Content uploaded by Phillip C Watts
Author content
All content in this area was uploaded by Phillip C Watts on Aug 18, 2015
Content may be subject to copyright.
1 23
Biological Invasions
ISSN 1387-3547
Volume 12
Number 11
Biol Invasions (2010)
12:3887-3903
DOI 10.1007/
s10530-010-9779-7
Patterns of spatial genetic structure and
diversity at the onset of a rapid range
expansion: colonisation of the UK by the
small red-eyed damselfly Erythromma
viridulum
1 23
Your article is protected by copyright and
all rights are held exclusively by Springer
Science+Business Media B.V.. This e-offprint
is for personal use only and shall not be self-
archived in electronic repositories. If you
wish to self-archive your work, please use the
accepted author’s version for posting to your
own website or your institution’s repository.
You may further deposit the accepted author’s
version on a funder’s repository at a funder’s
request, provided it is not made publicly
available until 12 months after publication.
ORIGINAL PAPER
Patterns of spatial genetic structure and diversity
at the onset of a rapid range expansion: colonisation
of the UK by the small red-eyed damselfly
Erythromma viridulum
Phillip C. Watts •Simon Keat •
David J. Thompson
Received: 19 August 2009 / Accepted: 17 May 2010 / Published online: 28 May 2010
ÓSpringer Science+Business Media B.V. 2010
Abstract Species’ geographic ranges may vary in
size in response to a change in environmental
conditions. The specific genetic consequences of
range expansions are context dependent, largely
depending upon the rate of colonisation as well as
the origins and numbers of founders, and the time
since colonisation. Like other ‘‘charismatic’’ taxa,
such as birds and lepidopterans, the distributions of
odonates (dragonflies and damselflies) are well-
known through substantial monitoring programmes
co-ordinated by various societies. The small red-eyed
damselfly Erythromma viridulum (Odonata: Zygop-
tera) has undergone a substantial, northward range
expansion in Europe in the last 30 years and has
recently-colonised two distinct areas in the UK. We
quantify the immediate genetic consequences of this
rapid colonisation by genotyping more than 1,400
E. viridulum from 39 sites across the northwest
margin of this species’ geographic range. Levels of
genetic diversity and spatial structure are impacted by
this species recent range expansion and non-equilib-
rium conditions that drive weak genetic divergence,
even at regional spatial scales. Populations of
E. viridulum become less diverse towards the edge
of this species’ distribution, presumably as a
consequence of colonisation through a series of
founder events. Specifically, there is a significant
reduction in genetic diversity in the smallest, most
recent focus of colonisation in the UK; however,
there are generally low levels of genetic diversity
across this E. viridulum’s northern range margin.
While most populations are generally poorly differ-
entiated, E. viridulum nonetheless consists of two
distinct lineages that broadly differentiate between
eastern and western Europe. Genetic divergence
between the two UK colonisation foci are indicative
of distinct immigration events from separate sources;
however a general lack of spatial structure prevents
us from pinpointing the specific origins of these
migrant damselflies.
Keywords Dispersal Genetic diversity
Gene flow Odonate
Introduction
Species’ geographic ranges reflect the outcome of
complex interactions between a number of factors,
such as the availability of suitable habitat, dispersal
ability, presence of competitors and natural enemies,
and historical contingency. Thus many species’
distributions are not fixed, but rather vary in response
to temporal fluctuations in such factors. In particular,
climate change and its associated consequences (such
as changes in habitat) has been associated with
P. C. Watts (&)S. Keat D. J. Thompson
School of Biological Sciences, University of Liverpool,
The Biosciences Building, Crown Street, Liverpool
L69 7ZB, UK
e-mail: p.c.watts@liv.ac.uk
123
Biol Invasions (2010) 12:3887–3903
DOI 10.1007/s10530-010-9779-7
Author's personal copy
altered geographic ranges in many insect species
(Parmesan and Yohe 2003). Irrespective of the
specific underlying mechanisms, rapid shifts in
distribution are expected for taxa with good dispersal
capabilities (Pearson and Dawson 2003), with such
changes in distributions likely impacting on current
patterns of biodiversity; for example, taxa that exploit
new environments may reduce the niche space of
native species and/or alter ecosystem function (Sakai
et al. 2001; Gabbard and Fowler 2007). An essential
part of biodiversity conservation, therefore, is to
identify the evolutionary patterns and processes that
are associated with successful colonisation (Pe
´rez
et al. 2006).
Molecular genetic markers are used extensively to
uncover such processes, as the pattern of spatial
genetic structure is determined principally by the
relative influences of migration, selection and genetic
drift (Wright 1931). Colonisation is expected to leave
a distinct genomic signature, with recently-established
populations differing from their source(s) depending
upon factors such as the number of founders, selective
pressures in the new environment and, crucially, the
time since colonisation. Typically, rapid colonisation
by relatively few individuals results in a substantial
loss of genetic diversity (reviewed by Puillandre et al.
2008) and weak or atypical patterns of spatial genetic
structure as a consequence of insufficient time to attain
migration-drift equilibrium (Slatkin 1993).
Odonates (Anisoptera and Zygoptera; dragonflies
and damselflies respectively) are an important com-
ponent of freshwater and terrestrial ecosystems and
are key bioindicator species. Many species are strong
fliers, capable of wide dispersal and like other insect
taxa (Hill et al. 1999; Parmesan and Yohe 2003;
Hickling et al. 2006) many odonate species are
altering their distributions, apparently in response to
climate change (Aoki 1997; Hassall et al. 2007).
British odonates are no exception, with 37 species
having shifted their range margin northward over the
last 40 years (Hickling et al. 2005). One consequence
of these directional changes in distributions is the
colonisation of the British Isles by two species of
formerly continental European odonates in the last
decade; the southern emerald damselfly Lestes barb-
arus (Nobes 2003) and the small red-eyed damselfly
Erythromma viridulum (Dewick and Gerussi 2000;
Cham 2002). Subsequent range expansion out of
these colonisation foci by L. barbarus has remained
limited, while, by contrast, E. viridulum has spread
extensively inland at a rapid rate.
Erythromma viridulum is a thermophilic, holomed-
iterranean species that has undergone significant
northward range expansion in Europe in the last
30 years (Ketelaar 2002), apparently exploiting the
Rhine valley as principal route of range expansion
through northern Germany and the Netherlands whilst
a slower sweep of range expansion has occurred across
central Europe into northwest-France. This species
was first recorded at coastal sites in the UK in 1999
(Dewick and Gerussi 2000) with continuing waves of
immigration from the continent occurring in subse-
quent years (Cham 2002). From these eastern UK
2004
2002
2004
2004
2006
eastern
UK
BRN
CRT
BVS,
BLT
LMD,
MGG
KDK, UAD
WEV, WGV
HET
2002
2002
2006
southern
UK
EAR
SSF
CAR
LWS
THM
TLS
FGH
SCV
COL
SVE
ML2, ML3
BFP SCP
PCP
WRP
BUR
BLH
PPP CLP
MRH
MOR
HOL
EVC
SFR
MAR
STL
Fig. 1 Locations of samples of the small red eyed damselfly,
Erythromma viridulum, collected from the UK and continental
Europe. Contour lines on map of UK indicate the limit of
E. viridulum’s UK range during 2002, 2004 and 2006 (S. Keat,
unpublished); the shaded area on the map of northern Europe
shows the current distribution of E. viridulum (adapted from
Askew 1988)
3888 P. C. Watts et al.
123
Author's personal copy
coastal sites, E. viridulum moved inland in a north-
westerly direction (Fig. 1), and the establishment of
breeding populations was confirmed by the presence
of larvae by 2003 (Cham 2004). A second colonisation
focus was identified in 2000 in the Isle of Wight,
southern UK (Cham 2001) that has spread northwards
subsequently. In contrast to the expansion inland from
eastern coastal sites the second colonisation has a
more restricted distribution (Fig. 1). The rapid and
recent range expansion of E. viridulum presents an
unprecedented opportunity to quantify the immediate,
within four generations, genetic consequences of
colonisation at the margin of a rapidly-expanding
population. We expect UK populations of E. viridulum
to display a genetic signature of population bottle-
necks, generally low levels of genetic diversity
relative to putative source populations from continen-
tal Europe and weak spatial genetic structure, which
are characteristic footprints of recent gene flow and
non-equilibrium conditions.
