UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Molecular systematic studies in commelinid monocots Saarela, Jeffery Michael 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2007-267899.pdf [ 11.22MB ]
Metadata
JSON: 831-1.0100351.json
JSON-LD: 831-1.0100351-ld.json
RDF/XML (Pretty): 831-1.0100351-rdf.xml
RDF/JSON: 831-1.0100351-rdf.json
Turtle: 831-1.0100351-turtle.txt
N-Triples: 831-1.0100351-rdf-ntriples.txt
Original Record: 831-1.0100351-source.json
Full Text
831-1.0100351-fulltext.txt
Citation
831-1.0100351.ris

Full Text

M O L E C U L A R S Y S T E M A T I C STUDIES IN C O M M E L I N I D M O N O C O T S by Jeffery Michael Saarela B. Sc. (Hons.), University of Manitoba, 2000 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in The Faculty of Graduate Studies (Botany) THE UNIVERSITY OF BRITISH COLUMBIA December 2006 © Jeffery Michael Saarela, 2006 A B S T R A C T To clarify phylogenetic relationships among the major commelinid monocot lineages, among families in the orders Commelinales and Poales, and among grass subfamilies (Poaceae), I surveyed multiple plastid protein-coding regions and associated noncoding regions from exemplar taxa sampled broadly across these lineages. I also characterized phylogenetic relationships in the grass genus Bromus, using two plastid regions and a nuclear region. In commelinids, phylogenetic inferences using parsimony and likelihood are generally congruent, and my analyses provide strong support for many aspects of commelinid relationships. Commelinales and Zingiberales are sister taxa, but the positions of Arecales and Dasypogonaceae are not clear. Commelinales includes a Commelinaceae-Hanguanaceae clade and a (Philydraceae, (Haemodoraceae-Pontederiaceae)) clade. Philydrella is the sister group of the rest of Philydraceae. Poales includes a cyperid clade (Thurniaceae, . (Cyperaceae-Juncaceae)), and a 'core Poales' clade consisting of a graminid clade (Flagellariaceae, (Joinvilleaceae, (Ecdeiocoleaceae-Poaceae))) and a restiid clade (Anarthriaceae, (Centrolepidaceae-Restionaceae)). Ecdeiocoleaceae are moderately supported as the sister group of Poaceae in analyses in which taxon.sampling is most dense. The position of the aquatic Mayacaceae differs between parsimony and likelihood analyses, perhaps reflecting a long-branch artifact. Rapateaceae are part of a clade that includes all Poales except Bromeliaceae and Typhaceae, and the positions of these latter two families at the base of the Poales subtree are unresolved. Within Poaceae, Anomochlooideae and the BEP clade are supported strongly, Anomochlooideae, Pharoideae, and Puelioideae are the ii successive sister groups (respectively) of the remaining grasses, and Bambusoideae and Pooideae are sister taxa. A generally accelerated substitution rate in Poaceae plastid genomes is shared with most, but not all, Poales lineages. Independent analyses of plastid and nuclear ribosomal data in Bromus indicate that several traditionally recognized sections are monophyletic, but sect. Bromopsis is not. There is some evidence of incongruence between plastid and nuclear linkage groups in Bromus, particularly with respect to relationships among major lineages. Based on plastid data, the dwarf aquatic family Hydatellaceae are not commelinids or (more broadly) monocots, but the sister group of water lilies (Nymphaeales), a result corroborated by nuclear data and morphology. iii T A B L E O F C O N T E N T S ABSTRACT , , ii T A B L E OF CONTENTS : iv LIST OF TABLES : ix LIST OF FIGURES xi PREFACE xiii ACKNOWLEDGEMENTS xiv DEDICATION xvi CO-AUTHORSHIP STATEMENT xvii CHAPTER 1. INTRODUCTION 1 1.1. Overview of Monocot Systematics 1 1.2. Objectives of the Thesis 3 1.3 References 4 CHAPTER 2. M O L E C U L A R PHYLOGENETICS OF BROMUS (POACEAE: POOIDEAE) BASED ON CHLOROPLAST AND N U C L E A R DNA SEQUENCE D A T A 8 2.1. Introduction 8 2.1.1. Phylogenetic Position 9 2.1.2. Taxonomy and Classification 9 2.1.3. Phylogenetic Relationships 15 2.1.4. Objectives of the Study 16 2.2. Materials and Methods 17 2.2.1. Taxon Sampling 17 iv 2.2.2. DNA Extraction, Amplification, and Sequencing 18 2.2.3. Phylogenetic Analyses 19 2.3. Results 20 2.3.1. Analyses of ITS Sequences 20 2.3.2. Analyses of Plastid Sequences 21 2.3.3. Incongruence Among Data Partitions 23 2.4. Discussion 24 2.4.1. Phylogenetic Utility of the Regions Examined 24 2.4.2. Incongruence Between Nuclear Ribosomal and Plastid Data Partitions... .25 2.4.3. Phylogeny and Classification 26 2.4.4. Conclusions and Future Directions 37 2.5. References 51 CHAPTER 3. PHYLOGENETIC RELATIONSHIPS A M O N G THE SUBFAMILIES OF GRASSES (POACEAE) AND R E L A T E D FAMILIES INFERRED FROM A L A R G E PLASTID D A T A SET 68 3.1. Introduction 68 3.1.1. The Grass Family (Poaceae) 68 3.1.2. The Sister-Group of Poaceae 70 3.1.3. Objectives of the Study 71 3.2. Materials and Methods 72 3.2.1. Taxonomic Sampling 72 3.2.2. Genomic Sampling 73 3.2.3. Data Assembly 74 v 3.2.4. Parsimony, Likelihood, and Bayesian Phylogenetic Analyses 75 3.3. Results 78 3.3.1. Sequence Characteristics 78 3.3.2. Phylogenetic Analyses 79 3.4. Discussion 83 3.4.1. Molecular Evolution and Phylogenetic Utility of the Plastid DNA Regions in Poaceae 83 3.4.2. Phylogenetic Relationships in Poaceae 85 3.4.3. The Sister Group of the Grasses 90 3.4.4. Congruence and Discordance in the Data 92 3.5. References 108 CHAPTER 4. HIGHER-ORDER PHYLOGENETIC RELATIONSHIPS IN THE COMMELINID MONOCOTS, WITH A FOCUS ON THE ORDERS COMMELINALES A N D POALES 127 4.1. Introduction 127 4.1.1. The Commelinid Monocots 127 4.1.2. Phylogenetic Relationships in the Order Commelinales 130 4.1.3. Phylogenetic Relationships in the Order Poales 132 4.1.4. Objectives of the Study 138 4.2. Materials and Methods 139 4.2.1. Taxonomic Sampling 139 4.2.2. Genomic Sampling 140 4.2.3. Data Assembly 141 4.2.4. Parsimony and Likelihood Phylogenetic Analyses 142 4.3. Results 144 4.3.1. Outgroup Relationships 144 4.3.2. Relationships in the Commelinid Monocots 145 4.3.3. Phylogenetic Relationships in Commelinales 146 4.3.4. Phylogenetic Relationships in Poales 146 4.4. Discussion 148 4.4.1. Phylogenetic Relationships Among the Major Lineages of Commelinid Monocots 148 4.4.2. Phylogenetic Relationships in Commelinales 150 4.4.3. Higher-Order Relationships in Poales 156 4.4.3.1. The Core Poales 157 4.4.3.2. The Graminid Clade 157 4.4.3.3. The Restiid Clade 164 4.4.3.4. The Cyperid Clade 166 4.4.3.5. The Xyrid Clade 169 4.4.3.6. The Basal Poales 173 4.4.4. Molecular Evolution in the Commelinid Monocots 175 4.5. References 197 CHAPTER 5 . H Y D A T E L L A C E A E A R E NOT MONOCOTS BUT A LINEAGE NEAR THE BASE OF ANGIOSPERM PHYLOGENY 226 5.1. Introduction 226 5.2. Materials and Methods 228 vii 5.2.1. Taxonomic and Genomic Sampling 228 5.2.2. Molecular Phylogenetic Analyses 230 5.2.3. Morphological Phylogenetic Analyses 232 5.3. Results 233 5.3.1. Molecular Analyses 233 5.3.2. Morphological Analyses 234 5.4. Discussion.... 235 5.5. References 253 CHAPTER 6. CONCLUSION 260 6.1. Conclusions and Future Directions....... 260 6.2. References 268 viii LIST O F T A B L E S Table 2.1. Summary of infrageneric classifications and generic segregations of Bromus following Smith (1970), Tsvelev (1976), and Stebbins (1981) 39 2.2. Number of species, morphological characteristics, and native geographic distribution of sections of Bromus. The classification follows Smith (1970, 1985a) and Scholz (1998) 40 2.3. Bromus sections (following Smith 1970 and Scholz 1998) and species sampled, sources of material, vouchers, and GenBank accession numbers for DNA sequences 42 3.1. GenBank accession numbers and vouchers for sequence data from exemplar species in Poaceae and Joinvilleaceae 94 3.2. Number of characters (aligned and unaligned), potentially parsimony informative characters (PPIC), percentage of total PPIC, and number of PPIC per unaligned nucleotide in various partitions of the data 96 3.3. Likelihood ratio test (LRT) assessing heterogeneity among hierarchical lineages of monocots using the GTR + Y + I model of evolution with and without enforcement of a molecular clock. Likelihood scores were estimated based on a randomly chosen most-parsimonious tree 97 ix 4.1. Families recognized currently in Commelinales and Poales (sensu APG II 2003), and their approximate number of genera and species. Indented families have been recognized recently in some treatments, but were included in the previous taxon by APG II (2003) 178 4.2. Voucher information and GenBank accession numbers for plastid regions for exemplar commelinid monocot taxa in Poales and Commelinales 181 4.3. Additional eudicot and monocot taxa included in phylogenetic analyses with GenBank accession numbers and references. All data are from complete plastid genomes, except for Ranunculus macranthus 185 5.1. Morphological characters and character states scored for Hydatellaceae 240 x LIST O F FIGURES Figure 2.1. Phylogram of one of 449 most-parsimonious (MP) trees found using data from the internal transcribed spacer (ITS) region of nuclear ribosomal DNA 47 2.2. Phylogram of one of 50,000 most-parsimonious (MP) trees using combined data from the trriL intron and the 3'-end of ndhF 49 3.1. Phylogram of one of six most-parsimonious trees inferred for Poaceae and other monocots, based on a large plastid data set (atpB, ndhB, ndhF, ten photosystem genes, rbcL, rps7, y-rpsYl, ndhB and various introns and noncoding regions) 98 3.2. One of two most parsimonious trees inferred from combined data from 17 chloroplast genes for Poaceae and related families in the commelinid monocots 100 3.3. Phylogenetic tree obtained from Bayesian analysis of combined data from 17 chloroplast genes for Poaceae and related families in the commelinid monocots.... 102 3.4. Ranked profiles of clade support from analyses of various data partitions, as measured by (A) bootstrap proportions from parsimony analyses; (B) bootstrap proportions from likelihood analyses; and (C) posterior probabilities from Bayesian analyses 104 4.1. One of 28 most-parsimonious trees inferred from a large plastid data set (atpB, ndhB, ndhF, rbcL, ten photosystem II genes, rpYl, rpsl, and V-rps\2, and several noncoding regions in the inverted repeat; see text) showing relationionships among the angiosperms 187 xi 4.2. One of 28 most-parsimonious trees inferred from a large plastid data set showing relationships among the major angiosperm lineages (eudicots, monocots, magnoliids, and basal angiosperms) 189 4.3. One of 12 most-parsimonious trees inferred from a large plastid data set showing relationships among the major monocot lineages 191 4.4. One of 12 most-parsimonious trees inferred from a large plastid data set showing relationships among the commelinid monocot lineages 193 4.5. Optimal tree from maximum likelihood analysis inferred from a large plastid data set showing relationships among commelinid monocots 195 5.1. Phylogenetic placement of Trithuria submersa (Hydatellaceae) in angiosperms according to plastid data (multiple protein-coding genes and several noncoding regions) 243 5.2. Phylogenetic placement of Trithuria submersa (Hydatellaceae) in angiosperms according to analyses of the nuclear phytochrome C locus (PHYC) 245 5.3. Phylogenetic placement of Hydatella inconspicua (Hydatellaceae) for analyses of the trnT-trnL-trnF region, including the trnL-trnF region for H. inconspicua 247 5.4. Most parsimonious position of Hydatellaceae based on morphology, with relationships of other taxa not constrained 249 5.5. Most parsimonious position of Hydatellaceae based on morphology (774 steps), with relationships of other taxa constrained 251 xii P R E F A C E " If organisms have reached their present state by a process of evolution, it follows that they have a built-in classification, and man's problem is to find it. " R. Holttum, 1967 "Systematics is . . . a synthetic interdisciplinary science that crowns the whole edifice of biology. " Armen Takhtajan, 1997 "Most major civilizations are based on the triploid endosperm of grasses. " David Mabberley, 1997 A C K N O W L E D G E M E N T S I am grateful to my Ph.D. supervisor, Dr. Sean W. Graham, for mentorship and guidance over the last five and a half years, for teaching me how to write, and for providing room (and money) to pursue my research interests in his laboratory. I also thank the members of my supervisory committees at the University of Alberta (Drs. Ruth Stockey and Felix Sperling) and the University of British Columbia (Drs. Jeannette Whitton, Quentin Cronk, and Wayne Maddison) for their contributions to my thesis work. Hardeep Rai and Marc McPherson generously showed me the ropes in the molecular laboratory, and Hardeep Rai has steadfastly helped me overcome numerous technical limitations. Special thanks to the many people who have joined me in the field on various collecting trips: Paul Peterson, Chris Sears, Stephen Smith, Jack Maze, Amanda Roe, Keith Todd, Heather Whitehouse, Philip Ferreira, and Diana Percy. Conversations with Jack Maze have been particularly stimulating. I am grateful to Kare Bremer, Barbara Briggs, Jacques Cayouette, Lynn Clark, Jerry Davis, Mel Duvall, Ryan Keane, Adam Marchant, and Paul Peterson for providing DNA and leaf tissue samples. I thank Dorothy Fabijan (ALTA), Cindy Sayre (UBC), and Linda Jennings (UBC) for support in the herbarium. My laboratory, herbarium, and field work has been supported by Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery grants to Sean Graham, Alberta Ingenuity, the University of Alberta, the University of British Columbia, the Department of Botany (University of British Columbia), the Robert Bateman Arctic Research Fund (Smithsonian Institution), the Lawrence R. Heckard Fund of the Jepson Herbarium (University of California, Berkeley), and Genome Canada. Personal financial support was xiv provided by a Postgraduate Scholarship (PGS-A) and a Canada Graduate Scholarship (CGS) from NSERC, a Walter H. Johns Fellowship (University of Alberta), an Alberta Ingenuity Incentive Award, the Faculty of Land and Food Systems (University of British Columbia), a University Graduate Fellowship (University of British Columbia), and an NSERC Discovery Grant to Sean W. Graham. xv DEDICATION For my Mom and Dad CO-AUTHORSHIP STATEMENT Chapter 2 is based on a manuscript that has been accepted for publication: Saarela, J. M . , Peterson, P. M. , Keane, R. M. , Cayouette, J., and Graham, S. W. In Press. Molecular phylogenetics of Bromus (Poaceae: Pooideae) based on chloroplast and nuclear DNA sequence data. In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative Biology and Evolution, Poales. Rancho Santa Ana Botanic Garden, Claremont, California, U. S. A. The project was suggested by P. M . Peterson. I conducted all laboratory work and data analyses, and wrote the manuscript. Most of the plant samples were contributed by P. M . Peterson, R. M . Keane, and J. Cayouette. Sean W. Graham provided insights into data analyses and contributed to the writing. Chapter 3 is a draft manuscript that will be submitted for publication: Saarela, J. M . , and Graham, S. W. Phylogenetic relationships among the subfamilies of grasses (Poaceae) and related families inferred from a large plastid data set. I designed the project in consultation with S. W. Graham, and I conducted all laboratory work and data analyses, and wrote the manuscript. Sean W. Graham provided insights into data alignment and analyses, and he contributed to the writing. Chapter 4 is a draft manuscript that will be submitted for publication: Saarela, J .M., Rai, H. S., Chang, Y., Prentis, P., Marchant, A., Briggs, B. G., and Graham, S. W. Higher-order phylogenetic relationships in the commelinid monocots, with a focus on the orders Commelinales and Poales. Sean W. Graham and I designed the project, and I carried out the xvii laboratory work (H. S. Rai contributed sequence data for three taxa included in this study, and Y. Chang contributed partial sequence data for one taxon). Critical DNA samples were provided by P. Prentis, A. Marchant, and B. G. Briggs. Sean W. Graham and I compiled alignments, and I conducted all analyses and wrote the manuscript, with contributions from S. W. Graham. Chapter 5 is based on a manuscript that is in review: Saarela, J. M . , H. S. Rai, J. A. Doyle, P. K. Endress, S. Mathews, A. D. Marchant, B. G. Briggs, and S. W. Graham. Hydatellaceae are not monocots but a lineage near the base of angiosperm phlogeny. I generated all of the plastid molecular data, except the plastid ndhF gene and short rbcL gene (generated by H. S. Rai) and the plastid trriL-trnF region (generated by A. Marchant and B. G. Briggs). The nuclear PHYC data was contributed by S. Mathews, and morphological characters were scored by J. A. Doyle and P. K. Endress. All authors contributed to analyses and preparation of the final manuscript. xviii C H A P T E R 1 Introduction 1.1 Overview of Monocot Systematics The monocots (Monocotyledoneae) are a large and diverse clade of flowering plants with -60,000 species in -92 families (APG II 2003), representing -25% of total angiosperm diversity. They include many familiar ornamental and horticultural taxa, such as bananas, bromeliads, cattails, grasses, lilies, onions, orchids, and palms. They are readily distinguished from other angiosperms by their (usually) parallel-veined leaves, sheathing leaf bases, stems with scattered bundles, no vascular cambium, sieve cell plastids with cuneate proteinaceous crystalloids (i.e., P2-type), and embryos with a single cotyledon (Judd et al. 2002; Chase 2004). Among flowering plants, monocots are arguably the best-studied major lineage (Chase 2004); nonetheless, substantial questions remain about several aspects of their evolutionary history. Molecular phylogenetic studies (e.g., Chase et al. 1995, 2000, 2006; Givnish et al. 1999, 2006; Fay et al. 2000; Bremer 2002; Caddick et al 2002; Davis et al. 2004; Tamura et al. 2004; Graham et al. 2006) have identified 11 major monocot lineages as orders (Acorales, Alismatales, Arecales, Asparagales, Commelinales, Dioscoreales, Liliales, Petrosaviales, Poales, and Zingiberales; APG II 2003, Chase 2004). The small family, Dasypogonaceae, remains unplaced to order (APG II 2003). Most studies agree that Acorus (Acorales) is the sister-group of the rest of the monocots, and that Alismatales and Petrosaviales respectively are the next successive sister, groups of a large clade that includes Asparagales and the commelinid monocots (e.g., Chase et al. 2006; Givnish et al. 2006; Graham et al. 2006). The 1 relative arrangements of Dioscoreales, Liliales, and Pandanales along the monocot backbone are still in question (e.g., Graham et al. 2006). The commelinid monocots - a large and heterogeneous group that includes about one third of monocot diversity - consist of the remaining four orders and Dasypogonaceae (Chase et al. 1995; APG II 2003). A roughly similar group of taxa was recognized by Dahlgren and Rasmussen (1983) based on morphology, and referred to subsequently as the Bromelianae-Commelinanae-Zingiberanae (BCZ) or Arecanae-Bromelianae-Commelinanae-Zingiberanae (ABCZ) clade (e.g., Clark et al. 1993). Although support for the commelinid clade is strong, several aspects of relationship among and within its major lineages are not clear, such as the relative arrangement of most of its five major clades, and relationships among families within Commelinales and Poales, the order that includes the grasses (Poaceae). The grasses are a large and taxonomically difficult commelinid family. Because of their economic and ecological importance, Poaceae have received substantial systematic study, and multiple major lineages (subfamilies) in the family have been identified. Relationships among several of these, however, are not clear (GPWG 2001; Duvall et al. 2006). Moreover, with 600 to 800 genera and approximately 11,000 species, only a fraction of the species has been included in molecular phylogenetic studies. One major genus is Bromus, a large (-160 species) and widespread taxon that includes important forage grasses (such as B. inermis Leyss.) and weeds of major economic concern [such as cheatgrass (B. tectorum L.) and ripgut grass (B. diandrus Roth.)], in addition to a number of species that have narrow distributions. Bromus is also the sister group of a commercially important clade (tribe Triticeae) that includes barley, rye, and wheat. 2 Reconstructions of monocot phylogeny at multiple hierarchical levels have provided evolutionary frameworks for identifying major monocot lineages and their inter-relationships, and for reconstructing the origins of several floral and other morphological characters (e.g., Furness and Rudall 1998, 2003; Smith and Harris 1999; Zona 2001; Prychid et al. 2003; Rudall and Bateman 2004; Linder and Rudall 2005; Gunawardena and Dengler 2006). Within Poaceae, phylogenies have permitted researchers to develop (and test) hypotheses about the origin and diversification of the grass spikelet, the organ in which cereal grains are produced (e.g., Rudall et al. 2005; Preston and Kellogg 2006). Better sampled and more robust inferences of monocot evolution should provide an even firmer footing for performing these types of studies. 1.2 Objectives of the Thesis The objectives of this thesis are to characterize phylogenetic relationships at a variety of hierarchical levels in commelinid monocots: among species in the grass genus, Bromus (Chapter 2); among subfamilies in the grass family, Poaceae (Chapter 3); and among and within major commelinid monocot lineages (Chapter 4). I also present a surprising discovery involving the reduced aquatic family Hydatellaceae, thought to belong to the commelinid monocots, but shown here to belong to the basal angiosperm grade that includes Amborellaceae and water lilies (Nymphaeales) (Chapter 5). 3 1.3 R E F E R E N C E S Angiosperm Phylogeny Group II [APG II]. 2003. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG II. Botanical Journal of the Linnean Society 141, 399-436. Bremer, K. 2002. Gondwanan evolution of the grass alliance of families (Poales). Evolution 56, 1374-1387. Caddick, L. R., Rudall, P. J., Wilkin, P., Hedderson, T. A. J., and Chase, M . W. 2002. Phylogenetics of Dioscoreales based on combined analyses of morphological and molecular data. Botanical Journal of the Linnean Society 138, 123-144. Chase, M . W. 2004. Monocot relationships: an overview. American Journal of Botany 91, 1645-1655. Chase, M . W., Stevenson, D. W., Wilkin, P., and Rudall, P. J. 1995. Monocot systematics: a combined analysis. Pp. 685-740 In: Rudall, P. J., Cribb, P. J., Cutler, D. F., and Humphries, C. J. [Eds.]. Monocotyledons: Systematics and evolution. Royal Botanic Gardens, Kew. Chase, M . W., Soltis, D. E. , Soltis, P. S., Rudall, P. J., Fay, M . F., Hahn, W. H., Sullivan, S., Joseph, J., Molvray, M . , Kores, P. J., Givnish, T. J., Sytsma, K. J., and Pires, J. C. 2000. Higher-level systematics of the monocotyledons: an assessment of current knowledge and a new classfication. Pp. 3-16 In: Wilson, K. L. , and Morrison, D. A. (Eds.). Monocots: Systematics and Evolution. CSIRO, Melbourne, Australia. Chase, M . W., Fay, M . F., Devey, D., Ronsted, N., Davies, J., Pillon, Y., Petersen, G., Seberg, O., Tamura, M . N., Asmussen, C. B., Hilu, K., Borsch, T., Davis, J. I., Stevenson, D. W., Pires, J. C , Givnish, T. J., Sytsma, K. J., McPherson, M . A., 4 Graham, S. W., and Rai, H. S. 2006. Multigene analyses of monocot relationships: A summary. In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Clark, W. D., Gaut, B. S., Duvall, M . R., and Clegg, M . T. 1993. Phylogenetic relationships of the Bromeliiflorae-Commeliniflorae-Zingiberiflorae complex of monocots based on rbcL sequence comparisons. Annals of the Missouri Botanical Garden 80, 987-998. Dahlgren, R. M . T., and Rasmussen, F. N. 1983. Monocotyledon evolution: characters and phylogenetic estimation. Evolutionary Biology 16, 255-395. Davis, J. I., Stevenson, D. W., Peterson, G., Seberg, O., Campbell, L. M . , Freudenstein, J. V. , Goldman, D. H., Hardy, C. R., Michelangeli, F. A., Simmons, M . P., Specht, C. D., Vergara-Silva, F., and Gandolfo, M . A. 2004. A phylogeny of the monocots, as inferred from rbcL and atpA sequence variation. Systematic Botany 29, 467-510. Duvall, M . R., Davis, J. I., Clark, L. G., Noll, J. D., Goldman, D. H., and Sanchez-Ken, J. G. 2006. Phylogeny of the grasses (Poaceae) revisited. In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. (Eds.). Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Fay, M . F., Rudall, P. J., Sullivan, S., Stobart, K. L. , de Bruijn, A. Y., Reeves, G., Qamaruz-Zaman, F., Hong, W.-P., Hahn, W. J., Conran, J. G., and Chase, M . W. 2000. Phylogenetic studies of Asparagales based on four plastid DNA regions. Pp. 360-371 5 In: Wilson, K. L., and Morrison, D. A. [Eds.]. Monocots: Systematics and Evolution. CSIRO Publishing, Collingwood, Victoria, Australia. Furness, C. A., and Rudall, P. J. 1998. The tapetum and systematics in monocotyledons. Botanical Review 64, 201-239. Furness, C. A., and Rudall, P. J. 2003. Apertures with lids: distribution and significance of operculate pollen in monocotyledons. International Journal of Plant Sciences 164, 835-854. Givnish, T. J., Evans, T. M . , Pires, J. C , and Sytsma, K. J. 1999. Polyphyly and convergent morphological evolution in Commelinales and Commelinidae: evidence from rbcL sequence data. Molecular Phylogenetics and Evolution 12, 360-385. Givnish, T. J., Pires, J. C , Graham, S. W., McPherson, M . A., Prince, L. M . , Patterson, T. B., Rai, H. S., Roalson, E. H., Evans, T. E. , Hahn, W. J., Millam, K. C , Meerow, A. W., Molvray, M . , Kores, P. J., O'Brien, H. E. , Hall, J. C , Kress, W. J., and Sytsma, K. J. 2006. Phylogenetic relationships of monocots based on the highly informative plastid gene ndW: evidence for widespread concerted convergence. Pp. 27-50 In: Columbus, J. T., Friar, E . A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Graham, S. W., Zgurski, J. M . , McPherson, M . A., Cherniawsky, D. M . , Saarela, J. M . , Home, E. C , Smith, S. Y., Wong, W. A., O'Brien, H., Biron, V. L., Pires, J. C , Olmstead, R. G., Chase, M . W., and Rai, H. S. 2006. Robust inference of monocot deep phylogeny using an expanded multigene plastid data set. Pp. 3-20 In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. 6 Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Grass Phylogeny Working Group [GPWG]. 2001. Phylogeny and subfamilial classification of the grasses (Poaceae). Annals of the Missouri Botanical Garden 88, 373-457 Gunawardena, A. H. L. A. N., and Dengler, N. G. 2006. Alternative modes of leaf dissection in monocotyledons. Botanical Journal of the Linnean Society 150, 25-44. Judd, W. J., Campbell, C. S., Kellogg, E. A., Stevens, P. F., and Donoghue, M . J. 2002. Plant Systematics - A Phylogenetic Approach. 2 n d Edition. Sinauer Associates, Sunderland, Massachusetts. Linder, H. P., and Rudall, P. J. 2005. Evolutionary history of Poales. Annual Review of Ecology, Evolution, and Systematics 35, 107-124. Prychid, C. J., Rudall, P. J., and Gregory, M . 2003. Systematics and biology of silica bodies in monocotyledons. Botanical Review 69, 377-440. Rudall, P. J., and Bateman, R. M . 2004. Evolution of zygomorphy in monocot flowers: iterative patterns and developmental constraints. New Phytologist 162, 25-44. Smith, B. G., and Harris, P. J. 1999. The polysaccharide composition of Poales cell walls: Poaceae cell walls are not unique. Biochemical Systematics and Ecology 27, 33-53. Tamura, M . N., Yamashita, J., Fuse, J., and Haraguschi, M . 2004. Molecular phylogeny of the monocotyledons inferred from combined analysis of plastid matK and rbcL gene sequences. Journal of Plant Research 117,109-120. Zona, S. 2001. Starchy pollen in commelinoid monocots. Annals of Botany 87, 109-116. 7 CHAPTER 2 1 Molecular Phylogenetics of Bromus (Poaceae: Pooideae) Based on Chloroplast and Nuclear DNA Sequence Data 2.1 INTRODUCTION Bromus L. is a large genus that is widely distributed in temperate and mountainous regions of the Northern and Southern hemispheres. Several species are important forage grasses (e.g., Ferdinandez and Coulman 2000; Ferdinandez et al. 2001; Puecher et al. 2001); some were important cereal crops in the past (Scholz and Mos 1994); and many are invasive weeds (e.g., Ainouche et al. 1999; Novak and Mack 2001; Keane 2002; Ogle et al. 2003). Bromus is distinguished from other grass genera by the combination of several morphological characters, including: leaf sheath margins that are connate for most of their length; awns that are almost always subapically inserted; hairy apical bilabiate appendages of the ovary; and simple starch grains (Wagnon 1952; Smith 1970). A version of this chapter has been accepted for publication: Saarela, J . M . , Peterson, P. M . , Keane, R. M . , Cayouette, J., and Graham, S. W. In Press. Molecular phylogenetics of Bromus (Poaceae: Pooideae) based on chloroplast and nuclear DNA sequence data. In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative Biology and Evolution (Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. 8 2.1.1. Phylogenetic Position The eastern Asian genus Littledalea Hemsl., with three species, was believed by Stebbins (1981) to be the closest living relative of Bromus; tribe Bromeae currently comprises these two genera (Smith 1970; Clayton and Renvoize 1986; Tsvelev 1989; Grass Phylogeny Working Group [GPWG] 2001). However, preliminary plastid sequence data indicate that Littledalea and Bromus do not form a clade, thus Bromeae may not be monophyletic (J. M . Saarela, unpubl. data). Other genera believed previously to be closely related to Bromus, based on morphological similarities, include Megalachne Steud., Metcalfia Conert, Pseudodanthonia Bor & C. E. Hubb., and Sinochasea Keng (Smith 1970; Stebbins 1981), but these genera are now considered distantly related (Clayton and Renvoize 1986; Soreng et al. 2003). Phylogenetic analyses of chloroplast DNA restriction site variation and DNA sequence data indicate that Bromeae are the sister group of Triticeae (e.g., Davis and Soreng 1993; Catalan et al. 1997; Hilu et al. 1999; Hsiao et al. 1999; Soreng and Davis 1998, 2000; GPWG 2001). 2.1.2. Taxonomy and Classification Bromus is a taxonomically difficult genus with a complex nomenclatural history [see Wagnon (1952), Smith (1970), and Acedo and Llamas (1999) for comprehensive reviews], and many species are difficult to distinguish due to their high degree of morphological similarity. As with many other genera of grasses, many species are polyploids, and hybridization is believed to have played an important role in the evolution of many species in the genus (Stebbins 1981). The complexity of Bromus is exemplified in the more than 1200 taxa that have been described, according to the International Plant Names Index (2004). The 9 most recent estimates of the number of species in the genus are 160 (Acedo and Llamas 2001) and 142 (Clayton et al. 2002 onwards), although estimates have ranged from around 100 (Gould and Shaw 1983) to 400 (Soderstrom and Beaman 1968). Several species complexes have been the subject of recent taxonomic investigations (e.g., Scholz 1981; Naranjo et al. 1990; Sales and Smith 1990; Sales 1993, 1994a; Smith and Sales 1993; Zajac 1996a, b; Allison et al. 2001; Bacic and Jogan 2001; Peterson et al. 2002; Spalton 2002a; J. M . Saarela and P. M . Peterson, unpubl. data), and new taxa continue to be collected and described (e.g., Smith 1985a; Veldkamp et al. 1991; Peterson and Planchuelo 1998; Scholz 1997, 1998; Acedo and Llamas 1997; Bomble and Scholz 1999; Holmstrom and Scholz 2000; Spalton 2001; J. M. Saarela and P. M . Peterson, unpubl. data). Because of its large size, taxonomical complexity, and wide geographic range, no comprehensive worldwide treatment of all the species in Bromus exists, but many floristic treatments and keys of Bromus have been published for various geographic regions in the New World (e.g., Shear 1900; Wagnon 1952; Mitchell 1967; Soderstrom and Beaman 1968; Pinto-Escobar 1981, 1986; Matthei 1986; Allred 1993; Pavlick 1995; Gutierrez and Pensiero 1998; Planchuelo and Peterson 2000) and the Old World (e.g., Veldkamp et al. 1991; Forde and Edgar 1995; Chen and Kuoh 2000; Spalton 2002b, 2004). Genetic variation within and among many species has been studied using data from isozymes (Kahler et al. 1981; Ainouche et al. 1995, 1999; Oja 1998, 1999, 2002a, b, 2005; Bartlett et al. 2002), as well as an array of DNA-based molecular techniques, including RAPDs and AFLPs (Ferdinandez et al. 2001; Massa et al. 2001; Puecher et al. 2001; Ferdinandez and Coulman 2002) and microsatellites (Green et al. 2001; Ramakrishnan et al. 2002). A physical map of the chloroplast genome has been constructed for one species, B. inermis (Pillay 1993). 10 The infrageneric classification of Bromus has received considerable study. The genus has been variously split into several groups that have been recognized as sections, subgenera, or generic segregates (Table 2.1). Smith (1970) reviewed the morphological characteristics and nomenclature of the commonly recognized groups in the genus, and accepted five distinct sections, characterized by minor differences in the structures of the spikelets. Using data from crossing experiments, Stebbins (1981) recognized seven subgenera (although one, subgenus Boissiera, is not validly published at this rank) based on their morphological distinctiveness and the apparent high degree of genetic divergence among them. He argued that the subgenera of Bromus are too distinct to be treated as sections, since they seemed more distantly related to one another than are several other genera of grasses. Other authors believe that each of these groups is sufficiently distinct to be regarded as genera (e.g., Tsvelev 1976). No taxonomic consensus exists, and infrageneric taxa in Bromus are recognized currently as distinct genera (e.g., Catalan et al. 1997; Green et al. 2001; Spalton 2002b, 2004), subgenera (e.g., Acedo and Llamas 1999), or sections (e.g., Smith 1985b; Pavlick 1995; Planchuelo and Peterson 2000). The sectional classification of Smith (1970) has been followed by most recent North American authors, and is employed here, incorporating the recent modifications of Smith (1985a) and Scholz (1998); all species mentioned below are species of Bromus. Each section of Bromus can be identified using a combination of several morphological characters, including the number of nerves on the first and second glumes, spikelet shape and compression, and lemma and awn morphology (Table 2.2). Additional data from embryo morphology (Kosina 1996), floral microstructures (Kosina 1999), micromorphology of the lemmas and paleas (Acedo and Llamas 2001), and anatomy (Acedo 11 and Llamas 1999) have recently been collected to aid in the infrageneric classification. Insights obtained from these studies generally agree with the classification schemes based on macromorphological evidence. Section Bromopsis is the largest section, comprising approximately 60 species that occur naturally in Eurasia, Africa, and North and South America, and thus occurs in all regions where brome grasses are native (Stebbins 1981; Armstrong 1991). The section includes diploids, tetraploids, hexaploids, octoploids, and decaploids (Stebbins 1981). Section Bromopsis comprises at least two geographically, morphologically, and cytologically distinct groups. North American taxa, and the B. ramosus complex from the Old World, are loosely tufted (non-rhizomatous), short-lived perennials or biennials [except B. texensis (Shear) A. S. Hitch., an annual] with small anthers and large chromosomes, and the majority are diploids (Wagnon 1952; Armstrong 1981, 1983, 1991; Stebbins 1981). Old World taxa and B. pumpellianus, which occurs in North America and the Old World, are densely tufted or rhizomatous long-lived perennials with large anthers and smaller chromosomes, and the majority are polyploids (tetra-, hexa-, octo-, and decaploids) (Wagnon 1952; Armstrong 1981, 1983, 1991; Stebbins 1981). Armstrong (1983, 1991) suggested that these two groups might have separate evolutionary histories, based on difficulties in crossing North American and Eurasian taxa, and noted that valid names are available at sectional rank for each of these groups if such recognition becomes appropriate. Cytology and evolutionary relationships of the South American species are poorly known (Stebbins 1981). Section Bromus comprises 30—40 diploid and tetraploid annual species native to Europe and Asia. One species, B. arenarius Labill., is thought by some authors to be the only native Bromus species in Australia; others believe the species is introduced there (Stebbins 12 1981). Many species are invasive and widely distributed in other regions of the world (Keane 2002). For example, the 11 species of sect. Bromus that occur in North America are all introduced (Pavlick 1995). Species in the section are morphologically similar (e.g., Smith and Sales 1993; Oja et al. 2003), and several subsectional classifications of sect. Bromus have been proposed (Smith 1972). The tetraploid species in sect. Bromus are believed to be allopolyploids (Stebbins 1981), and their putative intrasectional origins have been elucidated using data from serology (Smith 1972), allozymes (Ainouche et al. 1995; Oja 1998) and DNA sequence data (Ainouche and Bayer 1997; Ainouche et al. 1999). One group of tetraploid species, the B. pectinatus complex, is believed to be of hybrid origin between sects. Bromus and Genea (Scholz 1981; Stebbins 1956, 1981). Section Ceratochloa comprises 10-16 perennial species native to North and South America. All taxa in this section are polyploids (octo-, hexa-, and 12-ploid) (Stebbins 1981; Pavlick 1995). Species boundaries in sect. Ceratochloa are uncertain due to presumed hybridization and morphological intergradation among taxa, which have resulted in various taxonomic treatments (e.g., Soderstrom and Beaman 1968; Stebbins 1981; Pavlick 1995; Planchuelo and Peterson 2000). Some species complexes in sect. Ceratochloa have recently been revised. Based on genetical and morphological studies of six hexaploid and octoploid species from Patagonia, Massa et al. (2001, 2004) distinguished only two morphologically and genetically distinct taxa, which they treated as two species. Similar revisionary work is necessary to characterize morphological and molecular variation among North American taxa of sect. Ceratochloa. Section Genea comprises diploid, tetraploid, hexaploid, and octoploid annual species native to the Mediterranean, southwestern Asia, northern Europe, and northern Africa. 13 Several species are invasive [e.g., cheatgrass (B. tectorum L.), ripgut grass (B. diandrus Roth.), and red brome (B. madritensis subsp. rubens (L.) Husnot)] and have become widely distributed far beyond their native range (e.g., Pavlick 1995). Species in sect. Genea are highly variable morphologically, and many taxa have been proposed. Recent revisionary work has reduced the number of species to five, including several intraspecific taxa (Sales 1993, 1994a). Section Genea is thus the only geographically widespread section of Bromus that has received monographic-level taxonomic attention. Based on this taxonomic framework, Sales (1994b) proposed hypotheses for the origins of taxa and patterns of adaptive radiation that have occurred within the section. Isozyme data have indicated that the three diploid species of sect. Genea are putative donors of genomes in the origins of the polyploid species in the section (e.g., Oja 1998, 2002b,c). The remaining sections in Bromus (Boissiera, Neobromus, Nevskiella, and Triniusia) are individually small, but contribute substantially to morphological variation in the genus as a whole (Table 2.2). Section Neobromus comprises two annual hexaploid species native to the Pacific coasts of North and South America (Pavlick 1995; Matthei 1986). Sections Nevskiella (diploid; Armstrong 1991) and Boissiera (diploid [Smith 1972] or tetraploid [Oja and Jaaska 1998]) are both monotypic, while section Triniusia comprises two diploid species (Scholz 1998); species in these three sections are all annuals native to Asia and the eastern Mediterranean. Sections Boissiera and Triniusia were included within sect. Bromus by Smith (1970), but they were treated as distinct subgenera by Stebbins (1981; Table 2.1). They have both recently been recognized as distinct sections by Smith (1985a) and Scholz (1998), respectively (Table 2.1). 14 2.1.3. Phylogenetic Relationships Past attempts to understand phylogenetic relationships in Bromus among species and infrageneric taxa have been based largely on data from morphology (e.g., Shear 1900; Wagnon 1952), karyology (including chromosome number, satellite type, chromosome size, and DNA quantities) and hybridization experiments (e.g., Stebbins and Togby 1944; Stebbins 1947, 1956, 1981; Schulz-Schaeffer 1960; Wilton 1965; Armstrong 1975, 1981, 1983; Kozuharov et al. 1981; Naganowska 1993a,b), serology (Smith 1969, 1972), and allozymes (e.g., Oja 1998; Oja and Jaaska 1998; Oja 2005). Chromosome numbers, polyploidy, genome size, karyotypes, c-banding, cross-compatibility, and genome homology within Bromus have been summarized by Armstrong (1991). Five systematic studies have been conducted in Bromus using data from DNA, although the number of species included in each study was relatively limited. Pillay and Hilu (1990, 1995) studied cpDNA restriction site variation among 32 Bromus species, and identified two major clades: one comprising sects. Ceratochloa and Neobromus, the other comprising sects. Bromopsis, Bromus, and Genea. Species of sect. Bromopsis occurred in three different lineages, indicating that this taxon is not monophyletic, but the data did not support the New World/ Old World split hypothesized by Armstrong (1983) on the basis of morphology and chromosome pairing data. Sections Bromus and Genea were not monophyletic; species from both sections were intermixed in a single clade. Joachimiak et al. (2001) used RAPD data to portray relationships among nine species representing four infrageneric taxa in Bromus from the New and Old Worlds. Based on a phenetic analysis, they identified two distinct clusters: one comprising sect. Ceratochloa, and the other comprising sects. Bromopsis, Bromus, and Genea. However, because of their small sample 15 size, and the low level of molecular divergence detected, they were unable to make definitive statements regarding relationships in Bromus. Ainouche and Bayer (1997) and Ainouche et al. (1999) used sequence data from the internal transcribed spacer (ITS) regions of nuclear ribosomal DNA to study the phylogeny of sect. Bromus. Based on an analysis of 22 species from sect. Bromus (including sect, triniusia) and three species from other infrageneric groups, they found sect. Bromus to be monophyletic. They also studied the origin of some tetraploid species in the section. Little sequence heterogeneity was detected within tetraploid species, and they found that the inclusion of allotetraploid taxa with diploid taxa did not change the underlying topology of the trees obtained, compared to trees obtained from analyses of the diploid taxa alone. 2.1.4. Objectives of the Study To further characterize phylogenetic relationships in Bromus s. 1., I obtained new sequence data, from the chloroplast trnLQJAA) intron, the rapidly evolving 3' end of the chloroplast ndhF gene, and the nuclear ribosomal internal transcribed spacer regions, from 46 exemplar Bromus species that represent a large proportion of the geographical and morphological diversity in the genus. The specific objectives of this study were to use DNA sequence data to: (1) test the monophyly of the currently recognized infrageneric groups in Bromus; and (2) determine phylogenetic relationships among infrageneric groups and species. 16 2.2 M A T E R I A L S AND M E T H O D S 2.2.1. Taxon Sampling Exemplars from each of the currently recognized sections in Bromus were included in this study, except for the two monotypic sections, Boissiera and Nevskiella. Attempts to extract DNA from a herbarium specimen of B. gracillimus Bunge (sect. Nevskiella) were unsuccessful, and material of B. pumilio (Trin.) P. M . Sm. (sect. Boissiera) was not available. Table 2.3 lists the species sampled [following the classification schemes of Smith (1970) and Scholz (1998)], their geographic origins, vouchers, and GenBank accession numbers for the sequences. One individual of each species was examined, except for B. madritensis subsp. rubens, for which three individuals were sampled, and B. anomalus, for which two individuals were sampled. Samples were obtained from silica-gel dried leaf material from field collections, from plants grown in the greenhouse from seed obtained from the Western Regional Plant Introduction Station (United States Department of Agriculture, Pullman, Washington, USA) and Plant Gene Resources of Canada (Saskatoon Research Centre, Agriculture and Agri-Food Canada, Saskatoon, Saskatchewan, Canada), and from herbarium specimens. All taxonomic identifications were confirmed using the available world taxonomic literature of Bromus. Outgroup taxa from tribes Triticeae and Poeae were chosen based on previous molecular investigations of the grasses (see Catalan et al. 1997; GPWG 2001). The Bromus and Festuca breviglumis sequences used in this study are new. Sequence data for Hordeum vulgare and Triticum aestivum were obtained from GenBank. 17 2.2.2. DNA Extraction, Amplification, and Sequencing Total genomic DNA was extracted using a modified cetyltrimethylammonium bromide (CTAB) method (Doyle and Doyle 1987) with 2% B-mercaptoethanol added to each extraction. DNA extracts and PCR amplifications were purified using a Qiagen PCR Purification kit (Qiagen, Santa Clarita, California, USA) following manufacturer instructions. The trriL (UAA) intron was amplified and sequenced with primers developed by Taberlet et al. (1991). The region I refer to as ITS, which includes two spacer regions, ITS 1 and ITS 2, and the 5.8S rDNA locus, was amplified and sequenced using primers published by White et al. (1990), Hsiao et al. (1994), and Blattner (1999). The 3' end of ndh¥ was amplified and sequenced using primers designed by Olmstead and Sweere (1994) and Graham et al. (1998). Amplification reactions consisted of 26:5 ul sterile water, 5ul lOx buffer, 4 ul 10 mM dNTPs, 3 uL 25 mM MgCl 2 , 5 uL of each 5 pmol/ul primer, 1 ul of template DNA, and 0.5 ul of Taq DNA polymerase (1 unit). The thermal profile was: 1 cycle of 3 min at 94°C; 35 cycles of: 30 sec at 94°C, 1 min at 42.5°C, 2 min at 72°C; and 1 cycle of 5 min at 72°C. Sequencing products were generated using a BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, California, USA) with 50 ng of template DNA and the following thermal profile: 25 cycles of 10 sec at 96°C, 5 sec at 45°C, and 4 min at 60°C. For each sample, one or several duplicate sequencing reactions were included using a second DNA extract from the same source material. Sequencing reactions were analyzed using an Applied Biosystems Prism 377 automated DNA sequencer. 18 Sequence data were assembled and edited using Sequencher 4.1 (Genes Code Corporation, Ann Arbor, Michigan, USA). Consensus sequences were exported for each taxon and aligned manually using Se-Al ver. 1.0 alpha 1 (Rambaut 1998) according to guidelines outlined in Graham et al. (2000). Gaps in the final matrix were coded as missing data. Several inferred indels in the trnL intron were scored as binary characters. Alignments were imported into PAUP* version 4.0M0 (Swofford 2002). All sequence data have been submitted to Genbank (Table 2.3). 2.2.3. Phylogenetic Analyses For the ITS data set, a heuristic search was conducted with 100 random starting trees, tree-bisection-reconnection (TBR) branch swapping, and all character and character-state changes equally weighted. A two-tier approach was taken for the heuristic searches of the combined plastid data because an upper limit on the number of most-parsimonious trees was unattainable with the available computational resources and time: (1) 100 independent heuristic searches each with a random starting tree, saving 100 trees each (MaxTrees set to 100), were performed with the parameters noted above; and (2) another heuristic search, with the same parameters as above, was conducted, except that the shortest of the 10,000 trees from step 1 were used as starting trees, and MaxTrees was set to 50,000. I also implemented the parsimony ratchet (Nixon 1999) using PAUP Rat (Sikes and Lewis 2001) to search for shorter trees with the combined plastid data set. The incongruence length difference (ILD) test (Farris et al. 1994, 1995) was used to test for conflict among plastid and nuclear data partitions, with MaxTrees set to 500. In addition, trees and bootstrap analyses derived from plastid and nuclear data were compared visually to assess the robustness of topological 19 incongruence (e.g., Graham et al. 1998). I computed strict consensus trees from all of the most-parsimonious trees for each of the data partitions. I present phylograms of one randomly chosen tree from each of these analyses, and indicate which clades on the phylograms collapse in the strict consensus trees. Branch support was assessed using maximum parsimony bootstrap analysis (Felsenstein 1985) from 500 replicates using the heuristic search option, with one random starting tree, T B R branch swapping, and MaxTrees set to 500 per replicate. I use the terms 'weak', 'moderate', and 'we l l ' to refer to bootstrap values that range from <50-70, 71-90, and 91-100, respectively. 2.3 RESULTS 2.3.1. Analyses of ITS Sequences The boundaries of ITS 1, 5.8S, and ITS 2 follow Eckenrode et al. (1985), Yokota et al. (1989), and Kolosha and Fodor (1990). Lengths of ITS 1 and ITS 2 ranged from 216-219 and 213-216 base pairs (bp), respectively. The 5.8S r R N A gene was 163 bp in length. A small region of 10 bp (positions 108-117) in ITS 1 was difficult to align across taxa, and was excluded from all analyses. The ITS data matrix, without excluded sites, was 606 aligned nucleotides in length. O f these characters, 382 were constant, 224 were variable, and 125 (20.6%) were parsimony informative. Among the ingroup taxa, 437 characters were constant, 169 were variable, and 104 (17.2%) were parsimony informative. The heuristic searches of the ITS data set recovered 449 most-parsimonious trees (tree length = 380 steps, consistency index [CI] = 0.713, retention index [RI] - 0.826). Several clades receive good bootstrap support (Fig. 2.2.1). The monophyly of the genus is moderately supported (bootstrap proportion [BP] = 75%). Section Bromopsis is not 20 monophyletic. A well-supported clade (BP = 99%) consisting of two North American Bromopsis species, B. attenuatus and B. dolichocarpus, is the sister group of the rest of the genus, the latter clade with BP = 100%. The next major split in Bromus is between a well-supported clade (BP = 100%) of four South American species of sect. Bromopsis (B. lanatus, B. modestus, B. pellitus, and B. pflanzii) and all remaining species of Bromus. The latter clade is weakly supported (BP = 59%). A large and well-supported clade (BP = 96%) includes five species of sect. Bromopsis of Eurasian origin [B. erectus, B. inermis, B. korotkoyi, B. pumpellianus (which is also native in the New World), and B. riparius], one species of sect. Bromopsis from South American (B. brachyantherd), and the monophyletic sects. Ceratochloa, Genea, and Neobromus (BP = 92%, 100%, and 94% respectively). The Eurasian representatives of sect. Bromopsis are not united in a single clade. A second large, weakly-supported clade (BP < 50%) contains the remaining North American species of sect. Bromopsis, B. ramosus (a species classified in sect. Bromopsis from the Old World), and sects. Bromus and Triniusia (Fig. 2.1). Several well-supported relationships are evident among some species of sect. Bromopsis from North America. Section Triniusia is monophyletic (BP = 94%), and is part of a well-supported clade that otherwise only includes representatives of sect. Bromus (BP = 100%). 2.3.2. Analyses of Plastid Sequences The sequence data obtained for the 3' end of ndhF corresponds to positions 1441 to 2076 of ndhF in Oryza sativa (NC_00132). The sequenced portion of the 3' end of ndhF was 662 bp in length in all taxa, except for B. grandis, which had a six bp insertion. I was not 21 able to obtain data for ndhF from six taxa (Table 2.3). The unambiguously aligned ndhF matrix was 668 bp long; 576 nucleotides were constant, 92 were variable, and 44 (6.5%) were parsimony informative. Among the ingroup taxa, 616 characters were constant, 52 were variable, and 31 (4.6%) were parsimony informative. The trnL intron ranged in length from 568-586 bp. Several indels were present in the final data matrix; three of these were phylogenetically informative and were coded as binary characters in the analysis. In the trnL intron, two regions of 18 bp (positions 1397-1414) and 11 bp (positions 1755-1765) were homopolymer repeats of variable length that were difficult to align; these regions were excluded from all analyses. The aligned trnL intron matrix (including binary characters but without excluded sites) consisted of 646 aligned nucleotides; 578 were constant, 68 were variable, and 28 (4.9%) were parsimony informative. Among the ingroup taxa, 610 characters were constant, 36 were variable, and 23 (3.5%) were parsimony informative. No sequence data were obtained from either plastid locus for B. modestus and B. nottowayanus, and four species are represented solely by data from the trnL intron (Table 2.3). The heuristic search of the combined plastid data recovered 50,000 most parsimonious trees (tree length = 218 steps, CI =0.817, RI = 0.882). In the analyses of combined plastid data there is moderate phylogenetic structure that is supported by bootstrap analysis (Fig. 2.2). The monophyly of the genus Bromus is well supported (BP = 99%). Taxa classified in sects. Bromus, Genea, and Triniusia form a well-supported monophyletic group (BP = 91%), but none of the three sections is monophyletic. Bromus pectinatus (sect. Bromus) and B. diandrus (sect. Genea) comprise a well-supported clade (BP = 100%) that is weakly supported (BP = 56%) as the sister group of B. madritensis subsp. rubens (sect. Genea). Other species of sect. Bromus and species of sect. Triniusia are 22 mixed in a clade (BP = 86%). Species of sects. Ceratochloa, Neobromus, and B. brachyanthera (sect. Bromopsis) comprise a weakly-supported clade (BP = 62%). Section Neobromus is not monophyletic, and the monophyly of sect. Ceratochloa is weakly supported (BP = 63%). A large clade of 23 New and Old World species of sect. Bromopsis is weakly supported (BP = 68%). 