Abstract The fragmentation of landscapes has an

important impact on the conservation of biodiversity,

and the genetic diversity is an important factor for a

populations viability, influenced by the landscape

structure. However, different species with differing

ecological demands react rather different on the same

landscape pattern. To address this feature, we studied

three skipper species with differing habitat require-

ments (Lulworth Skipper Thymelicus acteon: a habitat

specialist with low dispersal ability, Small Skipper

Thymelicus sylvestris: a habitat generalist with low

dispersal ability, Essex Skipper Thymelicus lineola: a

habitat generalist with higher dispersal ability). We

analysed 18 allozyme loci for 1,063 individuals in our

western German study region with adjoining areas in

Luxembourg and north-eastern France. The genetic

diversity of all three species were intermediate in

comparison with other butterfly species. The FST was

relatively high for T. acteon (5.1%), low for T. sylves-

tris (1.6%) and not significant for T. lineola. Isolation

by distance analyses revealed a significant correlation

for T. sylvestris explaining 20.3% of its differentiation,

but no such structure was found for the two other

species. Most likely, the high dispersal ability of

T. lineola in comparison with T. sylvestris leads to a

more or less panmictic structure and hence impedes

isolation by distance. On the other hand, the isolation

of the populations of T. acteon seems to be so strict

that the populations develop independently. Although

no general genetic impoverishing was observed for the

endangered T. acteon, small populations had signifi-

cantly lower genetic diversities than big populations,

and therefore the high degree of isolation among

populations might threaten its local and regional sur-

vival.

Keywords Habitat fragmentation � Isolation by

distance � Allozyme electrophoreses � Thymelicus �Lepidoptera

Introduction

Loss and fragmentation of habitats is a major threat for

global biodiversity (Hanski 1999). For small popula-

tions in fragmented landscapes, increasing isolation

leads to reduced colonisation rates and enhanced risks

of extinction (Rosenzweig 1995; Hanski 1999). There-

fore, species richness of habitats decreases with

increasing habitat fragmentation. The same phenome-

non applies for the genetic diversity of populations too,

and many empirical studies show the correlation be-

tween habitat fragmentation and genetic diversity of

populations (e.g., Young et al. 1996; Buza et al. 2000;

Pedersen and Loeschcke 2001; Keller and Largiader

2003; Williams et al. 2003). Several examples show the

linkage between genetic diversity and the fitness of

individuals and populations (Frankham et al. 2002;

D. Louy (&) � J. C. Habel � T. Schmitt � P. MullerBiogeography, Fachbereich VI, University Trier,Am Wissenschaftspark 25-27, D-54296 Trier, Germanye-mail: D.Louy@web.de

D. Louy � T. AssmannInstitut of Ecology and Environmental Chemistry,University Lueneburg, D-21335 Lueneburg, Germany

M. MeyerMuseum of natural history Luxembourg, L-2160Luxembourg, Luxembourg

Conserv Genet (2007) 8:671–681

DOI 10.1007/s10592-006-9213-y

123

ORIGINAL PAPER

Strongly diverging population genetic patterns of three skipperspecies: the role of habitat fragmentation and dispersal ability

Dirk Louy Æ Jan Christian Habel Æ Thomas Schmitt ÆThorsten Assmann Æ Marc Meyer Æ Paul Muller

Received: 8 September 2005 / Accepted: 5 September 2006 / Published online: 2 March 2007� Springer Science+Business Media B.V. 2007

Hansson and Westerberg 2002; Reed and Frankham

2003; Schmitt and Hewitt 2004). Therefore, conserva-

tion biologists still pay increasing attention to the ef-

fects of habitat fragmentation on genetic diversity.

However, whether a landscape is fragmented strongly

depends on the respective species, its habitat require-

ment, dispersal ability, population density, habitat area

requirement, longevity etc.

The effects of a landscape on the population

genetics of species is testable by the analysis of closely

related taxa (e.g., species belonging to the same genus),

but with differing ecological demands and dispersal

abilities. A landscape therefore might be strongly

fragmented for a habitat specialist with poor dispersal

abilities and more or less continuous for a more

generalist species with higher dispersal capacity.

Generalist species with poor dispersal abilities as well

as specialist species with high dispersal abilities

might show some intermediate position. However,

such studies are largely missing (but compare

Johannesen et al. 1999; Schmitt and Seitz 2002; Schmitt

et al. 2003).

