Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Ecological genomics of invasive thistle, diffuse knapweed Turner, Kathryn Grace 2015

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

Item Metadata

Download

Media
24-ubc_2015_september_turner_kathryn.pdf [ 2.23MB ]
Metadata
JSON: 24-1.0166509.json
JSON-LD: 24-1.0166509-ld.json
RDF/XML (Pretty): 24-1.0166509-rdf.xml
RDF/JSON: 24-1.0166509-rdf.json
Turtle: 24-1.0166509-turtle.txt
N-Triples: 24-1.0166509-rdf-ntriples.txt
Original Record: 24-1.0166509-source.json
Full Text
24-1.0166509-fulltext.txt
Citation
24-1.0166509.ris

Full Text

ECOLOGICAL GENOMICS OF INVASIVE THISTLE, DIFFUSE KNAPWEED by  Kathryn Grace Turner  B.S., The University of Texas, 2005 B.A., The University of Texas, 2005  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Botany)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August 2015  © Kathryn Grace Turner, 2015  ii  Abstract  Invasive species are able to move into new environments, with new abiotic conditions and biotic interactions, survive, and even dominate, often within a hundred years. Much research in invasion biology has assessed whether populations of invasive plants show phenotypic or evolved genetic differences in growth and reproduction compared to their native range (such as the evolution of increased competitive ability hypothesis), in an effort to understand the causal drivers of invasion. Using diffuse knapweed (Centaurea diffusa) I assessed evolution in its invaded range in multiple ways. Chapter 2 describes two greenhouse common garden experiments that evaluated phenotypic and life history trait variation between the two ranges under benign and stressful conditions (drought, flooding, nutrient deficiency, and herbivory). Invasive individuals grew larger and flowered later in benign conditions, and performed as well or better under most of the tested stress conditions than native individuals. The strongest evidence for a trade-off in tolerance was exhibited under drought conditions, but only among native populations. Chapter 3 employs a field common garden to compare phenotypes, drought response, and adaptation to environmental conditions in a more natural setting. This study incorporates a large dataset of occurrence locations to look at the different relationships that populations in the two ranges have to their bioclimatic environments. I found that invasive C. diffusa individuals were larger, matured later, and have lost adaptation to environmental conditions apparent in native populations. More plastic invasive genotypes may have expanded the climatic niche inhabited in the invaded range. Chapter 4 attempts to identify a genetic mechanism underlying these phenotypic changes by comparing gene expression between the two  iii  ranges under benign and drought conditions. Genes were identified whose expression either varied constitutively or responded to drought stress differently between ranges. Based on these data, invasive populations may have constitutively higher levels of energy production, while native populations have a stronger cellular drought defense. This dissertation presents ample evidence of evolution in the invaded range and suggests that plasticity and rapid evolution had a significant impact on the successful invasion of North America by C. diffusa.  iv  Preface  I designed the study in Chapter 2, in collaboration with L.H. Rieseberg. The following people contributed to seed collections: R.A. Hufbauer, A. Blair, R. Marrs, D. Woods, P. Mraz, A. Shipunov, M. Bona, R. Sforza, O. Korniyenko, A. Norton, A. Guggisberg, K. Dlugosch, M. King, A. Stephens, J. Myers. I conducted and analyzed the greenhouse experiments and wrote the manuscript. L.H. Rieseberg and R.A. Hufbauer contributed to interpreting results and framing the manuscript. A version of Chapter 2 has been published:  Turner KG, Hufbauer RA, Rieseberg LH (2014) Rapid evolution of an invasive weed. New Phytologist, 202, 309–321. Chapter 3 was designed by myself and L.H. Rieseberg. With H. Freville, I organized and implemented the field experiment and collected data. I organized and implemented the greenhouse experiment, conducted the statistical analyses, and wrote the manuscript. L.H. Rieseberg and H. Freville contributed to interpreting results and framing the manuscript. A version of Chapter 3 has been accepted for publication.  Turner KG, Freville H, Rieseberg LH. Adaptation and niche expansion in an invasive thistle. Ecology and Evolution (accepted) ECE-2015-03-00175 . Chapter 4 was designed by myself and L.H. Rieseberg. I conducted the greenhouse experiment, produced the cDNA libraries, conducted the statistical analyses, and wrote the manuscript. L.H.  v  Rieseberg contributed to interpreting results and framing the manuscript. A manuscript of this gene expression analysis is in preparation. Genomic resources described in Appendix D were the result of several collaborative efforts. I produced two transcriptome libraries and, and with N.C. Kane and Z. Lai, assembled the resulting sequence and designed the microarray utilized in Chapter 4. The development of these microarrays has been published and is described in:  Lai Z, Kane NC, Kozik A et al. (2012) Genomics of Compositae weeds: EST libraries, microarrays, and evidence of introgression. American Journal of Botany, 99, 209–218. Gene annotations for the Centaurea diffusa transcriptomes utilized in Chapter 4 were annotated as part of a larger collaborative study by K.A. Hodgins, D.G. Bock, M.A. Hahn, S.M. Heredia, L.H. Rieseberg, and myself. This work has been published:  Hodgins KA, Bock DG, Hahn MA, Heredia SM, Turner KG, Rieseberg LH (2015) Comparative genomics in the Asteraceae reveal little evidence for parallel evolutionary change in invasive taxa. Molecular Ecology, DOI: 10.1111/mec.13026. I produced a whole genome sequence library and, with C.J. Grassa, assembled and annotated the plastid genome. This work is available as a pre-print:  Turner KG, Grassa CJ (2014) Complete plastid genome assembly of invasive plant Centaurea diffusa. bioRxiv, doi:10.1101/005900.  vi  Table of Contents  Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iv Table of Contents ......................................................................................................................... vi List of Tables ............................................................................................................................... xii List of Figures ............................................................................................................................. xiii Acknowledgements ......................................................................................................................xv Dedication .................................................................................................................................. xvii Chapter 1: Introduction ................................................................................................................1 1.1 Background: Evolution and invasive species ................................................................. 1 1.1.1 Phenotypic differentiation ........................................................................................... 2 1.1.2 Post-introduction evolution ........................................................................................ 2 1.1.3 Underlying genetic causes .......................................................................................... 5 1.2 Study system ................................................................................................................... 6 1.3 Research questions .......................................................................................................... 9 Chapter 2: Rapid Evolution of an Invasive Weed ....................................................................12 2.1 Introduction ................................................................................................................... 12 2.2 Materials and methods .................................................................................................. 15 2.2.1 Study species ............................................................................................................. 15 2.2.2 Seed collections ......................................................................................................... 17 2.2.3 Broad common garden .............................................................................................. 18  vii  2.2.4 Maternal common garden ......................................................................................... 20 2.2.5 Statistical analysis .................................................................................................... 21 2.3 Results ........................................................................................................................... 23 2.3.1 Broad common garden .............................................................................................. 23 2.3.2 Maternal common garden ......................................................................................... 29 2.3.3 Cross-generational analysis ..................................................................................... 33 2.4 Discussion ..................................................................................................................... 35 2.4.1 Evolution in the invaded range ................................................................................. 35 2.4.2 Test of EICA and trade-offs hypotheses .................................................................... 37 2.4.3 Conclusions ............................................................................................................... 40 Chapter 3: Phenotypic Plasticity and Niche Expansion in an Invasive Thistle .....................41 3.1 Introduction ................................................................................................................... 41 3.2 Materials and methods .................................................................................................. 45 3.2.1 Study species ............................................................................................................. 45 3.2.2 Populations ............................................................................................................... 46 3.2.3 Common garden experiment ..................................................................................... 47 3.2.4 Statistical analysis .................................................................................................... 48 3.2.5 Occurrence data and principal components analysis ............................................... 51 3.3 Results ........................................................................................................................... 52 3.3.1 Principal components analysis of experimental populations ................................... 52 3.3.2 Phenotypic differentiation ......................................................................................... 53 3.3.3 Drought response and resource allocation trade-offs .............................................. 57 3.3.4 Evidence of niche expansion ..................................................................................... 57  viii  3.4 Discussion ..................................................................................................................... 59 3.4.1 Phenotypic differentiation ......................................................................................... 59 3.4.2 Drought response and resource allocation trade-offs .............................................. 61 3.4.3 Environmental variation, plasticity, and niche expansion........................................ 62 3.4.4 Hybridization and phenotypic plasticity ................................................................... 64 3.4.5 Conclusion ................................................................................................................ 65 Chapter 4: Differential Gene Expression with Drought in Centaurea diffusa .......................67 4.1 Introduction ................................................................................................................... 67 4.2 Materials and methods .................................................................................................. 70 4.2.1 Study species ............................................................................................................. 70 4.2.2 Seed collection and production ................................................................................. 71 4.2.3 Treatment and sampling ........................................................................................... 72 4.2.4 RNA extraction .......................................................................................................... 73 4.2.5 cDNA preparation and hybridization ....................................................................... 73 4.2.6 Preprocessing and analysis ...................................................................................... 74 4.3 Results ........................................................................................................................... 78 4.3.1 Differential expression .............................................................................................. 78 4.3.2 Putative function of differentially expressed genes .................................................. 82 4.4 Discussion ..................................................................................................................... 88 4.4.1 Variation in gene expression .................................................................................... 89 4.4.2 Functional analysis ................................................................................................... 90 Chapter 5: Conclusion .................................................................................................................96 Concluding remarks ................................................................................................................ 101  ix  References ...................................................................................................................................102 Appendices ..................................................................................................................................119 Appendix A Chapter 2 Supplementary Material .................................................................... 119 A.1 Germination results of field collected Centaurea diffusa seed. .............................. 119 A.2 Native Centaurea diffusa population information. ................................................. 120 A.3 Invasive Centaurea diffusa population information. .............................................. 121 A.4 Test statistics for all traits measured in broad common garden, from range differentiation models of phenotype of Centaurea diffusa. ................................................ 122 A.5 Means and confidence intervals for all traits estimated from range differentiation models of phenotypes of Centaurea diffusa grown in a common garden. ......................... 123 A.6 Means and confidence intervals for all traits estimated from range differentiation models of phenotypes of Centaurea diffusa grown in a common garden, after one generation. ........................................................................................................................... 124 A.7 Test statistics for all traits measured in maternal common garden from range differentiation models of phenotypes of Centaurea diffusa grown in a common garden, after one generation in the glasshouse......................................................................................... 125 A.8 Tests of slope differences in origin-by-control treatment performance (control shoot mass) terms from explicit trade-off models of Centaurea diffusa grown in a common garden. ……………………………………………………………………………………..126 A.9 Effect of origin, latitude, and constitutive defense on herbivore preference of Centaurea diffusa grown in a common garden across two generations. ............................ 127 A.10 Tests of slope difference in origin-by-latitude terms from range differentiation models of Centaurea diffusa grown in a common garden. ................................................. 128  x  Appendix B Chapter 3 Supplementary Material ..................................................................... 129 B.1 Additional figures from the principal components analysis of environmental variables for C. diffusa populations used in the field experiment....................................... 129 B.2 Additional examples of morphological and stress tolerance trait divergence among Centaurea diffusa in the field experiment for traits with significant origin or origin-by-environment terms in range differentiation models. ........................................................... 130 B.3 Principal components analysis of environmental variables for all C. diffusa geo-referenced occurrences recorded in GBIF.org. ................................................................... 132 B.4 Principal components analysis of environmental variables for a subset of C. diffusa geo-referenced occurrences recorded in GBIF.org. ............................................................ 133 B.5 Centaurea diffusa experimental population information. ....................................... 135 B.6 Abiotic environmental data variables used in principal components analyses (Hijmans et al., 2005).......................................................................................................... 136 B.7 Test statistics for all traits measured in the Montpellier field common garden from range differentiation models of phenotype of Centaurea diffusa. ...................................... 137 B.8 Parameter estimates of fixed effects from range differentiation models of Centaurea diffusa grown in Montpellier common field experiment that included a significant origin term or significant interaction between origin and environment. ....................................... 138 Appendix C Chapter 4 Supplementary Material ..................................................................... 140 C.1 Heat map of genes with drought-induced expression differences between native and invasive Centaurea diffusa for the drought treatment from time point 2. .......................... 140 C.2 Heat map of genes with drought induced expression differences between native and invasive Centaurea diffusa for the control treatment from time point 2. ........................... 141  xi  C.3 Centaurea diffusa experimental population information. ....................................... 142 C.4 Results from likelihood ratio tests for genes with significant model terms. ........... 142 C.5 Gene Ontology terms for genes with a significant effect of drought treatment...... 143 C.6 Gene Ontology terms for gene with a significant effect of seed collection environment (PC1). ............................................................................................................. 144 Appendix D Genomic Resources for Centaurea diffusa ........................................................ 145 D.1 Centaurea diffusa EST libraries.............................................................................. 145 D.2 Microarray developed for Centaurea diffusa. ......................................................... 145 D.3 Gene Ontology terms overrepresented for rapidly evolving genes identified in invasive Centaurea diffusa. ................................................................................................ 146 D.4 Centaurea diffusa plastome assembly output numerics.......................................... 146 D.5 Annotated plastome for Centaurea diffusa. ............................................................ 147   xii  List of Tables  Table 2.1 Selected examples of means and confidence intervals estimated from range differentiation models of phenotypes of Centaurea diffusa grown in both common garden experiments. .................................................................................................................................. 24 Table 2.2 Test statistics from range differentiation models for selected phenotypic traits of Centaurea diffusa. ......................................................................................................................... 25 Table 2.3 Explicit trade-off models of phenotypes of Centaurea diffusa grown under stressful conditions in both common garden experiments. ......................................................................... 29 Table 2.4 Effect of origin and generation on phenotypes of Centaurea diffusa grown in a common garden, across two generations. ..................................................................................... 34 Table 3.1 Test statistics from range differentiation models of phenotypic measurements of Centaurea diffusa, for all traits measured in the field experiment with a significant origin or origin-by-environment term. ......................................................................................................... 55 Table 4.1 Gene Ontology terms for genes with a significant effect of origin. ............................. 83 Table 4.2 Gene Ontology terms for genes with significant origin by treatment interaction. ....... 85   xiii  List of Figures  Figure 2.1 Range and collection map of Centaurea diffusa in the northern hemisphere, by country. ......................................................................................................................................... 16 Figure 2.2 Phenotypic difference between ranges of Centaurea diffusa in broad common garden for control and all stress treatments that survived until harvest. ................................................... 26 Figure 2.3 Variation in stress tolerance and performance in benign conditions between ranges of Centaurea diffusa in broad common garden. ................................................................................ 27 Figure 2.4 Phenotypic difference between ranges of Centaurea diffusa in maternal common garden for control and all stress treatments that survived until harvest. ....................................... 31 Figure 2.5 Phenotypic differences between ranges of Centaurea diffusa across generations, for four native and four invasive populations in the control treatment. ............................................. 33 Figure 3.1 Range and population map of Centaurea diffusa in the Northern Hemisphere, by country, used in the field experiment. ........................................................................................... 46 Figure 3.2 Principal components analysis of abiotic environmental variables of sampling locations of experimental populations of Centaurea diffusa. ....................................................... 53 Figure 3.3 Selected examples of size and life history trait divergence among Centaurea diffusa ranges in the Montpellier experiment. .......................................................................................... 55 Figure 3.4 Putative climatic niche expansion as determined by principal components analysis of occurrence data in the native and invasive ranges of Centaurea diffusa. ..................................... 58 Figure 4.1 Heat map of genes with constitutive expression differences between native and invasive Centaurea diffusa. .......................................................................................................... 79  xiv  Figure 4.2 Heat map of genes with drought induced expression differences between native and invasive Centaurea diffusa. .......................................................................................................... 81    xv  Acknowledgements I would like to thank my advisor, L. Rieseberg, for building this lovely sandbox for me to play in; without this setting, these resources, these co-explorers, and his calming guidance, none of this work would have been possible. I would like to thank Loren for hiring me as his lab manager, sight unseen, for pursuing me as a graduate student, and for appearing confident throughout that I would manage it. I would like to thank my committee members, S. Otto, G. Crutsinger, and for a time, M. Vellend, for their encouragement, helpful advice, and rueful commiseration, as appropriate for the circumstance.  I would like to thank all the helpful people that have explained complex things to me in small words over the course of my dissertation. These wonderful people include A. Stephens, R. Fitzjohn, K. Samuk, D. Huang, H. Rowe, J. Chang, G. Conte, A. Guggisberg, C. Sears, W. Iles, H. Macfarlane, N. Kane, J. Ono, K. Dlugosch, R. Hufbauer, H. Freville, M. Stewart, and all the helpful folk of the Rieseberg lab and UBC Biodiversity. I would especially like to thank the RGSD (H. Dempewolf, G. Owens, G. Baute, C. Grassa, E. Drummond, D. Bock) and my lab BFFs (B. Moyers and K. Ostevik), without whom I would have gone crazy long ago. And finally I would like to thank my family and friends, near and far, for being excited when I was excited, optimistic when I was depressed, and for sharing a beer with me, when that is what needed to happen. I would like to thank my parents for fanning the flames of my curiosity, and for their unwavering certainty that all the best things would happen to me. I would like to thank my brother for keeping my wits sharp, and for being proud of me. I would like to thank Robson for being a perfectly normal human worm baby.  xvi  Specific acknowledgments for Chapter 2: I thank K. Nurkowski, E. Drummond, A. Cang, S. Lin, and V. Turner, for plant care and data collection help; for seed collections: A. Guggisberg, R. Sforza, A. Shipunov, O. Korniyenko (Europe); M. King, A. Blair, R. Marrs, D. Woods, A. Norton (North America); A. Stephens and J. Myers for weevil collection; A. South for map help; R. Colautti, A. MacDonald, and K. Hodgins for analysis help. Funding was provided by a National Science Foundation Graduate Research Fellowship and Natural Sciences and Engineering Research Council of Canada grant No. 353026. Specific acknowledgments for Chapter 3: I thank C. Devaux, A. Mignot, I. Olivieri, N. Espuno, S. Hello for plant care and data collection help; T. Mathieu, D. Deguedre for technical support at CEFE, Plateforme des Terrains d’expériences du Labex CeMEB; for seed collections, A. Guggisberg, R. Hufbauer, A. Blair, A. Shipunov, P. Mráz (Europe); M. King and A. Stephens (North America). This material is based upon work supported by the National Science Foundation, Grant No. 0541673. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSF. Additional funding provided by: NSF Graduate Research Fellowship to KGT, Natural Sciences and Engineering Research Council of Canada grant No. 353026 to LHR, and University of Montpellier SIBAGHE visiting scholar grant to KGT. Specific acknowledgments for Chapter 4: I thank K. Nurkowski for plant care, data collection, and cDNA library preparation help; A. Guggisberg and K. Hodgins for technical and experimental guidance; J. Huang for collecting the array hybridization data. Funding provided by: NSF Graduate Research Fellowship to KGT, Natural Sciences and Engineering Research Council of Canada grant No. 353026 to LHR.   xvii  Dedication     To my parents, who taught me be to be equal parts annoying and stubborn curious and persistent.   Plant your hope with good seeds, don’t cover yourself with thistle and weeds.   1  Chapter 1: Introduction  1.1 Background: Evolution and invasive species Invasive species, here defined as widespread non-indigenous species, are neither a recent nor a trivial problem. Invasive species pose a major threat to the biodiversity and health of ecosystems, rivaling habitat loss in their destructive effects (Sakai et al. 2001). They can lead to extirpations and lower biodiversity both within and between sites. The direct costs of invasive plant control (largely herbicides) and indirect costs of reduced crop production due to a handful of the most problematic weeds are as much as $35 billion annually in North America (Pimentel 2005). They therefore provide ample motivation to understand the processes and circumstances that enable invasion. In addition to a pragmatic impetus, invasive species also provide ecologists and evolutionary biologists with tractable systems in which to address some of the major questions in those fields. Invasions represent the ongoing assembly of novel biotic and abiotic interactions, as well as the evolutionary results of those interactions, unfolding over historical time scales, often replicated at many sites. These systems can yield valuable insights into the characteristics of species that successfully dominate or alter ecosystems, the genetic changes that underlie these characteristics, and the speed of adaptation to environmental change.  2  1.1.1 Phenotypic differentiation Despite constant human mediated transport, not all species that are transported to new habitats become established, and relatively few become widespread or ecologically disruptive in their new habitat (Williamson 1993).  The search for traits predictive of new problematic species has been extensive but not entirely conclusive. Baker (1965) described an “ideal weed” and included such characteristics as wide environmental tolerances and flexibility of pollination syndrome. Yet species lacking many of the traits listed by Baker and others may still be invasive, and which traits are beneficial may vary according to the stage of the invasion (Sakai et al. 2001; vanKleunen et al. 2010; Moles et al. 2012). Observing phenotypes over a single generation of native and invasive populations even in a common environment may be insufficient to demonstrate evolution in a novel habitat. Maternal environmental effects can have strong, even adaptive, impacts on offspring phenotype in some systems (Galloway 2005). Yet their influence on performance differences between native and invaded ranges is rarely experimentally controlled (except see Monty et al. 2009; Hodgins & Rieseberg 2011). Furthermore, increased size in invasive populations can be confounded by latitudinal origins or climate effects, factors that should be considered when comparing life-history traits of native and introduced populations (Colautti et al. 2009; Rypel 2014). 1.1.2 Post-introduction evolution Many successful biological invasions may require invasive species to adapt to novel contexts. Both observational and experimental studies have documented adaptive changes in invasive populations relative to native populations (Dlugosch & Parker 2008; Felker-Quinn et al. 2013;  3  Prentis et al. 2008; Whitney & Gabler 2008). Rates of adaptive phenotypic change may be higher in invasive species, relative to native species in the same environment, particularly in anthropogenic contexts (Hendry et al. 2008). Though not universal, increased growth rate or reproductive capacity is frequently reported from field observations and increasingly from common garden experiments of invasive populations (Elton 1958; Crawley 1987; Thebaud & Simberloff 2001; Parker et al. 2013; Felker-Quinn et al. 2013). This improved fecundity could contribute to rapid spread and population growth in the invaded range.  Some of the strongest evidence supporting contemporary evolution and post-invasion adaptation in invasive species is demonstrated by the recapitulation of clinal variation. Comparative studies of native and invasive populations that demonstrate parallel variation along geographic clines range from physiological adaptation to changes in life history traits. As invading populations adjust to local environments along latitudinal and environmental clines, adaptation can shift phenology, biomass, and other trait means (Colautti et al. 2010).  The causal drivers of increased performance of invasive populations relative to their native con-specifics in terms of size, fecundity, and competitive ability have been the subject of multiple hypotheses, several of which invoke the evolution of trade-offs between self-defense, growth and reproduction. All trade-off hypotheses are based on the assumption that organisms are unable to be both 1) highly competitive or have high reproductive output and 2) be highly tolerant of stressful conditions (Grime 1977). The most studied hypothesis in invasive plant species is the evolution of increased competitive ability (EICA) that posits that, in the absence of specialist herbivores, selection will favor genotypes with reduced resource allocation to herbivore defense and elevated reproductive output or competitive ability (Blossey & Nötzold 1995; Joshi &  4  Vrieling 2005). However, trade-offs in invasive plant species may not be limited to herbivore defense only, but may include tolerance to stressful abiotic conditions (Bossdorf et al. 2005; He et al. 2010). If novel habitats are less stressful than native ones, selection would favor individuals that shifted resources from stress tolerance to increased vigor and fecundity. Agricultural weeds may represent a simple example of this; in many ways, cultivated fields represent dramatically less stressful habitats than native habitats to many plants. Invasive individuals may evolve a increased competitive ability, vigor and/or fecundity predicated on lower tolerance to biotic or abiotic stress, and therefore will perform relatively poorly under stressful conditions (Hodgins & Rieseberg 2011; Lachmuth et al. 2011; Kumschick et al. 2012).  Unfortunately, detecting trade-offs can be complex and depend on testing performance under the appropriate stressors. Trade-offs may occur in multiple directions, including dispersal or competitive ability rather than a simple two-way relationship between performance and defense or tolerance (Burton et al. 2010).  