2.3.3. Incongruence Among Data Partitions The ILD test indicated significant incongruence between the nuclear ribosomal and plastid data partitions (p < 0.01). Overall, the trees derived from the nuclear ribosomal data were more resolved than trees derived from the plastid data. There were some well-supported clades whose positions differed substantially among trees, although not always with strong support. Topologically, the greatest differences between the plastid and nuclear ribosomal trees were the positions and monophyly of sects. Bromus, Trinusia, and Genea. In the nuclear trees, species from sects. Bromus and Triniusia formed a clade, and sect. Genea was well supported as monophyletic; a close relationship between these two clades was not inferred (Fig. 2.1). In the plastid trees, species from these three sections were intermixed in a well-supported clade (Fig. 2.2); for example, Bromus pectinatus (sect. Bromus) and B. diandrus (sect. Genea) comprise a well-supported clade. Other incongruencies involve relationships among species of sect. Bromopsis (Fig. 2.2). The plastid trees include species of sect. Bromopsis from the Old World in a weakly-supported clade with some North American species of sect. Bromopsis, while the nuclear ribosomal trees indicate a more distant relationship between these Old World (with the exception of B. ramosus) and North 23 American species. Because of these possible instances of intergenomic conflict, I did not conduct analyses of the combined nuclear ribosomal and plastid data. 2.4 DISCUSSION 2.4.1. Phylogenetic Utility of the Regions Examined Of the three regions examined, the nuclear ribosomal region was the most variable and accounted for 65.8% of the total number of parsimony-informative characters among all three data sets (among ingroup taxa). Resolution (number of bifurcated nodes in the strict consensus) was greater in the nuclear ribosomal phylogeny compared with the plastid phylogeny, probably because of the greater amount of variation in the former data set. The least parsimony-informative variation (among ingroup taxa) was observed in the trriL intron, which accounted for 14.5% of the total number of informative characters in all three data sets. Although this intron is commonly used for lower-level phylogenetic studies, several investigators have reported a paucity of phylogenetically informative characters in it to sufficiently resolve relationships among closely related grass genera and species (e.g., Hodkinson et al. 2002), arid a wide variety of other plant taxa (e.g., Bruneau et al. 2001; Klak et al. 2003; Shaw et al. 2005). The 3' end of ndhF provided 19.6% of the total parsimony-informative variation among all three data sets (among the ingroup taxa), 1.35 times as many parsimony informative characters as the trriL intron for approximately the same length. The complete ndhF region has been used in several phylogenetic studies of grasses at the familial, subfamilial, tribal and generic levels (e.g., Clark et al. 1995; Catalan et al. 1997; Spangler et al. 1999; Giussani et al. 2001; Aliscioni et al. 2003). The more rapidly evolving 3' end of ndhF has been used at the genus level in grasses (e.g., Catalan and Olmstead 2000) and other 24 plants (e.g., Graham et al. 1998; Davis et al. 2002; Winkworth et al. 2002; Graham and Barrett 2004). The greater level of sequence variation detected in the 3' end of ndhF compared with the variation detected in the more commonly used trnL intron indicates that the former warrants consideration for use in the resolution of relationships at similar taxonomic levels in other groups. 2.4.2. Incongruence Between Nuclear Ribosomal and Plastid Data Partitions Significant incongruence was detected between the nuclear ribosomal and plastid data partitions using the ILD test. The ILD test is commonly employed by systematists to examine congruence among data partitions, but there is growing evidence (e.g., Yoder et al. 2001; Barker and Lutzoni 2002) that the test can be misleading and should not be used to determine data combinability. Thus, I also visually compared trees derived independently from the plastid and nuclear data partitions for regions of incongruence, and found that each contained some moderately- to well-supported clades whose composition and position differed among trees. Because of this possible intergenomic conflict, I did not conduct analyses of the combined plastid and nuclear ribosomal data. Incongruence among trees is not uncommon in phylogenetic studies that employ multiple gene regions, particularly when the data are from different genomes (e.g., Hardig et al. 2000; Les et al. 2002). Although often viewed as a hindrance to reliable phylogenetic estimation, incongruence can potentially provide insight into past evolutionary events, such as hybridization, introgression, and lineage sorting (Wendel and Doyle 1998). The current data suggest that some of these phenomena may have been involved in the evolutionary history of Bromus. However, it is difficult to infer the exact evolutionary processes that have 25 led to the differing gene trees, as reticulate patterns of evolution are difficult to study in a cladistic framework, and ideally requires gene trees inferred from more than two linkage groups. Nonetheless, previous studies have indicated that hybridization, alloploidy, and introgression may have been prominent in the evolution of many Bromus species and sections [reviewed by Stebbins (1981) and Armstrong (1991)]. The implications of the different gene histories detected here in understanding the evolutionary history of infrageneric groups in Bromus are discussed below. Clarification of the contribution of these processes to the evolutionary history of Bromus will require better-supported phylogenetic trees from multiple genetic linkage groups. 2.4.3. Phylogeny and Classification In all analyses there is moderate to strong support for the monophyly of the genus Bromus s. 1., based on current outgroup and ingroup taxon sampling. These findings agree with Ainouche and Bayer's (1997) study of sect. Bromus, and broader studies of grass phylogeny that have included several species of Bromus s. 1. (e.g., Catalan et al. 1997; Hsiao et al. 1999), which all identified Bromus s. 1. as a monophyletic taxon. Sections Bromus, Triniusia, and Genea — The molecular evidence indicates that species of sect. Triniusia are nested within a clade that includes species of sect. Bromus (Figs. 2.1, 2.2). Section Triniusia was originally circumscribed as a monotypic section that included one species, B. danthoniae, characterized by three awns on each of the uppermost lemmas of the spikelets (Scholz 1998), but most authors have included this species in sect. Bromus (e.g., Smith 1970, 1972; Ainouche and Bayer 1997). A close relationship between B. danthoniae and B. pseudodanthoniae was not hypothesized until Scholz (1998) observed 26 that B. pseudodanthoniae sometimes has three awns on the uppermost lemmas of its spikelets, and that in portions of their ranges in the Middle East these two taxa intergrade. As a result, Scholz (1998) treated B. pseudodanthoniae as a subspecies of a polymorphic B. danthoniae, and re-circumscribed sect. Triniusia to include two morphologically similar species and several intraspecific taxa (B. danthoniae Trin. subsp. danthoniae, B. danthoniae subsp. pseudodanthoniae [Drobov] H. Scholz, B. danthoniae subsp. rogersii C. E. Hubb. ex H. Scholz, and B. turcomanicus H. Scholz). Scholz's (1998) recognition of sect. Triniusia is supported by isozyme data, which found B. danthoniae to be distinct from diploid members of sect. Bromus (Oja and Jaaska 1998), although serological evidence found B. danthoniae to be closely allied to species of sect. Bromus, including B. pumilio (classified currently in sect. Boissiera but often included in sect. Bromus), a species that also has multiple awns on the uppermost lemmas of its spikelets (Smith 1972). I did not sample B. turcomanicus, thus I was unable to fully test the monophyly of sect. Triniusia sensu Scholz (1998). However, my data confirm the close relationship hypothesized between B. danthoniae and B. pseudodanthoniae, and indicate that these species are nested phylogenetically within sect. Bromus (Fig. 2.1) or perhaps a somewhat broader clade (Fig. 2.2). These data are in accordance with the findings of Ainouche and Bayer (1997), who included B. danthoniae in their study of sect. Bromus. Recognition of sect. Triniusia renders sect. Bromus paraphyletic; it should therefore continue to be treated as a synonym of sect. Bromus, as past authors have done (e.g., Smith 1970, 1972; Ainouche and Bayer 1997). The distinct morphological characters that separate B. danthoniae and its close relatives from other species in sect. Bromus arose from within sect. Bromus. 27 The two sources of molecular evidence are in conflict in regards to the monophyly of sects. Bromus and Genea, due to the position of B. pectinatus. Sections Bromus (including sect. Triniusia; see above) and Genea (based on sampling only two of the approximately five species in the section) are each robustly supported as monophyletic in the nuclear ribosomal trees (Fig. 2.1). However, the plastid trees indicate that B. pectinatus (sect. Bromus) is the sister group of B. diandrus (sect. Genea; Fig. 2.2). Species of the B. pectinatus complex (only one species sampled here), a group of five tetraploid species that range from southern Africa to Tibet, classified in sect. Bromus, are morphologically similar to species of sect. Genea, with lemmas that taper toward the apex and paleas whose morphology is intermediate between the two sections (Smith 1972; Scholz 1981; Stebbins 1981; Sales 1993). A close relationship between B. pectinatus and sect. Genea is also supported by data from isozymes (Oja 2005) and embryo structure (Kosina 1996). Based on its morphological intermediacy, Stebbins (1956, 1981) suggested that the B. pectinatus complex (represented by B. arenarius in his studies) may be an intersectional amphidiploid that originated via a hybridization event between species of sects. Genea and Bromus. The conflicting positions of B. pectinatus in my plastid and nuclear ribosomal trees lend support to this hypothesis, indicating that the genome donors in the origin(s) of the complex were likely from sects. Genea and Bromus. Sampling of additional species of the B. pectinatus complex, and additional linkage groups, would be valuable, and may provide further insight into their origin(s). If B. pectinatus is a species of hybrid origin that arose after sects. Bromus and Genea initially diversified and it is excluded from consideration, then sects. Bromus and Genea are monophyletic. The morphological characteristics outlined in Table 2.2, widely employed in taxonomic keys to separate these two lineages (e.g., Pavlick 1995), constitute possible synapomorphies for these 28 clades; however, validation of these hypotheses will require rigorous reconstructions of character evolution on robustly supported and fully resolved gene trees. The plastid and nuclear ribosomal data sets infer different relationships between sects. Bromus and Genea. The nuclear ribosomal data do not infer a close relationship between sects. Bromus and Genea (Fig. 2.1), while the plastid data (Fig. 2.2) strongly support a clade containing all taxa from both sections. The placement of species from sects. Genea and Bromus together in a clade in the plastid trees corrobates the study of Pillay and Hilu (1995), although they did not detect sufficient chloroplast DNA variation to distinguish sects. Bromus and Genea as distinct monophyletic groups. Pillay and Hilu (1995) suggested that the similarity in chloroplast genomes among sects. Genea and Bromus may be the result of chloroplast transfer by hybridization and phylogenetic sorting. A close relationship between sects. Genea and Bromus is further supported by data from floral microstructural variation (Kosina 1999), and by their life histories. Both include only annual species (most other sections of Bromus comprise mostly perennial species), and both include many weedy species (Stebbins 1981). Stebbins (1981) also hypothesized a close relationship between sects. Genea and Bromus, and suggested that their origins probably involved different species of sect. Bromopsis as genome donors. My nuclear ribosomal data are potentially consistent with this hypothesis, as sect. Bromus is nested within a clade that includes species of sect. Bromopsis from North America and B. ramosus from the Old World, while sect. Genea is closely related to species of sect. Bromopsis from the Old World (excluding B. ramosus). These species groups of sect. Bromopsis, respectively, are potential candidates for genome donors in the origins of sects. Bromus and Genea. 29 Within sect. Genea, my data indicate a fairly substantial amount of genetic variability among individuals of B. madritensis subsp. rubens, in line with the results of a previous isozyme study (Kahier et al. 1981). The high genetic variation observed here seems to correspond with morphological variation that was high enough to result in the gross misidentification of one seed bank accession (see Table 2.3). The genetic variation observed in B. madritensis subsp. rubens raises the possibility that similar high levels of variation may be present in at least some other Bromus species. Section Bromopsis — Section Bromopsis, the largest section currently recognized in Bromus, comprises several independent lineages and is not monophyletic in any of my analyses. These results are congruent with Pillay and Hilu (1995), who found members of sect. Bromopsis to occur in three distinct lineages (based on the plastid genome but with less taxon sampling). Based on my nuclear ribosomal data, B. attenuatus and B. dolichocarpus, two North American species of sect. Bromopsis native to northeastern Mexico and southern Mexico and Guatemala, respectively (Wagnon 1952; Soderstrom and Beaman 1968), are the sister group of the rest of Bromus. Four South American species of sect. Bromopsis (B. lanatus, B. modestus, B. pellitus, B. pflanzii) comprise a well-supported clade that is resolved as part of the second-deepest split in the genus. The plastid data alone do not support these phylogenetic placements, possibly because of insufficient variation; however, the plastid trees do indicate that these species are not part of the clade that includes other New and Old World species of sect. Bromopsis (Fig. 2.2), and they do not strongly rule out the relationships seen for the nuclear ribosomal data. Although they define deep splits on the molecular trees, the morphological characteristics of these species are not sufficiently distinct compared with other New World species of sect. Bromopsis for previous workers to have 30 considered them as major evolutionary lineages. However, Wagnon (1952) suspected that B. attenuatus and B. dolichocarpus are closely related to each other, and that they are distantly related to other North American species of sect. Bromopsis. The molecular data agree with this hypothesis. Further study is necessary to identify possible morphological synapomorphies for a B. attenuatus/ B. dolichocarpus clade as well as a putative clade of South American species that may be part of the second deepest split in Bromus. The deep phylogenetic positions of these two clades in the nuclear ribosomal trees suggest that the crown clade of Bromus originated in the New World. In contrast, Stebbins (1981) suggested that Bromus originated in Eurasia, with sects. Neobromus, Ceratochloa, and Bromopsis being the first to differentiate and subsequently spread to North and South America, followed by the evolution of sects. Bromus, Genea, and Boissiera. The molecular evidence suggests that the remaining South American species of sect. Bromopsis sampled, B. brachyanthera (a hexaploid; Schifino and Winge 1983), is closely related to Old World species of sect. Bromopsis and sects. Ceratochloa and Neobromus. In the plastid trees, B. brachyanthera is the sister group of a clade corresponding to sect. Ceratochloa (Fig. 2.2), whereas in the nuclear ribosomal trees B. brachyanthera is the sister group of a clade comprised of sects. Neobromus and Ceratochloa (Fig. 2.1). Despite the close molecular relationship, Bromus brachyanthera is morphologically distinct from species in sect. Ceratochloa and Neobromus, having dorsiventrally flattened spikelets typical of other species in sect. Bromopsis, and straight awns. Stebbins (1981) suggested that some members of sect. Bromopsis may have donated genomes during the origin of sect. Ceratochloa. In line with this hypothesis, the phylogenetic affinities of B. brachyanthera and the Old World species of sect. Bromopsis with sect. Ceratochloa suggest that they, their 31 close relatives, or their immediate ancestors, are among the most likely candidates as possible genome donors. Inclusion of the six unsampled native species of sect. Bromopsis from South America (Planchuelo and Peterson 2000) in future studies should provide further insight into the evolution and relationships of this group of poorly understood species. Armstrong (1983) hypothesized that the North American and Old World members of sect. Bromopsis may represent distinct evolutionary lineages. Species from North America are generally diploids (a few are tetraploids), and all have large chromosomes with pinhead satellites, whereas Old World species are generally polyploids with smaller chromosomes lacking pinhead satellites (Armstrong 1983). Exceptions are B. pumpellianus, which is native in North America and Eurasia and morphologically and cytologically similar to Old World taxa, and the B. ramosus complex of the Old World (represented here by B. ramosus), which is morphologically and cytologically similar to North American species of sect. Bromopsis (Armstrong 1983). Differences in floral microstructural variation further support the distinctiveness of these morphologically and cytologically differentiated groups (Kosina 1999). My nuclear ribosomal data may partly support these hypotheses, as species of sect. Bromopsis from North America (excluding B. attenuatus and B. dolichocarpus) and B. ramosus from the Old World form a clade that does not include the other Old World species (Fig. 2.1). Old World species of sect. Bromopsis (including B. pumpellianus) comprise several independent but closely related lineages that are part of a weakly-supported clade that also includes B. brachyanthera (sect. Bromopsis) and sects. Ceratochloa and Neobromus (Fig. 2.1). These relationships are consistent with the findings of Kosina (1996), who observed similarity in the embryo structure of species of sects. Ceratochloa and Old World species of sect. Bromopsis. In contrast, the plastid data include Old World species of sect. 32 Bromopsis in a large, weakly-supported clade with many North American species of sect. Bromopsis (Fig. 2.2). The gene trees thus indicate that most Old World and North American lineages of sect. Bromopsis share a similar plastid genome, but have conflicting nuclear ribosomal histories. The Old World species (most of which are polyploids) may have originated via a hybridization event, with a diploid member of sect. Bromopsis contributing the plastid genome. Additional representatives of sect. Bromopsis as traditionally circumscribed from the Old World, and improved genomic samplings, will be required to provide further insight into the evolution and relationships of these species. If it becomes desirable to formally recognize these Old World lineages, the sectional name Pnigma Dumort. is available for the clade that contains B. inermis Leyss. (Armstrong 1983). Within the clade of North American species of sect. Bromopsis and B. ramosus, several weakly- to well-supported clades of two to five species are evident (Figs. 1, 2). Wagnon (1952) suggested several groupings of North American species of sect. Bromopsis, based on geographical distribution: (1) an Arctic group, (2) a Rocky Mountain Mexican Highland group, (3) a Pacific Slope group, and (4) an East-Midwest group. My trees are largely congruent with the East-Midwest group that Wagnon (1952) defined to include B. ciliatus, B. kalmii, B. nottowayanus, B. pubescens, B. purgans (here treated as B. latiglumis), and B. texensis. All of these species, except B. texensis, comprise a moderately-supported clade in the nuclear ribosomal trees; there is insufficient variation in the plastid data alone to support or reject such close species relationships. Wagnon (1952) noted that the placement of B. texensis might seem out of place in this group, since its geographic range is intermediate between other members of the East-Midwest group and members of the Rocky Mountain Mexican Highland group, but he included it because the morphology of its ligule is 33 similar to other members of the group. My data neither support nor reject Wagnon's (1952) other groups, as the phylogenetic relationships of many of these species are unresolved. The short branch lengths and lack of resolution among many of the North American species of sect. Bromopsis indicate that the species in this group likely diversified during a recent rapid radiation. There has been much confusion about the species status of B. richardsonii and B. ciliatus (North American species of sect. Bromopsis). Bromus richardsonii is often treated as a synonym of B. ciliatus (e.g., Hitchcock and Chase 1951; Soderstrom and Beaman 1958; Allred 1993), although recent taxonomic study has indicated that these taxa are sufficiently distinct morphologically, cytologically, and genetically to warrant specific recognition (Peterson et al. 2002). My nuclear ribosomal data confirm that B. richardsonii is a distinct species, closely related to B. mucroglumis (although the species status of B. mucroglumis is also controversial; Wagnon 1952, Peterson et al. 2002); these two taxa do not share an immediate common ancestor with B. ciliatus (Figs. 2.1, 2.2). It is clear from both plastid and nuclear ribosomal data that sect. Bromopsis is an artificial assemblage of species. The morphological characteristics traditionally used to circumscribe the section (Table 2.2) may therefore be a mixture of characters that are homoplasious of that represent symplesiomorphies of larger clades. The recognition of Bromopsis as a distinct section, subgenus, or genus (Table 2.1) is clearly not appropriate. Sections Ceratochloa and Neobromus — Section Ceratochloa is weakly supported as monophyletic in the plastid trees (Fig. 2.2), and robustly supported as monophyletic in the nuclear ribosomal trees (Fig. 2.1). None of the sequence data is sufficiently variable to resolve relationships among species in the section, and several species are genetically 34 identical at the loci examined. Similarly, Pillay and Hilu (1990, 1995) found no chloroplast restriction-site variation among species of sect. Ceratochloa. The plastid and nuclear ribosomal data are in conflict regarding the monophly of sect. Neobromus. In the plastid trees (Fig. 2.2), the two species of sect. Neobromus comprise a grade, in which B. berteroanus is the sister group of a clade comprising B. gunckelli, sect. Ceratochloa and B. brachyanthera (sect. Bromopsis). However, sect. Neobromus is a well-supported monophyletic group in the nuclear ribosomal trees (Fig. 2.1), a relationship not strongly rejected by the plastid data, and clearly both taxa are closely related. Both species are morphologically similar, sharing strongly twisted and divaricate awns (Table 2.2), hence their classification as a section. Bromus berteroanus (syn = B. trinii Desvaux) is morphologically similar to other grass genera because of its large glumes and a lemma that is deeply bilobed apically (Stebbins 1981), and in the past the species has been confused as a species of the genus Trisetum Pers. (Louis-Marie 1928), although this classification has not been followed by recent authors. My data confirm that B. berteroanus is a species of Bromus. The weakly supported relationship between sects. Ceratochloa and Neobromus (Figs. 2.1, 2.2) agrees with previous hypotheses that these taxa share common ancestry. Stebbins (1981) reported a weak affinity between one genome of sects. Neobromus and Ceratochloa, and Pillay and Hilu (1995) found that these taxa shared eight synapomorphies based on chloroplast DNA restriction site variation. Unfortunately, neither of these studies included representatives of South American species of sect. Bromopsis, one of which appears here to be closely related to sects. Ceratochloa and Neobromus (Figs. 2.1, 2.2). Stebbins (1981) hypothesized that sects. Neobromus and Ceratochloa evolved early within Bromus because 35 of their small chromosome size and spikelets that resemble those in genera that he thought were derived from the ancestral complex from which Bromus originated (including Littledalea, Megalachne, Metcalfia, and Pseudodanthonia). The plastid data neither reject or support this hypothesis (Fig. 2.2), but the nuclear ribosomal data indicate that B. attenuatus and B. dolichocarpus (sect. Bromopsis), species that are morphologically distinct from taxa in sects. Ceratochloa and Neobromus, are part of a deep lineage that diverged early in the history of the genus (see above; Fig. 2.1). Sections Ceratochloa and Neobromus are nested deep within the nuclear ribosomal trees. Current knowledge, favouring distant phylogenetic positions of the morphologically similar genera thought previously to be closely related to sects. Ceratochloa and Neobromus (e.g., Soreng et al. 2003), further discounts Stebbins' (1981) hypotheses. Sections Boissiera and Nevskiella — Material of B. pumilio (sect. Boissiera) was not available, and available material of B. gracillimus (sect. Nevskiella) was recalcitrant to molecular study, thus the phylogenetic positions of these taxa remain uncertain. Bromus pumilio was originally classified in its own genus, Boissiera, but was transferred to Bromus based on serological and morphological similarities to other Bromus species (Smith 1969). It has since been treated either in sect. Bromus (e.g., Smith 1970) or within its own section, Boissiera (Smith 1985a; Table 2.1), because of its unique morphology, having five to nine awns on each lemma (Table 2.2). Based on allozyme evidence, Oja and Jaaska (1998) found B. pumilio to be distinct from members of sect. Bromus, supporting its placement in its own section. The phylogenetic position of B. gnacillimus (sect. Nevskiella), characterized by awns that are four to six times the length of the lemma (Table 2.2), remains unknown. It would be valuable to include both species in future molecular studies. 36 2.4.4. Conclusions and Future Directions My study provides genus-wide phylogenetic hypotheses of relationships in Bromus s. 1., based on DNA sequence data from the plastid genome and the nuclear ribosomal internal transcribed spacer regions, and provides a foundation for further phylogenetic study of the genus. Based on the nuclear ribosomal data, sects. Bromus (including sect. Triniusia), Genea, Neobromus, and Ceratochloa are monophyletic, and sect. Bromopsis comprises several distinct lineages. Plastid trees indicate that sects. Genea and Bromus are closely related, and the incongruence detected between the plastid and nuclear ribosomal data support a hybrid origin for the B. pectinatus complex (here represented by a single exemplar) between sects. Bromus and Genea. Plastid trees indicate a close relationship between Old World and some North American species of sect. Bromopsis, and the plastid and nuclear ribosomal data indicate that one South American species of sect. Bromopsis is not closely related to North American and Eurasian species traditionally classified in the same section. Most species of Bromus sampled had levels of sequence variation too low to allow complete resolution of relationships among close relatives at the species level (e.g., among North American members of sect. Bromopsis and within sect. Ceratochloa). Sequence data from additional nuclear loci, such as the granule-bound starch synthase gene (waxy; e.g., Mason-Gamer 2001), AFLPs (e.g., Beardsley et al. 2003; Despres et al. 2003), or microsatellites (e.g., Alvarez et al. 2001) may provide further insight into species-level relationships in Bromus. Adding data from the plastid genome (e.g., Shaw et al. 2005) and the nuclear ribosomal region [the external transcribed spacer (ETS) of nuclear rDNA (e.g., Baldwin and Markos 1998; Markos and Baldwin 2002; Starr et al. 2003)] would also be valuable to improve resolution and support of trees inferred from these two linkage groups. 37 Recognition of the brome grasses as one distinct genus, Bromus, is in agreement with the molecular data, but current classification schemes do not satisfactorily reflect phylogenetic relationships within the genus, particularly with respect to the circumscription of sect. Bromopsis. However, before a revised infrageneric classification of Bromus is proposed, I advocate that substantially better sampling should be conducted of (1) DNA sequence regions, to obtain better support for phylogenetic relationships among taxa, and to further clarify incongruence among nuclear and plastid data partitions; and (2) taxa, to more adequately sample the molecular, morphological, and geographical variability in the genus. Although this is the largest study of Bromus phylogeny conducted thus far, all conclusions are based on a sample of less than one-third of the recognized species, mostly with one individual per taxon, and it is plausible that addition of other species will further contribute to and change understanding of evolution and phylogeny in this genus. 38 Table 2.1. Summary of infrageneric classificatons and generic segregations ofBromus following Smith (1970), Tsvelev (1976), and Stebbins (1981). Equivalent circumscriptions are aligned horizontally. Indented names were treated by the author as synonyms of the taxon above. Sections Subgenera Genera (Smith 1970) (Stebbins 1981) (Tsvelev 1976) Bromopsis Dumort. (as sect. Pnigma Dumort.) Festucaria Gren. & Godr. Bromopsis (Dumort) Fourr. Bromus Bromus Bromus L. Triniusia (Steud.) Nevski" Triniusia (Steud.) Penzes. Triniusia Steud. Boissiera (Hochst. ex Steud.) P. M . Smithb Boissiera nom. inval.. Boissiera Hochst. ex Steud. Ceratochloa (P. Beauv.) Griseb. Ceratochloa (P. Beauv.) Hack. Ceratochloa P. Beauv. Genea Dumort. Stenobromus Hack. Anisantha C. Koch Nevskiella (Krecz. & Vved.) Tournay Nevskiella (Krecz. & Vved.) Krecz. & Vved. Nevskiella Krecz. & Vved. Neobromus (Shear) Hitchc. Neobromus Shear Trisetobromus Nevski a Given sectional status by Scholz (1998). b Given sectional status by Smith (1985a). Table 2.2. Number of species, morphological characteristics, and native geographic distribution of sections in Bromus (after Smith 1970). The classification follows Smith (1970, 1985a) and Scholz (1998). Section No. Species 1st Glume Nerves 2nd Glume Nerves Spikelet Shape Lemmas Native Geographic Distribution Boissiera 1 3 5-9 Linear-lanceolate to oblong; terete Oblong; awns five-nine Asia, E Mediterranean Bromopsis -60 1(3) 3(5) Narrow, lanceolate; terete Rounded or slightly keeled; awn single, usually shorter than length of lemma, rarely absent Eurasia, Africa, N and S America Bromus 30-40 3-5 5-9 Ovate to ovate-lanceolate; terete to slightly compressed Rounded; awn single, equalling or slightly exceeding length of lemma, rarely absent Europe, Asia Ceratochloa 10-16 3-5 5-7 Ovate or ovate-lanceolate; strongly compressed Strongly keeled; awn single, short, often absent N and S America Genea 6 1 3 Cuneate, wider at top Narrow and elongate; awns single, less than three times length of lemma Mediterranean, SW Asia, N Europe 4^  O Table 2.2 continued Section No. Species 1 st Glume Nerves 2nd Glume Nerves Spikelet Shape Lemmas Native Geographic Distribution Neobromus 2 1 3-5 Narrowly elliptic Deep apical sinus and two long, Pacific coast of N and S America • narrow teeth; awn single, longer than .- length of lemma, geniculate Nevskiella 1 1 3 Ovate-lanceolate to cuneiform, Rounded; awn single, 4-6 times Central Asia, Iran, Afghanistan wider above; terete to slightly length of lemma compressed Triniusia 2 3-5 5-9 Ovate to lanceolate; Rounded; upper lemma with three E Mediterranean, SW Asia compressed awns; irregular apical notches Table 2.3. Sections [following Smith (1970) and Scholz (1998)] and species sampled, sources of material, vouchers, and GenBank accession numbers for DNA sequences. WRPIS = Western Regional Plant Introduction Station (United States Department of Agriculture, Pullman, Washington, USA); PGRC = Plant Gene Resources of Canada. All numbers preceded by PI or CN are seed accession numbers. Missing GenBank accession numbers indicate that no sequence was obtained. Sequence data for two outgroup taxa (Hordeum vulgare L. L. and Triticum aeslivum L.) were obtained from GenBank (accessions Z l 159, X757505, U2003, AF521903, X75709, and NC 0027). Accessions ofB. madritensis subsp. rubens are numbered 1—3. GenBank accession number Taxon Geographic origin/ source ' Voucher ITS IrnL intron ndhF Section Bromopsis B. anomalus Rupr. ex Foum. (acc. 1) B. anomalus Rupr. ex Fourn. (acc. 2) B. attenuatus Swallen B. brachyanthera Doll. B. ciliatus L. B. dolichocarpus Wagnon B. erectus Huds. B. exallatus Bernh. B. frondosus (Shear) Woot. et Standi. USA; PI 232199 (WRPIS) Mexico: Tamaulipas Mexico: Tamaulipas/Nuevo Leon border Bolivia: La Paz Canada: Quebec Mexico: Michoacan . Turkey; PI 337652 (WRPIS), received as B. benekenii (Lange) Trimen Mexico: Jalisco Mexico: Durango Keane 49 (ALTA) Peterson 15918 & Valdes-Reyna (US) Peterson 15926 & Valdes-Reyna (US) de Ros 9497 (US) Cayouette C8272 & Lavoie (DAO) Peterson 16128 (US) Keane 8 (ALTA) Peterson 16087 & Rosales (US) Peterson 15418 et al. (US) AY367905 AY367955 AY368004 AY367906 AY367956 AY368005 AY367910 AY367960 AY368009 AY367908 AY367958 AY368007 AY367909 AY367959 AY368008 AY367911 AY367961 AY367907 AY367957 AY368006 AY367912 AY367962 AY368010 AY367913 AY367963 AY368011 4^  Table 2.3 continued Taxon Geographic origin/ source Voucher ITS trnL intron ndhF B. ramosus Huds. UK Keane 101 (ALT A) AY367929 AY367978 AY368023 B. richardsonii Link USA: Arizona Peterson 15282 & Cayouette (US) AY367930 AY367979 AY368024 B. riparius Rehmann Czech Republic/ Slovakia; PI 598590 (WRPIS) Keane 26 (ALTA) AY367931 AY367980 AY368025 B. suksdorfii Vasey USA: Washington Soreng 6352 & Soreng (US) AY367934 AY367983 AY368028 B. texensis (Shear) A. S. Hitchc. USA: Texas Cayouette 668135 (DAO) AY367935 AY367984 AY368029 Section Bromus B. japonicus Thunb. Russia; PI 283198 CPI 24193 (WRPIS), Keane 24 (ALTA) AY367940 AY367989 AY368034 received as B. popovii Drob. B. pectinatus Thunb. Belgium; PI 442453 (WRPIS) Keane 23 (ALTA) AY367939 AY367988 AY368033 B. scoparius L. Turkey; PI 204425 (WRPIS) Keane 28 (ALTA) AY367932 AY367981 AY368026 Section Ceratochloa B. carinatus Hook, et Am. Mexico: Durango Peterson 15421 et al. (US) AY367948 AY367997 AY368042 B. catharticus Vahl Argentina; PI 578719 RGI441 (WRPIS), Keane 5 (ALTA) AY367954 AY368003 AY368048 received as B. araucanus Phil. B. cebadilla Steud. Chile; PI 202696 (WRPIS), Keane 13 (ALTA) AY367944 AY367993 AY368038 received as B. coloratus Steud. B. coloratus Steud. Chile: Region I Peterson 15746 & Soreng (US) AY367943 AY367992 AY368037 4^  Table 2.3 continued Taxon Geographic origin/ source Voucher ITS trnL intron ndhY B. grandis (Shear) A. S. Hitchc. USA: California Cayouette 7947a (DAO) AY367914 AY367964 AY368012 B. inermis Leyss. USA: Arizona* Peterson 15295 & Cayouette (US) AY367915 AY367965 AY368013 B. kalmii A. Gray Canada: Ontario; CN 51222 C7099 (PGRC) Keane 55 (ALTA) AY367916 AY367966 AY368014 B. korotkoyi Drob. China: Inner Mongolia Soreng 5160 et al. (US) AY367998 AY367988 AY368043 B. laevipes Shear USA: California Peterson 14840 et al. (US) AY367917 AY367967 AY368015 B. lanatipes (Shear) Rydb. USA: Arizona Peterson 15270 & Cayouette (US) AY367918 AY367968 B. lanatus Kunth Chile: Region I Peterson 15747 & Soreng (US) AY367919 AY367969 AY368016 B. latiglumis (Shear) A. S. Hitchc. Canada: Ontario Cayouette 4336-1 (DAO) AY367920 AY367970 AY368017 B. modestus Renvoize Bolivia: La Paz Peterson 12639 et al. (US) AY367921 B. mucroglumis Wagnon USA: Arizona Peterson 152 73 & Cayouette (US) AY367922 AY367972 AY368019 B. nottowayanus Fern. USA: Illinois Chase 13512 (US) AY367923 B. pellilus Hack. Argentina: Santa Cruz Peterson 17267 et al. (US) AY367951 AY368000 AY368045 B. pflanzii Pilg. Bolivia: La Paz Luteyn & Dorr 13828 (US) AY367924 AY367973 B. porteri (Coult.) Nash USA: Arizona Peterson 15245 & Cayouette (US) AY367925 AY367974 AY368020 B. pseudolaevipes Wagnon USA: California Cayouette C7987 (DAO) AY367926 AY367975 AY368021 B. pubescens Muhl. ex Willd. USA: Virginia Peterson 15776 & Saarela (US) AY367927 AY367976 B. pumpellianus Scribn. Mongolia; PI 610833 (WRPIS) Keane 17 (ALTA) AY367928 AY367977 AY368022 4^  Table 2.3 continued Taxon Geographic origin/ source Voucher ITS trnh intron ndhF B. marginatus Nees ex Steud. USA: Oregon Soreng 6360 & Soreng (US) AY367921 AY367971 AY368018 B. striatus Hitchc. France; PI 477988 974 (WRPIS) Keane 6 (ALTA) AY367945 AY367994 AY368039 B. subvelutinus Shear. Uzbekistan; PI 392355 (WRPIS) Keane 35 (ALTA) AY367953 AY368002 AY368047 Section Genea B. diandrus Roth. Germany; CN 31600 PGR 2848 (PGRC) Keane 25 (ALTA) AY367936 AY367985 AY368030 B. madrilensis L. Iraq; PI 253735 (WRPIS), Keane 16 (ALTA) AY367937 AY367986 AY368031 subsp. rubens (L.) Husnot (acc. 3) received as B. fasciculatus C. Presl. B. madritensis L. Iran; PI 239722 (WRPIS), Keane 20 (ALTA) AY367938 AY367986 AY368032 subsp. rubens (L.) Husnot (acc. 2) received as B. madritensis L. B. madrilensis L. Australia: Western Australia" Peterson 14534 et al. (US) AY367950 AY367987 AY368044 subsp. rubens (L.) Husnot (acc. 1) Section Neobromus B. gunckelli Matthei Chile: Region I Peterson 15697 & Soreng (US) AY367947 AY367996 AY368041 B. berteroanus Colla Chile; PI 224789 (WRPIS) Keane 37 (ALTA) AY367946 AY367994 AY368040 Section Triniusia B. danthoniae Trin. Turkey; PI 598455 TU85-028-01 (WRPIS) Keane 15 (ALTA) AY367941 AY367990 AY368035 B. pseudodanthoniae Drobov Turkey; PI 204424 (WRPIS) Keane 21 (ALTA) AY367942 AY367991 AY368036 Table 2.3 continued Taxon Geographic origin/ source Voucher ITS trnh intron ndhY Outgroup Festuca breviglumis Swallen • Mexico: Durango Peterson et al. 16924 (US) AY367952 AY368001 AY368046 1 Collected outside native range. 4^ ON Figure 2.1. Phylogram of one of 449 most-parsimonious (MP) trees found using data from the internal transcribed spacer (ITS) region of nuclear ribosomal DNA. MP trees are each 380 steps, with a CI of 0.713 and RI of 0.826. Bootstrap support values greater than 50% are indicated. Nodes that collapse in the strict consensus tree are indicated with an arrow. Sections in Bromus [following Smith (1970) and Scholz (1998)] are indicated to the right of the tree. Native geographic distributions of species in sect. Bromopsis are indicated: light-shading = New World, dark shading = Old World, boxed species (B. pumpellianus) = New and Old World. 47 Fig. 2.1 81 laevipes ciliatus h kalmii latiglumis nottowayanus pubescens ir— anomalus 2 \*— grandis r exaltatus 100 Bromopsis frondosus 51 [L— suksdorfii I j lanatipes W- porter! 11 pseudolaevipes 941- mucroglumis ~ richardsonii 941 texensis ramosus 59 trr-^brachya'nthe~ra~ \j marginatus . — carinatus 7 5 1 I coloratus striatus subvelutinus H I ' catharticus 4.6£ cebadilla I r - berteroanus 94" gunckelli diandrus 100 96 55 92 scoparius japonicus pectinatus danthoniae • pseudodanthoniae \ Bromopsis j Bromopsis Ceratochloa Neobromus Bromus Triniusia I I i madritensis subsp. rubens 1 IP madritensis subsp. rubens 3 67 — madritensis subsp. rubens 2 erectus Genea ripanus mermis ir pumpellianus\ Bromopsis 100 98 i — | _ i — moi 79» ne 99 lanatus pflanzii destus 2 ^ — p llitus attenuatus Bromopsis 4 dolichocarpus • Triticum aestivum — Hordeum vulgare ^—^— Festuca breviglumis . 5 changes 48 Figure 2.2. Phylogram of one of 50,000 most-parsimonious (MP) trees found using combined plastid data from the trnL intron and the 3' end of ndhF. MP trees are each 218 steps, with a CI of 0.817 and PJ of 0.882. Bootstrap support values greater than 50% are indicated. Nodes that collapse in the strict consensus tree are indicated with an arrow. Sections in Bromus [following Smith (1970) and Scholz (1998)] are indicated to the right of the tree. Native geographic distributions of species in sect. Bromopsis are indicated: light shading = New World, dark shading = Old World, boxed species (B. pumpellianus) = New and Old World. 49 2.2 56 ./I 68 59 991 - anomalus 1 95|— erectUs " npanus pubescens anomalus 2 r ciliatus pseudolaevipes |- pumpellianus ramosus 95 exaltatus i-v frondosus J ~ i laevipes 11 lanatipes grandis mucroglumis mermis kalmii \— latiglumis porteri richardsonii pt texensis | L korotkoyi suksdorffi scopanus Bromopsis 91 japonicus pseudodanthoniae danthoniae Bromus I IUU I i l| . ' pectinatus | M madritensis subsp. rubens 1 56|__n— madritensis subsp. rubens 3 86* madritensis subsp. rubens 2 L pflanzii Triniusia Genea Bromus Genea 85r— lanatus """«- pellitus |- dohchocarpus - attenuatus Bromopsis 89 76 62 63 p brachyanthera marginatus l h coloratus cebadilla • striatus carinatus subvelutinus — catharticus L gunckelli berteroanus Triticum aestivum Hordeum vulgare | Bromopsis Ceratochloa Neobromus 5 changes Festuca breviglumis 50 2.6 R E F E R E N C E S Acedo, C , and Llamas, F. 1997. Bromus cabrerensis and Bromus nervosus, two new species from the Iberian Peninsula. Willdenowia 27, 47-55. Acedo, C , and Llamas, F. 1999. The genus Bromus L. (Poaceae) in the Iberian Peninsula. Monographiae Phanerogamerum 22, 1-293. Acedo, C , and Llamas, F. 2001. Variation of micromorphological characters of lemma andpalea in the genus Bromus (Poaceae). Annales Botanici Fennici 38, 1-14. Ainouche, M . L., Misset, M . - T . , and Huon, H. 1995. Genetic diversity in Mediterraneandiploid and tetraploid Bromus L. (section Bromus Sm.) populations. Genome 38, 879-888. Ainouche, M . L. , and Bayer, R. J. 1997. On the origins of the tetraploid Bromus species {sQciionBromus, Poaceae): insights from the internal transcribed spacer sequences of nuclear ribosomal DNA. Genome 40, 730-743. Ainouche, M . L., Bayer, R. J., Gourret, J.-P., Defontaine, A., and Misset, M.-T. 1999. The allotetraploid invasive weed Bromus hordeaceus L. (Poaceae): genetic diversity, origin and molecular evolution. Folia Geobotanica 34, 405-419. Aliscioni, S. S., Giussani, L. M . , Zuloaga, F. O., and Kellogg, E. A. 2003. A molecular phylogeny of Panicum (Poaceae: Paniceae): tests of monophyly and phylogenetic placement within the Panicoideae. American Journal of Botany 90, 796-821. Allison, K. J., Ashton, D. B., and Darbyshire, S. J. 2001. Identification of florets of meadow brome (Bromus riparius) and smooth brome (Bromus inermis). Seed Science and Technology 29, 99-108. 51 Allred, K. W. 1993. Bromus, section Pnigma, in New Mexico, with a key to the bromegrasses of the state. Phytologia 74, 319-345. Alvarez, A. E. , van de Wiel, C. C. M . , Smulders, M . J. M . , and Vosman, B. 2001. Use of microsatellites to evaluate genetic diversity and species relationships in the genus Lycopersicon. Theoretical and Applied Genetics 103, 1283-1292. Armstrong, K. C. 1975. Genome relationships in Bromus erectus, B. pumpellianus ssp. dicksonii, and B. pumpellianus. Canadian Journal of Genetics and Cytology 17, 391-394. Armstrong, K. C. 1981. The evolution of Bromus inermis and related species of Bromus sect. Pnigma. Botanische Jahrbucher fur Systematik, Pflanzengeschichte und Pflanzengeographie 102, 427-443. Armstrong, K. C. 1983. The relationship between some Eurasian and American species of Bromus section Pnigma as determined by the karyotypes of some F l hybrids. Canadian Journal of Botany 61, 700-707. Armstrong, K. C. 1991. Chromosome evolution of Bromus. Pp. 336-377. In: Tsuchiya, T. and Gupta, P. K. [Eds.], Chromosome engineering in plants: genetics, breeding, evolution, part B. Elsevier, Amsterdam, The Netherlands. Bacic, T., and Jogan, N. 2001. Multivariate morphometric study of the Bromus erectus group (Poaceae-Bromeae) in Slovenia. Phyton (Horn) 41, 295-311. Baldwin, B. G., and Markos, S. 1998. Phylogenetic utility of the external transcribed spacer (ETS) of 18S-26S rDNA: congruence of ETS and ITS trees of Calycadenia (Compositae). Molecular Phylogenetics and Evolution 10, 449-463. 52 Barker, F. K., and Lutzoni, F. M . 2002. The utility of the incongruence length difference test. Systematic Biology 51, 625-637. Bartlett, E. , Novak, S. J., and Mack, R. N. 2002. Genetic variation in Bromus tectorum (Poaceae): differentiation in the eastern United States. American Journal of Botany 89, 602-612. Beardsley, P. M . , Yen, A., and Olmstead, R. G. 2003. AFLP phylogeny of Mimulus section Eryanthe and the evolution of hummingbird pollination. Evolution 57, 1397-1410. Blattner, F. R. 1999. Direct amplification of the entire ITS region from poorly preserved plant material using recombinant PCR. Biotechniques 27, 1180-1186. Bomble, W., and Scholz, H. 1999. Eine neue Unterart des Bromus secalinus (Gramineae) -ein sekundares Unkraut. Feddes Repertorium 110, 425-438. Bruneau, A., Forest, F., Herendeen, P. S., Klitgaard, B. B., and Lewis, G. P. 2001. Phylogenetic relationships in the Caesalpinioideae (Leguminosae) as inferred from chloroplast trnL intron sequences. Systematic Botany 26, 487-514. Catalan, P., Kellogg, E. A., and Olmstead, R. G. 1997. Phylogeny of subfamily Pooideae based on ndhF gene sequences. Molecular Phylogenetics and Evolution 8, 150-166. Catalan, P., and Olmstead, R. G. 2000. Phylogenetic reconstruction of the genus Brachypodium P-Beauv. (Poaceae) from combined sequences of chloroplast ndhF gene and nuclear ITS. Plant Systematics and Evolution 220, 1-19. Chen, C.-H., and Kuoh, C.-S. 2000. The genus Bromus L. (Poaceae) in Taiwan: a D E L T A database for generating key and descriptions. Taiwania 45, 311-322. Clark, L. G., Zhang, W., and Wendel, J. F. 1995. A phylogeny of the grass family (Poaceae) based on ndhF sequence data. Systematic Botany 20, 436-460. 53 Clayton, W. D., and Renvoize, S. A. 1986. Genera graminum: grasses of the World. Kew Bulletin, Additional Series 13, 1-389. London. Clayton, W. D., Harman, K. T., and Williamson, H. 2002 onwards. World grass species: descriptions, identification, and information retrieval, http://www.kew.org/data/ grasses-db.html (Nov 2005). Davis, C. C , Fritsch, P. W., Li , J., and Donoghue, M . J. 2002. Phylogeny and biogeography of Cercis (Fabaceae): evidence from nuclear ribosomal ITS and chloroplast ndhF sequence data. Systematic Botany 27, 289-302. Davis, J. I., and Soreng, R. J. 1993. Phylogenetic structure in the grass family (Poaceae) as inferred from chloroplast DNA restriction site variation. American Journal of Botany 80, 1444-1454. Despres, L., Gielly, L. , Redoutet, B., and Taberlet, P. 2003. Using AFLP to resolve phylogenetic relationships in a morphologically diversified plant species complex when nuclear and chloroplast sequences fail to reveal variability. Molecular Phylogenetics and Evolution. 27, 185-196. Doyle, J. J., and Doyle, J. L. 1987. A rapid isolation procedure for small quantities of fresh tissue. Phytochemical Bulletin of the Botatnical Society of America 19, 11-15. Eckenrode, V. K., Arnold, J., and Meagher, R. B. 1985. Comparison of the nucleotide sequence of soybean 18S rRNA with the sequences of other small-subunit rRNAs. Journal of Molecular Evolution 21, 259-269. Farris, J. S., Kallersjo, M , Kluge, A. G., and Bult, C. 1994. Testing significance of incongruence. Cladistics 10, 315-319. 54 Farris, J. S., Kallersjo, M , Kluge, A. G., and Bult, C. 1995. Constructing a significance test for incongruence. Systematic Biology 44, 570-572. Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39, 783-791. Ferdinandez, Y. S. N., and Coulman, B. E. 2000. Characterization of meadow x smooth bromegrass hybrid populations using morphological characteristics. Canadian Journal of Pant Science 80, 551-557. Ferdinandez, Y. S. N., Somers, D. J., and Coulman, B. E. 2001. Estimating the genetic relationship of hybrid bromegrass to smooth bromegrass and meadow bromegrass using RAPD markers. Plant Breeding (New York) 120, 149-153. Ferdinandez, Y. S. N., and Coulman, B. E. 2002. Evaluating genetic variation and relationships among two bromegrass species and their hybrid using RAPD and AFLP markers. Euphytica 125,281-291. Forde, M . B., and Edgar, E. 1995. Checklist of pooid grasses naturalised in New Zealand. 3. Tribes Bromeae and Brachypodieae. New Zealand Journal of Botany 33, 35-42. Giussani, L. M . , Cota-Sanchez, J. H., Zuloaga, F. O., and Kellogg, E. A. 2001. A Molecular phylogeny of the grass subfamily Panicoideae (Poaceae) shows multiple origins of C4 photosynthesis. American Journal of Botany 88, 1993-2012. Gould, F. W., and Shaw, R. B. 1983. Grass systematics, ed. 2. Texas A & M University Press, College Station, Texas, USA. 416 p. Graham, S. W., Kohn, J. R., Morton, B. R., Eckenwalder, J. E. , and Barrett, S. C. H. 1998. Phylogenetic congruence and discordance among one morphological and three molecular data sets from Pontederiaceae. Systematic Biology 47, 545-567. 55 Graham, S. W., Reeves, P. A., Burns, A. C. E. , and Olmstead, R. G. 2000. Microstructural changes in noncoding chloroplast DNA: interpretation, evolution, and utility of indels and inversions in basal angiosperm phylogenetic inference. International Journal of Plant Sciences 161 (6 Suppl.), S83-S96. Graham, S. W., and Barrett, S. C. H. 2004. Phylogenetic reconstruction of the evolution of stylar polymorphisms in Narcissus (Amaryllidaceae). American Journal of Botany 91, 1007-1021. Grass Phylogeny Working Group [GPWG]. 2001. Phylogeny and subfamilial classification of the grasses (Poaceae). Annals of the Missouri Botanical Garden 88, 373-457. Green, J. M . , Barker, J. H. A., Marshall, E. J. P., Froud-Williams, R. J., Peters, N. C. B., Arnold, G. M . , Dawson, K., and Karp, A. 2001. Microsatellite analysis of the inbreeding grass weed Barren Brome (Anisantha sterilis) reveals genetic diversity at the within- and between-farm scales. Molecular Ecology 10, 1035-1045. Gutierrez, H. F., and Pensiero, J. F. 1998. Sinopsis de las especies Argentinas del genero Bromus (Poaceae). Darwiniana 35, 75-114. Hardig, T. M . , Soltis, P. S., and Soltis, D. E. 2000. Diversification of the North American shrub genus Ceanothus (Rhamnaceae): conflicting phylogenies from nuclear ribosomal DNA and chloroplast DNA. American Journal of Botany 87, 108-123. Hilu, K. W., Alice, L. A., and Liang, H. 1999. Phylogeny of Poaceae inferred from matK sequences. Annals of the Missouri Botanical Garden 86, 825-851. Hitchcock, A. S., and Chase, A. 1951. Manual of grasses of the United States, Second Edition. USDA Miscellaneous Publications 200, 1-1051. 56 Hodkinson, T. R., Chase, M . W., Lledo, M . D., Salamin, N., and Renvoize, S. A. 2002. Phylogenetics of Miscanthus, Saccharum and related genera (Saccharinae, Andropogoneae, Poaceae) based on DNA sequences from ITS nuclear ribosomal DNAand plastid trnL intron and trnL-F intergenic spacers. Journal of Plant Research 115,381-392. Holmstrom, G., and Scholz, H. 2000. Bromus bidentatus spec, nova and B. lepidus, two odd brome-grasses (Gramineae). Botanische Jahrbucher fur Systematik, Pflanzengeschichte und Pflanzengeographie 122, 195-200. Hsiao, C , Chatterton, N. J., Asay, K. H., and Jensen, K. B. 1994. Phylogenetic relationships of 10 grass species: an assessment of phylogenetic utility of the internal transcribed spacer region in nuclear ribosomal DNA in monocots. Genome 37, 112-120. Hsiao, C , Jacobs, S. W. L., Chatterton, N. J., and Asay, K. H. 1999. A molecular phylogeny of the grass family (Poaceae) based on the sequences of nuclear ribosomal DNA (ITS). Austral. Systematic Botany 11,667-688. International Plant Names Index [IPNI]. 