GitBucket
Newer
zeynalig
<?xml version="1.0" ?> <tei> <teiHeader> <fileDesc xml:id="_ele0017-1464"/> </teiHeader> <text xml:lang="en"> <front>R E V I E W A N D <lb/>S Y N T H E S I S <lb/> Hybridisation is associated with increased fecundity and size in <lb/>invasive taxa: meta-analytic support for the hybridisation-<lb/>invasion hypothesis <lb/> Stephen M. Hovick 1 * and Kenneth <lb/>D. Whitney 2 <lb/> 1 Department of Evolution, Ecology, <lb/>and Organismal Biology The Ohio <lb/>State University Columbus, <lb/>OH,43210, USA <lb/> 2 Department of Biology University <lb/>of New Mexico Albuquerque, <lb/>NM,87131, USA <lb/>*Correspondence: <lb/>E-mail: hovick.2@osu.edu <lb/> Abstract <lb/> The hypothesis that interspecific hybridisation promotes invasiveness has received much recent <lb/>attention, but tests of the hypothesis can suffer from important limitations. Here, we provide the <lb/>first systematic review of studies experimentally testing the hybridisation-invasion (H-I) hypothesis <lb/>in plants, animals and fungi. We identified 72 hybrid systems for which hybridisation has been <lb/>putatively associated with invasiveness, weediness or range expansion. Within this group, 15 sys-<lb/>tems (comprising 34 studies) experimentally tested performance of hybrids vs. their parental spe-<lb/>cies and met our other criteria. Both phylogenetic and non-phylogenetic meta-analyses <lb/>demonstrated that wild hybrids were significantly more fecund and larger than their parental taxa, <lb/>but did not differ in survival. Resynthesised hybrids (which typically represent earlier generations <lb/>than do wild hybrids) did not consistently differ from parental species in fecundity, survival or <lb/>size. Using meta-regression, we found that fecundity increased (but survival decreased) with gener-<lb/>ation in resynthesised hybrids, suggesting that natural selection can play an important role in <lb/>shaping hybrid performance – and thus invasiveness – over time. We conclude that the available <lb/>evidence supports the H-I hypothesis, with the caveat that our results are clearly driven by tests in <lb/>plants, which are more numerous than tests in animals and fungi. <lb/> Keywords <lb/> Adaptive evolution, colonisation, hybridisation, introgression, invasion genetics, phylogenetic <lb/>meta-analysis, polyploidy, range expansion, weeds. <lb/> Ecology Letters (2014) 17: 1464–1477 <lb/></front> <body>INTRODUCTION <lb/> The study of what determines invasion success in non-native <lb/>organisms has a long history (Elton 1958; Baker 1974), <lb/>although not always a successful one (Perrins et al. 1992; <lb/>Mack 1996). Genetic and evolutionary factors have been con-<lb/>sidered important (Baker & Stebbins 1965; Crawford & <lb/>Whitney 2010), including the hypothesis that interspecific hy-<lb/>bridisation may promote invasiveness (Stebbins 1985; Abbott <lb/>1992). This Hybridisation-Invasion (H-I) hypothesis has <lb/>received particular attention in the past decade-and-a-half fol-<lb/>lowing the publication of Ellstrand & Schierenbeck (2000). In <lb/>addition to summarising the mechanisms by which hybridisa-<lb/>tion could enhance invasiveness, Ellstrand and Schierenbeck <lb/>compiled preliminary support for the hypothesis via a list of <lb/>28 taxa in which hybridisation has preceded invasiveness. This <lb/>original list has since been expanded to 35 hybrid taxa <lb/>(Schierenbeck & Ellstrand 2009), contributing to the increas-<lb/>ing acceptance of the idea that hybridisation can be a driver <lb/>of biological invasions (e.g. Darling et al. 2008; Le Roux & <lb/>Wieczorek 2009; but see Whitney et al. 2009). As species are <lb/>transported globally with increasing frequency, invasive <lb/>hybrids will continue being introduced to new regions, and <lb/>new opportunities for interspecific mating will undoubtedly <lb/>result in the formation of novel hybrid taxa. The hypothesised <lb/>connection between hybridisation and invasiveness is, <lb/>therefore, likely to remain a critical issue, both for conserva-<lb/>tion and for better understanding the evolutionary ecology of <lb/>colonising species. However, despite rapidly growing interest <lb/>in this topic, we currently lack a comprehensive assessment of <lb/>the evidence either supporting or refuting a causal link <lb/>between hybridisation and invasiveness. <lb/>Ellstrand & Schierenbeck (2000) outlined several, non-mutu-<lb/>ally exclusive, mechanisms by which hybridisation could <lb/>enhance invasiveness; we highlight them briefly here (see also <lb/>Rieseberg et al. 2007). First, hybridisation can create novel <lb/>phenotypes relative to the parental taxa, increasing the likeli-<lb/>hood of survival and establishment success in novel habitats. <lb/>Such novelty includes transgressive phenotypes, where hybrids <lb/>exhibit trait values that fall outside the range of their parents <lb/>(Rieseberg et al. 1999), as well as novel combinations of <lb/>parental phenotypic traits (Hovick et al. 2012). Second, hy-<lb/>bridisation can lead to increased phenotypic and genetic varia-<lb/>tion relative to the parental taxa, which may help hybrids <lb/>better cope with environmental stochasticity and increase their <lb/>evolutionary potential (Anderson & Stebbins 1954; Stebbins <lb/>1959). Heterosis is a special case of increased genetic varia-<lb/>tion, where hybrids (particularly F 1 s) experience performance <lb/>gains due to increased heterozygosity. When hybridisation is <lb/>accompanied by mechanisms that stabilise heterotic lineages <lb/>(i.e. polyploidy, clonal growth or agamospermy), the resulting <lb/>hybrids may experience increased invasiveness (e.g. Parepa <lb/></body> <front> © 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/>This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and <lb/>distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. <lb/></front> <front>Ecology Letters, (2014) 17: 1464–1477 <lb/>doi: 10.1111/ele.12355 <lb/></front> <body>et al. 2014). Finally, if the parental taxa are relatively isolated <lb/>and occur in small populations, hybridisation could lead to <lb/>the purging of genetic load, and the resulting fitness boost <lb/>could increase invasiveness (Ellstrand & Schierenbeck 2000). <lb/>To our knowledge, this final mechanism has not yet been <lb/>demonstrated empirically. <lb/>Establishing a causal relationship between hybridisation and <lb/>invasiveness can be challenging. A first challenge is that, <lb/>because many hybrid taxa are sterile, performance assessments <lb/>cannot always rely on fecundity as an indicator of potential <lb/>population growth rates. Second, the outcomes of perfor-<lb/>mance comparisons between hybrids and parents frequently <lb/>depend on which hybrid class (F 1 , BC 1 , etc.) is examined <lb/>(Arnold & Hodges 1995; Arnold & Martin 2010). For exam-<lb/>ple, hybrids experiencing heterosis in the F 1 generation may <lb/>experience hybrid breakdown as segregating genes in subse-<lb/>quent generations lead to decreasing fitness (Hooftman et al. <lb/> 2007). Even within the same hybrid class, hybrid individuals <lb/>can be highly variable, resulting in biologically meaningful dif-<lb/>ferences among lineages with differing parental backgrounds <lb/>(Py sek et al. 2003; Hartman et al. 2013). Third, the environ-<lb/>mental context can determine relative performance of hybrids <lb/>vs. parents (Arnold & Martin 2010), and in at least one case, <lb/>hybrids outperform their parental taxa by way of different <lb/>traits in different environments (Hovick et al. 2012). These <lb/>considerations suggest that the performance metric, the choice <lb/>of hybrid material and the environmental setting all need to <lb/>be carefully considered to provide meaningful tests of the H-I <lb/>hypothesis. <lb/>A further challenge in evaluating the H-I hypothesis is that <lb/>observational studies documenting increasing frequencies of <lb/>hybrids relative to their parental taxa tell us nothing definitive <lb/>about relative hybrid performance and therefore invasiveness. <lb/>Even in the absence of a hybrid advantage, interspecific mat-<lb/>ing will often lead to at least low levels of introgression of <lb/>selectively neutral alleles, creating hybrid swarms (e.g. Hassel-<lb/>man et al. 2014). Yet, some studies of hybrid invaders inter-<lb/>pret high abundances of hybrids relative to their parental taxa <lb/>as evidence that hybridisation has promoted invasiveness in <lb/>that taxon (e.g. Urbanska et al. 1997; Moody & Les 2002; <lb/>Tavalire et al. 2012; Wu et al. 2013). This practice can result <lb/>in unsubstantiated claims regarding the role of hybridisation <lb/>being propagated throughout the literature. <lb/>Because of these challenges, testing the H-I hypothesis <lb/>clearly requires an experimental approach. At the most basic <lb/>level, these experiments must compare hybrid performance <lb/>with that of the parental taxa in a common environment in <lb/>order to remove the effects of environmental variation and <lb/>historical chance while focusing on genetically based perfor-<lb/>mance differences among taxa. In any such experiment, the <lb/>proper parental taxa must be used as benchmarks for com-<lb/>parison with hybrid performance. Ideally, both parents <lb/>would be included in any such tests, but at a minimum the <lb/>more invasive parental taxon must be. For example, if an <lb/>invasive hybrid results from hybridisation between a rela-<lb/>tively benign native species and an introduced invasive <lb/>species, then performance assessments that exclude the inva-<lb/>sive parent are unlikely to be informative: the invasive par-<lb/>ent may have been capable of spreading just as rapidly and <lb/>achieving the same densities even if hybridisation had not <lb/>occurred. Unfortunately, a surprising number of studies fail <lb/>to include comparisons to the more invasive parent (see <lb/>Discussion, below). <lb/>Previous compilations of taxa that are both invasive and <lb/>hybrid-derived (Ellstrand & Schierenbeck 2000; Schierenbeck <lb/>& Ellstrand 2009) represent a vital first step in determining <lb/>whether hybridisation and invasiveness are causally related. <lb/>However, these compilations were not intended to be exhaus-<lb/>tive. More importantly, given that hybridisation is not rare <lb/>(animals: Schwenk et al. 2008; plants: Whitney et al. 2010a; <lb/>fungi: Giraud et al. 2008), we should expect some fraction of <lb/>invasive taxa to be hybrids, even in the absence of an under-<lb/>lying causal relationship. Our aims are to clarify the criteria <lb/>needed for evaluating the H-I hypothesis and to evaluate <lb/>tests meeting these criteria that have been conducted to date. <lb/>We perform a systematic review to address three questions: <lb/>(1) based on a comprehensive evaluation of the literature, <lb/>how frequently has hybridisation been associated with inva-<lb/>siveness in plants, animals, and fungi? (2) to what extent has <lb/>the H-I hypothesis been evaluated experimentally? and (3) <lb/>do those evaluations support or refute the H-I hypothesis? <lb/>We expand earlier compilations of invasive hybrids (Schier-<lb/>enbeck & Ellstrand 2009) by including a number of new <lb/>hybrid taxa. In addition, we present recommendations for <lb/>designing future tests of whether hybridisation promotes <lb/>invasiveness. <lb/> METHODS <lb/> To investigate the H-I hypothesis, we carried out a systematic <lb/>review (sensu Liberati et al. 2009) that included meta-analy-<lb/>ses. In doing so, we