UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Molecular phylogenetics of mosses and relatives Chang, Ying 2011

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

Item Metadata

Download

Media
24-ubc_2011_fall_chang_ying.pdf [ 7.83MB ]
Metadata
JSON: 24-1.0072202.json
JSON-LD: 24-1.0072202-ld.json
RDF/XML (Pretty): 24-1.0072202-rdf.xml
RDF/JSON: 24-1.0072202-rdf.json
Turtle: 24-1.0072202-turtle.txt
N-Triples: 24-1.0072202-rdf-ntriples.txt
Original Record: 24-1.0072202-source.json
Full Text
24-1.0072202-fulltext.txt
Citation
24-1.0072202.ris

Full Text

MOLECULAR PHYLOGENETICS OF MOSSES AND RELATIVES! by! Ying Chang! ! ! 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! (Vancouver)!   July 2011! © Ying Chang, 2011 ! ! ""! ABSTRACT! Substantial ambiguities still remain concerning the broad backbone of moss phylogeny. I surveyed 17 slowly evolving plastid genes from representative taxa to reconstruct phylogenetic relationships among the major lineages of mosses in the overall context of land-plant phylogeny. I first designed 78 bryophyte-specific primers and demonstrated that they permit straightforward amplification and sequencing of 14 core genes across a broad range of bryophytes (three of the 17 genes required more effort). In combination, these genes can generate sturdy and well- resolved phylogenetic inferences of higher-order moss phylogeny, with little evidence of conflict among different data partitions or analyses. Liverworts are strongly supported as the sister group of the remaining land plants, and hornworts as sister to vascular plants. Within mosses, besides confirming some previously published findings based on other markers, my results substantially improve support for major branching patterns that were ambiguous before. The monogeneric classes Takakiopsida and Sphagnopsida likely represent the first and second split within moss phylogeny, respectively. However, this result is shown to be sensitive to the strategy used to estimate DNA substitution model parameter values and to different data partitioning methods. Regarding the placement of remaining nonperistomate lineages, the [[[Andreaeobryopsida, Andreaeopsida], Oedipodiopsida], peristomate mosses] arrangement receives moderate to strong support. Among peristomate mosses, relationships among Polytrichopsida, Tetraphidopsida and Bryopsida remain unclear, as do the earliest splits within sublcass Bryidae. A [Funariidae, [Timmiidae, [Dicranidae, Bryidae]]] arrangement is strongly supported, as are major relationships within subclasses Funariidae and Dicranidae. I also reconstructed the phylogeny of ! """! the nonperistomate moss family Andreaeaceae, with a focus on costate taxa, using two complementary sets of plastid markers and taxa. The major subgenera (Andreaea and Chasmocalyx) and sections of Andreaea (Andreaea and Nerviae) are rejected as monophyletic. Well-supported lineages include clades comprising: (1) Andreaea nivalis and A. rigida (northern hemisphere members of subgenus Chasmocalyx) and A. blyttii (section Nerviae); (2) most of the remainder of Nerviae; (3) a mixture of costate and ecostate species from Chasmocalyx, Nerviae, all sampled members of section Andreaea, and subgenus Acroschisma. Relationships among the major lineages, including the root of the family, are all well supported. ! ! "#! PREFACE! A version of Chapter 2 has been published: Chang, Y., and S. W. Graham (2011), Bryophyte-specific primers for retrieving plastid genes suitable for phylogenetic inference. American Journal of Botany 98: e109-e113. I carried out the laboratory work, designed the primers, and wrote the manuscript. Sean W. Graham provided valuable guidance with respect to primer design and writing. A version of Chapter 3 has been published: Chang, Y., and S. W. Graham (2011), Inferring the higher-order phylogeny of mosses (Bryophyta) and relatives using a large, multigene plastid data set. American Journal of Botany 98: 839-849. I carried out the laboratory work and data analyses, and wrote the manuscript. Sean W. Graham provided insights into taxonomic sampling, data analyses and contributed to the writing. A version of Chapter 4 will be submitted for publication: Chang, Y., and S. W. Graham,! Backbone phylogeny of peristomate and nonperistomate mosses based on multiple plastid markers. I carried out the laboratory work and data analyses, and wrote the manuscript. Sean W. Graham provided insights into taxonomic sampling, data analyses and contributed to the writing. A version of Chapter 5 will be submitted for publication: Chang, Y., and S. W. Graham, Molecular phylogenetics of Andreaeaceae. I carried out the laboratory work and data analyses, and wrote the manuscript. Sean W. Graham provided insights into data analyses and contributed to the writing. ! ! #! TABLE OF CONTENTS Abstract.......................................................................................................................................... ii Preface........................................................................................................................................... iv Table of Contents .......................................................................................................................... v List of Tables ................................................................................................................................. x List of Figures.............................................................................................................................. xii Acknowledgements .................................................................................................................... xiv Dedication .................................................................................................................................... xv 1 Introduction............................................................................................................................. 1 1.1  Overview of early land-plant phylogeny .......................................................................... 1 1.2 Overview of backbone relationships of bryophytes .......................................................... 2 1.3  Objectives of the thesis ..................................................................................................... 6 2 Bryophyte-specific primers for retrieving plastid genes suitable for phylogenetic inference....................................................................................................................................... 11 2.1 Brief synopsis .................................................................................................................. 11 2.2 Introduction ..................................................................................................................... 11 2.3 Methods and results ......................................................................................................... 12 ! #"! 2.3.1 “Universal” primer design and testing...................................................................... 12 2.3.2 DNA extraction, amplification and sequencing ....................................................... 13 2.3.3 Sequence editing and alignment ............................................................................... 13 2.4 Conclusions ..................................................................................................................... 15 3.  Inferring the higher-order phylogeny of mosses (Bryophyta) and relatives using a large, multigene plastid data set ................................................................................................ 24 3.1 Brief synopsis .................................................................................................................. 24 3.2 Introduction ..................................................................................................................... 25 3.3 Materials and methods..................................................................................................... 27 3.3.1 Taxonomic sampling strategy................................................................................... 27 3.3.2 Sequence recovery and alignment ............................................................................ 28 3.3.3 Phylogenetic analyses—Maximum likelihood and parsimony analyses.................. 29 3.3.4 Effect of rapidly evolving sites on MP inference ..................................................... 31 3.4 Results ............................................................................................................................. 32 3.4.1 Phylogenetic inference—ML analysis...................................................................... 32 3.4.2 Phylogenetic inference—MP analysis...................................................................... 33 3.4.3 Exploration of conflicts between ML and MP analysis ........................................... 33 ! #""! 3.5 Discussion........................................................................................................................ 35 3.5.1 A well supported backbone of bryophyte phylogeny............................................... 35 3.5.2 Nonperistomate mosses and the earliest splits in moss phylogeny .......................... 36 3.5.3 The sister-group of land plants ................................................................................. 38 3.6 Conclusions and future work........................................................................................... 39 4. Backbone phylogeny of peristomate and nonperistomate mosses based on multiple plastid markers............................................................................................................................ 51 4.1 Brief synopsis .................................................................................................................. 51 4.2 Introduction ..................................................................................................................... 51 4.3 Materials and methods..................................................................................................... 54 4.3.1 Taxonomic and genomic sampling........................................................................... 54 4.3.2 Sequence recovery and alignment ............................................................................ 55 4.3.3 Phylogenetic analysis ............................................................................................... 56 4.4 Results ............................................................................................................................. 58 4.4.1 Phylogenetic relationships inferred by different analyses ........................................ 58 4.4.1.1 ML analyses using ML-estimated base frequencies.......................................... 58 4.4.1.2 ML analyses using empirical estimates of base frequencies ............................. 60 ! #"""! 4.4.1.3 MP analysis........................................................................................................ 60 4.4.2 Exploration of the conflicting placements of Takakia and Sphagnum ........................ 61 4.5 Discussion........................................................................................................................ 62 4.5.1 Length variation of the noncoding regions in mosses .............................................. 62 4.5.2  Towards a robust backbone phylogeny of peristomate mosses............................... 63 4.5.2.1 Nematodontous mosses ..................................................................................... 63 4.5.2.2 Diplolepideous-opposite taxa and relatives....................................................... 64 4.5.2.3 Diplolepideous-alternate and haplolepideous taxa (Bryidae and Dicranidae) .. 65 4.5.3 The earliest splits in moss phylogeny....................................................................... 67 4.5.4 Placement of Takakia and Sphagnum in moss phylogeny ....................................... 68 5 Molecular phylogenetics of Andreaeaceae ......................................................................... 82 5.1 Brief synopsis .................................................................................................................. 82 5.2 Introduction ..................................................................................................................... 82 5.3 Materials and methods..................................................................................................... 85 5.4 Results ............................................................................................................................. 88 5.5 Discussion........................................................................................................................ 90 ! "$! 5.5.1 Major lineages of Andreaea ..................................................................................... 92 5.5.2 Species monophyly and species delimitation ........................................................... 93 5.5.3 Implications of Andreaea phylogeny for morphological evolution ......................... 94 5.5.4 The placement of Andreaea wilsonii and A. fuegiana.............................................. 96 5.6 Conclusions and future directions ................................................................................... 97 6 Conclusions and future directions..................................................................................... 107 6.1 Major conclusions.......................................................................................................... 107 6.2 Future directions ............................................................................................................ 112 References .................................................................................................................................. 116 Appendices................................................................................................................................. 137 Appendix 1:  Source information of bryophyte taxa used in primer design and testing ....... 137 Appendix 2:  Source information and GenBank numbers for taxa included in Chapter 3 ..... 139 Appendix 3. Source information and GenBank numbers for taxa included in Chapter 4 ..... 141 Appendix 4: Source information and GenBank numbers for taxa included in Chapter 5 ...... 144 ! ! $! LIST OF TABLES! Table 2.1. Details of the amplification and sequencing products for 14 core plastid genes from the 11 taxa used in primer testing ......................................................................................... 16 Table 3.1. Selected recent multigene studies of higher-level phylogenetic relationships in various bryophyte lineages……………………………………………………………..…………..41 Table 3.2. Summary of bootstrap support for multi-family clades of bryophytes and relatives (excluding vascular plants) based on different data combinations and sub-partitions in Chapter 3............................................................................................................................... 42 Table 3.3. Approximately Unbiased (AU) tests of whether suboptimal arranegements concerning early moss phylogeny and the sister group of land plants are significantly worse than the optimal one............................................................................................................................ 44 Table 3.4. Bootstrap support for contrasting arrangements at the base of mosses, and concerning the sister group of land plants ............................................................................................... 45 Table 3.5. Support for multi-family clades of bryophytes and relatives, excluding vascular plants, across selected studies........................................................................................................... 46 Table 4.1a. Support for relationships among the four major land plant lineages and algal relatives in Chapter 4........................................................................................................................... 71 Table 4.1b. Support for the multi-family clades of mosses in Chapter 4 ..................................... 72 ! $"! Table 4.2. The placement of Takakia and Sphagnum in various ML analyses (bootstrap support in parentheses) ...................................................................................................................... 73 Table 4.3. Approximately Unbiased (AU) tests of whether suboptimal arrangements of Takakia and Sphagnum are significantly worse than the optimal one................................................ 74 Table 4.4. DNA substitution model parameters estimated for different data partitions ............... 75 Table 4.5. Variation in lengths of noncoding regions in nonperistomate vs. peristomate mosses ............................................................................................................................................... 76 Table 5.1. The taxonomy of the Andreaea species sampled in this study and the number of specimens included for each species..................................................................................... 99 Table 5.2. Primers used to amplify and sequence the targeted regions ...................................... 100 Table 5.3. Statistics for the various MP and ML analyses conducted on the six-region dataset.101 Table 5.4. Approximately Unbiased (AU) test of whether suboptimal arrangements constraining monophyly of several species are significantly worse than the paraphyletic arrangements in the optimal tree. .................................................................................................................. 102 ! $""! LIST OF FIGURES! ! Fig. 1.1. Selection of alternative hypotheses regarding land-plant relationships ........................... 8 Fig. 1.2a. Summary of higher-order relationships of hornworts (Anthocerotophyta). ................... 9 Fig. 1.2b. Summary of higher-order relationships of liverworts (Marchantiophyta) ..................... 9 Fig. 1.3. Summary of higher-order relationships of mosses ......................................................... 10 Fig. 2.1. Primer map of atpB region, and sequences of primers used to amplify and sequence this region .................................................................................................................................... 17 Fig. 2.2. Primer map of rbcL region, and sequences of primers used to amplify and sequence this region .................................................................................................................................... 18 Fig. 2.3. Primer map of ndhF region, and sequences of primers used to amplify and sequence this region .................................................................................................................................... 19 Fig. 2.4. Primer map of psbD-psbC region, and sequences of primers used to amplify and sequence this region.............................................................................................................. 20 Fig. 2.5. Primer map of psbB-psbT-psbN-psbH region, and sequences of primers used to amplify and sequence this region ....................................................................................................... 21 Fig. 2.6. Primer map of rpl2 region, and sequences of primers used to amplify and sequence this region .................................................................................................................................... 22 ! $"""! Fig. 2.7. Primer map for 3’-rps12-rps7-ndhB region, and sequences of primers used to amplify and sequence  this region ...................................................................................................... 23 Fig. 3.1. Phylogram of best ML tree based on combined analysis of 17 genes and associated noncoding regions................................................................................................................. 47 Fig. 3.2. Phylogram of the single best MP tree based on combined analysis of 17 genes and associated noncoding regions ............................................................................................... 49 Fig. 4.1. The relationships of land plants and algal relatives, inferred from ML analysis of an unpartitioned matrix with ML-estimates of base frequencies using Garli............................ 77 Fig. 4.2. Phylogram of moss phylogeny inferred from ML analysis of an unpartitioned data matrix with ML-estimates of base frequencies using Garli .................................................. 78 Fig. 4.3. Support for multi-family clade of mosses from this study and selected studies ............ 80 Fig. 5.1. Phylogeny of Andreaea inferred from the partitioned 6-region plastid matrix using maximum likelihood ........................................................................................................... 103 Fig. 5.2. Phylogenetic placement of Andreaeaceaea in moss evolution and the relationships of the major Andreaea clades based on ML analysis of the 14-gene matrix ................................ 105 ! $"#! ACKNOWLEDGEMENTS  First and foremost I want to thank my Ph.D supervisor, Dr. Sean W. Graham, for his all his contributions of time, ideas, and funding. His guidance through my Ph.D. period has been extremely valuable to me. I also thank other members of my supervisory committee (Drs. Wilf Schofield, Mary Berbee, and Wayne Maddison) for their help to my thesis work. I want to express my gratitude to Dr. Barbara Murray. Not only did she provide me with valuable specimens, she has also generously shared her understanding of the morphology and evolution of Andreaeaceae with me. I would like to thank former and current members from Graham lab, especially Vivienne Lam and Will Iles, for all the discussions, suggestions and help during the course of my degree. I owe my thanks to Shona Ellis, whose love and enthusiasm for bryophytes are contagious. It would not have been possible to write this doctoral thesis without the help and support of many other colleagues and friends. Thank you! Finally, I would like to thank my parents for all their love and encouragement, and my husband Jun, who has always been loving, patient, and encouraging.   ! $#! Dedication         Dr. Wilfred B. Schofield   !! %! 1 INTRODUCTION 1.1  Overview of early land-plant phylogeny Land plants (embryophytes) comprise four major extant lineages: liverworts (Marchantiophyta), mosses (Bryophyta), hornworts (Anthoceratophyta), and vascular plants (Tracheophyta). The former three are collectively referred as bryophytes. All land plants have a sporic life cycle, i.e., an alternation of generations. Unlike vascular plants, bryophytes are characterized by a gametophyte-dominant life history: the gametophyte generation is the major photosynthesizing phase, while the sporophyte generation is at least partially dependent on the female gametophyte for nutrients. The sporophytes of all extant bryophytes are unbranched and bear a single sporangium, in contrast to the branched (polysporangiate) sporophytes found in vascular plants. These attributes were considered to be evidence of the unique common ancestry shared by all bryophytes (Vanderpoorten and Goffinet, 2009). However, since the middle 1900s, various authors have challenged the monophyly of bryophytes and proposed that they represent independent evolutionary lines (e.g., Bold, 1956; Steere 1969; Crandall-Stotler, 1980). Early cladistic studies on morphological and biochemical characters by Mishler and Churchill (1984, 1985) demonstrated the paraphyly of bryophytes. Since then, an increasing number of studies has formed a consensus that each bryophyte lineage is monophyletic and that they form a grade subtending vascular plants (e.g., Kenrick & Crane, 1997; Garbary and Renzaglia, 1998; Qiu et al., 1998; however, Nishiyama, 2004; Goremykin and Hellwig, 2005). In contrast, the branching order of the three bryophyte lineages, in relation to each other and the vascular plants, still remains controversial (Qiu, 2008). Various hypotheses have been proposed over the years based !! &! on multiple lines of data, but rarely with convincing bootstrap support (e.g., Lewis et al., 1997; Nickrent et al., 2000; Kelch et al., 2004; see also Fig. 1.1). An unambiguous reconstruction of early land-plant phylogeny is essential for understanding plant evolution, as it would help in reconstructions of the early morphological innovations that developed in land plants after the invasion of the terrestrial environment, and also provide a comparative framework for studying the genome evolution of early land plants. 1.2 Overview of backbone relationships of bryophytes Among the three bryophyte lineages, hornworts are the smallest group, with approximately 150 species (Renzaglia et al., 2009). Hornworts are so named because of their horn-like sporophytes. Unlike the sporophytes of liverworts and mosses, the hornwort sporophyte possesses a basal intercalary meristem and so has indeterminate growth (in most species). The gametophyte is always thalloid and does not differentiate into stems and leaves. Hornworts are also known for their symbiotic relationship with nitrogen-fixing cyanobacteria and endomycorrizhal fungi (Vanderpoorten and Goffinet, 2009). Compared to the other two bryophyte groups, hornworts remain relatively unexplored. The number of species and genera found in hornworts has only been crudely estimated (Duff et al., 2007), and due to different opinions on the interpretation of some morphological characters, hornwort classification schemes have been highly variable (reviewed in Duff et al., 2004). In the past decade, molecular markers have been used to elucidate the phylogenetic relationships of hornworts (e.g., Stech et al., 2003; Duff et al., 2007). A broad degree of consensus has now been reached on this (Fig. 1.2a), corroborated by subsequent ultrastructural and developmental studies (Renzaglia et al., 2007). The monotypic Leiosporoceros is found to be highly divergent compared to the other hornworts !! '! (Duff et al., 2004) and is now recognized as a class, Leiosporocerotopsida (Duff et al., 2004; Renzaglia et al., 2009). The remaining taxa are included in the other hornwort class, Anthocerotopsida. Three well-segregated clades are found within Anthocerotopsida (Duff et al., 2004 and 2007), each treated as a subclass (Anthocerotidae, Notothylatidae and Dendrocerotidae). There are about 5,000 described liverwort species in total. Unlike the uniform thalloid gametophyte found in hornworts, the gametophytes of liverworts demonstrate a remarkable morphological diversity. They can be leafy in form with different leaf insertion patterns, and different numbers of lobes, or thalloid, with or without much internal differentiation. The sporophytes of liverworts, on the other hand, are more or less uniform among different liverwort species. Two major evolutionary lines were traditionally recognized in liverworts, based mainly on gametophytic characters (Schofield, 1985). The liverworts have been divided into two groups: marchantioids (= complex thalloid liverworts), and jungermannioids (= leafy and simple thalloid liverworts) (Crandall-Stotler et al., 2009). A large number of molecular studies have been carried out on the higher-order phylogeny of liverworts. These studies have brought new insights into the backbone relationships of liverworts (e.g., Davis, 2004; Forrest et al., 2006; He-Nygén, et al., 2006; Wilson et al., 2007). One of the major findings in these studies is the recovery of a clade comprising Haplomitrium and Treubia, two leafy taxa, as the sister-group of all other liverworts. Blasia and Cavicularis, two simple thalloid liverworts, then comprise a clade (Blasiidae) that is the sister group of the complex thalloid liverworts (Marchantiidae), and the clade consisting of Marchantiidae and Blasiidae (now collectively recognized as Marchantiopsida) is the sister group of the Jungermannioids (Jungermanniopsida). Within Jungermanniopsida, Pelliidae and !! (! Metzgeriidae form a grade of simple thalloid liverworts subtending the leafy taxa (Jungermanniidae) (Fig. 1.2b). With about 13,000 species, mosses are the largest bryophyte group and the second largest major plant group after flowering plants (Cox et al., 2010). Unlike liverworts and hornworts, the gametophyte of extant mosses is always leafy, with leaves usually spirally arranged around the stem. A typical moss sporophyte is characterized by having a robust seta and a ring or two of peristome teeth aligned with the opening of the capsule (sporangium), although multiple nonperistomate lineages (which have capsules opening without teeth) represent the earliest splits in moss phylogeny (Fig. 1.2). The nonperistomate mosses are mostly species-poor (one species each in Oedipodiopsida and Andreaeobryopsida, two species in Takakiopsida, ~50 species in Andreaeopsida, and ~200 species in Sphagnopsida), but they represent ancient and morphologically distinctive evolutionary lines whose relationships to each other are still not clear (Goffinet and Buck, 2004). Our understanding of the overall classification and relationships of the peristomate mosses has traditionally relied heavily on characters associated with the architecture and development of peristome teeth, due to their (presumed) conservative nature in moss evolution (e.g., Fleischer, 1904-1923; Dixon, 1932; Vitt, 1984). There are two major types of peristome teeth: nematodontous and arthrodontous teeth. The former type is made up of whole cells, while the latter is composed of cell-wall fragments. Nematodontous taxa are further arranged into two classes, Tetraphidopsida and Polytrichopsida. Both classes are species-poor, with only five species in Tetraphidopsida and ~200 species in Polytrichopsida (Bell and Hyvonen, 2010). In contrast, the only arthrodontous moss class, Bryopsida, harbours about 90% of all moss species !! )! (~12,000 species) and shows great diversity in peristome morphology (Goffinet and Buck, 2004). There are three major types arthrodontous teeth: diplolepideous-opposite, diplolepideous- alternate and haplolepideous. These major types of peristome teeth respectively characterize three subclasses: Funariidae, Bryidae and Dicranidae. However, not all types of peristome teeth fall nicely into one of the three categories. For example, the teeth found in Buxbaumia (Buxbaumiidae) appear transitional between nematodontous and arthrodontous types, and Timmia (Timmiidae) possesses teeth with a unique architecture that is hard to interpret (Budke et al., 2007). Despite extensive studies on the development and structure of peristome teeth, their evolutionary transitions are still poorly understood (Goffinet et al., 2001). Moreover, the morphologies of peristome teeth are generally too conserved to provide information about relationships within major groups (Vitt et al., 1998; Magombo, 2003). Many gametophytic characters, on the other hand, are known to be highly homoplastic and plagued with reversals and convergences, making it potentially problematic to use them in the inference of moss backbone relationships (Renzaglia et al., 2007; Goffinet et al., 2009) In the past two decades, DNA sequence data, or a combination of DNA sequence and morphological data, have been used in a considerable number of phylogenetic studies to investigate higher-order relationships among major moss groups (peristomate and nonperistomate) (e.g., Newton et al., 2000; Beckert et al., 2001; Goffinet et al., 2001; Magombo 2003; Cox et al., 2004; Wahrmund et al., 2010). Despite multiple examples of concordance among different phylogenetic studies, there are also multiple points along the backbone of moss phylogeny that do not have satisfactory resolution–i.e., those that are either polytomous, poorly !! *! supported, contradictory among studies, or unsatisfactorily corroborated by morphological synapomorphies. Higher-order moss phylogeny is therefore in need of substantial additional investigation (Goffinet et al., 2009). This is the major focus of my thesis (Chapters 2-4). Among nonperistomate classes that represent the early splits in moss evolution, only two comprise more than two species: Andreaeopsida and Sphagnopsida. Unlike Sphagnopsida, which has been well studied (e.g., Ligrone and Duckett, 1998; Vitt, 2006; Shaw et al., 2010), the systematics of the monofamililial Andreaeopsida is mostly understudied. There has been no treatment of the whole family since the early work of Roth (1903-1904, 1910-1911), and phylogenetic relationships within the family have not been addressed. A robust phylogeny of the family would not only help improve our understanding of evolutionary transitions in the family, but it would also help to evaluate the naturalness of its current classification, and provide evidence for future taxonomic revisions of the family, if needed. Furthermore, it would help to place the family more firmly in early moss phylogeny and improve our understanding of the deepest splits in moss evolution. This is the focus of the research described in Chapter 5. 