Population structure and the impact of regional and local habitat isolation upon levels of genetic...

Population structure and the impact of regional and localhabitat isolation upon levels of genetic diversity of theendangered damselfly Coenagrion mercuriale (Odonata:Zygoptera)

PHILLIP C. WATTS, ILIK J . SACCHERI, STEPHEN J. KEMP AND DAVID J. THOMPSONSchool of Biological Sciences, Liverpool University, Crown Street, Liverpool, L69 7ZB, U.K.

SUMMARY

1. Coenagrion mercuriale is one of Europe’s most threatened damselflies. There is concern forthe long-term persistence of many of its U.K. colonies because adult lifetime movement islimited, making isolated populations susceptible to extinction.2. Using 14 microsatellite loci we characterised levels of genetic diversity, evidence for arecent decline and the spatial genetic structure for C. mercuriale population in Wales, U.K.3. Spatial isolation is not an absolute predictor of low genetic diversity at either local orregional scales.4. One population inhabiting a remote, edge of range site is genetically impoverished withlevels of variability (at microsatellite loci) among the lowest reported for any insect species.5. Agricultural land and high ground are physical barriers to dispersal by adults.6. Consistent with work from elsewhere, movement by mature C. mercuriale inPembrokeshire is sufficient to prevent significant genetic differentiation throughout ahabitat matrix of some 3–4 km if the suitable habitat sites are <2 km apart and lack barriersto movement. Even within a good habitat matrix, however, genetic isolation by distancedevelops within 10 km.

Keywords: biodiversity, bottleneck, conservation, dispersal, population structure

Introduction

Destruction and fragmentation of habitat are major

causes of biodiversity loss [World Conservation

Monitoring Centre (WCMC), 1992]. Organisms sensi-

tive to the effects of habitat fragmentation are likely to

have some combination of low natural abundance/

high area requirement, large population fluctuations,

low intrinsic growth rate, specialised habitat require-

ments and/or poor dispersal capability (Henle et al.,

2004). Particularly by impacting upon the capability of

individuals to disperse, habitat fragmentation can

lead to a reduction in population size that will

heighten extinction risk by increasing vulnerability

to stochastic demographic changes or reducing gen-

etic diversity (Frankham, 1995; Frankham, Ballou &

Briscoe, 2002). Maintaining genetic diversity is im-

portant as it is associated with population viability

(Saccheri et al., 1998; Madsen et al., 1999; Spielman

et al., 2004a; Spielman, Brook & Frankham, 2004b) and

the evolutionary potential of a species to respond to

environmental change (Frankham et al., 2002; Reed &

Frankham, 2003; Schmitt & Hewitt, 2004). European

wetlands, which are inherently discontinuous, are

becoming increasingly patchy, largely because of

pollution, eutrophication or changes in land use. A

decline in movement among habitat patches may be

expected for many freshwater species, with the

expected consequences of genetic erosion and popula-

tion decline.

Correspondence: Phill Watts, Marine and Freshwater Biology

Research Group, The Biosciences Building, School of Biological

Sciences, Liverpool University, Crown Street, Liverpool,

L69 7ZB, U.K.

E-mail: [email protected]

Freshwater Biology (2006) 51, 193–205 doi:10.1111/j.1365-2427.2005.01478.x

! 2005 The Authors, Journal compilation ! 2005 Blackwell Publishing Ltd 193

As successful conservation of insect taxa is largely

achieved by maintaining areas of suitable habitat it is

imperative that species’ dispersal capabilities are

characterised so that an appropriate level of connec-

tivity within a habitat network can be established in

order to ameliorate the detrimental effects of frag-

mentation. Characterising dispersal can be problem-

atic for freshwater insects as many can disperse as

aquatic larvae and/or terrestrial adults. Nevertheless,

a number of studies have directly tracked dispersal of

aquatic insects (e.g. Miller, Blinn & Keim, 2002; Elliott,

2003; Petersen et al., 2004; Watts et al., 2004a). Such

direct observations, however, are likely to overlook

rare, long-distance movement events and, import-

antly, fail to inform on the successful transfer of

gametes that affect levels of diversity within popula-

tions (Slatkin, 1985). Many studies, therefore, have

used the spatial distribution of neutral genetic mark-

ers to make indirect appraisals of the dispersal

characteristics of a variety of freshwater insect taxa

(e.g. Geenen et al., 2000; Smith & Collier, 2001; Kelly,

Rundle & Bilton, 2002; Miller et al., 2002; Hughes,

Hillyer & Bunn, 2003a; Hughes et al., 2003b; Wilcock,

Nichols & Hildrew, 2003; Wishart & Hughes, 2003;

Watts et al., 2004a).

Odonates (damselflies and dragonflies) are a key

component of many freshwater ecosystems that are

generally perceived to be good fliers, potentially

capable of wide dispersal and possibly less sensitive

to habitat fragmentation than other freshwater taxa.

Numerous odonate species are declining, however,

with nearly 40% of indigenous European odonates

classified as endangered, vulnerable or rare under

IUCN red book categories (Van Tol & Verdonk, 1988).

The axiom of wide dispersal may hold for many

anisopterans (Corbet, 1999; Freeland et al., 2003) but

zygopterans are weaker fliers, with many species not

dispersing beyond several kilometres (Conrad et al.,

1999; Geenen et al., 2000; Purse et al., 2003; Watts et al.,

2004a). Thus, it is somewhat surprising that genetic

differentiation has been reported among zygopteran

populations separated by up to 100 km (Andres,

Sanchez-Guill _en & Cordero Rivera, 2000, 2002) but

not shorter distances (Geenen et al., 2000; Wong,

Smith & Forbes, 2003; but cf. Watts et al., 2004a).

This paper presents an analysis of genetic variab-

ility and population structure of an endangered

damselfly, Coenagrion mercuriale (Charpentier, 1840)

(Odonata: Zygoptera). Coenagrion mercuriale, restricted

to the south and west of Europe, is considered extinct

or on the edge of extinction in seven European

countries (Grand, 1996) and is listed in Appendix II

of the Berne Convention of 1979 (Grand, 1996), in

Appendix II of the European Community Habitat and

Species Directive of 1992 (Jackson & McLeod, 2000)

and as rare in the British Red Data Book. This species

is protected by the Wildlife and Countryside Act

(TSO, 1981) within the U.K. where, largely because of

decline/fragmentation of suitable habitat, it has suf-

fered a 30% reduction in distribution during the last

100 years (Thompson, Rouquette & Purse, 2003).

Coenagrion mercuriale is now limited to a few popula-

tions in southern England and Wales, with core sites

in the New Forest and along the River Itchen (both in

England) and in Pembrokeshire (Wales). There is

concern for the long-term persistence of C. mercuriale

because adult lifetime movement rarely exceeds 2 km

and is typically <100 m (Hunger & Roske, 2001; Purse

et al., 2003; Watts et al., 2004a), restricting (re-) colo-

nisation of suitable habitat to localities in close

proximity to extant sites. A previous study in the

Itchen Valley demonstrated that limited movement by

C. mercuriale leads to significant, local spatial genetic

structuring (Watts et al., 2004a). Here, we extend our

investigation into the population structure of C.

mercuriale to include all remaining population centres

in Wales (U.K.) that contrast with sites along the

Itchen Valley by exhibiting a greater level of spatial

isolation. In addition to exploring the pattern of

spatial genetic structure, we ask:

(i) What is the effect of habitat isolation upon

genetic diversity at local and regional scales? and

(ii) Do populations of C. mercuriale show a genetic

signature of demographic decline?