followed the PRISMA statement (Liberati <lb/> et al. 2009) to the extent possible; note that PRISMA was <lb/>developed for systematic reviews in medicine and thus not all <lb/>provisions are relevant here. Our first step was to conduct an <lb/>exhaustive literature review in an attempt to identify all plant, <lb/>animal, and fungal taxa for which interspecific hybridisation <lb/>is postulated to have enhanced invasiveness or contributed to <lb/>geographic range expansions. We then assessed the extent to <lb/>which this hypothesis has been tested experimentally, and then <lb/>quantified the effects of hybridisation on invasiveness using <lb/>meta-analysis. Following Richardson et al. (2000), we define <lb/>invasiveness as a measure of the ability of a species, without <lb/>human assistance, to increase in population size and spatial <lb/>distribution following introduction. However, here we have <lb/>relaxed the 'introduction' criterion to also include a handful <lb/>of systems where both parents are native to the region from <lb/>which a hybrid has subsequently spread. Thus, our dataset <lb/>includes putatively invasive hybrids originating via a number <lb/>of possible scenarios: (1) hybridisation in the native range, fol-<lb/>lowed by local or regional range expansion, (2) hybridisation <lb/>in the native range, followed by introduction to a novel <lb/>region, (3) hybridisation in the novel range between two (or <lb/>more) introduced parents, and (4) hybridisation in the novel <lb/>range between one introduced and one native parent. These <lb/>scenarios are not mutually exclusive because in some systems <lb/>hybrids have been introduced repeatedly to novel regions and/ <lb/>or the parental taxa recurrently hybridise. <lb/> <note place="footnote"> © 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/></note> <note place="headnote">Review and Synthesis <lb/>The hybridisation-invasion hypothesis</note> <page>1465 <lb/></page> Literature search criteria <lb/> On May 10, 2012 (and again on August 5, 2013 to update our <lb/>dataset), we searched Thomson Reuters Web of Science <lb/>(http://apps.webofknowledge.com) for the keywords 'hybrid*' <lb/>and 'inva*', subsequently limiting our records to those in the <lb/>following categories (to minimise references from the medical <lb/>literature): Ecology, Evolutionary Biology, Plant Sciences, <lb/>Marine and Freshwater Biology, Entomology, Zoology, <lb/>Fisheries and Agronomy. This search yielded 945 references. <lb/>On the same days, we also conducted a forward citation <lb/>search on three seminal papers that address the implications <lb/>of hybridisation for biological invasions (Stebbins 1985; Ab-<lb/>bott 1992; Ellstrand & Schierenbeck 2000); this added 593 ref-<lb/>erences to our list. We read all 1538 abstracts and every study <lb/>in which an invasive or weedy species had been determined to <lb/>be a hybrid, retaining studies in which hybridisation was <lb/>linked to invasiveness. Additional relevant references cited by <lb/>these studies were added to our database as we encountered <lb/>them. We kept a running list of hybrid systems that fit our <lb/>criteria (see following paragraph), and once we had finished <lb/>reviewing the studies in our database we conducted one addi-<lb/>tional taxon-specific search for each of the 72 hybrid systems <lb/>identified (last updated August 5, 2013) to ensure we had not <lb/>overlooked any experimental parent-hybrid comparisons in <lb/>those systems. These study selection steps are summarised in <lb/>Appendix 1, and Appendix 2 presents the resulting list of 72 <lb/>systems in which hybridisation has a presumed link to inva-<lb/>siveness. We then confirmed that the use of a single search <lb/>engine did not result in appreciable numbers of relevant stud-<lb/>ies being missed (Appendix 3). <lb/>Our search included hybrid plant, animal and fungal taxa <lb/>(including the 'fungoid' stramenopiles; bacteria were omitted). <lb/>We excluded hybrids that have only been synthesised experi-<lb/>mentally (i.e. have no analogues in nature), and we excluded <lb/>transgenic hybrids where hybrid advantage is linked to trans-<lb/>formed genes such as those providing herbicide resistance. <lb/>While there are many fascinating systems where intraspecific <lb/>hybridisation (i.e. 'wide crossing') may have enhanced inva-<lb/>siveness (Lavergne & Molofsky 2007; Culley & Hardiman <lb/>2009; Geiger et al. 2011), we limited our review to interspecific <lb/>hybridisation. Finally, one interspecific hybrid system identi-<lb/>fied previously (Schierenbeck & Ellstrand 2009) as a potential <lb/>example of the H-I process was excluded because recent evi-<lb/>dence argues against its hybrid origin ('strawhull' rice: Reagon <lb/> et al. 2011; Sun et al. 2013). <lb/> Assessing the literature: what do we know about hybrid invaders? <lb/> For each hybrid system we recorded basic biological informa-<lb/>tion (see Appendix 2) including growth form (for plants: tree, <lb/>herb, grass), life history (e.g. annual, perennial) and ploidy of <lb/>the parental and hybrid taxa. We recorded whether the paren-<lb/>tal taxa had commercial origins (i.e. horticulture, agriculture <lb/>or the pet trade), the provenance of parental taxa (native or <lb/>introduced) and whether hybridisation occurred before or <lb/>after introduction to a novel region. <lb/>Lastly, for each system we collected information about the <lb/>strength of evidence supporting the H-I hypothesis. We noted <lb/>whether such inferences were based on ad hoc observations, <lb/>observational studies (measurements taken in naturally occur-<lb/>ring populations), lab-or greenhouse-based experimental <lb/>studies, or field-based experimental studies conducted in a <lb/>common environment (including common garden and meso-<lb/>cosm experiments). We recorded the metrics by which hybrid <lb/>performance was assessed and the parental taxa to which <lb/>hybrid performance was compared. Finally, we recorded <lb/>which hybrid class(es) were investigated. <lb/> Meta-analysis: do hybrids outperform parents? <lb/> For the meta-analysis, we identified every study in our dataset <lb/>conducting an experimental hybrid-parent performance com-<lb/>parison, including field-, greenhouse-, and lab-based experi-<lb/>ments. We excluded non-experimental observations from <lb/>natural populations, as this approach does not permit taxon-<lb/>specific (i.e. genetic) effects to be separated from environmen-<lb/>tal effects. We also excluded performance comparisons that <lb/>were unable to test the H-I hypothesis because of uninforma-<lb/>tive benchmarks (e.g. studies failing to compare a hybrid with <lb/>its most invasive parental taxon). We extracted performance <lb/>data for hybrids and their parental taxa, grouping perfor-<lb/>mance metrics into three categories: fecundity (including for <lb/> Lactuca, the product of germination rate, survival rate and <lb/>seed output plant À1 ; see Hooftman et al. 2005), survival and <lb/>organism size. We consider these individual performance met-<lb/>rics to be potential components of population growth rates <lb/>and thus proxies for invasiveness, while acknowledging that <lb/>they can be imperfect indicators (see Discussion). For plants, <lb/>roughly half of the reported fecundity estimates included <lb/>values of zero for individuals that died before setting seed <lb/>(Helianthus, Lactuca and Sorghum), while other estimates <lb/>excluded pre-reproduction mortality (Reynoutria, Senecio <lb/>vulgaris and Senecio squalidus); Raphanus fecundity estimates <lb/>in our database represent a combination of both approaches. <lb/>Estimates including pre-reproduction mortality are likely to <lb/>be more robust indicators of population growth rates (see <lb/>Discussion), but we included both types of estimates for meta-<lb/>analyses. For the two hybrid fungal pathogens in our dataset <lb/>only pathogenicity was reported, and we considered this a <lb/>fecundity proxy for our meta-analysis. We extracted means <lb/>and standard deviations (or the data needed to calculate <lb/>them) from tables and figures, using in the latter case the <lb/>object measurement tool in Adobe Acrobat Professional <lb/>version 7 (Adobe Systems Inc., San Jose, CA, USA). <lb/>We used Hedges' d as our effect size metric: <lb/> d ¼ <lb/> Y Hyb À <lb/> Y Par <lb/> s <lb/>J; <lb/> where <lb/> Y Hyb and <lb/> Y Par are mean performance of the hybrid and <lb/>parental taxa, s is the pooled standard deviation and J is a <lb/>small-sample correction factor (Rosenberg et al. 2013). Posi-<lb/>tive values for d thus indicate hybrids that outperform their <lb/>parents, as predicted by the H-I hypothesis. If multiple hybrid <lb/>classes were considered in a single experiment, we extracted <lb/>data separately by hybrid class (e.g. resynthesised F 1 vs. wild <lb/>hybrids). Some of the studies in our meta-analysis reported <lb/>data from both parental taxa, in which case we calculated <lb/> <note place="footnote"> © 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/></note> <page>1466</page> <note place="headnote">S. M. Hovick and K. D. Whitney <lb/> Review and Synthesis <lb/></note> Hedges' d twice, once for the hybrid relative to the more <lb/>'invasive' parent (i.e. that with higher fecundity, larger size or <lb/>higher survival), and a second time for the hybrid relative to <lb/>mean performance of both parents. <lb/>We performed a series of data-aggregation steps to mini-<lb/>mise non-independence in our dataset. When multiple perfor-<lb/>mance metrics fitting one of our categories were reported in <lb/>a single study, we selected the most comprehensive (e.g. total <lb/>biomass instead of aboveground biomass). For studies con-<lb/>ducted using multiple experimental treatments, we selected <lb/>the treatment most closely approximating natural conditions <lb/>in the field (i.e. herbivores present); lacking an obvious selec-<lb/>tion on that basis, we chose one at random. We used fixed-<lb/>effect meta-analysis with inverse variance weighting to aggre-<lb/>gate effect sizes from multiple sites or habitats within a sin-<lb/>gle study (Mengersen et al. 2013), doing so separately for <lb/>each parental taxon and hybrid class. We used weighted <lb/>means to aggregate data from multiple hybrid classes within <lb/>an experiment, using two alternative aggregation approaches. <lb/>In the first, we simply combined data from all hybrid classes; <lb/>this can be used to calculate an overall hybrid effect, but it <lb/>ignores meaningful distinctions between the types of hybrids <lb/>used across studies. Thus, in the second approach we distin-<lb/>guished three hybrid categories: (1) naturally occurring, or <lb/>'wild' hybrids usually assumed to be advanced-generation <lb/>hybrids, (2) resynthesised F 1 hybrids from experimental <lb/>crosses, and (3) resynthesised post-F 1 hybrids from experi-<lb/>mental crosses (a broad category, ranging from F 2 and BC 1 <lb/> to F 10 hybrids). Finally, for hybrid taxa that had been stud-<lb/>ied in multiple studies, we aggregated effect sizes using one <lb/>final fixed-effects meta-analysis, yielding no more than one <lb/>effect size per species in every analysis. We conducted ran-<lb/>dom-effect meta-analyses with inverse variance weighting on <lb/>these data in R (version 3.0.2; R Development Core Team <lb/>2013) using the metafor package (version 1.9-3; Viechtbauer <lb/>2010). As a measure of heterogeneity in our analyses we <lb/>present I 2 values for each