Rather than specialize in local novel environments through adaptation, invasive species may instead benefit from generalist strategies; a phenotypically plastic, ‘general-purpose’ genotype may perform well in many novel environments (Baker, 1965). Adaptive phenotypic plasticity, whereby a genotype can maintain fitness in stressful or unfavorable environments and/or increase fitness in favorable environments by adjusting its phenotype (Richards et al. 2006), has played an important role in many invasions (Hahn et al. 2012; Zenni et al. 2014b; Bock et al. 2015). Theory suggests that stable environments will favor local adaptation, whereas frequent, especially anthropogenic, disturbance will rapidly select for plastic genotypes in either the native or invaded ranges, well suited for colonizing novel habitats (Meyers et al. 2005; Lande 2009; Hufbauer et al. 2012). By exposing populations to extreme environmental changes, the process  5  of invasion itself can result in the rapid evolutionary increase of plasticity in the early stages of invasion, though this benefit is typically transient (Lande 2009; Lande 2015). Phenotypic plasticity may therefore enable the expansion of the environmental tolerances realized in the invaded range of a species. 1.1.3 Underlying genetic causes The potential molecular mechanisms that underlie the success of invasive plants are actively being investigated by invasion biologists. Considerable effort has focused on the genetics of phenotypic traits that may enable invasiveness, including fast growth, high seed set, and phenotypic plasticity in crop plants and their weedy or wild relatives (Basu et al. 2004, Stewart et al. 2009). Ecological genomic approaches offer the ability to uncover the molecular genetic changes that have led to the evolution of complex traits, such as those associated with invasiveness (Stewart et. al 2009).  The genetics of invasive species may be shaped by many forces. Colonization events commonly result in losses of genetic variation, or bottlenecks, potentially limiting population growth and persistence, though significant amounts of genetic variation can sometimes be maintained via multiple introductions (Sakai et al. 2001; Dlugosch & Parker 2008). Hybridization may have an important role in the success of some biological invasions (Ellstrand & Schierenbeck 2000; Rieseberg et al. 2007). Inter- and intraspecific hybridization has been proposed to stimulate invasiveness through the generation of novel genotypes by recombination, potentially generating heterosis, increased fitness, and/or decreased genetic load (Schierenbeck & Ellstrand 2009).  To identify which of these forces  are most influential and the genetic mechanisms underlying invasion, genomic tools are necessary (Stewart et al. 2009). Genome-scale transcriptional  6  profiling can identify loci that are differentially expressed between native and invasive genotypes and suggest hypotheses to explain phenotypic variation, physiological trade-offs, and the origin of diversity, biological novelty, and adaptation (Whitehead & Crawford 2006; Lai et al. 2008; Hodgins et al. 2013).   1.2 Study system The sunflower family (Asteraceae) is arguably the largest angiosperm family and includes many well known weedy species, such as thistles, dandelions, and ragweed. The genus Centaurea, containing starthistles and knapweeds, has contributed 30 non-native species to North America, 11 of which are noxious weeds (USDA NCRS 2013) and is one of only 15 plant genera in the United States containing more weedy species than expected by chance (Kuester et al. 2014).  The five Centaurea species with the greatest impact, including C. diffusa (diffuse knapweed), have invaded millions of hectares of grassland, making it the most abundant noxious weed genus in the western United States (LeJeune & Seastedt 2001). Knapweed invasions can form dense monocultures that reduce the quality of forage for livestock and wildlife and alter soil and water resource availability (Sheley & Larson 1996). These impacts occur partially through competing with native species for limiting phosphorus in disturbed grasslands (Sheley 1998, LeJeune & Seastedt 2001). Centaurea diffusa is a tap-rooted, rosette-forming, facultative biennial with deeply lobed leaves (Thompson & Stout 1991). At maturity, once a threshold root crown diameter is reached, the plant bolts (typically one stem) and flowers (Powell & Myers 1988). An obligate out-crosser, C. diffusa is primarily pollinated by insects, predominantly the European honey bee, Apis mellifera (Harrod & Taylor 1995).  7  Native to parts of eastern Europe and western Asia, specifically the area surrounding the Black and Caspian Seas, C. diffusa is considered a naturalized alien throughout western Europe as well (Greuter 2009; Bleeker et al. 2007). Centaurea diffusa was first reported in North America in 1907 as a Bulgarian alfalfa seed contaminant in an agricultural field near Bingen, Washington (Sheley 1998). It then spread, initially as an alfalfa and hay seed contaminant, across western North America (Sheley 1998). It now occurs in six Canadian provinces and 25 of the United States (see Fig. 2.1, USDA NRCS 2013).  Patterns of chloroplast haplotype diversity suggest that C. diffusa has been introduced to North America at least twice, including once from Turkey (Hufbauer & Sforza 2008). Chloroplast haplotype diversity in native populations of C. diffusa is significantly higher than in invasive populations, and chloroplast intron allelic richness is more than double (Hufbauer & Sforza 2008), a common pattern for invasive species (Dlugosch & Parker 2008, Bossdorf et al. 2005). In contrast, polymorphic nuclear microsatellite markers for C. diffusa indicate that, although significantly more unique alleles are present in the native range of C. diffusa, allelic richness did not differ between the native and invaded ranges (Marrs et al. 2008). This study also found a trend toward higher levels of heterozygosity in the invaded range, suggesting that C. diffusa has not suffered from a genetic bottleneck, possibly due to multiple introductions (Marrs et al. 2008). Additionally, native populations show little population structure and are nearly panmictic (Marrs et al. 2008). A comparison of divergence between native and invasive C. diffusa transcriptomes in coding sequence at nonsynonymous and synonymous sites identified 104 genes as quickly evolving in the invaded range, and associated these genes with nine Gene Ontology terms, relating to cell division, translation, and photosynthesis (Appendix D; Hodgins et al. 2015).  8  Centaurea is a large genus, containing roughly 300 species, which have traditionally resisted taxonomic resolution (Garcia-Jacas et al. 2006), at least partially due to hybridization and introgression within the genus (Suarez-Santiago et al. 2007). Centaurea diffusa has been hypothesized to hybridize with the diploid form of spotted knapweed, or C. stoebe subsp. stoebe, supported largely by observations of co-occurrence of both species in some parts of the native range and individuals of intermediate floral morphology present in both ranges (Blair & Hufbauer 2009; Ochsmann 2001; Gayer 1909). Centaurea stoebe s.l. has both diploid and tetraploid forms in the native range, and typically has larger flowering heads with up to twice as many flowers per head, purple corollas, and a darkly pigmented spot on each bract, compared to C. diffusa (Ochsmann 2001, Sheley 1998). The tetraploid form (C. stoebe subsp. micranthos [Gugler] Hayek, historically referred to as C. maculosa) is another highly successful invader of North America, yet the diploid C. stoebe subsp. stoebe is absent from this range (Treier et al. 2009; Blair & Hufbauer 2010).  In contrast, C. diffusa is almost exclusively diploid, typically lacks pigmentation, and has a terminal spine on each bract, as well as smaller flowering heads, and white flowers (Ochsmann 2001). Introgression between C. diffusa and C. stoebe subsp. stoebe was confirmed by amplified fragment length polymorphism (AFLP), chloroplast haplotype, and transcriptome comparisons (Hufbauer & Sforza 2008; Blair & Hufbauer 2010; Blair et al. 2012; Lai et al. 2012; Mráz et al. 2012a). Results of these analyses indicate that roughly a quarter of C. diffusa individuals at North American invaded sites have a hybrid ancestry (Blair & Hufbauer 2010) and that the extent of introgression may be greater in invasive rather than native C. diffusa (Lai et al 2012).  Interestingly, introgression in the invaded range has been de-coupled from intermediate floral  9  morphology, perhaps the result of extensive  backcrossing between hybrid and ‘pure’ C. diffusa individuals in North America (Blair & Hufbauer 2010). Given the lack of diploid spotted knapweed in the invaded range, these data suggest that genotypes of C. diffusa with hybrid ancestry have colonized North America. Field surveys reveal that herbivores impact C. diffusa differently between the native and invaded range, with North American plants experiencing more seedhead herbivory, and European plants experiencing more root herbivory (Blair et al. 2008). Since 1970, 12 species of specialist insects have been introduced to western North America for biological control of C. diffusa, four of which are now widespread and abundant (Myers et al. 2009). Larinus minutus (Col.: Curculionidae) is the specialist weevil associated with the largest decline in knapweed density and the most recently introduced biocontrol agent. A decline in C. diffusa density following the establishment of L. minutus has been documented at several locations (Myers et al. 2009, Seastedt et al. 2005). Adult weevils overwinter in the soil at the root of the plant and emerge in spring to feed on the leaves, stems, and buds of the plant. Larvae feed on the developing seeds, and emerge as adults in late summer. Because L. minutus feeds on all life stages of C. diffusa, and both floral and leaf tissue, it can dramatically reduce seed production through both direct seed predation and damage to the vegetative and reproductive tissues of adult plants (Myers et al. 2009; Stephens & Myers 2013).  1.3 Research questions To understand the contribution of evolution to the invasion success of a weedy plant species, I asked the following questions:   10  Question 1: Do more invasive genotypes evolve in the invaded range of an invasive species?  As reported in other invasive plant species, I predicted that Centaurea diffusa from the invaded range would be larger, more robust, and have a higher reproductive output than plants from the native range in a benign common environment and that these differences would have a genetic basis. In Chapters 2 and 3 of this dissertation, I examine phenotypic differences between populations from the native range and the invaded range, using common garden experiments. Results indicate whether invasive populations of C. diffusa have higher growth rates than native populations. In Chapter 2 I describe two greenhouse common garden experiments that evaluate phenotypic and life history trait differences between the native and invaded ranges of C. diffusa under both benign common conditions. In Chapter 3, I employ a field common garden to compare phenotypes in a more realistic setting. Question 2: Do trade-offs between stress tolerance and growth/reproduction enhance the invasiveness of a species? I predicted that invasive Centaurea diffusa individuals would be less tolerant to some stressor characteristic of the native range than native individuals, though this stress might not be specialist herbivory, as predicted by the evolution of competitive ability hypothesis. In Chapter 2, I test for rapid evolution of stress-tolerance trade-offs by comparing growth rate in the presence or absence of experimentally applied stresses. Results indicate whether individual growth rate in invasive populations of C. diffusa is correlated with stress tolerance. In Chapter 2 I describe two greenhouse common garden experiments that evaluate phenotypic differences between the ranges of C. diffusa under stressful conditions, including drought, nutrient deficiency, and herbivory. In Chapter 3, I employ a field common garden to compare drought  11  stress response in a more realistic setting. In Chapter 4, I identify genes whose expression responds to drought stress differently between ranges.  Question 3: Do native and invasive individuals of a species relate differently to the environmental conditions they experience? I predicted that invasive Centaurea diffusa had rapidly evolved to reestablish similar relationships to environmental conditions apparent in the native range, as seen in some other invasive species. In Chapter 3,  I compare phenotypes in a common environment to the abiotic environment of each seed collection location, using 18 bioclimatic variables, altitude, and latitude, to assess adaptation to environmental conditions. This study incorporates a large dataset of occurrence locations from the native and invaded ranges of C. diffusa to look at the different relationships that populations in the two ranges have to their bioclimatic environments.  Question 4: What are the genetic mechanisms that underlie phenotypic and stress tolerance differences between native and invasive individuals? I predicted that native and invasive Centaurea diffusa populations would have both constitutive differences in gene expression and differences induced by drought stress and that these differences might underlie range-level differences in drought stress tolerance. In Chapter 4, I attempt to identify a genetic mechanism underlying these phenotypic changes by comparing gene expression profiles between the native and invaded ranges under benign and drought conditions. I identify genes whose expressions either vary constitutively or respond to drought stress differently between ranges. This work necessitated the development of genomic resources for this non-model system, some of which is still ongoing, (see Appendix D for a brief description).   12  Chapter 2: Rapid Evolution of an Invasive Weed  2.1 Introduction Invasions represent excellent opportunities to synthesize evolution, ecology, and genetics over contemporary time scales, and especially to address how human-induced changes in the environment and species distributions influence ecological and evolutionary processes. While extensive research focuses on the ecological processes that play a role in plant invasions, an understanding of the evolutionary processes involved is still maturing (Hendry et al. 2008; Catford et al. 2009; Buswell et al. 2011). It is clear, however, that the success of an invader may depend on adaptation to novel conditions, potentially many times, during the course of range expansion (Hufbauer et al. 2012).  Such adaptation should be observable as ecologically important differentiation between populations from the native versus invaded range. Increased growth rate or reproductive capacity is frequently reported from field observations in the invaded range (Elton 1958; Crawley 1987; Thebaud & Simberloff 2001; Parker et al. 2013), and increasingly from common garden experiments (reviewed in Felker-Quinn et al. 2013). This improved vigor could contribute to rapid spread and population growth in the invaded range. Multiple hypotheses have attempted to explain increased performance of invasive populations, including the evolution of increased competitive ability hypothesis (EICA; Blossey & Nötzold 1995). EICA posits that selection will favor genotypes with reduced allocation to herbivore defense and increased allocation to growth, reproductive output, or competitive ability in the absence of herbivores characteristic of the  13  native range (Blossey & Nötzold 1995). EICA thus predicts that due to trade-offs, individuals from the native range will produce less biomass than individuals from the introduced range, and specialist herbivores will show improved performance on individuals from the introduced range in a common environment (Blossey & Nötzold 1995; Joshi & Vrieling 2005). EICA is only supported if increases in growth are linked to decreases in defense. Though increased performance in invasive individuals relative to native is often observed in common garden experiments, shifts in defenses are less common (Kumschick et al. 2013), and thus evidence for EICA is equivocal (Felker-Quinn et al. 2013). Some authors have expanded ideas about trade-offs to include tolerance to other types of stress, not just herbivore pressure (Bossdorf et al. 2005; He et al. 2010). If novel habitats are less stressful than native ones, selection would favor individuals that shift resources from stress tolerance to increased vigor and fecundity. Invoking trade-offs assumes that plants are unable to be both highly tolerant to stressful environments and highly competitive or have high reproductive output (Grime 1977). Several species demonstrate such trade-offs. For example, tolerance to an abiotic stress, such as serpentine soils or drought, often comes at the expense of competitive ability or growth rate (Sambatti & Rice 2007). EICA and other trade-off hypotheses suggest that invasive individuals evolve a lower tolerance to biotic or abiotic stress, and therefore will perform relatively poorly under stressful conditions, as assessed in several studies (e.g. Hodgins & Rieseberg 2011; Lachmuth et al. 2011; Kumschick et al. 2012).  However, trade-offs can occur in multiple directions (e.g. Burton et al. 2010). Additionally, strategies favored by selection may change over time, between different phases of an invasion (Dietz & Edwards 2004) or depending on the habitats invaded (Lachmuth et al. 2011).  A shift in resource allocation under benign conditions is not the only explanation for increased growth and reproduction in a species’ invaded range. Other genetically-based changes  14  between the native and invaded ranges could lead to increased performance such as inter- or intra-specific hybridization (Rieseberg et al. 2007; Schierenbeck & Ellstrand 2008; Lai et al. 2012).  Observing the phenotypes of a single generation in a common environment may be insufficient to demonstrate adaptation to a novel habitat. Latitudinal clines can impede our ability to infer evolutionary change from common gardens (Colautti et al. 2009). Even within the invaded range of a species, local adaptation can vary along latitudinal and environmental clines as the invading populations adjust to local environments, shifting phenology, biomass, and other trait means (Colautti et al. 2010 ; Lachmuth et al. 2011). It is also necessary to rule out differences between the native and invaded range caused by maternal environmental effects. Maternal effects can have strong, even adaptive, impacts on offspring phenotype in some systems (Galloway 2005) which may mimic or obscure the impact of a trade-off on growth or reproduction. Yet their influence in shaping performance differences between ranges is only rarely experimentally controlled (except see Monty et al. 2009; Hodgins & Rieseberg 2011). To test trade-off hypotheses, and to look for evidence of rapid adaptation to novel habitats, we ask: (1) Do native and invasive populations show consistent phenotypic differences in growth and reproduction, and do such differences remain even after controlling for latitude and maternal effects? And (2), is there evidence of a trade-off between the ability to grow quickly under benign conditions and the ability to tolerate stressful conditions? To address these questions, we conducted two large glasshouse experiments with Centaurea diffusa Lam. (diffuse knapweed), one of North America’s worst weedy invaders (LeJeune and Seastedt 2001). The first experiment, hereafter the broad CG, included 28 native European and 18 invasive North  15  American populations of C. diffusa (Fig. 2.1), using one of the highest levels of population replication in studies of this type (see also Kumschick et al. 2013). The second experiment, hereafter the maternal CG, assessed whether patterns observed in the broad common garden were maintained after controlling the maternal environment by using seed produced from glasshouse crosses of four populations from each range. Both gardens included a benign control treatment and biotic and abiotic stress treatments (simulated herbivory, nutrient deficiency, drought, and flooding). Differences between the two experiments would suggest that the maternal environment (as well as experimental variation) exerts significant influence on phenotypic response. Similarity between them would suggest that genetic polymorphisms control the phenotypic divergence between native and invasive individuals (Moloney et al. 2009). Furthermore, if plants from the invaded range outperform plants from the native range in the control treatment, but not in the stress treatments, hypotheses invoking trade-offs would be supported. On the other hand, a lack of significant differences between the two ranges in stress tolerance may indicate that C. diffusa success in North America does not rely upon evolutionary differences between ranges.  2.2 Materials and methods 2.2.1 Study species Centaurea is the most abundant noxious weed genus in the western United States, and along with its congeners, diffuse knapweed (Centaurea diffusa Lam.) reduces the quality of forage for livestock and wildlife and alters soil and water resource availability (LeJeune and Seastedt  16  2001). Centaurea diffusa is a monocarpic biennial (Thompson & Stout 1991), native to eastern Europe and western Asia, and now found throughout western Europe as well (Fig. 2.1; Greuter 2009). It was first reported in North America in 1907 in Washington State (Sheley et al. 1998) and now occurs in nearly half of the United States and Canada (Fig. 2.1; USDA NCRS 2013). Field surveys of C. diffusa reveal differing impacts of herbivores between ranges, with North American plants experiencing more seedhead herbivory, and European plants experiencing more root herbivory (Blair et al. 2008). The seedhead weevil, Larinus minutus (Col.: Curculionidae), introduced in the late 1990s, is the biological control agent associated with the largest decline in density of knapweed species, including C. diffusa (Seastedt et al. 2005; Myers et al. 2009).  Figure 2.1 Range and collection map of Centaurea diffusa in the northern hemisphere, by country. Populations used in the various experiments are indicated by point shade. All populations were used in the broad common garden (open circles), and four from each origin in the maternal common garden (filled circles). Origin status in a particular country is indicated by color, and was determined from classification in Greuter (2009), ISSG.org (2013), Kartesz (2013), Tropicos.org (2013), and USDA NCRS (2013). ‘Present, status unknown’ also includes countries where Centaurea diffusa is considered naturalized. Degrees of latitude are indicated on dotted lines, and degrees of longitude are indicated on solid lines. CG = common garden experiment.  17  Population genetic surveys indicate that C. diffusa was introduced to North America multiple times and harbors comparable levels of diversity at molecular loci to the native range (Hufbauer & Sforza 2008; Marrs et al. 2008). Centaurea diffusa can hybridize with diploid spotted knapweed, or C. stoebe subsp. stoebe L. (Blair & Hufbauer 2010; Blair et al. 2012; Lai et al. 2012). The tetraploid form of spotted knapweed (C. stoebe subsp. micranthos [Gugler] Hayek, sometimes referred to as C. maculosa) is a successful invader of North America, but the diploid form does not occur there (Blair & Hufbauer, 2009; Treier et al. 2009; Müller-Schärer pers. comm.). Hybridization between the diploids has been confirmed by both amplified fragment length polymorphism and transcriptome comparisons (Blair et al. 2012; Lai et al. 2012), which indicate that hybrid ancestry occurs in approximately one quarter of North American C. diffusa individuals (Blair & Hufbauer 2010) and that the genomic extent of introgression may be greater in introduced than native individuals (Lai et al. 2012). Given the lack of diploid spotted knapweed in the invaded range, these data suggest that introgressed genotypes of C. diffusa colonized North America. 2.2.2 Seed collections Seeds were collected in a broad collaborative effort from 30 native European populations and 27 invasive North American populations of C. diffusa from 2001 to 2008. Samples with sufficient germination (46 populations) were included in the broad CG (Fig. 2.1). Some populations had lower germination success, and so were excluded from some analyses (Tables A.2, A.3).  18  2.2.3 Broad common garden Centaurea diffusa from field-collected seed were grown under benign (i.e. provided with ample resources) or stressful (i.e. removal of key resource, resulting in obvious physiological response) conditions in a glasshouse at the University of British Columbia. Seeds were germinated on filter paper in 1% plant preservative mixture and distilled H2O at room temperature. Within 12 days, c. 1500 seedlings, from three to six families from each population, were transplanted into 5 cm diameter cones filled with 80% potting mix and 20% silica sand.  Seedlings were randomly assigned to four stress treatments and a control group. Within each treatment, populations were represented by at least three families. In total, 498 invasive and 1047 native plants were included. Each stress treatment contained c. 215 plants, and the control group contained c. 400 plants. At the onset of the stress treatments, 213 plants were destructively sampled. Within a treatment, c. 35 plants were randomly assigned to a position within each block. Blocks were randomly ordered across a glasshouse bench, and rotated randomly each week. Supplemental lighting was maintained at 16 hr per day. Treatments began at 6 weeks (plants had an average of 10 leaves larger than 3 cm) and included three abiotic stresses (nutrient deficiency, drought, and flooding) and one biotic stress (simulated herbivory). Plants were watered with fertilizer until the treatments began, when all plants except those in the nutrient stress treatment were top dressed with Osmocote 13-13-13 and wand-watered from above. Drought blocks were no longer watered, and flood blocks were placed in individual bins that were filled with water to just above soil level, maintained by daily top ups, and changed weekly. Plants in drought and flooding treatments likely experienced differences in soil nutrient availability relative to the control group. However, we were interested in comparing  19  the relative responses of native and invasive populations rather than isolating responses to particular stresses.  In the herbivory treatment, the apical half of every other leaf of each rosette was cut, eliminating at least one quarter of the plants’ aboveground biomass. The plant was then sprayed with 1 mM methyl jasmonate, a plant hormone induced by insect herbivory (Norton et al. 2008). The following week, remaining leaves were cut, and each subsequent week, new leaves over a length of 3 cm were cut, each time immediately followed by a methyl jasmonate spray. This occurred weekly for the first 5 weeks of treatment when growth was rapid, then every other week as growth slowed.  Morphometric measurements were taken before treatments began (5 weeks after germination) and before harvest (16 - 18 weeks), and included shoot mass, root mass, root crown diameter, area of longest leaf, and number of basal leaves (Table A.4). Phenological and survival traits were assessed every two - three days throughout the experiment, and included bolting status, bolting date, date of first stress response (wilting, yellowing), date of advanced stress response (total wilting, root death) and date of death.  Responses to flooding were scored as follows. Roots were defined as dead when the plant detached from its soil, and shoots when all basal leaves detached from the stem under gentle pressure. At harvest, root biomass was washed to remove debris, and then whole roots and shoots were harvested, dried at 29 oC, and weighed.  We conducted leaf choice trials using L. minutus, a specialist weevil that feeds on floral and leaf tissue of C. diffusa (Stephens & Myers 2013). Recently emerged adults were collected from three sites in British Columbia (Table A.3), where L. minutus was introduced in 2000 (Myers et al. 2009). Weevils were collected several times and held, with knapweed material from their  20  collection site, for less than 5 days before trials. Leaf material was taken from herbivory treatment plants sampled immediately before the first treatment, and then 1 week later, and thus these plants could express both constitutive and induced defenses. For each trial, a single weevil was placed in a well with a moist cotton pad and allowed to acclimate for at least 1 hr. Fresh cut leaf discs 6.3 mm in diameter were placed on the cotton pad, one each from an invasive and a native plant. The weevil was then placed between the two leaf discs and left to eat until one was half eaten or up to 6 hr. Leaf discs were frozen and then scanned to quantify tissue consumption using ImageJ 1.45s (Rasband 2011). Each invasive population was represented in six trials paired with a sample from three to five randomly chosen native populations. 2.2.4 Maternal common garden To rule out maternal environmental effects as a cause for differences observed between native and invasive populations, glasshouse generated seeds were used in the second experiment. Plants used to produce seeds for the maternal CG were from eight populations in the broad CG (Fig. 2.1, Tables A.2, A.3) and were grown at the same time and conditions as the broad CG. For each family, germination traits were tracked for 10 days (overall average, c. 2 days to germinate). Two individuals from each of four or five families for each population were transplanted into 1 l pots containing potting soil. Plants that flowered between November 2009 and March 2010 were repeatedly reciprocally crossed with all individuals flowering simultaneously in the same population. Flower heads were bagged before anthesis and after hand pollination to prevent cross-fertilization. Seeds from these crosses were used in the maternal CG. In this experiment, treatments and measurements were the same as the broad CG. In total, this garden included 230 plants, equally divided between native and invasive populations. In each  21  stress treatment, all populations were represented by three individuals (24 plants total) from different sets of parents when possible. The control treatment included 111 plants, and the early harvest control included 23 plants. Phenotypic differences between native and invasive individuals were assessed as in the broad CG, except for the following. First, morphometric measurements were taken slightly later (before stress treatment, 15 leaves on average, 7 weeks after germination; and before harvest, 19 - 20 weeks). Second, additional leaves were sampled at harvest, one mature leaf to determine specific leaf area, and one young leaf for future genomic work. Third, root biomass was not measured because it was not logistically feasible. Finally, leaf choice trials were expanded to include four time points (one constitutive and three induced), each 1 week apart, and c. 350 additional trials run using every individual from the invasive populations. 2.2.5 Statistical analysis To determine if C. diffusa differs phenotypically between its native and invasive range, we compared morphological and life history traits between these two ranges. Using R 3.0.1 (R Core Team, 2013), we employed restricted maximum likelihood (REML) models with random effects using the lme4 package. Univariate response traits were modeled separately for each treatment. Because non-destructive early control measurements took place before treatments were initiated, they included both control and stress treatment plants, under control conditions. Gaussian distributions were fitted for continuous measures and natural log-transformed when necessary to improve normality of residuals. Poisson distributions were fitted for count data and binomial distributions were fit for binary data.   22  To assess the significance of each model term, we removed each term or interaction in a stepwise manner based on likelihood ratio tests (LRT). Chi-squared test statistic, degrees of freedom, and P values reported are from these LRTs. When random effects were non-significant, LRTs were run using generalized linear models (GLM) without random effects, and the results of these LRTs are reported. All non-Gaussian minimal GLMs were checked for over-dispersion. Because the effect of one variable depends on the condition of the other, it is not meaningful to test the significance of main effects that are included in significant interactions during step-wise model simplification (Crawley 2012) and so in some cases these are not reported. Least squares (LS) means are reported as a measure of effect size, and are estimated from models including all significant terms except interactions.  In range differentiation models, origin (native or invasive), latitude, and their interaction were included in all full models. Population (uniquely named), maternal lines nested within population, and population variance within each origin were used as random effects in all full models (e.g. Trait ~ Origin * Latitude + (Origin | Population/Maternal line)). Additional models were run to explicitly test for the existence of stress-tolerance trade-offs in stress treatments. In these trade-off models, a population mean measure of performance in the benign control treatment (population mean shoot mass) and its interaction with origin were included in models of stress tolerance traits, along with all of the terms included in the range differentiation models. A significant interaction indicates a different relationship between performance under benign and stressful conditions between the two origins and thereby supporting our prediction of evolved difference in stress response between the two ranges. When either range differentiation or trade- 23  off models had a significant interaction, direct tests of slope differences are conducted and  model estimates are reported (Tables A.8, A.10). Some mixed model analyses combined data from both experiments. To maximize statistical power, data from the leaf choice trials of both gardens were analyzed together, including generation as a random effect. In these models, the ‘generation’ term includes maternal effects as well as experimental variation between gardens.  Additionally, to explicitly test for similarity between the two gardens, origin, generation, and their interaction were included as fixed effects in models of several traits under control conditions from the populations and lines included in the maternal CG. Other fixed and random effects were included as described for range differentiation models.  