Potential limits to the benefits of admixture during biological invasion

Abstract

Species introductions often bring together genetically divergent source populations, resulting in genetic admixture. This geographic reshuffling of diversity has the potential to generate favorable new genetic combinations, facilitating the establishment and invasive spread of introduced populations. Observational support for the superior performance of admixed introductions has been mixed, however, and the broad importance of admixture to invasion questioned. Under most underlying mechanisms, admixture's benefits should be expected to increase with greater divergence among and lower genetic diversity within source populations, though these effects have not been quantified in invaders. We experimentally crossed source populations differing in divergence in the invasive plant Centaurea solstitialis. Crosses resulted in many positive (heterotic) interactions, but fitness benefits declined and were ultimately negative at high source divergence, with patterns suggesting cyto‐nuclear epistasis. We explored the literature to assess whether such negative epistatic interactions might be impeding admixture at high source population divergence. Admixed introductions reported for plants came from sources with a wide range of genetic variation, but were disproportionately absent where there was high genetic divergence among native populations. We conclude that while admixture is common in species introductions and often happens under conditions expected to be beneficial to invaders, these conditions may be constrained by predictable negative genetic interactions, potentially explaining conflicting evidence for admixture's benefits to invasion.
This article is protected by copyright. All rights reserved.

Full text

100
|
wileyonlinelibrary.com/journal/mec Molecular Ecology. 2019;28:100–113.
© 2018 John Wiley & Sons Ltd
1 | INTRODUCTION
Introduced non‐native species are now a common feature of ecosys‐
tems on Ear th, and they are recognized as being one of the dominant
sources of biodiversity change in the Anthropocene (Ellis, Antill, &
Kreft, 2012; Vellend et al., 2013). While many factors will shape the
success of species introductions and their ecological interactions
with recipient environments (Sakai et al., 2001), there is increasing
evidence that genetic factors can play a role in this process (Baker
& Stebbins, 1965; Colautti & Barrett, 2013; Cox, 2004; Ellstrand
& Schierenbeck, 2000; Lee, 2002; Mesgaran et al., 2016; Rius &
Darling, 2014; Whitney & Gering, 2015). Understanding when,
where, and how genetic changes influence the outcomes of colo‐
nization is likely to be crucial to resolving broader questions about
Received: 14 May 2017
|
Revised: 6 November 2018
|
Accepted: 9 November 2018
DOI : 10.1111 /mec.1495 8
ORIGINAL ARTICLE
Potential limits to the benefits of admixture during biological
invasion
Brittany S. Barker1,2 | Janelle E. Cocio1 | Samantha R. Anderson1 |
Joseph E. Braasch1 | Feng A. Cang1 | Heather D. Gillette1,3 | Katrina M. Dlugosch1
1Universi ty of Arizona, Tucson , Arizona
2United St ates Geo logical Survey, Boise,
Idaho
3Northern Arizona University, Flagstaf f,
Arizona
Correspondence
Katrin a M. Dlugosch, Universit y of Arizon a,
Tucson, AZ.
Email: kdlugosch@email.arizona.edu
Funding information
National Institute of General Medical
Sciences, Grant/Award Number:
K12GM0 00708; National Institute of Food
and Agriculture, Grant/Award Number:
2015‐67013‐23000 and 2017‐67011‐
26034; Division of Integrative Orga nismal
Systems, Grant/Award Number: 1750280
Abstract
Species introductions often bring together genetically divergent source populations,
resulting in genetic admixture. This geographic reshuffling of diversity has the poten‐
tial to generate favourable new genetic combinations, facilitating the establishment
and invasive spread of introduced populations. Observational support for the supe‐
rior performance of admixed introductions has been mixed, however, and the broad
importance of admixture to invasion questioned. Under most underlying mecha‐
nisms, admixture's benefits should be expected to increase with greater divergence
among and lower genetic diversity within source populations, though these effects
have not been quantified in invaders. We experimentally crossed source populations
differing in divergence in the invasive plant Centaurea solstitialis. Crosses resulted in
many positive (heterotic) interactions, but fitness benefits declined and were ulti‐
mately negative at high source divergence, with patterns suggesting cytonuclear
epistasis. We explored the literature to assess whether such negative epistatic inter‐
actions might be impeding admixture at high source population divergence. Admixed
introductions repor ted for plants came from sources with a wide range of genetic
variation, but were disproportionately absent where there was high genetic diver‐
gence among native populations. We conclude that while admixture is common in
species introductions and often happens under conditions expected to be beneficial
to invaders, these conditions may be constrained by predictable negative genetic
interactions, potentially explaining conflicting evidence for admixture's benefits to
invasion.
KEY WORDS
cytonuclear interactions, epistasis, genetic diversity, heterosis, invasiveness, multiple
introductions

|
101
BARKER E t Al.
when species introductions will lead to establishment and invasive
spread (Bock et al., 2015; Colautti & Lau, 2015; Dlugosch, Anderson,
Braasch, Cang, & Gillette, 2015; Forsman, 2014; Hufbauer, 2008,
2017; Lee & Gelembiuk, 2008; Ochocki & Miller, 2017; Rius &
Darling, 2014; Szűcs, Melbourne, Tuff, & Hufbauer, 2014; Szűcs,
Melbourne, Tuff, Weiss‐Lehman, & Hufbauer, 2017; Weiss‐Lehman,
Hufbauer, & Melbourne, 2017; Williams, Kendall, & Levine, 2016).
In particular, successful introductions often result from admix‐
ture between divergent genotypes originating from different source
populations (Dlugosch, Anderson, et al., 2015; Dlugosch & Parker,
2008a; Uller & Leimu, 2011). Admixture has the potential to facil‐
itate species invasions by creating unique opportunities for posi‐
tive genetic interac tions among previously isolated alleles and for
adaptive evolution of novel genotypes, which could increase the
fitness of admixed populations (Keller & Taylor, 2010; Kolbe et al.,
2004; Lavergne & Molofsky, 2007; Wagner, Ochocki, Crawford,
Compagnoni, & Miller, 2017). These same mechanisms are known
to have contributed directly to non‐native species est ablishment
and the rise of particularly invasive novel genotypes in cases in‐
volving hybridization between species, and the potential for similar
benefits of admix ture within species appears widespread (Drake,
2006; Ellstrand & Schierenbeck, 2000; Hovick & Whitney, 2014).
Consequently, there is intensifying interest in the potential for ge‐
netic admixture to provide a general mechanism by which many
non‐native species are able to establish and develop into invaders
(Bock et al., 2015; Dlugosch, Anderson, et al., 2015; Frankham,
2005; Hufbauer, 2008, 2017; Molofsky, Keller, Lavergne, Kaproth,
& Eppinga, 2014; Rius & Darling, 2014; Verhoeven, Macel, Wolfe, &
Biere, 2011).
Given that introduced species are often derived from multiple
source populations, genetic admixture could be a frequent path
to the evolution of invasiveness, but only if its fitness effects are
typically positive under conditions commonly experienced during
introductions. Positive correlations between fitness traits and
evidence of admixture have been identified in some invasions
(Keller, Fields, Berardi, & Taylor, 2014; Rius & Darling, 2014), and
experimental admixtures have performed better in several studies
(Turgeon et al., 2011; van Kleunen, Röckle, & Stift, 2015; Wagner
et al., 2017). On the other hand, studies that have found no as‐
sociation between admixture and increased invasiveness have
called into question whether mixing of divergent populations
can realistically be expected to contribute to increased fitness
and introduction success across many invaders (Chapple, Miller,
Kraus, & Thompson, 2013; Dutech et al., 2012; Wolfe, Blair, &
Penna, 20 07). These conflicting results are not necessarily sur‐
prising given that studies of native species have long demon‐
strated that mating across different populations can have fitness
effects ranging from positive to detrimental, depending upon
the mechanisms underlying the interactions between genotypes
and their fitness effects (Birchler, Yao, Chudalayandi, Vaiman, &
Veitia, 2010; Chen, 2013; Edmands, 1999; Frankham et al., 2011;
Keller & Waller, 2002; Lynch, 1991; Price & Waser, 1979; Reed &
Frankham, 2003).
There are several non‐mutually exclusive mechanisms that
could generate positive fitness effects as a consequence of either
the genetic interac tions that can result from bringing together new
combinations of alleles within individuals or the increase in genetic
diversity that should result from combining divergent populations
(Dlugosch, Anderson, et al., 2015; Hufbauer, 2017; Lynch, 1991):
1. Genetic Rescue. Also known as “directional dominance” (Birchler
et al., 2010; Chen, 2013), genetic rescue refers to the rescue
of deleterious inbred (homozygous) loci by outbreeding with
a divergent population (Tallmon, Luikart, & Waples, 2004).
Homozygous loci in introduced populations can be derived from
both historic genetic load already present in a native source
population, and additional fixation of deleterious variants during
founding events (Excoffier, Foll, & Petit, 2009). Multiple intro‐
ductions from divergent sources can therefore provide genetic
rescue to an establishing population by contributing/restoring
superior alleles that increase the fitness of introduced geno‐
types, potentially resulting in superior genotypes that transgress
the fitness of both parental populations (Ellstrand &
Schierenbeck, 2000; Hufbauer, 2008; Keller & Taylor, 2010;
Rius & Darling, 2014; van Kleunen et al., 2015). Genetic rescue
should benefit invasions when diversity within founding pop‐
ulations is low, and there is significant genetic load to rescue
(Lohr & Haag, 2015; Lynch, Conery, & Burger, 1995). The fitness
benefits should occur in the first generation of admixture and
scale positively with divergence between source populations,
until all loci are rescued and fitness gains plateau (Lynch, 1991).
2. Overdominance. Also known as “heterozygote advantage,” over‐
dominance occurs when heterozygous allele combinations have
higher fitness than any homozygous genot ype (Birchler et al.,
2010; Chen, 2013). These effects are expec ted to manifest pri‐
marily in the firs t (F1) genera ti on of an admix ture eve nt , but decay
quickly due to the increase in homozygosit y that occurs over sub‐
sequent generations, unless heterozygosity can be preserved by
asexual propagation, polyploidy, or other means (Drake, 2006;
Ellstrand & Schierenbeck, 2000; Facon, Pointier, Jarne, Sarda, &
David, 2008). These effec ts should be strongest when oppor tuni‐
ties for novel heterozygosity in admixed progeny are highest, that
is, when there are greater numbers of fixed differences between
source populations.
3. Epistasis. Epistasis occurs when alleles at different loci interact
(Phillips, 2008). Epistatic interactions that arise from admixture
are predicted to have increasing ef fect s on fitness as divergence
among source populations increases, due to natural selection for
locally co‐adapted allele combinations and/or to the build‐up of
Bateson–Dobzhansky–Muller incompatibilities from genetic drift,
acting separately in each source population (Lynch & Walsh, 1998;
Moyle & Nakazato, 2010). The fitness effects of epistatic interac‐
tions in the first generation can be positive (heterotic) but will be‐
come increasingly negative with greater genetic distance between
parents, particularly in later generations when co‐adapted
multilocus genotypes are broken apart by recombination (i.e.,