These questions are discussed with respect to the

general conservation of this species.

Methods

Description of study sites

We studied the remaining regions in Wales (Pem-

brokeshire, Gower and Anglesey) that host popula-

tions of C. mercuriale. This species does not inhabit

areas above 300 m in the U.K. (Purse, 2001) but the

Pembrokeshire populations (situated >150 m above

sea level) are at higher altitudes than other sites in

Wales (60–70 m; S. Coker, personal communication).

194 P.C. Watts et al.

! 2005 The Authors, Journal compilation ! 2005 Blackwell Publishing Ltd, Freshwater Biology, 51, 193–205

In Pembrokeshire, populations distributed in springs

on the Preseli Hills (Fig. 1) are undergoing a reduction

in size and distribution but still thought to be large

and extensive enough to be self-sustaining. Within

this area we collected samples from nine sites (Fig. 1)

that represent four of the five population centres that

have some separation by improved agricultural land;

northern [Brynberian (BRY), Clun Maen (CLM)],

eastern (not sampled), southern [Cors Tewgyll

(COT), Dolau Isaf (DOI), Gors Fawr (GOF), Pantithel

(PAN), Waun Isaf North (WAN)] and two colonies at

Waun Maes (WAM) and Waun Fawr (WAF). The

latter two sites are separated from the core areas by

relatively large distances (approximately 3 and 6 km,

respectively) and high (>300 m) ground. At 290 m in

altitude, WAM is 80–90 m higher than the other

Pembrokeshire sites and at the environmental limit of

this species (Purse, 2001). Although presently vigor-

ous, the WAF population is believed to be susceptible

to even subtle environmental changes because it is too

remote to receive migrants (Coker, 2001). All colonies

(except DOI and PAN which are on private land) in

Pembrokeshire are on common grazing land. Coenag-

rion mercuriale was presumably more widespread on

Gower but is now found at only two sites (Cefn Bryn

and Rhossili) that are separated by 7.5 km of farm-

land. At the extreme north of C. mercuriale’s U.K. (and

European) distribution and 150 km from the nearest

populations in Pembrokeshire, Nant Isaf on the island

of Anglesey represents an extremely remote site

(Fig. 1). There is concern for the small C. mercuriale

population at Anglesey because changes in the

grazing regime on this privately owned farmland

might have reduced habitat quality.

DNA extraction and polymerase chain reaction

Genomic DNA was extracted using a high salt

protocol (Sunnucks & Hales, 1996) from a tibia (stored

in a 1.5 mL tube containing 100% ethanol) for about

50 damselflies per site (Appendix). Loss of a leg does

not measurably affect fitness in damselflies (Fincke &

Hadrys, 2001) and we observed no significant effect of

sampling upon recapture rate elsewhere (D.J. Thomp-

son, unpublished data). We examined allelic variation

in 14 microsatellite loci (Appendix) characterised by

Watts, Thompson & Kemp (2004b) and Watts et al.

(2004c). Approximately 5 ng of DNA was used for a

10 lL polymerase chain reaction (PCR) containing

75 mMM Tris-HCl pH 8.9, 20 mMM (NH4)2SO4, 0.01% v/

v Tween-20, 0.2 mMM each dNTP, 3.0 mMM MgCl2,

20 pmol forward primer, 30 pmol reverse primer

and 0.25 U TAQTAQ polymerase (ABgene, Epsom, U.K.).

Thermal cycling conditions are described by Watts

et al. (2004b,c). PCR products were pooled into one of

two genotyping pools, determined by allelic size

range and the 5¢ fluorescent dye, along with a GENE-GENE-

SCANSCAN-500 LIZLIZ size standard (Applied Biosystems,

Fig. 1 Map of the location of three remaining population centres of C. mercuriale in Wales, U.K. Expanded section highlights geo-graphic location of sampled (filled circles) and unsampled (open circles) sites occupied by C. mercuriale in Pembrokeshire: northern[Brynberian (BRY), Clun Maen (CLM)], eastern (not sampled), southern [Cors Tewgyll (COT), Dolau Isaf (DOI), Gors Fawr (GOF),Pantithel (PAN), Waun Isaf North (WAN)], Waun Maes (WAM) and Waun Fawr (WAF); light grey shading indicates land >300 mhigh.

Genetic variation in an endangered damselfly 195


Foster City, CA, U.S.A.) and separated by capillary

electrophoresis through a denaturing acrylamide gel

on an ABI3100 automated sequencer (Applied Bio-

systems). Alleles were sized using the cubic model

of analysis in the GeneMapper analysis software

(Applied Biosystems).

Data analysis

Genetic variability. Genotypic linkage equilibrium

among all locus-pair combinations was assessed for

each population using the Fisher’s exact test imple-

mented by the online version (3.1c, http://

wbiomed.curtin.edu.au/genepop/) of GENEPOPGENEPOP (Ray-

mond & Rousset, 1995). Genetic diversity within each

sample was measured by the allelic richness (AR),

expected heterozygosity (He) and Wright’s (1951)

inbreeding coefficient (f) using FSTATFSTAT v.2.9.3 (Goudet,

1995);ARwas standardised at 31 individuals to account

for missing genotypes at one locus-sample combina-

tion (LIST4-002, CLM). We used FSTATFSTAT to test the

significance of any differences in AR, He and f among

sites at Pembrokeshire, Gower and Anglesey (the latter

site was randomly divided into two) by a permutation

procedure (5000 permutations of samples among

groups). The significance of any deviation from expec-

ted Hardy–Weinberg equilibrium (HWE) conditions

was estimated by making 5000 permutations of alleles

among individuals within samples using FSTATFSTAT.

Population bottleneck. We examined each sample for

evidence of a population decline (Thompson et al.,

2003) using two approaches. First, we used BOTTLE-BOTTLE-

NECKNECK v.1.2.02 software (Piry, Luikart & Cornuet,

1999) to estimate (assuming populations are in

mutation-drift equilibrium) the expected distribution

of heterozygosity from the observed number of alleles

under the infinite allele (IAM) and stepwise mutation

(SMM) models of mutation. A Wilcoxon signed-rank

test was used to test whether there was a significant

heterozygote excess (compared with that expected

given the sample’s allelic diversity) that is character-

istic of a population bottleneck (Cornuet & Luikart,

1996; Luikart & Cornuet, 1998). Second, for each

sample we calculated the average (over all loci) ratio

(M) of allelic size range (number of microsatellite

repeats) against the total number of alleles; values of

M above 0.82 indicate a stable population (Garza &

Williamson, 2001).

Population structure. We tested for heterogeneity of

genotype frequencies among all pairs of populations

using the exact test employed by FSTATFSTAT v.2.9.3

(Goudet, 1995); HWE within samples was not

assumed and genotypes were permuted 5000 times

among samples. Genetic differentiation between all

pairs of samples was also determined by calculating

Weir & Cockerham’s (1984) estimator of Wright’s

(1951) FST (h in Weir & Cockerham’s terminology)

using ARLEQUINARLEQUIN v.2.001 (Schneider, Roessli &

Excoffier, 2000). The significance of the estimates of

h from zero was assessed by making 1000 permuta-

tions of genotypes between populations. Heirarchical

analysis of molecular variance (AMOVAAMOVA; Schneider

et al., 2000) was used to partition the contribution to

genetic diversity arising from differences (i) between

regions (Pembrokeshire, Gower and Anglesey), (ii)

among sample sites within regions and (iii) among

individuals within sites. The significance of the

fixation indices was tested using 10 000 permuta-

tions.