1.3  Objectives of the thesis The overall goal of Chapters 2-4 is to provide a more resolved and better supported picture of the higher-order phylogeny of mosses and relatives, and in particular to improve our current understanding of the early branches of moss phylogeny. The inference of the higher- order relationships of bryophytes has relied on a limited number of molecular markers to date (Stech and Quandt, 2010). In this thesis, I develop 78 new bryophyte-specific primers to retrieve a suite of plastid markers (Chapter 2). The targeted plastid regions include four single genes (atpB, ndhF, rbcL and rpl2), and four multiple gene clusters (psbE-psbF-psbL-psbJ, psbD-psbC, !! +! psbB-psbT-psbN-psbH, rps12-rps7-ndhB). These regions have proved useful for elucidating higher-order relationships among various vascular plant groups (Graham and Olmstead, 2000) and should be suitable for reconstructing bryophyte backbone phylogenies. I assess the utility of these primers in retrieving these regions using a test set of exemplar species that represent a broad range of bryophyte taxa. I demonstrate the performance of the targeted DNA markers for inferring the broad phylogenetic backbone of mosses and their placement in land-plant phylogeny by first focusing on the nonperistomate moss lineages (Chapter 3). In this chapter I also examine whether identification and removal of the most rapidly evolving nucleotides in the plastid markers used here may have an impact on the phylogenetic inference using parsimony (as compared to maximum likelihood). I then substantially expand taxon sampling in peristomate mosses to characterize the overall phylogenetic relationships within this major moss clade (Chapter 4). In this chapter I also investigate the effect of DNA substitution model parameter settings on maximum likelihood (ML) based inference (Chapter 4). Finally, in Chapter 5 I present two complementary sets of plastid markers used at different densities of taxon sampling to characterize the broad phylogenetic structure of the granite moss family, Andreaeaceae, which represents the second largest group of nonperistomate mosses. I assess the monophyly of current infrageneric taxonomy (subgenera, sections and species) and summarize initial evidence on the evolution of some morphological characters and their potential utility in future classification schemes for the family (Chapter 5). !! ,! ! ! HL M V LH M V MH L V ML H V e.g., Mishler and Churchill, 1984 e.g., Nishiyama et al., 2004e.g., Qiu et al., 2006 e.g., Renzaglia et al., 2000 Fig. 1.1. Selection of alternative hypotheses regarding land-plant relationships. H: hornworts; L: liverworts; M: mosses; V: vascular plants !! -! ! Leiosporocerotopsida Anthocerotopsida Anthocerotidae Notothylatidae Dendrocerotidae Fig. 1.2a. Summary of higher-order relationships of hornworts (Anthocerotophyta). The classification follows Renzaglia et al. (2009). ! Haplomitriopsida Metzgeriidae Pelliidae Marchantiidae Blasiidae Jungermanniidae simple thalloid complex thalloid Marchantiopsida leafy Jungermanniopsida Treubiidae Haplomitridae Fig. 1.2b. Summary of higher-order relationships of liverworts (Marchantiophyta). The classification follows Crandall-Stotler et al. (2009). !! %.! ! Takakiopsida Oedipodiopsida Andreaeobryopsida Andreaeopsida Sphagnopsida Buxbaumiidae Polytrichopsida Bryopsida Tetraphidopsida Diphyciidae Funariidae Timmiidae Dicranidae Bryidae peristomate nonperistomate diplolepideous-opposite diplolepideous-alternate haplolepideous nematodontous arthrodontous Fig. 1.3. Summary of higher-order relationships of mosses. The classification follows Goffinet et al. (2009) !! %%!  2 BRYOPHYTE-SPECIFIC PRIMERS FOR RETRIEVING PLASTID GENES SUITABLE FOR PHYLOGENETIC INFERENCE 1 2.1 Brief synopsis I present here a total of 78 new bryophyte-specific primers that permit retrieval of 17 slowly evolving plastid genes and their associated introns and intergenic spacers. These regions were chosen to facilitate accurate phylogenetic inference across a broad range of mosses and other bryophytes. The primers were designed using an initial sampling of exemplar bryophytes and other plants. I assessed the ability of the new primers to amplify and sequence these regions using a test set of 11 exemplar bryophytes. The newly designed primers allowed ready retrieval of 14 of the 17 targeted genes from a broad range of bryophyte taxa, and should prove useful for future studies of bryophyte phylogeny. 2.2 Introduction The higher-order relationships of mosses (Bryophyta) are arguably understudied compared to other major embryophyte clades, and studies of higher-order moss phylogeny have focused on relatively few plastid genome regions to date. Among the most widely used plastid markers are the rbcL and rps4 loci and the trnL-trnF region (e.g., Tsubota et al., 2003; Cox et al., 2004). I have developed a series of new primers to amplify and sequence multiple additional  1 A version of this chapter has been published: CHANG, Y., AND S. W. GRAHAM. 2011. Bryophyte-specific primers for retrieving plastid genes suitable for phylogenetic inference. Am. J. Bot. 98:e109-e113  !! %&! conserved regions in the plastid genomes of bryophytes. The regions that can be surveyed with these primers are those used successfully in the inference of higher-order relationships at multiple levels of angiosperm, seed-plant and vascular-plant phylogeny (e.g., Graham and Olmstead, 2000). They comprise 10 photosystem II (psb) genes located in three major clusters (psbD-psbC, psbE-psbF-psbL-psbJ, and psbB-psbT-psbN-psbH), two NADH dehydrogenase subunit genes (ndhF and ndhB), three ribosomal protein genes (3’-rps12, rps7 and rpl2), an ATP synthase subunit gene (atpB), and the gene coding for the large subunit of Rubisco (rbcL). Among the 17 loci that I surveyed, only rbcL has been widely used in other studies of bryophyte phylogeny (e.g., Tsubota et al., 2003; Chang et al., 2006), although some of them have been sampled sporadically using (in part) primers previously designed for tracheophytes (e.g., the psbT-psbN-psbH region, Quandt et al., 2003). Here I document the utility of these primers in complementing the previously published primers for retrieving the targeted regions, based on a preliminary sampling of “exemplar” (representative) mosses and other bryophytes (hornworts and liverworts; Anthocerotophyta and Marchantiophyta). 2.3 Methods and results 2.3.1 “Universal” primer design and testing I chose candidate priming sites in plastid regions used previously to infer higher-order vascular-plant phylogeny, and designed oligonucleotides for amplification and sequencing of bryophytes using criteria described in Graham and Olmstead (2000). This design process involved several iterations of DNA sequence alignments comprising new and previously published plastid sequences (Appendix 1). I employed Oligo 4.0-s (National Biosciences Inc. !! %'! Plymouth, Minnesota, USA) to assess GC content and melting temperature of oligos, and to test for problematic hairpin structures or primer dimers (see Graham and Olmstead, 2000). Twelve taxa (Appendix 1) were newly sequenced during primer development. I also tested how well the new primers work in amplification and sequencing using an additional batch of “test” taxa (i.e., seven additional mosses, two liverworts and two hornworts, Appendix 1). 2.3.2 DNA extraction, amplification and sequencing I extracted total genomic DNA from exemplar taxa (Appendix 1), following the method of Doyle and Doyle (1987) or using a QIAGEN DNeasy plant mini kit (QIAGEN, Valencia, California, USA). The PCR profile includes an initial denaturation at 94 o C for 5 min, followed by 35-40 cycles of 94 o C for 1 min, 45 o , 43 o  or 41 o C for 1 min, 72 o C for 2 min, with a final extension at 72 o C for 10 min. I cleaned PCR products with the QIAquick  PCR purification kit (QIAGEN, Valencia, California, USA) and performed direct sequencing using an ABI BigDye Terminator kit version 3.1 (Applied Biosystems, Foster City, California, USA). The reactions were run on 377 or 3700 automated sequencers (Applied Biosystems, Foster City, California, USA). 2.3.3 Sequence editing and alignment I initiated this study by sequencing portions of the regions of interest using primers previously designed for vascular plants (Graham and Olmstead, 2000; Rai and Graham, 2010). I compiled chromatograms and called bases using Sequencher ver. 4.1 (Gene Codes Corp., Ann Arbor, Michigan, USA). I aligned the new sequences from each major region with a selection of previously published sequences from a phylogenetically diverse range of vascular plants and !! %(! charophycean algae with Se-Al ver. 2.0 (Rambaut, 2002), using criteria laid out in Graham et al. (2000). I included sequences only from the coding regions of the vascular plants and algae to minimize alignment ambiguities, but aligned all noncoding regions across bryophytes. I used these alignments to design more primers, and repeated this design process until I obtained satisfactory primer coverage for the targeted regions. In total I designed 77 new primers for the targeted regions (Figs. 2.1-2.7). For the region spanning psbE, psbF, psbL and psbJ, previously published primers (see Fig. 4 in Graham and Olmstead, 2000) generally worked well across the bryophytes that I surveyed. For ndhF (Fig. 2.3), the portion of the gene that I examined is the more conservative 5’-end (Olmstead and Sweere, 1994). After extensive primer design (Fig. 2.7) and testing I was unable to develop a core set of primers that work well across bryophytes for the rps12-rps7-ndhB region, and so I abandoned further primer design for it. I amplified and generated sequencing products for the remaining 14 core targeted regions for a further seven mosses, two liverworts and two hornworts (Appendix 1). Focusing on these taxa, I found amplifications typically worked with only one or two trials per amplification product. I could not amplify the ndhF region for the liverworts Haplomitrium and Treubia, the rpl2 region for the hornworts Leiosporoceros and Phaeoceros, or the psbE-psbF-psbL-psbJ region for the moss Tetrodontium, the liverwort Haplomitrium, and the hornwort Phaeoceros. The primers worked for sequencing with fewer trials on average for mosses, followed by liverworts and then hornworts (Table 2.1; this table is not a systematic exploration of all the new primers, but rather is intended to document my experience based on regular lab flow when working with each region. As a guide the most commonly used sequencing primers are noted in !! %)! Figs. 2.1-2.6 using an asterisk). This is not surprising, since much of my primer development work was based on alignments that included a high proportion of mosses. I demonstrated the ability of these conserved regions to infer splits along the backbone of moss phylogeny by using a subset of taxa I sampled in this study, together with new and previously published sequences (Chapter 3). These genes yield sturdy and well-resolved phylogenetic inferences using maximum likelihood (ML) analysis (Chapters 3, 4) that are largely congruent with other studies, showing that the targeted regions are suitable for inferring higher-order relationships among bryophytes. 2.4 Conclusions I find that the new primers designed in this study (Figs. 2.1-2.6) work well across a broad range of bryophytes, permitting straightforward amplification and sequencing of the targeted regions, except for the 3’–rps12-rps7-ndhB region (Fig. 2.7), which is more difficult to retrieve. These primers are good complements to those already described for vascular plants (e.g., Graham and Olmstead, 2000; Rai and Graham, 2010), and I expect that they will be suitable for retrieval of DNA sequence data from most other bryophytes for use in studies addressing bryophyte phylogeny. !! " # ! Table 2.1. Details of the amplification and sequencing products for 14 core plastid genes from the 11 taxa used in primer testing (psbDC = psbD-psbC; psbBTNH = psbB-psbT-psbN-psbH; psbEFLJ = psbE-psbF-psbL-psbJ). The average number of successful sequencing reactions is based on the number of sequences included in final contigs for 11 “test” taxa. Asterisk: data from a single species; n.a.: “not applicable” due to failed amplification.  1 Exhaustive list: the most commonly used amplification primers are noted in the figures. Average # of successful/failed sequencing reactions Region Length (kb) 1 Amplification primer pairs for test taxa Mosses  Liverworts Hornworts atpB 1.4  C91F/N98R 6/1 5/1.5 8/0.5 rbcL 1.4 80F/N86R 6/1.5 6/1 7/1.5 ndhF 1.3  N30F/36R or N31F/N34R 6/1 n.a 6/2 * psbDC 2.2  N40F or C40F/B47R, 44F/51R 8/1.2 10/0 13/1 * psbBTNH 2.2  60F/66R, 65F or N65F/B71R 10.5/1.5 9/3 11/1 psbEFLJ 0.7 55F or B55F/58R 4/0.5 3/0 * 4/1 rpl2 1.2  N20F or F20F/B25R or N21F/N24R 5/1 8/3 * n.a. !! "#! ! !"#$ %&'() *&+(,-&+() $&+. -&/.) -&0() -&1. -&2.) -&3.) !"#4 +0567# *&+(8699:99:9:;:%:9:;:::99:9966 66$&+.86:%%;%<;%:::;;::%%:9:::% -&+(86%;;%:::<%%=;::%;:;9;:%:9%:%%66 66666666666666-&/.86;;;;9:;%.:%::9<::;::;;%;:% >&?(86;;:9;:%:::;%;:99;;;:9;;;9;;:%66 66-&1.869:;9;;9::9;;:%:;;;99;:%<;% >&/(869;;9:::%:9%:::;::;99@;9;;:9%% 66-&2.86:%:;;::%:%%;A;;;9;;:%;:::9 -&0(86%;;::%;;9%;9::<;<9:;%%:9%66 66-&3.86:;%%;:;9%;;;;%%;%%%;::; (BCD!CE6#CFGHCI .HJHCIH6#CFGHCI >&/( &0(&?(,$&?(,>&?( Fig. 2.1. Primer map of atpB region, and sequences of primers used to amplify and sequence this region. All new primers were designed by Y. Chang, except for B92R (designed by H.S. Rai, University of British Columbia). Asterisks indicate the most commonly used sequencing primers for the first six regions. Primers starting with ‘N’ were designed to work for all bryophytes, ‘M’ for mosses, and ‘H’ for liverworts, and the remainder for vascular plants (e.g., Graham and Olmstead, 2000). Boxed primers are those used most frequently in amplification of the test taxa. The scale (bp) is relative to Physcomitrella patens (primers not to scale).  !! "#! ! !"#$% &"'(% !")( "*( !"+(% ,*-./0 !"#$1.23423455532545243533. !")(1.3355344255543354354342543 !"+(1.4555355533464355343442342 $789:8;.08<=>8? (>@>8?>.08<=>8? "-$ ")$%",$ Fig. 2.2. Primer map of rbcL region, and sequences of primers used to amplify and sequence this region. All new primers were designed by Y. Chang. Asterisks indicate the most commonly used sequencing primers for the first six regions. Primers starting with ‘N’ were designed to work for bryophytes, and the remainder for vascular plants (e.g., Graham and Olmstead, 2000). Boxed primers are those used most frequently in amplification of the test taxa. The scale (bp) is relative to Physcomitrella patens (primers not to scale).  !! "#! ! !"#$ %"#$& '"#$ %"($ %"() %"*$& %"*) %+,*"$ %+,-($ %"") ".) ".$& %".)& '",) %+(*,/) "0)& %+1-1$ *,#234 %"#$52667888787968868788887787992 %"()5297678787768786)68787778678776 !"#$52789977688787888677778988892 %"*)528779776767886687687788666 '"#$5287997988688988789788:78792 %"")5286777886777768766898:7866 %"($5298878978887879697879887:789862 %".)5279688889778779678989 %"*$52888899977887987997789892 '",)527797777988)8866898778899 $;<=><?24<@AB<C )BDB<CB24<@AB<C Fig. 2.3. Primer map of ndhF region, and sequences of primers used to amplify and sequence this region. Note that only the first ~1300 bp of ndhF is amplified. All new primers were designed by Y. Chang. Asterisks indicate the most commonly used sequencing primers for the first six regions. Primers starting with ‘N’ were designed to work for bryophytes, ‘M’ for mosses, ‘H’ for liverworts and the remainder for vascular plants (e.g., Graham and Olmstead, 2000). Boxed primers are those used most frequently in amplification of the test taxa. Scales (bp) are relative to Physcomitrella patens (primers not to scale).  !! "#! ! !"#$% &'#( &')(* (')( &')(+ '), &'-,* &'-( '-, ''(* &'"( ('", &'",* '.( /'0( /'0,* &'1( '2, &'2, &'2( "),* &'0(* &'#(345666756576567756758677865574 &'-,348587755676566678,88568588568 &')(346577677,677666585776584 &'",34776688566555757876668858 &')(+34755785855775756666586876677674 &'2,348857657758856555596858 &'-(3475677675677678555958566887678 &'"(3478667756778578985575685788 &'0(3488575558666575575568666888 &'1(346677677:75,77567756676657 &'2(347675566665677688658677 (;<=+<>4%<?@A<B ,ACA<BA4%<?@A<B %B$D %B$8 8'#( Fig. 2.4. Primer map of psbD-psbC region, and sequences of primers used to amplify and sequence this region. This region is recovered using overlapping amplifications: psbD-psbC. All new primers were designed by Y. Chang. Asterisks indicate the most commonly used sequencing primers for the first six regions. Primers starting with ‘N’ were designed to work for bryophytes, and the remainder for vascular plants (e.g., Graham and Olmstead, 2000). Boxed primers are those used most frequently in amplification of the test taxa. Scales (bp) are relative to Physcomitrella patens (primers not to scale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ig. 2.5. Primer map of psbB-psbT-psbN-psbH region, and sequences of primers used to amplify and sequence this region. All new primers were designed by Y. Chang. Asterisks indicate the most commonly used sequencing primers for the first six regions. Primers starting with ‘N’ were designed to work for all bryophytes, ‘H’ for liverworts, and the remainder for vascular plants (e.g., Graham and Olmstead, 2000). Note that psbN is on the complementary strand to psbB, psbT and psbH. Boxed primers are those used most frequently in amplification of the test taxa. Scales (bp) are relative to Physcomitrella patens (primers not to scale).  !! ""! ! !"#$% $"#$ &"'$% !"($ )"*+%!"(+%$"(+!",+ &""+ &"'+% "*#-. /.0"12345, /.0"12345" !"#$6177787799:89:77:77:897991 1111!",+6177:777:77::8777889878:87779 !",$6198:::788:::979:9899:::911 1111&""+61:988:::97887:877::;977:9 &"'$61988:88:779::7::79:777811 1111&"'+61::9:::78:77:778::7997998 !"($618::8778877:7:988:::79911 1111!"(+6188:7779987:7::99::9779 +2<2/=21./>?2/=$4/@A/B1./>?2/= >5C/45 !",$% Fig. 2.6. Primer map of rpl2 region, and sequences of primers used to amplify and sequence this region. All new primers were designed by Y. Chang. Asterisks indicate the most commonly used sequencing primers for the first six regions. Primers starting with ‘N’ were designed to work for all bryophytes, ‘M’ for mosses, and the remainder for vascular plants (e.g., Graham and Olmstead, 2000). Boxed primers are those used most frequently in amplification of the test taxa. Scales (bp) are relative to Physcomitrella patens (primers not to scale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ig. 2.7. Primer map for 3’-rps12-rps7-ndhB region, and sequences of primers used to amplify and sequence this region. This region is recovered using overlapping amplifications. Asterisks indicate the most commonly used sequencing primers for the first six regions. Primers starting with ‘N’ were designed to work for all bryophytes, ‘M’ for mosses, and ‘H’ for liverworts, and the remainder for vascular plants (e.g., Graham and Olmstead, 2000). Boxed primers are those used most frequently in amplification of the test taxa.!Scales (bp) are relative to Physcomitrella patens (primers not to scale).  !! "#! 3.  INFERRING THE HIGHER-ORDER PHYLOGENY OF MOSSES (BRYOPHYTA) AND RELATIVES USING A LARGE, MULTIGENE PLASTID DATA SET 1 3.1 Brief synopsis I surveyed 14–17 plastid genes from a broadly representative taxonomic sampling of the major bryophyte lineages, including all major lines of nonperistomate mosses. I inferred major clades with at least as strong support as other studies that used more taxa, and corroborated current views of overall embryophyte relationships—[liverworts, [mosses, [hornworts, tracheophytes]]]—with strong maximum likelihood (ML) bootstrap support. I also placed Zygnematales as the sister group of embryophytes with moderate ML bootstrap support. Within mosses, I confirmed Oedipodiaceae as the sister group of the large clade of peristomate taxa. Likelihood analysis firmly placed Takakiaceae as the sister group of all other mosses, a strong conflict with parsimony results, flagging a possible instance of long-branch attraction. Parsimony converged on the Takakia-sister result when rapidly evolving characters were removed, depending on the tree used to classify site rates. In general, however, my findings broadly support the utility of this 14-gene set from the plastome for future, more densely sampled phylogenetic studies of mosses and relatives. ! 1 A version of this chapter has been published: CHANG, Y., AND S. W. GRAHAM. 2011. Inferring the higher-order phylogeny of mosses (Bryophyta) and relatives using a large, multigene plastid data set. Am. J. Bot. 98: 839-849. !! "#! 3.2 Introduction The mosses (Bryophyta) comprise a significant portion of plant diversity (~13,000 extant species, Shaw, 2009). They are the direct descendents of the earliest divisions in land-plant phylogeny, and are functionally important in many terrestrial ecosystems (e.g., Rydin, 2009). The precise placement of mosses in plant phylogeny—i.e., their position relative to the two other bryophyte clades (liverworts, hornworts) and to the vascular plants—has been controversial (reviewed in Qiu, 2008). In addition, despite substantial progress in understanding the overall pattern of moss phylogeny, considerable room for improvement clearly remains (Goffinet et al., 2009). Thus, new tools for inferring phylogenetic relationships in mosses and other bryophytes would be valuable for more confidently resolving their higher-order relationships, and this in turn would facilitate more precise reconstructions of their evolution, and improve the stability and functionality of current classification schemes. Relationships among the multiple ancient and species-poor moss lineages that lack peristomes on their sporangia will be particularly critical to resolve, as they define the deepest splits in moss phylogeny. For example, while many molecular analyses recover a clade comprising the nonperistomate moss genera Sphagnum and Takakia, with this clade then as the sister group of all other mosses (e.g., Newton et al., 2000; Cox et al., 2004; Qiu et al., 2006, 2007; Shaw et al., 2010a, b), no morphological characters support this result (Newton et al., 2000; Frey and Stech, 2009), and there has been speculation that it represents a long-branch artefact (Shaw and Renzaglia, 2004). Subsequent splits in early moss phylogeny involve three distinctive nonperistomate genera: Andreaea, Andreaeobryum and Oedipodium (see Shaw and Renzaglia, 2004). Oedipodium may be the sister group of peristomate mosses (e.g., Cox et al., !! "#! 2004) but the precise relationship of this clade to Andreaea and Andreaeobryum, and of the latter genera to each other, are not clear (reviewed in Shaw and Renzaglia, 2004). Finally, major uncertainties in the backbone of moss phylogeny are by no means limited to the nonperistomate taxa. Within the peristomate clade, the precise relationships among the two nematodontous clades (Polytrichopsida and Tetraphidopsida) and the arthrodontous clade (Bryopsida), for example, have also not been resolved satisfactorily, as noted by Goffinet et al. (2009). Loci from the nuclear and mitochondrial genomes have contributed to our understanding of plant phylogeny, but plastid (chloroplast) genes have generally played a more central role to date. There are several reasons for this (Olmstead and Palmer, 1994). For example, the coding portions of plastid genes are generally highly conserved, and as plastid genes are single-copy or effectively so (the two copies of genes present in the plastid inverted repeat region are identical), the orthology of plastid genes from distantly or closely related taxa is uncontroversial, an advantage over most nuclear genes (Kellogg and Bennetzen, 2004). The plastid genome also does not suffer from lateral gene transfer or the considerable rate heterogeneity sometimes observed in mitochondrial genes (e.g., Palmer et al., 2000; Bergthorsson et al., 2004; Qiu et al., 2010). In addition, the generally highly conserved gene content and gene order of plastid genomes makes it relatively straightforward to design primers to retrieve orthologous plastid genes from distantly related taxa (e.g., Graham and Olmstead, 2000a). Studies of higher-order moss phylogeny have focused on relatively few plastid genome regions (e.g., Table 3.1; see also Stech and Quandt, 2010, for a recent review). I therefore developed a series of new primers to amplify and sequence 17 plastid conserved genes and associated regions (Chapter 2) that have been used successfully in the inference of higher-order relationships at multiple levels of !! "#! angiosperm, seed-plant and vascular-plant phylogeny (Graham and Olmstead, 2000a; Rai et al., 2003, 2008; Graham et al., 2006; Saarela et al., 2007; Zgurski et al., 2008; Rai and Graham, 2010). I focus here on using these plastid regions to infer the earliest splits in moss phylogeny and their position in embryophyte phylogeny, by using an exemplar-based approach that samples representatives of the major groups of mosses, liverworts and hornworts, including all lineages of nonperistomate mosses. I demonstrate that this sampling approach recovers relationships that are both well supported and broadly congruent with other studies, at least using maximum likelihood. Finally, I investigate whether identifying and removing the most rapidly evolving sites may help to ameliorate possible instances of mis-inference using parsimony. I restrict my inferences to maximum likelihood and parsimony analysis, as Bayesian inference can yield inflated or skewed branch support values in situations where ML or MP inference may be more conservative, such as the reconstruction of ancient phylogenetic splits (e.g., Suzuki et al., 2002; Cummings et al., 2003; Simmons et al., 2004; Kolaczkowski and Thornton, 2007; Susko, 2008). 3.3 Materials and methods 3.3.1 Taxonomic sampling strategy I generated new data from 19 bryophyte taxa (five liverworts, two hornworts, 12 mosses) for phylogenetic analysis in combination with previously published sequences from one liverwort, one hornwort, two mosses, 15 vascular plants and five charophycean algae (Appendix 2). Carefully focused exemplar-based sampling allows the broad outline of phylogenetic relationships to be inferred, even where the species surveyed represent a small fraction of species !! "#! numbers, as is the case in all current studies that survey the broad backbone of bryophyte phylogeny (Table 3.1). I carefully chose exemplar species that split the broad backbone of bryophyte phylogeny at major internal nodes of interest. In general, my relatively sparse sampling permits exact comparisons to other studies that may have different sampling densities in descendant clades. For example, I sampled all nonperistomate moss lineages by including at least one exemplar species from the major genera, and included multiple representative species that straddle the root nodes of major clades (such as liverworts, hornworts and peristomate mosses), so far as these are known. 