing a significance test for incongruence. Syst. Biol. 44: 570–572. Ford, V.S., Lee, J., Baldwin, B.G. & Gottlieb, L.D. 2006. Species divergence and relationships in Stephanomeria (Compositae): PGIC phylogeny compared to prior biosystematic studies. Amer. J. Bot. 93: 480–490. Gardner, C.S. 1982. A systematic study of Tillandsia subgenus Tillandsia. Dissertation, Texas A&M University. Gardner, C.S. 1986a. Inferences about pollination in Tillandsia (Bromeliaceae). Selbyana 9: 76– 87. Gardner, C.S. 1986b. Preliminary Classification of Tillandsia based on floral characters. Selbyana 9: 130–146. Gilmartin, A.J. 1983. Evolution of mesic and xeric habits in Tillandsia and Vriesea (Bromeliaceae). Syst. Bot. 8: 233–242. Gilmartin, A.J. & Brown, G.K. 1986. Cladistic tests of hypotheses concerning evolution of xerophytes and mesophytes within Tillandsia subg. Phytarrhiza (Bromeliaceae). Amer. J. Bot. 73: 387–397. Gilmartin, A.J., Brown, G.K., Varadarajan, G.S. & Neighbors, N. 1989. Status of Glomeropitcairnia within evolutionary history of Bromeliaceae. Syst. Bot. 14: 339–348. 186 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 Givnish, T.J., Barfuss, M.H.J., Van Ee, B., Riina, R., Schulte, K., Horres, R., Gonsiska, P.A., Jabaily, R.S., Crayn, D.M., Smith, J.A.C., Winter, K., Brown, G.K., Evans, T.M., Holst, B.K., Luther, H.E., Till, W., Zizka, G., Berry, P.E. & Sytsma, K.J. 2011. Phylogeny, adaptive radiation, and historical biogeography in Bromeliaceae: Insights from an eight-locus plastid phylogeny. Amer. J. Bot. 98: 872–895. Givnish, T.J., Millam, K.C., Berry, P.E. & Sytsma, K.J. 2007. Phylogeny, adaptive radiation, and historical biogeography of Bromeliaceae inferred from ndhF sequence data. Pp. 3–26 in: Columbus, J.T., Friar, E.A., Porter, J.M., Prince, L.M. & Simpson M.G. (eds.), Monocots: Comparative Biology and Evolution—Poales. Claremont: Rancho Santa Ana Botanic Garden (RSABG). Givnish, T.J., Millam, K.C., Evans, T.M., Hall, J.C., Pires, J.C., Berry, P.E. & Sytsma, K.J. 2004. Ancient vicariance or recent long-distance dispersal? Inference about phylogeny and South American-African disjunctions in Rapateaceae and Bromeliaceae based on ndhF sequence data. Int. J. Plant Sci. 165: S35–S54. Górniak, M., Paun, O. & Chase, M.W. 2010. Phylogenetic relationships within Orchidaceae based on a low-copy nuclear coding gene, XDH: Congruence with organellar and nuclear ribosomal DNA results. Molec. Phylog. Evol. 56: 784–795. Gortan, G. 1991. Narbenformen bei Bromeliaceen: Variationsmöglichkeiten und Überlegungen zu systematisch-taxonomischen Korrelationen. Dissertation, Universität Wien. Granados Mendoza, C.G. 2008. Sistemática del complejo Tillandsia macdougallii (Bromeliaceae). Dissertation, Universidad Nacional Autónoma de México. Grant, J.R. 1993a. New combinations in Mezobromelia and Racinaea (Bromeliaceae: Tillandsioideae). Phytologia 74: 428–430. Grant, J.R. 1993b. True tillandsias misplaced in Vriesea (Bromeliaceae: Tillandsioideae). Phytologia 75: 170–175. Grant, J.R. 1994a. The Tillandsia adpressa assemblage: A review and new combinations in Racinaea (Bromeliaceae: Tillandsioideae). Novon 4: 362–364. Grant, J.R. 1994b. The reduction of Platyaechmea under Hoplophytum, and a new name in Tillandsia (Bromeliaceae). Phytologia 77: 99–101. Grant, J.R. 1995a. Bromelienstudien. The resurrection of Alcantarea and Werauhia, a new genus. Trop. Subtrop. Pflanzenwelt 91: 1–57. Grant, J.R. 1995b. New combinations and new taxa in the Bromeliaceae. Phytologia 79: 254– 256. Grant, J.R. 2005 ("2004"). New combinations and names in Andean Pitcairnia, Tillandsia, and Werauhia (Bromeliaceae). Vidalia 2: 23–25. Grant, J.R. & Zijlstra, G. 1998. An annotated catalogue of the generic names of the Bromeliaceae. Selbyana 19: 91–1. Grisebach, A.H.R. 1864. Ueber die von Fendler in Venezuela gesammelten Bromeliaceen. Nachr. Königl. Ges. Wiss. Georg-August-Univ. [1]: 1–21. Gross, E. 1988. Bromelienstudien IV. Zur Morphologie der Bromeliaceen-Samen unter Berücksichtigung systematisch-taxonomischer Aspekte. Trop. Subtrop. Pflanzenwelt 64: 1– 215. Halbritter, H. 1988. Bromeliaceae: Pollenmorphologie und Systematik. Die Entwicklung des Pollens von Tillandsia sinuosa L.B. Smith. Dissertation, Universität Wien. Halbritter, H. 1992. Morphologie und systematische Bedeutung des Pollens der Bromeliaceae. Grana 31: 197–212. 187 BARFUSS, M.H.J. NEW CLASSIFICATION Hall, T.A. 1999. BioEdit: A user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl. Acids Symp. Ser., 41: 95–98. Harms, H. 1929. Bromeliaceae novae III. Notizbl. Bot. Gart. Berlin-Dahlem 10: 784–805. Harms, H. 1930. Bromeliaceae. Vol. 15a. Pp. 65–159 in: Engler, A. & Prantl, K. (eds.), Die natürlichen Pflanzenfamilien, 2nd edition. Leipzig: W. Engelmann. Hörandl, E. & Stuessy, T.F. 2010. Paraphyletic groups as natural units of biological classification. Taxon 59: 1641–1653. Horres, R., Zizka, G., Kahl, G. & Weising, K. 2000. Molecular phylogenetics of Bromeliaceae: evidence from trnL (UAA) intron sequences of the chloroplast genome. Pl. Biol. 2: 306– 315. Howarth, D.G. & Baum, D.A. 2002. Phylogenetic utility of a nuclear intron from nitrate reductase for the study of closely related plant species. Molec. Phylog. Evol. 23: 525–528. Hromadnik, L. 2005. Der Verwandtschaftskreis um Tillandsia tectorum/The Tillandsia tectorum family. Bromelie Sonderheft 5: 1–120. Huelsenbeck, J.P. & Ronquist, F. 2001. MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754–755. Jabaily, R.S. & Sytsma, K.J. 2010. Phylogenetics of Puya (Bromeliaceae): Placement, major lineages, and evolution of Chilean species. Amer. J. Bot. 97: 337–356. Kelchner, S.A. 2000. The evolution of noncoding chloroplast DNA and its application in plant systematics. Ann. Missouri Bot. Gard. 87: 482–498. Kessler, M. 2002. Environmental patterns and ecological correlates of range size among bromeliad communities of Andean forests in Bolivia. Bot. Rev. (Lancaster). 68: 100–127. Krömer, T., Kessler, M., Lohaus, G. & Schmidt-Lebuhn, A.N. 2008. Nectar sugar composition and concentration in relation to pollination syndromes in Bromeliaceae. Pl. Biol. 10: 502– 511. Lee, M.S.Y. 2001. Uninformative characters and apparent conflict between molecules and morphology. Molec. Biol. Evol. 18: 676–680. Leins, P. & Erbar, C. 2010. Flower and fruit: morphology, ontogeny, phylogeny, function and ecology. Stuttgart: E. Schweizerbart'sche Verlagsbuchhandlung. Leme, E.M.C. 2007. Improving taxa and character sampling to support generic and infrageneric status of Alcantarea. J. Bromeliad Soc. 57: 208–215. Leme, E.M.C. 2009. Notes on Alcantarea: a new medium-sized species and additions to A. roberto-kautskyi. J. Bromeliad Soc. 59: 19–27. Leme, E.M.C. & Brown, G.K. 2004. Four new lithophytic Vriesea species (Tillandsioideae) from southeastern Brazil. Vidalia 2: 3–11. León, B. & Sagástegui, A. 2008. General overview of Tillandsia subgenus Tillandsia in Peru: The three-pinnate species and the case of two endemic species/Sinopsis de Tillandsia subgenus Tillandsia en el Peru: las especies tri-pinnadas y el caso de dos especies endémicas. Revista Peruana Biol. 15: 25–30. Lewis, C.E. & Doyle, J.J. 2001. Phylogenetic utility of the nuclear gene malate synthase in the palm family (Arecaceae). Molec. Phylogen. Evol. 19: 409–420. Lewis, C.E. & Doyle, J.J. 2002. A phylogenetic analysis of tribe Areceae (Arecaceae) using two low-copy nuclear genes. Pl. Syst. Evol. 236: 1–17. Lindley, J. 1843. Edwards’s Botanical Register 29: pl. 1–66. Linné, C. 1753. Species Plantarum. Stockholm: Laurentius Salvius. 188 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 López-Ferrari, A.R. & Espejo-Serna, A. 2009. Nuevas combinaciones en monocotiledóneas mexicanas IV (Bromeliaceae, Orchidaceae). Acta Bot. Mex. 89. 43–46. Luther, H.E. 1998. Miscellaneous new taxa of Bromeliaceae (XII). Selbyana 19: 83–90. Luther, H.E. 2001. De Rebus Bromeliacearum III. Selbyana 22: 34–67. Luther, H.E. 2008. An alphabetical list of bromeliad binomials, 11th edition. The Bromeliad Society International. http://selby.org/research/papers/alphabetical-list-bromeliad-binomials. Luther, H.E. 2010. An alphabetical list of bromeliad binomials, 12th edition. Pp. I–IV, 1–45, Rabinowitz, L. & Holst, B.K. (eds.). Sarasota: The Sarasota Bromeliad Society & Marie Selby Botanical Gardens (MSBG). Luther, H.E. & Rabinowitz, L. 2010. De Rebus Bromeliacearum IV. Selbyana 30: 147–189. Luther, H.E. & Sieff, E. 1994. De Rebus Bromeliacearum I. Selbyana 15: 9–93. Luther, H.E. & Sieff, E. 1997. De Rebus Bromeliacearum II. Selbyana 18: 103–140. Manzanares, J.M. & Gouda, E.J. 2010. Four new species of the genus Racinaea (Bromeliaceae) from Ecuador. Phytotaxa 3: 1–18. Mathews, S. & Donoghue, M.J. 1999. The root of angiosperm phylogeny inferred from duplicate phytochrome genes. Science 286: 947–950. McNeill, J., Barrie, F.R., Burdet, H.M., Demoulin, V., Hawksworth, D.L., Marhold, K., Nicolson, D.H., Prado, J., Silva, P.C., Skog, J.E., Wiersema, J.H. & Turland, N.J. (eds.) 2006. International code of botanical nomenclature (Vienna Code): Adopted by the Seventeenth International Botanical Congress Vienna, Austria, July 2005. Regnum Vegetabile 146. Ruggell: Gantner. Mez, C. 1891–94. Bromeliaceae. Pp. 173–634, pl. 51–114 in: Martius, C.F.P., Eichler, A.G., Urban, I. (eds.), Flora brasiliensis 3 (3). Leipzig: F. Fleischer. Mez, C. 1896. Bromeliaceae, vol. 9. Pp. 1–990 in: de Candolle, C. (ed.), Monographiae phanerogamarum. Paris: Masson & Cie. Mez, C. 1934–35. Bromeliaceae. Pp. 1–667 in: Endler, A. (ed.), Das Pflanzenreich IV. 32. Leipzig: W. Engelmann. Morren, E. 1870. Description du Tillandsia (Wallisia) hamaleana Ed. Mn. Belgique Hort. 20: 97– 102, pl. s.n. Morren, E. 1871. Notice sur le Tillandsia lindeni, var. luxurians. Belgique Hort. 21: 289–290, pl. 20–21. Nickrent, D.L., Malécot, V., Vidal-Russell, R. & Der, J.P. 2010. A revised classification of Santalales. Taxon 59: 538–558. Nylander, J.A.A. 2004. MrModeltest v2. Program distributed by the author. Evolutionary Biology Centre, Uppsala University. Oxelman, B., Yoshikawa, N., McConaughy, B.L., Luo, J., Denton, A.L. & Hall, B.D. 2004. RPB2 gene phylogeny in flowering plants, with particular emphasis on asterids. Molec. Phylogen. Evol. 32: 462–479. Palací, C.A. 1997. A systematic revision of the genus Catopsis (Bromeliaceae). Dissertation, University of Wyoming, Laramie. Palma-Silva, C., dos Santos, D.G., Kaltchuk-Santos, E. & Bodanese-Zanettini, M.H. 2004. Chromosome numbers, meiotic behavior, and pollen viability of species of Vriesea and Aechmea genera (Bromeliaceae) native to Rio Grande do Sul, Brazil. Amer. J. Bot. 91: 804– 807. 189 BARFUSS, M.H.J. NEW CLASSIFICATION Ranker, T.A., Soltis, D.E., Soltis, P.S. & Gilmartin, A.J. 1990. Subfamilial phylogenetic relationships of the Bromeliaceae: evidence from chloroplast DNA restriction site variation. Syst. Bot. 15: 425–434. Rauh, W. 1973. Bromelienstudien I. Neue und wenig bekannte Arten aus Peru. Trop. Subtrop. Pflanzenwelt 3: 167–203. Read, R.W. 1968. A new combination in Vriesea (Bromeliaceae). Phytologia 16: 457–458. Reeves, G.M., Chase, M.W., Goldblatt, P., De Chies, T., Lejeune, B., Fay, M.F., Cox, A.V. & Ruddall, P.J. 2001. Molecular systematics of Iridaceae: evidence from four plastid DNA regions. Amer. J. Bot. 88: 2074–2087. Regel, E. 1869. VI. Annotationes botanicae. Index Seminum Hort. Bot. Petrop. 1868: 77–92. Ronquist, F. & Huelsenbeck, J.P. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574. Ruiz, H. & Pavón, J. 1802. Flora peruviana, et chilensis. 3: i–xxiv, 1–95, pl. 223–325. [Madrid]: Gabriél de Sancha. Russell, A., Samuel, R., Rupp, B., Barfuss, M.H.J., Safran, M., Besendorfer, V. & Chase, M.W. 2010a. Phylogenetics and cytology of a pantropical orchid genus Polystachya (Polystachyinae; Vandeae; Orchidaceae): evidence from plastid DNA sequence data. Taxon 59: 389– 404. Russell, A., Samuel, R., Klejna, V., Barfuss, M.H.J., Rupp, B. & Chase, M.W. 2010b. Reticulate evolution in diploid and tetraploid species of Polystachya (Orchidaceae) as shown by plastid DNA sequences and low-copy nuclear genes. Ann. Bot. (Oxford) 106: 37–56. Samuel, R., Kathriarachchi, H., Hoffmann, P., Barfuss, M.H.J., Wurdack, K.J., Davis, C.C. & Chase, M.W. 2005. Molecular phylogenetics of Phyllanthaceae: Evidence from plastid matK and nuclear PHYC sequences. Amer. J. Bot. 92:132–141. Schill, R., Dannenbaum, C. & Jentzsch, E.-M. 1988. Untersuchungen an Bromeliennarben/Investigations on the stigma of Bromeliaceae. Beitr. Biol. Pflanzen 63: 221–252. Schmidt-Lebuhn, A.N., Kessler, M. & Hensen, I. 2007. Hummingbirds as drivers of plant speciation? Trends Pl. Sci. 12: 329–331. Schulte, K., Barfuss, M.H.J. & Zizka, G. 2009. Phylogeny of Bromelioideae (Bromeliaceae) inferred from nuclear and plastid DNA loci reveals the evolution of the tank habit within the subfamily. Molec. Phylogen. Evol. 51: 327–339. Smith, L.B. 1935. Studies in the Bromeliaceae - VI. Proc. Amer. Acad. Arts 70: 147–224. Smith, L.B. 1951. Studies in the Bromeliaceae - XVI. Contr. U.S. Natl. Herb. 29: 429–520. Smith, L.B. & Downs, R.J. 1974. Pitcairnioideae (Bromeliaceae). Fl. Neotrop. Monogr. 14 (1): 1– 660. Smith, L.B. & Downs, R.J. 1977. Tillandsioideae (Bromeliaceae). Fl. Neotrop. Monogr.14 (2): 661–1492. Smith, L.B. & Downs, R.J. 1979. Bromelioideae (Bromeliaceae). Fl. Neotrop. Monogr.14 (3): 1493–2142. Smith, L.B. & Till, W. 1998. Bromeliaceae. Pp. 74–99 in: Kubitzki, K. (ed.), The families and genera of vascular plants, vol. 4. Berlin: Springer. Spencer, M.A. & Smith, L.B. 1993. Racinaea, a new genus of Bromeliaceae (Tillandsioideae). Phytologia 74: 151–160. Stefano, M., Papini, A. & Brighigna, L. 2008. A new quantitative classification of ecological types in the bromeliad genus Tillandsia (Bromeliaceae) based on trichomes. Revista Biol. Trop. 56: 191–203. 190 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 Swofford, D.L. 2003. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4.0b10. Sunderland, Massachusetts: Sinauer. Tardivo, R.C. 2002. Revisão taxonômica de Tillandsia L. subgênero Anoplophytum (Beer) Baker (Bromeliaceae). Dissertation, Universidade de São Paulo. Tel-Zur, N., Abbo, S., Myslabodski, D. & Mizrahi, Y. 1999. Modified CTAB procedure for DNA isolation from epiphytic cacti of the genera Hylocereus and Selenicereus (Cactaceae). Pl. Molec. Biol. Reporter 17: 249–254. Terry, R.G. & Brown, G.K. 1996. A study of evolutionary relationships in Bromeliaceae based on comparison of DNA sequences from the chloroplast gene ndhF. J. Bromeliad Soc. 46: 107–112, 123. Terry, R.G., Brown, G.K. & Olmstead, R.G. 1997a. Examination of subfamilial phylogeny in Bromeliaceae using comparative sequencing of the plastid locus ndhF. Amer. J. Bot. 84: 664–670. Terry, R.G., Brown, G.K. & Olmstead, R.G. 1997b. Phylogenetic relationships in subfamily Tillandsioideae (Bromeliaceae) using ndhF sequences. Syst. Bot. 22: 333–345. Thomas, M.M., Garwood, N.C., Baker, W.J., Henderson, S.A., Russell, S.J., Hodel, D.R. & Bateman, R.M. 2006. Molecular phylogeny of the palm genus Chamaedorea, based on the low-copy nuclear genes PRK and RPB2. Molec. Phylogen. Evol. 38: 398–415. Till, W. 1984. Sippendifferenzierung innerhalb Tillandsia subgenus Diaphoranthema in Südamerika mit besonderer Berücksichtigung des Andenostrandes und der angrenzenden Gebiete. Dissertation, Universität Wien. Till, W. 1992. Systematics and evolution of the tropical-subtropical Tillandsia subgenus Diaphoranthema (Bromeliaceae). Selbyana 13: 88–94. Till, W. 2000a. Tillandsioideae. Pp. 555–572 in: Benzing, D.H. (ed.), Bromeliaceae: profile of an adaptive radiation. Cambridge: Cambridge University Press. Till, W. 2000b. Tillandsia and Racinaea. Pp. 573–586 in: Benzing, D.H. (ed.), Bromeliaceae: profile of an adaptive radiation. Cambridge: Cambridge University Press. Till, W. & Barfuss, M.H.J. 2006. Progress towards a new classification of Tillandsioideae. J. Bromeliad Soc. 56: 253–259. Till, W., Halbritter, H. & Gortan, G. 1997. Some notes on the remarkable bromeliad genus Glomeropitcairnia. J. Bromeliad Soc. 47: 65–72. Ule, E. 1907. II. Beiträge zur Flora der Hylaea nach den Sammlungen von Ule’s Amazonas Expedition. Verh. Bot. Vereins Prov. Brandenburg 48: 129–150. Utley, J.F. 1978. A new combination in Guzmania (Bromeliaceae). Phytologia 40: 55–57. Utley, J.F. 1983. A revision of the Middle American thecophylloid Vriesea (Bromeliaceae). Tulane Stud. Zool. Bot. 24: 1–81. Utley, J.F. & Luther, H.E. 1991. Studies in Middle American Bromeliaceae II. Ann. Missouri Bot. Gard. 78: 270. Versieux, L.M. 2009. Sistemática, filogenia e morfologia de Alcantarea (Bromeliaceae). Dissertation, Universidade de São Paulo. Versieux, L.M., Barbará, T., Wanderley, M.d.G.L., Calvente, A., Fay, M.F. & Lexer, C. 2012. Molecular phylogenetics of the Brazilian giant bromeliads (Alcantarea, Bromeliaceae): implications for morphological evolution and biogeography. Molec. Phylog. Evol. 64: 177– 189. 191 Versieux, L.M., Elbl, P.M., Wanderley, M.G.L., de Menezes, N.L. 2010. Alcantarea (Bromeliaceae) leaf anatomical characterization and its systematic implications. Nordic J. Bot. 28: 385–397. Versieux, L.M. & Wanderleym M.G.L. 2007. A new species of Alcantarea (E. Morren ex Mez) Harms, Bromeliaceae. Hoehnea 34: 409–413. Weber, W. & Smith, L.B. 1983. A new combination in Mezobromelia. J. Bromeliad Soc. 33: 121–122. Whitten, W.M., Williams, N.H. & Chase, M.W. 2000. Subtribal and generic relationships of Maxillarieae (Orchidaceae) with emphasis on Stanhopeinae: combined molecular evidence. Amer. J. Bot. 87: 1842–1856. Wittmack, L. 1888. Bromeliaceae. Pp. 32–58 in: Engler, A. & Prantl, K. (eds.), Die natürlichen Pflanzenfamilien II. 4. Leipzig: W. Engelmann. Yoder, A.D., Irwin, J.A. & Payseur, B.A. 2001. Failure of the ILD to determine data combinability for slow Ioris phylogeny. Syst. Biol. 50: 408–424. Appendix Figure 2. (p. 193) Selected equally most-parsimonious phylogram found in total combined analysis supplemented with BP and PP mostly below branches, and branch length is given above branches. A small arrowhead below branches indicates BP and PP positions where space was insufficient. Alternative topologies of terminal branches and backbone relationships from Bayesian analysis are displayed in boxes mostly in front of the corresponding clades or on the left side. Their positions within the complete tree are highlighted by capital letters (A–P: terminals, Q–S: backbones). Relevant type species for each taxonomic unit and name-giving taxa for informal clades are underlined. For samples which are similar to a already described species (aff.), with a vague identification (cf.), or undetermined species (sp.), as well as for clearly polyphyletic species voucher details are given in addition to the DNA reference number. Alternative tree topologies from Bayesian analysis are (1) terminals: A = Catopsis p.p., B = Werauhia p.p., C = Mezobromelia, D and E = Alcantarea p.p., F = Guzmania p.p., G = Racinaea p.p., H = Rothowia, I and J = T. subg. Tillandsia p.p., K = T. subg. Viridantha p.p., L = Purpurea clade, M = Biflora clade p.p., N = T. subg. Anoplophytum p.p., O = T. subg. Diaphoranthema p.p., P = Xiphioides clade; (2) backbones: Q = CipuropsisMezobromelia clade, R = Tillandsieae, S = T. subg. Anoplophytum s.l. plus T. subg. Phytarrhiza and T. subg. Diaphoranthema. Supplementary Data. (pp. 195–205) Accessions included in this study arranged in hierarchical, followed by alphabetical order (according to the chapter "Classification system of Tillandsioideae"). Attributes ofamplified fragments of PHYC and PRK are listed. SNP = single nucleotide polymorphism, SSR = single sequence repeat (homopolymers). 