model (Higgins & Thompson <lb/>2002). I 2 ranges from 0 to 100% and can be interpreted as <lb/>the proportion of variability in a given effect size estimate <lb/>due to between-study heterogeneity. <lb/>Because we expected the group of resynthesised post-F 1 <lb/> hybrids to be a heterogeneous sample, we also created a sub-<lb/>set of our full dataset for more detailed analyses using meta-<lb/>regression (Mengersen et al. 2013). For the few hybrid taxa in <lb/>which performance data from multiple hybrid classes were <lb/>reported, we re-aggregated effect sizes within and among stud-<lb/>ies (as above) to yield a single effect size per taxon and per <lb/>hybrid class. This gave distinct effect sizes for wild hybrids, <lb/>F 1 s, generation 2 hybrids (including F 2 , BC 1 and progeny of <lb/>selfed F 1 s), generation 3 hybrids (including F 3 , BC 2 and prog-<lb/>eny of selfed F 2 s) and so on. We included hybrid class as a <lb/>continuous, fixed effect in meta-regressions, which tested for <lb/>significant increases or decreases in effect size across genera-<lb/>tions. Because hybrid class is unknown for wild hybrids in the <lb/>dataset, we omitted them from these analyses. To account for <lb/>multiple effect sizes per species, we used a multivariate <lb/>approach in metafor (the rma.mv command), specifying hybrid <lb/>taxon as a random factor and adding a study-level random <lb/>variable to achieve a random-effects meta-regression. Where <lb/>the hybrid class effect was significant it was sometimes incon-<lb/>sistent across taxa, so in these cases we re-ran random-effects <lb/>meta-regressions separately for each hybrid taxon in the data-<lb/>set (note that data are lacking to assess taxon 9 class interac-<lb/>tions statistically). All meta-regressions included an intercept, <lb/>although we report only on inferences regarding significance <lb/>of the slopes. <lb/>Because all species have a shared evolutionary history, <lb/>non-independence among species is an important consider-<lb/>ation for meta-analyses where effect sizes can be mapped <lb/>onto phylogenies (Lajeunesse 2009; Chamberlain et al. 2012). <lb/>We conducted phylogenetic meta-analyses using PhyloMeta <lb/>version 1.3 (Lajeunesse 2011) to account for this non-inde-<lb/>pendence. A phylogeny was constructed as follows: a base <lb/>topology of plant taxa was derived from Phylomatic v.3 <lb/>(Webb & Donoghue 2005) using tree R20120829; resolution <lb/>within Asteraceae was added using topologies in Funk et al. <lb/> (2009); animal and fungal taxa were added based on Maddi-<lb/>son & Schulz (2007). We aged internal nodes for the phylog-<lb/>eny of plant taxa based on Wikstr€ om et al. (2001), using the <lb/>bladj algorithm in Phylocom to interpolate ages of undated <lb/>nodes (Webb et al. 2008). We added ages to all remaining <lb/>nodes using the Expert Result dates from Hedges et al. <lb/> (2006). This base tree with all taxa and node ages is pre-<lb/>sented in Appendix 4. Phylogenetic meta-analysis results did <lb/>not differ qualitatively from non-phylogenetic results. <lb/>Because of this concordance and because PhyloMeta does <lb/>not support meta-regression, for the sake of consistency and <lb/>ease of interpretation we present results only from non-phy-<lb/>logenetic meta-analyses and meta-regressions in the main <lb/>text; see Appendix 4 for phylogenetic meta-analysis results. <lb/>The results we present make no distinction between experi-<lb/>ments conducted in the field vs. lab or greenhouse experi-<lb/>ments. Where sample sizes were large enough to run <lb/>analyses using only field-collected data, outcomes were simi-<lb/>lar to those using the more complete dataset (data not <lb/>shown); heterogeneity tests also indicated similar effect sizes <lb/>in field-vs. lab-based investigations (Appendix 5; see Table 1 <lb/>for Field/Lab designations). Because of limited sample sizes, <lb/>we present comparisons between hybrids and their 'more <lb/>invasive' parent only. Comparisons of hybrid performance to <lb/>the average performance of both parents yielded similar <lb/>results, although with slightly larger effect sizes (data not <lb/>shown); the results presented below can therefore be consid-<lb/>ered conservative. <lb/> RESULTS <lb/> Including those that have been identified in previous reviews <lb/>(Ellstrand & Schierenbeck 2000; Schierenbeck & Ellstrand <lb/>2009), we found 72 hybrid systems (i.e. a hybrid taxon and <lb/>its two or more parental species) for which hybridisation has <lb/>been putatively associated with invasiveness, weediness or <lb/>range expansion (Appendix 2). The vast majority are plants <lb/>(n = 59 systems), followed by fungal pathogens (including <lb/>fungi and stramenopiles; n = 8) and animals (n = 5). Plant <lb/>hybrids in our database come from 20 families, primarily <lb/>Asteraceae (n = 16), Poaceae (n = 8) and Brassicaceae <lb/>(n = 6). <lb/> <note place="footnote"> © 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/></note> <note place="headnote">Review and Synthesis <lb/>The hybridisation-invasion hypothesis</note> <page>1467 <lb/></page> Table 1 <lb/> Experimental studies included in meta-analyses. <lb/> Hybrid Taxon <lb/> Parental Taxa <lb/> Field/ <lb/> Lab <lb/> Hybrid Classes § <lb/> Fecundity <lb/> Survival <lb/> Size <lb/> Reference <lb/> Plant taxa <lb/> Carpobrotus <lb/> hybrids ¶ <lb/> C. edulis E <lb/> 9 C. chilensis E <lb/> Field <lb/> Wild <lb/> W>P1>P2 † <lb/> P2=P1=W <lb/> Vila & D'Antonio 1998b <lb/> Field <lb/> Wild <lb/> W=P2=P1 <lb/> W=P1=P2 <lb/> Vila & D'Antonio 1998a <lb/> Field <lb/> Wild <lb/> P2 a <lb/> =P1 ab <lb/> =W b <lb/> Weber & D'Antonio 1999 <lb/> Helianthus annuus <lb/> texanus <lb/> H. a. annuus N <lb/> 9 H. debilis N <lb/> Field <lb/> Wild, BC <lb/> 1 <lb/> W>P1 a <lb/> =P2 ab <lb/> =G 2 <lb/> b <lb/> W=G <lb/> 2 <lb/> =P2>P1 <lb/> P1=W=G <lb/> 2 <lb/> >P2 † <lb/> Whitney et al. <lb/> 2006, <lb/> 2010b <lb/> Field <lb/> Wild, BC <lb/> 1 , BC <lb/> 7 <lb/> G <lb/> 7 <lb/> =W=P1>P2>G 2 W a <lb/> =P2 ab <lb/> =G 2 <lb/> ab <lb/> =P1 b <lb/> >G 7 W=P1>G <lb/> 7 <lb/> >G 2 <lb/> >P2 <lb/> Hovick et al. <lb/> unpub. <lb/> (see Appendix 6) <lb/> Field <lb/> Wild, BC <lb/> 1 , BC <lb/> 4 <lb/> W=P1=G <lb/> 4 <lb/> >G 2 <lb/> P1 a <lb/> =G 2 <lb/> a <lb/> =G 4 <lb/> ab <lb/> =W b <lb/> P1 a <lb/> =G 4 <lb/> ab <lb/> = G <lb/> 2 <lb/> b <lb/> =W b <lb/> Whitney et al. <lb/> unpub. <lb/> (see Appendix 6) <lb/> Lactuca <lb/> hybrids <lb/> L. serriola N <lb/> 9 L. sativa E <lb/> Field <lb/> BC <lb/> 1 , S <lb/> 1 <lb/> G <lb/> 2 <lb/> >P2=P1 <lb/> G <lb/> 2 <lb/> >P2>P1 <lb/> Hooftman et al. <lb/> 2005 <lb/> Field <lb/> F <lb/> 1 , S <lb/> 1-3 , BC <lb/> 1 S <lb/> 1 , BC <lb/> 1-3 <lb/> G <lb/> 1 <lb/> a <lb/> =G 2 <lb/> ab <lb/> =G 3 <lb/> bc <lb/> =P1 bc <lb/> =G 4 <lb/> c <lb/> >P2 <lb/> G <lb/> 1 <lb/> a <lb/> =G 2 <lb/> ab <lb/> =G 3 <lb/> b <lb/> =G 4 <lb/> b <lb/> =P1 b <lb/> >P2 <lb/> Hooftman et al. <lb/> 2007; <lb/> Field <lb/> BC <lb/> 1 S <lb/> 1 <lb/> P1=G <lb/> 3 <lb/> >P2 <lb/> G <lb/> 3 <lb/> =P1>P2 <lb/> P2=G <lb/> 3 <lb/> =P1 <lb/> Hartman et al. <lb/> 2013 <lb/> Myriophyllum <lb/> hybrids ¶ <lb/> M. spicatum E <lb/> 9 M. sibiricum E <lb/> Lab <lb/> Wild <lb/> W>P1 <lb/> LaRue et al. <lb/> 2013 <lb/> Raphanus <lb/> hybrids <lb/> R. raphanistrum E <lb/> 9 R. sativus E <lb/> Field <lb/> G <lb/> 3 <lb/> P1>G <lb/> 3 * <lb/> G <lb/> 3 <lb/> >P1* <lb/> Campbell & Snow 2007 <lb/> Field <lb/> G <lb/> 4 <lb/> G <lb/> 4 <lb/> >P1 <lb/> G <lb/> 4 <lb/> >P1 <lb/> G <lb/> 4 <lb/> >P1 <lb/> Campbell et al. <lb/> 2006; <lb/> Field <lb/> G <lb/> 4 <lb/> G <lb/> 4 <lb/> >P1* <lb/> P1=G <lb/> 4 * <lb/> Hovick et al. <lb/> 2012 <lb/> Field <lb/> Wild <lb/> W>P1>P2 <lb/> W>P1>P2 † <lb/> W=P2>P1 <lb/> Ridley & Ellstrand 2009 <lb/> Field <lb/> F <lb/> 1 <lb/> P1>G <lb/> 1 * <lb/> P 1 <lb/> >G 1 * , † <lb/> Snow et al. <lb/> 2001 <lb/> Field <lb/> F <lb/> 2 , F <lb/> 10 <lb/> P1=G <lb/> 2 ; G <lb/> 10 <lb/> =P1 <lb/> Snow et al. <lb/> 2010 ‡ <lb/> Reynoutria <lb/> 9 bohemica ¶ <lb/> R. japonica E <lb/> 9 R. sachalinensis E <lb/> Field <lb/> Wild <lb/> P1=W=P2* , † <lb/> P2=W=P1 <lb/> Py sek et al. <lb/> 2003 <lb/> Field <lb/> Wild <lb/> P1=W>P2* <lb/> Brabec & Py sek 2000 <lb/> Field <lb/> F <lb/> 1 , BC <lb/> 1 , F <lb/> 2 <lb/> G <lb/> 2 <lb/> =G 1 <lb/> =P1 <lb/> Gammon et al. <lb/> 2010 <lb/> Field <lb/> Wild <lb/> W=P1=P2* , † <lb/> W>P2=P1* <lb/> Parepa et al. <lb/> 2014 <lb/> Lab <lb/> Wild <lb/> W=P1* <lb/> Richards et al. <lb/> 2008 <lb/> Lab <lb/> Wild <lb/> P1=W* <lb/> Rouifed et al. <lb/> 2011 <lb/> Senecio vulgaris <lb/> var. hibernicus <lb/> S. squalidus E <lb/> 9 <lb/> S. vulgaris <lb/> var. vulgaris N <lb/> Field <lb/> Wild <lb/> P1=W=P2 <lb/> W=P1=P2* <lb/> P2=W=P1 <lb/> Hawkes et al. <lb/> 2010 <lb/> Senecio squalidus <lb/> S. aethnensis E <lb/> 9 <lb/> S. chrysanthemumifolius E <lb/> Lab <lb/> Wild <lb/> P2=W>P1 <lb/> P2 a <lb/> =W ab <lb/> =P1 b <lb/> Brennan et al. <lb/> 2012 <lb/> Sorghum almum <lb/> S. halepense E <lb/> 9 S. bicolor E <lb/> Field <lb/> F1 <lb/> G <lb/> 1 <lb/> =P1 <lb/> G <lb/> 1 <lb/> =P1* <lb/> Arriola & Ellstrand 1997 <lb/> Spartina <lb/> hybrids ¶ <lb/> S. alterniflora E <lb/> 9 S. foliosa N <lb/> Lab <lb/> Wild <lb/> W=P1>P2 <lb/> Ayres et al. <lb/> 2004 <lb/> Typha <lb/> 9 glauca ¶ <lb/> T. angustifolia E <lb/> 9 T. latifolia N <lb/> Field <lb/> Wild <lb/> W>P2=P1 <lb/> Sullivan et al. <lb/> 2010 <lb/> (continued) <lb/> <note place="footnote"> © 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/></note> <page>1468</page> <note place="headnote">S. M. Hovick and K. D. Whitney <lb/>Review and Synthesis <lb/></note> Experimental tests of the H-I hypothesis <lb/> From our full list of 72 hybrids, we found 18 hybrid taxa for <lb/>which experimental hybrid-parent performance comparisons <lb/>utilizing the relevant parent(s) have been conducted (reported <lb/>in 40 studies and two unpublished datasets), either in the field <lb/>(24 studies plus our unpublished datasets) or in the lab or <lb/>greenhouse (16 studies). Only 14 of the 59 plant hybrids in our <lb/>dataset (23.7%) have been assessed experimentally in such a <lb/>way as to test the H-I hypothesis (Appendix 2). Nine of the 14 <lb/>hybrid systems have been assessed in the field, with the remain-<lb/>der tested only in the lab or greenhouse. These experiments <lb/>compare hybrid performance to that of the relevant parental <lb/>taxa, either both parents (e.g. Vila & D'Antonio 1998b; Hooft-<lb/>man et al. 2005; Sullivan et al. 2010) or, if one parent is clearly <lb/>more invasive or a superior competitor (i.e. if the inferior par-<lb/>ent has not become naturalised despite opportunity), the supe-<lb/>rior parent (e.g. Hovick et al. 2012). For 11 of the remaining <lb/>45 taxa that have not been adequately tested, observational or <lb/>incomplete experimental data have been reported that are sug-<lb/>gestive of hybrid performance gains; these include some studies <lb/>comparing hybrids with their inferior competitor parent only <lb/>(e.g. Grosholz 2010) and some studies comparing hybrids with <lb/>only one parent, either because the second is not clearly infe-<lb/>rior (Van Grunsven et al. 2009), or because the second parent <lb/>is not yet known (Henery et al. 2010; Hahn et al. 2012). For <lb/>the remaining 34 of these 45 plant hybrids, hybridisation has <lb/>been confirmed and also suggested as the underlying driver of <lb/>weediness or invasiveness, but we have found no data on which <lb/>to judge the latter claim. <lb/>For fungal pathogens, two