102
|
BARKER Et Al.
“hybrid breakdown” and “outbreeding depression”; Bomblies et
al., 2007; Frankham et al., 2011; Lynch, 1991; Orr & Turelli, 2001).
Negative epistatic effects are thought to be one of the most im‐
portant paths to reproductive isolation and speciation, and could
impede admixture of divergent sources during multiple introduc‐
tions, though the influence of these effects on species invasions is
rarely discussed and largely unknown (Carroll, Dingle, & Famula,
2003; Dlugosch, Anderson, et al., 2015).
4. Complementarity. As diversity increases within a population, dif‐
ferent genotypes may occupy somewhat different and comple‐
mentary niches, sometimes increasing mean fitness across the
population as a whole (Chen, Wang, Jiao, & Schmid, 2015;
Crawford & Whitney, 2010). Complementarity will occur immedi‐
ately under multiple introductions, even before interbreeding. It
should be most beneficial at low genetic diversity within a focal
population, where niche diversity among genotypes is low, and
become increasingly likely as divergence between admixing popu‐
lations increases and greater numbers of genotypes are combined
(Le Roux, Blignaut, Gildenhuys, Mavengere, & Berthouly‐Salazar,
2014; Wang et al., 2012), though this ef fect may plateau with di‐
versity after niches are exhausted (Ellers, Rog, Braam, & Berg,
2011).
5. Evolutionary Rescue. Finally, across multiple generations and
longer timescales, populations that will go extinct or fail to spread
because they lack adaptation to local conditions could be rescued
by inputs of additional genetic variation, a scenario known as
“evolutionary rescue” (Carlson, Cunningham, & Westley, 2014).
For introduced species in particular, additional inputs of genetic
diversit y could facilitate both adapt ation to the novel environ‐
ment of introduction and adaptation in traits that facilitate coloni‐
zation in general, such as increased dispersal ability (Cox, 2004;
Holt, Barfield, & Gomulkiewicz, 2005; Phillips, Brown, Webb, &
Shine, 20 06; Prentis, Wilson, Dormontt, Richardson, & Lowe,
2008; Thompson, 1998; Weiss‐Lehman et al., 2017). Evolutionary
rescue should be most likely in populations with low genetic di‐
versity (such that adaptive variation is limiting), and increasingly
impactful as the divergence between admixing populations in‐
creases and combines a greater numbers of unique alleles
(Ochocki & Miller, 2017; Rieseberg et al., 2007; Szűcs, Melbourne,
et al., 2017; Wagner et al., 2017).
Based on these mechanisms, fitness benefits from admixture
should be expected to vary in predictable ways (Figure 1a,b). As diver‐
gence between sources increases, opportunities for rescue of genetic
load, the creation of overdominant heterozygotes, epistatic interac‐
tions among loci, complementarity and evolutionary rescue should
all increase. These interactions should all be positive for fitness at
low divergence, though benefits of most effects should ultimately
plateau, and epistatic interactions will become increasingly negative,
with increasing divergence (Figure 1a). Benefits of all mechanisms
other than epistasis should also be strongest where founding pop‐
ulations harbour low within‐population diversity (Figure 1b), espe‐
cially to the extent that this represents fixation of deleterious alleles
(Lohr & Haag, 2015) (with the caveat that species with a history of
inbreeding may have purged genetic load [Crnokrak & Barrett, 2002]
and therefore stand to benefit only from evolutionary rescue, com‐
plementarity and/or overdominance when at low genetic variation).
Thus, the fitness effects of admixture are expected to vary in mag‐
nitude under different scenarios, but to be either positive or neutral
under most mechanisms other than epistasis (Dlugosch, Anderson,
et al., 2015; Hufbauer, 2017; Lynch, 1991), consistent with the
idea that admixture could be almost universally beneficial to invad‐
ers (Frankham, 2005; Hufbauer, 2008, 2017; Rius & Darling, 2014;
Verhoeven et al., 2011).
With these considerations in mind, where do admixture events
and their benefit s to introduced species fall in the parameter space
of divergence among sou rce populations and genetic variation wit hin
founding populations? To date, there has been little study of the di‐
vergence among potentially admixing populations during species in‐
troductions (Dlugosch, Anderson, et al., 2015). The question of how
much genetic diversity is available in introduced populations has
received considerably more attention. Previous studies have shown
that introduced populations generally do not experience large re‐
ductions in diversity relative to native populations (though certainly
many exceptions exist) and that low levels of marker diversity do
not prevent adaptation (Dlugosch & Parker, 2008a, 2008b; Szűcs,
Melbourne, et al., 2017; Uller & Leimu, 2011). Nevertheless, a lack of
strong founder/bottleneck effects does not preclude the presence
of historical genetic load and/or low diversity derived from source
populations. Manipulations of genetic diversity have shown a range
of effects on the performance of experimental invading populations,
and the potential benefits of admixture in this regard remain an ac‐
tive area of research (Crawford & Whitney, 2010; Hufbauer, 2017;
Ochocki & Miller, 2017; Szűcs et al., 2014; Szűcs, Melbourne, et al.,
2017; Wagner et al., 2017; Weiss‐Lehman et al., 2017; Williams et
al., 2016). Species introductions might thus create particularly abun‐
dant opportunities for genetic interactions (mechanisms 1–3 above)
which can provide immediate fitness benefits in the first generation
of admixture and will be especially relevant to establishing admixed
populations.
In this study, we experimentally test for fitness effects of ge‐
netic interactions in controlled crosses of Centaurea solstitialis L.
(“yellow starthistle”; Asteraceae), a highly invasive annual plant in
the Americas. We cross native populations that span a range of ge‐
netic divergence and diversity to test for associations between the
fitness of admixed progeny and these factors, which are expected
to shape the outcome of genetic interactions that will manifest in
early generations of admixed mating. We ask whether fitness bene‐
fits increase as expected. We then explore the literature to put our
results into a broader context by asking whether reported cases of
admixture in introduced plants are being realized under conditions
in which we might expect admixture to be favourable for many in‐
vaders, and whether there might be limits to these benefits due to
negative genetic interactions. We interpret our findings with respect
to the likelihood that admixture is a general mechanism promoting
the invasiveness of introduced species.

|
103
BARKER E t Al.
2 | MATERIALS AND METHODS
2.1 | Study system
We experimentally tested for genetic interactions in controlled
crosses among native populations of C. solstitialis spanning a range
of genetic diversity and divergence. This species was introduced
in large numbers to the Americas as a seed contaminant of alfalfa
stock imported from the Old World, where it escaped agricultural
fields and became a major pest of grasslands (Gerlach, 1997). A
lineage in western Europe appears to be the product of ancient
(a)
(c) (d)
(b)
(e) (f)