192 1 1 *Tillandsia bermejoensis B0034 *Tillandsia bermejoensis B0034 2 86/100 Tillandsia diaguitensis B0429 Tillandsia diaguitensis B0429 Tillandsia koehresiana B0371 Tillandsia koehresiana B0371 1 1 6 Tillandsia cochabambae B0392 Tillandsia lorentziana B0374 6 -/61 Tillandsia lorentziana B0374 Tillandsia aff. arequitae W. Till 6068 B0387 2 1 Tillandsia cardenasii B0409 Tillandsia cochabambae B0392 4 Tillandsia aff. arequitae W. Till 6068 B0387 Tillandsia cardenasii B0409 1 5 4 Tillandsia gerdae B0411 Tillandsia xiphioides subsp. prolata B0418 27 1 -/93 4 Usneoides clade Tillandsia xiphioides subsp. prolata B0418 4 Tillandsia muhriae B0438 1 99/100 -/93 -/79 3 60/97 1 Tillandsia muhriae B0438 Tillandsia aff. lotteae P. Lechner 20-451 B0808 6 1 2 -/94 97/100 Tillandsia yuncharaensis B0417 Tillandsia aff. lotteae P. Lechner 20-451 B0808 1 2 *Tillandsia argentina B0087 *Tillandsia xiphioides subsp. xiphioides B0040 12 -/76 2 2 -/65 -/72 Tillandsia argentina B0797 Tillandsia yuncharaensis B0417 1 Tillandsia rosarioae B0403 Tillandsia gerdae B0411 3 9 3 62/89 Tillandsia markusii B0399 Tillandsia argentina B0087 22 60/97 6 10 4 Tillandsia zecheri var. cafayatensis B0361 Tillandsia argentina B0797 -/54 -/72 100/100 9 4 *Tillandsia xiphioides subsp. xiphioides B0040 Tillandsia rosarioae B0403 1 62/89 Tillandsia aff. lotteae P. Lechner 20-451 B0808 Tillandsia zecheri var. cafayatensis B0361 -/100 5 97/100 Tillandsia aff. lotteae P. Lechner 20-451 B0809 Tillandsia markusii B0399 5 91/100 4 1 Tillandsia ramellae B0430 Tillandsia ramellae B0430 11 2 Tillandsia aff. vernicosa W. Till 6045 B0355 6 83/100 1 89/100 Tillandsia buchlohii B0395 18 -/96 1 Tillandsia recurvata 11 Tillandsia rosacea B0405 100/100 67/98 Xiphioides clade 3 2 *Tillandsia caulescens B0071 2 64/92 *Tillandsia didisticha B0038 10 98/100 Tillandsia comarapaensis B0407 *Tillandsia usneoides B0109 71/96 2 *Tillandsia usneoides B0109 2 71/96 6 6 Tillandsia usneoides B0451 Tillandsia usneoides B0451 92/100 2 92/100 2 *Tillandsia usneoides B0083 *Tillandsia usneoides B0083 27 22 8 1 99/100 Capillaris clade 26 Tillandsia mollis B0206 Tillandsia landbeckii subsp. andina B0423 6 -/99 3 9 99/100 8 91/100 Tillandsia cf. usneoides W. Till 21024 B0539 Tillandsia landbeckii subsp. andina B0423 3 -/54 1 9 62/81 Tillandsia mollis B0206 Tillandsia cf. usneoides W. Till 21024 B0539 1 Tillandsia duratii B0737 2 -/76 1 2 Tillandsia duratii B0740 62/88 7 1 1 *Tillandsia duratii B0088 56/99 63/100 1 3 Tillandsia aff. streptocarpa L. Hromadnik 20056 B0226 23 5 23 -/100 Tillandsia paleacea B0404 99/100 100/100 3 Tillandsia kirschnekii B0384 4 Tillandsia recurvata B0529 18 1 100/100 Tillandsia recurvata B0540 Tillandsia virescens B0516 4 90/100 1 13 Tillandsia capillaris B0518 100/100 24 Tillandsia kuehhasii B0396 1 4 17 -/99 -/64 Tillandsia esseriana 9 *Tillandsia funebris B0089 99/100 *Tillandsia albertiana B0033 67/92 1 1 7 Tillandsia albertiana B0225 2 92/100 6 Tillandsia albertiana B0212 Tillandsia edithae B0425 8 100/100 Tillandsia edithae B0436 16 *Tillandsia esseriana B0069 9 15 2 84/100 3 Tillandsia edithae 99/100 Tillandsia esseriana B0200 -/56 100/100 2 Tillandsia araujei B0394 2 14 1 Tillandsia aff. araujei E.C.M. Leme 491 B0422 7 4 4 Tillandsia recurvifolia var. subsecundifolia B0401 63/100 4 *Tillandsia tenuifolia B0026 5 4 *Tillandsia bergeri B0097 91/100 3 10 *Tillandsia bergeri B0097 5 2 91/100 *Tillandsia bergeri cv. Alba B0110 3 6 4 *Tillandsia bergeri cv. Alba B0110 7 5 5 *Tillandsia tenuifolia B0026 -/53 1 Tillandsia albertiana -/90 Tillandsia jonesii B0389 6 92/100 Tillandsia recurvifolia var. subsecundifolia B0401 12 *Tillandsia ixioides B0043 5 7 *Tillandsia pohliana B0080 2 22 5 15 -/71 84/100 *Tillandsia pohliana B0080 22 16 -/55 4 Tillandsia aff. araujei E.C.M. Leme 491 B0422 2 99/100 Tillandsia burle-marxii B0393 -/61 1 12 4 *Tillandsia ixioides B0043 2 1 Tillandsia aff. tenuifolia W. Till 11082 B0414 -/90 2 -/76 Tillandsia araujei B0394 5 12 **Tillandsia stricta s.l. B0081 & B0410 1 -/77 Tillandsia jonesii B0389 6 6 Tillandsia neglecta B0339 19 7 Gardneri clade *Tillandsia brachyphylla B0082 13 100/100 4 84/100 16 Tillandsia globosa B0419 18 100/100 Gardneri clade *Tillandsia gardneri B0041 Tillandsia australis B0203 13 100/100 Tillandsia australis B0759 1 Tillandsia lajensis B0242 1 Tillandsia lajensis B0242 60/99 60/99 Tillandsia lajensis B0546 Tillandsia lajensis B0546 5 3 3 86/100 86/100 Tillandsia lajensis B0554 Tillandsia lajensis B0554 -/55 2 Tillandsia lajensis B0555 Tillandsia lajensis B0555 4 Tillandsia buseri var. nubicola B0561 Tillandsia buseri var. nubicola B0561 7 7 -/80 6 2 100/100 100/100 Tillandsia cf. superba J.M. Manzanares 7787 B0602 Tillandsia cf. superba J.M. Manzanares 7787 B0602 -/80 62/96 5 5 Tillandsia orbicularis B0375 Tillandsia orbicularis B0375 2 2 2 1 -/63 Tillandsia sp. B0634 Tillandsia sp. J.M. Manzanares 8146 B0634 3 72/98 72/98 -/80 Tillandsia cf. buseri var. nubicola R. Wülfinghoff s.n. B0441 2 Tillandsia cf. buseri var. nubicola R. Wülfinghoff s.n. B0441 7 2 2 2 2 Tillandsia cf. buseri var. nubicola R. Wülfinghoff s.n. B0442 Tillandsia cf. buseri var. nubicola R. Wülfinghoff s.n. B0442 57/100 68/99 8 8 11 Tillandsia fendleri W. Till s.n. B0445 Tillandsia fendleri W. Till s.n. B0445 11 57/100 8 5 -/63 Tillandsia glauca B0618 Tillandsia glauca B0618 6 -/63 90/100 -/85 Tillandsia fosteri B0639 Tillandsia fosteri B0639 4 3 7 4 *Tillandsia demissa B0075 *Tillandsia demissa B0075 15 12 *Tillandsia fendleri H. & L. Hromadnik 2052 B0009 *Tillandsia fendleri H. & L. Hromadnik 2052 B0009 1 5 -/57 9 6 -/57 *Tillandsia baliophylla B0101 *Tillandsia baliophylla B0101 37 3 2 2 *Tillandsia brevilingua B0056 Tillandsia confinis B0587 3 8 100/100 1 -/93 2 Tillandsia rudolfii B0654 34 100/100 Tillandsia brevilingua B0252 100/100 6 6 Tillandsia brenneri B0236 *Tillandsia brevilingua B0056 8 100/100 1 Tillandsia brevilingua B0252 31 Tillandsia deppeana B0457 -/87 5 16 99/100 6 100/100 5 15 Tillandsia deppeana B0458 Tillandsia brenneri B0236 -/55 5 99/100 7 1 *Tillandsia heterophylla B0047 Tillandsia deppeana B0457 -/63 5 99/100 17 8 16 *Tillandsia multicaulis B0107 Tillandsia deppeana B0458 13 90/100 1 7 99/100 1 Tillandsia imperialis B0292 *Tillandsia heterophylla B0047 -/63 61/100 1 16 6 2 *Tillandsia kauffmannii B0074 *Tillandsia multicaulis B0107 -/55 4 99/100 90/100 1 2 4 Tillandsia roezlii B0764 Tillandsia imperialis B0292 92/78 8 1 11 Tillandsia subconcolor B0227 18 *Tillandsia kauffmannii B0074 4 99/100 3 -/62 2 6 Tillandsia confinis B0587 100/100 Tillandsia roezlii B0764 2 8 92/78 -/93 1 Tillandsia rudolfii B0654 10 Tillandsia subconcolor B0227 4 3 11 Tillandsia stenoura B0635 Tillandsia stenoura B0635 21 Tillandsia somnians B0301 93/100 13 95/100 100/100 4 Tillandsia somnians B0312 -/93 6 Tillandsia pastensis B0521 3 -/88 1 5 Tillandsia maculata B0574 1 2 5 -/55 Tillandsia tovarensis B0605 71/100 14 *Tillandsia biflora B0090 3 23 1 -/50 Tillandsia ionochroma B0600 31 1 Tillandsia complanata B0562 7 100/100 12 2 80/100 Tillandsia complanata B0563 -/68 4 Tillandsia complanata B0244 1 18 8 5 Biflora clade Tillandsia denudata var. vivipara B0750 1 5 100/100 -/96 *Tillandsia macbrideana B0070 3 4 6 1 Tillandsia krahnii B0235 2 6 *Tillandsia pseudomacbrideana B0036 17 80/100 3 Tillandsia incarnata B0522 5 81/100 Tillandsia latifolia var. leucophylla B0228 62/98 3 -/52 2 9 Tillandsia latifolia var. latifolia B0433 2 58/100 1 *Tillandsia latifolia var. divaricata B0068 4 Tillandsia confertiflora B0614 2 -/81 13 1 Tillandsia turneri B0650 11 Tillandsia floribunda B0351 14 7 *Tillandsia rauhii B0092 11 68/100 2 100/Tillandsia teres B0201 13 Rauhii clade 4 57/Tillandsia ecarinata B0237 9 8 100/84 Tillandsia ferreyrae B0241 3 3 Tillandsia purpurea B0246 Tillandsia purpurea B0246 8 8 2 2 99/100 3 99/100 3 Tillandsia straminea B0247 Tillandsia straminea B0247 82/99 82/99 2 Purpurea clade 2 2 11 14 *Tillandsia cacticola B0044 *Tillandsia cacticola B0044 66/94 5 100/100 100/100 4 4 25 24 Tillandsia aurea B0250 Tillandsia aurea B0250 -/72 3 57/11 Tillandsia nana B0343 Tillandsia sphaerocephala B0366 9 4 99/100 Tillandsia pseudomicans B0347 18 *Tillandsia tortilis B0049 3 94/100 85/100 2 5 Tillandsia lepidosepala B0219 72/100 1 Tillandsia mauryana B0238 1 3 4 *Tillandsia plumosa B0086 90/100 8 82/98 6 59/99 Tillandsia atroviridipetala B0215 5 2 Tillandsia ignesiae B0222 71/100 2 4 Tillandsia oblivata B0205 Tillandsia oblivata B0205 3 4 63/100 63/100 16 6 4 Tillandsia balsasensis B0221 Tillandsia balsasensis B0221 1 2 -/97 -/51 Tillandsia heteromorpha B0224 51/100 2 Tillandsia reducta B0209 3 1 73/100 4 Tillandsia stellifera B0239 Tillandsia chusgonensis B0210 9 23 51/100 1 73/100 Tillandsia heteromorpha B0224 4 2 Tillandsia reducta B0209 97/100 Singularis clade 8 3 99/100 1 Tillandsia chusgonensis B0210 Tillandsia stellifera B0239 97/100 -/93 5 95/100 3 3 2 Tillandsia tomekii B0218 Tillandsia tomekii B0218 2 -/99 Tillandsia tectorum var. tectorum f. tectorum B0249 Tillandsia tectorum var. tectorum f. tectorum B0249 3 Tillandsia tectorum var. globosa B0248 14 4 92/100 6 1 7 Tillandsia tectorum var. tectorum f. gigantea B0253 66/100 100/100 7 14 *Tillandsia rupicola B0039 64/98 10 7 Tillandsia tectorum var. viridula B0245 Chrysostachys clade 2 12 85/100 Tillandsia sphaerocephala B0366 Tillandsia hitchcockiana B0468 2 72/100 Tillandsia hitchcockiana B0629 4 96/100 Tillandsia cereicola B0134 1 Tillandsia cereicola B0476 2 14 1 -/100 Tillandsia espinosae B0143 6 22 -/88 97/100 1 Tillandsia espinosae B0462 4 Vriesea tuerckheimii B0148 2 1 -/98 89/100 Tillandsia andreettae B0373 3 3 80/100 Tillandsia cf. werneriana W. Rauh 40226 B0514 6 Tillandsia cf. werneriana BGBM 109-40-74-83 B0144 1 6 Tillandsia fragrans B0632 -/86 3 Tillandsia spathacea B0565 1 -/52 6 3 Tillandsia tequendamae B0569 -/100 14 2 2 Tillandsia strobeliae B0198 -/88 3 Splendens clade 7 -/92 100/100 *Tillandsia werneriana B0067 1 -/93 Tillandsia peruviana B0470 3 17 **Tillandsia barclayana B0028 & B0142 *Tillandsia appenii B0066 2 9 -/70 2 77/100 Tillandsia appenii B0464 10 3 Tillandsia seleriana B0413 1 *Tillandsia barthlottii B0035 21 -/53 6 5 2 Tillandsia ehlersiana B0431 99/100 Tillandsia barthlottii B0716 -/56 7 4 3 3 *Tillandsia ionantha B0084 -/97 Tillandsia cf. porphyrocraspeda W. Till 13080 B0350 68/99 1 21 8 Tillandsia aff. caput-medusae HBV B00B81-1 B0424 2 Tillandsia myriantha B0760 3 -/100 -/90 10 8 1 Tillandsia praschekii B0364 Tillandsia frank-hasei B0315 6 2 7 Tillandsia mitlaensis var. tulensis B0402 Tillandsia petraea B0630 5 -/94 4 89/100 9 Tillandsia pueblensis B0398 6 Tillandsia arpocalyx B0552 5 4 77/100 Tillandsia schatzlii B0416 *Tillandsia ionantha B0084 5 1 6 21 -/60 Tillandsia guerreroensis B0349 Tillandsia aff. caput-medusae HBV B00B81-1 B0424 2 -/97 14 10 Tillandsia variabilis B0308 2 Tillandsia seleriana B0413 2 1 13 6 -/70 Tillandsia cryptopoda B0787 -/53 Tillandsia ehlersiana B0431 3 7 3 -/57 10 Tillandsia cf. rhomboidea W. Till 19009 B0306 -/100 3 Tillandsia praschekii B0364 90/100 3 5 2 Tillandsia cf. rhomboidea W. Till 15048 B0372 1 -/100 Tillandsia mitlaensis var. tulensis B0402 3 4 1 Tillandsia cf. beutelspacheri R. Ehlers M890701 B0360 -/55 Tillandsia pueblensis B0398 5 5 4 9 89/100 5 Tillandsia exserta B0390 77/100 Tillandsia schatzlii B0416 -/99 2 5 6 *Tillandsia fasciculata B0076 Tillandsia guerreroensis B0349 4 -/60 5 14 -/58 Tillandsia glabrior B0340 3 Tillandsia variabilis B0308 2 -/60 9 13 4 Tillandsia baileyi B0421 -/70 Tillandsia cryptopoda B0787 3 10 -/79 7 4 Tillandsia durangensis B0365 Tillandsia cf. rhomboidea W. Till 19009 B0306 3 90/100 2 -/99 2 3 -/83 Tillandsia tricolor B0785 2 8 Tillandsia cf. rhomboidea W. Till 15048 B0372 -/100 12 3 -/69 5 5 *Tillandsia caput-medusae B0046 Tillandsia cf. beutelspacheri R. Ehlers M890701 B0360 4 -/81 8 1 Tillandsia plagiotropica B0386 7 *Tillandsia fasciculata B0076 1 8 8 98/100 Tillandsia velutina B0427 Tillandsia plagiotropica B0386 7 4 8 *Tillandsia remota B0072 98/100 5 12 Tillandsia velutina B0427 8 14 98/100 Tillandsia pseudosetacea B0333 *Tillandsia caput-medusae B0046 6 1 5 *Tillandsia guatemalensis HBV B260/96 B0008 2 Tillandsia glabrior B0340 3 -/67 -/60 6 9 11 Tillandsia selleana B0243 3 Tillandsia baileyi B0421 -/79 6 -/72 10 35 Tillandsia leiboldiana B0323 2 12 Tillandsia durangensis B0365 -/99 91/100 2 89/100 3 *Tillandsia guatemalensis H. & L. Hromadnik 14257 B0103 5 13 Tillandsia tricolor B0785 8 5 97/100 *Tillandsia guatemalensis L. Hromadnik 15127 B0104 *Tillandsia remota B0072 8 22 7 *Tillandsia punctulata B0061 98/100 Tillandsia pseudosetacea B0333 2 -/58 11 1 *Tillandsia andrieuxii B0063 *Tillandsia guatemalensis HBV B260/96 B0008 2 2 4 -/67 6 8 Tillandsia moronesensis B0380 11 2 Tillandsia selleana B0243 -/75 9 -/72 6 Tillandsia achyrostachys B0408 Tillandsia leiboldiana B0323 9 76/98 6 2 89/100 Tillandsia huajuapanensis B0713 *Tillandsia guatemalensis H. & L. Hromadnik 14257 B0103 13 5 8 Tillandsia juerg-rutschmannii B0715 97/100 *Tillandsia guatemalensis L. Hromadnik 15127 B0104 4 2 Tillandsia moronesensis B0380 76/98 2 12 9 2 Tillandsia achyrostachys B0408 91/100 60/6 11 Tillandsia huajuapanensis B0713 3 3 6 *Tillandsia andrieuxii B0063 -/75 19 *Tillandsia punctulata B0061 5 Tillandsia juerg-rutschmannii B0715 8 2 Tillandsia exserta B0390 Tillandsia aff. utriculata J. Lautner 95/48 B0743 7 Tillandsia aff. utriculata J. Lautner 95/48 B0747 100/100 1 20 Tillandsia utriculata B0805 100/100 *Tillandsia utriculata B0100 4 6 2 99/100 62/96 Tillandsia utriculata B0807 17 1 Tillandsia filifolia B0788 89/100 6 1 8 80/100 Tillandsia filifolia B0790 15 100/100 5 5 Tillandsia fuchsii B0391 67/94 12 Tillandsia aff. cucaensis W. Rauh 70802 B0732 4 Tillandsia propagulifera B0310 1 -/57 1 4 Tillandsia cf. mima var. mima W. Rauh 53543 B0709 -/96 Tillandsia spiraliflora B0762 3 81/100 74/100 1 Tillandsia mima var. chiletensis B0708 4 Tillandsia mima var. chiletensis B0710 6 4 79/100 -/70 Tillandsia extensa B0712 19 Tillandsia adpressiflora B0597 5 11 55/77 Tillandsia hildae B0763 Tillandsia secunda B0527 Tillandsia secunda B0528 Tillandsia secunda B0549 Tillandsia secunda B0528 20 17 1 100/100 100/100 3 Tillandsia secunda B0551 Tillandsia secunda B0549 35 1 26 Tillandsia secunda B0527 Tillandsia secunda B0551 -/75 6 17 12 Tillandsia malzinei B0145 *Tillandsia paniculata B0102 31 5 3 100/100 Tillandsia malzinei B0492 90/100 17 *Tillandsia paniculata B0102 *Tillandsia disticha B0048 2 94/99 52 Tillandsia disticha B0202 9 100/100 Tillandsia disticha B0233 Tillandsia macropetala B0742 7 100/100 9 Tillandsia macropetala B0748 11 99/100 *Tillandsia viridiflora B0006 21 2 99/100 Tillandsia grandis B0124 6 99/100 Tillandsia grandis B0125 1 Lemeltonia triglochinoides B0723 78/94 2 1 10 Lemeltonia triglochinoides B0725 85/100 2 3 Lemeltonia triglochinoides B0724 -/97 3 11 Lemeltonia scaligera HBV B128/96 B0327 70/100 3 9 Lemeltonia scaligera L. Hromadnik 17033 B0234 67/100 4 1 Lemeltonia monadelpha B0745 -/86 1 Lemeltonia acosta-solisii B0626 10 36 1 100/100 Lemeltonia cornuta B0744 3 100/100 *Lemeltonia dodsonii B0016 9 6 100/100 Lemeltonia dodsonii B0721 *Rothowia wagneriana B0058 *Rothowia wagneriana B0058 5 5 100/100 100/100 Rothowia wagneriana B0217 Rothowia wagneriana B0217 1 1 Rothowia platyrhachis allele 2 B0753 Rothowia platyrhachis allele 2 B0753 -/86 7 20 Rothowia laxissima B0294 Rothowia platyrhachis allele 1 B0753 9 17 100/100 81/100 Rothowia laxissima B0757 Rothowia laxissima B0294 81/100 8 6 100/100 Rothowia platyrhachis allele 1 B0753 Rothowia laxissima B0757 1 Racinaea pallidoflavens B0369 8 3 98/100 2 Racinaea diffusa B0480 -/76 7 Racinaea tetrantha var. tetrantha B0367 Racinaea tetrantha var. aurantiaca B0496 Racinaea tetrantha var. aurantiaca B0496 9 9 2 2 100/100 100/100 4 4 Racinaea pseudotetrantha B0624 Racinaea pseudotetrantha B0624 1 -/56 -/84 2 -/84 2 *Racinaea seemannii B0024 *Racinaea seemannii B0024 4 4 1 2 -/88 *Racinaea elegans B0051 *Racinaea elegans B0051 7 57/95 1 5 4 10 57/95 7 1 Racinaea riocreuxii B0440 Racinaea riocreuxii B0440 -/79 9 8 1 -/79 Racinaea tetrantha var. caribaea B0448 Racinaea tetrantha var. caribaea B0448 93/100 4 71/100 2 2 20 *Racinaea ropalocarpa B0057 *Racinaea ropalocarpa B0057 4 4 74/100 74/100 -/81 1 1 21 1 Racinaea parviflora B0335 Racinaea parviflora B0335 74/97 74/97 3 3 5 Racinaea tillii B0572 Racinaea tillii B0572 55/98 3 Racinaea pugiformis B0613 5 8 76/100 Racinaea sinuosa B0617 3 51/91 9 Racinaea pectinata B0520 2 *Racinaea spiculosa B0099 9 2 100/100 Racinaea spiculosa B0304 Racinaea multiflora var. multiflora B0426 8 13 100/100 61/100 9 Racinaea multiflora var. decipiens B0478 100/100 3 Racinaea multiflora var. tomensis B0437 4 2 84/99 Racinaea insularis B0494 2 64/100 2 7 Racinaea fraseri J.M. Manzanares 7815 B0616 4 96/100 1 53/100 Racinaea fraseri W. Till 21049a B0547 13 1 Racinaea dyeriana B0151 20 100/100 100/100 Racinaea dyeriana B0456 1 Racinaea venusta B0729 1 69/87 2 16 Racinaea venusta B0731 100/100 1 *Racinaea venusta B0007 21 2 95/100 Racinaea hamaleana B0251 7 5 Racinaea hamaleana B0730 98/100 Racinaea hamaleana B0749 1 **Josemania lindenii B0023 & B0746 4 1 95/100 Josemania cyanea B0780 4 Josemania cyanea B0782 85/100 Josemania anceps B0741 5 3 66/99 99/100 Josemania anceps B0781 30 6 Josemania pretiosa B0231 3 100/100 Josemania umbellata B0216 10 99/100 Josemania umbellata B0758 2 *Guzmania donnell-smithii B0053 11 2 6 100/100 Guzmania donnell-smithii B0126 9 4 67/100 Guzmania flagellata B0475 -/64 5 Guzmania rauhiana B0297 5 12 Guzmania lingulata var. lingulata B0444 2 -/69 -/71 8 1 Guzmania lingulata var. minor B0778 7 6 Guzmania glomerata B0131 97/100 3 Guzmania conglomerata B0314 7 1 18 -/72 *Guzmania patula B0011 3 90/100 8 1 100/100 *Guzmania rhonhofiana B0096 55/93 3 Guzmania sphaeroidea B0506 10 13 Guzmania roezlii B0286 2 58/77 2 87/100 5 1 Guzmania strobilantha B0504 -/74 8 Guzmania coriostachya B0584 3 *Guzmania musaica var. musaica B0014 6 1 3 97/100 Guzmania musaica var. discolor B0489 8 6 *Guzmania herrerae B0010 91/100 10 Guzmania cf. scherzeriana W. Till 15011 B0295 1 *Guzmania multiflora B0094 1 -/50 1 1 *Guzmania multiflora B0094 3 3 Guzmania multiflora B0606 -/50 1 4 1 Guzmania multiflora B0606 Guzmania polycephala B0651 