3.3.2 Sequence recovery and alignment  With a few minor exceptions, I recovered data from 14 plastid genes and associated non- coding regions from all taxa (Appendix 2). For most taxa I also generated sequence data for a region comprising three additional plastid genes (3’rps12, rps7 and ndhB) that proved significantly harder to retrieve (Chapter 2). This region was not recovered or attempted for a subset of taxa (i.e., Buxbaumia aphylla, Bryum capillare, Treubia lacunose, Haplomitrium hookeri, Metzgeria conjugata, Pleurozia purpurea, Leiosporoceros dussii and Phaeoceros carolinianus). Methods for DNA extraction, amplification, sequencing and base-calling follow Graham and Olmstead (2000a) and Chapter 2. The targeted markers were amplified and sequenced as eight distinct regions (four multi-gene clusters, psbD-psbC, psbE-psbF-psbL-psbJ, psbB-psbT-psbN-psbH, and 3’-rps12-rps7-ndhB, and four single-gene regions, ndhF, rpl2, rbcL, and atpB, respectively) using primers and protocols described in Graham and Olmstead (2000a), Rai and Graham (2010) and Chapter 2. I recovered sequences from replicate amplification products generated from independent DNA extracts to detect pipetting errors or cross- !! "#! contamination. I detected none, although I observed several clear instances of polymorphisms from extracts involving multiple individuals from the same species, including Andreaea rupestris (for psbD), Dicranum scoparium (for ndhB, ndhF, psbF and rbcL) and Orthotrichum lyellii (for the ndhB, rps7, rps12 introns and the rps12-rps7 spacer region). I recorded these within- population polymorphisms using ambiguity codes in the consensus sequences. All regions were sequenced at least twice to generate contigs that are largely bidirectional. I used the published sequences of Physcomitrella patens (Sugiura et al., 2003; Miyata and Sugita, 2004) to define gene and exon boundaries, and excluded noncoding regions from the vascular plants and algae to minimize alignment ambiguities. Sequences were aligned for each major region using Se-Al ver. 2.0 (Rambaut, 2002), according to criteria set out in Graham et al. (2000); see also Kelchner (2000) and Simmons and Ochoterena (2000). I offset hard-to-align noncoding regions that were frequently limited to single taxa in a staggered manner (Graham et al., 2006), although some of these regions include aligned blocks with multiple taxa, potentially providing phylogenetic information. The final alignment for all regions is 25,064 bp in length. This approach contributes significantly to the length of the alignment (Table 3.1), which is derived from ~10-13 kb of unaligned data for the taxa from which I attempted to sequence all 17-gene regions, and from ~8-10 kb of unaligned data where I attempted to sequence a subset of 14-gene regions (i.e., excluding 3’-rps12-rps7-ndhB region). 3.3.3 Phylogenetic analyses—Maximum likelihood and parsimony analyses I performed maximum likelihood (ML) and parsimony (MP) analyses using Garli 1.0 (Zwickl, 2006) and PAUP* 4.0 b10 (Swofford, 2002) respectively, on concatenated matrices or subsets of them representing individual data partitions. The MP analyses used default settings !! "#! with TBR branch-swapping, but with 100 random addition replicates. I used the GTR + ! + I model in ML analyses, with all model parameters estimated and two independent search replicates per analysis. I assessed branch support in ML and MP analyses using non-parametric bootstrapping (Felsenstein, 1985) with 300 bootstrap replicates, using simple taxon addition for MP, and otherwise using the search conditions described above. In the following discussion, I consider bootstrap support less than 70% as weak, between 70% and 90% as moderate, and higher than 90% as strong (e.g., Graham et al., 1998). I also considered several alternative tree arrangements (see Results) by constraining the monophyly of clades of interest that were not seen in the optimal trees, searching for the best ML tree satisfying a given constraint, and evaluating the significance of the difference in shortest unconstrained vs. constrained trees using the Approximately Unbiased (AU) test (Shimodaira, 2002) as implemented in CONSEL (Shimodaira and Hasegawa, 2001). All tree comparisons for a given data set were made simultaneously (see Goldman et al., 2000) and considered the full data set. I ran individual ML analyses on eight subpartitions of the data comprising individual genes (atpB, rbcL, ndhF, rpl2) or local clusters of genes that included intergenic spacer regions (psbD-psbC, psbE-psbF-psbL-psbJ, psbB-psbT-psbN-psbH, 3’rps12-rps7-ndhB). In all but one instance (ndhB in 3’rps12-rps7-ndhB) the multigene clusters comprise cotranscribed genes with related function (i.e., photosystem II or ribosomal protein genes). Three genes include introns (one each in rpl2, 3’rps12 and ndhB). I found two cases of possible strong conflict among regions, considering ML bootstrap support values. These concerned the relative arrangement of Atrichum and Buxbaumia for the psbD-psbC region (listed as clades p vs. p2 in Table 3.2), and an arrangement of Chara and Zygnema relative to land plants for the psbE-psbF-psbL-psbJ !! "#! region (clades a and a2 in Table 3.2). These conflicts were not significant according to the AU test (P > 0.05), and so I concatenated all plastid regions into a combined data set using four different permutations: all 17 genes combined (with or without associated noncoding regions for bryophytes), and 14 genes combined, either with or without the associated noncoding regions for bryophytes. The 14- vs. 17-gene comparisons were included to examine whether the absence of the three-gene region that was difficult to obtain in some taxa (the rps12-rps7-ndhB region) had a substantial effect on bootstrap support; comparisons with and without noncoding regions were performed to assess the effect of including these regions. For MP analysis I considered only the case with all 17 genes combined, including noncoding regions. For this combination 6,280 aligned sites are potentially parsimony informative, and 1,856 sites are variable but parsimony uninformative (for comparison, for the 14-gene combination including noncoding data, 4,708 aligned sites are potentially parsimony informative, and 1,093 sites are variable but parsimony uninformative). 3.3.4 Effect of rapidly evolving sites on MP inference I observed several moderate to strong conflicts between the ML and MP analysis of the concatenated data (see Results). I therefore investigated the effects of very rapidly evolving sites on MP-based inference by using HyPhy (Pond et al., 2005) to classify nucleotide sites (for the full 17-gene version, with coding and noncoding regions) into nine discrete ML rate classes, which I then set up as character sets. I repeated the site-rate assignments using the best ML and MP trees, respectively, to assess whether the tree considered for these rate classifications affected subsequent phylogenetic inference (Graham and Iles, 2009). The rate classes (RC) range from RC0 to RC8 (i.e., slowest to fastest; Burleigh and Mathews, 2004). I ran MP analysis on !! "#! three pooled subsets of rate classes: RC0-RC5, RC0-RC6, and RC78 (i.e., two slow subsets and one fast subset of the data). 3.4 Results 3.4.1 Phylogenetic inference—ML analysis The results for vascular plants have been reported elsewhere (e.g., Graham and Olmstead, 2000; Rai et al., 2003), and so I focus here on phylogenetic results for bryophytes and algal outgroups. The best ML trees inferred from the various concatenated 14- and 17-gene matrices are identical to the topology shown in Fig. 3.1. Not surprisingly, the trees inferred from these combined analyses are better supported than in the analyses of individual data subpartitions (Table 3.2), which are nonetheless almost entirely congruent with each other (two exceptions are noted in the Materials and Methods). ML bootstrap support for most clades from the concatenated 14- and 17-gene data sets is generally highly comparable across analyses (Table 3.2), and almost all multi-family clades that involve bryophytes receive 80-100% ML bootstrap support from all four ML analyses (i.e., clades b-k, l, n-p, q, s, t, v, w; Fig. 1, Table 3.2); these clades represent almost all of the early backbone of bryophyte phylogeny. A clade supporting Zygnema (representing Zygnematales) as the sister group of land plants (clade a, Fig. 3.1) is moderately well supported by ML analyses based on the 17-gene data sets, with or without noncoding regions (Table 3.2). Only one clade is weakly supported in all ML analyses (clade u, representing a branch within Bryidae). Inclusion vs. exclusion of the hard-to-retrieve 3’rps12- rps7-ndhB region seems to have little effect at this taxon sampling. A minor exception is clade r, which is slightly weaker when this region is excluded. Excluding the noncoding regions also has !! ""! a relatively minor effect, with only two clades differing noticeably. One of these, clade m (Andreaeaceae-Andreaeobryaceae), is 16-21% less well supported with the noncoding regions included; the other, clade r (corresponding to the most-recent common ancestor of Dicranidae and Bryidae), is 30-34% better supported with the noncoding regions included. 3.4.2 Phylogenetic inference—MP analysis Most bryophyte clades are identical in shortest trees inferred by the MP and ML analyses (Fig. 3.2), with three major exceptions. For MP, the first split in moss phylogeny is defined by a clade comprising Sphagnum and Takakia. This arrangement strongly contradicts the ML results (i.e., clades k2 vs. k, respectively, in Table 3.2). In addition, the liverworts are not monophyletic on the best MP tree, although their lack of monophyly is not well supported by MP bootstrap analysis (Fig. 3.2). A moderately well supported clade comprising Zygnema (Zygnematales) and Chara (Charales) contradicts the ML results (i.e., clades a2 and a, both have 72% bootstrap support, see Figs. 3.1, 3.2; Table 3.2). The MP results therefore weakly contradict the ML analysis concerning the sister group of land plants, as the clade comprising embryophytes, Chara and Zygnema has only 64% MP bootstrap support (Fig. 3.2). Seven multifamily clades have substantially reduced bootstrap support from MP compared to their generally strong support in ML bootstrap analysis (i.e., clades e, i, l-p), and only one clade is substantially stronger (clade u, which is still only moderately supported by MP bootstrap analysis, Table 3.2). 3.4.3 Exploration of conflicts between ML and MP analysis  The AU test rejects the MP arrangement for the deepest split in moss phylogeny (Fig. 2): when Sphagnum and Takakia are constrained as a clade, the resulting ML tree is significantly !! "#! longer than the best ML arrangement, in which Takakia is the sister group of all other mosses (Table 3.3). A third possible arrangement, with Sphagnum sister to all other mosses, is also rejected by the AU test (Table 3.3). An ML tree resulting from a search in which Zygnema and Chara are constrained as a clade (as in the best MP tree, Fig. 3.2) is not significantly longer than the moderately well supported arrangement found in the unconstrained ML tree in Fig. 1 (Table 3.3); note that several other possible outgroup arrangements assessed here cannot be rejected. The effect of rate filtering on MP inference of the first split in moss phylogeny is sensitive to the tree used to make site-rate classifications. MP analysis of rate-filtered data recovered the ML arrangement, with Takakia sister to mosses for RC0-RC5, but only when the ML tree was used to generate rate classifications, and then with only moderate support (Table 3.4). In all other rate-filtered MP analyses, the Sphagnum-Takakia clade was recovered with moderate to strong MP bootstrap support. MP analysis recovers the ML arrangement for the sister-group of land plants when the most rapidly evolving sites are removed from consideration, with weak to strong bootstrap support. This depended in part on the precise set of characters considered (see RC0-RC5 and RC0-RC6 results concerning Zygnema and Chara in Table 3.4). For the RC0-RC5 results, this did not depend on whether the tree used for rate classifications was the ML or MP tree.  !! "#! 3.5 Discussion 3.5.1 A well supported backbone of bryophyte phylogeny The ML relationships I recovered for the three major clades of bryophytes, with liverworts strongly supported as the sister group of other land plants, and hornworts as the sister group of vascular plants (Fig. 3.1, Table 3.5), agree with other recent studies (reviewed in Qiu, 2008). These relationships have been controversial until relatively recently, but although multiple lines of evidence now support them, they have not always been well supported by bootstrap analysis (e.g., a maximum of 87% ML or MP bootstrap support among studies summarized in Table 3.5, compared to 94-100% here for ML; clades c and d in Table 3.2). Other major relationships reported here within liverworts, hornworts and peristomate mosses are also consistent with recent studies that include broader samples of taxa for fewer genes, generally with at least as good ML bootstrap support here as reported elsewhere (Table 3.5). My exemplar-based taxon-sampling typically permits me to make exact comparisons of the support for major clades to other recent studies (see Materials and Methods, and see footnote 2 in Table 3.5). Twenty of the 23 multi-family clades labeled in Fig. 3.1 have at least 90% ML bootstrap support (based on the 17-gene matrix that includes noncoding data in bryophytes; Tables 3.2, 3.5). The plastid genes employed here have also been used successfully in inferring the deep phylogeny of various vascular plant groups (e.g., Graham and Olmstead, 2000a; Graham et al., 2006; Saarela et al., 2007; Rai et al., 2008; Zgurski et al., 2008; Rai and Graham, 2010). A region comprising three of the 17 genes was difficult to retrieve for some bryophytes (Chapter 2). However, excluding these three genes (and associated noncoding regions) generally !! "#! had minimal effect on phylogenetic inference: the topology recovered using ML was the same for various 14- and 17-gene combinations in ML analyses, and support values were generally comparable (Table 3.2). This suggests that the 14-gene combination will generally provide sufficient information to infer major splits in bryophyte deep phylogeny with strong support. The gene set surveyed here is the largest attempted to date for this taxonomic breadth of bryophytes (Table 3.1) and our results indicate that they should be useful in future studies of higher-order relationships in the major clades of liverworts, hornworts and peristomate mosses. 3.5.2 Nonperistomate mosses and the earliest splits in moss phylogeny Our ML analyses strongly support the monotypic Oedipodium as the sister group of peristomate mosses, congruent with the results of Cox et al. (2004), and consistent with the suggestions of Newton et al. (2000) and Goffinet et al. (2001). Support for a sister-group relationship between Andreaea and the monotypic Andreaeobryum is consistently found in our ML analysis (Table 3.2), and also in other studies that include both taxa (summarized in Cox et al., 2004; see also Table 3.5). This clade is moderately to strongly supported here in the 14- and 17-gene analyses that included coding regions only, but oddly, including noncoding regions reduced support by 16-21%. Volkmar and Knoop (2010) recently recovered this relationship with strong support, based on combined evidence from five plastid and mitochondrial genes. In the future I will sample Andreaea more densely for the current wholly plastid-based sampling. Sampling more densely within the clade comprising Sphagnum and its generic segregates would also be useful, and may help break up this branch substantially (Shaw et al., 2010a, b). !! "#! The most striking result here is the finding in ML analysis that Takakia is the sister group of all other mosses, with the next split being between Sphagnum and the remainder (Fig. 3.1). The position of Takakia in bryophyte phylogeny is considered problematic, and indeed before a sporophyte generation was known it was placed with liverworts based on its simple gametophyte morphology (reviewed in Kenrick and Crane, 1997; Cox et al., 2004). Published studies disagree on the position of Takakia, with some recovering the Takakia-sister arrangement (Yatsenyuk, 2001; Forrest et al., 2006; Volkmar and Knoop, 2010). Forrest et al. obtained strong support for this placement using Bayesian inference (see Table 3.5); Volkmar and Knoop recovered this from analysis of mitochondrial but not plastid data. Other researchers have obtained a contrasting arrangement with Takakia sister to Sphagnum (e.g., Cox et al., 2004; Qiu et al., 2007). I observed a strong conflict between my ML and MP analyses, the former strongly preferring Takakia as the sister group of all other mosses, the latter placing it strongly with Sphagnum, with these two then weakly supported as the sister group of other mosses (Fig. 3.2). The contrasting placements of Takakia may therefore be indicative of a long-branch attraction problem; ML analyses are known to be less prone than MP to systematic error (e.g., Chang, 1996; Huelsenbeck, 1997, 1998; Sullivan and Swofford, 2001; Swofford et al., 2001). However, the Takakia-Sphagnum arrangement has also been recovered with moderately strong support from ML bootstrap analysis in other published studies (e.g., Table 3.5). I examined this question further by repeating the MP analysis after removing nucleotides that evolve rapidly according to an ML-based rate classification, since these characters may be more prone to saturation effects (e.g., Felsenstein, 1983). While I recovered a moderately well- supported Takakia-sister result using this data transformation, this result was sensitive to the !! "#! reference tree used for rate classifications, which suggests that this method of rate filtering may not be suitable for ameliorating long-branch problems in MP analysis of bryophyte phylogeny (see also Graham and Iles, 2009, who came to a similar conclusion for angiosperms). The strong conflict reported here concerning the first split in moss phylogeny underlines that this question warrants further careful study. However, the majority of the long terminal branches among nonperistomate mosses can not be broken up by denser taxon sampling, as I sampled all known extant species for three of five of these lineages (i.e., Andreaeobryum, Oedipodium, Takakia). 3.5.3 The sister-group of land plants The closest algal relative to land plants also remains controversial. Charales, Coleochaetales, Zygnematales, or various combinations of these taxa are posited as the sister group of land plants (Becker and Marin, 2009). Recent studies based on DNA and protein-based analyses of whole plastid genomes (Turmel et al., 2006, 2007) recovered Zygnematales as the sister group of embryophytes, generally with strong ML support. In contrast, Finet et al. (2010) analyzed a large number of nuclear ribosomal protein genes and found strong ML support for Coleochaete (Coleochaetales) as the sister group of land plants, with Chaetosphaeridium, another member of Coleochaetales, nested in Zygnematales. However, an expanded nuclear gene analysis supports Zygnematales as the sister group of embryophytes (Wodniok et al., 2011). Although my sampling of outgroups to land plants is relatively limited, I included representatives of all major clades here, and my ML analysis moderately supports the Zygnematales-sister hypothesis (Fig. 3.1); note the moderate conflict with my MP results (cf. Figs. 3.1, 3.2). When I filtered out the highest rate sites based on the best ML or MP tree, considering the slowest sites (i.e., RC0-5), MP recovered the arrangement with Zygnematales as the sister group of land plants !! "#! with moderate to strong bootstrap support (96% and 79% respectively, depending on whether the best ML and MP tree was used for site rate assignment; Table 3.4). However, the AU test could not reject three of four alternative hypotheses for the sister group of land plants here using the full data set (Table 3.3; only Charales as sister is rejected at P < 0.05). A Zygnematales-sister arrangement has been questioned as a long-branch artefact by Qiu (2008). This may be true, and the existence of long-branch artefacts may be supported by the moderate ML vs. MP conflict observed here (clades a vs. a2 in Table 3.5). An expanded taxon sampling would certainly be useful in all current phylogenetic studies that examine the transition from charophytes to bryophytes. 3.5 Conclusions and future work Phylogenetic congruence (the recovery of the same or similar tree topologies using different genes and/or taxon samplings) is a key criterion in the systematist’s toolkit for assessing the accuracy of recovered relationships, as it supports the idea that the inferred branching relationships reflect evolutionary history rather than random error (e.g., Penny et al., 1982; Chase et al., 1993; Hillis, 1995; Graham et al., 1998). In contrast, topological discordance among studies or between likelihood and parsimony analysis may be indicative of systematic error, such as long-branch attraction (Felsenstein, 1978). I have shown that a concatenated subset of 14 plastid regions, that are straightforward to amplify and sequence (Chapter 2), collectively permit inference of the broad backbone of bryophyte phylogeny in the broader context of early land- plant evolution. These loci corroborate corresponding clades in other studies, generally with strong bootstrap support here, and in several cases provide the highest support values observed among these studies. I therefore conclude that these plastid regions will be valuable for !! "#! addressing further surveys of deep and recent phylogenetic relationships in bryophytes, and I do so in Chapter 4 for the most species-rich group of bryophytes, the peristomate mosses. Finally, it should be straightforward to use amplification products from my primers in multiplexed next- generation sequencing reads, as has been done for some whole plastome studies (e.g., Cronn et al., 2008). This could facilitate more rapid retrieval of this gene set from large numbers of taxa, including those represented by small or degraded samples (many of my current samples came from limited or herbarium samples). The regions explored here are straightforward to align and annotate, and my current study also suggests that the gene set examined here may be sufficient to recover many deep splits with confidence without recourse to recovering whole plastomes for all taxa. Our approach should therefore complement ongoing efforts to collect whole-plastid genome data sets for mosses and relatives. !! " # ! Table 3.1. Selected recent multigene studies of higher-level phylogenetic relationships in various bryophyte lineages. ML = maximum likelihood; MP = parsimony; BI = Bayesian inference; NJ = Neighbor joining; nu = nuclear; pt = plastid; mt= mitochondrial. ! Authors Primary focus group (no. of exemplar species) Other groups (no. of  exemplar species) No. of genes Aligned nucleotides Genes Method(s) of analysis This study Bryophytes (23) Green algae (5) Vascular plants (15) 14-17 25,064 a  All pt: The 14-gene set comprises atpB, ndhF, psbB, psbC, psbD, psbE, psbF, psbJ, psbL, psbN, psbH, psbT, rbcL, rpl2. The 17-gene set adds ndhB, rps7, 3’-rps12 ML, MP Cox et al. (2004) Mosses (30) Liverworts (4)  8 11,071 mt: nad5, nad7 nu: 18S rDNA, 26S rDNA pt:  trnL-trnF, psbA, rbcL, rps4 ML, MP, BI Forrest et al. (2006) Liverworts (173) Algae (4) Mosses (10) Hornworts (5) Vascular plants (2) 5 8,245 mt: nad5 nu: 26S rDNA pt: psbA, rbcL, rps4 MP, BI Qiu et al. (2006) Land plants (184) Green algae (9) 6 13,631 mt: LSU nu: 18S rDNA pt: atpB, LSU, SSU, rbcL  ML, MP Qiu et al. (2007) Land plants (181) Green algae (9) 7 14,553 mt: atp1, LSU nu: 18S rDNA pt: atpB, LSU, SSU, rbcL ML, MP Volkmar and Knoop (2010) Mosses (33) Liverwort (11) 4 5,453 mt: cox1 intron, nad5 pt: rbcL, rps4 ML, BI Wahrmund et al. (2009) Mosses (50) -- 5 7,549 mt: nad2, nad5 pt: rbcL, rps4 MP, BI, NJ Wahrmund et al. (2010) Mosses (56) -- 6 8,921 mt: cob intron, nad2, nad5 pt: rbcL, rps4 ML, BI a  Aligned length for the taxon subset considered in the present study, including offsets (staggered regions typically limited to unalignable portions for single taxa). The corresponding unaligned length ranges from ~8-13 kb per taxon in bryophytes, depending in part on whether we attempted to retrieve 14 or 17 genes (see Appendix 1 for list of genes retrieved by taxon). Noncoding regions were excluded for algae and vascular plants. ! !! " # ! Table 3.2. Summary of bootstrap support for multi-family clades of bryophytes and relatives (excluding vascular plants) based on different data combinations and sub-partitions. Clade labels correspond to branches (taxon bipartitions) in Figs. 3.1 and 3.2, except branch p2 (which defines a clade comprising Atrichum and Buxbaumia). All listed clades have at least 70% support from at least one analysis; bootstrap values <50% are noted as “--". Some taxa have missing data for individual regions (see Results); the corresponding branch support is labeled as “na.” Abbreviations: psbDC = psbD-psbC; psbBTNH = psbB-psbT-psbN-psbH; psbEFLJ = psbE-psbF-psbL-psbJ; individual regions include coding and noncoding regions where relevant.  !! " # !    Combined data sets Individual regions  14 genes (coding only) 14 genes + noncoding regions 17 genes (coding only) 17 genes + noncoding regions 17 genes + noncoding regions atpB rbcL ndhF psbDC psbEFLJ psbBTNH rpl2 3’rps12- rps7-ndhB  ML ML ML ML MP ML ML ML ML ML ML ML ML a -- -- 76 72 -- -- -- -- -- -- -- -- 93 a2 -- -- -- -- 72 -- -- -- -- 72 -- -- -- b 100 100 100 100 100 78 82 100 95 -- 96 -- 100 c 99 99 100 100 93 74 76 -- -- -- -- -- 75 d 96 94 97 97 90 61 67 -- 63 -- 78 -- 70 e 100 100 100 100 57  -- -- na 75 57 96 55 na f 100 98 96 94 89 57 63 na 78 na 98 -- na g 100 100 100 100 96 55 60 99 95 96 90 -- 100 h 100 100 100 100 97 99 99 98 90 62 100 -- na i 91 93 86 93 75 -- 52 na 77 82 58 94 na j 88 93 100 100 85 -- -- 96 -- -- -- -- 100 k 81 89 97 99 -- -- -- 82 -- -- -- na 94 k2 -- -- -- -- 99 -- -- -- -- 64 -- na -- l 100 100 100 100 -- -- -- 94 -- -- -- -- 100 m 95 79 85 64 -- -- -- 63 -- -- 79 -- -- n 100 100 100 100 62 63 58 97 -- -- 87 -- na o 100 100 100 100 78 -- 50 79 -- -- 97 -- 100 p 95 95 94 93 81 68 66 92 -- 53 94 -- na p2 -- -- -- -- -- -- -- -- 98 -- -- -- -- q 100 100 100 100 100 91 93 100 83 58 100 61 100 r 53 87 65 95 100 -- -- -- 55 90 -- 58 77 s 100 100 100 100 100  57 60 -- 95 94 75 85 100 t 100 100 100 100 100  -- -- 100 87 81 82 74 na u -- -- -- 54 80 -- 60 -- -- 51 72 69 100 v 100 100 100 100 100 100 100 100 100 95 100 na na w 100 100 100 100 99 92 90 93 64 na 100 na na !! " " ! Table 3.3#!Approximately Unbiased (AU) tests of whether suboptimal arrangements concerning early moss phylogeny and the sister group of land plants are significantly worse than the optimal one. In the best likelihood tree, Takakia is the sister group of all other mosses, and Zygnema the sister group of embryophytes. !L = increase in –lnL compared to the best tree. The analyses consider all data combined (i.e., all 17 genes and conservative noncoding regions). The charophyte orders represented here are: Charales (Chara); Coleochaetales (Chaetosphaeridium); Zygnematales (Zygnema).  Arrangement of interest    !L p value (Takakia, (all other mosses))  best ML tree (Sphagnum, (all other mosses))   24.3441 0* First split in mosses ((Sphagnum, Takakia), (all other mosses)) 19.9754 0.0289* Zygnema best ML tree Chaetosphaeridium  12.9220 0.1529 Chaetosphaeridium-Zygnema  15.151 0.1656 Chara-Zygnema  14.0910 0.3137 Sister group of land plants Chara 19.9806 0.030* Notes:!* P < 0.05: significantly worse. ! !! " # ! Table 3.4. Bootstrap support for contrasting arrangements at the base of mosses, and concerning the sister group of land plants. The analyses considered all data combined (i.e., all 17 genes and associated noncoding regions) or subsets involving different rate classes (RC; see text) as determined using different reference trees. Bootstrap values with <50% support are indicated as ‘--'.   Bootstrap support determined by:   ML (MP) (unfiltered) MP only (using rate-filtered data)   Tree for rate classification 1  n/a (1) Best ML tree (2) Best MP tree   Rate classes 2  n/a RC0-RC5 RC0-RC6 RC78 RC0-RC5 RC0-RC6 RC78  Arrangement of interest: Takakia sister to other mosses 100 (--) 81 -- -- -- -- -- Takakia sister to Sphagnum (--) 99 -- 84 100 73 96 98  Zygnema sister to land plants 72 (--) 96 85 -- 79 -- -- Zygnema sister to Chara 3  -- (72) -- -- 78 -- 51 52   1 Site rate classifications determined using the best ML or MP tree, respectively. 2  Subsets of sites used in MP inference, according to the rate classification. RC0-RC5 includes the slowest six classes. RC78 includes the two fastest rate classes. 3  [Zygnema + Chara] then sister to land plants, except for RC78, where Chaeotosphaeridium (Coleochaetales) is sister to land plants. ! !! " # ! Table 3.5. Support for multi-family clades of bryophytes and relatives, excluding vascular plants, across selected studies, as determined by likelihood bootstrap (ML), parsimony bootstrap (MP) or posterior probability (PP, expressed as a percentage here). Support values for each study are for the largest combinations of genes noted in Tables 3.1 and 3.2 (e.g., 17 genes and noncoding regions for the current study). Labels in first column correspond to branches in Figs. 3.1 and 3.2 (except a3). Author abbreviations: C04, Cox et al. 2004; F06, Forrest et al. (2006); Q07, Qiu et al. 2007; V10, Volkmar and Knoop (2010); W10, Wahrmund et al. (2010). Clade composition comparisons are relative to Fig. 1 in Qiu et al. (2007) for clades b-d, Fig. 1 in Forrest et al. (2006) for clades e-i, Fig. 3 in Cox et al. (2004) for clades j and k2-n, Fig. 3 in Wahrmund et al. (2010) for clades o-u, and Fig. 2 in Duff et al. (2007) for clades v, w. MRCA = most recent common ancestor.   Taxon bipartition (clade on rooted tree) This study ML (MP) C04 ML (MP, PP 1 ) F06 MP (PP 1 ) Q07 ML V10 ML (PP) W10 ML (PP) a [Zygnematales + land plants] 72 (<50)  -- -- -- -- -- a2 [Zygnematales + Charales] <50 (72) -- -- -- -- -- a3 [Charales + land plants] <50 (<50) -- 100 (--) 93 -- -- b Land plants 100 (100) -- 99 (--) 100 -- -- c Landplants excl. liverworts 100 (93) -- 59 (--) 87 -- -- d [Hornworts + vascular plants] 97 (90) -- 61 (100–100) 87 -- -- e Liverworts 100 (57)  -- 100 (100–100) 100 100 (100)  -- f Haplomitriopsida 94 (89) -- 99 (100–100) 83 100 (100) 2  -- g Liverworts excl. Haplomitriopsida 100 (96) -- 98 (100–100) 92 100 (100) -- h Jungermanniopsida 100 (97) 2 95 (75, 100–100) 2 100 (100–100) 100 100 (100) -- i Metzgeriidae 93 (75)  -- 99 (100–100) 100 2 -- -- j Mosses 100 (85) 100 (100, 100–100) 98 (100–100) 100 100 (100) -- k Mosses excl. Takakiaceae 99 (<50) -- -- (58–97) --  -- -- k2 [Takakiaceae + Sphagnaceae] <50 (99) 86 (100, 99–100) -- 86 79 (--) -- l Mosses excl. Takakiaceae & Sphagnaceae    Sphagnupsida AandanSSSSSSphagnopsida 100 (<50) 67 (--,  99–100) 96 (100–100) 100 94 (100) 100 (100) m [Andreaeaceae + Andreaeobryaceae] 64 (<50) -- (--, 86–90) -- -- 95 (100) -- n [Oedipodiaceae + peristomate mosses] 100 (62) 91 (--, 99–100) -- -- -- -- o Peristomate mosses  100 (78)  72 (--, 99–100) 94 (100–100)  100 98 (100) 100 (100) p Bryopsida 93 (81) 83 (--, 99–100) (= t)  100 95 (100)    100 (100) q Bryopsida excl. Buxbaumiaceae 100 (100) 2 100 (77, 100–100) (= t) 100 100 (100) 100 (100) r MRCA of Dicranidae  & Bryidae 95 (100) 78 (79, 100–100) -- 100 100 (100) <50 (--) s Dicranidae  100 (100) 2 100 (100, 100–100) 2  -- 100 2 91 (100) 2  98 (100)  t Bryidae  100 (100) 2 100 (100, 100–100) 100 (100–100) 2 100 100 (100) 100 (100) u MRCA of Orthotrichaceae & Brachytheciaceae 54 (80) -- (--, 53–85) -- <50 -- 96 (100) v Hornworts 100 (100) -- 100 (100–100) 100 -- -- w Anthocerotopsida 100 (99) -- 75 (91-100) 85 -- -- --: Precise support not noted or branch not applicable given the taxa sampled. 