2.3 Results 2.3.1 Broad common garden For 11 of the 31 traits measured across all treatment levels and time-points, native and invasive individuals exhibit different phenotypes, demonstrated by a significant origin or origin-by-latitude term. For every measure of size with a significant difference between the ranges, invasive individuals were larger (Fig. 2.2a, b). Life history also differentiated the ranges: when significant, invasive individuals matured later (i.e. bolted; Fig. 2.2c). Latitude was a significant covariate for most traits (19 traits), though rarely had a significant interaction with origin (three traits). Significant random effects were common, especially population (20 traits). Seed weight and seed collection age were not significantly different between native and invasive mothers.  24  Table 2.1 Selected examples of means and confidence intervals estimated from range differentiation models of phenotypes of Centaurea diffusa grown in both common garden experiments.   Origin   Native Invasive  Trait Distribution and model Estimate CI Estimate CI P Broad Common Garden       Control       No. of basal leaves Poisson, REML 12.43 11.04 – 13.99 16.28 15.31 – 17.32 *** Bolting status (harvest) (%) Binomial, REML 24.30 17.49 – 32.69 2.07 0.62 – 6.76 [**] Herbivory       Root mass (g) Gaussian (loge), REML 1.03 0.88 – 1.21 1.37 1.09 – 1.72 * Bolting status (harvest) (%) Binomial, GLM 26.71 20.17 – 34.47 6.45 2.44 – 15.97 [***] Maternal Common Garden      Control       No. of basal leaves Poisson, REML 11.74 8.51 – 16.21 20.41 14.81 – 28.14 * Bolting status (harvest) (%) Binomial, GLM 75.09 54.53 – 88.34 7.57 2.51 – 20.67 [*] Nutrient       Area of longest leaf (cm2) Gaussian, GLM 56.45 41.76 – 71.13 82.74 68.05 – 97.42 ** Herbivory       Bolting status (harvest) (%) Binomial, GLM 84.70 36.30 – 98.18 0.44 0.01 – 24.82 [**] Least squares (LS) means from restricted maximum likelihood (REML) models including origin and all significant terms (interactions were excluded). Where no random effects were significant, LS means are estimated from generalized linear models (GLM). Significance of origin term (or origin-by-latitude in square brackets []) indicated in right-most column: ., P < 0.1; *, P < 0.05; **, P < 0.01;  ***, P < 0.001. CI = 95% confidence interval. Loge indicates natural log transformation of data. Nt = Not tested, due to significant interaction term.  Under benign control conditions, latitude was a significant covariate or had a significant interaction with origin, for most traits (Table A.4). Invasive individuals matured later than native (Fig. 2.2c, Table 2.2); at the time of harvest, 24% of native individuals had bolted, while only 2% of invasive individuals had bolted (Table 2.1). Origin-by-latitude was significant for this trait such that in native populations, increasing latitude marginally decreased the proportion of bolted  25  individuals within a population, while in invasive populations, the proportion bolted did not vary with latitude (Tables 2.2, A.10). Root crown diameter (a measure of developmental stage) did not predict bolting, in contrast to Powell & Myers (1988) (root crown diameter, χ2 = 0.48, P = 0.487). Few invasive individuals had bolted by harvest, making the analysis of bolting date problematic, though it showed a similar trend. Invasive individuals also had c. three more basal leaves at harvest than natives (Fig. 2.2b). Origin-by-latitude was significant for this trait; leaf number increased with latitude, but more steeply in invasive than native populations (Tables 2.2, A.10). Origin was marginally significant for two additional size traits (root crown diameter at c. 6 weeks, and root mass at harvest; Table 2.2), and in each case, invasive individuals were larger. The remaining traits (Table A.4) showed no effect of origin.  Table 2.2 Test statistics from range differentiation models for selected phenotypic traits of Centaurea diffusa.  Fixed effects Random effects Origin Latitude Origin-by- Latitude Populations Maternal lines Populations within each Origin Trait χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P Broad Common Garden      Control       No. of basal leaves 13.08 (1) *** 5.91 (1) * 1.50 (1) 66.22 (1) *** 257.29 (3) *** 37.14 (4) *** Bolting status (harvest) nt nt 7.14 (1) ** 0.57 (1) 10.71 (1) *** 4.07 (4) Herbivory       Root mass 4.33 (1) * 15.82 (1) *** 0.06 (1) 13.04 (1) *** <0.01 (1) 0.84 (4) Bolting status (harvest) nt nt 14.46 (1) *** <0.01 (1) 0.32 (1) 6.24 (4) Maternal Common Garden      Control       No. of basal leaves 4.20 (1) * 0.64 (1) 3.75 (1) . 4.48 (1) * 215.7 (1) *** 4.22 (4) Bolting status (harvest) nt nt 5.78 (1) * <0.01 (1) 0.11 (1) 0.68 (4) Nutrient       Area of longest leaf 6.89 (1) ** 1.12 (1)  1.32 (1) 0.13 (1) <0.01 (1) 1.47 (4)  26   Fixed effects Random effects Origin Latitude Origin-by- Latitude Populations Maternal lines Populations within each Origin Trait χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P Herbivory       Bolting status (harvest) nt nt 7.34 (1) ** <0.01 (1) <0.01 (1) <0.01 (4) Effect of origin and latitude on phenotypes of Centaurea diffusa grown in a common environment, for both the broad and maternal effects common garden. Results are presented from restricted maximum likelihood (REML) models. Where no random effects were significant, generalized linear models (GLM) were used to test fixed effects. Significance of term indicated by symbol: ., P < 0.1; *, P < 0.05; **, P < 0.01;  ***, P < 0.001. (df) = degrees of freedom. χ2 = chi-squared test statistic. Nt = Not tested, due to significant interaction term.       Figure 2.2 Phenotypic difference between ranges of Centaurea diffusa in broad common garden for control and all stress treatments that survived until harvest. Significance level of origin (or origin-by-latitude in box) in range differentiation models (Table 2.2) is indicated by red asterisks and large red dots: •, P < 0.1; *, P < 0.05; **, P < 0.01; ***, P < 0.001.  Sample size by treatment: Early control (varies by trait), Control (Invasive = 125, Native = 261), Nutrient (Invasive = 81, Native = 180), Herbivory (Invasive = 62, Native = 146). Two size traits (a) root mass, and (b) number of basal leaves at harvest, by origin. Early life stage control included for comparison; early control sample size for root mass (Invasive = 54, Native = 145) and leaf number (Invasive = 498, Native = 1047). Whiskers extend from the median to the highest and lowest values within 1.5 * the inter-quartile range. c) Proportion of individuals that bolted by harvest.  27    Points represent population means. Shaded area represents standard error. Text and asterisks indicate significance level of origin and origin-by-control mass (population mean shoot mass in control treatment) in trade-off model analyses (Table 2.3); red asterisks: *, P < 0.05; **, P < 0.01.  a) Days to first wilt in drought treatment (Invasive = 13 populations [65 individuals], Native = 26 populations [145 individuals]). b) Days to death in flood treatment (Invasive = 12 populations [49 individuals], Native = 25 populations [122 individuals]).  Under drought conditions, invasive plants wilted significantly faster than native plants, though there was no effect of origin on time to total wilt or death in range differentiation models (Table A.4). Trade-off models of these drought tolerance traits revealed a significant interaction between origin and performance in the benign control in each case (Fig. 2.3a, Table 2.3). For all three traits, the pattern is the same. Native populations exhibit a trade-off between drought tolerance and size in benign conditions; populations that produce larger individuals in benign conditions are less tolerant to drought. Invasive populations do not exhibit this trade-off. On average, invasive individuals may be somewhat less tolerant than native (days to first wilt in range differentiation models: native mean 3.73 [3.43 – 4.05] days vs invasive mean 3.14 [2.76 – 3.58] days, P < 0.05 ; Table A.5 ). That said, invasive Figure 2.3 Variation in stress tolerance and performance in benign conditions between ranges of Centaurea diffusa in broad common garden.  28  individuals do not appear to trade-off drought tolerance with size (Figure 2.3a), and tolerance even marginally increases with size in days to total wilt (explicit trade-off models; Table A.8).  Under flood conditions, there was no effect of origin on any trait in range differentiation models. However, in trade-off models, there was a significant interaction between origin and control performance in days to death. For both native and invasive populations, populations that produced larger plants in benign conditions survived longer under flooding conditions; however, larger size was significantly more advantageous in invasive populations (Fig. 2.3b, Tables 2.3, A.8). Under nutrient deficient conditions, both shoot and root mass exhibited a marginally significant difference between the ranges (Fig. 2.2a). Invasive individuals had more below-ground and above-ground biomass than native individuals (Table A.5). In the herbivory treatment, origin had a significant effect on both size and life history of the plants in range differentiation models (Fig. 2.2, Table 2.2). Invasive individuals had more root mass than native individuals (Table 2.1). As in the control, more native individuals had bolted by the end of the experiment than invasive (27% of native versus 6% of invasive individuals; Table 2.1).  Origin-by-latitude was significant for this trait, because plants bolted in only the three northern-most invasive populations. In the native populations, likelihood of bolting did not vary with latitude (Table A.10).   29  Table 2.3 Explicit trade-off models of phenotypes of Centaurea diffusa grown under stressful conditions in both common garden experiments.  Fixed effects Random effects Origin Control mass Origin-by- Control mass Latitude Populations Maternal lines Trait χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P Broad Common Garden      Drought       Date of 1st wilt nt nt 4.84 (1) * 3.31 (1) . 2.33 (1) <0.01 (1) Date of total wilt nt nt 4.17 (1) * 5.99 (1) * 2.83 (1) . <0.01 (1) Death date nt nt 4.36 (1) * 3.55 (1) . <0.01 (1) <0.01 (1) Flood       Death date nt nt 9.89 (1) ** 4.80 (1) * 4.51 (1) * 4.35 (1) * Maternal Common Garden      Drought       Date of 1st wilt 2.10 (1) 0.88 (1) 1.31 (1)  0.48 (1)  <0.01 (1) <0.01 (1) Date of total wilt nt nt 5.63 (1) * 0.31 (1) <0.01 (1) <0.01 (1) Death date 0.60 (1) 0.46 (1) 0.03 (1) 1.75 (1) <0.01 (1) <0.01 (1) Flood       Death date 0.01 (1) 1.27 (1) 0.33 (1) 0.17 (1) <0.01 (1) <0.01 (1) Interaction between stress tolerance phenotypes and performance in benign control conditions (‘Control mass’, or the population mean shoot mass in control treatment) for Centaurea diffusa grown in a common environment, for both the broad and maternal effects common gardens. The three-way interaction and the random effect of population variation within each origin were also tested, but were never significant. Results are presented from restricted maximum likelihood (REML) models. Where no random effects were significant, generalized linear models (GLM) were used to test fixed effects. Significance of term indicated by symbol: ., P < 0.1; *, P < 0.05; **, P < 0.01;  ***, P < 0.001. (df) = degrees of freedom. χ2 = chi-squared test statistic. Nt = Not tested, due to significant interaction term.  2.3.2 Maternal common garden Native and invasive phenotypes were different or marginally different for 13 of the 27 traits measured across all treatment levels and time-points (Table A.7), demonstrated by a significant origin or origin-by-latitude term. As before, for every measure of size with a significant difference between the ranges, invasive individuals were larger (Fig. 2.4a, Table A.6). Life history traits again differentiated the ranges: when significant, invasive individuals matured later (Fig. 2.4c). Origin-by-latitude was significant for all of the same traits, and one additional trait (four traits total: leaf number in early control, leaf area in control, and bolting status in control  30  and herbivory; Tables 2.2, A.7).  However, unlike the previous garden, latitude was rarely a significant covariate (one trait: average germination date; Table A.7). Because of the subset of populations available for use in the maternal CG, latitude and origin are statistically confounded. The populations included in the maternal CG do not exhibit the same variation in latitude as the broad CG, and the latitudinal ranges do not overlap (40⁰ - 44⁰ native, 45⁰ - 49⁰ invasive). However, there is no reason to expect origin and latitude are biologically confounded (i.e. that phenotypes differ between ranges simply because one range is more southerly than the other), as general results are congruent with the broad CG, where these two terms are not confounded. For some traits, a significant effect of latitude may be obscured by the confounded origin term, though latitude is still important for several traits (7 traits; Table A.7). Random effects are less common than in the previous garden, and population was significant for only four traits. From intra-population crosses in the glasshouse, invasive individuals produced more seed, and had a higher germination rate, than their native counterparts (Table A.7). Origin had a significant effect on germination date in the maternal CG; invasive seed germinated earlier than native. Seed weight did not exhibit an effect of origin.  31   Figure 2.4 Phenotypic difference between ranges of Centaurea diffusa in maternal common garden for control and all stress treatments that survived until harvest. Significance level of origin (or origin-by-latitude in box) in range differentiation models (Table 2.2) is indicated by red asterisks: *, P < 0.05; **, P < 0.01.  Sample sizes by treatment as follows: Early control (Invasive = 115, Native = 115), Control (Invasive = 55, Native = 56); Herbivory (Invasive = 12, Native = 12); Nutrient (Invasive = 12, Native = 12). Two size traits, (a) approximate area of longest leaf and (b) number of basal leaves at harvest, by origin. Early life stage control included for comparison. Whiskers extend from the median to the highest and lowest values within 1.5 * the inter-quartile range. Herbivory treatment not represented in leaf size measure because leaf biomass was experimentally altered. c) Proportion of individuals that bolted by harvest.  Under control conditions, most size traits showed a significant effect of origin or origin-by-latitude in the maternal CG (Fig. 2.4a,b; Table A.7). As in the previous garden, invasive individuals had more basal leaves at harvest than native individuals (Tables 2.1, 2.2). Origin-by-latitude was significant for number of basal leaves at 7 weeks; number of basal leaves increased with latitude among native individuals but marginally decreased with latitude among invasive individuals (Table A.10). Additionally, invasive individuals were larger than natives for several measures: area of longest leaf, shoot mass at 7 weeks, and shoot mass at harvest. Area of longest leaf at harvest exhibited a significant effect of origin-by-latitude; area increased with latitude  32  among native individuals, and decreased with latitude among invasive individuals (Table A.10). Bolting status at harvest again showed a significant effect of origin-by-latitude. At harvest, 67% more native individuals had bolted than invasive (Fig. 2.4c, Table 2.1). As in the broad CG, the likelihood of bolting in invasive populations did not vary with latitude. However, for natives the trend reversed; likelihood of bolting increased with latitude (Table A.10). Among plants that bolted, invasive individuals bolted later (native mean 59.61[53.63 – 66.25] days, invasive mean 71.27 [61.81 – 82.16]days; Table A.6), though the significance of origin in this model was marginal. Under nutrient stress, leaves of invasive individuals were nearly 50% larger than native (Tables 2.2, A.6). No trait measured in the drought condition showed a significant difference between native and invasive individuals in range differentiation models. However in the trade-off model of days to first wilt, there was a significant interaction between origin and control performance, though only for this trait (Table 2.3). As before (Fig. 2.3a), native populations demonstrated a trade-off between size in the control treatment and drought tolerance that was not apparent in the invasive populations. Invasive drought tolerance even increased with size in one trait (days to total wilt) and a similar trend is seen in another (days to first wilt; Table A.8). Origin was not significant for any trait in the flood treatment in range differentiation models. No significant interactions were evident in trade-off models of the flood treatment, though the trend for days to death was similar as in the broad CG (Fig. 2.3b). In the herbivory treatment, bolting status at harvest had a significant effect of origin-by-latitude (Fig. 2.4c).  For invasive populations, likelihood of bolting does not vary with latitude, and for native, it may increase with latitude,  33  though because only three of eight populations included bolting individuals in this treatment, this result should be treated with caution (Table A.8). 2.3.3 Cross-generational analysis Analysis of the combined leaf choice trial data from both gardens showed no difference between origins or experiments for either preference or amount consumed, and weevils’ preference was evenly divided (Table A.9). Generation was not a significant random effect in these models. There was a significantly negative effect of week on the amount of leaf disc eaten by the weevils, among the post-stress time points, suggesting a cumulative effect of increased herbivore defense and decreased palatability in response to the treatment.     Data presented are only from these eight populations. Text and asterisks indicate significance level of origin and origin-by-generation in range differentiation models (Tables 2.2, 2.4); red asterisks and large red dots: ns, not significant; •, P < 0.1; *, P < 0.05; **, P < 0.01; ***, P < 0.001.  The text in italics is there as a reminder of the significance (or not) of origin in mixed model analyses from the broad common garden (Generation 0; model including all populations in that experiment), maternal common garden (Generation 1), and combined datasets (middle).  Bars represent 95% confidence intervals from mixed model analyses. a) Number of basal leaves at harvest (Invasive = 96, Native = 91, in combined dataset). b) Bolting date, for the subset of plants that bolted before harvest (Invasive = 11, Native = 52, in combined dataset). CG = common garden experiment. Figure 2.5 Phenotypic differences between ranges of Centaurea diffusa across generations, for four native and four invasive populations in the control treatment.  34  A subset of traits measured in control conditions in both experiments were explicitly examined for the effect of generation (i.e. maternal effects as well as experimental variation between gardens). Differences were common; of the six traits analyzed this way, only one showed no effect of generation (bolting status; Table 2.4). Origin-by-generation was significant for two traits: number of basal leaves at harvest significantly increased with generation in invasive populations, while it decreased in native, and for bolting date, invasive populations did not significantly vary with generation, while native populations bolted earlier after a generation in the glasshouse (Fig. 2.5). For all other traits origin either remained significant as in the two gardens separately, or gained significance. Latitude was not included in this analysis because it is confounded with origin for these populations.  Table 2.4 Effect of origin and generation on phenotypes of Centaurea diffusa grown in a common garden, across two generations. Cross-generational Analysis Fixed effects Random effects Origin Generation Origin-by- Generation Populations Maternal lines Populations within Origin Trait χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P Early Control       No. of basal leaves 0.32 (1) 44.72 (1) *** 0.79 (1) 2.47 (1) 1.29 (1) 0.59 (4) Control       Shoot mass 0.77 (1) 7.92 (1) ** 0.06 (1) 13.35 (1) *** 0.76 (1) 0.91 (4) Area of longest leaf 7.93 (1) ** 41.97 (1) *** 0.02 (1)  1.30 (1) <0.01 (1) 0.26 (4) No. of basal leaves nt nt 15.10 (1) *** 19.61 (1) *** 185.07 (3) *** 11.72 (4) * Bolting status (harvest) 46.33 (1) *** 1.80 (1) 0.22 (1) 1.29 (1) 2.40 (1) 0.04 (4) Bolting date nt nt 8.21 (1) ** 2.05 (1) 69.19 (1) *** 1.21 (4) Results are presented from restricted maximum likelihood (REML) models. Where no random effects were significant, generalized linear models (GLM) were used to test fixed effects. Significance of term indicated by symbol: ., P < 0.1; *, P < 0.05; **, P < 0.01;  ***, P < 0.001. (df) = degrees of freedom. χ2 = chi-squared test statistic. Nt = Not tested, due to significant interaction term.   35  2.4 Discussion 2.4.1 Evolution in the invaded range These experiments reveal genetically-based phenotypic differences in plant growth and reproduction between the native and invaded ranges of C. diffusa (Question 1). Under benign common conditions, C. diffusa demonstrates life history trait differences and greater growth and reproductive potential in the invaded compared to the native range, though populations in these ranges are separated by little more than 100 years. Invasive individuals are larger than natives and, when maternal effects are controlled for, this effect is even more apparent.  Invasive individuals also have delayed maturity, suggesting an adaptation to a longer growing season, or a shift towards decreased use of the facultative annual strategy. Such a shift could position individuals to optimally utilize seasonal resources, such as precipitation. Delayed maturity could also have strong impacts on the invasion success of these populations; if delayed bolting means the plant is larger at flowering, then it has the potential to produce more flower heads, more seeds, and more progeny. This potential for increased fitness is demonstrated from the results of our glasshouse crosses: the larger invasive individuals produced more seed than their native counterparts. Under nutrient limited conditions in both gardens, minimal bolting occurred (Fig 2.2c, 2.4c), suggesting that overall the species is not adapted to low nutrient conditions and that adaptation to this environment does not vary between ranges, a pattern consistent with other characteristics of opportunistic species. In the face of human-induced selective pressures and novel environments (Hufbauer et al. 2012), C. diffusa has demonstrated the capacity to rapidly adapt, which may have contributed to its successful invasion of North America.  36  Performance traits showed a significant effect of origin across broad population sampling, after an attempt to control for maternal environment, and even, in some cases, across stressful treatment environments. Phenotypes often varied with origin, even when latitude was included in statistical models, indicating that differences in phenotype are not due solely to latitude, a concern raised by Colautti et al. (2009). Broad patterns of growth and life history were consistent across both gardens.  These patterns also held in the combined dataset, a balanced dataset that minimized maternal effects, and supported results after a generation in a common environment within particular family lines. When differences in maternal effects were minimized, some traits varied with origin, while others lost this effect; this difference may be due to the maternal environment influencing results in the broad CG. However, our seed collections covered a broad geographic and latitudinal range, no sampling biases between ranges for collection date are apparent (Fig. A.1), and seed weight, which is commonly used as a proxy for maternal effects, did not vary between the ranges (Table A.4). The reduction of maternal environmental ‘noise’ in the maternal CG may have allowed more traits to show a significant effect of origin in the control treatment. But while this experiment may have decreased the effects of the maternal environment, due to limited population and seed sample size, some analyses may have lacked power to detect differences that were significant in the broad CG. Insufficient power may be a concern in the stress treatments, each of which only contained 24 individuals. Other potential explanations for differences between the two gardens include the inadvertent reduction of genetic variation by crossing only a subset of plants from each population, or more environmental variation occurring over the course of one experiment than the other.   37  The focus of this study on broad sampling and controlling maternal environment may have allowed us to detect range variation not seen in previous studies of this species (Blumenthal & Hufbauer 2007; Blair et al. 2012). Other aspects may affect the analysis of our data. Population relatedness has been shown to have a strong effect on the analysis of phenotypic variation in common gardens, and the implications of this variation for adaptive evolution (Lachmuth et al. 2011), however relatedness was not assessed here. Population genetic analyses of these populations are ongoing. Also, growth or biomass production measured without competition may not translate to performance (particularly competitive ability) in the field (see Vila & Weiner 2004; Gruntman et al. 2013). Blair et al. (2012) compared performance of native and introduced C. diffusa experiencing competition from a North American native grass in a glasshouse, and found no effects of range. Additionally, further analysis of the differences in abiotic environment between the two ranges is needed to elucidate the relationship between invasiveness and abiotic stress tolerance in this species and may have more explanatory value than latitude alone. 2.4.2 Test of EICA and trade-offs hypotheses Investigations into trade-offs in C. diffusa reveal a trade-off in performance between plant growth in a benign environment and tolerance to some stresses (Question 2) in the native, but not the invaded range. In the case of herbivory, none of the predictions of the EICA hypothesis were supported. Invasive individuals did as well or better than natives in both benign control and simulated herbivory treatments. Palatibility was not significantly different between the ranges in the leaf choice trials, suggesting no defense trade-off against a specialist herbivore in the invaded range, a pattern also seen with a generalist herbivore by Blair et al. (2008). A possible caveat is that the specialist used has been present as a biological control agent at some locations in the  38  invaded range for up to 15 years – this brief re-exposure may be enough to select for renewed herbivore defense in some populations. Weevil and seed collection overlapped at only two locations (Table A.3), however, L. minutus is widespread in North America (Hufbauer, pers. obs.). A recent meta-analysis of experiments testing EICA found evidence for genetically based changes between the native and introduced ranges in defense, growth, and competitive traits (Felker-Quinn et al. 2013), but not necessarily in the direction predicted by EICA, and without evidence for trade-offs between growth and defense against herbivores. Our results here are consistent with this conclusion. Our results do, however, provide some evidence for trade-offs between growth and tolerance to abiotic stressors, including both drought and flooding, which vary in their strength between the two ranges in this species. Under drought stress, native populations exhibit a trade-off between drought tolerance and size in benign conditions – populations that produce smaller plants are more drought tolerant. The invasive populations appear to have lost this trade-off, as tolerance to drought does not decrease with size in either experiment in any of the three traits measured, and even increases with size in days to total wilt in the maternal CG. However, it is interesting to note that invasives may be somewhat less tolerant than the natives on average at least in terms of the mean time first signs of wilt in the broad CG (but trendlines cross, see Fig 2.3a). The evidence for a flood tolerance trade-off is less well supported, but suggestive. In the broad CG, for both native and invasive populations, populations that produced larger plants in benign conditions took nearly a day longer to die under flooding conditions; however this trade-off among populations within a range was significantly steeper for invasive populations (Fig 2.3b). In the maternal CG, the trend for this trait was similar in direction, but non-significant. Taken  39  together, these results suggest the prevalence of “all-purpose genotypes” (Baker 1965), which do well in a wide array of environments, in invasive populations. These “all-purpose genotypes” may be released from the constraint apparent in the native range, where native population may produce large individuals or tolerant individuals but not both. There is some evidence that invasive populations have the potential to produce both very large and very tolerant individuals. Other explanations are possible for these data. For example, controlled glasshouse conditions are not the same as those experienced by the plant in the field (for example, in terms of biotic interactions with soil organisms), and may fail to reveal important interactions observable in field conditions, especially nutrient-water interactions. Genetically based changes between the native and invaded ranges could include the effect of inter- or intra-specific hybridization, and has been documented in C. diffusa (Blair et al. 2012; Lai et al. 2012). Hybridization is proposed to stimulate invasiveness through the generation of novel genotypes by recombination allowing hybrid individuals to benefit from heterosis, increased fitness, and/or decreased genetic load (Schierenbeck & Ellstrand 2008). In several species, intra-specific genetic admixture has resulted in higher genetic diversity in invasive populations compared to native populations, as well as hybrid vigor and loss of inbreeding depression (Kolbe et al. 2004; Keller & Taylor 2010; Verhoeven et al. 2011). Hybridization may even result in stress tolerant “all-purpose genotypes” (Baker 1965). While a causal link between invasion and hybridization has been demonstrated rigorously in some cases (Facon et al. 2005; Lavergne & Molofsky 2007), it is generally not clear if hybridization is a cause or consequence of range expansion, or if it has an impact on ‘invasiveness’ per se.  However, because this study does not explicitly address the genetic  40  background of the populations involved, further work is required to establish the impact of inter- or intra-specific hybridization in this species. 2.4.3 Conclusions Our data suggest that post-invasion evolution may have played a significant role in the invasion success of C. diffusa in North America but do not rule out the impact of pre-invasion hybridization with C. stoebe subsp. stoebe. Field surveys of C. diffusa in its native and invaded ranges found larger, more robust plants in North America (Blair & Hufbauer 2009). As Blair et al. (2012) suggest, some of these differences likely result from variation in environmental field conditions. Our data indicate, however, that variation in the maternal environment associated with field-collected seeds may confound common garden studies involving C. diffusa. We controlled for effect of latitude and minimized maternal effects by producing seeds under controlled environmental conditions, which revealed evidence for post-invasion evolution. While our data do not follow all of the necessary predictions of EICA, they do lend support for trade-offs for some abiotic stress tolerances, especially drought tolerance, in the native but not the invasive range. Invasive populations may instead avoid drought through delayed maturity. While the genetic basis for these patterns is currently unknown, future common gardens and genetic work will explore the possibility that intra- or inter-specific hybridization has resulted in the prevalence of “all-purpose genotypes” in the invaded range and aim to identify the genetic changes that enabled C. diffusa to become one of the most successful weedy invaders of western North America.   41  Chapter 3: Phenotypic Plasticity and Niche Expansion in an Invasive Thistle  3.1 Introduction Much recent research in invasion biology has assessed whether populations of invasive plants show heritable phenotypic differences in growth and reproduction between their native and invaded ranges in an effort to understand the causal drivers of invasion (Thébaud & Simberloff 2001; Hinz & Schwarzlaender 2004; Bossdorf et al. 2005; Felker-Quinn et al. 2013). Where such differences are not found, species which successfully invade may be pre-adapted. Pre-adapted species may already be well-suited to the (typically anthropogenically) disturbed conditions found in the novel habitat, perhaps fueled by adaptation to frequent disturbance or human–altered habitats in the native range (Lee & Gelembiuk 2008; Hufbauer et al. 2012; Mráz et al. 2012b). Indeed, a species is more likely to establish a self-sustaining population in a new location if there is at least some degree of environmental overlap with the native range (Bock et al. 2015; Mesgaran et al. 2014). Yet invasion success may depend on the capacity of a species to adapt to novel environmental conditions, and rapid adaptive change through microevolution has been documented in many invasive species (reviewed in Dlugosch & Parker 2008; Felker-Quinn et al. 2013), often occurring over very short time spans (Whitney & Gabler 2008; Buswell et al. 2011). This rapid evolution is often understood to be the result of environmental differences between the ranges generating strong selective pressures (Bock et al. 2015). Clinal genetically-based phenotypic variation demonstrated amongst invasive populations represents some of the best evidence for rapid evolution in the invaded range, including  42  adaptation to latitudinal and altitudinal clines (Bock et al. 2015; Alexander et al. 2009), though multiple introductions and admixture may play an underappreciated role in driving clinal variation (Kao et al., 2015). Local adaptation can quickly shape phenotypic variation during range expansion along selective climate gradients as the invading populations adjust to local environments, shifting phenology, biomass, and other trait means (Colautti et al. 2010). Such local adaptation often results in important differentiation between native and invasive populations and may have a stronger benefit to the fitness of an invasive species than genetic constraints, enemy release, or the evolution of increased competitive ability (Colautti & Barrett 2013; Zenni et al. 2014a).  