To analyse the effect of a landscape on the population

genetic structure of ecologically differing taxa, we se-

lected the three Central European skipper species of the

genus Thymelicus. Although the caterpillars of these

species all feed on various grass species (Ebert and

Rennwald 1991; Asher et al. 2001), their ecological de-

mands, in general, differ strongly: The Lulworth Skipper

(Thymelicus acteon Rottemburg 1775) is a xerothermo-

philic species restricted to hot grassland habitats,

whereas the Small Skipper (Thymelicus sylvestris Poda

1761) and the Essex Skipper (Thymelicus lineola Och-

senheimer 1808) are widely distributed in a great variety

of flower-rich open habitats (Ebert and Rennwald 1991;

Tolman and Lewington 1998; Asher et al. 2001). The

dispersal abilities of T. acteon and T. sylvestris are smaller

than in T. lineola (Bink 1992). Thymelicus acteon has a

restricted distribution and is decreasing and therefore is

considered endangered in many European countries,

whereas T. sylvestris and T. lineola are widespread and

stable in most parts of Europe (van Swaay and Warren

1999). These characteristics make these three species a

suitable model system to test for the divergent effects of

landscape fragmentation.

As study area, we selected the Rhineland-Palatinate

and the Saarland (south-west Germany) with some

adjacent areas in Luxembourg and north-eastern

France. This region has several calcareous grassland

areas (Negendank 1974) with populations of T. acteon

(Schmidt-Koehl 1977; Kraus 1993). Thymelicus lineola

and T. sylvestris are widely distributed all over this

region and are found at most flower-rich places

including meadows, fallow land, forest skirts etc.

(Schmidt-Koehl 1977; Kraus 1993).

For each species, we analysed 8–11 locations from

our study region. We predict that this landscape is

fragmented for the little dispersing calcareous grass-

land specialist T. acteon resulting in considerable ge-

netic differentiation among the samples analysed. On

the other hand, the availability of many habitats for

T. lineola and its stronger dispersal capacity might

make the study region a more or less continuous hab-

itat for this species, with no or only rather weak genetic

differentiation. As T. sylvestris is a habitat generalist as

T. lineola, but a bad disperser as T. acteon, the habitat

availability for this species might not be sufficient to

allow continuous gene flow all over the study region,

but nevertheless should allow strong gene flow among

neighbouring patches, resulting in relatively weak and

maybe distance dependent genetic differentiation. If

so, the comparison of the genetic diversities among the

three species might allow predictions of the quality of

the analysed calcareous grassland areas for the con-

servation of xerothermophilic species. High conserva-

tion priority thus may be attributed to the areas with

genetically viable populations. Futhermore, action

plans may be elaborated for patches with less geneti-

cally diverse populations.

Materials and methods

Imagoes of three Thymelicus species (T. acteon,

T. sylvestris, T. lineola) were collected at 12 localities

(Fig. 1), and stored in liquid nitrogen until electro-

phoresis. The samples were taken in the Rhineland-

Palatinate (RP) and the Saarland (SL) (Western

Germany) and adjacent regions in Luxembourg and

France. We analysed two out group samples: Thyme-

licus sylvestris from Kresna Gorge (Bulgaria) and

Thymelicus lineola from Vallee des Glaciers (French

Alps). A total of 1,063 individuals (457 T. sylvestris,

440 T. lineola, 166 T. acteon) were analysed with

sample sizes ranging from 17 to 44 individuals. We

distinguished the sample locations in large and small

habitats with 15 ha being the threshold. More detailed

classifications were not possible as (i) the boundaries of

the habitats were not clear in all cases, (ii) the habitat

sizes vary among the three species and (iii) the popu-

lation densities are unknown.

Half of the abdomen of the imagoes were homog-

enised in Pgm-buffer (Harris and Hopkinson 1978) by

ultrasound and centrifuged at 8,000 g for 4 min. The

remaining tissue was stored for further analysis. We

ran electrophoreses on cellulose acetate plates (Hebert

672 Conserv Genet (2007) 8:671–681

123

and Beaton 1993). We analysed 15 enzyme systems

representing 18 loci (Table 1).

The alleles were labelled according to their relative

mobility, starting with ‘‘1’’ for the slowest. The allele

frequencies and genetic distances (Nei 1978) were cal-

culated with the package G-Stat (Siegismund 1993).