hybrids have been assessed suffi-<lb/>ciently to yield experimental tests of the H-I hypothesis: the <lb/>rust Melampsora x columbiana (Newcombe et al. 2000, 2001) <lb/>and a Pythium hybrid found on European Phragmites austral-<lb/>is (Nechwatal & Mendgen 2009). All such investigations have <lb/>been conducted in the lab, thus the relative performance of <lb/>hybrid pathogens in the field is not known. <lb/>Of the five invasive animal hybrids in our dataset, two sys-<lb/>tems have been investigated experimentally. Ambystoma <lb/> hybrids have been assessed both in field (Ryan et al. 2009) <lb/>and lab tests (Johnson et al. 2010) and Cyprinella hybrids <lb/>have been assessed in the lab only (Blum et al. 2010). <lb/>For our meta-analyses, we excluded eight of the 42 studies <lb/>reporting experimental data, either because they failed to <lb/>report some measure of variability or because they only <lb/>reported data (percent germination, photosynthesis-related <lb/>traits and walking speed) not fitting with our three perfor-<lb/>mance categories; this exclusion eliminated three hybrid plant <lb/>taxa from subsequent analyses (Sarcocornia hybrids, Sphag-<lb/>neticola hybrids and Viola x tatrae), cutting our final list to 15 <lb/>hybrid taxa (11 plant, two animal and two fungal hybrids; see <lb/>Table 1 and Appendix 4). <lb/> Meta-analysis: does hybridisation lead to hybrid performance <lb/>gains? <lb/> Many of the 34 studies included in our meta-analyses used <lb/>multiple hybrid classes, reported data from multiple unique <lb/>experiments or established experiments across multiple loca-<lb/> Table 1. (continued) <lb/> Hybrid Taxon <lb/> Parental Taxa <lb/> Field/ <lb/> Lab <lb/> Hybrid Classes § <lb/> Fecundity <lb/> Survival <lb/> Size <lb/> Reference <lb/> Fungal taxa <lb/> Melampsora <lb/> 9 <lb/> columbiana <lb/> M. medusae N <lb/> 9 M. occidentalis N <lb/> Lab <lb/> Wild <lb/> P2>W>P1 <lb/> Newcombe et al. <lb/> 2000 <lb/> Lab <lb/> Wild <lb/> W=P1>P2 <lb/> Newcombe et al. <lb/> 2001 <lb/> Pythium <lb/> hybrids <lb/> P. arrhenomanes E <lb/> (or similar) <lb/> 9 P. phragmitis N <lb/> Lab <lb/> Wild <lb/> W=P1>P2* <lb/> Nechwatal & <lb/> Mendgen 2009 <lb/> Animal taxa <lb/> Ambystoma <lb/> hybrids <lb/> A. tigrinum mavortium E <lb/> 9 A. californiense N <lb/> Field <lb/> F <lb/> 1 , F <lb/> 2 , BC <lb/> 1 <lb/> P2=P1=G <lb/> 1 <lb/> =G 2 <lb/> P1=G <lb/> 1 <lb/> =G 2 <lb/> >P2 <lb/> Ryan et al. <lb/> 2009 <lb/> Lab <lb/> Wild, F <lb/> 1 , F <lb/> 2 , BC <lb/> 1 <lb/> P1>G <lb/> 1 <lb/> =G 2 <lb/> >P2>W <lb/> P 1 <lb/> >G 2 <lb/> a <lb/> =W ab <lb/> =G 1 <lb/> b <lb/> >P2 <lb/> Johnson et al. <lb/> 2010 <lb/> Cyprinella <lb/> hybrids <lb/> C. lutrensis E <lb/> 9 C. venusta N <lb/> Lab <lb/> F <lb/> 1 <lb/> G <lb/> 1 <lb/> >P1=P2* <lb/> Blum et al. <lb/> 2010 <lb/> *Significant differences as reported in references; otherwise, significance reflects means separated by <lb/> >2 SEM, based on our aggregated data. <lb/> †Excluded from meta-analysis because variability not reported. <lb/> ‡G 2 and G <lb/> 10 hybrids assessed in separate experiments. <lb/> §All hybrid classes used (S <lb/> x <lb/> = selfed F <lb/> x s; BC <lb/> x <lb/> = backcrossed F <lb/> x s; BC <lb/> 1 S <lb/> 1 <lb/> = selfed BC <lb/> 1 s). <lb/> ¶Capable of clonal growth (Flora of North America Editorial Committee 1993+). <lb/> For each entry, significant differences are indicated by inequality signs and superscript letters, where necessary (taxa sharing the same letter do not differ). Differences are from referenced studies <lb/> or our data aggregations (see footnote) and were not used in meta-analysis. Parental taxa are denoted P1 and P2, reflecting their order in the 'Parental Taxa' column; P1 is the more invasive/ <lb/> higher performing parent used for meta-analyses. Parents are denoted as native (N) or exotic (E), relative to the region where hybrids are invasive. <lb/> <note place="footnote"> © 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/></note> <note place="headnote">Review and Synthesis <lb/>The hybridisation-invasion hypothesis</note> <page>1469 <lb/></page> tions. Thus, we calculated a total of 47 fecundity-based, 52 <lb/>size-based and 42 survival-based effect sizes from experimental <lb/>hybrid-parent performance comparisons. In all analyses, <lb/>effect-size heterogeneity (I 2 ) among taxa was substantial, <lb/>which likely contributed to non-significant differences in het-<lb/>erogeneity tests comparing F 1 , post-F 1 and wild hybrids <lb/>(fecundity: Q = 3.32, P = 0.190, I 2 = 86.1%; survival: <lb/>Q = 1.18, P = 0.553, I 2 = 93.3%; size: Q = 2.54, P = 0.280, <lb/>I 2 = 96.2%; all d.f.=2). We thus conducted analyses on all <lb/>hybrid classes combined; however, because of an a priori <lb/> interest in relative performance by each group and because <lb/>these categories represent hybrid populations with distinct his-<lb/>tories of genotypic filtering and selective pressures, we also <lb/>conducted analyses separately for each hybrid class. This <lb/>approach is also supported by (less conservative) fixed-effects <lb/>analyses, which found significant differences between hybrid <lb/>classes (results not shown). We found no indication of publi-<lb/>cation bias based on funnel plots and rank correlation tests <lb/>for funnel plot asymmetry (see Appendix 5). <lb/> Do hybrids out-reproduce their parental taxa? <lb/> Meta-analyses of fecundity indicated that naturally occurring <lb/>wild hybrids significantly outperformed their 'more invasive' <lb/>parental taxon (Z = 2.79, P = 0.005, n = 6, I 2 = 72.1%; see <lb/>'Overall M-A results' in Fig. 1a). Within individual taxa, wild <lb/>hybrids were uniformly more fecund than their more invasive <lb/>parents, with the effect reaching statistical significance in Raph-<lb/>anus and Senecio squalidus (and nearly reaching statistical sig-<lb/>nificance in Helianthus, with a 95% confidence interval of <lb/>[À0.034, 0.281], Fig. 1a). For all hybrid classes considered <lb/>together, hybrids were also significantly more fecund than their <lb/>more invasive parent (Z = 2.04, P = 0.042, n = 9, I 2 = 64.1%; <lb/>Fig. 1a). However, resynthesised F 1 and later-generation (post-<lb/>F 1 ) hybrids did not have significantly enhanced fecundity (F 1 : <lb/>Z = À0.73, P = 0.468, n = 4, I 2 = 92.7%; post-F 1 : Z = 0.72, <lb/> P = 0.470, n = 4, I 2 = 85.2%; Fig. 1a). <lb/>Our meta-regression of fecundity responses indicated a mar-<lb/>ginally significant generation effect across all three taxa, with <lb/>relative hybrid fecundity increasing over subsequent genera-<lb/>tions (b gen AE SE = 0.106 AE 0.058, Z = 1.83, P = 0.067, <lb/>I 2 = 95.6%), although this relationship varied across taxa <lb/>(Fig. 2a). When taxa were assessed separately with linear <lb/>models, the generation effect was marginally significant and <lb/>positive for Helianthus, not significant for Raphanus, and sig-<lb/>nificantly negative for Lactuca (Helianthus: b gen = <lb/> 0.218 AE 0.117, Z = 1.87, P = 0.062, I 2 = 93.2%; Raphanus: <lb/> b gen = 0.132 AE 0.108, Z = 1.223, P = 0.221, I 2 = 93.7%; <lb/> Lactuca: b gen = À0.067 AE 0.032, Z = À2.13, P = 0.033, <lb/>I 2 = 74.6%). Despite apparent nonlinearity in the scatterplot <lb/>(Fig. 2a), a polynomial relationship was not supported for all <lb/>three taxa combined (polynomial model P = 0.101, b gen <lb/> P = 0.133 and b gen <lb/>2 P = 0.286). In contrast, a polynomial <lb/>model did fit the combined Helianthus and Raphanus data well <lb/>(polynomial model P = 0.004; b gen =0.688 AE 0.259, Z = 2.66, <lb/> P = 0.008; b gen <lb/>2 =À0.049 AE 0.023, Z = À2.162, P = 0.031, <lb/>I 2 = 93.9%), indicating that hybrid relative fecundity for these <lb/>two taxa increased rapidly over the first few generations <lb/>before reaching a point where hybrids outperformed their par-<lb/>ents (Fig. 2a). <lb/> Do hybrids outlive their parental taxa? <lb/> Meta-analyses of survival indicated that none of our three <lb/>hybrid classes differed from their parental taxa (wild: <lb/>Z = 0.36, P = 0.718, n = 4, I 2 = 81.9%; F 1 : Z = 0.76, <lb/> P = 0.446, n = 3, I 2 = 92.0%; post-F 1 : Z = À0.12, P = 0.901, <lb/> n = 4, I 2 = 96.2%; Fig. 1b), but we note that small sample <lb/>sizes for these analyses limit our power to detect any true <lb/>effects. The combination of all hybrid classes gave the same <lb/>non-significant <lb/>result <lb/>(Z = 0.79, <lb/> P = 0.429, <lb/> n = 8, <lb/>I 2 = 93.9%). Unlike our other performance metrics, relative <lb/>survival of hybrids tended to vary more among taxa than <lb/>among hybrid classes within taxa. Regardless of hybrid class, <lb/>all Lactuca, Raphanus, Senecio vulgaris and Cyprinella hybrids <lb/>had significantly higher survival than their parents, whereas <lb/>all Reynoutria and Ambystoma hybrids had significantly lower <lb/>survival (Fig. 1b). <lb/>Across the three systems for which data are available, rela-<lb/>tive hybrid survival significantly decreased across generations <lb/>(b gen = À0.122 AE 0.048, Z = À2.55, P = 0.011; I 2 = 93.3%; <lb/>Fig. 2b). This pattern was similar for Helianthus and Lactuca <lb/> when analysed separately (Helianthus: b gen = À0.097 AE 0.022, <lb/>Z = À4.41, P < 0.001, I 2 = 0.0%; Lactuca: b gen = <lb/> À0.077 AE 0.042, Z = À1.82, P = 0.068, I 2 = 86.8%; note that <lb/> Ambystoma had too few data points to be assessed sepa-<lb/>rately). <lb/> Are hybrids larger than their parental taxa? <lb/> In addition to being more fecund, wild hybrids were also sig-<lb/>nificantly larger than their more invasive parent (Z = 2.00, <lb/> P = 0.046, n = 9, I 2 = 89.4%; Fig. 1c). In contrast to the <lb/>fecundity data, effect size variation among wild hybrid taxa <lb/>for organism size was much greater (89.4 vs. 72.1%), with a <lb/>broader range of effect size estimates (Fig. 1c) and two influ-<lb/>ential outlier genera with large wild hybrids relative to their <lb/>parents (the plants Typha and Myriophyllum; Cook's <lb/>D > 0.34 and Dffits > 0.65 for both). Because relative size <lb/>advantages may be particularly relevant for species that per-<lb/>sist and spread by clonal growth, we re-analysed the wild <lb/>hybrid size data by clonal growth category (present vs. absent; <lb/>see Table 1). Although effect sizes for these two groups were <lb/>not significantly different from each other (Q = 2.36, d.f.