104
|
BARKER Et Al.
admixture between populations from eastern Europe and Asia and
has served as a bridgehead for invasions in the Americas (Barker,
Andonian, Swope, Luster, & Dlugosch, 2017). Several invading popu‐
lations show evidence of additional recent admixture with other na‐
tive populations (Barker et al., 2017; Dlugosch, Lai, Bonin, Hierro,
& Rieseberg, 2013; Eriksen et al., 2014). Phenotypic studies have
revealed evolutionary increases in plant size during this range ex‐
pansion, which is associated with higher fitness (Dlugosch, Cang,
et al., 2015; Eriksen, Desronvil, Hierro, & Kesseli, 2012; Widmer,
Guermache, Dolgovskaia, & Reznik, 2007). Experimental crosses in
C. solstitialis provide an oppor tunity to better understand how ad‐
mixture might be influencing the success of invading lineages.
2.2 | Collections
Genotypes of C. solstitialis were collected from 21 native sites
(Figure 2a; Supporting Information Table S1) spanning multiple po‐
tential source regions across western and eastern Europe, Asia and
southern Greece (Barker et al., 2017; Gerlach, 1997; Tutin, Heywood,
& Burges, 2010). Seeds were collected from wild plants during
August–September 2008, from each of 9 to 22 maternal plants lo‐
cated at least 1 m apart along a linear transect at each site. In Asia,
seeds of mothers at each site were combined into bulk collections by
site. The species was identified by the authors according to the Flora
Europaea ( Tutin et al., 2010) and vouchers from each sampling site
are available at the University of Arizona herbarium (ARIZ, Accession
nos in Supporting Information Table S1).
2.3 | Genetic variation within and divergence
among native populations
We previously identified four geographically structured, genetically
divergent populations in the native range (Figure 2a) using popula‐
tion genomic analyses of double‐digest restriction site‐associated se‐
quences (dd R ADseq; Bar ke r et al., 2017 ). Her e, we use singl e nucleoti de
polymorphism (SNP) information from these ddRADseq reads to quan‐
tify genome‐wide sequence divergence among genotypes from these
four populations (N = 155 individuals; Supporting Information Table S1;
NCBI sequ ence re ad arch ive BioProje ct PR JNA 275986). Det ail ed meth ‐
ods for sequencing and SNP generation are described in Barker et al.
(2017). Briefly, tota l genomic DNA was isolated using a CTAB/PVP DNA
extraction protocol (Webb & Knapp, 1990), and digested with enzymes
PstI and MseI to create fragments for ddRADseq (Peterson, Weber, Kay,
Fisher, & Hoekstra, 2012). Unique combinations of individual P1 and
P2 barcoded adapters were annealed to each sample, and resulting li‐
braries size‐selected for fragments 350–650 bp. Size‐selected libraries
FIGURE 1 Identifying the conditions under which admixture might be most favourable during invasions. Shown are predicted
relationships for the performance of admixed populations as a function of (a) the genetic divergence between their source populations and
(b) the genetic variation within their initial founding populations, under different genetic and evolutionary mechanisms. In a linear model
of progeny performance in experimental crosses of the invasive plant Centaurea solstitialis (excluding heterotic outlier [Asia × southern
Greece]), genetic divergence (interpopulation π) had a significant negative effect (c, p < 0.01) and genetic variation (intrapopulation π) in
maternal populations had a significant positive effect (d, p < 0.001) on the deviation of progeny from mid‐parent genotype expectations.
Data points in (c, d) show progeny deviation from mid‐parent as partial residuals after taking into account all other effects in the model. In
the literature, genetic admixture in introduced plant populations is reported more often at intermediate genetic divergence (FST and related
metrics for microsatellite markers) among potential source populations in the native range (e) and shows no relationship with average genetic
variation (HE) within native populations (f)
FIGURE 2 Sampling sites and
sequence variation in the native range
of Centaurea solstitialis. (a) Sampling sites
for this study (large dots) span genetically
divergent populations in western Europe
(WE, blue), eastern Europe (EE, purple),
southern Greece (SG, red) and Asia (A S,
green) as previously identified from
population genomic analyses (previous
sampling indicated by large and small
dots; Barker et al., 2017). (b) Average
intrapopulation nucleotide diversity (π)
across the total leng th of all ddRADseq
reads within each sampling site (dots) and
for all individuals pooled within a region
(bars). (c) Average interpopulation π in
pairwise comparisons between individuals
from different populations [Colour figure
can be viewed at wileyonlinelibrary.com]
km
(a)
(b) (c)

|
105
BARKER E t Al.
were enriched using 14 PCR cycles and sequenced on an Illumina HiSeq
2000 or 2500 to generate 100‐bp paired‐end reads. Reads were qual‐
ity‐filtered and demultiplexed using SɴᴏWʜɪᴛᴇ 2.0.2 (Dlugosch et al.,
2013). R1 reads were trimmed to a uniform length of 76 bp for final
SNP analysis. The denovo_map.pl pipeline program in stacks 1.20
(Catchen, Amores, Hohenlohe, Cresko, & Postlethwait, 2011; Catchen,
Hohenlohe, Bassham, Amores, & Cresko, 2013) was used to merge
identical reads into “stacks,” identify polymorphic sites, create a catalog
of loci across individuals and determine the allelic state at each locus
in each individual (Barker et al., 2017). Here, the “populations” module
in stacks was used to expor t pol ymor phisms for individual sampl ing sit es
(N = 35) and for the four native‐range populations. For both exports,
a locus was required to be genotyped in at least 70% of the samples in
each population (−r 0.7) and have a minimum stack depth (−m) of 10. We
also required a locus to be observed in at least 80% of sampling sites (−p
28) when exporting polymorphisms from the individual sampling sites,
and 100% of populations for the four population expor t. This resulted
in 2,287 polymorphic ddRADseq tags across individual sampling sites,
including 9,301 to 12,403 SNPs within each sampling site, and 1,585
polymorphic tags across the four populations, including 9,036 to 9,094
SNPs within each population. To quantify genetic variation, we ob‐
tained intrapopulation nucleotide diversity (π) as the average number of
nucleotide differences among alleles (including invariant sites) across all
ddRADseq tags using the “populations” module in stacks. To quantify di‐
vergence between two populations, we used a custom script to extract
pairwise interpopulation π (i.e., Dxy) between alleles at the same locus
from different regions (Dryad https://doi.org/10.5061/dryad.11j4k4v).
2.4 | Experimental crosses
Experimental crosses were conducted to compare the performance
of matings within and among the four native populations, using par‐
ents reared in a common environment (Supporting Information Figure
S1). We reared parents to flowering in a glasshouse at the Universit y
of Arizona (as in Dlugosch, Cang, et al., 2015). Bet ween 9 and 22 par‐
ents/site were reared from seeds of different field mothers, or from
bulk collections at A sian sites (N = 332; Supporting Information Table
S1). Flowering heads (capitula) were covered with fine mesh bags
while in bud, and hand‐pollinated using a single pollen donor when
a large fraction of florets were receptive. Strong self‐incompatibility
in this species was verified by manual self‐pollination and by bagging
unmanipulated capitula (yielding 0% seed set). Seeds were collected
at maturit y from a total of 349 successful crosses within and among
populations (Dryad https://doi.org/10.5061/dryad.11j4k4v).
Progeny (N = 523, including 1–3 per cross) were reared for growth
measurements in the common glasshouse environment. Increased
growth is a fundamental metric of heterosis in experimental crosses
(Birchler et al., 2010; Chen, 2013), and it is a key trait whose evolution is
associated with increased fitness in invasions of C. solstitialis (Dlugosch,
Cang, et al., 2015; Eriksen et al., 2012). All size measurement s of both
“source” genot ypes (produced by within‐population crosses) and “ad‐
mixed” genotypes (produced by among population crosses) were made
in the same experiment, at both 4 and 5 weeks of age (Suppor ting
Information Figure S1). Size at both time points was measured using
a non‐destructive size index ([maximum leaf length × maximum leaf
width]1/2 × leaf number) that has been shown to have a strong linear
correlation with total biomass under glasshouse conditions in this spe‐
cies (Dlugosch, Cang, et al., 2015). Exponential grow th rates between
these two measurements for each plant were compared using REML
analyses of variance (ANOVA) with fixed effects of (a) the source pop‐
ulation of crossed genotypes; (b) observer (the person measuring the
plants); and nested effects of (c) cross direction (the source popula‐
tion of the maternal vs. paternal genotype); and (d) individual parental
combination. Least‐squares means (LSM) and standard errors were ex‐
tracted from these models for use in analyses below.
2.5 | Mid‐parent trait values
Source populations can be genetically divergent from one another
in growth rate due to local adapt ation or genetic drift, so we tested
for evidence of non‐additive genetic interactions during admixture
by comparing admixed genotypes to mid‐parent expec tations that
assume additivity in the trait (Dlugosch, Cang, et al., 2015). To calcu‐
late these expect ations, both parental and admixed genot ypes must
be products of crosses conducted in a common environment, to
minimize transgenerational plasticity effects on the traits of interest.
Moreover, it is essential to compare measurements of parental and
admixed genotypes at the same life stage within the same experi‐
ment, because growth rate is highly sensitive to experimental condi‐
tions. To accomplish this for C. solstitialis, we generated a distribution
of pseudo‐mid‐parent values by randomly drawing combinations of
“source” growth phenotypes from the progeny of within‐popula‐
tion crosses (i.e., representatives of parental lineages, produced by
crosses in the common glasshouse environment, as described above)
that were reared and measured at the same time as admixed geno‐
types (Supporting Information Figure S1). Pseudo‐parental combina‐
tions with observer effects removed were drawn with replacement
1,00 0 times, and their average (mid‐parent) growth rates calculated
to create a distribution of pseudo‐mid‐parent values that were com‐
pared to grow th rates of admixed genotypes using two‐tailed t tests.
2.6 | Relationship between progeny
performance and parental population
diversity and divergence
To examine the relationship between growth rates in our experimen‐
tal crosses and both genetic variation within source populations and
genetic divergence between source populations, we used a linear
model to explain deviation of growth patterns from mid‐parent ex‐
pectations (Ymp). The model included fixed effects in the form:
where πm a nd πp are maternal and p aternal source int rapopulation
π, respectively, which are predicted to scale negatively with growth
deviation due to the benefits of admixture at low within‐population
Ymp
=μ+𝜋
m
+𝜋
p
+𝜋
mp
+
gm
+
gp
+
emp