Here, we quantify the patterns of genetic variation
and spatial genetic structure of the small red-eyed
damselfly, Erythromma viridulum (Odonata: Zygop-
tera) across its range in the British Isles and in
potential source populations in northwest continental
Europe. The main aims of this paper are to (1)
contrast levels of genetic variation in UK and
continental European populations and (2) quantify
the spatial genetic signature of recent colonisation.
Materials and methods
Sampling, DNA extraction and genotyping
Adult E. viridulum were sampled between 2002 and
2006 from 28 sites in southern UK and 11 sites (from
Germany, Netherlands, Belgium and France) in north-
west continental Europe (Fig. 1, Table 1). DNA was
extracted using a high salt protocol (Aljanabi and
Martinez 1997) from either a single tibia (samples from
2002 to 2004) or thoracic muscle (samples from 2005
to 2006). All samples were genotyped at 10 microsat-
ellite loci (Keat et al. 2005). Approximately 5 ng of
DNA was used in a 10 ll PCR containing 75 mMTris–
HCl pH 8.9, 20 mM(NH
4
)
2
SO
4
, 0.01% Tween-20,
0.2 mMeach dNTP, 3.0 mMMgCl
2
, 2 pmol of each
primer, 0.25 UTaq polymerase (ABgene) and pub-
lished thermal cycling (Keat et al. 2005). PCR products
were pooled into one of two genotyping panels
(depending on allelic size range and the 50fluorescent
dye—either 6-FAM,NED,PET or VIC) along with
GENESCAN-500 LIZ size standard (Applied Biosystems)
and separated using capillary electrophoresis through a
denaturing polymer (POP4) on an ABI3100 (Applied
Biosystems). Allele sizes were determined using the
cubic model in GENEMAPPER v.3.0 software (Applied
Biosystems).
Analysis of genetic diversity
Tests for linkage disequilibrium between all locus-
pair combinations were carried out using GENEPOP
v.3.1d (Raymond and Rousset 1995) (Markov chain
parameters were 1,000 dememorisation, 100 batches
and 1,000 iterations per batch). Deviations from
Hardy–Weinberg Equilibrium (HWE) conditions
were quantified using FSTAT v.2.9.3 (Goudet 1995)
by making 2,000 permutations of alleles among
individuals within samples. FSTAT was used to calcu-
late allelic richness (A
R
) standardised to 18 individ-
uals, expected heterozygosity (H
e
) and Wright’s
(1951) inbreeding coefficient (f) within each sample
that contained 19 or more individuals (Table 1)to
avoid small sample size bias. Standard errors about
these genetic diversity parameters were calculated by
jackknifing over loci (Sokal and Rohlf 1995). The
permutation procedure implemented by FSTAT (2,000
permutations of samples among groups) was used to
test whether A
R
,H
e
, and fdiffered between three
groups of samples: southern UK, eastern UK and
continental Europe (Fig. 1). Because the smallest and
most recent colonisation occurred in southern UK,
and both UK regions were colonised more recently
than continental Europe, we used a one-sided test
under the hypotheses that (1) samples from southern
UK are less diverse than those from eastern UK and
(2) samples from both regions in UK are less diverse
than those from continental Europe.
Temporal genetic variation
Some samples from within the UK were collected
over several years (2002–2006, see Table 1). Anal-
ysis of molecular variance (AMOVA) was used to
partition the contribution to genetic diversity arising
from spatial variation with that occurring among
these successive sampling periods. Only large
Genetic consequences of range expansion 3889
123
Author's personal copy
Table 1 Summary of sample sizes and collection dates of E. viridulum sites
Country Site name Code Year of sampling Total
‘02 ‘03 ‘04 ‘05 ‘06
UK (eastern sites) Bedfords Park BFP 7 7
London Wetland Centre LWC 9 9
Shenfield Common pond SCP 3 3
Silver End SVE 10 10
Maldon – 2 ML2 10 59 44 113
Maldon – 3 ML3 3 3
Salcott-cum-Virley SCV 53 54 107
Colchester COL 13 13
Fingringhoe FGH 9 9
Thorpe-Le-Soken TLS 17 17
Thorpeness Meare THM 3 3
Lowestoft LWS 9 17 58 52 136
Strumpshaw Fen SSF 4 4
Carr House CAR 10 10
East Ruston EAR 21 50 33 104
Priory Country Park PCP 1 27 68 54 150
Wrest Park WRP 46 75 121
UK (southern sites) Beaulieu Heath BLH 4 4
Bursledon BUR 38 38
Stag Lane STL 4 45 49
Marvel Farm MAR 9 31 37 77
Stone Farm Reservoir SFR 7 7
Hollier Farm HOL 11 11
East View Cottage EVC 33 33
Parsonage Peat Pond PPP 10 30 39 79
Morton Pond MOR 27 32 39 98
Marsh House MRH 10 1 11
Carpenters Lane Pond CLP 6 6
70 25 153 594 390 1,232
Belgium Zoutleeuw, Het Vinne HET 10 10
Wijvenheide Eutroof Ven WEV 19 19
Wijvenheide Grote Vijver WGV 5 5
France Blangy-Tronville BLT 25 25
Boves BVS 6 6
Courtemont Varennes CRT 10 10
Le Marais de Guines: 1 LMD 3 3
Le Marais de Guines: 2 MGG 7 7
Germany Braunschweig BRN 53 53
Netherlands Koudekerke KDK 19 19
Alphen Ann Den Rijn UAD 19 19 38
72 123 195
Samples with C19 individuals are highlighted bold as these sites only were used for population-based analyses (see ‘‘Materials and
methods’’)
3890 P. C. Watts et al.
123
Author's personal copy
samples (nC19) were used for this analysis and thus
the groups of samples consisted of: 2004 (PCP,
MAR, MOR and EAR), 2005 (PCP, WRP, ML2,
SCV, BUR, EVC, MAR, MOR, PPP, STL, EAR and
LWS) and 2006 (PCP, WRP, ML2, SCV, MOR, PPP,
EAR and LWS). The significance of the fixation
indices were tested using the permutation procedure
(10,000 permutations) implemented by ARLEQUIN
v.2.000 (Schneider et al. 2000).
Spatial genetic structure
Except where stated, genotype data collected from
different dates but within a site were pooled (see
Results of AMOVA below) and only larger samples
(nC19) were used to quantify inter-site differentia-
tion to limit any effect of sample size. Allele frequency
differences among samples were quantified using the
exact test employed by FSTAT v.2.9.3 (Goudet 1995);
HWE within samples was not assumed and genotypes
were permuted 2,000 times among samples. FSTAT was
used also to calculate F
ST
(Weir and Cockerham 1984)
between samples, with the significance of estimates of
F
ST
from zero assessed through 2,000 permutations of
genotypes between populations. The pattern of genetic
relationships among populations was summarised by
multidimensional scaling analysis of the values of
pairwise F
ST
using SPSS v.13.0 (SPSS Inc. Chicago).
Isolation by distance (IBD) was assessed by spatial
autocorrelation (Vekemans and Hardy 2004). We used
SPAGEDI v.1.2 (Hardy and Vekemans 2002) to calculate
the correlation in average kinship (F
ij
; Loiselle et al.
1995) relative to the whole data set between pairs of
E. viridulum separated (linear distances) by a range of
increasing spatial scales; Loiselle et al.’s (1995)F
ij
was used as it is more powerful at detecting spatial
genetic structure than other estimators (Vekemans and
Hardy 2004); all genotype data were included for this
analysis. To avoid a bias in the correlation coefficient
arising because of unequal sample sizes within each
spatial category, SPAGEDI was used to assign distance
categories that contained a similar number of pairwise
comparisons (Hardy and Vekemans 2002). Ninety-
five percent confidence intervals for F
ij
at each
distance class were generated from the distribution
of 2,000 permutations of spatial group locations
among the spatial groups. The pattern of spatial
genetic structure was examined for (1) the entire UK
data set and (2) groups of samples within southern UK,
eastern UK and from continental Europe separately.