1  Range of posterior possibilities noted when multiple Bayesian analyses were done. 2  MRCA of exemplar taxa is less inclusive than in corresponding reference studies.! ! !! "#!  Fig. 3.1. Phylogram of best ML tree based on combined analysis of 17 genes and associated noncoding regions (-lnL = 181836.287). ML bootstrap values indicated as numbers beside branches; those with <50% support indicated as ‘--'. Letter labels indicating major clades of bryophytes and early land plants referred to in Tables 3.2 and 3.5 and the text.  !! "#! ! Treubia Haplomitrium Pleurozia Metzgeria Andreaeobryum Andreaea Oedipodium Leiosporoceros Adiantum Zamia !"#$%& Podocarpus Pinus Gnetum Nymphaea Amborella Trimenia Magnolia Nicotiana Lilium '(')*+,-+.".,."&#+/+".0 !"#$%&'"#() *+)),) -")./'"0%&'"#() 1,,$%&'"#() Mesostigma Chlorokybus Chara Chaetosphaeridium Zygnema Marchantia Scapania !23,04+0() Takakia ceratophylla Takakia lepidozioides Sphagnum Atrichum Buxbaumia Physcomitrella Dicranum Syntrichia Bryum Orthotrichum Homalothecium 5,02)(+6"(, 6+)),) Anthoceros Huperzia Psilotum Angiopteris 7+0#4+0()Phaeoceros a b c d e f g h i j k l m n o p q 95 s t u v 72 100 94 93 99 64 100 100 100 100 100 100 100 100 100 100 100 93 97 100 r 100 100 100 100 100 100 69 86 97 -- 72 58 100 100 -- 54 w 8+#&,02)(+6"(, 6+)),) 100 100 100 !! "#! Fig. 3.2. Phylogram of the single best MP tree based on combined analysis of 17 genes and associated noncoding regions (tree length = 36527 steps). MP bootstrap values indicated as numbers beside branches; those with <50% support indicated as ‘--'. Letter labels indicating major clades of bryophytes and early land plants referred to in Tables 3.2 and 3.5 and the text.  !! "#! ! Mesostigma Chara Zygnema Treubia Marchantia Pleurozia Metzgeria Scapania Sphagnum Andreaea Oedipodium Atrichum Buxbaumia Dicranum Bryum Phaeoceros Anthoceros Huperzia Podocarpus Zamia Ginkgo Nicotiana Amborella Nymphaea Trimenia Magnolia Lilium Gnetum Psilotum Adiantum Angiopteris !""#$%&'()* Chlorokybus Chaetosphaeridium Haplomitrium Takakia ceratophylla Takakia lepidozioides Andreaeobryum Physcomitrella Syntrichia Orthotrichum Homalothecium Leiosporoceros Pinus 100 100 100 100 100 100 100 100 b c d f e2 g h i j k2 l m n o p q r s t u v a2 64 72 89 -- 96 97 75 85 99 -- -- 62 78 81 80 93 90 99 89 73 100 76 59 58 -- 100 -- -- -- -- 94 94 !"#$%&'"#() *+)),) -")./'"0%&'"#() 1,,$ &'"#() !23,04+0() 5,02)(+6"(, 6+)),) 7+0#4+0() w 8+#&,02)(+6"(, 6+)),) !! "#! 4. BACKBONE PHYLOGENY OF PERISTOMATE AND NONPERISTOMATE MOSSES BASED ON MULTIPLE PLASTID MARKERS 4.1 Brief synopsis To further investigate relationships among major groups of peristomate mosses, I retrieved 14-17 genes from a total of 35 peristomate species, and additional representatives from the nonperistomate lineages. The phylogenetic relationships recovered here generally were consistent with other studies on peristomate mosses, and I recovered improved support for multiple clades, including several that have not been satisfactorily resolved to date. For example, within Bryopsida, Timmiidae was found to be the sister group of a clade comprising Dicranidae and Bryidae, with Funariidae then the sister group of Bryidae-Dicranidae-Timmiidae, all with strong maximum likelihood (ML) bootstrap support. Higher-order relationships within Funariidae and Dicranidae were also generally well resolved and supported, while relationships within Bryidae were less well supported, possibly reflecting a rapidly early radiation in this major clade of mosses. Conflicting arrangements of Takakiaceae and Sphagnaceae at the base of moss phylogeny were obtained with ML inference, depending on the method of data partitioning and how base frequencies were parameterized in the DNA substitution model. 4.2 Introduction The relatively simple morphology of mosses results in a paucity of characters suitable for the inference of higher-order phylogenetic relationships. While sporophyte characters like those !! "#! related to the morphology and development of peristome teeth are useful for defining major taxonomic groups, due to their presumed evolutionary stability, they are not numerous or ubiquitous enough to permit confident inference of all aspects of the broad backbone of moss phylogeny (Goffinet et al., 2001; Cox et al., 2004). Gametophytic characters, on the other hand, appear to suffer from frequent secondary loss and convergence, which may lead to incorrect inference of higher-order relationships (Goffinet et al., 2009). For the past two decades, studies based on molecular data have therefore provided important new evidence for improving our understanding of phylogenetic relationships among and within major moss groups (e.g., Goffinet et al., 2001, La Farge et al., 2002; also Fig. 4.1). While a considerable degree of consensus has been reached, substantial ambiguities still remain, including multiple relationships along the major backbones of moss phylogeny (Renzaglia et al., 2007; Goffinet et al., 2009). The earliest splits in moss phylogeny are defined by nonperistomate mosses. In one plausible scenario, Takakia (Takakiopsida: Takakiaceae) is the sister group of the remaining mosses, with Sphagnum and relatives (Sphagnopsida) the next successive sister group (e.g., Forrest et al., 2006; Chapter 3); alternatively a clade comprising Takakiopsida and Sphagnopsida may be the sister group to all other mosses (e.g., Cox et al., 2004, Qiu et al., 2007; Chapter 3). A clade comprising Andreaeopsida and Andreaeobryopsida may then represent the next successive sister group (Frey and Stech, 2009; Chapter 3), with the monotypic Oedipodiopsida the sister group of all peristomate mosses (Cox et al., 2004; Wahrmund et al., 2009; Chapter 3). Within the peristomate moss clade, two lineages (Polytrichopsida and Tetraphidopsida) have solely nematodontous peristome teeth, and one has arthrodontous peristome teeth (Bryopsida). The relationships among these three lineages remain uncertain (e.g., Qiu et al., 2006, 2007; Volkmar !! "#! and Knoop, 2010), although evidence from peristome development may support a close relationship between Tetraphidopsida and Bryopsida (Shaw and Anderson, 1988). The arthrodontous taxa (Bryopsida), which comprise ~90% of all mosses (Shaw et al., 2003), include several lineages that are somewhat intermediate in peristome form and development between arthrodontous and nematodontous types, such as Buxbaumia (Buxbaumiidae). Most of the remaining “true” arthrodontous taxa can be further defined by their possession of diplolepideous- opposite, diplolepideous-alternate or haplolepideous peristomes. The diplolepideous-alternate mosses are classified in subclass Bryidae, and haplolepideous mosses are placed in Dicranidae. These two subclasses may be sister groups to each other (Newton et al., 2000; Cox et al., 2004; Qiu et al., 2006). The interpretation of the peristome in Timmia (Timmiidae) and Encalypta (Funariidae) remains controversial (Shaw and Renzaglia, 2004); the phylogenetic position of these taxa relative to the diplolepideous-opposite mosses (Funariales), and Bryidae-Dicranidae is also unclear (e.g., Volkmar and Knoop, 2010; Wahrmund et al., 2010). The earliest splits in moss evolution likely happened more than 300 million years ago (Newton et al., 2009) during relatively rapid early radiations that left correspondingly short internodes between major groups (Goffinet et al., 2009). Many lineages that arose at this time (e.g., nonperisomate mosses) are also quite species poor. As a consequence, even densely sampled phylogenies have substantially long branches in early moss phylogeny that are connected by short internodes. This combination may be particularly challenging for phylogenetic inference (e.g., Felsenstein, 1978, Penny and Hendy, 1989). Here I sample a relatively large number of conservative DNA markers from the plastid genome to reconstruct the backbone relationships of mosses. Slowly evolving markers may be less susceptible to any long !! "#! branch problems (Chapters 2, 3, and see Graham and Olmstead 2000 for a more detailed discussion). The plastid markers I sample here are atpB, ndhF, rbcL and rpl2, and four clusters of genes with associated intergenic spacers (i.e., psbD-psbC, psbE-psbF-psbL-psbJ, psbB-psbT- psbN-psbH, and 3’-rps12-rps7-ndhB). I developed bryophyte-specific primers for these genes and demonstrated their utility in inferring higher-order relationships of mosses based on a preliminary taxonomic sampling (Chapters 2, 3). Here I use these genes to infer the deep phylogenetic relationships of mosses using a substantially expanded taxa sampling (especially of peristomate mosses). I use different phylogenetic criteria for tree inference: maximum parsimony (MP) and several implementations of maximum likelihood (ML). I also show that at least one major relationship in moss phylogeny is sensitive to the stategy used to estimate model parameters for likelihood analysis. 4.3 Materials and methods 4.3.1 Taxonomic and genomic sampling  I sampled 75 taxa in total from various land-plant groups and algal relatives (Appendix 3). Within mosses, I sampled at least one species from each order, except for Ambuchananiales and Gigaspermales (the classification used here is that of Goffinet et al., 2009). A total of 43 exemplar mosses is included, among which 29 taxa are newly sequenced for this study. In addition, I included six liverworts, three hornworts and 15 vascular plants to represent the other major land-plant clades, and six charophycean and two chlorophycean algae as outgroups (e.g., Turmel et al., 2007). !! ""!  Among all the targeted markers, the 3’rps12-rps7-ndhB region proved the most difficult to retrieve (Chapter 2). I therefore focussed on a subset of 14 genes and associated noncoding regions from all the taxa (Chapters 2, 3). However, I was able to add 3’rps12-rps7-ndhB for several taxa that may occupy potentially critical positions in the phylogeny (i.e., Sphagnum squarrosum, Alophosia, Timmia, Funaria, Bartramia, Rhytidiadelphus and Hookeria) based on results from previous studies (e.g., Newton et al., 2000; Cox et al., 2004; Qiu et al., 2007; Wahrmund et al., 2009 and 2010), or preliminary analyses based on the 14-gene matrix in this study. 4.3.2 Sequence recovery and alignment Methods for DNA extraction, amplification, sequencing and base-calling follow Graham and Olmstead (2000) and Chapters 2, 3. I amplified and sequenced the targeted markers in eight distinct regions (comprising respectively four multi-gene clusters, psbD-psbC, psbE-psbF-psbL- psbJ, psbB-psbT-psbN-psbH, and 3’-rps12-rps7-ndhB, and four single-gene regions, ndhF, rpl2, rbcL and atpB) using primers and protocols described in Graham and Olmstead (2000), Rai and Graham (2010) and Chapter 2. The newly obtained data were aligned to a previously generated alignment (Chapter 3) with Se-Al (Rambaut, 2002), following alignment criteria outlined in Graham et al. (2000). The alignments include several noncoding regions and introns that I aligned across mosses. I excluded these regions from other green plants (Chapter 3). The final alignment is 29,266 bp in length, derived from ~10-13 kb of unaligned data for the taxa from which I attempted to sequence all 17-gene regions, and from ~8-10 kb of unaligned data where I attempted to sequence a subset of 14-gene regions (i.e., excluding 3’-rps12-rps7-ndhB region). Across all taxa included in this study, 7,014 aligned sites are potentially parsimony informative, !! "#! and 1,824 are variable but parsimony uninformative; across mosses, 4,228 aligned sites are potentially parsimony informative, and 1,855 variable but parsimony uninformative. 4.3.3 Phylogenetic analysis  I first conducted phylogenetic analysis based on the eight regions described above, using maximum likelihood (ML). I searched for the best ML tree in each case using Garli 1.0 (Zwickl, 2006), with the GTR + ! + I DNA substitution model, using ML estimates of the model parameters (i.e., with these optimized as part of tree inference). I assessed branch support using non-parametric bootstrapping (Felsenstein, 1985) with 500 bootstrap replicates for each region, with all model parameters estimated. I compared ML bootstrap support values from individual regions and found no strong conflicts (i.e., no conflicting results with bootstrap values both higher than 70%; data not shown). In the previous chapter (Chapter 3), there were two moderate to strong conflicts regarding the relationship between Atrichum (Polytrichopsida, Polytrichaceae) and Buxbaumia (Bryopsida, Buxbaumiaceae), and the relative arrangement of Zygnema, other charophytes and the land plants. I no longer saw the conflicts among regions reported in Chapter 2 with the expanded taxon sampling used here. I concatenated all of the regions for a combined ML analysis that considered two alternative data partitionings for the analysis [i.e., either by individual region (the eight regions described above, including noncoding regions with relevant regions, where present), or by “codon position” (i.e., first, second and third, with noncoding regions included as a fourth partition)]. I also performed a combined ML analysis with no subpartitions, and analyzed the combined matrix using maximum parsimony (MP). !! "#! I used PAUP* 4.0 b10 (Swofford, 2002) for MP heuristic searches, using TBR branch- swapping and 100 random addition replicates to search for the shortest MP trees. For ML heuristic searches, I used Garli 1.0 (for nonpartitioned analyses), Garli-Part-0.97 (for partitioned analyses) and RAxML-7.2.6 (Stamatakis, 2006; Stamatakis et al., 2008). Ten independent replicates were carried out for each ML search. The RAxML analyses were executed using the CIPRES portals (Miller et al., 2009). In preliminary analyses, I noticed an apparent conflict among the earliest branches in moss phylogeny in analyses using these two programs (see Results). As base frequencies can be estimated in Garli using ML optimization or empirically from terminal taxa only, and RAxML allows only the latter, I explored the possible impact of these two estimation methods on phylogenetic inference by re-running Garli using empirical etimates of base frequencies. In summary, for each matrix and data partitioning (unpartitioned, partitioned by codons or regions), I ran ML analyses once using RAxML (with empirical base frequencies) and twice using Garli (once each with ML-estimated vs. empirical base frequencies) (Table 4.1). I used bootstrap analysis with 500 pseudo-replicates for MP and the various ML analyses to assess branch support. For the bootstrap analyses, I used simple taxon addition for MP analysis, but otherwise used the search conditions described above. I also considered three alternative tree arrangements concerning the earliest split in moss phylogeny: the optimal arrangement in the unpartitioned ML analysis, and two suboptimal arrangements (see Results). In the latter two cases I used Garli to perform heuristic searches with topological constraints that reflect each suboptimal arrangement (considering ML estimates of base frequencies and with no data partitions). I then used CONSEL (Shimodaira and Hasegawa, !! "#! 2001) to assess whether suboptimal trees satisfying the constraints were significantly less optimal then the optimal tree, using the Approximately Unbiased (AU) test (Shimodaira, 2002). To do this test I determined individual site likelihoods using PAUP*, with base frequencies either estimated using ML or determined empirically, and again with no data partitions. 4.4 Results 4.4.1 Phylogenetic relationships inferred by different analyses I discussed relationships among the four major land-plant lineages in Chapter 3. My results here for these lineages are consistent with what I reported there, and so here I focus on presenting multi-family relationships within the mosses. I first discuss the results from the Garli analyses that used ML estimates of base frequencies (Figs. 4.1, 4.2, 4.3; column 3-5 in Table 4.1), followed by a comparison of results from the other ML analyses (i.e., analyses using empirical estimates of base frequency in Garli and RaxML; columns 4-9 in Table 4.1), and the MP analysis (column 10 in Table 4.1). In the following discussion, I consider bootstrap support less than 70% as weak, between 70% and 90% as moderate, and higher than 90% as strong (e.g., Graham et al., 1998; Chapter 3). For the interpretation of the results below, note that the likelihood scores estimated by different programs should not be compared directly, due to the different methods of ML optimization used by different programs (Sundberg et al., 2008), 4.4.1.1 ML analyses using ML-estimated base frequencies The results from the three Garli analyses using ML estimates of base frequencies are generally highly consistent with each other (see the first three Garli analyses in Table 4.1; one unpartitioned ML analysis; the two partitioned ML analyses). The best ML trees have essentially !! "#! the same topology (one is shown in Figs. 4.1-4.3), with only minor differences regarding the relative placements of the “early-diverging” Bryidae taxa such as Tetraplodon (Splachnales), Bartramia (Bartramiales) and Hedwigia (Hedwigiales). For comparison, among the 41 clades of interest recovered in the optimal tree based on the unpartitioned ML analysis, 33 clades receive moderate or strong in all three analyses; five clades are weakly supported in all analyses; three clades are moderately supported by analysis based on the matrix partitioned by codon-position but are weakly supported by the other two analyses (Table 4.1; Figs. 4.1, 4.2). The monophyly of land plants as a whole, and of each of the four land-plant lineages (liverworts, mosses, hornworts, vascular plants), and the relationships among these lineages (Table 4.1a; clades b-h), are among the branches that are strongly supported in all three ML analyses. The monophyly of all bryophyte classes and subclasses, and the majority of relationships among these major clades (e.g., clades j, l, m, o, p, q, r, s, t, u, and cc) are also recovered with moderate or strong support from all three analyses. Other branching patterns receiving moderate or strong support in all three analyses include the relationships among some moss orders and families (Table 4.1b; Fig. 4.2). The relationships recovered for nonperistomate mosses are consistent with those seen in Chapter 3, again with mostly strong support from all three analyses; noteably Takakia is recovered as the sister group of all other mosses. Within the perisomate mosses, the relationships among Tetraphidopsida and Bryopsida (clade n), the relationships among Grimmiales, Pottiales, and Dicranales s.s. (clade z), and the sister relationship between Hypopterygiaceae and Hypnales (clade mm) receive moderate support from the ML analysis with partitioning by codon position, but only weak support from the other two ML analyses. The clades with weak support from all !! "#! analyses mainly concern relationships within Bryidae, especially the placements of “early diverging” taxa (Splachnales, Hedwigiales, Bartramiales vs. the remaining Bridae; clades dd and ee), and the local placement of Orthotrichales (clade gg) and Ptychomniales (clade jj). The closest algal relative of land plants (clade a) also remains uncertain (Table 4.1) but is recovered as Zygnematales in the best ML trees in all three analyses. 4.4.1.2 ML analyses using empirical estimates of base frequencies The results from the other ML analyses (i.e., ML analyses using Garli and RaxML in which base frequencies are estimated empirically rather than using ML) are generally highly consistent with the analyses in which all parameters are ML estimates (Table 4.1), except for the relative placements of Takakia (Takakiopsida) and Sphagnum (Sphagnopsida). In the unpartitioned analysis and the analysis partitioned by the eight major regions, Takakia is recovered as the sister group of Sphagnum (the Takakia-Sphagnum arrangement) using either Garli or RAxML (with 69-97% support, Table 4.1, see also Table 4.2). However, in the analysis partitioned by codon position, the Takakia-sister arrangement is recovered in the optimal tree, with 100% bootstrap support in both Garli and RAxML analyses. In all ML analyses, the optimal trees inferred when the data are partioned by codon position have better likelihood scores than the optimal trees inferred from unpartitioned data, or when data partitioning is by major region (Table 4.1). 4.4.1.3 MP analysis Overall the results from the MP analysis are consistent with the results from ML analyses using ML estimates of base frequencies, except for the placements of Takakia and Sphagnum !! "#! (Table 4.1). As in Chapter 3, the MP analysis supports the Takakia-Sphagnum arrangement (here clade i2) with strong support (100%). For the other clades of interest, the MP bootstrap support is generally comparable to but lower than that observed in the ML analyses. Only 21 of the 33 clades with moderate or strong ML support also receive moderate or strong bootstrap support from MP analysis. The remaining 12 clades, most of which concern branching order within Dicranidae and Bryidae (e.g., clades ff and kk), in addition to the monophyly of liverworts (clade f) and the earliest splits mosses (e.g., clades j and k), receive only weak support in MP analyses. For the clades with weak ML support, the MP support is also generally weak, except for clade ee2 (Bartramiales-Hedwigiales; Table 4.1), which is strongly supported by MP analysis, but receives only weak to moderate support from the various ML analysis. 4.4.2 Exploration of the conflicting placements of Takakia and Sphagnum The optimal trees found in ML analyses (with ML estimates of base frequencies) depict Takakia as the sister group of other mosses. I examined whether two alternative arrangements of Takakia and Sphagnum were significantly less optimal than the best tree by comparing the optimal arrangement to trees satisfying these alternative arrangements (i.e., either with Takakia sister to Sphagnum; or Sphagnum as the sister group of other mosses). The test used site likelihoods estimated with no data partitions for the whole matrix, but used two different ways of estimating base frequencies (ML vs. empirical estimates). For the AU tests in which I used ML estimates of base frequencies the two suboptimal arrangements were strongly rejected (for Takakia-Sphagnum, and Sphagnum-sister, !lnL = 28.476 and 31.621, respectively; P < 0.001; Table 4.3). When I used empirical estimates of base frequencies, the Takakia-Sphagnum arrangement was the best scoring of the three considered, !! "#! but the AU test could not reject the Takakia-sister arrangement (!lnL = 6.802; P = 0.998; Table 4.3). However, the Sphagnum-sister arrangement was still rejected (!lnL = 26.341; P < 0.001; Table 4.3). The observed base frequencies differ substantially among codon positions. For example, the GC content of the first and second codon positions is more than twice of that of the third codon position and the noncoding regions (Table 4.4). Further comparison of the values of other nucleotide substitution model parameters reveals substantial differences among different codon positions (Table 4.4). I ran ML analyses using Garli with only the first and second codon positions included, or with only the third codon position and noncoding sites included, with either ML or empirical estimates of base frequencies. I recovered the Takakia-sister arrangement as the optimal tree in all cases, with moderate to strong bootstrap support (Table 4.2). 4.5 Discussion 4.5.1 Length variation of the noncoding regions in mosses The noncoding plastid regions examined here for mosses are not obviously more conserved in length than in the seed plants (cf. Table 2 in Rai et al. 2003 and Table 4.5 here), a crown clade of comparable age. A few studies have reported that the “early diverging” mosses have longer intergenic spacers (IGS) or introns in various mitochondrial and plastid regions (e.g., Quandt and Stech, 2004; Wahrmund et al., 2009 and 2010). This has been considered indicative of the presence of larger plastid and mitochondrial genomes in ancestral mosses and of a subsequent streamlining in “derived” (peristomate) lineages (Volkmar and Knoop, 2010). My data, however, suggest that noncoding regions are not evolving consistently in terms of their !! "#! length across the major moss lineages and plastid regions examined (Table 4.5). Comparing mean lengths for the intergenic spacer regions and introns examined here, four of six spacer regions (between psb genes) are shorter in the nonperistomate mosses than the peristomate mosses, whereas five other noncoding regions are longer, and there is substantial variation in the lengths of most of the noncoding regions (see ranges and standard deviations in Table 4.5). In general there is little generality to patterns of change in the lengths of these regions between nonperistomate and peristomate mosses (the ndhB intron is a possible exception; it is substantially shorter in the peristomate mosses considered here). 4.5.2  Towards a robust backbone phylogeny of peristomate mosses 4.5.2.1 Nematodontous mosses One of the major unresolved questions in moss evolution concerns the placement of two nematodontous lineages, the monofamilial classes Polytrichopsida and Tetraphidopsida. Both lineages are characterized by peristome teeth made from whole cells. However, the structure and development of the teeth are different between the two classes (Shaw and Renzaglia, 2004). In addition, Polytrichopsida and Tetraphidopsida differ substantially in many gametophytic features so it has been proposed that they are not closely related (e.g., Crum and Anderson, 1981; Shaw and Anderson, 1988). In contrast, the early stages of the development of peristome teeth in Tetraphidopsida is “virtually indistinguishable” from the patterns found in arthrodontous mosses (Bryopsida), which may be indicative of a close affinity between them (Shaw and Anderson, 1988). To date, however, a sister-group relationship between Tetraphidopsida and Bryopsida has not received strong support from any phylogenetic study (e.g., only weak support in Cox et al., !! "#! 2004; Fig. 4.3), and although I recovered this relationship in most of the ML analyses here (data not shown), it does not have strong support in any analysis (clade n; Table 4.1, see also Fig 4.2). This is despite a broadly representative taxon sampling that includes exemplar species from the deepest splits of these major clades: I sampled species representing the core polytrichids and the two earliest splits in Polytrichopsida (Bell and Hyvonen 2008) and included both genera of Tetraphidopsida. A confident resolution of this issue may therefore require whole plastid genome data or genomic structural characters. 4.5.2.2 Diplolepideous-opposite taxa and relatives While taxa with typically diplolepideous-opposite teeth are included in Funariidae, species in Encalyptales (Funariidae) have diverse peristome teeth morphologies whose interpretation has been controversial, as has been the order’s phylogenetic placement (Vitt, 1984; Newton et al., 2000; Magombo, 2003). Here I recovered a well-supported clade comprising Encalypatales (represented by Encalypta) and Funariales (represented by Funaria and Physcomitrella). This result is consistent with recent studies based on multi-gene analyses (e.g., Qiu et al., 2007; Wahrmund et al., 2010; Fig. 4.3). The recent discovery of a 71-kb inversion in the large single copy region of plastid genome in Encalyptales and Funariales (excluding Gigaspermaceae), which is absent in the genomes from other mosses (Goffinet, 2007), also supports this hypothesis. The interpretation of the peristome teeth of Timmia (the sole genus of Timmiidae) is also controversial (e.g., Vitt, 1984; Budke et al., 2007). Timmia possesses two rings of arthrodontous teeth, with the exostome being diplolepideous. Unlike most of the other diplolepideous taxa, the endostome of Timmiidae has no regular segment (instead it has 64 cilia). Timmia has been !! "#! classified with the diplolepideous-alternate taxa (Bryidae), because the latter sometimes also possess cilia (e.g., Vitt, 1984). However, recent phylogenetic studies indicate that Timmia may be placed outside Bryidae, although generally without strong support (e.g., Newton et al., 2000; Cox et al., 2004; Volkmar and Knoop, 2010). Here I find strong support for a sister-group relationship between Timmia and a clade comprising Bryidae and Dicranidae, with Funariidae well supported as the sister-group of Timmiidae-Bryidae-Dicranidae (clades q, s and t in Fig. 4.2). These clades are all consistently well-supported in the different ML analyses (Table 4.1). This key relationship between Timmiidae and Bryidae-Dicranidae has not been recovered with strong support before (Fig. 4.3). A recent study on the peristome development of Timmia megapolitana shows symmetric division of its eight inner peristomial layer (IPL) cells (Budke et al., 2007), similar to Funariales. Asymmetric cell division in the IPL is found in other  “early diverging” peristomate taxa (e.g., Diphyscium, Tetraphis), and in the haplolepideous (Dicranidae) and diplolepideous-alternate taxa (Bryidae) (Budke et al., 2007). The placement of Timmia here therefore implies that symmetric division of IPL cells was gained independently in Funariales and Timmiales, or that it was gained before the divergence of Funariales, and then lost before the origin of the diplolepidous-alternate and haplolepideous mosses. 