12 3 *Guzmania acorifolia B0052 2 2 *Guzmania condensata B0055 -/62 4 2 3 6 *Guzmania condensata B0055 Guzmania roseiflora B0291 -/75 1 -/96 3 -/66 6 3 Guzmania roseiflora B0291 7 2 -/75 Guzmania squarrosa B0607 4 5 -/77 2 Guzmania polycephala B0651 -/75 *Guzmania acorifolia B0052 8 7 9 Guzmania squarrosa B0607 -/75 Guzmania retusa B0498 9 10 Guzmania retusa B0498 Guzmania brasiliensis B0303 2 1 -/57 2 2 Guzmania marantoidea B0487 -/56 1 1 Guzmania cf. killipiana J.M. Manzanares 8127 B0627 -/64 4 2 Guzmania farciminiformis B0796 -/87 6 1 1 Guzmania tarapotina B0591 -/67 4 Guzmania conifera B0138 9 Guzmania acuminata B0300 3 5 Guzmania ekmanii B0447 4 90/100 4 Guzmania virescens B0718 11 Guzmania xanthobracteata B0623 2 6 *Guzmania graminifolia allele 1 B0120 3 83/100 57/96 7 16 *Guzmania graminifolia allele 2 B0120 13 67/100 Guzmania pearcei B0610 6 Guzmania danielii B0510 2 -/68 10 8 Guzmania sp. J.M. Manzanares 8131 B0628 3 -/72 8 *Guzmania wittmackii B0012 91/74 14 6 **Guzmania variegata B0015 & B0646 7 -/50 9 Guzmania calamifolia B0483 1 11 51/74 -/69 Guzmania mosquerae B0582 3 Guzmania sanguinea var. sanguinea B0495 13 5 100/100 Guzmania sanguinea var. brevipedicellata B0500 52/79 9 21 Guzmania fusispica B0289 64/91 12 13 17 *Guzmania angustifolia B0093 94/100 39 68/100 3 *Guzmania monostachia B0022 67/61 4 *Guzmania melinonis B0032 22 28 5 100/100 Guzmania melinonis B0593 100/100 10 Guzmania nicaraguensis B0479 36 *Guzmania hutchisonii B0003 2 3 Alcantarea imperialis B0135 Alcantarea imperialis B0135 2 2 Alcantarea imperialis B0154 1 Alcantarea imperialis B0154 -/99 32 Alcantarea regina B0136 *Alcantarea imperialis B0001 3 1 94/100 1 100/100 Alcantarea cf. geniculata L.F. Carvalho 5 B0751 Alcantarea edmundoi B0171 3 90/100 1 3 3 *Alcantarea imperialis B0001 90/100 Alcantarea edmundoi B0161 3 Alcantarea edmundoi B0161 Alcantarea regina B0136 13 4 1 1 Alcantarea edmundoi B0171 73/100 94/100 Alcantarea cf. geniculata L.F. Carvalho 5 B0751 2 Alcantarea nevaresii E.M.C. Leme 2299 B0169 2 12 2 Alcantarea nevaresii E.M.C. Leme 2299 B0169 1 73/100 Alcantarea geniculata B0153 1 -/58 2 Alcantarea geniculata B0153 2 66/99 3 Alcantarea nevaresii E.M.C. Leme 2227 B0163 3 66/99 Alcantarea nevaresii E.M.C. Leme 2227 B0163 2 -/78 1 *Alcantarea duarteana B0059 3 96/100 1 Alcantarea duarteana B0165 -/73 9 7 Alcantarea benzingii B0166 7 Alcantarea roberto-kautskyi B0158 Alcantarea nahoumii B0156 1 Alcantarea nahoumii B0156 69/13 1 -/88 Alcantarea sp. E.M.C. Leme 3660 B0160 Alcantarea sp. E.M.C. Leme 3660 B0160 3 1 2 100/100 -/81 Alcantarea aff. extensa E.M.C. Leme 1942 B0164 Alcantarea burle-marxii B0159 3 2 2 59/-/78 Alcantarea glazioviana B0167 Alcantarea aff. extensa E.M.C. Leme 1942 B0164 7 4 8 Alcantarea burle-marxii B0159 Alcantarea sp. E.M.C. Leme 4803 B0168 91/100 91/100 1 5 1 Alcantarea sp. E.M.C. Leme 4803 B0168 Alcantarea glazioviana B0167 -/62 Alcantarea heloisae B0155 7 2 4 88/100 52/90 Alcantarea odorata B0157 3 Alcantarea farneyi B0170 1 *Vriesea jonghei B0065 6 8 2 96/100 Vriesea platynema B0146 12 Vriesea pabstii B0357 3 4 Vriesea flava B0322 6 3 10 51/99 Vriesea maxoniana B0490 7 4 29 *Vriesea correia-araujoi B0095 8 Vriesea procera var. tenuis B0296 7 100/100 5 69/100 Vriesea neoglutinosa B0311 5 6 *Vriesea psittacina allele 1 B0020 3 -/96 2 3 *Vriesea psittacina allele 2 B0020 -/96 6 2 Vriesea erythrodactylon B0309 -/97 8 4 12 Vriesea sucrei B0302 5 -/100 Vriesea scalaris B0147 4 60/96 1 Vriesea simplex B0505 10 3 *Vriesea saundersii B0004 Vriesea unilateralis B0316 18 100/100 2 Vriesea roethii B0511 -/59 11 1 1 100/100 Vriesea ruschii B0719 7 Vriesea sparsiflora B0484 20 4 Vriesea longicaulis B0290 8 89/100 Vriesea plurifolia B0376 5 89/100 9 6 Vriesea aff. croceana G. Croce s.n. B0801 11 66/100 Vriesea croceana B0802 1 Mezobromelia sp. J.M. Manzanares 7640 B0194 Mezobromelia schimperiana B0589 1 3 -/61 1 Mezobromelia schimperiana B0611 Mezobromelia schimperiana B0612 3 1 3 -/93 Mezobromelia schimperiana B0334 Mezobromelia aff. schimperiana J.M. Manzanares 7714 B0197 1 Mezobromelia bicolor B0578 Mezobromelia bicolor B0583 17 2 1 57/72 Mezobromelia schimperiana B0589 Mezobromelia schimperiana B0334 1 7 8 2 93/100 Mezobromelia aff. schimperiana J.M. Manzanares 7714 B0197 Mezobromelia schimperiana B0611 99/100 99/100 2 2 2 Mezobromelia schimperiana B0612 Mezobromelia sp. J.M. Manzanares 7640 B0194 1 2 2 2 Mezobromelia bicolor B0583 66/94 66/94 Mezobromelia cf. pleiosticha J.M. Manzanares 7766 B0330 2 Mezobromelia cf. pleiosticha J.M. Manzanares 7766 B0330 Mezobromelia bicolor B0578 1 1 Mezobromelia pleiosticha B0274 Mezobromelia pleiosticha B0274 5 5 19 99/100 2 Mezobromelia pleiosticha B0649 99/100 2 85/100 Mezobromelia pleiosticha B0649 1 *Cipuropsis zamorensis B0045 5 93/100 3 4 1 71/100 *Cipuropsis cv. Elan B0117 82/98 1 Cipuropsis amicorum B0318 14 4 *Tillandsia singularis B0064 22 17 99/100 Tillandsia asplundii B0588 Singularis clade 14 18 Vriesea tuerckheimii B0148 66/100 17 *Vriesea chrysostachys B0005 10 6 32 85/100 *Vriesea ospinae B0054 Chrysostachys clade -/88 3 *Vriesea splendens B0037 14 10 100/100 Vriesea glutinosa B0130 Splendens clade 2 Werauhia hygrometrica B0598 10 100/100 2 Werauhia hygrometrica B0621 72/98 11 4 Werauhia haltonii B0580 67/93 4 *Werauhia ororiensis B0025 6 17 8 69/100 20 Werauhia sintenisii B0450 -/99 16 26 Werauhia pedicellata B0752 100/100 100/100 24 *Werauhia insignis B0017 2 11 Werauhia viridiflora B0285 5 89/100 2 1 2 Werauhia viridiflora B0325 Werauhia viridiflora B0285 5 89/100 3 3 Werauhia ampla B0132 10 Werauhia viridiflora B0325 7 19 5 99/100 4 Werauhia gladioliflora B0149 -/80 13 100/100 Werauhia gladioliflora B0149 8 11 4 5 98/100 Werauhia gladioliflora B0509 98/100 Werauhia gladioliflora B0509 99/100 21 17 3 *Werauhia ringens B0019 21 Werauhia ampla B0132 9 60/99 100/100 Werauhia guadelupensis B0313 16 *Werauhia tarmaensis B0018 18 12 99/100 Werauhia werckleana B0497 10 *Catopsis nutans B0002 6 9 19 82/100 Catopsis delicatula B0461 52/100 9 2 Catopsis sessiliflora B0459 18 4 *Catopsis subulata B0105 10 *Catopsis nutans B0002 12 11 6 Catopsis nitida B0463 9 70/100 4 82/100 Catopsis delicatula B0461 14 52/100 4 *Catopsis juncifolia B0029 9 17 Catopsis sessiliflora B0459 -/88 5 -/55 *Catopsis morreniana B0106 27 13 11 Catopsis nitida B0463 2 11 96/100 99/100 Catopsis morreniana B0488 22 99 70/100 *Catopsis subulata B0105 7 Catopsis minimiflora B0467 17 *Catopsis juncifolia B0029 100/100 26 44 Catopsis paniculata B0507 2 *Glomeropitcairnia erectiflora B0030 19 100/100 1 83 100/100 Glomeropitcairnia erectiflora B0128 9 100/100 *Glomeropitcairnia penduliflora B0013 S 10 83/100 Xiphioides clade 86/100 2 P Tillandsia subg. Phytarrhiza P O O Tillandsia subg. Anoplophytum Tillandsia subg. Phytarrhiza & Tillandsia subg. Diaphoranthema Tillandsia subg. Anoplophytum N N S M Tillandsia subgg./clades R Tillandsia subg. Pseudovriesea Lemeltonia M Racinaea subg. Racinaea Racinaea subg. Pseudophytarrhiza Josemania Rothowia Guzmania L Mezobromelia Q L Tillandsia subg. Viridantha Cipuropsis K K Tillandsia subg. Pseudovriesea J J Tillandsia subg. Tillandsia Tillandsia I I Tillandsia subg. Pseudalcantarea Lemeltonia H Rothowia H G Core Tillandsieae G Racinaea Racinaea subg. Racinaea Racinaea subg. Pseudophytarrhiza Josemania Tillandsieae R F F Guzmania Core Tillandsioideae E E Alcantarea D D Vrieseinae Vriesea Mezobromelia Tillandsioideae C Vrieseeae CipuropsisMezobromelia clade C Cipuropsis Q Cipuropsidinae Werauhia B B Catopsis A Pogospermeae A Glomeropitcairnia 5 changes Glomeropitcairnieae PHYC PRK postulated ploidy level Geno types Length Glomeropitcairnia erectiflora Mez B0030 1 2 1177 Glomeropitcairnia erectiflora Mez B0128 0 1 1177 0 0 0 1 1198 1198 1198 2n = 2x Glomeropitcairnia penduliflora (Griseb.) Mez B0013 0 1 1177 0 0 0 1 1205 1205 1205 2n = 2x Catopsis delicatula L.B. Sm. B0461 3 2 1177 0 0 0 1 927 927 927 2n = 2x Catopsis juncifolia Mez & Wercklé ex Mez B0029 0 1 1177 0 0 0 1 975 975 975 2n = 2x Indels SSR Consensus length Allele 2 length (a2) SNPs SNPs Geno types Allele 1 length (a1) DNA no. Taxon Bromeliaceae tribe Glomeropitcairnieae Harms Glomeropitcairnia (Mez) Mez 7 0 0 2 1198 1198 1198 2n = 2x Bromeliaceae tribe Pogospermeae Brongn. Catopsis Griseb. Catopsis minimiflora Matuda B0467 0 1 1177 7 1 0 2 1005 1005 1004 2n = 2x Catopsis morreniana Mez B0106 0 1 1177 3 0 0 2 969 969 969 2n = 2x Catopsis morreniana Mez B0488 0 1 1177 4 0 0 2 969 969 969 2n = 2x Catopsis nitida (Hook.) Griseb. B0463 0 1 1177 0 0 0 1 968 968 968 2n = 2x Catopsis nutans (Sw.) Griseb. B0002 0 1 1177 0 0 0 1 927 927 927 2n = 2x Catopsis paniculata E. Morren B0507 0 1 1177 0 0 0 1 892 892 892 2n = 2x Catopsis sessiliflora (Ruiz & Pav.) Mez B0459 1 2 1177 6 0 0 2 927 927 927 2n = 2x Catopsis subulata L.B. Sm. B0105 2 2 1177 6 0 0 2 969 969 969 2n = 2x Bromeliaceae tribe Tillandsieae Rchb. Guzmania Ruiz & Pav. Guzmania acorifolia (Griseb.) Mez B0052 0 1 1192 0 0 0 1 1021 1021 1021 2n = 2x Guzmania acuminata L.B. Sm. B0300 0 1 1177 1 0 0 2 992 992 992 2n = 2x Guzmania angustifolia (Baker) Wittm. B0093 1 2 1177 9 2 0 2 1050 1050 1018 2n = 2x Guzmania brasiliensis Ule B0303 0 1 1177 0 0 0 1 1037 1037 1037 2n = 2x Guzmania calamifolia André ex Mez B0483 1 2 1177 1 0 0 2 1033 1033 1033 2n = 2x Guzmania condensata Mez & Wercklé B0055 0 1 1192 0 0 0 1 918 918 918 2n = 2x Guzmania conglomerata H. Luther B0314 0 1 1177 8 0 0 2 967 967 967 2n = 2x Guzmania conifera (André) André ex Mez B0138 2 2 1177 0 0 0 1 993 993 993 2n = 2x Guzmania coriostachya (Griseb.) Mez B0584 1 2 1177 5 1 0 2 926 920 926 2n = 2x Guzmania danielii L.B. Sm. B0510 1 2 1177 2 0 0 2 1056 1056 1056 2n = 2x Guzmania donnell-smithii Mez B0053 0 1 1177 0 0 0 1 1040 1040 1040 2n = 2x Guzmania donnell-smithii Mez B0126 0 1 1177 0 0 0 1 1040 1040 1040 2n = 2x Guzmania ekmanii (Harms) Harms ex Mez B0447 0 1 1177 0 0 0 1 1026 1026 1026 2n = 2x Guzmania farciminiformis H. Luther B0796 1 2 1177 3 0 0 2 1008 1008 1008 2n = 2x Guzmania flagellata S. Pierce & J.R. Grant B0475 0 1 1177 0 0 0 1 1035 1035 1035 2n = 2x Guzmania fusispica Mez & Sodiro B0289 0 1 1177 0 0 0 1 1050 1050 1050 2n = 2x Guzmania glomerata Mez & Wercklé B0131 0 1 1177 0 0 0 1 967 967 967 2n = 2x Guzmania graminifolia (André ex Baker) L.B. Sm. B0120 2 2 1177 9 4 1 2 n.u. 1209! 1185! 2n = 2x Guzmania herrerae H. Luther & W.J. Kress B0010 1 2 1177 13 1 1 2 1220 1220 1219 2n = 2x Guzmania hutchisonii (L.B. Sm.) Barfuss & W. Till B0003 1 2 1177 0 0 0 1 1129 1129 1129 2n = 2x Guzmania cf. killipiana L.B. Sm. B0627 0 1 1177 8 0 0 2 1020 1020 1020 2n = 2x Guzmania lingulata (L.) Mez var. lingulata B0444 0 1 1177 0 0 0 1 1003 1003 1003 2n = 2x Guzmania lingulata var. minor (Mez) L.B. Sm. & Pittendr. B0778 0 1 1177 0 0 0 1 949 949 949 2n = 2x Guzmania marantoidea (Rusby) H. Luther B0487 6 2 1177 6 2 0 2 1022 1018 939 2n = 2x Guzmania melinonis Regel B0032 0 1 1177 0 0 0 1 1132 1132 1132 2n = 2x Guzmania melinonis Regel B0593 1 2 1177 0 0 1 1 1154 1154 1154 2n = 2x Guzmania monostachia (L.) Rusby ex Mez B0022 0 1 1177 0 0 0 1 1005 1005 1005 2n = 2x 195 BARFUSS, M.H.J. NEW CLASSIFICATION PHYC PRK postulated ploidy level SNPs Geno types Length Guzmania mosquerae (Wittm.) Mez B0582 0 1 1180 0 0 0 1 1055 1055 1055 2n = 2x Guzmania multiflora (André) André ex Mez B0094 0 1 1177 0 0 0 1 926 926 926 2n = 2x Guzmania multiflora (André) André ex Mez B0606 0 1 1177 0 0 1 1 1098 1098 1098 2n = 2x Guzmania musaica var. discolor H. Luther B0489 0 1 1177 4 0 1 2 1225 1225 1225 2n = 2x Guzmania musaica (Linden & André) Mez var. musaica B0014 1 2 1177 1 0 1 2 1225 1225 1225 2n = 2x Guzmania nicaraguensis Mez & C.F. Baker B0479 0 1 1177 0 0 0 1 1127 1127 1127 2n = 2x Guzmania patula Mez & Wercklé B0011 0 1 1177 0 0 0 1 831 831 831 2n = 2x Guzmania pearcei (Baker) L.B. Sm. B0610 0 1 1177 0 0 1 1 1232 1232 1232 2n = 2x Guzmania polycephala Mez & Wercklé ex Mez B0651 0 1 1177 0 0 0 1 917 917 917 2n = 2x Indels SSR Consensus length Allele 2 length (a2) DNA no. SNPs Geno types Allele 1 length (a1) Taxon Guzmania rauhiana H. Luther B0297 0 1 1177 0 0 0 1 1325 1325 1325 2n = 2x Guzmania retusa L.B. Sm. B0498 1 2 1177 0 0 0 1 995 995 995 2n = 2x Guzmania rhonhofiana Harms B0096 0 1 1177 0 0 0 1 831 831 831 2n = 2x Guzmania roezlii (E. Morren) Mez B0286 0 1 1177 0 0 0 1 884 884 884 2n = 2x Guzmania roseiflora Rauh B0291 6 2 1177 1 1 0 2 934 934 928 2n = 2x Guzmania sanguinea var. brevipedicellata Gilmartin B0500 0 1 1177 3 0 0 2 1050 1050 1050 2n = 2x Guzmania sanguinea (André) André ex Mez var. sanguinea B0495 2 2 1177 4 0 0 2 1050 1050 1050 2n = 2x Guzmania cf. scherzeriana Mez B0295 0 1 1177 0 0 1 1 1218 1218 1218 2n = 2x Guzmania sphaeroidea (André) André ex Mez B0506 0 1 1177 1 0 0 2 838 838 838 2n = 2x Guzmania squarrosa (Mez & Sodiro) L.B. Sm. & Pittendr. B0607 0 1 1177 0 0 0 1 929 929 929 2n = 2x Guzmania strobilantha (Ruiz & Pav.) Mez B0504 0 1 1177 0 0 0 1 880 880 880 2n = 2x Guzmania tarapotina Ule B0591 7 2 1177 14 2 0 2 1038 ? ? 2n = 2x Guzmania variegata L.B. Sm. B0015 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. Guzmania variegata L.B. Sm. B0646 8 2 1177 0 0 0 1 1015 1015 1015 2n = 2x Guzmania virescens (Hook.) Mez B0718 5 2 1177 3 1 0 2 1035 1035 1029 2n = 2x Guzmania wittmackii (André) André ex Mez B0012 0 1 1177 0 0 0 1 1030 1030 1030 2n = 2x Guzmania xanthobractea Gilmartin B0623 0 1 1177 0 0 0 1 1033 1033 1033 2n = 2x Guzmania sp. B0628 1 2 1177 2 0 0 2 1043 1043 1043 2n = 2x Josemania anceps (Lodd.) Barfuss & W. Till B0741 0 1 1177 0 0 0 1 1048 1048 1048 2n = 2x Josemania anceps (Lodd.) Barfuss & W. Till B0781 0 1 1177 0 0 0 1 1048 1048 1048 2n = 2x Josemania cyanea (Linden ex K. Koch) Barfuss & W. Till B0780 2 2 1177 4 1 0 2 1049 1049 1048 2n = 2x Josemania cyanea (Linden ex K. Koch) Barfuss & W. Till B0782 5 2 1177 6 0 0 2 1049 1049 1049 2n = 2x Josemania Barfuss & W. Till Josemania lindenii (Regel) Barfuss & W. Till B0023 n.d n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. Josemania lindenii (Regel) Barfuss & W. Till B0746 0 1 1177 7 0 0 2 1049 1049 1049 2n = 2x Josemania pretiosa (Mez) Barfuss & W. Till B0231 3 2 1177 0 0 0 1 1049 1049 1049 2n = 2x Josemania umbellata (André) Barfuss & W. Till B0216 1 2 1177 9 1 1 2 1071 1071 1070 2n = 2x Josemania umbellata (André) Barfuss & W. Till B0758 5 2 1177 14 1 1 2 1071 1071 1070 2n = 2x Lemeltonia acosta-solisii (Gilmartin) Barfuss & W. Till B0626 0 1 1177 0 0 1 1 1103 1103 1103 2n = 2x Lemeltonia cornuta (Mez & Sodiro) Barfuss & W. Till B0744 0 1 1177 0 0 1 1 1103 1103 1103 2n = 2x Lemeltonia dodsonii (L.B. Sm.) Barfuss & W. Till B0016 3 2 1177 0 0 1 1 1137 1137 1137 2n = 2x Lemeltonia dodsonii (L.B. Sm.) Barfuss & W. Till B0721 1 2 1177 0 0 1 1 1116 1116 1116 2n = 2x Lemeltonia monadelpha (E. Morren) Barfuss & W. Till B0745 0 1 1177 0 0 1 1 1120 1120 1120 2n = 2x Lemeltonia scaligera (Mez & Sodiro) Barfuss & W. Till B0234 0 1 1177 9 0 1 2 1119 1119 1119 2n = 2x Lemeltonia scaligera (Mez & Sodiro) Barfuss & W. Till B0327 1 2 1177 4 0 1 2 1120 1120 1120 2n = 2x Lemeltonia triglochinoides (C. Presl) Barfuss & W. Till B0723 5 2 1177 7 0 1 2 1118 1118 1118 2n = 2x Lemeltonia Barfuss & W. Till 196 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 PHYC PRK postulated ploidy level SNPs Geno types Length Lemeltonia triglochinoides (C. Presl) Barfuss & W. Till B0724 3 2 1177 0 0 1 1 1126 1126 1126 2n = 2x Lemeltonia triglochinoides (C. Presl) Barfuss & W. Till B0725 3 2 1177 4 1 1 2 1125 ? ? 2n = 2x Racinaea hamaleana (E. Morren) Barfuss & W. Till B0251 0 1 1177 2 1 1 2 1122 1122 1121 2n = 2x Racinaea hamaleana (E. Morren) Barfuss & W. Till B0730 0 1 1177 6 1 1 2 1118 1118 1119 2n = 2x Indels SSR Consensus length Allele 2 length (a2) DNA no. SNPs Geno types Allele 1 length (a1) Taxon Racinaea M.A. Spencer & L.B. Sm. Racinaea subg. Pseudophytarrhiza Barfuss & W. Till Racinaea hamaleana (E. Morren) Barfuss & W. Till B0749 1 2 1177 8 1 1 2 1122 1122 1118 2n = 2x Racinaea venusta (Mez & Wercklé) Barfuss & W. Till B0007 0 1 1177 4 0 1 2 1131 1131 1131 2n = 2x Racinaea venusta (Mez & Wercklé) Barfuss & W. Till B0729 2 2 1177 2 2 1 2 1129 ? ? 2n = 2x Racinaea venusta (Mez & Wercklé) Barfuss & W. Till B0731 0 1 1177 2 2 1 2 1129 ? ? 2n = 2x Racinaea diffusa (L.B. Sm.) M.A. Spencer & L.B. Sm. B0480 2 2 1177 0 0 0 1 844 844 844 2n = 2x Racinaea dyeriana (André) Barfuss & W. Till B0151 0 1 1177 0 0 0 1 953 953 953 2n = 2x Racinaea subg. Racinaea Racinaea dyeriana (André) Barfuss & W. Till B0456 0 1 1177 4 0 0 2 953 953 953 2n = 2x Racinaea elegans (L.B. Sm.) M.A. Spencer & L.B. Sm. B0051 1 2 1177 0 0 0 1 881 881 881 2n = 2x Racinaea fraseri (Baker) M.A. Spencer & L.B. Sm. B0547 0 1 1177 0 0 1 1 1144 1144 1144 2n = 2x Racinaea fraseri (Baker) M.A. Spencer & L.B. Sm. B0616 0 1 1177 0 0 1 1 1140 1140 1140 2n = 2x Racinaea insularis (Mez) M.A. Spencer & L.B. Sm. B0494 0 1 1177 0 0 0 1 1139 1139 1139 2n = 2x B0478 1 2 1177 0 0 1 1 1141 1141 1141 2n = 2x B0426 0 1 1177 0 0 1 1 1141 1141 1141 2n = 2x Racinaea multiflora var. decipiens (André) M.A. Spencer & L.B. Sm. Racinaea multiflora (Benth.) M.A. Spencer & L.B. Sm. var. multiflora Racinaea multiflora var. tomensis (L.B. Sm.) M.A. Spencer & L.B. Sm. B0437 0 1 1177 7 0 0 2 1142 1142 1142 2n = 2x Racinaea pallidoflavens (Mez) M.A. Spencer & L.B. Sm. B0369 1 2 1177 5 0 0 2 838 838 838 2n = 2x Racinaea parviflora (Ruiz & Pav.) M.A. Spencer & L.B. Sm. B0335 4 2 1177 5 0 0 2 885 885 885 2n = 2x Racinaea pectinata (André) M.A. Spencer & L.B. Sm. B0520 0 1 1177 0 0 0 1 920 920 920 2n = 2x Racinaea pseudotetrantha (Gilmartin & H. Luther) J.R. Grant B0624 0 1 1177 0 0 0 1 870 870 870 2n = 2x Racinaea pugiformis (L.B. Sm.) M.A. Spencer & L.B. Sm. B0613 1 2 1177 0 0 0 1 902 902 902 2n = 2x Racinaea riocreuxii (André) M.A. Spencer & L.B. Sm. B0440 3 2 1177 3 0 0 2 885 885 885 2n = 2x Racinaea ropalocarpa (André) M.A. Spencer & L.B. Sm. B0057 5 2 1177 9 1 0 2 885 885 881 2n = 2x Racinaea seemannii (Baker) M.A. Spencer & L.B. Sm. B0024 1 2 1177 7 3 0 2 869 ? ? 