Several other evolutionary hypotheses invoke trade-offs in resource allocation to account for genetically-based phenotypic differences between the native and invaded ranges. If novel habitats are less stressful, either biotically, for example, due to absence of specialist herbivores (evolution of increased competitive ability (EICA); Blossey & Notzold 1995), or abiotically, for example when resources are abundant (Bossdorf et al. 2005; He et al. 2010), selection would favor individuals that shift resource allocation from stress tolerance to increased vigor and fecundity, and therefore invasiveness. Such trade-offs and their role in the invasion process have been assessed by several studies (Hodgins & Rieseberg 2011; Lachmuth et al. 2011; Kumschick et al. 2013; Chapter 2), but these attempts are complicated by the variability of favored strategies between different habitats (Lachmuth et al. 2011).  Rather than specialize in local novel environments through a micro-evolutionary response, invasive species may instead benefit from generalist strategies, where a plastic, ‘general-purpose’ genotype performs similarly well in many environments (Baker 1965). Phenotypic  43  plasticity, whereby a genotype can adjust its phenotype to maintain fitness in stressful or unfavorable environments and/or increase fitness in favorable environments (Richards et al. 2006), has been shown to be a causal driver of many invasions (Hahn et al. 2012; Zenni et al. 2014b; Bock et al. 2015). Theoretical work suggests that while stable environments will favor local adaptation, frequent disturbance (especially anthropogenic) will select for phenotypically and developmentally plastic genotypes, in either the native or invaded ranges, well suited for colonizing novel habitats (Meyers et al. 2005; Lande 2009; Hufbauer et al. 2012). Yet, at least among Holarctic invasive plants, evidence of species adapting to environments outside of those experienced in the native range, i.e. shifts in either the realized or fundamental niche is rare (Petitpierre et al. 2012, but see Webber et al. 2012).  Here, we report on a field-based common garden study of genetically based phenotypic variation in Centaurea diffusa (diffuse knapweed), one of North America’s most problematic weedy invaders (Lejeune & Seastedt 2001). Such changes, demonstrated in a previous greenhouse study (Chapter 2), could be due to local adaptation to new local environmental conditions in the invaded range (such as trade-offs) or to the evolution of a generalist strategy that produces high performance under a wide range of conditions.  To distinguish between these two hypotheses, we examine phenotypic differences between populations from the native and invaded ranges, using a common garden experiment in the naturalized range of C. diffusa in Montpellier, France, to test for evidence of local adaptation or increased phenotypic plasticity. The naturalized range represents an area known to be within the physiological tolerances of C. diffusa (it is reported there, though rarely; Greuter 2009), and yet external to both the native and invaded ranges. If trait divergence is at all due to local adaptation, then investigating performance within either  44  range could favor local populations if individuals experience biotic or abiotic conditions more typical of their ‘home range’ (Colautti et al. 2009).  Thus, our experimental design allowed us to examine performance under more realistic field conditions (as compared to the previous greenhouse study in Chapter 2), while reducing potential ‘home range’ advantage.  Next we test whether patterns of genetically based phenotypic variation are in agreement with the resource allocation trade-offs hypothesis by comparing plant traits from both ranges in the presence or absence of experimentally applied drought stress, a trade-off implicated in a previous study (Turner et al. 2014). We test whether invasive populations of C. diffusa perform better than native populations in a field setting in the naturalized range and whether performance is correlated with drought tolerance. If the evolution of a trade-off between drought tolerance and growth rate is a causal driver of invasion, then invasive populations should perform significantly less well than native populations under drought stress. Apart from the resource allocation trade-offs hypothesis, we also investigate evidence for local adaptation by measuring phenotypes in a common environment of plants sampled from native and invaded ranges from a variety of environmental conditions. In the absence of local adaptation (including trade-offs), selection for increased phenotypic plasticity in the invaded range may instead explain the spread and dominance of invasive populations over many habitats. Therefore, using publically available occurrence data, we further examine the climate space inhabited by native and invasive C. diffusa at a larger spatial scale than our sampled populations, to test the prediction that plasticity in environmental tolerance should expand the realized climatic niche in the invaded range of C. diffusa.   45  3.2 Materials and methods 3.2.1 Study species Within a large family containing many weeds (Asteraceae), the genus Centaurea has contributed 30 non-native species to North America, including 11 noxious weeds (USDA NCRS 2013) and is one of only 15 plant genera in the US to contain more weedy species than expected by chance (Kuester et al. 2014).  The five Centaurea species with the greatest impact, including Centaurea diffusa Lam. (diffuse knapweed), have invaded millions of hectares of grassland, making it the most abundant noxious weed genus in the western United States (LeJeune &Seastedt 2001). Centaurea diffusa is typically a monocarpic, facultative biennial (Thompson & Stout 1991) that forms a basal rosette, then bolts and dies after reproducing. Native to parts of eastern Europe and western Asia, C. diffusa is found sparsely throughout Western Europe, where it is considered a naturalized alien (Fig. 3.1; Greuter 2009; Bleeker et al. 2007). First reported in North American more than 100 years ago (Sheley et al. 1998), it now occurs in roughly half of Canada and the US (Fig. 3.1; USDA NCRS 2013). Surveys of genetic diversity in this species suggest that 1) C. diffusa has been introduced to North America multiple times (at least once from Turkey); 2) comparable genetic diversity exists within each range; and 3) little population structure is evident in the native range (Hufbauer & Sforza 2008; Marrs et al. 2008).   46    Figure 3.1 Range and population map of Centaurea diffusa in the Northern Hemisphere, by country, used in the field experiment. Origin of sampled population (invasive or native C. diffusa) indicated by point shape. Origin status in each country is indicated by color. Degrees of latitude are indicated on dotted lines, longitude on solid lines. Modified with permission from Turner et al., 2014. 3.2.2 Populations Seeds were collected in a broad collaborative effort from eight native European populations and six invasive North American populations of C. diffusa as part of a large sampling scheme (see Chapter 2). Collection dates ranged from 2005 to 2010, with the majority of populations collected in 2008 (Table B.5).   47  3.2.3 Common garden experiment To compare the phenotypes of invasive C. diffusa to native C. diffusa and look for evidence of local adaptation or increased phenotypic plasticity in the invaded range, we measured phenotypes in a common environment. In the spring of 2011 we initiated a field common garden in the naturalized range of C. diffusa at Montpellier (CEFE, Plateforme des Terrains d’expériences du Labex CeMEB), Languedoc-Roussillon, France, near the northern edge of the Mediterranean Sea. Seeds collected from four to six mothers at each of 14 collection locations were used, collected between 2005 - 2010 (Table B.5). In total, this common garden included 263 individuals, with a mortality rate of 14%, so that 225 survived until harvest. Seeds were germinated on filter paper in distilled H2O in a temperature controlled chamber, which was maintained to a diurnal range of 12 - 22 °C. Supplemental lighting maintained a 16 hr day. Within 15 days, c. 400 seedlings were transplanted into 8 cm square peat pots (Jiffy Products International B.V., Moerdijk, Netherlands) filled with 33% tomato potting mix, 33% silica sand and 33% steam sterilized field soil. Seedlings were grown in a climate-controlled glasshouse and watered daily. When plants reached the median eight-leaf stage (4 weeks after germination), they were transplanted to the field. Individuals were planted into an experimental field in a randomized block design, with 0.5 m between rows and between plants within a row and each row assigned a treatment. Plants from each population were randomly assigned to a treatment. Treatments included an irrigated control and non-irrigated ‘drought’ treatment. Because drought stress tolerance was only a subset of one of our hypotheses, roughly twice as many plants were assigned to the control group than to the drought treatment. Non-experimental plants were planted around the edge of the plot to lessen  48  edge effects. After transplantation, all plants received supplemental watering every 12 hr. Once treatment began, 2 weeks after field planting, water addition to drought treatment rows ceased. Thus, plants from the drought treatment received water from only natural rainfall. Morphometric and life history measurements were taken several times over the course of four months, and these measurements were taken blind.  Repeated morphometric measurements were taken before treatment began (2 weeks after transplantation), during treatment (4 weeks after transplantation), and at harvest (at bolting or four months after transplantation for those plants that did not bolt) and included common size measurements for rosette-forming plants, such as length and width of longest leaf, number of basal leaves >3 cm long, and maximum diameter of basal rosette. Four weeks after transplantation a subset of plants (126) were sampled for specific leaf area (SLA); one leaf per plant was harvested and image scanned while fresh, and leaf area calculated using ImageJ (Rasband 2011). Life history traits were assessed weekly and included bolting probability, bolting date, date of first stress response (wilting or yellowing), and mortality. When a plant bolted, but before it flowered, it was measured and harvested to avoid release of pollen or seed from potentially invasive genotypes. Additional measurements were taken at harvest, including shoot mass, root crown diameter, and approximate rosette area (maximum diameter × perpendicular diameter × π/4). The leaf sampled for SLA and the harvested shoot material for each plant were stored separately in paper bags and oven-dried at approximately 65 °C for at least three days, and then weight measurements taken. 3.2.4 Statistical analysis To determine if C. diffusa differs phenotypically between its native and invasive range we compared morphological and life history traits among C. diffusa individuals. Using R 3.0.1 (R  49  Core Team 2014), we employed restricted maximum likelihood (REML) models with random effects using the lme4 package. Univariate response traits included root crown diameter, rosette area, shoot mass, specific leaf area, bolting probability, bolting date, date of first wilting, and mortality rate. Repeated measurements of a trait (leaf count, length and width of longest leaf, and rosette diameter) were analyzed together, and measurement date and individual were included as random effects. Gaussian distributions were fit for continuous measures, and trait values were natural log-transformed when necessary to improve normality of residuals. Poisson distributions were fit for count data, and binomial distributions were fit for binary data. Data were scaled when necessary to improve model performance. To account for phenotypic variation arising from environmental differences between sampled locations, each full model included a composite abiotic environmental covariate determined by a principal component analysis (PCA) of altitude, latitude, and 18 bioclimatic variables of each seed collection location taken from the WorldClim database of current climatic conditions (hereafter, the “experimental PCA”; Table B.6; Hijmans et al. 2005). The principal component that explained the most variance among collection locations (PC1) was used in all trait models as the environment term. Models were also run using a second composite environmental variable (PC2) or only latitude in place of the environment term, but this did not substantially alter results (not shown).  To test whether invasive populations of C. diffusa perform better than native populations, we ran range differentiation models, where origin (native or invasive), environment (PC1 from the experimental PCA), and their interaction, as well as treatment, were included as fixed effects in all full models. Population (uniquely named) and maternal lines nested within population were  50  used as random effects in all full models (e.g. Trait ~ Origin * Environment + Treatment + (1 | Population/Maternal line)). When a range differentiation model had a significant origin-by-environment term, slopes of regression lines from model estimates are reported. To test for resource allocation trade-offs, we assessed differences in morphological and life history traits between treatments. Trade-off models explicitly tested for a trade-off between performance in benign conditions and tolerance to drought stress (as in Chapter 2). In these models, a population mean measure of performance in the benign control treatment (log transformed population mean shoot mass) and its interaction with origin were included in models using only individuals in the drought treatment, along with all of the terms included in the range differentiation models (Trait ~ Origin * Control population mean mass + Environment + (1 | Population/Maternal line)). A significant interaction would indicate a different relationship between performance under benign and stressful conditions between the two origins.   To assess the significance of each model term, we removed each term or interaction in a stepwise manner based on likelihood ratio tests (LRT). All LRTs were corrected for multiple comparisons using the false discovery rate (FDR) procedure implemented in the ‘qvalue’ package v.1.40.0, with an FDR cut-off value of 5% and the ‘bootstrap’ method (Storey et al. 2004). However, because this correction did not change the significance of any fixed effect (and only four random effects out of 65 LRTs; Table B.7), and because model terms were included based on the P value of each LRT, only significance based on P value is reported. Chi-squared test statistic, degrees of freedom, and significance (P value < 0.05) are reported from these LRTs. When all random effects were non-significant, generalized linear models (GLM) were used, and the results of these LRTs are reported. All non-Gaussian minimal GLMs were checked for over-dispersion.  51  For models with significant origin or origin-by-environment terms, model estimates are reported for fixed effects. Because the effect of one variable depends on the condition of the other, it is not meaningful to test the significance of main effects that are included in significant interactions during step-wise model simplification (Crawley 2012) and so these are not reported. 3.2.5 Occurrence data and principal components analysis To determine if differing relationships between phenotype and environment between origins observed in our dataset are reflected in a difference between the realized climatic niches of the species ranges at a spatial scale larger than our sampling area, we investigated evidence of a climatic niche expansion in the invaded range of C. diffusa. Five hundred and ninety-two geo-referenced occurrence locations for C. diffusa from North America, Europe, and western Asia were retrieved from the Global Biodiversity Information Facility, using the R package ‘rgbif’ (GBIF 2014; Chamberlain et al. 2014). This was combined with 70 seed collection locations from previous sampling efforts (Chapter 2). For each occurrence record, corresponding climate data were retrieved from the WorldClim database as above (Table B.6; Hijmans et al. 2005). This dataset was then used in a principal components analysis of the climate, altitude, and latitude of all occurrence locations (hereafter, the “occurrence PCA”). The magnitude and statistical significance of the niche shift between the occurrence centroids in the invaded and native ranges in the PCA graph were assessed using a between-class analysis with the R package ‘ade4’ yielding a between-class inertia percentage (Broennimann et al. 2007; Dray et al. 2007). This ratio was further tested with a Monte-Carlo randomization test (999 iterations; Dray et al. 2007). In addition, 99% confidence ellipses describing the cluster for each range using the bivariate t-distribution are presented. Because the GBIF data used here may not be error free, we  52  reran this analysis using only populations within two standard deviations away from the mean of PC1 and PC2 of the full dataset to verify results.  3.3 Results 3.3.1 Principal components analysis of experimental populations The first two components obtained by the experimental PCA of abiotic environmental variables characteristic of each seed collection location explain 33% and 28% of variance among the collection locations used in the Montpellier experiment, respectively (Fig. 3.2, B.1). Axis 1 was correlated most strongly with maximum temperature of the warmest month, annual precipitation, and precipitation during the wettest periods (BIO5, BIO12, BIO13, BIO16; Table B.6) and can be conceptualized in terms of “aridity”, with small values associated with dry, hot summers. Axis 2 was correlated most strongly with minimum and mean temperature of the coldest periods, annual mean temperature, and temperature seasonality (BIO6, BIO11, BIO1, BIO4) and can be thought of as “harshness of winter”, with small values associated with cold winters. A comparison of invasive and native C. diffusa sample locations for these two axes indicates a substantial degree of overlap of the sampled climatic niches of these populations (Fig. 3.2; 95% confidence ellipses group most populations into a single cluster, not shown). The degree of dispersion among populations may indicate that native populations were sampled from a narrower range of environments. Nevertheless, later analyses had sufficient power to detect significant differences between native and invasive populations and their relationship to environmental conditions.  53          Figure 3.2 Principal components analysis of abiotic environmental variables of sampling locations of experimental populations of Centaurea diffusa. Climate data from WorldClim database (Hijmans et al. 2005). Variables defined in Table B.6.  3.3.2 Phenotypic differentiation In the Montpellier dataset, of the 13 traits assessed for range differentiation, nine exhibited significant phenotypic differentiation between the native and invaded ranges of C. diffusa (leaf number, leaf width, root crown diameter, rosette area, shoot mass, bolting probability, bolt date, wilt date, and yellowing date; Table 3.1, B.7).  In each case origin had an effect on trait values, often via an interaction, but sometimes not. Origin was significant for leaf number and marginally significant for leaf width and rosette diameter. Random effects were significant in every model that differentiated the two ranges, except bolting probability. Specific leaf area and  54  mortality rate did not differ significantly between treatments or ranges. For every measure of size that varied significantly or marginally significantly between the two ranges, invasive individuals were larger (Table B.8, Fig. 3.3a, b, f, B.2). For example, invasive rosettes had approximately 32 grams more shoot mass than natives in the control treatment (observed means and standard errors: invasive 83.96 ± 7.66, native 51.46 ± 6.27, Fig. 3.3a, b). All size traits with a significant origin-by-environment term displayed a similar trend: for invasive populations, size did not significantly vary with environment, whereas for native populations, size significantly changed along the axis of collection location environmental conditions (increase: root crown diameter, rosette area, shoot mass; decrease: leaf width; Table B.8, Fig. 3.3f, B.2). In other words, for native populations, the hotter and drier the climate experienced at the source location, the smaller the individual the population produced. Leaf width is an interesting exception to this trend; leaf width in native populations appears to decrease along the axis of environmental conditions for the first two measurements, but it increases during the third, resulting in an overall negative slope (Fig. B.2, Table B.8). Invasive populations have lost the relationship to this axis of environmental conditions. Life history traits also differentiated the two ranges; invasive individuals were less likely to bolt during the course of the experiment (observed mean and standard error in control treatment: invasive 29.5 ± 5.9%, native 52.4 ± 5.5%; Fig. 3.3c, d). Both native and invasive individuals exhibited a significant relationship to the environment at the collection location for bolting, although in the opposite direction. Moving along the aridity axis (PC1) wetter climates with milder summers, native individuals decreased their probability of bolting whereas invasive individuals increase their probability of bolting. Though the subset of plants that bolted during the course of the experiment was less than half (28 invasive and 68 native individuals), there was a significant interaction between origin and environment for bolt  55  date, such that the milder and wetter the climate experienced by native populations, the later the bolting date, while bolting date in invasive individuals had no significant relationship to environmental conditions of collection location (Fig. 3.3e). Table 3.1 Test statistics from range differentiation models of phenotypic measurements of Centaurea diffusa, for all traits measured in the field experiment with a significant origin or origin-by-environment term.  Fixed effects Random effects  Origin Env Origin-by-Env Treatment Population Maternal line Repeat measure Trait χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P Number of basal leavesǂ 5.82 (1) * 0.49 (1)  0.03 (1) 0.12 (1)  0.81 (1)  0 (1)  552.87 (3) *** Width of longest leaf nt nt 8.50 (1) ** 0.82 (1) 1.02 (1)  0.85 (1) 132.38 (3) *** Root crown diameter nt nt  9.88 (1) ** 0.82 (1) 14.89 (1) *** 16.33 (1) *** --- Rosette area at harvest nt nt  8.35 (1) ** 5.23 (1) * 24.23 (1) *** 3.16 (1) . --- Shoot mass nt nt  14.44 (1) *** 1.71 (1) 14.82 (1) *** 9.71 (1) ** --- Bolting probability nt nt 37.19 (1) *** 0.06 (1) 0 (1) 0 (1) --- Bolt date nt nt 9.34 (1) ** 0.07 (1) 4.84 (1) *  0 (1) --- Wilt date nt nt 6.28 (1) * 21.42 (1) *** 4.76 (1) * 0 (1) --- Yellow date nt nt  25.89 (1) . 0.46 (1) 4.45 (1) *  0 (1) --- Results are presented from restricted maximum likelihood (REML) models. Significance of term indicated by symbol: ., P < 0.1; *, P < 0.05; **, P < 0.01;  ***, P < 0.001. Env = environment term. (df) = degrees of freedom. χ2 = chi-squared test statistic. Nt = not tested because of significant interaction term. ǂData scaled when necessary to improve model performance.  Figure 3.3 Selected examples of size and life history trait divergence among Centaurea diffusa ranges in the Montpellier experiment. Next page. All figures are from observed data; model parameters are described in Tables 3.1 and B.8. Shaded area represents standard error. (a) Shoot mass at harvest by origin. (b) Population mean shoot mass (log transformed) along the axis of environmental conditions (significant origin-by-environment interaction). (c) Proportion of each group that had matured (i.e. bolted) by harvest. (d) Population mean bolting probability along the axis of environmental conditions (significant origin-by-environment interaction). (e) Population mean bolt date, among plants that bolted, along the axis of environmental conditions (significant origin-by-environment interaction). (f) Population mean root crown diameter along the axis of environmental conditions (significant origin-by-environment interaction). In (b), (d), (e), and (f) origin is indicated by point shape; invasive C. diffusa as circles, native C. diffusa as triangles.  56    57  3.3.3 Drought response and resource allocation trade-offs Only two traits demonstrated a significant effect of treatment in range differentiation models (rosette area, date of wilting), suggesting a limited impact of water addition (Table B.7). Explicit trade-off models of drought treatment plants revealed no significant interactions between origin and population mean performance in the benign control treatment for any trait (not shown). Total natural rainfall at the field location during the duration of the drought treatment (June – September) was 170.4 mm. Unfortunately flow rate for the water addition treatment was not recorded.  3.3.4 Evidence of niche expansion Principal components analysis of the climatic data of all C. diffusa occurrences (376 invasive and 286 native occurrences) in the occurrence dataset defined the realized environmental space according to two significant axes of variation. The first two components explained 32% and 27% of variance among occurrences respectively (Fig. 3.4, B.3). Axis 1 is correlated most strongly with precipitation during driest month and quarter and mean diurnal temperature range (BIO14, BIO17, BIO2; Table B.6). Axis 2 is correlated most strongly with precipitation during the coldest and wettest quarters and mean temperature during the coldest quarter (BIO19, BIO16, BIO11). Niche centroids in this dataset differ slightly but significantly between ranges (between group inertia: 6.85%; P=0.001). Though the 99% confidence ellipses of the invaded range covers most of the climate space in the native range ellipse, the invaded niche appears to have shifted into more arid climates (towards lower values of PC1) and expanded significantly into habitats with a broader range of precipitation during cold and wet periods (expanded in both directions along PC2; Fig 3.4). After subdividing the data to include only populations within two standard  58  deviations of the mean of PC1 and PC2 (613 populations), the pattern of putative range shift and expansion remained (Fig. B.4), and niche centroids differed slightly more (between group inertia: 9.31%; P=0.001). This pattern may in part be due to the highly clumped nature of the GBIF occurrence data, which may oversample some locations (though note that proximity in climate space does not necessarily imply proximity in geographic space).  Figure 3.4 Putative climatic niche expansion as determined by principal components analysis of occurrence data in the native and invasive ranges of Centaurea diffusa. Shaded area represents 99% confidence ellipse for each range. Centroid of niche marked by large point.   59  3.4 Discussion 3.4.1 Phenotypic differentiation Assuming maternal effects do not strongly influence these results (discussed below), these experiments confirm previously reported rapid evolution between the native and invaded ranges of C. diffusa over a timescale of a century (Chapter 2). Under benign common conditions in a field experiment in the naturalized range of the species, C. diffusa consistently demonstrated morphological and life history trait differences with greater growth and vegetative biomass for invasive compared to native individuals. This differentiation between invasive and native individuals occurred even in the more realistic field setting, which allowed for biotic and soil interactions, and some degree of conspecific competition (although these were not experimentally controlled).  However, several important biological realities, such as maternal effects, could influence our results and these are discussed below. As reported in many other studies (Müller-Schärer et al. 2004; Williams 2009; Kumschick et al. 2013), invasive individuals grew faster and were thus larger than natives on average in this common garden study. In a previous study including these populations, invasive individuals were both larger and produced more seed in a common environment than natives (Chapter 2), a positive relationship that has been demonstrated between size and fitness for other monocarpic species (reviewed in Metcalf et al. 2003). Given that, this result demonstrates a potential for increased fitness, and therefore invasiveness, amongst North American populations. The populations included here are only a subset of all populations; it is therefore possible that the native source populations that generated the invasive populations are not included and that trait differences between the groups evolved prior to introduction. However, a greenhouse study that  60  included many more populations reported similar phenotypic differentiation as was observed here (Chapter 2). Ultimately, genotypic information is needed to establish the identity of the source populations for this invasion. Invasive individuals also exhibitied delayed maturity, a result that may suggest either an adaptation to a longer growing season, or a shift towards bienniality. Such a shift in reproductive strategies has already been documented in the facultative monocarp houndstongue Cynoglossum officinale L., where iteroparity was much more common in the introduced range than in the native range (Williams 2009). We did not detect any difference in mortality between ranges during the first growing season. If this result were to hold true after a second growing season, delayed maturity might tentatively represent a fitness increase in the invaded range. If delayed bolting means the plant is larger at maturity, then it has the potential to produce more flower heads, more seed, and more progeny. To test this hypothesis, additional experiments are needed to link demographic parameters measured at the individual level such as survival and growth to population dynamics.  Although we cannot rule out the existence of maternal effects in the present experiment, it is very unlikely that such effects are the only source of phenotypic differentiation between native and invasive C. diffusa. First, previous work reported similar results for seed collections from natural populations versus those derived from C. diffusa plants grown under common conditions in the greenhouse, thereby controlling for maternal effects (Chapter 2). Second, maternal effects in plants are mostly apparent for early traits in the life cycle such as during the germination stage (see Weiner et al. 1997 in Centaurea stoebe subsp. micranthos) and are thus unlikely to affect traits such as those measured at harvest for which we observed significant differentiation  61  between native and invasive populations. Additionally, results from the control treatment in this experiment are qualitatively similar to those of previous greenhouse experiments (Chapter 2). Altogether, these results suggest that the observed phenotypic differentiation has a genetic basis. 3.4.2 Drought response and resource allocation trade-offs Many attempts to explain invasion success have invoked trade-offs between growth or resource allocation and tolerance to stresses, which are characteristic of the native range, such as biotic (Blossey & Notzold 1995; Joshi & Vrieling 2005) or abiotic stresses (Bossdorf et al. 2005; He et al. 2010). For instance, a pattern consistent with a trade-off between growth and tolerance to drought stress has been shown in Ambrosia artemisiifolia (Hodgins & Rieseberg 2011). In terms of drought response, we found no evidence of variation in resource allocation and drought tolerance between the ranges in the present study, in contrast to Turner et al. (2014; Chapter 2). That said, only two of 13 measured traits demonstrated a significant effect of treatment, suggesting a limited impact of water addition on plants. Such limited effects may arise from the difficulty to control for water supply in an open field setting. For instance, it may be explained by sufficient natural rainfall during the duration of the experiment or insufficient separation between treatment rows in the experimental setup allowing for potential water flow from control to drought rows. Alternatively, trade-offs with drought stress reported previously might be an artifact of having conducted the drought evaluations in the greenhouse (e.g., plants might have become root bound) and therefore unlikely to materialize under field conditions.  62  3.4.3 Environmental variation, plasticity, and niche expansion Individuals from the native and invaded ranges of C. diffusa varied not just in phenotype, but in how that phenotype relates to abiotic environmental variation. The composite environmental covariate used here from the experimental PCA had greater explanatory power than latitude alone; latitude contributed only 5% to the variance of PC1 (not shown). Native populations demonstrated a significant relationship with these abiotic environmental conditions for several size and life history traits, which was typically non-significant in invasive populations. This may indicate that while native populations are locally adapted to their environments, invasive populations have yet to fully adapt to the invaded range, although the climatic niches inhabited by these populations largely overlap along the two main axes of climate variation in the PCA of experimental populations (Fig. 3.2, Table B.6). Alternatively, this may indicate that invasive populations have adapted along an axis of environmental conditions not seen in the native range. In fact, invasive phenotypes have a significant relationship to the environmental conditions used here for only two traits, bolting probability and wilting date, and for both traits, though the slope is weaker, the opposite trend was seen between ranges. Though niche-based distribution models assume that invasive species’ responses to environmental gradients (i.e. their ecological niche) are conserved between ranges (Peterson 2003), some studies suggest that responses can vary among the ranges (Fitzpatrick et al. 2007; Broennimann et al. 2007) though this is rare for terrestrial plant invaders (Petitpierre et al. 2012). The lack of local adaptation to environmental conditions seen in the invaded range of C. diffusa, a pattern which may be common in the genus (Hahn et al. 2012; Broennimann et al. 2014), contrasts sharply with the strong, adaptive, latitudinal or altitudinal clines in traits related to  63  growth, phenology, and life history that appear to be common in other introduced plants (Huey et al. 2000; Maron et al. 2004; Alexander et al. 2009; Colautti et al. 2009). Neither is it the case that C. diffusa was too recently introduced to develop a clinal relationship; for example a strong clinal pattern apparent in Lythrum salicaria, which was introduced at approximately the same time (Colautti et al. 2010). Novel abiotic conditions in the invaded range (such as those experienced in the areas of putative niche expansion seen in occurrence PCA for C. diffusa), biotic interactions (Keane & Crawley 2002), or genetic composition (Ellstrand & Schierenbeck 2000; Bossdorf et al. 2005; Taylor & Keller 2007) could alter or limit plant responses to similar environmental gradients between ranges (Alexander et al. 2009). There is little evidence that reductions in genetic variation have limited local adaptation in the invaded range since comparable genetic diversity exists in both ranges (Hufbauer & Sforza 2008). The lack of apparent local adaptation is consistent with the prominence of ‘general-purpose’ genotypes in the invaded range, which have a plastic, robust performance across environments and have not been selected to specialize in any particular environment (Baker 1965; Hahn et al. 2012).  