Isolation by distance genetic structure (Wright,

1943) was examined for (i) all populations and (ii)

the Pembrokeshire samples separately by a regression

of pairwise estimates of genetic differentiation

[h/(1 ) h)] against the corresponding natural log-

arithm of the geographic distance (m) separating the

populations (Rousset, 1997). For simplicity we calcu-

lated the Euclidian distance among pairs of samples.

Some sites in Pembrokeshire are separated by high

ground that may limit dispersal more than expected

from their linear separation (Fig. 1); therefore, we also

measured the shortest routes that were restricted to

terrain <300 m high (i.e. that circumvented the central

ridge) between pairs of Pembrokeshire sites. A Mantel

test (1000 permutations of population locations

among all locations) was used to assess the signifi-

cance of any correlation between pairwise genetic

differentiation and geographic distance using the

online version (3.1c) of GENEPOPGENEPOP (Raymond &

Rousset, 1995). Mantel tests were applied to all pairs

of samples as well as the colonies from Pembrokeshire

only; for the latter analysis we considered both

Euclidian distance among samples as well as the

routes that avoid high ground.

Multiple testing. Where appropriate a sequential Bon-

ferroni correction (Rice, 1989) was applied to adjust

the significance of (k) multiple tests.



Results

Genetic diversity

Because only two loci demonstrated significant link-

age disequilibrium within any sample (LIST4-035-

LIST4-066 in GOF, P < 0.05, k ! 91) all loci were

retained for subsequent genetic analyses. Individual

sample-locus results for measures of genetic diversity

are presented in Appendix. Almost all samples met

expected HWE conditions with four and one sample-

locus combination(s) having a significant (P < 0.05,

k ! 14) heterozygote deficit or excess respectively

(Appendix).

Raw values of allelic richness (AR), expected hetero-

zygosity (He) and Wright’s (1951) inbreeding coeffi-

cient (f) are provided in Appendix. Particularly

notable is that five loci (36%) were monomorphic at

Anglesey, while only one locus (LIST4-037) was

invariable in two of the samples elsewhere (WAM &

WAF). Average levels of genetic diversity are gen-

erally similar among samples from Pembrokeshire

and Gower, but substantially lower in Anglesey

(Fig. 2a–c). Permutation analyses demonstrated that

AR, He or f did not significantly differ between groups

of samples from Pembrokeshire and Gower (P !0.234, 0.954 and 0.108, respectively), although when

Anglesey was included there were significant differ-

ences in AR (P ! 0.001) and He (P ! 0.001) but not f

(P ! 0.396) among groups.

Bottleneck analysis

Only C. mercuriale from Gower and Pembrokeshire

(five samples) demonstrated a significant (P < 0.05)

heterozygote excess from that expected under an IAM

ofmutation [but none did under the SMM– see Luikart

& Cornuet (1998) for discussion on mutational mod-

els]. This heterozygote excess remained significant

(P < 0.05, k ! 12) after a sequential Bonferroni correc-

tion (Rice, 1989) for CEB and RHO only. The ratio M

was 0.72 or less for all samples, with the lowest value

(M ! 0.61) at RHO (Table 1); M was 0.753 (variance

0.052) for the combined Pembrokeshire population.

Genetic differentiation between samples

Sites at Anglesey and Gower had significantly differ-

ent (P < 0.05, k ! 45) genotype frequencies from all

1.0

2.0

3.0

4.0

5.0(a)

Alle

lic ri

chne

ss (A

R)

0.0

0.1

0.2

0.3

0.4

0.5

0.6(b)

Exp

ecte

d he

tero

zygo

sity

(He)

–0.2

–0.1

0.0

0.1

0.2

0.3

0.4

PAN

GO

F

CO

T

WA

N

DO

I

CL

M

BR

Y

WA

M

WA

F

Cef

n B

ryn

Rho

ssili

Ang

lese

y

(c)

Wri

ght's

(195

1) in

bree

ding

coe

ffic

ient

(f)

Fig. 2 Variation (mean ± 95% CI) in (a) allelic richness (AR)standardised for a sample size of 31 individuals (b) expectedheterozygosity (He) and (c) Wright’s (1951) inbreeding coeffi-cient (f) among samples of C. mercuriale from the Pembrokeshire(PAN-WAF), Gower and Anglesey. See Methods and Fig. 1 forfurther description of site names and locations.



other samples (Table 1). Among populations from

Pembrokeshire, there was significant (P < 0.05, k !66) heterogeneity in genotype frequency between

samples separated by the central hill ridge (except

between GOF and CLM) and also among all four

northern sites. The southern Pembrokeshire samples

were mostly similar, with significant (P < 0.05, k !66) genotype differences between PAN and GOF, and

PAN and WAN only (Table 1).

All pairwise estimates of genetic differentiation (h)between either Anglesey and the Gower sites and all

other samples were significantly (P < 0.05, k ! 66)

different from zero, with comparisons involving

Anglesey notably high (h ! 0.35–0.44). Genetic differ-

entiation among the southern Pembrokeshire samples

was either low (h ! 0.016 and 0.018, PAN-COT and

PAN-WAN, respectively, P < 0.05, k ! 66) or non-

significant (P > 0.05, k ! 66). With the exception of

two comparisons involving CLM, all northern Pem-

brokeshire samples were significantly dissimilar to all

other samples from Pembrokeshire irrespective of

their geographic location. Estimates of h between

WAM and other samples from Pembrokeshire are

relatively large compared with h between other pairs

of samples from this area. Likewise, h between the

Gower colonies is also greater than that generally

observed between similarly spaced sites in Pembroke-

shire (Table 2). Hierarchical AMOVAAMOVA revealed signifi-

cant (P > 0.05) genetic structure at all levels but with

almost all variance (>97%) attributed to differences

within samples and among regions (Table 3).

There is a significant pattern of genetic isolation by

distance at a broad scale between all of the samples

(y ! 0.120x ) 0.974, P ! 0.0002, r2 ! 0.63) and also

locally among the Pembrokeshire samples (Fig. 3). For

the latter samples, although there is a positive and

significant regression between genetic differentiation

[h/(1 ) h)] and Euclidian distance (ln m) between

samples (y ! 0.013x ) 0.077, P ! 0.016, r2 ! 0.24) the

fit of this relationship (y ! 0.014x ) 0.087, P ! 0.003,

Table 1 Results of bottleneck analysis for samples of C. mercu-riale from Pembrokeshire, Gower and Anglesey (see Methodsand Fig. 1 for full sample site names and their location)

Region Population

Bottleneck M

IAM SMM Mean Variance

Pembrokeshire PAN 0.2508 0.9406 0.6549 0.0733GOF 0.0209 0.8794 0.6760 0.0630COT 0.0338 0.4758 0.6694 0.0758WAN 0.0595 0.8917 0.6542 0.0653DOI 0.0863 0.9031 0.6404 0.0545CLM 0.0101 0.5961 0.6695 0.0578BRY 0.0209 0.7869 0.6328 0.0564WAM 0.0090 0.2367 0.6182 0.0837WAF 0.1082 0.4197 0.6378 0.0733

Gower Cefn Bryn 0.0003* 0.1955 0.7245 0.0820Rhossili 0.0017* 0.3129 0.6120 0.0862

Anglesey 0.0645 0.1250 0.7000 0.0620

Bottleneck: *indicates a significant (P < 0.05, k ! 12) heterozy-gote excess for microsatellite loci under infinite allele (IAM) andstepwise (SMM) models of mutation.M: ratio of allelic size range (number of microsatellite repeats)against the total number of alleles.