2004. http://www.ipni.org/index.html [November 2004]. Joachimiak, A., Sutkowska, A., and Mitka, J. 2001. RAPD studies in Bromus (Poaceae) from the Old and New worlds-preliminary results. Acta Biologica Hungarica 43, 79-86. Kahler, A. L. , Krzakowa, M . , and Allard, R. W. 1981. Isozyme phenotypes in five species of Bromus sect. Genea. Botanische Jahrbucher fur Systematik, Pflanzengeschichte und Pflanzengeographie 102, 401-409. 57 Keane, R. M . 2002. Predicting exotic plant invasions: a comparative approach using the genus Bromus L. Ph.D. dissertation, Imperial College London, UK. Klak, C , Hedderson, T. A., and Linder, H. P. 2003. A molecular systematic study of the Lampranthus group (Aizoaceae) based on the chloroplast trnL-tmV and nuclear ITS and 5S NTS sequence data. Systematic Botany 28, 70-85. Kolosha, V. O., and Fodor, I. 1990. Nucleotide sequence of Citrus limon 26S rRNA gene and secondary structure model of its RNA. Plant Molecular Biology 14, 147-161. Kosina, R. 1996. Embryo morphology in the genus Bromus (Poaceae). Fragmenta Floristica et Geobotanica41, 563-576. Kosina, R. 1999. Patterns of flower microstructural variation within the genus Bromus. Acta Societatis Botanicorum Poloniae 68, 221-226. Kozuharov, S., Petrova, A., and Ehrendorfer, F. 1981. Evolutionary patterns in brome grass species (Bromus, Gramineae) of the Balkan Peninsula. Botanische Jahrbucher fur Systematik, Pflanzengeschichte und Pflanzengeographie 102, 381-391. Les, D., Crawford, D. J., Landolt, E. , Gabel, J. D., and Kimball, R. T. 2002. Phylogeny and systematics of Lemnaceae, the duckweed family. Systematic Botany 27, 221-240. Louis-Marie, F. 1928. The genus Trisetum in America. Rhodora 30: 209-223. Markos, S., and Baldwin, B. G. 2002. Structure, molecular evolution, and phylogenetic utility of the 5' region of the external transcribed spacer of 18S-26S rDNA in Lessingia (Compositae, Astereae). Molecular Phylogenetics and Evolution 23, 214-228. 58 Mason-Gamer, R. 2001. Origin of North American Elymus (Poaceae: Triticeae) allotetraploids based on granule-bound starch synthase gene sequences. Systematic Botany 26, 757-768. Massa, A. N., Larson, S. R., Jensen, K. B., and Hole, D. J. 2001. AFLP variation in Bromus section Ceratochloa germplasm of Patagonia. Crop Science (Madison) 41, 1609-1616. Massa, A. N., Jensen, K. B., Larson, S. R., and Hole, D. J. 2004. Morphological variation in Bromus sect. Ceratochloa germplasm of Patagonia. Canadian Journal of Botany 82: 136-144. Matthei, O. 1986. The genus Bromus L. (Poaceae) in Chile. Gayana Botanica 43, 47-110. Mitchell, W. W. 1967. Taxonomic synopsis of Bromus section Bromopsis (Gramineae) in Alaska. Canadian Journal of Botany 45, 1309-1313. Naganowska, B. 1993a. Karyotypes of five Bromus species of Genea section. Genetica Polonica 34, 197-213. Naganowska, B. 1993b. Cytophotometric analysis of DNA content in cell nuclei of five Bromus species of Genea section. Genetica Polonica 34, 215-221. Naranjo, C. A., Arias, F. H., Gil, F. E. , and Soriano, A. 1990. Bromus pictus of the B. setifolius complex (section Pnigma): numerical taxonomy and chromosome evidence for species rank. Canadian Journal of Botany 68, 2493-2500. Nixon, K. C. 1999. The Parsimony Ratchet, a new method for rapid parsimony analysis. Cladistics 15,407-414. Novak, S. J., and Mack, R. N. 2001. Tracing plant introduction and spread: genetic evidence from Bromus tectorum (cheatgrass). Bioscience 51, 114-122. 59 Ogle, S. M . , Reiners, W. A., and Gerow, K. G. 2003. Impacts of exotic annual brome grasses (Bromus spp.) on ecosystem properties of northern mixed grass prairie. American Midland Naturalist 149, 46-58. Oja, T. 1998. Isoenzyme diversity and phylogenetic affinities in the section Bromus of the grass genus Bromus (Poaceae). Biochemical Systematics and Ecology 26, 403-413. Oja, T. 1999. Allozyme diversity and interspecific differentiation of the two diploid bromegrass species, Bromus tectorum L. and B. sterilis L. (Poaceae). Plant Biology (Stuttgart) 1, 679-686. Oja, T. 2002a. Genetic divergence and interspecific differentiation in the Bromus madritensis complex (Poaceae) based on isozyme data. Biochemical Sysematics and Ecology 30, 433-449. Oja, T. 2002b. Bromus fasciculatus Presl-a third diploid progenitor of Bromus section Genea allopolyploids (Poaceae). Hereditas 137, 113-118. Oja, T. 2002c. Comparative study of the ploidy series Bromus sterilis, B. diandrus and B. rigidus (Poaceae) based on chromosome numbers, morphology and isozymes. Plant Biology (Stuttgart) 4, 484-491. Oja, T. 2006. The origin and diploid progenitors of Bromus section Genea (Poaceae) allopolyploids: isozyme evidence. In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative Biology and Evolution, Poales. Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Oja, T., and Jaaska, V. 1998. Allozyme diversity and phylogenetic relationships among diploid annual bromes (Bromus, Poaceae). Annates Botanici Fennici 35, 123-130. 60 Oja, T., Jaaska, V., and Vislap, V. 2003. Breeding system, evolution and taxonomy of Bromus arvensis, B. japonicus and B. squarrosus (Poaceae). Plant Systematics and Evolution 242, 101-117. Olmstead, R. G., and Sweere, J. A. 1994. Combining data in phylogenetic systematics: an empirical approach using three molecular data sets in the Solanaceae. Systematic Biology 43,467-481. Pavlick, L. E. 1995. Bromus L. of North America. Royal British Columbia Museum, Victoria, British Columbia, Canada. 161 p. Peterson, P. M. , and Planchuelo, A. M . 1998. Bromus catharticus in South America (Poaceae: Bromeae). Novon 8, 53-60. Peterson, P. M. , Cayouette, J., Ferdinandez, Y. S. N., Coulman, B. E. , and Chapman, R. E. 2001 (2002). Recognition of Bromus richardsonii and B. ciliatus: evidence from morphology, cytology, and DNA fingerprinting (Poaceae: Bromeae). Aliso 20, 21-36. Pillay, M . 1993. Chloroplast genome organization of bromegrass, Bromus inermis Leyss. Theoretical and Applied Genetics 86, 281-287. Pillay, M . , and Hilu, K. W. 1990. Chloroplast DNA variation in diploid and polyploid species of Bromus (Poaceae) subgenera Festucaria and Ceratochloa. Theoretical and Appied Genetics 80, 326-332. Pillay, M . , and Hilu, K. W. 1995. Chloroplast-DNA restriction site analysis in the genus Bromus (Poaceae). American Journal of Botany 82, 239-249. Pinto-Escobar, P. 1981. The genus Bromus in northern South America. Botanische Jahrbucher fur Systematik, Pflanzengeschichte und Pflanzengeographie 102, 445-457. 61 Pinto-Escobar, P. 1986. El genero Bromus en Los Andes Centrales de Suramerica. Caldasia 15, 15-34. Planchuelo, A. M. , and Peterson, P. M . 2000. The species of Bromus (Poaceae: Bromeae) in South America. Pp. 89-101 In: Jacobs, S. W. L. and J. Everett [eds.], Grasses: systematics and evolution. CSIRO, Melbourne, Victoria, Australia. Puecher, D. I., Robredo, C. G., Rios, R. D., and Rimieri, P. 2001. Genetic variability measures among Bromus catharticus Vahl populations and cultivars with RAPD and AFLP markers. Euphytica 121, 229-236. Ramakrishnan, A. P., Coleman, C. E. , Meyer, S. E. , and Fairbanks, D. J. 2002. Microsatellite markers for Bromus tectorum (cheatgrass). Molecular Ecology Notes 2, 22-23. Rambaut, A. 1998. Se-Al (sequence alignment editor vers. 1.0 alpha 1). Department of Zoology, University of Oxford, Oxford, UK. Sales, F. 1993. Taxonomy and nomenclature of Bromus sect. Genea. Edinburgh Journal of Botany 50, 1-31. Sales, F. 1994a. A reassessment of the Bromus madritensis complex (Poaceae): a multivariate approach. Israel Journal of Plant Science 42, 245-255. Sales, F. 1994b. Evolutionary tendencies in some annual species of Bromus (Bromus L. sect. Genea Dum. (Poaceae)). Botanical Journal of the Linnean Society 115, 197-210. Sales, F., and Smith, P. 1990. A new species in the genus Bromus. Edinburgh Journal of Botany 47, 361-366. 62 Schifino, M . T., and Winge, H. 1983. Karyotypes and nuclear DNA content of species of the Briza complex and some other genera of Poeae (Gramineae). Revista Brasileira de Genetica 6, 245-259. Scholz, H. 1981. Der Bromus-pectinatus-Komplex (Gramineae) im Nahen und Mittleren Osten. Botanische Jahrbucher fur Systematik, Pflanzengeschichte und Pflanzengeographie 102, 471-495. Scholz, H. 1997. A new annual brome-grass (Bromus) from Greece. Feddes Repertorium 108, 97-100. Scholz, H. 1998. Notes on Bromus danthoniae and relatives (Gramineae). Willdenowia 28, 143-150. Scholz, H., and Mos, U. 1994. Status and short history of the extinct cultivated cereal Bromus mango E. Desv. - and the origin of Bromus secalinus L. Flora 189, 215-222. Schulz-Schaeffer, J. 1960. Cytological investigations in the genus Bromus. III. The cytotaxonomic significance of the satellite chromosomes. Journal of Heredity 51, 269-277. Shaw, J., Lickey, E. B., Beck, J. T., Farmer, S. B., Liu, W., Miller, J., Siripun, K. C., Winder, C. T., Schilling, E. E. , and Small, R. L. 2005. The tortoise and the hare II: relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis. American Journal of Botany 92, 142-166. Shear, C. L. 1900. A revision of the North American species of Bromus occurring north of Mexico. Bulletin U.S.D.A. (1895-1901) 23, 1-66. 63 Sikes, D. S., and Lewis, P. O. 2001. Beta software, version 1. PAUPRat: PAUP* implementation of the parsimony ratchet. Distributed by the authors. Department of Ecology and Evolutionary Biology, University of Connecticut, Storrs, USA. Smith, P. 1969. Serological relationships of Bromus L. and Boissiera Hochst. ex Steud. Feddes Repertorium 79, 337-345. Smith, P. 1970. Taxonomy and nomenclature of the brome-grasses {Bromus L. s. 1.). Notes from the Royal Botanic Garden, Edinburgh 30, 361-375. Smith, P. 1972. Serology and species relationships in annual bromes {Bromus L. sect. Bromus). Annals of Botany (Oxford) 36, 1-30. Smith, P. 1985a. Observations on Turkish brome-grasses. 1. Some new taxa, new combinations and notes on typification. Notes from the Royal Botanic Garden, Edinburgh 42, 491-501. Smith, P. 1985b. Bromus. Pp. 145-301 In: Davis, P. H. [Ed.], Flora of Turkey and the east Aegean Islands, volume 9. Edinburgh University Press, Edinburgh, UK. Smith, P., and Sales, F. 1993. Bromus L. sect. Bromus: taxonomy and relationships of some species with small spikelets. Edinburgh Journal of Botany 50, 149-171. Soderstrom, T. R., and Beaman, J. H. 1968. The genus Bromus (Gramineae) in Mexico and Central America. Publications of the Museum, Michigan State University, Biological Series 3, 465-520. Soreng, R. J., and Davis, J. I. 1998. Phylogenetics and character evolution in the grass family (Poaceae): simultaneous analysis of morphological and chloroplast DNA restriction site character sets. Botanical Review (Lancaster) 64, 1-85 64 Soreng, R. J., and Davis, J. I. 2000. Phylogenetic structure in Poaceae subfamily Pooideae as inferred from molecular and morphological characters: misclassification versus reticulation. Pp. 61-74. In: Jacobs, S. W. L., and Everett, J. [Eds.]. Grasses: Systematics and Evolution. CSIRO, Melbourne, Victoria, Australia. Soreng, R. J., Peterson, P. M . , Davidse, G., Judziewicz, E. J., Zuloaga, F. O., Filgueiras, T. S., and Morrone, O. 2003. Catalogue of New World grasses (Poaceae): IV. Subfamily Pooideae. Contributions from the U.S. National Herbarium 48, 1-730. Spalton, L. M . 2001. A new subspecies of Bromus hordeaceus L. (Poaceae). Watsonia 23, 525-531. Spalton, L. M . 2002a. An analysis of the characters of Bromus racemosus L. , B. commutatus Schrad., and B. secalinus L. (Poaceae). Watsonia 24, 193-202. Spalton, L. M . 2002b. A new key to the tribe Bromeae in Britain and Ireland. BSBI News 91, 15-17. Spalton, L. M . 2004. A key to Bromeae in the Mediterranean climatic zones of southern Europe, southwest Asia, and north Africa. BSBI News 95, 22-26. Spangler, R., Zaitchik, B., Russo, E. , and Kellogg, E. 1999. Andropogoneae evolution and generic limits in Sorghum (Poaceae) using ndhF sequences. Systematic Botany 24, 267-281. Starr, J. R., Harris, S. A., and Simpson, D. A. 2003. Potential of the 5' and 3' ends of the intergenic spacer (IGS) of rDNA in the Cyperaceae: new sequences for lower-level phylogenies in sedges with an example from Uncinia Pers. International Journal of Plant Sciences 164, 213-227. 65 Stebbins, G. L. 1947. The origin of the complex of Bromus carinatus and its phytogeographic implications. Contributions of the Gray Herbarium 165, 42-55. Stebbins, G. L. 1956. Cytogenetics and evolution in the grass family. American Journal of Botany 43, 890-905. Stebbins, G. L. 1981. Chromosomes and evolution in the genus Bromus (Gramineae). Botanische Jahrbucher fur Systematik, Pflanzengeschichte und Pflanzengeographie. 102,359-379. Stebbins, G. L., and Togby, H. A. 1944. The cytogenetics of hybrids in Bromus L. hybrids within the section 'Ceratochloa'. American Journal of Botany 31, 1-11. Swofford, D. L. 2002. PAUP*: phylogenetic analysis using parsimony (and other methods), vers. 4.0b 10. Sinauer, Sunderland, Massachusetts, USA. Taberlet, P., Gielly, L. , Pautou, G., and Bouvet, J. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Molecular Biology 17,1105-1109. Tsvelev, N . N . 1976. Grasses of the Soviet Union. Part I. Nauka Publishers, Leningrad, Russia. 788 p. Tsvelev, N. N. 1989. The system of grasses (Poaceae) and their evolution. Botanical Review (Lancaster) 55, 141-203. Veldkamp, J. F., Eriks, M . , and Smit, S. S. 1991. Bromus (Gramineae) in Malesia. Blumea 35, 483-497. Wagnon, H. K. 1950. Three new species and one new form in Bromus. Leaflets of Western Botany 6, 64-49. 66 Wagnon, H. K. 1952. A revision of the genus Bromus, section Bromopsis, of North America. Brittonia 7, 415-480. Wendel, J. F., and Doyle, J. J. 1998. Phylogenetic incongruence: window into genome history and speciation. Pp. 265-296 In: Soltis, P. S., Soltis, D. E. , and Doyle, J. J. [Eds.]. Molecular Systematics of Plants. Chapman and Hall, New York, USA. White, T. J., Bruns, T., Lee, S., and Taylor, J. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. Pp. 315-322 In: Innis, M . , Gelfand, D., Sninsky, J., and White, T. [Eds.]. PCR protocols: a guide to methods and applications. Academic Press, San Diego, California, USA. Wilton, A. C. 1965. Phylogenetic relationships of an unknown tetraploid Bromus, section Bromopsis. Canadian Journal of Botany 43, 723-730. Winkworth, R. C , Grau, J., Robertson, A. W., and Lockhart, P. J. 2002. The origins and evolution of the genus Myosotis L. (Boraginaceae). Molecular Phylogenetics and Evolution 24, 180-193. Yoder, A. D., Irwin, J. A., and Payseur, B. A. 2001. Failure of the ILD to determine data combinability for slow loris phylogeny. Systematic Biology 50, 408-424. Yokota, Y., Kawata, T., Iida, Y., Kato, A., and Tanifuji, S. 1989. Nucleotide sequences of the 5.8S rRNA gene and internal transcribed spacer regions in carrot and broad bean ribosomal DNA. Journal of Molecular Evolution 29, 294-301. Zajac, E. U. 1996a. Kompleks Bromus mollis (Poaceae) w Polsce. Fragmenta Floristica et Geobotanica 3,3-9. Zajac, E. U. 1996b. Interspecies differentiation of Bromus hordeaceus (Poaceae) in Poland. Fragmenta Floristica et Geobotanica 41, 577-578. 67 C H A P T E R 3 1 Phylogenetic Relationships Among the Subfamilies of Grasses (Poaceae) and Related Families Inferred from a Large, Multigene Plastid Data Set 3.1. INTRODUCTION 3.1.1. The Grass Family (Poaceae) The grasses (Poaceae) are the fourth largest family of angiosperms, with some 11,000 species classified in 600 to 900 genera (Campbell 1985; Clayton and Renvoize 1986; Tzvelev 1989). Grasses provide many of the world's major crops, and they are important components of most global ecosystems. As a result, the family has received substantial systematic study. Over the last two centuries, numerous classifications of Poaceae have been proposed [e.g., Brown 1810; Hitchcock and Chase 1951; Beetle 1955; Stebbins 1957; Hilu and Wright 1982; Clayton and Renvoize 1986; Tsvelev 1989; Watson and Dallwitz 1992; see Grass Phylogeny Working Group (GPWG; 2001) for a comprehensive review], based principally on data from morphology (e.g., Brown 1810), anatomy (e.g., Reeder 1957; Brown 1958; Tateoka et al. 1959), and cytology (e.g., Avdulov 1931; Brown and Emery 1957). Recent phylogenetic studies of morphological (e.g., Kellogg and Campbell 1987; Kellogg Chapter 3 represents a draft manuscript that will be submitted for publication: Saarela, J . M . , and Graham, S. W. Phylogenetic relationships among the subfamilies of grasses (Poaceae) and related families inferred from a large, multigene plastid data set. 68 and Watson 1993) and molecular data (e.g., Barker et al. 1995; Columbus et al. 2000; Soreng and Davis 2000; Giussani et al. 2001; Ge et al. 2002; Mathews et al. 2002; Aliscioni et al. 2003; Chapter 2; see additional refs. below) have clarified taxonomic circumscriptions and hypotheses of evolutionary relationships throughout the family. Several family-level phylogenetic investigations of Poaceae have been conducted with plastid and nuclear data. Plastid studies have been based on restriction-site variation (Davis and Soreng 1993; Soreng and Davis 1998, 2000) and DNA sequence data from single protein-coding genes (Duvall and Morton 1996; Cummings et al. 1994; Nadot et al. 1994; Clark et al. 1995, Catalan et al. 1997; Liang and Hilu 1996; Hilu and Alice 1999; Hilu et al. 1999), a noncoding region (Zhang 2000), and complete plastid genomes (Matsuoka et al. 2002). Nuclear studies have been conducted with various coding and noncoding regions (Hamby and Zimmer 1988; Hsiao et al. 1994, 1999; Mason-Gamer et al. 1998; Mathews and Sharrock 1996; Mathews et al. 2000). These, as well as additional studies with narrower taxonomic foci (e.g., Catalan et al. 1997; Barker et al. 2000; Gomez-Martinez and Culham 2000; Soreng and Davis 2000; Zhang and Clark 2000; Hilu and Alice 2000, 2001; Ge et al. 2002), have identified several major lineages in Poaceae with moderate to robust support. As a result, the broad phylogenetic structure of Poaceae is fairly well established. The GPWG (2000, 2001) based a new subfamilial classification on a fairly robust phylogeny generated using combined data from seven molecular datasets and morphology, although a few taxa (Eriachne R. Br., Gynerium Willd. ex P. Beauv., Micraira F. Muell, and Streptogyna P. Beauv.) were treated as incertae sedis because of unclear and poorly supported placements. Recent work, however, has placed these latter taxa in Poaceae with moderate to strong support (Duvall et al. 2006). Poaceae include three small subfamilies (Anomochlooideae, 69 Pharoideae, and Puelioideae) that are the successive sister groups (respectively) of a large clade that is divided into two major lineages. One lineage includes subfamilies Panicoideae [including Centothecoideae (Soreng et al. 2003; Duvall et al. 2006)], Aristidoideae, Chloridoideae, Arundinoideae, Danthonioideae, and the recently recognized Micrairoideae [the "PACMAD" clade (Duvall et al. 2006), called the "PACC" or "PACCAD" clade in earlier studies], and the other includes subfamilies Bambusoideae, Ehrhartoideae, and Pooideae (the "BEP" clade, also called the "BOP" clade in some earlier studies; GPWG 2001). The BEP clade, however, has not been uniformly recovered; several analyses indicate weakly that Pooideae may instead be the sister group of the P A C M A D clade (e.g., Soreng and Davis 1998; GPWG 2001; Caetono-Anolles 2005a; Duvall et al. 2006). Relationships among subfamilies within these major lineages generally have been resolved inconsistently and are supported poorly (e.g., Duvall and Morton 1996; Hilu et al. 1999; Hsiao et al. 1999; Zhang 2000; GPWG 2001; Duvall et al. 2006). This is unfortunate, as understanding evolutionary patterns among the major lineages in Poaceae is critical for providing a robust framework for macro-evolutionary and comparative genomic investigations of extant grasses (including cereals) and their relatives (e.g., Kellogg 1998; Gaut 2000; Caetano-Anolles 2005b; Rudall et al. 2005), for the reconstruction of the timing and pattern of diversification of major clades in the family (e.g., Stromberg 2005; Prasad et al. 2005), and for improved understanding of the grass fossil record"(e.g., Cfepef and Feldman 1991; Prasad et al. 2005). 3.1.2. The Sister Group of Poaceae Correct inference of the identity of the sister group of the grasses has also been problematic. Phylogenetic studies of morphological and molecular data indicate that the 70 closest living relatives of Poaeae are Joinvilleaceae (e.g., Campbell and Kellogg 1987; Doyle et al. 1992; GPWG 2001), Ecdeiocoleaceae (Bremer 2002; Michelangeli et al. 2003; Davis et al. 2004), or a lineage that includes both families (Bremer 2000); relationships among these three families were unresolved in a recent analysis of the ndhF gene (Givnish et al. 2006). Additional molecular studies have indicated that either of these families is the sister-group of Poaceae, but none sampled both families (e.g., Davis and Soreng 1993; Davis 1995; Katayama and Ogihara 1996; Soreng and Davis 1998; Linder et al. 2000; Briggs et al. 2000; Hilu et al. 2003; Graham et al. 2006). In contrast to the large size and cosmopolitan distribution of Poaceae, Joinvilleaceae are monogeneric, with two species distributed throughout Malaysia and the south Pacific (Newell 1969), and Ecdeiocoleaceae are bigeneric, with two species distributed in southwestern Australia (Cutler and Shaw 1965; Briggs and Johnson 1998). A third undescribed species from western Australia, morphologically similar to Ecdeiocolea monostachya, has recently been discovered (B. G. Briggs, pers. comm.). Resolution of the sister group of Poaceae is important for making inferences about the origins and evolution of reproductive characteristics in Poaceae (e.g., Rudall and Bateman 2004; Rudall et al. 2005) and to clarify broader patterns of diversification within Poales. 3.1.3. Objectives of the Study One strategy to obtain more fully resolved and better supported phylogenetic trees is to examine more characters per taxon (e.g., Hillis 1998; Poe and Swofford 1999; Graham and Olmstead 2000a; Rokas et al. 2003; Wortley et al. 2005; Graham et al. 2006). I take this approach here to clarify relationships among subfamilies within Poaceae, and to address the 71 identity of their sister group. I perform phylogenetic analyses of a large, plastid data set from representatives of most of the grass subfamilies. A similar complement of plastid regions has been used to resolve deep phylogenetic relationships among basal angiosperms (Graham and Olmstead 2000a,b; Graham et al. 2000), cycads (Rai et al. 2003; Zgurski 2004), and monocots as a whole (McPherson 2003; Zgurski 2004; Graham et al. 2006), but this combination of genes and intergenic spacers has not previously been sampled intensively within a single family. Within-family phylogenetic studies generally focus on an examination of a few rapidly evolving (often noncoding) regions, but larger genomic samples are becoming increasingly feasible and affordable to examine (Graham et al. 2006), and the slower rates in some of the regions examined are likely to be compensated for by the large number of regions examined, and by the generally elevated substitution rates found in Poaceae (e.g., Bousquet et al. 1992; Gaut et al. 1992, 1997). 3.2 M A T E R I A L S AND M E T H O D S 3.2.1. Taxonomic Sampling I included at least one exemplar taxon from ten of the twelve Poaceae subfamilies according to GPWG (2001) and Duvall et al. (2006) (Table 3.1). Material representing subfamilies Micrairoideae and Puelioideae was not available for this study. It is arguably appropriate (e.g., Chase et al. 2006) to use one or a few exemplar species to represent each subfamily in this way, as the monophyly of each subfamily has been recovered repeatedly in most previous analyses (see earlier references). I obtained new sequence data for ten grass species, and used published sequence data from GenBank for Oryza sativa (Hiratsuka et al. 1989), Triticum aestivum (Ogihara et al. 2000) and Zea mays (Maier et al. 1995) (Table 3.1) 72 I obtained the rbcL sequence for Pharus from GenBank for a congeneric species (P. latifolius L.). Outside of Poaceae, I included new sequence data from one species of Joinvilleaceae (Table 3.1), and previously published data from exemplar taxa representing 59 other monocot families (Graham and Olmstead 2000a; Graham et al. 2006). This broad taxon sampling represents all orders of monocots, and includes Dasypogonaceae, a family currently unplaced to order (APG II 2003; Chase 2004). 3.2.2. Genomic Sampling For most taxa, I extracted DNA from silica gel-dried leaf material using the CTAB protocol of Doyle and Doyle (1987), with the addition of 2% y5-mercaptoethanol to each extraction. I amplified and sequenced multiple plastid genes involved in photosynthesis (i.e., atpB, psbB, psbC, psbD, psbE, psbF, psbH, psbJ, psbL, psbH, psbT, rbcL), chlororespiration (ndhF, ndhB), and plastid translation (rpll, V-rpsXl, rpsl), and their associated noncoding regions [i.e., six intergenic regions in two of the photosystem II gene clusters, one intron each in rpll, 3 ' -rpsl2 and ndhB, and intergenic spacer regions between 3 ' -rpsl2, rpsl, ndhB, and trnL(CAA); see Graham and Olmstead (2000a) and Graham et al. (2006) for further details]. Most of the regions examined are in the single copy (SC) regions of the plastid genome, but four (rpll, 3 ' -rpsl2, rpsl, and ndhB) are in the large inverted repeat (IR) region in most land plants, including grasses examined thus far. I amplified and sequenced DNA using the methods and primers described in Graham and Olmstead (2000a), McPherson (2003), and Graham et al. (2006). I sequenced all regions in the forward and reverse directions, and obtained sequencing products from duplicate amplification products obtained 73 from independent DNA extractions to detect pipetting errors or cross-contamination (I observed none). 3.2.3. Data Assembly I performed basecalling and contig assembly using Sequencher 4.1 (Genes Code Corporation, Ann Arbor, MI), determining gene boundaries by comparison to sequences of Nicotiana tobacum L. (Shinozaki et al. 1986) and Ginkgo biloba L. These sequences were added to previously published alignments that include representatives from across the seed plants (Graham and Olmstead 2000a; Rai et al. 2003; Graham et al. 2006) using Se-Al version 1.0 alpha (Rambaut 1998), according to guidelines outlined in Graham et al. (2000). With few exceptions, all taxa are represented nearly completely for all regions (Table 3.1). The length of the aligned data matrix is 28,062 bp. The unaligned sequence lengths of exemplars from the commelinid monocots examined range from -13.9 kb in Strelitzia Ait. (see Graham et al. 2006) to -15.8 kb in Oryza L. , with a mean unaligned length of-15.3 kb. The aligned data matrix is substantially larger than the unaligned length for any single taxon because of substantial gaps and/ or unalignable regions, including some from non-angiosperm taxa that are part of the matrix but not included in this study. Those noncoding regions that were too difficult to align were set aside as staggered gapped regions primarily consisting of unique sequences. The unique regions are effectively ignored for parsimony-based tree searches and scores (Graham et al., 2006), and should have only minimal effect for model-based methods (e.g., on estimation of base frequency parameter values). I coded gap cells in the data as "missing data" and did not attempt to score inferred insertion/deletion (indel) events. 74 3.2.4. Parsimony, Likelihood, and Bayesian Phylogenetic Analyses I examined two different taxon sets, a large 79-taxon (58 family) set using maximum parsimony alone, and a smaller 28-taxon set (consisting of the 13 grasses, 10 exemplar Poales and several representatives of Commelinales and Zingiberales as outgroup taxa) with three different phylogenetic inference methods: (1) maximum parsimony (MP); (2) maximum likelihood (ML) analysis using P H Y M L version 2.4.4 (Guindon and Gascuel 2003), and; (3) Bayesian inference using MrBayes 3.1 (Ronquist and Huelsenbeck 2003). The smaller taxon set was used for model-based approaches. For the monocot-wide taxon set I accepted Acorus as the probable sister group of the rest of the monocots (e.g., Duvall et al. 1993b; Chase et al. 2000; Graham et al. 2006); exemplar species from Commelinales and Zingiberales were used as outgroups for the reduced taxon set. The MP searches were performed using PAUP* version 4.0M0 (Swofford 2002), with all characters and character-state changes equally weighted, using tree bisection-reconnection (TBR) branch-swapping, and with 100 random addition replicates performed for each search, but otherwise using default conditions. To examine whether the results are affected by data partitions that have different DNA substitutional dynamics, I analyzed four data subpartitions (defined as CHARSETs in the Nexus file) for the reduced taxon set using MP: (i) codon positions 1 and 2; (ii) codon position 3; (iii) all protein-coding regions, and; (iv) all noncoding regions. For the model-based phylogenetic inference methods I chose optimal DNA substitution models using the hierarchical likelihood ratio test (hLRT) and the Akaike Information Criterion (AIC), as implemented in ModelTest ver. 3.7 (Posada and Crandall 1998). The optimal model was GTR + Y + I [general time reversible rate (GTR) model with 75 a proportion of invariant sites (I) considered and a gamma (r) distribution used to account for among-site rate variation)] for the combined data set, the protein-coding partition, and the noncoding partition; the T V M + F + I model is optimal for the partitions consisting of codon positions 1 and 2, and codon position 3 (for model details see ModelTest documentation; Posada and Crandall 1998). These models were used for the M L analyses of the respective data partitions, with model parameters estimated from the data, and otherwise using default settings in PHYML. For Bayesian analyses, I used the GTR + F + I model of evolution in each analysis. For all analyses, I ran two parallel sets of four chains (three heated and one cold) simultaneously, starting each Markov chain from a random tree. I stopped runs after the average standard deviation of split frequencies (a convergence diagnostic) was less than 0.01. I sampled trees every 1,000 generations, discarding the first 25% of the trees as burn-in. I conducted Bayesian analyses for the data partitions analyzed using MP and M L , but also conducted analyses of the full data set with substitution-model parameters estimated independently ('unlinked') for each of several data subpartitions (Ronquist and Huelsenbeck 2003). Unlinking may be desirable when partitions of the data have different substitutional dynamics, as is clearly the case for different codon positions and between the IR (inverted repeat) and SC (single copy) regions in monocot plastid genomes (e.g., Graham et al. 2006). In total, I conducted four unlinked analyses (unlinked partitions are separated by a comma): (i) codon position 1, codon position 2, and codon position 3; (ii) SC codon positions 1 and 2, SC region codon position 3, IR codon positions 1 and 2, IR codon position 3; (iii) SC and IR protein-coding regions, IR noncoding regions, SC noncoding regions; and (iv) SC codon positions 1 and 2, SC codon position 3, SC noncoding regions, IR codon positions 1 and 2, 76 IR codon position 3, and IR noncoding regions. These variants exclude (i, ii) or include (iii, iv) noncoding data, and consider SC vs. IR variants of different codon positions as separate partitions (i.e., variants ii and iv), or lump these together (i, iii). To estimate posterior probabilities of individual clades for each analysis, I computed a majority-rule consensus tree in PAUP4.0blO. I estimated branch support for MP and M L using nonparametric bootstrap analysis (Felsenstein 1985). For MP bootstrap analysis, I used 100 (79-taxon data set) and 1,000 (28-taxon data sets) bootstrap replicates, with the same search parameters as the initial search, but using ten random addition replicates per bootstrap replicate. For M L bootstrap analysis I used 500 bootstrap replicates. Clade posterior probabilies, estimated using MrBayes, are expressed as percentages here. With recognition of the arbitrariness in placing levels of bootstrap support into defined classes, I consider "strongly" supported or "robust" branches to have bootstrap support of 90% or more, "moderately" supported branches to have bootstrap support 70-89%), and "weakly" or "poorly" supported branches to have bootstrap support < 70%. To compare support values among the various maximum parsimony, maximum likelihood, and Bayesian analyses conducted with the 28-taxon data set, I determined and graphed support values for all clades within Poaceae and among Ecdeiocoleaceae-Joinvilleaceae-Poaceae that were recovered with bootstrap or posterior probability support greater than 50% in at least one of the 19 analyses (see Fig. 4 in Graham et al. 1998). I recovered support values <50% from the bootstrap log for each analysis, ignoring values <5%. I identified each clade with a unique letter [clades "a" to "q" refer to branches recovered in Figs. 3.2, 3.3, "r" to "aa" to other branches, noted in legend to Fig. 3.4]. 77 To assess heterogeneity in the rates of evolution among four monocot lineages, I conducted a series of likelihood ratio tests (LRT) at various hierarchical levels: 1) all monocots (n = 79); 2) commelinid monocots (excluding Dasypogonaceae and Arecaceae; n = 28); 3) Poales (n = 23); and 4) Poaceae (n =13). I conducted the LRT test using the GTR + r +1 model with and without a molecular clock enforced. 3.3 R E S U L T S 3.3.1. Sequence Characteristics In the 28-taxon commelinid data set, there are 2,407 potentially parsimony informative characters (PPIC; Table 3.2). Among commelinid taxa, -85% of the PPIC are from the single copy regions, whereas -92% of the PPIC in Poaceae are from the single copy regions. Within the single copy regions, combined data from atpB, ndhF, and rbcL provide -1.4 times as many PPIC in the commelinids and Poaceae compared with the combined photosystem II regions, for a similar length of sequence examined [a consequence of a lower subsitution rate, also reflected in fewer PPIC per unaligned nucleotide, and consistent with substitution rates noted for some of these genes in Olmstead and Palmer (1994)]. In single-copy regions, codon position 3 provides the greatest percentage of PPIC, and more PPIC per unaligned character than the noncoding regions. Noncoding nucleotides in the IR evolve at the same rate (Poaceae), or are more slowly evolving (commelinids) than IR codon position 3 (i.e., fewer PPIC per unaligned nucleotide in the latter case), but nonetheless provide more PPIC in total, because of the greater length (unaligned) of sequence examined (Table 3.2). The latter estimates may be more prone to sampling error, given the small amount of variation encountered within the inverted regions in Poaceae. 78 3.3.2. Phylogenetic Analyses Maximum parsimony (MP) analysis of the large taxon set recovers six most-parsimonious trees. In the strict consensus tree, all nodes outside of Poaceae are resolved (branches collapsing in the strict consensus are noted in Fig. 3.1). Visual inspection of the phylogram indicates that members of the clade that includes Poaceae, Cyperaceae, Ecdeiocoleaceae, Mayacaceae, Restionaceae, and Xyridaceae generally have very long terminal branches compared with most other monocots, including other Poales, although there also appears to be considerable variation within this clade (cf. the relative lengths of the terminal branches ending in Cyperus and Ecdeiocolea to those ending in Flagellaria and Joinvillea; Fig. 3.1). Likelihood ratio tests with the GTR + T + I model of evolution with and without enforcement of a molecular clock confirm these observations. There is significant heterogeneity in the rate of evolution in the regions examined across the monocots, across the commelinid monocots, across Poales, and across Poaceae (P < 0.05; Table 3.3). MP analysis of the complete 28-taxon data set recovers two most-parsimonious trees (one is shown in Figure 3.2). Likelihood analysis of this data set recovers a tree that is identical in topology to the parsimony tree, with the exception of relationships among Ecdeiocoleaceae-Joinvilleaceae-Poaceae (Fig. 3.2). A tree inferred from Bayesian analyses of the complete, unpartitioned 28-taxon data set is shown in Figure 3.3. In general, clades that receive moderate to high bootstrap support in maximum parsimony and maximum likelihood analyses receive posterior probabilities of greater value in Bayesian analyses, in line with recent findings of empirical and simulation studies (e.g., Rannala and Yang 1996; Leache and Reeder 2002; Suzuki et al. 2002; Whittingham 2002; Wilcox et al. 2002; Cummings et al. 2003; Erixon et al. 2003; Simmons et al. 2004). Clades that receive low 79 bootstrap support in parsimony analyses are not usually recovered by likelihood and Bayesian analyses, and vice versa. Within Poaceae, there is general congruence among trees inferred from most data partitions in parsimony, likelihood, and Bayesian analyses (Fig. 3.4). Analysis of the noncoding data partition, however, provides instances where there is moderate to strong support for several clades not recovered or well supported by analyses of other data partitions (Fig. 3.4: clades "r," "t," "u," "v," and "x"), and which conflict with moderately to strongly supported clades inferred from those other partitions. Topologies and support values do not vary substantially between the various partitioned and corresponding unpartitioned Bayesian analyses, except in parts of the tree that are generally weakly supported by the complete data set (Fig. 3.4c). Relationships inferred along most of the backbone of Poaceae are supported robustly in all analyses, but support for relationships among several of the subfamilies varies (Figs. 3.2-3.4). In analyses involving all data sets, partitions and partitionings, and all phylogenetic methods, I recover strong support for the monophyly of Poaceae [maximum parsimony bootstrap proportion ( B P M P ) = 100%, maximum likelihood BP ( B P M L ) ; Bayesian posterior probability, P P = 100%; clade "a"] and moderate to strong support for the monophyly of Anomochlooideae (Anomochloa and Streptochaeta; clade "c": B P M P = 78-100%), B P M L = 90-100%, P P > 95%; Figs. 3.2-3.4. In parsimony, Bayesian and most likelihood analyes, I generally infer Anomochlooideae to be the sister group of the rest of the grasses (clade "b": B P M P = 89-100%, B P M L = 83-100%, P P = 100%; Figs. 3.2-3.4), and Pharoideae to be the next successive sister group of the rest of the family (clade "d": B P M P = 100%, P P = 100%; Figs. 3.2-3.4). Likelihood anlayses of codon position 3, however, place Anomochlooideae and Pharoideae in a weakly supported clade (clade "aa": B P M L = 69%; Fig. 3.4b) that is the 80 sister group of the rest of the grasses (i.e., the clade that includes the BEP and P A C M A D subclades). The latter clade is robustly supported (clade "d": B P M P = 100%, B P M L = 100%, PP = 100%; Figs. 3.2-3.4). The BEP (Bambusoideae, Ehrhartoideae, Pooideae) clade is supported to different degrees by different data partitions (clade "e": B P M P - 39-97%, B P M L = 32-94%, PP = 88-100%; Figs. 3.2-3.4); but the P A C M A D (Panicoideae, Arundinoideae, Chloridoideae, Aristidoideae, Danthonioideae) clade consistently has strong support (clade "h": B P M P = 99-100%, B P M L = 99-100%, PP = 100%; Figs. 3.2-3.4). As might be expected, where support for individual clades is weaker across partitions, it is generally from partitions that have the least number of informative characters (Table 3.2) Within the BEP clade, the two sampled taxa of Pooideae constitute a robustly supported clade in most analyses (clade "g", Fig. 3.4). When all data are examined, Bambusoideae are consistently inferred to be the sister group of Pooideae with moderate support (clade "f, Figs. 3.2, 3.3) and Ehrhartoideae are inferred to be the sister group of these two subfamilies (clade "e", Figs. 3.2, 3.3). Support for these relationships is again generally lower when data partitions with fewer PPIC per unaligned nucleotide are examined for parsimony and likelihood analyses, but not for most Bayesian analyses (for example, compare the different support values for clade "f" in Fig 3.4). Within the P A C M A D clade, the monophyly of Panicoideae receives weak to strong support with various data partitions (clade "k", Fig. 3.4). In most cases, phylogenetic relationships among subfamilies in the P A C M A D clade are not consistently resolved or supported. Internal branches are generally relatively short within this clade (Fig. 3.1). A clade consisting of Aristidoideae and Chloridoideae (clade "1", Fig. 3.4) is recovered with moderate to strong support in most 81 analyses, with less support from the smaller and less variable data partitions (differences in support values for this clade among partitions are less pronounced for Bayesian analyses). Several hypotheses of relationship from analyses of the noncoding data partition conflict with analyses of most other data partitions or subpartitions. These include a sister-group relationship between Triticum (Pooideae) and Ehrhartoideae (clade "v": BPMP = 71%, BPML = 74%, PP = 100%; Fig . 3.4), a Chloridoideae-Danthonioideae-Panicoideae clade (clade "t": BPMP = 55%, BPML = 65%, PP = 98%; Fig. 3.4), and an Aristidoideae-Arundinoideae clade (clade "u": BPMP = 57%, BPML = 69%, PP = 99%; Fig. 3.4). A l l analyses indicate that Ecdeiocoleaceae, Joinvilleaceae, and Poaceae are a strongly supported clade (clade "n": BPMP = 100%, BPML = 100%; PP = 100%; Figs. 3.2-3.4). A l l three possible arrangements among the three families are inferred in different analyses, often with low support, even considering all of the data combined. For example, in the monocot-wide parsimony analyses, Joinvilleaceae and Ecdeiocoleaceae are a weakly supported clade (BPMP = 43%, Fig . 3.1) that is the sister group of Poaceae, whereas in parsimony analyses restricted to commelinid taxa, Joinvilleaceae are weakly inferred to be the sister group of Poaceae (clade "m": BPMP = 31%; Fig. 3.2). In contrast, most likelihood and Bayesian analyses infer a sister-group relationship between Ecdeiocoleaceae and Joinvilleaceae, with weak to strong support (clade "o": BPML = 64-100%, PP = 53-100%; Fig . 3.4), and this two-family clade is robustly supported as the sister group of Poaceae (clade "n": Figs. 3.2-3.4). When the noncoding data are considered separately, Ecdeiocoleaceae are inferred to be the sister group of the grasses with moderate to robust support (clade "r": BPMP = 73%, BPML= 80%, PP = 86%, 96%>; Fig . 3.4). In general, all three arrangements among Poaceae and relatives (clades "r", "o" and "m") have equally poor support in most M P analyses. 82 However, with the exception of analyses involving the noncoding data partition (and one of the data partitionings used in the Bayesian analysis) clade "o" (Ecdeiocoleaceae-Joinvilleaceae) is the only arrangement with reasonably strong support in the model-based analyses. This support is strongest for analyses that focus on coding regions (Fig. 3.4). 3.4. DISCUSSION 3.4.1. Molecular Evolution and Phylogenetic Utility of the Plastid DNA Regions in Poaceae Visual inspection of branch variation in the phylogram here (Fig. 3.1) corroborates the findings of earlier studies on the accelerated substitution rate of grass plastid genomes compared with other monocots (e.g., Wilson et al. 1990; Bousquet et al. 1992; Gaut et al. 1992, 1996; Graham and Olmstead 2000a), and demonstrates that this increased substitution rate is not unique to Poaceae (Graham et al. 2006). The plastid rate elevation occurs in some, but not all, families within Poales [I have examined exemplar taxa here for only three fifths of Poales families, sensu APG II (2003), so this rate heterogeneity will require further characterization]. Comparison of likelihood models with and without a molecular clock enforced confirms that there is significant heterogeneity at various hierarchical levels among the monocots (Table 3.3). Most earlier studies did not characterize or comment on broader rate-heterogeneity within Poales, but published phylograms show generally increased branch lengths in the order compared with other monocots (e.g., Duvall et al. 1993a; Bremer 2002; Givnish et al. 2006), and grass long branches have been blamed on anomalous results in angiosperm-wide phylogenetic inference involving limited taxon sampling of whole plastid genomes (Stefanovic et al. 2004; Leebens-Mack et al. 2005). Various explanations have 83 been proposed for such rate heterogeneity, including a relaxation of purifying selection, demographic factors, and variable mutation rates among lineages (e.g., due to changes in DNA repair efficiency that may affect whole genomes or parts of them), or a combination of these factors (e.g., Cho et al. 2004; Parkinson et al. 2005; Young and dePamphilis 2005). Givnish et al. (1999) also observed this accelerated plastid substitution rate in Poales, and suggested that such genetic divergence may have arisen due to the limited seed dispersal abilities of many of the rate-accelerated lineages (i.e., wind or gravity dispersed), which might increase isolation of populations by distance, thereby promoting possible speciation and substantial genetic divergence. Two of the three closest relatives of Poaceae, Joinvilleaceae and Flagellariaceae (Figs. 3.1-3.3), are both fleshy-fruited, and both terminate branches here that appear to be less divergent than the predominantly non-animal dispersed Poaceae and Ecdeiocoleaceae (Fig. 3.1). The regions sampled represent most of the diversity of evolutionary rates in the plastid genome, with the exception of the most rapidly evolving plastid regions (e.g., Small et al. 1998; Shaw et al. 2005). The combined plastid data contain sufficient phylogenetically informative variation to resolve several relationships in Poaceae with strong support, including some that were problematic previously, but they do not provide enough characters to robustly resolve some relationships that have consistently proven recalcitrant. The genes atpB, ndhF, and rbcL have been used commonly for phylogenetic reconstruction at higher taxonomic levels (e.g., Savolainen et al. 2000; Givnish et al. 2006). In Poaceae, rbcL and ndhF alone yield insufficient variation to provide strong support for deeper nodes in the family (Clark et al. 1995; Duvall and Morton 1996). In my analyses, rbcL, ndhF, and atpE 84 combined contribute approximately half of the parsimony informative variation in the data set (Table 3.2). Inclusion of data from the more slowly evolving photosystem and inverted repeat regions of the plastid genome provides conservative characters. It has been suggested that they may be useful in reconstructing higher-level phylogenies in instances where long-branch attraction might otherwise complicate phylogenetic inference (e.g., Graham and Olmstead 2000a). These might also have, better-than-expected phylogenetic utility in "lower-level" (within-family) phylogenetic analysis in Poaceae and relatives, because of the relatively accelerated plastid substitution rate in many Poales (Fig. 3.1). The photosystem II regions contribute approximately a third of the useful variation in the data set here (Table 3.2). Some of these have been used in recent phylogenetic studies (e.g., Sanderson et al. 2000; He-Nygren and Sinikka 2003; Schiitze et al. 2003; Cameron and Molina 2006), but most have been largely underutilized and underexplored in studies below the family level. The plastid inverted repeat regions (IR; 3,-rpsl2-rps7-ndhB, and rpll) are by far the most conservative examined here, consistent with earlier studies (e.g., Wolfe et al. 1987; Goremykin et al. 1996; Graham and Olmstead 2000a; Graham et al. 2000; Masood et al. 2004). Most informative characters in the examined regions of the IR come from the noncoding portions, a function of the total amount of nucleotides examined for these portions (Table 3.2). 3.4.2. Phylogenetic Relationships in Poaceae There is general consensus on most aspects of branching order at the base of the Poaceae subtree, here and elsewhere. All analyses here involving larger data partitions corroborate most earlier molecular studies that include Anomochloa and Streptochaeta in a 85 clade (Anomochlooideae; Clark and Judziewicz 1996) that is the sister group of the rest of Poaceae. It should be noted that I have not yet sampled Puelioideae [a small subfamily variously resolved as the sister group of all of Poaceae except Anomochlooideae and Pharoideae (Clark et al. 2000; GPWG 2001), or as the sister group of the BEP clade (Zhang 2000)]. Some previous analyses indicate that Anomochlooideae are not a natural group (e.g., Hilu et al. 1999; Mathews et al. 2000; Zhang 2000), but my data strongly support its monophyly. GPWG (2001) suggested that Anomochloa and Streptochaeta might be spuriously attracted in phylogenetic analyses due to long-branch attraction. This hypothesis could be tested in future work using a simulation approach (e.g., Sanderson et al. 2000). In agreement with previous studies (e.g., Clark et al. 1995; Soreng and Davis 1995; Mathews et al. 2000; Zhang 2000; GPWG 2001), two branches ("b" and "d", Figs. 3.2-3.4) that collectively support a sister group relationship between Pharus (Pharoideae; Clark and Judziewicz 1996) and the B E P - P A C M A D clade, are strongly supported in all analyses here. One possible moderate conflict to this finding from my data is that likelihood analysis of the third codon position data identifies an Anomochlooideae-Pharoideae clade ("aa" in Fig. 3.4), but this clade has very poor support from all other partitions and analytical methods, and may represent a spurious conflict. The "deep" positions of Anomochlooideae and Pharoideae in Poaceae in this and previous molecular analyses are consistent with substantial morphological differences that characterize these taxa compared with most grasses (see GPWG 2001). The rest of the grasses form a very strongly supported clade (clade "d", the "core grasses", or BEP-PACMAD clade) in all of my analyses. This crown clade contains most grass diversity, and is subtended by a relatively long branch. Multiple putative 86 morphological synapomorphies have been identified for this major lineage (loss of pseudopetiole, reduction to two lodicules, loss of the inner whorl of stamens, and loss of arm cells and fusoid cells in the mesophyll), although several of these have undergone subsequent reversals in the clade (see GPWG 2001). Phylogenetic relationships inferred among the nine subfamilies in the core grasses have varied among previous studies. Several analyses unite Bambusoideae, Ehrhartoideae, and Pooideae in a clade (the BEP clade) with poor (e.g., Clark et al. 1995; Cummings et al. 1994; Mason-Gamer et al. 1998; Zhang 2000) or moderate support (e.g., Mathews et al. 2000; GPWG 2001); other studies have found alternate topologies, including a weakly-supported sister-group relationship between Pooideae and the P A C M A D clade (e.g., Duvall and Morton 1996; Soreng and Davis 1998; Hsiao et al. 1999; Caetano-Anolles 2005a; Duvall et al. 2006). No morphological synapomorphies have been identified for a BEP clade (GPWG 2001); six have been proposed for a Pooideae-PACMAD clade (Soreng and Davis 1998). My data recover the BEP clade with strong support in most analyses (clade "e", Figs. 3.2-3.4), an improvement for plastid data compared with the earlier multigene study [e.g., BPMP = 95% here for the combined data, vs. 62% for plastid data in GPWG 2001); in contrast, I find only poor support for a Pooideae-PACMAD clade; the highest support for this clade (clade "s") is very weak (BPMP = 52% here for codon positions 1 and 2 combined; Fig. 3.4). The remainder of the sampled grasses here form the P A C M A D clade (clade "h"), in agreement with all previous plastid analyses (e.g., Soreng and Davis 1998; Cummings et al. 1994; Clark et al. 1995; Duvall and Morton 1996; Zhang 2000), although I have not sampled the recently recognized subfamily Micrairoideae (Duvall et al. 2006). A substantial amount 87 of molecular evolution has clearly occurred in the plastid genome along the branch that subtends the P A C M A D crown clade (e.g., Fig. 3.1), which has undoubtedly contributed to the consistent recovery of this clade in previous analyses (even using limited amounts of data), and the consistently strong support found here across all analyses and data partitions (Fig. 3.4). However, only two putative morphological synapomorphies (presence of an elongated mesocotyl internode, and loss of the epiblast) have been identified for the P A C M A D grasses (GPWG 2001). All possible topologies have been inferred among the three subfamilies that form the BEP clade. I find moderate to strong support for a sister-group relationship between Bambusoideae and Pooideae for several of my analyses (clade "f', Figs. 3.2-3.4). This relationship was also recovered in the analyses of Mathews and Sharrock (1996) and Zhang (2000), but with substantially lower support than is found here. In the previous multigene studies (GPWG 2001; Duvall et al. 2006); a sister-group relationship between Bambusoideae and Ehrhartoideae was recovered. In most cases here, support for a Bambusoideae-Ehrhartoideae clade is < 50%, but I find this topology with moderate support in one analysis (clade "z": BPML = 72% for analysis of codon positions 1 and 2 combined, Fig. 3.4), representing a moderate conflict with the other data partitions and analyses. A stronger conflict is also evident, as analysis of noncoding data places Ehrhartoideae and Triticum as a clade with moderate to strong support (clade "v": BPMP = 71%; BPML = 74%; PP = 100%, Fig. 3.4). This arrangement was also weakly supported in an analysis of nuclear data from phytochrome B (Mathews et al. 2000). The cause of this incongruence is unknown. Despite sampling an expanded data set, clear inference of phylogenetic structure within the P A C M A D clade has not emerged from this study, although several aspects of 88 relationship are moderately to strongly supported. Chasmanthium and Zea are a well supported clade, consistent with the recent inclusion of the rather ill-defined Centothecoideae (GPWG 2001) in Panicoideae (e.g., Soreng et al. 2003; Duvall et al. 2006). In most instances, I infer a sister-group relationship between Aristidoideae and Chloridoideae, with moderate to robust support (clade "1": Figs. 3.2-3.4). This clade was inferred previously in an rbch analysis (Duvall and Morton 1996). It is not in conflict with most of those inferred for the P A C M A D clade in previous studies, as internal relationships in this clade generally have very weak support (GPWG 2001; Duvall et al. 2006). Analysis of the noncoding data partition here instead indicates a conflicting sister-group relationship between Arundinoideae and Aristidoideae, with moderate to strong support (clade "u"; Fig. 3.4), a topology that has not been seen in other studies (e.g., GPWG 2001). Bayesian analysis identifies a strongly supported clade that includes Arundinoideae, Aristidoideae, Chloridoideae, and Danthonioideae (clade "p", Figs. 3.3, 3.4). This clade was recovered in a previous rbcL study (Duvall and Morton 1996), but it does not receive bootstrap support higher than 50% in parsimony or likelihood analyses here (Fig. 3.4). Most of the moderate to strong conflict observed here among data partitions involves noncoding regions. These may reflect substantially different "process partitions" for these regions (e.g., Gielly and Taberlet 1994; Shaw et al. 2005), or may be a simple function of more poorly aligned regions. The regions that are most difficult to align in the current data set all belong to the single-copy noncoding regions (i.e., the intergenic spacer regions between the various photosystem II genes). It may therefore be reasonable to exclude these regions from consideration, and thus ascribe the conflicting clades (i.e., clades "v", "u", "t", etc.) to alignment difficulties or inflated Bayesian support values. If so, greater weight 89 should be placed on the relationships inferred using only protein-coding regions (and perhaps the very conservative noncoding regions in the inverted repeat, not reported on separately here) using MP and M L . Aside from these conflicts, several aspects of relationship, particularly with the P A C M A D clade, resist satisfactory resolution. These continuing difficulties may reflect the relatively short branches observed at deeper levels in the clade (Fig. 3.1), possibly reflecting a rapid initial radiation of each of the major lineages (GPWG 2001). These relationships are possibly difficult to resolve because too little information has been recovered to do so satisfactorily. Strong (high support values) and clear (congruent) inference of relationships in the P A C M A D clade may require substantially more data per taxon (from the plastid or other genomes). 