2n = 2x Racinaea sinuosa (L.B. Sm.) M.A. Spencer & L.B. Sm. B0617 4 2 1177 1 0 0 2 940 940 940 2n = 2x Racinaea spiculosa (Griseb.) M.A. Spencer & L.B. Sm. B0099 0 1 1177 0 0 0 1 979 979 979 2n = 2x Racinaea spiculosa (Griseb.) M.A. Spencer & L.B. Sm. B0304 0 1 1177 0 0 1 1 1008 1008 1008 2n = 2x B0496 0 1 1177 0 0 0 1 870 870 870 2n = 2x B0448 0 1 1177 0 0 0 1 906 906 906 2n = 2x B0367 2 2 1177 5 1 0 2 906 906 904 2n = 2x B0572 0 1 1177 3 1 0 2 908 908 906 2n = 2x Rothowia laxissima (Mez) Barfuss & W. Till B0294 0 1 1177 0 0 1 1 1144 1144 1144 2n = 2x Rothowia laxissima (Mez) Barfuss & W. Till B0757 0 1 1177 0 0 1 1 1144 1144 1144 2n = 2x Rothowia platyrhachis (Mez) Barfuss & W. Till B0753 2 2 1177 10 3 1 2 n.u. 1149 1143 2n = 2x Rothowia wagneriana (L.B. Sm.) Barfuss & W. Till B0058 0 1 1177 0 0 1 1 1149 1149 1149 2n = 2x Rothowia wagneriana (L.B. Sm.) Barfuss & W. Till B0217 0 1 1177 0 0 1 1 1149 1149 1149 2n = 2x B0394 9 2 1177 6 0 1 2 1293 1293 1293 2n = 2x Racinaea tetrantha var. aurantiaca (Griseb.) M.A. Spencer & L.B. Sm. Racinaea tetrantha var. caribaea (L.B. Sm.) M.A. Spencer & L.B. Sm. Racinaea tetrantha (Ruiz & Pav.) M.A. Spencer & L.B. Sm. var. tetrantha Racinaea tilli Manzan. & Gouda Rothowia Barfuss & W. Till Tillandsia L. Tillandsia subg. Anoplophytum (Beer) Baker Tillandsia araujei Mez 197 BARFUSS, M.H.J. NEW CLASSIFICATION PHYC PRK postulated ploidy level SNPs Geno types Length Tillandsia aff. araujei Mez B0422 0 1 1177 2 0 1 2 1112 1112 1112 2n = 2x Tillandsia bergeri Mez B0097 7 2 1177 0 0 0 1 1058 1058 1058 2n = 2x Tillandsia bergeri Mez cv. Alba B0110 2 2 1177 0 0 1 1 1089 1089 1089 2n = 2x Indels SSR Consensus length Allele 2 length (a2) DNA no. SNPs Geno types Allele 1 length (a1) Taxon Tillandsia burle-marxii Ehlers B0393 0 1 1177 0 0 1 1 1133 1133 1133 2n = 2x Tillandsia ixioides Griseb. B0043 0 1 1177 1 0 1 2 1129 1129 1129 2n = 2x Tillandsia jonesii Strehl B0389 3 2 1177 13 1 1 2 1113 1113 1109 2n = 2x Tillandsia neglecta E. Pereira B0339 9 2 1177 3 0 1 2 1120 1120 1120 2n = 2x Tillandsia pohliana Mez B0080 0 1 1177 0 0 0 1 990 990 990 2n = 2x Tillandsia recurvifolia var. subsecundifolia (W. Weber & Ehlers) W. Till B0401 1 2 1177 4 0 1 2 1131 1131 1131 2n = 2x Tillandsia stricta var. albifolia H. Hrom. & Rauh B0410 1 2 1177 2 0 1 2 1113 1113 1113 2n = 2x Tillandsia stricta Sol. ex Sims var. stricta B0081 n.d. n.d. n.d. n.d. n.d. n.d. n.d n.d. n.d. n.d. n.d. Tillandsia tenuifolia L. B0026 5 2 1177 10 0 1 2 1112 1112 1112 2n = 2x Tillandsia aff. tenuifolia L. B0414 0 1 1177 2 0 1 2 1115 1115 1115 2n = 2x Tillandsia capillaris Ruiz & Pav. B0518 11 4 1180 13 2 1 2 1054 ? ? 2n = 4x! Tillandsia funebris A. Cast. B0089 1 2 1177 9 3 1 2 1141 1141 ? 2n = 2x Tillandsia kuehhasii W. Till B0396 1 2 1180 7 1 1 2 1053 1053 1051 2n = 2x Tillandsia landbeckii subsp. andina W. Till B0423 0 1 1180 0 0 1 1 1131 1131 1131 2n = 2x Tillandsia mollis H. Hrom. & W. Till B0206 4 2 1177 8 2 1 2 1133 ? ? 2n = 2x Tillandsia recurvata (L.) L. B0529 0 1 1177 0 0 0 1 1145 1145 1145 2n = 2x Tillandsia recurvata (L.) L. B0540 0 1 1177 0 0 1 1 1145 1145 1145 2n = 2x Tillandsia usneoides (L.) L. B0083 4 2 1177 3 1 1 2 1130 1130 1129 2n = 2x Tillandsia usneoides (L.) L. B0109 3 2 1177 8 1 1 2 1133 1134 1133 2n = 2x Tillandsia usneoides (L.) L. B0451 3 2 1177 9 1 1 2 1133 1133 1132 2n = 2x Tillandsia cf. usneoides (L.) L. B0539 0 1 1177 0 0 1 1 1128 1128 1128 2n = 2x Tillandsia virescens Ruiz & Pav. B0516 3 4 1180 9 1 1 2 1056 1056 1054 2n = 4x! Tillandsia duratii Vis. B0088 5 2 1180 7 0 0 2 1134 1134 1134 2n = 2x Tillandsia duratii Vis. B0737 4 2 1180 8 0 1 2 1134 1134 1134 2n = 2x Tillandsia duratii Vis. B0740 5 2 1180 9 1 1 2 1135 1135 1134 2n = 2x Tillandsia kirschnekii Rauh & W. Till B0384 1 2 1180 4 1 1 2 1134 1134 1133 2n = 2x Tillandsia subg. Diaphoranthema (Beer) Baker Tillandsia subg. Phytarrhiza (Vis.) Baker Tillandsia paleacea C. Presl B0404 4 2 1180 11 1 1 2 1137 1137 1134 2n = 2x Tillandsia aff. streptocarpa Baker B0226 2 2 1180 3 1 1 2 1135 1135 1134 2n = 2x Tillandsia grandis Schltdl. B0124 2 2 1177 2 0 1 2 1141 1141 1141 2n = 2x Tillandsia grandis Schltdl. B0125 1 2 1177 9 2 1 2 1148 1148 1147 2n = 2x Tillandsia macropetala Wawra B0742 2 2 1177 3 3 0 2 1142 1142 1136 2n = 2x Tillandsia macropetala Wawra B0748 2 2 1177 1 1 1 2 1139 1139 1136 2n = 2x Tillandsia viridiflora (Beer) Baker B0006 0 1 1177 0 1 1 2 1139 1139 1138 2n = 2x B0373 0 1 1177 0 0 0 1 1136 1136 1136 2n = 2x Tillandsia subg. Pseudalcantarea Mez Tillandsia subg. Pseudovriesea Barfuss & W. Till Tillandsia andreettae (Rauh) J.R. Grant Tillandsia appenii (Rauh) J.R. Grant B0066 0 1 1177 4 0 1 2 1133 1133 1133 2n = 2x Tillandsia appenii (Rauh) J.R. Grant B0464 3 2 1177 1 1 1 2 1115 1115 1114 2n = 2x Tillandsia arpocalyx André B0552 0 1 1177 1 1 1 2 1176 1176 1132 2n = 2x Tillandsia barclayana Baker B0028 n.d. n.d. n.d. n.d. n.d n.d. n.d. n.d. n.d. n.d. n.d. Tillandsia barclayana Baker B0142 3 2 1177 4 2 1 2 1137 1137 1134 2n = 2x 198 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 PHYC PRK postulated ploidy level SNPs Geno types Length Tillandsia barthlottii Rauh B0035 3 2 1177 0 1 1 2 1133 1133 1130 2n = 2x Tillandsia barthlottii Rauh B0716 1 2 1177 3 1 1 2 1134 1134 1133 2n = 2x Tillandsia cereicola Mez B0134 1 2 1177 7 1 1 2 1137 1138 1137 2n = 2x Tillandsia cereicola Mez B0476 0 1 1177 2 1 1 2 1142 1142 1136 2n = 2x Tillandsia espinosae L.B. Sm. B0143 0 1 1177 0 0 0 1 1135 1135 1135 2n = 2x Tillandsia espinosae L.B. Sm. B0462 0 1 1177 2 0 1 2 1136 1136 1136 2n = 2x Indels SSR Consensus length Allele 2 length (a2) DNA no. SNPs Geno types Allele 1 length (a1) Taxon Tillandsia fragrans André B0632 2 2 1177 10 1 0 2 1132 1132 1131 2n = 2x Tillandsia frank-hasei J.R. Grant B0315 0 1 1177 0 0 1 1 1131 1131 1131 2n = 2x Tillandsia hitchcockiana L.B. Sm. B0468 0 1 1177 2 0 0 2 1137 1137 1137 2n = 2x Tillandsia hitchcockiana L.B. Sm. B0629 2 2 1177 0 0 0 1 1137 1137 1137 2n = 2x Tillandsia myriantha Baker B0760 0 1 1177 0 0 0 1 1102 1102 1102 2n = 2x Tillandsia peruviana J.R. Grant B0470 0 1 1177 2 0 1 2 1136 1136 1136 2n = 2x Tillandsia petraea L.B. Sm. B0630 1 2 1177 0 0 0 1 1110 1110 1110 2n = 2x Tillandsia cf. porphyrocraspeda J.R. Grant B0350 0 1 1177 3 2 1 2 1132 1132 1132 2n = 2x Tillandsia spathacea Mez & Sodiro B0565 0 1 1177 0 0 0 1 1130 1130 1130 2n = 2x Tillandsia strobeliae (Rauh) J.R. Grant B0198 2 2 1177 4 1 1 2 1131 1131 1130 2n = 2x Tillandsia tequendamae André B0569 0 1 1177 0 0 0 1 1131 1131 1131 2n = 2x Tillandsia werneriana J.R. Grant B0067 0 1 1177 2 2 1 2 1137 1137 1135 2n = 2x Tillandsia cf. werneriana J.R. Grant B0144 2 2 1177 2 1 0 2 1149 1149 1148 2n = 2x Tillandsia cf. werneriana J.R. Grant B0514 3 2 1177 7 0 0 2 1136 1136 1136 2n = 2x Tillandsia achyrostachys E. Morren ex Baker B0408 6 2 1177 4 1 0 2 1103 1103 1099 2n = 2x Tillandsia adpressiflora Mez B0597 2 2 1180 0 0 1 1 1139 1139 1139 2n = 2x Tillandsia andrieuxii (Mez) L.B. Sm. B0063 1 2 1177 6 1 0 2 1098 1098 1080 2n = 2x Tillandsia baileyi Rose ex Small B0421 5 2 1180 8 1 0 2 1140 1140 1139 2n = 2x Tillandsia cf. beutelspacheri Matuda B0360 2 2 1177 9 2 0 2 1138 1138 1137 2n = 2x Tillandsia caput-medusae E. Morren B0046 0 1 1180 0 0 0 1 1142 1142 1142 2n = 2x Tillandsia aff. caput-medusae E. Morren B0424 5 2 1180 1 0 2 2 1162 1162 1162 2n = 2x Tillandsia subg. Tillandsia Tillandsia cryptopoda L.B. Sm. B0787 0 1 1177 0 0 1 1 1133 1133 1133 2n = 2x Tillandsia aff. cucaensis Wittm. B0732 15 2 1177 0 0 0 1 1135 1135 1135 2n = 2x Tillandsia durangensis Rauh & Ehlers B0365 6 2 1180 0 0 1 1 1140 1140 1140 2n = 2x Tillandsia ehlersiana Rauh B0431 1 2 1180 2 0 1 2 1140 1140 1140 2n = 2x Tillandsia exserta Fernald B0390 1 2 1177 5 0 0 2 1135 1135 1135 2n = 2x Tillandsia extensa Mez B0712 1 2 1177 9 1 1 2 1130 1131 1130 2n = 2x Tillandsia fasciculata Sw. B0076 2 2 1177 18 3 0 2 1136 1131! 1134! 2n = 2x Tillandsia filifolia Schltdl. & Cham. B0788 8 2 1177 9 3 1 2 1121 1121 1120 2n = 2x Tillandsia filifolia Schltdl. & Cham. B0790 5 2 1177 12 2 1 2 1123 1123 1121 2n = 2x Tillandsia fuchsii W. Till B0391 3 2 1177 11 1 0 2 1120 1120 1118 2n = 2x Tillandsia glabrior (L.B. Sm.) López-Ferr., Espejo & I. Ramírez B0340 5 2 1180 2 2 0 2 1138 1138 1137 2n = 2x Tillandsia guatemalensis L.B. Sm. B0008 4 2 1177 11 1 0 2 1129 1129 1128 2n = 2x Tillandsia guatemalensis L.B. Sm. B0103 3 2 1177 7 0 0 2 1128 1128 1128 2n = 2x Tillandsia guatemalensis L.B. Sm. B0104 8 2 1177 1 0 0 2 1129 1129 1129 2n = 2x Tillandsia guerreroensis Rauh B0349 6 2 1177 8 3 1 2 1137 ? ? 2n = 2x Tillandsia hildae Rauh B0763 2 2 1180 7 1 0 2 1130 1130 1116 2n = 2x Tillandsia huajuapanensis Ehlers & Lautner B0713 3 2 1177 6 1 0 2 1130 1130 1129 2n = 2x Tillandsia ionantha Planch. B0084 3 2 1180 7 0 2 2 1136 1136 1136 2n = 2x 199 BARFUSS, M.H.J. NEW CLASSIFICATION PHYC PRK postulated ploidy level SNPs Geno types Length Tillandsia juerg-rutschmannii Rauh B0715 1 2 1177 4 1 0 2 1137 1137 1136 2n = 2x Tillandsia leiboldiana Schltdl. B0323 4 2 1177 0 1 0 2 1116 1116 1106 2n = 2x Tillandsia malzinei (E. Morren) Baker B0145 0 1 1177 0 0 0 1 1125 1125 1125 2n = 2x Indels SSR Consensus length Allele 2 length (a2) DNA no. SNPs Geno types Allele 1 length (a1) Taxon Tillandsia malzinei (E. Morren) Baker B0492 0 1 1177 3 1 0 2 1125 1125 1110 2n = 2x Tillandsia mima var. chiletensis Rauh B0708 2 2 1177 1 0 1 2 1130 1130 1130 2n = 2x Tillandsia mima var. chiletensis Rauh B0710 2 2 1177 1 0 1 2 1130 1130 1130 2n = 2x Tillandsia cf. mima L.B. Sm. var. mima B0709 5 2 1177 0 0 0 1 1130 1130 1130 2n = 2x Tillandsia mitlaensis var. tulensis Lautner & Ehlers B0402 0 1 1177 8 0 1 2 1141 1141 1141 2n = 2x Tillandsia moronesensis Ehlers B0380 0 1 1177 1 0 0 2 1137 1137 1137 2n = 2x Tillandsia paniculata (L.) L. B0102 0 1 1177 0 0 1 1 1130 1130 1130 2n = 2x Tillandsia plagiotropica Rohweder B0386 0 1 1177 0 0 0 1 1116 1116 1116 2n = 2x Tillandsia praschekii Ehlers & Willinger B0364 0 1 1177 3 2 1 2 1140 1142 1140 2n = 2x Tillandsia propagulifera Rauh B0310 2 2 1177 8 2 1 2 1131 1130 1130 2n = 2x Tillandsia pseudosetacea Ehlers & Rauh B0333 0 1 1177 0 0 0 1 1135 1135 1135 2n = 2x Tillandsia pueblensis L.B. Sm. B0398 2 2 1180 8 2 1 2 1144 1146 1144 2n = 2x Tillandsia punctulata Schltdl.& Cham. B0061 0 1 1177 0 0 0 1 1138 1138 1138 2n = 2x Tillandsia remota Wittm. B0072 4 2 1177 4 0 0 2 1135 1135 1135 2n = 2x Tillandsia cf. rhomboidea André B0306 0 1 1177 0 0 0 1 1134 1134 1134 2n = 2x Tillandsia cf. rhomboidea André B0372 0 1 1177 0 0 1 1 1134 1134 1134 2n = 2x Tillandsia schatzlii Rauh B0416 0 1 1177 0 0 1 1 1144 1144 1144 2n = 2x Tillandsia secunda Kunth B0527 0 1 1177 3 0 1 2 1129 1129 1129 2n = 2x Tillandsia secunda Kunth B0528 0 1 1177 3 1 1 2 1130 1129 1128 2n = 2x Tillandsia secunda Kunth B0549 3 2 1177 2 1 1 2 1130 1130 1129 2n = 2x Tillandsia secunda Kunth B0551 3 2 1177 3 0 1 2 1129 1129 1129 2n = 2x Tillandsia seleriana Mez B0413 5 2 1180 0 0 1 1 1144 1144 1144 2n = 2x Tillandsia selleana Harms B0243 0 1 1177 0 0 0 1 1128 1128 1128 2n = 2x Tillandsia spiraliflora Rauh B0762 4 2 1177 8 0 1 2 1130 1130 1130 2n = 2x Tillandsia tricolor Schltdl. & Cham. B0785 3 2 1180 4 1 0 2 1131 1131 1114 2n = 2x Tillandsia utriculata L. B0100 0 1 1180 0 0 1 1 1143 1143 1143 2n = 2x Tillandsia utriculata L. B0805 0 1 1180 0 0 1 1 1143 1143 1143 2n = 2x Tillandsia utriculata L. B0807 0 1 1180 0 0 1 1 1140 1140 1140 2n = 2x Tillandsia aff. utriculata L. B0743 0 1 1180 0 0 1 1 1143 1143 1143 2n = 2x Tillandsia aff. utriculata L. B0747 0 1 1180 0 0 1 1 1143 1143 1143 2n = 2x Tillandsia variabilis Schltdl. B0308 0 1 1180 0 0 0 1 1129 1129 1129 2n = 2x Tillandsia velutina Ehlers B0427 0 1 1177 4 2 0 2 1123 ? ? 2n = 2x Tillandsia atroviridipetala Matuda B0215 0 1 1177 3 0 1 2 1082 1082 1082 2n = 2x Tillandsia balsasensis Rauh B0221 0 1 1177 0 0 1 1 1127 1127 1127 2n = 2x Tillandsia chusgonensis L. Hrom. B0210 5 2 1177 5 1 1 2 1085 1085 1084 2n = 2x Tillandsia heteromorpha Mez B0224 0 1 1177 0 0 0 1 1081 1081 1081 2n = 2x Tillandsia subg. Viridantha (Espejo) Barfuss & W. Till Tillandsia ignesiae Mez B0222 7 2 1177 7 1 1 2 1084 1084 1081 2n = 2x Tillandsia lepidosepala L.B. Sm. B0219 1 2 1177 0 0 1 1 1083 1083 1083 2n = 2x Tillandsia mauryana L.B. Sm. B0238 3 2 1177 2 2 1 2 1084 1084 1082 2n = 2x Tillandsia oblivata L. Hrom. B0205 2 2 1177 0 1 1 2 1081 1081 1067 2n = 2x Tillandsia plumosa Baker B0086 2 2 1177 2 1 1 2 1082 1082 1081 2n = 2x Tillandsia reducta L.B. Sm. B0209 0 1 1177 4 1 1 2 1084 1084 1080 2n = 2x 200 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 PHYC Taxon DNA no. SNPs Geno types PRK Length SNPs Indels SSR Geno types Consensus length Allele 1 length (a1) Allele 2 length (a2) postulated ploidy level Tillandsia rupicola Baker B0039 0 1 1177 0 0 0 1 1096 1096 1096 2n = 2x Tillandsia stellifera L. Hrom. B0239 0 1 1177 0 0 0 1 1081 1081 1081 2n = 2x Tillandsia tectorum var. globosa L. Hrom. B0248 0 1 1177 2 0 0 2 1094 1094 1094 2n = 2x Tillandsia tectorum var. tectorum f. gigantea L. Hrom. B0253 4 2 1177 2 0 0 2 1097 1097 1097 2n = 2x Tillandsia tectorum E. Morren var. tectorum f. tectorum B0249 0 1 1177 4 3 0 2 1131 1131 1127 2n = 2x Tillandsia tectorum var. viridula L. Hrom. B0245 0 1 1177 1 0 1 2 1127 1127 1127 2n = 2x Tillandsia tomekii L. Hrom. B0218 1 2 1177 6 1 1 2 1128 1128 1127 2n = 2x Tillandsia tortilis Klotzsch ex Baker B0049 2 2 1177 1 0 1 2 1087 1087 1087 2n = 2x Tillandsia baliophylla Harms B0101 0 1 1177 0 0 1 1 1108 1108 1108 2n = 2x Tillandsia biflora Ruiz & Pav. B0090 0 1 1177 0 0 1 1 1641 1641 1641 2n = 2x Tillandsia brenneri Rauh B0236 0 1 1180 0 0 1 1 1122 1122 1122 2n = 2x Tillandsia brevilingua Mez ex Harms B0056 0 1 1180 0 1 0 2 1124 1124 1123 2n = 2x Tillandsia brevilingua Mez ex Harms B0252 1 2 1180 0 1 0 2 1124 1124 1123 2n = 2x Tillandsia buseri var. nubicola Gilmartin B0561 0 1 1177 1 0 1 2 1101 1101 1101 2n = 2x Tillandsia cf. buseri var. nubicola Gilmartin B0441 1 2 1177 0 0 1 1 1124 1124 1124 2n = 2x Biflora clade Tillandsia cf. buseri var. nubicola Gilmartin B0442 4 2 1177 1 0 1 2 1117 1117 1117 2n = 2x Tillandsia complanata Benth. B0244 4 2 1177 1 1 0 2 1131 1131 1129 2n = 2x Tillandsia complanata Benth. B0562 1 2 1177 8 1 0 2 1133 1133 1129 2n = 2x Tillandsia complanata Benth. B0563 3 2 1177 6 1 0 2 1129 1129 1128 2n = 2x Tillandsia confertiflora André B0614 4 2 1177 0 0 1 1 1119 1119 1119 2n = 2x Tillandsia confinis L.B. Sm. B0587 2 2 1177 0 2 1 2 1134 1134 1127 2n = 2x Tillandsia demissa L.B. Sm. B0075 3 2 1177 0 0 0 1 1124 1124 1124 2n = 2x Tillandsia denudata var. vivipara Rauh B0750 2 2 1177 1 0 0 2 1135 1135 1135 2n = 2x Tillandsia deppeana Steud. B0457 0 1 1177 3 0 0 2 1111 1111 1111 2n = 2x Tillandsia deppeana Steud. B0458 0 1 1177 2 1 0 2 1111 1111 1109 2n = 2x Tillandsia fendleri Griseb. B0009 0 1 1177 0 2 0 2 1116 1116 1114 2n = 2x Tillandsia fendleri Griseb. B0445 0 1 1177 0 0 1 1 1117 1117 1117 2n = 2x Tillandsia floribunda Kunth B0351 1 2 1180 4 1 1 2 1068 1068 1067 2n = 2x Tillandsia fosteri Gilmartin B0639 3 2 1177 12 2 1 2 1123 1123 1117 2n = 2x Tillandsia glauca L.B. Sm. B0618 1 2 1177 0 0 1 1 1117 1117 1117 2n = 2x Tillandsia heterophylla E. Morren B0047 0 1 1177 0 0 0 1 1110 1110 1110 2n = 2x Tillandsia imperialis E. Morren ex Roezl B0292 1 2 1177 8 1 1 2 1113 1113 1111 2n = 2x Tillandsia incarnata Kunth B0522 0 1 1177 0 0 1 1 1113 1113 1113 2n = 2x Tillandsia ionochroma André ex Mez B0600 0 1 1177 0 0 1 1 1692 1692 1692 2n = 2x Tillandsia kauffmannii Ehlers B0074 0 1 1177 0 0 1 1 1114 1114 1114 2n = 2x Tillandsia krahnii Rauh B0235 0 1 1177 0 0 0 1 1120 1120 1120 2n = 2x Tillandsia lajensis André B0242 0 1 1177 1 0 1 2 1120 1120 1120 2n = 2x Tillandsia lajensis André B0546 0 1 1177 0 0 1 1 1120 1120 1120 2n = 2x Tillandsia lajensis André B0554 0 1 1177 1 0 1 2 1120 1120 1120 2n = 2x Tillandsia lajensis André B0555 0 1 1177 0 0 1 1 1120 1120 1120 2n = 2x Tillandsia latifolia var. divaricata (Benth.) Mez B0068 2 2 1177 7 2 0 2 1115 1116! 1100! 2n = 2x Tillandsia latifolia Meyen var. latifolia B0433 3 2 1177 4 1 1 2 1115 1115 1098 2n = 2x Tillandsia latifolia var. leucophylla Rauh B0228 1 2 1177 6 1 1 2 1115 1115 1098 2n = 2x Tillandsia macbrideana L.B. Sm. B0070 2 2 1177 13 2 1 2 1113 ? ? 2n = 2x Tillandsia maculata Ruiz & Pav. B0574 4 2 1177 1 1 0 2 1124 1124 1116 2n = 2x 201 BARFUSS, M.H.J. NEW CLASSIFICATION PHYC Taxon DNA no. SNPs Geno types PRK Length SNPs Indels SSR Geno types Consensus length Allele 1 length (a1) Allele 2 length (a2) postulated ploidy level Tillandsia multicaulis Steud. B0107 0 1 1177 0 0 0 1 1115 1115 1115 2n = 2x Tillandsia orbicularis L.B. Sm. B0375 0 1 1177 0 0 1 1 1085 1085 1085 2n = 2x Tillandsia pastensis André B0521 0 1 1177 0 0 0 1 1095 1095 1095 2n = 2x Tillandsia pseudomacbrideana Rauh B0036 3 2 1177 8 1 1 2 1129 1129 1116 2n = 2x Tillandsia roezlii E. Morren B0764 1 2 1177 1 0 1 2 1115 1115 1115 2n = 2x Tillandsia rudolfii E. Gross & Hase B0654 0 1 1177 7 2 1 2 1113 1113 1113 2n = 2x Tillandsia somnians L.B. Sm. B0301 1 2 1177 8 0 1 2 1100 1100 1100 2n = 2x Tillandsia somnians L.B. Sm. B0312 0 1 1177 0 0 1 1 1100 1100 1100 2n = 2x Tillandsia stenoura Harms B0635 0 1 1177 10 2 1 2 1114 1114 1109 2n = 2x Tillandsia subconcolor L.B. Sm. B0227 0 1 1177 0 0 1 1 1115 1115 1115 2n = 2x Tillandsia cf. superba Mez & Sodiro B0602 3 2 1177 0 0 1 1 1101 1101 1101 2n = 2x Tillandsia tovarensis Mez B0605 1 2 1177 0 0 0 1 1104 1104 1104 2n = 2x Tillandsia turneri Baker B0650 0 1 1177 0 0 1 1 1081 1081 1081 2n = 2x Tillandsia sp. B0634 6 2 1177 8 0 1 2 1088 1088 1088 2n = 2x Tillandsia brachyphylla Baker B0082 2 2 1177 2 0 1 2 1097 1097 1097 2n = 2x Tillandsia gardneri Lindl. B0041 0 1 1177 0 0 1 1 1092 1092 1092 2n = 2x Tillandsia globosa Wawra B0419 5 2 1177 10 2 0 2 1084 ? ? 2n = 2x Gardneri clade Purpurea clade Tillandsia aurea Mez B0250 0 1 1177 6 0 0 2 1096 1096 1096 2n = 2x Tillandsia cacticola L.B. Sm. B0044 0 1 1177 1 0 0 2 1096 1096 1096 2n = 2x Tillandsia purpurea Ruiz & Pav. B0246 0 1 1177 0 0 0 1 1095 1095 1095 2n = 2x Tillandsia straminea Kunth B0247 3 2 1177 0 0 0 1 1096 1096 1096 2n = 2x Tillandsia ecarinata L.B. Sm. B0237 1 2 1180 5 1 1 2 1134 1134 1132 2n = 2x Tillandsia ferreyrae L.B. Sm. B0241 1 2 1177 2 2 1 2 1132 1132 1130 2n = 2x Tillandsia rauhii L.B. Sm. B0092 0 1 1177 0 0 0 1 1128 1128 1128 2n = 2x Tillandsia teres L.B. Sm. B0201 3 2 1177 8 0 0 2 1128 1128 1128 2n = 2x Tillandsia aff. arequitae (André) André ex Mez B0387 0 1 1177 0 0 1 1 1132 1132 1132 2n = 2x Tillandsia argentina C.H. Wright B0087 2 2 1177 9 3 1 2 1129 1127! 1125! 2n = 2x Tillandsia argentina C.H. Wright B0797 3 2 1177 4 0 1 2 1129 1129 1129 2n = 2x Tillandsia bermejoensis H. Hrom. B0034 1 2 1177 0 0 1 1 1130 1130 1130 2n = 2x Tillandsia buchlohii Rauh B0395 5 2 1177 4 0 1 2 1129 1129 1129 2n = 2x Tillandsia cardenasii L.B. Sm. B0409 0 1 1177 4 1 1 2 1138 1138 1136 2n = 2x Tillandsia caulescens Brongn. ex Baker B0071 0 1 1177 0 0 1 1 1130 1130 1130 2n = 2x Tillandsia cochabambae E. Gross & Rauh B0392 6 2 1177 8 1 1 2 1132 1132 1131 2n = 2x Tillandsia comarapaensis H. Luther B0407 5 2 1177 2 0 1 2 1129 1129 1129 2n = 2x Tillandsia diaguitensis A. Cast. B0429 7 2 1177 8 0 1 2 1129 1129 1129 2n = 2x Tillandsia didisticha (E. Morren) Baker B0038 1 2 1177 0 0 1 1 1130 1130 1130 2n = 2x Rauhii clade Xiphioides clade Tillandsia gerdae Ehlers B0411 6 2 1177 8 3 1 2 1131 ? ? 2n = 2x Tillandsia koehresiana Ehlers B0371 1 2 1177 5 2 1 2 1131 1131 1127 2n = 2x Tillandsia lorentziana Griseb. B0374 0 1 1177 4 3 1 2 1139 ? ? 2n = 2x Tillandsia aff. lotteae H. Hrom. B0808 0 1 1177 0 0 1 1 1131 1131 1131 2n = 2x Tillandsia aff. lotteae H. Hrom. B0809 0 1 1177 0 0 1 1 1131 1131 1131 2n = 2x Tillandsia markusii L. Hrom. B0399 1 2 1177 7 3 1 2 1129 1129 1126 2n = 2x 202 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 PHYC PRK postulated ploidy level SNPs Geno types Length Tillandsia muhriae W. Weber B0438 0 1 1177 6 0 1 2 1130 1130 1130 2n = 2x Tillandsia ramellae W. Till & S. Till B0430 0 1 1177 0 2 1 2 1138 ? ? 