=1, <lb/> P = 0.124), only wild hybrids capable of clonal growth were <lb/>larger than their parental taxa (clonal: Z = 2.18, P = 0.029, <lb/> n = 5, I 2 = 88.0%; non-clonal: Z = 0.55, P = 0.581, n = 4, <lb/>I 2 = 90.8%; Fig. 1c). At the level of individual clonal taxa, <lb/>hybrids were significantly larger than parents in Myriophyl-<lb/>lum, Reynoutria and Typha but not in Carpobrotus or Sparti-<lb/>na. For resynthesised F 1 s and post-F 1 s, hybrids were smaller <lb/>than their parental taxa on average, but with large confidence <lb/>intervals that overlapped zero (F 1 : Z = À0.99, P = 0.325, <lb/> n = 2, I 2 = 97.8%; post-F 1 : Z = À0.77, P = 0.444, n = 4, <lb/>I 2 = 98.5%). When all hybrid classes were considered <lb/>together, hybrid and parental size did not differ (Z = 0.50, <lb/> P = 0.620, n = 12, I 2 = 97.1%). <lb/>Among the three resynthesised hybrid taxa that have been <lb/>assessed, relative hybrid size did not vary across generations <lb/>(b gen = À0.023 AE 0.078, Z = À0.30, P = 0.768, I 2 = 98.1%; <lb/>Fig. 2c). <lb/> <note place="footnote">© 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/></note> <page>1470</page> <note place="headnote">S. M. Hovick and K. D. Whitney <lb/>Review and Synthesis <lb/></note> DISCUSSION <lb/> We found moderate support for the hypothesised connection <lb/>between interspecific hybridisation and increased invasiveness. <lb/>Quantitative data from experimental performance compari-<lb/>sons that could be used to test the H-I hypothesis were avail-<lb/>able for roughly 21% of all invasive or weedy hybrid taxa in <lb/>our dataset. Not surprisingly, among-study heterogeneity was <lb/> Effect Size (Hedges' d) <lb/> Overall <lb/> M-A <lb/>results <lb/> all <lb/>F 1 <lb/> post-F 1 <lb/> Lactuca <lb/> Ambystoma <lb/> all <lb/>F 1 <lb/> post-F 1 <lb/> Cyprinella (F 1 ) <lb/> Reynoutria (wild) <lb/> * <lb/> * <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/> all <lb/>post-F 1 <lb/> wild <lb/> Helianthus <lb/>Carpobrotus (wild) <lb/> * <lb/> All Hybrids (n = 8) <lb/>F 1 <lb/> (n = 3) <lb/>Post-F 1 <lb/> (n = 4) <lb/>Wild <lb/>(n = 4) <lb/> Raphanus (post-F 1 ) <lb/> * <lb/>* <lb/> Senecio vulgaris (wild) <lb/> (c) Size <lb/> –2.5 –2.0 –1.5 –1.0 –0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 <lb/>–2.5 –2.0 –1.5 –1.0 –0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 <lb/>–2.5 –2.0 –1.5 –1.0 –0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 <lb/> all <lb/>post-F 1 <lb/> wild <lb/> Helianthus <lb/> all <lb/>F 1 <lb/> post-F 1 <lb/> all <lb/>F 1 <lb/> post-F 1 <lb/> wild <lb/> Lactuca <lb/>Raphanus <lb/> all <lb/>F 1 <lb/> post-F 1 <lb/> Reynoutria <lb/>Senecio vulgaris (wild) <lb/> Overall <lb/>M-A <lb/>results <lb/> Senecio squalidus (wild) <lb/> Sorghum (F 1 ) <lb/> Melampsora (wild) <lb/> Pythium (wild) <lb/> All Hybrids (n = 9) <lb/>F 1 <lb/> (n = 4) <lb/>Post-F 1 <lb/> (n = 4) <lb/>Wild <lb/>(n = 6) * <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/> Overall <lb/>M-A <lb/>results <lb/> all <lb/>post-F 1 <lb/> wild <lb/> Helianthus <lb/> all <lb/>post-F 1 <lb/> wild <lb/> Raphanus <lb/>Ambystoma <lb/> † Reynoutria (wild) <lb/> Senecio vulgaris (wild) <lb/> Sorghum (F 1 ) <lb/> † Spartina (wild) <lb/> † Typha (wild) <lb/>all <lb/>F 1 <lb/> post-F 1 <lb/> Lactuca (post-F 1 ) <lb/> † Myriophyllum (wild) <lb/> † Carpobrotus (wild) <lb/> Senecio squalidus (wild) <lb/> * <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/>* <lb/> All Hybrids (n = 12) <lb/>F 1 <lb/> (n = 2) <lb/>Post-F 1 <lb/> (n = 4) <lb/>Wild <lb/>(n = 9) <lb/> † Wild clonal, n = 5 <lb/>Wild nonclonal, n = 4 <lb/> * <lb/> (a) Fecundity <lb/> (b) Survival <lb/> Figure 1 Meta-analysis results, showing effect sizes and 95% confidence intervals for hybrid vs. parental performance for (a) fecundity, (b) survival and (c) <lb/>size. Effects greater than zero (dotted vertical line) indicate greater hybrid performance relative to their 'more invasive' parental taxon. Confidence intervals <lb/>excluding zero are noted with an asterisk along the left side of the panel. Overall effect sizes shown in the non-shaded region at the top of each panel are <lb/>from random effect meta-analyses conducted separately for (1) all hybrid classes combined, (2) resynthesised F 1 hybrids, (3) resynthesised post-F 1 hybrids <lb/>and (4) wild hybrids. In panel C, separate overall effect sizes are also shown for wild hybrids that are capable vs. incapable of clonal growth. Individual <lb/>effect sizes are shown beneath the solid horizontal line, with taxa separated by alternate shading and symbols indicating different hybrid classes. <lb/> <note place="footnote"> © 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/></note> <note place="headnote">Review and Synthesis <lb/>The hybridisation-invasion hypothesis</note> <page>1471 <lb/></page> high in all cases, reflecting species-specific responses to hy-<lb/>bridisation in addition to likely variation in the design of <lb/>experiments and performance assessments. The strongest evi-<lb/>dence supporting the H-I hypothesis was found in fecundity-<lb/>based performance assessments involving wild (as opposed to <lb/>resynthesised) hybrids; evidence was strongest in plants but <lb/>weaker in the few animal and fungal taxa for which data were <lb/>available. Although fecundity relative to parents was on <lb/>average low in resynthesised hybrids, it increased significantly <lb/>over time in two of the three systems for which we could <lb/>assess the pattern; in these systems, resynthesised hybrids out-<lb/>performed their parents by approximately the fifth generation <lb/>post-hybridisation. Hybrid survival was more variable among <lb/>taxa, and although no overall effect was seen in the meta-<lb/>analysis, hybrid survival relative to parents tended to be high <lb/>for plants and inconsistent between the two animal systems <lb/>evaluated (note that tests in animal systems have only focused <lb/>on resynthesised, instead of wild, hybrids). For resynthesised <lb/>hybrids, there was an overall pattern of decreasing relative <lb/>survival over time post-hybridisation. Finally, with respect to <lb/>size, the meta-analysis revealed that wild hybrids were larger <lb/>than their parental taxa, a pattern that was particularly strong <lb/>for taxa in which hybrids are capable of clonal growth. <lb/> Designing experiments to test the H-I hypothesis: two basic criteria <lb/> One of the clearest patterns from our review is that many stud-<lb/>ies appearing to test the H-I hypothesis have not adequately <lb/>done so. At a minimum, we argue that an informative test <lb/>must meet two basic criteria. First, it must compare the perfor-<lb/>mance of hybrids to that of the most invasive parent to ensure <lb/>that hybrid performance is assessed relative to the correct <lb/>benchmark. Where two introduced or two native species have <lb/>hybridised, typically both parents must be included in the <lb/>assessment to determine whether hybridisation has enhanced <lb/>invasiveness. Where hybrids result from crossing one native <lb/>taxon with one introduced and invasive taxon, the introduced <lb/>parent must be included in any assessment for meaningful <lb/>comparisons. If the most relevant parent is omitted, then infer-<lb/>ences regarding the H-I hypothesis are impossible. For exam-<lb/>ple, California Spartina is often cited as a case where <lb/>hybridisation is associated with invasiveness (e.g. Ayres et al. <lb/> 2008; Grosholz 2010). However, we found no experimental <lb/>tests comparing fecundity or survival of the hybrid to the inva-<lb/>sive parent, and a test of size found no difference between the <lb/>taxa (Table 1). A test of resistance to goose herbivory found <lb/>decreased palatability in the hybrid relative to the native par-<lb/>ent S. foliosa (Grosholz 2010). This pattern was interpreted as <lb/>support for the H-I hypothesis, via the suggestion that hybrids <lb/>'may pose a greater risk to natural systems than the parent <lb/>species.' Again, however, no comparison was made to the <lb/>invasive parent S. alterniflora. It could be that the invasive <lb/> 10 <lb/> 5 <lb/>0 <lb/> Effect Size (Hedges' d) <lb/> –3.0 <lb/>–2.5 <lb/>–2.0 <lb/>–1.5 <lb/>–1.0 <lb/>–0.5 <lb/>0.0 <lb/>0.5 <lb/>1.0 <lb/> Helianthus <lb/> Lactuca <lb/>Raphanus <lb/> Linear fit (all 3 spp.) <lb/>Polynomial fit <lb/>(Helianthus & Raphanus only) <lb/> Wild † <lb/> Effect Size (Hedges' d) <lb/> –4 <lb/>–3 <lb/>–2 <lb/>–1 <lb/>0 <lb/>1 <lb/> Helianthus <lb/>Lactuca <lb/>Ambystoma <lb/> Post-hybridization generation <lb/>Effect Size (Hedges' d) <lb/> –3 <lb/>–2 <lb/>–1 <lb/>0 <lb/>1 <lb/>2 <lb/> Helianthus <lb/>Raphanus <lb/>Ambystoma <lb/> (c) Size <lb/> (a) Fecundity <lb/> (b) Survival <lb/> 10 <lb/>5 <lb/>0 <lb/> Wild † <lb/> 10 <lb/>5 <lb/>0 <lb/> Wild † <lb/> Figure 2 Meta-regression results, showing effect sizes (hybrid vs. parental <lb/>performance) by hybrid generation for (a) fecundity, (b) survival and (c) <lb/>size. Effects greater than zero (above the dotted horizontal line) indicate <lb/>greater hybrid performance relative to their 'more invasive' parental taxa. <lb/>Error bars indicate 95% confidence intervals. The best-fit lines in panels <lb/>A and B show significant relationships between hybrid generation and <lb/>Hedges' d for fecundity and survival based on meta-regression parameter <lb/>estimates; the meta-regression for size in panel C was not statistically <lb/>significant. In addition to the linear best-fit line, panel A also depicts the <lb/>nonlinear relationship between hybrid generation and relative fecundity <lb/>supported for Helianthus and Raphanus (considered apart from Lactuca; <lb/> see Results). Note that effect sizes for naturally occurring wild hybrids <lb/>were not included in the analysis, but are shown in the figure for <lb/>comparison. <lb/> <note place="footnote"> © 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/></note> <page>1472</page> <note place="headnote">S. M. Hovick and K. D. Whitney <lb/>Review and Synthesis <lb/></note> parent has equal (or even lower) palatability relative to the <lb/>hybrid, in which case hybridisation could have nothing to do <lb/>with performance or invasiveness in this system. <lb/>In some cases, accurate identification of both parental taxa <lb/>is a barrier to our first criterion for testing the H-I hypothesis. <lb/>For example, a tetraploid cytotype of spotted knapweed <lb/>(Centaurea stoebe s.l.) has been recorded as invasive in North <lb/>America, with molecular analyses pointing to an allopolyploid <lb/>origin between the diploid Centaurea stoebe s. str. and an <lb/>unknown parent (Mraz et al. 2012). Robust common garden <lb/>experiments have been conducted in this system (Henery et al. <lb/> 2010; Hahn et al. 2012), but without knowing the second par-<lb/>ent's identity inferences regarding the H-I hypothesis are <lb/>moot. Similar challenges exist in other systems (e.g. Kim et al. <lb/> 2008; Inderbitzin et al. 2011), emphasising the continued <lb/>importance of phylogenetic and taxonomic research in puta-<lb/>tive cases of hybridisation-induced invasiveness. <lb/>We suggest that a second criterion for testing the H-I hypo-<lb/>thesis is that performance assessments be conducted experimen-<lb/>tally in a common environment, as observations of hybrid <lb/>performance or abundance in naturally occurring populations <lb/>are not unambiguously interpretable. This criterion is not satis-<lb/>fied by a handful of key studies often cited in support of the <lb/>H-I hypothesis that document increased abundances of hybrids <lb/>relative to their parental taxa (Urbanska et al. 1997; Moody & <lb/>Les 2002). Abundant hybrids could indeed result from <lb/>enhanced invasiveness, but the same patterns could also result <lb/>from chance dispersal events or simply from the introgression <lb/>of neutral alleles from one parental genome into the other. <lb/> Fitness components vs. population growth rates as measures of <lb/>invasiveness <lb/> In our meta-analyses, the degree to which the H-I hypothesis <lb/>was supported varied with the fitness component (fecundity, <lb/>survival or size) examined. In some cases, this pattern may <lb/>reflect expected life-history tradeoffs: for example, resyn-<lb/>thesised Helianthus hybrids displayed an increase in relative <lb/>fecundity but a decrease in relative survival over time post-<lb/>hybridisation (Fig. 2a,b). This issue highlights the importance <lb/>of knowing how individual fitness components are related to <lb/>the population growth rate (k), which determines invasiveness. <lb/>While fitness components are expected to generally be posi-<lb/>tively correlated with k, the details matter. For instance, if <lb/>populations