3.4.3. The Sister Group of Poaceae Previous phylogenetic analyses of morphological and molecular data agree that Joinvilleaceae, Ecdeiocoleaceae, and Poaceae are a clade, but the identity of the sister group of the grasses has been variously resolved and remains unclear (e.g., Campbell and Kellogg 1987; Bremer 2000, 2002; Michelangeli et al. 2003; Davis et al. 2004). In my MP analyses, different taxon samplings yield different resolutions of relationships within this clade [(Poaceae, (Joinvilleaceae-Ecdeiocoleaceae)), Fig. 3.2; (Ecdeiocoleaceae, (Joinvilleaceae-Poaceae), Fig. 3.1], but this is not surprising, as branch support in the various parsimony analyses is consistently low and relatively evenly split among the different possibilities (clades "r", "o" and "m", Fig. 3.4a). In most likelihood and Bayesian analyses, however, Joinvilleaceae and Ecdeiocoleaceae are a clade (clade "o"; Figs. 3.2-3.4) with relatively strong support, particularly when protein-coding data are considered alone. By contrast, 90 when the noncoding data are considered independently (or in one case when they are considered as a separate partition in unlinked Bayesian analysis), Ecdeiocoleaceae are placed as the sister group of Poaceae, with moderate support in parsimony and likelihood analyses, and strong support in Bayesian analyses (clade "r"; Fig. 3.4). For the reasons discussed above, it may be reasonable to place more weight on the results from analyses that exclude some or all of the noncoding data (i.e., the finding of a sister-group relationship between Ecdeiocoleaceae and Joinvilleaceae). Indeed, for the noncoding data partition the terminal branch subtending Ecdeiocoleaceae is substantially longer than that for Joinvilleaceae and all of the grasses (data not shown), suggesting that there might be unrecognized alignment problems in the noncoding data for Ecdeiocolea (a subsequent review of the matrix could not identify substantial problems), or an unusual rate of molecular evolution for this data partition and taxon. The Ecdeiocoleaceae-Joinvilleaceae clade seen in most model-based analyses of the larger data partitions here (that exclude noncoding data) has the highest support levels obtained to date for putative relationships among these three families. Before this hypothesis can be accepted, however, it clearly requires corroboration by additional lines of molecular evidence (for example, additional taxa and substantially more data may be necessary to recover an Ecdeiocoleaceae-Joinvilleaceae clade in parsimony analyses, if it is correct). It would also be valuable to identify possible synapomorphies (e.g., morphological, developmental, anatomical, chemical, molecular) that might define an Ecdeiocoleaceae-Joinvilleaceae clade. 91 3.4.4. Congruence and Discordance in the Data Most relationships inferred here are either well supported (particularly among analyses performed on the larger data partitions; Fig. 3.4) and congruent among the various types of analyses performed, or are simply poorly supported across all analyses (e.g., several branches in the P A C M A D clade). Broad-scale congruence is one of the best sources of evidence that phylogenetic inferences have correctly predicted historical relationships (e.g., Penny et al. 1982; Graham et al. 1998). In the face of this reassuring picture, instances of incongruence may suggest underlying problems with the data (e.g., that there may be alignment problems) or with particular analytical methods (e.g., the tendency for Bayesian analyses to produce inflated support values compared to bootstrap analysis, a result that seems to be generally supported here: Fig. 3.4). Because all the data considered here come from the same genetic linkage group (the plastid genome) it would be desirable to collect data from other linkage groups to detect sources of incongruence between gene and organismal phylogenies, such as lineage sorting of ancestral polymorphisms and ancient hybridization events (e.g., Maddison 1997; Maddison and Knowles 2006). However, the latter effects might be expected to have little consequence for most of the inferences made here, because of the long time periods separating most of the evolutionary splits examined. In the current study I focussed on adding more data per exemplar taxon for making inferences of relationship with Poaceae and relatives, as this is expected to be an effective strategy for performing accurate phylogenetic inference (e.g., Swofford and Poe 1999; see Graham et al. 2006). Studies have also demonstrated that accurate inference of phylogenetic relationships can be enhanced by increasing the density of taxon sampling (e.g., Hillis 1998; Pollock et al. 2002; Zwickl and Hillis 2002). To estimate relationships among the major 92 clades of Poaceae for the current (or improved) genomic samplings, improving the taxon density may therefore also prove effective. One simple improvement would be to sample at least a few taxa per subfamily, chosen to straddle the root node of each of these major clades. However, with few exceptions [e.g., Puelioideae (see Chapter 4), and Micrairoideae within the P A C M A D clade (Duvall et al. 2006)], it is doubtful that further additions can be made along the deepest branches of Poaceae phylogeny, or the branch between Poaceae and its closest relatives (Ecdeiocoleaceae and Joinvilleaceae). 93 Table 3.1. GenBank accession numbers and vouchers for sequence data from exemplar species in Poaceae and Joinvilleaceae. The rbcL sequence for Pharus (underlined) was obtained from GenBank; this sequence is from a closely related species. Data for three grasses (Ehrhartoideae: Oryza sativa L.; Pooideae: Triticum aestivum L.; and Panicoideae: Zea mays L. are from whole plastid genomes deposited on GenBank (accessions NC_001320, NC 002762 nad NC_001666.2, respectively). Taxon Voucher Gene/ Region Length (unaligned) atpB ndhF psbB-T- N-H psbD-C psbE-F-L-J rbcL 3'rps\2, rpll rpsl, and ndhH All Protein-coding nucleotides nucleotides POACEAE Anomochlooideae Anomochloa marantoidea Brongn. Clark 1299 (ISC) 15.6 kb 12.2 kb Streptochaeta angustifolia Soderstrom Clark 1304 (ISC) 15.6 kb 12.2 kb Pharoideae Pharus lappulaceus Aubl. Clark 1329 (ISC) AY357724 15.4 kb 12.1 kb Bambusoideae Pseudosasa japonica (Siebold & Zucc. Saarela 265 (ALTA) 15.7 kb 12.3 kb ex Steud.) Makino ex Nakai Pooideae Brachyelytrum aristosum (Michx.) Saarela 50 (ALTA) 15.4 kb 12.0 kb P. Beauv. ex Trel. Aristidoideae 15.4 kb 12.1 kb Aristida adscensionis L. Peterson et al. 16679 (US) 4^ Table 3.1 continued Taxon Voucher Gene/ Region Length (unaligned) 3'rpsl2, atpB ndhF psbB-T- N-H psbD-C psbE-F-L-J rbcL rpl2 rpsl, and ndhB All Protein-coding nucleotides nucleotides Arundinoideae 15.6 kb 12.2 kb Phragmites australis (Cav.) Trin. ex Steud. Saarela 264 & Ferreira (ALTA) Danthonioideae Danthonia californica Bol. Saarela 267, Greif, & 15.4 kb 12.0 kb Hernandez-Castillo (ALTA) Panicoideae Chasmanthium latifolium (Michx.) H. 0. Peterson 17532 (US) 15.4 kb 12.1 kb H. 0. Yates Chloridoideae Bouteloua curtipendula (Michx.) Torr. Saarela 20 (ALTA) 15.2 kb 11.8 kb JOINVILLEACEAE Joinvillea plicata (Hook, f.) Newell & Stone Thien 84 (NO) 14.8 kb 12.0 kb 3Partly sequenced genes by taxon, relative to sequences considered in Graham and Olmstead (2000a): (1) rbcL: Brachyelytrum, 195 bp missing at 3'-end; (2)atpB: Chasmanthium and Aristida, 193 bp missing at 5'-end, Bouteloua, missing 194 bp at 5'-end and 311 bp at 3'-enA,Danthonia, missing 255 bp at 5'-end, Phragmites, missing 75 bp at 5'-end, Anomochloa, missing 57 bp at 5'-end, Streptochaeta, missing 52 bp at 5'-end, and Brachyelytrum, missing 71 bp at the 3'-end; (3) ndhF: Pharus, missing 176 bp at 3'-end; (4)psAE-F-L-J: Aristida, missing 58 bp at 3'-end. Table 3.2. Number of characters (aligned and unaligned), potentially parsimony informative characters (PPIC), percentage of total PPIC, and number of PPIC per unaligned nucleotide in various partitions of the data. Number of characters Number of characters % of total Number of PPIC per unaligned nucleotide Data partition (aligned) (unaligned)a PPIC PPIC examined Combined IRb Commelinids 11427 -5.7 kb 366 15.2 0.064 Poaceae 11427 -5.7 kb 61 8 0.011 atpB, ndhF, andrbcL Commelinids 5375 -5.0 kb 1193 49.6 0.239 Poaceae 5375 -5.0 kb 414 54.5 0.083 Combined photosytem II genes0 Commelinids 11260 -5.0 kb 848 35.2 0.17 Poaceae 11260 -5.0 kb 285 37.5 0.057 All data Commelinids 28062 2407d Poaceae 28062 760d IR codon positions 1 and 2 Commelinids 2456 -1.9 kb 99 4.1 0.052 Poaceae 2456 -1.9 kb 10 1.3 0.005 IR codon position 3 Commelinids 1230 -0.9 kb 87 3.6 0.097 Poaceae 1230 -0.9 kb 12 1.6 0.013 IR noncoding Commelinids 7741 -3.0 kb 180 7.5 0.06 Poaceae 7741 -3.0 kb 39 5.1 0.013 SC codon positions 1 and 2 Commelinids 6714 -6.4 kb 567 23.5 0.089 Poaceae 6714 -6.4 kb 177 23.3 0.028 SC codon position 3 Commelinids 3386 -3.2 kb 1331 55.3 0.415 Poaceae 3386 -3.2 kb 483 63.4 0.015 SC noncoding Commelinids 6588 -0.5 kb 144 6 0.288 Poaceae 6588 -0.5 kb 40 5.3 0.08 All data Commelinids 28062 2408d 100 Poaceae 28062 761d 100 a Unaligned lengths were determined for a reference taxon, Oryza sativa L . , for which complete data are available. b Including genes, introns, and intergenic spacers. 0 Including genes and intergenic spacers. d Total PPIC vary between different combined analyses because the 53 bp overlap between psbD andpsbC is considered in analyses of both SC codon positions 1 and 2 and SC codon position 3. This overlap was not considered twice when determining total number of characters. 96 Table 3.3. Likelihood ratio test (LRT) assessing heterogeneity among hierarchical lineages of monocots using the G T R + T - M a model of evolution with and without enforcement of a molecular clock. Likelihood scores were estimated based on a randomly chosen most-parsimonious tree. - ln likelihood Taxon set Number of taxa (df)c G T R + r + I G T R + r + I + molecular clock - 2 1 n Ad P Monocots 79 (77) 158371.77 162234.24 7724.94 <0.05 Commelinid monocotsb 28 (25) 76899.36 78885.92 3973.12 <0.05 Poales 23 (21) 67608.54 69115.64 3014.2 <0.05 Poaceae 13(11) 39216.3 39541.78 650.96 <0.05 "Abbreviations: G T R , general time-reversible; T, Gamma; I, proportion of invariable sites. b Excluding Dasypogonaceae and Arecaceae. c df = number of taxa - 2. d Likelihood ratio test statistic. 97 Figure 3.1. Phylogram of one of six most parsimonius trees inferred for Poaceae and other monocots, based on a large plastid data set (atpB, ndhB, ndhF, ten photosystem genes, rbcL, rps7, V-rpsYl, ndhB and various introns and noncoding regions; see text). Tree length = 26,761 steps, consistency index = 0.370, retention index = 0.553. The clades marked with an arrow within Poaceae collapse in the strict consensus tree. 98 Fig. 3.1 Chasmanthium ' Phragmites Aristida ; Bouteloua Dantrionia , ; , Jriticum Bra'chyelytrum. "q Pseudoi O r y z a ' Cyperus Xyris Philydrum Ensete Strelitzia Roystonea Chlorophytum Yucca Muilla Muscari Asparagus Lomandra 7 Allium | Dasypogonaceae | Commelinales j Zingiberales • Arecales Asparagales Poales 100 changes Sagittaria Scheuchzeria Tofieldia Spathiphyllum Acorus Liliales Pandanales Dioscoreales Petrosaviales Alismatales I Acorales 99 Figure 3.2. One of two most parsimonious trees inferred from combined data from 17 chloroplast genes for Poaceae and related families in the commelinid monocots. Subfamilies within Poaceae, the BEP and P A C M A D clades, and names of closely related families are indicated. Tree length = 10,318 steps, consistency index = 0.585, retention index = 0.611. Bootstrap values from parsimony and likelihood analyses are indicated above the branches (parsimony to the left, likelihood to the right). Clades marked with arrows collapse in the strict consensus tree. An Ecdeiocoleaceae-Joinvilleaceae clade inferred in the M L analysis (clade "o"; BP = 64% in likelihood analysis, 38% in parsimony analysis) is not shown here. In the likelihood tree, -ln = -77085.77. Letters "a" to "n" identify branches scored for analyses from partitions of the data. 100 Fig. 3.2 Poaceae ^ 93,99, ^ < 3 1 , " Poales 100,100 100,100" Zingiberales • Zea 1 •(Chasmanthium • Phragmites - -j Anstida Bouteloua ' Danthonia friticum; ' Brachyelytrum'':*: Pseudosasa-:'>] Oryza Pharus Anomochloa Streptochaeta,^ Joinvillea Ecdeiocolea Elegia Flagellaria Cyperus Xyris Mayaca Ananas Typha Sparganium Hydrothrix Xiphidium Philydrum Ensete Strelitzia Panicoideae I Arundinoideae I Aristidoideae I Chloridoideae I Oanthonioideae Pooideae I Bambusoideae I Ehrhartoideae I Pharoideae Anomochlooideae I Joinvilleaceae I Ecdeiocoleaceae I Restionaceae I Flagellariaceae I Cyperaceae I Xyridaceae I Mayacaceae I Bromeliaceae Typhaceae Pontederiaceae Haemodoraceae Philydraceae Zingiberaceae Strelitziaceae TJ > O > D 00 m "0 101 Figure 3.3. Phylogenetic tree obtained from Bayesian analysis of combined data from 17 chloroplast genes for Poaceae and related families in the commelinid monocots. The analysis was conducted with the GTR + G + I model of evolution, with all data considered and an unpartitioned analysis. Numbers above branches are posterior probabilities. Subfamilies within Poaceae, the BEP and P A C M A D clades, and names of closely related families are indicated. Letters ("a" to "q") identify branches scored for analyses from partitions of the data. 102 Fig. 3.3 100 Poales 59 ^ 100 {Chasmanthium Aristida Bouteloua ... banthonia .,; Phragmites Triticum •Brachyelytrum.;: Pseudosasa Oryza Phar'us ' Anomochloa Panicoideae Chloridoideae Aristidoideae Danthonioideae Arundinoideae Pooideae Bambusoideae Ehrhartoideae Pharoideae Anomochlooideae TJ > O > CD m Commelinales ioo r~~ 100 Zingiberales Streptochaeta Joinvillea I Joinvilleaceae Ecdeiocolea I Ecdeiocoleaceae Flagellaria I Flagellariaceae Elegia I Restionaceae Xyris I Xyridaceae Cyperus I Cyperaceae Mayaca Mayacaceae Typha Sparganium Typhaceae Ananas I Bromeliaceae Hydrothrix • Pontederiaceae Xiphidium 1 Haemodoraceae Philydrum i Philydraceae Ensete • Zingiberaceae Strelitzia • Strelitziaceae 103 Figure 3.4. Ranked profiles of clade support from analyses of various data partitions, as measured by (A) bootstrap proportions from parsimony analyses; (B) bootstrap proportions from likelihood analyses; and (C) posterior probabilities from Bayesian analyses. Letters "a" to "q" correspond to clades in Figs. 3.2, 3.3; "r" = (Ecdeiocoleaceae-Poaceae); "s" = (PACMAD-Pooideae); "t" = (Chloridoideae-Danthonioideae-Panicoideae); "u" = (Aristidoideae- Arundinoideae); "v" = (rr/rtcww-Ehrhartoideae); "w" = Aristidoideae-Arundinoideae -Chloridoideae-Zea); "x" = (Chloridoideae-Danthoniodeae); "y" = (Aristidoideae- Chloridoideae-Danthoniodeae-Zea); "z" = (Bambusoideae-Ehrhartoideae); and "aa" = (Anomochlooideae-Pharoideae)]. Data are from 17 chloroplast genes from 28 taxa representing Poaceae and several closely related commelinid monocot families. Data partitions examined are indicated in each figure, f indicates analyses where substitution model parameters were estimated independently for each partition (partitions separated by semicolons). Clades with bootstrap proportions or posterior probabilities <5% are not indicated here. 104 Fig. 3.4A Parsimony a c d h n b g k e l f I r o m j q p z x s t u v w y a a Clade Fig. 3.4B Likelihood b a c d h n k e g f I o p q r x I j z t v s m a a u w y ^ Clade o as Fig. 3.4C Bayesian a b c d h l e g f k n p o q j r y w z s l m t u v x a a - Clade o - J 3.5 R E F E R E N C E S Aliscioni, S. S., Giussani, L. M . , Zuloaga, F. O., and Kellogg, E. A. 2003. A molecular phylogeny of Panicum (Poaceae: Paniceae): tests of monophyly and phylogenetic placement within the Panicoideae. American Journal of Botany 90, 796-821. Angiosperm Phylogeny Group (APG II). 2003. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants. Botanical Journal of the Linnean Society 141, 399-436. Avdulov, N. P. 1931. Kario-sistematicheskoye issledovaniye semeystva zlakov. Supplement 44, Bulletin of Applied Botany, Genetics and Plant Breeding. Leningrad. Barker, N. P., Linder, H. P., and Harley, E. H. 1995. Polyphyly of Arundinoideae (Poaceae): Evidence from rbcL sequence data. Systematic Botany 20, 423-435. Barker, N. P., Morton, C. M . , and Linder, H. P. 2000. The Danthonieae: generic composition and relationships. Pp. 221-229. In: Jacobs, S. W. L. , and Everett, J. [Eds.]. Grasses: Systematics and Evolution. CSIRO, Melbourne. Beetle, A. A. 1955. The four subfamilies of the Gramineae. Bulletin of the Torrey Botanical Club 82, 196-197. Bousquet, J., Strauss, S. H., Doerksen, S. H., and Price, R. A. 1992. Extensive variation in evolutionary rate of rbcL gene sequences among seed plants. Proceedings of the National Academy of Sciences U.S.A. 89, 7844-7848. Bremer, K. 2000. Early Cretaceous lineages of monocot flowering plants. Proceedings of the National Academy of Sciences U.S.A. 97, 4707-4711. Bremer, K. 2002. Gondwanan evolution of the grass alliance of families (Poales). Evolution 56, 1374-1387. 108 Briggs, B. G., and Johnson, L. A. S. 1998. Georgeantha hexandra, a new genus and species of Ecdeiocoleaceae (Poales) from Western Australia. Telopea 7, 307-312. Briggs, B. G., Marchant, A. D., Gilmore, S., and Porter, C. L. 2000. A molecular phylogeny of Restionaceae and allies. Pp. 661-671 In: Wilson, K. L. , and Morrisson, D. A. [Eds.], Monocots: Systematics and Evolution, vol. II. CSIRO, Melbourne, Australia. Brown, R. 1810. Prodromus florae Novae Hollandiae et Insulae Van-Diemen. Vol. 1. J. Johnson, London, UK. Brown, W. V. 1958. Leaf anatomy in grass systematics. Botanical Gazette 119, 170-178. Brown, W. V., and Emery, W. H. P. 1957. Persistent nucleoli and grass systematics. American Journal of Botany 44, 585-590. Caetano-Anolles, G. 2005a. Grass evolution inferred from chromosomal rearrangements and geometrical and statistical features in RNA structure. Journal of Molecular Evolution 60, 635-652. Caetano-Anolles, G. 2005b. Evolution of genome size in the grasses. Crop Science 45, 1809-1816. Cameron, K. M . , and Molina, M . C. 2006. Photosytem II gene sequences ofpsbB andpsbC clarify the phylogenetic position of Vanilla (Vanilloideae, Orchidaceae). Cladistics 22, 239-248. Campbell, C. S. 1985. The subfamilies and tribes of Gramineae (Poaceae) in the southeastern United States. Journal of the Arnold Arboretum 66, 123-199. Campbell, C. S., and Kellogg, E. A. 1997. Sister group relationships of the Poaceae. Pp. 217-224 In: Soderstron, T. R., Hilu, K. W., Campbell, C. S., and Barkworth, M . E. 109 [Eds.]. Grass Systematics and Evolution. Smithsonian Institution Press, Washington, D. C. Catalan, P., Kellogg, E. A., and Olmstead, R. G. 1997. Phylogeny of subfamily Pooideae based on chloroplast ndhF gene sequences. Molecular Phylogenetics and Evolution 8, 150-166. Chase, M . W. 2004. Monocot relationships: an overview. American Journal of Botany 91, 1645-1655. Chase, M . W., Soltis, D. E. , Soltis, P. S., Rudall, P. J., Fay, M . F., Hahn, W. H., Sullivan, S., Joseph, J., Molvray, M . , Kores, P. J., Givnish, T. J., Sytsma, K. J., and Pires, J. C. 2000. Higher-level systematics of the monocotyledons: an assessment of current knowledge and a new classification. Pp. 3-16 In: Wilson, K. L., and Morrison, D. A. [Eds.]. Monocots: Systematics and Evolution. CSIRO Publishing, Collingwood, Victoria, Australia. Chase, M . W., Fay, M . F., Devey, D., R0nsted, N., Davies, J., Pillon, Y., Petersen, G., Seberg, O., Tamura, M . N., Asmussen, C. B., Hilu, K., Borsch, T., Davis, J. I, Stevenson, D. W., Pires, J. C , Givnish, T. J., Sytsma, K. J., McPherson, M . A., Graham, S. W., and Rai, H. S. 2006. Multigene analyses of monocot relationships: a summary. In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , & Simpson, M . G. [Eds.], Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanical Garden, Claremont, California. In Press. Cho, Y., Mower, M . P., Qiu, Y. -L. , and Palmer, J. D. 2004. Mitochondrial substitution rates are extraordinarily elevated and variable in a genus of flowering plants. Proceedings of the National Academy of Sciences U.S.A. 101, 1774-17746. 110 Clark, L. G., Zhang, W., and Wendel, J. F. 1995. A phylogeny of the grass family (Poaceae) based on ndhF sequence data. Systematic Botany 20, 436-460. Clark, L. G;, and Judziewicz, E. J. 1996. The grass subfamilies Anomochlooideae and Pharoideae (Poaceae). Taxon 45, 641-645. Clark, L. G., Kobayashi, M . , Mathews, S., Spangler, R. E. , and Kellogg, E. A. 2000. The Puelioideae, a new subfamily of Poaceae. Systematic Botany 2000, 181-187. Clayton, W. D., and Renvoize, S. A. 1986. Genera Graminum: Grasses of the World. Londong, Her Majesty's Stationary Office. Columbus, J. T., Kinney, M . S., Delagado, M . E. S., and Porter, J. M . 2000. Phylogenetics of Bouteloua and relatives (Gramineae: Chloridoideae): Cladistic parsimony analysis of internal transcribed spacer (nrDNA) and trnL-F (cpDNA) sequences. Pp. 189-194 In: Jacobs, S. W. L., and Everett, J. [Eds.]. Grasses: Systematics and Evolution. CSIRO, Melbourne, Australia. Crepet, W. L. , and Feldman, G. D. 1991. The earliest remains of grasses in the fossil record. American Journal of Botany 78, 1010-1014. Cummings, M . P., King, L. M . , and Kellogg, E. A. 1994. Slipped strand mispairing in a plastid gene: rpoC2 in grasses (Poaceae). Molecular Biology and Evolution 11, 1-8. Cummings, M . P., Handley, S. A., Myers, D. S., Reed, D. L. , Rokas, A., and Winka, K. 2003. Comparing bootstrap and posterior probability values in the four-taxon case. Systematic Biology 52, 477-487. Cutler, D. F., and Shaw, H. K. A. 1965. Anarthriaceae and Ecdeiocoleaceae: two new monocotyledonous familes, separated from the Restionaceae. Kew Bulletin 19, 489-555. I l l Davis, J. I. 1995. A phylogenetic structure for the monocotyledons, as inferred from chloroplast DNA restriction site variation, and a comparison of measures of clade support. Systematic Botany 20, 503-527. Davis, J. I, and Soreng, R. J. 1993. Phylogenetic structure in the grass family (Poaceae) as inferred from chloroplast DNA restriction site variation. American Journal of Botany 80, 1444-1454. Davis, J. I, Stevenson, D. W., Petersen, G., Seberg, O., Campbell, L. M . , Freudenstein, J. V., Goldman, D. H., Hardy, C. R., Michelangeli, F. A., Simmons, M . P., Specht, C. D., Vergara-Silva, F., and Gandolof, M . 2004. A phylogeny of the monocots, as inferred from rbcL and atpA sequence variation, and a comparison of methods for calculating jackknife and bootstrap values. Systematic Botany 29, 467-510. Doyle, J. J., and Doyle, J. L. 1987. A rapid isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19, 11-15. Doyle, J. J., Davis, J. I, Soreng, R. J., Garvin, D., and Anderson, M . J. 1992. Chloroplast inversions and the origin of the grass family (Poaceae). Proceedings of the National Academy of Sciences U.S.A. 89, 7722-7726. Duvall, M , Clegg, M . , Chase, M. , Clark, W. D., Kress, W. J., Hills, H., Eguiarte, L., Smith, J., Gaut, B., Zimmer, E. , and Learn, G. Jr. 1993a. Phylogenetic hypotheses for the monocotyledons constructed from rbcL sequence data. Annals of the Missouri Botanical Garden 80, 607-619. Duvall, M . R., Learn, G. H. Jr., Eguiarte, L. E. , and Clegg, M . T. 1993b. Phylogenetic analysis of rbcL sequences identifies Acorus calamus as the primal extant 112 monocotyledon. Proceedings of the National Academy of Sciences U.S.A. 90, 4641-4644. Duvall, M . R., and Morton, B. R. 1996. Molecular phylogenetics of Poaceae: an expanded analysis of rbcL sequence data. Molecular Phylogenetics and Evolution 5, 352-358. Duvall, M . R., Davis, J. I., Clark, L. G., Noll, J. D., Goldman, D. H., and Sanchez-Ken, J. G. 2006. Phylogeny of the grasses (Poaceae) revisited. In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Erixon, P., Svennblad, B., Britton, T., and Oxelman, B. 2003. Reliability of Bayesian posterior probabilities and bootstrap frequencies in phylogenetics. Systematic Biology 52, 665-673. Felsenstein, J. 1985. Confidence limits on phylogeny: an approach using the bootstrap. Evolution 39, 783-791. Gaut, B. S., Muse, S. V., Clark, W. D., and Clegg. M . T. 1992. Relative rates of nucleotide substitution at the rbcL locus of monocotyledonous plants. Journal of Molecular Evolution 35, 292-303. Gaut, B. S., Morton, B. R., McCaig, B. C , and Clegg, M . T. 1996. Substitution rate comparisons between grasses and palms: Synonymous rate differences at the nuclear gene Adh parallel rate differences at the plastid gene rbcL. Proceedings of the National Academy of Sciences U. S. A.93, 10274-10279. 113 Gaut, B. S., Clark, L. G., Wendel, J. F., and Muse, S. V. 1997. Comparisons of the molecular evolutionary process at rbcL and ndhF in the grass family (Poaceae). Molecular Biology and Evolution 14, 769-777. Gaut, B. S. 2000. Evolutionary dynamics of grass genomes. New Phytologist 154, 15-28. Ge, S., Li , A., Lu, A., Zhang, S.-Z., and Hong, D.-Y. 2002. A phylogeny of the rice tribe Oryzeae (Poaceae) based on matK sequence data. American Journal of Botany 89, 1967-1972. Gielly, L. P, and Taberlet, P. L. 1994. The use of chloroplast DNA to resolve plant phylogenies: noncoding versus rbcL sequences. Molecular Biology and Evolution 11, 769-777. Giussani, L. M . , Cota-Sanchez, J. H., Zuloaga, F. O., and Kellogg, E. A. 2001. A molecular phylogeny of the grass subfamily Panicoideae (Poaceae) shows multiple origins of C4 photosynthesis. American Journal of Botany 88, 1993-2012. Givnish, T. J., Evans, T. M . , Pire, J. C , and Sytsma, K. J. 1999. Polyphyly and convergent morphological evolution in Commelinales and Commelinidae: evidence from rbcL sequence data. Molecular Phylogenetics and Evolution 12, 360-385. Givnish, T. J., Pires, J. C , Graham, S. W., McPherson, M . A., Prince, L. M . , Patterson, T. B., Rai, H. S., Roalson, E. H., Evans, T. E. , Hahn, W. J., Millam, K. C , Meerow, A. W., Molvray, M . , Kores, P. J., O'Brien, H. E. , Hall, J. C , Kress, W. J., and Sytsma, K. J. 2006. Phylogenetic relationships of monocots based on the highly informative plastid gene ndhF: evidence for widespread concerted convergence. Pp. 27-50 In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . 114 G. [Eds.]. Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Gomez-Martinez, R., and Culham, A. 2000. Phylogeny of the subfamily Panicoideae with emphasis on the tribe Paniceae: evidence from the trnL-V cpDNA region. Pp. 136-140 In: Jacobs, S. W. L., and Everett, J. [Eds.]. Grasses: Systematics and Evolution. CSIRO, Melbourne. Goremykin, V., Bobrova, V., Pahnke, J., Troitsky, A., Antonov, A., and Martin, W. 1996. Noncoding sequences from the slowly evolving chloroplast inverted repeat in addition to rbcL data do not support gnetalean affinities of angiosperms. Molecular Biology and Evolution 13, 383-396. Graham, S. W., Kohn, J. R., Morton, B. R., Eckenwalder, J. E. , and Barrett, S. C. H. 1998. Phylogenetic congruence and discordance among one morphological and three molecular data sets from Pontederiaceae. Systematic Biology 47, 545-567. Graham, S. W., and Olmstead, R. G. 2000a. Utility of 17 chloroplast genes for inferring the phylogeny of the basal angiosperms. American Journal of Botany 87, 1712-1730. Graham, S. W., and Olmstead, R. G. 2000b. Evolutionary significance of an unusual chloroplast DNA inversion found in two basal angiosperm lineages. Current Genetics 37, 183-188. Graham, S. W., Reeves, P. A., Burns, A. C. E. , and Olmstead, R. G. 2000. Microstructural changes in noncoding chloroplast DNA: interpretations, evolution, and utility of indels and inversion in basal angiosperm phylogenetic inference. International Journal of Plant Sciences 161 (6 Suppl), S83-S96. 115 Graham, S. W., Zgurski, J. M . , McPherson, M . A., Cherniawsky, D. M . , Saarela, J. M . , Home, E. S. C., Smith, S. Y., Wong, W. A., O'Brien, H. E. , Biron, V. L. , Pires, J. C , Olmstead, R. G., Chase, M . W., and Rai, H. S. 2006. Robust inference of monocot deep phylogeny using an expanded multigene plastid data set. Pp. 3-20 In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M., and Simpson, M . G. [Eds.]. Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Grass Phylogeny Working Group [GPWG]. 2000. A phylogeny of the grass family (Poaceae), as inferred from eight character sets. Pp. 3-7 In: Jacobs, S. W. L. , and Everett, J. [Eds.]. Grasses: Systematics and Evolution. CSIRO, Melbourne. Grass Phylogeny Working Group [GPWG]. 2001. Phylogeny and subfamilial classification of the grasses (Poaceae). Annals of the Missouri Botanical Garden 88, 373-457. Guindon, S., and Gascuel, O. 2003. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Systematic Biology 52, 696-704. Hamby, R. K., and Zimmer, E. A. 1988. Ribosomal RNA sequences for inferring phylogeny within the grass family (Poaceae). Plant Systematics and Evolution 160, 29-37. He-Nygren, X, and Sinikka, P. 2003. Phylogenetic relationships of the generic complex Chiloscyphus—Lophoclea—Heteroscyphus (Geocalycaceae, Hepaticae): insights from three chloroplast genes and morphology. Annals Botanica Fennica 40, 317-329. Hillis, D. 1998. Taxonomic sampling, phylogenetic accuracy, and investigator bias. Systematic Biology 42, 182-192. Hilu, K. W., and Wright, K. 1982. Systematics of Gramineae: a cluster analysis. Taxon 31, 9-36. 116 Hilu, K. W., Alice, L. A., and Liang, H. 1999. Phylogeny of Poaceae inferred from matK sequences. Annals of the Missouri Botanical Garden 86, 835-851. Hilu, K. W., and Alice, L. A. 1999. Evolutionary implications of matK indels in Poaceae American Journal of Botany 86, 1735-1741. Hilu, K. W., and Alice, L. A. 2000. Phylogenetic relationships in subfamily Chloridoideae (Poaceae) based on matK sequences: a preliminary assessment. Pp. 173-179 In: Jacobs, S. W. L., and Everett, J. [Eds.]. Grasses: Systematics and Evolution. CSIRO, Melbourne. Hilu, K. W., and Alice, L. A. 2001. A phylogeny of Chloridoideae (Poaceae) based on matK sequences. Systematic Botany 26, 386-405. Hilu, K. W., Borsch, T., Muller, K., Soltis, D. E. , Soltis, P. S., Savolainen, V., Chase, M.W., Powell, M . P., Alice, L. A., Evans, R., Sauquet, H., Neinhuis, C , Slotta, T. A. B., Rohwer, J. G., Campbell, C. S., and Chatrou, L. W. 2003. Angiosperm phylogeny based on matK sequence information. American Journal of Botany 90, 1758-1776. Hiratsuka, J., Shimada, H., Whittier, R., et al. (16 co-authors). 1989. The complete sequence of the rice (Oryza sativa) chloroplast genome: intermolecular recombination between distinct tRNA genes accounts for a major plastid DNA inversiion during the evolution of the cereals. Molecular Genomics and Genetics 262, 462-467. Hitchcock, A. S. 1951. Manual of the Grasses of the United States (ed. 2, revised by A. Chase). Washington: U. S. Department of Agriculture. Hsiao, C , Chatterton, N. J., Asay, K. H., and Jensen, K. B. 1994. Phylogenetic relationships of 10 grass species: an assessment of phylogenetic utility of the internal transcribed spacer region in nuclear ribosomal DNA in monocots. Genome 37, 112-120. 117 Hsiao, C , Jacobs, S. W. L. , Chatterton, N. J., and Asay, K. H. 1999. A molecular phylogeny of the grass family (Poaceae) based on sequences of nuclear ribosomal DNA (ITS). Australian Systematic Botany 11, 667-688. Huelsenbeck, J. P., and Ronquist, F. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572-1574. Katayama, H., and Ogihara, Y. 1996. Phylogenetic affinities of the grasses to other monocots as revealed by molecular analysis of chloroplast DNA. Current Genetics 29,572-581. Kellogg, E. A., and Campbell, C. S. 1987. Phylogenetic analyses of the Gramineae. Pp. 310-322 In: Soderstrom, T. R., Hilu, K. W., Campbell, C. S., and Barkworth, M . E. [Eds.]. Grass Systematics and Evolution. Smithsonian Institution, Washington, DC, USA. Kellogg, E. A. 1998. Relationships of cereal crops and other grasses. Proceedings of the National Academy of Sciences U.S.A 95, 2005-2010. Kellogg, E. A., and Watson, L. 1993. Phylogenetic studies of a large data set. I. Bambusoideae, Andropogonodae, and Pooideae. Botanical Review 59, 273-343. Leache, A. D., and Reeder, T. W. 2002. Molecular systematics of the eastern fence lizard (Sceloporus undulatus): a comparison of parsimony, likelihood, and Bayesian approaches/Systematic Biology 51, 44-68. Leebens-Mack, J., Raubeson, L. A., Cui, L. , Kuehl, J. V., Fourcade, M . H., Chumley, T. W., Boore, J. L. , Jansen, R. K., and dePamphilis, C. W. 2005. Identifying the basal angiosperm node in chloroplast genome phylogenies: sampling one's way out of the Felsenstein zone. Molecular Biology and Evolution 22, 1948-1963. 118 Liang, H., and Hilu, K. W. 1996. Application of the matK gene sequences to grass systematics. Canadian Journal of Botany 74, 125-134. Linder, H. P., Briggs, B. G., and Johnson, L. A. S. 2000. Restionaceae: a morphological phylogeny. Pp. 653-660 In: Wilson, K. L., and Morrison, D. A. [Eds.]. Monocots: Systematics and Evolution, vol. II. CSIRO, Melbourne, Australia. Maier, R. M . , Neckermann, K., Iglioli, G. L. , and Kossel, H. 1995. Complete sequence of the maize chloroplast genome: gene content, hotspots of divergence and fine tuning of genetic information for transcript editing. Journal of Molecular Biology 251, 614-628. Maddison, W. P. 1997. Gene trees in species trees. Systematic Biology 46, 523-536. Maddison, W. P., and Knowles, L. L. 2006. Inferring phylogeny despite incomplete lineage sorting. Systematic Biology 55, 21-30. Mason-Gamer, R. J., Weil, C. F., and Kellogg, E. A. 1998. Granule -bound starch synthase: Structure, funtion and phylogenetic utility. Molecular Biology and Evolution 15, 1658-673. Masood, M . S., Nishikawa, T., Fukuoka, S.-L, Njenga, P. K., Tsudzuki, T., and Kadowaki, K.-I. 2004. The complete nucleotide sequence of wild rice (Oryza nivara) chloroplast genome: first genome wide comparative sequence analysis of wild and cultivated rice. Gene 340, 133-139. Mathews, S. R., and 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. Molecular Biology and Evolution. 13, 1141-1150. 119 Mathews, S., Tsai, R. C , and Kellogg, E. A. 2000. Phylogenetic structure in the grass family (Poaceae): evidence from the nuclear gene phytochrome B. American Journal of Botany 87, 96-107. Mathews, S., Spangler, R. E. , Mason-Gamer, R. J., and Kellogg, E. A. 2002. Phylogeny of Andropogoneae inferred from phytochrome B, GBSSI, and ndhF. International Journal of Plant Sciences 163, 441-450. Matsuoka, Y., Yamazaki, Y., Ogihara, Y., and Tsunewaki, K., 2002. Whole chloroplast genome comparison of rice, maize, and wheat: implications for chloroplast gene diversification and phylogeny of cereals. Molecular Biology and Evolution 19, 2084-2091. McPherson, M . A. 2003. Phylogenetics of Asparagales (Liliopsida). M . Sc. Thesis. University of Alberta, Edmonton, Alberta, Canada. Michelangeli, F. A., Davis, J. I, and Stevenson, D. W. 2003. Phylogenetic relationships among Poaceae and related families as inferred from morphology, inversions in the plastid genome, and sequence data from the mitochondrial and plastid genomes. American Journal of Botany 90, 93-106. Nadot, S., Bajon, R., and Lejeune, B. 1994. The chloroplast gene rps4 as a tool for the study of Poaceae phylogeny. Plant Systematics and Evolution 191, 27-38. Newell, T. K. 1969. A study of the genus Joinvillea (Flagellariaceae). Journal of the Arnold Arboretum 50, 527-555. Olmstead, R. G., and Palmer, J. D. 1994. Chloroplast DNA systematics: a review of methods and data analysis. American Journal of Botany 81, 1205-1224. 120 Ogihara, Y., Isono, K., Kojima, T., Endo, A., Hanaoka, M . , Shiina, T., Terachi, T., Utsugi, S., Murata, M . , Mori, N., Takumi, S., Ikeo, K., Gojobori, T., Murai, R., Murai, K., Matsuoka, Y., Ohnishi, Y., Tajiri, H., and Tsunewaki, K. 2000. Chinese spring wheat (Triticum aestivum L.) chloroplast genome: complete sequence and contig clones. Plant Molecular Biology Reporter 18, 243-253. Parkinson, C. L. , Mower, J. P., Qiu, Y. -L. , Shirk, A. J., Song, K., Young, N. D., dePamphilis, C. W., and Palmer, J. D. 2005. Multiple major increases and decreases in mitochondrial substitution rates in the plant family Geraniaceae. B M C Evolutionary Biology 5, 73. doi: 10.1186/1471-2148-5-73. Penny, D., Foulds, R., and Hendy, M . D. 1982. Testing the theory of evolution by comparing phylogenetic trees constructed from five different protein sequences. Nature 297, 197-200. Poe, S., and Swofford, D. L. 1999. Taxon sampling revisited. Nature 398, 299-300. Pollock, D. D., Zwickl, D. J., McGuire, J. A., and Hillis, D. M . 2002. Increased taxon sampling is advantageous for phylogenetic inferences. Systematic Biology 51, 664-671. Posada, D., and Crandall, K. A. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14, 817-818. Prasad, V., Stromberg, C. A. E. , Alimohammadian, H., and Sahni, A. 2005. Dinosaur coprolites and the early evolution of grasses and grazers. Science 310, 1177-1180. Rai, H. S., O'Brien, H. E. , Reeves, P. A., Olmstead, R. G., and Graham, S. W. 2003. Inference of higher-order relationships in the cycads from a large chloroplast data set. Molecular Phylogenetics and Evolution 29, 350-359. 121 Rambaut, A. 1998. Se-Al (sequence alignment editor version 1.0 alpha 1). Department of Zoology, University of Oxford, Oxford. Rannala, B., and Yang, Z. 1996. Probability distribution of molecular evolutionary trees: a new method of phylogenetic inference. Journal of Molecular Evolution 43, 304-311. Reeder, J. R. 1957. The embryo in grass systematics. American Journal of Botany 44, 756-769. Rokas, A., Williams, B. L. , King, N., and Carroll, S. B. 2003. Genome-scale approaches to resolving incongruence in molecular phylogenies. Nature 425, 798-804. Ronquist, F., and Huelsenbeck, J. P. 2003. MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572-1574. Rudall, P. J., and Bateman, R. M . 2004. Evolution of zygomorphy in monocot flowers: iterative patterns and developmental constraints. New Phytologist 162, 25-44. Rudall, P. J., Stuppy, W., J. Cunniff, J., Kellogg, E. A., and Briggs, B. G. 2005. Evolution of reproductive structures in grasses (Poaceae) inferred by sister-group comparison with their putative closest living relatives, Ecdeiocoleaceae. American Journal of Botany 92: 1432-1443. Sanderson, M . J., Wojciechowski, M . F., Hu, J.-M., Sher Khan, T., and Brady, S. G. 2000. Error, bias, and long branch attraction in data for two chloroplast photosystem genes in seed plants. Molecular Biology and Evolution 17, 782-797. Savolainen, V., Chase, M . W., Hoot, S. B., Morton, C. M . , Soltis, D. E. , Bayer, C , Fay, M . F., de Bruijn, A. Y., Sullivan, S., and Qiu, Y . -L . 2000. Phylogenetics of flowering 122 plants based on combined analysis of plastid atpB and rbcL gene sequences. Systematic Biology 49, 306-362. Schutze, P., Freita, G., and Weising, K. 2003. An integrated molecular and morphological study of the subfamily Suaedoideae Ulbr. (Chenopodiaceae). Plant Systematics and Evolution 239, 257-286. Shaw, J., Lickey, E. B., Beck, J. T., Farmer, S. B., Liu, W., Miller, J., Siripun, K., Winder, C. T., Schilling, E. E. , and Small, R. L. 2005. The tortoise and the hare II: relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis. American Journal of Botany 92, 142-166. Shinozaki, K., Ohme, M . , Tanaka, M . , Wakasugi, T., Hayashida, N., Matsubayashi, T., Zaita, N., Chunwongse, J., Obokata, J., Yamaguchi-Shinozaki, K., Ohto, C , Torazawa, K., Meng, B. Y., Sugita, M . , Deno, H., Kamogashira, T., Yamada, K., Kusuda, J., Takaiwa, F., Kato, A., Tohdoh, N., Shimada, H., and Sugiura, M . 1986. The complete nucleotide sequence of tobacco chloroplast genome: its gene organization and expression. EMBO Journal 5, 2043-2049. Simmons, M . P., Pickett, K. M . , Miya, M . 2004. How meaningful are Bayesian support values? Molecular Biology and Evolution 2, 188-199. Small, R. L., Ryburn, J. A., Cronn, R. C , Seelanan, T., and Wendel, J. F. 1998. The tortoise and the hare: choosing between noncoding plastome and nuclear Adh sequences for phylogeny reconstruction in a recently diverged plant group. American Journal of Botany 85, 1301-1315. 123 Soreng, R. J., and Davis, J. I. 1998. Phylogenetics and character evolution in the grass family (Poaceae): simultaneous analysis of morphological and chloroplast DNA restriction site character sets. Botanical Review 64, 1-85. Soreng, R. J., and Davis, J. I. 2000. Phylogenetic structure in Poaceae subfamily Pooideae as inferred from molecular and morphological characers: misclassification versus reticulation. Pp. 61-74. In: Jacobs, S. W. L., and Everett, J. [Eds.]. Grasses: Systematics and Evolution. CSIRO, Melbourne. Soreng, R. J., Peterson, P. M . , Davidse, G., Judziewicz, E. J., Zuloaga, F. O., Filgueiras, T. S., and Morrone, O. [Eds.]. 2003. Catalogue of New World Grasses (Poaceae): IV. subfamily Pooideae. Contributions from the United States National Herbarium 48, 1-730. Stebbins, G. L. 1957. Cytogenetics and evolution of the grass family. American Journal of Botany 43, 890-905. Stefanovic, S., Rice, D., and Palmer, J. D. 2004. Long branch attraction, taxon sampling, and the earliest angiosperms: Amborella or Monocots? B M C Evolutionary Biology 4, 35. doi:10.1186/1471-2148-4-35. Stromberg, C. A. E. 2005. Decoupled taxonomic radiation and ecological expansion of open-habitat grasses in the Cenozoic of North. Proceedings of the National Academy of Sciences U.S.A. 102, 11980-11984. Suzuki, Y., Glazko, G. V., and Nei, M . , 2002. Overcredibility of molecular phylogenetics obtained by Bayesian phylogenetics. Proceedings of the National Academy of Sciences U.S.A. 99, 16138-16143. 124 Swofford, D. L. 2002. PAUP* version 4.0M0. Computer program and documentation distributed by Sinaur, Sunderland, Mass. Tateoka, T., Inoue, S., and Kawano, S. 1959. Notes on some grasses. IX. Systematic significance of bicellular microhairs of leaf epidermis. Botanical Gazette 121, 80-91. Tzvelev, N. N. 1989. The system of grasses (Poaceae) and their evolution. Botanical Review 55, 141-203. Watson, L, and Dallwitz, M . J. 1992. The Grass Genera of the World. C A G International, Wallingford, United Kingdom. Whittingham, L. A., Slikas, B., Winkler, D. W., and Sheldon, F. H. 2002. Phylogeny of the tree swallow genus, Tachycineta (Aves: Hirundinidae), by Bayesian analysis of mitochondrial DNA sequences. Molecular Phylogenetics and Evolution 22, 43-441. Wilcox, T. P., Zwickl, D. J., Heath, T. A., and Hillis, D. M . 2002. Phylogenetic relationships of the dwarf boas and a comparison of Bayesian and bootstrap measures of phylogenetic support. Molecular Phylogenetics and Evolution 25, 361-371. Wilson, M . A., Gaut, B., and Clegg, M . T. 1990. Chloroplast DNA evolves slowly in the palm family (Arecaceae). Molecular Biology and Evolution 7, 303-314. Wolfe, K. H., Li , W.-H., and Sharp, P. M . 1987. Rates of nucleotide substitution vary greatly among plant mitochondrial, chloroplast, and nuclear DNAs. Proceedings of the National Academy of Sciences U.S.A. 84, 9054-9058. Wortley, A. H., Rudall, P. J., Harris, D. J., and Scotland, R. W. 2005. How much data are needed to resolve a difficult phylogeny? Case study in Lamiales. Systematic Biology 54, 697-709. 125 Young, N. D., and dePamphilis, C. W. 2005. Rate variation in parasitic.plants: correlated and uncorrected patterns among plastid genes of different funtion. B M C Evolutionary Biology 5, 16. doi: 10.1186/1471-2148-5-16. Zgurski, J. 2004. Analysis of multiple levels of vascular-plant phylogeny using a nuclear RNA polymerase subunit gene (RPB2) and a large plastid data set: Studies of the tracheophytes, cycads and the lilies and relatives. M . Sc. Thesis. University of Alberta, Edmonton, Alberta, Canada. Zhang, W. 2000. Phylogeny of the grass family (Poaceae) from rpl\6 intron sequence data. Molecular Phylogenetics and Evolution 15, 135-146. Zhang, W., and Clark, L. G. 2000. Phylogeny and classification of the Bambusoideae (Poaceae). Pp. 35-42 In: Jacobs, S. W. L., and Everett, J. [Eds.]. Grasses: Systematics and Evolution. CSIRO, Melbourne, Australia. Zwickl, D. J., and Hillis, D. M . 2002. Increased taxon sampling greatly reduces phylogenetic error. Systematic Biology 51, 588-598. 126 CHAPTER 4 1 Higher-Order Phylogenetic Relationships in the Commelinid Monocots, with a Focus on the Orders Commelinales and Poales. 4.1. INTRODUCTION 4.1.1. The Commelinid Monocots The commelinid monocots (Commelinidae) are a strongly supported clade (e.g., Givnish et al. 1999, 2006; Chase et al. 2000, 2006; Hilu et al. 2003; Davis et al. 2004; Graham et al. 2006) that includes approximately 32,000 species and 28 families - about half of all extant monocot species and one third of all monocot families. A similar group of taxa was identified in the first cladistic monocot study based on morphological characters (Dahlgren and Rasmussen 1983). Commelinids are an ancient angiosperm lineage; a recent estimate places the age of their stem node in the mid-Cretaceous at 122 mya (Janssen and Bremer 2004). They are morphologically and ecologically diverse, and include representatives of most major plant life forms, such as aquatics, submergents, epiphytes, xeromorphs, annuals, perennials, herbs, shrubs, and even trees. Commelinid monocots include several species of major and minor economic concern, such as the cereal grasses (e.g., Hordeum L. , Orzya L., Sacchamm L. , Sorghum Moench, Triticum L. , Zea L.), bananas 1 Chapter 4 represents a draft of a manuscript that will be submitted for publication: Saarela, J. M., Rai, H. S., Prentis, P., Briggs, B. G, Marchant, A. D., and Graham, S. W. Phylogenetic relationships among the commelinid monocots, with a focus on the orders Commelinales and Poales. 127 (Musa L. spp.), pineapples (Ananas spp.), ginger (Zingiber spp.), water chestnut (Eleocharis dulcis Hensch.), the horticultural bird-of-paradise (Strelitzia Banks ex Dryander spp.), and the aquarium plant Mayaca Aublet. The monophyly of the commelinid monocots is supported by several putative non-DNA synapomorphies, including the presence of ultraviolet fluorescing compounds (e.g., ester-linked ferulic acid and coamuric acid) in unlignified cell walls (Harris and Hartley 1976, 1980; Rudall and Caddick 1994; Rudall and Chase 1996), Strelitzia-like epicuticular waxes (long, often curly wax rodlets) (Barthlott and Frolich 1983; Frolich and Barthlott 1988), copiously starchy endosperm and possibly starchy pollen (Dahlgren and Rasmussen 1983; Dahlgren et al. 1985; Zona 2001), and the presence of silica bodies in leaves (Linder and Kellogg 1995). Several of these characters, however, are not present in all taxa, and some also occur outside the commelinid lineage. An additional cell-wall characteristic has recently been identified that may also be synapomorphic for all or most commelinids. Among nine commelinid families examined, the predominant noncellulosic polysaccharides in cell walls in all but Arecaceae are arabinose, xylose, and guluronic acids (highly substituted heteroxylans collectively referred to as glucuronoarabinoxylans [GAXs]; Harris etal. 1997). Multiple molecular analyses have identified five major lineages with moderate to strong support within the commelinid monocots. In the most recent angiosperm classification, four of these are treated as orders (Arecales, Commelinales, Poales, and Zingiberales), and one family, Dasypogonaceae, is unplaced at the ordinal level (APG II 2003). Close relationships among several of the families in Commelinales, Poales, and Zingiberales have been hypothesized in the past (e.g., Dahlgren et al. 1985; Cronquist 1981, 128 1988). Affinities of the palms (Arecaceae, the one family in Arecales) — a morphologically distinct and easily recognizable group of trees, shrubs or lianas, with usually unbranched trunks and trimerous flowers — have also been hypothesized previously to be with other commelinid monocots, but also with the screw pines (Pandanaceae), the Panama-hat palm family (Cyclanthaceae), and aroids (Araceae) (Uhle et al. 1995). Placement of Dasypogonaceae, a small and morphologically diverse family of xeromorphic herbs, shrubs, and trees, with usually small and inconspicuous flowers (Dahlgren et al. 1985), among the commelinid monocots (Chase et al. 1995) was largely unexpected. Dasypogonaceae Dumort. includes four Australian genera (Baxteria R. Br. ex Hook., Calectasia R. Br., Dasypogon R. Br., and Kingia R. Br.) that were previously included in the superficially similar family Xanthorreaceae (Asparagales sensu APG II 2003), but morphological and anatomical data indicate clearly that they are distinct (Dahlgren et al. 1985; Rudall and Caddick 1994; Rudall and Chase 1996). Baxteria and the tinsel lily genus, Calectasia, have sometimes been included in their own families, Baxteriaceae Takht. and Calectasiaceae Endlicher, respectively (e.g., Dahlgren et al. 1985; Takhtajan 1997). Nonetheless, molecular data indicate that the four dasypogonoid genera are a monophyletic group (Chase et al. 1995; Davis et al. 2004), and place them among other commelinids, a position supported by the presence of UV-flourescent compounds in their cell walls (e.g., Chase et al. 1993, 1995, 2006; Rudall and Caddick 1994; Rudall and Chase 1996; Givnish et al. 2006; Graham et al. 2006). Dasypogonaceae do not, however, have Strelitzia-like epicuticular wax and possibly starchy pollen, as do most other commelinid lineages (Barthlott and Frolich 1983; Frolich and Barthlott 1988; Rudall and Chase 1996; Zona 2001). 