Spatial genetic structure was assessed for the entire
data set using the model-based clustering approach
implemented by STRUCTURE v.2.0 (Pritchard et al.
2000), which simultaneously identifies populations
(clusters) and assigns individuals to populations using
a Bayesian approach. Briefly, STRUCTURE models K
populations that are characterised by a set of allele
frequencies at each locus and probabilistically assigns
individuals to populations on the basis of their
multilocus genotypes. The most likely number of
populations may be estimated from the value of Kthat
maximises the posterior probability of the data for a
given posterior probability distribution Pr(K|X) that is
calculated from the posterior distribution of Pr(X|K)(X
is the multilocus genotypes of sampled individuals); in
STRUCTURE output, this criterion ‘Ln P(D)’ is the
average of the log likelihood of the data at each step of
the Markov Chain Monte Carlo (MCMC) and then
half their variance is subtracted from the mean.
STRUCTURE also calculates the proportion of member-
ship (Q) of each individual in each cluster. The true
number of populations is identified by the model of K
that returns the maximal value of Ln P(D), although if
there is weak spatial structure this may not be obvious.
Under such circumstances the value of Kat the
beginning of a ‘plateau’ of estimates of Ln P(D) can be
selected, i.e. the smallest value of Kthat captures the
major structure of the data set (Pritchard and Wen
2003). Alternatively one can calculate an ad hoc
measure, DK, which is the second order rate of change
of Ln P(D) with respect to K; the modal value of DK
corresponds to the most pronounced partition of the
data set (Evanno et al. 2005). Five independent runs of
STRUCTURE were made, to assess output consistency
and to calculate DK, for values of Kfrom 1 up to 18
using the admixture model and correlated allele
frequencies. All model runs were based on 500,000
iterations after an initial 10% burn-in period (i.e.
50,000 iterations), which was sufficient to ensure
convergence of the MCMC (see Pritchard et al. 2000;
Pritchard and Wen 2003 for details).
Finally, we used the Bayesian model developed by
Ciofi et al. (1999; software 2MOD available at http://
www.rubic.rdg.ac.uk/*mab/software.html) to deter-
mine the relative likelihoods of (1) population diver-
gence occurring by genetic drift (drift model) against
an alternative extreme (2) of the spatial distribution of
Genetic consequences of range expansion 3891
123
Author's personal copy
allele frequencies being determined by a balance
between migration and drift (gene flow model).
Analyses were made on four groups of populations
using samples that contained 19 or more individuals:
(1) eastern UK 2005, (2) eastern UK 2006, (3) south-
ern UK 2005 and (4) continental Europe 2006. We ran
the Markov Chain Monte Carlo (MCMC) searches
twice (on different PCs) for 200,000 iterations to
ensure that the posterior probabilities had converged.
The first 10% of the runs were then discarded to avoid
the effects of initial run parameters when estimating
the probabilities of each model.
Results
Genetic diversity
One locus pair (out of 45 comparisons) demonstrated
significant linkage disequilibrium in the total sample
(LIST14-021 and LIST14-042, P\0.001) but not for
comparisons within populations, so all loci were
assumed to be unlinked. Most samples met expected
HWE conditions with only 26 out of the 620 (*4%)
possible sample-locus combinations having a signif-
icant heterozygote deficit after correction for multiple
tests (k=10 tests per sample); these deficits were not
associated with any particular locus or sample (data
not shown) and all loci were retained for subsequent
analyses.
Genetic diversity was moderate to low with up to 8
alleles per sample (at LIST14-002). The least diverse
locus, LIST14-021, was monomorphic in many sam-
ples (9 sites, data pooled over all years) and a second
locus, LIST14-040, was also weakly polymorphic,
with no more than 3 alleles per sample (data not
shown). Average A
R
is generally lowest in southern
UK sites, intermediate in the eastern UK and greatest
in continental Europe, notably the German sample
BRN (Fig. 2a); there is less variation in average H
e
among samples, but the pattern described for A
R
is
evident nonetheless (Fig. 2b). Wright’s (1951)f,by
contrast, is variable and mostly positive for all samples
from the UK and generally close to zero for samples
from continental Europe, but otherwise displays no
consistent differences between regions (Fig. 2c).
Hence, it is not surprising that A
R
and H
e
was
significantly lower in southern UK compared with
the eastern UK (P=0.027 and 0.003 respectively) but
not for f(P=0.399). Also, there was a significant
reduction in A
R
and H
e
(P\0.001 and =0.028
respectively) in the southern UK compared with
continental Europe, but not between the latter region
and eastern UK (Table 2). For all significant tests only
comparisons between eastern UK and southern UK
(H
e
) and between the latter and continental Europe
(A
R
) remain significant after a sequential Bonferroni
correction (Rice 1989) (Table 2).
Temporal genetic variation
No significant genetic differences were attributed to
variation among the temporal groups of samples
(P=0.254), with just 0.2% of the total genetic
variance attributable to temporal genetic variation.
Approximately 6.5% of the total genetic variance
(F
ST
=0.066, P\0.001) was attributed to differ-
ences among spatial groups within each sample
period, and the majority of genetic variation occurred
within populations (Table 3). The absence of tempo-
ral variation among samples justifies pooling samples
from different years at the same site for subsequent
analyses of spatial structure.
Spatial genetic structure
There is a clear pattern of significantly different
(P\0.05, k=153) genetic divergence (genotypic
differences and pairwise F
ST
) between samples from
southern England and those from elsewhere, but not
between samples within southern England (Table 4).
Average pairwise F
ST
was greatest for comparisons
between samples from southern England and those
from eastern England and continental Europe (F
ST
=
0.066 and 0.064, respectively), intermediate for
comparisons between the eastern England (EED)
and continental Europe (CEE) (EED–CEE F
ST
=
0.0187; CEE–CEE F
ST
=0.0169; EED–EED F
ST
=
0.0157) and low among populations from the south-
ern England only (F
ST
=0.004). Given the clear
distinction between the three regions, it is not
surprising that the multidimensional scaling analysis
of pairwise F
ST
generates two clusters: (1) a tight
group of samples comprising samples from southern
England and (2) a loose cluster that contains the
samples from eastern England and those from
continental Europe (Fig. 3).
3892 P. C. Watts et al.
123
Author's personal copy
Average kinship (F
ij
) among UK E. viridulum
generally decreases with increasing distance classes
such that pairs of individuals separated by *150 km
or less have significantly (P\0.05) positive values
of F
ij
while average F
ij
among pairs of individuals
separated by greater distances is significantly (P\
0.05) less than zero (Fig. 4). This decline in F
ij
does
not demonstrate a classic IBD pattern (i.e. a gradual
decline in relatedness), but rather the sharp reduction
in F
ij
is indicative of genetic divergence between
the southern an eastern UK. Indeed, geographic
separation has no significant effect upon values of F
ij
when the autocorrelation is limited to pairs of
individuals within a particular region, even among
samples from continental Europe (Fig. 5).
Independent runs of STRUCTURE generated similar
clustering solutions up to values of K=11 (data not
shown). Values of L(K) were broadly similar for model
runs of kbetween 1 and 6 but with peaks at K=2 and
K=5 (Fig. 6a). There was a clear modal value of DK
at K=2 (Fig. 6b), supporting a bipartition of northern
European E. viridulum that broadly correlates with
0.20
0.30
0.40
0.50
0.60
2.00
3.00
4.00
-0.20
0.00
0.20
0.40
Expected heterozygosity (He)Allelic richness (AR)
Inbreeding (f)
(a)
(b)
(c)
BUR
STL
MAR
EVC
PPP
MOR
ML2
SCV
LWS
EAR
PCP
WRP
BRN
KDK
UAD
BEV
BLT
BUR
STL
MAR
EVC
PPP
MOR
ML2
SCV
LWS
EAR
PCP
WRP
BRN
KDK
UAD
BEV
BLT
BUR
STL
MAR
EVC
PPP
MOR
ML2
SCV
LWS
EAR
PCP
WRP
BRN
KDK
UAD
BEV
BLT
southern UK eastern UK continental
Europe
Fig. 2 Geographic variation in genetic diversity (mean ±
standard error) for populations of the small red-eyed damselfly,
E. viridulum, from northern Europe; aallelic richness (A
R
),
bexpected heterozygosity (H
e
) and cinbreeding coefficient (f)
(Wright 1951). Multiple values at a location represent samples
collected over several years (see Table 1)
Genetic consequences of range expansion 3893
123
Author's personal copy
geographic location: cluster 1 broadly comprises
samples from the southern UK and cluster 2 generally
consists of individuals from eastern UK and continen-
tal Europe (Fig. 6a, Table 5). Average membership
coefficients for cluster 1 are generally high ([0.6) for
most of the southern UK samples (with the exception of
two small samples). Similarly, average membership
coefficients for the eastern UK and continental Euro-
pean samples are mostly high for cluster 2, however a
greater percentage of populations, in particular three of
the French samples, have roughly equal probabilities of
membership to either simulated cluster (Table 5).