Hardy–Weinberg equilibrium (Louis and Dempster

1987), genetic disequilibrium (Weir 1991), locus by locus

F-statistics and AMOVA variance analyses were calcu-

lated with the package Arlequin 2.000 (Schneider et al.

2000) with FIS representing the genetic variance

component among individuals within populations, FST

the genetic variance component among populations and

FCT the genetic variance component among groups of

populations. The phenograms were calculated from Nei’s

(1978) genetic distances, using the package PHYLIP

(Felsenstein 1993). We used the neighbor joining method

(Saitou and Nei 1987). Mantel test were calculated using

XLSTAT version 2006.3 (Addinsoft 2006). Differences

between means were analysed by Friedmann ANOVAs,

Wilcoxon matched pairs tests or Man–Whitney U tests

using STATISTICA (Stat Soft inc. 1993).

Fig. 1 Location of the 12sample stations of Thymelicuslineola (black), T. sylvestris(white) and T. acteon (grey)in western Germany andadjoining France andLuxembourg. 1: Weinsheimbei Prum; 2: SchoneckerSchweiz; 3: Romerskopfchen;4: Scharren bei Bettingen; 5:Our-Tal bei Wallendorf; 6:Echternacherbruck; 7:Aarnescht bei Niederanven(L); 8: Perfeist beiWasserliesch; 9: Eiderberg beiFreudenburg; 10: Montenach(F); 11: Badstube beiMimbach; 12: Himsklamm beiNiedergailbach

Table 1 Electrophoresisconditions for the differentenzymes analysed for thethree Thymelicus species

TC: Tris–citrate pH 8.2(Richardson et al. 1986); TG:Tris–glycine pH 8.5 (Hebertand Beaton 1993); TM: Tris–maleic acid pH 7.0 (adjustedfrom TM pH 7.8 (Richardsonet al. 1986)). All buffers wererun at 200 V

Enzyme EC-No.

Number ofloci

Buffer Homogenateapplications

Running time(min)

MDH 1.1.1.37 2 TC 2 40G6PDH 1.1.1.49 1 TC 2 45ACON 4.2.1.3 1 TC 4 45MPI 5.3.1.8 1 TC 4 30AAT 2.6.1.1 2 TG 4 40FUM 4.2.1.2 1 TG 3 40PGI 5.3.1.9 1 TG 1 30ME 1.1.1.40 1 TG 3 40HBDH 1.1.1.30 1 TG 3 30APK 2.7.3.3 1 TG 2 30PGM 5.4.2.2 1 TG 2 406PGDH 1.1.1.44 1 TM 3 45IDH 1.1.1.42 2 TM 3 45GPDH 1.1.1.8 1 TM 4 40PEPPhe-Pro 3.4.11/13 1 TM 3 30

Conserv Genet (2007) 8:671–681 673

123

Results

All 18 loci were polymorphic for all species, with the

exception of ME in T. lineola. Furthermore, polymor-

phisms in GPDH of T. sylvestris were restricted to the

out group sample from Bulgaria.

No remarkable deviations from Hardy–Weinberg

equilibrium were detected. Only few cases of significant

deviations exist after Bonferroni correction (T. lineola:

6PGDH in Echternacherbruck (6); T. sylvestris: AAT1

in Wasserliesch (8) and Scharren bei Bettingen (4),

HBDH in Romerskopfchen (3) and Weinsheim bei

Prum (1); T. acteon: IDH1 in Wasserliesch (8), Scharren

bei Bettingen (4) and Echternacherbruck (6)). In gen-

eral, no linkage disequilibrium exists. However, some

evidence of linkage between PGI and FUM was found

for T. lineola. As FUM is mostly monomorphic (only

four polymorphic populations with a second allele

representing solely 1.3–2.6%), this possible linkage has

almost no consequences for further analyses. Therefore,

standard algorithms of population genetic analyse could

be applied.

The mean numbers of alleles per locus (A), the per-

centages of expected heterozygosity (He), the percent-

age of observed heterozygosity (Ho), the absolute

percentage of polymorphic loci (Ptot) and the percent-

age of polymorphic loci with the most common allele

not exceeding 95% (P95) are given in Table 2. These

values refer to the first 20 individuals of each population.