129 Phylogenetic relationships among the five major commelinid lineages have varied in previous studies (e.g., Chase et al. 1993, 1995, 2000, 2006; Givnish et al. 1999, 2006; Davis et al. 2004; Tamura et al. 2004; Graham et al. 2006). Most analyses infer a sister group relationship between Commelinales (see below), and Zingiberales. The latter order are a large clade that includes eight families of large, tropical, herbaceous monocots: the 'banana-families' (Lowiaceae, Heliconiaceae, Musaceae, and Strelitziaceae) with large, banana-like leaves and generally five fertile stamens, and the 'ginger-families' (a clade that includes Cannaceae, Costaceae, Marantaceae, and Zingiberaceae) generally with one fertile stamen (Kress et al. 2001). Arecales and Dasypogonaceae have been variously and weakly [i.e., bootstrap proportion (BP) < 70%] placed with respect to Commelinales-Zingiberales and Poales (e.g., Chase et al. 2000, 2006; Hilu et al. 2003; Davis et al. 2004; Tamura et al. 2004; Givnish et al. 1999, 2006; Graham et al. 2006). Some studies indicate that Arecales and Dasypogonaceae are the first two successive lineages among the five lineages; each has been identified as the sister-group of the rest of the commelinids (Chase et al. 2000; Davis et al. 2004). Two recent multigene studies suggest different sets of relationships, although support for these is consistently weak. A seven-gene study placed Dasypogonaceae as the sister group of Poales, and Arecales as the sister group of Commelinales-Zingiberales (Chase et al. 2006). In contrast, a 17-gene study placed Arecales as the sister group of Poales, and Dasypogonaceae as the sister group of Commelinales-Zingiberales (Graham et al. 2006). 4.1.2. Phylogenetic Relationships in the Order Commelinales Most molecular studies of monocots have identified a clade that includes the families Commelinaceae, Haemodoraceae, Hanguanaceae, Philydraceae, and Pontederiaceae (e.g., 130 Chase et al. 1993, 1995, 2000, 2006; Savolainen et al. 2000; Soltis et al. 2000; Givinish et al. 1999, 2006; Graham et al. 2002; Davis et al. 2004). In the most recent angiosperm classification, this clade is recognized as the order Commelinales (APG II 2003). In previous morphology-based classifications, these five families were not thought to be closely related, and were variously allied with or placed in taxa that are now included in most orders of monocots, sensu APG II 2003 (e.g., Cronquist 1981, 1988; Dahlgren and Rasmussen 1982; Dahlgren et al. 1985; Thorne 1992a,b; Takhtajan 1997). For example, Dahlgren et al. (1985) included Haemodoraceae, Philydraceae, and Pontederiaceae in their own monofamilial orders, Hanguanaceae in Asparagales, and Commelinaceae in an order (Commelinales) with the distantly related families Eriocaulaceae, Mayacaceae, Rapateaceae, and Xyridaceae. Commelinales are relatively small in terms of numbers of genera and species (Table 4.1), but the order is geographically and ecologically diverse. Commelinaceae and Haemodoraceae are largely terrestrial families, Hanguanaceae generally occur in marshy habitats, Philydraceae occur in wet areas, and Pontederiaceae are freshwater aquatics. Within-family phylogenetic relationships have been characterized recently in Commelinaceae (Evans et al. 2000, 2003), Haemodoraceae (Simpson 1990; Hopper et al. 1999), and Pontederiaceae (Eckenwalder and Barrett 1986; Graham and Barrett 1995; Kohn et al. 1996; Graham et al. 1998, 2002), but no study has sampled the four genera of Philydraceae. Inferred relationships among families in Commelinales have also varied. Most studies have included only one or a few Commelinales representatives (e.g., Chase et al. 1993; Davis 1995; Soltis et al. 2000; Tamura et al. 2004; Davis et al. 2004; Graham et al. 2006), but four studies have attained complete family-level sampling for the order. Based on the plastid 131 rbcL gene and the mitochondrial atpA gene, Davis et al. (2004) inferred a sister-group relationship between Philydraceae and Haemodoraceae, and found a clade that identified Hanguanaceae as the sister-group of Pontederiaceae-Commelinaceae. In contrast, analyses based on the plastid genes ndhF and rbcL inferred a sister group relationship between Philydraceae and Pontederiaceae, and found a clade that identified Haemodoraceae as the sister-group of Commelinaceae-Hanguanaceae (Graham et al. 2002). All relationships in these two-gene studies received bootstrap support less than 50%. Based on the plastid gene ndhF but with much denser sampling across monocots compared with Graham et al. (2002), Givnish et al. (2006) identified two well supported clades (Haemodoraceae-Pontederiaceae and Commelinaceae-Hanguanaceae), with Philydraceae inferred weakly to be the sister group of the rest of the order. Finally, in a recent study based on multiple mitochondrial, nuclear, and plastid regions, Chase et al. (2006) inferred a Commelinaceae-Hanguanaceae clade that is the sister group of a (Philydraceae (Haemodoraceae-Pontederiaceae)) clade. Both two-family clades received strong support (i.e., BP > 90%) in this study, but the position of Philydraceae received only moderate support (i.e., BP = 70-89%). No study has robustly inferred relationships among all Commelinales families. 4.1.3. Phylogenetic Relationships in the Order Poales The clade currently classified as the order Poales (APG II 2003) was identified in the first angiosperm- and monocot-wide molecular phylogenetic studies (e.g. Chase et al. 1993; Duvall et al. 1993), and it has since been recovered with moderate to strong support in several studies (e.g., Givnish et al. 1999, 2006; Tamura et al. 2004; Chase et al. 2006; Graham et al. 2006). According to the most recent monocot classification, Poales includes 132 17 families (Chase 2004; Table 4.1). In earlier classifications, several of these families were variously included in multiple orders (e.g., Cronquist 1981, 1988; Dahlgren et al. 1985), often with families with which they are now known to be related to only distantly. Close relationships among other families currently included in the order, however, have been hypothesized previously. For example, with the exception of Commelinaceae (see above), all families that Cronquist (1988) included in his subclass Commelinidae are currently part of Poales. Hydatellaceae, a family that has traditionally been difficult to place among any of the orders of monocots (e.g., Dahlgren et al. 1985; Hamann 1998) was included in Poales by APG II (2003), but recent work has demonstrated that it is not part of the clade, or even monocots as a whole (see Chapter 5). Poales includes several taxa that are among the most readily-identifiable to family -and most taxonomically difficult - monocot groups, such as the grasses (Poaceae), sedges (Cyperaceae), and pineapples (Bromeliaceae). Many of the monocot families with highly modified, wind-pollinated flowers are part of Poales, although the order also includes several families with showy perianths or bracts, and insect pollination. Poales includes several small families with only one or a few genera and species, (e.g., Ecdeiocoleaceae, Thurniaceae), and others that are among the largest angiosperm families, with hundreds of genera, thousands of species, and cosmopolitan distributions (e.g., Poaceae, Cyperaceae) (Table 4.1). In addition to the 16 families currently included in the order (excluding Hydatellaceae; Table 4.1), several additional small, segregate families have been described and recognized recently (e.g., Hopkinsiaceae, Lyginiaceae, Mapaniaceae, Prioniaceae) (APG 1998; Briggs and Johnson 1998; Munro and Linder 1998; Shipunov 2003). These are included in more inclusive taxa (Table 4.1) following APG II (2003). 133 Molecular studies have clarified several aspects of relationship among Poales families (e.g., Givnish et al. 1999, 2006; Bremer 2002; Michelangeli et al. 2003; Davis et al. 2004; Chase et al. 2006; Graham et al. 2006). Most studies reveal a basal grade of families that includes Bromeliaceae, Rapateaceae, and Typhaceae [the 'basal Poales' sensu Linder and Rudall (2005)], but the exact divergence pattern of these families is not well established. Most previous analyses indicate variously that Bromeliaceae and Typhaceae are a clade (Bremer et al. 2002; Chase et al. 2006; Givnish et al. 2006) or that they are respectiively the successive sister groups of the rest of Poales (Michelangeli et al. 2003; Graham et al. 2006). Rapateaceae have been variously inferred to be the sister group of Bromeliaceae (Clark et al. 1993; Chase et al. 1995; Linder and Kellogg 1995; Givnish et al. 1999), the sister group all Poales (Chase et al. 2000; Bremer et al. 2002), the sister group of the xyrid clade (Davis et al. 2004), or the sister group of all Poales except Bromeliaceae and Typhaceae (Chase et al. 2006; Givnish et al. 2006). The remainder of Poales includes 13 families and four major clades [the 'cyperid', 'graminid', 'restiid', and 'xyrid' clades, sensu Linder and Rudall (2005)], although a few families have not been unequivocally placed in one of these groups. Most of these families have a substantially accelerated rate of plastid evolution compared with other monocots, but the reason for this is not yet clear (Givnish et al. 1999, 2006; Graham et al. 2006; Chapter 3). A close relationship has been inferred between the graminid and restiid clades (e.g., Michelangeli et al. 2003; Chase et al. 2006; Givnish et al. 2006; Graham et al. 2006;), and this large clade has been referred to as the 'core Poales' (Linder and Rudall 2005; I use this designation here) or the 'graminoid' clade (Bremer 2002). Relationships among the core 134 Poales and the cyperid and xyrid clades, however, are not yet fully established (e.g., Bremer 2002; Chase et al. 2006; Givnish et al. 2006; Graham et al. 2006). The restiid clade includes the 'southern rush' families Anarthriaceae, Centrolepidaceae, and Restionaceae (Linder and Rudall 2005; Table 4.1). Analyses of different plastid genes variously resolve Centrolepidaceae as the sister group of Restionaceae (Briggs et al. 2000; Michelangeli et al. 2003), or include it within Restionaceae (Linder and Rudall 1993; Bremer 2002), the latter circumscription supported by morphological data (e.g., Hamann 1975; Linder and Kellogg 1.995; Linder et al. 2000). It is not clear if Centrolepidaceae are a distinct lineage, or if they are part of a paraphyletic Restionaceae. The western Australian genera, Hopkinsia and Lyginia, were recently segregated from Restionaceae on the basis of several morphological and anatomical differences, and placed in their own monogeneric families, Hopkinsiaceae and Lyginiaceae (Briggs and Johnson 2000; Linder et al. 2000). The few molecular studies that have sampled these genera place them with the monogeneric Anarthriaceae in a clade that is the sister group of Centrolepidaceae/ Restionaceae (e.g., Briggs et al. 2000; Bremer 2002). Following recommendations by APG II (2003) not to recognize small, monogeneric families that are clearly sister taxa, Hopkinisia and Lyginia are now included with Anarthria in Anarthriaceae (APG II 2003). The graminid clade includes Ecdeiocoleaceae, Joinvilleaceae, Poaceae, and sometimes Flagellariaceae (e.g., Michelangeli et al. 2003; Tamura et al. 2004; Davis et al. 2004; Linder and Rudall 2005; Chase et al. 2006). An Ecdeiocoleaceae-Joinvilleaceae-Poaceae clade has been consistently recovered with strong support, but relationships among these families remain largely equivocal (e.g., Campbell and Kellogg 1987; Doyle et al. 1992; Linder and Rudall 1993; Bremer 2000, 2002; Michelangeli et al. 2003; Davis et al. 2004; 135 Chapter 3). The identity of the sister-group of the grasses - the world's most economically important family of plants - is unfortunately not clear (Chapter 3). Resolution of this phylogenetically critical trichotomy has important implications for characterizing homologies of grass reproductive structures among monocots (e.g., Rudall et al. 2005). Flagellariaceae have been variously inferred to be the sister group of Ecdeiocoleaceae-Joinvilleaceae-Poaceae, a larger restiid-graminid clade, or the xyrid clade (e.g., Bremer 2002; Davis et al. 2004; Chase et al. 2006; Graham et al. 2006; Givnish et al. 2006). The cyperid clade - equivalent in circumscription to the former Cyperales sensu Dahlgren et al. (1985) - includes the large and cosmopolitan sedge (Cyperaceae) and rush (Juncaceae) families, and Thurniaceae (e.g., Givnish et al. 1999, 2006; Chase et al. 2000, 2006; Bremer 2002; Michelangeli et al. 2004; Linder and Rudall 2005). A sister-group relationship between Cyperaceae and Juncaceae is supported by morphological and molecular data, with Thurniaceae then resolved as their sister group. Cyperaceae are morphologically diverse, and recent molecular work on the family has identified two major sedge clades (e.g., Muasya et al. 1998, 2000; Simpson et al. 2003, 2006) which are recognized as subfamilies Mapaniodeae C. B. Clark and Cyperoideae in the most recent Cyperaceae classification (Simpson 2006). Shipunov (2003) recognized Mapanioideae at family rank (Mapaniaceae Shipunov), but recognition of two sedge families has not been accepted by cyperologists. The monotypic South African genus, Prionium, was traditionally included in Juncaceae on the basis of morphological characteristics (Cutler 1969; Dahlgren and Clifford 1982; Dahlgren et al. 1985; Simpson 1995), but early molecular analyses indicated that it is not part of Juncaceae and it was therefore placed in its own family, Prioniaceae (Munro and Linder 1998; APG 1998). Subsequent molecular studies have 136 placed Prionium in a clade with Thurnia (Thurniaceae) (e.g., Chase et al. 1993; Munro and Linder 1998; Bremer 2002; Davis et al. 2004), and it is now included in Thurniaceae (APG II 2003; see rationale for combining monogeneric families discussed above). Several molecular studies have indicated that Juncaceae may not be monophyletic, because the South American cushion-forming genus Oxychloe (traditionally included in Juncaceae) was placed as the sister group of, or within, Cyperaceae (e.g., Chase et al. 1993; Plunkett et al. 1995; Muasya et al. 1998, 2000). However, these earlier studies were based on misidentified material; subsequent analyses have found that Oxychloe is part of Juncaceae, indicating that the family as traditionally circumscribed is natural (Kristiansen et al. 2005; Roalson 2005). Family composition of the xyrid clade has varied among studies, and support for the clade and relationships within it have generally been low (e.g., Chase et al. 2000, 2006; Bremer 2002; Michelangeli et al. 2003; Davis et al. 2004; Givnish et al. 2006). Linder and Rudall (2005) suggested that the xyrid clade probably contains the families Eriocaulaceae, Hydatellaceae, Mayacaceae, and Xyridaceae. Several studies have inferred a close relationship between Eriocaulaceae and Xyridaceae (Stevenson and Loconte 1995; Chase et al. 2000, 2006; Bremer 2002; Givnish et al. 2006). Some studies indicate that Xyridaceae may not be monophyletic (e.g., Davis et al. 2004), a finding that is possibly consistent with the sometimes-recognized segregate family, Abolbodaceae Nakai (e.g., Steyermark 1984; APG 1998). Mayacaceae have been variously inferred to be the sister group of Xyridaceae (Bremer 2002; Davis et al. 2004), the cyperid clade (Chase et al. 2006), or the cyperid clade and Eriocaulaceae-Xyridaceae (Givnish et al. 2006). Hydatellaceae have been placed among other xyrid families in some previous studies, but this was based on a contaminant rbcL sequence that is a chimera of a grass and a moss (Chapter 5). I have demonstrated that 137 Hydatellaceae are not part of Poales (Chapter 5), and I therefore do not discuss this family further here. 4.1.4. Objectives of the Current Study Major outstanding questions in higher commelinid monocot phylogenetics include the relative branching orders of Arecales, Dasypogonaceae, Commelinales-Zingiberales, and Poales; relationships among the five families in Commelinales, particularly the position of Philydraceae; and relationships among families in Poales, particularly the relative branching orders of Bromeliaceae, Rapateaceae, and Typhaceae, the positions of Flagellariaceae and Mayacaceae, the compositions and interrelationships of the xyrid, cyperid, restiid, and graminid clades, and the identity of the sister group of the grasses. Strategies to obtain more resolved and better supported phylogenetic trees include examining more characters per taxon (e.g., Hillis 1998; Poe and Swofford 1999; Graham and Olmstead 2000a; Rokas et al. 2003; Graham et al. 2006), and sampling major lineages more densely (e.g., Hillis 1998; Swofford et al. 2001). The former approach has helped resolve several aspects of deep relationship among basal angiosperms (Graham and Olmstead 2000a, b; Graham et al. 2000), cycads (Rai et al. 2003), Liliales (Zgurski 2004), Asparagales (McPherson 2003), Poaceae (see Chapter 3), and monocots as a whole (Graham et al. 2006). To clarify relationships in the commelinid monocots, I take this approach here, supplemented with denser taxon sampling than attempted before. My approach is to sample multiple plastid genes and associated noncoding regions (Graham and Olmstead 2000a) from multiple exemplar taxa (in most cases) for each of the commelinid families, with a focus on the orders Commelinales and Poales. 138 4.2 M A T E R I A L S AND M E T H O D S 4.2.1. Taxonomic Sampling The current taxon sampling includes representatives from each of the five major commelinid lineages, including Dasypogonaceae, Arecales, and all families in Commelinales and Poales. I sampled multiple representatives from most families in Commelinales and Poales, and single exemplars from Eriocaulaceae, Restionaceae, and the monogeneric Flagellariaceae, Hanguanaceae, and Mayacaceae (Table 4.2). This taxon sampling also includes exemplars from each of the segregate families that have been recognized recently (i.e., Abolbodaceae, Hopkinsiaceae, Lyginiaceae, Mapaniaceae, Prioniaceae; see Introduction). To maximize the morphological, molecular, and taxonomic diversity sampled in large families, I attempted to choose exemplar taxa from different clades (often classified as subfamilies or tribes in rank-based classifications) that span the root node of each family, based on earlier phylogenetic studies (e.g., Simpson 1990; Bruhl 1995; Terry et al. 1997; Graham et al. 1998, 2002; Hopper et al. 1999; Horres et al. 2000; Givnish et al. 2000, 2006; GPWG 2001; Evans et al. 2003; Drabkova et al. 2003; Linder et al. 2003; Simpson et al. 2003; Campbell 2004; Roalson 2005). In the current study, 28 taxa are new, compared to the eleven new taxa examined in Chapter 3. In Commelinales, taxon sampling is greatest in Philydraceae, where I sampled a representative from each of its three to four genera. I attempted to sample each of the six species in Philydraceae, but DNA for two species was not available. Material from three Helmholtzia novoguineensis (K. Krausse) Skotts. herbarium specimens [Brass 13431 (GH), Brass 12859 (GH), and Hoogland & Craven 11068 (GH)] and a collection of Philydrella pygmaea (R. Br.) Caruel [Morrison s.n. (BRI)] was recalcitrant to molecular study. Within 139 Poales, taxon sampling is most dense within Poaceae, where I include 13 taxa examined previously (Chapter 3), recently published sequence data from three Poaceae plastid genomes (Table 4.3) and new sequence data for Puelia oylriformis (Puelioideae), one of two grass subfamilies not included in my earlier study (Chapter 3). 4.2.2. Genomic Sampling I obtained DNA from various sources, or extracted DNA from field-collected, silica gel-dried leaf material, using the C T A B protocol of Doyle and Doyle (1987), with the addition of 2% /?-mercaptoethanol to each extraction. For a subset of taxa, two different species (Anigozanthos, Mapania) or two different individuals (Anarthria) were used to obtain the plastid data (Table 4.2). I amplified and sequenced multiple plastid genes involved in a range of functions, including photosynthesis (atpB, psbB, psbC, psbD, psbE, psbF, psbH, psbJ, psbL, psbN, psbT, rbch), chlororespiration (ndhB and ndhF), and plastid translation (rp\2, rpsl, 3'-rpsl2), and associated noncoding regions (intergenic spacers that spanps^B-T-N-H,psbE-F-L-i, and V-rps\2-rpsl gene clusters, and introns in rp/2, 3'-rpsYl, and ndhB), using the methods and primers described in Graham and Olmstead (2000a), McPherson (2003), and Graham et al. (2006). I used additional primers from Zurawski et al. (1984), Olmstead and Sweere (1994), Hoot et al. (1995), Kim and Jansen (1995); Neyland and Urbatsch (1996a,b), and Graham et al. (1998), and the following two previously unpublished primers designed by H. S. Rai (UBC): C91F: 5'-T T G T G A G G T A C A R C A A T T A T T A G G - 3 ' (atpB), and F45R: 5 ' - C A T T A A A G A G C G TTTCCAC-3' (psbD). I sequenced all regions in forward and reverse directions, and in most cases I obtained sequencing products from duplicate PCR products amplified from independent DNA extractions to detect pipetting errors or cross-contamination (I observed 140 none). A subset of sequences for regions that I was unable to amplify and sequence satisfactorily are from GenBank (Table 4.2). 4.2.3. Data Assembly I performed basecalling and contig assembly using Sequencher 4.1 (Genes Code Corporation, Ann Arbor, MI), and determined gene boundaries by direct comparisons with sequences of Nicotiana tabacum L. (Shinozaki et al. 1986) and Ginkgo biloba L. I added sequences to a previously published alignment that includes representatives from across extant seed plants, including monilophytes, conifers, cycads, gnetophytes, basal angiosperms, magnoliids, eudicots and monocots (Graham and Olmstead 2000a; Rai et al. 2003; McPherson 2003; Zgurski 2004; Graham et al. 2006). The matrix includes two exemplar taxa from each of Dasypogonaceae and Zingiberales (Musaceae and Strelitziaceae), and a single exemplar from Arecaceae. I added data for portions of the psbL-psbi intergenic spacer and psbJ missing from a previous version of the matrix for Acorus calamus L. (positions 63750-63950; NC_007407; Goremykin et al. 2005), 195 bp from the 3'-end of rbcL that was not included in previous versions of the matrix for Triticum aestivum (NC_002762; Ogihara et al. 2000) and Zea mays (NC_001666; Maier et al. 1995), and data for the current plastid genome sampling for 20 taxa representing 17 families and 14 orders of eudicots (APG II 2003), and four monocot taxa (three grasses and an orchid), most from recently published plastid genomes (Table 4.3). The eudicot exemplars included here are an expansion from the seven orders represented in Graham and Olmstead (2000a). The ndhF gene is absent from the genome of Phalaenopsis afrodite (Chang et al. 2006), and ndhF data are not available for Ranunculus macranthus. The final matrix includes 221 taxa. 141 I aligned sequences manually using Se-Al version 1.0 alpha (Rambaut 1998) according to guidelines outlined in Graham et al. (2000). The total length of the aligned data matrix is 31, 285 base pairs (bp). The unaligned total sequence length of exemplar taxa added here range from -10.4 kb in Anarthria to -15.1 kb in Aphelia; one major exception is Lyginia, for which I was only able to obtain -6.4 kb of sequence data. For multiple taxa I was unable to amplify and sequence -200 bp at the 3'-end of rbcL and the second exon of 3'-rps\2. Aside from these instances, all taxa are nearly completely represented for all regions. The aligned data matrix is substantially larger than the unaligned length for any single taxon because of substantial gaps and/ or unalignable regions, including some from non-angiosperm taxa that are part of the matrix but not included in this study. I set aside noncoding regions that were too difficult to align as staggered unique gapped regions. The unique regions are effectively ignored for parsimony-based tree searches and scores (Graham et al. 2006), and should have only minimal effect for model-based methods (e.g., on estimation of base frequency parameter values). I did not include insertion/ deletion (indel) events in the analysis, and coded gaps in the data as 'missing data'. Further details on the alignment and the data matrix are outlined in Graham and Olmstead (2000a), Rai et al. (2003), McPherson (2003) and Graham et al. (2006). 4.2.4. Parsimony and Likelihood Phylogenetic Analyses In a previous study, I observed several instances of moderate to strong conflict between protein coding and noncoding partitions of the data examined here for fewer monocots (Chapter 3). Until the source of this conflict can be resolved, it seems reasonable to exclude most of these noncoding regions from consideration. In my analyses here, I 142 therefore include only protein-coding regions and some very conservative noncoding regions in the inverted repeat, including three introns in rpl2, V-rpsYl, and ndhB, and intergenic spacers between y-rpsYl-rpsl, rpsl-ndhB, and ndhB-trnL(CAA) (i.e., intergenic spacers between the various photosystem II genes are excluded). In Chapter 3,1 also observed substantial inflation of Bayesian posterior probabilities compared with bootstrap support from parsimony and likelihood analyses, consistent with the findings of several studies (e.g., Suzuki et al. 2002; Cummings et al. 2003; Erixon et al. 2003); I therefore do not include Bayesian analyses here. I examined two different taxon sets with maximum parsimony: a large angiosperm-wide 159-taxon set, and a smaller 113-taxon set representing all major lineages of monocots, with dense sampling in commelinids. For maximum likelihood analysis, I examined a 69-taxon subset of this smaller taxon set that included all commelinid taxa and five outgroup taxa from Asparagales and Liliales. For the 159-taxon set, I accepted Amborella as the probable sister group of the rest of the angiosperms (e.g., Graham and Olmstead 2000a; Hilu et al. 2003; Chapter 5) and used it to root the trees. For the 113-taxon set I accepted Acorus as the probable sister group of the rest of the monocots (e.g., Duvall et al. 1993b; Chase et al. 2000; Graham et al. 2006), and for the 69-taxon set I rooted the trees with the exemplars from Asparagales and Liliales. I performed MP searches using PAUP* version 4.0M0 (Swofford 2002), with all characters and character-state changes equally weighted, using tree bisection-reconnection (TBR) branch-swapping, and with 10 random addition replicates performed for each search, but otherwise using default conditions. I performed the maximum likelihood searches using PHYML version 2.4.4 (Guindon and Gascuel 2003). I chose an optimal DNA substitution 143 model using the hierarchical likelihood ratio test (hLRT) and the Akaike Information Criterion (AIC), as implemented in ModelTest ver. 3.7 (Posada and Crandall 1998). The optimal model was GTR + F + I [general time reversible rate (GTR) model with a proportion of invariant sites (I) considered and a gamma (T) distribution used to account for among-site rate variation]. I estimated all model parameters from the data, and otherwise used default settings in PHYML. I estimated branch support for MP and M L using nonparametric bootstrap analysis (Felsenstein 1985). For MP bootstrap analysis, I used 100 bootstrap replicates, with the same search parameters as the initial search, but using a single random addition replicate per bootstrap replicate. For M L bootstrap analysis I used 100 bootstrap replicates. I consider "strongly" supported or "robust" branches to have bootstrap support of 90% or more, "moderately" supported branches to have bootstrap support 70-89%, and "weakly" or "poorly" supported branches to have bootstrap support < 70%. I obtained bootstrap support values <50% from the bootstrap log in PAUP*. 4.3 RESULTS 4.3.1. Outgroup Relationships Outgroup relationships for basal angiosperms, monocots as a whole, Asparagales, and Liliales are discussed elsewehere (Graham and Olmstead 2000a; Graham et al. 2000, 2006; McPherson 2003; Zgurski 2004), therefore I do not consider these further here. My angiosperm-wide analysis (Fig. 4.1) included 20 eudicot exemplars not included in these previous analyses. Within the eudicots, I generally identify the same higher order lineages (i.e., rosids and asterids), Saxifragales, and a grade of 'basal' lineages (Ranunculales, 144 Proteales, Trochodendrales, Gunnerales) that have been inferred in previous angiosperm-wide studies (Fig. 4.2; reviewed in Soltis et al. 2005). However, relationships among the rosid lineages differ in several aspects compared with a recent analysis that considered data for complete plastid genomes for a smaller complement of taxa and with no Rosales exemplars (Cai et al. 2006). Interrelationships among Vitales, asterids, and rosids are not clear here (Fig. 4.2). I infer monocots to be the sister group of eudicots and Ceratophyllum, but this relationship receives very weak support here [maximum parsimony bootstrap proportion (BPMP) < 50%; Fig. 4.2). Eudicots are strongly supported as monophyletic (BPMP = 100%; Fig. 4.2). The monophyly of the monocots is only weakly supported (BPMP = 56%, Fig. 4.2; but see Graham et al. (2006) for details on how exclusion of Ceratophyllum from analyis substantially increases bootstrap support for monocots as a whole). 4.3.2. Relationships in the Commelinid Monocots Within the monocots, commelinid monocots are strongly supported as a clade [BPMP = 100%) and maximum likelihood bootstrap proportion (BPML) = 100%; Figs. 4.3-4.5] that is the sister-group of Asparagales (BPMP = 99%; Figs. 4.3, 4.4). Commelinales, Poales, and Zingiberales sensu APG II (2003) are each strongly supported as monophyletic in parsimony and likelihood analyses [BPMP = 100% (for each order) and BPML = 100% (for Poales and Zingiberales) and 94% for Commelinales; Figs. 4.4, 4.5]. Commelinales and Zingiberales are strongly supported as a clade (BPMP = 100% and BPML = 100%; Figs. 4.3-4.5), but the relative positions of Arecales, Dasypogonaceae, Commelinales-Zingiberales, and Poales are not clear here. In parsimony and likelihood analyses, Arecales are variously placed as the sister group of Commelinales-Zingiberales [BPMP = 46% (Fig. 4.4) and BPML = 30 % (data 145 not shown) or the sister group of all commelinids except Dasypogonaceae [BPMP = 3 7 % (data not shown) and BPML = 3 7 % , Fig. 4 . 5 ] , and Dasypogonaceae are variously placed as the sister-group of all remaining commelinids [BPMP = 1 3 % (data not shown) and BPML= 3 4 % , Fig. 4 . 5 ) ] , the sister group of all commelinids except Arecales (BPMP = 3 2 % and BPML = 3 4 % , data not shown), or the sister group of Poales [BPMP = 5 3 % (Fig. 4 . 4 ) and BPML = 1 3 % (data not shown)]. Maximum likelihood analysis provides weak support for a sister group relationship between Commelinales-Zinbigerales and Poales (BPML = 6 4 % ; Fig 4 . 5 ) ; this relationship receives very weak support from parsimony analysis (BPMP = 3 7 % ; data not shown). 4.3.3. Phylogenetic Relationships in Commelinales Commelinaceae and Hanguanaceae are strongly supported as sister taxa (BPMP and BPML = 1 0 0 % ; Figs 4 . 4 , 4 . 5 ) . I find moderate support for the branch supporting a sister-group relationship between Philydraceae and Haemodoraceae-Pontederiaceae (BPMP = 8 0 % and BPML = 8 3 % , Figs. 4 . 4 , 4 . 5 ) , and moderate to strong support for the branch within Philydraceae that identifies a sister group relationship between Philydrella and Philydrum-Helmholtzia (BPMP = 9 4 % and BPML = 8 4 % ; Figs. 4 . 4 , 4 . 5 ) ; Philydraceae are monophyletic (BPMP and BPML = 1 0 0 % ; Figs. 4 . 4 , 4 . 5 ) . 4.3.4. Phylogenetic Relationships in Poales All families in Poales sampled here with more than one exemplar per family are monophyletic for the current taxon sampling and receive 1 0 0 % bootstrap support from parsimony and likelihood analyses (Figs. 4 . 4 , 4 . 5 ) . Within Poales, I find a moderately to 1 4 6 strongly supported clade that includes all Poales families except Bromeliaceae and Typhaceae (BPMP = 78% and BPML = 98%; Figs. 4.4, 4.5); the relative positions of these latter two families at the base of Poales are not clear here. Rapateaceae are weakly identified as the sister group of the rest of this large clade (BPMP = 59% and BPML = 35%; Figs. 4.4, 4.5). I also find 99-100%) bootstrap support from parsimony and likelihood analyses for several groups in Poales, including a cyperid clade (Thurniaceae, (Juncaceae-Cyperaceae)), a restiid clade (Anarthriaceae (Centrolepidaceae-Restionaceae)), and an Ecdeiocoleaceae-Joinvilleaceae-Poaceae clade (Figs. 4.4, 4.5). Flagellariaceae and these latter three families are a clade with moderate support (BPMP = 87%; Fig. 4.4) to strong (BPML = 98%; Fig. 4.5), and these four families (the graminid clade) are strongly supported as the sister group of the restiid clade (Figs. 4.4, 4.5). The graminid and restiid clades together represent the core Poales. Within the graminid clade, Ecdeiocoleaceae are weakly inferred to be the sister group of Poaceae (BPMP - 62% and BPML = 62%; Figs. 4.4,4.5). Relationships among subfamilies within Poaceae are generally consistent with those inferred in my earlier study (Chapter 3), and Puelioideae are inferred here with strong support (BPMP and BPML = 100%; Figs. 4.4, 4.5) to be the third of three successive lineages (following Anomochlooideae and Pharoideae, respectively) that are the sister groups of the remaining grasses. Relationships among Eriocaulaceae, Mayacaceae, and Xyridaceae differ between maximum parsimony and maximum likelihood analyses. In parsimony analysis, a clade that includes these three families receives very weak support (BPMP = 56%; Fig. 4.4), and Eriocaulaceae and Mayaceae are moderately supported as sister taxa (BPMP = 78%; Fig. 4.4). The position of this xyrid clade with respect to the cyperid clade or the core Poales is not clear here (Fig. 4.4). By contrast, likelihood analyses identify a moderately supported 147 Eriocaulaceae-Xyridaceae clade ( B P M L = 89%; Fig. 4.5) that is strongly supported as the sister group of the core Poales ( B P M L = 93%; Fig. 4.5). Mayacaceae are very weakly inferred to be the sister group of this large clade ( B P M L = 12%; Fig. 4.5). 4.4 DISCUSSION 4.4.1. Phylogenetic Relationships Among the Major Lineages of Commelinid Monocots My phylogenetic analysis of multiple plastid genes and conservative noncoding regions sampled broadly across multiple monocot taxa provides strong support for the commelinid monocot clade (Figs 4.4, 4.5). Several putative non-DNA synapomorphies have been identified for this clade (see Introduction). Within the commelinid monocots, I identify a strongly supported sister-group relationship between Commelinales and Zingiberales (Figs. 4.4, 4.5), consistent with the findings of several recent studies (e.g., Chase et al. 2000, 2006; Soltis et al. 2000; Hilu et al. 2003; Givnish et al. 2006; Graham et al. 2006). This large clade includes 13 families [five in Commelinales (see below), and eight in Zingiberales (see Kress et al. 2001)]. The Zingiberales are well supported by morphological and molecular data (e.g., Kress 1995; Kress et al. 2001; Givnish et al. 2006). A few possible non-DNA synapomorphies for a Commelinales-Zingiberales clade have been identified, including the presence of phenylphenalenones (known in Haemodoraceae, Musaceae, Pontederiaceae, and Strelitziaceae; see Otalvaro et al. 2002), inflorescences with many-flowered cincinnal branches, and invasive or plasmodial tapetum (Stevens 2006). Higher-order relationships among Arecales, Dasypogonaceae, Commelinales-Zingiberales, and Poales are not consistently resolved here, and no particular relationships receive strong support. In parsimony and likelihood analyses, Arecales and Dasypogonaceae 148 are variously placed among Cornmelinales-Zingiberales and Poales. In some of the shortest trees from parsimony, Dasypogonaceae are weakly placed in a clade with Poales ( B P M P = 53%; Fig. 4.4); this topology receives only 13% bootstrap support from maximum likelihood analysis (data not shown). Chase et al. (2006) found this same placement for Dasypogonaceae, with a similar level of support from parsimony. Consistent difficulty in placing Arecales and Dasypogonaceae among commelinid monocots may reflect a rapid initial radiation of the major commelinid lineages (i.e., divergence of major lineages within a relatively short time). Indeed, Jansenn and Bremer (2004) dated the crown nodes of Arecaceae and Dasypogonaceae to the mid-Cretaceous at 100-110 Mya, a period in which angiosperms were undergoing major radiation according to the fossil record (e.g., Herendeen and Crane 1995). Typical of rapid radiations, the lengths of these deep internal commelinid branches are relatively short (Figs. 4.4, 4.5). Inability to clearly (i.e., no collapsed branches in a strict consensus tree and no conflicts among analyses) and strongly (i.e., high support) infer deep commelinid relationships is therefore likely a function of sampling error; not enough characters have been examined to characterize the divergence pattern of these ancient lines for the regions examined (Graham et al. 2006). Additional sampling from the plastid or other genomes will likely be necessary to place Arecales and Dasypogonaceae with strong support, in combination with denser taxon sampling, particularly in these two lineages. My current sampling includes two (Dasypogon and Kingia) of the four Dasypogonaceae genera, representing each of two putative clades in the family (e.g., Davis et al. 2004), but Baxteria and Calectasia should also be included in future work. I have also only included a single exemplar from the large family Arecaceae. Recent work has identified two major clades in Arecaceae that differ in their rates of plastid 149 evolution. One clade (subfamily Arecoideae), which includes Roystonea (examined here), has a relatively slow rate of plastid evolution, and the other clade (subfamily Calamoideae) has a more rapid rate of plastid evolution (Asmussen 2006). Inclusion of exemplars from Calamoideae in future studies may help place Arecaceae among commelinids. 4.4.2. Phylogenetic Relationships in Commelinales Until recently, Commelinaceae, Hanguanaceae, Haemodoraceae, Philydraceae, and Pontederiaceae (Commelinales) were not all considered to be closely related, although close affinities among some of these families had been suggested previously (e.g., Cronquist 1981, 1988; Dahlgren et al. 1985; Thorne 1992a, b). Previous molecular data support their inclusion in a single clade (e.g., Chase et al. 1995, 2000, 2006; Graham et al. 2002; Davis et al. 2004; Givnish et al. 1999, 2006), although only a subset of these studies included all five families. I also find strong support for this five-family lineage here (Figs. 4.4, 4.5). Only a few potential non-DNA synapomorphies have been identified for Commelinales in its current circumscription, including abundant helobial endosperm, tannin cells in the perianth, and the presence of sclereids in the placenta (e.g., Givnish et al. 1999; Judd et al. 2002; Soltis et al. 2005) . Haemodoraceae-Philydraceae-Pontederiaceae — Several workers have suggested close relationships among Haemodoraceae, Philydraceae, and Pontederiaceae (e.g., Simpson 1990, 1993; Kress 1995; Linder and Kellogg 1995). Only a single molecular study places these three families in a clade, and this topology receives only moderate support (Chase et al. 2006) . Here I also find a moderately supported Haemodoraceae-Philydraceae-Pontederiaceae clade (Figs. 4.4. 4.5). Several putative non-DNA synapomorphies have been 150 identified for this three-family lineage, including the presence of styloids, tanniniferous tepals that are generally persistent in fruit (e.g., Soltis et al. 2005), and tepalar perianths. The aquatic or semi-aquatic habit shared by Philydraceae and Pontederiaceae (Haemodoracaee are terrestrial) might have arisen only once in the lineage with a reversal in Haemodoraceae, or it may have arisen independently in both lineages (Barrett and Graham 1997). The three families also include species with enantiostyly, a unique floral asymmetry polymorphism in which the style is variously deflected to the left or right side of a flower that is relatively rare in angiosperms (Graham and Barrett 1995; Jesson and Barrett 2003). This character has been suggested previously as a possible synapomorphy for this three-family clade (Graham and Barrett 1995; Givnish et al. 1999). Among monocots, enantiostyly occurs in multiple genera in Haemodoraceae, Philydraceae, and Pontederiaceae, but also in the closely related Commelinaceae and the distantly related Tecophilaeaceae (Asparagales) (see Graham and Barrett 1995). Dimorphic enantiostyly, one of two enantiostyly subtypes in which left- and right-styled flowers occur on different individuals as a genetic polymorphism (Barrett et al. 2000), has been positively confirmed in monocots only in Haemodoraceae and Pontederiaceae (Jesson and Barrett 2002a, b; Jesson and Barrett 2003). Given the limited occurrence of dimorphic enantiostyly in both families (Jesson and Barrett 2003), it is possible that it originated in parallel in these lineages. Within the Haemodoraceae-Philydraceae-Pontederiaceae clade, I infer Philydraceae to be the sister-group of Haemodoraceae-Pontederiaceae (Figs. 4.4, 4.5), with moderate support, similar to that found by Chase et al. (2006). All three families have an inner and outer tepaloid whorl, but in Philydraceae, the outer petaloid whorl is substantially smaller than the inner petaloid whorl, whereas in Haemodoraceae and Pontederiaceae, the outer and 151 innter petaloid whorls are similar in size (Soltis et al. 2005). The latter character (equal tepal size) may therefore constitute a synapomorphy for a Haemodoraceae-Pontederiaceae clade. A sister group relationship between Pontederiaceae and Haemodoraceae is further consistent with aspects of their pollen morphology (Simpson 1987). Commelinaceae-Hanguanaceae — In the current analyses I find a strongly supported sister-group relationship between Commelinaceae and Hanguanaceae (Figs. 4.4, 4.5), consistent with the findings of most monocot tree of life work (Givnish et al. 1999, 2006; Chase et al. 2006; but see Tamura 2004). The spiderwort or wandering jew family, Commelinaceae, consists of herbs with cymose inflorescences, short-lived flowers with differentiated perianths (i.e., petals and sepals) and moniliform anthers, and the fruit is a capsule or berry (Dahlgren et al. 1985; Faden 1998; Soltis et al. 2005). Hanguanaceae are herbs with (sometimes broad) petiolate leaves, paniculate inflorescences with sessile, small flowers, and the fruit is a berry (Bayer et al. 1998; Stevens 2006). Commelinaceae were previously included in an order (Commelinales) with Eriocaulaceae, Mayacaceae, Rapateaceae, and Xyridaceae based on their shared petaloid and sepaloid perianths (Dahlgren et al. 1985), but Commelinaceae are not closely related to these families and this character is therefore homoplasious or symplesiomorphic. Hanguana was originally included in Flagellariaceae (Backer 1951), but removed to its own monogeneric family on the basis of differences in vegetative morphology and anatomy (Airy Shaw 1965). It has been variously allied with or placed in taxa that are now included in five orders of monocots, including Asparagales (Dahlgren et al. 1985), Liliales (Cronquist 1981), Poales (Backer 1951), and Zingiberales (Rudall et al. 1999). From a biodiversity perspective, Hanguanaceae is probably the most poorly known family in Commelinales and even commelinid monocots; 152 ongoing research indicates that there are multiple undescribed Hanguana species in Borneo and western Malaysia that need to be evaluated (Bayer et al. 1998; P. M . Boyce, pers. comm.). A sister-group relationship between Commelinaceae and Hanguanaceae was largely unexpected. Only one putative non-DNA synapomorphy (largely non-photosynthetic cotyledons; Tillich 1996; Tillich and Sill 1999) has been identified for a Commelinaceae-Hanguanaceae clade (Stevens 2006). A 5-bp insertion in matK shared by Monochoria (Pontederiaceae) and Hanguana, which Tamura et al. (2004) suggested as a possible synapomorphy for a Hanguanaceae-Pontederiaceae clade, likely is homoplasy. A recent morphological analysis found that Hanguana shares many morphological characteristics with Zingiberales (Rudall et al. 1999); these characters may be symplesiomorphies. Other similarities between Commelinaceae and Hanguanaceae have been observed recently, but it is not clear if these are symplesiomorphies, synapomorphies, or parallelisms. Tillich (1996) noted that an undescribed species of Hanguana (subsequently named H, bogneri; Tillich and Sill 1999) with broad, petiolate leaves, is similar in habit to some Commelinaceae genera, such as Palisota. In both Hanguanaceae and Commelinaceae the seed coats are testal in origin, although this character state occurs in other commelinid families (Tillich 1996). Also, Hanguanaceae and Cartonema - the putative sister-group of the rest of Commelinaceae (e.g., Evans et al. 2000, 2003; Figs. 4.4, 4.5) that has sometimes been recognized in its own family, Cartonemataceae Pichon (e.g., Tomlinson 1969) - have inconspicuous root collars (Tillich 1996). Rudall and Bateman (2004) observed that Hanguana and several species of Cartonema have actinomorphic flowers, and suggested that actinomorphic flowers may have 153 arisen only once in Hanguanaceae and Commelinaceae, and undergone a subsequent reversal in the remainder of Commelinaceae. Philydraceae — Philydraceae are a small monocot family characterized by a perennial, rhizomatous or cormose habit, with basal and distichous ensiform leaves, bracteate inflorescences, indeterminate racemose inflorescences with flowers subtended by spathelike bracts, and zygomorphic flowers with four perianth parts and a single stamen (Simpson 1985; Hamann 1998). Three or four genera are currently recognized in Philydraceae. Generic circumscriptions of Philydrum (one species) and Philydrella (two species) are without controversy [but see Adams (1987) for nomenclatural issues with the name Philydrella Caruel]. Nonetheless, in this study I could not test the monophyly of Philydrella with molecular data because I was not able to sample P. pygmaea, the other species in the genus. The generic circumscription of Helmoltzia has varied previously. The genus Helmholtzia F. Muell. was established in 1865 to include a single species, H. acorifolia F. Muell. A second species, H glaberrima (Hook f.) Caruel, was originally described as a species of Philydrum (P. glaberrimum Hook, f, 1873) but later transferred to Helmholtzia. Skottsberg (1932), however, placed this taxon in a monotypic genus, Orthothylax (Hook, f.) Skottsb. [O. glaberrimus (Hook, f.) Skottsb.], on the basis of several morphological differences with H. acorifolia [free perianth segments (vs. a conspicuous perianth tube in H. acorifolia); zygomorphic pistil (vs. actinomorphic pistil); partly unilocular ovary (vs. trilocular ovary); and dry, loculicidally dehiscent capsule (vs. an ovary that does not split)] (Skottsberg 1934), and he described a second species, H novoguineensis (K. Krause) Skottsb., from Papua New Guinea. Subsequent authors have variously recognized Orthothylax as a distinct genus (e.g., Cronquist 1981; Hamann 1966, 1998; Simpson 1985) or included it within Helmholtzia (e.g., 154 Adams 1987; Prentis et al. 2006). The molecular data presented here indicate that H. acorifolia and H glaberrima are sister taxa, consistent with their possible inclusion in a single genus. The current data are not, however, inconsistent with the recognition of two genera, with the caveat that I am not able to test the monophyly of Helmholtzia s.s., because I have not sampled H. novoguineensis, the only remaining taxon in this species complex. My complete genus-level sampling for Philydraceae indicates that the family is a well-supported natural taxon (Figs. 4.4, 4.5), consistent with the conclusions of a previous morphological investigation (Hamann 1966). Hamann (1966, 1998) recognized two ecologically distinct groups in Philydraceae: one including Helmholtzia and Orthothylax, distributed in montane rainforests; the other including Philydrum and Philydrella, distributed in lowland areas. My data support a close relationship between Helmholtzia and Orthothylax. Philydrella is the sister group of a clade that includes Philydrum and Helmholtzia— Orthothylax (Figs. 4.4, 4.5); the latter relationship receives moderate to strong support. A sister-group relationship of Philydrella with the rest of the family is possibly consistent with several morphological, biogeographical, and ecological differences among Philydrella and Helmholtzia, Orthothylax, and Philydrum. Hamann (1966) observed that Philydrella has stomata with two lateral subsidiary cells that are smaller than other epidermal cells, Philydrum has two or sometimes four subsidiary cells arranged like those in Philydrella, whereas Helmholtzia and Orthothylax have four or more subsidiary cells arranged polarly or laterally, that are generally the same size as other epidermal cells. Additionally, Helmholtzia, Orthothylax, and Philydrum occur in wet habitats mostly in eastern Australasia, and they are each characterized by distichous, isobifacial-ensiform leaves, rhizomes, long woolly hairs on the flowering stem and bracts, and inflorescences of a terminal spike with 155 many flowers and lateral spikes (Hamann 1966, 1998). In contrast, Philydrella occurs in the drier climate in western Australia, and is characterized by a single terete leaf with an acuminate blade, corrns, glabrous flowering stems and bracts, and a terminal spike with few flowers and no lateral spikes (Hamann 1966, 1998). Simpson (1985) noted that pollen grains of Helmholtzia, Orthothylax and Philydrum have a reticulate nonapertural wall sculpting, whereas Philydrella differs in having a regulate nonapertural sculpturing. This minor character may constitue an additional synapomorphy for a Helmholtzia-Orthothylax-Philydrum clade. The complete genus-level phylogeny for the family determined here should provide a framework to test hypotheses of evolution in Philydraceae (P. Prentis, pers. comm.). 4.4.3. Higher-Order Relationships in Poales Poales sensu APG II (2003) is strongly supported as monophyletic in my analyses (Figs. 4.4, 4.5). I find moderate to strong support here for a major lineage that includes all Poales families except Bromeliaceae and Typhaceae (Figs. 4.4, 4.5); the positions of these latter two families at the base of the Poales subtree are not resolved here. Rapateaceae are weakly inferred to be the sister group of the rest of this large clade. These findings are generally consistent with three recent studies (Chase et al. 2006; Givnish et al. 2006; Graham et al. 2006), although the latter study did not include Rapateaceae. Most previous studies did not recover this 14-family clade at all, or recovered it with only weak support (e.g., Chase et al. 2000; Michelangeli et al. 2003; Davis et al. 2004). Despite the increasingly strong molecular evidence for this clade, no non-DNA synapomorphies for it are known at present (Soltis et al. 2005). Within this major Poales lineage, I infer a xyrid clade with differing 156 compositions in parsimony vs. likelihood analyses, a cyperid clade, and the core Poales which consists of the graminid and restiid clades. Relationships among the core Poales, the xyrids, and the cyperids are not resolved in all analyses here, and it is not clear if Mayacaceae belongs to the xyrid clade. 4.4.3.1. The Core Poales The core Poales (Anarthriaceae, Centrolepidaceae, Ecdeiocoleaceae, Flagellariaceae, Joinvilleaceae, Poaceae, and Restionaceae) have long been considered to be closely related. These families constituted Poales in Dahlgren et al's (1985) classification, and they form a clade in most recent analyses (e.g., Bremer 2002; Michelangeli et al. 2003; Davis et al. 2004; Givnish et al. 2006; Chase et al. 2006). I find strong support here (Figs. 4.4, 4.5) for the core Poales, an improvement on the moderately strong support found for the clade by Chase et al. (2006). Several putative synapomorphies for a core Poales clade have been identified, including distichous leaves with a (usually) open sheath, stomata with dumbbell-shaped guard cells (although this is not the case in Flagellaria; Sack 1994), monoporate pollen with an annular ring, apical placentation, and nuclear endosperm development (e.g., Kellogg and Linder 1995; Soltis et al. 2005). 