Realized niche shifts cannot alone indicate adaptation into novel habitats (i.e. change in the fundamental niche) in the invaded range. Ordination analysis, though likely to quantify niche overlap more accurately overall than ecological niche modeling alone (Broennimann et al. 2014), brings with it several caveats (Guisan et al. 2014). First, occurrence data of the type used in the occurrence PCA (which does not include absence data) very likely underestimate the distribution of the species or alternately over-sample some areas in either range. It is possible that collection or reporting effort may vary between ranges, and therefore bias the results seen here. Second, this analysis makes no attempt to assess the availability of analog versus non-analog habitats  64  between the two ranges, and can therefore suggest only the possibility of the evolution of a climatic niche expansion. Finally, we can only assess the realized, not the fundamental niche of this species using occurrence data. Biotic interactions and dispersal may limit the realized niche in the native range, and these limiting factors may shape occurrences in the two ranges differently. However, coupled with evidence of genetically based phenotypic change, the putative shift in the realized niche of C. diffusa in the invaded range suggests that phenotypic change may have coincided with the evolution of increased physiological tolerance (Guisan et al. 2014). Though rare (Peitipierre et al. 2012), niche shifts or expansions have been demonstrated in some invasive plants. For example, some populations of Pinus taeda, grown in replicated common gardens outside its native range, were more invasive in climate niche spaces distinct from those of their native source range (Zenni et al. 2014a). Perhaps the best supported example of a realized niche expansion occurring in the invaded range of a plant is from the closely related spotted knapweed. Spotted knapweed (Centaurea stoebe subsp. micranthos) has expanded  its realized niche, demonstrated from  two replicated spatio-temporal invasion routes through North America, to eventually encompass wetter, drier, and warmer conditions than those experienced in the native range (Broennimann et al. 2014). 3.4.4 Hybridization and phenotypic plasticity Genetic changes induced by inter- or intra-specific hybridization have been hypothesized to promote invasiveness (Schierenbeck & Ellstrand 2009). Centaurea diffusa, a diploid, has a history of hybridization with diploid spotted knapweed, C. stoebe subsp. stoebe L., which is also native to eastern Europe (Blair & Hufbauer 2010; Blair et al. 2012; Lai et al. 2012; Mráz et al. 2012a). Though C. stoebe subsp. stoebe does not occur in North America (Treier et al. 2009;  65  Blair & Hufbauer 2010), the tetraploid form (C. stoebe subsp. micranthos [Gugler] Hayek, sometimes referred to as C. maculosa) has invaded the US with dramatic success. Hybrids between C. diffusa and C. stoebe sp. stoebe have been reported in both ranges (Blair et al. 2012; Lai et al. 2012).  The lack of reestablishment of adaptation to environmental conditions  among invasive populations could be the result of increased introgression from C. stoebe subsp. stoebe in the invaded range. Indeed, hybridization may play a role in the prominence of plastic, stress-tolerant ‘general-purpose’ genotypes (Schierenbeck & Ellstrand 2009; Blair et al. 2012; Parepa et al. 2014). Though the process of invasion alone, by exposing populations to extreme environmental changes, can result in the rapid evolutionary increase of plasticity in the early stages of invasion (Lande 2015; Bock et al. 2015), this benefit may be transient, and selection may then favor a locally adapted fixed phenotype if there is a cost associated with maintaining plasticity (Lande 2009; Lande 2015). Heterosis resulting from hybridization, however, is known to stabilize fitness across environments (Lippman & Zamir 2007; Bock et al. 2015). This type of stabilization could be observed as the loss of adaptation to environmental conditions in the invaded range and may also enhance invasiveness by providing an advantage over parental taxa (Burke & Arnold 2001). We do not know the level of introgression of the majority of populations used in this study (but see Table B.5). Further comparisons, including comprehensive genomic studies of admixture, are thus needed to assess the extent of introgression and its impact on the performance of invasive populations of C. diffusa. 3.4.5 Conclusion The invaded range of C. diffusa is dominated by larger individuals with delayed maturity and a more generalist relationship to climate, relative to the native range. While local adaptation in the  66  form of a resource allocation trade-off for drought tolerance is equivocal, local adaptation to abiotic climatic conditions is evident in the native range. However, invasive populations do not show such a relationship between phenotypic variation and climate. Instead, a plastic, generalist strategy may have been favored in the invaded range, resulting in the expansion of the species into a greater diversity of environments. This could make climatic niche-based predictive distribution models built on data from the native range potentially uninformative for this species (Broennimann et al. 2007). Future work will attempt to address the role of hybridization in the production of the hugely successful plastic phenotypes in the invaded range of C. diffusa.   67  Chapter 4: Differential Gene Expression with Drought in Centaurea diffusa  4.1 Introduction Rapid evolution in novel contexts can be readily observed in invasive species, and the literature includes many well-documented examples (Dlugosch & Parker 2008; Whitney & Gabler 2008; Hodgins & Rieseberg 2011; Buswell 2011; Felker-Quinn et al. 2013 for a few). Introduced individuals are exposed to novel biotic and abiotic conditions that can act as strong selective pressures, affecting fitness in the novel habitat and acting as barriers to invasion, unless the species is able to evolve to avoid demographic stasis or extinction (Guggisberg et al. 2013; Bock et al. 2015).  Some hypotheses invoke the evolution of trade-offs in resource allocation to explain phenotypic differences between the native and invaded ranges. If novel habitats offer a release from a selective pressure characteristic of the native range, either biotically, for example, due to absences of specialist herbivores (evolution of increased competitive ability (EICA); Blossey & Notzold 1995), or abiotically, for example when resources are abundant (Bossdorf et al. 2005; He et al. 2010), selection in the invaded range should favor individuals that shift resource allocation from stress tolerance to increased growth and reproduction, and presumably, invasiveness. The role of trade-offs in the invasion process has been the focus of several studies (Hodgins & Rieseberg 2011; Lachmuth et al. 2011; Kumschick et al. 2013; Chapter 2), but these attempts can be complicated by the variability of  strategy favored between different habitats (Lachmuth et al. 2011).   68  The genes that mechanistically underlie these trade-offs may share similar pathways, as in Arabidopsis thaliana, where alleles that increased pathogen resistance were also found to decrease leaf mass (Todesco et al. 2010). Mutations or expression levels that are advantageous in one environment maybe disadvantageous in another due to competing demands on energy supplies, such as the trade-off of nitrogen allocation between photosynthesis and defensive cell walls demonstrated in invasive Ageratina adenophora (Feng et al. 2009).  Yet adaptation to the novel habitat may not be necessary for a successful invasion. Rather than adapt to specific novel selection pressures, invasive species may instead benefit from a generalist strategy. Plastic, ‘general-purpose’ genotypes, perhaps fueled by adaptation to human–disturbed habitats in the native range (Lee & Gelembiuk 2008; Hufbauer et al. 2012; Mráz et al. 2012b), may perform similarly (i.e. well) in many environments (Baker 1965). Plasticity may allow an individual to maintain fitness in stressful or unfavorable environments and/or increase fitness in favorable environments (Richards et al. 2006) and is a causal driver of many invasions (Hahn et al. 2012; Zenni et al. 2014b; Bock et al. 2015). By exposing populations to extreme environmental changes, the process of invasion can result in the rapid evolutionary increase of plasticity in the early stages of invasion (Lande 2015). This benefit is typically transient if there is a cost associated with maintaining plasticity (Lande 2009; Lande 2015), though in some cases it may be fixed by heterosis (Lippman & Zamir 2007; Bock et al. 2015). The mechanistic underpinnings of such invasion hypotheses have only recently begun to be addressed (Stewart et al. 2009; Hodgins et al. 2013; Guggisberg et al. 2013). Extensive work has suggested rapid evolution based on phenotypic differences observed in common garden experiments, but few studies investigate invasions at the genomic level and therefore cannot  69  identify molecular mechanisms underlying emergence of new adaptive attributes. Of the studies that do, many focus on inferring levels of genetic diversity and the geographical origins of invasive source populations. By identifying molecular changes involved in rapid adaptation to novel environments – such as those involved in phenotypic plasticity, herbivore resistance, stress tolerance, or growth rate – we can clarify the role that each hypothesis may play in a given invasion, and eventually, its importance to invasion overall. Candidate genes can be identified by genomic approaches (such as identifying selective sweeps or genome-wide gene expression), even in non-model organisms. Gene expression studies may be especially useful to address resource allocation shifts in response to stress (Hodgins et al. 2013); gene expression differences have been associated with phenotypic differences between weedy and wild populations of the common sunflower (Helianthus annuus) in a common environment, caused by either a change in regulation or a loss of function mutation (Lai et al. 2008).  Here we investigate change in transcriptional response to drought stress as a possible mechanism underlying a shift in drought stress response previously reported (Chapter 2) between native and introduced populations of diffuse knapweed (Centaurea diffusa).  We also identify candidate genes underlying this shift, which may contribute to an invasive advantage for introduced populations in this species. Several previous experiments have documented phenotypic differences between native and invasive populations (Chapter 2, 3). In benign control environments, invasive individuals grow larger and delay maturity relative to native individuals. Some evidence suggests that this may be due to increased plasticity in the invaded range, and a loss of a trade-off between drought tolerance and size evident in the native range. To assess differences in gene expression between native and invasive populations of C. diffusa, we used  70  tissue from plants grown for a previous generation in a common environment, and then subjected to extreme drought stress and benign control conditions at three time points. We then used species-specific microarrays (Lai et al. 2012) to identify genes (1) showing a constitutive difference between ranges (whose expression varies by range across both treatments), and (2) showing an induced response (whose expression varies between ranges with treatment). We ask 1) What are the differences between native and invasive populations in response to drought stress? 2) Are these differences consistent with a trade-off in drought tolerance in the invaded range? This analysis will determine if there is evidence for expression differences between native and invasive populations that may have evolved during range expansion.  4.2 Materials and methods 4.2.1 Study species The genus Centaurea is one of only 15 plant genera in the United States to have more invasive species than expected by chance (Kuester et al. 2014) and is the most abundant noxious weed genus in western North America (LeJeune & Seastedt 2001). In invaded rangelands, diffuse knapweed (Centaurea diffusa Lam.) alters soil and water resource availability and reduces the quality of forage for livestock and wildlife (LeJeune & Seastedt 2001). A monocarpic biennial (Thompson & Stout 1991), C. diffusa is native to western Asia and eastern Europe and is now found throughout western Europe (Chapter 2; Greuter 2009). It was first reported in North America in 1907 in Washington State (Sheley et al. 1998) and currently occurs in roughly half of the United States and Canada (Chapter 2; USDA NCRS 2014).   71  Multiple common gardens (Chapter 2, 3) provide evidence for genetic differentiation in quantitative traits between native and invasive populations of C. diffusa. Across many environments, including benign greenhouse, open field, and stress treatments, including flooding, nutrient deficiency, and herbivory, invasive individuals grew larger and mature later than native individuals. Only in the case of drought stress is there evidence for a trade-off in tolerance for growth in the greenhouse (Chapter 2), though its importance in a more realistic field setting in unclear (Chapter 3).  This trade-off was, however, only observed for individuals from the native range. Invasive populations may have instead evolved increased phenotypic plasticity, thus allowing them to invade novel habitats (Chapter 3). In either the case of a drought tolerance trade-off or increased plasticity, invasive individuals should vary in regards to their expression profile from native individuals when exposed to drought. 4.2.2 Seed collection and production Seeds were collected in a broad collaborative effort (described in Chapter 2) from 3 native European populations and 3 invasive North American populations of C. diffusa from 2006 to 2008, and stored in dry, room-temperature conditions (Table C.3).  To rule out maternal environmental effects as a cause for differences observed between native and invasive populations, greenhouse generated seed was used in this microarray experiment. Field-collected seeds were germinated on filter paper in 1% plant preservative mixture and distilled H2O at room temperature. Within 10 days, seedlings were transplanted into 5 cm diameter cones filled with 80% potting mix and 20% silica sand and placed in a greenhouse.  After 10 weeks, two individuals from each of four or five maternal families for each population were transplanted into 1 l pots containing potting soil. Plants were watered with fertilizer for the first six weeks, and  72  then all were top dressed with Osmocote 13-13-13 and wand-watered from above. Plants that flowered between November 2009 and March 2010 were repeatedly reciprocally crossed with all individuals flowering simultaneously in the same population. Flower heads were bagged before anthesis and after hand pollination to prevent cross-fertilization. Progeny from these crosses were divided between this microarray experiment and the maternal common garden (Chapter 2) and were grown at the same time under the same conditions for both of these experiments. 4.2.3 Treatment and sampling Greenhouse-generated seeds were germinated and transplanted to cones as above in June 2010. Seedlings were randomly assigned to either a drought stress treatment or a control group. Grand-maternal lines are represented in both treatments as much as possible due to seed limitation. In total, 102 plants were included in this experiment (54 invasive and 48 native individuals). Each treatment group contained half of the plants, which were randomly assigned to a position within each block. Blocks were randomly ordered across a glasshouse bench and rotated randomly each week. Supplemental lighting maintained a 16 hr day. Treatment began at ~7 weeks after transplantation (plants had an average of 15 leaves larger than 3 cm). Plants were watered with fertilizer until the treatments began. Thereafter, control blocks were wand-watered from above with tap water. Drought blocks were no longer watered. Leaf tissue sampling occurred at three time points: pre-stress (time point 0: Jul 29, 2010, 5 days before treatment began), ~1.5 days after last watering (time point 1: Aug 4, 2010), and ~3.5 days after last watering (time point 2:Aug 6, 2010). For the pre-stress sampling, the 3rd newest completely unfurled leaf was clipped from each plant between 14:00 – 16:00, placed in paper envelopes, flash frozen in liquid N, and then stored in a -80C freezer.  For plants with smaller leaves, two  73  leaves of similar ages were taken to ensure sufficient leaf material for analysis. Later time points sampled the remaining completely unfurled young leaves.  Before extraction, leaf tissue samples from 102 plants were pooled into 3 pools per population-by-treament combination, with 3 individuals per pool, except for population RU008 which had only 2 plants in each pool. To minimize individual differences, which are prominent in natural populations, individuals were randomly assigned to pools; the focus of this study is variation at the range and population levels. 4.2.4 RNA extraction Using the TRIzol reagent (Invitrogen)/RNeasy (Qiagen) approach as described in Lai et al. (2006) and Hodgins et al. (2013), we extracted total RNA from each tissue pool. Concentration and quality was verified by Nanodrop ND-1000 to ensure that each sample had a concentration of >1.0 µg/µL, a A260/A280 > 1.8, and a A260/A230 > 1.8 (Lai et al. 2006). 4.2.5 cDNA preparation and hybridization In collaboration with NimbleGen, we developed a high-density customized expression 12-plex array with 60-mer probes from the transcriptomes of two C. diffusa individuals (see Lai et al. 2012 for details), primarily from one native individual (population TR001), with some contribution from one invasive individual (population US022). In total, 61,024 unigenes were represented on the array with two or three probes per unigene for a total of 136,906 features. We prepared the double-stranded cDNA according to the array manufacturer’s instructions (Roche NimbleGen; NimbleGen Arrays User’s Guide v.5.1). Following Hodgins et al. (2013), we added 1 µL SUPERase (Ambion) to the first-strand synthesis reaction to prevent RNA degradation. We  74  obtained cDNA yields with the Nanodrop and Qbit and assessed cDNA quality using the Agilent Bioanalyzer. We labeled cDNA and performed the hybridization on a 12-bay NimbleGen Hybridization System, using the array washes recommended by NimbleGen (NimbleGen Arrays User Guide v.5.1). Using a GenePix Professional 4200A scanner (Axon Instruments), we acquired individual array images independently adjusting the PMT gain for each image as recommended. Following this, we analyzed the images with the NimbleScan v.2.6 software (NimbleGen) and exported feature intensities as .xys files. 4.2.6 Preprocessing and analysis Each array was manually inspected and all artifacts larger than 8 X 8 features in size were removed, causing the removal of some features due to quality (41 out of 108 arrays were affected, with an average 7550 features removed per array) using NimbleScan v.2.6. The microarray design and experimental data for this experiment will be deposited in a public repository (such as ArrayExpress or Dryad) upon publication. As our method of analysis did not allow for missing data, we replaced the missing intensity scores using the k-nearest neighbor method implemented by the ‘impute’ package v.1.40.0 in R 3.1.1 (KNN method, 10 nearest neighbors; Troyanskaya et al. 2001; R Core Team 2014). We then preprocessed the microarray data using the ‘oligo’ package v.1.30.0 from Bioconductor (release 3.0; Carvalho & Irizarry 2010) in R, applying the robust multi-chip averaging method (‘oligo’ package v.1.30.0, Bolstad et al. 2003; Irizarry et al 2003a; Irizarry et al. 2003b). Specifically, the data were log2-transformed, and then the probe-level data for each microarray were background-corrected independently using a probabilistic model (Irizarry et al 2003a; Irizarry et al. 2003b). For  75  normalization and summarization, we applied the quantile normalization method followed by robust median averaging. We then used the ‘lme4.0’ package (Bates et al. 2014) to identify genes with significantly different expression over three time points and across two treatments between native and invasive populations. We employed restricted maximum likelihood (REML) models with random effects. Expression intensities for each gene were modeled as univariate responses, and models were fit using Gaussian distributions. Origin (native or invasive), treatment (control or drought), and their interaction were included as fixed effects in all full models. To assess the importance of environmental differences between sampled locations in gene expression levels, each full model included as a covariate a composite abiotic environmental variable determined by a principal component analysis (PCA) of altitude, latitude, and 18 bioclimatic variables of each seed collection location taken from the Worldclim database (Table B.6; Hijmans et al. 2005), following Chapter 3, but including only populations used in the microarray experiment. The principal component that explained the most variance among collection locations (PC1) was used in all models as the environment term (Table C.3). To analyze the data as repeated measures, sampling date and pool were included as random effects. Additionally, population (uniquely named) was also included as random effects in all models (e.g. Gene expression intensity ~ Origin * Treatment + Environment + (Time point | Pool ID) + (1 | Population)).  To assess the significance of each fixed effect and the random effect of population, we preformed likelihood ratio tests (LRT) on nested models, comparing full models with models missing one effect term or interaction. Because the effect of one variable depends on the condition of the other, interpretation of the significance of main effects that are included in  76  significant interactions during step-wise model simplification requires caution (Crawley 2012) and so genes with a significant interaction were removed from downstream analyses of genes with significant origin or treatment effects, following Hodgins et al. (2013). We then implemented the false discovery rate (FDR) procedure, using the ‘qvalue’ package v.1.40.0 to correct for multiple comparisons using an FDR cut-off value of 5% (Storey 2002). The resulting q-values are reported for every gene with at least one significant fixed effect term (Table C.4). This identified genes that were differentially expressed between the ranges or reacted to drought stress differentially across the ranges, as well as genes that had a treatment or environmental effect. For genes that had a significant effect of origin-by-treatment or origin, dendrograms were constructed to cluster samples using Euclidean distances and hierarchical clustering in order to identify putative subgroups among investigated samples, using the function heatmap.2 in the R package gplots v.2.16.0 (Warnes et al. 2015). Additionally, we functionally annotated all the genes represented in the microarray, and asked whether any Gene Ontology terms were over-represented among those genes with a significant effect of origin-by-treatment, origin, treatment, or environment. Following Hodgins et al. (2015), both transcriptomes used to design the microarray were annotated, and Gene Ontology terms assigned. In short, open reading frames were translated and annotated through BLASTP to the TAIR 10 database (The Arabidopsis Information Resource; arabidopsis.org). We assigned GO terms to each gene based on the GO A. thaliana mappings to the top hits and removed redundant GO terms. To identify the biological processes associated with genes identified as differentially expressed, we performed a GO enrichment analysis using topGO (Alexa & Rahnenfuhrer 2009). In separate analyses, all genes that were significantly differentially expressed in tests of origin- 77  by-treatment, origin, treatment, or PC1 were used as genes of interest, and all other expressed genes on the array were used as background. Significance for each individual GO-identifier was computed for all significant genes for a given model term (such as origin) with Fisher's exact test (α = 0.01). As GO terms are non-independent, we used the parent-child method that determines overrepresentation of terms in the context of annotations to the term's parents, thus reducing the dependencies between the individual terms, and avoiding false-positives and the need for further multiple comparison corrections (Grossmann et al. 2007; Alexa & Rahnenfuhrer 2009). Because origin effects were our main effect of interest, our analyses of GO terms for genes with a significant effect of origin-by-treatment or origin included all GO types (biological processes, molecular function, and cellular components). Analyses of genes with significant treatment or environment effect were limited to the type ‘biological processes’ as the set most likely to have implications for the evolution and ecology of this system.  To determine relative levels of expression, several subsets of genes were used in separate toGO analyses. Genes with a significant origin term were further divided into those that had higher expression levels on average for a range, using samples from time point 0, before the onset of treatment. These subsets did not include genes whose average expression was nearly equal between ranges (i.e. within 0.1 units). Similarly, genes with a significant interaction term were further divided into those that had higher expression levels on average for a range, using samples from the drought treatment at time point 2 when treatment effect was the strongest. Additional comparisons were made within each range for these genes, comparing drought treatment samples from time point 0 to time point 2. These subsets did not include genes whose average expression was nearly equal between time points.  78  4.3 Results 4.3.1 Differential expression Among the genes included on the array, expression of 585 genes showed a significant effect of origin after FDR correction (Table C.4). In other words, these genes showed constitutive differences in expression between native and invasive populations. Of genes with a significant origin effect overall, 311 genes had greater expression in invasive populations at time point 0, prior to onset of drought treatment, and 274 had greater expression in native populations. For this set of genes, dendrograms were constructed from all samples at time point 0, before the onset of treatment. In this dendrogram, samples cluster completely by range, and largely by population (Fig. 4.1).     79   Figure 4.1 Heat map of genes with constitutive expression differences between native and invasive Centaurea diffusa.  A heat map of the 585 constitutively differently expressed genes for all samples from time point 0, before onset of treatment. All genes had a significant effect of origin determined by mixed model analysis. Normalized, scaled, expression values are displayed. Heat maps were drawn in R with heatmap.2 and both the genes (rows) and samples (columns) were clustered using dendrograms. Dendrograms were constructed using Euclidean distances and hierarchical clustering. Origin is displayed in the colored bar above the heat map (red for native populations, and black for invasive).  80  The origin by treatment interaction was significant for 227 genes after the FDR correction indicating that expression levels for these genes responded differently to drought stress in native versus invasive population pools (Table C.4). For example, at time point 2 of the control treatment, 69 of these genes were more strongly expressed in the invasive than native populations. At the same time point in the drought treatment, 167 genes were more strongly expressed in the invasive relative to the native populations. For this set of genes, dendrograms were constructed for samples from time point 2, when effect of treatment would be strongest. When both treatments were included in a single dendrogram, samples clustered largely, but not completely, by treatment (Fig. 4.2). When split into separate treatments, samples largely cluster by range, with some exceptions. In the drought treatment, two samples of invasive population US001 cluster within the native range samples, close to samples from populations in Turkey and Bulgaria (Fig. C.1). Among control treatment samples, all invasive populations cluster within a branch that contains samples from native populations in Bulgaria, Turkey, and Russia (Fig. C.2).         81                Figure 4.2 Heat map of genes with drought induced expression differences between native and invasive Centaurea diffusa.  A heat map of the 227 induced differently expressed genes for all samples from time point 2. All genes had a significant origin by treatment interaction, determined by mixed model analysis. Normalized, scaled, expression values are displayed. Heat maps were drawn in R with heatmap.2 and both the genes (rows) and samples (columns) were clustered using dendrograms. Dendrograms were constructed using Euclidean distances and hierarchical clustering. Origin is displayed in the colored bar above the heat map (red for native populations, and black for invasive). Treatment is indicated by colored bar below sample name (blue for control, white for drought).  82  Additional genes were found to be significant for the remaining model terms (Table C.4). Among genes that did not have a significant interaction between origin and treatment, 9578 showed a significant effect of treatment. Environment of the collection location (PC1; Table C.3) was a significant covariate for 1111 genes. Random effect of population was significant for 9988 genes (not shown). 4.3.2 Putative function of differentially expressed genes Of the 61,024 genes on the array, 37,373 could be annotated to ‘biological processes’ (BP) GO terms based on mappings to the A. thaliana TAIR 10 database, 37,190 to ‘molecular function’ (MF), and 33,904 to ‘cellular component’ (CC) GO terms, respectively. Of the 585 genes with a significant origin effect, most could be mapped to GO annotations (369 to BP terms, 365 to MF terms, and 340 to CC terms). When this set was analyzed for GO term enrichment, 25 GO terms were significantly overrepresented (Table 4.1), including several metabolic and biosynthetic processes. To highlight the constitutive differences between native and invasive expression profiles, genes that had higher expression levels in invasive relative to native populations among time point 0 samples (pre-drought treatment, including all samples) were then analyzed separately. For genes in this set that had greater expression in invasive populations, enriched GO terms among pre-treatment samples included flavonoid and L-ascorbic acid biosynthesis, mitochondrion organization, and photosynthetic components (14 GO terms, Table 4.1). Enriched GO terms for genes in this set that had greater expression in native populations among pre-treatment samples included signal transduction, ethylene-activated signaling pathway, and nucleobase transportation (9 GO terms, Table 4.1). Some GO terms were associated both with  83  genes that greater expression and genes that had lesser expression in a given range (2 GO terms, indicated as ‘both’ in Table 4.1). Table 4.1 Gene Ontology terms for genes with a significant effect of origin. GO ID Term P-value Range comparison, Time pt 0 Biological processes   GO:0009812 flavonoid metabolic process < 0.0001 both GO:0007165 signal transduction 0.00034 Native GO:0043086 negative regulation of catalytic activity 0.00098 Invasive GO:0009813 flavonoid biosynthetic process 0.0015 Invasive GO:0019853 L-ascorbic acid biosynthetic process 0.0016 Invasive GO:0019852 L-ascorbic acid metabolic process 0.00198 Invasive GO:0032989 cellular component morphogenesis 0.00219 Native GO:0009690 cytokinin metabolic process 0.00366 Native GO:0009873 ethylene-activated signaling pathway 0.00425 Native GO:0071369 cellular response to ethylene stimulus 0.00429 Native GO:0044723 single-organism carbohydrate metabolic process 0.00511 Invasive GO:0015851 nucleobase transport 0.00655 Native GO:0007005 mitochondrion organization 0.00726 Invasive GO:0006040 amino sugar metabolic process 0.00743 Invasive GO:0009699 phenylpropanoid biosynthetic process 0.00927 Native Molecular function   GO:0016705 oxidoreductase activity, acting on paired donors, with incorporation or reduction of molecular oxygen 0.00015 both GO:0004022 alcohol dehydrogenase (NAD) activity 0.00108 Native GO:0008233 peptidase activity 0.00446 Invasive GO:0051213 dioxygenase activity 0.00512 Invasive GO:0015932 nucleobase-containing compound  transmembrane transporter activity 0.00964 Native Cellular component   GO:0030095 chloroplast photosystem II 0.00054 Invasive GO:0009654 photosystem II oxygen evolving complex 0.00112 Invasive GO:0009521 photosystem 0.00326 Invasive GO:0000315 organellar large ribosomal subunit 0.00484 Invasive GO:0000311 plastid large ribosomal subunit 0.00926 Invasive Results of Fisher’s exact tests examining the number of gene associations with GO terms are presented. Genes of interest were determined in mixed model analysis to have a significant constitutive differences between native and invasive populations. Ontology type is indicated for each set of terms (biological process , molecular function, cellular component). The analysis was conducted using the Fisher algorithm and parent-child method in topGO, and P-values for this test are reported. Range level comparison of expression was determined by a GO term appearing in separate topGO analyses of the subset of genes which had higher average expression levels in either range for all time point 0 samples, before onset of treatment. If a GO term was overrepresented in both ‘higher expression in invasive samples’ and ‘higher expression in native samples’ subsets, it is listed as ‘both’.  