4.5.2.3 Diplolepideous-alternate and haplolepideous taxa (Bryidae and Dicranidae) The existence of a common uniform cell division pattern in peristome teeth development of the haplolepideous mosses (subclass Dicranidae) supports the idea that this taxon represents a natural group (Edwards, 1979; Shaw, 1989). I found strong support for the monophyly of Dicranidae here. This is consistent with previous studies (e.g., La Farge et al., 2000; Tsubota et !! ""! al., 2003; Hedderson, 2004; Wahrmund et al., 2009; and see clade u in Fig. 4.3). The earliest phylogenetic splits within Dicranidae are represented here by Scouleria (Scouleriales) and Bryoxiphium (Bryoxiphiales). The relationships recovered here for these taxa (clades v and w; Fig. 4.1) are well suppported across most analyses (Table 4.1) and have been found with good support in only a few other studies (Fig. 4.3). The monophyly of the species-rich order Dicranales is strongly rejected here (Fig. 4.2), as Leucobryum (Dicranales) is strongly and consistently supported as the sister-group of Archidium (Archidales) (clade x; Table 4.1, Fig. 4.2) and two other members of Dicranales (Dicranum and Schistostega) form a grade subtending Syntrichia (Pottiales) with moderate to strong support from the different ML analyses (clade bb; Table 4.1, Fig. 4.2). Similar relationships have also been recovered in other studies (e.g., Qiu et al., 2006 and 2007; Wahrmund et al., 2009 and 2010, see also Fig. 4.3), supporting the need for a revised classification of Dicranales as a whole (La Farge et al., 2000; Hedderson et al., 2004). The monophyly of Bryidae (the diplolepideous-alternate group) has been well established by various studies (e.g., Goffinet et al., 2001; Qiu et al., 2006; see also Fig. 4.3). Splachnales (Tetraplodon), Bartramiales (Bartramia), and Hedwigiales (Hedwigia) represent the first splits in Bryidae, although the exact branching order among them is poorly supported (Fig. 4.2, Table 4.1) as it is in other studies (Fig. 4.3). The branching order of the remaining deep splits in Bryidae finds improved support here in some cases (e.g., clades ff, hh, ii; Fig. 4.3), and the core pleurocarpous mosses (as defined in Bell et al., 2007; clade kk here) are moderately well supported as monophyletic. Substantial ambiguities still remain regarding the relationships of the Bryidae taxa that form the grade subtending the core pleurocarps, as various branching orders have been recovered in previous studies, generally with poor support (e.g., Goffinet et al., 2001; !! "#! Bell et al., 2007; O’Brien, 2007). Rhizogoniales and the core pleurocarps have been referred to collectively as “pleurocarpids” (Bell at el., 2007), but the monophyly of Rhizogoniales (and therefore pleurocarpids) is weakly rejected here and elsewhere, as Rhizogonium (Rhizogoniales) is weakly supported as the sister group of Orthotrichium (Orthotrichales) (see clade gg in Figs. 4.2, 4.3). Robust inference of relationships among the “early-diverging” Bryidae lineages has proved to be difficult. In the best trees recovered here, there are short internodes defining the first splits, which in turn lead to long terminal lineages. Inference of these relationships may be especially susceptible to long-branch attraction. In addition, as there is only a limited number of synapomorphies that can be mapped onto those short internodes, sampling error may lead to the generally low observed branch support for these relationships (clades ee and dd). These relationships may become more clearly resolved by substantial expansion of taxon sampling (to break up the relatively long terminal branches) and addition of more data. Including additional genes with higher substitution rates may help, but this should be done cautiously, given the observed imbalance between short internal nodes and longer terminal branches (Fig. 4.2), and the possibility that this might contribute to long-branch attraction if faster genes are sampled. Improved taxon sampling may help particularly in the core pleurocarps, which represent about half of all moss species (Newton et al., 2007). 4.5.3 The earliest splits in moss phylogeny  The early splits in mosses are represented by the nonperistomate lineages. The branching order among these early-diverging taxa has long been of interest to bryologists. Two major unanswered questions concern relationships between Andreaeobryopsida and Andreaeopsida, !! "#! and between Takakiopsida and Sphagnopsida, in relation to other mosses. Here I sampled species representing every genus from the five early-diverging lineages, except for the extremely rare Ambuchanania (Sphagnopsida). Consistent with many other studies, Takakia (Takakiopsida) and Sphagnum (Sphagnopsida) are inferred here to represent the first split(s) among mosses. However, two conflicting placements of Takakia and Sphagnum are inferred by different ML analyses, as well as in various other studies (e.g., Yatsenyuk, 2001; Cox et al., 2004; Volkmar and Knoop, 2010; Chapter 3). The placements of Takakia and Sphagnum are heavily affected by different base frequency settings here (Table 4.1; see discussion below). In contrast, the placement of Andreaeobryum (Andreaeobryopsida) and Andreaea (Andreaeopsida) is recovered consistently among different ML analyses (Table 4.1): they form a clade that defines the next phylogenetic split in moss phylogeny. The sister-group relationship between Andreaea and Andreaeobryum receives consistently moderate to strong ML bootstrap support here (78% - 94%; Table 4.1). The long branch leading to each genus indicates an early split between these two taxa. The branch connecting Andreaea and Andreaeobryum to other mosses  (branch j; Fig. 4.1), on the other hand, is quite short. This combination of long terminal branches and a short internode may explain the general lack of support for this relationship in most other studies to date (Fig. 4.3). 4.5.4 Placement of Takakia and Sphagnum in moss phylogeny The backbone phylogenetic relationships inferred here for peristomate and nonperistomate mosses here are generally consistent among anlayses performed using different phylogenetic criteria, except for for the placements of Takakia and Sphagnum. I found that the base frequency settings can have striking effects on how the deepest split in moss phylogeny is !! "#! resolved (Table 4.1). Equilibrium base frequencies are usually estimated by optimizing them during inference of the best tree, or are calculated based on the observed base frequencies of terminal taxa. Here, application of ML-estimated base frequencies seems to outperform the application of empirical base frequencies, at least in terms of the overall consistency of results. When ML-estimated base frequencies are applied, the same arrangement (Takakia-sister) is recovered with moderate to strong support (>85%) using different data partitionings and partitions (Tables 4.1, 4.2). In contrast, strongly conflicting results are found in analyses of different data partitioning when empirical estimates of base frequencies are used. The GC content of the genes examined here varies greatly among different codon positions and the noncoding regions (Table 4.4), but it does not differ substantially among different major lineages or among the different regions sampled here (data not shown). The GC content of the first and second codon positions is about twice that observed for third codon position and the noncoding sites (Table 4.4). Base composition heterogeneity among different sites may be only one aspect of different tempos and modes of nucleotide substitution in these plastid regions. For example, the values of the among-site rate variation (!) and invariant site (I) parameters also vary substantially among data partitions (Table 4.4), as do site-site substitution rate parameters (data not shown). Applying one set of parameters to all sites (or even within major partitions) may over-simplify the pattern of DNA substitution sufficiently to misdirect subsequent phylogenetic inference, at least concerning the very first split in moss phylogeny. The observation that ML vs. empirical estimates of base frequencies can have an effect on anaysis here may be only one symptom of a more complex inference problem, as ML analyses that examine each of the major codon positions individually (i.e., first and second, vs. third and !! "#! noncoding combinations), recover the Takakia-sister arrangement with moderately strong to strong support, regardless of how base frequencies are estimated (Table 4.2: columns 4, 5, 6), and despite the very different model parameters inferred here (Table 4.4). ML inference may therefore be performing adequately on these distinct partitions alone, whereas a strong interaction among data partitions and their estimated parameter values may be responsible for strongly conflicting inferences made using the full data set, in some cases. Resolving the relationships among Takakia, Sphagnum and the remaining mosses is an important outstanding issue in moss phylogeny. I have demonstrated conflicting results here based on different MP and ML analyses, and earlier demonstrated that rate filtering based on ML rate classes is apparently not effective in ameliorating long-branch problems in MP analysis (Chapter 3). In this chapter I showed that ML analyses can be sensitive to parameter settings and data partition methods, in a way that can lead different phylogenetic programs to lead to strongly conflicting inferences. The results of the subsequent AU test, a common method to choose among competing hypotheses, are also conditional on different parameter settings. This suggests that extra caution should be taken when investigating and interpreting the Takakia-Sphagnum problem. However, my results suggest that Takakia is more likely to represent the deepest split in moss phylogeny, since this is the result that has the best ML value (when data are partitioned by codons, when subsets of these partitions are analyzed individually, or when ML-estimates of base frequencies are used). Nonetheless, it is clear that further investigation of this problem is needed. !! " # ! Table 4.1a. Bootstrap support for relationships among the four major land plant lineages and algal relatives. Labels in the first column correspond to individual branches (taxon bipartitions) in Fig. 4.1. ML analyses were performed using Garli and RAxML, with likelihood-based or empirical estimates of base frequencies. ML analysis abbreviations: Un-P, unpartitioned; P-codon: partitioned by codon position; P-region: partitioned by major region. The likelihood values of the optimal trees were calculated using different programs (i.e., Garli or RAxML) and so are not directly comparable (Sundberg et al., 2008). Bootstrap support less than 50% is represented by --.    ML MP  Garli, estimated base frequencies Garli, empirical base frequencies RAxML, empirical base frequencies  Un-P P-codon P-region Un-P P-codon P-region Un-P P-codon P-region                      Log likelihood of optimal tree 247052 239161 244660 248329 239534 246678 249033 242990 248127 n/a  Taxon bipartition (= clade on rooted tree) a [Zygnematales + land plants] -- 65 -- -- 63 -- 62 77 -- -- b Land plants 100 100 100 100 100 100 100 100 100 100 c Landplants excl. liverworts 100 100 100 100 100 100 100 100 100 94 d [Hornworts + vascular plants] 98 100 98 100 100 100 100 92 100 90 e Liverworts 100 100 100 99 100 100 100 100 100 50 f Mosses 100 100 100 99 100 99 100 100 100 100 g Hornworts 100 100 100 100 100 100 100 100 100 100 h Vascular plants 96 97 93 96 97 98 100 85 100 75 !! " # ! Table 4.1b. Bootstrap support for the multi-family clades of mosses. Labels in first column correspond to branches (taxon bipartitions) in Figs. 4.2 and 4.3 (except i2 and ee2). Clade composition comparisons are relative to Fig. 3 of Cox et al. (2004) for clades i, j-q, Fig. 3 of Wahrmund et al. (2010) for clades r-bb, and Fig. 2 of Bell et al. (2007) for clades cc-oo. MRCA = most recent common ancestor. ML analyses were performed using Garli and RAxML, with likelihood-based or empirical estimates base frequencies. ML analysis abbreviations: Un-P, unpartitioned; P-codon: partitioned by codon position; P-region: partitioned by region. The likelihood values of the optimal trees were calculated using different programs (i.e., Garli or RAxML) and so are not directly comparable (Sundberg et al., 2008). Bootstrap support less than 50% is represented by --.   ML MP  Garli, estimated base frequencies Garli, empirical base frequencies RAxMl, empirical base frequencies  Un-P P-codon P-region Un-P P-codon P-region Un-P P-codon P-region                     Log likelihood of optimal tree 247052 239161 244660 248329 239534 246678 249033 242990 248127  Taxon bipartition (= clade on rooted tree) i Mosses excl. Takakiopsida 100 100 99 -- 100 -- -- 100 -- -- i2 Takakiopsida + Sphagnopsida -- -- -- 70 -- 72 69 -- 97 100 j Andreaeopsida + Andreaeobryopsida 81 85 78 92 83 93 94 92 78 -- k Mosses excl. Takakiopsida and Sphagnopsida    Sphagnupsida AandanSSSSSSphagnopsida 100 100 100 100 100 100 100 100 100 -- l [Oedipodopsida + peristomate mosses] 100 100 100 100 100 100 100 100 100 97 m Peristomate mosses 100 100 100 100 100 100 100 100 100 68 n [Tetraphidopsida + Bryopsida] 56 72 -- 63 73 59 -- 85 68 -- o Bryopsida 100 100 100 100 100 100 100 100 100 99 p Bryopsida excl. Buxbaumiidae 100 100 100 100 100 100 100 100 100 100 q Bryopsida excl. Buxbaumiidae & Diphysciidae 100 100 100 100 100 100 100 100 100 100 r Funariidae 1 99  99  99 99 99 99 84 100 100 89 1 s [Timmiidae + Dicranidae + Bryidae] 99 98 98 99 98 98 100 100 100 76 t [Dicranidae + Bryidae] 94 89 96 93 90 95 86 100 92 89 u Dicranidae 2 100  100  100 100 100 100 100 100 100 100  v Core Dicranidae 3  + Bryoxiphiales 91 96 89 82 93 81 76 100 60 -- w Core Dicranidae 3  99 99 98 96 98 93 100 100 87 -- x [Leucobryum + Archidales] 100 100 100 100 100 100 100 100 100 100 y Grimmiales 100 100 100 100 100 100 100 100 100 100 z [Grimmiales + Pottiales + Dicranales s.s. 4 ] 63 74 66 -- 74 57 69 85 54 -- aa Pottiales + Dicranales s.s. 4 93 87 92 92 89 92 100 100 94 59 bb MRCA of Schistostega & Pottiales 79 76 84 89 76 88 92 92 78 71 cc Bryidae 100 100 100 100 100 100 100 100 100 100 dd [Bartramiales + Splachnales] -- -- -- -- -- -- -- -- -- -- ee Bryidae excl. Hedwigiales -- -- -- -- -- -- -- -- -- -- ee 2 [Bartramiales + Hedwigiales] -- -- -- 74 -- 73 76 62 86 98 ff [Bryales + Orthotrichales + pleurocarpids] 84 83 90 80 80 80 76 -- 73 67 gg [Orthotrichales + Rhizogonium] -- 53 -- -- -- -- -- -- -- 67 hh [Orthotrichales + pleurocarpids 5 ] 73 83 80 69 76 73 52 91 68 -- ii [Aulacomnium + core pleurocarps 6 ] 91 87 91 84 86 79 84 68 92 67 jj [Ptychomniales + Hypnodendrales] -- 55 -- -- -- -- -- -- -- 72 kk Core pleurocarps 6 80 85 87 68 85 68 72 69 92 --  ll [Hookeriales + Hypnales] 100 100 100 100 100 100 100 100 100 -- m m [Hypopterygiaceae + Hypnales] -- 70 51 59 73 57 60 77 54 -- nn Hypnales 100 100 95 100 100 100 100 100 100 90 oo MRCA of Brachytheciaceae and Hypnaceae 94 92 95 95 90 96 100 85 84 92 !! "#! Table 4.2. The placement of Takakia and Sphagnum in various ML analyses (bootstrap support in parentheses), as determined using Garli, considering either likelihood or empirical estimates of base frequencies. ‘T-sister’: Takakia-sister to other mosses (including Sphagnum); ‘T-S clade’: clade comprising Takakia and Sphagnum. Note that several bootstrap support values (columns 2-4) are also cited in Table 4.1. Matrix used in Garli Base frequency estimates in Garli All (unpartitioned) All (partitioned by major regions) All (partitioned by three codon positions and noncoding regions) First and second codon positions only Third codon position and noncoding regions ML estimates T-sister (100%) T-sister (99%) T-sister (100%) T-sister (86%) T-sister (100%) Empirical T-S clade (70%)  T-S clade (72%)  T-sister (100%)  T-sister (82%)  T-sister (100%)  !! "#!  Table 4.3. Approximately Unbiased (AU) tests of whether suboptimal arrangements of Takakia and Sphagnum are significantly worse than the optimal one. Tree searches (with or without constraint) were based on the unpartitioned matrix using Garli, with likelihood-estimates of all DNA substitution model parameters. Likelihood values used in AU tests perfirormed in CONSEL were calculated in PAUP* (tree scores not comparable with Garli estimates in Table 4.1). Likelihood estimates of base frequencies (PAUP*)  Empirical  estimates of base frequencies (PAUP*)  -lnL (PAUP*) AU test result -lnL (PAUP*) AU test result Takakia sister to other mosses 245901.643 Best tree 248346.914 P = 0.999 Takakia-Sphagnum 245930.119 P < 0.001 248340.113 Best tree Sphagnum sister to other mosses 245933.264 P < 0.001 248366.454 P < 0.001 !! "#!  Table 4.4. DNA substitution model parameters estimated for different data partitions. All parameter values are likelihood estimates, but the ! and I values were estimated given empirical estimates of base frequencies. Noncoding regions from green algae and vascular plants were not included in the current matrix. The base frequencies of non-coding regions are the average over bryophyte taxa sampled in this study.  Observed base frequencies ML estimate values Pi (C) Pi (G) Pi (G+C) ! I All sites 0.1624 0.2025 0.3649 0.7249 0.2844 1 st  codons 0.1715  0.3320  0.5025  0.6941 0.3365 2 nd  codons 0.2337  0.1970  0.4307  0.5578 0.3991 3 rd  codons 0.1022  0.1078  0.2100  2.0568 0.0338 Noncoding 0.1049  0.1178  0.2227  0.9554 0.1026 !! "#!  Table 4.5. Variation in lengths of noncoding regions in nonperistomate vs. peristomate mosses (lengths rounded to nearest bp).     Region Mean length Standard deviation Range  nonperistomate / peristomate nonperistomate / peristomate nonperistomate / peristomate psbE-psbF-psbL-psbJ      psbE-psbF 9/9 0/0 9/9-10      psbF-psbL 26/35 7/1 22-37/34-36      psbL-psbJ 112/117 4/2 108-115/113-122 psbB-psbT-psbN-psbH     psbB-psbT 135/155 3/23 132-140/100-210     psbT-psbN 108/83 8/10 94-114/64-104     psbN-psbH 84/94 2/9 83-88/88-120 rpl2 intron 621/618 6/22 616-628/566-666 3’rps12-rps7-ndhB*     rps12 intron 571/528 10/53 559-582/474-588     rps12-rps7 46 /44 14/15 25-53/20-55     rps7-ndhB 281/204 83/13 204-369/190-216     ndhB intron 765/641 58/22 722-855/614-671 * Data for peristomate mosses based on eight exemplar species (Alophosia azorica, Atrichum selwynii, Tetraphis pellucida, Funaria hygrometrica, Timmia austriaca, Dicranum scoparium, Orthotrichum lyellii and Rhytidiadelphus loreus). ! !! ""!  !!"!#$%&'%()(&()*+%,%)(- Zygnema Liverworts Mosses Hornworts Vascular Plants Chlamydomonas Chlorella Nephroselmis Mesostigma Chlorokybus Chaetosphaeridium Chara 100 100 100 100 100 100 98 -- -- 100 100 100 95 a c d b e f g h Fig. 4.1. The relationships of land plants and algal relatives, inferred from ML analysis of an unpartitioned matrix with ML-estimates of base frequencies, using Garli (-lnL = 245901.624). The numbers by the branches are bootstrap support values. Support values lower than 50% are represented by a dash (--). Species names of land plants, bootstrap support for relationships within each land-plant lineage have been removed, and branches condensed vertically for simplicity. Clade labels are those used in Table 4.1.  !! "#!   Fig. 4.2. Phylogram of moss phylogeny inferred from ML analysis of an unpartitioned data matrix with ML-estimates of base frequencies, using Garli (this phylogram corresponds to the portion labelled as mosses in Fig. 4.1; the stem lineage has been shortened). The classification follows Goffinet et al. (2009); orders are indicated for exemplar taxa in Bryidae, Dicranidae and Funariidae. Major peristome types: nematodontous – Tetraphidopsida and Polytrichopsida; arthrodontous – Bryopsida; diplolepideous-opposite – Funariidae; diplolepideous-alternate – Bryidae; haplolepideous – Dicranidae. Other major clades: core Dicranidae – clade w; pleurocarpids – clade hh (if it excluded Orthotrichum); core pleurocarps – clade kk (see Table 4.1 for clade definitions and support in other analyses). Thickened lines represent strong ML bootstrap support (>90%). ML Support values between 50% and 90% are labeled above or beside the branches; those with <50% support are indicated with a dash (--). !! "#! ! !"#"#$"%&'(")*+,-.." !"#"#$"%.'+$/*0$*$/'1 2+,"3456%1+7 2+,"3456%185"((*156 94/('"'*:(-56 94/('"'"%(5+'1)($1 94/('"'"%4$;".$1 <'/$+*/$56 9.*+,*1$" ="()("6$*+1$1 >*.-)($&,56 9)($&,56 !')("+,$1 !')(*/*4)$56 =5?:"56$" @$+,-1&$56 A4&".-+)"%B!"#$%&'($%)*C D54"($"%B+,"$-.$%)*C >,-1&*6$)('.."%B+,"$-.$%)*C !$66$" Scouleria (Scouleriales) Bryoxiphium (Bryoxiphiales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i k f j l m n o p q r s t u v w x y z aa bb cc dd ee ff gg hh ii jj kk ll mm nn oo 81 63 79 -- --84 73 80 -- -- 56 Bryopsida Tetraphidopsida Polytrichopsida Peristomate mosses Andreaeobryopsida Andreaeopsida Oedipodiopsida Sphagnopsida Takakiopsida / .# -$ " .2 $) :-&.2$) + , " $- ..2 $) /.'1&*#..2$) :,@A$,4..2$) B.44..2$) !! "#! Fig. 4.3. Support for interfamilial relationships among mosses from this study and selected studies. The cladogram topology is from the ML analysis of an unpartitioned data matrix with ML-estimates of base frequencies, using Garli (Fig. 4.1, 4.2; see Table 4.1 for clade definitions and support in other analyses here). In several studies, the MRCA of the exemplar taxa is less inclusive than the relevant reference phylogeny. The least inclusive clades are shown; more inclusive interpretations are indicated in parentheses; in several cases, my exemplar-based sampling is also less inclusive than the reference (indicated with an asterisk). References: 1 -- Cox et al., 2004 (ML); 2 -- Bell et al., 2007 (MP); 3 -- Qiu et al., 2007 (ML); 4 -- Stech and Frey, 2008 (MP); 5 -- Volkmar and Knoop, 2010 (ML); 6 -- Wahrmund et al., 2010 (ML).  !! "#! Strong support Weak support Moderate support BranchBar !"#"#$"%&'(")*+,-.." !"#"#$"%.'+$/*0$*$/'1 2+,"3456%1+7 2+,"3456%185"((*156 94/('"'*:(-56 94/('"'"%(5+'1)($1 94/('"'"%4$;".$1 <'/$+*/$56 9.*+,*1$" ="()("6$*+1$1 >*.-)($&,56 9)($&,56 !')("+,$1 !')(*/*4)$56 =5?:"56$" @$+,-1&$56 A4&".-+)" B54"($" >,-1&*6$)('.." !$66$" Scouleria Bryoxiphium C'5&*:(-56 9(&,$/$56 =.$4/$" D($66$" @$&("456 2-4)($&,$" 2&,$1)*1)'3" E'/F$3$" ="()("6$" !')("+.*/*4 =(-56 <(),*)($&,56 G,$0*3*4$56 95."&*64$56 >)-&,*64$*4 G"&*+$.56 E**#'($" E-+*+)'(-3$56 G,-)$/$"/'.+,51 E*6".*),'&$56 E-+456 12 5 f i k j 12 56 5 l m 1 12 56 n o p 1 12 564 12 564 1256 s q 12 r 14 56 u 145 6(2) t cc 123456 ee 35 6 ff 46 hh 2 46 ii 36 kk 3(126) ll 231 4 nn mm oo3 4632 2 dd jj gg 3 3 2 3 4 bb 6 aa 2 6 x y 2 4 2 46 w 56 z 6 v 2 (4) !! "#! 5 MOLECULAR PHYLOGENETICS OF ANDREAEACEAE 5.1 Brief synopsis The monogeneric family Andreaeaceae is one of the major nonperistomate moss lineages and represents a very early split in moss phylogeny. To examine its phylogeny and test the current infrageneric taxonomy (subgenera and sections), I sampled six plastid DNA markers from a total of 39 specimens (representing 22 different species), with an emphasis on the costate taxa. I identified multiple well-supported major clades within the genus, but none of the currently recognized infrageneric taxa was supported. Relationships among the clades were also generally strongly supported by maximum likelihood (ML) bootstrap analysis. The lack of monophyly (or substantial intraspecific molecular divergences) of several species indicates possible cryptic speciation for these species. I sub-sampled five species that represent the major lineages of the family for the 14-gene set employed in earlier chapters. ML analysis of this expanded gene-set confirmed the overall relationships in Andreaeaceae found with the six-gene data set, and the sister-group relationship of Andreaeaceae with Andreaeobryaceae. 5.2 Introduction Andreaea Hedw., the only genus of the granite moss family, Andreaeaceae (Bryophyta), comprises 50 ~ 75 species. Andreaea  is mainly distributed in high latitude and high altitude regions of both hemispheres, with its greatest diversity in cool-temperate regions (Schofield, 1985; Murray, 2006). Andreaea is commonly found growing on granitic or other acidic rocks. In contrast to most mosses, Andreaea lacks peristome teeth to aid in spore release; instead, the sporangium of Andreaea opens up along four to eight longitudinal lines that extend from the !! "#! sporangium base up to its apex. When dry, the dehisced mature sporangium resembles a Chinese lantern, which gives Andreaea its other name, the Chinese lantern moss. The sporophyte lacks a seta and the sporangium is instead held up by a pseudopodium derived from gametophytic tissue. Collectively, these unusual morphological features support the view that Andreaea represents an isolated lineage in moss phylogeny (Mishler and Churchill, 1984; Murray 1988a). Recent phylogenetic studies that considered evidence from morphological and various molecular markers support an early divergence of the lineage leading to Andreaea in the evolution of mosses, as one of several nonperistomate lineages that arose before the origin of the large peristomate clade of mosses (e.g., Mishler and Chruchill 1984; Cox et al. 2004). A proposed sister-group relationship between Andreaea and the monotypic family Andreaeobryaceae has been confirmed by various phylogenetic studies, although generally with weak support until the recent studies by Volkmar and Knoop (2010) and Chapters 3, 4. Despite advances in understanding its placement in moss phylogeny, the taxonomy of Andreaea needs further work. The most comprehensive recent treatments of the genus were performed by Schultze-Motel (1970) and Murray (1988b). The former treated only the costate species, which comprise approximately one third of species in the genus. The latter provided a detailed treatment of the Andreaea species of the British Isles, which has a similar species composition to North America (~11-12 species). The greatest diversity of Andreaea is found in the Southern Hemisphere (Murray, 1986), where it is understudied, except for the Australian taxa. Andreaea has generally been divided into two to three subgroups that have been treated at various ranks (e.g., Braithwaite, 1880-1887; Sainsbury, 1955; Schultze-Motel, 1970). The taxonomy used here largely follows Murray (1988a), with three subgenera recognized: subgenus !! "#! Acroschisma Hook. f. & Wilson (1 species), subgenus Chasmocalyx (Braithw.) Limpr. (~ 5 species), and subgenus Andreaea (~ 45 species). The only member of subgenus Acroschisma, A. wilsonii Hook. f., stands out from other Andreaea species. Due to its unique sporangium morphology, some authors treated A. wilsonii as its own genus (e.g., Ochyra et al., 2003). Subgenera Andreaea and Chasmocalyx are each defined based mainly on characteristics of their perichaetial leaves. Chasmocalyx has costate perichaetial leaves that are similar in shape to vegetative leaves and are generally not convoluted; in contrast, the perichaetial leaves of subgenus Andreaea are strongly differentiated from vegetative leaves and are convoluted. Subgenus Andreaea is further divided into two sections: sections Nerviae Card. ex Broth. (9-10 species) and Andreaea (~ 35 or more species), based mainly on the presence vs. absence of costa in the vegetative leaf, respectively. The ecostate section Andreaea was believed to be a heterogeneous group comprising several independent evolutionary lines, and requires further taxonomic work (Murray, 1988b). Phylogenetic relationships among different Andreaea species have also not been assessed, except for an unpublished phylogenetic study based on two plastid regions (Hedderson, pers. comm.). A robust phylogeny would be useful for future taxonomic revision of the genus and permit reconstructions of ancestral character states, dates and biogeography. In this study, I attempted to reconstruct an overall framework for Andreaea phylogeny using multiple plastid markers. The major goals of this study are to 1) infer phylogenetic relationships using an exemplar-based sampling and; 2) test the monophyly of intra-generic taxa. The taxonomy of the costate species has been better studied; there is no worldwide treatment of the ecostate species. I focused my sampling efforts on the costate taxa, but sampled a broad selection of ecostate !! "#! species for which material was available, choosing ones that are relatively well defined morphologically. 5.3 Materials and methods I used two different sets of plastid genes to examine phylogenetic relationships within Andreaea. For the first of these, I recovered six plastid regions with a variety of rates to assess deep and recent phylogenetic splits: atpB, atpB-rbcL, ndhF, rbcL, rpoB and atpF-atpH. The rpoB locus and the atpF-atpH intergenic spacer region have been used or proposed for DNA barcoding studies (e.g., Chase et al. 2007; Wang et al., 2010), the atpB-rbcL intergenic region has been widely used in bryophyte phylogenetic studies (Stech and Quant, 2010), and the atpB, ndhF, and rbcL genes have been used extensively in phylogenetic studies of various plant groups, including those involving higher-order relationships (e.g., Tsubota et al., 2003; Saarela et al., 2007; Smith et al., 2010; Chapters 3, 4). I sampled a total of 39 specimens for these markers from four species in subgenus Chasmocalyx, nine in section Nerviae, seven in section Andreaea, one in section Acroschisma, and one species (A. fuegiana [Cardot] S. W. Greene) with an uncertain taxonomic placement (Table 5.1, Appendix 4). Only two costate species were not sampled here: A. depressinervis Cardot. (with uncertain subgeneric placement) and A. pachyphylla (Müll. Hal.) Broth. (subgenus Chasmocalyx). No recent material is available for either species, as they have very restricted distribution and have been rarely collected in recent years. I included multiple samples from different geographical ranges for most species, where material was available, especially for the costate species (among the 13 sampled costate species, the majority of species were sampled with two or more accessions from different populations, except for A. schofieldiana, A. crassinervia and A. megistospora; Table 5.1). I used !! "#! Andreaeobryum macroporum Steere & B. M. Murray (Andreaeobryaceae) as an outgroup for this gene set. Relatively long branches connect Andreaea and its putative sister group, Andreaeobryum (Volkmar and Knoop 2010; Chapters 3, 4). Rooting of major clades may be problematic when outgroups are distantly related to the ingroup (Graham et al., 2002), and this appears to be the case here, as the deepest split in Andreaea is only moderately well supported by bootstrap analysis of the 22-species, 6-region data set (see Results). Long-branch problems associated with distant outgroups may potentially be ameliorated by adding additional taxa to break up long outgroup branches (e.g., Graham et al., 2002; Graham and Iles, 2009). I therefore sampled multiple outgroups and added additional slowly evolving plastid markers to confirm overall root placement in Andreaea and relationships among the major subclades of Andreaea. I focused on five exemplar species of Andreaea that represent these lineages according to the 22-species data set mentioned above, and sampled and aligned 14 plastid regions (atpB, rbcL, and ndhF and 11 additional plastid genes: psbB, psbC, psbD, psbE, psbF, psbH, psbJ, psbL, psbN, psbT, and rpl2, and associated noncoding regions, Chapter 2) to an existing matrix consisting of sequences from other major moss groups and relatives (the 14-gene matrix used in Chapter 4); these effectively served as additional outgroups. I was also interested in assessing whether including multiple Andreaea species would have a substantial effect on the strength of the sister-group relationship between Andreaea and Andreaeobryum, as only one Andreaea species has been included in most previous studies (e.g., Cox et al., 2004; Volkmar and Knoop, 2010; Chapters 3, 4), and this relationship was often only moderately well supported for the gene set considered here (Chapter 3, 4). !! "#! Protocols for DNA extraction, amplification and sequencing of targeted regions and for base calling and sequence alignment followed those described in Chapters 2 and 3. Primers used to amplify and sequence the targeted regions were previously published (Table 5.2). I used jModelTest 0.1.1 (Guindon and Gascuel, 2003; Posada, 2008) to choose a DNA substitution model for each individual region using the standard Akaike Information Criterion (AIC; Akaike, 1974), the AIC corrected for small samples (AICc), and the Bayesian Information Criterion (BIC; Schwarz, 1978). When different models were selected according to different criteria, I chose a model that fell within the 95% confidence interval for all criteria. I carried out both maximum parsimony (MP) and maximum likelihood (ML) analyses. I used PAUP* 4.0 b10 (Swofford, 2002) for MP analyses, with tree-bisection-reconnection (TBR) branch-swapping for 100 random addition replicates, and otherwise with default settings. I used Garli 1.0 and Garli-Part-0.97 (Zwickl, 2006) for unpartitioned and partitioned ML analyses, respectively. In ML analyses, I ran ten independent replicates to search for the best tree, with the model parameters estimated. To assess branch support in MP and ML analyses, I performed non- parametric bootstrapping (Felsenstein, 1985) with 500 bootstrap replicates, using simple taxon addition for MP, but otherwise using the search conditions described above. For the six-region sampling, I first did maximum likelihood analysis using Garli 1.0 (Zwickl, 2006) on individual regions. The results based on different individual regions showed no strong conflicts, and so I concatenated all regions into one matrix. I then performed MP and unpartitioned and partitioned ML analyses. For the 14-gene matrix, I performed unpartitioned ML analysis on the combined dataset, with the associated noncoding regions included or excluded. I used GTR + ! + I as the DNA substitution model for both analyses. Throughout the !! ""! following discussion, I consider bootstrap support less than 70% as weak, between 70% and 90% as moderate, and higher than 90% as strong (e.g., Graham et al., 1998). 