4.4.3.2. The G r a m i n i d Clade Within the core Poales, I identify a strongly supported graminid clade that includes Ecdeiocoleaceae, Flagellariaceae, Joinvilleaceae, and Poaceae (Figs. 4.4, 4.5). This Ecdeiocoleaceae-Joinvilleaceae-Poaceae clade has been recovered with strong support in several molecular analyses (e.g., Bremer 2002; Michelangeli et al. 2003; Davis et al. 2004; 157 Chapter 3), and the monophyly of this three-family group is further supported by a shared 6-kb plastid inversion that is not found in other Poales lineages (e.g., Doyle et al. 1992; Katayama and Ogihara 1996; Michelangeli et al. 2003). Despite strong support for this clade, relationships among these three families have been difficult to resolve unambiguously. Identification of the sister-group of the grasses is critical to homologize Poaceae reproductive structures with those of their closest relatives (Rudall et al. 2005; see Chapter 3), and to facilitate reconconstruction of the sequence of events that lead to the origin of the highly modified grass spikelet. Several recent studies have inferred Ecdeiocoleaceae to be the sister-group of the grasses, although this topology has generally received only weak support (e.g., Bremer 2002; Michelangeli et al. 2003; Davis et al. 2004). Here, I also identify Ecdeiocoleaceae as the sister-group of the grasses. Although I find only weak support for an Ecdeiocoleaceae-Poaceae clade (Figs. 4.4, 4.5), this level of support is greater than that inferred for this clade in most previous studies (e.g., Bremer 2002; Michelangeli et al. 2003; Davis et al. 2004). In a previous study I identified an Ecdeiocoleaceae-Joinvilleaceae clade with weak support from parsimony analyses and moderate support from model based analyses, particularly when protein-coding data were considered alone (Chapter 3). An Ecdeiocoleaceae-Joinvilleaceae clade here receives only 11 and 27% bootstrap support from parsimony and likelihood analyses, respectively (data not shown). Increased support for an Ecdeiocoleaceae-Poaceae clade here compared with my earlier study and other previous work is likely based on a combination of more data and increased taxon sampling in the former family (I have sampled both genera), which has helped break up the relatively long terminal Ecdeiocolea branch (see Chapter 3). It is possible that with even more data per 158 taxon, it will be possible to achieve strong support for this clade (if it is true), and unambiguously identify the sister-group of the grasses. Possible non-DNA synapormorphies for an Ecdeiocoleaceae-Poaceae clade include scrobiculate pollen (Campbell and Kellogg 1987; Michelangeli et al. 2003) and operculate pollen pores (Michelangeli et al. 2003). The presences of six anthers may also constitute a synapomorphy for this clade (Rudall et al. 2005). Some 'early-diverging' grasses have six anthers (e.g., Anomochlooideae; GPWG 2001), and one genus (Georgeantha) of Ecdeiocoleaceae has this character (Ecdeiocolea has four anthers) (Rudall et al. 2005). Flagellariaceae — Flagellariaceae are a small, paleotropical family of rattan-like climbers with dichotomous branching, leaves terminated in tendrils, and large, paniculate inflorescences with perfect flowers (Stevens 2006). In its original circumscription, Flagellariaceae included the superficially similar genera Flagellaria, Hanguana, and Joinvillea (e.g., Backer 1951), but the latter two genera have since been segregated to their own families based on differences in vegetative morphology and anatomy (Airy Shaw 1965, Tomlinson and Smith 1970). Although Flagellariaceae have most characters typical of core Poales (see above), they have been difficult to place in core Poales (e.g., Takhtajan 1997) because they are unique in several aspects. Autopomorphies for Flagellariaceae include embryo-sac development of the Allium type, binucleate pollen, crassinucellate ovules, multicellular stigmatic papillae, and a unique guard-cell morphology (e.g., Campbell and Kellogg 1987; Sack 1994; Appel and Bayer 1998). In molecular analyses, Flagellariaceae have been variously placed as the sister group of Ecdeiocoleaceae-Joinvilleaceae-Poaceae (e.g., Chase et al. 2000, 2006; Givnish et al. 2006), the sister group of the rest of core Poales 159 (e.g., Michelangeli et al. 2003; Graham et al. 2006), or the sister group of the xyrid clade (Davis et al. 2004). Here I infer Flagellaria to be part of the graminid clade; they are the sister-group of Ecdeiocoleaceae-Joinvilleaceae-Poaceae (Figs. 4.4, 4.5), with moderate to strong support, similar to that found by Chase et al. (2006) for the same position. Placement of Flagellariaceae as the sister-group of a Restionaceae-Ecdeiocoleaceae-Poaceae clade by Graham et al. (2006), who considered a subset of the taxa data that I examine here, may be an artifact of incomplete taxon sampling; they included only four of the seven core Poales families, mostly with single exemplars. A sister-group relationship between Flagellariaceae and Ecdeiocoleaceae-Joinvilleaceae-Poaceae is consistent with the presence of a 28-kb plastid inversion found in these four families (although this inversion also occurs in some Restionaceae) (Doyle et al. 1992; Michelangeli et al. 2003). Flagellariaceae lacks the 6-kb inversion shared by Ecdeiocoleaceae-Joinvilleaceae-Poaceae (Doyle et al. 1992; Michelangeli et al. 2003). Inclusion of Flagellariaceae in the graminid clade might also be consistent with a microstructural cell-wall characteristic. Poaceae and Flagellariaceae have a similar cell-wall-labelling pattern of (1—>3), (1—>4)-B-D-glucans that is distinct from the pattern observed in other major Poales lineage, although only a few Poales families have been examined for this character (Trethewey et al. 2005). Ecdeiocoleaceae — Ecdeiocoleaceae are strongly xeromorphic herbs with leaves reduced to sheaths, spicate inflorescences with male and female flowers, and fruits that are achenes or nutlets (Briggs 2005). The tussocky cord rush, Ecdeiocolea monostachys, was previously included in Restionaceae, but subsequently segregated to its own family on the basis of multiple vegetative and reproductive characteristics (Cutler and Shaw 1965). 160 Ecdeiocoleaceae were thought previously to be closely related to Restionaceae (e.g., Dahlgren et al. 1985; Manning and Linder 1990; Kellogg and Linder 1995; Stevenson and Loconte 1995) largely based on their superficially similar habit, but it is now clear that they are related more distantly (see above). Similarity between these two families may be a convergence in response to their xeromorphic habitat. A second genus and species of Ecdeiocoleaceae, Georgeantha hexandra, was described only recently, differing from E. monostachya in its rhizomatous habit, culms with multiple internodes, branched inflorescences with up to three spikelets (each with five to eleven flowers), tricarpellate ovules, six stamens, and loculicidally dehiscent fruits (Briggs and Johnson 1998; Briggs 2005; Rudall et al. 2005). By contrast, E. monostachya has a caespitose habit, no culm internodes, unbranched inflorescences (with up to 50 flowers in a single spikelet), unicarpellate ovules, four stamens, and indehiscent fruits (Briggs and Johnson 1998; Rudall et al. 2005). The two genera are further distinguished by unique flavonoid patterns (Williams et al. 1997). Both species have been included in two molecular analyses (Briggs et al. 2000; Bremer 2002) which robustly supported the monophyly of Ecdeiocoleaceae. In the current study, I also identify an Ecdeiocolea-Georgeantha clade, with strong support (Figs. 4.4, 4.5). I observe substantial plastid variation between these two taxa that is possibly correlated with the strong morphological differences between the genera. A second undescribed species of Ecdeiocolea from western Australia has recently been discovered (B. G. Briggs, personal communication). This unnamed taxon differs from E. monostachya in having a more elongated rhizome, more spaced-out culms, and a generally stouter form (B. G. Briggs, unpublished data). The new taxon appears to be much less widespread than E. monostachya, but over most of its range the two species are extremely 161 common and grow together with no apparent ecological distinction (B. G. Briggs, unpublished data). There is currently no molecular data for this new species, and it should be a priority for inclusion in future studies to test the monophyly of Ecdeiocolea and Georgeantha, and to further characterize levels of plastid variation among these taxa. Inclusion of data for this species may also provide additional insight into relationships among Ecdeiocoleaceae-Joinvilleaceae-Poaceae. Joinvilleaceae — The reed-like joinvillea family, Joinvilleaceae, which includes robust, grass-like plants with strongly plicate leaves and terminal panicles of small, perfect, flowers (Stevens 2006), were thought previously to be the sister group of Poaceae. In the current study, this relationship receives only 25 and 11% bootstrap support from parsimony and likelihood analysis, respectively (data not shown), and is clearly not favoured. One possible synapomorphy for a Joinvilleaceae-Poaceae clade that has been hypothesized is the presence of long and short cells in the leaf epidermal cells of both families (e.g., Campbell and Kellogg 1987). Long and short epidermal cells, however, have also been noted in a few palm (Arecaceae) and sedge (Cyperaceae) genera (Tomlinson 1969, Metcalfe 1971), indicating that this character may be a symplesiomorphy, or that it may have originated multiple times across Poales. Michelangeli et al. (2003) also recently reported the presence of long and short epidermal cells in Ecdeiocoleaceae. It may therefore unite Ecdeiocoleaceae, Joinvilleaceae, and Poaceae. Within Joinvilleaceae, I have sampled Joinvillea plicata and J. ascendens, the two currently recognized species in the monogeneric family (Newell 1969). I detect minimal plastid variation between these two species (Figs. 4.4, 4.5). The low level of variation is 162 possibly consistent with the limited morphological variation that distinguishes them; they differ only in tepal shape and the frequency of bristles on the laminar surface (Newell 1969). P o a c e a e — The grass family - the largest Poales family, with ~11,000 species (Table 4.1) - has received substantial systematic attention, and its general phylogenetic structure is fairly well established (see GPWG 2001; Chapter 3). The most recent classification of the family includes twelve subfamilies (GPWG 2001; Soreng et al. 2006; Duvall et al. 2006). In my previous study, I examined phylogenetic relationships among ten of these major lineages with the same complement of plastid data examined here. Here I infer the same relationships found in the previous study, generally with similar levels of support. In this study I have added data for a species of Puelia (subfamily Puelioideae), one of two major lineages not included in my previous study, and three additional exemplar taxa from available plastid genomes [Oryza nivara (Ehrhartoideae); Saccharum officinarum and Sorghum bicolor (Panicoideae)]. I have not sampled the recently recognized subfamily Micrairoideae (Duvall et al. 2006). The previously unsampled subfamily, Puelioidoeae, is inferred here with strong support (Figs. 4.4, 4.5) to be the third of three successive lineages (i.e., including Anomochlooideae and Pharoideae) that are the respective sister groups of the rest of the grasses). This placement is consistent with most earlier studies that have sampled Puelioideae (e.g., Clark et al. 2000; GPWG 2001), although one study placed Puelia with moderate support as the sister group of the BEP clade (Zhang 2000). A Puelioideae-BEP clade here receives < 5% bootstrap support from parsimony and likelihood analyses (data not shown). Concerning the other taxa added in comparison to Chapter 3 (i.e., Oryza nivara, Saccharum, and Sorghum), I detect only minimal plastid variation among wild and cultivated 163 rice, the two sampled Oryza species, consistent with previous plastid genome comparisons of these taxa (Masood et al. 2004). In the P A C M A D clade, my within-subfamily taxon sampling is most dense in Panicoideae, where I sampled four exemplar taxa. I find Chasmathium [currently included in tribe Centotheceae Ridl. (Soreng et al. 2006), but formerly included in a separate subfamily, Centothecoideae (GPWG 2001)] to be the sister group of a clade that includes Zea, Saccharum, and Sorghum (these latter genera are currently included in various subtribes in tribe Andropogoneae Dumort.; Soreng et al. 2006). This strongly supported set of relationships is consistent with other recent studies of the subfamily (e.g., Spangler et al. 1999; GPWG 2001; Mathews et al. 2002). Other higher-level relationships within the P A C M A D clade receive very weak support here (Figs. 4.4, 4.5), consistent with most of my previous parsimony and likelihood analyses (Chapter 3). 4.4.3.3. The Restiid Clade The remaining families in the core Poales (Anarthriaceae, Centrolepidaceae, and Restionaceae) are strongly supported here as the restiid clade (sensu Linder and Rudall 2005), as previous studies have found (e.g., Bremer 2002; Chase et al. 2006; Michelangeli et al. 2003; Davis et al. 2004; Givnish et al. 2006). Putative non-DNA morphological synapomorphies for this clade include dioecy (although this is also found in other Poales lineages, such as Cyperaceae and Poaceae), chlorenchyma with peg cells, and dorsifixed anthers (Soltis et al. 2005). Restionaceae and Anarthriaceae — The cape-reed or restio family, Restionaceae, are xeromorphic herbs with reduced leaves, and small, imperfect flowers aggregated into spikelets (Stevens 2006). Restionaceae are the largest family in the restiid clade, and second 164 largest family among core Poales (Table 4.1). Family-level circumscriptions of Restionaceae have been controversial. The genus Anarthria was originally included in Restionaceae, but placed in its own family, Anarthriaceae, on the basis of multiple anatomical and reproductive differences (Cutler and Airy Shaw 1965; Linder and Rudall 1993). Previous molecular studies have placed Anarthria as the sister-group of Restionaceae (e.g., Briggs et al. 2000; Bremer 2002; Michelangeli et al. 2003; Chase et al. 2006), consistent with its placement in its own family. Two additional genera, Hopkinsia and Lyginia, were traditionally included in Restionaceae, but have long been recognized as abberant in that family based on multiple morphological and embryological differences (Briggs and Johnson 2000). Lyginia and Hopkinsia also have unique flavonoid patterns with respect to each other, Anarthria, and other Restionaceae (Williams et al. 1997). Briggs and Johnson (2000) proposed the new families, Hopkinsiaceae and Lyginiaceae, for these genera. Molecular studies indicate that Hopkinsia and Lyginia form a clade with Anarthria (Briggs et al. 2000; Bremer 2002), and the three genera are now included in Anarthriaceae (APG II 2003). Here I also strongly infer a sister group relationship between Hopkinsia and Lyginia, and these genera are the sister group of Anarthria (Fig. 4.4, 4.5). A possible non-DNA synapomorphy for Anarthriaceae s. 1. is tetrathecate anthers (Linder and Rudall 1993). Centrolepidaceae — The family status of Centrolepidaceae, a small Australasian group of dwarf, aquatic annuals or cushion-forming perennials with minute, spike-like bi- or unisexual inflorescences comprised of sterile bracts that surround highly reduced unisexual flowers (Cooke 1998), has also been controversial. Centrolepidaceae are morphologically similar to Restionaceae in several respects, and the two have been considered closely related (e.g., Dahlgren and Clifford 1982). The superficially similar genera Hydatella and Trithuria 165 were previously included in Centrolepidaceae, but are now included in Hydatellaceae (Hamann 1976; see Chapter 5). Workers have variously suggested that Centrolepidaceae should be included in Restionaceae (e.g., Hamann 1975) or that these families should not be considered closely related (e.g., Cutler 1969; Linder and Ferguson 1985). In previous molecular analyses, Centrolepidaceae have been included in Restionaceae (e.g., Briggs 2000; Bremer 2002), or inferred to be the sister group of Restionaceae (e.g., Michelangeli et al. 2003; Chase et al. 2006). Here, I have sampled only a single exemplar from Restionaceae [Elegia, representing an African clade, one of two major clades in the family; Linder et al. 2003], thus I am unable to test the monophyly of Restionaceae with respect to Centrolepidaceae. Nonetheless, I infer the two Centrolepidaceae taxa that I have sampled to be a clade that is the sister group of this single exemplar (Figs. 4.4, 4.5). Broader sampling in Restionaceae, including representatives from the Australian clade, is a priority for future work. Both Centrolepidaceae exemplars sampled here appear to have a substantially elevated rate of plastid evolution compared with Anarthriaceae and Elegia Restionaceae. 4.4.3.4. The Cyperid Clade My analyses identify a well-supported cyperid clade that includes Cyperaceae, Juncaceae, and Thurniaceae, consistent with the findings of earlier studies (e.g., Chase et al. 2000, 2006; Michelangeli et al. 2003; Davis et al. 2004; Givnish et al. 2006). Several synapomorphies have been identified for this three-family lineage, including solid stems with tristichous leaves, pollen shed in tetrads, and chromosomes with diffuse centromeres (Plunkett et al. 1995; Munro and Linder 1998; Soltis et al. 2005). Within the cyperid clade, I find Thurniaceae to be the sister group of a Cyperaceae-Juncaceae clade. Close affinities 166 between Cyperaceae and Juncaceae have been hypothesized previously (e.g., Dahlgren et al. 1985; Thorne 1992a, b), and several potential synapomorphies for this clade have been identified (e.g., absence of calcium oxalate raphides, simultaneous microsporogenesis, usually paracytic stomata; Simpson 1995). In parsimony and likelihood analyses, the cyperid clade is very weakly supported as the sister group of a large clade that includes the core Poales and Eriocaulaceae, Mayacaceae, and Xyridaceae (Figs. 4.4, 4.5). This relationship contrasts with several recent studies (Chase et al. 2006; Givnish et al. 2006; Graham et al. 2006) that found moderate support for a clade that includes the cyperid and xyrid families. Precise placement of the cyperid clade among Poales will likely require additional data. Thurniaceae — The South African genus Prionium, which occurs on stream and river margins (Munro and Linder 1998), was previously included in Juncaceae, although it has long been recognized as aberrant in that family because of its woody, aerial rhizomes and unique leaf anatomy (e.g., Dahlgren et al. 1985; Balslev 1998). Cutler (1969) suggested that Prionium should not be included in Juncaceae, but he did not formally remove it from the family. The first molecular evidence for Prionium inferred the genus to be the sister group of Cyperaceae-Juncaceae (e.g., Chase et al. 1993; Plunkett et al. 1995; Munro and Linder 1998), and it was therefore recently segregated to its own monogeneric family, Prioniaceae (Munro and Linder 1998). The small South American genus, Thurnia, is a rhizomatous, sedge-like herb with small, tepaloid flowers arranged in a dense inflorescence, and subulate-pointed seeds (Dahlgren et al. 1985). Thurnia has been variously included in Rapateaceae or Juncaceae (see Cutler 1965), but it was placed in its own family on the basis of distinct leaf anatomy, in addition to seed and embryo morphology (e.g., Cutler 1965; Dahlgren et al. 167 1985; Kubitzki 1998), with suggested affinities to Cyperaceae and Juncaceae (e.g., Cutler 1969; Cronquist 1981) and possibly Rapateaceae or Xyridaceae (Dahlgren et al. 1985). In some morphological analyses, Thurnia has been placed as the sister group of Juncaceae and Prionium (e.g., Stevenson and Loconte 1995; Rudall et al. 1999), and in others, Prionium and Thurnia are inferred to be sister taxa nested within Juncaceae (Simpson 1995). Recent molecular studies place Thurnia and Prionium in a clade that is the sister group of Cyperaceae-Juncaceae (e.g., Michelangeli et al. 2003; Davis et al. 2004; Givnish et al. 1999, 2006). Because these genera are closely related and morphologically similar (i.e., both genera have serrate leaf margins, erect stems and pentacyclic inflorescences; Stevens 2006) and because they are monogeneric sister taxa, Prioniaceae was included in Thurniaceae in the most recent angiosperm classification (APG II 2003). Consistent with this recent work, I also find a strongly supported sister group relationship between Prionium—Thurnia, and find Thurniaceae to be the sister-group of Cyperaceae-Juncaceae (Figs. 4.4, 4.5). Cyperaceae — The sedge family, Cyperaceae, is a large and morphologically diverse group characterized by usually three-angled stems, reduced flowers surrounded by sterile bracts, and achene fruits (Stevens 2006). The most recent classification of the family recognizes two subfamilies (Mapanioideae and Cyperoideae) and twelve tribes (Simpson et al. 2006). Here, I have sampled representatives from each subfamily and four tribes in Cyperoideae. The relationships that I infer among these (Figs. 4.4, 4.5) are consistent with relationships found in previous studies (e.g., Muasya et al. 1998, 2000; Givnish et al. 2006; Simpson et al. 2006), although my sampling is currently much sparser. 168 4 . 4 . 3 . 5 . The Xyrid Clade Support levels and family composition for a xyrid clade are different in maximum parsimony and maximum likelihood analyses here, because of conflicting placements for Mayacaceae. In my parsimony analysis, I find a weakly supported xyrid clade (Fig 4.4) that includes Eriocaulaceae, Mayacaceae, and Xyridaceae; the relationships of this lineage with other Poales is not clear (Fig. 4.4). In contrast, in my likelihood analysis, I find a moderately supported xyrid clade that includes only Eriocaulaceae and Xyridaceae (Fig. 4.5), and this lineage is strongly supported as the sister group of the core Poales (Fig. 4.5); Mayacaceae are not part of this clade (Fig. 4.5). The aquatic bogmoss family, Mayacaceae, has a Lycopodium-like habit, and is characterized by spirally arranged leaves, solitary actinomorphic flowers, a showy perianth, and tricarpellate ovaries (Faden 1985; Venturelli and Bouman 1986). By contrast, the pipewort family, Eriocaulaceae, are characterized by a tufted basal rosette of leaves with capitula arranged in racemes (Unwin 2004), and Xyridaceae are petaloid monocots with variously arranged equitant leaves, a bracteate conelike-spike or capitulum-like inflorescence situated on a scape, and a two-whorled, tripartite perianth (Krai 1998). Mayacaceae have been variously included in a clade with Eriocaulaceae-Xyridaceae (e.g., Michelangeli et al. 2003), Bromeliaceae-Rapateaceae (Givnish et al. 1999), or placed as the sister group of Xyridaceae (e.g., Bremer 2002), a cyperid clade (e.g., Chase et al. 2006), or a xyrid-cyperid clade (e.g., Givnish et al. 2006). In parsimony analysis, I find Mayacaceae and Eriocaulaceae to be moderately supported as sister taxa (Fig. 4.4), and Xyridaceae to be their sister group. No previous study has inferred the Eriocaulaceae-Mayacaceae clade that I find here. Most studies place Eriocaulaceae and Xyridaceae in a robustly supported clade [e.g., Chase et al. 2006; Givnish et al. 2006; but see 169 Michelangeli et al. (2003) and Davis et al. (2004)]. In contrast to my parsimony analysis, I find this latter relationship from maximum likelihood analysis with moderate support (Fig. 4.5). Several morphological synapomorphies have been identified for an Eriocaulaceae-Xyridaceae clade, including a differentiated perianth, a basal rosette growth habit, paracytic stomata, dimerous flowers, ovules with thin-walled megasporangia, and spinulate/ echinate pollen (Dahlgren et al. 1985; Soltis et al. 2005). However, if Eriocaulaceae and Mayacacae are in fact sister taxa, and Xyridaceae are their sister group, then these putative Eriocaulaceae-Xyridaceae synapomorphies might actually be deeper symplesiomorphies. Alternatively, they may represent synapomorphies for a broader Eriocaulaceae-Mayacaceae-Xyridaceae clade, but this hypothesis necessitates substantial subsequent morphological change in the Mayaca lineage. Difficulty inferring relationships clearly and strongly among the xyrid families is likely a function of their substantially accelerated plastid substitution rates, particularly in Mayacaceae. In the current study, branches for Eriocaulaceae, Mayacaceae, and Xyridaceae are very long; three of the four exemplars {Eriocaulon, Mayaca, and Xyris) are the longest among Poales taxa examined here (Figs 4.4, 4.5). It is well known that bootstrap analysis might be misleading for taxa with long branches (e.g., Felsenstein 1978; Hendy and Penny 1989), thus long branch attraction might be distorting inference of relationships among Eriocaulaceae, Mayacaceae, and Xyridaceae. The possibility of a long-branch artifact here is further substantiated by the moderately conflicting topologies inferred from parsimony and likelihood analyses. Likelihood analyses are known to be generally less prone to long- branch artifacts compared with parsimony analyses (e.g., Swofford et al. 2001; Sanderson and Shaffer 2002). Given previous difficulties placing Mayacaceae among Poales (e.g., Bremer 170 2002), Givnish et al. (2006) suggested that this taxon might act like a 'wild-card' in phylogenetic analyses; the current analyses uphold this hypothesis. Now that data are available for Eriocaulaceae, Mayacaceae, and Xyridaceae, as well as related families in Poales, future work could use a simulation approach (e.g., Huelsenbeck 1997; Sanderson et al. 2000) to determine if difficulty in placing Mayacace is a function of systematic bias (i.e., long branch attraction). Increased taxon sampling in each family may also address this issue, by dividing long branches. For Mayacaceae, most phylogenetic studies have included sequence data only for Mayaca fluviatilis, the widest ranging species in the genus (Lourteig 1952); no studies have included more than one Mayaca species, and levels of plastid variation among the four or so species in the family are unknown. Recent morphological and molecular studies of Eriocaulaceae have identified two major clades in the family, recognized as subfamilies Eriocauloideae (two genera) and Paepalanthoideae (multiple genera) (Giulietti et al. 1995; Unwin 2004). I have included only a single representative for this family here, from the generally aquatic subfamily Eriocauloideae (Unwin 2004). Future work should sample additional Mayaca and Eriocauloideae species, and also sample from from the morphologically diverse and generally terrestrial Paepalanthoideae. Xyridaceae — The monophly of Xyridaceae s. 1. has been questioned in several studies. Most species in Xyridaceae are in the widespread genus Xyris, and one or a few species are in Abolboda Humb. & Bonpl., Achlyphila Maguira & Wurdack, Aratitiyopea Steyerm. & Berry, and Orectanthe Maguire (see Krai 1988, 1992, 1998; Campbell and Stevenson 2005). Xyridaceae are morphologically diverse, and some authors have placed Abolboda, Aratitiyopea, and Orectanthe in a segregate family, Abolbodaceae (e.g., A P G 171 1998). Achlyphila, however, is morphologically intermediate between Xyris and Abolbodaceae; thus, most authors recognize only a single family (e.g., Dahlgren et al. 1985; Krai 1992, 1998). Nonetheless, some phylogenetic analyses indicate that the family is not monophyletic (e.g., Michelangeli et al. 2003; Davis et al. 2004), and others support Xyridaceae s. 1. as a natural unit (e.g., Bremer 2002; Givnish et al. 1999, 2006; Chase et al. 2006), though sometimes with only weak support. In a preliminary study of morphological characters, Rudall and Sajo (1999) were unable to resolve relationships among Xyridaceae genera. They noted substantial variation in several embryological characters, and suggested that recognition of two families may be justified. In a more comprehenseive morphological study, Campbell (2004) inferred two major clades in Xyridaceae: one incuding Xyris (subfamily Xyridoideae s. s.), the other including Achlyphila (traditionally included in subf. Xyridoideae) and subfamily Abolboideae (Aratitiyopea—Orectanthe and Abolboda). In the current study I have sampled a single exemplar from each of these two major groups (i.e., Aratitiyopea and Xyris), and I recover strong support for a monophyletic Xyridaceae (Figs. 4.4, 4.5). Nonetheless, given the lack of monophyly for Xyridaceae in some studies and limited taxon sampling in the family in most studies (including the current one), extensive sampling of all five Xyridaceae genera, potentially including multiple species from Xyris and Abolboda (the two largest genera in the family; Krai 1988, 1992), should be carried out in future work, to further test the monophyly of the family and the phylogenetic hypotheses presented by Campbell (2004). At present, Xyridaceae are the largest family in Poales for which no family-wide phylogenetic hypotheses based on molecular data exist. 172 4.4.3.6. The 'Basal' Poales Here I find Rapateaceae to be moderately supported as part of a clade that includes all of Poales except Bromeliaceae and Typhaceae (Figs. 4.4, 4.5), consistent with the recent analyses of Chase et al. (2006) and Givnish et al. (2006). Rapateaceae are a family of terrestrial, epiphytic, or amphibious herbs characterized by aerial stems with unifacial or bifacial leaves, and pedunculate inflorescences with petaloid flowers (Stevenson et al. 1998). Their superficial similarities to Eriocaulaceae and Xyridaceae are likely parallelisms (Soltis et al. 2005). No non-DNA synapomorphies are known for the large clade that includes Rapateaceae and the rest of Poales except Bromeliaceae and Typhaceae. Bromeliaceae and Typhaceae have been consistently placed near the base of the Poales subtree, but their relationship to each other and the rest of Poales is not clear (e.g., Bremer 2002; Michelangeli et al. 2003; Chase et al. 2000, 2006; Graham et al. 2006; Givnish et al. 1999, 2006). The cattail family, Typhaceae, includes temperately-distributed rhizomatous herbs with erect leaves, and densely cylindrical inflorescences with reduced staminate and pistillate flowers. Typhaceae have been variously circumscribed to include or exclude the bur-reed genus, Sparganium (Sparganiaceae) (e.g., Cronquist 1981; Thorne 1992a, b; APG 1998; APG II 2003; Chase 2004). Typha and Sparganium are ecologically, cytologically, embryologically, and morphologically similar (e.g, Thieret and Luken 1996), and in most molecular studies (including the current study; Figs. 4.4, 4.5) they are a clade (e.g., Givnish et al. 1999, 2006; Graham et al. 2006; Chase et al. 2006), consistent with their inclusion in a single family. By contrast, the bromeliad family, Bromeliaceae, is a morphologically, physiologically, and ecologically variable neotropical group characterized 173 by a (usually) basal rosette of leaves, often coloured inflorescence bracts, and a petaloid perianth. Here, Bromeliaceae and Typhaceae are a grade at the base of Poales, with either Bromeliaceae or Typhaceae very weakly inferred to be the sister group of the rest of Poales (Figs. 4.4, 4.5). By contrast, a Bromeliaceae-Typhaceae clade has been inferred with weak to moderate support in several studies (e.g., Bremer 2002; Chase et al. 2006; Givnish et al. 2006); here, a Bromeliaceae-Typhaceae clade receives only 16 and 37% bootstrap support from parsimony and likelihood analyses (data not shown). Regardless of their exact branching orders, 'deep' positions for Bromeliaceae and Typhaceae in Poales are consistent with cell-wall structural data. All sampled Poales families except Bromeliaceae and Typhaceae have (1—>3), (1—>4)-B-D-glucans in their cell walls (Trethewey et al. 2005). Given the difficulties observed here and elsewhere for inferring the exact positions of Bromeliaceae and Typhaceae at the base of Poales, additional sampling of plastid regions and possibly taxa may help clarify their positions. In this study I have sampled two exemplars that likely span the root of Bromeliaceae (e.g., Terry et al. 1997; Givnish et al. 2006), but given the low levels of plastid variation observed across Bromeliaceae (e.g., Givnish et al. 2006), it seems unlikely that further taxon samplings from the family will substantially affect inference of its position in Poales. Overall levels of plastid variation in Typhaceae are not known, as few species have been included in molecular studies. However, relatively high morphological variation in Typha (i.e., viz. T. latifolia L. vs. T. minima Hoppe.) and Sparganium (see Cook and Nichols 1986, 1987) suggests the possibility of additional molecular variation that might be useful in placing the family. 174 Few non-DNA characters are known that support the possible branching orders for Bromeliaceae and Typhaceae at the based of Poales. A close affinity between these families was hypothesized by Dahlgren et al. (1985), based on the presence of helobial endosperm in both families, but this character also occurs in Commelinales, Juncaceae, and Zingiberales (Linder and Kellogg 1995). Both Bromeliaceae and Typhaceae lack the mitochondrial gene sdhA (Adams et al. 2002), and some authors have suggested that this loss might represent a synapomorphy for a Bromeliaceae-Typhaceae clade (Soltis et al. 2005; Stevens 2006). However, this should be viewed with caution, as sdh and ribosomal protein genes in the mitochondrial genome have been lost across the angiosperms multiple times (Adams et al. 2002; Ong and Palmer 2006). 4.4.4. Molecular Evolution in the Commelinid Monocots Poaceae have been known to have an accelerated plastid substitution rate among monocots since some of the earliest plant molecular phylogenetic studies (Gaut et al. 1992), but more recent work has demonstrated that this accelerated substitution also occurs in several closely related Poales lineages (e.g., Bremer 2002; Givnish et al. 2006; Graham et al. 2006; Chapter 3). Here I have sampled representatives from all Poales families. Examination of branch variation in the order suggests that all families except Bromeliaceae, Rapateaceae, and Typhaceae (the basal Poales), and possibly Flagellariaceae, share this rate acceleration to some extent (Figs. 4.4, 4.5). Branch lengths in most Poales are substantially longer than those observed here in most other monocots, most eudicots, and basal angiosperms (Fig. 4.1). Within this 'rate-accelerated' Poales clade, there is also evidence of substantial variation in branch lengths among examined taxa (Chapter 3). Flagellariaceae 175 and Joinvilleaceae have shorter branches, whereas those for Centrolepidaceae, Eriocaulaceae, Mayacaceae, and Xyridaceae are substantially longer. The implications of these long branches for phylogenetic analyses are not evident. Although strong concordance for most Poales branches in parsimony and likelihood analyses here suggests that long branches are not causing misinference of relationships, discordance in the placement of Mayacaceae between parsimony and likelihood analyses (see above) suggest that they may do so, at least with respect to this taxon. Various explanations have been proposed for plastid rate heterogeneity among plants, including one, or a combination of, a relaxation of purifying selection, demographic factors, and variable mutation rates among lineages (e.g., due to changes in DNA repair efficiency that may affect whole genomes or parts of them) (e.g., Cho et al. 2004; Parkinson et al. 2005; Young and dePamphilis 2005). Large-scale sequencing projects involving multiple taxa and genomic regions have often provided insight into molecular evolution of particular loci. For example, here I have detected an intronless version of the inverted repeat 3'-r/«12 locus in Luzula (Juncaceae). The V-rpsM intron is present in most land plant taxa that have been examined so far (e.g., Graham and Olmstead 2000a, b, 2006; McPerson 2003; Rai et al. 2003; Zgurski 2004), but multiple, putatively independent losses for this locus have been documented in several taxa in Fabaceae (Graham et al. 2000; Lin et al. unpub. data, GenBank number: NC_003119.6), monilophytes (H. S. Rai et al. unpub. data), Cuscuta europaea L. (Convolvulaceae; Freyer et al. 1995), Anomene L. (Ranunculaceae; Hoot and Palmer 1994), and the Asparagales families Asphodelaceae and Hemerocallidaceae (McPherson et al. 2004). An intronless V-rpsM in Luzula is the first report for this loss in monocots outside Asparagales (McPherson et al. 2004), but it is possible that further monocot sampling will reveal additional instances for it. 176 The V-rps\2 intron is present in Juncus, the other Juncaceae exemplar examined here, indicating that the intron loss is not characteristic of all of Juncaceae, although it might be a synapomorphy for the putatively monophyletic Luzula or broader clades that also include members of the non-monophyletic Juncus (Drabkova et al. 2003; Roalson et al. 2005). Future work should more thoroughly explore and characterize the distribution of the V-rps\2 intron in Juncaceae. 177 Table 4.1. Families recognized currently in Commelinales and Poales (sensu A P G II2003), their generic and species level diversity, and distribution. Indented families have been recognized recently in some treatments, but are included in the previous taxon by A P G II (2003). ' Order and Family No. Genera/ Species Distribution Reference(s) Commelinales Commelinaceae R. Br. Haemodoraceae R. Br. Hanguanaceae Airy Shaw Philydraceae Link 41 / - 6,500 14/107 1/3-15 3-4/6 Pontederiaceae Kunth. in Humb., Bonpl. & Kunth 3-4 / 30-35 Poales Africa, Asia, Australia, North America, South America Australia,W and S North America, South America, S Africa Indo-Malaysia, N Australia SE Asia, Australia Africa, Australia, Eurasia, North America, South America Cronquist 1981; Faden 1998 Simpson 1990; Hopper et al. 1999 Bayer et al. 1998; P. M . Boyce, pers. comm. Hamann 1966, 1998; Adams 1987 Eckenwalder and Barrett 1986; Graham and Barrett 1995 Anarthriaceae Cutler and Airy Shaw 1/5 W Australia Cutler and Shaw 1965 Hopkinsiaceae B. G. Briggs & L.A.S. Johnson 1 / 2 W Australia Briggs and Johnson 2000 Lyginiaceae B. G. Briggs & L.A.S. Johnson Bromeliaceae Juss. Centrolepidaceae Endl. 1/3 56/2,600 3/35 W Australia S North America, S America, W Africa SE Asia, Australia, S South America Briggs and Johnson 2000 Smith and Till 1998 Cooke 1998; Stevens 2006 oo Table 4.1 continued Order and Family No. Genera/ Species Distribution Reference(s) Cyperaceae Juss. -120/-5,000 Worldwide Bruhl 1995 Ecdeiocoleaceae Cutler and Airy Shaw 21 3 W Australia Cutler and Shaw 1965; Briggs and Johnson 1998; Briggs, pers. comm. Eriocaulaceae Desv. 11 / - 1,110 SE Asia, Africa, N America, S America, Australia Stutzel 1998; Sano 2004; Unwin 2004 Flagellariaceae Dum. 1 /4 Africa, SE Asia, N E Australia Backer 1951 Joinvilleaceae Tomlinson and A . C . Sm. 1/2 SE Asia Newell 1969; Tomlinson and Smith 1970 Juncaceae Juss. 7/442 Worldwide Kirschener 2002a-c Mayacaceae Kunth 1 /4 W Africa, SE North America, South America Lourteig 1952 Poaceae Barnhart 600-800/ - 11,000 Worldwide Clayton et al. 2002 onwards; P. M . Peterson, pers. comm. Rapateaceae Dum. 1 6 / - 100 W Africa, South America Givnish et al. 2000 Restionaceae R. Br. 55 / - 4 9 0 S Africa, SE Asia, Australia, S South America Linder et al. 1998 . Thurniaceae Engl. 1/3 N South America Cutler 1965; Dahlgren et al. 1985 Prioniaceae S. L. Munro & H. P. Linder 1 /1 S Africa Munro and Linder 1998 Table 4.1 continued Order and Family No. Genera/ Species Distribution Reference(s) Typhaceae Juss. 1/8-13 Worldwide Thieret and Luken 1996; Mavrodiev 2002 Sparganiaceae Schultz-Schultzenst. 1 / 14 N Africa, Australia, Eurasia, North America Cook and Nichols 1986, 1987 Xyridaceae C. A. Agardh. 5/-385 SE Asia, Africa, N America, Australia, S America Krai 1988, 1992, 1998; Campbell 2004a Albolbodaceae Nakai 3/23 South America Krai 1998 oo o Table 4.2. Voucher information and GenBank accession numbers for plastid regions for exemplar commelinid monocot taxa in Poales and Commelinales. Source details and GenBank information for previously published taxa used in this study (including several commelinid taxa) can be found in Graham et al. (2006), Saarela et al. (in press), and Chapter 3. Alternative source details (e.g., living collection accession numbers) are given in place of voucher information in two cases. GenBank numbers for sequences from other studies are underlined; taxon identity and references for these sequences are included as footnotes. Regions for which no sequence was obtained are indicated by a hyphen. Gene or Region psbB, T, psbE, F, 3'rps\2, Taxon Voucher Information atpB ndhF N, and p ,J L, and rbcL rpll rpsl, and psbU P psb] ndhB P O A L E S Anarthriaceae Anarthria proliferaR. Br. Anarthria scabraR. Br. Hopkinsia anoectocolea(F. Muell.) D. F. Cutler Hopkinsia anoectocolea(¥. Muell.) D. F. Cutler Lyginia imberbisR. Br. Bromeliaceae Brocchinia micrantha(Baker) Mez Centrolepidaceae Aphelia brizulaV. Muell. Centrolepis monogyna(Hook.. f.) Benth. i—* oo Brummitt 213376(H) Briggs 9581 (NSW) Briggs 93 76 (NSW) Briggs 9475 (NSW) Briggs 9477 (NSW) AJ41913Q1 No voucher; U.Wisconsin Botany Greenhouse living collection Hopper 8532 (KPBG) Linder 5689 (K) AF148760.f AF14877f AF148787" Table 4.2 continued Gene or Region Taxon Voucher Information atpB ndhF psbFS, T, N , and psbW psbD, psbC psbF, F, L , and psbi rbcL rpll i'rps\2, rpsl, and ndhB C y p e r a c e a e Carex cordillerana Saarela & B. A. Ford Eleocharis palustris (L.) Roemer & JA Schultes Gahnia baniensis Benl. Mapania meditensis D. A. Simpson Mapania cf. pubisquama Cherm E c d e i o c o l e a c e a e Georgeantha hexandra B .G. Briggs & L.A.S . Johnson E r i o c a u l a c e a e Eriocaulon compressum Lam. J o i n v i l l e a c e a e Joinvillea ascendens Gaudich. ex. Brongn. & Gris. J u n c a c e a e Juncus effusus L . Luzula sp. Saarela 196 (ALTA) Saarela 258 (UBC) Simpson s.n. (K) Simpson et al. 25J5 (K) Walters et al. 563 (MO) Briggs 9530 (NSW) Unwin 241 (MU) Weston 2501 (NSW) Rai 1004 ( A L T A ) Peterson, Saarela, and Smith 18634 (US) L126727 DO058337 AF148772.L oo Table 4.2 continued Gene or Region Taxon Voucher Information atpB ndhF psbB, T, N, and ^ psbW p s b C psbB, F, L, and psbi rbcL rpll I'rpsXl, rpsl, and ndhB Poaceae Puelia olyriformis (Franch.) Clayton Rapateaceae Rapatea sp. Stegolepis sp. Thurniaceae Prionium serratum (L. f.) Dr.ge Thurnia sphaerocephala (Rudge) Hook. f. Xyridaceae Aratitiyopea lopezii (L.B. Sm.) Steyerm. & P.E. Berry Xyris jupicai Rich. C O M M E L I N A L E S Commelinaceae Cartonema philydroides F. Muell. Palisota bogneri Brenan Bradley et al. 1060 (MO) Chase 195 (K) Kubitzki et al. 97-30 (HBG) No voucher; National Botanic Garden of Belgium living collection (Acc. 198800031 Kelloffet al. 1335 (US) AF164870" AY123242" AY2089862 AY12323912 van der Werff, Vasquez, and Gray 16131 (MO) Goldman 1766 (BH) AY465541 AF547017 AY465566 AY465670 AY465594 AY465698 AY465722 AY465622 Hort. Munich Bot. Gard. s.n. Stockey and Rothwell s. n. (ALTA) AY147602 AY147767 AY147508 AY147647 AY147555 AY149352 AY147694 AY147459 AY147606AY147771 AY147513 AY147652 AY147560 AY149356 AY147699 AY147464 oo Table 4.2 continued Gene or Region Taxon Voucher Information atpB ndhF psbB, T, N, and psbH psbD, psbC psbF, F, L, and psbl rbcL rp!2 2'rps 12, rpsl, and ndhB Tradescantia ohiensis Raf. Haemodoraceae Anigozanthos flavidus DC Anigozanthos flavidus DC Hanguanaceae Hanguana malayana (Jack) Merr. Philydraceae Helmholtzia acorifolia F. Muell. Helmholtzia glaberrima (Hook, f.) Caruel Philydrella drummondii L . G. Adams Pontederiaceae Eichhornia crassipes (Mart.) Solms Saarela 321 (UBC) Neyland 1884 (MCN) Chase 159 (NCU) Stockey and Rothwell 1 (ALTA) Prentis s. n. (BRI acc. A Q 741268) Prentis s. n. (BRI) Davis 10650 (PERTH) Barrett 814 (TKF) AF31228" AM110254'4 U41599.23 Previously published sequence data: 'Lyginia barbata, Bremer (2002); 2Thurnia sphaerocephala, Givnish et al. (2006);3Eichhornia crassipes, Graham et al. (1998),4Anarthria polyphylla, Briggs et al. (2000);5 Hopkinsia adscendens, Briggs et al. (2000);6 Lyginia barbata, Briggs et al. (2000);7 Carex hostiana, Chase et al. (1993);8 Mapania cuspidata, Verboom et al. (2006);9 Georgeantha hexandra, Briggs et al. (2000); 1 0 Puelia ciliata, Clark et al. (2000);'1 Stegolepisparvipetala, Michelangeli et al. (2003); 1 2 Thurnia polycephala, Michelangeli et al. (2003),13 Hanguana malayana, Chase et al. (2000); 1 4 Tradescantia soconuscana, Evans et al. (2003). oo 4^ Table 4.3. Additional eudicot and monocot taxa included in phylogenetic analyses with GenBank accession numbers and references. All data are from complete plastid genomes, except for Ranunculus macranthus. Taxon GenBank Accession Reference Eudicots Apiales Apiaceae, Daucus carota L. Araliaceae, Panax schinseng Nees. Asterales Asteraceae, Helianthus annuusL. Asteraceae, Lactuca sativa L. Caryophyllales Chenopodiaceae, Spinacia oleracea L. Cucurbitales Cucurbitaceae, Cucumis sativus L. Fabales Fabaceae, Glycine max (L.) Merr. Fabaceae, Lotus japonicus L. Malphigiales Salicaceae, Populus alba L. Malvales Malvaceae, Gossypium hirsutum L. Myrtales Myrtaceae, Eucalyptus globulus Labill. Onagraceae, Oenothera elata Kunth Proteales Platanaceae, Platanus occidentalis L. Rosales Moraceae, Morus indica L. Ranunculales Berberidaceae, Nandina domestica Thunb. NC 008325 NC_006290 NC_007977 NC_007578 NC 002202 NC_007144 NC_007942 NC_002694 NC_008235 NC 002694 DQ345959 AY780259 NC_002693 DQ923116 NC_008359 DQ923117 Ruhlman et al. (2006) Kim and Lee (2004) Timme et al. (unpublished) Kanamoto et al. (unpublished) Schmitz-Linneweber et al. (2001) Plader et al (unpublished); Kim et al. (2006)1 Saski et al. (2005) Kato et al. (2005) Okumura et al. (unpublished) Kato et al. (2006) Lee et al. (2006) Steane(2005) Hupferetal. (2000) Moore et al. (2006) Ravi et al. (unpublished) Moore et al. (2006) oo Table 4.3 continued Taxon GenBank Accession Reference Ranunculaceae, Ranunculus macranthus Scheele DQ069337-DQ0069702 Leebens-Mack et al. (2005) Sapindales Rutaceae, Citrus sinensis (L.) Osbeck DQ864733 Bausher et al. (2006) Solanales Solanaceae, Solanum lycopersicum L. DQ347959 Daniell et al. (2006) Solanaceae, Atropa belladonna L. NC_004561 Schmitz-Linneweber et al. (2001) Vitales Vitaceae, Vitis vinifera L. NC_007957 Jansen et al. (2006) Monocots Asparagales . Orchidaceae, Phalaenopsis aphrodite Rchb. f. NC_007499 Chang et al. (2005) Poales Poaceae, Saccharum officinarum L. NC_006084 Asano et al. (2004) Poaceae, Oryza nivaraS. D. Sharma & Shastry NC_005973 Masood et al. (2004) Poaceae, Sorghum bicolor L. AC 144549 Wiley et al. (unpublished) psbDC obtained from this genome. o o Figure 4.1. One of 28 most-parsimonious trees inferred from a large plastid data set (atpB, ndhB, ndhF, rbcL, ten photosystem II genes, rpll, rpsl, and 3'-rpsl2, and several noncoding regions in the inverted repeat; see text) showing relationships among 159 angiosperm taxa. The tree is presented as a phylogram (ACCTRAN optimization). 187 Fig 4.1 Angiosperms Commelinid Monocots 100 changes Monocots Eudicots Magnoliids "Basa l" Angiosperms 188 Figure 4.2. One of 28 most-parsimonious trees inferred from a large plastid data set (atpB, ndhB, ndhF, rbcL, ten photosystem II genes, rp/2, rpsl, and V-rps\2, and several noncoding regions in the inverted repeat; see text) showing relationships among the major angiosperm lineages (eudicots, monocots, magnoliids, and basal angiosperms). The analysis considered 159 taxa. Monocots are collapsed into a single branch for clarity. Small arrows indicate branches that collapse in a strict consensus tree. Numbers above branchs are results of bootstrap analysis. 189 E u d i c o t s i ids Angiosperms o o Figure 4.3. One of 12 most-parsimonious trees inferred from a large plastid data set (atpB, ndhB, ndhF, rbcL, ten photosystem II genes, rpll, rpsl, and 3'-rpsl 2, and several noncoding regions in the inverted repeat; see text) showing relationships among all major monocot lineages (orders). The analysis considered 113 taxa. Commelinid monocots are collapsed into a single branch for clarity. Small arrows indicate branches that collapse in a strict consensus tree. Numbers above branchs are results of bootstrap analysis. Orders and families follow APG II (2003) using the optional "bracketed" system in Asparagales, except Petrosaviaceae are recognized as an order (see Chase 2004). 191 Fig. 4.3 Monocots Agavaceae Themidaceae Hyacinthaceae Asparagaceae s.s. Lomandraceae Ruscaceae Alliaceae s.s. Agapanthaceae Amaryllidaceae Xanthorrheaceae Xanthorrheaceae Asphodelaceae Xeronemataceae Lanariaceae Tecophilaeaceae Ixioliriaceae Doryanthaceae Hypoxidaceae Lanariaceae Asteliaceae Blandfordiaceae Alaniaceae Orchidaceae Alstroemeriaceae Colchicaceae Petermanniaceae Melanthiaceae Liliaceae Smilacaceae Pandanceae Cyclanthaceae Stemonaceae Veltoziaceae Dioscoreaceae Burmanniaceae Petrosaviaceae Alismataceae Scheuchzeriaceae Araceae Tofieldiaceae Acoraceae Asparagales Liliales Pandanales Dioscoreales Petrosaviales Alismatales Acorales 192 Figure 4.4. One of 12 most-parsimonious trees inferred from a large plastid data set (atpB, ndhB, ndhF, rbcL, ten photosystem II genes, rp/2, rpsl, and 3'-rpsl2, and several noncoding regions in the inverted repeat; see text) showing relationships among the commelinid monocots. The analysis included 113 monocot taxa. All major lineages (orders) outside the commelinid monocots are collapsed into single branches for clarity. Small arrows indicate branches that collapse in a strict consensus tree. Numbers above or below branchs are results of bootstrap analysis. Orders and families follow APG II (2003), except Petrosaviaceae are recognized as an order (see Chase 2004). Subfamilies in Poaceae follow GPWG (2001) and Soreng et al. (2006). 193 Fig. 4.4 MP 100 - . \j Saccharum 100 |L so rghum Chasmanth ium Core Poales Arist idoideae Chlor idoideae Danthonioideae Arundinoideae Poo ideae Bambuso ideae Ehrharto ideae Anomoch loo ideae Ecdeiocoleaceae Joinvilleaceae Flagellariaceae Centrolepidaceae I Restionaceae Anarthriaceae I Eriocaulaceae I Mayacaceae Xyridaceae Cyperaceae Juncaceae | Thurniaceae Commelinid 53 Monocots - A C O R A L E S • A L I S M A T A L E S Commelinales Zingiberales i Arecales 100 changes TJ O SL CD 194 Figure 4.5. Optimal tree from maximum likelihood analysis inferred from a large plastid data set (atpB, ndhB, ndhF, rhcL, ten photosystem II genes, rpll, rpsl, and V-rpsXl, and several noncoding regions in the inverted repeat; see text) showing relationships among commelinid monocots. The analysis considered 69 taxa, including four exemplars from Asparagales and Liliales. The tree was inferred using the GTR + T + I model of evolution. Numbers above or below branchs are results of maximum likelihood bootstrap analysis. Orders and families follow APG II (2003), and subfamilies in Poaceae follow GPWG (2001) and Soreng et al. (2006). 195 Fig 4.5 ML Zea Saccharum Sorghum Chasmanthium Aristida Danthonia Core Poales | Rapateaceae Bromeliaceae | Typhaceae I Hanguanaceae Commelinaceae Pontederiaceae Haemodoraceae Philydraceae | Musaceae i Strelitziaceae I Arecaceae | Dasypogonaceae - Chlorophytum • Sisyrinchium | ASPARAGALES | LlLIALES Panicoideae Aristidoideae Danthoniodeae Arundinoideae Chloridoideae | Pooideae Bambusoideae | Ehrhartoideae Puelioideae Pharoideae I Anomochlooideae | Ecdeiocoleaceae | Joinvilleaceae I Flagellariaceae j Restionaceae Apheiia Centrolepidaceae Anarthriaceae i Eriocaulaceae | Xyridaceae i Mayacaceae Cyperaceae Juncaceae Thurniaceae Commelinales Zingiberales Arecales o SL n> CO 196 4.5 R E F E R E N C E S Airy Shaw, H. K. 1965. Diagnoses of new families, new names, etc., for the seventh edition of Willis's 'Dictionary'. Kew Bulletin 18, 249-273. Adams, L. G. 1987. Philydraceae. In: George, A. S. [Ed.]. Flora of Australia, Vol. 45: Hydatellaceae to Liliaceae. Australian Government Publishing Service, Canberra, Australia. Adams, K. L. , Qiu, Y . -L . , Stoutemyer, M . , and Palmer, J. D. 2002. Punctuated evolution of mitochondrial gene content: high and variable rates of mitochondrial gene loss and transfer to the nucleus during angiosperm evolution. Proceedings of the National Academy of Sciences U .S.A. 99, 9905-9912. Angiosperm Phylogeny Group [APG]. 1998. An Ordinal Classification for the Families of Flowering Plants. Annals of the Missouri Botanical Garden 85, 531-553. Angiosperm Phylogeny Group II [APG II]. 2003. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: A P G II. Botanical Journal of the Linnean Society 141, 399-436. Appel, O., and Bayer, C. 1998. Flagellariacaee. Pp. 208-211 In: Kubitzki, K. [Ed.]. The Families and Genera of Vascular Plants. Vol. 4. Springer-Verlag, Berlin. Asano, T., Tsudzuki, T., Takahashi, S., Schimada, H. , and Kadowaki, K. 2004. Complete nucleotide sequence of the sugarcane (Saccharum officinarum) chloroplast genome: a comparative analysis of four monocot chloroplast genomes. DNA Research 11, 93-99. 197 Asmussen, C. B., Dransfield, J., Deickmann, V., Barford, A. S., Pintaud, J . - C , and Baker, W. J. 2006. A new subfamily classification of the palm family (Arecaceae): evidence from plastid DNA phylogeny. Botanical Journal of Linnean Society 151, 15-38. Backer, C. A. 1951. Flagellariaceae. Pp. 245-250 In: van Steenis, C. G. G. J. [Ed.]. Flora Malesiana I. 4. Noordhoff-Kolff N. V., Djakarta. Balslev, H. 1998. Juncaceae. Pp. 252-260 In: Kubitzki, K. [Ed.] The Families and Genera of Vascular Plants. Vol. 4. Springer-Verlag, Berlin. Barrett, S. C. H., and Graham, S. W. 1997. Adaptive radiation in the aquatic plant family Pontederiaceae: insights from phylogenetic analysis. Pp. 225-258 In: Givnish, T. J., and Sytsma, K. J. [Eds.]. Molecular Evolution and Adaptive Radiation. Cambridge University Press, Cambridge, United Kingdom. Barrett, S. C. H., Jesson, L. K., and Baker, A. M. 2000. The evolution and function of stylar polymorphism in flowering plants. Annals of Botany 85, 253-265. Barthlott, W., and Frolich, D. 1983. Mikromorphologie und Orientierungsmuster epicuticularer Wachse-Kristalloide: Ein neues systematiches Merkmal bei Monokotylen. Plant Systematics and Evolution 142, 171-185. Bausher, M . G., Singh, N. D., Lee, S. -B., Jansen, R. K., and Daniell, H. 2006. The complete chloroplast genome sequence of Citrus sinensis (L.) Osbeck var. 'Ridge Pineapple': organization and phylogenetic relationships to other angiosperms. B M C Plant Biology 6, 21. doi:10.1186/1471-2229-6-21. Bayer, C , Appel, O., and Rudall, P. J. 1998. Hanguanaceae. Pp. 223-225 In: Kubitzki, K. [Ed.]. The Families and Genera of Vascular Plants. Vol'.-4. Springer-Verlag, Berlin. 