2n = 2x Tillandsia rosacea L. Hrom. & W. Till B0405 4 2 1177 1 0 1 2 1128 1128 1128 2n = 2x Indels SSR Consensus length Allele 2 length (a2) DNA no. SNPs Geno types Allele 1 length (a1) Taxon Tillandsia rosarioae L. Hrom. B0403 5 2 1177 5 2 1 2 1129 1129 1127 2n = 2x Tillandsia aff. vernicosa Baker B0355 0 1 1177 11 0 1 2 1129 1129 1129 2n = 2x Tillandsia xiphioides subsp. prolata H. Luther B0418 1 2 1177 6 2 1 2 1130 ? ? 2n = 2x Tillandsia xiphioides Ker Gawl. subsp. xiphioides B0040 3 2 1177 8 1 1 2 1130 1130 1129 2n = 2x Tillandsia yuncharaensis W. Till B0417 0 1 1177 7 0 1 2 1129 1129 1129 2n = 2x Tillandsia zecheri var. cafayatensis Palací & G.K. Br. B0361 5 2 1177 11 3 1 2 1138 ? ? 2n = 2x Tillandsia albertiana F. Verv. B0033 2 2 1177 0 0 1 1 1138 1138 1138 2n = 2x Tillandsia albertiana F. Verv. B0212 0 1 1177 0 0 1 1 1136 1136 1136 2n = 2x Tillandsia albertiana F. Verv. B0225 0 1 1177 4 0 1 2 1141 1141 1141 2n = 2x Tillandsia australis Mez B0203 0 1 1177 0 0 1 1 1115 1115 1115 2n = 2x Tillandsia australis Mez B0759 0 1 1177 0 0 1 1 1115 1115 1115 2n = 2x Tillandsia disticha Kunth B0048 0 1 1177 2 1 1 2 1131 1131 1130 2n = 2x Tillandsia disticha Kunth B0202 3 2 1177 1 1 1 2 1130 1131 1130 2n = 2x Tillandsia disticha Kunth B0233 2 2 1177 0 0 1 1 1134 1134 1134 2n = 2x Tillandsia edithae Rauh B0425 2 2 1177 2 0 1 2 1136 1136 1136 2n = 2x Incertae sedis Tillandsia edithae Rauh B0436 0 1 1177 0 0 1 1 1136 1136 1136 2n = 2x Tillandsia esseriana Rauh & L.B. Sm. B0069 0 1 1177 7 2 1 2 1157 1157! 1128! 2n = 2x Tillandsia esseriana Rauh & L.B. Sm. B0200 3 2 1177 0 0 1 1 1129 1129 1129 2n = 2x Tillandsia nana Baker B0343 0 1 1177 1 1 1 2 1133 1133 1132 2n = 2x Tillandsia pseudomicans Rauh B0347 3 2 1177 3 0 1 2 1115 1115 1115 2n = 2x Tillandsia sphaerocephala Baker B0366 1 2 1177 2 1 0 2 1135 1135 1134 2n = 2x B0318 1 2 1177 0 0 0 1 903 903 903 2n = 2x Bromeliaceae tribe Vrieseeae W. Till & Barfuss Bromeliaceae subtribe Cipuropsidinae Barfuss & W. Till Cipuropsis Ule Cipuropsis amicorum (I. Ramírez & Bevil.) Barfuss & W. Till Cipuropsis zamorensis (L.B. Sm.) Barfuss & W. Till B0045 0 1 1177 4 0 0 2 915 915 915 2n = 2x Cipuropsis cv. Elan B0117 5 2 1177 7 1 0 2 918 918 915 2n = 2x Mezobromelia bicolor L.B. Sm. B0578 3 2 1177 3 2 0 2 902 ? ? 2n = 2x Mezobromelia bicolor L.B. Sm. B0583 1 2 1177 6 0 0 2 901 901 901 2n = 2x Mezobromelia pleiosticha (Griseb.) Utley & H. Luther B0274 0 1 1177 0 0 0 1 923 923 923 2n = 2x Mezobromelia L.B. Sm. Mezobromelia pleiosticha (Griseb.) Utley & H. Luther Mezobromelia cf. pleiosticha (Griseb.) Utley & H. Luther Mezobromelia schimperiana (Wittm.) Barfuss & Manzan. Mezobromelia schimperiana (Wittm.) Barfuss & Manzan. Mezobromelia schimperiana (Wittm.) Barfuss & Manzan. Mezobromelia schimperiana (Wittm.) Barfuss & Manzan. Mezobromelia aff. schimperiana (Wittm.) Barfuss & Manzan. Mezobromelia sp. B0649 0 1 1177 0 0 0 1 923 923 923 2n = 2x B0330 5 2 1177 5 1 0 2 901 901 899 2n = 2x B0334 2 2 1177 4 1 0 2 901 901 899 2n = 2x B0589 1 2 1177 8 0 0 2 899 899 899 2n = 2x B0611 2 2 1177 3 1 0 2 901 901 899 2n = 2x B0612 0 1 1177 6 1 0 2 901 901 899 2n = 2x B0197 0 1 1177 0 0 0 1 886 886 886 2n = 2x B0194 1 2 1177 5 0 0 2 901 901 901 2n = 2x B0132 4 2 1177 7 0 0 2 1015 1015 1015 2n = 2x Werauhia J.R. Grant Werauhia ampla (L.B. Sm.) J.R. Grant 203 BARFUSS, M.H.J. NEW CLASSIFICATION PHYC PRK postulated ploidy level SNPs Geno types Length Werauhia gladioliflora (H. Wendl.) J.R. Grant B0149 0 1 1177 0 0 0 1 1026 1026 1026 2n = 2x Werauhia gladioliflora (H. Wendl.) J.R. Grant B0509 0 1 1177 0 0 0 1 1026 1026 1026 2n = 2x Werauhia guadelupensis (Baker) J.R. Grant B0313 0 1 1177 0 0 0 1 1079 1079 1079 2n = 2x Indels SSR Consensus length Allele 2 length (a2) DNA no. SNPs Geno types Allele 1 length (a1) Taxon Werauhia haltonii (H. Luther) J.R. Grant B0580 0 1 1180 0 0 0 1 975 975 975 2n = 2x Werauhia hygrometrica (André) J.R. Grant B0598 0 1 1180 0 0 0 1 975 975 975 2n = 2x Werauhia hygrometrica (André) J.R. Grant B0621 0 1 1180 0 0 0 1 975 975 975 2n = 2x Werauhia insignis (Mez) W. Till, Barfuss & R. Samuel B0017 0 1 1159 0 0 0 1 1031 1031 1031 2n = 2x Werauhia ororiensis (Mez) J.R. Grant B0025 2 2 1180 9 2 0 2 962 962 954 2n = 2x Werauhia pedicellata (Mez & Wercklé) J.R. Grant B0752 0 1 1180 0 0 0 1 1016 1016 1016 2n = 2x Werauhia ringens (Griseb.) J.R. Grant B0019 0 1 1177 0 0 0 1 1045 1045 1045 2n = 2x Werauhia sintenisii (Baker) J.R. Grant B0450 0 1 1180 0 0 0 1 1031 1031 1031 2n = 2x Werauhia tarmaensis (Rauh) J.R. Grant B0018 0 1 1177 0 0 0 1 1081 1081 1081 2n = 2x Werauhia viridiflora (Regel) J.R. Grant B0285 0 1 1177 0 0 0 1 908 908 908 2n = 2x Werauhia viridiflora (Regel) J.R. Grant B0325 7 2 1177 3 0 0 2 908 908 908 2n = 2x Werauhia werckleana (Mez) J.R. Grant B0497 3 2 1177 8 0 0 2 1083 1083 1083 2n = 2x Vriesea chrysostachys E. Morren B0005 0 1 1177 0 0 0 1 1029 1029 1029 2n = 2x Vriesea ospinae H. Luther B0054 0 1 1177 4 1 0 2 1036 1036 1035 2n = 2x Tillandsia asplundii L.B. Sm. B0588 0 1 1177 2 2 0 2 921 921 918 2n = 2x Tillandsia singularis Mez & Wercklé B0064 0 1 1180 0 0 0 1 914 914 914 2n = 2x Chrysostachys clade Singularis clade Splendens clade Vriesea glutinosa Lindl. B0130 0 1 1177 0 0 0 1 1067 1067 1067 2n = 2x Vriesea splendens (Brongn.) Lem. B0037 0 1 1177 0 0 1 1 1069 1069 1069 2n = 2x B0148 0 1 1177 1 0 0 2 1090 1090 1090 2n = 2x Alcantarea benzingii Leme B0166 0 1 1177 0 0 0 1 1125 1125 1125 2n = 2x Alcantarea burle-marxii (Leme) J.R. Grant B0159 2 2 1177 11 1 1 2 1134 1134 1133 2n = 2x Alcantarea duarteana (L.B. Sm.) J.R. Grant B0059 0 1 1177 4 0 0 2 1133 1133 1133 2n = 2x Alcantarea duarteana (L.B. Sm.) J.R. Grant B0165 0 1 1177 3 1 0 2 1133 1133 1132 2n = 2x Alcantarea edmundoi (Leme) J.R. Grant B0161 2 2 1177 1 0 1 2 1134 1134 1134 2n = 2x Alcantarea edmundoi (Leme) J.R. Grant B0171 0 1 1177 0 0 0 1 1135 1135 1135 2n = 2x Alcantarea aff. extensa (L.B. Sm.) J.R. Grant B0164 4 2 1177 7 1 1 2 1138 1138 1133 2n = 2x Alcantarea farneyi (Martinelli & A.F. Costa) J.R. Grant B0170 0 1 1177 4 0 1 2 1135 1135 1135 2n = 2x Alcantarea geniculata (Wawra) J.R. Grant B0153 2 2 1177 0 0 0 1 1133 1133 1133 2n = 2x Alcantarea cf. geniculata (Wawra) J.R. Grant B0751 0 1 1177 0 0 1 1 1134 1134 1134 2n = 2x Alcantarea glazioviana (Lem.) Leme B0167 0 1 1177 10 2 0 2 1134 ? ? 2n = 2x Alcantarea heloisae J.R. Grant B0155 9 2 1177 4 2 0 2 1135 ? ? 2n = 2x Alcantarea imperialis (Carrière) Harms B0001 2 2 1177 5 1 1 2 1135 1134! 1135! 2n = 2x Alcantarea imperialis (Carrière) Harms B0135 0 1 1177 0 0 1 1 1135 1135 1135 2n = 2x Alcantarea imperialis (Carrière) Harms B0154 4 2 1177 0 2 1 2 1137 ? ? 2n = 2x Incertae sedis Vriesea tuerckheimii (Mez) L.B. Sm. Bromeliaceae subtribe Vrieseinae W. Till & Barfuss Alcantarea (E. Morren ex Mez) Harms Alcantarea nahoumii (Leme) J.R. Grant B0156 0 1 1177 0 0 1 1 1133 1133 1133 2n = 2x Alcantarea nevaresii Leme B0163 1 2 1177 0 0 0 1 1134 1134 1134 2n = 2x Alcantarea nevaresii Leme B0169 1 2 1177 2 0 1 2 1134 1134 1134 2n = 2x 204 OF TILLANDSIOIDEAE (BROMELIACEAE) PART 3, CHAPTER 5 PHYC Taxon DNA no. SNPs Geno types PRK Length SNPs Indels SSR Geno types Consensus length Allele 1 length (a1) Allele 2 length (a2) postulated ploidy level Alcantarea odorata (Leme) J.R. Grant B0157 0 1 1177 0 0 0 1 1133 1133 1133 2n = 2x Alcantarea regina (Vell.) Harms B0136 0 1 1177 0 0 0 1 1134 1134 1134 2n = 2x Alcantarea roberto-kautskyi Leme B0158 2 2 1177 2 1 1 2 1140 1140 1138 2n = 2x Alcantarea sp. B0160 0 1 1177 1 1 1 2 1134 1134 1133 2n = 2x Alcantarea sp. B0168 5 2 1177 0 0 1 1 1133 1133 1133 2n = 2x Vriesea correia-araujoi E. Pereira & I.A. Penna B0095 0 1 1177 12 3 0 2 1118 1118 1116 2n = 2x Vriesea croceana Leme & G.K. Br. B0802 1 2 1177 7 0 0 2 1134 1134 1134 2n = 2x Vriesea aff. croceana Leme & G.K. Br. B0801 0 1 1177 0 0 1 1 1133 1133 1133 2n = 2x Vriesea Lindl. Vriesea erythrodactylon (E. Morren) E. Morren ex Mez B0309 0 1 1177 0 0 0 1 1133 1133 1133 2n = 2x Vriesea flava A.F. Costa, H. Luther & Wand. B0322 4 2 1180 0 0 0 1 999 999 999 2n = 2x Vriesea jonghei (K. Koch) E. Morren B0065 2 2 1177 3 2 1 2 1115 1115 1112 2n = 2x Vriesea longicaulis (Baker) Mez B0290 1 2 1177 7 2 1 2 1117 1117 1115 2n = 2x Vriesea maxoniana (L.B. Sm.) L.B. Sm. B0490 0 1 1180 0 0 0 1 1135 1135 1135 2n = 2x Vriesea neoglutinosa Mez B0311 7 2 1177 8 0 0 2 1121 1121 1121 2n = 2x Vriesea pabstii McWill. & L.B. Sm. B0357 0 1 1177 0 0 0 1 1115 1115 1115 2n = 2x Vriesea platynema Gaudich. B0146 0 1 1183 0 0 0 1 1135 1135 1135 2n = 2x Vriesea plurifolia Leme B0376 0 1 1177 0 0 1 1 1132 1132 1132 2n = 2x Vriesea procera var. tenuis L.B. Sm. B0296 0 1 1180 0 0 0 1 1138 1138 1138 2n = 2x Vriesea psittacina (Hook.) Lindl. B0020 6 2 1177 7 5 0 2 n.u. 1067! 999! 2n = 2x Vriesea roethii W. Weber B0511 1 2 1177 2 0 1 2 1117 1117 1117 2n = 2x Vriesea ruschii L.B. Sm. B0719 5 2 1177 11 0 0 2 1116 1116 1116 2n = 2x Vriesea saundersii (Carrière) E. Morren ex Mez B0004 0 1 1177 0 0 0 1 1133 1133 1133 2n = 2x Vriesea scalaris E. Morren B0147 0 1 1177 0 0 0 1 1027 1027 1027 2n = 2x Vriesea simplex (Vell.) Beer B0505 4 2 1177 5 1 0 2 1027 1027 1026 2n = 2x Vriesea sparsiflora L.B. Sm. B0484 2 2 1177 0 0 0 1 1115 1115 1115 2n = 2x Vriesea sucrei L.B. Sm. & Read B0302 1 2 1177 0 0 0 1 1024 1024 1024 2n = 2x Vriesea unilateralis (Baker) Mez B0316 2 2 1177 4 0 0 2 1117 1117 1117 2n = 2x 205 Appendix Abstracts 2005–2012 (Conferences) Oral presentations and posters Botany 2012 Adaptive radiation, correlated evolution, and determinants of net diversification rates in Bromeliaceae: Test of an a priori model Thomas J. Givnish 1, Michael H.J. Barfuss 2, Benjamin Van Ee 3, Ricarda Riina 4, Katharina Schulte 5, Ralf Horres 6, Philip A. Gonsiska 1, Rachel S. Jabaily 1, Darren M. Crayn 5, J. Andrew C. Smith 7, Klaus Winter 8, Gregory K. Brown 9, Timothy M. Evans 10, Bruce K. Holst 11, Harry E. Luther 11, Walter Till 2, Georg Zizka 12, Paul E. Berry 4 & Kenneth .J. Sytsma 1 1 University of Wisconsin-Madison, Department of Botany, 430 Lincoln Drive, Birge Hall, Madison, WI, 53706, USA; 2 University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria; 3 Black Hills State University, Spearfish, USA; 4 University of Michigan Herbarium, Ecology and Evolutionary Biology, 3600 Varsity Dr, Ann Arbor, Michigan, 48108, USA; 5 James Cook University, Australian Tropical Herbarium, Cairns, QLD, 4870, Australia; 6 GenXPro GmbH, Altenhöferallee 3 60438, Frankfurt/Main, Germany; 7 University of Oxford, Department of Plant Sciences, South Parks Road, Oxford, OX1 3RB, United Kingdom; 8 Smithsonian Tropical Research Institute, P.O. Box 2072, Balboa, Ancón, Republic of Panama; 9 University of Wyoming, Department of Botany, 3165, 1000 E University Avenue, Laramie, Wyoming, 82071, USA; 10 Grand Valley State University, Biology Department, 1 Campus Drive, Allendale, Michigan, 49401, USA; 11 Marie Selby Botanic Gardens, 811 South Palm Avenue, Sarasota, Florida, 34236-7726, USA; 12 Research Institute Senckenberg, Department of Botany, Frankfurt/Main, Germany Author for correspondence: Thomas J. Givnish, givnish@wisc.edu Oral presentation: (638) in: Botany 2012: The Next Generation (abstracts). Columbus, Ohio: Greater Columbus Convention Center, 7–11 July 2012. Abstract We present an integrative model predicting associations among epiphytism, the tank habit, entangling seeds, C3 vs. CAM photosynthesis, avian pollinators, life in fertile, moist montane habitats, and species richness and net rates of species diversification in the monocot family Bromeliaceae. We test the predictions of this model by overlaying individual character-states on a molecular phylogeny, relating evolutionary shifts to time and reconstructed shifts in geographic distribution; by quantifying patterns of correlated and contingent evolution among pairs of traits; and by analyzing the apparent impact of individual traits on diversification within subfamilies. All patterns of correlated evolution among pairs of traits and environmental conditions predicted by our model were significant. The pattern and timing of shifts in phenotype and expansion of distributions outside the Guayana Shield also generally accorded with the model's predictions. Patterns of contingent evolution were largely consistent with the model. Species richness and net rates of species diversification were most closely tied to life in fertile, moist, geographically extensive cordilleras, with weaker ties to epiphytism, avian pollination, and the tank habit. The highest rates of diversification were seen in the core tillandsioids, associated with the Andes, and especially the tank-epiphyte clade of bromelioids, associated with the Serra do Mar and nearby ranges. Six adaptive radiations and associated speciation - including one clade of CAM epiphytes, one of predominantly C3 epiphytes, three of CAM terrestrials in arid habitats (in Central America, high elevations in the Andes and low elevations in the Brazilian Shield), and one of C3 terrestrials in rain- and cloud-forest understories - account for > 80% of total species number in the family. This integrative study is among the first to test a priori hypotheses about the relationships among phylogeny, phenotypic evolution, geographic spread, and net species diversification. 210 XVIII IBC 2011 Systematics, evolution, and phylogeography of Tillandsia (Bromeliaceae) and related genera Michael H.J. Barfuss, Walter Till & Rosabelle Samuel University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria Author for correspondence: Michael H.J. Barfuss, michael.h.j.barfuss@univie.ac.at Oral presentation: P. 189 (SYM004) in: XVIII IBC 2011: XVIII International Botanical Congress (abstracts). Melbourne, Australia: Melbourne Congress and Exhibition Centre, 23–30 July 2011. Abstract Tillandsia L. is the largest genus within Bromeliaceae comprising more than 660 accepted species in six recognized subgenera (Tillandsia subgen. Allardtia, T. subgen. Anoplophytum, T. subgen. Diaphoranthema, T. subgen. Phytarrhiza, T. subgen. Pseudalcantarea, T. subgen. Tillandsia). Eight additional genera have been associated with Tillandsia in subfamily Tillandsioideae, i.e., Alcantarea, Catopsis, Glomeropitcairnia, Guzmania, Mezobromelia, Racinaea, Vriesea, and Werauhia. The segregation of a group of Tillandsia species occurring in Mexico (Viridantha) and the transfer of xerophytic Vriesea species to Tillandsia is not accepted by all bromeliad researchers based on long-established morphological characterizations of traditionally accepted genera. Habit shifts from phytotelms to extreme xerophytes, life form transitions from terrestrials to epiphytes and lithophytes, shifts in pollination syndromes, and seed and stigma morphology are key events and characters for the evolution and systematics of genera and infrageneric units, but they can only be interpreted in the context of DNA sequence data, since some of these features have evolved independently within different phylogenetic lineages. Therefore we conducted DNA sequence analyses from nuclear genes (i.e., PRK and PHYC, totaling about 2500 bp) and published chloroplast markers (with about 6500 bp) in combination with a reevaluation of morphological characters. Based on these results three new genera with a distinctive stigma and seed morphology are segregated from Tillandsia, i.e., Josemania, Lemeltonia, and Rothowia. Species of all three belonged to the former subgenus Phytarrhiza, which turned out to be highly polyphyletic. Circumscription of the remaining Tillandsia subgenera also changes significantly and new infrageneric taxa and groups emerge (e.g., T. subgen. Pseudovriesea, T. subgen. Viridantha). DNA data suggest that the subfamily had its origin in the geologically old parts of northern South America. From there the two earliest diverging lineages migrated into the Caribbean (Catopsis, Glomeropitcairnia) with some taxa extending also into the Andes and into Central America (Catopsis). The next diverging clade (Vrieseeae) splits into two subgroups, one radiating into eastern Brazil (Alcantarea, Vriesea), the other spreading into the Andes (Cipuropsis, Werauhia). Within Tillandsieae the earliest lineage is Guzmania with a predominately Andean distribution, but also extending into Central America and the Caribbean. Early diverging taxa of the core Tillandsieae have their current distribution mainly in the Andes of northern Peru, Ecuador and Colombia, which seems to be the ancient area for the whole Tillandsia s.l. complex. Tillandsia s.str. has two centers of diversity, one in the Northern and Central Andes, the other in the mountain systems of Northern Central America, which proves to have reached Mexico and adjacent areas during at least three independent colonization events from the South (T. subgen. Pseudalcantarea, T. subgen. Tillandsia p.p., and T. subgen. Viridantha p.p.). 212 XVIII IBC 2011 New taxonomic implications in Tillandsioideae (Bromeliaceae) based on DNA data and morphology Michael H.J. Barfuss, Walter Till & Rosabelle Samuel University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria Author for correspondence: Michael H.J. Barfuss, michael.h.j.barfuss@univie.ac.at Poster: Pp. 568–569 (P0730) in: XVIII IBC 2011: XVIII International Botanical Congress (abstracts). Melbourne, Australia: Melbourne Congress and Exhibition Centre, 23–30 July 2011. Abstract In order to establish a natural and stable classification system for Tillandsioideae we conducted phylogenetic analyses using newly generated single-copy nDNA sequences (i.e., PHYC and PRK) and published cpDNA data (i.e., atpB-rbcL spacer, matK, rbcL, partial rbcL-accD spacer, rps16 intron, partial trnK intron, trnL intron and trnL-trnF spacer). Derived phylogenetic units were then characterized using previously used, neglected, and new morphological characters. A new classification for Tillandsioideae is urgently needed, since modern bromeliad taxonomists rely on a more than 30 years old monograph. Subsequent taxonomic changes were only partly summarized in Floras or recent taxonomic treatments of certain groups. Our results highlight the presence of two new subtribes (Cipuropsidinae, Vrieseinae), three new genera (Josemania, Lemeltonia, Rothowia), and three new subgenera (Racinaea subg. Pseudophytarrhiza, T. subg. Viridantha, and T. subg. Pseudovriesea). Beside species from newly erected taxa, several others are also reclassified where initial placement was controversial among bromeliad researchers. Parsimony analysis revealed the following, mostly well-supported general branching pattern: ((Glomeropitcairnia, Catopsis) (((Alcantarea, Vriesea) (Werauhia, Cipuropsis–Mezobromelia clade)) (Guzmania, ((Josemania, Racinaea) (Rothowia, (Lemeltonia, Tillandsia); Bayesian analysis in contrast revealed a slightly different branching pattern within core Tillandsieae: (Guzmania, ((Josemania, Rothowia) (Racinaea (Lemeltonia, Tillandsia))). Relationships of subgenera and informal clades within Tillandsia are mostly unsupported, but most clades themselves receive good statistical support. Although much progress has been achieved in exploring phylogenetic relationships of Tillandsioideae by combining both cpDNA and nDNA sequence data, a final and conclusive classification for the whole subfamily cannot be presented here. The most critical groups that remain to be solved using new molecular markers are the Cipuropsis–Mezobromelia alliance and the genus Tillandsia. 