Table 2 Genetic differentiation among samples of C. mercuriale from Pembrokeshire, Gower and Anglesey; see Methods and Fig. 1 forfull sample site names and their location. Exact probability of genotypic differentiation between samples (below diagonal) andpairwise estimates of genetic differentiation (h) (above diagonal).

Pembrokeshire Gower

AngleseyPAN GOF COT WAN DOI CLM BRY WAM WAF Cefn Bryn Rhossili

PAN 0.0112 0.0158* 0.0178* 0.0144 0.0302* 0.0410* 0.0764* 0.0349* 0.1521* 0.1531* 0.4014*GOF 0.00076* 0.0031 0.0022 0.0070 0.0146 0.0229* 0.0678* 0.0212* 0.1137* 0.1293* 0.4055*COT 0.00379 0.18485 0.0076 0.0096 0.0140 0.0293* 0.0657* 0.0282* 0.1511* 0.1564* 0.4149*WAN 0.00076* 0.02121 0.02348 0.0059 0.0165* 0.0250* 0.0653* 0.0339* 0.1383* 0.1514* 0.4262*DOI 0.00152 0.01136 0.02500 0.02803 0.0155* 0.0191* 0.0456* 0.0224* 0.1270* 0.1462* 0.4037*CLM 0.00076* 0.00152 0.00076* 0.00076* 0.00076* 0.0180* 0.0604* 0.0105 0.1523* 0.1662* 0.4244*BRY 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.0336* 0.0187* 0.1489* 0.1491* 0.4076*WAM 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.0558* 0.1834* 0.1913* 0.3482*WAF 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.1383* 0.1521* 0.4123*Cefn Bryn 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.0844* 0.4844*Rhossili 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.4425*Anglesey 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076* 0.00076*

*Indicates a significant (P < 0.05, k ! 45) genetic difference between populations.



r2 ! 0.36) is improved if the distance measured

between sites is restricted to a route below 300 m

altitude.

Discussion

Molecular genetic studies have provided detailed

insights into factors affecting the spatial genetic

structure of aquatic insects (e.g. Geenen et al., 2000;

Smith & Collier, 2001; Kelly et al., 2002; Miller et al.,

2002; Hughes et al., 2003a,b; Wilcock et al., 2003;

Wishart & Hughes, 2003; Watts et al., 2004a). How-

ever, most studies focus on dispersal rather than

processes affecting levels of genetic variability, de-

spite an increasing recognition that conservation

strategies should prevent or minimise genetic erosion

and if possible restore background levels of variability

(Couvet, 2002; Frankham et al., 2002). In this study the

large Pembrokeshire populations are the most gen-

etically diverse, although somewhat surprisingly the

smaller, separate Gower colonies have resisted sub-

stantial genetic erosion. The contrast in levels of

diversity at these sites with that observed at Anglesey

(Fig. 2a,b) may be attributed to one or a combination

of several causes.

Impact of habitat isolation and population decline

Anglesey, some 150 km north from Pembrokeshire

(Fig. 1) and probably a remnant from when C.

mercuriale extended along the lowland coastal areas

of Wales, meets several (not mutually exclusive)

criteria associated with genetically depauperate insect

populations: isolated (Keller & Largiader, 2003; Wil-

liams, Brawn & Paige, 2003; Krauss et al., 2004), small

size (Schmitt & Seitz, 2002; Harper, MacLean &

Goulson, 2003; Joyce & Pullin, 2003) and/or at the

edge of a species’ range (Schmitt & Hewitt, 2004). The

contrast in diversity at the isolated sites of Anglesey

and Gower underlines the point that genetic impov-

erishment is not simply an unavoidable consequence

of habitat fragmentation but rather reflects the effect

of population size upon the rate of genetic drift (see

Frankham, 1996; Frankham et al., 2002). For example,

using the relationship:

Ht ! Ho"1# 1=2Ne$t

0.0

0.2

0.4

0.6

0.8

1.0

6 7 8 9 10 11 12 13Geographic distance (m)

Gen

etic

dif

fere

ntia

tion

[q/(

1–q)

]

Pembrokeshire samples only

Fig. 3 Relationship between geographicdistance (ln m) separating samples of C.mercuriale from Pembrokeshire, Gowerand Anglesey and the correspondingestimate of pairwise genetic differenti-ation as defined by [h/(1 ) h)]. Expandedplot of the Pembrokeshire samples in thedashed frame highlights the isolation bydistance relationship at this site.

Table 3 Hierarchical analysis of molecular variance (AMOVAAMOVA)of microsatellite loci among samples of C. mercuriale from threeregions in Wales, U.K.: Pembrokeshire, Gower and Anglesey

Source ofvariation

Fixationindex P-value

Percentagevariation

Among regions 0.2343 0.0012 20.96Among sampleswithin regions

0.0312 0.0000 2.47

Within samples 0.2097 0.0000 76.57

P ! probability of a more extreme variance component than thatobserved.



where Ht ! expected heterozygosity after t genera-

tions, Ho ! initial expected heterozygosity and Ne !effective population size (Crow & Kimura, 1970), the

present level of expected heterozygosity at Anglesey

(from an initial Ht ! 0.5) is equivalent to Ne ! 10 or

1000 for about 20 or 1900 generations, respectively. It

may seem remarkable that there is no evidence for a

bottleneck at Anglesey but this is consistent with the

poor statistical power of (i) the ratio test when the

starting population size is small and (ii) both tests

with increasing time following the bottleneck (Luikart

& Cornuet, 1998; Garza & Williamson, 2001). A useful

feature of the ratio test, however, is that allelic

diversity remains low in permanently reduced popu-

lations, although the value of M partially recovers

(Garza & Williamson, 2001). The Anglesey colony

evidently suffered a severe reduction in size. While

the colony at Anglesey may have been founded by

one/few gravid females (see e.g. Joyce & Pullin, 2003

for genetic diversity in an introduced, isolated

population of Euphydryas aurinia Rott.) this is unlikely

given this species’ poor dispersal capability (Hunger

& Roske, 2001; Purse et al., 2003; Watts et al., 2004a).

Adaptation to an extreme, edge of range environment

may further contribute to the low genetic diversity at

Anglesey, but the consistency across loci suggests that

genome-wide drift is the principal cause.