198 Bremer, K. 2000. Early Cretaceous lineages of monocot flowering plants. Proceedings of the National Academy of Sciences U.S.A. 97, 4707-4711. Bremer, K. 2002. Gondwanan evolution of the grass alliance of families (Poales). Evolution 56, 1374-1387. Briggs, B. G., and Johnson, L. A. S. 1998. Georgeantha hexandra, a new genus and species of Ecdeiocoleaceae (Poales) from western Australia. Telopea 7, 307-312. Briggs, B. G., and Johnson, L. A. S. 2000. Hopkinsiaceae and Lyginiaceae, two new families of Poales in western Australia, with revisions of Hopkinsia and Lyginia. Telopea 8, 477-502. Briggs, B. G., Marchant, A. D., Gilmore, S., and Porter, C. L. 2000. A molecular phylogeny of Restionaceae and allies. Pp. 661-671 In: Wilson, K. L. , and D. A. Morrison [Eds.], Monocots—Systematics and Evolution. CSIRO, Melbourne. Briggs, B. G. 2004. Restionaceae (Poales) in the footsteps of Robert Browne. Telopea 10, 499-503. Briggs, B. G. 2005. Switching sex along the spike —- synchronized monoecy in Ecdeiocoleaceae, sister of Poaceae. Pp. 298 In: Abstracts, XVII International Botanical Congress, Vienna, Austria. Bruhl, J. J. 1995. Sedge genera of the world, relationships, and a new classification of the Cyperaceae. Australian Systematic Botany 8, 125-305. Cai, Z., Penaflor, C , Kuehl, J. V. , Leebens-Mack, J., Carlson, J. E. , dePamphilis, C. W., Boore, J. L, and Jansen, R. K. 2006. Complete plastid genome sequences of Drimys, Liriodendron, and Piper, implications for the phylogenetic relationships of magnoliids. BMC Evolutionary Biology 6, 77. doi: 10.1186/1471-2148-6-77. 199 Campbell, C. S., and Kellogg, E. A. 1987. Sister group relationships of the Poaceae. Pp. 217-224 In: Soderstrom, T. R., Hilu, K. W., Campbell, C. S., and Barkworth, M . E. [Eds.]. Grass Systematics and Evolution. Smithsonian Institution Press, Washington, D. C. Campbell, L. M . 2004a. Xyridaceae. Pp. 492-493 In: Smith, N., Mori, S. A., Henderson, A., Stevenson, D. W., and Heald, S. V. [Eds.]. Flowering Plants of the Neotropics. Princeton University Press, Princeton, New Jersey. Campbell, L . M . 2004b. Anatomy and systematics of Xyridaceae with special reference to Aratitiyopea. Unpublished Ph.D. Dissertation. City University of New York. Campbell, L. M . , and Stevenson, D. W. 2005. Vegetative anatomy of Aratitiyopea lopezii (Xyridaceae). Acta Botanica Venezuala 28, 2 [online, cited 30 June 2006], p. 395-408. Chang C.-C, Lin, H . - C , Lin, I.-P., Chow,T.-Y., Chen, H.-H., Chen ,W.-H., Cheng, C.-H., Lin C.-Y., Liu S.-M., Chang, C . - C , and Chaw, S.-M. 2006. The chloroplast genome of Phalaenopsis aphrodite (Orchidaceae): comparative analysis of evolutionary rate with that of grasses and its phylogenetic implications. Molecular Biology and Evolution 23,279-291. Chase, M . W., Soltis, D. E. , Olmstead, R. G., Morgan, D., Les, D. H., Mishler, B. D., Duvall, M . R., Price, R. A., Hills, H. G., Qui, Y. L. , Kron, K. A., Rettig, J. H., Conti, E. , Palmer, J. D., Manhart, J. R., Sytsma, K. J., Michael, H. J., Kress, W.J . , Karol, K. A., Clark, W. D., Hedren, M . , Gaut, B. S., Jansen, R. K., Kim, K. J., Wimpee, C. F., Smith, J. F., Furnier, G. R., Strauss, S. H., Xiang, Q. Y., Plunkett, G. M . , Soltis, P. S., Swensen, S. M . , Willaims, S. E. , Gadek, P. A., Quinn, C. J., Eguiarte, L. E. , 200 Golenberg, E. , Learn, G. H., Graham, S. W., Barrett, S. C. H., Dayanandan, S., and Albert, V. A. 1993. Phylogenetics of seed plants: an analysis of nucleotide sequences from the plastid gene rbcL. Annals of the Missouri Botanical Garden 80, 528-580. Chase, M . W., Stevenson, D. W., Wilkin, P., and Rudall, P. J. 1995. Monocot systematics: a combined analysis. Pp. 685-740 In: Rudall, P. J., Cribb, P. J., Cutler, D. F., and Humphries, C. J. [Eds.]. Monocotyledons: Systematics and Evolution. Royal Botanic Gardens, Kew. Chase, M . W., Soltis, D. E. , Soltis, P. S., Rudall, P. J., Fay, M . F., Hahn, W. H., Sullivan, S., Joseph, J., Molvray, M . , Kores, P. J., Givnish, T. J., Sytsma, K. J., and Pires, J. C. 2000. Higher-level systematics of the monocotyledons: an assessment of current knowledge and a new classfication. Pp. 3-16 In: Wilson, K. L , and Morrison, D. A. [Eds.]. Monocots: Systematics and Evolution. CSIRO, Melbourne, Australia. Chase, M . W. 2004. Monocot relationships: an overview. American Journal of Botany 91, 1645-1655. Chase, M . W., Duvall, M . R., Hills, H. G., Conran, J. G., Cox, A. V. , Eguiarte, L. E. , Hartwell, J., Fay, M . F., Caddick, L. R., Cameron, K. M . , and Hoot, S. 1995. Molecular phylogenetics of Lilianae. Pp. 109-137. In: Rudall, P. J., Cribb, P. J., Cutler, D. F., and Humphries, C. J. [Eds.]. Monocotyledons: Systematics and evolution. Royal Botanic Gardens, Kew. Chase, M . W., Fay, M . F., Devey, D., Rjansted, N. , Davies, J., Pillon, Y., Petersen, G., Seberg, O., Tamura, M . N., Asmussen, C. B., Hilu, K., Borsch, T., Davis, J. I., Stevenson, D. W., Pires, J. C , Givnish, T. J., Sytsma, K. J., McPherson, M . A., 201 Graham, S. W., and Rai, H. S. 2006. Multigene analyses of monocot relationships: A summary. In: Columbus, J. T., Friar, E. A. , Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Cho, Y., Mower, M . P., Qiu, Y . -L . , and Palmer, J. D. 2004. Mitochondrial substitution rates are extraordinarily elevated and variable in a genus of flowering plants. Proceedings of the National Academy of Sciences U.S.A. 101, 1774-17746. Clark, W. D., Gaut, B. S., Duvall, M . R., and Clegg, M . T. 1993. Phylogenetic relationships of the Bromeliiflorae-Commeliniflorae-Zingiberiflorae complex of monocots based on rbcL sequence comparisons. Annals of the Missouri Botanical Garden 80, 987-998. Clark, L. G., Kobayashi, M . , Mathews, S., Spangler, R. E. , and Kellogg, E. A. 2000. The Puelioideae, a new subfamily of Poaceae. Systematic Botany 2000, 181-187. Clayton, W. D., Harman, K. T., and Williamson, H. 2002 onwards. World grass species, descriptions, identification, and information retrieval, http://www.kew.org/data/ grasses-db.html [Sep 2006]. Cook, C. D. K., and Nicholls, M . S. 1986. A monographic study of the genus Sparganium (Sparganiaceae). 1. Subgenus Xanthosparganium Holmberg. Botanica Helvetica 96, 213-267. Cook, C. D. K., and Nicholls, M . S. 1987. A monographic study of the genus Sparganium (Sparganiaceae). 2. Subgenus Sparganium. Botanica Helvetica 97, 1-44. Cooke, D. A. 1998. Centrolepidaceae. Pp. 106-109 In: Kubitzki, K. [Ed.]. The Families and Genera of Vascular Plants, Vol. IV. Springer-Verlag, Berlin. 202 Cronquist, A. 1981. An integrated system of classification of flowering plants. Columbia University Press, New York. Cronquist, A. 1988. The Evolution and Classification of Flowering Plants. Second edition. New York Botanical Garden, New York. Cummings, M . P., Handley, S. A., Myers, D. S., Reed, D. L. , Rokas, A:, and Winka, K. 2003. Comparing bootstrap and posterior probability values in the four-taxon case. Systematic Biology 52, 477-487. Cutler, D. F. 1965. Vegetative anatomy of Thurniaceae. Kew Bulletin 19, 431-441. Cutler, D. F. 1969. Juncales. In: Metcalfe, C. R. [Ed.]. Anatomy of the Monocotyledons IV. Clarendon Press, Oxford. Cutler, D. F., and Airy Shaw, H. K. 1965. Anarthriaceae and Ecdeiocoleaceae: two new monocotyledonous families, separated from the Restionaceae. Kew Bulletin 19, 489-499. Dahlgren, R. M . T., and Clifford, H. T. 1982. The Monocotyledons: a Comparative Study. Academic Press, London. Dahlgren, R. M . T., and Rasmussen, F. N. 1983. Monocotyledon evolution: characters and phylogenetic estimation. Evolutionary Biology 16, 255-395. Dahlgren, R. M . T., Clifford, H. T., and Yeo, P. F. 1985. The Families of Monocotyledons: Structure, Evolution and Taxonomy. Springer-Verlag, Berlin. Daniell, H., Lee, S.-B., Grevich, J., Saski, C , Quesada-Vargas, T., Guda, C , Tomkins, J., and Jansen, R. K. 2006. Complete chloroplast genome sequences of Solarium bulbocastanum, Solarium lycopersicon and comparative analyses with other Solanaceae genomes. Theoretical and Applied Genetics 112, 1503-1518. 203 Davis, J. I. 1995. A phylogenetic structure for the monocotyledons, as inferred from chloroplast DNA restriction site variation, and a comparison of measures of clade support. Systematic Botany 20, 503-527. Davis, J. I., Stevenson, D. W., Peterson, G., Seberg, O., Campbell, L. M . , Freudenstein, J. V., Goldman, D. H., Hardy, C. R., Michelangeli, F. A., Simmons, M . P., Specht, C. D., Vergara-Silva, F., and Gandolfo, M . A. 2004. A phylogeny of the monocots, as inferred from rbcL and atpA sequence variation. Systematic Botany 29, 467-510. Doyle, J. J., and Doyle, J. L. 1987. A rapid isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19, 11-15. Doyle, J. J., Davis, J. I., Soreng, R. J., Garvin, D., and Anderson, M . J. 1992. Chloroplast inversions and the origin of the grass family (Poaceae). Proceedings of the National Academy of Sciences U.S.A. 89, 7722-7726. Drabkova, L., Kirschner, J., Seberg, O., Petersen, G., and Vlacek, C. 2003. Phylogeny of the Juncaceae based on rbcL sequences, with special emphasis on Luzula DC. and Juncus L. Plant Systematics and Evolution 240, 133-147. 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., and Learn Jr., G. H. 1993. Phylogenetic hypotheses for the monocotyledons constructed from rbcL sequence data. Annals of the Missouri Botanical Garden 80, 607-619. Duvall, M . R., Davis, J. I., Clark, L. G., Noll, J. D., Goldman, D. H., and Sanchez-Ken, J. G. 2006. Phylogeny of the grasses (Poaceae) revisited. In: Columbus, J. T., Friar, E. A., Porter, J. M. , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative 204 Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Eckenwalder, J. E. , and Barrett, S. C. H. 1986. Phylogenetic systematics of Pontederiaceae. Systematic Botany 11, 373-391. Erixon, P., Svennblad, B., Britton, T., and Oxelman, B. 2003. Reliability of Bayesian posterior probabilities and bootstrap frequencies in phylogenetics. Systematic Biology 52, 665-673. Evans, T. M. , Faden, R. B., Simpson, M . G., and Sytsma, K. J. 2000. Phylogenetic relationships in the Commelinaceae: I. A cladistic analysis of morphological data. Systematic Botany 25, 668-691. Evans, T. M . , Sytsma, K. J., Faden, R. B., and Givnish, T. J. 2003. Phylogenetic relationships in the Commelinaceae: II. A cladistic analysis of rbcL sequences and morphology. Systematic Botany 28, 270-292. Faden, R. B. 1985. Mayacaceae. Pp. 387-388 In: Dahlgren, R. M . , Clifford, H. T., and Yeo, P. F. [Eds.]. The Families of the Monocotyledons: Structure, Evolution and Taxonomy. Springer, Berlin, Germany. Faden, R. B. 1998. Commelinaceae. Pp. 109-128 In: Kubitzki, K. [Ed.]. The Families and Genera of Vascular Plants. Vol. 4. Springer-Verlag, Berlin. Felsenstein, J. 1978. Cases in which parsimony or compatibility methods will be positively misleading. Systematic Zoology 29, 401-410. Felsenstein, J. 1985. Confidence limits on phylogeny: an approach using the bootstrap. Evolution 39, 783-791. 205 Freyer, R., Neckermann, K., Maier, R. M. , and Kossel, H. 1995. Structural and functional analysis of plastid genomes from parasitic plants: loss of an intron within the genus Cuscuta. Current Genetics 27, 580-586. Frolich, D., and Barthlott, W. 1988. Mikromorphologie der epicuticularen Wachse und das System der Monokotylen. Tropische und Subtropische Pflanzenwelt 63, 1-135. Gaut, B. S., Muse, S. V., Clark, W. D., and Clegg, M . T. 1992. Relative rates of nucleotide substitution at the rbcL locus of monocotyledonous plants. Journal of Molecular Evolution 35, 292-303. Gilg-Benedict, T. 1930. Centrolepidaceae. Pp. 27-33 In: Engler & Prantl, Naturlichen Pflanzenfamilien 2 n d ed., 15a. Leipzig, W. Engelmann. Giulietti, A. M.,'Amaral, M . C , and Bittrich, V: 1995. .Phylogenetic analysis of inter and infrageneric relationships of Leiothrix Ruhland (Eriocaulaceae). Kew Bulletin 50, 55-71. Givnish, T. J., Evans, T. M . , Pires, J. C , and Sytsma, K. J. 1999. Polyphyly and convergent morphological evolution in Commelinales and Commelinidae: evidence from rbcL sequence data. Molecular Phylogenetics and Evolution 12, 360-385. Givnish, T. J., Evans, T. M . , Zjhra, M . L. , Patterson, T. B., Berry, P. E . , and Sytsma, K. J. 2000. Molecular evolution, adaptive radiation, and geographic diversification in the amphiatlantic family Rapateaceae: evidence from ndhF sequences and morphology. Evolution 54, 1915-1937. Givnish, T. J., Pires, J. C , Graham, S. W., McPherson, M . A., Prince, L. M . , Patterson, T. B., Rai, H. S., Roalson, E. H., Evans, T. E. , Hahn, W. J., Millam, K. C , Meerow, A. W., Molvray, M . , Kores, P. J., O'Brien, H. E. , Hall, J. C , Kress, W. J., and 206 Sytsma, K. J. 2006. Phylogenetic relationships of monocots based on the highly informative plastid gene ndhF: evidence for widespread concerted convergence. Pp. 27-50 In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Goetghebeur, P. 1998. Cyperaceae. Pp. 141-190 In: Kubitzki, K. [Ed.]. The Families and Genera of Vascular Plants. Vol. 4. Springer-Verlag, Berlin. Goremykin, V. V., Hollan, B., Hirsch-Ernst, K. I., and Hellwig, F. H. 2005. Analysis of Acorus calamus chloroplast genome and its phylogenetic implications. Molecular Biology and Evolution 22, 1813-1822. Graham, S. W., and Barrett, S. C. H. 1995. Phylogenetic systematics of Pontederiales: ~ implications for breeding-system evolution. Pp. 415-551 In: Rudall, P. J., Cribb, P. J., Cutler, D. F., and Humphries, C. J. [Eds.]. Monocotyledons: Systematics and Evolution. Royal Botanic Gardens, Kew. Graham, S. W., Kohn, J. R., Morton, B. R., Eckenwalder, J. E. , and Barrett, S. C. H. 1998. Phylogenetic congruence and discordance among one morphological and three molecular data sets from Pontederiaceae. Systematic Biology 47, 545-567. Graham, S. W., and Olmstead, R. G. 2000a. Utility of 17 chloroplast genes for inferring the phylogeny of the basal angiosperms. American Journal of Botany 87, 1712-1730. Graham, S. W., and Olmstead, R. G. 2000b. Evolutionary significance of an unusual chloroplast DNA inversion found in two basal angiosperm lineages. Current Genetics 37, 183-188. Graham, S. W., Reeves, P. A., Burns, A. C. E. , and Olmstead, R. G. 2000. 207 Microstructural changes in noncoding chloroplast DNA: interpretations, evolution, and utility of indels and inversion in basal angiosperm phylogenetic inference. International Journal of Plant Sciences 161 (6 Suppl.), S83-S96. Graham, S. W., Olmstead, R. G., and Barrett, S. C. H. 2002. Rooting phylogenetic trees with distant outgroups: a case study from the commelinoid monocots. Molecular Biology and Evolution 19, 1769-1781. Graham, S. W., Zgurski, J. M . , McPherson, M . A., Cherniawsky, D. M . , Saarela, J. M . , Home, E. C , Smith, S. Y., Wong, W. A., O'Brien, H., Biron, V. L. , Pires, J. C., Olmstead, R. G., Chase, M . W., and Rai, H. S. 2006. Robust inference of monocot deep phylogeny using an expanded multigene plastid data set. Pp. 3-20 In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Grass Phylogeny Working Group [GPWG]. 2001. Phylogeny and subfamilial classification of the grasses (Poaceae). Annals of the Missouri Botanical Garden 88, 373-457. Guindon, S., and Gascuel, O. 2003. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Systematic Biology 52, 696-704. Hamann, U. 1964. Centrolepidaceae. Pp. 558-559 In Engler, A. [Ed.]. Syllabus der Pflanzenfamilien 12th ed. Vol. 2. Borntraeger, Berlin. Hamann, U. 1966. Embryologische, morphologisch-anatomische und systematische Untersuchungen an Philydraceen. Willdenowia 4, 1-178. 208 Hamann, U. 1975. Neue untersuchungen zur embryologie und systematic der Centrolepidaceae. Botanische Jahrbucher fur Systematik, Pflanzengeschichte und Pflanzengeographie 96, 154-191. Hamann, U. 1976. Hydatellaceae - a new family of Monocotyledoneae. New Zealand Journal of Botany 14, 193-196. Hamann, U. 1998. Philydraceae. Pp. 389-394. In: Kubitzki, K. (Ed.). The Families and Genera of Vascular Plants. Vol.4. Springer-Verlag, Berlin. Harris, P. J., and Hartley, R. D. 1976. Detection of bound ferulic acid in cell walls of the Gramineae by ultraviolet fluorescence microscopy. Nature 259, 508. Harris, P. J., and Hartley, R. D. 1980. Phenolic constituents of the cell walls of monocotyledons. Biochemical Systematics and Ecology 8, 53-160. Harris, P. J., Kelderman, M . R., Kendon, M . F., and Mckenzie, R. J. 1997. Monosaccharide compositions of unlignified cell walls of monocotyledons in relation to the occurrence of wall-bound ferulic acid. Biochemical Systematics and Ecology 25, 167-179. Hendy, M . D., and Penny, D. 1989. A framework for the quantitative study of evolutionary trees. Systematic Zoology 38, 297-309. Herendeen, P. S., and Crane, P. R. 1995. Pp. 1-21 In: Rudall, P. J., Cribb, P. J., Cutler, D. F., and Humphries, C. J. [Eds.]. Monocotyledons: Systematics and Evolution. Royal Botanic Gardens, Kew. Hillis, D. 1998. Taxonomic sampling, phylogenetic accuracy, and investigator bias. Systematic Biology 42, 182-192. 209 Hilu, K. W., Borsch, T., Muller, K., Soltis, D. E. , Soltis, P. S., Savolainen, V., Chase, M . W., Powell, M . P., Alice, L. A., Evans, R., Sauquet, H., Neinhuis, C , Slotta, T. A. B., Rohwer, J. G., Campbell, C. S., and Chatrou, L. W. 2003. Angiosperm phylogeny based on matK sequence information. American Journal of Botany 90, 1758-1776. Hoot, S. B., and Palmer, J. D. 1994. Structural rearrangements, including parallel inversion, within the chloroplast genome of Anemone and related genera. Journal of Molecular Evolution 38, 274-281. Hoot, S. B., Culham, A., and Crane, P. R. 1995. The utility of atpB gene sequences in resolving phylogenetic relationships: comparison with rbcL and 18S ribosomal DNA sequences in Lardizabalaceae. Annals of the Missouri Botanical Garden 82, 194-207. Hopper, S. D., Fay, M. F., Rossetto, M . , and Chase, M. W. 1999. A molecular phylogenetic analysis of the bloodroot and kangaroo paw family, Haemodoraceae: taxonomic, biogeographic and conservation implications. Botanical Journal of the Linnean Society 131,285-299. Horres, R., Zizka, G., Kahl, G., and Weising, K. 2000. Molecular phylogenetics of Bromeliaceae: evidence from trnL(UAA) intron sequences of the chloroplast genome. Plant Biology 2, 306-315. Huelsenbeck, J. P. 1997. Is the Felsenstein zone a fly trap? Systematic Biology 46, 69-74. Hupfer, H., Swaitek, M . , Hornung, S., Herrmann, R. G., Maier, R. M . , Chiu, W. L., and Sears, B. 2000. Complete nucleotide sequence of the Oenothera elata plastid chromosome, representing plastome 1 of the five distinguishable Euoenthera plastomes. Molecular Genomics and Genetics 263, 581-585. 210 Hutchinson, J. 1973. The Families of Flowering Plants. 1st edition, Volume 2. Monocotyledons. Oxford University Press, London. Jansenn, T., and Bremer, K. 2004. The age of major monocot groups inferred from 800+ rbcL sequences. Botanical Journal of the Linnean Society 146, 385-398. Jansen, R. K., Kaittanis, C , Saski, C , Lee, S. -L . , Tomkins, J., Alverson, A. J., and Daniell, H. 2006. Phylogenetic analyses of Vitis (Vitaceae) based on complete chloroplast genome sequences: effects of taxon sampling and phylogenetic methods on resolving relationships among rosids. B M C Evolutionary Biology 6, 32. doi: 10.1186/1471-2148-6-32. Jesson, L. K., and Barrett, S. C. H. 2002a. The genetics of mirror-image flowers. Proceedings of the Royal Society of London B 269, 1835-1839. Jesson, L. K., and Barrett, S. C. H. 2002b. Enantiostyly in Wachendorfia (Haemodoraceae): the influence of reproductive systems on the maintenance of the polymorphism. American Journal of Botany 89, 253-262. Jesson, L. K., and Barrett, S. C. H. 2003. The comparative biology of mirror-image flowers. International Journal of Plant Sciences 164, S237-S249. Judd, W. J., Campbell, C. S., Kellogg, E. A., Stevens, P. F., and Donoghue, M . J. 2002. Plant Systematics - A Phylogenetic Approach. 2 n d Edition. Sinauer Associates, Sunderland, Massachusetts. Katayama, H., and Ogihara, Y. 1996. Phylogenetic affinities of the grasses to other monocots as revealed by molecular analysis of chloroplast DNA. Current Genetics 29, 572-581. 211 Kato, T., Kaneko, T., Sato, S., Nakamura, Y., and Tabata, S. 2000. Complete structure of the chloroplast genome of a legume, Lotus japonicus. DNA Research 7, 323-330. Kellogg, E. A., and Linder, H. P. 1995. Phylogeny of Poales. Pp. 511-542. In: Rudall, P. J., Cribb, P. J., Cutler, D. F., and Humphries, C. J. [Eds.]. Monocotyledons: Systematics and Evolution. Royal Botanic Gardens, Kew. Kim, J.-S., Jung, J. D., Lee, J. -A., Park, H.-W., Oh, K. -H. , Jeong, W.-J., Choi, D.-W., Liu, J. R., and Cho, K. Y. 2006. Complete sequence and organization of the cucumber (Cucumis sativus L. cv. Baekmibaekdadagi) chloroplast genome. Plant Cell Reports 25, 334-340. Kim, K.-J., and Lee, H.-L. 2004. Complete chloroplast genome sequences from Korean ginseng (Panax schinseng Nees) and comparative analysis of sequence evolution among 17 vascular plants. DNA Research 11, 247-261. Kirschener, J. 2002a. Juncaceae 1: Rostkovia to Luzula, Species Plantarum: Flora of the world, part 6. ABRS, Canberra. Kirschener, J. 2002b. Juncaceae 2: Juncus subg. Juncus, species plantarum: Flora of the world, part 7. ABRS, Canberra. Kirschener, J. 2002c. Juncaceae 3: Juncus subg. Agathryon, species plantarum: Flora of the world, part 8. ABRS, Canberra. Kohn, J. R., Graham, S. W., Morton, B. R., Doyle, J. J., and Barrett, S. C. H. 1996. Reconstruction of the evolution of reproductive characters in Pontederiaceae using phylogenetic evidence from chloroplast DNA restriction-site variation. Evolution 50, 1454-1469. 212 Krai, R. 1988. The genus Xyris (Xyridaceae) in Venezuala and contiguous northern South America. Annals of the Missouri Botanical Garden 75, 522-722. Krai, R. 1992. A treatment of American Xyridaceae exclusive of Xyris. Annals of the Missouri Botanical Garden 79, 819-885. Krai, R. 1998. Xyridaceae. Pp. 461-469 In: Kubitzki, K. [Ed.]. The Families and Genera of Vascular Plants. Vol. 4. Springer-Verlag, Berlin. Kress, W. J. 1995. Phylogeny of the Zingiberanae: morphology and molecules. Pp.443-460 In: Rudall, P. J., Cribb, P. J., Cutler, D. F., and Humphries, C. J. [Eds.]. Monocotyledons: Systematics and Evolution. Royal Botanic Gardens, Kew. Kress, W. J., Prince, L. M . , Hahn, W. J., and Zimmer, E. A. 2001. Unraveling the evolutionary radiation of the families of the Zingiberales using morphological and molecular evidence. Systematic Biology 50, 926-944. Kristiansen, K. A., Cilieborg, M . , Drabkova, L., J0rgensen, T., Petersen, G., and Seberg, O. 2005. DNA Taxonomy - the riddle of Oxychloe (Juncaceae). Systematic Botany 30, 284-289. Lee, S. -B., Kaittanis, C , Jansen, R.K., Hostetlerm J. B., Tallon, L.J., Town, C D . , Daniell H. 2006. The complete chloroplast genome sequence of Gossypium hirsutum: organization and phylogenetic relationships to other angiosperms. B M C Genomics 7, 61. doi: 10.1186/1471 -2-164-7-61. Leebens-Mack, J., Raubeson, L. A., Cui, L. , Kuehl, J. V., Fourcade, M . H., Chumley, T. W., Boore, J. L. , Jansen, R. K., and dePamphilis, C. W. 2005. Identifying the basal angiosperm node in chloroplast genome phylogenies: sampling one's way out of the Felsenstein zone. Molecular Biology and Evolution 22, 1948-1963. 213 Linder, H. P., and Ferguson, I. K. 1985. On the pollen morphology and phylogeny of the Restionales and Poales. Grana 24, 65-76. Linder, H. P., and Rudall, P. J. 1993. The megagametophyte in Anarthria (Anarthriaceae, Poales) and its implications for the phylogy of the Poales. American Journal of Botany 80,1455-1464. Linder, H. P., and Kellogg, E. A. 1995. Phylogenetic patterns in the commelinid clade. Pp. 473-496 In: Rudall, P. J., Cribb, P. J., Cutler, D. F., and Humphries, C. J. [Eds.]. Monocotyledons: Systematics and Evolution. Royal Botanic Gardens, Kew. Linder, H. P., Briggs, B. G., and Johnson, L. A. S. 1998. Restionaceae. Pp. 425-444 In: Kubitzki, K. [Ed.]. The Families and Genera of Vascular Plants. Vol. 4. Springer-Verlag, Berlin. Linder, H. P., Briggs, B. G., and Johnson, L. A. S. 2000. Restionaceae—a morphological phylogeny. Pp. 653-660 In: Wilson, K. L., and Morrison, D. A. [Eds.]. Monocots: Systematics and Evolution. CSIRO, Melbourne, Australia. Linder, H. P., Eldenas, P., and Briggs, B. G. 2003. Contrasting patterns of radiation in Australian and African Restionaceae. Evolution 57, 2688-2702. Linder, H. P., and Rudall, P. J. 2005. Evolutionary history of Poales. Annual Review of Ecology, Evolution, and Systematics 35, 107-124. Lourteig, A. 1952. Mayacaceae. Notulae Systematicae (Paris) 14, 234-248. Manning, J. C , and Linder, H. P. 1990. Cladistic analyses of patterns of endothecial thickenings in the Poales/ Restionales. American Journal of Botany 77, 196-210. Masood, M . S., Nishikawa, T., Fukuoka, S., Njenga, P. K., Tsudzuki, T., and Kadowaki, K. 2004. The complete nucleotide sequence of wild rice (Oryza nivara) chloroplast 214 genome: first genome wide comparative sequence analysis of wild and cultivated rice. Gene 340, 133-139. Mathews, S., Spangler, R. E. , Mason-Gamer, R. J., and Kellogg, E. A. 2002. Phylogeny of Andropogoneae inferred from phytochrome B, GBSSI, and ndhF. International Journal of Plant Sciences 163, 441-450. Mavrodiev, E. V. 2002. Two new species of Typha L. (Typhaceae Juss.) from the far east of Russia and from Mongolia. Feddes Repertorium 113, 281-288. McPherson, M . A. 2003. Phylogenetics of Asparagales (Liliopsida). M . Sc. Thesis. University of Alberta, Edmonton, Alberta, Canada. McPherson, M . A., Fay, M . F., Chase, M . W., and Graham, S. W. 2004. Parallel loss of a slowly evolving intron from two closely related families in Asparagales. Systematic Botany 29, 296-307. Metcalfe, C. R. 1971. Anatomy of the Monocotyledons. V. Cyperaceae. Clarendon Press, Oxford. Michelangeli, F. A., Davis, J. I., and Stevenson, D. W. 2003. Phylogenetic relationships among Poaceae and related families as inferred from morphology, inversions in the plastid genome, and sequence data from the mitochondrial and plastid genomes. American Journal of Botany 90, 93-106. Moore, M . J., Dhingra, A., Soltis, P. S., Shaw, R., Farmerie, W. G., Folta, K. M . , and Soltis, D. E. 2006. Rapid and accurate pyrosequencing of angiosperm plastid genomes. BMC Plant Biology 6, 17. doi: 10.1186/147-2229-6-17. 215 Muasya, A. M . , Simpson, D. A., Chase, M . W., and Culham, A. 1998. An assessment of suprageneric phylogeny in Cyperaceae using rbcL DNA sequences. Plant Systematics and Evolution 211, 257-271. Muasya, A. M . , Bruhl, J. J., Simpson, D. A., Culham, A., and Chase, M . W. 2000. Suprageneric phylogeny of Cyperaceae: a combined analysis. Pp. 593-601 In: Wilson, K. L., and Morrison, D. A. [Eds.]. Monocots: Systematics and Evolution. CSIRO, Melbourne, Australia. Munro, S. L., and Linder, H. P. 1998. The phylogenetic position of Prionium (Juncaceae) within the order Juncales based oh morphological and rbcL sequence data. Systematic Botany 23, 43-55. Newell, T. K. 1969. A study of the genus Joinvillea (Flagellariaceae). Journal of the Arnold Arboretum 50, 527-555. Neyland, R., and Urbatsch, L. E. 1996a. Phylogeny of subfamily Epidendroideae (Orchidaceae) inferred from ndhF chloroplast gene sequences. American Journal of Botany 83, 1195-1206. Neyland, R., and Urbatsch, L. E. 1996b. The ndhF chloroplast gene detected in all vascular plant divisions. Planta 200, 213-211. Ogihara, Y., Isono, K., Kojima, T., Endo, A., Hanaoka, M . , Shiina, T., Terachi, T., Utsugi, S., Murata, M . , Mori, N., Takumi, S., Ikeo, K., Gojobori, T., Murai, R., Murai, K., Matsuoka, Y., Ohnishi, Y., Tajiri, H., and Tsunewaki, K. 2000. Chinese spring wheat (Triticum aestivum L.) chloroplast genome: complete sequence and contig clones. Plant Molecular Biology Reporter 18, 243-253. 216 Olmstead, R. G., and Sweere, J. A. 1994. Combining data in phylogenetic systematics: an empirical approach using three molecular data sets in Solanaceae. Systematic Biology 43,467-481. Ong, H. C , and Palmer, J. D. 2006. Pervasive survival of expressed mitochondrial rps\4 pseudogenes in grasses and their relatives for 80 million years following three functional transfers to the nucleus. B M C Evolutionary Biology 6, 55. doi: 10.1186/1471-2148-6-55. Otalvaro, F., Gorls, H., Holscher, D., Schmitt, B, Echeverrie, F., Quifiones, and W., Schneider, B. 2002. Dimeric phenylphenalenones from Musa acuminata and various Haemodoraceae species. Crystal structure of anigorootin. Phytochemistry 60, 61-66. Parkinson, C. L., Mower, J. P., Qiu, Y. -L. , Shirk, A. J., Song, K., Young, N. D., dePamphilis, C. W., and Palmer, J. D. 2005. Multiple major increases and decreases in mitochondrial substitution rates in the plant family Geraniaceae. B M C Evolutionary Biology 5, 73. doi: 10.1186/1471-2148-5-73. Plunkett, G. M . , Soltis, D. E. , Soltis, P. S., and Brooks, R. E. 1995. Phylogenetic relationships between Juncaceae and Cyperaceae based on rbcL sequence data. American Journal of Botany 82, 520-525. Poe, S., and Swofford, D. L. 1999. Taxon sampling revisited. Nature 398, 299-300. Posada, D., and Crandall, K. A. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14,817-818. Prentis, P. J., Meyers, N. M . , and Mather, P. B. 2006. Significance of post-germination buoyancy in Helmholtzia glaberrima and Philydrum lanuginosum (Philydraceae). Australian Journal of Botany 54, 11-16. 217 Rai, H. S., O'Brien, H. E. , Reeves, P. A., Olmstead, R. G., Graham, S. W. 2003. Inference of higher-order relationships in the cycads from a large chloroplast data set. Molecular Phylogenetics and Evolution 29, 350-359. Rambaut, A. 1998. Se-Al (sequence alignment editor version 1.0 alpha 1). Department of Zoology, University of Oxford, Oxford. Roalson, E. H. 2005. Phylogenetic relationships in the Juncaceae inferred from nuclear ribosomal DNA internal transcribed spacer sequence data. International Journal of Plant Sciences 166, 397-413. Rokas, A., Williams, B. L., King, N. , and Carroll, S. B. 2003. Genome-scale approaches to resolving incongruence in molecular phylogenies. Nature 425, 798-804. Rudall, P. J., and Caddick, L. R. 1994. Investigation of the presence of phenolic compounds in monocotyledonous cell walls, using U V fluorescence microscopy. Annals of Botany 74, 483-491. Rudall, P. J., and Chase, M . W. 1996. Systematics of Xanthorrhoeaceae sensu lato: evidence for polyphyly. Telopea 6, 629-647. Rudall, P. J., and Sajo, M . G. 1999. Systematic position of Xyris: flower and seed anatomy. International Journal of Plant Sciences 160, 795-808. Rudall, P. J., Stevenson, D. W., and Linder, H. P. 1999. Structure and systematics of Hanguana, a monocotyledon of uncertain affinity. Australian Systematic Botany 12, 311-330. Rudall, P. J., and Bateman, R. M . 2004. Evolution of zygomorphy in monocot flowers: iterative patterns and developmental constrainsts. New Phytologist 162, 25-44. 218 Rudall, P. J., Stuppy, W., Cunniff, J., Kellogg, E. A., and Briggs, B. G. 2005. Evolution of reproductive structures in grasses (Poaceae) inferred by sister-group comparison with their putative closest living relatives, Ecdeiocoleaceae. American Journal of Botany 92, 1432-1443. Ruhlman, T., Lee, S.-B., Jansen, R. K., Hostetler, J. B., Talon, L. J., Town, C. D., and Daniell, H. 2006. Complete plastid genome sequence of Daucus carota: implications for biotechnology and phylogeny of angiosperms. B M C Genomics 7, 222. doi 10.1186/1471-2164-7-222. Sack, F. D. 1994. Structure of the stomatal complex of the monocot Flagellaria indica. American Journal of Botany 81, 339-344. Sanderson, M . J., Wojciechowski, M . F., Hu, J.-M., Sher Khan, T., and Brady, S. G. 2000. Error, bias and long branch attraction in data for two chloroplast genes in seeds plants. Molecular Biology and Evolution 17, 782-797. Sanderson, M . J., and Shaffer, H. B. 2002. Troubleshooting molecular phylogenetic analyses. Annual Review of Ecology and Systematics 33, 49-72. Saski, C , Lee, S. Daniell, H., Wood, T., Tomkins, J., Kim, H.-G., and Jansen, R. K. 2005. Complete chloroplast genome sequence of Glycine max and comparative analyses with other legume genomes. Plant Molecular Biology 59, 309-322. Savolainen, V., Chase, M . W., Hoot, S. B., Morton, C. M . , Soltis, D. E. , Bayer, C , Fay, M . F., de Bruijn, A. Y., Sullivan, S., and Qiu, Y . -L . 2000. Phylogenetics of flowering plants based on combined analysis of plastid atpB and rbcL gene sequences. Systematic Biology 49, 306-362. 219 Schmitz-Linneweber, C , Maier, R. M . , Alcaraz, J. -P., Cottet, A., Herrmann, R. G., and Mache, R. 2001. The plastid chromosome of spinach (Spinacia oleracea): complete nucleotide sequence and gene organization. Plant Molecular Biology 45, 307-315. Schmitz-Linneweber, C., Regel, R., Du, T. G., Hupfer, H., Herrmann, R. G., and Maier, R. M . 2002. The plastid chromosome of Atropa belladonna and its comparison with that of Nicotiana tobacum: the role of RNA editing in generating divergence in the process of plant speciation. Molecular Biology and Evolution 19,1602-1612. Schultes, R. E. 1951. Plantae Austro-Americanae VII. Botanical Museum Leaflets 15, 29-78. Shinozaki, K., Ohme, M . , Tanaka, M . , Wakasugi, T., Hayashida, N., Matsubayashi, T., Zaita, N., Chunwongse, J., Obokata, J., Yamaguchi-Shinozaki, K., Ohto, C., Torazawa, K., Meng, B. Y., Sugita, M . , Deno, H., Kamogashira, T., Yamada, K., Kusuda, J., Takaiwa, F., Kato, A., Tohdoh, N., Shimada, H., and Sugiura, M . 1986. The complete nucleotide sequence of tobacco chloroplast genome: its gene organization and expression. EMBO Journal 5, 2043-2049. Shipunov, A. B. 2003. The system of flowering plants: synthesis of classical and molecular approaches. Journal of General Biology 64, 501-510. Simpson, M . G. 1985. Pollen ultrastructre of the Philydraceae. Grana 24, 23-31. Simpson, M . G. 1987. Pollen ultrastructure of the Pontederiaceae: evidence for exine homology with the Haemodoraceae. Grana 26, 113-126. Simpson, M . G. 1990. Phylogeny and classification of the Haemodoraceae. Annals of the Missouri Botanical Garden 77, 722-784. 220 Simpson, M . G. 1993. Septal nectary anatomy and phylogeny of the Haemodoraceae. Systematic Botany 18, 593-613. Simpson, D. A. 1995. Relationships within Cyperales. Pp. 497-509. In: Rudall, P. J., Cribb, P. J., Cutler, D. F., and Humphries, C. J. [Eds.]. Monocotyledons: Systematics and Evolution. Royal Botanic Gardens, Kew. Simpson, D. A., Furness, C. A., Hodkinson, T. R., Muasya, A. A., and Chase, M . W. 2003. Phylogenetic relationships in Cyperaceae subfamily Mapaniodeae inferred from pollen and plastid DNA sequence data. American Journal of Botany 90, 1071-1086. Simpson, D. A., Muasya, A. M. , Alves, M . , Bruhl, J. J., Dhooge, S., and Chase, M . W. 2006. Phylogeny of Cyperaceae based on DNA sequence data—a new rbcL analysis. In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.] Monocots: Comparative Biology and Evolution (Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In press. Skottsberg, C. 1932. Bemerkungen uber die Philydraceae. Botanische Jahrbucher fur Systematik, Pflanzengeschichte und Pflanzengeographie. 66, 261-265. Skottsberg, C. 1934. Notes on Orthothylax. Kew Bulletin 3, 97-99. Smith, B. G., and Harris, P. J. 1999. The polysaccharide composition of Poales cell walls: Poaceae cell walls are not unique. Biochemical Systematics and Ecology 27, 33-53. Smith, L. B., and Till, W. 1998. Bromeliaceae. Pp. 74-99 In: Kubitzki, K. [Ed.]. The Families and Genera of Vascular Plants. Vol. 4. Springer-Verlag, Berlin. Soltis, D. E. , Soltis, P. S., Chase, M . W., Mort, M . E., Albach, D. C , Zanis, M . , Savolainen, V., Hahn, W. H., Hoot, S. B., Fay, M . F., Axtell, M . , Swensen, S. M . , Prince, L. M . , Kress, W. J., Nixon, K. C , and Farris, J. S. 2000. Angiosperm phylogeny inferred 221 from 18S rDNA, rbcL, and atpB sequences. Botanical Journal of the Linnean Society 133,381-461. Soltis, D. E. , Soltis, P. S., Endress, P. K., and Chase, M . W. 2005. Phylogeny and Evolution of Angiosperms. Sinauer Associate, Inc., Sunderland, Massachusetts. Soreng, R. J., Davidse, G., Peterson, P. M . , Zuloaga, F. O., Judziewicz, E. J., Filgueiras, T. S., and Morrone, O. 2006. Catalogue of New World Grasses (Poaceae). http://mobot. mobot.org/W3T/Search/nwgc.html [Sept 2006]. Spangler, R., Zaitchik, B, Russo, E. , and Kellogg, E. 1999. Andropogoneae evolution and generic limits in Sorghum (Poaceae) using ndhF sequences. Systematic Botany 24, 267-281. Steane, D. A. 2005. Complete nucleotide sequence of the chloroplast genome from the Tasmanian blue gum, Eucalyptus globulus (Myrtaceae). DNA Research 12, 215-220. Stevens, P. F. 2006. Angiosperm Phylogeny Website, version 7. http://www.mobot. org/mobot/research/apweb/. Stevenson, D. W., and Loconte, H. 1995. Cladistic analysis of monocot families. Pp. 543-578 In: In: Rudall, P. J., Cribb, P. J., Cutler, D. F., and Humphries, C. J. [Eds.]. Monocotyledons: Systematics and Evolution. Royal Botanic Gardens, Kew. Stevenson, D. W., Colella, M . , and Boom, B. 1998. Rapateaceae. Pp. 415-424 In: Kubitzki, K. [Ed.]. The Families and Genera of Vascular Plants, Vol. 4. Springer-Verlag, Berlin. Steyermark, J. A. 1984. Flora of the Venezualan Guyana -1 . Annals of the Missouri Botanical Garden 71, 297-340. 222 Stutzel, T. 1998. Eriocaulaceae. Pp. 197-207 In: Kubitzki, K. [Ed.]. The Families and Genera of Vascular Plants, Vol. 4. Springer-Verlag, Berlin. Suzuki, Y., Glazko, G. V., and Nei, M . , 2002. Overcredibility of molecular phylogenetics obtained by Bayesian phylogenetics. Proceedings of the National Academy of Sciences U.S.A. 99, 16138-16143. Swofford, D. L. , Waddell, P. J., Huelsenbeck, J. P., Foster, P. G., Lewis, P. O., and Rogers, J. S. 2001. Bias in phylogenetic estimation and its relevance to the choice between parsimony and likelihood methods. Systematic Biology 50, 525-539. Swofford, D. L. 2002. PAUP* version 4.0b 10. Computer program and documentation distributed by Sinauer, Sunderland, Mass. Tamura, M . N., Yamashita, J., Fuse, J., and Haraguschi, M . 2004. Molecular phylogeny of the monocotyledons inferred from combined analysis of plastid matK and rbcL gene sequences. Journal of Plant Research 117,109-120. Takhtajan, A. 1997. Diversity and classification of flowering plants. Columbia University Press, New York. Terry, R. G., Brown, G. K., and Olmstead, R. G. 1997. Examination of subfamilial phylogeny in Bromeliaceae using comparative sequencing of the plastid locus ndhF. American Journal of Botany 84, 664-670. Thieret, J. W.,.and Luken, J. O. 1996. The Typhaceae in the Southeastern United States. Harvard Papers in Botany 8, 27-56. Thorne, R. T. 1976. A phylogenetic classification of the Angiospermae. Evolutionary Biology 9, 35-106. 223 Thome, R. T. 1992a. Classification and geography of the flowering plants. Botanical Review 58, 225-348. Thorne, R. F. 1992b. An updated classification of the flowering plants. Aliso 13, 365-389. Tillich, H. J. 1996. Seeds and seedlings in Hanguanaceae and Flagellariaceae (Monocotyledons). Sendtnera3, 187-197. Tillich, H. J., and Sill, E. 1999. Morphological and anatomical systematics of Hanguana Blume (Hanguanaceae) and Flagellaria L. (Flagellariaceae) with a description of a new species, Hanguana bogneri spec. nov. Sendtnera 6, 215-238. Tomlinson, P. B. 1969. Anatomy of the Monocotyledons III. Commelinales-Zingiberales. Clarendon Press, Oxford. Tomlinson, P. B. and Smith, A. C. 1970. Joinvilleaceae, a new family of monocotyledons. Taxon 19, 887-889. Trethewey, J. A. K., Campbell, L. M . , and Harris, P. J. 2005. (1—>3), (l-»4)-B-D-glucans in the cell walls of the Poales (sensu lato): an immunogold labeling study using a monoclonal antibody. American Journal of Botany 92, 1660-1674. Uhle, N. W., Dransfield, J., Davis, J. I., Luckow, M . A., Hansen, K. S., and Doyle, J. J. 1995. Phylogenetic relationships among palms: cladistic analyses of morphological and chloroplast DNA restriction site variation. Pp. 623-661 In: Rudall, P. J., Cribb, P. J., Cutler, D. F., and Humphries, C. J. [Eds.]. Monocotyledons: Systematics and Evolution. Royal Botanic Gardens, Kew. Unwin, M . T. 2004. Molecular systematics of the Eriocaulaceae Martinov. Unpublished PhD dissertation. Miami University, Florida. 224 Venturelli, M . , and Boumann, F. 1986. Embryology and seed development in Mayaca fluviatilis (Mayacaceae). Acta Botanica Neerlandica 35, 497-516. Williams, C. A., Harborne, J. B., Greenham, J., Briggs, B. G., and Johnson, L. A. S. 1997. Flavonoid evidence and the classification of the Anarthriaceae within the Poales. Phytochemistry 46, 1189-1196. Young, N. D., and dePamphilis, C. W. 2005. Rate variation in parasitic plants: correlated and uncorrelated patterns among plastid genes of different function. B M C Evolutionary Biology 5, 16. doi: 10.1186/1471-2148-5-16. Zgurski, J. 2004. Analysis of multiple levels of vascular-plant phylogeny using a nuclear RNA polymerase subunit gene (RPB2) and a large plastid data set: Studies of the tracheophytes, cycads and the lilies and relatives. M . Sc. Thesis. University of Alberta, Edmonton, Alberta, Canada. Zhang, W. 2000. Phylogeny of the grass family (Poaceae) from rp/16 intron sequence data. Molecular Phylogenetics and Evolution 15, 135-146. Zona, S. 2001. Starchy pollen in commelinoid monocots. Annals of Botany 87, 109-116. Zurawski, G., Clegg, M . T., and Brown, A. H. D. 1984. The nature of nucleotide sequence divergence between barley and maize chloroplast DNA. Genetics 106, 735-749. 225 C H A P T E R 5 1 Hydatellaceae are Not Monocots but a Lineage Near the Base of Angiosperm Phylogeny 5.1 INTRODUCTION Although the relationship of angiosperms to other seed plants remains poorly resolved, great progress has been made in identifying the earliest extant splits in flowering-plant phylogeny with the discovery that the New Caledonian shrub Amborella, the water lilies (Nymphaeales), and the woody Austrobaileyales constitute a basal grade of lines that diverged before the main angiosperm radiation (e.g., Graham and Olmstead 2000; Mathews and Donoghue 2000; Zanis et al. 2002; Borsch et al. 2003; Qiu et al. 2005; Leebens-Mack et al. 2005). By focusing attention on these ancient lines, this finding has re-written our understanding of angiosperm structural and reproductive biology, physiology, ecology, and taxonomy (e.g., Doyle and Endress 2000; APG II 2003; Williams and Friedman 2004; Feild et al. 2004). The discovery of a new basal lineage would lead to further re-evaluation of the initial angiosperm radiation, but would also be unexpected, as nearly all of the -460 flowering-plant families have been surveyed in molecular studies (APG II 2003). 1 A version of this chapter is in review: Saarela, J. M., Rai, H. S., Doyle, J. A., Endress, P. K., Mathews, S., Marchant, A. D., Briggs, B. G., and Graham, S. W. Hydatellaceae are not monocots but a lineage near the base of angiosperm phylogeny. 226 One such group that has resisted placement is Hydatellaceae, a small, aquatic family of two genera (Hydatella Diels and Trithuria Hook, f.) restricted to Australasia and India. Hydatella and Trithuria were traditionally included in their own tribe Diplanthereae (=Trithurieae) (Hieronymous 1887; Gilg-Benedict 1930; Hamann 1964) in the superficially similar Centrolepidaceae, a family of highly reduced monocots. The minute reproductive structures of both groups are interpreted as multi-flowered inflorescences subtended by bracts; individual flowers are unisexual, consisting of a single stamen or carpel, with no associated perianth or bract. Despite these similarities, a number of structural differences warranted separation of Hydatellaceae from Centrolepidaceae (Hamann 1976). In fact, many features in Hydatellaceae are rare or unknown among 'core Poales' (a clade consisting of Centrolepidaceae, grasses and five other families; see Chapter 4), including monosulcate pollen, completely anatropous ovules, and typically starchy perisperm (seed storage tissue of nucellar origin). Other characteristics are virtually unknown in monocots (e.g., cellular endosperm development, restricted to the basal monocot genus Acorus). Based on morphological characteristics, a placement of Hydatellaceae within any of the major clades of monocots has therefore been viewed as problematic (e.g., Dahlgren et al. 1985; Hamann 1998). A single rbcL sequence for one species (Trithuria submersa Hook, f.) has been included in several studies, which generally places the family in Poales, seeming to confirm conventional views (e.g., Stevenson et. al. 2000; Michelangeli et al. 2003; Davis et al. 2004; Chase et al. 2006). Most molecular studies, however, have excluded the family due to unavailability of material and/ or poor quality DNA (e.g., Bremer 2002; Givnish et al. 2006). As part of a large-scale phylogenetic survey of grasses and relatives (Chapter 4), we re-examined patterns of variation along the published rbcL sequence from T. submersa, and 227 discovered that it is likely a PCR-based chimera of a grass and a moss sequence (confirmed by Dr. Jerrold Davis, Cornell University; pers. comm.). The objectives of this study are to infer the position of Hydatellaceae among the flowering plants. Using morphology and molecular data for Hydatella and Trithuria, we show that Hydatellaceae have been misinterpreted as monocots, and are instead a highly modified and previously unrecognized ancient lineage of angiosperms. 5.2 M A T E R I A L S AND M E T H O D S 5.2.1. Taxonomic and Genomic Sampling We obtained plastid sequence data from Hydatella (H. inconspicua Cheeseman) and Trithuria (T. submersa), and nuclear sequence data from Trithuria. Methods of DNA extraction, amplification, sequencing, and alignment follow Graham and Olmstead (2000), Borsch et al. (2003), and Graham et al. (2006) for plastid data, and Mathews and Donoghue (2000) for nuclear data. For T. submersa, we sequenced multiple plastid regions involved in photosynthesis (atpB, psbB, psbT, psbN, psbH, rbcL), chlororespiration (ndhF, ndhB), and translation (rpYl, V-rps\2, rpsl) (GenBank accession numbers: DQ915185-DQ915186, DQ915188-DQ915189) using DNA extracted by the Royal Botanic Gardens, Kew [J. G. Conran 961 & P. J. Rudall (ADU)]. Sequences for ndhF (AF547020) and the nuclear gene phytochrome C (PHYC) (DQ981794), as well as an additional rbcL sequence (DQ915187), were sequenced from an extraction prepared by K. Bremer, Uppsala University [A. Doust 1123 (MELU)]. The latter extraction was also used to generate an atpB sequence for T. submersa in an earlier study (AJ419142; Bremer 2002). The two new rbcL sequences, one from each source of DNA, are identical for their 483 bp shared portion; we use the longer of 228 the two here (DQ915188; Conran and Rudall collection). The problematic published rbcL sequence for T. submersa [AF458076.1 (Stevenson et al. 2000; Michelangeli et al. 2003)] was obtained using a different DNA extraction generated from the Doust collection; that DNA extraction was not used here. We obtained sequence data from H. inconspicua for the plastid noncoding regions spanning trnL(UAA)-trnF(GAA) (DQ916291), from cultivated material sourced from Lake Rotokawau, New Zealand [P. D Chapman s.n. (NSW accession 428712)]. We added plastid data for Trithuria submersa (except for trnL-trnF; see below) to an existing alignment used in previous studies (Graham and Olmstead 2000; Rai et al. 2003; Graham et al. 2006) that includes basal angiosperms, eudicots, and monocots. Voucher information and GenBank numbers not previously published for taxa included here are: [Arabidopsis thaliana (L.) Heynh. (Brassicaceae), R.G. Olmstead98-59 (WTU), GenBank numbers: AF238063, AY007488, AF239774, AY007473, AY007458, AF238049, excepts. thaliana atpB and rbcL sequences are from the whole plastid genome (AP000423; Sato et al. 1999)], [Ascarina lucida Hook f. (Chloranthaceae), Peter de Lange 3594 (AK), GenBank numbers: AF238064, AY116647, AF239776, AY007474, AY007459, AF239775, AF238050, AF238051], [Chloranthus japonicus Siebold (Chloranthaceae), M. W. Chase 204 (NCU), GenBank numbers: AF238066, AY007490, AF239778, AY007476, AY007461, AF238053, except C. japonicus atpB (AJ235431.2; Savolainen et al. 2000) and rbcL (L12640; Qiu et al. 1993)], [Gunnera chilensis Lam. (Gunneraceae), no voucher, DNA U. Washington greenhouse (DNA = R.G. Olmstead #98-33), GenBank numbers: AF238068, AY007491, AF239781, AY007478, AY007463, AF238054, except Gunnera atpB (AF093374) and rbcL (AF093724) sequences are from G. hamiltonii Kirk ex W.S. Ham. 229 (Hoot et al. 1999)], and [Schisandra chinensis (Turcz.) Baill. (Schisandraceae), Reznicek 10720 (MICH), GenBank numbers: AF238075, AY007498, AF239791, AY007485, AY007470, AF239790, AF238061, AF238062]. We added the trnL-trnF region from H. inconspicua to a larger published matrix for trnF-trnL-trnF (Borsch et al. 2003), and excluded several mutational hotspots (Borsch et al. 2003). We aligned the T. submersa PHYC sequence with previously published (Mathews and Donoghue 2000; Duvall et al 2006) and two unpublished sequences [(Schisandra sphenathera A. C. Smith (Schisandraceae), Mathews 576 (A), GenBank number DQ981793), and (Brasenia schreberi J.F. Gmelin (Cabombaceae), no voucher, GenBank number DQ981792; see also Mathews and Donoghue (2000))]. 5.2.2. Molecular Phylogenetic Analyses For the plastid-based analyses (excluding trnL-trnF) involving T. submersa, we included all protein-coding genes sampled and several conservative noncoding regions [i.e., introns in rpl2, V-rps\2 and ndhB; intergenic spacers between Y-rps\2-rpsl and rpsl-ndhB (Graham and Olmstead 2000)]. We excluded intergenic spacer from two photosystem gene regions (psbB-psbH; psbE-psbT). The combined analyses also included several regions not yet obtained for Trithuria (i.e., psbD-psbC, psbE-psbJ; intron and second exon of ndhB). For analyses of individual regions (subpartitions atpB, ndhF', psbB-psbH, rbcL, rpl2, and 3'-rps\2-ndhB), we excluded outgroups (Cycas and Ginkgo) to minimize the effect of rooting uncertainty (Graham and Olmstead 2000) on inferred bootstrap support values, as we were interested primarily in the structure of the angiosperm subtree in these analyses. We excluded one variable region in the PHYC alignment from all analyses. 230 We performed heuristic maximum-parsimony (MP) and maximum-likelihood (ML) searches using PAUP* v. 4.0b 10 (Swofford et al. 2002) and P H Y M L v. 2.4.4 (Guindon and Gascuel 2003), respectively, with default settings (e.g., TBR branch-swapping), except that MP searches used 100 random addition replicates. We determined optimal M L models using the likelihood ratio test and the Akaike information criterion, as implemented in ModelTest v. 3.7 (Posada and Crandall 1998); the two methods concurred in each case examined, except for PHYC. The optimal models favoured by the likelihood ratio test and the Akaike information criterion were GTR + Y + I for rbcL, atpB, and combined data; T V M + T + I for ndhY',psbB-K, and 3:' rps\2-ndhB\ and K81uF + T + I for rpll (see ModelTest documentation for details of individual DNA substitution models; Posada and Crandall 1998). The optimal models favoured by the likelihood ratio test and the Akaike information criterion for PHYC were TrN + T + I and GTR + T + I, respectively. The latter is a more general case of the former model, and it is the one that we used for the M L analysis of PHYC. We computed optimal model parameters with PAUP*, based on an MP tree from the corresponding analysis (e.g., Fig. 5.1), and used these values in the heuristic M L searches, except that we determined base frequencies empirically. We estimated branch support using bootstrap analysis, with 200 bootstrap replicates in each case (each with one random addition replicate in the MP analyses). For the large plastid multigene data set that included T. submersa, we also assessed whether several suboptimal root positions (see arrows and asterisks in Fig. 5.1) could be distinguished from the optimal one. We used the Shimodaira-Hasegawa (Shimodaira and Hasegawa 1999) in PAUP* to (simultaneously) compare alternative root placements for the angiosperm subtree in Fig. 5.1. We performed this test using R E L L (resampling estimated 231 log-likelihood) with 1,000 bootstrap replicates, and with all DNA substitution model parameters estimated from the data, except for base frequencies, which were calculated empirically. Results were evaluated as a one-tailed test, with an alpha-level of 0.05. We also analysed six individual plastid regions (atpB, ndh¥,psbB-psbH, rbcL, rpll, and V-rps\l-ndhB) in separate unrooted MP and M L analyses to infer relationships in the absence of a root. To demonstrate that the absence of several regions examined for taxa in the large multigene analysis, but not yet obtained for Trithuria submersa (i.e., psbD, psbC, psbE-psbF-psbh-psM; the intron and second exon of ndhB), has little effect on basic tree inference, we re-ran the heuristic searches of the combined plastid data, considering only the regions obtained for Trithuria. 0 5.2.3. Morphological Phylogenetic Analyses For morphological analyses, we added a consensus of Hydatella and Trithuria to a published data matrix (Doyle and Endress 2000) with modifications (Doyle 2005), and supplemented by data on four- vs. eight-nucleate female gametophytes (Williams and Friedman 2004). Scoring of most characters was based on Hamann (1975, 1976, 1998) and Dahlgren et al. (1985), with data from other sources on anatomy (Cutler 1969; Cheadle and Kosakai 1975),palynology (Ladd 1977), and seedling germination (Cooke 1983b). Characters and character scorings for Hydatellaceae are presented in Table 5.1. Analysis of this data set without Hydatellaceae gives some weakly supported results that are strongly overruled by molecular data [e.g., Nymphaeales are linked with monocots and not located in the basal grade (Doyle and Endress 2000)]. We therefore analyzed this 232 data set using PAUP*, both without constraints and with relationships of other taxa constrained in heuristic MP searches using two backbone trees: one that corresponds to the MP tree from the previous combined morphological and molecular analysis (Doyle and Endress 2000), and a second (see Fig. 5.3) with modifications that take account increasingly robust molecular evidence on relationships in (eu)magnoliids (Zanis et al. 2002; Qiu et al. 2005) (Piperales, Canellales, Magnoliales, Laurales), eudicots (Kim et al. 2004), and monocots (Graham et al. 2006; see also Fig. 5.1). These constraints allow the position of Hydatellaceae to be determined by morphology. For the constrained analysis, we evaluated less than optimal arrangements of Hydatellaceae by searching for trees up to seven steps less parsimonious than the optimal tree length, and by moving taxa manually with MacClade v. 4.03 (Maddison and Maddison 2001). Apart from enforcing topological constraints and performing 100 random addition replicates per analysis, we used default settings for all heuristic searches. For the bootstrap analysis (performed without constraints), we used the settings described for the molecular analysis. We determined unambiguous synapomorphies using the "Trace all changes" tool in MacClade. 