Of the 227 genes with a significant origin-by-treatment interaction, most could be mapped to GO annotations and used in the functional analysis (180 to BP terms, 187 to MF terms, and 178 to  84  CC terms). When this set was analyzed for GO term enrichment, 74 GO terms were significantly overrepresented (Table 4.2). We found that this set contained many basic cellular processes, such as translation, gene expression, and ribosomal activity. It also included response to osmotic stress, indicating the possibility of a trade-off in drought tolerance between native and invasive populations. To highlight the differences between native and invasive response to drought stress, genes with a significant interaction between origin and treatment were divided into sets of greater and lesser expressed genes between the ranges for drought samples and time point 2 and analyzed for GO term enrichment separately. For genes that had greater expression in invasive relative to native populations, enriched GO terms included translation, gene expression, macromolecule biosynthesis and metabolism, ribosomal components, mitochondrial membrane protein organization and transport, ncRNA processing and metabolism, and response to osmotic stress (70 GO terms, Table 4.2). For the set of genes which had greater expression in native populations among time point 2 drought treatment samples, enriched GO terms included glycolipid biosynthesis and metabolism, alanine-glyoxylate transaminase activity, and vacuole and associated memberane (5 GO terms, Table 4.2). Some GO terms were associated both with genes that greater expression and genes that had lesser expression between the two ranges (3 GO terms, indicated as ‘both’ in Table 4.2). For the 227 genes with a significant origin-by-treatment effect, more genes decreased in expression and fewer genes increased between time point 0 (before stress) and time point 2 (after stress) in invasive than in native samples (68 genes increased and 127 decreased in invasive samples, 81 genes increased and 88 genes decreased in native samples, and some genes did not vary by more than 0.1 unit and were considered nearly equal). When comparing drought  85  treatment samples within each range, the overall pattern for most GO terms associated with genes with a significant origin-by-treatment effect was decreased expression in response to drought in both ranges  (65 GO terms in invasive samples, 47 GO terms in native sample; Table 4.2).  Of the 65 GO terms with decreased expression in invasive samples, all but six still had higher expression in invasive samples compared to native samples. For the 47 GO terms that demonstrated decreased expression in response to drought in both ranges, expression levels were lower in native samples for all but one term (intracellular organelle; Table 4.2).   Some GO terms were associated both with genes that were down-regulated and genes that were up-regulated in response to drought (18 GO terms in both ranges, indicated as ‘both’ in Table 4.2).  Table 4.2 Gene Ontology terms for genes with significant origin by treatment interaction. GO ID Term P-value Range comparison, Drought, Time pt 2. Higher in… Change over time, Invasive drought samples Change over time, Native drought samples Biological processes     GO:0006412 translation < 0.0001 Invasive down both GO:0010467 gene expression < 0.0001 Invasive down down GO:0009059 macromolecule biosynthetic process < 0.0001 Invasive down down GO:0019538 protein metabolic process < 0.0001 Invasive down down GO:0044267 cellular protein metabolic process < 0.0001 Invasive down down GO:0034645 cellular macromolecule biosynthetic process < 0.0001 Invasive down down GO:0022613 ribonucleoprotein complex biogenesis < 0.0001 Invasive down down GO:0044249 cellular biosynthetic process < 0.0001 Invasive down down GO:0009058 biosynthetic process < 0.0001 Invasive down down GO:0008152 metabolic process < 0.0001 both both both GO:0044260 cellular macromolecule metabolic process < 0.0001 Invasive down down GO:0071840 cellular component organization or biogenesis < 0.0001 Invasive down down GO:1901576 organic substance biosynthetic process < 0.0001 Invasive down down GO:0009987 cellular process < 0.0001 Invasive down both GO:0044237 cellular metabolic process < 0.0001 Invasive down both GO:0045039 protein import into mitochondrial inner membrane < 0.0001 Invasive down down GO:0043170 macromolecule metabolic process < 0.0001 Invasive down down GO:0044085 cellular component biogenesis < 0.0001 Invasive down down GO:0090151 establishment of protein localization to mitochondrial membrane < 0.0001 Invasive down down  86  GO ID Term P-value Range comparison, Drought, Time pt 2. Higher in… Change over time, Invasive drought samples Change over time, Native drought samples GO:0007006 mitochondrial membrane organization 0.00014 Invasive down down GO:0044238 primary metabolic process 0.00015 Invasive both up GO:0034470 ncRNA processing 0.0005 Invasive down both GO:0006626 protein targeting to mitochondrion 0.00093 Invasive down down GO:0034660 ncRNA metabolic process 0.00148 Invasive down down GO:0070585 protein localization to mitochondrion 0.00216 Invasive down down GO:0072655 establishment of protein localization to mitochondrion 0.0025 Invasive down down GO:0071704 organic substance metabolic process 0.00413 Invasive up up GO:0009247 glycolipid biosynthetic process 0.00439 Native down both GO:0065002 intracellular protein transmembrane transport 0.0055 Invasive down down GO:0071806 protein transmembrane transport 0.00705 Invasive down down GO:0006664 glycolipid metabolic process 0.00734 Native down both GO:0044802 single-organism membrane organization 0.00806 Invasive down down GO:0006970 response to osmotic stress 0.00834 Invasive   Molecular function     GO:0005198 structural molecule activity < 0.0001 Invasive down both GO:0003735 structural constituent of ribosome < 0.0001 Invasive down down GO:0003723 RNA binding < 0.0001 Invasive both both GO:0016675 oxidoreductase activity, acting on a heme group of donors < 0.0001 Invasive   GO:0015002 heme-copper terminal oxidase activity < 0.0001 Invasive   GO:0004129 cytochrome-c oxidase activity < 0.0001 Invasive   GO:0008453 alanine-glyoxylate transaminase activity 0.0023 Native up up Cellular component     GO:0044445 cytosolic part < 0.0001 Invasive down both GO:0044391 ribosomal subunit < 0.0001 Invasive down down GO:0030529 ribonucleoprotein complex < 0.0001 Invasive down both GO:0005840 ribosome < 0.0001 Invasive down both GO:0043228 non-membrane-bounded organelle < 0.0001 Invasive down both GO:0043232 intracellular non-membrane-bounded organelle < 0.0001 Invasive down both GO:0032991 macromolecular complex < 0.0001 Invasive down both GO:0044422 organelle part < 0.0001 Invasive down down GO:0005829 cytosol < 0.0001 Invasive down both GO:0043226 organelle < 0.0001 Invasive down both GO:0044446 intracellular organelle part < 0.0001 Invasive down down GO:0022626 cytosolic ribosome < 0.0001 Invasive down down GO:0031974 membrane-enclosed lumen < 0.0001 Invasive down down GO:0030054 cell junction < 0.0001 Invasive down down GO:0055044 symplast < 0.0001 Invasive down down GO:0044424 intracellular part < 0.0001 Invasive down down GO:0044437 vacuolar part < 0.0001 both down up GO:0016020 membrane < 0.0001 Invasive down down GO:0005622 intracellular < 0.0001 Invasive down down GO:0005773 vacuole < 0.0001 Native down up GO:0044464 cell part < 0.0001 Invasive down down GO:0005623 cell < 0.0001 Invasive down down GO:0031090 organelle membrane < 0.0001 Invasive down down  87  GO ID Term P-value Range comparison, Drought, Time pt 2. Higher in… Change over time, Invasive drought samples Change over time, Native drought samples GO:0005740 mitochondrial envelope 0.00011 Invasive down down GO:0005774 vacuolar membrane 0.00017 Native down up GO:0044428 nuclear part 0.00165 Invasive down down GO:0043229 intracellular organelle 0.00177 both down down GO:0031978 plastid thylakoid lumen 0.00388 Invasive down down GO:0009543 chloroplast thylakoid lumen 0.00413 Invasive down down GO:0031970 organelle envelope lumen 0.00483 Invasive down down GO:0005737 cytoplasm 0.00565 Invasive down down GO:0005758 mitochondrial intermembrane space 0.00667 Invasive down down GO:0005852 eukaryotic translation initiation factor 3 complex 0.00675 Invasive  up GO:0044444 cytoplasmic part 0.00803 Invasive down down Results of Fisher’s exact tests examining the number of gene associations with GO. Genes of interest were determined in mixed model analysis to have a significant origin effects induced by drought treatment. Ontology type is indicated for each set of terms (biological process, molecular function , cellular component). The analysis was conducted using the Fisher algorithm and parent-child method in topGO, and P-values for this test are reported.  Range level and time course comparison of expression were determined by a GO term appearing in the separate topGO analyses of the subsets of genes which had higher average expression levels in drought treatment samples, either between ranges, or between timepoint witing a range. If a GO term was overrepresented in both subsets in a given comparison, it is listed as ‘both’. If a GO term did not appear within the top 100 terms in either subset within a given comparison, it is left blank.   In addition to identifying differences between native and invasive populations of C. diffusa, this dataset can contribute initial clues to some of the basic biology in this and related systems. An understanding of the traits and genes that confer adaptation to abiotic environments, and particularly drought stress, is need for many applied and theoretical questions in ecology and evolution (as in the case of domesticated crops, invasive species, and conservation in the face of climate change).  Of the 9578 genes with a significant treatment effect, 6480 had ‘biological processes’ GO annotations (Table C.4). When this set was analyzed for GO term enrichment, 236 GO terms were significantly over represented (Table C.5). As expected, this set included many regulation related terms, and response to stimuli, such as desiccation, heat, light, and osmotic stress. Of the 1111 genes with a significant environmental covariate (PC1), 734 had ‘biological processes’ GO annotations (Table C.4). When this set was analyzed for GO term  88  enrichment, 22 GO terms were significantly over represented (Table C.6). We found that this set included many metabolic processes, and as well as organ senescence, aging, and protein repair.  4.4 Discussion To complement studies of phenotypic variation between native and invasive populations, this work investigates the genomic basis of invasiveness in C. diffusa, a developing topic in invasion biology (Stewart et al. 2009). We searched for expression differences between native and invasive populations of C. diffusa under both control and drought stress, with the goal of identifying gene networks and candidate genes that might contribute to invasiveness. Phenotypic differences were previously found to exist between these same native and invasive populations grown at the same time under the same conditions (maternal common garden, Chapter 2), including differences in growth, days to maturity, and drought tolerance. Many of the candidate genes identified here may potentially underlie the observed growth differences. Specifically we found 585 genes with constitutive differences in expression between the ranges, and a further 227 genes that responded to the drought treatment differently between the native and invasive populations (i.e. had a significant interaction between origin and treatment). These differences in expression were found to be significant after accounting for maternal environment, climate variation among collection locations, and the random effects of population, sampling time, and the effect of repeated sampling (i.e. wounding).  89  4.4.1 Variation in gene expression Whether the differences in gene expression reported here, and phenotype reported previously (Chapter 2, 3), are driven by directional selection during or after invasion or reflect the colonization history of diffuse knapweed are yet to be determined. Population genetic surveys indicate that C. diffusa was introduced to North America multiple times and harbors comparable levels of diversity at nuclear loci to the native range (Hufbauer & Sforza 2008, Marrs et al. 2008). As such, founder effects and population bottlenecks probably do not play a large role in the expression differences observed here. When invasive populations did not cluster completely in expression dendrograms, some invasive populations clustered with populations from Turkey, Bulgaria, and Russia. Turkey has already been suggested as a source location for invasive populations, based on chloroplast marker data (Hufbauer & Sforza 2008). Bulgaria has also been implicated, based on historical records; C. diffusa was first recorded in North America as a Bulgarian alfalfa seed contaminant. The Russian population used here is also from the region surrounding the Black Sea; populations from this region are nearly panmictic (Marrs et al. 2008). Though we found several genes that varied between ranges, this may not represent the whole picture. Limited population sampling within each range (3 native and 3 invasive populations) could reduce our capacity to detect associations given the high level of technical and biological variation typically found in microarray studies (Whitehead & Crawford 2006). The timing and location of sampling may also effect results, as we sampled leaves of plants well before reproductive maturity, and so expression differences associated with reproduction (such as bolting date) may not be apparent.  90  Because of greenhouse-produced seed limitation, and the chosen random pooling strategy, we can only address variation partitioned among ranges, and to a lesser extent, among populations, with this dataset. However, it is well known that substantial variation exists within populations (Hamrick & Godt 1996), as has been demonstrated in weedy Asteraceae (Lai et al. 2008; Hodgins et al. 2013). Because of seed limitation, in some cases it was necessary to pool individuals from the same grand maternal line into different pools, meaning that these biological replicates are not truly independent of one another, and could artificially reduce variation among pools. However, our primary level of interest was among regions. To address variation within populations, pooling samples would be undesirable. 4.4.2 Functional analysis Though a popular and fast method to characterize functional information about genes that can be quantified and compared across genomes, GO annotation is not without challenges (discussed in Hodgins et al. 2013), especially in non-model organisms where annotations are typically based on homology to a model species, as we have done here. As such, this analysis represents a coarse investigation into the molecular basis of ecologically important differences between native and invasive populations of C. diffusa. None of the enriched GO terms identified here for genes with a significant origin-by-treatment or origin effect were identified as rapidly evolving in invasive C. diffusa in a previous analysis (Hodgins et al. 2015), which used the same transcriptomes from which our microarrays were constructed (Lai et al. 2012). Given these necessary caveats, we discuss the potential function for some selected differentially expressed genes below. Many constitutive differences are apparent between native and invasive populations of C. diffusa. Genes involved in both signal transduction and nucleobase (including RNA) cross- 91  membrane transport are expressed at a constitutively lower level in invasive populations, which may affect the speed with which invasive individuals can respond to environmental changes. In particular, the expression of the ethylene response pathway is decreased in invasive populations, suggesting that invasive populations may react less quickly to both external and internal ethylene producing events. Ethylene is a plant hormone that can be induced through environmental stresses such as drought and wounding (Yang & Hoffman 1984) but is also produced endogenously to induce fruit ripening and flowering. Delayed maturity has been observed in invasive C. diffusa (Chapter 2, 3); this reduced response to the ethylene pathway may be the underlying genetic mechanism. Importantly for drought response, ethylene also induces leaf abscission (to reduce water loss) and root hair growth (to increase water uptake), suggesting that reduced drought tolerance and delayed maturity in invasive populations may have a common cause.  Invasive populations have increased constitutive expression for genes relating to both mitochondrial and chloroplast activity. Several significant GO terms that have greater expression in invasive samples relate to photosynthetic activity (Table 4.1). Invasive populations also have increased expression of genes involved in mitochondrion organization (the assembly, replication, morphogenesis and distribution of mitochondria), suggesting that invasive individuals have increased mitochondrial activity/respiration.  Both of these changes may contribute to the rapid growth rate of invasive populations.  We speculate that this combination of a possible delay in response to environmental changes and an increase in cellular energy production could result in a less environmentally sensitive, but potentially more robust phenotype, relative to the native range.  92  Across both ranges, drought treatment had a significant effect on many genes (9578 genes; Table C.4), though only roughly two thirds could be mapped to ‘biological processes’ GO annotations using sequence similarity to Arabidopsis thaliana (6480 genes; Table C.4). This dataset can therefore contribute initial clues to the biological role of these genes in a species where functional work is limited. The drought treatment also induced differences between the ranges for many genes. Genes that demonstrate a significant interaction between origin and treatment could have important effects on any environment specific differences in phenotype including those resulting from trade-offs in performance. The majority of these genes had greater expression in the invasive rather than the native populations, even though most of these genes decreased in response to drought in samples from both ranges. This may point to a weaker drought response in invasive populations, which fail to turn down expression relative to time point 0 as much as native populations. Phenotypic evidence suggests that while native populations demonstrate a trade-off between size and drought tolerance, invasive populations have lost this trade-off (Chapter 2, Fig 2.3). If invasive individuals are less responsive to drought, then they may ‘ignore’ mild short term fluctuations in water availability, which native individuals would react to with strong changes in expression levels, only to turn around and return them to ‘normal’ levels once water became available. Because invasive individuals react less strongly, they may maintain biomass production, until the drought stress becomes too extreme. In this way they can maintain fitness across moderate environmental variation (i.e. fitness homeostasis, or Jack-of-all-trades adaptive plasticity; Richards et al. 2006). In this case, the ‘general-purpose’ genotype will have limits, and may only be beneficial under cases of moderate rather than extreme stress. This pattern of greater expression in invasive drought samples than in native samples  may also be reflected in the enrichment of regulatory GO terms  93  such as gene expression, translation, non-coding RNA processing and metabolism, and ribosome activity, all of which had greater expression in drought samples from invasive rather than native populations. Though some non-coding RNAs have been linked with stress response, and particularly drought response in plants, the role they play is not yet clear (Guleria et al. 2012; Hackenberg et al. 2014; Zhang et al. 2014). In invasive populations, drought treatment maintained more cellular activity relating to their constitutively high levels of mitochondria-related gene expression. Enriched GO terms for genes with a significant interaction between origin and treatment, which had greater expression in the invasive samples from the drought treatment time point 2, included protein import into mitochondrial inner membrane, establishment of protein localization to mitochondria and mitochondrial membranes, mitochondrial membrane organization, protein targeting and localization to mitochondrion, and general intracellular transmembrane transport of proteins. This may indicate higher respiration levels even during drought stress in invasive compared to native populations. Leaf mitochondrial respiration typically decreases in response to drought stress, though in some plants, including sunflower (Helianthus annuus, Asteraceae), a bi-phasic response is observed (Ghashghaie et al. 2001; Atkin & Macharel 2009). In these species, leaf mitochondrial respiration initially decreases in response to water limitation but increases after prolonged or more severe water stress (Flexas et al. 2005; Atkin & Macharel 2009).  Changes in respiration in response to drought have been linked to overall drought tolerance. Mediterranean plant species with a lower water stress resistance have also been shown to have a low resistance to cell membrane destabilization induced by drought, which leads to increased respiratory flux in order to sustain a higher plant maintenance costs (Varone & Grantani 2015). Species with lower  94  drought tolerance may suffer from significant metabolic damage and electrolyte leakage under severe water stress, which increases the demand for respiratory products (such as ATP and reducing equivalents) and may force the use of stored energy materials (such as sucrose and starch), a highly energetically expensive process (Butler & Landsberg 1981; Varone & Grantani 2015). This, in turn, increases respiratory activity to meet the need for higher ATP levels (Rizhsky et al.,2002; Atkin & Macherel 2009).  Like the related sunflower, C. diffusa may have a bi-phasic respiratory response to drought stress (Ghashghaie et al. 2001). If this is the case, our gene expression results would be consistent with an earlier occurrence of the increased respiratory phase occurring in invasive relative to native populations. If invasive individuals are less responsive to drought, then less ameliorative activity occurs, and severe water stress may occur more rapidly. In addition to maintaining expression of genes related to mitochondrial function, invasive populations also had greater expression of genes related to osmotic stress response than native populations. Lower drought tolerance in terms of time to wilt and time to death has already been observed in invasive populations (Chapter 2), and these data may indicate the underlying cellular mechanism.  The native individuals, by contrast, responded to the drought treatment with relatively higher levels of biosynthesis and metabolism of glycolipids. Quantitative changes in cellular membrane lipids, such as the level of glycolipid saturation, that occur due to water stress could affect membrane fluidity (Quartacci et al. 2002). In drought tolerant grape cultivars, drought stress can induce an increase in total lipid content and reduce fatty acid saturation and length (Toumi et al. 2008). In particular, changes to the ratio of glycolipids may be a cellular mechanism that contributes to the maintenance of membrane fluidity and stability under drought stress in tolerant  95  genotypes (Toumi et al. 2008; Quartacci et al. 1995). In C. diffusa, this increased level of glycolipid biosynthesis and metabolism may contribute to higher drought tolerance in native populations. Native populations, unlike invasive, also increased expression of genes relating to the vacuole, which may play an important role in drought response. By increasing the ion concentration in the vacuole, for example, the vacuole membrane transporters can function in osmotic homeostasis (Zhu 2002). We have identified many genes that have constitutive differences in expression between the native and invaded ranges of C. diffusa and more that demonstrate induced differences upon exposure to drought. With these differentially expressed genes as candidates, future investigations can identify genetic changes that may have occurred during the invasion of North America by C. diffusa and further evaluate the molecular and biological role of the candidate weedy genes.   96  Chapter 5: Conclusion  Centaurea diffusa has rapidly evolved over the last century to be one of the most problematic invasive plant species in western North America. The studies included in this dissertation have contributed to our understanding of the phenotypic and genetic changes that have made invasive C. diffusa so formidable. Here, I revisit the primary research questions outlined in the Introduction and discuss how this work has contributed to answering them. Question 1: Do more invasive genotypes evolve in the invaded range of an invasive species? In Chapters 2 and 3, I found that under benign conditions, phenotypic differences in growth and reproduction have arisen between native and invasive populations of C. diffusa. These phenotypic differences are observed in a common environment, implying that they are the result of genetic differentiation between the two ranges, which may have evolved since introduction. Support for this conclusion is strengthened by efforts to control for several possibly confounding non-genetic effects, including a large and broad sampling effort, maternal environment, population latitude, variation in experimental context, and across multiple common gardens, the influences of which are infrequently considered in similar studies (but see Monty et al. 2009; Colautti et al. 2009; Hodgins & Rieseberg 2011; Rypel 2014). Chapter 2 in particular represents one of the highest levels of population replication in similar studies of invasive plants to date (but see Kumschick et al. 2013). Rapid phenotypic change, as demonstrated here in C. diffusa, may be common in invasive relative to native populations, particularly in the anthropogenic context in which invasive populations are commonly found (Hendry et al. 2008; Hufbauer et al.  97  2012). For example, Buswell et al. (2011) found significantly more morphological changes in introduced relative to native Australian species using herbarium specimens spanning 150 years.  Although in the absence of genotypic information, these phenotypic data support a conclusion of rapid evolution in the invaded range of C. diffusa. Question 2: Do trade-offs between stress tolerance and growth/reproduction enhance the invasiveness of a species?  In Chapters 2 - 4, I investigated variation between native and invasive populations of C.diffusa induced by stress. Tested stresses include both biotic (simulated herbivory and specialist herbivore leaf choice) and abiotic stresses (extreme drought, flood, nutrient deficiency), with drought providing the most convincing evidence of a physiological trade-off in native, but not invasive populations. Chapter 3 investigated the response to moderate water limitation in a field setting, though the treatment effect was not readily apparent, suggesting the applied stress was insufficient for phenotypic response. In the literature, equivocal evidence has been found to support the most well tested of the trade-off hypotheses, the evolution of increased competitive ability (EICA; Blossey & Nötzold 1995; Joshi & Vrieling 2005). For example, invasive individuals of Ageratina adenophora, a subtropical noxious forb in the sunflower family, have evolved increased nitrogen allocation to photosynthesis and growth, and reduced nitrogen allocation to cell walls, resulting in weaker structural herbivore defense (Feng et al. 2009). Though invasive individuals frequently out-perform native individuals in common environments, shifts in defenses are less common (Kumschick et al. 2013). EICA is supported only if increases in plant growth co-occur with decreases in defense; thus, the work presented here does not  98  support EICA in this system, and evidence for EICA in general remains equivocal (Felker-Quinn et al. 2013). Abiotic stress tolerance trade-offs have previously been demonstrated in invasion contexts. For example, in common ragweed, drought tolerance comes at the expense of competitive ability and growth rate (Hodgins & Rieseberg 2011). Similarly, the findings presented here suggest that, at least among native populations of C. diffusa, increased growth rate comes at the expense of reduced drought tolerance (Chapter 2). However, among invasive populations, while their mean drought tolerance is lower than native populations, they do not exhibit the same trade-off between drought tolerance and size in benign conditions. Among invasive populations, increased growth rate does not result in decreased drought tolerance, if anything the trend is in the opposite direction (Fig. 2.3a).  Question 3: Do native and invasive individuals of a species relate differently to the environmental conditions they experience? In Chapter 3 I look more closely at the relationship between performance in a benign environment and the climate of a populations’ home location. Some of the best evidence for rapid evolution in the invaded range comes from comparative studies of native and invasive populations that demonstrate parallel variation along geographic clines. For example, extensive work has demonstrated that physiological and genetic changes in invasive Drosophila populations have evolved parallel to native ones across geographic, altitudinal and latitudinal gradients (Huey et al. 2000, Hoffmann et al. 2002, Gilchrist et al. 2004, Gilchrist et al. 2008; Folguera et al. 2008), though new evidence indicates that this pattern may be enhanced by multiple waves of migration (Kao et al. 2015). Similarly, changes in growth and reproductive  99  traits along altitudinal gradients were found among native and invasive populations in several species of Asteraceae, although trait variation was also influenced by environmental context resulting in strong local adaptation (Alexander et al. 2009). Surprisingly, in the study presented in Chapter 3, I found that while native populations of C. diffusa demonstrated strong correlations between phenotype and their home climatic and geographic environment, invasive populations had lost this relationship to their environment. Indeed, the slopes with respect to the major axis of environmental variation differed significantly between invasive and native populations for most traits measured (Table B.8) and were often opposite in sign (Fig. 3.3). Adaptive plasticity has played an important role in many invasions by allowing a genotype to buffer its fitness across environmental variation (i.e. fitness homeostasis, or Jack-of-all-trades plasticity; Richards et al. 2006, Hahn et al. 2012; Zenni et al. 2014b; Bock et al. 2015). In C. diffusa, adaptive plasticity may enable the expansion of the environmental tolerances realized in the invaded range, a pattern observed in some other invasive plants including the closely related Centaurea stoebe subsp. micranthos (Hahn et al. 2012; Broennimann et al. 2014). Over the course of two distinct North America invasion routes, C. stoebe subsp. micranthos has expanded its realized niche to encompass wetter, drier, and warmer conditions than those experienced in the native range (Broennimann et al. 2014). In Pinus taeda, some populations grown in replicated common gardens outside its native range are more invasive in climate niche spaces distinct from those of their native source range (Zenni et al. 2014a). Yet at least among Holarctic invasive plants, evidence of invasive populations (potentially enabled by adaptive plasticity) spreading to environments outside of those experienced in the native range, i.e. shifts in either the realized or fundamental niche, is rare (Petitpierre et al. 2012, but see Webber et al. 2012).  100  While we now have some evidence that North America has been invaded by genotypes of C. diffusa with enhanced adaptive plasticity, further work is needed to establish the source of this increased plasticity, and whether it arose before or after introduction. Question 4: What are the genetic mechanisms that underlie phenotypic and stress tolerance differences between native and invasive individuals? In Chapter 4, I use genome-wide transcriptional profiling to identify loci that are differentially expressed between native and invasive genotypes, an approach useful to explain phenotypic variation, physiological trade-offs, and the origin of diversity and adaptation (Whitehead & Crawford 2006; Lai et al. 2008; Stewart et al. 2009; Hodgins et al. 2013). For example, microarray studies of weedy and wild populations of common sunflower (Helianthus annuus) indicate that genes contributing to biotic or abiotic stress tolerance are mostly down-expressed in weedy populations, supporting the hypothesis that weedy populations have evolved faster growth rates by reducing investment in stress tolerance (Lai et al. 2008). Both Canada thistle (Cirsium arvense) and common ragweed (Ambrosia artemisiifolia) also show differential expression of stress response genes between native and invasive populations (Guggisberg et al. 2013; Hodgins et al. 2013). Chapter 4 reveals more about the potential genetic mechanisms underlying the differences in performance in benign conditions and drought tolerance observed between native and invasive populations of C. diffusa. Invasive populations have constitutively higher levels of expression of genes relating to energy production and react less strongly to drought than native populations. This pattern supports the expectation of invasive populations consisting of generalist genotypes capable of fitness homeostasis, which may have allowed the successful invasion of many environments in North America.  