Also in contrast to expectations were the similar

levels of genetic diversity at Pembrokeshire and

Gower. Similarly, Monaghan et al. (2002) found no

effect of fragmentation on the genetic diversity of

stream insects because population sizes were large

enough to minimise the rate of genetic drift and/or

there is gene flowamong populations. TheGower sites,

50 kmsouth of thePembrokeshire colonies, are likely to

have been isolated for a shorter duration thanAnglesey

and may have experienced more recent immigration

(contemporary gene flow to the Gower is unlikely, see

below) and/or historically maintained larger effective

population sizes. The genetic signature of recent

population decline correspondswith qualitative obser-

vations of a reduction in distribution of C. mercuriale at

Gower (D.J. Thompson, personal observations) that

will accelerate a reduction in gene diversity.

Within Pembrokeshire, localised differences in gen-

etic variability are indicative of fine-scale variation in

demography that is typical of many insect populations

(e.g. Schmitt & Seitz, 2002; Wynne et al., 2003). Of note

is the likely recent decline in size at WAM (cf. Table 1)

and concurrent reduction in AR (it is possible that low

diversity reflects a selective response to altitude); any

loss of variability at WAM will be sustained by

restricted immigration (see below). By contrast, the

level of variability at WAF further highlights the point

that the rate of genetic drift in large populations is

generally low, such that substantial levels of diversity

may be retained even when a population is isolated

(see also Keller, Nentwig & Largiader, 2004). Else-

where in Pembrokeshire, connectivity between the

southern, eastern and northern populations may

buffer individual colonies against genetic erosion

associated with demographic fluctuations.

Population genetic structure and movement among sites

The substantial level of genetic differentiation that may

be attributed to differences among regions (Table 3;

Fig. 3) undoubtedly corresponds to breaks in distribu-

tion that far outreachC.mercuriale’s dispersal capability

(Hunger & Roske, 2001; Purse et al., 2003; Watts et al.,

2004a). The pattern of genetic differentiation among

populations with distance, however, provides more

insight into processes affecting population structure

(Miller et al., 2002). Weak and/or variable regional

genetic differences and absence of isolation by distance

that have been reported for some aquatic insects are

indicative of, for example, wide dispersal, genetic

disequilibrium, balancing selection or poor marker

resolution (e.g. Smith &Collier, 2001;Miller et al., 2002;

Monaghan et al., 2002;Hughes et al., 2003b). In contrast,

an apparent pattern of isolation by distance (Fig. 3) is a

feature of restricted dispersal and relatively small

populations that facilitates rapid population-differen-

tiation through genetic drift (see also Miller et al., 2002;

Wilcock et al., 2003).

At a local scale, genetic differentiation between

the Gower sites is greater than that observed

between similarly spaced sites in continuously

distributed C. mercuriale populations (Table 2; Watts

et al., 2004a) indicating that dispersal is disrupted by

agricultural improvement. Indeed, restricted gene

flow has been associated with anthropogenic chan-

ges to habitat for several terrestrial (Keller &

Largiader, 2003; Williams et al., 2003; Keller et al.,

2004) and aquatic insects (Smith & Collier, 2001;

Watts et al., 2004a). Similarly, the improved fit of the

regression based on movement among sites within

Pembrokeshire that avoids high ground corresponds



with observations that, like many freshwater insects

(e.g. Petersen et al., 2004), adult C. mercuriale move

exclusively between areas of suitable habitat (Purse

et al., 2003); the hill ridge is a natural barrier to

dispersal (see also Wishart & Hughes, 2003). Low or

non-significant genetic differences among adjacent

areas in Pembrokeshire (Table 2) indicate dispersal

by C. mercuriale over 1–2 km, although this does not

prevent genetic isolation by distance (see Fig. 3,

insert). Increased genetic differentiation at WAM

(Table 2) is a likely consequence of isolation by high

ground, but directional selection or drift during the

recent demographic reduction may also accentuate

genetic differences between this site and elsewhere

(see Hedrick, Gutierrez-Espeleta & Lee, 2001).

Variable but often low genetic differences between

the central sites and WAF do not indicate con-

nectivity, but reflect a historical genetic signature

with the large WAF population not yet achieving

migration-drift equilibrium since its isolation.

Neither Geenen et al. (2000) or Wong et al. (2003)

identified fine-scale genetic structure in other spe-

cies of zygopteran but the pattern and magnitude of

local genetic differences in Pembrokeshire parallels

that observed for this species in the Itchen Valley

(Watts et al., 2004a).

Implications for conservation

Population size is an important correlate of extinction

risk (O’Grady et al., 2004) and, accordingly, the

current U.K. strategy for C. mercuriale conservation is

based upon habitat protection to maintain large

populations. However, even apparently healthy pop-

ulations can be at long-term risk if they lack genetic

diversity (Saccheri et al., 1998; Frankham et al., 2002;

Keller & Waller, 2002; Reed & Frankham, 2003;

Schmitt & Hewitt, 2004; Spielman et al., 2004a,b).

Improving connectivity among fragmented popula-

tions will help maintain/restore genetic diversity and

preserve evolutionary patterns and processes (Moritz,

1999): such a strategy is not possible between regions

but is generally appropriate within each area, especi-

ally the Gower sites. Somewhat ironically, the histor-

ical lack of management (agricultural improvement)

imposed by its common land status has helped limit

genetic erosion in the Pembrokeshire, although a

recent decline in sheep grazing has unfortunately

reduced habitat quality. It is relevant, therefore, that

M < 0.68, characteristic of each Pembrokeshire sam-

ple, is symptomatic of a recent reduction in popula-

tion size (Garza & Williamson, 2001); even when

genotype data for all Pembrokeshire sites are com-

bined to constitute a single population the calculated

value of M ! 0.75 is still below that observed in

demographically stable populations (M ! 0.82, Garza

& Williamson, 2001). Data on levels of genetic

variability in C. mercuriale populations towards the

centre of its range would provide further context to

the variation observed in Pembrokeshire and Gower.

Some combination of habitat isolation, small popula-

tion size and/or directional selection has lead to

localised reduction in genetic variability at WAM and

a striking level of genetic impoverishment at Anglesey

that is more typical of threatened mammal (e.g.

Hedrick et al., 2001; Whitehouse & Harley, 2001; Jones

et al., 2004) than fragmented insect (e.g. Keller &

Largiader, 2003; Williams et al., 2003; Keller et al.,

2004) populations. There is particular concern for

C. mercuriale at Anglesey which may be (i) locally

adapted and/or (ii) likely to suffer as a result of no

immigration (see Couvet, 2002; Keller & Waller, 2002),

for example, from inbreeding depression (Saccheri

et al., 1998) or increased susceptibility to disease

(Spielman et al., 2004b). The former implies that there

should be focus on habitatmanagementwhile the latter

suggests that the population may benefit from aug-

mentation (e.g. Madsen et al., 1999). Because we have

not yet addressed either scenario it would be unwise to

suggest augmentation (see Moritz, 1999; Stockwell,

Hendry & Kinnison, 2003). It is interesting to note that,

if the decision was taken for management plans to

prioritise the conservation of genetic diversity per se, for

example to allow future evolutionary processes, rather

than current adaptive novelties (seeMoritz, 1999), then

any effort directed towardsmanaging theAnglesey site

should perhaps be diminished and efforts focused on

the larger populations in Pembrokeshire and

Hampshire.