106
|
BARKER Et Al.
variation; πmp is interpopulation π between parental source popu‐
lations, predicted to influence growth deviation positively where
genetic load is increasingly rescued, and negatively where there
are epistatic incompatibilities accumulating at high divergence; gm
and gp are, respectively, maternal and paternal source population
LSM grow th rate phenotype (“source” lineages grown in the same
experiment as admixed progeny, as described above), predicted to
correlate with growth deviation positively where there are trans‐
generational plasticity/epigenetic effects; and emp is the residual
error. All interaction terms were non‐significant (p > 0.1) and are not
shown. All non‐significant terms were removed from the final model.
2.7 | Literature survey
We explored the literature to ask whether reported cases of admix‐
ture are being realized under conditions in which we might expect
admixture to be favourable to many invaders, and whether there
might be limits on these benefits due to negative genetic interac‐
tions. Using papers that repor ted the distribution of molecular ge‐
netic variation in native populations and in populations introduced
from the native range to new areas (i.e., “primar y” introductions),
we extracted metrics of genetic variation and divergence for native
populations and recorded whether admixture was reported in the
introduced populations. We analysed an older data set of this type
(Dlugosch, Anderson, et al., 2015) to identify the effect s of study
design on reports of admixture (Supporting Information Methods
S1). Based on this information, we focused our survey on micros‐
atellite‐based studies and controlled for the number of sites sur‐
veyed in the native range (see Suppor ting Information Methods S1
and Results S1). We examined the relationship between reports of
admixture and metrics of genetic divergence and within‐popula‐
tion genetic variation across native sites using logistic regressions
with the number of sites sampled as a covariate (ln‐transformed).
We quantified potential source population divergence using FST and
related metrics (ɸST, GST) reported among all native populations in
a study. We quantified genetic variation as the mean across all na‐
tive sampling sites and loci (within a single type of genetic marker)
of expected heterozygosity (HE) and obser ved heterozygosit y (HO).
All statistical tests were performed in j mp 11 (SAS Institute, Cary,
NC, USA).
3 | RESULTS
3.1 | Experimental crosses in C. solstitialis
Nucleotide diversity (π) varied within and across native C. solsti‐
tialis populations and was consistently higher between pairs of
populations than within populations (Figure 2). Intrapopulation π
ranged from 0.004 to 0.005 average SNPs/site, with the highest
value in western Europe and the lowest in Asia. Interpopulation π
increased to 0.005–0.008 average SNPs/site. The largest values of
interpopulation π occurred in comparisons of alleles from south‐
ern Greece to those from other populations, consistent with our
previous observation of a highly differentiated lineage occupying
the Apennine–Balkan Peninsulas (Barker et al., 2017).
Growth rates differed significantly among different admixture
combinations of parental source populations (ANOVA F5,11 = 3.59,
p = 0.005), spanning an order of magnitude in exponential growth
rates (Supporting Information Figure S2). Of the 12 combinations of
crosses between genotypes from dif ferent maternal and paternal
source populations, we found that seven deviated significantly from
pseudo‐mid‐parent expectations, and all but one of these were in
the positive (heterotic) direction (Figure 3). Progeny of the mater‐
nal source population from western Europe experienced multiple
heterotic interactions with those from other populations. The single
negative interaction occurred in crosses with a maternal genot ype
originating from eastern Europe and a paternal genotype originating
from the highly divergent population in southern Greece. In con‐
trast, maternal genotypes originating from southern Greece showed
positive genetic interactions with paternal genot ypes from other re‐
gions, including those from eastern Europe.
Holding maternal source population constant, crosses to in‐
creasingly divergent paternal source populations showed a wide va‐
riety of trends in growth performance (Figure 3), including positive
(Asia), negative (eastern Europe) and curvilinear relationships peak‐
ing at intermediate values (western Europe and southern Greece).
Using data from all crosses, the deviations of admixed progeny from
additive expect ations were not predicted by parental source pop‐
ulation divergence (interpopulation π), parental source intrapop‐
ulation π or parental lineage phenotypes in a linear model (model
p = 0.24). Removing the extreme heterotic data point in the cross
of Asia × southern Greece (see Figure 3a, Supporting Information
Figure S3), however, yielded a highly significant model, strongly
predicting growth deviation (
r2
adj
= 0.92; F4,10 = 28.6, p = 0.005), in
which the direction of main effects was most consistent with epi‐
static interactions among loci (Figure 1c,d). In particular, deviations
in growth were negatively associated with divergence between
parental source populations (Figure 1c; πmp effect p = 0.006) and
positively associated with maternal source intrapopulation π such
that high genetic variation in the maternal source population made
heterotic interactions stronger (Figure 1d; πm effect p = 0.0007).
Progeny performance depended significantly on the source
of the maternal versus the paternal genotype in the cross
(Figure 4a; ANOVA of progeny performance with nested ef‐
fect of cross direction: F6,5 = 0.014, p = 0.02). This result could
suggest that transgenerational maternal effects influenced the
growth of progeny, in which case a positive relationship would
be predic ted between maternal lineage phenotype and progeny
deviation from mid‐parent expectations. Yet, maternal lineage
growth rate negatively predicted deviation from mid‐parent
expectations (Figure 4b; gm effect p < 0.0001), and it s interac‐
tion with genetic divergence was also negative (πmp × gm effect
p = 0.0009). This result is not consistent with maternal effects,
but could be explained by non‐additive genetic interac tions be‐
tween the maternally inherited cytoplasmic genome and the bi‐
parentally inherited nuclear genome.

|
107
BARKER E t Al.
3.2 | Literature survey
In total, we acquired data from introductions of 34 species (Dr yad
https://doi.org/10.5061/dryad.11j4k4v), including reports of ad‐
mixture in 15 (44%) of these. We found no monotonic relationships
between admixture and either genetic divergence (ln‐transformed
values: N = 26,
𝜒2
2
= 0.42, p = 0.81), or within‐population diversit y
measured as HE (N = 29,
𝜒2
2
= 2.56, p = 0.28; Figure 1f) or HO (N = 24,
𝜒2
2
= 0.03, p = 0.98), which was also the case in the older data set
of 167 plant and non‐plant species (Supporting Information Results
S1). Inspection of the data revealed a strong cur vilinear relationship
between admixture and genetic divergence (Figure 1f).
4 | DISCUSSION
Introduced species have opportunities to gain fitness advantages
from increases in genetic diversity and novel genetic combinations
that are associated with admixture (Dlugosch, Anderson, et al.,
2015; Frankham, 2005; Hufbauer, 2008; Rius & Darling, 2014). The
fitness benefits of admix ture are expected to grow as the divergence
between source populations increases and as the genetic variation
within source populations decreases, under a variety of mechanisms
(Dlugosch, Anderson, et al., 2015; Hufbauer, 2017; Lynch, 1991). We
found significant genetic interactions in over 50% of our controlled
crosses among native C . solstitialis populations, and all but one were
positive (heterotic) in the first generation. Heterotic effects declined
in magnitude as divergence among populations increased, however,
in a manner consistent with epistatic cytonuclear interactions. Such
interactions could limit the benefits of admixture for invaders gener‐
ally. We found that while admix ture has been frequently reported in
the literature, it is dispropor tionately lacking from systems with high
genetic divergence among potential source populations. Our find‐
ings are consistent with the idea that admix ture might often occur
under conditions that are likely to be favourable for introduced spe‐
cies, but also that these conditions should be bounded by negative
epistatic interactions at high levels of divergence.
Epistasis is unique in its potential to generate both positive and
negative interactions during admixture. Any positive interactions
are expected to diminish and become negative in later genera‐
tions, as recombination breaks up co‐adapted allele combinations
(Lynch, 1991). Thus, benefits from epistasis during admix ture are
expected to be transient over time. Experiments with first‐genera‐
tion crosses should generally demonstrate the maximum benefit of
FIGURE 3 For each maternal region, panels (a–d) show growth rates of Centaurea solstitialis progeny from experimental crosses to
fathers from other regions and the mid‐parent expectations for each cross. Grow th rates are shown versus average pairwise nucleotide
diversity (π) between parental populations. Colour codes for each cross are as in Figure 2. Within‐region crosses are indicated by solid
colours, and between‐region crosses by paternal region colour outlined by maternal region colour. Mid‐parent expectations are shown as
means (black bars) ± SEM (grey bars). Growth rates are least‐squares means ± SEM, and significant deviations from mid‐parent distributions
are indicated as: ***p < 0.0001; **p < 0.001; *p < 0.05 [Colour figure can be viewed at wileyonlinelibrary.com]
(a) (b)
(c) (d)

108
|
BARKER Et Al.
epistasis and will be conservative with respect to identif ying neg‐
ative interactions that might ultimately constrain interbreeding of
divergent populations (see Introduction 3. Epistasis). Impor tantly,
transient increases in fitness can be highly beneficial when found‐
ing populations are struggling to establish (i.e., the “catapult ef‐
fect”; Drake, 2006). Further, negative epistatic interactions from
hybrid breakdown in later generations can also be avoided by
backcrossing with resident parental populations (Dlugosch, Cang,
et al., 2015), and high fitness of early generation admixed geno‐
types can be maintained through clonal propagation or polyploidy
in some cases (Facon et al., 2008). Indeed, clonal or polyploid
spread of hybrids has provided some of our best examples to date
of the evolution of invasiveness (Ellstrand & Schierenbeck, 2000;
Hufbauer, 2008).
The complexity of these outcomes from epistasis alone argues
that we should expect admixture to var y in its contribution to the
establishment and invasion of introduced populations, and early
generation crosses should be particularly useful for gauging its most
powerful effects on fitness. Notably, some of the strongest evidence
of heterotic interactions in our crosses of C. solstitialis (with some of
the highest resulting plant growth rates) occurred when the west‐
ern European population ser ved as the maternal parent. This lineage
appears to be a primary contributor of invasions into the Americ as
(Barker et al., 2017; Eriksen et al., 2014). Strong heterotic interac‐
tions in this lineage suggest that a possible contributor to its success
as an invasion “bridgehead” (an invasion which initiates many sub‐
sequent invasions; Lombaert et al., 2011), could be its propensity to
foster beneficial admixture events (e.g., Turgeon et al., 2011).
Both the identity of the maternal parent population of origin and
genetic diversity within the maternal population were important to
the performance of C. solstitialis crosses, which strongly suggests
that the divergence of maternal (cytoplasmic) DNA from the pater‐
nal nuclear genome is the underlying cause of the interactions that
we observed. There is increasing evidence that cytonuclear inter‐
actions within species are common and can have impor tant phe‐
notypic effects in both animals and plants (Ballard & Melvin, 2010;
Bock, Andrew, & Rieseberg, 2014). The additional effec t of genetic
diversit y within the maternal population would seem to suggest that
maternal heterozygosity in some way reflected the potential for pos‐
itive interactions between maternally inherited cytoplasmic genes
and nuclear genes from a divergent paternal population. We are not
aware of an established mechanism that could account for this pat‐
tern, though it seems plausible that a recent evolutionary history
with a greater diversity of nuclear backgrounds might predispose
the maternal cytotype to have favourable interactions with novel pa‐
ternal genotypes. Intriguingly, crosses between maternal genotypes
from the invaded range in California (USA) and paternal genotypes
from the invasion's origin in Spain showed reduced seed set in a pre‐
vious study (Montesinos, Santiago, & Callaway, 2012). Compared
to other populations across the species range, those from western
Europe and California have some of the lowest levels of genetic
divergence between them, but the highest phenotypic divergence
(Barker et al., 2017), suggesting that adaptation might be driving the
accumulation of negative epistatic interactions in this case.
Negative epistatic interactions are expected to have increasing
fitness costs as populations diverge, ultimately resulting in repro‐
ductive isolation and speciation (Orr & Turelli, 2001). Given multi‐
ple introductions of divergent material, pre‐ or postzygotic isolation
among particularly divergent source populations could prevent
the formation and establishment of admixed genotypes (Rius &
Darling, 2014). The only significant negative interaction among our
FIGURE 4 (a) Growth rates of Centaurea solstitialis progeny from
experimental crosses with all possible combinations of mother and
father population of origin. Colour codes for each cross indicate
maternal and paternal population as in Figure 2, with paternal
region colour outlined by maternal region colour. Growth rates are
least‐squares means ± SEM, and significant differences between
reciprocal crosses of the same parental populations are indicated
*p < 0.05. (b) In a linear model of progeny performance (excluding
heterotic outlier [Asia × southern Greece]), growth rate of genotypes
from the maternal region had a significant negative relationship
(p < 0.0001) with the deviation of progeny from mid‐parent
expectations. Data in (b) show progeny deviation from mid‐parent
as partial residuals after taking into account all other effects in the
model [Colour figure can be viewed at wileyonlinelibrary.com]
Maternal growth rate (In mm/day)
Progeny growth vs. mid-parent
0.01
0.02
–0.01
–0.02
(a)
(b)