5.4 Results In this section, I first report results for the six-region dataset. I recovered the four coding regions (atpB, ndhF, rbcL, and rpoB; see Table 5.3) for most specimens. Due to difficulties with amplification, data were missing from the two intergenic spacer regions, atpB-rbcL and especially atpF-atpH, for some specimens (Table 5.3). However, the majority of specimens had all six regions sequenced or only one region missing. Only one specimen (one of the Andreaea microvaginata accessions) had three missing regions (rbcL, atpB-rbcL and atpF-atpH spacers). I found no strong conflicts among regions for ML analysis of individual regions. I therefore performed ML (unpartitioned and partitioned) and MP analyses on a combined matrix for all six regions. The combined six-locus matrix has 5,551 aligned sites of which 364 sites are parsimony-informative (Table 5.3). The unpartitioned and partitioned best ML trees are identical topologically and have comparable bootstrap support. ML trees inferred from individual regions are generally consistent with the trees based on the combined analyses but with poorer support (Table 5.3). Best MP trees (not shown) are very similar topologically to the best ML trees, with no strong or moderate conflicts. In the following discussion, I mainly focus on the results from the partitioned ML analysis and the MP analysis based on the combined dataset. Twenty-six of 35 clades recovered in the optimal ML tree based on the partitioned six- region matrix receive moderate or strong support in the partitioned ML and MP analyses of the combined dataset (Fig. 5.1 and Table 5.3). Twenty-two of 35 clades have at least 90% bootstrap support. Seven of 35 clades are weakly supported in both analyses. One clade (clade G in Fig. !! "#! 5.1) is strongly supported by MP analysis but only weakly supported by the partitioned ML analyses. The two subgenera (Andreaea, Chasmocalyx) and the two sections of subgenus Andreaea (Andreaea and Nerviae) are not recovered as monophyletic in the best ML tree. However, four notable major clades within the genus are defined by major portions of these taxa (clades A-D in Fig. 5.1). These clades all have strong support (93%-100% bootstrap support in partitioned ML and MP analyses). A sister-group relationship between clades C and D is strongly supported (clade F in Fig. 5.1; 100% bootstrap support in partitioned ML and MP analyses). However, because the sister-group relationship between clades A and B receives only moderate support (clade E in Fig. 5.1; 77% and 78% bootstrap support in partitioned ML and MP analyses), the root of Andreaea as a whole is only moderately well supported. Clades A and B are each composed of only one or two species. Andreaea blyttii (section Nerviae) is the only member of clade A. Andreaea nivalis and A. rigida, both from subgenus Chasmocalyx, comprise clade B. All six species resolved in clade C are from section Nerviae. There are three distinct lineages within clade C: A.heinemannii alone, a clade comprising A. schofieldiana and A. frigida, and a clade comprising A. rothii, A. crassinervia and A. megistospora (Fig. 5.1), which I refer to here as the rothii complex. Relationships among these three subclades of clade C are well supported: the schofieldiana-frigida clade is the sister group of the rothii complex. Clade D is heterogeneous, as it includes species from all four sub-generic groups. Andreaea wilsonii (section Acroschisma) is weakly supported as the sister group of the rest of clade D, with a small clade of two Chasmocalyx species, A. nitida and A. australis, the next successive sister group. The seven species sampled from section Andreaea do not form a clade, as a small clade of two section Nerviae species, A. subulata and A. microvaginata, is !! "#! nested in a grade of species in section Andreaea (including A. fuegiana). The relationships among these small clades within clade D are mostly weakly supported. The monophyly of most of the 12 species with multiple accessions is strongly supported at the current taxon population samplings (i.e., Andreaea alpina, A. australis, A. blyttii, A. heinemannii, A. microvaginata, A. rigida, A. subulata, A. frigida; >98% ML and MP bootstrap support). However, A. nivalis and A. rupestris are found to be paraphyletic with moderate support. The monophyly of the remaining species and subspecies that are sampled for multiple accessions is only weakly supported (A. rothii spp. falcata) or rejected (A. nitida, A. rothii and A. rothii spp. rothii). The AU test cannot reject the monophyly of any of the species (or subspecies) found to be paraphyletic on the best ML topology (Table 5.4). The best ML trees for the 14-gene matrix (with noncoding included or excluded) have an identical topology and comparable bootstrap support (Fig. 5.2). A clade of Andreaeobryum and Andreaea is recovered with moderate support (80% and 97% with and without the noncoding regions, respectively; Fig. 5.2), and is the sister group of a clade composed of Oediopodium and the peristomate mosses (100% bootstrap support; Fig. 5.2; see Chapters 3, 4). The monophyly of Andreaea is strongly supported (100% bootstrap support; Fig. 5.2). The root of Andreaea is consistent with the six-region analysis, but is strongly supported for this gene sampling (Fig. 5.2, branches E and F). 5.5 Discussion Apart from an unpublished study by T. Hedderson (University of Cape Town, pers. comm.), this study provides the first formal phylogeny analysis of Andreaea, a highly distinctive nonperistomate moss genus, and the sole extant member of Andreaeaceae. Its monophyly is !! "#! strongly supported here for the two different gene and taxon samplings that I considered (Figs. 5.1, 5.2); more broadly, a large multigene analyses confirms the placement of Andreaea among the early diverging nonperistomate lineages mosses, and provides moderately strong to strong support for its sister-group relationship with Andreaeobryum (Andreaeobryaceae), consistent with previous studies (e.g., Cox et al., 2004; Chapters 3, 4). The deep phylogenetic split observed here between Andreaea and Andreaeobryum (Fig. 5.2) is consistent with a very early divergence between these lineages, although a formal estimate of the timing of this split remains to be done. Relationships among four major clades of Andreaea inferred here (i.e., clades A-D; Fig. 5.1) are mostly well-supported in the 6-region analyses (Fig. 5.1); the 14-gene analysis more firmly places the root node of Andreaea recovered by the 6-region analyses (cf. Figs. 5.1, 5.2). For the 6-region analysis I sampled a broad selection of ecostate species (section Andreaea in subgenus Andreaea) and included nearly all of the costate species (i.e., species in subgenus Chasmocalyx, and section Nerviae of subgenus Andreaea). While the ecostate taxa need to be sampled more heavily in the future, it is clear that none of the currently recognized subgeneric taxa are monophyletic as currently circumscribed. Subgenus Chasmocalyx and section Nerviae are each divided into two distantly related clades that are separated by ecostate lineages, and the ecostate section Andreaea is part of a large, well-supported clade (clade D) that includes portions (subclades) of Chasmocalyx and Nerviae (Fig. 5.1). Assuming that the major phylogenetic results uncovered here using plastid data are an accurate reflection of species relationships, my study indicates the need for a substantial re-classification of the genus above the species level, in addition to a re-consideration of the homology of the characters that have traditionally been used to define current taxa at this level. I address the significance of the broad !! "#! phylogenetic backbone of Andreaea recovered here below, but defer formal changes to the classification of evolutionary reconstructions until a broader sampling of ecostate species can be included and the current results corroborated with data from additional genetic linkage groups (e.g., multiple nuclear genes). 5.5.1 Major lineages of Andreaea I recover four major well-supported clades in this study (clades A-D; Fig. 5.1). The sole member of clade A, Andreaea blyttii, is currently included in section Nerviae. The isolated placement of A. blyttii from other Nerviae species here is consistent with earlier morphological observations that A. blyttii is distinct from other members of section Nerviae (Murray 1988b). The two members of clade B, A. nivalis and A. rigida (currently in subgenus Chasmocalyx), are also morphologically distinctive within Andreaea, having lanceolate leaves that are denticulate and papillose, and strong and well-defined costa that are percurrent or excurrent (Schultze-Motel, 1970). Three distinctive lineages can be found in clade C, which comprises the majority of the species in section Nerviae. Andreaea heinemannii stands apart from other Nerviae because of its small size, its sporangium morphology, and its growth as a short turf on its usual habitats of basalt rocks (Murray, 2006). The European endemic A. frigida is sister to the North American endemic A. schofieldiana. Both species have lanceolate or gradually tapering leaves with strong costae (Murray 1987a). The three species within the rothii-complex are characterized by strongly costate leaves that have a more or less abrupt contraction from the leaf base to upper the part of the leaf (Murray 1987a). !! "#! Clade D is a large and heterogeneous group that includes species from all three sub- generic units, and includes all of the ecostate species sampled here. I can identify no obvious morphological synapomorphies for this group. However, all taxa in this clade are defined by a unique 8-base indel (inferred to be a deletion) in the atpF-atpH region. Based on our current understanding of Andreaea morphology, I predict that most, if not all, of the unsampled ecostate taxa belong to this clade. 5.5.2 Species monophyly and species delimitation Molecular phylogenetic evidence is frequently used to aid in species circumscriptions (e.g., history-based phylogenetic species concepts, e.g., Baum and Donoghue, 1995), supplementing morphological definitions of species and in some cases revealing substantial non- monophyly of species in mosses (e.g., Vanderpoorten and Goffinet, 2006). High genetic variation (or lack of monophyly) accompanied by limited morphological differentiation can be indicative of cryptic speciation (Shaw et al., 2003), which appears to be common for bryophytes, especially those with wide distribution ranges (e.g., Fernandez et al., 2006). However, Vanderpoorten and Shaw (2010) caution that multiple linkage groups need to be considered when using molecular data to assess species boundaries, and remind us that species may often be paraphyletic, at least initially (consistent with the plesiospecies concept of Olmstead, 1995). Nonetheless, the current phylogenetic data provide initial molecular information on the naturalness of individual species that can then be followed up with more intensive molecular and morphological study, and additional population sampling. Twelve species sampled here included multiple accessions; when I sampled multiple populations per species these were generally well separated geographically (see Fig. 5.1). At the current sampling intensity, eight of these were !! "#! recovered as monophyletic with strong support (>97% ML and MP bootstrap support; Fig. 5.1). Among these, Andreaea alpina and A. heinemanii have a broad distribution and occur in both hemispheres. Andreaea alpina is found widely in the Southern Hemisphere and in northwestern Europe and Greenland, while A. heinemannii has a disjunct distribution in the Northern Hemisphere, and has recently been recognized in Australia (Murray, 2006), a sample of which was included here. The genetic variation within each of these species, however, is quite different. The distance between the two A. alpina accessions is similar to the average intra-species genetic distance of the six other monophyletic species that have limited distributions (see branch lengths in Fig. 5.1). The average intra-species genetic distance of A. heinemannii, in contrast, is about five times higher than that of A. alpina, which suggests that cryptic speciation may have occurred. The remaining four species with multiple accessions are recovered as monophyletic with weak support, or paraphyletic with moderate support (none are polyphyletic at current samplings). However, the AU test cannot reject the monophyly of any of the four species or the two subspecies of A. rothii (Table 5.4); constraining the monophyly of each species does not result in trees that are significantly longer than the best ML tree. Sampling of additional populations from across the range of all of the species examined here will be required to further assess their monophyly (or lack). Andreaea nitida and A. rupestris would be good targets for initial study as both species are morphologically variable (Murray, 1988b and pers. comm). 5.5.3 Implications of Andreaea phylogeny for morphological evolution I did not perform formal ancestral-state reconstructions here due to the relatively thin sampling of ecostate taxa in this study. Nonetheless, Andreaea phylogeny shows character- distribution patterns that are worth attention in light of the current phylogeny, including the !! "#! evolution of the sexual system. Our knowledge of sexuality in Andreaea is incomplete, but based on our current understanding, the majority of Andreaea appear to be monoicous (i.e., with sperm and eggs produced on the same gametophyte). The phylogenetic root of Andreaea defined here (Figs. 5.1, 5.2) divides the genus into a dioicous clade (clades A+B) and a primarily monoicous clade (clades C+D): the only known dioicous species are A. blytii, A. nivalis, and probably A. australis (Murray, 1988b and per comm.). As the sister group of Andreaea (Andreaeobryum) is dioicous (as is Takakia and some members of Sphagnum), I predict that dioicy is the ancestral sexual system in the genus, and that monoicy is a derived condition defining clade C and clade D (which contain the majority of Andreaea species). All members of clades A, B, and C have leaves with costae. Clade D is more heterogeneous, as it includes both costate and ecostate taxa. Andreaeobryum is costate, and so I predict that leaves with costae were the ancestral condition in Andreaea as a whole, and that there were subsequently multiple transitions (gains and/or losses) within clade D. The presence or absence of a costa is clearly not a reliable marker considering current subgeneric taxonomy. It is not uncommon to observe that costa development and morphology vary within individual moss genera (e.g., Allen, 1987; Vanderpoorten, 2002), but few genera are comparable to Andreaea in terms of the diversity of morphology of their costa and costa-like structures. Within the genus, the costa may be prominent throughout the leaf (e.g., in A. nivalis), restricted to the lower part of the leaf (e.g., in A. nitida), or absent towards the leaf base (e.g., in A. microvaginata). In addition, the weak or weaker part of the costa can be two or three layers of cells, with little cell differentiation (Murray, 1988b; Frey and Stech, 2010). In some taxa considered to be ecostate, bi- or multi-stratose stripes can be found near the leaf base or along the !! "#! leaf apex that resemble a highly reduced costa (e.g., A. wilsonii) (pers. obs.). An extreme example is the multistratose stripes found along the leaf margin of A. fuegiana, which receives controversial interpretations from different authors (e.g., Schultz-Motel, 1970; Ochyra et al., 2003). The diversity of costa form clearly deserves more attention in a phylogenetic context. Another intriguing pattern is the suite of characters that are shared by members of Chasmocalyx. Besides having large but undifferentiated and non-convolute perichaetial leaves, all species in this subgenus have a pseudopodium that develops from both archegonial and stem tissues. They also have auriculate leaf bases and stems with very small outer cells and a sharp transition to the larger inner cortex cells (Murray, 1988b). However, the four Chasmocalyx species sampled in this study are divided between two distantly related clades. This suggests that this suite of characters co-evolved in these two widely isolated lineages, although there is no obvious structural or developmental linkage among these characters. 5.5.4 The placement of Andreaea wilsonii and A. fuegiana Among all the Andreaea species sampled in this study, A. wilsonii and A. fuegiana are perhaps the most distinctive in terms of their morphology. Andreaea fuegiana has two bi- or multi-stratose stripes near the leaf margin, starting from the base of the leaf and extending all the way to the apex. Whether these stripes are a form of costa remains debatable (e.g., Schultze- Motel 1970; Murray, 1988b; Ochyra et al. 2003). To emphasize the unique leaf morphology, Ochyra et al. (2003) proposed a new genus, Bicosta, to accommodate this species. Andreaea wilsonii stands out from other Andreaea species due to its unique capsule morphology. The sporangium is long-cylindric, similar to the typical bryopsid sporangium, and opens by eight divisions that are confined only to the apex of the sporangium. Based on these features, A. !! "#! wilsonii has been treated as its own section, subgenus or even as a separate genus (Acroschisma) by some authors (e.g., Wijk et al., 1959; Frey and Stech, 2009; Goffinet et al., 2009). However, my phylogenetic results recover both species as deeply nested within the genus Andreaea. Andreaea fuegiana is well resolved as a member of clade D. Its closest relatives, A. remotifolia and A. alpina (Fig. 5.1) are typical ecostate taxa, and so its “bi-costa”-like stripes likely have an independent origin from other costa-like structures in the genus. The closest relatives of A. wilsonii remain uncertain. However, this species is clearly part of clade D, although it is also a relatively isolated lineage. This suggests that the unique capsule found in this species is a modification of the typical valved capsule of Andreaea. However, the relatively long branch leading to A. wilsonii (Fig. 5.1) may reflect in part an early divergence within clade D, and may be consistent with its relatively drastic change in capsule morphology. 5.6 Conclusions and future directions A multigene analysis based on larger set of plastid markers confirms the early divergence of the lineage leading to Andreaea in moss evolution, and places it as the sister group of the monotypic Andreaeobryum with moderately strong support, consistent with previous studies (e.g., Cox et al., 2004; Volkmar and Knoop, 2010; Chapters 3, 4). It also confirms the monophyly of the genus Andreaea and supports relationships among and within four major subgeneric clades, permitting inference of the placement of the root node with strong support. I further investigated phylogenetic relationships within the genus using a set of six plastid markers. The phylogenetic relationships inferred from plastid data do not support the current subgeneric taxonomy, and hint at the paraphyly of several species, possibly indicative of cryptic speciation in some cases. Future elaborations of the phylogenetic backbone inferred here should !! "#! focus on sampling additional ecostate species and on adding data from the nuclear genome to corroborate or refute it. This will be essential for further improving our understanding of the evolution of this fascinating lineage of mosses. !! ""!  Table 5.1. The taxonomy of the Andreaea species sampled in this study and the number of specimens included for each species. *: the taxonomic treatment of A. fuegiana and A. wilsonii is controversial (e.g., Vitt, 1984; Ochyra, 2003; Goffinet et al., 2009;).  Subgenus Section Species # of specimens A. australis Mitt. 2 A. nitida  Hook. f. & Wilson 3 A. nivalis Hook.  2 Chasmocalyx costate; ~5 species   A. rigida Wilson 2 A. blyttii Schimp. 3 A. crassinervia Bruch 1 A. frigida Hüb. 2 A. heinemannii Hampe & Müll. Hal. 3 A. megistospora B. M. Murray 1 A. microvaginata  Müll. Hal. 2 A. rothii Web. & Mohr. 4 A. schofieldiana B. M. Murray 1 Nerviae costate; ~9 species A. subulata Harv. 2 A. acutifolia Hook.f. & Wilson 1 A. acuminata Mitt. 1 A. alpestris (Thed.) Schimp. 1 A. alpina Hedw. 2 A. mutabilis Hook. f. & Wilson 1 A. remotifolia Dusèn 1 Andreaea Andreaea ecostate; ~35 species A. rupestris Hedw.  2 Acroschisma ecostate; 1 species  A. wilsonii* Hook. f. 1   A. fuegiana* (Cardot) S. W. Greene 1 !! " # # ! Table 5.2. Primers used to amplify and sequence the rpoB gene, and the intergenic spacers (IGS) between atpB and rbcL and between atpF and atpH. The 14-gene regions were amplified with the set of primers cited in Chapter 2 (Figs. 2.1 – 2.7). Citations for the primers: atpB-1 and rbcL-1 -- Chiang et al. (1998); atpF, atpH, rpoBajfF1, and rpoBajfR1 -- Fazekas et al. (2008); B92R -- Chapter 2.  Forward primers Reverse primers rpoB rpoBajfF1: TCTAATATGCAICGTCAAGC rpoBajfR1: GAGGIGTTAITTIACCTAC IGS atpB-rbcL atpB-1: CCAACCGGAACACTTAARGGAGC rbcL-1: GGTTGAGGAGTCATTCGAAATG  B92R: TCCACYACTTTAATTCCTGTTTC IGS atpF-atpH atpF: ACTCGCACACACTCCCTTTCC atpH: GCTTTTATGGAAGCTTTAACAAT !! " # " ! Table 5.3. Statistics for the various MP and ML analyses conducted on the six-region dataset.   Regions Combined atpB ndhF rbcL rpoB atpB- rbcL atpF-atpH Number of accessions included  a  40 40 40 39 39 35 21 Aligned length    5551 1479 985 1369 562 596 559 Number of variable characters  622 124 133 114 74 39 138 Number of parsimony informative characters  364 65 79 67 55 24 74 Number of nodes with BS > 80% in MP analyses b 25  c na na na na na na DNA substitution model used in ML analysis d GTR+!+I Partitioned e TIM2+ ! +I TPM3uf + ! +I GTR+ ! +I TPM1uf+ I TPM1uf+ I TPM3uf Likelihood score (best tree) -12818.86 -12584.12 -3065.51 -2467.67 -2915.51 -1367.78 -1004.26 -1503.16 Number of nodes with BS > 80% in ML analysis b    24 c     23 c  11 10 8 5 5 4  Number of nodes with BS > 70% in ML analysis b  27 27 12 12 10 13 5 5  a  39 samples from Andreaea and one from Andreaeobryum. b  BS = bootstrap support.! c 22, 21 and 22 nodes with BS > 90% in the MP analysis, partitioned and unpartitioned ML analyses, respectively d See Posada (2008) for specifics of DNA substitution models. e In the partitioned ML analysis, individual regions were assigned the individual DNA substitution model shown in the rest of the table.  ! !! "#$! Table 5.4. Approximately Unbiased (AU) test of whether suboptimal arrangements constraining monophyly of several species are significantly worse than the paraphyletic arrangements in the optimal tree. !L = increase in –lnL compared to the best ML tree. The analyses consider the combined six-region dataset.  All the analyses were based on the partitioned matrix. Arrangement of interest !lnL P value* Monophyly of A. nivalis 4.27 0.08 Monophyly of A. rupestris 6.95 0.21 Monophyly of A. rothii 4.30 0.20 Monophyly of A. rothii spp. rothii 8.86 0.08 Monophyly of A. nitida 0.00 0.85 Notes:!* The monophyly is rejected when P < 0.05. ! !! "#$!   Fig. 5.1. Phylogeny of Andreaea inferred from the partitioned 6-region plastid matrix using maximum likelihood. The outgroup (Andreaeobryum macrosporum) is removed for clarity, with the stem lineage shortened.  Numbers beside branches are bootstrap support values from partitioned ML analysis and MP analysis. The bootstrap support values from unpartitioned ML analysis (not shown) are largely comparable to those from the partitioned ML analysis. “--” represents support less than 50%. Labels by the dashed lines indicated current taxonomy: subgenus Chasmocalyx, sections Nerviae and Andreaea (in subgenus Andreaea). !! "#$! ! blyttii NAm blyttii NAm blyttii EU nivalis EU nivalis NAm rigida AS rigida AS heinemannii NAf heinemannii AU heinemannii NAm schofieldiana NAm frigida EU frigida EU crassinervia EU megistospora NAm rothii rothii NAm rothii rothii EU rothii falcata EU rothii falcata EU wilsonii   SAmnitida SAm nitida SAm nitida AU australis AU australis AU alpina EU alpina SAm remotifolia SAm *fuegiana SAm rupestris NAm alpestris NAm rupestris NAm mutabilis AS acutifolia AU acuminata AU subulata SAf subulata SAm microvaginata AU microvaginata AU0.001 substitutions/site rothii- complex clade A c la d e  D c la d e  C clade B E 100/100 100/99 100/100 77/78 60/72 93/100 83/87 99/99 92/94 58/58 54/52 55/-- 77/66 97/96 99/94 99/100 94/95 91/90 75/70 99/97 77/84 98/98 95/92 --/-- 52/-- 65/98 --/54 Nerviae (costate) Chasmocalyx     (costate) Nerviae Andreaea Chasmocalyx Nerviae 100/100 100/100 100/100 100/100 100/100 100/100 100/100 99/99 F G (costate) (costate) (costate) (ecostate) (ecostate) Sexuality: Dioicous Monoicous Unknown Geographical locality: AS: Asia AU: Australia EU: Europe NAm: North America NAf: North Africa SAm: South America SAf: South Africa !! "#$! Fig. 5.2. Phylogenetic placement of Andreaeaceae in moss evolution and the relationships of the major Andreaea clades based on ML analysis of the 14-gene matrix. Taxa from other plant groups (including algal outgroups) are removed for clarity, with the stem lineage shortened. Species names of peristomate mosses are removed and branches condensed vertically for simplicity. Only the branches representing the nonperistomate mosses are shown with bootstrap support: numbers beside branches are ML bootstrap values with and without the noncoding regions included, respectively. Clade labels correspond to those in Fig. 5.1.  !! "#$! ! Takakia ceratophylla Takakia lepidozioides Sphagnum sp. Sphagnum squarrosum Andreaeobryum macrosporum Andreaea blyttii Andreaea nivalis Andreaea rothii Andreaea rupestris Andreaea wilsonii Oedipodium griffithianum 98/99 100/100 100/100 100/100 80/97 99/98 99/96 100/100 100/100 0.05 substitutions/site p e ri s to m a te  m o s s e s (Clade A) (Clade D) (Clade C) (Clade B) F E 100/100 100/100 100/100 !! "#$! 