To conclude, populations of C. mercuriale exhibit

local and regional variation in levels of genetic

diversity that is associated with their population

demographics. Rather than directly eliciting a loss of

variation, habitat fragmentation reinforces genetic

impoverishment in small populations by preventing

augmentation via gene flow. Consistent with the

pattern of spatial genetic structure from elsewhere in

the U.K., dispersal by mature C. mercuriale in Pem-



brokeshire is sufficient to prevent significant genetic

differentiation throughout a habitat matrix of about

3–4 km if the sites are <2 km apart and lack

intervening barriers. Nevertheless, dispersal by

C. mercuriale is sufficiently restricted so that isolation

by distance genetic structure develops within 10 km.

Acknowledgments

Coenagrion mercuriale is a protected species under

Schedule 5 of the Wildlife and Countryside Act (1981).

All work was carried out under licence from the

Countryside Council for Wales. We are grateful to the

NERC (grant no. NER/A/S/2000/01322) for provi-

sion of funds and the landowners for allowing us

access to their land. Stephen Coker and Adrian

Fowles provided helpful comments on the manu-

script. Jac Pearson, Sarah Perry, Carri Westgarth and

Kerry Woodbine were excellent field assistants.

References

Andres R.A., Sanchez-Guill _en R.A. & Cordero Rivera A.(2000) Molecular evidence for selection on female colorpolymorphism in the damselfly Ischnura graellsii.Evolution, 54, 2156–2161.

Andres R.A., Sanchez-Guill _en R.A. & Cordero Rivera A.(2002) Evolution of female color polymorphism indamselflies: testing the hypotheses. Animal Behaviour,63, 677–685.

Coker S. (2001) A Long-Term Plan for Coenagrion mercurialein North-East Pembrokeshire. Report published privatelyby author, 8 pp.

Conrad K.F., Willson K.H., Harvey I.F., Thomas C.J. &Sherratt T.N. (1999) Dispersal characteristics of sevenodonate species in an agricultural landscape. Ecogra-phy, 22, 524–531.

Corbet P.S. (1999) Dragonflies: Behaviour and Ecology ofOdonata. Harley Books, Colchester.

Cornuet J.M. & Luikart G. (1996) Description and poweranalysis of two tests for detecting recent populationbottlenecks from allele frequency data. Genetics, 144,2001–2014.

Couvet D. (2002) Deleterious effects of restricted geneflow in fragmented populations. Conservation Biology,16, 369–376.

Crow J.F. & Kimura M. (1970) An Introduction toPopulation Genetics Theory. Harper & Row, New York.

Elliott J.M. (2003) A comparative study of the dispersal of10 species of stream invertebrates. Freshwater Biology,48, 1652–1668.

Fincke O.M. & Hadrys H. (2001) Unpredictable offspringsurvivorship in the damselfly, Megaloprepus coerulatus,shapes parental behavior, constrains sexual selection,and challenges traditional fitness estimates. Evolution,55, 762–772.

Frankham R. (1995) Conservation genetics. Annual Re-views in Genetics, 29, 305–327.

Frankham R. (1996) Relationship of genetic variation topopulation size in wildlife. Conservation Biology, 10,1500–1508.

Frankham R., Ballou J.D. & Briscoe D.A. (2002) Introduc-tion to Conservation Genetics. Cambridge UniversityPress, Cambridge.

Freeland J.R., May M., Lodge R.L. & Conrad K.F. (2003)Genetic diversity and widespread haplotypes in amigratory dragonfly, the common green darner Anaxjunius. Ecological Entomology, 28, 413–421.

Garza J.C. & Williamson E.G. (2001) Detection ofreduction in population size using data from micro-satellite loci. Molecular Ecology, 10, 305–318.

Geenen S., Jordaens K., de Block M., Stoks R. & de BruynL. (2000) Genetic differentiation and dispersal amongpopulations of the damselfly Lestes viridis (Odonata).Journal of the North American Benthological Society, 19,321–328.

Goudet J. (1995) FSTATFSTAT (vers. 1.2): a computer program tocalculate F-statistics. Journal of Heredity, 86, 485–486.

Grand D. (1996) Coenagrion mercuriale (Charpentier,1840). In: Background Information on Invertebrates of theHabitats Directive and the Bern Convention – Part II:Mantodea, Odonata, Orthoptera, and Arachnida, Vol. 80(Eds P.J. van Helsdingen, L. Willemse & M.C.D.Speight ), Council of Europe Publishing, Strasbourg.

Harper G.L., MacLean N. & Goulson D. (2003) Micro-satellite markers to assess the influence of populationsize, isolation and demographic change on the geneticstructure of the UK butterfly Polyommatus bellargus.Molecular Ecology, 12, 3349–3357.

Hedrick P.W., Gutierrez-Espeleta G.A. & Lee R.N. (2001)Founder effect in an island population of bighornsheep. Molecular Ecology, 10, 851–857.

Henle K., Davies K.F., Kleyer M., Margules C. & Settele J.(2004) Predictors of species sensitivity to fragmenta-tion. Biodiversity and Conservation, 1, 207–251.

Hughes J.M., Hillyer M. & Bunn S.E. (2003a) Small-scalepatterns of genetic variation in the mayfly Bungonanarilla (Ephemeroptera: Baetidae) in rainforest streams,south-east Queensland. Freshwater Biology, 48, 709–717.

Hughes J.M., Mather P.B., Hillyer M.J., Cleary C. &Peckarsky B. (2003b) Genetic structure in a montanemayfly Baetis bicaudatus (Ephemeroptera: Baetidae),from the Rocky Mountains, Colorado. FreshwaterBiology, 48, 2149–2162.



Hunger H. & Roske W. (2001) Short-range dispersal ofthe southern damselfly (Coenagrion mercuriale: Odo-nata) defined experimentally using UV fluorescent ink.Zeitschrift fur Okologie und Naturshutz, 9, 181–187.

Jackson D.L. & McLeod C.R. (2000) Handbook on the UKstatus of EC Habitats Directive interest features:provisional data on the UK distribution and extent ofAnnex I habitats and the UK distribution and popula-tion size of Annex II species. JNCC Report 312, 180 pp.,Peterborough, U.K.

Jones M.E., Paetkau D., Geffen E. & Moritz C. (2004)Genetic diversity and population structure of Tasma-nian devils, the largest marsupial carnivore. MolecularEcology, 13, 2197–2209.

Joyce D.A. & Pullin A.S. (2003) Conservation implica-tions of the distribution of genetic diversity at differentscales: a case study using the marsh fritillary butterfly(Euphydryas aurinia). Biological Conservation, 114, 453–461.

Keller I. & Largiader C.R. (2003) Recent habitat fragmen-tation due to major roads leads to reduction of geneflow and loss of genetic variability in ground beetles.Proceedings of the Royal Society of London, Series B, 270,417–423.

Keller L.F & Waller D.M. (2002) Inbreeding effects inwild populations. Trends in Ecology and Evolution, 17,230–241.

Keller I., Nentwig W. & Largiader C.R. (2004) Recenthabitat fragmentation due to roads can lead to signi-ficant genetic differentiation in an abundant flightlessground beetle. Molecular Ecology, 13, 2983–2994.

Kelly L.C., Rundle S.D. & Bilton D.T. (2002) Geneticpopulation structure and dispersal in Atlantic Islandcaddisflies. Freshwater Biology, 47, 1642–1650.

Krauss J., Schmitt T., Seitz A., Steffan-Dewenter I. &Tscharntke T. (2004) Effects of habitat fragmentationon the genetic structure of the monophagous butterflyPolyommatus coridon along its northern range margin.Molecular Ecology, 13, 311–320.