101  Concluding remarks  This body of work tested hypotheses about physiological trade-offs and  rapid evolution in the invaded range and represents some of the first evolutionary investigations into one of western North America’s most problematic invasive species. This research contributes to invasion genetics and more broadly to our understanding of evolutionary and ecological responses to environmental changes, which may occur due to range expansion or in situ. This work is also an effort to integrate information across multiple levels of biological organization in a non-model organism, from biotic or abiotic stressors, to phenotype, to gene expression, and eventually, to genotype (using recently developed genomic resources; see Appendix D). The information provided here is necessary to test evolutionary theories about how and why species become invasive. Such information may be used to minimize evolutionary potential, for example, when selecting a biocontrol agent, or to predict which species pose the greatest threat for future invasions.  102  References Alexander JM, Edwards PJ, Poll M, Parks CG, Dietz H (2009) Establishment of parallel altitudinal clines in traits of native and introduced forbs. Ecology, 90, 612–622. Alexa A, Rahnenführer J (2009) Gene set enrichment analysis with topGO. Available at http://rgm.ogalab.net/RGM-files/R_BC/download/topGO/inst/doc/topGO.pdf.  Atkin OK, Macherel D (2009) The crucial role of plant mitochondria in orchestrating drought tolerance. Annals of Botany, 103, 581–597. Baker HG, Stebbins GL (1965) The genetics of colonizing species: proceedings. Academic Press. Basu C, Halfhill MD, Mueller TC, Stewart Jr CN (2004) Weed genomics: new tools to understand weed biology. Trends in Plant Science, 9, 391–398. Bates D, Mächler M, Bolker B, Walker S (2014) Fitting linear mixed-effects models using lme4. arXiv:1406.5823 [stat]. Blair AC, Blumenthal D, Hufbauer RA (2012) Hybridization and invasion: An experimental test with diffuse knapweed (Centaurea diffusa Lam.). Evolutionary Applications, 5, 17–28. Blair AC, Hufbauer RA (2009) Geographic patterns of interspecific hybridization between spotted knapweed (Centaurea stoebe) and diffuse knapweed (C. diffusa). Invasive Plant Science and Management, 2, 55–69. Blair AC, Hufbauer RA (2010) Hybridization and invasion: one of North America’s most devastating invasive plants shows evidence for a history of interspecific hybridization. Evolutionary Applications, 3, 40–51. Blair AC, Schaffner U, Häfliger P, Meyer SK, Hufbauer RA (2008) How do biological control and hybridization affect enemy escape? Biological Control, 46, 358–370.  103  Bleeker W, Schmitz U, Ristow M (2007) Interspecific hybridisation between alien and native plant species in Germany and its consequences for native biodiversity. Biological Conservation, 137, 248–253. Blossey B, Notzold R (1995) Evolution of increased competitive ability in invasive nonindigenous plants: a hypothesis. Journal of Ecology, 83, 887–889. Blumenthal DM, Hufbauer RA (2007) Increased plant size in exotic populations: a common-garden test with 14 invasive species. Ecology, 88, 2758–2765. Bock DG, Caseys C, Cousens RD et al. (2015) What we still don’t know about invasion genetics. Molecular Ecology, DOI: 10.1111/mec.13032. Bolstad BM, Irizarry RA, Åstrand M, Speed TP (2003) A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics, 19, 185–193. Bossdorf O, Auge H, Lafuma L et al. (2005) Phenotypic and genetic differentiation between native and introduced plant populations. Oecologia, 144, 1–11. Broennimann O, Mráz P, Petitpierre B, Guisan A, Müller-Schärer H (2014) Contrasting spatio-temporal climatic niche dynamics during the eastern and western invasions of spotted knapweed in North America. Journal of Biogeography, 41, 1126–1136. Broennimann O, Treier UA, Müller-Schärer H et al. (2007) Evidence of climatic niche shift during biological invasion. Ecology Letters, 10, 701–709. Burke JM, Arnold ML (2001) Genetics and the fitness of hybrids. Annual Review of Genetics, 35, 31–52. Burton OJ, Phillips BL, Travis JMJ (2010) Trade-offs and the evolution of life-histories during range expansion. Ecology Letters, 13, 1210–1220.  104  Buswell JM, Moles AT, Hartley S (2011) Is rapid evolution common in introduced plant species? Journal of Ecology, 99, 214–224. Butler DR, Landsberg JJ (1981) Respiration rates of apple trees, estimated by CO2-efflux measurements. Plant, Cell & Environment, 4, 153–159. Carvalho BS, Irizarry RA (2010) A framework for oligonucleotide microarray preprocessing. Bioinformatics, 26, 2363–2367. Catford JA, Jansson R, Nilsson C (2009) Reducing redundancy in invasion ecology by integrating hypotheses into a single theoretical framework. Diversity and Distributions, 15, 22–40. Chamberlain S, Ram K, Barve V, Mcglinn D (2014) rgbif: Interface to the Global Biodiversity Information Facility API. [WWW document] URL https://github.com/ropensci/rgbif. [accessed 1 June 2014]. Colautti RI, Barrett SCH (2013) Rapid adaptation to climate facilitates range expansion of an invasive plant. Science, 342, 364–366. Colautti RI, Eckert CG, Barrett SC. (2010) Evolutionary constraints on adaptive evolution during range expansion in an invasive plant. Proceedings of the Royal Society B: Biological Sciences, 277, 1799. Colautti RI, Maron JL, Barrett SCH (2009) Common garden comparisons of native and introduced plant populations: latitudinal clines can obscure evolutionary inferences. Evolutionary Applications, 2, 187–199. Crawley MJ (1987) What makes a community invasible?. In: Colonization, succession, and stability: the 26th Symposium of the British Ecological Society held jointly with the Linnean Society of London, pp. 429–453. Blackwell Scientific Publications, Oxford.  105  Crawley MJ (2012) The R book. John Wiley & Sons, West Sussex, UK. Dlugosch KM, Parker IM (2008) Founding events in species invasions: Genetic variation, adaptive evolution, and the role of multiple introductions. Molecular Ecology, 17, 431–449. Dray S, Dufour A-B, others (2007) The ade4 package: implementing the duality diagram for ecologists. Journal of statistical software, 22, 1–20. Ellstrand NC, Schierenbeck KA (2000) Hybridization as a stimulus for the evolution of invasiveness in plants? Proceedings of the National Academy of Sciences, 97, 7043–7050. Elton CS (1958) The ecology of invasions by plants and animals. Methuen, London. Facon B, Jarne P, Pointier JP, David P (2005) Hybridization and invasiveness in the freshwater snail Melanoides tuberculata: hybrid vigour is more important than increase in genetic variance. Journal of Evolutionary Biology, 18, 524–535. Felker-Quinn E, Schweitzer JA, Bailey JK (2013) Meta-analysis reveals evolution in invasive plant species but little support for Evolution of Increased Competitive Ability (EICA). Ecology and Evolution, 3, 739–751. Feng YL, Lei YB, Wang RF, Callaway RM, Valiente-Banuet A (2009) Evolutionary tradeoffs for nitrogen allocation to photosynthesis versus cell walls in an invasive plant. Proceedings of the National Academy of Sciences, 106, 1853. Fitzpatrick MC, Weltzin JF, Sanders NJ, Dunn RR (2007) The biogeography of prediction error: Why does the introduced range of the fire ant over-predict its native range? Global Ecology and Biogeography, 16, 24–33. Flexas J, Galmes J, Ribas-Carbo M, Medrano H (2005) The effects of water stress on plant respiration. In: Plant Respiration: Advances in photosynthesis and respiration. (eds Lambers H, Ribas-Carbo M), pp. 85–94. Springer Netherlands.  106  Folguera G, Ceballos S, Spezzi L, Fanara JJ, Hasson E (2008) Clinal variation in developmental time and viability, and the response to thermal treatments in two species of Drosophila. Biological Journal of the Linnean Society, 95, 233–245. Galloway LF (2005) Maternal effects provide phenotypic adaptation to local environmental conditions. New Phytologist, 166, 93–100. Garcia-Jacas N, Uysal T, Romashchenko K et al. (2006) Centaurea revisited: A molecular survey of the Jacea group. Ann Bot, 98, 741–753. Gayer G (1909) Vier neue Centaureen der Flora Botanikai von Ungarn. Magyar Botanikai Lapok, 8, 58–61. GBIF (2014) Global Biodiversity Information Facility Data Portal. [WWW document] URL http://gbif.org [accessed Dec 20, 2014]. Ghashghaie J, Duranceau M, Badeck F-W et al. (2001) δ13C of CO2 respired in the dark in relation to δ13C of leaf metabolites: Comparison between Nicotiana sylvestris and Helianthus annuus under drought. Plant, Cell & Environment, 24, 505–515. Gilchrist GW, Huey RB, Balanya J, Pascual M, Serra L (2004) A time series of evolution in action: a latitudinal cline in wing size in South American Drosophila subobscura. Evolution, 58, 768–780. Gilchrist GW, Jeffers LM, West B et al. (2008) Clinal patterns of desiccation and starvation resistance in ancestral and invading populations of Drosophila subobscura. Evolutionary Applications, 1, 513–523. Gray AJ, Crawley MJ, Edwards PJ (1987) Colonization, succession, and stability: the 26th Symposium of the British Ecological Society held jointly with the Linnean Society of London. Blackwell Scientific Publications, Oxford.  107  Greuter W (2009) Compositae (pro parte majore). In: Greuter, W. & Raab-Straube, E. von (ed.): Compositae. Euro+Med Plantbase - the information resource for Euro-Mediterranean plant diversity. [WWW document] URL http://ww2.bgbm.org/EuroPlusMed/. [accessed 9 June 2013]. Grime JP (1977) Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. The American Naturalist, 111, 1169. Grossmann S, Bauer S, Robinson PN, Vingron M (2007) Improved detection of overrepresentation of Gene-Ontology annotations with parent–child analysis. Bioinformatics, 23, 3024–3031. Guggisberg A, Lai Z, Huang J, Rieseberg LH (2013) Transcriptome divergence between introduced and native populations of Canada thistle, Cirsium arvense. New Phytologist, 199, 595–608. Guisan A, Petitpierre B, Broennimann O, Daehler C, Kueffer C (2014) Unifying niche shift studies: insights from biological invasions. Trends in Ecology & Evolution, 29, 260–269. Guleria P, Goswami D, Mahajan M et al. (2012) MicroRNAs and their role in plants during abiotic stresses. In: Environmental adaptations and stress tolerance of plants in the era of climate change (eds Ahmad P, Prasad MNV), pp. 265–278. Springer New York. Hackenberg M, Gustafson P, Langridge P, Shi B-J (2015) Differential expression of microRNAs and other small RNAs in barley between water and drought conditions. Plant Biotechnology Journal, 13, 2–13. Hahn MA, van Kleunen M, Müller-Schärer H (2012) Increased phenotypic plasticity to climate may have boosted the invasion success of polyploid Centaurea stoebe. PLoS ONE, 7, e50284.  108  Hamrick JL, Godt MJW (1996) Effects of life history traits on genetic diversity in plant species. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences, 351, 1291–1298. Harrod RJ, Taylor RJ (1995) Reproduction and pollination biology of Centaurea and Acroptilon species, with emphasis on C. diffusa. Northwest Science, 69. Hendry A, Farrugia T, Kinnison M (2008) Human influences on rates of phenotypic change in wild animal populations. Molecular Ecology, 17, 20–29. He W-M, Thelen GC, Ridenour WM, Callaway RM (2010) Is there a risk to living large? Large size correlates with reduced growth when stressed for knapweed populations. Biological Invasions, 12, 3591–3598. Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A (2005) Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology, 25, 1965–1978. Hinz HL, Schwarzlaender M (2004) Comparing invasive plants from their native and exotic range: What can we learn for biological control? Weed Technology, 18, 1533–1541. Hodgins KA, Bock DG, Hahn MA et al. (2015) Comparative genomics in the Asteraceae reveals little evidence for parallel evolutionary change in invasive taxa. Molecular Ecology, DOI: 10.1111/mec.13026. Hodgins KA, Lai Z, Nurkowski K, Huang J, Rieseberg LH (2013) The molecular basis of invasiveness: differences in gene expression of native and introduced common ragweed (Ambrosia artemisiifolia) in stressful and benign environments. Molecular Ecology, 22, 2496–2510.   109  Hodgins KA, Rieseberg L (2011) Genetic differentiation in life‐history traits of introduced and native common ragweed (Ambrosia artemisiifolia) populations. Journal of Evolutionary Biology, 24, 2731–2749. Hoffmann AA, Anderson A, Hallas R (2002) Opposing clines for high and low temperature resistance in Drosophila melanogaster. Ecology Letters, 5, 614–618. Hovick SM, Campbell LG, Allison A. Snow, Whitney KD (2012) Hybridization alters early life-history traits and increases plant colonization success in a novel region. The American Naturalist, 179, 192–203. Huey RB, Gilchrist GW, Carlson ML, Berrigan D, Serra L (2000) Rapid evolution of a geographic cline in size in an introduced fly. Science, 287, 308–309. Hufbauer RA, Facon B, Ravigné V et al. (2012) Anthropogenically induced adaptation to invade (AIAI): Contemporary adaptation to human-altered habitats within the native range can promote invasions. Evolutionary Applications, 5, 89–101. Hufbauer RA, Sforza R (2008) Multiple introductions of two invasive Centaurea taxa inferred from cpDNA haplotypes. Diversity and Distributions, 14, 252–261. Irizarry RA, Bolstad BM, Collin F et al. (2003a) Summaries of Affymetrix GeneChip probe level data. Nucleic acids research, 31, e15–e15. Irizarry RA, Hobbs B, Collin F et al. (2003b) Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics, 4, 249–264. ISSG (2013) Centaurea diffusa Lam. Global Invasive Species Database. [WWW document] URL http://www.issg.org/database/species/distribution.asp?si=313&fr=1&sts=&lang=EN. [accessed 9 June 2013].  110  Joshi J, Vrieling K (2005) The enemy release and EICA hypothesis revisited: incorporating the fundamental difference between specialist and generalist herbivores. Ecology Letters, 8, 704–714.  Kao JY, Zubair A, Salomon MP, Nuzhdin SV, Campo D (2015) Population genomic analysis uncovers African and European admixture in Drosophila melanogaster populations from the south-eastern United States and Caribbean Islands. Molecular Ecology, 24, 1499–1509. Kartesz J (2013) Centaurea diffusa Lam. The Biota of North America Program (BONAP). North American Plant Atlas.  [WWW document] URL http://bonap.org/BONAPmaps2010/Centaurea.html. [accessed 9 June 2013]. Van Kleunen M, Weber E, Fischer M (2010) A meta-analysis of trait differences between invasive and non-invasive plant species. Ecology Letters, 13, 235–245. Kuester A, Conner JK, Culley T, Baucom RS (2014) How weeds emerge: a taxonomic and trait-based examination using United States data. New Phytologist, 202, 1055–1068. Kumschick S, Hufbauer RA, Alba C, Blumenthal DM (2013) Evolution of fast-growing and more resistant phenotypes in introduced common mullein (Verbascum thapsus). Journal of Ecology, 101, 378–387. Lachmuth S, Durka W, Schurr FM (2011) Differentiation of reproductive and competitive ability in the invaded range of Senecio inaequidens: the role of genetic Allee effects, adaptive and nonadaptive evolution. New Phytologist, 192, 529–541. Lai Z, Gross BL, Yizou J, Rieseberg L (2006) Microarray analysis reveals differential gene expression in hybrid sunflower species. Molecular ecology, 15, 1213. Lai Z, Kane NC, Kozik A et al. (2012) Genomics of Compositae weeds: EST libraries, microarrays, and evidence of introgression. American Journal of Botany, 99, 209–218.  111  Lai Z, Kane NC, Zou Y, Rieseberg LH (2008) Natural variation in gene expression between wild and weedy populations of Helianthus annuus. Genetics, 179, 1881. Lande R (2009) Adaptation to an extraordinary environment by evolution of phenotypic plasticity and genetic assimilation. Journal of Evolutionary Biology, 22, 1435–1446. Lande R (2015) Evolution of phenotypic plasticity in colonizing species. Molecular Ecology, DOI: 10.1111/mec.13037. Lavergne S, Molofsky J (2007) Increased genetic variation and evolutionary potential drive the success of an invasive grass. Proceedings of the National Academy of Sciences, 104, 3883–3888. Lee CE, Gelembiuk GW (2008) Evolutionary origins of invasive populations. Evolutionary Applications, 1, 427–448. Lejeune KD, Seastedt TR (2001) Centaurea species: the forb that won the west. Conservation Biology, 15, 1568–1574. Lohse M, Drechsel O, Kahlau S, Bock R (2013) OrganellarGenomeDRAW--a suite of tools for generating physical maps of plastid and mitochondrial genomes and visualizing expression data sets. Nucleic Acids Research, 41, W575–W581. Marrs RA, Sforza R, Hufbauer RA (2008) When invasion increases population genetic structure: a study with Centaurea diffusa. Biological Invasions, 10, 561–572. Metcalf JC, Rose KE, Rees M (2003) Evolutionary demography of monocarpic perennials. Trends in Ecology & Evolution, 18, 471–480. Meyers LA, Ancel FD, Lachmann M (2005) Evolution of genetic potential. PLoS Comput Biol, 1, e32.  112  Moles AT, Flores-Moreno H, Bonser SP et al. (2012) Invasions: The trail behind, the path ahead, and a test of a disturbing idea. Journal of Ecology, 100, 116–127. Moloney KA, Holzapfel C, Tielbörger K, Jeltsch F, Schurr FM (2009) Rethinking the common garden in invasion research. Perspectives in Plant Ecology, Evolution and Systematics, 11, 311–320. Monty A, Lebeau J, Meerts P, Mahy G (2009) An explicit test for the contribution of environmental maternal effects to rapid clinal differentiation in an invasive plant. Journal of Evolutionary Biology, 22, 917–926. Mráz P, Garcia-Jacas N, Gex-Fabry E et al. (2012a) Allopolyploid origin of highly invasive Centaurea stoebe sl (Asteraceae). Molecular Phylogenetics and Evolution, 62, 612–623. Mráz P, Španiel S, Keller A et al. (2012b) Anthropogenic disturbance as a driver of microspatial and microhabitat segregation of cytotypes of Centaurea stoebe and cytotype interactions in secondary contact zones. Annals of Botany, 110, 615–627. Müller-Schärer H, Schaffner U, Steinger T (2004) Evolution in invasive plants: Implications for biological control. Trends in Ecology & Evolution, 19, 417–422. Myers JH, Jackson C, Quinn H, White SR, Cory JS (2009) Successful biological control of diffuse knapweed, Centaurea diffusa, in British Columbia, Canada. Biological Control, 50, 66–72. Norton AP, Blair AC, Hardin JG, Nissen SJ, Brunk GR (2008) Herbivory and novel weapons: No evidence for enhanced competitive ability or allelopathy induction of Centaurea diffusa by biological controls. Biological Invasions, 10, 79–88.  113  Ochsmann J (2001) An overlooked hybrid in North America: Centaurea x psammogena Gayer (diffuse knapweed x spotted knapweed). In: The First International Knapweed Symposium of the Twenty-First Century (ed Smith L), pp. 33–41. Couer d’Alene, ID. Parepa M, Fischer M, Krebs C, Bossdorf O (2014) Hybridization increases invasive knotweed success. Evolutionary Applications, 7, 413–420. Parker JD, Torchin ME, Hufbauer RA et al. (2013) Do invasive species perform better in their new ranges? Ecology, 94, 985–994. Peterson AT (2003) Predicting the geography of species’ invasions via ecological niche modeling. The Quarterly review of biology, 78, 419–433. Petitpierre B, Kueffer C, Broennimann O et al. (2012) Climatic niche shifts are rare among terrestrial plant invaders. Science, 335, 1344–1348. Pimentel D, Zuniga R, Morrison D (2005) Update on the environmental and economic costs associated with alien-invasive species in the United States. Ecological Economics, 52, 273–288. Powell RD, Myers JH (1988) The effect of Sphenoptera jugoslavica Obenb.(Col., Buprestidae) on its host plant Centaurea diffusa Lam.(Compositae). Journal of Applied Entomology, 106, 25–45. Prentis PJ, Wilson JRU, Dormontt EE, Richardson DM, Lowe AJ (2008) Adaptive evolution in invasive species. Trends in Plant Science, 13, 288–294. Quartacci MF, Glišić O, Stevanović B, Navari‐Izzo F (2002) Plasma membrane lipids in the resurrection plant Ramonda serbica following dehydration and rehydration. Journal of Experimental Botany, 53, 2159–2166.  114  Quartacci MF, Pinzino C, Sgherri C, Navari-Izzo F (1995) Lipid composition and protein dynamics in thylakoids of two wheat cultivars differently sensitive to drought. Plant Physiology, 108, 191–197. Rasband W (2011) ImageJ. U. S. National Institutes of Health. [WWW document] URL http://imagej.nih.gov/ij/ [accessed 29 October 2011], Bethesda, Maryland, USA. R Core Team (2014) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, 2012. ISBN 3-900051-07-0. Richards CL, Bossdorf O, Muth NZ, Gurevitch J, Pigliucci M (2006) Jack of all trades, master of some? On the role of phenotypic plasticity in plant invasions. Ecology Letters, 9, 981–993. Ridley CE, Ellstrand NC (2009) Evolution of enhanced reproduction in the hybrid-derived invasive, California wild radish (Raphanus sativus). Biological Invasions, 11, 2251–2264. Rieseberg LH, Archer MA, Wayne RK (1999) Transgressive segregation, adaptation and speciation. Heredity, 83, 363–372. Rieseberg LH, Kim SC, Randell RA et al. (2007) Hybridization and the colonization of novel habitats by annual sunflowers. Genetica, 129, 149–165. Rizhsky L, Liang H, Mittler R (2002) The combined effect of drought stress and heat shock on gene expression in tobacco. Plant Physiology, 130, 1143–1151. Rypel AL (2014) Do invasive freshwater fish species grow better when they are invasive? Oikos, 123, 279–289. Sakai AK, Allendorf FW, Holt JS et al. (2001) The population biology of invasive species. Annual Review of Ecology and Systematics, 32, 305–332. Sambatti J, Rice KJ (2007) Functional ecology of ecotypic differentiation in the Californian serpentine sunflower (Helianthus exilis). New Phytologist, 175, 107–119.  115  Schierenbeck K, Ellstrand N (2009) Hybridization and the evolution of invasiveness in plants and other organisms. Biological Invasions, 11, 1093–1105. Seastedt TR, Suding KN, LeJeune KD (2005) Understanding invasions: The rise and fall of diffuse knapweed (Centaurea diffusa) in North America. In: Invasive plants: ecological and agricultural aspects, pp. 129–139. Birkhauser Verlag, Basel, CH. Sheley RL, Jacobs JS, Carpinelli MF (1998) Distribution, Biology, and Management of Diffuse Knapweed (Centaurea diffusa) and Spotted Knapweed (Centaurea maculosa). Weed Technology, 12, 353–362. Sheley RL, Larson LL (1996) Emergence date effects on resource partitioning between diffuse knapweed seedlings. Journal of Range Management, 49, 241–244. Stephens AE, Myers JH (2013) Influences of two life-history stages of the weevil, Larinus minutus, on its host plant Centaurea diffusa. Ecological Entomology, 38, 40–48. Stewart CN, Tranel PJ, Horvath DP et al. (2009) Evolution of weediness and invasiveness: Charting the course for weed genomics. Weed Science, 57, 451–462. Storey JD (2002) A direct approach to false discovery rates. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 64, 479–498. Storey JD, Taylor JE, Siegmund D (2004) Strong control, conservative point estimation and simultaneous conservative consistency of false discovery rates: A unified approach. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 66, 187–205. Suárez-Santiago VN, Salinas MJ, Garcia-Jacas N et al. (2007) Reticulate evolution in the Acrolophus subgroup (Centaurea L., Compositae) from the western Mediterranean: Origin and diversification of section Willkommia Blanca. Molecular Phylogenetics and Evolution, 43, 156–172.  116  Thébaud C, Simberloff D (2001) Are plants really larger in their introduced ranges? The American Naturalist, 157, 231–236. Thompson DJ, Stout DG (1991) Duration of the juvenile period in diffuse knapweed (Centaurea diffusa). Canadian Journal of Botany, 69, 368–371. Todesco M, Balasubramanian S, Hu TT et al. (2010) Natural allelic variation underlying a major fitness trade-off in Arabidopsis thaliana. Nature, 465, 632–636. Toumi I, Gargouri M, Nouairi I et al. (2008) Water stress induced changes in the leaf lipid composition of four grapevine genotypes with different drought tolerance. Biologia Plantarum, 52, 161–164. Treier UA, Broennimann O, Normand S et al. (2009) Shift in cytotype frequency and niche space in the invasive plant Centaurea maculosa. Ecology, 90, 1366–1377. Tropicos (2013) Centaurea diffusa Lam.: Literature based distribution. Missouri Botanical Garden. [WWW document] URL http://www.tropicos.org/Name/2700419 [accessed 10 June 2013]. Troyanskaya O, Cantor M, Sherlock G et al. (2001) Missing value estimation methods for DNA microarrays. Bioinformatics, 17, 520–525. Turner KG, Grassa CJ (2014) Complete plastid genome assembly of invasive plant, Centaurea diffusa. bioRxiv, 005900, DOI: 10.1101/005900. Turner KG, Hufbauer RA, Rieseberg LH (2014) Rapid evolution of an invasive weed. New Phytologist, 202, 309–321. USDA NCRS (2013) Centaurea diffusa Lam. The PLANTS Database, National Plant Data Team. [WWW document] URL http://plants.usda.gov/. [accessed 9 June 2013].  117  Varone L, Gratani L (2015) Leaf respiration responsiveness to induced water stress in Mediterranean species. Environmental and Experimental Botany, 109, 141–150. Vilà M, D’Antonio CM (1998) Hybrid vigor for clonal growth in Carpobrotus (Aizoaceae) in coastal California. Ecological Applications, 8, 1196–1205. Warnes GR, Bolker B, Bonebakker L et al. (2015) gplots: Various R programming tools for plotting data. [WWW Document] URL http://CRAN.R-project.org/package=gplots. [accessed 12 December 2014]. Webber BL, Maitre DCL, Kriticos DJ (2012) Comment on “Climatic niche shifts are rare among terrestrial plant invaders.” Science, 338, 193–193. Weiner J, Martinez S, Muller-Scharer H, Stoll P, Schmid B (1997) How important are environmental maternal effects in plants? A study with Centaurea maculosa. Journal of Ecology, 85, 133–142. Whitehead A, Crawford DL (2006) Variation within and among species in gene expression: Raw material for evolution. Molecular Ecology, 15, 1197–1211. Whitney KD, Ahern JR, Campbell LG, Albert LP, King MS (2010) Patterns of hybridization in plants. Perspectives in Plant Ecology, Evolution and Systematics, 12, 175–182. Whitney KD, Gabler CA (2008) Rapid evolution in introduced species,’invasive traits’ and recipient communities: Challenges for predicting invasive potential. Diversity and Distributions, 14, 569–580. Williams JL (2009) Flowering life‐history strategies differ between the native and introduced ranges of a monocarpic perennial. The American Naturalist, 174, 660–672.  118  Williams JL, Auge H, Maron JL (2008) Different gardens, different results: Native and introduced populations exhibit contrasting phenotypes across common gardens. Oecologia, 157, 239–248. Williamson M (1999) Invasions. Ecography, 22, 5–12. Yang SF, Hoffman NE (1984) Ethylene biosynthesis and its regulation in higher plants. Annual Review of Plant Physiology, 35, 155–189. Zenni RD, Bailey JK, Simberloff D (2014a) Rapid evolution and range expansion of an invasive plant are driven by provenance–environment interactions. Ecology letters, 17, 727–735. Zenni RD, Lamy J-B, Lamarque LJ, Porté AJ (2014b) Adaptive evolution and phenotypic plasticity during naturalization and spread of invasive species: Implications for tree invasion biology. Biological Invasions, 16, 635–644. Zhang W, Han Z, Guo Q et al. (2014) Identification of maize long non-coding RNAs responsive to drought stress. PLoS ONE, 9, e98958. Zhu J-K (2002) Salt and Drought Stress Signal Transduction in Plants. Annual review of plant biology, 53, 247–273.   119  Appendices  Appendix A   Chapter 2 Supplementary Material A.1 Germination results of field collected Centaurea diffusa seed.  Seed was collected between 2001 and 2008, but no significant effect of age was detected on germination rate, and no sample bias between ranges is apparent. Germination rate is percent of seeds that germinated per maternal plant for wild collected seed. Sample size as follows: Invasive = 135, Native = 152, field collected mothers.     120  A.2 Native Centaurea diffusa population information. Population ID Latitude Longitude Voucher Accession Included in the following experiments (treatments) BG001 43.38194 28.4575 V236763 B, M GR001 40.15667 22.54806 V236766 B GR002 40.62139 23.07861 V232702, V236769 B, M GR003 40.85083 25.79306 V232674, V236768 B HU001 47.64194 18.7825 V236764 B RO001 44.11028 28.63694 V236767 B RO002 43.90243 28.57392 V232701 B RO003 44.38967 28.52664 V232721, V232723 B RO004 45.57618 29.51805 V232699, V232732, V232724 B RO005 45.49786 27.91181 V232686, V232683 B RU001 51.38333 56.8  B RU002 49.17917 53.24226  B (hndf) RU003 38.1625 54.8  B RU004 60.16667 55.01667  B (chndf) RU005 49.18767 53.25347  B (ecndf) RU008 44.05 43.06 V232687 B, M TR001 41.75111 27.24778 V236765 B, M TR003 38.76722 37.00389  B TR004 38.36722 42.77361  B TR005 39.7825 41.07361  B (chndf) UA001 48.10067 37.81611 V232729 B UA002 48.15233 37.84058 V232684 B (chndf) UA003 50.48075 30.48631 V232717 - V232720 B  UA004 48.64603 30.77508 V232728 B UA005 47.862 38.46069 V232700, V232725 B UA006 45.70833 33.34972  B (chndf) UA007 48.09222 30.74972 V232703, V232726 B (non-destructive early control only) UA008 46.75722 32.76611  B (chndf) Populations are included in all treatments and analyses in a given experiment unless otherwise stated. All populations in a given experiment were included in non-destructive early control measurements (number of basal leaves, area of longest leaf). Experiments are indicated by an upper-case letter and treatments and analyses are indicated by a lower-case letter as in the following code: broad common garden (B), maternal effects common garden (M), early control measurements that required destructive sampling such as biomass (e), control (c), herbivory (h), nutrient deficiency (n), drought (d), flood (f). Vouchers are located in the UBC Herbarium.     121  A.3 Invasive Centaurea diffusa population information. Population ID Latitude Longitude Voucher Accession Included in the following experiments (analyses) CA001 49.01494 -122.882 V232677- V232679, V232704 - V232706 B, M CA006 49.320181 -119.630  Weevils only CA007 49.2961 -118.474  B (chn), Weevils CA008 49.01208 -118.646 V232722 B, Weevils US001 45.61523 -120.788 V232694 - V232697 B, M US002 46.18227 -118.826 V232669 - V232671 B, M US003 46.60415 -116.642  B, M US011 43.38553 -106.937  B (hndf) US012 40.52815 -104.849  B (chndf) US013 40.5612 -104.865 V232713 - V232716, V232680 - V232682 B (non-destructive early control only) US014 40.12227 -101.28 V232730, V232698 B US015 42.72536 -118.001 V232672 B US017 40.42111 -122.555  B (non-destructive early control only) US018 43.54677 -118.914  B (nd) US020 46.98399 -119.58  B (chndf) US021 46.617 -110.092  B (cnd) US022 45.74515 -119.785  B US023 39.22428 -103.122  B (chndf) US026 40.37191 -104.473 V232733 B (chdf) Populations are included in all treatments and analyses in a given experiment unless otherwise stated. All populations in a given experiment were included in non-destructive early control measurements (number of basal leaves, area of longest leaf). Experiments are indicated by an upper-case letter and treatments and analyses are indicated by a lower-case letter as in the following code: broad common garden (B), maternal effects common garden (M), early control measurements that required destructive sampling such as biomass (e), control (c), herbivory (h), nutrient deficiency (n), drought (d), flood (f). Weevils were collected for the leaf choice trial from locations indicated by “Weevils.” Vouchers are located in the UBC Herbarium.     122  A.4 Test statistics for all traits measured in broad common garden, from range differentiation models of phenotype of Centaurea diffusa. Broad Common Garden Fixed effects Random effects Origin Latitude Origin-by- Latitude Populations Maternal lines Populations within each Origin Trait χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P Germination       Seed weight by family 1.38 (1) 0.11 (1) 1.02 (1) 76.54 (1) *** -- 0.54 (2) Early Control       Shoot mass 2.57(1) 10.80 (1) *** 0.05 (1) 13.10 (1) *** 14.31 (1) *** 0.57(4) Root mass 0.58 (1) 12.31 (1) *** 1.12 (1) 13.43 (1) *** 7.79 (1) ** 3.95 (4) Root crown diameter 3.05 (1) . 3.86 (1) * 0.61 (1) 18.97 (1) *** 3.69 (1) . 0.30 (4) Area of longest leaf 0.34 (1) 3.14 (1) . 0.25 (1) 96.67 (1) *** 59.04 (1) *** 2.82 (4) Number of basal leaves nt nt 6.23 (1) * 160.9 (1) *** 8.78 (1) ** 1.19 (4) Control       Shoot mass 0.18 (1) 5.99 (1) * 0.06 (1) 49.74 (1) *** 2.90 (1) . 4.81 (4) Root mass 3.31 (1) . 4.90 (1) * 0.11 (1) 41.15 (1) *** 0.23 (1) 3.86 (4) Root crown diameter 0.05 (1) 9.64 (1) ** 0.08 (1) 48.89 (1) *** 5.49 (3) 18.22 (4) ** Area of longest leaf 1.25 (1) 8.90 (1) ** 0.29 (1) 3.82 (1) . 7.78 (1) ** 9.48 (4) . Number of basal leaves 13.08 (1) *** 5.91 (1) * 1.50 (1) 66.22 (1) *** 257.29 (3) *** 37.14 (4) *** Bolting status at harvest nt nt 7.14 (1) ** 0.57 (1) 10.71 (1) *** 4.07 (4) Bolting date 1.59 (1) 3.54 (1) . 0.005 (1) 154.7 (1) *** 184.2 (1) *** 9.25 (4) . Nutrient       Shoot mass 3.22 (1) . 3.83 (1) * <0.01 (1) 15.96 (1) *** 0.05 (1) 2.30 (4) Root mass 3.30 (1) . 7.84 (1) ** 0.68 (1) 42.65 (1) *** 0.14 (1) 5.43 (4) Root crown diameter 0.89 (1) 2.83 (1) . <0.01 (1) 26.05 (1) *** 0.68 (1) 9.10 (4). Area of longest leaf 0.71 (1) 2.05 (1) <0.01 (1) 7.97 (1) ** <0.01 (1) 0.23 (4) Number of basal leaves 1.97 (1) 3.71 (1) . 2.44 (1) 62.57 (1) *** 48.11 (3) *** 10.33 (4) * Herbivory       Root mass 4.33 (1) * 15.82 (1) *** 0.06 (1) 13.04 (1) *** <0.01 (1) 0.84 (4) Root crown diameter 0.81 (1) 14.66 (1) *** 0.21 (1) 10.85 (1) *** <0.01 (1) 2.22 (4) Number of basal leaves 5.91 (1) * 13.46 (1) *** 0.02 (1) 166.6 (1) *** 237.03 (3) *** 20.54 (4) *** Bolting status at harvest nt nt 14.46 (1) *** <0.01 (1) 0.32 (1) 6.24 (4) Bolting date 1.91 (1) 5.34 (1) * <0.01 (1) 53.36 (1) *** 80.88 (1) *** 8.03 (4) . Drought       Date of 1st wilt 4.76 (1) *  6.34 (1) * 0.50 (1) 2.12 (1) 0.27 (1) 3.71 (4) Date of total wilt 1.31 (1) 7.68 (1) ** 1.19 (1) 1.11 (1) <0.01 (1) 3.41 (4) Death date 2.68 (1) 12.46 (1) *** 0.16 (1) <0.01 (1) <0.01 (1) <0.01 (4) Flood       Date of 1st yellow leaf 0.38 (1) 0.63 (1) 0.06 (1) <0.01 (1) <0.01 (1) <0.01 (4) Root death date 2.58 (1) 17.87 (1) *** 0.36 (1) <0.01 (1) 0.04 (1) 0.64 (4) Death date 0.80 (1) 7.06 (1) ** 0.01 (1) 7.01 (1) ** 3.68 (1) . 