The gene flow model had substantially more
support than the model of divergence by drift for
all comparisons of groups of samples within the
eastern and southern UK (P[gene flow] [0.98,
Bayes factor [50 for all model runs). By contrast,
separate runs of 2MOD did not provide strong support
for either model among populations of E. viridulum
from continental Europe (Table 6).
Discussion
By genotyping E. viridulum from 39 sites at the
northern edge of this species’ geographic range
we have identified a reduction in genetic diversity
in the smallest, most recent focus of colonisation in
the UK and generally low levels of genetic diversity
throughout this species’ northern range margin.
Moreover, at the onset of this species’ colonisation
in the UK E. viridulum is characterised by weak
spatial genetic structure, consistent with its dispersal
capability, but nonetheless consists of two distinct,
eastern and western, genetic units.
Levels of genetic diversity
An east–west gradient in genetic diversity points to a
loss of variation in more peripheral locations and may
be explained by range expansion through colonisation
of new areas via sequential founder populations
(Ramachandran et al. 2005; Herborg et al. 2007). A
reduction in genetic variability through founder
effects and bottlenecks is often a feature of invasive
populations (Tsutsui et al. 2000; Grapputo et al. 2005;
Herborg et al. 2007; Schmid-Hempel et al. 2007;
Dlugosch and Parker 2008; Puillandre et al. 2008;
Ahern et al. 2009). Specifically, because rare alleles
are lost faster than gene diversity, a bottlenecked
population typically has an excess of heterozygotes
Table 2 Average values of allelic richness (A
R
), expected
heterozygosity (H
e
) and Wright’s (1951) inbreeding coefficient
(f) for samples of E. viridulum from three geographic regions:
eastern UK, southern UK and continental Europe
Population group Measure of genetic variability
A
R
H
e
f
Eastern UK 2.930 0.432 0.131
Southern UK 2.728 0.389 0.111
P0.027 0.003 0.399
Eastern UK 2.930 0.432 0.131
Continental Europe 3.048 0.429 0.065
P0.845 0.456 0.166
Southern UK 2.728 0.389 0.111
Continental Europe 3.048 0.429 0.065
P0.001 0.028 0.520
P, 1-sided probability of genetic diversity parameter differing
between groups of samples under the hypotheses that (1)
samples from southern UK are less diverse than those from
eastern UK and (2) that all samples from UK are less diverse
than samples from continental Europe. Comparisons that
remain significant after sequential Bonferroni correction
(Rice 1989)(P\0.05, k=4) are highlighted bold
Table 3 Analysis of molecular variance (AMOVA) of samples of E. viridulum from the UK
Source of variation df SS Variance
components
% Variance Fixation
index
P
Among temporal groups 2 31.6 0.005 0.21 0.0021 0.254
Among populations within
temporal groups
21 297.2 0.139 6.55 0.0656 0.000
Within populations 2066 4,098.5 1.984 93.24 0.0676 0.000
Total 2089 4,427.3 2.128 100.00
Variation is quantified among three temporal groups of samples (collected during 2004, 2005 and 2006), among samples within each
temporal group and within populations. P, probability of a more extreme variance component than that observed
3894 P. C. Watts et al.
123
Author's personal copy
Table 4 Pairwise estimates of genotypic differentiation (above diagonal) and F
ST
(below diagonal) between samples of E. viridulum from the UK and continental Europe
Eastern UK Southern UK Continental Europe
ML SCV LWC LWS EAR PCP WRP BUR STL MAR EVC PPP MOR BRN KDK UAD WEV BLT
ML 0.2007 0.0029* 0.0010* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0510 0.0003* 0.0026* 0.1863
SCV 0.0041* 0.0121* 0.0516 0.0003* 0.0003* 0.0624 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0007* 0.1033 0.0003* 0.0029* 0.8131
LWC 0.0036 0.0030 0.0016* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0232* 0.0003* 0.0010* 0.2219
LWS 0.0043* 0.0009 0.0134* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0007* 0.0356* 0.0003* 0.0095* 0.1775
EAR 0.0233* 0.0198* 0.0228* 0.0132* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0078* 0.0003* 0.0003* 0.0062*
PCP 0.0366* 0.0302* 0.0325* 0.0235* 0.0038* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003*
WRP 0.0118* 0.0036* 0.0117* 0.0084* 0.0255* 0.0337* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0003* 0.0013* 0.0003* 0.1422
BUR 0.0498* 0.0554* 0.0617* 0.0485* 0.0588* 0.0804* 0.0699* 0.0049* 0.0007* 0.0085 0.0092* 0.0043* 0.0003* 0.0003* 0.0003* 0.0007* 0.0003*
STL 0.0538* 0.0543* 0.0613* 0.0455* 0.0533* 0.0722* 0.0675* -0.0049 0.4366 0.1039 0.3118 0.5578 0.0003* 0.0003* 0.0003* 0.0003* 0.0003*
MAR 0.0684* 0.0639* 0.0813* 0.0533* 0.0659* 0.0838* 0.0681* 0.0058 -0.0010 0.2709 0.0088* 0.3418 0.0003* 0.0003* 0.0003* 0.0003* 0.0003*
EVC 0.0600* 0.0632* 0.0799* 0.0487* 0.0656* 0.0770* 0.0715* 0.0075 0.0055 0.0036 0.0673 0.1850 0.0003* 0.0003* 0.0003* 0.0003* 0.0003*
PPP 0.0657* 0.0642* 0.0723* 0.0545* 0.0668* 0.0888* 0.0797* 0.0007 -0.0023 0.0049 0.0140* 0.3797 0.0003* 0.0003* 0.0003* 0.0003* 0.0003*
MOR 0.0668* 0.0671* 0.0754* 0.0568* 0.0693* 0.0867* 0.0739* 0.0034 -0.0022 -0.0004 0.0063 0.0030 0.0003* 0.0003* 0.0003* 0.0003* 0.0003*
BRN 0.0138* 0.0096* 0.0204* 0.0089* 0.0260* 0.0282* 0.0199* 0.0447* 0.0419* 0.0515* 0.0368* 0.0556* 0.0557* 0.0118* 0.0003* 0.0010* 0.0317*
KDK 0.0122* 0.0131 0.0147 0.0122* 0.0278* 0.0445* 0.0426* 0.0549* 0.0565* 0.0823* 0.0765* 0.0573* 0.0790* 0.0181* 0.0003* 0.7157 0.3389
UAD 0.0161* 0.0069 0.0113 0.0109* 0.0128* 0.0232* 0.0083* 0.0617* 0.0549* 0.0637* 0.0681* 0.0683* 0.0651* 0.0213* 0.0292* 0.0003* 0.0768
WEV 0.0171* 0.0188* 0.0317* 0.0131* 0.0331* 0.0481* 0.0377* 0.0486* 0.0533* 0.0658* 0.0703* 0.0519* 0.0657* 0.0362* 0.0002 0.0323* 0.2392
BLT 0.0068 -0.0028 -0.0034 0.0064 0.0207* 0.0289* 0.0086 0.0743* 0.0690* 0.0839* 0.0891* 0.0769* 0.0841* 0.0165* 0.0023 0.0031 0.0095
* Indicates significant difference between population pairs (P\0.05); values that remain significant (P\0.05, k=153) after sequential Bonferroni correction for multiple tests (Rice 1989) are
highlighted bold
Genetic consequences of range expansion 3895
123
Author's personal copy
compared with that expected at equilibrium condi-
tions (Cornuet and Luikart 1996). Despite a greater
reduction in A
R
than H
e
(cf. Fig. 2a, b) there was no
evidence for a bottleneck in any sample (P[0.05;
analysis using BOTTLENECK v.1.2.02; Piry et al. 1999—
Appendix Table 7). Although this genetic feature is
transient (persisting for just a few generations after the
demographic reduction), any effect should be detect-
able in our samples as potentially they were collected
between 2 and 4 generations (this species is predom-
inantly semivoltine in the UK) after E. viridulum
established itself in the UK. Alternatively, E. viridu-
lum may have occupied sites in the UK for some time
prior to the 1999 (first) records, but given the large
effort directed towards monitoring odonates in the UK
(see http://www.dragonflysoc.org.uk/index.html) this
seems unlikely. The lack of an apparent bottleneck
effect likely reflects successful colonisation by large
numbers of migrants that reached UK coastal sites
(Cham 2002; Parr 2005). Indeed, many studies of
range expansions/invasive species have uncovered
relatively high levels of genetic diversity in intro-
duced populations, sometimes even exceeding that of
the source(s), which is a putative outcome of mul-
tiple introductions and/or large numbers of colonis-
ers (Johnson and Starks 2004; Kolbe et al. 2004;
Chen et al. 2006; Dlugosch and Parker 2008). The
slight difference in genetic diversity between the
eastern UK and northwest continental Europe is a
likely effect of (1) repeated waves of migrants
reaching the UK and/or (2) a persistent genetic
signature of the relatively recent expansion into
northern continental Europe itself. By contrast, the
significant reduction in genetic diversity in popula-
tions in southern England is indicative of more
limited immigration. However, as the southern UK
may have been colonised from unsampled sites in
western France (see discussion below) it is possible
that the source population(s) per se lacks diversity
relative to those further east.