|
109
BARKER E t Al.
crosses occurred when eastern European maternal genotypes were
crossed with paternal genotypes originating from an adjacent area in
southern Greece. The population in southern Greece is particularly
divergent from other populations and might belong to a distinct sub‐
species (Barker et al., 2017). The geographic boundary separating
southern Greek and eastern European populations could be a region
in which early speciation dynamics might fruitfully be studied. These
directional negative interactions raise questions about whether
certain lineages are serving as the maternal parents in admixture
events between dif ferentiated source populations across the range
of C. solstitialis.
Epistasis has rarely been discussed as a mechanism of major im‐
portance to admixture's role in invasions, and there have been few
previous tests for epistatic interactions in introduced species. A
small number of empirical studies have found evidence of epistasis
underlying both negative and positive interactions. Keller, Kollmann
and Edwards (2000) crossed native populations of three widespread
agricultural weeds and found evidence of negative epistatic interac‐
tions in either the first‐generation (F1) cross or the backcross (F2) in
each. Johansen‐Morris and Latta (20 06) crossed two invading geno‐
types of Avena barbata in their introduced range and found evidence
for epistasis underlying both hybrid vigour in early generation hy‐
brids (F2) and reduced fitness in later (F6) generations. Notably, later
generations were highly variable and some individual lines showed
potential for outperforming parental genotypes, revealing rare op‐
portunities for beneficial admixture even when there is an overall
pattern of deleterious hybrid breakdown.
Other authors have pointed out that admixed progeny might
also suffer from the loss of local adaptation to the introduced
range, due to introgression of a divergent source that is not as well
adapted to local conditions (Verhoeven et al., 2011). This has been
demonstrated in Mimulus guttatus, where first‐generation crosses
among and between native and invading populations are gener‐
ally heterotic (though with reduced benefits in the F2; Li, Stift,
& Kleunen, 2018; van Kleunen et al., 2015), but crosses between
ranges are maladapted to the invaded environment (Pantoja,
Timothy Paine, & Vallejo‐Marín, 2018). In general, negative epis‐
tasis could contribute to such outbreeding depression, as a result
of divergence due to both local adaptation and genetic drift. In
the absence of negative epistatic interactions, it becomes much
more likely that natural selection will favour admixed progeny that
retain locally adapted alleles, and introgression at some level will
occur. Invaders might also be uniquely poised to avoid problems
with loss of local adaptation from the native range during natu‐
ral admixture events, given that they occupy a novel environment
(Rius & Darling, 2014).
Our survey of the literature revealed that while admix ture is
commonly reported in phylogeographic studies of invaders, it is
disproportionately observed at intermediate values of divergence.
At low values of divergence, admixture might not be particularly
favourable, given that the fitness benefits of introgression in‐
crease with divergence under all proposed underlying mechanisms
(Figure 1a,b). Perhaps more importantly, admixture is also likely to
be underestimated at low levels of divergence due to low power
to identif y multiple source populations (Dlugosch, Anderson, et al.,
2015), and indeed, we found evidence for a variet y of study design
effects that suggest power is an important issue (see Supporting
Information Results S1). In contrast, the decline in observations of
admixture at the highest levels of divergence, where the ability to
detect sources of admixture should be strong, suggests that intro‐
gression between increasingly divergent sources is less likely.
A disproportionate lack of admixture at high levels of diver‐
gence could be due either to poor establishment of admixture
events, as we have argued should be expected under epistasis, or
to a lack of opportunity for admixture events to occur between
highly divergent sources in the first place. How the sources of in‐
troduced species are distributed across the native range, particu‐
larly as a function of the genetic divergence of those sources, has
not been investigated to our knowledge. If introductions tend not
to be sampled from across geographic barriers that are also barriers
to gene flow, this could result in a lack of vectors for admixture at
high genetic divergence among native population, though multiple
introductions from different parts of the native range appear to be
frequent within species introductions (Dlugosch & Parker, 200 8a;
Uller & Leimu, 2011). This possibility could be studied given knowl‐
edge of specific source populations for both failed and successful
multiple introductions, as well as the geographic distribution of
genetic variation within the native range. Such data sets might be
available (or possible to accumulate) for groups with particularly
well‐documented introduction attempts, such as birds (Maitner,
Rudgers, Dunham, & Whitney, 2012).
Finally, we note that the fitness effects of admixture are also
expected to be affected by genetic diversity and genetic load
within the admixing populations. In our experimental crosses, we
found that genetic diversity in the maternal lineage was positively
associated with heterosis, opposite of our predictions, and there
was no effect of diversity in the paternal lineage on performance.
We also found that admixture was not identified more often in the
literature for species with lower levels of heterozygosit y. Several
factors may obscure our ability to interpret whether introduced
species are often in a situation to benefit from an increase in ge‐
netic diversity. Heterozygosity and haplotype diversity are less
sensitive measures of genetic variation than metrics that better
capture rare alleles, such as allelic richness. On the other hand,
heterozygosity should reflect the history of standing variation in a
population (Lohr & Haag, 2015), and rare alleles, by virtue of being
rare, will not strongly influence most of the mechanisms that un‐
derlie the effects of admixture (with the notable exception of evo‐
lutionary rescue).
We also emphasize that we have measured diversity in potential
source populations rather than in founding populations that would
have experienced admixture, but this information is unattainable for
already admixed populations. Previous surveys have indicated that
genetic bottlenecks are not typically severe during founding events,
such that founder population genetic diversity largely resembles
native population diversity, though this is not true for all invaders