2.06 (4) Results are presented from restricted maximum likelihood (REML) models. Where no random effects were significant, generalized linear models (GLM) were used to test fixed effects. Significance of term indicated by symbol: ., P < 0.1; *, P < 0.05; **, P < 0.01;  ***, P < 0.001. (df) = degrees of freedom. χ2 = chi-squared test statistic. Nt = Not tested, due to significant interaction term.   123  A.5 Means and confidence intervals for all traits estimated from range differentiation models of phenotypes of Centaurea diffusa grown in a common garden. Broad Common Garden  Origin    Native Invasive  Trait Distribution and model Estimate CI Estimate CI P Early Control       Shoot mass (g) Gaussian, REML 0.41 0.36 – 0.45 0.48 0.40-0.55  Root mass (g) Gaussian (loge), REML 0.25 0.20 – 0.31 0.29 0.20 – 0.43  Root crown diameter (mm) Gaussian, REML 3.40 3.14 – 3.67 3.88 3.44 – 4.31 . Area of longest leaf (cm2) Gaussian, REML 45.18 42.13 – 48.23 46.67 42.60 – 50.74  Number of basal leaves Poisson, REML 9.83 9.28 – 10.41 9.60 8.90 – 10.36 [*] Control       Shoot mass (g) Gaussian, REML 3.13 2.87 – 3.39 3.23 2.88 – 3.60  Root mass (g) Gaussian (loge), REML 1.40 1.17 – 1.68 1.84 1.44 – 2.37 . Root crown diameter (mm) Gaussian, REML 7.80 7.38 – 8.22 7.87 7.59 – 8.14  Area of longest leaf (cm2) Gaussian, REML 148.43 140.44 – 156.43 155.85 144.71 – 166.99  Number of basal leaves Poisson, REML 12.43 11.04 – 13.99 16.28 15.31 – 17.32 *** Bolting status at harvest (%) Binomial, REML 24.30 17.49 – 32.69 2.07 0.62 – 6.76 [**] Bolting date Poisson, REML 58.13 51.59 – 65.49 74.42 51.88 – 106.76  Nutrient       Shoot mass (g) Gaussian, REML 1.29 1.19 – 1.39 1.44 1.30 – 1.57 . Root mass (g) Gaussian (loge), REML 2.55 2.23 – 2.92 3.14 2.60 – 3.79 . Root crown diameter (mm) Gaussian, REML 5.58 5.33 – 5.82 5.78 5.43 – 6.12  Area of longest leaf (cm2) Gaussian, REML 69.66 63.98 – 75.33 73.75 65.67 – 81.86  Number of basal leaves Poisson, REML 14.45 13.04 – 16.01 15.79 14.79 – 16.87  Herbivory       Root mass (g) Gaussian (loge), REML 1.03 0.88 – 1.21 1.37 1.09 – 1.72 * Root crown diameter (mm) Gaussian, REML 6.81 6.52 – 7.09 7.03 6.62 – 7.45  Number of basal leaves Poisson, REML 21.58 19.26 – 24.18 25.71 23.82 – 27.74 * Bolting status at harvest (%) Binomial, GLM 26.71 20.17 – 34.47 6.45 2.44 – 15.97 [***] Bolting date Poisson, REML 44.55 38.81 – 51.14 63.08 39.64 – 100.40  Drought       Date of 1st wilt Poisson, GLM 3.73 3.43 – 4.05 3.14 2.76 – 3.58 * Date of total wilt Poisson, GLM 6.13 5.74 – 6.54 5.73 5.20 – 6.31  Death date Poisson, GLM 12.82 12.25 – 13.41 11.98 11.21 – 12.82  Flood       Date of 1st yellow leaf Poisson, GLM 6.47 6.07 – 6.90 6.23 5.63 – 6.90  Root death date Poisson, GLM 16.27 15.33 – 17.26 17.67 16.29 – 19.17  Death date Poisson, REML 18.06 17.08 – 19.08 18.90 17.41 – 20.52  Least squares (LS) means from restricted maximum likelihood (REML) models which include origin and all significant terms (interactions were excluded). Where no random effects were significant, LS means are estimated from generalized linear models (GLM). Significance of origin term (or origin-by-latitude in square brackets []) indicated in right-most column: ., P < 0.1; *, P < 0.05; **, P < 0.01; ***, P < 0.001.  CI = 95% confidence interval. Loge indicates natural log transformation of data. Nt = Not tested, due to significant interaction term.   124  A.6 Means and confidence intervals for all traits estimated from range differentiation models of phenotypes of Centaurea diffusa grown in a common garden, after one generation. Maternal Common Garden Origin    Native Invasive  Trait Distribution and model Estimate CI Estimate CI P Germination       Seed count by family Poisson, REML 3.35 2.56 – 4.38 7.21 5.48 – 9.48 ** Rate by family (% of global average seed count) Poisson, GLM 59.19  51.05 – 68.62 79.85  71.16 – 89.59 ** Average germination date by family Gaussian, GLM 2.36  2.21 – 2.51 2.07  1.94 – 2.20 * Average seed weight by family (mg) Gaussian, GLM 1.00 1.00-1.00 1.00 1.00 – 1.00  Early Control       Shoot mass (g) Gaussian, GLM 0.88 0.73 – 1.03 1.24 1.10 – 1.38 *** Root crown diameter (mm) Gaussian, GLM 5.05 4.64 – 5.47 4.85 4.46 – 5.24  Area of longest leaf (cm2) Gaussian, REML 56.04 40.96 – 71.11 75.80 60.88 – 90.73 * Number of basal leaves Poisson, REML 15.06 13.09 – 17.33 14.76 12.83 – 17.00 [*] Control       Shoot mass (g) Gaussian, GLM 3.37 3.11 – 3.63 3.79 3.52 – 4.05 * Root crown diameter (mm) Gaussian, REML 0.88 0.70 – 1.06 0.87 0.78 – 0.96  Area of longest leaf (cm2) Gaussian, GLM 102.48 90.06 – 114.89 120.97 109.19 – 132.74 [*] Number of basal leaves Poisson, REML 11.74 8.51 – 16.21 20.41 14.81 – 28.14 * Specific leaf area (mm2/mg) Gaussian (loge), GLM 175.53 162.0 – 190.2 169.73 157.6 – 182.8  Bolting status at harvest (%) Binomial, GLM 75.09 54.53 – 88.34 7.57 2.51 – 20.67 [*] Bolting date Poisson, REML 59.61 53.63 – 66.25 71.27 61.81 – 82.16 . Nutrient       Shoot mass (g) Gaussian, GLM 1.66 1.30 – 2.02 1.97 1.61 – 2.33  Root crown diameter (mm) Gaussian, GLM 0.70 0.59 – 0.81 0.75 0.64 – 0.86  Area of longest leaf (cm2) Gaussian, GLM 56.45 41.76 – 71.13 82.74 68.05 – 97.42 ** Number of basal leaves Poisson, REML 20.74 15.82 – 27.20 17.51 13.29 – 23.07  Herbivory       Root crown diameter (mm) Gaussian, GLM 0.88 0.62 – 1.13 1.04 0.79 – 1.30  Number of basal leaves Poisson, REML 31.84 24.80 – 40.88 32.32 25.15 – 41.54  Bolting status at harvest (%) Binomial, GLM 84.70 36.30 – 98.18 0.44 0.01 – 24.82 [**] Drought       Date of 1st wilt Poisson, GLM 2.00 1.34 – 2.98 1.25 0.75 – 2.07  Date of total wilt Poisson, GLM 4.17 3.16 – 5.50 3.83 2.87 – 5.12  Death date Poisson, GLM 8.17 6.70 – 9.95 7.00 5.65 – 8.67  Flood       Date of 1st yellow leaf Poisson, GLM 6.50 5.21 – 8.12 6.00 4.76 – 7.56  Root death date Poisson, GLM 11.67 9.23 – 14.75 12.14 9.82 – 15.02  Death date Poisson, GLM 13.5 11.57 – 15.75 13.67 11.73 – 15.93  Least squares (LS) means from restricted maximum likelihood (REML) models which include origin and all significant terms (interactions were excluded). Where no random effects were significant, LS means are estimated from generalized linear models (GLM). Significance of origin term (or origin-by-latitude in square brackets []) indicated in right-most column: ., P < 0.1; *, P < 0.05; **, P < 0.01; ***, P < 0.001. CI = 95% confidence interval. Loge indicates natural log transformation of data. Nt = Not tested, due to significant interaction term.  125  A.7 Test statistics for all traits measured in maternal common garden from range differentiation models of phenotypes of Centaurea diffusa grown in a common garden, after one generation in the glasshouse. Maternal Common Garden Fixed effects Random effects Origin Latitude Origin-by- Latitude Populations Maternal lines Populations within each Origin Trait χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P Germination       Seed count by family 8.56 (1) ** 2.91 (1) . 2.09 (1) 2.38 (1) 79.57 (1) *** 1.04 (4) Rate by family 10.73 (1) ** 0.78 (1) 0.85 (1) <0.01 (1) 2.99 (1) . 0.86 (4) Average germination date by family 5.51 (1) * 5.56 (1) * 0.85 (1) <0.01 (1) <0.01 (1) 9.04 (4) . Average seed weight by family 0.01 (1) 0.24 (1) 0.26 (1) 1.30 (1) <0.01 (1) 0.06 (4) Early Control       Shoot mass 13.66 (1) *** 1.40 (1) 0.93 (1) 0.14 (1) 2.26 (1) 3.39 (4) Root crown diameter 0.56 (1) 0.34 (1) <0.01 (1) <0.01 (1) <0.01 (1) 3.31 (4) Area of longest leaf 4.76 (1) * 0.35 (1) <0.01 (1) 16.09 (1) *** 38.54 (1) *** 1.78 (4) Number of basal leaves nt nt 6.24 (1) * 1.45 (1) 16.87 (1) *** 0.05 (4) Control       Shoot mass 4.85 (1) * 3.12 (1) . 0.32 (1) 1.37 (1) 0.29 (1) 3.14 (4) Root crown diameter 0.08 (1) 0.76 (1) 0.44 (1) 0.82 (1) 12.88 (1) *** 0.68 (4) Area of longest leaf 4.60 (1) * 1.79 (1)  4.28 (1) * 1.30 (1) 0.52 (1) 3.12 (4) Number of basal leaves 4.20 (1) * 0.64 (1) 3.75 (1) . 4.48 (1) * 215.7 (1) *** 4.22 (4) Specific leaf area 0.38 (1) <0.01 (1) 2.76 (1) . <0.01 (1) 2.32 (1) 1.43 (4) Bolting status at harvest nt nt 5.78 (1) * <0.01 (1) 0.11 (1) 0.68 (4) Bolting date 3.50 (1) . <0.01 (1) 0.26 (1) 1.82 (1) 55.18 (1) *** 0.37 (4) Nutrient       Shoot mass 1.56 (1) <0.01 (1) 0.63 (1) 1.30 (1) <0.01 (1) 0.08 (4) Root crown diameter 0.43 (1) 1.42 (1) 2.18 (1) 0.24 (1) 3.82 (1) . 4.51 (4) Area of longest leaf 6.89 (1) ** 1.12 (1)  1.32 (1) 0.13 (1) <0.01 (1) 1.47 (4) Number of basal leaves 0.70 (1) 0.90 (1) 2.15 (1) 7.72 (1) ** 1.28 (1) 0.33 (4) Herbivory       Root crown diameter 1.01 (1) 0.04 (1) 0.03 (1) 0.36 (1) <0.01 (1) 1.34 (4) Number of basal leaves <0.01 (1) 1.73 (1) 0.63 (1) 16.18 (1) *** 63.44 (1) *** 0.02 (4) Bolting status at harvest nt nt 7.34 (1) ** <0.01 (1) <0.01 (1) <0.01 (4) Drought       Date of 1st wilt 2.10 (1) 0.11 (1) <0.01 (1) <0.01 (1) <0.01 (1) 0.09 (4) Date of total wilt 0.17 (1) 0.39 (1) 0.73 (1) <0.01 (1) <0.01 (1) <0.01 (4) Death date 1.08 (1) 2.24 (1) 0.57 (1) <0.01 (1) <0.01 (1) <0.01 (4) Flood       Date of 1st yellow leaf 0.24 (1) 0.46 (1) <0.01 (1) <0.01 (1) <0.01 (1) <0.01 (4) Root death date 0.06 (1) 1.19 (1) 0.18 (1) <0.01 (1) <0.01 (1) <0.01 (4) Death date 0.01 (1) 0.97 (1) 0.14 (1) <0.01 (1) <0.01 (1) <0.01 (4) Results are presented from restricted maximum likelihood (REML) models. Where no random effects were significant, generalized linear models (GLM) were used to test fixed effects. Significance of term indicated by symbol: ., P < 0.1; *, P < 0.05; **, P < 0.01;  ***, P < 0.001. (df) = degrees of freedom. χ2 = chi-squared test statistic. Nt = Not tested, due to significant interaction term.  126  A.8 Tests of slope differences in origin-by-control treatment performance (control shoot mass) terms from explicit trade-off models of Centaurea diffusa grown in a common garden. Explicit trade-off models Sample size     Trait Coefficient Invasive Native Estimate SE Z ratio P Broad Common Garden       Drought        Days to first wilt Invasive mean (intercept) 65 145 0.32 0.53 0.61   Difference in means   1.44 0.56 2.58 **  Invasive slope over control shoot mass   0.23 0.16 1.46   Difference in slopes over control shoot mass   -0.38 0.17 -2.24 *         Days to total wilt Invasive mean (intercept) 65 145 1.86 0.49 3.76 ***  Difference in means   0.93 0.41 2.24 *  Invasive slope over control shoot mass   0.22 0.12 1.84 .  Latitude   -0.02 0.01 -2.43 *  Difference in slopes over control shoot mass   -0.26 0.13 -2.07 *         Days to death Invasive mean (intercept) 65 145 2.33 0.27 8.63 ***  Difference in means   0.65 0.29 2.29 *  Invasive slope over control shoot mass   0.04 0.08 0.54   Difference in slopes over control shoot mass   -0.18 0.09 -2.11 * Flood        Days to death Invasive mean (intercept) 122 49 1.28 0.39 3.32 ***  Difference in means   1.01 0.32 3.17 **  Invasive slope over control shoot mass   0.32 0.09 3.43 ***  Latitude   0.01 0.006 2.22 *  Difference in slopes over control shoot mass   -0.34 0.10 -3.41 *** Maternal Common Garden       Drought        Days to total wilt Invasive mean (intercept) 12 12 -3.06 2.20 -1.39   Difference in means   5.35 2.31 2.32 *  Invasive slope over control shoot mass   1.15 0.57 2.03 *  Difference in slopes over control shoot mass   -1.40 0.60 -2.33 * Z ratios are presented from explicit trade-off models that include a significant interaction term only, described in Table 2.3. Results are presented from restricted maximum likelihood (REML) models. Where no random effects were significant, generalized linear models (GLM) were used to test fixed effects. Models include all significant terms. All random effects were non-significant, except for days to death in the flood treatment of the broad common garden (maternal line: χ2 = 4.35 (1)*; population: χ2 =4.51 (1)*). Estimates and standard errors (SE) are untransformed. Significance of term indicated by symbol: ., P < 0.1; *, P < 0.05; **, P < 0.01;  ***, P < 0.001. (df) = degrees of freedom. χ2 = chi-squared test statistic.     127  A.9 Effect of origin, latitude, and constitutive defense on herbivore preference of Centaurea diffusa grown in a common garden across two generations.  Leaf choice trials Fixed effects Random effects  Origin Defense Latitude Populations Maternal lines Populations within Origin Trait χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P Most preferred 0.75 (1) <0.01 (3) 1.00 (1) <0.01 (1) <0.01 (1) 0.32 (4) Area consumed 0.09 (1) >250 (3) *** 3.06 (1) . 1.04 (1) <0.01 (1) 1.92 (4) Results are presented from restricted maximum likelihood (REML) models. Generation was included as a random effect, but was never significant. Where no random effects were significant, generalized linear models (GLM) were used to test fixed effects. Significance of term indicated by symbol: ., P < 0.1; *, P < 0.05; **, P < 0.01;  ***, P < 0.001. (df) = degrees of freedom. χ2 = chi-squared test statistic.     128  A.10 Tests of slope difference in origin-by-latitude terms from range differentiation models of Centaurea diffusa grown in a common garden. Range differentiation models Sample size     Trait Coefficients Invasive Native Estimate SE Z ratio P Broad Common Garden        Early Control        No. of basal leaves Invasive mean (intercept) 490 1045 0.29 0.50 0.58   Difference in means   1.51 0.57 2.64 **  Invasive slope over latitude   0.04 0.01 3.91 ***  Difference in slopes over latitude   -0.03 0.01 -2.61 ** Control        Bolting status (harvest) Invasive mean (intercept) 125 261 -25.67 17.19 -1.44   Difference in means   32.03 17.95 1.78 .  Invasive slope over latitude   0.47 0.37 1.25   Difference in slopes over latitude   -0.64 0.38 -1.68 . Herbivory        Bolting status (harvest) Invasive mean (intercept) 62 146 -172.96 242.39 -0.71   Difference in means   173.31 242.40 0.72   Invasive slope over latitude   3.51 4.94 0.71   Difference in slopes over latitude   -3.54 4.94 -0.72  Maternal Common Garden       Early Control        No. of basal leaves Invasive mean (intercept) 115 115 5.27 1.36 3.87 ***  Difference in means   -4.80 1.80 -2.67 **  Invasive slope over latitude   -0.05 0.03 -1.89 .  Difference in slopes over latitude   0.11 0.04 2.68 ** Control        Area of longest leaf (cm2) Invasive mean (intercept) 50 45 624.06 210.21 2.97 **  Difference in means   -656.30 299.52 -2.19 *  Invasive slope over latitude   -10.75 4.49 -2.39 *  Difference in slopes   13.91 6.73 2.07 *         Bolting status (harvest) Invasive mean (intercept) 55 56 9.53 16.04 0.59   Difference in means   -39.85 19.61 -2.03 *  Invasive slope over latitude   -0.24 0.34 -0.70   Difference in slopes over latitude   0.96 0.43 2.22 * Herbivory        Bolting status (harvest) Invasive mean (intercept) 12 12 28.68 62.74 0.46   Difference in means   -1204 275700 -0.004   Invasive slope over latitude   -0.07 1.358 -0.49   Difference in slopes over latitude   27.78 6355 0.004  Z ratios are presented from range differentiation models described in Tables A.4 and A.7 that include a significant interaction term only. Results are presented from restricted maximum likelihood (REML) models. Where no random effects were significant, generalized linear models (GLM) were used to test fixed effects. Models include all significant terms. Estimates and standard errors (SE) are untransformed.     129  Appendix B  Chapter 3 Supplementary Material B.1 Additional figures from the principal components analysis of environmental variables for C. diffusa populations used in the field experiment. Axis 1 and 2 are discussed within the main text. (a) Scree plot of eigenvalues for the top ten principal components. (b) PC1 versus PC3. Axis 3 explains 14% of variance, and is correlated with mean temperature of warmest quarter, annual mean temperature, and maximum temperature of the warmest month (BIO10, BIO1, BIO5; Table B.6).      130  B.2 Additional examples of morphological and stress tolerance trait divergence among Centaurea diffusa in the field experiment for traits with significant origin or origin-by-environment terms in range differentiation models. All figures are from observed data; model parameters are described in Table 3.1 and Table B.8. Shaded area represents standard error. (a) Population mean basal leaf number for three time points (>3 cm in length; significant origin term). (b) Population mean width of longest leaf along environmental cline, for three time points (significant origin-by-environment term). (c) Population mean rosette area at harvest along environmental cline (significant effect of origin-by-environment and treatment terms). (d) Population mean wilt date, among plants which wilted, along environmental cline (significant effect of origin-by-environment and treatment terms). (e) Population mean yellowing date, among plants which yellowed, along environmental cline (significant origin-by-environment term).   131    132  B.3 Principal components analysis of environmental variables for all C. diffusa geo-referenced occurrences recorded in GBIF.org. Axis 1 and 2 are discussed within the main text. (a) Scree plot of eigenvalues for the top ten principal components. (b) PC1 versus PC2. (c) PC1 versus PC3, including 95% confidence ellipse. Axis 3 explains 15% of variance, and is correlated with mean temperature of warmest quarter, annual mean temperature, and mean temperature of wettest quarter (BIO10, BIO1, BIO8; Table B.6).    133  B.4 Principal components analysis of environmental variables for a subset of C. diffusa geo-referenced occurrences recorded in GBIF.org. Because the GBIF data used here may not be error free, data were subdivided to include only populations within two standard deviations of the mean of PC1 and PC2 (613 populations) in the larger analysis of occurrence data. Axis 1 explains 35% of variance and correlated with the same top three bioclimatic variables as in the larger occurrence dataset (BIO17, BIO14, BIO2; Table B.6). Axes 2 and 3 explain 22% and 19% of variance respectively.  Relative to the larger occurrence dataset, top loadings for Axis 2 and Axis 3 are reversed (Axis 2: BIO11, BIO1, BIO6; Axis 3: BIO19, BIO16, BIO13). (a) Scree plot of eigenvalues for the top ten principal components. (b) PC1 versus PC2. Centroid of niche marked by large point, and significantly varied between ranges (between group inertia: 9.31%; P=0.001). (c) PC1 versus PC3. (b) and (c) include 95% confidence ellipses.   134      135  B.5 Centaurea diffusa experimental population information. Population ID Origin Collection year Latitude Longitude Voucher Accession BG001 Native C. diffusa 2008 43.38194 28.4575 V236763 GR001 Native C. diffusa 2008 40.15667 22.54806 V236766 GR002 Native C. diffusa 2008 40.62139 23.07861 V232702, V236769 HU001 Native C. diffusa 2008 47.64194 18.7825 V236764 RO001 Native C. diffusa 2008 44.11028 28.63694 V236767 RO005 Native C. diffusa 2005 45.49786 27.91181 V232686, V232683 RU008* Native C. diffusa 2006 44.05 43.06 V232687 TR001* Native C. diffusa 2008 41.75111 27.24778 V236765 CA001 Invasive C. diffusa 2008 49.01494 -122.882 V232677- V232679, V232704 - V232706 CA009 Invasive C. diffusa 2010 49.2961 -118.474  CA010 Invasive C. diffusa 2010 49.320181 -119.630  US001 Invasive C. diffusa 2008 45.61523 -120.788 V232694 - V232697 US002 Invasive C. diffusa 2008 46.18227 -118.826 V232669 - V232671 US003 Invasive C. diffusa 2008 46.60415 -116.642  Vouchers are located in the UBC Herbarium. * Indicates additional information on levels of introgression in that population, as follows. Population RU008 was verified as diploid (RU119; Blair et al., 2012) and 0 of 4 individuals from that population were determined to be introgressed from AFLP analysis (Blair and Hufbauer, 2010). An individual from population TR001 was used by Lai et al. (2012), to determine approximate level of introgression when compared to a tetraploid individual of Centaurea stoebe subsp. micranthos. In that study, the individual from TR001 was found to have a lower extent of introgression than an individual from the invaded range.         136  B.6 Abiotic environmental data variables used in principal components analyses (Hijmans et al., 2005). Bioclim variable Code Annual mean temperature BIO1 Mean diurnal temperature range (mean of monthly (max temp – min temp)) BIO2 Isothermality (BIO2/BIO7)*100 BIO3 Temperature seasonality (standard deviation*100) BIO4 Max temperature of warmest month BIO5 Min temperature of coldest month BIO6 Temperature annual range (BIO5-BIO6) BIO7 Mean temperature of wettest quarter BIO8 Mean temperature of driest quarter BIO9 Mean temperature of warmest quarter BIO10 Mean temperature of coldest quarter BIO11 Annual precipitation BIO12 Precipitation of wettest month BIO13 Precipitation of driest month BIO14 Precipitation seasonality (coefficient of variation) BIO15 Precipitation of wettest quarter BIO16 Precipitation of driest quarter BIO17 Precipitation of warmest quarter BIO18 Precipitation of coldest quarter BIO19 Altitude ALT Latitude LAT Data taken from current conditions (interpolations of observed data, representative of years 1950-2000) from ~1km tiles numbers 6, 7, 11,12,13,15,16,17,18.        137  B.7 Test statistics for all traits measured in the Montpellier field common garden from range differentiation models of phenotype of Centaurea diffusa.  Fixed effects Random effects Origin Env Origin-by-Env Treatment Population Maternal lines Repeat measure Trait χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P χ2 (df) P Number of basal leavesǂ 5.82 (1) * 0.49 (1)  0.03 (1) 0.12 (1)  0.81 (1)  0 (1)  552.87 (3) *** Length of longest leaf 3.14 (1) . 0 (1) 0.39 (1) . 2.13 (1) 7.03 (1) ** 12.76 (1) *** 134.42 (3) *** Width of longest leaf nt nt 8.50 (1) ** 0.82 (1) 1.02 (1)  0.85 (1) 132.38 (3) *** Rosette diameter 3.70 (1) . 0 (1)  0.39 (1) . 1.45 (1) 4.58 (1) *£  13.03 (1) *** 161.68 (3) *** Root crown diameter nt nt  9.88 (1) ** 0.82 (1) 14.89 (1) *** 16.33 (1) *** --- Rosette area at harvest nt nt  8.35 (1) ** 5.23 (1) * 24.23 (1) *** 3.16 (1) . --- Bolting probability nt nt 37.19 (1) *** 0.06 (1) 0 (1) 0 (1) --- Mortality 0.12 (1) 1.14 (1) 0.01 (1)  0.66 (1) ǂ1.30 (1)  0 (1) --- Specific leaf area 2.49 (1) 0.05 (1) 1.05 (1) 0.02 (1) 1.30 (1) 0 (1) --- Bolt date nt nt 9.34 (1) ** 0.07 (1) 4.84 (1) *£  0 (1) --- Shoot mass nt nt  14.44 (1) *** 1.71 (1) 14.82 (1) *** 9.71 (1) ** --- Wilt date nt nt 6.28 (1) * 21.42 (1) *** 4.76 (1) *£ 0 (1) --- Yellow date nt nt  25.89 (1) *** 0.46 (1) 4.45 (1) *£  0 (1) --- Results are presented from restricted maximum likelihood (REML) models. Significance of term indicated by symbol: ., P < 0.1; *, P < 0.05; **, P < 0.01;  ***, P < 0.001. Env = environment term. (df) = degrees of freedom. χ2 = chi-squared test statistic. Nt = not tested because of significant interaction term. ǂData scaled when necessary to improve model performance. £ Non-significant after FDR correction.      138  B.8 Parameter estimates of fixed effects from range differentiation models of Centaurea diffusa grown in Montpellier common field experiment that included a significant origin term or significant interaction between origin and environment. Trait Distribution, model Fixed effect Estimate SE Test statistic P Number of basal leaves Gaussian (scaled), REML Invasive mean (intercept) 0.37 0.02 15.86 ***   Difference between origins -0.07 0.03 -2.42 *        Leaf width Gaussian, REML Invasive mean (intercept) 3.32 0.11 30.15 ***   Difference between origins -0.31 0.14 -2.16 *   Invasive slope along environmental cline 0.07 0.04 1.86 .   Difference in slopes along environmental cline -0.22 0.07 -3.00 **        Root crown diameter Gaussian (loge), REML Invasive mean (intercept) 3.18 0.11 28.91 ***   Difference between origins -0.23 0.15 -1.60    Invasive slope along environmental cline -0.02 0.03 -0.65    Difference in slopes along environmental cline 0.23 0.07 3.27 **        Rosette area at harvest Gaussian (loge), REML Invasive mean (intercept) -1.36 0.12 -11.53 ***   Difference between origins -0.31 0.15 -2.03 .   Invasive slope along environmental cline -0.01 0.04 -0.39    Difference between treatments -0.15 0.07 -2.29 *   Difference in slopes along environmental cline 0.22 0.08 2.92 *        Shoot mass Gaussian (loge), REML Invasive mean (intercept) 3.99 0.19 20.50 ***   Difference between origins -0.53 0.26 -2.05 .   Invasive slope along environmental cline -0.05 0.06 -0.89    Difference in slopes along environmental cline 0.54 0.13 4.30 **        Bolting probability Binomial, REML Invasive mean (intercept) -0.94 0.24 -3.93 ***   Difference between origins 0.91 0.31 2.93 **   Invasive slope along environmental cline 0.19 0.09 2.16 *   Difference in slopes along environmental -1.11 0.22 -4.99 ***  139  Trait Distribution, model Fixed effect Estimate SE Test statistic P cline Bolt date Poisson, REML Invasive mean (intercept) 3.89 0.03 133.95 ***   Difference between origins -0.02 0.04 -0.46    Invasive slope along environmental cline 0.003 0.01 0.39    Difference in slopes along environmental cline 0.04 0.01 3.04 **        Wilt date Poisson, REML Invasive mean (intercept) 3.80 0.02 210.06 ***   Difference between origins 0.02 0.02 0.80    Invasive slope along environmental cline -0.02 0.01 -2.65 **   Difference between treatments -0.11 0.02 -4.60 ***   Difference in slopes along environmental cline 0.04 0.01 2.49 *        Yellow date Poisson, REML Invasive mean (intercept) 3.88 0.02 246.85 ***   Difference between origins -0.002 0.02 -0.11    Invasive slope along environmental cline -0.003 0.01 -0.46    Difference in slopes along environmental cline 0.08 0.02 5.01 *** Test statistics are presented from range differentiation models, described in Table B.7. Results are presented from restricted maximum likelihood (REML) models. For Gaussian distributed models, T values are reported for the test statistic; for Binomial and Poisson distributed models, Z ratios are reported. Models include all significant terms. Significance of term indicated by symbol: ., P < 0.1; *, P < 0.05; **, P < 0.01;  ***, P < 0.001.     140  Appendix C  Chapter 4 Supplementary Material C.1 Heat map of genes with drought-induced expression differences between native and invasive Centaurea diffusa for the drought treatment from time point 2. A heat map of the 227 induced differently expressed genes for drought samples only from the final time point. All genes had a significant origin by treatment interaction, determined by mixed model analysis. Normalized, scaled, expression values are displayed. Heatmaps were drawn in R with heatmap.2 and both the genes (rows) and samples (columns) were clustered using dendrograms. Dendrograms were constructed using Euclidean distances and hierarchical clustering. Origin is displayed in the colored bar above the heatmap (red for native populations, and black for invasive).   141  C.2 Heat map of genes with drought induced expression differences between native and invasive Centaurea diffusa for the control treatment from time point 2. A heat map of the 227 induced differently expressed genes for control samples only from the final time point. All genes had a significant origin by treatment interaction, determined by mixed model analysis. Normalized, scaled, expression values are displayed. Heatmaps were drawn in R with heatmap.2 and both the genes (rows) and samples (columns) were clustered using dendrograms. Dendrograms were constructed using Euclidean distances and hierarchical clustering. Origin is displayed in the colored bar above the heatmap (red for native populations, and black for invasive).     142  C.3 Centaurea diffusa experimental population information. Population ID Origin Collection year Latitude Longitude PC1 value Voucher Accession BG001 Native 2008 43.38194 28.4575 -0.4691047 V236763 RU008 Native 2006 44.05 43.06 2.5006967 V232687 TR001 Native 2008 41.75111 27.24778 0.1627519 V236765 CA001 Invasive  2008 49.01494 -122.882 -5.6572239 V232677- V232679, V232704 - V232706 US001 Invasive  2008 45.61523 -120.788 2.4523743 V232694 - V232697 US003 Invasive  2008 46.60415 -116.642 1.0105056  Vouchers are located in the UBC Herbarium. PC1 value was determined by principal components analysis described in text, and was used in all mixed model analyses.   C.4 Results from likelihood ratio tests for genes with significant model terms. Contig Name TAIR Accessions Origin-by- Treatment Origin Treatment PC1 GO Terms Contig100 AT1G51470,AT1G21065 0.579277 1 0.001302 1 GO:0008150, GO:0009507, GO:0004553, GO:0008422, GO:0009651, GO:0009651, GO:0012505, GO:0019137 Contig10001 AT5G39820,AT3G59140 0.356843 1 0.007906 0.64572 GO:0003700, GO:0005575, GO:0006355, GO:0006355, GO:0007275, GO:0005774, GO:0005774, GO:0005774, GO:0006810, GO:0042626, GO:0042626, GO:0055085, GO:0055085 LRTs are reported for each fixed effect term in mixed effect models of gene expression, for all 11,047 genes with at least one significant fixed effect (α=0.05). Q values corrected for multiple comparisons are reported for each LRT. TAIR 10 accessions (The Arabidopsis Information Resource; arabidopsis.org) and GO term mappings are included where possible. [Sample table only. Full table available at http://dx.doi.org/10.6084/m9.figshare.1372537 (Turner 2015).]      143  C.5 Gene Ontology terms for genes with a significant effect of drought treatment. GO ID Term P-value GO:2000112 regulation of cellular macromolecule bio... < 0.0001 GO:0010468 regulation of gene expression < 0.0001 GO:0031326 regulation of cellular biosynthetic proc... < 0.0001 GO:0010556 regulation of macromolecule biosynthetic... < 0.0001 GO:0009889 regulation of biosynthetic process < 0.0001 GO:0032774 RNA biosynthetic process < 0.0001 Results of Fisher’s exact tests examining the number of gene associations with GO terms. Genes of interest were determined in mixed model analysis to have a significant differences between control and drought treatments. Ontology type is indicated for each term; here, biological process only (BP). The analysis was conducted using the Fisher algorithm and parent-child method in topGO, and P-values for this test are reported. Direction of expression was determined by relative range level averages of expression all first time point samples, before onset of treatment, and is expressed in terms of invasive relative to native samples. Separate topGO analyses of up or down regulated genes were then used to confirm association with GO terms. [Sample table only. Full table available http://dx.doi.org/10.6084/m9.figshare.1372537 (Turner 2015).]      144  C.6 Gene Ontology terms for gene with a significant effect of seed collection environment (PC1). GO.ID Term P-value GO:0006506 GPI anchor biosynthetic process 1.70E-05 GO:0009247 glycolipid biosynthetic process 7.10E-05 GO:0010260 organ senescence 0.0001 GO:0030091 protein repair 0.00055 GO:0007568 aging 0.00111 GO:0042780 tRNA 3'-end processing 0.00185 GO:0007154 cell communication 0.00188 GO:0000255 allantoin metabolic process 0.00239 GO:0006664 glycolipid metabolic process 0.00388 GO:0010467 gene expression 0.00407 GO:0060560 developmental growth involved in morphog... 0.00415 GO:0006661 phosphatidylinositol biosynthetic proces... 0.00424 GO:0016070 RNA metabolic process 0.00495 GO:0006629 lipid metabolic process 0.00525 GO:0010275 NAD(P)H dehydrogenase complex assembly 0.00539 GO:0006505 GPI anchor metabolic process 0.0062 GO:0007165 signal transduction 0.0065 GO:0050665 hydrogen peroxide biosynthetic process 0.00691 GO:0044700 single organism signaling 0.00695 GO:0006354 DNA-templated transcription, elongation 0.00727 GO:1903409 reactive oxygen species biosynthetic pro... 0.00765 GO:0055114 oxidation-reduction process 0.00819 Results of Fisher’s exact tests examining the number of gene associations with GO terms. Genes of interest were determined in mixed model analysis to have significant have a significant effect of composite abiotic environment for seed collection location (PC1). Ontology type is indicated for each term; here, biological process only (BP). The analysis was conducted using the Fisher algorithm and parent-child method in topGO, and P-values for this test are reported. Direction of expression was determined by relative range level averages of expression all first time point samples, before onset of treatment, and is expressed in terms of invasive relative to native samples. Separate topGO analyses of up or down regulated genes were then used to confirm association with GO terms.     145  Appendix D  Genomic Resources for Centaurea diffusa Over the course of this dissertation, and with the collaboration of several others (indicated from citation for each table or figure, and described in the Preface), I produced some of the first genomic resources for this non-model weed system. These include: D.1 Centaurea diffusa EST libraries.  Sample range Collection ID Coordinates No. reads Total sequence (Mbp) No. unigenes Total assembly length (Mbp) Dryad doi Native  DK TR001-1L 41.7511, 27.2478 407817 183 48936 31 10.5061/dryad.cm7td/5 Invasive DK US022-31E 45.745, -119.785 631874 308 61749 43 10.5061/dryad.cm7td/6 Described in Lai et al. 2012. Both libraries were produced from normalized, double-stranded libraries made from leaf tissue, and sequenced using 454.  D.2 Microarray developed for Centaurea diffusa. Platform Collection ID No. unigenes No. features NimbleGen, 12-plex Primarily DK TR001-1L, some contribution from DK US022-31E 61024 136906 Described in Lai et al. 2012. Microarray used in Chapter 4.         146  D.3 Gene Ontology terms overrepresented for rapidly evolving genes identified in invasive Centaurea diffusa. GO.ID Term P-value GO:0017126 nucleologenesis 0.0075 GO:0010019 chloroplast-nucleus signaling pathway 0.01 GO:0051302 regulation of cell division 0.0248 GO:0009451 RNA modification 0.0281 GO:0051202 phytochromobilin metabolic process 0.0294 GO:0033014 tetrapyrrole biosynthetic process 0.0355 GO:0006997 nucleus organization 0.0383 GO:0010024 phytochromobilin biosynthetic process 0.04 GO:0006413 translational initiation 0.0482 Described in Hodgins et al. 2015. Gene Ontology terms identified from 104 loci with dN/dS > 1 for pairwise comparisons between native and introduced transcriptomes. Significance of Gene Ontology terms assigned using the Fisher test and parent-child method, as in Chapter 4 (P < 0.05).  D.4 Centaurea diffusa plastome assembly output numerics.  Contigs Contigs Scaffolds Scaffolds Size range ≥ 100 nt ≥ 500 nt ≥ 100 nt ≥ 500 nt Number 21 14 14 7 Total length 128037 126739 129501 128203 Average 6097 9052 9250 18314 N50 23320 23320 27489 27489 Median 1234 2960 1112 11940 Largest 52890 52890 52890 52890 Described in Turner & Grassa 2014[PRE-PRINT]. Assembly numerics of C. diffusa chloroplast genome, using Ray assembly method.   147  D.5 Annotated plastome for Centaurea diffusa.  Described in Turner & Grassa 2014 [PRE-PRINT]. Map of annotated Centaurea diffusa chloroplast genome, produced using OGDraw (Lohse et al., 2013). 

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:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0166509/manifest

Comment

Related Items