A comparison of levels of gene diversity in various
odonate species indicates that these populations of
E. viridulum have low levels of genetic diversity
(Watts 2009). One corollary is that this is a persistent
signature of a rapid range expansion, although geno-
type data of E. viridulum from southern locations are
required to determine the full extent and pattern of
genetic erosion associated with this species’ range
expansion. Interestingly, the damselfly Coenagrion
mercuriale also reaches the northern limit of its
geographic range in the UK and presents a sharp
demographic contrast with E. viridulum as the former
is a poor disperser that has a contracting and highly
-0.5
0.0
0.5
-1.0 -0.5 0.0 0.5 1.0
Dimension 1
Dimension 2
Fig. 3 Multidimensional scaling plot of values of pairwise F
ST
between samples of E. viridulum from (1) southern UK (black),
(2) eastern UK (grey) and (3) continental Europe (white)
g
eo
g
ra
p
hic distance (km)
Kinship (Fij)
-0.04
-0.02
0.00
0.02
0.04
0 50 100 150 200 250 300
* * * * * * * *
Fig. 4 Correlogram profiles of the variation in kinship (F
ij
)
(Loiselle et al. 1995) as a function of the average distance
separating pairs of E. viridulum from locations in the UK. 95%
confidence intervals are twice the standard error, obtained by
jackknifing over 10 microsatellite loci. * Indicates value of F
ij
that is significantly (P\0.05) different from zero as assessed
by permutation procedure
-0.02
-0.01
0.00
0.01
0.02
01234567
g
eo
g
ra
p
hic distance (ln km)
Kinship (Fij)
Fig. 5 Spatial variation in average kinship (F
ij
) (Loiselle et al.
1995) among pairs of E. viridulum from eastern UK (grey),
southern UK (black) and continental Europe (white)
3896 P. C. Watts et al.
123
Author's personal copy
fragmented distribution in the UK. With the notable
exception of an extremely isolated population that is
monomorphic at many microsatellite loci (Watts et al.
2006), levels of genetic diversity are similar, and often
greater, in UK populations of C. mercuriale compared
with E. viridulum (Watts et al. 2004,2005,2007).
Together these studies demonstrate that considerably
different demographic processes may generate
comparable levels of genetic diversity.
The role of genetic diversity per se in determining
colonisation success remains equivocal (Pe
´rez et al.
2006; Dlugosch and Parker 2008). Certainly loss of
genetic diversity can be associated with reduced
fitness (Saccheri et al. 1998) yet under some circum-
stances a population may benefit by passing through a
bottleneck (Tsutsui et al. 2000; Schmid-Hempel et al.
2007). Like many successful colonisers, a relative
genetic diversity has not prevented an extensive
range expansion by E. viridulum. This may imply
that the viability of new colonies, at least during the
short-term, is determined more by demographic
processes, such as large founding populations and
repeated immigration, than genetic factors (but see
also Pe
´rez et al. 2006). Certainly multiple invasions
are a common, but not ubiquitous (Grapputo et al.
2005), theme in invasion biology (Kolbe et al. 2004).
Thus, it is interesting E. viridulum migrations tend to
take place towards the end of the flying season when
population densities, and numbers of potential
migrants, are greatest (Cham 2002). Following the
future success of E. viridulum populations should
provide new insights into the interplay between
genetic diversity and adaptive potential as this
species is challenged by a range of selective pressures
posed in novel environments (Ahern et al. 2009).
Spatial genetic structure
F-statistics cannot be interpreted as direct indicator of
the absolute level of gene where populations have not
0
10
20
30
40
50
1 3 5 7 9 1113151719
-28,000
-26,000
-24,000
-22,000
1 3 5 7 9 1113151719
L(K)
(a)
(b)
K
ΔK [L(K)]
K
HOL
EVC
SFR
MOR
MAR
MRH
PPP
CLP
STL
BLH
BUR
BFP MLH
MLT
TLS FGH
COL
SCV
SVE SCP
LWC
LWS
THM
EAR
CAR SSF
PCP
WRP
BRN
KDK
UAD WEV
WGV HET
CRT BVS
BLT LMD
MGG
Southern UK Eastern UK Continental
Europe
INSET:
Fig. 6 Probabilities of individual membership of E. viridulum
from northern Europe to hypothetical clusters in a two cluster
simulation. Each bar represents an individual and the
proportion of the bar that is black or white represents the
proportion of assignment to cluster 1 or 2 respectively. Inset:
(a) Mean L(K)(±95% CI) over 5 independent runs for each K
value. (b)DK, the second order rate of change of Ln P(D) with
respect to K; the modal value of this distribution corresponds to
the true value of Kor the uppermost level of genetic structure
(see Methods and Evanno et al. 2005 for further details)
Genetic consequences of range expansion 3897
123
Author's personal copy
reached genetic equilibrium conditions (Whitlock and
McCauley 1999), which, given this species’ recent
range expansion, seems likely for E. viridulum across
much of northern Europe. Nevertheless, typically low
values of F
ST
are consistent with observations that
E. viridulum is capable of relatively long distance
dispersal over inhospitable habitat; for example,
some 100 km of sea separate the southern UK sites
from the French coast, and this species has expanded
its range throughout a wide area of UK landscape by
an average rate of 28 km per year (S. Cham, personal
communication, see also Fig. 1). Thus it is not
surprising that a model of gene flow in the UK is
supported unequivocally. Moreover, our data indicate
that migrants reached the eastern UK from localities
somewhere between northern Germany and the
northeast coast of France (Tables 4,5; Fig. 6a). It is
possible that E. viridulum arrived at the eastern UK
from various sites across northwest Europe, as the
eastern and continental samples form a somewhat
loose genetic cluster (Fig. 3); however an alternative
scenario is that genetic non-equilibrium conditions
and persistent weak spatial structure in northwest
continental Europe prevents a specific source(s) from
being located. Assuming that E. viridulum exploits
prevailing winds for dispersal then the sources may
be similar to that used by other insects that have
colonised the UK recently (e.g. Culicoides sp. Gloster
et al. 2007). Despite an increased effect of drift
(Table 6), relatively low genetic divergence (average
F
ST
=0.017) between distant continental sites
(Table 4; Fig. 5) is probably a persistent footprint
of relatively recent immigration (rather than contin-
ued high gene flow) into northern Europe;
E. viridulum colonised the Netherlands and Belgium
during the 1970s and 1990s, respectively (De Kniff
et al.2001; Ketelaar 2002). Rapid colonisation by
Table 5 Average probability of membership for populations
of Erythromma viridulum from the UK and continental Europe
to one of two simulated clusters derived using STRUCTURE
Region Population Probability of assignment
Cluster 1 Cluster 2
Southern UK HOL 0.790 0.210
EVC 0.798 0.202
SFR 0.579 0.421
MOR 0.785 0.215
MAR 0.792 0.208
MRH 0.683 0.317
PPP 0.756 0.244
CLP 0.695 0.305
STL 0.757 0.243
BLH 0.503 0.497
BUR 0.707 0.293
Eastern UK BFP 0.629 0.371
MLH 0.299 0.701
MLT 0.375 0.625
TLS 0.302 0.698
FGH 0.378 0.622
COL 0.390 0.610
SCV 0.376 0.624
SVE 0.289 0.711
SCP 0.547 0.453
LWC 0.484 0.516
LWS 0.404 0.596
THM 0.180 0.820
EAR 0.363 0.637
CAR 0.823 0.177
SSF 0.366 0.634
PCP 0.316 0.684
WRP 0.329 0.671
Germany BRN 0.479 0.521
Holland KDK 0.353 0.647
UAD 0.332 0.668
Belgium WEV 0.423 0.577
WGV 0.488 0.512
HET 0.262 0.738
France CRT 0.253 0.747
BVS 0.597 0.403
BLT 0.365 0.635
LMD 0.584 0.416
MGG 0.450 0.550
Populations with average membership [0.6 to a particular
cluster are highlighted bold
Table 6 Probabilities of two alternate (gene flow versus drift)
models of population structure in samples of the small-red eyed
damselfly E. viridulum from three regions in northwest Europe:
EUK, eastern UK; SUK, southern UK; CEU, northwest
continental Europe (see Fig. 1for details of sample locations)
EUK
(2005)
EUK
(2006)
SUK
(2005)
CEU
(2006)