5.3 RESULTS 5.3.1. Molecular Analyses Maximum parsimony (MP) and maximum likelihood (ML) analyses of multiple plastid genes and associated noncoding regions identify Hydatellaceae as the sister group of water lilies (Nymphaeales) (BPMP = 80% and B P M L = 100%; Fig 5.1). These analyses include several regions not yet obtained for Trithuria, but exclusion of these regions does not substantially affect tree inference or support values (data not shown). The topology that we 233 recovered for the full data set is also recovered for this reduced data set in the MP analysis (one of two MP trees) and the M L analysis. To corroborate this result using evidence from another genome, we sampled a portion of the nuclear gene phytochrome C (PHYC) from T. submersa and analyzed this sequence with orthologous sequences from other angiosperms that include most of the lineages in Fig. 5.1. In this analysis, we again observe a sister-group relationship between Hydatellaceae and water lilies, with moderate to strong bootstrap support (BPMP = 70% and B P M L = 90%; Fig. 5.2). We also obtained plastid data from Hydatella for two noncoding regions that span the plastid tRNA genes trriL(\] AA) and trnF(GAA) of H. inconspicua. Despite being based on a relatively limited amount of data, phylogenetic analyses again depict Hydatellaceae and Nymphaeales as sister taxa, with moderate bootstrap support (BPMP = 75% and B P M L = 69%; Fig. 5.3). 5.3.2. Morphological Analyses Despite the extensive structural reduction of Hydatellaceae, we were able to score 59% of the morphological characters in the matrix (see Table 5.1). Analysis of the morphological data also showed that Hydatellaceae and Nymphaeales are sister taxa when relationships are not constrained, with 83% bootstrap support (Fig. 5.4); the clade consisting of Hydatellaceae and Nymphaeales is then sister to monocots, but with only 24% bootstrap support (Fig. 5.4). When backbone-relationships among other taxa are constrained to a tree according to recent molecular studies (see Methods), we also find that Hydatellaceae are the sister group of Nymphaeales (Fig. 5.5). We also find similar results when the original topology (Doyle and Endress 2000) is used. 234 A sister-group relationship between Hydatellaceae and Nymphaeales is supported by nine unequivocal synapomorphies (lack of a vascular cambium, anomocytic stomata, truncate anther connective, boat-shaped pollen, inner integument with two cell layers, sclerotic exotesta, seed operculum formed by cell enlargement in the inner integument, perisperm, hypogeal germination). Although most of these (except the seed operculum character state) occur in other taxa, they are not consistently associated. Hydatellaceae remain linked with Nymphaeales in all trees up to five steps less parsimonious than the shortest tree. The next-best position for Hydatellaceae, six steps less parsimonious than the shortest tree, is nested within monocots, as the sister group of Tofieldia and Butomus. A sister-group relationship to all monocots is seven steps less parsimonious. In the latter case, the relationship of Hydatellaceae and monocots would be supported by only three unequivocal synapomorphies (no cambium, boat-shaped pollen, two-layered inner integument), all of which occur as convergences in Nymphaeales. Linear leaves and P2 sieve tube plastids would also support this relationship (they originate twice when Hydatellaceae are linked with Nymphaeales), but would be equivocal as synapomorphies, due to ambiguity concerning positions of character-state changes. Placement of Hydatellaceae in core Poales would require additional reversals (e.g., from orthotropous to anatropous ovules, porate to sulcate pollen, and nuclear to cellular endosperm development). 5.4 DISCUSSION Molecular evidence has been particularly useful in clarifying the phylogenetic positions of groups with highly modified morphologies, such as holoparasites (e.g., Rafflesiaceae, Barkman et al. 2004), mycoheterotrophs (e.g., Petrosavia Becc, Cameron et al. 2003), and 235 certain aquatics (e.g., Hydrostachys Thou., Albach et al. 2001; Podostemaceae, Ueda et al. 1997). Hydatellaceae are also highly modified (i.e., extremely reduced) and have been difficult to place in morphology-based classifications. We find that combined analysis of multiple plastid regions from Trithuria submersa identify Hydatellaceae as the sister group of the water lilies (Nymphaeales; Fig. 5.1), with strong bootstrap support from maximum parsimony (MP) and maximum likelihood (ML) analyses. This surprising result is corroborated by data for PHYC from the nuclear genome (Fig. 5.2). Plastid data from H. inconspicua, representing the other genus in the family, for two noncoding regions that span the tRNA genes trnh(\JAA) and trnY(GAA), also place Hydatellaceae as the sister group of Nymphaeales (Fig. 5.3). Based on these three independent lines of evidence, the placement of Hydatellaceae as the sister group of Nymphaeales indicates that water lilies are part of a larger lineage that evolved more extreme and diverse modifications for life in an aquatic habitat than previously recognized. There is some uncertainty concerning the root of flowering-plant phylogeny (e.g., Soltis et al. 1999; Graham and Olmstead 2000; Mathews and Donoghue 2000; Zanis et al. 2002; Borsch et al. 2003; Qiu et al. 2005; Leebens-Mack et al. 2005), likely a function of the relatively long branch connecting angiosperms to other seed plants. As misrooting can lead to misinference of ingroup relationships (Graham et al. 2002), we determined plausible roots for the combined plastid tree using the Shimodaira-Hasegawa test (see Methods). A root on the Trithuria branch is among those rejected, but trees rooted at two positions not significantly worse than the optimal one (see arrows in Fig. 5.1) also indicate a sister-group relationship between Hydatellaceae and Nymphaeales. Additionally, when each of six distinct plastid regions (atpB, ndhF,psbB-psb¥L, rbcL, rpTl, and V-rpsYl-ndhB) is analysed 236 in separate unrooted MP and M L analyses (see Methods), we consistently observe a branch separating Hydatellaceae and Nymphaeales from all other angiosperms with strong bootstrap support (90-100%; data not shown). Morphological data identify nine unequivocal synapomorpies supporting a sister-group relationship between Hydatellaceae and Nymphaeales (see Results; Fig. 5.5). Some of these features are among those originally used to segregate Hydatellaceae from Centrolepidaceae (Hamann 1976). Several other similarities support a link between Hydatellaceae and Nymphaeales, and a position among the most basal angiosperms. These include ascidiate carpels and a four-nucleate embryo sac, although the latter needs evaluation in the context of studies among related lineages (Williams and Friedman 2004). Both of these features are considered ancestral in angiosperms (Doyle and Endress 2000; Williams and Friedman 2004). Several character states found in Hydatellaceae are scattered across various monocot lineages, including perisperm (e.g., Acorus L. , some Zingiberales; Rudall and Furness 1997), cellular endosperm (e.g., Acorus, but this requires further review; see Rudall and Furness 1997), and monosulcate pollen [e.g., Acorus, Petrosaviaceae (Cameron et al. 2003), some Asparagales (Penet et al. 2005)]. If Hydatellaceae were linked with monocots, the presence of these characters in Hydatellaceae would likely represent parallelisms or reversals. Better information on other characters (see Table 5.1) could affect support for the relationships inferred from morphology here. For example, cotyledon number is unknown in Hydatellaceae, which have a "minute, lens-shaped, incompletely developed embryo" (Hamann 1998), as in some Nymphaeales. Our results have little effect on previous inferences of the growth habit and ecology of the common ancestor of extant angiosperms (Feild et al. 2004), because the closely related 237 Hydatellaceae and Nymphaeales are both aquatic herbs. A rooting of the flowering plants near Amborella and/or water lilies (Nymphaeales) has been robustly supported by numerous analyses of individual and combined sequences from the plastid and other genomes. One contrary result, however, is likely a function of long-branch attraction and low taxon sampling (see Leebens-Mack et al. 2005). If Amborella alone, or a clade consisting of Amborella, Nymphaeales and Hydatellaceae is sister to other angiosperms (see arrows in Fig. 5.1), parsimony optimization implies that the first angiosperms were woody and terrestrial, and that Hydatellaceae and Nymphaeales were an early line that invaded aquatic habitats. The third plausible rooting, in which Hydatellaceae and Nymphaeales are the sister group of all other angiosperms, would imply that the first angiosperms could be either woody and terrestrial or herbaceous and aquatic (Feild et al. 2004). It would be misleading to view Hydatellaceae merely as reduced water lilies. First, Hydatellaceae occupy a different aquatic niche from Nymphaeales. No member of Nymphaeales has a phenotype that comes close to the minute, submergence-tolerant, moss-like habit of Hydatellaceae, evidently a convergence with the distantly related Centrolepidaceae. Hydatella inconspicua can be found growing and flowering at depths of a metre or more. In other species, plants are usually initially submerged, but may flower under water or on drying mud at the edges of seasonal pools or swamps. Second, Hydatellaceae have inflorescences rather than solitary flowers like Nymphaeales, indicating that their common ancestor with Nymphaeales could have had either condition. The Early Cretaceous aquatic genus, Archaefructus Sun, Dilcher, Zheng & Zhou, has been interpreted as the sister group of all extant angiosperms (Sun et al. 2002), or as a reduced aquatic nested within crown-group angiosperms (Friis et al. 2003). A link between Archaefructus and 238 Nymphaeales has been rejected (Crepet et al. 2004), but the possibility of a relationship of Archaefructus with Hydatellaceae should be investigated, since both taxa have inflorescences of naked, unisexual flowers. One recently described species of Archaefructus (A. eoflora) has a denser infloresence structure that is more comparable to that of Hydatellaceae than other species of Archaefructus (Ji et al. 2004) Our current knowledge of Hydatellaceae is limited. The family has only recently been discovered in India (Yadav and Janarthanam 1995), and there is speculation that it has been overlooked elsewhere (Hamann 1998). Half of the ten or so species were described in the past 25 years (Cooke 1981, 1983; Yadav and Janarthanam 1995), and there appears to be substantial morphological variation among them; more species may therefore await discovery. It is clear that monographic work should be a high priority. Phylogenetic relationships need to be characterized within the family, in addition to studies of the developmental morphology, aquatic ecology, biogeography, and conservation status of these curious and neglected basal angiosperms. 239 Table 5.1 Morphological characters and character states scored for Hydatellaceae. See Doyle and Endress (2000) and Doyle (2005) for more information on characters and character states. Character: character states Hydatellaceae 1 Habit: Tree or shrub, rhizomatous or scandent rhizomatous or scandent 2 Stele: eustele, siphonostele, monocot type ? 3 Inverted vascular bundles: absent, present ? 4 Protoxylem lacunae: absent, present siphonostele 5 Cambium: present, absent absent 6 Storied structure: absent, present ? 7 Vessels: tracheids, porose pits, vessels porose pits 8 Vessel grouping: solitary, multiples ? 9 Vessel perforations: scalariform, mixed, simple only scalariform 10 Fiber pits: bordered, reduced ? 11 Ray width: narrow, wide ? 12 Paratracheal parenchyma: absent or scanty, well developed . ? 13 Tangential parench bands: absent, present ? 14 Pith: uniform, septate uniform 15 Secondary phloem: simple, stratified 16 Sieve tube plastids: starch, PI type, P2 type P2 type 17 Pericycle: separate fiber bundles, continuous fiber ring, composite with U sclereids, no sclerenchyma ? 18 Laticifers: absent, present absent 19 Raphide idioblasts: absent, present absent 20 Phyllotaxy: spiral, distichous, opposite ? 21 Nodes: multilacunar, one trace unilacunar, two trace unilacunar, trilacunar ? 22 Prophylls: paired lateral, single ? 23 Stipules: absent, adaxial axillary, interpetiolar absent 24 Axillary squamules: absent, present ? 25 Leaf blade: bifacial, unifacial ? 26 Blade shape: obovate to elliptical, ovate, linear linear 27 Major venation: pinnate, palmate or basally crowded ? 28 Base of blade: not peltate, peltate ? 29 Leaf dissection: simple, lobed, or dissected simple 30 Marginal teeth: absent, chloranthoid, monimioid, platanoid ? 31 Stomata: paracytic, laterocytic, anomocytic, stephanocytic anomocytic 32 Midrib vasculature: simple arc, arc with adaxial plate ring ? 33 Palisade parenchyma: absent, present absent 34 Asterosclerids: absent, present absent 35 Oil cells: absent, present absent 36 Mucilage cells: absent, present absent 37 Inflorescence: solitary, raceme or spike or botryoid, richly branched raceme spike botryoid 38 Sex of flowers: bisexual, bisexual and unisexual, unisexual unisexual 240 Table 5.1 (continued) Character: character states Hydatellaceae 39 Hypanthium: absent, present, inferior ovary 40 Perianth phyllotaxy: spiral, whorled 41 Perianth whorls: more than two, two, one, absent 42 Perianth merosity: irregular, threes, twos fours or fives 43 Outer perianth cycle: undifferentiated, sepaloid 44 Calyx fusion: separate or basally fused, calyptrate 45 Nectar petals: absent, present 46 Stamen phyllotaxy: spiral, whorled, irregular 47 Stamen merosity: irregular, threes, twos fours or fives 48 Stamen fusion: free, connate 49 Stamen base: short, long and wide, long and narrow filament 50 Paired stamen glands: absent, present 51 Connective apex: extended, truncate or rounded 52 Microsporangia: four, two 53 Pollen sacs: protruding, embedded 54 Orientation of dehiscence: introrse, latrorse to slightly introrse, extrorse 55 Dehiscence mode: longitudinal slit, H valvate, upward opening flaps 56 Connective hypodermis: unspecialized, endothecial 57 Tapetum: secretory, amoeboid 58 Microsporogenesis: simultaneous, successive 59 Pollen unit: monads, tetrads 60 Pollen shape: boat shaped, globose 61 Apertures: monosulcate etc., inaperturate sulcuate, Garside tri or hexacolpate, tricolpate 62 Pollen size: large >50 micrometers, medium, small <20 micrometers 63 Infratectum: granular, intermediate, columellar 64 Tectum: continuous or microperforate, perforate to semitectate, reduced 65 Striate muri: absent, present 66 Supratectal spinules: absent, present 67 Prominent spines: absent, present 68 Aperture membrane: smooth, sculptured 69 Nexine stratification: homogeneous foot layer, footlayer and endexine, foliated nexine 70 Nexine thickness: absent or discontinuous, thin (less than 1:3), thick (1:3 or more) 71 Inner staminodes: absent, present 72 Carpel number: more than one, one 73 Carpel form: ascidiate, intermediate, plicate 74 Carpel sealing: by secretion, mixed partial PGF, mixed full PGF, postgenital fusion 75 Pollen tube transm. tissue: not differentiated, one layer diffi, multilayered ? ? absent ? ? ? ? ? ? ? long and narrow filament absent truncate or rounded four ? latrorse to slightly introrse longitudinal slit unspecialized ? ? monads boat shaped monosulcate small ? continuous absent present absent ? 9 one ascidiate ? 9 241 Table 5.1 continued Character: character states Hydatellaceae 76 Style: absent - sessile stigma, present 77 Stigma: extended, restricted 78 Stigma papillae: unicellular or smooth, uniseriate pericellular, pluriseriate 79 Compitum: absent, extragynoecial 80 Carpel fusion: apocarpous, paracarpous, eusyncarpous 81 Oil cells in carpels: not visible, intrusive 82 Septal nectaries: absent, present 83 Ovule number: one, mostly two, more than two 84 Placentation: linear, laminar, diffuse 85 Ovule direction: pendent, horizontal, ascendent 86 Ovule curvature: anatropous, orthotropous, hemitropous 87 Integuments: two, one 88 Outer integ shape: semiannular, annular 89 Outer integ lobation: unlobed, lobed 90 Outer integ thickness: 2 cells, 2 and 3 to 4, 4 and 5 or more 91 Inner integ thickness: 2 cells, 2 and 3 or 3, 3 and more 92 Chalaza: unextended, pachychalazal, perichalazal 93 Nucellus: crassinucellar, tenuinucellar, or pseudocrassi 94 Fruit wall: fleshy, fleshy with endocarp, dry 95 Fruit dehiscence: indehiscent, dehiscent 96 Testa: nonmultiplicative, multiplicative 97 Exotesta: normal, palisade or shorter sclerotic, tabular 98 Mesotesta: unspecialized, sclerotic, fibrous, sarcotesta spongy 99 Endotesta: unspecialized single lignified layer, multiple lignified layer, tracheidal, palisade or shorter sclerotic 100 Tegmen: unspecialized, both layers sclerotic, fibrous, exotegmen 101 Ruminations: absent, testal, tegminal, or chalazal 102 Operculum: absent, present 103 Ari l : absent, present 104 Endosperm development: cellular, nuclear, helobial 105 Endosperm in seed: present, absent 106 Perisperm: absent, present 107 Embryo: minute, large 108 Cotyledons: two, one 109 Germination: epigeal, hypogeal 110 Embryo sac: 4 nucleate, 8 or 9 nucleate absent-sessile stigma ? uniseriate pericellular apocarpous not visible absent one ? pendent anatropous two ? ? 2 cells 2 cells unextended crassinucellar dry indehiscent/ dehiscent nonmultiplicative palisade or shorter sclerotic ? unspecialized single absent absent present absent cellular present present minute ? hypogeal 4 nucleate 242 Figure 5.1. Phylogenetic placement of Trithuria submersa (Hydatellaceae) in angiosperms according to plastid data (multiple protein-coding genes and several noncoding regions; see text). MP and M L analyses yield the same topology (12,607 steps; -InL = 88,820.974). Branch lengths are MP estimates (ACCTRAN optimization). Two outgroup taxa, Cycas and Ginkgo, have been trimmed for clarity (with the stem lineage shortened). Bootstrap support values are noted near branches (MP above or to the left, M L below or to the right). The optimal and two suboptimal roots (indicated with arrows) could not be distinguished from each other in a Shimodaira-Hasegawa test that simultaneously considered these and four additional candidate roots (indicated with asterisks; the latter four cases rejected at P < 0.05). 243 5.1 Amborella Trithuria 1°P_ 100 i Cabomba Nymphaea Austrobai leya Tr imenia 100 changes | Ambore l laceae I Hydatellaceae Nymphaeales (water lilies) Austrobaileyales Chloranthaceae Magnoliids (eumagnoliids) | Ceratophyllaceae Eudicots Arab idops is Oryza Typha Monocots Poa les 2 4 4 Figure 5.2. Phylogenetic placement of Trithuria submersa (Hydatellaceae) in angiosperms according to analyses of the nuclear phytochrome C locus (PHYC). The best MP tree (tree length: 4,074 steps) is shown; the M L tree (tree length: -InL = 17,094.891) had the following general topology: (Amborella, ((Hydatellaceae, Nymphaeales), (Austrobaileyales, (Chloranthaceae, ((Canellales, (Laurales,Magnoliales)), (Monocots, (Piperales,Eudicots)) ))))), and with Nelumbo sister to (Arabidopsis, Solarium), but otherwise with the same structure within each clade as the MP tree. Branch lengths are MP estimates (ACCTRAN optimization). Bootstrap support values are noted near branches (MP above or to the left, M L below or to the right). 245 Fig 5.2 Amborella 70 90 99 100 100 100 Trithuria — Nymphaea Brasenia Cabomba CI i Amborellaceae • Hydatellaceae Nymphaeales (water lilies) 100 87 83 62 86 50 changes 100 63 56 66 98 100 99 100 16 12 26 16 23 78 67 30 51 20,27 97 86 89 Pleea • Tofieldia Dioscorea Sagittaria 81 63 — Lilium Oryza Austrobaileya — Illicium — Schisandra 92 90 44 65 • Asarum — Saururus Pteridophyllum 82 92 Nelumbo 88 Solanum — Arabidopsis Canella 100 100 Drimys Hedycarya Knema 92 90 • Eupomatia • Magnolia 100 100 Hedyosmum Sarcandra Monocots Austrobaileyales Eudicots Magnoliids (eumagnoliids) Chloranthaceae 246 Figure 5.3. Phylogenetic placement of Hydatella inconspicua (Hydatellaceae) for analyses of the trnT-trnL-trnF region, including the trnL-trnF region for H. inconspicua. One of two MP trees (tree length: 3,292 steps) is shown; the arrow indicates a branch that collapses in the strict consensus of the MP trees. The M L tree (tree length: -InL = 19,968.977) has several additional differences in topology (i.e., Nuphar is depicted as sister to the remaining water lilies; Chloranthaceae sister to magnoliids). Branch lengths are MP estimates (ACCTRAN optimization). Three outgroup taxa (Araucaria, Ginkgo and Pinus) have been trimmed for clarity, with the stem lineage shortened. Bootstrap support values are noted near branches (MP above or to the left, M L below or to the right). 247 Fig. 5.3 100_ "100 A m b o r e l l a 75 84 78 69 98 83 84 94 99 97 50 changes 100 Hydatella — Cabomba 100 100 100 100 64,56 100 100 45 Brasenia Nuphar lutea Nuphar advena 1001 Victoria 100 I Nymphaea - Austrobaileya 1 lllicium • Schisandra Amborellaceae Hydatellaceae Nymphaeales (water lilies) Austrobaileyales 72 57 61 91 94 . Ceratophyllum Acorus Nypa 86 -Tofieldia -Orontium • Chloranthus 100 78 Dicentra „ f r J H Platanus - Nelumbo • Trochodendron Buxus Aextoxicon Calycanthus Umbellularia Myristica 100 100 • Annona u I™ 67,100|_r-1 6 100,1001. 62,50 97 78 Asimina Liriodendron Magnolia Michelia Drimys Canella Saruma Aristolochia Lactoris 1001 [Tool "I 100 Saururus Houttuynia j Piper angustum Piper sp. Ceratophyllaceae Monocots 1 Chloranthaceae Eudicots Magnoliids (eumagnoliids) 248 Figure 5.4. Most parsimonious position of Hydatellaceae based on morphology, with relationships of other taxa not constrained (see Methods). Bootstrap support values are noted near branches. Arrows indicate branches that collapse in a strict consensus tree. 249 Fig. 5.4 Amborella 52 i Hedyosmum P t^— Ascarina 8 1 *— Sarc/Chloranthus 30 Trimenia — Austrobaileya - Illicium <5 13 Schisandraceae i Winteraceae - Canellaceae Amborel laceae Chloranthaceae Austrobaileyales Myristicaceae Magnoliaceae Annonaceae Degeneria Eupomatia 23 ' Galbulimima Nelumbo Euptelea - Buxaceae Monimioideae Mollinedioideae Gomortega j — Lauraceae -\ I— Hernandioideae g4 I— Gyrocarpoideae 100 Menispermaceae Lardizabalaceae Berberidaceae Ranunculaceae Papaverales Aristolochioideae Asaroideae Lactoris 5 changes C- Piperaceae Saururaceae Hydatellaceae Cabombaceae Nuphar ^ 8 4 Dioscoreales Tofieldia Butomus \ i Barclaya Nymphaeoideae — Acorus Araceae Canel la les Magnoliales Laurales Eudicots Piperales Hydatellaceae Nymphaeales (water lilies) Monocots 2 5 0 Figure 5.5. Most parsimonious position of Hydatellaceae based on morphology (774 steps), with relationships of other taxa constrained (see Methods). Numbers of unequivocal changes (with respect to the root indicated) are noted near each branch. For clarity, most phylogenetic structure within magnoliids (i.e., within Canellales, Piperales, Laurales and Magnoliales) has been excluded [see Doyle and Endress (2000) for details]. 251 Fig. 5.5 Ambore l la I Hydatellaceae I Cabombaceae Nuphar Barclaya Nymphaeoideae Austrobaileya Trimenia icium Schisandraceae Hedyosmum Sarc/Chloranthus Ascarina Trochodendraceae Buxaceae Nelumbo Platanaceae Proteaceae Euptelea Papaverales Lardizabalaceae Menispermaceae Berberidaceae Ranunculaceae Acorus Dioscoreales Araceae Butomus Tofieldia Cane l la les P ipera les Magno l ia les Laura les Amborel laceae Hydatellaceae Nymphaeales (water lilies) Austrobaileyales Chloranthaceae Eudicots Monocots Magnoli ids (eumagnoliids) 2 5 2 5.4 R E F E R E N C E S Albach, D. C , Soltis, D. E. , Chase, M . W., and Soltis, P. S. 2001. Phylogenetic placement of the enigmatic angiosperm Hydrostachys. Taxon 50, 781-805. Angiosperm Phylogeny Group (APG II). 2003. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants. Botanical Journal of the Linnean Society 141, 399-436. Barkman, T. J., Lim, S. H., Salleh, K. M . , and Nais, J. 2004. Mitochondrial DNA sequences reveal the photosynthetic relatives of Rafflesia, the world's largest flower. Proceedings of the National Academy of Sciences U.S.A. 101, 787-792. Behnke, H. -D. 2000. Forms and sizes of sieve-element plastids and evolution of the Monocotyledons. Pp. 163-188 In: Wilson, K. L., and Morrison, D. A. [Eds.]. Monocots: Systematics and Evolution. CSIRO Publishing, Collingwood, Australia. Borsch, T., Hilu, K., W., Quandt, D., Wilde, V. , Neinhuis, C , and Barthlott, W. 2003. Noncoding plastid tmT-trnY sequences reveal a well resolved phylogeny of basal angiosperms. Journal of Evolutionary Biology 16, 558-576. Bremer, K. 2002. Gondwanan evolution of the grass alliance of families (Poales). Evolution 56, 1374-1387. Burleigh, J. G., and Mathews, S. 2004. Phylogenetic signal in nucleotide data from seed plants: implications for resolving the seed plant tree of life. American Journal of Botany 91, 1599-1613. Cameron, K. M . , Chase, M . W., and Rudall, P. J. 2003. Recircumscription of the monocotyledonous family Petrosaviaceae to include Japonolirion. Brittonia 55, 214-225. 253 Cheadle, V. I., and Kosakai, H. 1975. Vessels in Juncales: II. Centrolepidaceae and Restionaceae. American Journal of Botany 62, 1017-1026. Cooke, D. A. 1981. New species of Schoenus (Cyperaceae) and Trithuria (Hydatellaceae). Muelleria 4, 299-303. Cooke, D. A. 1983a. Two western Australian Hydatellaceae. Muelleria 5, 123-125. Cooke, D. A. 1983b. The seedling of Trithuria (Hydatellaceae). Victorian Naturalist (Melbourne) 100, 68-69. Crepet, W. L., Nixon, K. C , and Gandolfo, M . A. 2004. Fossil evidence and phylogeny: the age of major angiosperm clades based on mesofossil and macrofossil evidence from Cretaceous deposits. American Journal of Botany 91, 1666-1682. Cutler, D. F. 1969. Anatomy of the Monocotyledons. IV. Juncales. Oxford University Press, United Kingdom. Dahlgren, R. M . T., Clifford, H. T. and Yeo, P. F. 1985. The Families of the Monocotyledons: Structure, Evolution, and Taxonomy. Springer-Verlag, Berlin. Doyle, J. A. 2005. Early evolution of angiosperm pollen as inferred from molecular and morphological phylogenetic analyses. Grana 44, 227-251. Doyle, J. A., and Endress, P. K. 2000. Morphological phylogenetic analysis of basal angiosperms: comparison and combination with molecular data. International Journal of Plant Sciences 161 (6 Suppl.), S121-S153. Duvall, M . , Mathews, S., Mohammad, N . A., and Russell, T. 2006. Placing the monocots; conflicting signal from trigenomic analyses. Pp. 77-88 In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative 254 Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California. In press. Feild, T. S., Arens, N. C , Doyle, J. A., Dawson, T. E. , and Donoghue, M . J. 2004. Dark and disturbed: a new image of early angiosperm ecology. Paleobiology 30, 82-107. Friis, E. M . , Doyle, J. A., Endress, P. K., and Leng, Q. 2003. Archaefructus - angiosperm precursor or specialized early angiosperm? Trends in Plant Science 8, 369-373. Gilg-Benedict, T. 1930. Centrolepidaceae. In: Engler and Prantl, Centrolepidaceae Nat. Pflanzenfam 2 n d edn, 15a: 27-33. Leipzig, W. Engelmann. Graham, S. W., and Olmstead, R. G. 2000. Utility of 17 chloroplast genes for inferring the phylogeny of the basal angiosperms. American Journal of Botany 87, 1712-1730. Graham, S. W., Olmstead, R. G., and Barrett, S. C. H. 2002. Rooting phylogenetic trees with distant outgroups: a case study from the commelinoid monocots. Molecular Biology and Evolution 19, 1769-1781. Graham, S. W., Zgurski, J. M . , McPherson, M . A., Cherniawsky, D. M . , Saarela, J. M . , Home, E. S. C , Smith, S. Y., Wong, W. A., O'Brien, H., Biron, V. L., Pires, J. C , Olmstead, R. G., Chase, M . W., and Rai, H. S. 2006. Robust inference of monocot deep phylogeny using an expanded multigene plastid data set. Pp. 3-20 In: Columbus, J. T., Friar, E. A., Porter, J. M. , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California. Guindon, S., and Gascuel, O. 2003. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Systematic Biology 52, 696-704. 255 Hamann, U. 1975. Neue Untersuchungen zur Embryologie und Systematik der Centrolepidaceae. Botanische Jahrbucher fur Systematik, Pflanzengeschichte und Pflanzengeographie 96, 154-191. Hamann, U. 1976. Hydatellaceae-a new family of Monocotyledoneae. New Zealand Journal of Botany 14, 193-196. Hamann, U. 1998. Hydatellaceae. Pp. 231-234 In: Kubitzki, K. [Ed.]. Flowering Plants. Monocotyledons: Lilianae (except Orchidaceae). Springer-Verlag, Berlin. Hoot, S. B., Magallon, S., and Crane, P. R. 1999. Phylogeny of basal eudicots based on three molecular datasets: atpB, rbcL, and 18S nuclear ribosomal DNA sequences. Annals of the Missouri Botanical Garden 86, 1-32. Ji, Q., Li , H. Q., Bowe, L. M . , Liu, Y. S., Taylor, D. W. 2004. Early cretaceous Archaefructus eojlora sp. nov. with bisexual flowers from Beipiao, Western Liaoning, China. Acta Geologica Sinica - English Edition 78, 883-896. Kim, S., Soltis, D. E. , Soltis, P. S., Zanis, M . J., and Suh, Y. 2004. Phylogenetic relationships among early-diverging eudicots based on four genes: were the eudicots ancestrally woody? Molecular Phylogenetics and Evolution 31, 16-30. Ladd, P. G. 1977. Pollen morphology of some members of the Restionaceae and related families, with notes on the fossil record. Grana 16, 1-14. Leebens-Mack, J., Raubeson, L. A., Cui, L. , Kuehl, J. V., Fourcade, M . H., Chumley, T. W., Boore, J. L. , Jansen, R. K., and dePamphilis, C. W. 2005. Identifying the basal angiosperm node in chloroplast genome phylogenies: sampling one's way out of the Felsenstein zone. Molecular Biology and Evolution 22, 1948-1963. 256 Maddison, D. R., and Maddison, W. P. 2001. MacClade 4: Analysis of Phylogeny and Character Evolution, Version 4.03. Sinauer Associates, Sunderland, Massachusetts Mathews, S., and Donoghue, M . J. 2000. Basal angiosperm phylogeny inferred from duplicate phytochromes A and C. International Journal of Plant Sciences 161 (6 Suppl.), S41-S55. Michelangeli, F. A., Davis, J. I., and Stevenson, D. W. 2003. Phylogenetic relationships among Poaceae and related families as inferred from morphology, inversions in the plastid genome, and sequence data from the mitochondrial and plastid genomes. American Journal of Botany 90, 93-106. Penet, L., Nadot, S., Ressayre, A. , Forchioni, A. , Dreyer, L. , and Gouyon, P. H. 2005. Multiple developmental pathways leading to a single morph: monosulcate pollen (Examples from the Asparagales). Annals of Botany 95, 331-343. Posada, D., and Crandall, K. A. 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14, 817-818. Qiu, Y. L., Chase, M . W., Les, D. H., and Parks, C. R. 1993. Molecular phylogenetics of the Magnoliidae - cladistic analyses of nucleotide-sequences of the plastid gene rbcL. Annals of the Missouri Botanical Garden 80, 587-606. Qiu, Y . -L , Dombrovska, O., Lee, J., Li , L. , Whitlock, B. A., Bernasconi-Quadroni, F., Rest, J. S., Davis, C. C , Borsch, T., Hilu, K. W., Renner, S. S., Soltis, D. E. , Soltis, P. S., Zanis, M . J., Cannone, J. J., Gutell, R. R., Powell, M . , Savolainen, V. , Chatrou, L. W., and Chase, M . W. 2005. Phylogenetic analyses of basal angiosperms based on nine plastid, mitochondrial, and nuclear genes. International Journal of Plant Sciences 166, 815-842. 257 Rai, H.S., O'Brien, H E . , Reeves, P. A., Olmstead, R. G., and Graham, S. W. 2003. Inference of higher-order relationships in the cycads from a large chloroplast data set. Molecular Phylogenetics and Evolution 29, 350-395. Rudall, P. J., and Furness, C. A. 1997. Systematics of Acorus: ovule and anther. International Journal of Plant Sciences 158, 640-651. Sato, S. S., Nakamura, Y., Kaneko, T., Asamizu, E. , and Tabata, S. 1999. Complete structure of the chloroplast genome of Arabidopsis thaliana. DNA Research 6, 283-290. Savolainen, V., Chase, M . W., Hoot, S. B., Morton, C M . , Soltis, D. E. , Bayer, C , Fay, M . F., De Bruijn, A., Sullivan, S., and Qiu, Y . -L . 2000. Phylogenetics of flowering plants based upon a combined analysis of plastid atpB and rbcL gene sequences. Systematic Biology 49, 306-362. Shimodaira, H., and Hasegawa, M . 1999. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Molecular Biology and Evolution 16, 1114-1116. Soltis, P. S., Soltis, D. E., and Chase, M . W. 1999. Angiosperm phylogeny inferred from multiple genes as a tool for comparative biology. Nature 402, 402-404. Stevenson, D. W., Davis, J. I., Freudenstein, J. V., Hardy, C. R., Simmons, M . P., and Specht, C. D. 20000. Pp. 17-24 In: Wilson, K. L., and Morrison, D. A [Eds.]. Monocots: Systematics and Evolution. CSIRO Publishing, Collingwood, Australia. Sun, G., Qi, J., Dilcher, D. L., Zheng, S., Nixon, K. C , and Wang, X. 2002. Archaefructaceae, a new basal angiosperm family. Science 296, 899-904. 258 Swofford, D. L. 2002. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Computer program and documentation. Sinauer, Sunderland, Massachusetts. Ueda, K., Hanyuda, T., Nakano, A., Shiuchi, T., Seo, A, Okubo, H., and Hotta, M . 1997. Molecular phylogenetic position of the Podostemaceae, a marvelous aquatic flowering plant family. Journal of Plant Research 110, 87-92. Williams, J. H., and Friedman, W. E. 2004. The four-celled female gametophyte of Illicium (Illiciaceae; Austrobaileyales): implications for understanding the origin and early evolution of monocots, eumagnollids, and eudicots. American Journal of Botany 91, 332-351. Yadav, S. R., and Janarthanam, M . K. 1995; Trithuria konkanensis (Hydatellaceae), eine neu Art aus Indien. Aqua-Planta 3,91-101. Zanis, M . J., Soltis, D. E., Soltis, P. S., Mathews, S., and Donoghue, M . J. 2002. The root of angiosperms revisited. Proceedings of the National Academy of Sciences U.S.A. 99, 6848-6853. 259 C H A P T E R 6 Conclusion 6.1 CONCLUSIONS AND F U T U R E DIRECTIONS Most previous phylogenetic studies of monocots have been based on only one or a few genes or regions, and multiple branches of the monocot tree of life have not yet been resolved clearly or supported strongly. In this thesis, I have used an expanded plastid data set to make phylogenetic inferences among major commelinid monocot lineages, among families within the commelinid orders Commelinales and Poales, and among major lineages in the grass family, Poaceae. I also conducted a phylogenetic study of the grass genus Bromus, using plastid and nuclear data. In line with several previous studies that used the same plastid regions examined here (Graham and Olmstead 2000; McPherson 2003; Zgurski 2004; Graham et al. 2006), my analyses of higher-order relationships in the commelinid monocots generally provide increased support for several relationships that have been inferred previously (e.g., Chase et al. 2006; Givnish et al. 2006; Graham et al. 2006), and provide support for some relationships not inferred previously, including the position of Hydatellaceae outside of the monocots. My study of Bromus has provided new insights into the identities of major lineages in the genus, their interrelationships, and possible insight into past evolutionary processes acting in the genus, such as hybridization and polyploidy. Nonetheless, several higher-order commelinid relationships that have been difficult to infer in the past, and most aspects of recent relationships in Bromus, remain recalcitrant to clear and strong phylogenetic inference here. 260 In Chapter 2,1 characterized phylogenetic relationships in the morphologically diverse and taxonomically complex grass genus Bromus, using sequence data from two plastid loci and a nuclear locus for 46 species (a third of known species-level diversity in the genus). Based on the nuclear ribosomal data, sects. Bromus (including sect. Triniusia), Genea, Neobromus, and Ceratochloa are monophyletic, and sect. Bromopsis comprises several distinct lineages. Plastid trees indicate that sects. Genea and Bromus are closely related, and the incongruence detected between the plastid and nuclear ribosomal data possibly supports a hybrid origin for the B. pectinatus complex between sects. Bromus and Genea. Plastid trees indicate a close relationship between Old World and some North American species of sect. Bromopsis, and the plastid and nuclear ribosomal data indicate that one South American species of sect. Bromopsis is not closely related to North American and Eurasian species traditionally classified in the same section. Most species of Bromus sampled had levels of sequence variation too low to allow complete resolution of relationships among close relatives at the species level. Recognition of the brome grasses as one distinct genus, Bromus, is in agreement with my molecular data, but current classification schemes do not satisfactorily reflect phylogenetic relationships within the genus, particularly with respect to the circumscription of sect. Bromopsis. However, before a revised infrageneric classification of Bromus is proposed, substantially better sampling should be conducted of different genomic regions, to obtain better support for phylogenetic relationships among taxa, and of taxa, to further clarify incongruence among different nuclear and plastid data partitions to more adequately sample the molecular, morphological, and geographical variability in the genus. Through this phylogenetic work on Bromus, I also discovered, described, and characterized the phylogenetic position of a new species from 261 Peru, Bromus ayacuchensis Saarela and P. M . Peterson in ed. (Saarela et al. in press), and I identified and clarified the taxonomic status of an endemic Californian species [Bromus hallii (Hitchc.) Saarela & P. M . Peterson; Saarela et al. 2005). In Chapters 3 and 4,1 characterized phylogenetic relationships among 10-11 of the twelve major lineages (subfamilies) of grasses. I observe general concordance in topology among most partitions of the data, and among different methods of phylogenetic analysis, but there are several instances of moderate to strong conflict (primarily between the protein coding and the noncoding data partitions) in various parts of the tree. My data provide strong support for relationships along much of the spine of Poaceae, congruent with the findings of a recent multigene study (GPWG 2001). Major findings of this study include strong support for the monophyly of Anomochlooideae, strong support for the position of Puelioideae as the third of three successive lineages that are the sister group of the rest of the grasses, substantially increased support from plastid data for the monophyly of the BEP clade, and moderate support for a sister-group relationship between Pooideae and Bambusoideae. Given this latter finding, I suggest that comparative genomic studies of the grasses (e.g., Xu et al. 2005) could benefit from analysis of a bamboo genome, which may provide additional insight into the evolution of Pooideae cereal crop genomes (e.g., Avena L. , Hordeum L. , Triticum L. , Secale L.) compared with the more distantly related rice genome (Ehrhartoideae). As in previous studies, most relationships among subfamilies within the P A C M A D clade are not inferred consistently or strongly here, perhaps a consequence of short internal branches. An Aristidoideae-Chloridoideae clade is recovered with moderate to strong support in most analyses, but the noncoding data do not recover this group. I have not sampled Micrairoideae, a recently recognized subfamily that is part of the P A C M A D clade 262 (Duvall et al. 2006). Future work should sample this lineage for the regions examined here. Nonetheless, substantial work remains to unravel the phylogenetic history of the diverse P A C M A D clade. A simple and potentially effective addition to the current analysis would be to sample more than one taxon per subfamily of Poaceae in future studies, and more data might be necessary in a subset of cases (e.g., within the P A C M A D clade). In Chapter 4,1 broadly characterized phylogenetic relationships in the commelinid moncots, a large clade that includes about a third of monocot diversity. In line with recent studies (e.g., Chase et al. 2006; Givnish et al. 2006; Graham et al. 2006), I inferred a strongly supported sister group relationship between Commelinales and Zingiberales. As in earlier studies, I was not able to infer the relative positions of Arecales, Dasypogonaceae, Commelinales-Zingiberales, and Poales with clear or strong support. Placements for these ancient lineages will likely require substantially more data per taxon, and possibly complete plastid genomes, which have helped clarify the positions of some taxa (e.g., Vitaceae; Jansen et al. 2006). Within Commelinales, I identified two major clades: one comprising Commelinaceae and Hanguanceae, the other Haemodoraceae, Philydraceae, and Pontederiaceae. In the latter clade, I inferred Philydraceae - a family whose exact position in Commelinales has been most controversial - to be the sister group of Haemodoraceae-Pontederiaceae, with moderate support, similar to that found in the only previous study to also infer this position (Chase et al. 2006). I also sampled each of the four genera in Philydraceae, and provide the first complete genus-level phylogeny for the family. I found Philydrella to be the sister group of a Helmholtzia-Philydrum clade with moderate support, a relationship that is seemingly consistent with several aspects of morphology in the family. 263 Within Poales, I identified a major clade that includes all Poales families except Bromeliaceae and Typhaceae, consistent with the findings of recent studies (e.g., Chase et al. 2006; Givnish et al. 2006; Graham et al. 2006). As in previous studies, the exact positions of Bromeliaceae and Typhaceae at the base of the Poales subtree are not resolved here. In line with previous studies, my analyses infer a strongly supported cyperid clade, in which Thurniaceae (Thurnia and Prionium) are the sister group of Cyperaceae-Juncaceae. However, the relationship of the cyperid clade to other Poales remains unresolved. The composition and relationships among taxa in the xyrid clade differ between my parsimony and likelihood analyses. With parsimony I find a weakly supported xyrid clade that includes Xyridaceae as the sister group of a more-strongly supported Eriocaulaceae-Mayacaceae clade, whereas with likelihood I find a strongly supported Eriocaulaceae-Xyridaceae clade that is the sister group of the core Poales, and the position of Mayacaceae is not resolved. These different relationships are likely a function of long branches in each of these families, particularly in the aquatic genus Mayaca. As likelihood methods are less prone to long-branch artifacts, it seems reasonable to have more confidence in the less certain placement for Mayacaceae inferred by this method. The close relationship inferred between Eriocaulaceae and Xyridaceae in likelihood analysis here has also been found in numerous other studies, and it is also supported by several putative morphological synapomorphies. Clearly, the position of Mayacaceae remains unestablished. A possible solution for this problem would be to sample additional species in each of these three families that potentially span the root node in each family, which might help break up long branches. I identify a core Poales clade with strong support that includes a graminid clade and a restiid clade. Within the restiid clade, Anarthriaceae sensu APG II (2003) (i.e., including 264 Hopkinsia and Lyginia) are strongly supported as the sister group of Centrolepidaceae-Restionaceae. With only one exemplar sampled from Restionaceae, I am unable to determine if Centrolepidaceae are a distinct lineage, or if they are part of a broader Restionaceae, as has been suggested previously. Further sampling within Restionaceae will be necessary to address this unresolved question. I identify Flagellariaceae, a family whose position has been difficult to unambiguously resolve with both morphological and molecular data, to be strongly supported as the sister group of the rest of the graminid clade. The graminid clade also includes a strongly supported Ecdeiocoleaceae-Joinvilleaceae-Poaceae clade. In my first study (Chapter 3), I found relationships among Ecdeiocoleaceae, Joinvilleaceae, and Poaceae to be resolved variously, generally with low support. However, if the possibly conflicting noncoding data partition is excluded from consideration, a clade consisting of Ecdeiocoleaceae and Joinvilleaceae is inferred by model-based analyses to be the sister-group of Poaceae, with moderate to strong support. I expanded taxon sampling in and around this clade in Chapter 4, and found weak support for an Ecdeiocoleaceae-Poaceae clade, a topology that has also been inferred in several recent studies (e.g., Bremer 2002; Michelangeli et al. 2003; Davis et al. 2004; Chase et al. 2006). Despite exclusion of rapidly evolving noncoding regions, my parsimony and likelihood analyses in Chapter 4 did not infer the Ecdeiocoleaceae-Joinvilleaceae clade that I found in Chapter 3. This is likely a function of increased taxon sampling in Chapter 4, in which I sampled all Poales lineages, and also included Georgeantha, the other genus in Ecdeiocoleaceae (inclusion of Georgeantha appears to have divided the relatively long branch in Ecdeiocolea in Chapter 3). Nonetheless, support for an Ecdeicoloeaceae-Poaceae clade here remains unsatisfactorily weak. Increased support for this relationship (if it is true) will require more data, and may 265 benefit from inclusion of a second recently discovered species of Ecdeiocolea (B. G. Briggs, personal communication). In Chapters 4 and 5,1 demonstrated that several closely related families in Poales have a substantially accelerated plastid substitution rate compared with most other monocots. Now that several aspects of higher-order relationships in Poales are clear and well supported, the molecular evolution of the plastid regions examined here should be examined more closely. It would be worthwhile to characterize and compare rates of synonymous and non-synonymous substitutions in protein-coding regions within and across Poales and other monocot lineages, to distinguish among possible causes for this rate variation (i.e., relaxation of purifying selection, the role of genetic drift, changes in the underlying mutation rate, or a decrease in DNA repair efficiency; Young et al. 2005). I also found an intronless version of the 3'-rps\2 locus in Luzula - the only report for this intron loss in monocots outside of Asparagales. Future work should characterize the distribution of this intron loss across Juncaceae. Finally, in Chapter 5,1 demonstrated that the putative monocot family Hydatellaceae, a small and highly reduced group of plants that has been difficult to place among monocots, is not a commelinid (Poales) at all, but the sister group of the water lilies (Nymphaeales). Previous placements for the family in Poales with molecular data were based on a single rbcL sequence, which I found to be a contaminant, chimeric sequence of a grass and a moss. New data for multiple plastid loci and the nuclear PHYC locus strongly place Trithuria submersa as the sister group of Nymphaeales. Plastid data for the ^r«L(UAA)-rr«F(GAA) region from Hydatella, the other genus in the family, place Hydatellaceae in the same position. Morphological data are strongly congruent with these molecular results. Although 266 only 59% of 110 morphological characters used in previous studies of basal angiosperms could be scored for Hydatellaceae (e.g., Doyle 2005), analyses of morphological characters also identify a sister group relationship between Hydatellaceae and Nymphaeales, supported by nine unequivocal synapomorphies. Discovery of this new lineage of basal angiosperm has substantial implications for understanding early patterns of angiosperm diversification, and should provide a solid evolutionary framework for future work on the family. 267 6.2 R E F E R E N C E S Chase, M . W., Fay, M . F., Devey, D., R0nsted, N., Davies, J., Pillon, Y., Petersen, G., Seberg, O., Tamura, M . N., Asmussen, C. B., Hilu, K., Borsch, T., Davis, J. I., Stevenson, D. W., Pires, J. C , Givnish, T. J., Sytsma, K. J., McPherson, M . A., Graham, S. W., and Rai, H. S. 2006. Multigene analyses of monocot relationships: A summary. In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: comparative biology and evolution (excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Davis, J. I., Stevenson, D. W., Peterson, G., Seberg, O., Campbell, L. M . , Freudenstein, J. V., Goldman, D. H., Hardy, C. R., Michelangeli, F. A., Simmons, M . P., Specht, C. D., Vergara-Silva, F., and Gandolfo, M . A. 2004. A phylogeny of the monocots, as inferred from rbcL and atpA sequence variation. Systematic Botany 29, 467-510. Doyle, J. A. 2005. Early evolution of angiosperm pollen as inferred from molecular and morphological phylogenetic analyses. Grana 44, 227-251. Duvall, M . R., Davis, J. I., Clark, L. G., Noll, J. D., Goldman, D. H., and Sanchez-Ken, J. G. 2006. Phylogeny of the grasses (Poaceae) revisited. In: J. T., Columbus, E. A. Friar, J. M . Porter, L. M . Prince, and M . G. Simpson (eds.). Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Givnish, T. J., Pires, J. C , Graham, S. W., McPherson, M . A., Prince, L. M . , Patterson, T. B., Rai, H. S., Roalson, E. H., Evans, T. E. , Hahn, W. J., Millam, K. C , Meerow, A. W., Molvray, M . , Kores, P. J., O'Brien, H. E. , Hall, J. C , Kress, W. J., and Sytsma, K. J. 2006. Phylogenetic relationships of monocots based on the highly 268 informative plastid gene ndhF: evidence for widespread concerted convergence. Pp. 27-50 In: Columbus, J. T., Friar, E . A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Graham, S. W., and Olmstead, R. G. 2000. Utility of 17 chloroplast genes for inferring the phylogeny of the basal angiosperms. American Journal of Botany 87, 1712-1730. Graham, S. W., Zgurski, J. M . , McPherson, M . A., Cherniawsky, D. M . , Saarela, J. M . , Home, E. C , Smith, S. Y., Wong, W. A., O'Brien, H., Biron, V. L. , Pires, J. C , Olmstead, R. G., Chase, M . W., and Rai, H. S. 2006. Robust inference of monocot deep phylogeny using an expanded multigene plastid data set. Pp. 3-20 In: Columbus, J. T., Friar, E. A., Porter, J. M . , Prince, L. M . , and Simpson, M . G. [Eds.]. Monocots: Comparative Biology and Evolution (Excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. In Press. Grass Phylogeny Working Group [GPWG]. 2001. Phylogeny and subfamilial classification of the grasses (Poaceae). Annals of the Missouri Botanical Garden 88, 373-457. Jansen, R. K., Kaittanis, C , Saski, C , Lee, S. -L . , Tomkins, J., Alverson, A. J., and Daniell, H. 2006. Phylogenetic analyses of Vitis (Vitaceae) based on complete chloroplast genome sequences: effects of taxon sampling and phylogenetic methods on resolving relationships among rosids. B M C Evolutionary Biology 6, 32. doi: 10.1186/1471-2148-6-32. McPherson, M . A. 2003. Phylogenetics of Asparagales (Liliopsida). M . Sc. Thesis. University of Alberta, Edmonton, Alberta, Canada. Michelangeli, F. A., Davis, J. I., and Stevenson, D. W. 2003. Phylogenetic relationships 269 among Poaceae and related families as inferred from morphology, inversions in the plastid genome, and sequence data from the mitochondrial and plastid genomes. American Journal of Botany 90, 93-106. Saarela, J. M , Peterson, P. M . , and Cayouette, J. 2005. Bromus hallii (Poaceae), a new combination for California, U. S. A., and taxonomic notes on Bromus orcuttianus and Bromus grandis. Sida 21, 1997-2003. Saarela, J. M , Peterson, P. M . , and Refulio-Rodriguez, N. F. 2006. Bromus ayacuchenis (Poaceae: Pooideae: Bromeae), a new species from Peru, with a key to Bromus in Peru. Sida 22, 915-926. In Press. Xu, Y., McCouch, S. R., and Zhang, Q. 2005. How can we use genomics to improve cereals with rice as a reference genome. Plant Molecular Biology 59, 7-26. Young, N. D., and dePamphilis, C. W. 2005. Rate variation in parasitic plants: correlated and uncorrelated patterns among plastid genes of different function. B M C Evolutionary Biology 5, 16. doi: 10.1186/1471-2148-5-16. f Zgurski, J. 2004. Analysis of multiple levels of vascular-plant phylogeny using a nuclear RNA polymerase subunit gene (RPB2) and a large plastid data set: Studies of the tracheophytes, cycads and the lilies and relatives. M . Sc. Thesis. University of Alberta, Edmonton, Alberta, Canada. 270 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0100351/manifest

Comment

Related Items