214 XVIII IBC 2011 Origin, phylogeny, adaptive radiation, and geographic diversification of Bromeliaceae Thomas J. Givnish 1, Michael H.J. Barfuss 2, Benjamin Van Ee 3, Ricarda Riina 4, Katharina Schulte 5, Ralf Horres 6, Philip A. Gonsiska 1, Rachel S. Jabaily 1, Darren M. Crayn 5, J. Andrew C. Smith 8, Klaus Winter 9, Bruce K. Holst 10, Harry E. Luther 10, Walter Till 2, Georg Zizka 11, Paul E. Berry 4, Ann Arbor 1, & Kenneth .J. Sytsma 1 1 University of Wisconsin-Madison, Department of Botany, 430 Lincoln Drive, Birge Hall, Madison, WI, 53706, USA; 2 University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria; 3 Black Hills State University, Spearfish, USA; 4 University of Michigan Herbarium, Ecology and Evolutionary Biology, 3600 Varsity Dr, Ann Arbor, Michigan, 48108, USA; 5 James Cook University, Australian Tropical Herbarium, Cairns, QLD, 4870, Australia; 6 GenXPro GmbH, Altenhöferallee 3, 60438 Frankfurt/Main, Germany; 8 University of Oxford, Department of Plant Sciences, South Parks Road, Oxford, OX1 3RB, United Kingdom; 9 Smithsonian Tropical Research Institute, P.O. Box 2072, Balboa, Ancón, Republic of Panama; 10 Marie Selby Botanic Gardens, 811 South Palm Avenue, Sarasota, Florida, 34236-7726, USA; 11 Research Institute Senckenberg, Department of Botany, Frankfurt/Main, Germany Author for correspondence: Thomas J. Givnish, givnish@wisc.edu Oral presentation: Pp. 186–187 (SYM004) in: XVIII IBC 2011: XVIII International Botanical Congress (abstracts). Melbourne, Australia: Melbourne Congress and Exhibition Centre, 23–30 July 2011. Abstract Sequence variation in eight rapidly evolving plastid regions for 90 bromeliad species from 45 of 56 described genera confirm that the eight bromeliad subfamilies are related to each other in ladder-like fashion: (Brocchinoidieae, (Lindmanioideae, (Tillandsioideae, (Hechtioideae, (Navioideae, (Pitcairnioideae, (Puyoideae, Bromelioideae). Puya shows a basal split between species found primarily at higher elevations in the Andes and those found near sea level in Chile. The earliest-divergent bromelioids are mostly restricted to the southern Andes and Chilean coast. We calibrated this phylogeny against the ages of fossil monocots using penalized likelihood, and assessed patterns of biogeographic spread using maximum parsimony, Bayesian inference, and S-DIVA. Bromeliads appear to have arisen in the Guayana Shield roughly 100 Mya, spread centrifugally in the New World, and reached tropical West Africa via long-distance dispersal 9.2 Mya. Modern lineages began to diverge from each other 19 Mya, with invasions of drier or higher peripheral areas in Central America and northern South America beginning 16–13 Mya, coincident with a major adaptive radiation – the 'bromeliad revolution' – involving the repeated evolution of epiphytism, the tank habit, CAM photosynthesis, and avian pollination, as well as several features of leaf and trichome anatomy, and an accelerated pace of species diversification. This revolution coincided with the uplift of the northern Andes and its invasion by epiphytic tillandsioids and by ancestors of the pitcairnioids, puyoids, and bromelioids. Bromelioids invaded the Serra do Mar and nearby mountains in southeastern Brazil, most likely from southern Chile, starting 9.1 Mya. A major radiation of epiphytic bromelioids in Brazil began 5.7 Mya. Calculations of net rates of species diversification for subfamilies or clades of similar rank range from 0.16 sp sp-1 My-1 in brocchinioids to 1.12 sp sp-1 My-1 in the tank-epiphytic bromelioids. Acquisition of the epiphytic habit and related traits appears to have accelerated net species diversification by favoring seed traits that increased attachment to epiphytic perches, and coincidentally increased the ability of tillandsioids and bromelioids to colonize extensive montane regions in the Andes and Central America, and in the Serra do Mar and nearby mountains in coastal Brazil, permitting geographic speciation to proceed in massively parallel fashion as epiphytes occupied a cloud-forest landscape dissected by numerous drier, lower valleys that could act as extrinsic barriers to gene flow. Avian pollination (mainly by hummingbirds) evolved at least twice, coincident with the invasion of cool, wet montane habitats. Net diversification rates were significantly higher in hummingbird-pollinated clades, perhaps reflecting the rise of gullet-shaped flowers adapted to such pollinators and the resulting opportunity for rapid partitioning of another rapidly speciating montane clade based on differences in bill length. Entomophily is the ancestral condition in bromeliads, and bat pollination appears to have evolved several times from hummingbird pollination; chiropterophily is associated with warmer and/or drier conditions then ornithophily. 216 BioSystematics 2011 Conflicting phylogenetic signal within the nuclear marker PRK highlights the importance of hybridization events in the diversification of Bromeliaceae Ingo Michalak 1,2,3, Daniele Silvestro 1,2,3, David Brie 1,2, Michael H.J. Barfuss 4, Katharina Schulte 5 & Georg Zizka 1,2,3 1 Senckenberg Research Institute, Department of Botany and Molecular Evolution, 253 Senckenberganlage 25, 60325 Frankfurt/Main, Germany 2 Goethe University, Institute of Ecology, Evolution and Diversity, Senckenberganlage 25, 60325 Frankfurt/Main, Germany 3 Biodiversity and Climate Research Centre (BiK-F), Senckenberganlage 25, 60325 Frankfurt/Main, Germany 4 University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria 5 James Cook University, Australian Tropical Herbarium, Sir Robert Norman Building (E2), PO Box 6811, Cairns 4870, QLD, Australia Author for correspondence: Ingo Michalak, imichalak@senckenberg.de; Daniele Silvestro, dsilvestro@senckenberg.de Poster: Pp. 252–253 (II-74) in: BioSystematics Berlin 2011: 7th International Congress of Systematic and Evolutionary Biology (ICSEB VII), 12th Annual Meeting of the Society of Biological Systematics (Gesellschaft für Biologische Systematik, GfBS), 20th International Symposium “Biodiversity and Evolutionary Biology” of the German Botanical Society (DBG) (abstracts). Berlin, Germany: 21–27 February 2011. Abstract The family Bromeliaceae includes 57 genera and over 1700 species which have a predominantly Neotropical distribution (1 species in West Africa), and underwent several major radiation events over the past 20 million years. Despite the remarkable morphological diversity within the family only little variation is found in plastid DNA sequences, severely hampering the inference of interspecific relationships based on these markers. Nuclear markers generally yield a higher proportion of informative sites compared to plastid sequences, but are more difficult to amplify and problems in distinguishing paralogous and orthologous sequences may arise. We produced nuclear (PRK) and plastid alignments for representatives of the subfamilies Bromelioideae and Pitcairnioideae s.str. (with focus on the genus Fosterella) and investigated their contrasting evolutionary histories. Here we show that plastid and nuclear phylogenies yield incongruent topologies and that complex patterns within the nuclear phylogenetic reconstructions based on the PRK dataset are best explained with the occurrence of paralogous copies but also with past hybridization events within Bromelioideae as well as within Fosterella. The importance of hybridization in the rapid radiation of Bromeliaceae is discussed as well as potential strategies to extract phylogenetic information from these and similar data sets. 218 Systematics 2009 The genus Deuterocohnia Mez (Bromeliaceae): Conflicting data in phylogenetic analysis Nicole Schuetz 1, Michael H.J. Barfuss 2, Kurt Weising 1 & Georg Zizka 3 1 Universität Kassel, Institut für Biologie, AG Systematik und Morphologie der Pflanzen, Heinrich-Plett-Straße 40, 34132, Kassel, Germany 2 University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria 3 Senckenberg Research Institute, Department of Botany and Molecular Evolution, 253 Senckenberganlage 25, 60325 Frankfurt/Main, Germany Author for correspondence: Nicole Schuetz, nschuetz@uni-kassel.de Oral presentation: P. 118 in: Systematics 2009, 7th Biennial Conference of the Systematics Association (abstracts). Leiden, The Netherlands: National Herbarium of The Netherlands and National Museum of Natural History Naturalis, Leiden University Medical Centre, 10–14 August 2009. 220 Systematics 2009 Phylogenetic utility of the nuclear marker PRK on a low taxonomic level: A case study in the genus Fosterella (Bromeliaceae) David Brie 1,2, Katharina Schulte1, Michael H.J. Barfuss 4 & Georg Zizka 1,2,3 1 Senckenberg Research Institute, Department of Botany and Molecular Evolution, 253 Senckenberganlage 25, 60325 Frankfurt/Main, Germany 2 Goethe University, Institute of Ecology, Evolution and Diversity, Senckenberganlage 25, 60325 Frankfurt/Main, Germany 3 Biodiversity and Climate Research Centre (BiK-F), Senckenberganlage 25, 60325 Frankfurt/Main, Germany 4 University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria Author for correspondence: D. Brie, dbrie@senckenberg.de Oral presentation: P. 118 in: Systematics 2009: 7th Biennial Conference of the Systematics Association (abstracts). Leiden, The Netherlands: National Herbarium of The Netherlands and National Museum of Natural History Naturalis, Leiden University Medical Centre, 10–14 August 2009. 222 Botany & Mycology 2009 Classification, adaptive radiation, and geographic diversification in Bromeliaceae: Insights from a new multi-locus phylogeny Thomas J. Givnish 1, Benjamin Van Ee 2, Michael H.J. Barfuss 3, Ricarda Riina 4, Katharina Schulte 5, Ralf Horres 6, Philip A. Gonsiska 1, Rachel .S. Jabaily 1, Darren M. Crayn 7, J. Andrew C. Smith 8, Klaus Winter 9, Gregory K. Brown 10, Timothy M. Evans 11, Bruce K. Holst 12, Harry E. Luther 12, Walter Till 3, Georg Zizka 5, Paul .E. Berry 13 & Kenneth J. Sytsma 8 1 University of Wisconsin-Madison, Department of Botany, 430 Lincoln Drive, Birge Hall, Madison, WI, 53706, USA; 2 Harvard University, Botanical Museum, 26 Oxford St, Cambridge, Massachusetts, 02138, USA; 3 University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria; 4 University of Michigan, Herbarium, 3600 Varsity Drive, Ann Arbor, Michigan, 48108, USA; 5 Research Institute Senckenberg, Department of Botany, Frankfurt/Main, Germany; 6 GenXPro GmbH, Altenhöferallee 3 60438, Frankfurt/Main, Germany; 7 James Cook University, Australian Tropical Herbarium, Cairns, QLD, 4870, Australia; 8 University of Oxford, Department of Plant Sciences, South Parks Road, Oxford, OX1 3RB, United Kingdom; 9 Smithsonian Tropical Research Institute, P.O. Box 2072, Balboa, Ancón, Republic of Panama; 10 University of Wyoming, Department of Botany, 3165, 1000 E University Avenue, Laramie, Wyoming, 82071, USA; 11 Grand Valley State University, Biology Department, 1 Campus Drive, Allendale, Michigan, 49401, USA; 12 Marie Selby Botanic Gardens, 811 South Palm Avenue, Sarasota, Florida, 34236-7726, USA; 13 University of Michigan Herbarium, Ecology and Evolutionary Biology, 3600 Varsity Dr, Ann Arbor, Michigan, 48108, USA Author for correspondence: Thomas J. Givnish, givnish@wisc.edu Oral presentation: (753) In: Botany & Mycology 2009 (abstracts). Snowbird, Utah, USA: Snowbird Center, 25–29 July 2009. Abstract The slow rate of molecular evolution in Bromeliaceae, frequent morphological homoplasy, and extensive molecular and morphological divergence from the families most closely to it have hampered progress toward an understanding of evolutionary relationships within the family. Here we present a molecular phylogeny based on more than 9500 aligned bases from eight rapidly genes and spacers, for 90 ingroup taxa and three outgroups. The eight-subfamily classification recently advanced by Givnish et al. is supported. Brocchinioideae is sister to all other extant bromeliads; Lindmanioideae, also endemic to the Guayana Shield, diverged next, followed by Tillandsioideae. It appears that xeromorphic Hechtioideae diverged from the bromeliad spine next, followed by Navioideae (endemic to the Guayana Shield, with one species on the Brazilian Shield), Pitcairnioideae, with Puyoideae and Bromelioideae being sister to each other. Calibration of the bromeliad molecular tree against dates corresponding to nonbromeliad fossils indicates that bromeliads began to divergence from other monocots 70 Mya, and that the extant bromeliad genera began to diverge from each other only in the last 19 My. Fifty-one million years between the origin of bromeliads and initial divergence of surviving lineages helps explain the difficulty in identifying their closest relatives. Extant species of Brocchinia began to diverge from each other about 17 Mya, before almost any other genera began to diverge, helping explain why Brocchinia shows such a wide range of adaptive types. The pace of diversification accelerated greatly 13 Mya, coincident with the rise of several morphological and physiological adaptations to dry or epiphytic conditions arose, and invasion of areas peripheral to the Guayana Shield. This “bromeliad revolution” corresponds to the time of multiple origins of CAM photosynthesis, epiphytism, bird pollination, tank habit (following its initial origin in Brocchinia), and absorptive trichomes. A new model for determinants of bromeliad diversity is presented and discussed. 224 Monocots IV Molecular phylogenetics of Tillandsia (Bromeliaceae) and related genera Michael H.J. Barfuss, Rosabelle Samuel & Walter Till University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria Author for correspondence: Michael H.J. Barfuss, michael.h.j.barfuss@univie.ac.at Oral presentation: In: Monocots IV: The 4th international conference on the comparative biology of the monocotyledons & the 5th international symposium on grass systematics and evolution (abstracts). Copenhagen, Denmark: University of Copenhagen, H.C. Ørsted Institute, 11– 15 August 2008. Abstract Tillandsia is the most species-rich genus of subfamily Tillandsioideae (Bromeliaceae). Its traditional delimitation and internal classification are the subjects of much debate and conflict with recent molecular studies. Previously investigated plastid regions have shown little variation across Tillandsioideae, due to rapid radiation and subsequent genetic isolation/speciation. Until now it has not been possible to unambiguously infer phylogenetic relationships within the core Tillandsioideae, especially the Tillandsia s. l. (incl. Racinaea and Viridantha) and Vriesea s. l. (incl. Alcantarea and Werauhia) backbones. In this study we have analysed additional chloroplast and nuclear DNA markers to obtain more phylogenetic information, with the particular aim of resolving these backbones better. Results from the difficult ITS (Internal transcribed spacer) region show ITS 1 to be highly variable, but the ITS region as a whole contains only little phylogenetic signal. Therefore, sequences from the single-copy nuclear genes PRK (Phosphoribulokinase) and PHYC (Phytochrome C) are here combined with ITS and chloroplast sequence data to provide a well-supported molecular phylogeny. Only now we can propose a novel classification for Tillandsia and its closely related genera that were only partially reflected by previous chloroplast studies. 226 Monocots IV Adaptive radiation and diversification in Bromeliaceae: Inferences from a new multigene phylogeny Thomas J. Givnish 1, Ricarda Riina 2, Benjamin Van Ee 3, Paul .E. Berry 4, Kenneth J. Sytsma 5, Michael H.J. Barfuss 6, Walter Till 6, Ralf Horres 7, Katharina Schulte 8 & Georg Zizka 8 1 University of Wisconsin-Madison, Department of Botany, 430 Lincoln Drive, Birge Hall, Madison, WI, 53706, USA; 2 University of Michigan, Herbarium, 3600 Varsity Drive, Ann Arbor, Michigan, 48108, USA; 3 Harvard University, Botanical Museum, 26 Oxford St, Cambridge, Massachusetts, 02138, USA; 4 University of Michigan Herbarium, Ecology and Evolutionary Biology, 3600 Varsity Dr, Ann Arbor, Michigan, 48108, USA; 5 University of Oxford, Department of Plant Sciences, South Parks Road, Oxford, OX1 3RB, United Kingdom; 6 University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria; 7 GenXPro GmbH, Altenhöferallee 3 60438, Frankfurt/Main, Germany; 8 Research Institute Senckenberg, Department of Botany, Frankfurt/Main, Germany Author for correspondence: Thomas J. Givnish, givnish@wisc.edu Oral presentation: In: Monocots IV: The 4th international conference on the comparative biology of the monocotyledons & the 5th international symposium on grass systematics and evolution (abstracts). Copenhagen, Denmark: University of Copenhagen, H.C. Ørsted Institute, 11– 15 August 2008. Abstract Bromeliads show an unusually slow rate of molecular evolution, substantial homoplasy in morphological character-states, and substantial molecular and morphological divergence from other monocots, making inference of relationships within the family difficult. Here we present the most comprehensive analysis of bromeliad phylogeny to date, based on sequencing three rapidly evolving plastid genes (matK, ndhF, rps16) and three rapidly evolving plastid spacers (psbA-trnH, rbcL-atpB, trnL-trnF) across nearly 100 bromeliad taxa and outgroups drawn from six other families of Poales. We use these data to test the monophyly and recently proposed relationships among the eight subfamilies of Bromeliaceae; trace the evolution of several characters crucial for the evolution of the epiphytic habit under parsimony, maximum likelihood, and Bayesian frameworks; infer the broad-scale geographic patterns of geographic diversification within the family; and produce a calibrated chronology of evolution within the family. Together, these findings allow us to identify the timing of the “bromeliad revolution”, during which several traits crucial to extensive speciation within the family apparently arose, accompanying its breakout from the exceedingly humid, infertile environments of the Guayana Shield. 