Gene flow 0.989 0.981 0.999 0.583
Drift 0.011 0.019 0.001 0.417
3898 P. C. Watts et al.
123
Author's personal copy
relatively few long distance migrants can create a
‘leading edge’ that is comprised of a relatively
homogenous genome, particularly where expansion
occurs via serial founder events. Under such condi-
tions, once the available niche space has become fully
occupied, successive waves of immigration may fail
to have a significant genomic impact enabling weak
spatial structure to persist for some time (Hewitt
2000). Indeed, there was no strong temporal compo-
nent to the genetic variation among our samples.
Significant genetic differences between the south-
ern UK and continental Europe indicates that the
source of colonisation into the former area lies further
west of our sample range, most likely in the Cherbourg
Peninsula as this is the nearest French coastal area to
the Isle of Wight; certainly the greater degree of
clustering and generally lower values of F
ST
among
the southern UK samples points to a more specific
origin of migrants reaching this area. Despite these
differences and the convincing partition of the dataset
(Fig. 6b) however, the average probability of mem-
bership of UK samples to either of the two model
clusters was rarely definitive, particularly for conti-
nental samples (Table 5). It is possible that genetic
drift among founder populations (Grapputo et al.
2005; Herborg et al. 2007) may have driven diver-
gence between eastern and southern UK sites, espe-
cially as the colonisation into the latter area appears to
be more limited. Additional sampling to the west and
south is required to deconfound the contributions of
(1) source genomic structure and (2) founder effects to
the present pattern of spatial genetic structure dis-
played by E. viridulum. Similarly, it is not possible to
determine whether the apparently anomalous genetic
signature in the eastern UK sample CAR, which is one
of the earliest founder populations, reflects genetic
drift or a divergent source population.
Where dispersal is spatially restricted, IBD genetic
structure should be present at migration-drift equi-
librium (Rousset 1997). Conversely, failure to detect
IBD can be indicative of non-equilibrium conditions,
for example due to a recent change in migration
pattern or a recent range expansion (Slatkin 1993).
Clearly the outcome of (historical and contemporary)
extensive gene flow (cf. Table 6) is that populations
of E. viridulum are characterised by quite weak
values of pairwise F
ST
in continental Europe, eastern
UK and southern UK (average F
ST
=0.0167, 0.0157
and 0.0036 respectively; Table 4); in addition, this
species lacks IBD genetic structure over a wide range
of spatial scales, an effect observed in other recently-
established populations (Tsutsui and Case 2001;
Genton et al. 2005; Chen et al. 2006). A pattern
approaching that expected under IBD within the UK
is driven by the genetic divergence between eastern
and western UK and which leads to a sharp increase
in the level of genetic differentiation among samples
separated by approximately 150 km (Fig. 4). In a
similar vein, other studies of introduced species have
reported significant genetic divergence, and some-
times an apparent IBD structure, that may reflect
divergence between populations established from
multiple sources (Goodisman et al. 2001; Shoemaker
et al. 2006; Herborg et al. 2007).
Compared with anispopterans (dragonflies) which
are typically powerful fliers, many species of zyg-
opteran (damselflies) are relatively sedentary (Utzeri
et al. 1984; Michiels and Dhondt 1991; Watts et al.
2004) and, accordingly, such species tend to exhibit
quite strong spatial genetic structure (Andre
´s et al.
2002; Svensson et al. 2004; Watts et al. 2004,2005,
2006; Chaput-Bardy et al. 2008). There has been little
time for these northern populations to diverge
following the initial phase of expansion and coloni-
sation, but two future scenarios are plausible. First,
E. viridulum may become less vagile after this phase
of colonisation, leading to greater spatial structure
that will develop at a rate determined by the relative
impacts of some long-distance dispersal, the ability of
immigrants to impact on established genomic struc-
ture, genetic drift and selective pressures. Alterna-
tively, E. viridulum may maintain weak spatial
structure by continuing to disperse widely, similar
to metapopulations of other damselflies such as
Ischnura pumilio or Lestes barbarus that inhabit
early successional or temporary water bodies. Cer-
tainly the contrast between the UK and northwest
continental Europe in the demographic model that is
favoured (Table 6) may reflect an increased effect of
drift and approach to genetic equilibrium continents
on the continent, although the extent to which this is
influenced by the greater spatial or temporal scale is
unclear. An indication of near-equilibrium levels of
spatial genetic structure could be made by sampling
populations of E. viridulum toward the centre of this
species’ distribution.
To conclude, this study has quantified the spatial
genetic structure of a species undergoing a range
Genetic consequences of range expansion 3899
123
Author's personal copy
expansion. Both levels of genetic diversity and spatial
structure are impacted by high dispersal ability and
non-equilibrium conditions that particularly drive a
lack of genetic divergence even at regional spatial
scales. Genetic differences between the two UK
colonisation foci are indicative of distinct colonisation
events from separate sources, though we are unable to
pinpoint the origins of migrant damselflies and cannot
exclude the possibility that divergence is a conse-
quence of genetic drift. Populations of E. viridulum
become less diverse towards the northern edge of this
species’ range margin, presumably as a consequence
of colonisation by a series of founder events.
Acknowledgments We thank the following people for their
help collecting samples: Lorna Bousfield, Caroline Daguet,
Dave Dana, Laurent Gavory, John Horne, Samantha Jacobs,
Sebastien Legris, Paul Richardson, Chris Slatcher, Robby
Stoks, Frank Su
¨hling, Ce
´dric Vanappelghem, Peter De Knijff,
Anthony Nixon, Errol Newman and Arjen Van’t Hof. The
work was funded by a NERC studentship (NER/S/A/2003/
11287) to Simon Keat and by the Environment Agency. We
thank Tim Sykes for his support throughout and Steve Cham
for assistance in finding suitable sites.
Appendix
See Table 7.