110
|
BARKER Et Al.
and expansion front s (Dlugosch, Anderson, et al., 2015; Dlugosch &
Parker, 2008a; Excoffier et al., 2009; Uller & Leimu, 2011). For ex‐
ample, several studies have found evidence consistent with inbreed‐
ing depression in natural invading populations and its rescue by
admixture (Bailey & McCauley, 2006; Keller et al., 2014; Mullarkey,
Byers, & Anderson, 2013; Nolte, Gompert, & Buerkle, 2009; Rius
& Darling, 2014; van Kleunen et al., 2015). It may be that founding
populations with particularly low genetic variation and high genetic
load, which would be predicted to benefit most from admixture, are
relatively rare across species introductions as a whole, although ad‐
mixture could provide substantial benefits to founding populations
when severe genetic bottlenecks do occur, and these benefits may
be realized over evolutionary time, which is not readily observed in
experimental crosses (Szűcs et al., 2014; Szűcs, Melbourne, et al.,
2017; Szűcs, Vahsen, et al., 2017; Wagner et al., 2017).
5 | CONCLUSIONS
Early discussions of admixture in species introductions recog‐
nized its potential to provide genetic and evolutionary rescue,
and favourable genetic interactions (Ellstrand & Schierenbeck,
2000; Novak & Mack, 1993). These benefits were hailed as the
resolution to what has been called the “genetic paradox” of inva‐
sions, wherein invaders are somehow successful despite experi‐
encing founding event s that are expected to have unfavourable
effects on genetic diversity and fitness (Allendorf & Lundquist,
2003; Hufbauer, 2008; Kolbe et al., 2004). Particularly given that
multiple introductions seemed to be common in successful inva‐
sions (Dlugosch, Anderson, et al., 2015), admixture offered the
potential to resolve the genetic paradox for many if not most in‐
vaders. We now know that introduced species do not often suffer
from large reductions in genetic diversity during founding events
(Dlugosch, Anderson, et al., 2015; Dlugosch & Parker, 2008a; Uller
& Leimu, 2011). Admixture will still be beneficial to introduced
species under a variety of mechanisms (Drake, 2006; Ellstrand
& Schierenbeck, 2000; Facon et al., 2008; Frankham, 2005), but
there should be limits to these benefits, par ticularly at high diver‐
gence among source populations due to negative epistasis. This
pattern is a general prediction of epistatic interactions among loci
that have been diverging among populations (C arroll et al., 2003;
Phillips, 2008), though epistasis is rarely mentioned in the context
of invasions (Rius & Darling, 2014). Here, we observed increased
performance of experimentally admixed progeny which diminished
and became negative as divergence between parental populations
increased, consistent with epistasis. Additional experimental in‐
vestigations across many taxa will be essential to understanding
the mechanisms shaping the fitness effects of admixture in natural
systems. Also needed are studies of systems with known introduc‐
tion attempts, wherein it is possible to identify opportunities for
admixture, whether those opportunities have been realized, and
whether the prevalence of admixture is associated with the de‐
gree of fitness benefits. In general, our analyses argue that it may
be possible to resolve conflicting evidence for the benefits of ad‐
mixture during invasion by examining genetic divergence between
and diversity within source populations. If the different outcomes
of admixture are generally predictable, it will be possible to clarif y
whether we should realistically expect that admixture is a general
mechanism for the success of introduced species.
ACKNOWLEDGEMENTS
We thank M.S. Barker and A. Guggisberg for seed collections; K.
Gibson, C. Patterson and S.W. Smith for assistance with plant propa‐
gation and measurements; L.K. Honeker and S. Tran for assistance
with genot yping; A. Zeeger for greenhouse logistics; and reviewers
for helpf ul comments on d rafts of thi s manuscript . This study wa s sup‐
ported by the National Institute of General Medical Sciences of the
National Institutes of Health under Award #K12GM000708 through
the Center for Insect Science at UA to BSB, USDA ELI Fellowship
#2017‐67011‐26034 to JEB, USDA grant #2015‐67013‐23000 and
NSF grant #1750280 to KMD. The content is solely the responsi‐
bility of the authors and does not necessarily represent the official
views of the National Institutes of Health.
DATA ACCESSIBILITY
Original sequence data are available at NCBI sequence read archive
BioProject PRJNA275986.
Script to obtain sequence divergence bet ween regions from
ddRADseq polymorphism data is available at Dryad https://doi.
org/10.5061/dryad.11j4k4v.
Data from experimental crosses are available at Dr yad https://
doi.org/10.5061/dryad.11j4k4v.
Data from the literature survey are available at Dryad https://
doi.org/10.5061/dryad.11j4k4v.
AUTHOR CONTRIBUTIONS
K.M.D. conceived the study. S.R.A., J.E.B., F.A.C., H.D.G. and K.M.D.
performed the literature review. B.S.B. performed the genomic anal‐
yses. J.E.C. performed the experimental crosses. B.S.B. and K.M.D.
analysed the data and wrote the manuscript, which was edited by
all authors.
ORCID
Brittany S. Barker https://orcid.org/0000‐0002‐2198‐8287
Joseph E. Braasch https://orcid.org/0000‐0001‐7502‐1517
Katrina M. Dlugosch https://orcid.org/0000‐0002‐7302‐6637
REFERENCES
Allendorf, F. W., & Lundquist, L. L. (2003). Introduction: Population biol‐
ogy, evolution, and control of invasive species. Conservation Biology,
17, 24–30. https://doi.org/10.1046/j.1523‐1739.2003.02365.x

|
111
BARKER E t Al.
Bailey, M. F., & McCauley, D. E. (2006). The effects of inbreeding, out‐
breeding and long‐distance gene flow on survivorship in North
American populations of Silene vulgaris. Journal of Ecolog y, 94, 98–
109. https://doi.org/10.1111/j.1365‐2745.2005.01090.x
Baker, H. G., & G. L. Stebbins (Eds.) (1965). The genetics of colonizing spe‐
cies. New York, NY: Academic Press.
Ballard, J. W. O., & Melvin, R. G. (2010). Linking the mitochondrial geno‐
type to the organismal phenotype. Molecular Ecology, 19, 1523–1539.
https://doi.org/10.1111/j.1365‐294X.2010.04594.x
Barker, B. S., Andonian, K., Swope, S. M., Luster, D. G., & Dlugosch, K. M.
(2017). Population genomic analyses reveal a histor y of range expan‐
sion and trait evolution across the nati ve and invaded range of yellow
starthistle (Centaurea solstitialis). Molecular Ecology, 26, 1131–1147.
Birchler, J. A., Yao, H., Chudalayandi, S., Vaiman, D., & Veitia, R. A. (2010).
Heterosis. The Plant Cell, 22, 2105–2112. htt ps://d oi.org /10.1105/
tpc.110.076133
Bock, D. G., Andrew, R. L ., & Rieseberg, L. H. (2014). On the adaptive
value of cytoplasmic genomes in plants. Molecular Ecology, 23, 489 9–
4911. https ://doi. org/10.1111/me c.1 292 0
Bock, D. G., Caseys, C., Cousens, R. G., Hahn, M. A., Heredia, S. M.,
Hübner, S., … Rieseberg, L. H. (2015). What we still don’t know about
invasion genetics. Molecular Ecology, 24, 2277–2297. https://doi.
org /10.1111/me c.13 03 2
Bomblies, K., Lempe, J., Epple, P., Warthmann, N., Lanz, C., Dangl, J. L.,
& Weigel, D. (20 07). Autoimmune response as a mechanism for a
Dobzhansky‐Muller‐type incompatibility syndrome in plants. PLoS
Biology, 5, e236. https://doi.org/10.1371/journal.pbio.0050236
Carlson, S. M., Cunningham, C. J., & Westley, P. A. H. (2014). Evolutionar y
rescue in a changing world. Trends in Ecology & Evolution, 29, 521–
530. https://doi.org/10.1016/j.tree.2014.06.0 05
Carroll , S. P., Dingle, H., & Famula, T. R. (2003). Ra pid appearance of epis‐
tasis during adaptive divergence following colonization. Proceedings
of the Royal Society B: Biological Sciences, 270(Suppl), S80–S83.
https://doi.org/10.1098/rsbl.2003.0019
Catchen, J. M., Amores, A., Hohenlohe, P., Cresko, W., & Postlethwait, J.
H. (2011). Stacks: Building and genotyping loci de novo from short‐
read sequences. G3, 1, 17 1–1 8 2 .
Catchen, J., Hohenlohe, P. A., Bassham, S., Amores, A., & Cresko, W. A.
(2013). Stac ks: An analysi s tool set for pop ulation genom ics. Molecular
Ecology, 22, 3124–3140. https://doi.org/10.1111/mec.12354
Chapple, D. G., Miller, K. A., Kraus, F., & Thompson, M. B. (2013).
Divergent introduction histories among invasive populations of t he
delicate skink (Lampropholis delicata): Has the importance of genetic
admixture in the success of biological invasions been overempha‐
sized? Diversity and Distributions, 19, 1 34 –14 6.
Chen, X.‐Y., Wang, X.‐Y., Jiao, J., & Schmid, B. (2015). Complementarity
effec ts do not necessarily result in signif icant transgressive over‐
performance in mixtures. Biological Invasions, 17, 529–535. https://
doi.org /10.10 07/s10530‐014‐0755‐5
Chen, Z. J. (2013). Genomic and epigenetic insights into the molecular
bases of heterosis. Nature Reviews Genetics, 14, 471–482. https://doi.
org/10.1038/nrg3503
Colautti, R. I., & Barrett, S. C. H. (2013). Rapid adaptation to climate fa‐
cilitates range expansion of an invasive plant. Science, 342, 364–366.
https://doi.org/10.1126/science.1242121
Colautti, R. I., & Lau, J. A. (2015). Contemporary evolution during inva‐
sion: Evidence for differentiation, natural selection, and local adap‐
tation. Molecular Ecology, 24, 19 99–2017. https://doi.o rg /10.1111 /
mec .13162
Cox, G. W. (20 04). Alien specie s and evolution: The evolutionar y ecol‐
ogy of exotic p lants, a nimals, microbes, and intera cting native species.
Washington, DC: Island Press.
Crawford, K. M., & Whitney, K. D. (2010). Population genetic diversity
influences colonization success. Molecular Ecology, 19, 1253–1263.
https://doi.org/10.1111/j.1365‐294X.2010.04550.x
Crnokrak, P., & Barrett, S. C. H. (2002). Perspective: Purging the genetic
load: A review of the experimental evidence. Evolution, 56, 2 347–
2358. https://doi.org/10.1111/j.0 014‐3820.2002.tb00160.x
Dlugosch, K. M., Anderson, S. R., Braasch, J., Cang, F. A., & Gillette, H.
D. (2015). The devil is in the details: Genetic variation in introduced
populations and its contributions to invasion. Molecular Ecology, 24,
2095–2111.
Dlugosch, K. M., Cang, F. A., Barker, B. S., Andonian, K., Swope, S. M., &
Rieseberg, L . H. (2015). Evolution of invasiveness through increased
resource use in a vacant niche. Nature Plants, 1, 15066.
Dlugosch, K. M., Lai, Z., Bonin, A ., Hierro, J., & Rieseberg, L. H. (2013).
Allele identification for transcriptome‐based population genomics in
the invasive plant Centaurea solstitialis. G3, 3, 3 59 –3 67.
Dlugosch, K. M., & Parker, I. M. (20 08a). Founding events in species inva‐
sions: Genetic var iation, adaptive evolution, and the role of multiple
introductions. Molecular Ecology, 17, 4 31–449.
Dlugosch, K. M., & Parker, I. M. (2008b). Invading populations of an orna‐
mental shrub show r apid life history evolution despite genetic bot‐
tlenecks. Ecology Letters, 11, 701–709.
Drake, J. M. (2006). Heterosis, the catapult effect and establishment
success of a colonizing bird. Biology Letters, 2, 3 04–307. https://doi.
org/10.1098/rsbl.2006.0459
Dutech, C., BarrÈs, B., Bridier, J., Robin, C., Milgroom, M. G., & Ravigné,
V. (2012). The chestnut blight fungus world tour: Successive in‐
troduc tion events from diverse origins in an invasive plant fun‐
gal pathogen. Molecular Ecology, 21, 3931–3946. ht tps://doi.
org /10.1111/j.1365 ‐294X .2012.0 5575.x
Edmands, S. (1999). Heterosis and outbreeding depression in interpop‐
ulation crosses spanning a wide range of divergence. Evolution, 53,
1757–1768. ht tp s://doi.org/10.1111/j.1558‐564 6.19 99.tb 04560 .x
Ellers, J., Rog, S., Braam, C., & Berg, M. P. (2011). Genotypic richness and
phenotypic dissimilarity enhance population performance. Ecology,
92, 1605–1615. https://doi.org/10.1890/10‐2082.1
Ellis, E. C., Antill, E. C., & Kreft, H. (2012). All is not loss: Plant biodiversity
in the anthropocene. PLoS ONE, 7, e30535. https://doi.org/10.1371/
journal.pone.0030535
Ellstrand, N. C., & Schierenbeck, K. A. (200 0). Hybridization as a stimulus
for the evol ution of invasivene ss in plants? Proceedings of the National
Academy of Sciences, 97, 7043–7050. https://doi.org/10.1073/
pn a s .9 7.1 3 .7 0 4 3
Eriksen, R. L., Desronvil, T., Hierro, J. L ., & Kesseli, R. (2012).
Morphological differentiation in a common garden experiment
among native and non‐native specimens of the invasive weed yellow
starthistle (Centaurea solstitialis). Biological Invasions, 14, 1459–1467.
https://doi.org/10.1007/s10530‐012‐0172‐6
Eriksen, R. L., Hierro, J. L., Eren, Ö., Andonian, K., Török, K., Becerra, P. I.,
… Kesseli, R. (2014). Dispersal pathways and genetic differentiation
among worldwide populations of the invasive weed Centaurea solsti‐
tialis L. (As teraceae). PLoS ONE, 9, e114786. https://doi.org/10.1371/
journal.pone.0114786
Excoffier, L., Foll, M., & Petit, R. J. (2009). Genetic consequences
of range expansions. Annual Review of Ecology, Evolution, a nd
Systematics, 40, 481–501. https://doi.org/10.1146/annurev.
ecolsys.39.110707.173414
Facon, B., Pointier, J.‐P., Jarne, P., Sarda, V., & David, P. (2008). High ge‐
netic variance in life‐history str ategies within inv asive population s by
way of multiple introductions. Current Biolog y, 18, 363–367. https://
doi.org/10.1016/j.cub.2008.01.063
Forsman, A. (2014). Effects of genotypic and phenotypic variation on
establishment are important for conservation, invasion, and infec‐
tion biology. Proceedings of the National Ac ademy of Sciences of the
United States of America, 111, 302–307. https://doi.org/10.1073/
pna s.1317745111
Frankham, R. (2005). Resolving the genetic paradox in invasive species.
Heredity, 94, 385. https://doi.org/10.1038/sj.hdy.6800634