are establishment (microsite) limited rather than <lb/>propagule limited (Poulsen et al. 2007), then beyond a certain <lb/>point, increased fecundity will not necessarily result in <lb/>increased k. <lb/> Female fecundity alone could be misleading in tests of the <lb/>H-I hypothesis if male fitness (i.e. pollen or sperm viability) is <lb/>severely depressed in hybrids relative to their parents. Because <lb/>we only found pollen viability data for four taxa representing a <lb/>variety of hybrid classes (Sorghum, Reynoutria, Raphanus and <lb/> Senecio squalidus), a meaningful meta-analysis was not possi-<lb/>ble. However, for resynthesised Raphanus hybrids the change <lb/>in pollen viability over time is similar to the changes in Raph-<lb/>anus fecundity we identified using meta-regression; that is, pol-<lb/>len viability is depressed in the F 1 generation but then recovers <lb/>by approximately generation 3–6, eventually reaching values <lb/>equal to or greater than that of its parental taxon (Snow et al. <lb/> 2010). It remains unknown whether male fitness responds simi-<lb/>larly in other invasive hybrids and to what extent such patterns <lb/>might influence population growth rates. <lb/>Four studies (of two hybrid systems) included in our meta-<lb/>analyses do report k for hybrids vs. parent populations (Hooft-<lb/>man et al. 2005, 2007; Campbell et al. 2006; Hartman et al. <lb/> 2013). In Raphanus, k did not differ between paired field popu-<lb/>lations of resynthesised hybrids and their better performing <lb/>parental taxon over the second through fourth years post-hy-<lb/>bridisation in Michigan (Campbell et al. 2006). From the per-<lb/>spective of testing the H-I hypothesis, this finding concurs with <lb/>a common garden experiment in the same region where fourth-<lb/>generation hybrids did not outperform their parental taxon <lb/>(hybrids were significantly less fecund and had similar survival <lb/>to the parent; Campbell et al. 2006). In Lactuca, k was esti-<lb/>mated from experimental populations of hybrids and both <lb/>parental taxa by multiplying germination and survival rates by <lb/>seed output plant À1 (Hooftman et al. 2005, 2007; Hartman <lb/> et al. 2013); in all cases, differences in k between taxa reflected <lb/>differences in survival but not in seed output, suggesting that <lb/>for Lactuca, survival is a better indicator of k than is fecundity. <lb/>Unfortunately, without estimates of k and individual fitness <lb/>components from other systems, questions remain about the <lb/>degree to which these alternative performance metrics may <lb/>indicate differences in population growth rates. Overall, future <lb/>tests of the H-I hypothesis for sexually-reproducing species <lb/>would be better served if k (rather than fitness components <lb/>alone) were estimated as the measure of invasiveness. <lb/>For some taxa, size (or vegetative growth rate) may be the <lb/>most relevant measure of invasiveness. Many invasive plants <lb/>reproduce primarily via clonal spread (Py sek & Richardson <lb/>2007); and indeed clonality has been identified as an impor-<lb/>tant mechanism by which heterosis achieved in the F 1 (which <lb/>is normally broken down by recombination in sexual species) <lb/>can be stabilised in hybrid lineages, potentially increasing <lb/>invasiveness (Ellstrand & Schierenbeck 2000). Invasive popu-<lb/>lations of hybrid Myriophyllum, Reynoutria and Typha can <lb/>reproduce clonally, have been shown to be largely composed <lb/>of F 1 individuals (Kirk et al. 2011; Bailey 2013; LaRue et al. <lb/> 2013), and in our analyses were all significantly larger than <lb/>their parents (Fig. 1c). For taxa such as these, size may be a <lb/>relevant measure of invasive potential; thus a pattern in which <lb/>hybrids are larger than parents can be interpreted as support <lb/>for the H-I hypothesis. Conversely, although Carpobrotus and <lb/> Spartina hybrids also exhibit clonal growth, sexual reproduc-<lb/>tion (and thus recombination) appears to be common in these <lb/>systems (Gallagher et al. 1997; Ayres et al. 2008), and hybrids <lb/>were not larger than their parental taxa (Fig. 1c). The con-<lb/>trasting patterns expressed by these two groups of hybrid taxa <lb/>illustrate the critical importance of measuring performance in <lb/>a way that captures invasive potential appropriately for each <lb/>particular system. <lb/> How does within-taxon variation in relative hybrid performance <lb/>affect our understanding of the H-I hypothesis? <lb/> Hybrid taxa are often highly variable, and our review high-<lb/>lights three key aspects of this variability that have direct <lb/> <note place="footnote">© 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/> </note> <note place="headnote">Review and Synthesis <lb/>The hybridisation-invasion hypothesis</note> <page>1473 <lb/></page> bearing on how the H-I hypothesis is conceived, interpreted <lb/>and tested. These non-mutually exclusive sources of variation <lb/>can be distinguished as habitat effects, generation effects and <lb/>lineage effects. <lb/> Habitat effects: genotype by environment (G 9 E) interactions <lb/>influence hybrid invasiveness <lb/> Habitat-specific performance advantage can lead to a compli-<lb/>cated mosaic of peaks in hybrid relative performance. For <lb/>example, Carpobrotus hybrids had higher relative growth rates <lb/>than both of their parental taxa in backdune (but not bluff <lb/>scrub) habitats, yet hybrid survival in response to herbivory <lb/>was greatest in the bluff scrub (Vila & D'Antonio 1998b). At <lb/>a larger geographic scale (North America), Raphanus hybrids <lb/>have been shown to outperform their weedy parental taxa in <lb/>California, where they are currently invasive (Campbell et al. <lb/> 2006; Ridley & Ellstrand 2009), and in Texas, where they do <lb/>not yet occur naturally (Hovick et al. 2012), but not in Michi-<lb/>gan (Campbell et al. 2006). Although hybrids were tradition-<lb/>ally thought to occur only in intermediate hybrid zones <lb/>(Anderson 1948; Barton & Hewitt 1985), mounting evidence <lb/>suggests that genome shuffling following hybridisation can <lb/>result in phenotypes that are able to thrive in non-intermedi-<lb/>ate and even non-parental habitats (Rieseberg et al. 1999, <lb/>2007). Site-specific variation in hybrid relative performance <lb/>will make it more difficult to reject the null hypothesis of no <lb/>difference between hybrids and their parents if some common <lb/>(but not ubiquitous) combination of biotic and abiotic factors <lb/>exists that favours hybrid invaders. These considerations sug-<lb/>gest there may be great value in assessing performance across <lb/>a range of conditions. <lb/> Generation effects: not all hybrid generations are created equal <lb/> In most cases, assessments of hybrid relative performance <lb/>depend greatly on which hybrid class(es) are investigated. Our <lb/>meta-analyses indicate that wild (presumably later-generation) <lb/>hybrids are more likely to outperform their parental taxa than <lb/>are resynthesised (mostly early-generation) hybrids, on aver-<lb/>age. This pattern conforms to long-held expectations that <lb/>hybrid performance should increase in post-F 1 hybrid classes <lb/>as poorly adapted genotypes are filtered from the population <lb/>by natural selection (reviewed by Arnold & Hodges 1995). <lb/>Similarly, based on our meta-regressions, relative fecundity <lb/>increased over subsequent generations post-hybridisation <lb/>(Fig. 2a), and in many cases, the earliest-generation (F 1 ) <lb/>hybrids performed worse than their parents and all other <lb/>hybrid classes (Fig. 1a–c). Hybrid survival also varied by gen-<lb/>eration but in the opposite direction, decreasing over time <lb/>post-hybridisation (Fig. 2b). This pattern may reflect a sur-<lb/>vival-fecundity life history tradeoff (cf. Helianthus from <lb/>Fig. 2a and b). <lb/>These considerations mean that the choice of hybrid class <lb/>used in testing the H-I hypothesis is a crucial one. On one <lb/>hand, resynthesising hybrid lineages for testing the H-I <lb/>hypothesis is the cleanest way to isolate the effects of hy-<lb/>bridisation from other evolutionary factors such as popula-<lb/>tion admixture and bottlenecks. On the other, tests using <lb/>early-generation hybrid populations will often include poorly <lb/>adapted genotypes (e.g. Snow et al. 2010) and thus may give <lb/>misleading answers about whether hybrids are capable of <lb/>outperforming the parental species. Early-generation hybrid <lb/>populations often have low average fitness but with high <lb/>variance and a small number of well-adapted individuals <lb/>(e.g. Whitney et al. 2006). The composition of these popula-<lb/>tions can therefore change substantially in only a few gener-<lb/>ations, making tests with only early-generation hybrids <lb/>suspect. These problems are avoided if resynthesised lineages <lb/>are first allowed to evolve in natural conditions post-hybridi-<lb/>sation (Campbell et al. 2006; Whitney et al. 2006), thus let-<lb/>ting natural selection filter out poorly adapted genotypes <lb/>and mimicking naturally occurring hybridisation/invasion <lb/>events. <lb/> Lineage effects: variable outcomes among hybridisation events <lb/> Hybrid lineages derived from the same pair of parental spe-<lb/>cies, but formed with input from different parental individu-<lb/>als, can differ substantially in traits and performance (Py sek <lb/> et al. 2003; Hartman et al. 2013). Such variability can also <lb/>play out at the population level, which means that source <lb/>populations must be chosen carefully for experimental com-<lb/>parisons (or for resynthesising hybrid lineages). For instance, <lb/>a hybrid-parent comparison that draws parental individuals <lb/>from far beyond the supposed range of hybrid formation may <lb/>be uninformative, particularly where parental taxa are them-<lb/>selves highly variable. <lb/> Strong associations between hybridisation and polyploidy for weedy <lb/>and invasive hybrids <lb/> In many systems where the H-I hypothesis has been invoked, <lb/>hybridisation is also closely associated with polyploidy (Steb-<lb/>bins 1985). In our dataset, 37 of 59 plant hybrids (63%) are <lb/>reported to be polyploid, and of these nearly half have <lb/>increased ploidy relative to their parental taxa (16 of 37; see <lb/>Appendix 2). Polyploidy is not as well studied in fungal <lb/>pathogens (Albertin & Marullo 2012), but two fungal hybrids <lb/>from our database are also reported to be polyploid (Verticil-<lb/> lium longisporum and Phytophthora alni alni). Because poly-<lb/>ploidy alone may contribute to invasiveness (Pandit et al. <lb/> 2011; te Beest et al. 2012), an important future research direc-<lb/>tion is disentangling the effects of hybridisation from those of <lb/>polyploidy. <lb/>A robust 'gold standard' design for experimentally separat-<lb/>ing the effects of hybridisation from those of polyploidy in <lb/>systems experiencing both phenomena would be to create re-<lb/>synthesised hybrid lineages in which both effects are manipu-<lb/>lated independently. That is, experimental crosses would be <lb/>exposed (or not) to chromosome doubling agents such as col-<lb/>chicine to create homoploid hybrid descendants, allopolyploid <lb/>(hybrid) descendants, and autopolyploid (non-hybrid) descen-<lb/>dants from the same set of parents. Performance could then <lb/>be compared among parental taxa and their descendant lin-<lb/>eages to disentangle the phenotypic