6 CONCLUSIONS AND FUTURE DIRECTIONS 6.1 Major conclusions A major goal of this thesis was to improve our understanding of the overall phylogenetic backbone of mosses in the context of early land-plant phylogeny. Chapter 2 focused on the development of tools (primers) for retrieving plastid genes from a broad range of bryophyte taxa, yielding a large set of loci suitable for higher-order phylogenetic inference. Subsequent chapters focused on different aspects of phylogenetic inference using these genes, including a demonstration of their utility for inferring very broad-scale relationships among the major embryophyte and moss lineages (Chapter 3), and an expanded taxon sampling focusing on the peristomate moss clade that includes most extant species (Chapter 4). Finally, I developed a well-supported phylogenetic backbone of one of the nonperistomate moss lineages, the granite- moss family, Andreaeaceae, using different sets of plastid genes and taxon samplings (Chapter 5). These phylogenetic backbones should provide frameworks for developing more natural classification schemes and to help shed light on one of the most intriguing sets of events during the evolution of life on earth, the origin and diversification of the early land plants. However, future inferences made about their morphological and genomic evolution in a phylogenetic framework will depend in part on the accuracy of tree inference. It is therefore important that the trees are as well corroborated and as strongly supported as possible. The inferences made in Chapters 3 and 4 largely corroborate those recovered in other studies using different taxon and gene/genomic samplings, and generally provide well-supported tree structure, in several cases with improved support here compared to other studies. The phylogenetic backbone of !! "#$! Andreaeaceae, the subject of Chapter 5, has not been studied in detail before. These backbones should be useful for additional evolutionary and systematic research on the bryophyte branches of the tree of life. The study of moss phylogeny has focused on extant taxa because of the paucity of records for fossil bryophytes (Shaw and Renzaglia, 2004). DNA sequence data have played a key role (Goffinet et al., 2009), but compared to vascular plants there has been a relatively limited number of plastid DNA markers used for studying backbone relationships (Stech and Quandt, 2010). In Chapter 2 I developed 78 new primers that permit ready retrieval of 14-17 plastid genes and associated noncoding regions (~8-13 kb unaligned bp) for phylogenetic inference of mosses and relatives. Most of the targeted regions have not been used in the inference of these backbone relationships before. I tested the utility of these primers for retrieving the targeted regions from a broad range of bryophyte taxa. One region comprising three genes (the 3’–rps12- rps7-ndhB region) proved more difficult to retrieve than the others, but in general the primers allowed straightforward amplification and sequencing of the targeted regions from a broad variety of bryophytes. These regions generally have a conservative rate of evolution (plastid DNA is considerably more slowly evolving than nuclear genes; Wolfe et al., 1987; Olmstead and Palmer, 1994) but span a range of evolutionary rates (e.g., coding vs. noncoding regions), and so should be useful at multiple levels of phylogenetic inference in moss and bryophyte phylogeny. I demonstrated the utility of these regions in phylogenetic inference of the earliest splits in land-plant (embryophyte) phylogeny, even when using a relatively sparse taxon sampling of the major lineages of bryophytes (Chapter 3). The major land-plant groups were resolved as monophyletic with strong support from maximum likelihood (ML) analysis, as were the !! "#$! relationships among them, corroborating studies based on other multi-gene analyses and plastome structural data (e.g., Kelch et al., 2004; Wolf et al., 2005; Qiu et al., 2006, 2007): in particular, I recovered the liverworts as the sister group of all other land plants, and the hornworts as the sister group of vascular plants, both with strong support. The deepest phylogenetic splits within liverworts and hornworts were also well supported here, and were consistent with current understanding of the backbone phylogeny in each group (e.g., Renzaglia et al., 2007). The charophycean algal lineage Zygnematales (represented by Zygnema) was recovered as the sister group of land plants with moderate ML bootstrap support, consistent with some other recent large-scale analyses of plastid and nuclear data (Turmel at el., 2006, 2007; Wodniok et al., 2011). Relationships among the major lineages of nonperistomate mosses were also generally well supported. A sister-group relationship between Andreaeobryopsida (Andreaeobryum) and Andreaeopsida (Andreaea) received moderate to strong support from various ML analyses, and Oedipodiopsida (Oedipodium) was recovered the sister group of the peristomate mosses with strong support, consistent with previous studies (e.g., Cox et al., 2004). The very first evolutionary split in moss phylogeny has not been resolved satisfactorily to date. Although some studies have recovered a clade comprising Sphagnum and Takakia as the sister group of the remaining moss taxa (e.g., Cox et al., 2004; Qiu et al., 2006), others find Takakia as the sister group of all other mosses (e.g., Forrest et al., 2006; Volkmar and Knoop, their Fig. 3). No morphological synapomorphies satisfactorily support either hypothesis (e.g., Newton et al., 2001), and inference of this important feature of moss phylogeny may be prone to long-branch attraction (Shaw and Renzaglia, 2004). In Chapter 3 I recovered two conflicting hypotheses !! ""#! regarding the earliest splits in moss phylogeny: Takakia as the sister group of Sphagnum (using maximum parsimony, MP), or Takakia as the sister group of all other mosses (using ML), both with strong bootstrap support. The latter ‘Takakia-sister’ result may be more trustworthy, as ML is known to be less prone to long-branch attraction than MP, in general (Huelsenbeck, 1995). More rapidly evolving sites are expected to be more prone to saturation effects (Felsenstein, 1983). When I identified and removed these sites using a likelihood-based site-rate classification, I was able to recover the ‘Takakia-sister’ result using MP analysis. However, this result was sensitive to the initial tree topology used to classify site rates, suggesting that ML-based rate filtering is not a general solution to tree mis-inference in MP analysis. I explored the Takakia problem further in Chapter 4, a study that sampled additional exemplar taxa across the moss phylogeny. This study had a major focus on the peristomate mosses, but also included several additional nonperistomate taxa. I found that ML inferences concerning the position of Takakia can be highly sensitive to how base frequencies are handled in DNA substitution models (i.e., empirical vs. ML estimates of these parameters), and to the type of data partitioning used. I documented substantial differences in the tempo and mode of DNA substitution in different data partitions, but combined vs. separate analysis of these partitions did always not yield straightforward (predictable) differences concerning which arrangement of Takakia was favoured by ML analysis, suggesting that interactions among different data partitions may be at last partly responsible for misdirecting phylogenetic inference. ML analyses that partitioned the plastid data by codon partition (considering noncoding data as a fourth partition) yielded trees with higher likelihood scores than unpartitioned ML analyses or the ML analyses that partitioned the data by major plastid region, regardless of the program used !! """! for ML inference. The codon-based partitioning recovered the Takakia-sister arrangement. Although the first split in moss phylogeny remains controversial, the strong conflict observed in ML analysis here for this relationship suggests that using ML estimates of base frequencies may be less misleading than using empirical estimates (the latter method is used as a short-cut in one widely used ML program, the current version of RAxML, Stamatakis, 2006). This conclusion assumes that the Takakia-sister arrangement is correct. The major goal of Chapter 4 was to improve our understanding of relationships among the major lineages of peristomate mosses. The major clades recognized as classes and subclasses in current moss classification (Goffinet et al., 2009) were recovered as monophyletic at current taxon samplings with strong support, as were many of the relationships among them. For example, within Bryopsida, the sub-classes Buxbaumiidae and Diphysciidae were recovered as successive sister-groups to a major clade comprising Funariidae, Timmiidae, Bryidae and Dicranidae, with strong ML bootstrap support. The placement of Funariidae and Timmiidae with respect to the latter two clades has not previously been convincingly resolved. Here I recovered Timmia (Timmiidae, which has a unique peristome teeth morphology) with strong support as the sister group of the clade comprising Bryidae (diplolepideous-alternate peristomes) and Dicranidae (haplolepideous peristomes). The diplolepideous-opposite Funariidae was well supported as the sister group of the other three lineages. Relationships within Dicranidae and Funariidae were also generally resolved here, with improved support. However, within Bryidae, substantial ambiguities remain especially concerning the grade of lineages that subtends the core pleurocarps. Relationships between the two nematodontous lineages of mosses (classes !! ""#! Polytrichopsida and Tetraphidopsida) relative to the arthrodontous Bryopsida, also eluded satisfactory resolution here. The two most species-rich lineages of nonperistomate mosses are the peat mosses, Sphagnopsida, and the granite mosses, Andreaeaceae. Phylogenetic relationships within the former clade have been relatively well studied (e.g., Cronberg, 1996; Natcheva, 2007; Shaw et al., 2010a), but the latter are more poorly known. I characterized phylogenetic relationships within Andreaeaceae using two complementary taxon and plastid gene samplings (17 vs. 6 plastid loci). The 17-gene data set included a handful of exemplar taxa from the major lineages of Andreaeaceae (as defined by the 6-region analysis), and confirmed a sister-group relationship of the family to the monotypic family Andreaeobryaceae with moderate to strong ML bootstrap support. It also resolved the root split in the family with strong support. The more densely sampled 6-locus analysis included nearly all costate species and a broad selection of ecostate species. The current subgeneric taxonomy of Andreaeaceae, and the utility of the characters used to support them (e.g., the presence or absence of costa on leaves) were rejected, as I identified several major lineages that each includes species from different subgeneric units. The monophyly of each lineage and the relationships among them were strongly supported, indicating that a substantial revision of the genus above the species level is needed, and could be based on the clades identified here if these can be confirmed by other data sources. In addition, a subset of species were recovered as non-monophyletic according to the plastid data, with weak to moderate support (e.g., Andreaea nivalis and A. rupestris), implying that there may have been cryptic speciation in at least these lineages. !! ""#! 6.2 Future directions  The availability of large-scale genomic data sets promises to transform systematic and evolutionary research in bryology, permitting fairly ready retrieval of whole plastid genomes or transcriptome-based data sets, for example (Cronn et al. 2008; Wodniak et al. 2011). However, there is still a place for large-scale inferences of bryophyte phylogeny using more focused collections of plastid genes, not least to cleanly connect phylogenetic inference in these early branches of land-plant phylogeny to the extensive plastid-based inferences that have already been made in vascular plants (e.g., APG III 2009), and to permit retrieval of minimal DNA sequence data from fragmentary or degraded museum-based samples of poorly collected clades. It is not clear that sampling whole plastid genomes offers significant advantages to the smaller plastid gene set considered here, as the latter alone appear to permit inference of a well- supported phylogenetic backbone; nonetheless, some of the more poorly supported branches identified here would likely benefit from expansion of the plastid gene set used in phylogenetic inference.  The orthology of plastid-based genes is unambiguous (Olmstead and Palmer, 1994): the same cannot be said for nuclear genes, although reconciling gene tree conflicts among different linkage groups (plastid, mitochondrial and especially nuclear) can constitute an additional sources of phylogenetic information (e.g., Ness et al., in press). On the other hand, recovering the same clades using different linkage groups is a powerful tool for systematic inference, by permitting cross-validation (corroboration) of results obtained from individual genomes. It is therefore important to obtain a broadly sampled, well-supported plastid-based phylogenetic estimate as a starting point for comparison. This had not been done before for many of the !! ""#! lineages investigated here. But clearly, more intensive taxon sampling would also be useful, particularly in species-rich clades like Bryidae that still have poor resolution of many early, key evolutionary splits. Continuing the taxon sampling started here in peristomate mosses for the regions surveyed here would be valuable, and should be relatively straightforward to do using ‘next-generation’ sequencing technology (e.g., by multiplexing sequencing products from different species for the gene set considered here).  The surprising result obtained here for Takakia, where different method of phylogenetic analysis (or even different ways of estimating basic parameters) led to conflicting phylogenetic estimates, underlines that we should remain vigilant for phylogenetic mis-inference. Given the great age of some the clades investigated here, it is perhaps surprising that more conflicts were not uncovered. However, it should also be borne in mind that lack of conflict in different analysis is no guarantee that any analysis is free from systematic bias.  Finally, clearly considerable additional work is needed for Andreaeaceae, one of a small handful of evolutionarily isolated lineages of nonperistomate mosses that collectively define the first branches of moss phylogeny. It is straightforward to identify what is needed next for molecular phylogenetic study of this family: more taxa/samples (particularly so for the ecostate species) and more genes (especially additional linkage groups). It will be important to link this work back to the growing body of work being developed using classical systematic methods that focus on morphology (e.g., Murray 1988b and 2006) to refine our understanding of the family’s classification. But the deep divergences and lack of monophyly observed for some individual species underlines that future research in Andreaeaceae needs to blend the toolkits of classical !! ""#! and molecular methods of moss systematics, and the insights that it is possible to gain from them. !! ""#! REFERENCES AKAIKE, H. 1974. A new look at the statistical model identification. IEEE Transactions on Automatic Control 19: 716-723. ALLEN, B. H. 1987. A systematic account of Pulchrinodus inflatus (Musci: Pterobryaceae), genus novum. New Zealand Journal of Botany 25: 335-342. ANGIOSPERM PHYLOGENY GROUP. 2009. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III. Botanical Journal of the Linnean Society 161: 105-121. BAUM, D. A., AND M. J. DONOGHUE. 1995. Choosing among alternative "phylogenetic" species concepts. Systematic Botany 20: 560-573. BECKER, B., AND B. MARIN. 2009. Streptophyte algae and the origin of embryophytes. Annals of Botany 103: 999-1004. BECKERT, S., H. MUHLE, D. PRUCHNER, AND V. KNOOP. 2001. The mitochondrial nad2 gene as a novel marker locus for phylogenetic analysis of early land plants: a comparative analysis in mosses. Molecular Phylogenetics and Evolution 18: 117-126. BELL, N. E., AND J. HYVONEN. 2008. Rooting the Polytrichopsida: The phylogenetic position of Atrichopsis and the independent origin of the polytrichopsid peristome. In H. Mohamed, B. B. Baki, A. Nasrulhaq-Boyce, and P. K. Y. Lee [eds.], Bryology in the new millenium, 227 – 239. University of Malaya, Kuala Lumpur, Malaysia. !! ""#! BELL, N. E., AND J. HYVONEN. 2010. A phylogenetic circumscription of Polytrichastrum (Polytrichaceae): Reassessment of sporophyte morphology supports molecular phylogeny. American Journal of Botany 97: 566-578. BELL, N. E., D. QUANDT, T. J. O'BRIEN, AND A. E. NEWTON. 2007. Taxonomy and phylogeny in the earliest diverging pleurocarps: square holes and bifurcating pegs. The Bryologist 110: 533-560. BERGTHORSSON, U., A. O. RICHARDSON, G. J. YOUNG, L. R. GOERTZEN, AND J. D. PALMER. 2004. Massive horizontal transfer of mitochondrial genes from diverse land plant donors to the basal angiosperm Amborella. Proceedings of the National Academy of Sciences USA 101: 17747-17752. BRAITHWAITE, R. 1880-1887. The British Moss Flora. Vol. I. London. BUDKE, J. M., C. S. JONES, AND B. GOFFINET. 2007. Development of the enigmatic peristome of Timmia megapolitana (Timmiaceae; Bryophyta). American Journal of Botany 94: 460- 467. BURLEIGH, J. G., AND S. MATHEWS. 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. CHANG, J. T. 1996. Inconsistency of evolutionary tree topology reconstruction methods when substitution rates vary across characters. Mathematical Biosciences 134: 189-215. CHANG, Y., C. W. KHIONG, S. W. GRAHAM, AND B. C. TAN. 2006. Molecular evidence for the systematic positions of two enigmatic mosses: Pterogonidium pulchellum !! ""#! (Sematophyllaceae, Musci) and Piloecium pseudorufescens (Myuriaceae, Musci). Canadian Journal of Botany 84: 501-507. CHASE, M. W., D. E. SOLTIS, R. G. OLMSTEAD, D. MORGAN, D. H. LES, B. D. MISHLER, M. R. DUVALL, R. A. PRICE, H. G. HILLS, Y.-L. QIU, K. A. KRON, J. H. RETTIG, E. CONTI, J. D. PALMER, J. R. MANHART, K. J. SYTSMA, H. J. MICHAELS, W. J. KRESS, K. G. KAROL, W. D. CLARK, M. HEDREN, S. G. BRANDON, R. K. JANSEN, K.-J. KIM, C. F. WIMPEE, J. F. SMITH, G. R. FURNIER, S. H. STRAUSS, Q.-Y. XIANG, G. M. PLUNKETT, P. S. SOLTIS, S. M. SWENSEN, S. E. WILLIAMS, P. A. GADEK, C. J. QUINN, L. E. EGUIARTE, E. GOLENBERG, G. H. LEARN, JR., S. W. GRAHAM, S. C. H. BARRETT, S. DAYANANDAN, AND V. A. ALBERT. 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., R. S. COVWAN, P. M. HOLLINGSWORTH, C. VAN DE BERG, AND E. AL. 2007. A proposal for a standardised protocol to barcode all land plants. Taxon 56: 295-299. COX, C. J., B. GOFFINET, A. J. SHAW, AND S. B. BOLES. 2004. Phylogenetic relationships among the mosses based on heterogeneous Bayesian analysis of multiple genes from multiple genomic compartments. Systematic Botany 29: 234-250. COX, C. J., B. GOFFINET, N. WICKETT, S. B. BOLES, AND A. J. SHAW. 2010. Moss diversity: A molecular phylogenetic analysis of genera. Phytotaxa 9: 175-195. CRANDALL-STOTLER, B. 1980. Morphogenetic designs and a theory of bryophyte origins and divergence. BioScience 30: 580-585. !! ""#! CRANDALL-STOTLER, B., R. E. STOTLER, AND D. G. LONG. 2009. Morphology and classificaiton of the Marchantiophyta. In B. Goffinet and A. J. Shaw [eds.], Bryophyte Biology, 1-54. Cambridge University Press, New York. CRONBERG, N. 1996. Isozyme evidence of relationships withinSphagnum sect.Acutifolia (Sphagnaceae, Bryophyta). Plant Systematics and Evolution 203: 41-64. CRONN, R, A. LISTON, M. PARKS, D. S. GERNANDT, R. SHEN, AND R MOCKLER. 2008. Multiplex sequencing of plant chloroplast genomes using Solexa sequencing-by-synthesis technology. Nucleic Acids Research 36: e122. CRUM, H. A., AND L. E. ANDERSON. 1981. Mosses of Eastern North America. Columbia University Press, New York. CUMMINGS, M. P., S. A. HANDLEY, D. S. MYERS, D. L. REED, A. ROKAS, AND K. WINKA. 2003. Comparing bootstrap and posterior probability values in the four-taxon case. Systematic Biology 52: 477-487. DAVIS, C. 2004. A molecular phylogeny of leafy liverworts (Jungermanniidae, Marchantiophyta). In B. Goffinet, V. Hollowell, and R. Magill [eds.], Molecular Systematics of Bryophytes, 61-86. Missouri Botanical Garden Press, St. Louis. DIXON, H. N. 1932. Classification of mosses. In F. Verdoorn [ed.], Manual of Bryology, 397-412. Martius Nijhoff, The Hague. DOYLE, J. J., AND J. L DOYLE. 1987. A rapid isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11–15. DUFF, R., D. C. CARGILL, J. C. VILLARREAL A., AND K. RENZAGLIA. 2004. Phylogenetic relationships of the hornworts based on rbcL sequence data: novel relationships and new !! "#$! insights. In B. Goffinet, V. Hollowell, and R. Magill [eds.], Molecular Systematics of Bryophytes, 41-58. Missouri Botanical Garden Press, St. Louis. DUFF, R. J., J. C. VILLARREAL, D. C. CARGILL, AND K. S. RENZAGLIA. 2007. Progress and challenges toward developing a phylogeny and classification of the hornworts. The Bryologist 110: 214-243. EDWARDS, S. R. 1979. Taxonomic implications of cell patterns in haplolepidous moss peristomes. In G. C. S. Clarke and J. G. Duckett [eds.], Bryophyte Systematics. Academic Press, New York. FAZEKAS, A. J., K. S. BURGESS, P. R. KESANAKURTI, S. W. GRAHAM, S. G. NEWMASTER, B. C. HUSBAND, D. M. PERCY, M. HAJIBABAEI, AND S. C. H. BARRETT. 2008. Multiple multilocus DNA barcodes from the plastid genome discriminate plant species equally well. PLoS ONE 3: e2802. FELSENSTEIN, J. 1978. Cases in which parsimony or compatibility methods will be positively misleading. Systematic Zoology 27: 401-410. ______. 1983. Parsimony in systematics: biological and statistical issues. Annual Review of Ecology and Systematics 14: 313-333. ______. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791. FERNANDEZ, C. C., J. R. SHEVOCK, A. N. GLAZER, AND J. N. THOMPSON. 2006. Cryptic species within the cosmopolitan desiccation-tolerant moss Grimmia laevigata. Proceedings of the National Academy of Sciences of the United States of America 103: 637-642. !! "#"! FINET, C., R. E. TIMME, C. F. DELWICHE, AND F. MARLÉTAZ. 2010. Multigene phylogeny of the green lineage reveals the origin and diversification of land plants. Current Biology 20: 2217-2222. FLEISCHER, M. 1904-1923. Die Musci der Flora von Buitenzorg (Zugleich Laubmoosflora von Java). Vols. 1-4. E. J. Brill, Leiden. FORREST, L. L., E. C. DAVIS, D. G. LONG, B. J. CRANDALL-STOTLER, A. CLARK, AND M. L. HOLLINGSWORTH. 2006. Unraveling the evolutionary history of the liverworts (Marchantiophyta): multiple taxa, genomes and analyses. The Bryologist 109: 303-334. FREY, W. AND M. STECH. 2009. Bryophyta (Musci, Mosses). In W. Frey (ed.). Syllabus of Plant Families A. Engler's Syllabus der Pflanzenfamilien. Part 3. Bryophytes and Seedless Vascular Plants. 13th Edition, 116-257. Gebr. Borntraeger Verlagsbuchhandlung, Stuttgart, Germany. GARBARY, D. J., AND K. S. RENZAGLIA. 1998. Bryophyte phylogeny and the evolution of land plants: evidence from development and ultrastructure. In J. W. Bates, N. W. Ashton, and J. G. Duckett [eds.], Bryology for the twenty-first century, 45-63. The British Bryological Society and Maney Publishing, Leeds. GOFFINET, B., AND W. R. BUCK. 2004. Systematics of the Bryophyta (mosses) from molecules to a revised classification. In B. Goffinet, V. Hollowell, and R. Magill [eds.], Molecular Systematics of Bryophytes, 204-239. Missouri Botanical Garden Press, St. Louis. GOFFINET, B., C. J. COX, A. J. SHAW, AND T. A. J. HEDDERSON. 2001. The Bryophyta (Mosses): systematic and evolutionary inferences from an rps4 gene (cpDNA) phylogeny. Annals of Botany 87: 191-208. !! "##! GOFFINET, B., N. J. WICKETT, O. WERNER, R. M. ROS, A. J. SHAW, AND C. J. COX. 2007. Distribution and phylogenetic significance of the 71-kb inversion in the plastid genome in Funariidae (Bryophyta). Annals of Botany 99: 747-753. GOFFINET, B., W. R. BUCK, AND A. J. SHAW. 2009. Morphology, anatomy, and classification of the Bryophyta. In B. Goffinet and A. J. Shaw [eds.], Bryophyte Biology, 55-138. Cambridge University Press, New York. GOREMYKIN, V. V., AND F. H. HELLWIG. 2005. Evidence for the most basal split in land plants dividing bryophyte and tracheophyte lineages. Plant Systematics and Evolution 254: 93- 103.! GOLDMAN, N., J. P. ANDERSON, AND A. G. RODRIGO. 2000. Likelihood-based test of topologies in phylogenetics. Systematic Biology 49: 652-670. GRAHAM, S. W., AND W. J. D. ILES. 2009. Different gymnosperm outgroups have (mostly) congruent signal regarding the root of flowering plant phylogeny. American Journal of Botany 96: 216-227. GRAHAM, S. W., AND R. G. OLMSTEAD. 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 R. G. OLMSTEAD. 2000b. Evolutionary significance of an unusual chloroplast DNA inversion found in two basal angiosperm lineages. Current Genetics 37: 183–188. GRAHAM, S. W., J. R. KOHN, B. R. MORTON, J. E. ECKENWALDER, AND S. C. H. BARRETT. 1998. Phylogenetic congruence and discordance among one morphological and three molecular data sets from Pontederiaceae. Systematic Biology 47: 545-567. !! "#$! GRAHAM, S. W., P. A. REEVES, A. C. E. BURNS, AND R. G. OLMSTEAD. 2000. Microstructural changes in noncoding chloroplast DNA: interpretation, evolution, and utility of indels and inversions in basal angiosperm phylogenetic inference. International Journal of Plant Sciences 161: S83-S96. GRAHAM, S. W., R. G. OLMSTEAD, AND S. C. H. BARRETT. 2002. Rooting phylogenetic trees with distant outgroups: a case study from the commelinoid monocots. Molecular  Biology and Evolution 19: 1769-1781. GRAHAM, S. W., J. M. ZGURSKI, M. A. MCPHERSON, D. M. CHERNIAWSKY, J. M. SAARELA, E. F. C. HORNE, S. Y. SMITH ET AL. 2006. Robust inference of monocot deep phylogeny using an expanded multigene plastid data set. In J. T. Columbus, E. A. Friar, J. M. Porter, L. M. Prince, and M. G. Simpson [eds.], Monocots: comparative biology and evolution (excluding Poales). Rancho Santa Ana Botanic Garden, Claremont, California, USA. GUINDON, S. P., AND O. GASCUEL. 2003. A Simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Systematic Biology 52: 696-704. HEDDERSON, T. A., D. J. MURRAY, C. J. COX, AND T. L. NOWELL. 2004. Phylogenetic relationships of haplolepideous mosses (Dicranidae) inferred from rps4 gene sequences. Systematic Botany 29: 29-41. HENDY, M.D., AND D. PENNY. 1989. A framework for the quantitative study of evolutionary trees. Systematic Zoology 38: 297–309. HE-NYGRÉN, X., A. JUSLÉN, I. AHONEN, D. GLENNY, AND SINIKKA PIIPPO. 2006. Illuminating the evolutionary history of liverworts (Marchantiophyta): towards a natural classification. Cladistics 22: 1-31. !! "#$! HILLIS, D. M. 1995. Approaches for assessing phylogenetic accuracy. Systematic Biology 55: 44: 3-16. HUELSENBECK, J. P. 1995. Performance of Phylogenetic Methods in Simulation. Systematic Biology 44: 17-48. ______. 1997. Is the Felsenstein zone a fly trap? Systematic Biology 46: 69-74. ______. 1998. Systematic bias in phylogenetic analysis: Is the Strepsiptera problem solved? Systematic Biology 47: 519-537. KELCH, D. G., A. DRISKELL, AND B. D. MISHLER. 2004. Inferring phylogeny using genomic characters: a case study using land plant plastomes. In B. Goffinet, V. Hollowell, and R. Magill [eds.], Molecular systematics of bryophytes, 3-11. Missouri Botanical Garden Press, St. Louis. KELCHNER, S. A. 2000. The evolution of non-coding chloroplast DNA and its application in plant systematics. Annals of the Missouri Botanical Garden 87: 482-498. KELLOGG, E. A., AND J. L. BENNETZEN. 2004. The evolution of nuclear genome structure in seed plants. American Journal of Botany. 91: 1709-1725. KENRICK, P., AND P. R. CRANE. 1997. The origin and early diversification of land plants: a cladistic study. Smithsonian Press, Washington, D.C., USA. KOLACZKOWSKI, B., AND J. W. THORNTON. 2007. Effects of branch length uncertainty on Bayesian posterior probabilities for phylogenetic hypotheses. Molecular Biology and Evolution 24: 2108-2118. !! "#$! LA FARGE, C., A. J. SHAW, AND D. H. VITT. 2002. The circumscription of the Dicranaceae (Bryopsida) based on the chloroplast regions trnL-trnF and rps4. Systematic Botany 27: 435-452. LA FARGE, C., B. D. MISHLER, J. A. WHEELER, D. P. WALL, K. JOHANNES, S. SCHAFFER, AND A. J. SHAW. 2000. Phylogenetic relationships within the haplolepideous mosses. The Bryologist 103: 257-276. LEWIS, L. A., B. D. MISHLER, AND R. VILGALYS. 1997. Phylogenetic relationships of the liverworts (Hepaticae), a basal embryophyte lineage, inferred from nucleotide sequence data of the chloroplast gene rbcL. Molecular Phylogenetics and Evolution 7: 377-393. LIGRONE, R., AND J. G. DUCKETT. 1998. Development of the leafy shoot in Sphagnum (Bryophyta) involves the activity of both apical and subapical meristems. New Phytologist 140: 581-595. MAGOMBO, Z. L. K. 2003. The phylogeny of basal peristomate mosses: evidence from cpDNA, and implications for peristome evolution. Systematic Botany 28: 24-38. MILLER, M. A., M. T. HOLDER, R. VOS, P. E. MIDFORD, T. LIEBOWITZ, L. CHAN, P. HOOVER, AND T. WARNOW 2009. The CIPRES Portals. URL: http://www.phylo.org/sub_sections/portal. (Archived by WebCite(r) at http://www.webcitation.org/5imQlJeQa) MISHLER, B. D., AND S. P. CHURCHILL. 1984. A cladistic approach to the phylogeny of the "bryophytes". Brittonia 36: 406-424. ______. 1985. Transistion to a land flora: phylogenetic relationships of the green algae and bryophytes Cladistics 1: 305-328. !! "#$! MIYATA, Y., AND M. SUGITA. 2004. Tissue- and stage-specific RNA editing of rps14 transcripts in moss (Physcomitrella patens) chloroplasts. Journal of Plant Physiology 161: 113-115. MURRAY, B. M. 1987a. Andreaea schofieldiana and A. megistospora, species novae, and taxonomic criteria for sect. Nerviae (Andreaeopsida). The Bryologist 90: 15-26. ______. 1987b. Illustrated moss flora of arctic North America and Greenland. Meddelelser om Groønland, Bioscience 23: 15-26. ______. 1988a. Systematics of the Andreaeopsida (Bryophyta): two orders with links to Takakia. Beihefte zur Nova Hedwigia 90: 289-336. ______. 1988b. The genus Andreaea in Britain and Ireland. Journal of Bryology 15: 17-82. ______.  2006. Andreaeaceae. Flora of Australia 51: 108-123, 390-391, 406-409. NATCHEVA, R., AND N. CRONBERG. 2007. Recombination and introgression of nuclear and chloroplast genomes between the peat mosses, Sphagnum capillifolium and  Sphagnum quinquefarium. Molecular Ecology 16: 811-818. NESS, R. W., S. W. GRAHAM, AND S. C. H. BARRETT. In press. Reconciling gene and genome duplicaiton events: using multiple nuclear gene families to infer the phylogeny of the aquatic plant family Pontederiaceae. Molecular  Biology and Evolution. NEWTON, A. E., C. J. COX, J. G. DUCKETT, J. A. WHEELER, B. GOFFINET, T. A. J. HEDDERSON, AND B. D. MISHLER. 2000. Evolution of the major moss lineages: phylogenetic analyses based on multiple gene sequences and morphology. The Bryologist 103: 187-211. NEWTON, A. E., C. J. COX, J. G. DUCKETT, J. A. WHEELER, B. GOFFINET, T. A. J. HEDDERSON, AND B. D. MISHLER. 2000. Evolution of the major moss lineages: phylogenetic analyses based on multiple gene sequences and morphology. The Bryologist 103: 187-211. !! "#$! NEWTON, A. E., N. WILKSTRÖM, AND A. J. SHAW. 2009. Mosses (Bryophyta). In S. B. Hedges and S. Kumar [eds.], The Timetree of Life, 138-145. Oxford University Press, Oxford, UK. NICKRENT, D. L., C. L. PARKINSON, J. D. PALMER, AND R. J. DUFF. 2000. Multigene phylogeny of land plants with special reference to bryophytes and the earliest land plants. Molecular Biology and Evolution 17: 1885-1895. NISHIYAMA, T., P. G. WOLF, M. KUGITA, R. B. SINCLAIR, M. SUGITA, C. SUGIURA, T. WAKASUGI, K. YAMADA, K. YOSHINAGA, K. YAMAGUCHI, K. UEDA, AND M. HASEBE. 2004. Chloroplast phylogeny indicates that bryophytes are monophyletic. Molecular Biology and Evolution 21: 1813-1819. O'BRIEN, T. J. 2007. The phylogenetic distribution of pleurocarpous mosses: evidence from cpDNA sequences. In A. E. Newton and R. S. Tangney [eds.], Pleurocarpous mosses: systematics and evolution, 41–64. CRC Press, Boca Raton. OCHYRA, R., J. ZARNOWIEC, AND H. BEDNAREK-OCHYRA. 2003. Census catalogue of Polish mosses. Biodiversity of Poland 3: 1-372. OLMSTEAD, R. G. 1995. Species concepts and plesiomorphic species. Systematic Botany 20: 623- 630. OLMSTEAD, R. G., AND J. D. PALMER. 1994. Chloroplast DNA systematics: A review of methods and data analysis. American Journal of Botany 81: 1205–1224. OLMSTEAD, R. G., AND J. A. SWEERE. 1994. Combining data in phylogenetic systematics: an empirical approach using three molecular data sets in the Solanaceae. Systematic Biology 43: 467-481. !! "#$! OLSSON, S., V. BUCHBENDER, J. ENROTH, L. HEDEN!§S, S. HUTTUNEN, AND D. QUANDT. 2009. Phylogenetic analyses reveal high levels of polyphyly among pleurocarpous lineages as well as novel clades. The Bryologist 112: 447-466. PALMER, J. D., K. L. ADAMS, Y. CHO, C. L. PARKINSON, Y.-L. QIU, AND K. SONG. 2000. Dynamic evolution of plant mitochondrial genomes: Mobile genes and introns and highly variable mutation rates. Proceedings of the National Academy of Sciences USA 97: 6960- 6966. PENNY, D., L. R. FOULDS, AND M. D. HENDY. 1982. Testing the theory of evolution by comparing phylogenetic trees constructed from five different protein sequences. Nature 297: 197- 200. POND, K. S. L., S. D. W. FROST, AND S. V. MUSE. 2005. HyPhy: hypothesis testing using phylogenies. Bioinformatics 21: 676-679. POSADA, D. 2008. jModelTest: phylogenetic model averaging. Molecular Biology and Evolution 25: 1253-1256. QIU, Y.