Table 7 Probability values for tests for a significant hetero-
zygote excess indicative of a population bottleneck for samples
of Erythromma viridulum using three models of microsatellite
allele mutation (IAM, infinite allele model; TPM, two phase
model; SMM, stepwise model) and two methods of analysis
(Sign test and Wilcoxon sign-rank test)
Region Site Year Sign test Wilcoxon test
IAM TPM SMM IAM TPM SMM
Isle of Wight
and Hampshire
EVC 2005 0.1700 0.2106 0.2196 0.0156* 0.0234* 0.2188
MOR 2004 0.2798 0.3278 0.6050 0.3262 0.4102 0.8203
MOR 2005 0.2732 0.3828 0.0375* 0.3672 0.6328 0.8750
MOR 2006 0.3820 0.4872 0.2373 0.3125 0.5391 0.8623
MAR 2004 0.3557 0.4204 0.4635 0.0967 0.2783 0.3848
MAR 2005 0.2455 0.3197 0.3657 0.1250 0.3672 0.4551
PPP 2005 0.2675 0.5921 0.3577 0.1504 0.5898 0.8496
PPP 2006 0.0934 0.5717 0.3594 0.0645 0.4102 0.8496
STL 2005 0.4058 0.5184 0.4012 0.2783 0.6523 0.7539
BUR 2005 0.0874 0.3434 0.3662 0.0117* 0.0391 0.4063
Essex ML2 2005 0.1701 0.0903 0.0789 0.0371* 0.7695 0.9863
ML2 2006 0.1075 0.1760 0.1696 0.1016 0.6328 0.9863
SCV 2005 0.2044 0.2589 0.0117* 0.0977 0.3203 0.9805
SCV 2006 0.2889 0.3976 0.0493* 0.0645 0.5000 0.9180
East Anglia
and Bedfordshire
LWS 2005 0.0410* 0.2065 0.0770 0.0059* 0.1563 0.9023
LWS 2006 0.3971 0.5309 0.0931 0.1875 0.7217 0.9346
EAR 2004 0.2066 0.2501 0.2054 0.0977 0.3203 0.8750
EAR 2005 0.0242* 0.1464 0.0406* 0.0645 0.0820 0.9932
EAR 2006 0.0476* 0.0698 0.1980 0.0195* 0.0273* 0.6289
PCP 2004 0.0233* 0.1352 0.3654 0.0049* 0.0244* 0.6328
PCP 2005 0.1634 0.5419 0.4898 0.0967 0.3477 0.8838
PCP 2006 0.0156* 0.2941 0.4279 0.0068* 0.0820 0.6328
WRP 2005 0.0504 0.2385 0.5466 0.0137* 0.1563 0.5781
WRP 2006 0.0888 0.3065 0.1679 0.0645 0.2481 0.8984
Germany BRN 2005 0.6065 0.1135 0.0274* 0.5000 0.9180 0.9971
Belgium HET 2006 0.3082 0.6267 0.2976 0.1504 0.5000 0.8750
WEV 2006 0.4507 0.0997 0.0762 0.5391 0.9033 0.9951
3900 P. C. Watts et al.
123
Author's personal copy
References
Ahern RG, Hawthorne DJ, Raupp MJ (2009) Founder effects
and phenotypic variation in Adelges cooleyi, an insect pest
introduced to the eastern United States. Biol Invasions
11:959–971
Aljanabi SM, Martinez I (1997) Universal and rapid salt-
extraction of high quality genomic DNA for PCR-based
techniques. Nucleic Acids Res 22:4692–4693
Andre
´s RA, Sanchez-Guille
¨n RA, Cordero-Rivera A (2002)
Evolution of female color polymorphism in damselflies:
testing the hypotheses. Anim Behav 63:677–685
Aoki T (1997) Northward expansion of Ictinogomphus pertinax
(Selys) in eastern Shikoku and western Kinki districts, Japan
(Anisoptera: Gomphidae). Odonatologica 26:121–133
Askew R (1988) Dragonflies of Europe. Harley Books, Martins,
UK
Cham S (2001) The status of small red-eyed damselfly Erythromma
viridulum in the British Isles. Atropos 12:7–9
Cham S (2002) Range expansion of the small red-eyed dam-
selfly Erythromma viridulum (Charp.) in the British Isles.
Atropos 15:3–9
Cham S (2004) Observations on an inland population of the
small red-eyed damselfly Erythromma viridulum (Char-
pentier) with notes on the discovery of the first larvae in
Britain. J Br Dragonfly Soc 20:31–34
Chaput-Bardy A, Lemaire C, Picard D, Secondi J (2008)
Spatial genetic structure in a large river network. Mol
Ecol 17:3496–3505
Chen YH, Opp SB, Berlocher SH, Roderick GK (2006) Are
bottlenecks associated with colonization? Genetic diver-
sity and diapauses variation of native and introduced
Rhagoletis completa populations. Oecologia 149:656–667
Ciofi C, Beaumont MA, Swingland IR, Bruford MW (1999)
Genetic divergence and units for conservation in the
Komodo dragon Varanus komodoensis. Proc R Soc Lond
B266:2269–2274
Cornuet JM, Luikart G (1996) Description and power analysis
of two tests for detecting recent population bottlenecks
from allele frequency data. Genetics 144:2001–2014
De Kniff G, Anselin A, Goffart P (2001) Trends in dragonfly
occurrence in Belgium (Odonata). Proceedings of the 13th
international colloquium European invertebrate survey,
Leiden
Dewick S, Gerussi R (2000) Small red-eyed damselfly
Erythromma viridulum (Charpentier) found breeding in
Essex—the first British Records. Atropos 9:3–4
Dlugosch KM, Parker IM (2008) Founding events in species
invasions: genetic variation, adaptive evolution, and the
role of multiple introductions. Mol Ecol 17:431–449
Evanno G, Regnaut S, Goudet J (2005) Detecting the number
of clusters of individuals using the software structure: a
simulation study. Mol Ecol 14:2611–2620
Gabbard B, Fowler N (2007) Wide ecological amplitude of
a diversity-reducing invasive grass. Biol Invasions 9:
149–160
Genton BJ, Shykoff JA, Giraud T (2005) High genetic diversity
in French invasive populations of common ragweed,
Ambrosia artemisiifolia, as a result of multiple sources of
introduction. Mol Ecol 14:4275–4285
Gloster J, Burgin L, Witham C, Athanassiadou M, Mellor PS
(2007) Bluetongue in the United Kingdom and northern
Europe in 2007 and key issues for 2008. Vet Rec 162:
298–302
Goodisman MAD, Matthews RW, Crozier RH (2001) Hierar-
chical genetic structure of the introduced wasp Vespula
germanica in Australia. Mol Ecol 10:1423–1432
Goudet J (1995) FSTAT (vers. 1.2): a computer program to
calculate F-statistics. J Hered 86:485–486
Grapputo A, Boman S, Lindstro
¨m L, Lyytinen A, Mappes J
(2005) The voyage of an invasive species across continents:
genetic diversity of North American and EuropeanColorado
potato beetle populations. Mol Ecol 14:4207–4219
Hardy OJ, Vekemans X (2002) SPAGeDi: a versatile computer
program to analyse spatial genetic structure at the indi-
vidual or population levels. Mol Ecol Notes 2:618–620
Hassall C, Thompson DJ, French GC, Harvey IF (2007) His-
torical changes in the phenology of British Odonata are
related to climate. Glob Change Biol 13:933–941
Herborg L-M, Weetman D, van Oosterhout C, Hanfling B
(2007) Genetic population structure and contemporary
dispersal patterns of a recent invader, the Chinese mitten
crab, Eriocheir sinensis. Mol Ecol 16:231–242
Hewitt GM (2000) The genetic legacy of the Quaternary ice
ages. Nature 405:907–913
Hickling R, Roy DB, Hill JK, Thomas CD (2005) A northward
shift of range margins in British Odonata. Glob Change
Biol 11:502–506
Table 7 continued
Region Site Year Sign test Wilcoxon test
IAM TPM SMM IAM TPM SMM
Holland KDK 2006 0.3108 0.3415 0.1450 0.1504 0.4551 0.8203
UAD 2005 0.1028 0.1188 0.4124 0.1016 0.2129 0.5449
UAD 2006 0.0044* 0.0484* 0.2193 0.0020* 0.0039* 0.0371*
France CRT 2006 0.1684 0.2112 0.2728 0.0977 0.2734 0.3203
BLT 2006 0.0669 0.2627 0.2933 0.0244* 0.0820 0.7871
* Indicates a significant (P\0.05) heterozygote excess, however none of the tests remain significant (P\0.05) after a sequential