112
|
BARKER Et Al.
Frankham, R., Ballou, J. D., Eldridge, M. D. B., Lacy, R. C., Ralls, K.,
Dudash, M. R., & Fenster, C. B. (2011). Predicting the probability of
outbreeding depression. Conservation Biology, 25, 465–475. https://
doi .or g/10.1111/ j.1523‐1739.2011.016 62.x
Gerlach, J. D. (1997). How the West was lost: Reconstructing the inva‐
sion dynamics of yellow star thistle and other plant invaders of west‐
ern rangelands an d natural areas. Proceedings. California Exotic Pest
Plant Council. Symposium, 3, 67–72.
Holt, R. D., Barfield, M., & Gomulkiewicz, R. (2005). Theories of niche
conser vatism and evolution: Could exotic species be potential tests.
In D. F. Sax, J. J. Stachowicz, & S. D. Gaines (Eds.), Species invasions:
Insights i nto ecology, evolution, and biogeography (pp. 259–290).
Sunderland, MA: Sinauer Associates Inc.
Hovick, S. M., & Whitney, K. D. (2014). Hybridisation is associated with
increased fecundity and size in invasive taxa: Meta‐analytic support
for the hybridisation‐invasion hypothesis. Ecology Letters, 17, 1464–
1477. ht tps://doi.or g/10 .1111/ele.1 2355
Hufbauer, R. A. (2008). Biological invasions: Paradox lost and paradise
gained. Current Biology, 18, R246–R247. https://doi.org/10.1016/j.
cub.2008.01.038
Hufbauer, R. A. (2017). Admixture is a driver rather than a passenger in
experimental invasions. Journal of Animal Ecology, 86, 4–6. https://
doi .or g/10.1111/1365‐2656.12600
Johansen‐Morris, A. D., & Latta, R. G. (2006). Fitness consequences of
hybridization between ecotypes of Avena barbata: Hybrid break‐
down, hybrid vigor, and transgressive segregation. Evolution, 60,
1585–1595. https://doi.org/10.1111/j.0 014‐3820.2006.tb00503.x
Keller, L. F., & Waller, D. M. (2002). Inbreeding effects in wild populations.
Trends in Ecolog y & Evolution, 17, 23 0–241. ht tps ://doi.org/10 .1016/
S0169‐5347(02)0 2489‐8
Keller, M., Kollmann, J., & Edwards, P. J. (2000). Genetic introgres‐
sion from distant provenances reduces fitness in local weed
populations. Journal of Applied Ecology, 37, 647–659. https://doi.
org/10.1046/j.1365‐2664.2000.00517.x
Keller, S. R., Fields, P. D., Berardi, A. E., & Taylor, D. R. (2014). Recent
admixture generates heterozygosity‐fitness correlations during
the range expansion of an invading species. Journal of Evolutionary
Biology, 27, 616–627. https://doi.org/10.1111/jeb.12330
Keller, S. R., & Taylor, D. R. (2010). Genomic admix ture increases f itness
during a biological invasion. Journa l of Evo lutionary Biology, 23, 1720–
1731. https://doi.org/10.1111/j.1420‐9101.2010.02037.x
Kolbe, J. J., Glor, R. E., Rodríguez Schettino, L., Lara, A. C., Larson, A., &
Losos, J. B. (200 4). Genetic variation increases during biological inva‐
sion by a Cuban lizard. Nature, 431, 177–181. ht tp s://doi.or g/10 .1038/
nature02807
Lavergne, S., & Molofsky, J. (2007). Increased genetic variation and evo‐
lutionary potential drive the success of an invasive grass. Proceedings
of the National Academy of Sciences of the United States of Americ a,
104, 3883–3888. htt ps://doi.org /10.1073/p nas.06 07324104
Le Roux, J. J., Blignaut, M., Gildenhuys, E., Mavengere, N., & Berthouly‐
Salazar, C. (2014). The molecular ecology of biological invasions:
What do we kn ow about non‐additive genot ypic effect s and invasion
succes s? Biological Invasions, 16, 9 97–1001. ht tps://doi.org/10.1007/
s10530‐013‐0568‐y
Lee, C. E . (2002). Evolutionary genetics of invasive species. Tr en ds
in Ecology & Evolution, 17, 386–391. https://doi.org/10.1016/
S0169‐5 347( 02) 02554‐5
Lee, C. E ., & Gelembiuk, G . W. (2008). Evolutionary origins of invasive
populations. Evolutionary Applications, 1, 427–448. ht tps://doi.
org /10.1111/j.1752‐4571. 20 08 .0 0 039.x
Li, Y., Stift, M., & van Kleunen, M. (2018). Admixture increases per‐
formance of an invasive plant beyond first‐generation heterosis
(S Bonser, Ed,). Journal of Ecology, 106 , 1595–1606. https://doi.
org /10.1111/1365‐2745.12926
Lohr, J. N., & Haag, C. R. (2015). Genetic load, inbreeding depression,
and hybrid vigor covary with population size: An empirical evalua‐
tion of theoretical predictions. Evolution, 69, 3109–3122 . https://doi.
org /10.1111/evo.128 02
Lombaert, E., Guillemaud, T., Thomas, C. E., Lawson handley, L. J., Li,
J., Wang, S., … Estoup, A. (2011). Inferring the origin of populations
introduced from a genetically structured native range by approx‐
imate Bayesian computation: C ase study of the invasive ladybird
Harmonia axyridis. Molecular Ecology, 20, 4654–4670. https://doi.
org /10.1111/j.1365 ‐294X .2011.05 322 .x
Lynch, M. (1991). The genetic interpretation of inbreeding depression
and outbreeding depression. Evolution, 45, 622–629. https://doi.
org /10.1111/j.1558 ‐56 46 .1991 .tb 0433 3. x
Lynch, M., Conery, J., & Burger, R. (1995). Mutation accumulation and
the extinction of small populations. The American Naturalist, 146,
489–518. https://doi.org /10.108 6/285812
Lynch, M., & Walsh, B. (1998). Genetics and analysis of quantitative traits.
Sunderland, MA: Sinauer Associates Inc.
Maitner, B. S., Rudgers, J. A ., Dunham, A. E., & Whitney, K. D.
(2012). Patterns of bird invasion are consistent with envi‐
ronmental filtering. Ecography, 35, 614–623. https://doi.
org /10.1111/j.160 0‐0587.2011.07176. x
Mesgaran, M. B., Lewis, M. A., Ades, P. K., Donohue, K., Ohadi, S., Li,
C., & Cousens, R. D. (2016). Hybridization can facilitate species in‐
vasions, even without enhancing local adaptation. Proceedings of
the National Academy of Sciences of the United State s of Americ a, 113,
10210–10214. h ttps://doi.org /10.1073/pnas.1605626113
Molofsky, J., Kell er, S. R., La ve rg ne, S., Kaproth, M. A., & Ep pinga, M. B.
(2014). Human‐aided admixture may fuel ecosystem transforma‐
tion during biological invasions: Theoretic al and experimental evi‐
dence. Ecology and Evolution, 4, 899–910 . https://doi.or g/10.1002 /
ece3.966
Montesinos, D., Santiago, G., & Callaway, R. M. (2012). Neo‐allopatr y and
rapid reproductive isolation. The American Natu ralist, 180, 529–533.
htt ps://doi.org /10.1086/667585
Moyle, L. C., & Nakazato, T. (2010). Hybrid incompatibility “snowballs”
between Solanum species. Science, 329, 1521–1523. ht tps://doi.
org/10.1126/science.1193063
Mullarkey, A. A., Byers, D. L., & Ander son, R. C. (2013). Inbreeding de‐
pression and par titioning of genetic load in the invasive biennial
Alliaria petiolata (Brassicaceae). American Journal of Botany, 100,
50 9–518.
Nolte, A. W., Gompert, Z., & Buerkle, C. A . (2009). Variable patterns of
introgression in two sculpin hybrid zones suggest that genomic iso‐
lation differs among populations. Molecular Ecology, 18, 2615–2627.
https://doi. org/10.1111/j .1365‐294X. 20 09.0 4208.x
Novak, S. J., & Mack, R. N. (1993). Genetic variation in Bromus tectorum
(Poaceae): Comparison between native and introduced populations.
Heredity, 71, 167–176 . ht tps ://doi.org/10 .1038/hdy.1993.121
Ochocki, B. M., & Miller, T. E. X. (2017). Rapid evolution of disper‐
sal ability makes biologic al invasions faster and more variable.
Nature Communications, 8, 14315. https: //doi.or g/10.10 38/
ncomms14315
Orr, H. A., & Turelli, M. (2001). The evolution of postzygositc isolation:
Accumulating Dobzhansky‐Muller incompatibilities. Evolution, 55,
1085–1094.
Pantoja, P. O., Timothy Paine, C. E., & Vallejo‐Marín, M. (2018). Natural
selection and outbreeding depression suggest adaptive differentia‐
tion in the invasive range of a clonal plant. Proceedin gs of the Royal
Society B, 285, 20181091.
Peterson, B. K., Weber, J. N., Kay, E. H., Fisher, H. S., & Hoekstra, H. E.
(2012). Double digest RADseq: An inexpensive method for de novo
SNP discovery and genotyping in model and non‐model species. PLoS
ONE, 7, e37135. https://doi.org/10.1371/journal.pone.0037135