228 Systematics 2008 The backbone problem: Hunting for informative characters in DNA sequences of Tillandsioideae (Bromeliaceae) for reconstructing a well-resolved phylogeny Michael H.J. Barfuss, Walter Till & Rosabelle Samuel University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria Author for correspondence: Michael H.J. Barfuss, michael.h.j.barfuss@univie.ac.at Oral presentation: P. 356 in: Gradstein, R. & al. (eds.), Systematics 2008: 10th Annual Meeting of the Gesellschaft für Biologische Systematik, 18th International Symposium "Biodiversity and Evolutionary Biology" of the German Botanical Society (DBG) (abstracts). Göttingen, Germany: University of Göttingen, 7–11 April 2008. Abstract Plastid Genomes have shown to be very uniform in recently diverged plant groups (like in Bromeliaceae) mainly caused by rapid radiation (and subsequent quick speciation / isolation). In addition to the less variable chloroplast regions which we analysed in our previous studies, we sequenced another set of cpDNA markers, which are supposed to be phylogenetically much more informative. Nevertheless, avoiding results of still insufficiently resolved phylogenies derived from this additional chloroplast data we also established primers for nuclear DNA regions, especially with the target of getting well resolved backbones in our trees. Attempts of several research groups to amplify the nrDNA region ITS (Internal Transcribed Spacer) in Bromeliaceae failed so far, because of very strong bounded secondary structures caused by repetitive GC-rich motives in the ITS 1 region. While we were able to successfully amplify and sequence over this difficult region for a set of species of Bromeliaceae by modifying PCR programs and sequencing chemistries, results of sequence analysis again showed very poor amounts of phylogenetically informative characters among taxa, having a similar substitution rate as it can be found in the investigated chloroplast regions (which is in contrast to many other plant groups). All resulting trees (even from combined analyses) did not yet allow an unambiguous interpretation of phylogenetic relationships of the whole subfamily Tillandsioideae; especially not at the backbone of very species-rich genera like Tillandsia and close relatives (Racinaea, Viridantha), and Vriesea and associates (Alcantarea, Werauhia). Therefore, sequences (including exons as well as introns) of the assumed single-copy nDNA genes PRK (Phosphoribulokinase), PHYC (Phytochrom C) and MS (Malate Synthase) have been analysed, and results compared with phylogenies derived from ITS and more than 10 different plastid regions; this nuclear markers seem to be much more promising in giving higher phylogenetic signals per sequenced base pair than any of the other previously investigated DNA regions does. 230 XVII IBC 2005 Phylogenetics of subfamily Tillandsioideae (Bromeliaceae): A comparison of plastid and nuclear DNA sequence with morphological data Michael H.J. Barfuss, Walter Till & Rosabelle Samuel University of Vienna, Faculty of Life Sciences, Department of Systematic and Evolutionary Botany, Faculty Center of Biodiversity, Rennweg 14, 1030 Vienna, Austria Author for correspondence: Michael H.J. Barfuss, michael.h.j.barfuss@univie.ac.at Oral presentation: P. 150 (9.7.3.) in: XVII IBC 2005: XVII International Botanical Congress (abstracts). Vienna, Austria: Austria Center Vienna, 17–23 July 2005. Abstract Part of the low-copy nuclear gene phosphoribulokinase (PRK) and a multicopy nuclear rDNA spacer region (ITS) were used to clarify relationships of subfamily Tillandsioideae (Bromeliaceae). These data are compared with the available cpDNA sequences of coding rbcL and matK and noncoding trnL intron, trnL-F intergenic spacer, atpB-rbcL intergenic spacer, rps16 intron, and partial 5' and 3' trnK intron. We added taxa of Tillandsia, Guzmania and Vriesea to get better resolution within these genetically convergent genera. Molecular phylogenetic data is compared with various morphological characters to test their usefulness for a new classification of Tillandsioideae. These characters are optimised onto a strict consensus tree of the combined matrices of plastid and nuclear DNA sequences. Preference has been given to characters likely to be little affected by adaptive constraints, i.e., pollen, stigma, ovules, anther, seed, and nectary morphology. 232 Conclusions The current study demonstrates that the analyses of additional plastid and nuclear DNA markers are able to provide better resolution and support in phylogenetic trees of Bromeliaceae. The inclusion of more taxa significantly improved the tree topologies, especially where more distantly related taxa were sampled in order to break long branches, while sampling of lineages which were not included before tend to decrease the support for certain clades. However, in all cases several unknown relationships could be resolved. Analyses of plastid DNA sequences including more markers and more species clarified several uncertain relationships which existed among subfamilies of Bromeliaceae. The eight subfamilies are confirmed, but uncertainty about a monophyletic origin of subfamily Puyoideae persists. Previously proposed hypotheses concerning biogeography, origin and age of Bromeliaceae could be re-analysed and refined with the enlarged data set. Rapid radiation and relatively recent diversification within the last 8.7–4 My are most likely the explanation for the low sequence divergence. The investigation of the low-copy nuclear gene PRK has provided useful insights into phylogenetic relationships of Bromelioideae. The amplified region is at least 3-times more variable than any plastid regions, thus making PRK suitable for phylogenetic reconstructions. However, the amplified fragment alone did not provide sufficient information for resolving relationships completely, only the combined analysis with plastid markers clarified several previously unresolved relationships at deeper nodes (“basal bromelioids”, “eu-bromelioids”). Genera within core bromelioids remain problematic. Most parsimonious reconstructions for tank habit and the sepal symmetry revealed strong evolutionary significance and indicated diagnostic utility for future classifications. The comparative studies of nuclear DNA sequences within Tillandsioideae showed that some nuclear markers are able to provide more information and a higher degree of resolution in phylogenetic trees than plastid markers. However, their utility cannot be measured only by PICs, but depends also on methodological challenges. Nuclear ITS nrDNA is not recommended for phylogenetic investigations of Bromeliaceae due to the presence of strong secondary structures, which create problems in amplification and sequencing, as well as few PICs to resolve deeper nodes. Some other amplified fragments of genes were to short and showed few potentially PICs. The selected nuclear DNA regions of PRK and PHYC are able to provide reasonable support in some parts of phylogenetic trees, but several relationships still remain unresolved. So far, the search for “the ideal” phylogenetic marker for Tillandsioideae continues and perhaps it is impossible to find such a marker with standard PCR and Sanger-sequencing. However, progress has been achieved in exploring phylogenetic relationships of Tillandsioideae by combining nuclear DNA sequence Data with already published plastid DNA sequence data. DNA results together with re-evaluated morphological characters allowed the recircumscription of existing genera and description of three new genera and subgenera, respectively. However, taxonomic problems in certain groups persist due to poor sampling, limited 233 BARFUSS, M.H.J. MOLECULAR STUDIES resolution and variable morphology. To get a fully resolved picture of relationships within Tillandsioideae, more DNA data is needed as well as a careful revision of morphological characters which are already indicated to be useful to circumscribe taxonomic groups. A major challenge that remains within Bromeliaceae is the great number of species, for which it is sometimes impossible to get material for both DNA and morphological investigations. Many species are not cultivated in ex situ collections, because either they are simply not available or require very special growing conditions. Several species were collected only once in the wild or are known from a single locality. If plant material is available, it is rather simple to generate additional DNA sequence data for phylogenetic studies, but collecting morphological characters is still challenging. Many useful characters are missing on herbarium specimens and flowering material is sometimes hardly available for certain species. Without having flowering or fruiting specimens (ideally both) in good conditions, a careful taxonomic revision is impossible. The future direction will be to establish improved international collaborations with taxonomists in the countries of origin of Bromeliaceae to access important plant material more easily. The continuously improving next generation sequencing (NGS) technologies will play important roles in future phylogenetic investigations of Bromeliaceae. Sequencing and analysing of plastoms has already been successfully implemented for monocots; this will definitely help to clarify relationships and uncover the evolutionary history of the plastid genome in Bromeliaceae. The analysis of the entire 18S-5.8S-26S nrDNA repeat unit can also provide useful insights into the evolution of nrDNA in bromeliads and uncover the ecological significance of the strong secondary structure. NGS of nrDNA might also be useful to differentiate species and therefore could be a potential candidate marker for DNA barcoding. Furthermore, the possibility to gain huge amounts of allelic nuclear data will help to identify nuclear regions that are useful for inferring relationships and uncover hybridisation events. This work aimed to provide a starting point for a biological classification of Bromeliaceae apart from the already well supported subfamilies. It intends to stimulate researchers interested in Bromeliaceae to focus on certain taxa, collect morphological information and apply newest sequencing technologies to answer several, to date unresolved evolutionary questions in Bromeliaceae. 234 Curriculum Vitae Mag. Michael H.J. Barfuss Name Michael Harald Johannes Barfuss Academic degree Mag.rer.nat. Date of birth, nationality 15.03.1977, Austria Home address Schillerstraße 12 Haus 12 2351 Wiener Neudorf, Austria Corresponding address University of Vienna, Faculty of Life Sciences, Faculty Center of Biodiversity Department of Systematic and Evolutionary Botany, Rennweg 14, 1030, Vienna, Austria Phone: +43-(0)1-4277-54130 (office) +43-(0)1-4277-54129 (lab) Fax: +43-(0)1-4277-9541 E-mails: michael.h.j.barfuss@univie.ac.at a9601611@unet.univie.ac.at ariocarpus@gmx.at Employment since 2006 Technical laboratory assistant, molecular systematics laboratory (lab management and sequencing technology); research associate, Department of Systematic and Evolutionary Botany, University of Vienna Academic career 2006–2012 Dr. student, Botany, University of Vienna, Austria 2006 Diplom (Mag.rer.nat), supervisors ao. Univ.-Prof. Dr. Rosabelle Samuel, o. Univ.-Prof. Dr. Tod Stuessy, final exam (Diplomprüfung) 25. August 2006, title of diploma thesis Molecular studies in Bromeliaceae: Phylogenetic relationships in subfamily Tillandsioideae based on evidence from plastid sequences 1996–2006 Diploma student, Biology (Botany), University of Vienna, Austria 1996 Qualification for university entrance (Matura) 3. June 1996, Höhere Bundeslehr- und Versuchsanstalt für Gartenbau, Vienna, Austria 235 BARFUSS, M.H.J. MOLECULAR STUDIES Project leadership since 2011 Evolution of the Tillandsia capillaris complex (Bromeliaceae): Adaptive radiation in a semi-arid environment?, project leaders Dr. Jorge Chiapella (Argentinia), Ass.-Prof. Dr. Walter Till, Mag. Michael H.J. Barfuss (Austria) (BMWF, OeAD, WTZ) since 2008 Molecular Phylogeny of Gymnocalycium (Cactaceae), project leaders Mag. Michael. H.J. Barfuss, Ass.-Prof. Dr. Walter Till (GÖK, AGG) 2007–2008 Evolution und Sekundärstruktur der nrDNS-Region ITS von Tillandsia s. l. (Bromeliaceae): Zusammenhänge zwischen Sequenz-Alignments, Phylogenie und ökologisch-adaptiver Variabilität, project leaders Ass.-Prof. Dr. Walter Till, Mag. Michael. H.J. Barfuss (ÖAW, KIÖS no. 2007-02) Project collaboration since 2012 Phylogeny and evolution of Tillandsia subgenus Tillandsia (Bromeliaceae), project leaders Ass.-Prof. Dr. Walter Till, Mag. Michael H.J. Barfuss; postdoc Juan Pablo Pinzón since 2003 Phylogeny, adaptive radiation, and historical biogeography in Bromeliaceae, The Bromeliad Phylogeny Group, group leader Prof. Dr. Thomas J. Givnish 2003 Molecular phylogeny of Bromeliaceae subfamily Tillandsioideae: Insights from the nuclear and chloroplast genome, project leader Ass.-Prof. Dr. Walter Till (ÖAW, KIÖS) 2002 Molekulare Analyse der Verwandtschaftsbeziehungen innerhalb der Gattung Tillandsia (Bromeliaceae, Unterfamilie Tillandsioideae), project leader ao. Univ.-Prof. Dr. Rosabelle Samuel (Hochschuljubiläumsstiftung der Stadt Wien) Project assistance since 2010 Evolution and Biodiversity of New Caledonian Diospyros (Ebenaceae), project leader ao. Univ.-Prof. Dr. Rosabelle Samuel (FWF no. P22159) 2007–2011 Taxonomische Überarbeitung der Gattung Leontopodium, project leader o. Univ.-Prof. Dr. Hermann Stuppner (FWF no. P19480) 2006–2010 Species delineation and autecology of Spirogyra, project leader ao. Univ.-Prof. Mag. Dr. Michael Schagerl (FWF no. P18465) 2006–2010 Molecular Phylogeny, genome size and chromosomal evolution in genus Polystachya Hooker (Orchidaceae), project leader ao. Univ.-Prof. Dr. Rosabelle Samuel (FWF no. P19108) 2005–2006 Chromosomes and evolution of Melampodium (Asteraceae), project leader o. Univ.-Prof. Dr. Tod F. Stuessy (FWF no. P18201) 236 IN BROMELIACEAE CURRICULUM VITAE 2004–2007 Phylogeny and Historical Biogeography of Ebenaceae, project leader ao. Univ.-Prof. Dr. Rosabelle Samuel (FWF no. P17094) 2002 Phylogenie and historical biogeography of Phyllanthaceae, project leader ao. Univ.-Prof. Dr. Rosabelle Samuel (FWF no. P15333) 2001–2002 Intraspecific phylogeography of alpine plants, project leader Univ.-Prof. Dr. Harald Niklfeld (FWF no. P13874) 2000–2002 Generic delimitations and molecular phylogeny of subfamily Tillandsioideae (Bromeliaceae), project leader Ass.-Prof. Dr. Walter Till (FWF no. P13690) Grants (donator) 2011 XVIII IBC 2011, congress (University of Vienna) 2009 Jodrell Laboratory, Royal Botanic Gardens, Kew (Synthesys) 2008 Systematics 2008, conference (University of Vienna) 2008 Monocots IV, conference (ÖFB) 2003 Monocots III, conference (ÖFB) Student tutorials 2008–2009 Project practical course in DNA markers and chromosomes in plant systematics and evolutionary research 2005–2007 Practical course in population biology (AFLP, SSR) 2004–2010 Practical course in DNA sequencing analysis and molecular phylogeny 2001–2002 Practical course in enzyme analysis in plant systematics 2001 Plant systematic-morphological introductory excursions 2000 Plant systematic-morphological practical course Teaching experience since 2012 Project practical course in molecular and karyological methods in evolution and ecology since 2011 Project practical course in DNA barcoding - a new approach to species identification in ecology and biodiversity research since 2010 Practical course in DNA sequencing analysis and molecular phylogeny since 2010 Lectures in macromolecules and molecular phylogeny in plant systematics and evolutionary research 2010–2011 Project practical course in DNA markers and chromosomes in plant systematics and evolutionary research 237 BARFUSS, M.H.J. MOLECULAR STUDIES International field trips 2006 Ecuador 2006 USA (SW) 2002 Dominican Republic 2001 Borneo (Brunei Darussalam, Malaysia) 2000 USA (SW) 2000 Costa Rica Invited talks 2009 Implications of five nuclear regions on the phylogenetic relationships of Tillandsioideae (Bromeliaceae); Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, UK 2006 Tillandsien und verwandte Gattungen: Kann uns die DNA in der Systematik die fehlenden Antworten liefern? Ein Überblick über unseren derzeitigen Wissenstand; Biologiezentrum Linz der OÖ. Landesmuseen, Linz, Austria International conferences 2011 XVIII IBC 2011, Melbourne, Australia; oral presentation, poster 2010 Biodiversity and Evolutionary Biology, Vienna, Austria 2008 Monocots IV, Copenhagen, Denmark; oral presentation 2008 Sytematics 2008, Göttingen, Germany; oral presentation 2007 9. Jahrestagung der GfBS, Vienna, Austria 2005 XVII IBC 2005, Vienna, Austria; oral presentation 2003 Monocots III, Claremont, California, USA; oral presentation 2001 Systematics 2001, London, UK; oral presentation Scientific memberships Since 2006 American Society of Plant Taxonomists, ASPT Since 2006 International Association for Plant Taxonomy, IAPT Since 2005 Botanical Society of America, BSA Special interests All topics concerning the angiosperm families Bromeliaceae and Cactaceae; succulent plants; Asteraceae, Orchidaceae; special focus on molecular techniques (sequencing), (molecular) systematics, taxonomy, nomenclature, and field botany; plant cultivation and propagation, protection and preservation of endangered plant species (CITES). 238 IN BROMELIACEAE CURRICULUM VITAE Publications 00. Barfuss, M.H.J., W. Till & R. Samuel. In prep. A new classification of Bromeliaceae subfamily Tillandsioideae inferred from DNA sequences data of two genomes and morphology. 00. Barfuss, M.H.J., R. Samuel, M.W. Chase & F. Forest. In prep. Optimizing eight nuclear DNA markers for phylogenetic studies in recently diverged angiosperms: A case study in Bromeliaceae subfamily Tillandsioideae. 00. Givnish, T.J., M.H.J. Barfuss, B. Van Ee, R. Riina, K. Schulte, R. Horres, P.A. Gonsiska, R.S. Jabaily, D.M. Crayn, J.A.C. Smith, K. Winter, G.K. Brown, T.M. Evans, B.K. Holst, H.E. Luther, W. Till, G. Zizka, P.E. Berry & K.J. Sytsma. In prep. Adaptive radiation, correlated and contingent evolution, and net species diversification in Bromeliaceae. 00. Sehr, E.M., M.H.J. Barfuss, G.E. Barboza, R. Samuel, E.A. Moscone & F. Ehrendorfer. In prep. DNA sequences from the plastid genome and the single-copy nuclear gene waxy suggest two dysploid karyotype changes in the phylogeny of Capsicum (Solanaceae). 17. Chen, C., M.H.J Barfuss, T. Pröschold & M. Schagerl. 2012. Hidden genetic diversity in the green alga Spirogyra (Zygnematophyceae, Streptophyta). B.M.C. Evolutionary Biology (12): 77, doi 1186/1471-2148-12-77. 16. Murillo-A., J., E. Ruiz-P., L.R. Landrum, T.F. Stuessy & M.H.J. Barfuss. 2012. Phylogenetic relationships in Myrceugenia (Myrtaceae) based on plastid and nuclear DNA sequences. Molecular Phylogenetics and Evolution, 62 (2): 764–776. 15. Demaio, P.H., M.H.J. Barfuss, R. Kiesling, W. Till & J. O. Chiapella. 2011. Molecular phylogeny of Gymnocalycium (Cactaceae): Assessment of alternative infrageneric systems, a new subgenus, and trends in the evolution of the genus. American Journal of Botany 98 (11): 1841–1854. 14. Givnish, T.J., M.H.J. Barfuss, B. Van Ee, R. Riina, K. Schulte, R. Horres, P.A. Gonsiska, R.S. Jabaily, D.M. Crayn, J.A.C. Smith, K. Winter, G.K. Brown, T.M. Evans, B.K. Holst, H.E. Luther, W. Till, G. Zizka, P.E. Berry & K.J. Sytsma. 