effects of hybridisation <lb/>and polyploidy. Such experiments with invasive polyploid taxa <lb/>would benefit by considering the same sources of variation <lb/>discussed above (habitat, generation and lineage effects), mak-<lb/>ing strong inferences on the roles of hybridisation and poly-<lb/>ploidy for polyploid invaders possible. <lb/> <note place="footnote"> © 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/> </note> <page>1474</page> <note place="headnote">S. M. Hovick and K. D. Whitney <lb/>Review and Synthesis <lb/></note> The H-I hypothesis: conclusions and future directions <lb/> Based on experiments that have been reported to date, the <lb/>H-I hypothesis is moderately supported across taxa, with <lb/>strong support in some systems (particularly plants) and little <lb/>support in others for which it has been postulated (e.g. Sene-<lb/> cio vulgaris var. hibernicus). Overall, our results suggest an <lb/>important but variable effect of hybridisation in triggering or <lb/>allowing invasions, although we note that these inferences rely <lb/>on estimates of fecundity, survival and size as proxies of pop-<lb/>ulation growth rates and thus of invasiveness. Variability in <lb/>outcomes (i.e. whether hybridisation triggers invasiveness or <lb/>not) is consistent with a main theme of the current invasion <lb/>literature emphasising that invasion is multicausal (Rejm anek <lb/> et al. 2005); as historic, species-specific and environment-spe-<lb/>cific factors can predominate, different invasions across differ-<lb/>ent taxa and locations are unlikely to be linked to a single <lb/>'smoking gun.' <lb/>The list of invasive hybrids initially compiled by Ellstrand <lb/>& Schierenbeck (2000) continues to grow (see Appendix 2); <lb/>however, in most systems we still lack well-designed experi-<lb/>ments to test the H-I hypothesis. Critical components of <lb/>future tests include ensuring that: (1) hybrid performance is <lb/>compared to the correct benchmark, i.e. the more invasive <lb/>parent, (2) tests are experimental and carried out in a com-<lb/>mon environment, (3) species-specific performance metrics are <lb/>used that accurately reflect the invasiveness of species with <lb/>different life histories, and (4) more tests in animal and fungal <lb/>taxa are attempted. <lb/> </body> <back> <div type="acknowledgement">ACKNOWLEDGEMENTS <lb/> Thanks to Allison Snow, Loren Rieseberg, Jim Grover, J. <lb/>Chris Pires, the Rudgers-Whitney lab group and anonymous <lb/>reviewers for comments on the manuscript, advice and discus-<lb/>sion. Sincere thanks to the authors of the original studies <lb/>included in the meta-analyses. This work was supported by <lb/>NSF grants DEB 0716868 and DEB 1257965 (to KDW) and <lb/>DEB 1146203 (to SMH and KDW). <lb/></div> <div type="annex">AUTHORSHIP <lb/> Designed the study, reviewed the literature, reconstructed phy-<lb/>logenies, and wrote the manuscript: SMH and KDW. Com-<lb/>piled data for the meta-analyses and performed all statistical <lb/>analyses: SMH. <lb/></div> <listBibl> REFERENCES <lb/> Abbott, R.J. (1992). Plant invasions, interspecific hybridization and the <lb/>evolution of new plant taxa. Trends Ecol. Evol., 7, 401–405. <lb/>Albertin, W. & Marullo, P. (2012). Polyploidy in fungi: evolution after <lb/>whole-genome duplication. Proc. Biol. Sci., 279, 2497–2509. <lb/>Anderson, E. (1948). Hybridization of the habitat. Evolution, 2, 1–9. <lb/>Anderson, E. & Stebbins, G.L. (1954). Hybridization as an evolutionary <lb/>stimulus. Evolution, 8, 378–388. <lb/>Arnold, M.L. & Hodges, S.A. (1995). Are natural hybrids fit or unfit <lb/>relative to their parents? Trends Ecol. Evol., 10, 67–71. <lb/>Arnold, M.L. & Martin, N.H. (2010). Hybrid fitness across time and <lb/>habitats. Trends Ecol. Evol., 25, 530–536. <lb/>Arriola, P.E. & Ellstrand, N.C. (1997). Fitness of interspecific hybrids in <lb/>the genus Sorghum: Persistence of crop genes in wild populations. Ecol. <lb/>Appl., 7, 512–518. <lb/>Ayres, D.R., Smith, D.L., Zaremba, K., Klohr, S. & Strong, D.R. (2004). <lb/>Spread of exotic cordgrasses and hybrids (Spartina sp.) in the tidal <lb/>marshes of San Francisco Bay, California, USA. Biol. Invasions, 6, <lb/>221–231. <lb/>Ayres, D.R., Zaremba, K., Sloop, C.M. & Strong, D.R. (2008). Sexual <lb/>reproduction of cordgrass hybrids (Spartina foliosa x alterniflora) <lb/>invading tidal marshes in San Francisco Bay. Divers. Distrib., 14, <lb/>187–195. <lb/>Bailey, J. (2013). The Japanese knotweed invasion viewed as a vast <lb/>unintentional hybridisation experiment. Heredity, 110, 105–110. <lb/>Baker, H.G. (1974). The evolution of weeds. Annu. Rev. Ecol. Syst., 5, <lb/>1–24. <lb/>Baker, H.G. & Stebbins, G.L. (1965). The Genetics of Colonizing Species. <lb/> Academic Press New York, NY. <lb/>Barton, N.H. & Hewitt, G.M. (1985). Analysis of hybrid zones. Annu. <lb/>Rev. Ecol. Syst., 16, 113–148. <lb/>te Beest, M., Le Roux, J.J., Richardson, D.M., Brysting, A.K., Suda, J., <lb/>Kubesova, M. et al. (2012). The more the better? The role of <lb/>polyploidy in facilitating plant invasions. Ann. Bot., 109, 19–45. <lb/>Blum, M.J., Walters, D.M., Burkhead, N.M., Freeman, B.J. & Porter, <lb/>B.A. (2010). Reproductive isolation and the expansion of an invasive <lb/>hybrid swarm. Biol. Invasions, 12, 2825–2836. <lb/>Brabec, J. & Py sek, P. (2000). Establishment and survival of three <lb/>invasive taxa of the genus Reynoutria (Polygonaceae) in mesic <lb/>mown meadows: a field experimental study. Folia Geobotanica, 35, <lb/>27–42. <lb/>Brennan, A.C., Barker, D., Hiscock, S.J. & Abbott, R.J. (2012). <lb/>Molecular genetic and quantitative trait divergence associated with <lb/>recent homoploid hybrid speciation: a study of Senecio squalidus <lb/>(Asteraceae). Heredity, 108, 87–95. <lb/>Campbell, L.G. & Snow, A.A. (2007). Competition alters life history and <lb/>increases the relative fecundity of crop-wild radish hybrids (Raphanus <lb/>spp.). New Phytol., 173, 648–660. <lb/>Campbell, L.G., Snow, A.A. & Ridley, C.E. (2006). Weed evolution after <lb/>crop gene introgression: greater survival and fecundity of hybrids in a <lb/>new environment. Ecol. Lett., 9, 1198–1209. <lb/>Chamberlain, S.A., Hovick, S.M., Dibble, C.J., Rasmussen, N.L., Van <lb/>Allen, B.G., Maitner, B.S. et al. (2012). Does phylogeny matter? <lb/>Assessing the impact of phylogenetic information in ecological meta-<lb/>analysis. Ecol. Lett., 15, 627–636. <lb/>Crawford, K.M. & Whitney, K.D. (2010). Population genetic diversity <lb/>influences colonization success. Mol. Ecol., 19, 1253–1263. <lb/>Culley, T.M. & Hardiman, N.A. (2009). The role of intraspecific <lb/>hybridization in the evolution of invasiveness: a case study of the <lb/>ornamental pear tree Pyrus calleryana. Biol. Invasions, 11, 1107–1119. <lb/>Darling, J.A., Bagley, M.J., Roman, J., Tepolt, C.K. & Geller, J.B. <lb/>(2008). Genetic patterns across multiple introductions of the globally <lb/>invasive crab genus Carcinus. Mol. Ecol., 17, 4992–5007. <lb/>Ellstrand, N.C. & Schierenbeck, K.A. (2000). Hybridization as a stimulus <lb/>for the evolution of invasiveness in plants? Proc. Natl Acad. Sci. USA, <lb/> 97, 7043–7050. <lb/>Elton, C.E. (1958). The Ecology of Invasions by Animals and Plants. <lb/> Methuen & Company Ltd, London, UK. <lb/>Flora of North America Editorial Committee (1993+).Flora of North <lb/>America North of Mexico. New York and Oxford. <lb/>Funk, V.A., Susanna, A., Stuessey, F.T. & Bayer, R.J. (2009). <lb/> Systematics, Evolution, and Biogeography of Compositae. International <lb/>Association for Plant Taxonomy Vienna, Austria. <lb/>Gallagher, K.G., Schierenbeck, K.A. & Dantonio, C.M. (1997). <lb/>Hybridization and introgression in Carpobrotus spp. (Aizoaceae) in <lb/>California. 2. Allozyme evidence.. Am. J. Bot., 84, 905–911. <lb/>Gammon, M.A., Baack, E., Orth, J.F. & Kesseli, R. (2010). Viability, <lb/>Growth, and Fertility of Knotweed Cytotypes in North America. <lb/> Invasive Plant Sci. Manage., 3, 208–218. <lb/></listBibl> <note place="footnote">© 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/></note> <note place="headnote">Review and Synthesis <lb/>The hybridisation-invasion hypothesis</note> <page>1475 <lb/></page> <listBibl>Geiger, J.H., Pratt, P.D., Wheeler, G.S. & Williams, D.A. (2011). Hybrid <lb/>vigor for the invasive exotic Brazilian peppertree (Schinus <lb/> terebinthifolius Raddi., Anacardiaceae) in Florida. Int. J. Plant Sci., <lb/> 172, 655–663. <lb/>Giraud, T., Refregier, G., Le Gac, M., de Vienne, D.M. & Hood, M.E. <lb/>(2008). Speciation in fungi. Fungal Genet. Biol., 45, 791–802. <lb/>Grosholz, E. (2010). Avoidance by grazers facilitates spread of an <lb/>invasive hybrid plant. Ecol. Lett., 13, 145–153. <lb/>Hahn, M.A., Buckley, Y.M. & Muller-Scharer, H. (2012). Increased <lb/>population growth rate in invasive polyploid Centaurea stoebe in a <lb/>common garden. Ecol. Lett., 15, 947–954. <lb/>Hartman, Y., Uwimana, B., Hooftman, D.A.P., Schranz, M.E., van de <lb/>Wiel, C.C.M., Smulders, M.J.M. et al. (2013). Genomic and <lb/>environmental selection patterns in two distinct lettuce crop-wild hybrid <lb/>crosses. Evol. Appl., 6, 569–584. <lb/>Hasselman, D.J., Argo, E.E., McBride, M.C., Bentzen, P., Schultz, T.F., <lb/>Perez-Umphrey, A.A. et al. (2014). Human disturbance causes the <lb/>formation of a hybrid swarm between two naturally sympatric fish <lb/>species. Mol. Ecol., 23, 1137–1152. <lb/>Hawkes, C.V., Douglas, A.E. & Fitter, A.H. (2010). Origin, local <lb/>experience, and the impact of biotic interactions on native and <lb/>introduced Senecio species. Biol. Invasions, 12, 113–124. <lb/>Hedges, S.B., Dudley, J. & Kumar, S. (2006). TimeTree: a public <lb/>knowledge-base of divergence times among organisms. Bioinformatics, <lb/> 22, 2971–2972. <lb/>Henery, M.L., Bowman, G., Mraz, P., Treier, U.A., Gex-Fabry, E., <lb/>Schaffner, U. et al. (2010). Evidence for a combination of pre-adapted <lb/>traits and rapid adaptive change in the invasive plant Centaurea stoebe. <lb/> J. Ecol., 98, 800–813. <lb/>Higgins, J.P.T. & Thompson, S.G. (2002). Quantifying heterogeneity in a <lb/>meta-analysis. Stat. Med., 21, 1539–1558. <lb/>Hooftman, D.A.P., Oostermeijer, J.G.B., Jacobs, M.M.J. & Den Nijs, <lb/>H.C.M. (2005). Demographic vital rates determine the performance <lb/>advantage of crop-wild hybrids in lettuce. J. Appl. Ecol., 42, 1086–1095. <lb/>Hooftman, D.A.P., Jong, M.J.D., Oostermeijer, J.G.B. & Den Nijs, H. <lb/>(2007). Modelling the long-term consequences of crop-wild relative <lb/>hybridization: a case study using four generations of hybrids. J. Appl. <lb/>Ecol., 44, 1035–1045. <lb/>Hovick, S.M., Campbell, L.G., Snow, A.A. & Whitney, K.D. (2012). <lb/>Hybridization Alters Early Life-History Traits and Increases Plant <lb/>Colonization Success in a Novel Region. Am. Naturalist, 179, 192–203. <lb/>Inderbitzin, P., Davis, R.M., Bostock, R.M. & Subbarao, K.V. (2011). <lb/>The Ascomycete Verticillium longisporum Is a Hybrid and a Plant <lb/>Pathogen with an Expanded Host Range. PLoS ONE, 6(3), e18260 <lb/>Johnson, J.R., Fitzpatrick, B.M. & Shaffer, H.B. (2010). Retention of <lb/>low-fitness genotypes over six decades of admixture between native and <lb/>introduced tiger salamanders. BMC Evol. Biol., 10, 147. <lb/>Kim, S.T., Sultan, S.E. & Donoghue, M.J. (2008). Allopolyploid <lb/>speciation in Persicaria (Polygonaceae): insights from a low-copy <lb/>nuclear region. Proc. Natl Acad. Sci. USA, 105, 12370–12375. <lb/>Kirk, H., Connolly, C. & Freeland, J.R. (2011). Molecular genetic data <lb/>reveal hybridization between Typha angustifolia and Typha latifolia <lb/>across a broad spatial scale in eastern North America. Aquat. Bot., 95, <lb/>189–193. <lb/>Lajeunesse, M.J. (2009). Meta-Analysis and the Comparative <lb/>Phylogenetic Method. Am. Naturalist, 174, 369–381. <lb/>Lajeunesse, M.J. (2011). phyloMeta: a program for phylogenetic <lb/>comparative analyses with meta-analysis. Bioinformatics, 27, 2603–2604. <lb/>LaRue, E.A., Grimm, D. & Thum, R.A. (2013). Laboratory crosses and <lb/>genetic analysis of natural populations demonstrate sexual viability of <lb/>invasive hybrid watermilfoils (Myriophyllum spicatum x M. sibiricum). <lb/> Aquat. Bot., 109, 49–53. <lb/>Lavergne, S. & Molofsky, J. (2007). Increased genetic variation and <lb/>evolutionary potential drive the success of an invasive grass. Proc. Natl <lb/>Acad. Sci. USA, 104, 3883–3888. <lb/>Le Roux, J. & Wieczorek, A.M. (2009). Molecular systematics and <lb/>population genetics of biological invasions: towards a better <lb/>understanding of invasive species management. Ann. Appl. Biol., 154, <lb/>1–17. <lb/>Liberati, A., Altman, D.G., Tetzlaff, J., Mulrow, C., Gotzsche, P.C., <lb/>Ioannidis, J.P.A. et al. (2009). The PRISMA statement for reporting <lb/>systematic reviews and meta-analyses of studies that evaluate healthcare <lb/>interventions: explanation and elaboration. Bmj-British Medical <lb/>Journal, 339, b2700. <lb/>Mack, R.N. (1996). Predicting the identity and fate of plant invaders: <lb/>emergent and emerging approaches. Biol. Conserv., 78, 107–121. <lb/>Maddison, D.R. & Schulz, K.