Luikart G. & Cornuet J.-M. (1998) Empirical evaluation ofa test for identifying recently bottlenecked populationsfrom allele frequency data. Conservation Biology, 12,228–237.

Madsen T., Shine R., Olsson M. & Wittzell H. (1999)Conservation biology – restoration of an inbred adderpopulation. Nature, 402, 34–35.

Miller M.P., Blinn D.W. & Keim P. (2002) Correlationsbetween observed dispersal capabilities and patternsof genetic differentiation in populations of four aquaticinsect species from the Arizona Mountains, U.S.A.Freshwater Biology, 47, 1660–1673.

Monaghan M.T., Spaak P., Robinson C.T. & Ward J.(2002) Population genetic structure of 3 alpine stream

insects: influences of gene flow, demographics, andhabitat fragmentation. Journal of the North AmericanBenthological Society, 21, 114–131.

Moritz C. (1999) Conservation units and translocations:strategies for conserving evolutionary processes. He-reditas, 130, 217–228.

O’Grady J.J., Reed D.H., Brook B.W. & Frankham R.(2004) What are the best correlates of predictedextinction risk? Biological Conservation, 118, 513–520.

Petersen I., Masters Z., Hildrew A.G. & Ormerod S.J.(2004) Dispersal of aquatic insects in catchments ofdiffering landuse. Journal of Applied Ecology, 41, 934–950.

Piry S., Luikart G. & Cornuet J.-M. (1999) BOTTLENECKBOTTLENECK.A Program for Detecting Recent Effective PopulationSize Reductions from Allele Frequency Data. http://www.ensam.inra.fr/URLB.

Purse B.V. (2001) The Ecology and Conservation of theSouthern Damselfly (Coenagrion mercuriale). PhDthesis, University of Liverpool, Liverpool.

Purse B.V., Hopkins G.W., Day K.J. & Thompson D.J.(2003) Dispersal characteristics and management of arare damselfly. Journal of Applied Ecology, 40, 716–728.

Raymond M. & Rousset F. (1995) GENEPOP, Version 1.2.Population genetics software for exact tests andecumenicisms. Journal of Heredity, 86, 248–249.

Reed D.H. & Frankham R. (2003) Fitness is correlatedwith genetic diversity. Conservation Biology, 17, 1–9.

Rice W.R. (1989) Analyzing tables of statistical tests.Evolution, 43, 223–225.

Rousset F. (1997) Genetic differentiation and estimationof gene flow from F-statistics under isolation bydistance. Genetics, 145, 1219–1228.

Saccheri I.J., Kuussaari M., Kankare M., Vikman P.,Fortelius W. & Hanski I. (1998) Inbreeding andextinction in a butterfly metapopulation. Nature, 392,491–494.

Schmitt T. & Hewitt G.M. (2004) The genetic pattern ofpopulation threat and loss: a case study of butterflies.Molecular Ecology, 13, 21–31.

Schmitt T. & Seitz A. (2002) Influence of habitatfragmentation on the genetic structure of Polyommatuscoridon (Lepidoptera: Lycaenidae): implications forconservation. Biological Conservation, 107, 291–297.

Schneider S., Roessli D. & Excoffier C. (2000) ArlequinVersion 2000, Genetics and Biometry. University ofGeneva, Switzerland.

Slatkin M. (1985) Gene flow in natural populations.Annual Review of Ecology and Systematics, 16, 393–430.

Smith P.J. & Collier K.J. (2001) Allozyme diversity andpopulation genetic structure of the caddisfly Orthop-syche fimbriata and the mayfly Acanthophlebia cruentatain New Zealand streams. Freshwater Biology, 46, 795–805.



Spielman D., Brook B.W. & Frankham R. (2004b) Mostspecies are not driven to extinction before geneticfactors impact them. Proceedings of the National Academyof Sciences, USA, 101, 15261–25264.

Spielman D., Brook B.W., Briscoe B.W. & Frankham R.(2004a) Does inbreeding and a loss of genetic diversitydecrease disease resistance? Conservation Genetics, 5,439–448.

Stockwell C.A., Hendry A.P. & Kinnison M.T. (2003)Contemporary evolution meets conservation biology.Trends in Ecology and Evolution, 18, 94–101.

Sunnucks P. & Hales D.F. (1996) Numerous transposedsequences of mitochondrial cytochrome oxidase I-II inaphids of the genus Sitobion (Hemiptera: Aphididae).Molecular Biology and Evolution, 13, 510–524.

Thompson D.J., Rouquette J.R. & Purse B.V. (2003)Ecology of the Southern Damselfly, Coenagrion mercuriale.Conserving Natura 2000 Rivers Ecology Series No.8.English Nature, Peterborough.

TSO (1981) Wildlife and Countryside Act. The StationaryOffice (TSO), House of Commons, 128 pp., London,U.K.

Van Tol J. & Verdonk M.J. (1988) Protection of Dragonfliesand Their Biotopes. Nature and Environment Series.Council of Europe, Strasburg.

Watts P.C., Thompson D.J. & Kemp S.J. (2004b) Cross-species amplification of microsatellite loci in someEuropean zygopteran species (Odonata: Coenagrioni-dae). International Journal of Odonatology, 7, 87–96.

Watts P.C., Rouquette J.R., Saccheri I.J., Kemp S.J. &Thompson D.J. (2004a) Molecular and ecologicalevidence for small-scale isolation by distance in anendangered damselfly, Coenagrion mercuriale. MolecularEcology, 13, 2931–2945.

Watts P.C., Wu J.H., Westgarth C., Thompson D.J. &Kemp S.J. (2004c) A panel of microsatellite loci for theSouthern Damselfly, Coenagrion mercuriale (Odonata:Coenagrionidae). Conservation Genetics, 5, 117–119.

Weir B.S. & Cockerham C.C. (1984) Estimating F-statistics for the analysis of population structure.Evolution, 38, 1358–1370.

Whitehouse A.M. & Harley E.H. (2001) Post-bottleneckgenetic diversity of elephant populations in SouthAfrica, revealed using microsatellite analysis. Molecu-lar Ecology, 10, 2139–2149.

Wilcock H.R., Nichols R.A. & Hildrew A.G. (2003)Genetic population structure and neighbourhoodpopulation size estimates of the caddisfly Plectrocnemiaconspersa. Freshwater Biology, 48, 1813–1824.

Williams B.L., Brawn J.D. & Paige K.N. (2003) Landscapescale genetic effects of habitat fragmentation on a highgene flow species: Speyeria idalia (Nymphalidae).Molecular Ecology, 12, 11–20.

Wishart M.J. & Hughes J.M. (2003) Genetic populationstructure of the net-winged midge, Elporia barnardi(Diptera: Blephariceridae) in streams of the south-western Cape, South Africa: implications for dispersal.Freshwater Biology, 48, 28–38.

Wong A., Smith M.L. & Forbes M.R. (2003) Differentia-tion between subpopulations of a polychromaticdamselfly with respect to morph frequencies, but notneutral genetic markers. Molecular Ecology, 12, 3505–3513.

World Conservation Monitoring Centre (WCMC) (1992)Global Biodiversity: Status of the Earth’s Living Resources.Chapman and Hall, London.