Bonferroni correction (Rice 1989) for k=32 multiple tests
Genetic consequences of range expansion 3901
123
Author's personal copy
Hickling R, Roy DB, Hill JK, Fox R, Thomas CD (2006) The
distributions of a wide range of taxonomic groups are
expanding polewards. Glob Change Biol 12:1–6
Hill JK, Thomas CD, Huntley B (1999) Climate and habitat
availability determine 20th century changes in a butter-
fly’s range margin. Proc R Soc Lond B266:1197–1206
Johnson RN, Starks PT (2004) A surprising level of genetic
diversity in an invasive wasp: Polistes dominulus in the
northeastern United States. Ann Entomol Soc Am 97:732–737
Keat S, Thompson DJ, Kemp SJ, Watts PC (2005) Ten micro-
satellite loci for the small red-eyed damselfly Erythromma
viridulum (Charpentier). Mol Ecol Notes 5:788–790
Ketelaar R (2002) The recent expansion of the Small Red-eyed
Damselfly Erythromma viridulum (Charpentier) in The
Netherlands. J Br Dragonfly Soc 18:1–9
Kolbe JJ, Glor RE, Schettino LRG, Lara AC, Larson A, Losos
JB (2004) Genetic variation increases during biological
invasion by a Cuban lizard. Nature 431:177–181
Loiselle BA, Sork VL, Nason J, Graham C (1995) Spatial
genetic structure of a tropical understorey shrub, Psy-
chotria officinalis (Rubiaceae). Am J Bot 82:1420–1425
Michiels NK, Dhondt AA (1991) Characteristics of dispersal in
sexually mature dragonfiies. Ecol Entomol 16:449–459
Nobes G (2003) Southern emerald damselfly Lestes barbarus
(Fabr.)-the first British record. Atropos 18:3–6
Parmesan C, Yohe G (2003) A globally coherent fingerprint of
climate change impacts across natural systems. Nature
421: 37–42
Parr AJ (2005) Migrant and dispersive dragonflies in Britain
during 2004. J Br Dragonfly Soc 21:14–20
Pearson RG, Dawson TP (2003) Predicting the impacts of
climate change on the distribution of species: are biocli-
mate envelope models useful? Glob Ecol Biogeogr 12:
361–371
Pe
´rez JE, Nirchio M, Alfonsi C, Mun
˜oz C (2006) The biology
of invasions: the genetic adaptation paradox. Biol
Invasions 8:1115–1121
Piry S, Luikart G, Cornuet J-M (1999) BOTTLENECK. A program
for detecting recent effective population size reductions from
allele frequency data. http://www.ensam.inra.fr/URLB
Pritchard JK, Wen W (2003) Documentation for structure
software: version 2. http://pritch.bsd.uchicago.edu
Pritchard JK, Stephens M, Donnelly P (2000) Inference of popu-
lation structure using multilocus genotype data. Genetics 155:
945–959
Puillandre N, Dupas S, Dangles O, Zeddam J-L, Capdevi-
elle-Dulac C, Barbin K, Torres-Leguizamon M, Silvain
J-F (2008) Genetic bottleneck in invasive species: the
potato tuber moth adds to the list. Biol Invasions 10:
319–333
Ramachandran S, Deshpande O, Roseman CC, Rosenberg NA,
Feldman MW, Cavalli-Sforza LL (2005) Support from the
relationship of genetic and geographic distance in human
populations for a serial founder effect originating in Africa.
Proc Natl Acad Sci USA 102:15942–15947
Raymond M, Rousset F (1995) GENEPOP (version 1.2):
population genetics software for exact tests and ecumen-
icism. J Hered 86:248–249
Rice WR (1989) Analyzing tables of statistical tests. Evolution
43:223–225
Rousset F (1997) Genetic differentiation and estimation of
gene flow from F-statistics under isolation by distance.
Genetics 145:1219–1228
Saccheri I, Kuussaari M, Kankare M, Vikman P, Fortelius W,
Hanski I (1998) Inbreeding and extinction in a butterfly
metapopulation. Nature 392:491–494
Sakai AK, Allendorf FW, Holt JS, Lodge DM, Molofsky J,
With KA, Baughman S, Cabin RJ, Cohen JE, Ellstrand
NC, McCauley DE, O’Neil P, Parker IM, Thompson JN,
Weller SG (2001) The population biology of invasive
species. Annu Rev Ecol Syst 32:305–332
Schmid-Hempel P, Schmid-Hempel R, Brunner PC, Seeman
OD, Allen GR (2007) Invasion success of the bumblebee,
Bombus terrestris, despite a drastic genetic bottleneck.
Heredity 99:414–422
Schneider S, Roessli D, Excoffier C (2000) Arlequin version
2000, genetics and biometry. University of Geneva,
Switzerland
Shoemaker DD, DeHeer CJ, Krieger MJB, Ross KG (2006)
Population genetics of the invasive fire ant Solenopsis
invicta (Hymenoptera: Formicidae) in the United States.
Ann Entomol Soc Am 99:1213–1233
Slatkin M (1993) Isolation by distance in equilibrium and
non-equilibrium populations. Evolution 47:264–279
Sokal RR, Rohlf (1995) Biometry. WH Freeman and Co, New York
Svensson EI, Kristoffersen L, Oskarsson K, Bensch S (2004)
Molecular population divergence and sexual selection on
morphology in the banded demoiselle (Calopteryx splendens).
Heredity 93:423–433
Tsutsui ND, Case TJ (2001) Population genetics and colony
structure of the argentine ant (Linepithema humile) in its
native and introduced ranges. Evolution 55:976–985
Tsutsui ND, Suarez AV, Holway DA, Case TJ (2000) Reduced
genetic variation and the success of an invasive species.
Proc Natl Acad Sci USA 97:5948–5953
Utzeri C, Carchini G, Falchetti E, Belfiore C (1984) Philopatry,
homing and dispersal in Lestes barbarus (Fabricius)
(Zygoptera: Lestidae). Odonatologica 13:573–584
Vekemans X, Hardy OJ (2004) New insights from fine-scale
spatial genetic structure analyses in plant populations.
Mol Ecol 13:921–935
Watts PC (2009) Characteristics of microsatellite loci in
Odonata. Int J Odonatol 12:275–286
Watts PC, Rouquette JR, Saccheri IJ, Kemp SJ, Thompson DJ
(2004) Molecular and ecological evidence for small-scale
isolation by distance in an endangered damselfly,
Coenagrion mercuriale. Mol Ecol 13:2931–2945
Watts PC, Kemp SJ, Saccheri IJ, Thompson DJ (2005)
Conservation implications of genetic variation between
spatially and temporally distinct colonies of the damselfly
Coenagrion mercuriale. Ecol Entomol 30:541–547
Watts PC, Saccheri IJ, Kemp SJ, Thompson DJ (2006) Popu-
lation structure and the impact of regional and local
habitat isolation upon levels of genetic diversity of the
endangered damselfly Coenagrion mercuriale (Odonata:
Zygoptera). Freshw Biol 51:193–205
Watts PC, Rousset F, Saccheri IJ, Leblois R, Kemp SJ,
Thompson DJ (2007) Compatible genetic and ecological
estimates of dispersal rates in insect (Coenagrion mer-
curiale: Odonata: Zygoptera) populations: analysis of
3902 P. C. Watts et al.
123
Author's personal copy
‘neighbourhood size’ using a more precise estimator. Mol
Ecol 16:737–751
Weir BS, Cockerham CC (1984) Estimating F-statistics for the
analysis of population structure. Evolution 38:1358–1370
Whitlock MC, McCauley DE (1999) Indirect measures of gene
flow and migration: F
ST
=1/(4Nm ?1). Heredity 82:
117–125
Wright S (1931) Evolution in mendelian populations. Genetics
16:97–159
Wright S (1951) The genetical structure of populations. Ann
Eugen 15:323–354
Genetic consequences of range expansion 3903
123
Author's personal copy