2011. Phylogeny, adaptive radiation, and historical biogeography in Bromeliaceae: Insights from an eight-locus plastid phylogeny. American Journal of Botany 98 (5): 872–895. 13. Demaio, P.H., M.H.J. Barfuss, W. Till & J. O. Chiapella. 2010. Phylogenetic relationships and infrageneric classification of the genus Gymnocalycium: Insights from molecular data. Gymnocalycium (Sonderausgabe 2010): 925–946. 12. Russell, A., R. Samuel, V. Klejna, M.H.J. Barfuss, B. Rupp & M.W. Chase. 2010. Reticulate evolution in diploid and tetraploid species of Polystachya (Orchidaceae) as shown by plastid DNA sequences and low-copy nuclear genes. Annals of Botany 106 (1): 37–56. 11. Russell, A., R. Samuel, B. Rupp, M.H.J. Barfuss, M. Šafran, V. Besendorfer & M.W. Chase. 2010. Phylogenetics and cytology of a pantropical orchid genus Polystachya (Polystachyinae, Vandeae, Orchidaceae): Evidence from plastid DNA sequence data. Taxon 59 (2): 389–404. 10. Blöch, C., H. Weiss-Schneeweiss, G.M. Schneeweiss, M.H.J. Barfuss, C.A. Rebernig, J.L. Villaseñor & T.F. Stuessy. 2009. Molecular phylogenetic analyses of nuclear and plastid DNA sequences support dysploid and polyploid chromosome number changes and reticulate evolution in the diversification of Melampodium (Millerieae, Asteraceae). Molecular Phylogenetics and Evolution 53 (1): 220–233. 239 BARFUSS, M.H.J. MOLECULAR STUDIES 9. Duangjai, S., R. Samuel, J. Munzinger, F. Forest, B. Wallnöfer, M.H.J. Barfuss, G. Fischer & M.W. Chase. 2009. A multi-locus plastid phylogenetic analysis of the pantropical genus Diospyros (Ebenaceae), with an emphasis on the radiation and biogeographic origins of the New Caledonian endemic species. Molecular Phylogenetics and Evolution 52 (3): 602– 620. 8. Gruenstaeudl, M., E. Urtubey, R.K. Jansen, R. Samuel, M.H.J. Barfuss & T. F. Stuessy. 2009. Phylogeny of Barnadesioideae (Asteraceae) inferred from DNA sequence data and morphology. Molecular Phylogenetics and Evolution 51 (3): 572–587. 7. Schulte, K., M.H.J. Barfuss & G. Zizka. 2009. Phylogeny of Bromelioideae (Bromeliaceae) inferred from nuclear and plastid DNA loci reveals the evolution of the tank habit within the subfamily. Molecular Phylogenetics and Evolution 51 (2): 327–339. 6. Möller, M., M. Pfosser, C.-G. Jang, V. Mayer, A. Clark, M.L. Hollingsworth, M.H.J. Barfuss, Y.-Z. Wang, M. Kiehn & A. Weber. 2009. A preliminary phylogeny of the 'didymocarpoid Gesneriaceae' based on three molecular data sets: Incongruence with available tribal classifications. American Journal of Botany 96 (5): 989–1010. 5. Till, W. & M.H.J. Barfuss. 2006. Progress towards a new classification of Tillandsioideae. Journal of the Bromeliad Society 56 (6): 253–259. 4. Barfuss, M.H.J., R. Samuel, W. Till & T.F. Stuessy. 2005. Phylogenetic relationships in subfamily Tillandsioideae (Bromeliaceae) based on DNA sequence data from seven plastid regions. American Journal of Botany 92 (2): 337–351. 3. Samuel, R, H. Kathriarachchi, P. Hoffmann, M.H.J. Barfuss, K.J. Wurdack, C.C. Davis & M.W. Chase. 2005. Molecular phylogenetics of Phyllanthaceae: Evidence from plastid matK and nuclear PHYC sequences. American Journal of Botany 92 (1): 132–141. 2. Barfuss, M.H.J., R. Samuel & W. Till. 2004. Molecular phylogeny in subfamily Tillandsioideae (Bromeliaceae) based on six cpDNA Markers: An update. Journal of the Bromeliad Society 54 (1): 9–17, 48. 1. Schönswetter, P., A. Tribsch, M.H.J. Barfuss & H. Niklfeld. 2002. Several Pleistocene refugia detected in the high alpine plant Phyteuma globulariifolium Sternb. & Hoppe (Campanulaceae) in the European Alps. Molecular Ecology 11 (12): 2637–2647. Abstracts (Presenter) 37. [Oral presentation] Turner, B., J. Munzinger, S. Duangjai, M.H.J. Barfuss, B. Wallnöfer, O. Paun, M.W. Chase & R. Samuel. 2012. Diversification of endemic New Caledonian Diospyros (Ebenaceae). P. 53 in: Berichte des naturwissenschaftlich-medizinischen Vereins in Innsbruck, Supplementum 2020, 15. Treffen der Österreichischen Botanikerinnen und Botaniker (abstracts). Innsbruck, Austria: Universität Innsbruck, Institut für Botanik, 27–29 September. 36. [Oral presentation] Ehrendorfer, F., V. Vladimirov, J.-F. Manen & M.H.J. Barfuss. 2012. Adaptive Radiation und weltweite Expansion der artenreichen Tribus Rubieae (Rubiaceae). P. 13 in: Berichte des naturwissenschaftlich-medizinischen Vereins in Innsbruck, Supplementum 2020, 15. Treffen der Österreichischen Botanikerinnen und Botaniker (abstracts). Innsbruck, Austria: Universität Innsbruck, Institut für Botanik, 27–29 September. 240 IN BROMELIACEAE CURRICULUM VITAE 35. [Oral presentation] Givnish, T.J., M.H.J. Barfuss, B. Van Ee, R. Riina, K. Schulte, R. Horres, P.A. Gonsiska, R.S. Jabaily, D.M. Crayn, J.A.C. Smith, K. Winter, G.K. Brown, T.M. Evans, B.K. Holst, H.E. Luther, W. Till, G. Zizka, P.E. Berry & K.J. Sytsma. 2012. Adaptive radiation, correlated evolution, and determinants of net diversification rates in Bromeliaceae: Test of an a priori model. (638) in: Botany 2012: The Next Generation (abstracts). Columbus, Ohio: Greater Columbus Convention Center, 7–11 July. 34. [Oral presentation] Barfuss, M.H.J., W. Till & R. Samuel. 2011. Systematics, evolution, and phylogeography of Tillandsia (Bromeliaceae) and related genera. P. 189 (SYM004) in: XVIII IBC 2011: XVIII International Botanical Congress (abstracts). Melbourne, Australia: Melbourne Congress and Exhibition Centre, 23–30 July. 33. [Poster] Barfuss, M.H.J., W. Till & R. Samuel. 2011. New taxonomic implications in Tillandsioideae (Bromeliaceae) based on DNA data and morphology. Pp. 568–569 (P0730) in: XVIII IBC 2011: XVIII International Botanical Congress (abstracts). Melbourne, Australia: Melbourne Congress and Exhibition Centre, 23–30 July. 32. [Oral presentation] Givnish, T.J., M.H.J. Barfuss, B. Van Ee, R. Riina, K. Schulte, R. Horres, P.A. Gonsiska, R.S. Jabaily, D.M. Crayn, J.A.C. Smith, K. Winter, B.K. Holst, H.E. Luther, W. Till, G. Zizka, P.E. Berry, Ann Arbor & K.J. Sytsma. 2011. Origin, phylogeny, adaptive radiation, and geographic diversification of Bromeliaceae. Pp. 186–187 (SYM004) in: XVIII IBC 2011: XVIII International Botanical Congress (abstracts). Melbourne, Australia: Melbourne Congress and Exhibition Centre, 23–30 July. 31. [Oral presentation] Ehrendorfer, F., V. Vladimirov & M.H.J. Barfuss. 2011. Worldwide radiation, phylogeography and paraphyly of the speciose Rubiaceae-Rubieae. Pp. 253–254 (SYM094) in: XVIII IBC 2011: XVIII International Botanical Congress (abstracts). Melbourne, Australia: Melbourne Congress and Exhibition Centre, 23–30 July. 30. [Poster] Russell, A., R. Samuel, M.H.J. Barfuss, B. Turner & M.W. Chase. 2011. Evolutionary inference from multiple incongruent DNA data matrices: Reticulate evolution of Polystachya (Orchidaceae). P. 665 (P1042) in: XVIII IBC 2011: XVIII International Botanical Congress (abstracts). Melbourne, Australia: Melbourne Congress and Exhibition Centre, 23–30 July. 29. [Poster] Samuel, R., B. Turner, S. Duangjai, J. Munzinger, B. Wallnöfer, M.W. Chase & M.H.J. Barfuss. 2011. Origin and evolution of New Caledonian Diospyros (Ebenaceae): A phylogenetic approach. P. 635 (P0947) in: XVIII IBC 2011: XVIII International Botanical Congress (abstracts). Melbourne, Australia: Melbourne Congress and Exhibition Centre, 23–30 July. 28. [Poster] Michalak, I., D. Silvestro, D. Brie, M.H.J. Barfuss, K. Schulte & G. Zizka. 2011. Conflicting phylogenetic signal within the nuclear marker PRK highlights the importance of hybridization events in the diversification of Bromeliaceae. Pp. 252–253 (II-74) in: BioSystematics Berlin 2011: 7th International Congress of Systematic and Evolutionary Biology (ICSEB VII), 12th Annual Meeting of the Society of Biological Systematics (Gesellschaft für Biologische Systematik, GfBS), 20th International Symposium “Biodiversity and Evolutionary Biology” of the German Botanical Society (DBG) (abstracts). Berlin, Germany: 21–27 February. 27. [?] Murillo-Aldana, J., E. Ruiz-P., L.R. Landrum, T.F. Stuessy & M.H.J. Barfuss. 2010. Relaciones filogeneticas en Myrceugenia(Myrtaceae). In: Latin American Botanical Congress. La Serena, Chile: October. 241 BARFUSS, M.H.J. MOLECULAR STUDIES 26. [Oral presentation] Russell, A., M.H.J. Barfuss, M.W. Chase & R. Samuel. 2010. Reticulate evolution in diploid and tetraploid species of Polystachya (Orchidaceae) revealed by multiple nuclear and plastid loci. P. 55 in: Albach, D. & Greimler, J. (eds), 19th International Symposium "Biodiversity and Evolutionary Biology" of the German Botanical Society (DBG) (abstracts). Vienna, Austria: University of Vienna, Faculty Centre of Biodiversity, 16–19 September. 25. [Poster] Duangjai, S., R. Samuel, J. Munzinger, B. Wallnöfer, F. Forest, M.H.J. Barfuss &. M.W. Chase. 2010. Biogeography origins and evolution of endemic species of Diospyros (Ebenaceae) from New Caledonia: A molecular phylogenetic perspective. P. 85 in: Albach, D. & Greimler, J. (eds), 19th International Symposium "Biodiversity and Evolutionary Biology" of the German Botanical Society (DBG) (abstracts). Vienna, Austria: University of Vienna, Faculty Centre of Biodiversity, 16–19 September. 24. [Poster] Stockenhuber, R.M., M.H.J. Barfuss, G. Cruz-Mazo & R. Samuel. 2010. Molecular phylogeny of Scorzoneroides. P. 116 in: Albach, D. & Greimler, J. (eds), 19th International Symposium "Biodiversity and Evolutionary Biology" of the German Botanical Society (DBG) (abstracts). Vienna, Austria: University of Vienna, Faculty Centre of Biodiversity, 16–19 September. 23. [?] Russell, A., R. Samuel, V. Klejna, M.H.J. Barfuss, B. Rupp & M.W. Chase. 2010. Analysis of multiple nuclear and plastid loci reveals reticulate evolution in diploid and tetraploid species of genus Polystachya (Orchidaceae). In: New Frontiers in Plant Systematics and Evolution (NFPSE 2010). Beijing, China: 7–9 July. 22. [Oral presentation] Schuetz, N., M.H.J. Barfuss, K. Weising & Georg Zizka. 2009. The genus Deuterocohnia Mez (Bromeliaceae): Conflicting data in phylogenetic analysis. P. 118 in: Systematics 2009, 7th Biennial Conference of the Systematics Association (abstracts). Leiden, The Netherlands: National Herbarium of The Netherlands and National Museum of Natural History Naturalis, Leiden University Medical Centre, 10–14 August. 21. [Oral presentation] Brie, D., K. Schulte, M.H.J. Barfuss & G. Zizka. 2009. Phylogenetic utility of the nuclear marker PRK on a low taxonomic level: A case study in the genus Fosterella (Bromeliaceae). P. 33 in: Systematics 2009, 7th Biennial Conference of the Systematics Association (abstracts). Leiden, The Netherlands: National Herbarium of The Netherlands and National Museum of Natural History Naturalis, Leiden University Medical Centre, 10–14 August. 20. [Oral presentation] Russell, A., R. Samuel, B. Rupp, M.H.J. Barfuss, M. Safran, V. Klejna & M.W. Chase. 2009. Phylogeny, cytology and biogeography of the pantropical orchid genus Polystachya. P. 109 in: Systematics 2009, 7th Biennial Conference of the Systematics Association. Leiden, The Netherlands: National Herbarium of The Netherlands and National Museum of Natural History Naturalis, Leiden University Medical Centre, 10–14 August. 19. [Oral presentation] Givnish, T.J., B. Van Ee, M.H.J. Barfuss, R. Riina, K. Schulte, R Horres, P.A. Gonsiska, R.S. Jabaily, D.M. Crayn, J.A.C. Smith, K. Winter, G.K. Brown, T.M. Evans, B.K. Holst, H.E. Luther, W. Till, G. Zizka, P.E. Berry & K.J. Sytsma. 2009. Classification, adaptive radiation, and geographic diversification in Bromeliaceae: Insights from a new multi-locus phylogeny. (753) In: Botany & Mycology 2009 (abstracts). Snowbird, Utah, USA: Snowbird Center, 25–29 July. 242 IN BROMELIACEAE CURRICULUM VITAE 18. [Poster] Russell, A., R. Samuel, M.H.J. Barfuss, B. Rupp, V. Klejna & M.W. Chase. 2009. Low copy nuclear genes reveal hybrid speciation in Polystachya (Orchidaceae). In: International conference on polypoidy, hybridisation and biodiversity (abstracts). Saint-Malo, France: 17–20 Mai. 17. [Poster] Russell, A., R. Samuel, B. Rupp, M.H.J. Barfuss & M.W. Chase. 2008. Phylogenetics, African biodiversity and intercontinental dispersal in Polystachya (Orchidaceae). Pp. 13–14 in: 10th Young Systematists’ Forum (abstracts). London, UK: Flett Theatre, Natural History Museum, 2 December. 16. [Oral presentation] Barfuss, M.H.J., R. Samuel & W. Till. 2008. Molecular phylogenetics of Tillandsia (Bromeliaceae) and related genera. In: Monocots IV: The 4th international conference on the comparative biology of the monocotyledons & the 5th international symposium on grass systematics and evolution (abstracts). Copenhagen, Denmark: University of Copenhagen, H.C. Ørsted Institute, 11–15 August. 15. [Oral presentation] Givnish, T.J., R. Riina, B. Van Ee, P.E. Berry, K.J. Sytsma, M.H.J. Barfuss, W. Till, R. Horres, K. Schulte & G. Zizka. 2008. Adaptive radiation and diversification in Bromeliaceae: Inferences from a new multigene phylogeny. In: Monocots IV: The 4th international conference on the comparative biology of the monocotyledons & the 5th international symposium on grass systematics and evolution (abstracts). Copenhagen, Denmark: University of Copenhagen, H.C. Ørsted Institute, 11–15 August. 14. [Oral presentation] Russell, A., R. Samuel, M.H.J. Barfuss, B. Rupp & M.W. Chase. 2008. Molecular systematics of Polystachya (Orchidaceae). In: Monocots IV: The 4th international conference on the comparative biology of the monocotyledons & the 5th international symposium on grass systematics and evolution (abstracts). Copenhagen, Denmark: University of Copenhagen, H.C. Ørsted Institute, 11–15 August. 13. [Poster] Russell, A., B. Rupp, M. Šafran, R. Samuel, M.H.J. Barfuss, V. Klejna, H. WeissSchneeweiss, V. Besendorfer, D. Reich & M.W. Chase. 2008. Molecular systematics of Polystachya (Orchidaceae). In: Botany 2008: Botany without borders (abstracts). Vancouver, Canada: University of British Columbia, 26–30 July. 12. [Oral presentation] Schneeweiss, G. M., M.H.J. Barfuss & M. Thiv. 2008. Evolving towards the tops: Phylogeny and evolution of the European endemic Phyteuma (Campanulaceae). P. 16 (O 15) in: 10th Symposium of the International Organization of Plant Biosystematists (abstracts). Štrbské Pleso, Vysoké Tatry, Slovak Republic: 2–4 July. 11. [Oral presentation] Barfuss, M.H.J., W. Till & R. Samuel. 2008. The Backbone Problem: Hunting for informative characters in DNA sequences of Tillandsioideae (Bromeliaceae) for reconstructing a well-resolved phylogeny. P. 356 in: Gradstein, R. & al. (eds.), Systematics 2008, 10th Annual Meeting of the Gesellschaft für Biologische Systematik, 18th International Symposium "Biodiversity and Evolutionary Biology" of the German Botanical Society (DBG) (abstracts). Göttingen, Germany: University of Göttingen, 7–11 April. 10. [Oral presentation] Blöch, C., H. Weiss-Schneeweiss, M.H.J. Barfuss, C.A. Rebernig, J.L. Villaseñor & T.F. Stuessy. 2008. Phylogeny of the genus Melampodium and the development of the x = 10 chromosomal line. P. 41 in: Gradstein, R. & al. (eds.), Systematics 2008, 10th Annual Meeting of the Gesellschaft für Biologische Systematik, 18th International Symposium "Biodiversity and Evolutionary Biology" of the German Botanical Society (DBG) (abstracts). Göttingen, Germany: University of Göttingen, 7–11 April. 243 9. [Oral presentation] Thiv, M., M.H.J. Barfuss & G. M. Schneeweiss. 2008. Evolving to the peaks: Phylogeny and habitat evolution in the European endemic Phyteuma (Campanulaceae). P. 141 in: Gradstein, R. & al. (eds.), Systematics 2008, 10th Annual Meeting of the Gesellschaft für Biologische Systematik, 18th International Symposium "Biodiversity and Evolutionary Biology" of the German Botanical Society (DBG) (abstracts). Göttingen, Germany: University of Göttingen, 7–11 April. 8. [Poster] Russell, A., B. Rupp, M. Šafran, R. Samuel, M.H.J. Barfuss, H. Weiss-Schneeweiss, V. Besendorfer, D. Reich & M.W. Chase. 2007. Molecular systematics of Polystachya (Orchidaceae). In: Orchid evolutionary biology and conservation: From Linnaeus to the 21st century (abstracts). Richmond, UK: Royal Botanic Gardens, Kew, 31 October–2 November. 7. [Oral presentation] Blöch, C., H. Weiss-Schneeweiss, M.H.J. Barfuss, T.F. Stuessy, C.A. Rebernig & J.L. Vil-laseñor. 2007. Molecular phylogeny and chromosomal changes in Melampodium. P. 45 (96) in: Systematics 2007: The 6th Biennial Conference of the Systematics Association (abstracts). UK, Edinburgh, Royal Botanic Garden Edinburgh: 28–31 August. 6. [Oral presentation] Till, W. & M.H.J. Barfuss. 2006. Progress towards a new classification of Tillandsioideae: Mapping promising morphological characters on a phylogenetic tree. In: Bromeliads on the Border: 17th World Bromeliad Conference (BSI WBC 2006) (abstracts). San Diego, California, USA: The Bromeliad Society International, The San Diego Bromeliad Society, Town and Country Resort & Convention Center, 6–11 July. 5. [?] Blöch, C., M.H.J. Barfuss, T.F. Stuessy, C.A. Rebernig, H. Weiss-Schneeweiss & J.L. Villaseñor. 2006. Molecular phylogeny and chromosome numbers in Melampodium (Asteraceae, Heliantheae). P. 92 in: Systematics and Evolution of the Compositae: The International Compositae Alliance (TICA) Symposium (abstracts). Barcelona, Spain: 3–10 July. 4. [Oral presentation] Barfuss, M.H.J., W. Till & R. Samuel. 2005. Phylogenetics of subfamily Tillandsioideae (Bromeliaceae): A comparison of plastid and nuclear DNA sequence with morphological data. P. 150 (9.7.3.) in: XVII International Botanical Congress (XVII IBC 2005) (abstracts). Vienna, Austria: Austria Center Vienna, 17–23 July. 3. [?] Kathriarachchi, H.S., R. Samuel, P. Hoffmann, M.H.J. Barfuss & M.W. Chase. 2003. Molecular phylogenetics of Phyllanthaceae based on plastid and nuclear sequence data. P. 26 in: Systematics 2003, 4th Biennial Conference of the Systematics Association (abstracts). Dublin, Ireland: Trinity College, 18–22 August. 2. [Oral presentation] Barfuss, M.H.J., R. Samuel & W. Till. 2003. Phylogenetic relationships in subfamily Tillandsioideae (Bromeliaceae) based on evidence from plastid trnL intron, trnLtrnF intergenic spacer, atpB-rbcL intergenic spacer, rps16 intron, partial 5’ and 3’ trnK intron, matK, and rbcL sequences. P. 4 in: Monocots III, The 3rd international conference on the comparative biology of the monocotyledons & the 4th international symposium on grass systematics and evolution (abstracts). Ontario, California, USA: Ontario Convention Centre, 31 March–4 April. 1. [Oral presentation] Barfuss, M.H.J., R. Samuel & W. Till. 2001. Generic delimitations and molecular phylogeny in subfamily Tillandsioideae (Bromeliaceae) based on cpDNA sequences of atpB-rbcL intergenic spacer, trnL (UAA) intron and trnL-trnF (GAA) spacer, rps16 intron, matK and rbcL gene. P. 47 in: Systematics 2001, 3rd biennial conference of the Systematics Association (abstracts). London, UK: Imperial College, 3–7 September. 244