Wright S. (1943) Isolationbydistance.Genetics, 28, 114–138.Wright S. (1951) The genetical structure of populations.Annals of Eugenics, 15, 323–354.

Wynne I.R., Loxdale H.D., Brookes C.P. & Woiwod I.P.(2003) Genetic structure of fragmented November moth(Lepidoptera: Geometridae) populations in farmland.Biological Journal of the Linnean Society, 78, 467–477.

(Manuscript accepted 29 September 2005)

Appendix Genetic diversity at 14 microsatellite loci for samples of C. mercuriale from the Pembrokeshire (PAN–WAF), Gower andAnglesey (see Methods and Fig. 1 for full sample site names and their location

Pembrokeshire Gower

Anglesey(n ! 52)

PAN(n ! 48)

GOF(n ! 48)

COT(n ! 48)

WAN(n ! 48)

DOI(n ! 48)

CLM(n ! 48)

BRY(n ! 47)

WAM(n ! 48)

WAF(n ! 48)

Cefn Bryn(n ! 48)

Rhossili(n ! 48)

4–002AR 2.000 2.000 2.000 2.000 2.000 2.000 2.000 2.000 2.000 2.000 2.000 1.000He 0.479 0.491 0.443 0.505 0.506 0.508 0.437 0.389 0.502 0.506 0.505 0.000f )0.001 )0.018 0.295 0.052 0.126 0.224 0.054 )0.017 0.103 0.012 )0.053 –

4–023AR 4.272 3.494 2.997 2.992 3.955 3.649 3.401 2.438 2.880 3.437 3.000 1.656He 0.555 0.573 0.584 0.573 0.615 0.573 0.546 0.514 0.540 0.564 0.627 0.039f 0.078 0.128 )0.263 0.009 0.081 )0.113 0.061 )0.136 0.074 )0.035 )0.097 1.000



Appendix (Continued)

Pembrokeshire Gower

Anglesey(n ! 52)

PAN(n ! 48)

GOF(n ! 48)

COT(n ! 48)

WAN(n ! 48)

DOI(n ! 48)

CLM(n ! 48)

BRY(n ! 47)

WAM(n ! 48)

WAF(n ! 48)

Cefn Bryn(n ! 48)

Rhossili(n ! 48)

4–024AR 3.989 2.447 2.994 2.972 2.994 2.987 2.730 2.992 2.000 2.553 3.000 1.000He 0.287 0.419 0.341 0.283 0.393 0.455 0.434 0.368 0.446 0.423 0.648 0.000f 0.149 )0.271 0.083 )0.127 )0.091 )0.098 0.057 )0.076 )0.259 )0.245 0.049 –

4–030AR 2.000 2.000 2.871 2.000 2.000 2.000 2.000 2.000 2.000 2.000 2.000 1.000He 0.501 0.504 0.527 0.490 0.499 0.391 0.494 0.498 0.463 0.368 0.426 0.000f )0.462 )0.112 )0.103 )0.065 )0.570† 0.011 )0.235 )0.266 )0.338 0.039 )0.027 –

4–031AR 3.753 3.500 3.871 3.447 3.707 3.707 3.512 3.000 3.512 3.437 3.000 1.000He 0.643 0.661 0.684 0.672 0.680 0.684 0.646 0.646 0.630 0.680 0.618 0.000f )0.185 0.244 0.048 0.304* )0.215 0.014 0.169 )0.135 0.187 0.142 0.090 –

4–034AR 1.993 2.000 2.000 1.999 2.000 2.000 1.994 2.000 2.000 2.000 2.000 1.420He 0.157 0.241 0.307 0.221 0.216 0.241 0.160 0.369 0.327 0.234 0.271 0.020f )0.082 0.029 )0.221 0.058 )0.125 0.029 )0.084 0.153 )0.073 0.118 )0.020 0.000

4–035AR 7.241 8.611 9.138 7.668 8.306 7.108 8.503 6.642 8.523 6.482 4.385 2.000He 0.735 0.798 0.809 0.765 0.771 0.779 0.781 0.561 0.772 0.768 0.398 0.499f )0.162 0.046 )0.158 )0.023 )0.091 )0.113 )0.086 )0.077 0.059 )0.069 0.092 0.121

4–037AR 2.826 2.864 2.693 2.853 1.871 2.000 1.457 1.000 1.000 2.000 3.000 2.000He 0.304 0.179 0.213 0.177 0.068 0.276 0.022 0.000 0.000 0.451 0.565 0.389f )0.029 )0.067 )0.100 0.292 )0.024 0.135 0.000 – – 0.211 0.097 )0.125

4–042AR 2.000 2.000 2.000 2.000 2.477 1.999 2.000 1.992 1.996 2.000 2.000 2.000He 0.367 0.410 0.345 0.438 0.361 0.205 0.369 0.155 0.167 0.401 0.507 0.465f )0.306 )0.061 )0.026 )0.165 0.118 )0.119 0.174 0.192 )0.089 )0.246 0.343 )0.119

4–060AR 2.877 2.500 3.357 2.835 2.456 2.000 2.953 2.985 2.500 2.000 2.477 2.000He 0.237 0.331 0.271 0.323 0.279 0.261 0.355 0.461 0.420 0.479 0.454 0.470f )0.105 )0.006 0.141 )0.054 )0.012 0.000 0.015 0.186 0.036 0.701* )0.051 0.044

4–062AR 2.992 2.844 2.985 2.997 2.961 2.948 3.000 3.000 2.753 2.686 2.000 2.000He 0.499 0.446 0.518 0.482 0.552 0.530 0.614 0.670 0.482 0.438 0.439 0.498f )0.085 )0.024 )0.086 )0.014 )0.006 0.096 0.292 )0.039 0.210 0.001 0.032 )0.045

4–063AR 3.853 3.731 2.972 3.648 3.951 3.590 3.688 2.438 2.999 1.888 2.447 1.447He 0.546 0.463 0.499 0.333 0.466 0.488 0.537 0.515 0.580 0.071 0.303 0.021f )0.059 0.009 0.040 )0.023 )0.001 0.042 )0.133 )0.092 0.224 )0.026 )0.053 0.000

4–066AR 5.240 6.598 6.898 6.086 6.160 6.538 6.776 3.993 4.678 2.000 4.757 2.000He 0.421 0.551 0.559 0.501 0.421 0.357 0.613 0.610 0.254 0.320 0.535 0.278f )0.187 0.050 0.002 0.045 )0.158 0.004 )0.074 )0.069 )0.082 )0.105 0.204 0.559*

4–067AR 7.243 6.961 8.951 7.551 8.009 6.084 6.729 4.590 9.154 9.026 7.186 1.000He 0.617 0.738 0.769 0.727 0.721 0.719 0.691 0.592 0.665 0.810 0.692 0.000f 0.020 0.019 )0.098 )0.107 0.029 )0.061 )0.184 )0.222 )0.165 0.021 0.278* –

Sample size (n), allelic richness (AR), expected heterozygosity (He) and Wright’s (1951) inbreeding coefficient (f ).* and † indicate a significant (P < 0.05, k ! 9–14) heterozygote deficit or excess, respectively – indicates monomorphic locus, so test notcarried out.



Population structure and the impact of regional and local habitat isolation upon levels of genetic...

Documents

Transcript of Population structure and the impact of regional and local habitat isolation upon levels of genetic...