-L. 2008. Phylogeny and evolution of charophytic algae and land plants. Journal of Systematics and Evolution 46: 287-306. QIU, Y.-L., Y. CHO, J. C. COX, AND J. D. PALMER. 1998. The gain of three mitochondrial introns identifies liverworts as the earliest land plants. Nature 394: 671-674. QIU, Y.-L., L. LI, B. WANG, Z. CHEN, V. KNOOP, M. GROTH-MALONEK, O. DOMBROVSKA, J. LEE, L. KENT, J. REST, G. F. ESTABROOK, T. A. HENDRY, D. W. TAYLOR, C. M. TESTA, M. AMBROS, B. CRANDALL-STOTLER, R. J. DUFF, M. STECH, W. FREY, D. QUANDT, AND !! "#$! C. C. DAVIS. 2006. The deepest divergences in land plants inferred from phylogenomic evidence. Proceedings of the National Academy of Sciences 103: 15511-15516. QIU, Y.-L., L. LI, B. WANG, Z. CHEN, O. DOMBROVSKA, J. LEE, L. KENT, R. LI, R. W. JOBSON, T. A. HENDRY, D. W. TAYLOR, C. M. TESTA, AND M. AMBROS. 2007. A nonflowering land plant phylogeny inferred from nucleotide sequences of seven chloroplast, mitochondrial, and nuclear genes. International Journal of Plant Sciences 168: 691-708.  QIU, Y.-L., L. LI, B. WANG, J.-Y. XUE, T. A. HENDRY, R.-Q. LI, J. W. BROWN, Y. LIU, G. T. HUDSON, AND Z.-D. CHEN. 2010. Angiosperm phylogeny inferred from sequences of four mitochondrial genes. Journal of Systematics and Evolution 48: 391-425. QUANDT, D., AND M. STECH. 2004. Molecular evolution of the trnTUGU-trnFGAA region in bryophytes. Plant Biology 6: 545-554. QUANDT, D., K. MÜLLER, AND S. HUTTUNEN. 2003. Characterisation of the chloroplast DNA psbT-H region and the influence of dyad symmetrical elements on phylogenetic reconstructions. Plant Biology 5: 400-410. RAI, H. S., AND S. W. GRAHAM. 2010. Utility of a large, multigene plastid data set in inferring higher-order relationships in ferns and relatives (monilophytes). American Journal of Botany 97: 1444-1456. RAI, H. S., H. E. O'BRIEN, P. A. REEVES, R. G. OLMSTEAD, AND S. W. GRAHAM. 2003. Inference of higher-order relationships in the cycads from a large chloroplast data set. Molecular Phylogenetics and Evolution 29: 350-359. !! "#$! RAI, H. S., P. A. REEVES, R. PEAKAL, R. G. OLMSTEAD, AND S. W. GRAHAM. 2008. Inference of higher-order conifer relationships from a multi-locus plastid data set. Botany 86: 658- 669. RAMBAUT, A. 2002. Se-Al sequence alignment editor, ver. 2.0a11. Oxford: University of Oxford. RENZAGLIA, K. S., R. JOEL DUFF, D. L. NICKRENT, AND D. J. GARBARY. 2000. Vegetative and reproductive innovations of early land plants: implications for a unified phylogeny. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences 355: 769-793. RENZAGLIA, K. S., S. SCHUETTE, R. J. DUFF, R. LIGRONE, A. J. SHAW, B. D. MISHLER, AND J. G. DUCKETT. 2007. Bryophyte phylogeny: Advancing the molecular and morphological frontiers. The Bryologist 110: 179-213. RENZAGLIA, K., J. C. VILLARREAL A., AND R. DUFF. 2009. New insights into morphology, anatomy, and systematics of hornworts. In B. Goffinet and A. J. Shaw [eds.], Bryophyte Biology, 139-171. Cambridge University Press, New York. ROTH, G. 1903-1904. Die europäischen Laubmoose, Vol. 1. Leipzig. ROTH, G. 1910-1911. Die aussereuropäischen Laubmoose, Vol. 1. Dresden. RYDIN, H. 2009. Population and community ecology. In B. Goffinet and A. J. Shaw [eds.], Bryophyte Biology, 393-444. Cambridge University Press, New York, New York, USA. SAARELA, J. M., H. S. RAI, J. A. DOYLE, P. K. ENDRESS, S. MATHEWS, A. D. MARCHANT, B. G. BRIGGS, AND S. W. GRAHAM. 2007. Hydatellaceae identified as a new branch near the base of the angiosperm phylogenetic tree. Nature 446: 312-315. !! "#"! SAINSBURY, G. O. K. 1955. A Handbook of the New Zealand Mosses. The Royal Society of New Zealand Bulletin 5: 1-490. SCHOFIELD, W. B. 1985. Introduction to bryology. Macmillan, New York, USA. SCHULTZE-MOTEL, W. 1970. Monographie der Laubmoosgattung Andreaea. 1. Die costaten Arten. Willdenowia 6: 25-110. SCHWARZ, G. E. 1978. Estimating the dimension of a model. Annals of Statistics 6: 461-464. SHAW, A. J. 2009. Bryophyte species and speciation.  In B. Goffinet and A. J. Shaw [eds.], Bryophyte Biology, 445-485. Cambridge University Press, New York, New York, USA. SHAW, J., AND L. E. ANDERSON. 1988. Peristome development in mosses in relation to systematics and evolution. II. Tetraphis pellucida (Tetraphidaceae). American Journal of Botany 75: 1019-1032. SHAW, J., AND K. RENZAGLIA. 2004. Phylogeny and diversification of bryophytes. American Journal of Botany 91: 1557-1581. SHAW, A. J., C. J. COX, B. GOFFINET, W. R. BUCK, AND S. B. BOLES. 2003. Phylogenetic evidence of a rapid radiation of pleurocarpous mosses (Bryophyta) Evolution 57: 2226- 2241. SHAW, A. J., N. DEVOS, C. J. COX, S. B. BOLES, B. SHAW, A. M. BUCHANAN, L. CAVE, AND R. SEPPELT. 2010a. Peatmoss (Sphagnum) diversification associated with Miocene Northern Hemisphere climatic cooling? Molecular Phylogenetics and Evolution 55: 1139-1145. SHAW, A. J., C. J. COX, W. R. BUCK, N. DEVOS, A. M. BUCHANAN, L. CAVE, R. SEPPELT, ET AL. 2010b. Newly resolved relationships in an early land plant lineage: Bryophyta class Sphagnopsida (peat mosses). American Journal of Botany 97: 1511-1531. !! "#$! SHIMODAIRA, H., AND M. HASEGAWA. 2001. CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinformatics 17: 1246-1247. SHIMODAIRA, H. 2002. An approximately unbiased test of phylogenetic tree selection. Systematic Biology 51: 492-508. SIMMONS M. P., AND H. OCHOTERENA. 2000. Gaps as characters in sequence-based phylogenetic analyses. Systematic Biology 49: 369–381. SIMMONS, M. P., K. M. PICKETT, AND M. MIYA. 2004. How meaningful are Bayesian support values? Molecular Biology and Evolution 21: 188-199. SMITH, S. A., J. M. BEAULIEU, AND M. J. DONOGHUE. 2010. An uncorrelated relaxed-clock analysis suggests an earlier origin for flowering plants. Proceedings of the National Academy of Sciences 107: 5897-5902. STAMATAKIS, A. 2006. RAxML-VI-HPC: Maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22: 2688-2690. STAMATAKIS, A., P. HOOVER, AND J. ROUGEMONT. 2008. A fast bootstrapping algorithm for the RAxML web-servers. Systematic Biology 57: 758-771. STECH, M., AND D. QUANDT. 2010. 20,000 species and five key markers: The status of molecular bryophyte phylogenetics. Phytotaxa 9: 196-228. STECH, M., D. QUANDT, AND W. FREY. 2003. Molecular circumscription of the hornworts (Anthocerotophyta) based on the chloroplast DNA trnL–trnF region. Journal of Plant Research 116: 389-398. STEERE, W. C. 1969. A new look at evolutionand phylogeny in bryophytes. In J. E. Gunkel [ed.], Current Topics in Plant Science, 134-143. Academic Press, New York, USA. !! "##! SUGIURA, C., Y. KOBAYASHI, S. AOKI, C. SUGITA, AND M. SUGITA. 2003. Complete chloroplast DNA sequence of the moss Physcomitrella patens: evidence for the loss and relocation of rpoA from the chloroplast to the nucleus. Nucleic Acids Research 31: 5324-5331. SULLIVAN, J., AND D. L. SWOFFORD. 2001. Should we use model-based methods for phylogenetic inference when we know that assumptions about among-site rate variation and nucleotide substitution pattern are violated? Systematic Biology 50: 723-729. SUNDBERG, K., H. CARROLL, Q. SNELL, AND M. CLEMENT. 2008. Incomparability of results between phylogenetic search programs. International Conference on Bioinformatics and Computational Biology, pp 81-84. SUSKO, E. 2008. On the distributions of bootstrap support and posterior distributions for a star tree. Systematic Biology 57: 602-612. SUZUKI, Y., G. V. GLAZKO, AND M. NEI. 2002. Overcredibility of molecular phylogenies obtained by Bayesian phylogenetics. Proceedings of the National Academy of Sciences USA 99: 16138-16143. SWOFFORD, D. L. 2002. PAUP*-Phylognetic Analyses Using Parsimony (* and Other Methods). Sinauer Associates, Sunderland. SWOFFORD, D. L., P. J. WADDELL, J. P. HUELSENBECK, P. G. FOSTER, P. O. LEWIS, AND J. S. ROGERS. 2001. Bias in phylogenetic estimation and its relevance to the choice between parsimony and likelihood methods. Systematic Biology 50: 525-539. THIERS, B. 2010 [continuously updated]. Index Herbariorum: A global directory of public herbaria and associated staff. New York Botanical Garden's Virtual Herbarium. http://sweetgum.nybg.org/ih/ !! "#$! TSUBOTA, H., Y. AGENO, B. ESTEBANEZ, T. YAMAGUCHI AND H. DEGUCHI. 2003. Molecular phylogeny of Grimmiales (Musci) based on chloroplast rbcL sequences. Hikobia 14: 55- 70. TURMEL, M., C. OTIS, AND C. LEMIEUX. 2006. The chloroplast genome sequence of Chara vulgaris sheds new light into the closest green algal relatives of land plants. Molecular Biololgy and Evolution 23: 1324-1338. TURMEL, M., J.-F. POMBERT, P. CHARLEBOIS, C. OTIS, AND C. LEMIEUX. 2007. The Green algal ancestry of land plants as revealed by the chloroplast genome. International Journal of Plant Sciences 168: 679-689. VANDERPOORTEN, A., AND B. GOFFINET. 2006. Mapping uncertainty and phylogenetic uncertainty in ancestral character state reconstruction: an example in the moss genus Brachytheciastrum. Systematic  Biology 55: 957-971. VANDERPOORTEN, A., AND B. GOFFINET. 2009. Introduction to Bryophytes. Cambridge University Press, Cambridge, UK. VANDERPOORTEN, A., L. HEDENÅS, C. J. COX, AND A. J. SHAW. 2002. Phylogeny and morphological evolution of the Amblystegiaceae (Bryopsida). Molecular Phylogenetics and Evolution 23: 1-21. VITT, D. H. 1980. A comparative study of Andreaea acutifolia, A. mutabilis, and A. rupestris. New Zealand Journal of Botany 18: 367-377. ______. 1984. Classification of the Bryopsida. In R. M. Schuster [ed.], New Manual of Bryology, 696-759. Hattori Botanical Laboratory, Nichinan, Japan. !! "#$! ______. 2006. Functional characteristics and indicators of boreal peatlands. In R. K. Wieder and D. H. Vitt [eds.], Boreal Peatland Ecology, 9-24. Springer-Verlag., Berlin. VITT, D. H., B. GOFFINET, AND T. A. HEDDERSON. 1998. The ordinal classification of the mosses: Questions and answers for the 1990’s. In N. J. W. Bates, W. Ashton, and J. G. Duckett [eds.], Bryology for the Twenty-first Century, 143-159. Maney Publishing and British Bryological Society, Leeds, U.K. VOLKMAR, U., AND V. KNOOP. 2010. Introducing intron locus cox1i624 for phylogenetic analyses in bryophytes: on the issue of Takakia as sister genus to all other extant mosses. Journal of Molecular Evolution 70: 506-518. WAHRMUND, U., T. REIN, K. MÜLLER, M. GROTH-MALONEK, AND V. KNOOP. 2009. Fifty mosses on five trees: comparing phylogenetic information in three types of non-coding mitochondrial DNA and two chloroplast loci. Plant Systematics and Evolution 282: 241- 255. WAHRMUND, U., D. QUANDT, AND V. KNOOP. 2010. The phylogeny of mosses - Addressing open issues with a new mitochondrial locus: Group I intron cobi420. Molecular Phylogenetics and Evolution 54: 417-426. WANG, W., Y. WU, Y. YAN, M. ERMAKOVA, R. KERSTETTER, AND J. MESSING. 2010. DNA barcoding of the Lemnaceae, a family of aquatic monocots. BMC Plant Biology 10: 205. WIJK, R. V. D., W. D. MARGADANT, AND P. A. FLORSCHÜTZ. 1959. Index Muscorum. 1 (A--C). WILSON, R., S. ROBBERT GRADSTEIN, H. SCHNEIDER, AND J. HEINRICHS. 2007. Unravelling the phylogeny of Lejeuneaceae (Jungermanniopsida): Evidence for four main lineages. Molecular Phylogenetics and Evolution 43: 270-282. !! "#$! WODNIOK, S., H. BRINKMANN, G. GLOCKNER, A. HEIDEL, H. PHILIPPE, M. MELKONIAN, AND B. BECKER. 2011. Origin of land plants: Do conjugating green algae hold the key? BMC Evolutionary Biology 11: 104. WOLF, P. G., K. G. KAROL, D. F. MANDOLI, J. KUEHL, K. ARUMUGANATHAN, M. W. ELLIS, B. D. MISHLER, D. G. KELCH, R. G. OLMSTEAD, AND J. L. BOORE. 2005. The first complete chloroplast genome sequence of a lycophyte, Huperzia lucidula (Lycopodiaceae). Gene 350: 117-128. WOLFE, K. H., W. H. LI, AND P. M. SHARP. 1987. Rates of nucleotide substitution vary greatly among plant mitochondrial, chloroplast, and nuclear DNAs. Proceedings of the National Academy of Sciences of the United States of America 84: 9054-9058. YATSENTYUK, S. P. 2001. Molecular phylogeny of the Bryophyta and Lycophyta, according to results from some sequences of chloroplast DNA. Ph.D Dissertation, Moscow State University, Russia. ZGURSKI, J. M., H. S. RAI, Q. M. FAI, D. J. BOGLER, J. FRANCISCO-ORTEGA, AND S. W. GRAHAM. 2008. How well do we understand the overall backbone of cycad phylogeny? New insights from a large, multigene plastid data set. Molecular Phylogenetics and Evolution 47: 1232-1237. ZWICKL, D. J. 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. Ph.D. Dissertation, The University of Texas at Austin, Texas, USA. !! "#$! APPENDICES Appendix 1:  Source information of bryophyte taxa used in primer design and testing in Chapter 2. Herbarium abbreviations: ABSH, Hepatic Herbarium, Southern Illinois University; CONN, The George Safford Torrey Herbarium; E, Royal Botanic Garden, Edinburgh; UBC, University of British Columbia.  A: Previously published data used in primer development. Adiantum capillus-veneris L., Amborella trichopoda Baill., Angiopteris evecta (G. Frost.) Hoffm., Anthoceros formosae Steph., Chaetosphaeridium globosum (Nordst.) Kleb., Chara vulgaris L., Mesostigma viride Lauterborn, Chlorokybus atmophyticus Geitler, Ginkgo biloba L., Gnetum gnemon L., Huperzia lucidula (Michx.) Trevisan, Lilium superbum L., Magnolia stellata Maxim., Marchantia polymorpha I., Nicotiana tabacum L., Nymphaea odorata Aiton, Physcomitrella patens (Hedw.) Bruch & Schimp, Pinus koraiensis Siebold & Zucc., Podocarpus chinensis Wall. ex. J. Forbes, Psilotum nudum (L.) Beauv., Sphagnum L. sp., Trimenia moorei (Oliv.) Philipson, Zygnema circumcarinatum Czurda.  B: Taxa sequenced during primer development. Bryophyta:  Andreaea rupestris Hedw. (Andreaeobryopsida, Andreaeaceae), Y. Chang 06-03, UBC. Atrichum selwynii Aust. (Polytrichopsida, Polytrichaceae), Y. Chang 06-01, UBC. Bryum capillare Hedw. (Bryopsida, Bryaceae), Y. Chang 07-01, UBC. Buxbaumia aphylla Hedw. (Bryopsida, Buxbaumiaceae), W. B. Schofield 126011A, UBC. Dicranum scoparium Hedw. (Bryopsida, Dicranaceae), D. Woods 5-04, UBC. Homalothecium fulgescens (Mitt. ex C. Muell.) Lawt. (Bryopsida, Brachytheciaceae), D. Woods 05-01, UBC. Orthotrichum lyellii Hook. & Tayl. (Bryopsida, Orthotrichaceae), D. Woods 05-03, UBC. Syntrichia norvegica F. Weber  (Bryopsida, Pottiaceae), Y. Chang 06-08, UBC. Takakia ceratophylla (Mitt.) Grolle (Takakiopsida, Takakiaceae), W. B. Schofield 125387, UBC. Takakia lepidozioides Hattori & H. Inoue (Takakiopsida, Takakiaceae), Y. Chang 05-02 UBC. !! "#$! Marchantiophyta: Metzgeria conjugata Lindb. (Jungermanniopsida, Metzgeriaceae), Y. Chang 07-02, UBC. Scapania bolanderi Aust. (Jungermanniopsida, Scapaniaceae), Y. Chang 06-05, UBC.  C. Taxa used in primer testing (see Table 2.1). Bryophyta: Aulacomnium palustre (Hedw.) Schwaegr. (Bryopsida, Aulacomniaceae), W.B. Schofield 123402, UBC. Bartramiopsis lescurii (James) Kindb. (Bryopsida, Bartramiaceae), W.B. Schofield 125406, UBC. Encalypta ciliata Hedw. (Bryopsida, Encalyptaceae), W.B. Schofield 117021, UBC. Oedipodium griffithianum (Dicks.) Schwaegr. (Oedipodiopsida, Oedipodiaceae), W.B. Schofield 115800, UBC. Ptychomnion cygnisetum (C. Mull.) Kindb. (Bryopsida, Ptychomniaceae), Bruce Allen 26391, UBC. Tetrodontium brownianum (Dicks.) Schwaegr. (Tetraphidopsida, Tetraphidaceae), D.G. Long 37600, E. Tetraplodon mnioides (Hedw.) B.S.G. (Bryopsida, Splachnaceae), W.B. Schofield 121611, UBC. Marchantiophyta: Haplomitrium hookeri (Sm.) Nees (Haplomitriopsida, Haplomitriaceae), D. G. Long 33741, E. Treubia lacunosa (Col.) Prosk. (Haplomitriopsida, Treubiaceae), Stotler & Crandall-Stotler 4561, ABSH. Anthoceratophyta: Leiosporoceros dussii (Steph.) Hassel (Leiosporocerotopsida, Leiosporocerotaceae), Villarreal & Rodríguez 852A, CONN. Phaeoceros carolinianus (Michx.) Prosk. (Anthocerotopsida, Notothyladaceae), Bernard Goffinet 9652, UBC.  !! "#$! Appendix 2:  Source information and GenBank numbers for taxa included in Chapter 3.  A: Data obtained in this study: For each taxa, the information is displayed in the following sequence: species (classification), voucher specimen, herbarium; GenBank accession: atpB, ndhF, psbB-T-N-H, psbD-C, psbE-F-L-J, rbcL, rpl2, rps12-rps7-ndhB. When there are two separate accessions for one region, the two accessions are connected by “&”. “—” represents missing data. The classification follows Goffinet et al. (2009). Herbarium abbreviations (Thiers, 2010): ABSH: Hepatic Herbarium, Southern Illinois University; ALA, University of Alaska; CONN, University of Connecticut; UBC, University of British Columbia.  Bryophyta: Andreaea rupestris Hedw. (Andreaeobryopsida, Andreaeaceae), Y. Chang 06-03, UBC; GU295845, GU295930, GU295893, GU295869, GU295857, GU295869, GU295919, GU295905. Andreaeobryum macrosporum Steere & B.M.Murray (Andreaeobryopsida, Andreaeobryaceae), O.M. Afonina 153, ALA; HQ412998, HQ413022, HQ413017, HQ413028, HQ412991, HQ413005, HQ413012, HQ412994. Atrichum selwynii Aust. (Polytrichopsida, Polytrichaceae), Y. Chang 06-01, UBC; GU295846, GU295931, GU295894, GU295870, GU295858, GU295870, GU295920, GU295906 & GU295907. Bryum capillare Hedw. (Bryopsida, Bryaceae), Y. Chang 07- 01, UBC; GU295850, GU295935, GU295898, GU295874, GU295862, GU295874, GU295924, —. Buxbaumia aphylla Hedw. (Bryopsida, Buxbaumiaceae), W. B. Schofield 126011A, UBC; GU295847, GU295932, GU295895, GU295871, GU295859, GU295871, GU295921, —. Dicranum scoparium Hedw. (Bryopsida, Dicranaceae), D. Woods 5-04, UBC; GU295848, GU295933, GU295896, GU295872, GU295860, GU295872, GU295922, GU295908 & GU295909. Homalothecium fulgescens (Mitt. ex C. Muell.) Lawt. (Bryopsida, Brachytheciaceae), D. Woods 05-01, UBC; GU295852, GU295937, GU295900, GU295876, GU295864, GU295876, GU295926, GU295914. Oedipodium griffithianum (Dicks.) Schwaegr. (Oedipodiopsida, Oedipodiaceae), W.B. Schofield 115800, UBC; HQ412999, HQ413023, HQ413018, HQ413029, HQ412992, HQ413006, HQ413013, —. Orthotrichum lyellii Hook. & Tayl. (Bryopsida, Orthotrichaceae), D. Woods 05-03, UBC; GU295851, GU295936, GU295899, GU295875, GU295863, GU295875, GU295925, GU295912 & GU295913. Syntrichia norvegica F. Weber  (Bryopsida, Pottiaceae), Y. Chang 06-08, UBC; GU295849, GU295934, GU295897, GU295873, GU295861, GU295873, GU295923, GU295910 & GU295911. Takakia ceratophylla (Mitt.) Grolle (Takakiopsida, Takakiaceae), W. B. Schofield 125387, UBC; GU295843, GU295928, GU295891, GU295867, GU295855, GU295867, GU295917, GU295902 & GU295903. Takakia lepidozioides Hattori & H. Inoue (Takakiopsida, Takakiaceae), Y. Chang 05-02 UBC; GU295844, GU295929, GU295892, GU295880, GU295856, GU295868, GU295918, GU295904. !! "#$!  Marchantiophyta: Haplomitrium hookeri (Sm.) Nees (Haplomitriopsida, Haplomitriaceae), D. G. Long 33741, E; HQ412996, —, HQ413015, HQ413026, —, HQ413003, HQ413010, —. Metzgeria conjugata Lindb. (Jungermanniopsida, Metzgeriaceae), Y. Chang 07-02, UBC; GU295842,  —, GU295890, GU295866, GU295854, GU295866, GU295916, —. Pleurozia purpurea Lindb. (Jungermanniopsida, Pleuroziaceae), W. B. Schofield 125173, UBC; HQ412997, HQ413021, HQ413016, HQ413027, HQ412990, HQ413004, HQ413011, —. Scapania bolanderi Aust. (Jungermanniopsida, Scapaniaceae), Y. Chang 06-05, UBC; GU295841, GU295927, GU295889, GU295865, GU295853, GU295865, GU295915, GU295901. Treubia lacunosa (Col.) Prosk. (Haplomitriopsida, Treubiaceae), Stotler & Crandall-Stotler 4561, ABSH; HQ412995, —, HQ413014, HQ413025, HQ412989, HQ413002, HQ413009, —.  Anthocerotophyta: Leiosporoceros dussii (Steph.) Hassel (Leiosporocerotopsida, Leiosporocerotaceae), Villarreal & Rodríguez 852A CONN; HQ413000, HQ413024, HQ413019, HQ413030, HQ412993, HQ413007, —, —.  Phaeoceros carolinianus (Michx.) Prosk. (Anthocerotopsida, Notothyladaceae), B. Goffinet 9652, UBC; HQ413001, —, HQ413020, HQ413031, —, HQ413008, —, —.  B: Previously published data:  Plastid genomes: Adiantum capillus-veneris L. (NC_004766), Angiopteris evecta (G. Frost.) Hoffm. (NC_008829), Anthoceros formosae Steph. (NC_004543), Chaetosphaeridium globosum (Nordst.) Kleb. (NC_004115), Chara vulgaris L. (NC_008097), Chlorokybus atmophyticus Geitler (NC_008822), Mesostigma viride Lauterborn (NC_002186), Huperzia lucidula (Michx.) Trevisan (NC_006861), Marchantia polymorpha I. (NC_001319), Psilotum nudum (L.) Beauv. (NC_003386), Physcomitrella patens (Hedw.) Bruch & Schimp (NC_005087), Pinus koraiensis Siebold & Zucc. (NC_004677), Nicotiana tabacum L. (NC_006581), and Zygnema circumcarinatum Czurda (NC_008117).  Previously published sequences: Amborella trichopoda Baill. Lilium superbum  L., Ginkgo biloba L., Gnetum gnemon L., Magnolia stellata Maxim., Nymphaea odorata Aiton, Podocarpus chinensis Wall. ex. J. Forbes., Sphagnum L. sp., and Trimenia moorei (Oliv.) Philipson. The source information and GenBank accession numbers are in Graham & Olmstead (2000a,b), Graham et al. (2002; 2006), Rai et al. (2003) and Rai & Graham (2010). !! "#"! Appendix 3. Source information and GenBank numbers for taxa included in Chapter 4. For each taxa newly sequenced, the information is displayed in the following sequence: species (classification), voucher specimen, herbarium; GenBank accession: atpB, ndhF, psbB-T-N-H, psbD-C, psbE-F-L-J, rbcL, rpl2, rps12-rps7-ndhB. When there are two separate accessions for one region, the two accessions are connected by “&”. “—” represents missing data. The classification follows Goffinet et al. (2009). Herbarium abbreviations (Thiers, 2010): ABSH, Hepatic Herbarium, Southern Illinois University; CONN, University of Connecticut; MO, Missouri Botanical Garden; UBC, University of British Columbia. For data previously published, see Appendices 1, 2. Sphagnopsida:  Sphagnum squarrosum Crome (Sphagnales, Sphagnaceae), W.B.Schofield 120039, UBC; JN162134, JN162175, JN162204, JN162233, JN162262, JN162289, JN162318, JN162163. Andreaeopsida: Andreaea nivalis Hook. (Andreaeales, Andreaeaceae), W.B.Schofield 102464, UBC; JN162135, JN162176, JN162205, JN162234, JN162263, JN162290, JN162319, —. Polytrichopsida: Alophosia azorica (Renauld & Cardot) Cardot (Polytrichales, Polytrichaceae), F.J.Rumsey 1998, UBC; JN162136, JN162177, JN162206, JN162235, JN162264, JN162291, JN162320, JN162164 & JN162165. Bartramiopsis lescurii (James) Kindb. (Polytrichales, Polytrichaceae), W.B.Schofield & S.S.Talbot 125406, UBC; JN162137, JN162178, JN162207, JN162236, JN162265, JN162292, JN162321, —. Polytrichum juniperinum Hedw. (Polytrichales, Polytrichaceae), Y.Chang 07-03, UBC; JN162138, JN162179, JN162208, JN162237, JN162266, JN162293, JN162322, —. Tetraphidopsida: Tetraphis pellucida Hedw. (Tetraphidales, Tetraphidaceaea), S. Ellis s.n., UBC; JN162139, JN162180, JN162209, JN162238, JN162267, JN162294, JN162323, JN162166 & JN162167. Tetrodontium brownianum (Dicks.) Schwägr. (Tetraphidales, Tetraphidaceaea), D.G.Long 37600, E; JN162140 JN162181 JN162210 JN162239, —, JN162295, JN162324, —. !! "#$! Bryopsida: Archidium hallii Austin (Archidales, Arcidaceae) S.P.Churchill & T. Florentin P20196, MO; JN162149, JN162190, JN162219, JN162248, JN162304, JN162333, —. Aulacomnium palustre (Hedw.) Schwägr. (Rhizogoniales, Rhizogoniaceae), W.B.Schofield & L.Boldwin 123403, UBC; JN162155, JN162196, JN162225, JN162254, JN162281, JN162310, JN162339, —. Bartramia pomiformis Hedw. (Bartramiales, Bartramiaceae), S. Ellis s.n., UBC; JN162152, JN162193, JN162222, JN162251, JN162278, JN162307, JN162336, JN162171 & JN162172,  Blindia acuta (Hedw.) Bruch & Schimp. (Grimmiales, Seligeriaceae), S. Joya 08-19-07- 1, UBC; JN162145, JN162186, JN162215, JN162244, JN162272, JN162300, JN162329, —. Bryoxiphium norvegicum (Brid.) Mitt. (Bryoxiphiales, Bryoxiphiaceae), S.S.Talbot KIS529, UBC; JN162150, JN162191, JN162220, JN162249, JN162276, JN162305, JN162334, —. Diphyscium foliosum(Hedw.) D. Mohr. (Diphysciales, Diphysciaceae), W.B.Scholfield 126011B, UBC; JN162141, JN162182, JN162211, JN162240, JN162268, JN162296, JN162325, —. Encalypta ciliata Hedw. (Encalyptales, Encalyptaceae), W.B.Schofield, J.Harpel, & Forest Service Personnel 117021, UBC; JN162143, JN162184, JN162213, JN162242, JN162270, JN162298, JN162327, —. Funaria hygrometrica Hedw. (Funariales, Funariaceae), Y.Chang 05-01, UBC; JN162144, JN162170, JN162185, JN162214, JN162243, JN162271, JN162299, JN162328. Grimmia pulvinata (Hedw.) Sm. (Grimmiales, Grimmiaceae), Y.Chang 08-02, UBC; JN162146, JN162187, JN162216, JN162245, JN162273, JN162301, JN162330, —. Hedwigia stellata Hedenås (Hedwigiales, Hedwigiaceae), S. Joya 03-26-07-1, UBC; JN162154, JN162195, JN162224, JN162253, JN162280, JN162309, JN162338, —. Hookeria lucens (Hedw.) Sm. (Hookeriales, Hookeriaceae), D.Woods 05-31, UBC; JN162159, JN162200, JN162229, JN162258, JN162285, JN162314, JN162343, JN162173. Hypnum circinale Hook. (Hypnales, Hypnaceae), Y.Chang 08-01, UBC; JN162162, JN162203, JN162232, JN162261, JN162288, JN162317, JN162346, —. !! "#$! Hypopterygium fauriei Besch. (Hookeriales, Hypopterigiaceae), K. Dillman 2005-109, UBC; JN162160, JN162201, JN162230, JN162259, JN162286, JN162315, JN162344, —. Leucobryum javense (Brid.) Mitt. (Dicranles, Leucobryaceae), Y.Chang 07-17, UBC; JN162147, JN162188, JN162217, JN162246, JN162274, JN162302, JN162331, —. Ptychomnion cygnisetum (Müll. Hal.) Kindb. (Ptychomniales, Ptychomniaceae), B.Allen 26391, UBC; JN162156, JN162197, JN162226, JN162255, JN162282, JN162311, JN162340, —. Racopilum cuspidigerum (Schwägr.) Ångström (Hypnodendrales, Racopilaceae), Y.Chang 07-20, UBC; Rhizogonium graeffeanum (Müll. Hal.) A. Jaeger (Rhizogoniales, Rhizogoniaceae), Y.Chang 07-18, UBC; XXXX, XXXX, XXXX, XXXX, XXXX, XXXX. Rhytidiadelphus loreus (Hedw.) Warnst. (Hypnales, Hylocomiaceae), D.Woods 05-05, UBC; JN162161, JN162202, JN162231, JN162260, JN162287, JN162316, JN162345, JN162174. Schistostega pennata (Hedw.) F. Weber & D. Mohr (Dicranales, Schistostegaceae), Y.Chang 07-01, UBC; JN162148, JN162189, JN162218, JN162247, JN162275, JN162303, JN162332, —. Scouleria aquatica Hook. (Scouleriales, Scouleriaceae), S. Joya 04-30-07-1, UBC; JN162151, JN162192, JN162221, JN162250, JN162277, JN162306, JN162335, —. Tetraplodon mnioides (Sw. ex Hedw.) Bruch & Schimp. (Splachnales, Splachnaceae), W.B.Schofield & S.S.Talbot 121611, UBC; JN162153, JN162194, JN162223, JN162252, JN162279, JN162308, JN162337, —. Timmia austriaca Hedw. (Timmiales, Timmiaceae), W.B.Schofield 123467, UBC; JN162142, JN162183, JN162212, JN162241, JN162269, JN162297, JN162326, JN162168 & JN162169.  !! "##! Appendix 4: Source information and GenBank numbers for taxa included in Chapter 5.  For each taxa, the information is displayed in the following sequence: species, voucher specimen, herbarium; GenBank accession: atpB, IGS atpB-rbcL, IGS atpF-atpH, ndhF, rbcL, and rpoB. “—” represents missing data Herbarium abbreviations (Thiers, 2010):  ALA, University of Alaska Museum of the North; E, Royal Botanic Garden, Edinburgh; University of Connecticut; NY, New York Botanical Garden; UBC, University of British Columbia.  Andreaea acuminata Mitt. W.B.Schofield 127663, UBC; JN388728, JN388896, JN388861, JN388765, JN388801, JN388840 Andreaea acutifolia Hook. f. & Wilson.  B.M.Murray 93-10, ALA; JN388723, JN388893, —, JN388760, JN388796, JN388835. Andreaea alpestris (Thed.) Schimp. F.M.Boas 2001-313A, ALA; JN388722, JN388892, JN388860, JN388759, JN388795, JN388834. Andreaea alpina Hedw. D.G.Long 33219, E; JN388720, JN388889, —, JN388757, JN388793, JN388831. W.R.Buck 46192, NY; JN388721, JN388890, JN388858, JN388758, JN388794, JN388832. Andreaea australis F. Muell. ex Mitt.  Streimann 53547, NY; JN388717, JN388886, —, JN388754, JN388790, JN388828. Streimann 53460, NY; JN388718, JN388887, JN388856, JN388755, JN388791, JN388829. Andreaea blyttii Schimp. W.B.Schofield 113953, UBC; JN388692, JN388862, JN388842, JN388729, JN388766, JN388802. Heegaard 399a, ALA; JN388694, —, —, JN388731, JN388768, JN388804. !! "#$! B.M.Murray 96-198, ALA; JN388693, JN388863, JN388843, JN388730, JN388767, JN388803. Andreaea crassinervia Bruch.  B.M.Murray 98-493, ALA; JN388704, JN388874, —, JN388741, JN388778, JN388815 Andreaea frigida Huebener.  B.M.Murray 98-497, ALA; JN388701, JN388871, —, JN388738, JN388775, JN388812.  D.G.Long & G.P.Rothero 36961, E; JN388702, JN388872, JN388848, JN388739, JN388776, JN388813 Andreaea fuegiana (Cardot) S.W. Greene W.R.Buck:47936, NY; JN388727 JN388895, —, JN388764, JN388800, JN388839. Andreaea heinemannii Hampe & Müll. Hal.  B.M.Murray 92-293, ALA; JN388698, JN388868, —, JN388735, JN388772, JN388809.  B.M.Murray 92-29, ALA; JN388699, JN388869, —, JN388736, JN388773, JN388810.  O’Brien 3023, ALA; JN388700, JN388870, —, JN388737, JN388774, JN388811. Andreaea megistospora B.M. Murray  W.B.Schofield 113960, UBC; JN388705, JN388875, —, JN388742, JN388779, JN388816. Andreaea microvaginata Müll. Hal.  B.M.Murray 93-7, ALA; JN388712, JN388882 —, JN388749, JN388786, JN388823.  H.Streimann 15014, NY; JN388713, —, —, JN388750, —, JN388824. Andreaea mutabilis Hook. f. & Wilson  J.Shevock 18076, ALA; JN388726, —, —, JN388763, JN388799, JN388838. Andreaea nitida Hook. f. & Wilson  W.R.Buck 47938, NY; JN388716, JN388885, —, JN388753, JN388789, JN388827. S.P.Churchill, Z.Magombo, M.Price19834, NY; JN388714, JN388883, JN388854, JN388751, JN388787, JN388825.  H.Streimann 51004, NY;!JN388715, JN388884, JN388855, JN388752, JN388788, JN388826. !! "#$! Andreaea nivalis Hook.  W.B.Schofield 120263, UBC; JN162135, JN388865, JN388845, JN162176, JN162290. D.G.Long 13547, E; JN388695, JN388864, JN388844, JN388732, JN388769, JN388805. Andreaea remotifolia Dusén  B.Allen 26668, NY; JN388725, —, —, JN388762, JN388798, JN388837. Andreaea rigida Wilson  D.G.Long 26602, E; JN388697, JN388867, JN388847, JN388734, JN388771, JN388808.  D.G.Long & J.Shevock 37179, E; JN388696, JN388866, JN388846, JN388733, JN388770, JN388807. Andreaea rothii spp. falcata (Schimp.) Lindb.  D.G.Long 27782, E; JN388707, JN388877, JN388851, JN388744, JN388781, JN388818.  B.M.Murray 98-480, ALA; JN388708, JN388878, —, JN388745, JN388782, JN388819. Andreaea rothii spp. rothii F. Weber & D. Mohr.  B.M.Murray 98-485, ALA; JN388709, JN388879, —, JN388746, JN388783, JN388820.  W.B.Schofield 124669, UBC; JN388706, JN388876, JN388850, JN388743, JN388780, JN388817. Andreaea rupestrisi Hedw.  Y.Chang 06-03, UBC; GU295930, JN388891, JN388859, GU295893, GU295869, JN388833. B.M.Murray 10-01, JN388724, JN388894, —, JN388761, JN388797, JN388836 Andreaea schofieldiana B.M.Murray  W.B.Schofield 112012, UBC; JN388703, JN388873, JN388849, JN388740, JN388777, JN388814. Andreaea subulata Harv.  T.A.Hedderson 13568, UBC; JN388710, JN388880, JN388852, JN388747, JN388784, JN388821.  W.R.Buck 45663, NY; JN388711, JN388881, JN388853, JN388748, JN388785, JN388822. !! "#$! Andreaea wilsonii Hook. f.  B.Allen 26335, NY; JN388719, JN388888, JN388857, JN388756, JN388792,JN388830. Andreaeobryum macrosporum Steere & B.M.Murray.  O.M. Afonina 153, ALA; HQ412998, —, JN388841, HQ413022, HQ413005, —.

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.24.1-0072202/manifest

Comment

Related Items