|
113
BARKER E t Al.
Phillips, B. L., Brown, G. P., Webb, J. K., & Shine, R. (2006). Invasion
and the evolution of sp eed in toads. Nature, 439, 80 3. https://doi.
org /10.103 8/4398 03a
Phillips , P. C. (2008). Epistasis–the essential role of gene interac tions
in the structure and evolution of genetic systems. Nature Reviews
Genetics, 9, 855–867. https://doi.org/10.1038/nrg2452
Prentis, P. J., Wilson, J. R. U., Dormontt, E. E., Richardson, D. M., & Lowe,
A. J. (2008). Adaptive evolution in invasive species. Trends in Pla nt
Science, 13, 288–294. https://doi.org/10.1016/j.tplants.2008.03.004
Price, M. V., & Waser, N. M. (1979). Pollen dispersal and optimal out‐
crossing in Delphinium nelsoni. Nature, 277, 294–297. https://doi.
org /10.103 8/277294a 0
Reed, D. H., & Frankham, R. (2003). Correlation between fitness and
genetic diversity. Conser vation Biology, 17, 230–237. https://doi.
org /10.1046/j.1523‐1739.2003 .01236. x
Rieseberg, L. H., Kim, S.‐C., Randell, R. A., Whitney, K. D., Gross, B. L.,
Lexer, C., & Clay, K. (2007). Hybridization and the colonization of
novel habitats by annual sunflowers. Genetica, 129, 149–165. https://
doi.org /10.10 07/s10709‐0 06 ‐9011‐y
Rius, M., & Darling, J. A. (2014). How important is intraspecific ge‐
netic admixture to the succe ss of colonising populations? Tr en ds
in Ecology & Evolution, 29, 233–242. https://doi.or g/10.1016/j.
tree.2014.02.003
Sakai, A. K., Allendorf, F. W., Holt, J. S., Lodge, D. M., Molofsky, J., With,
K. A., … Weller, S. G. (2001). The population biology of invasive spe‐
cies. Annual Review of Ecology and Systematics, 32, 305–332. https://
doi.org/10.1146/annurev.ecolsys.32.081501.114037
Szűcs, M., Melbourne, B. A., Tuff, T., & Hufbauer, R. A . (2014). The roles
of demography and genetics in the early stages of colonization.
Proceedi ngs of the Royal Society B: Biologica l Sciences, 281, 20141073.
htt ps://doi.org /10.1098/rsp b.2014.1073
Szűcs, M., Melbourne, B. A., Tuff, T., Weiss‐Lehman, C., & Hufbauer, R.
A. (2017). G enetic and demographic founder ef fect s have long‐term
fitness consequences for colonising populations. Ecology Letters, 20,
436–444.
Szűcs, M., Vahsen, M. L., Melbourne, B. A., Hoover, C., Weiss‐Lehman,
C., & Hufbauer, R . A. (2017). Rapid adaptive evolution in novel en‐
vironments acts as an architect of population range expansion.
Proceedi ngs of the National Academy of Sciences of the United States of
America, 114, 13501–13506.
Tallmon, D. A., Luikar t, G., & Waples, R . S. (2004). The alluring simplicit y
and complex reality of genetic rescue. Trends in Ecolog y & Evolution,
19, 489–496. https://doi.org/10.1016/j.tree.2004.07.003
Thompson, J. N. (1998). Rapid evolution as an ecological process.
Trends in Ecolog y & Evolution, 13, 329–332. htt ps://doi.org/10.1016/
S0169‐5347(98)01378‐0
Turgeon, J., Tayeh, A., Facon, B., Lombaert, E., De clercq, P., Berkvens,
N., … Estoup, A. (2011). Experimental evidence for the pheno‐
typic impact of admixture between wild and biocontrol Asian
ladybird (Harmonia axyridis) involved in the European inva‐
sion. Journal of Evolutionary Biology, 24, 1044–1052. htt ps://doi.
org /10.1111/j.1420 ‐9101 .2 011. 02234 .x
Tutin, T. G., Heywood, V. H., Burges, N. A., Moore, D. M., Valentine,
D. H., Walters, S. M., & Webb, D. A. (Eds.) (2010). Flora Europaea.
Cambridge, UK: Cambridge University Press.
Uller, T., & Leimu, R. (2011). Foun der events predict chan ge s in gene ti c diver‐
sity during human‐mediated range expansions. Global Change Biology,
17, 3478–3 485. https://doi.o rg/10.1111/j.1365‐2486 .2011. 025 09.x
van Kleunen, M., Röckle, M., & Stift , M. (2015). Admixture between na‐
tive and invasive populations may increase invasiveness of Mimulus
guttatus. Proceedings of the Royal Society B: B iological Science s, 282,
20 1 51487.
Vellend, M., Baeten, L., Myers‐Smith, I. H., Elmendorf, S. C., Beausejour,
R., Brown, C. D., … Wipf, S. (2013). Global meta‐analysis reveals no
net change in local‐scale plant biodiversit y over time. Proceedings of
the National Academy of Sciences of the United States of America, 110,
19456–19459. https://doi.org /10.1073/pnas.1312779110
Verhoeven, K. J. F., Macel, M., Wolfe, L. M., & Biere, A. (2011). Population
admixture, biological invasions and the balance bet ween loc al
adaptation and inbreeding depression. Proceedings of the Royal
Society B‐Biological Sciences, 278, 2–8. https://doi.org/10.1098/
rsp b.2 010.1272
Wagner, N. K., Ochocki, B. M., Crawford, K. M., Compagnoni, A., & Miller,
T. E. X. (2017). Genetic mix ture of multiple source populations accel‐
erates invasive range expansion. Journal of Animal Ecology, 86, 21–34.
https://doi. org/10.1111/1365‐265 6.12567
Wang, X. Y., Shen, D. W., Jiao, J., Xu, N. N., Yu, S., Zhou, X . F., … Chen, X.
Y. (2012). Genotypic diversity enhances invasive ability of Spartina
alterniflora. Molecular Ecology, 21, 2542–2551.
Webb, D. M., & Knapp, S. J. (1990). DNA extraction from a previously
recalcitrant plant genus . Plant Molecular Biology Reporter, 8, 180–185.
https://doi.org/10.1007/BF02669514
Weiss‐Lehman, C., Hufbauer, R. A., & Melbourne, B. A. (2017). Rapid
trait evolution drives increased speed and variance in experimen‐
tal range expansions. Nature Communications, 8, 14303. https://doi.
org/10.1038/ncomms14303
Whitney, K. D., & Gering, E. (2015). Five decades of invasion genetics.
New Phytologist, 205, 472– 475. h tt ps: //doi.org /10.1111/nph.13197
Widmer, T. L., Guermache, F., Dolgovskaia, M. Y., & Reznik, S. Y. (2007).
Enhanced growth and seed properties in introduced vs. native pop‐
ulations of yellow starthistle (Centaurea solstitialis). Weed Science, 55,
465–473. https://doi.org/10.1614/WS‐06‐211R.1
Williams, J. L., Kendall, B. E., & Levine, J. M. (2016). Rapid evolution ac‐
celerates plant population spread in fragmented experimental land‐
scapes. Science, 353, 482–485. https://doi.org/10.1126/science.
aaf6268
Wolfe, L. M., Blair, A. C ., & Penna, B. M. (20 07). Does intraspecific hy‐
bridization contribute to the evolution of invasiveness? an experi‐
mental test. Biological Invasions, 9, 51 5–521. http s://d oi.org /10.1007/
s10530‐006‐9046‐0
SUPPORTING INFORMATION
Additional suppor ting information may be found online in the
Suppor ting Information section at the end of the article.
How to cite this article: Barker BS, Cocio JE, Anderson SR, et
al. Potential limits to the benefits of admixture during
biological invasion. Mol Ecol. 2019;28:100–113. h t t p s: //d o i .
org /10.1111/mec.14958