-S. (2007). The Tree of Life Web Project. <lb/>http://tolweb.org. <lb/>Mengersen, K., Jennions, M.D. & Schmid, C.H. (2013). Statistical models <lb/>for the meta-analysis of nonindependent data. In: Handbook of Meta-<lb/>Analysis in Ecology and Evolution (eds Koricheva, J., Gurevitch, J., <lb/>Mengersen, K.). Princeton University Press, Princeton, NJ, pp. 255– <lb/>283. <lb/>Moody, M.L. & Les, D.H. (2002). Evidence of hybridity in invasive <lb/>watermilfoil (Myriophyllum) populations. Proc. Natl Acad. Sci. USA, <lb/> 99, 14867–14871. <lb/>Mraz, P., Garcia-Jacas, N., Gex-Fabry, E., Susanna, A., Barres, L. & <lb/>Muller-Scharer, H. (2012). Allopolyploid origin of highly invasive <lb/>Centaurea stoebe s.l. (Asteraceae). Mol. Phylogenet. Evol., 62, 612–623. <lb/>Nechwatal, J. & Mendgen, K. (2009). Evidence for the occurrence of <lb/>natural hybridization in reed-associated Pythium species. Plant. Pathol., <lb/> 58, 261–270. <lb/>Newcombe, G., Stirling, B., McDonald, S. & Bradshaw, H.D. (2000). <lb/>Melampsora xcolumbiana, a natural hybrid of M-medusae and M-<lb/>occidentalis. Mycol. Res., 104, 261–274. <lb/>Newcombe, G., Stirling, B. & Bradshaw, H.D. (2001). Abundant <lb/>pathogenic variation in the new hybrid rust Melampsora xcolumbiana <lb/>on hybrid poplar. Phytopathology, 91, 981–985. <lb/>Pandit, M.K., Pocock, M.J.O. & Kunin, W.E. (2011). Ploidy influences <lb/>rarity and invasiveness in plants. J. Ecol., 99, 1108–1115. <lb/>Parepa, M., Fischer, M., Krebs, C. & Bossdorf, O. (2014). Hybridization <lb/>increases invasive knotweed success. Evol. Appl., 7, 413–420. <lb/>Perrins, J., Williamson, M. & Fitter, A. (1992). Do annual weeds have <lb/>predictable characters? Acta Oecol.-Int. J. Ecol., 13, 517–533. <lb/>Poulsen, J.R., Osenberg, C.W., Clark, C.J., Levey, D.J. & Bolker, B.M. <lb/>(2007). Plants as reef fish: fitting the functional form of seedling <lb/>recruitment. Am. Naturalist, 170, 167–183. <lb/>Py sek, P. & Richardson, D.M. (2007). Traits associated with invasiveness <lb/>in alien plants: where do we stand? In: Biological Invasions. (ed <lb/>Nentwig, W.). Springer-Verlag Berlin, Germany, pp. 97–125. <lb/>Py sek, P., Brock, J.H., Bimova, K., Mandak, B., Jarosik, V., <lb/>Koukolikova, I. et al. (2003). Vegetative regeneration in invasive <lb/>Reynoutria (Polygonaceae) taxa: the determinant of invasibility at the <lb/>genotype level. Am. J. Bot., 90, 1487–1495. <lb/>R Development Core Team (2013). R: A Language and Environment for <lb/>Statistical Computing. (ed. R Foundation for Statistical Computing) <lb/>Vienna, Austria. <lb/>Reagon, M., Thurber, C.S., Olsen, K.M., Jia, Y.L. & Caicedo, A.L. <lb/>(2011). The long and the short of it: SD1 polymorphism and the <lb/>evolution of growth trait divergence in U.S. weedy rice. Mol. Ecol., 20, <lb/>3743–3756. <lb/>Rejm anek, M., Richardson, D.M. & Py sek, P. (2005). Plant invasions and <lb/>invasibility of plant communities. In: Vegetation Ecology. (ed van der <lb/>Maarel, E.). Blackwell, Oxford, pp. 332–355. <lb/>Richards, C.L., Walls, R.L., Bailey, J.P., Parameswaran, R., George, T. <lb/>& Pigliucci, M. (2008). Plasticity in salt tolerance traits allows for <lb/>invasion of novel habitat by Japanese knotweed s. l. (Fallopia japonica <lb/>and F-bohemica, Polygonaceae). Am. J. Bot., 95, 931–942. <lb/>Richardson, D.M., Py sek, P., Rejmanek, M., Barbour, M.G., Panetta, <lb/>F.D. & West, C.J. (2000). Naturalization and invasion of alien plants: <lb/>concepts and definitions. Divers. Distrib., 6, 93–107. <lb/>Ridley, C.E. & Ellstrand, N.C. (2009). Evolution of enhanced <lb/>reproduction in the hybrid-derived invasive, California wild radish <lb/>(Raphanus sativus). Biol. Invasions, 11, 2251–2264. <lb/></listBibl> <note place="footnote"> © 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/></note> <page>1476</page> <note place="headnote">S. M. Hovick and K. D. Whitney <lb/>Review and Synthesis <lb/></note> <listBibl> Rieseberg, L.H., Archer, M.A. & Wayne, R.K. (1999). Transgressive <lb/>segregation, adaptation and speciation. Heredity, 83, 363–372. <lb/>Rieseberg, L.H., Kim, S.C., Randell, R.A., Whitney, K.D., Gross, B.L., <lb/>Lexer, C. et al. (2007). Hybridization and the colonization of novel <lb/>habitats by annual sunflowers. Genetica, 129, 149–165. <lb/>Rosenberg, M.S., Rothstein, H.R. & Gurevitch, J. (2013). Effect sizes: <lb/>conventional choices and calculations. In Handbook of Meta-Analysis in <lb/>Ecology and Evolution. (eds Koricheva, J., Gurevitch, J., Mengersen, <lb/>K.). Princeton University Press Princeton, NJ, pp. 61–71. <lb/>Rouifed, S., Puijalon, S., Viricel, M.R. & Piola, F. (2011). Achene <lb/>buoyancy and germinability of the terrestrial invasive Fallopia x <lb/>bohemica in aquatic environment: a new vector of dispersion? <lb/> Ecoscience, 18, 79–84. <lb/>Ryan, M.E., Johnson, J.R. & Fitzpatrick, B.M. (2009). Invasive hybrid <lb/>tiger salamander genotypes impact native amphibians. Proc. Natl Acad. <lb/>Sci. USA, 106, 11166–11171. <lb/>Schierenbeck, K.A. & Ellstrand, N.C. (2009). Hybridization and the <lb/>evolution of invasiveness in plants and other organisms. Biol. Invasions, <lb/> 11, 1093–1105. <lb/>Schwenk, K., Brede, N. & Streit, B. (2008). Introduction. Extent, <lb/>processes and evolutionary impact of interspecific hybridization in <lb/>animals. Philos Trans R Soc B: Biol Sci, 363, 2805–2811. <lb/>Snow, A.A., Uthus, K.L. & Culley, T.M. (2001). Fitness of hybrids <lb/>between weedy and cultivated radish: implications for weed evolution. <lb/> Ecol. Appl., 11, 934–943. <lb/>Snow, A.A., Culley, T.M., Campbell, L.G., Sweeney, P.M., Hegde, S.G. <lb/>& Ellstrand, N.C. (2010). Long-term persistence of crop alleles in <lb/>weedy populations of wild radish (Raphanus raphanistrum). New <lb/>Phytol., 186, 537–548. <lb/>Stebbins, G.L. (1959). The role of hybridization in evolution. Proc. Am. <lb/>Philos. Soc., 103, 231–251. <lb/>Stebbins, G.L. (1985). Polyploidy, hybridization, and the invasion of new <lb/>habitats. Ann. Mo. Bot. Gard., 72, 824–832. <lb/>Sullivan, L., Wildova, R., Goldberg, D. & Vogel, C. (2010). Growth of <lb/>three cattail (Typha) taxa in response to elevated CO(2). Plant Ecol., <lb/> 207, 121–129. <lb/>Sun, J., Qian, Q., Ma, D.R., Xu, Z.J., Liu, D., Du, H.B. et al. (2013). <lb/>Introgression and selection shaping the genome and adaptive loci of <lb/>weedy rice in northern China. New Phytol., 197, 290–299. <lb/>Tavalire, H.F., Bugbee, G.E., LaRue, E.A. & Thum, R.A. (2012). <lb/>Hybridization, cryptic diversity, and invasiveness in introduced <lb/>variable-leaf watermilfoil. Evol. Appl., 5, 892–900. <lb/>Urbanska, K.M., Hurka, H., Landolt, E., Neuffer, B. & Mummenhoff, <lb/>K. (1997). Hybridization and evolution in Cardamine (Brassicaceae) at <lb/>Urnerboden, central Switzerland: biosystematic and molecular evidence. <lb/> Plant Syst. Evol., 204, 233–256. <lb/>Van Grunsven, R.H.A., Bos, F., Ripley, B.S., Suehs, C.M. & Veenendaal, <lb/>E.M. (2009). Release from soil pathogens plays an important role in <lb/>the success of invasive Carpobrotus in the Mediterranean. South <lb/>African J. Botany, 75, 172–175. <lb/>Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor <lb/>package. J. Stat. Softw., 36, 1–49. <lb/>Vila, M. & D'Antonio, C.M. (1998a). Fitness of invasive Carpobrotus <lb/>(Aizoaceae) hybrids in coastal California. Ecoscience, 5, 191– <lb/>199. <lb/>Vila, M. & D'Antonio, C.M. (1998b). Hybrid vigor for clonal growth in <lb/>Carpobrotus (Aizoaceae) in coastal California. Ecol. Appl., 8, <lb/>1196–1205. <lb/>Webb, C.O. & Donoghue, M.J. (2005). Phylomatic: tree assembly for <lb/>applied phylogenetics. Mol. Ecol. Notes, 5, 181–183. http:// <lb/>phylodiversity.net/phylomatic/. <lb/>Webb, C.O., Ackerly, D.D. & Kembel, S.W. (2008). Phylocom: software <lb/>for the analysis of phylogenetic community structure and trait <lb/>evolution. Bioinformatics, 24, 2098–2100. <lb/>Weber, E. & D'Antonio, C.M. (1999). Phenotypic plasticity in hybridizing <lb/>Carpobrotus spp. (Aizoaceae) from coastal California and its role in <lb/>plant invasion. Can. J. Bot.-Revue Canadienne De Botanique, 77, <lb/>1411–1418. <lb/>Whitney, K.D., Randell, R.A. & Rieseberg, L.H. (2006). Adaptive <lb/>introgression of herbivore resistance traits in the weedy sunflower <lb/>Helianthus annuus. Am. Naturalist, 167, 794–807. <lb/>Whitney, K.D., Ahern, J.R. & Campbell, L.G. (2009). Hybridization-<lb/>prone plant families do not generate more invasive species. Biol. <lb/>Invasions, 11, 1205–1215. <lb/>Whitney, K.D., Ahern, J.R., Campbell, L.G., Albert, L.P. & King, M.S. <lb/>(2010a). Patterns of hybridization in plants. Perspectives in Plant Ecol. <lb/>Evol. Systematics, 12, 175–182. <lb/>Whitney, K.D., Randell, R.A. & Rieseberg, L.H. (2010b). Adaptive <lb/>introgression of abiotic tolerance traits in the sunflower Helianthus <lb/>annuus. New Phytol., 187, 230–239. <lb/>Wikstr€ om, N., Savolainen, V. & Chame, M.W. (2001). Evolution of <lb/>angiosperms: calibrating the family tree. Proc. Biol. Sci., 268, <lb/>2211–2220. <lb/>Wu, W., Zhou, R.C., Ni, G.Y., Shen, H. & Ge, X.J. (2013). Is a new <lb/>invasive herb emerging? Molecular confirmation and preliminary <lb/>evaluation of natural hybridization between the invasive Sphagneticola <lb/>trilobata (Asteraceae) and its native congener S-calendulacea in South <lb/>China. Biol. Invasions, 15, 75–88. <lb/></listBibl> <div type="annex">SUPPORTING INFORMATION <lb/> Additional Supporting Information may be downloaded via <lb/>the online version of this article at Wiley Online Library <lb/>(www.ecologyletters.com). <lb/>Editor, Jessica Gurevitch <lb/>Manuscript received 12 May 2014 <lb/>First decision made 11 June 2014 <lb/>Manuscript accepted 14 August 2014 <lb/></div> <note place="footnote"> © 2014 The Authors. Ecology Letters published by John Wiley & Sons Ltd and CNRS. <lb/></note> <note place="headnote">Review and Synthesis <lb/>The hybridisation-invasion hypothesis</note> <page>1477</page> </back> </text> </tei>