They vary considerably among populations (Appendix

Tables 5, 6, 7). Comparing the means of the five

parameters of genetic diversity revealed significant dif-

ferences among the three species (Friedmann ANOVA:

P = 0.022). While T. sylvestris and T. lineola showed no

significant differences (Wilcoxon test: P = 0.50), these

two species had significantly lower means than T. acteon

(both Wilcoxon tests: P = 0.043). The genetic diversities

of the out groups for T. lineola (Vallee des Glaciers,

French Alps) and T. sylvestris (Kresna Gorge, south-

west Bulgaria) were not significantly different from the

means of the respective Central European samples

(both Wilcoxon tests: P > 0.10).

Separate statistical tests (Friedmann ANOVAs) of

the single parameters revealed no significant difference

for the mean number of alleles per locus among the

three species, but significance for the four other

parameters. Thymelicus lineola had lower means than

T. sylvestris and T. acteon for He and Ho, and the

means for Ptot and P95 were higher in T. acteon than in

the other two species (Table 2).

The size of the habitat influenced the genetic diver-

sity of the respective population only for T. acteon, but

not for the two other skipper species (Table 3). The

Lulworth Skipper had significantly higher means for the

mean number of alleles per locus (A), total percentage

of polymorphic loci (Ptot) and percentage of polymor-

phic loci with the most commonest allele not exceeding

95% (P95) and a marginally significantly higher mean

for the observed heterozygosity (Ho) on habitats of

15 ha and more than on smaller patches.

The total genetic variance was highest for T. acteon,

intermediate for T. sylvestris and lowest for T. lineola

(Table 4). The differentiation among populations of

T. acteon was considerable (FST 5.1%; 20 individuals per

population: FST 5.3%), low for T. sylvestris (FST 1.6%;

20 individuals per population: FST 2.3%) and not

Table 2 Means of five parameters of genetic diversity for allpopulations analysed of three Thymelicus species excluding outgroups

T. acteon T. sylvestris T. lineola P

A 1.88 ± 0.18 1.80 ± 0.10 1.78 ± 0.17 0.5404He 14.9 ± 2.9a 11.9 ± 1.5a 9.6 ± 2.1b 0.0183Ho 12.5 ± 2.6a 11.0 ± 1.4a 9.2 ± 2.1b 0.0024Ptot 66.0 ± 9.1b 42.9 ± 7.9a 52.0 ± 9.7a 0.0151P95 49.3 ± 13.4b 32.3 ± 4.2a 36.4 ± 9.4a 0.0263

Tests for significant differences are performed using FriedmannANOVAs. Individual pairs were tested by U-test and significantdifferences (P < 0.05) are indicated by different characters. Alldata are given for the first 20 individuals of the respective sample

Abbreviation: number of alleles per locus (A), percentage ofexpected heterozygosity (He), percentage of observed hetero-zygosity (Ho), total percentage of polymorphic loci (Ptot), per-centage of polymorphic loci with the most commonest allele notexceeding 95% (P95)

Table 3 Influence of the habitat size on the genetic diversity of the three Thymelicus species

T. acteon T. sylvestris T. lineola

A 1.97 ± 0.14 M 1.70 ± 0.03 (P = 0.024) 1.82 ± 0.08 M 1.78 ± 0.13 (P = 0.65) 1.81 ± 0.16 M 1.73 ± 0.19 (P = 0.118)He 16.0 ± 2.3 M 12.9 ± 3.1 (P = 0.101) 11.7 ± 0.8 M 12.1 ± 2.1 (P = 0.068) 10.5 ± 1.8 M 8.6 ± 2.1 (P = 0.78)Ho 13.6 ± 2.5 M 10.6 ± 1.7 (P = 0.053) 11.5 ± 1.0 M 10.3 ± 1.7 (P = 0.31) 9.5 ± 1.8 M 8.9 ± 2.6 (P = 0.23)Ptot 71.1 ± 7.3 M 57.4 ± 3.2 (P = 0.037) 41.7 ± 7.7 M 44.4 ± 8.8 (P = 0.78) 53.7 ± 9.7 M 50.0 ± 10.4 (P = 0.20)P95 56.7 ± 10.0 M 37.0 ± 8.5 (P = 0.034) 30.6 ± 3.0 M 43.4 ± 4.6 (P = 0.62) 36.1 ± 8.4 M 36.7 ± 11.5 (P = 0.64)

The sample sites are distinguished into large and small ones (threshold 15 ha). Means are given with their SD. The respective firstvalues represent the larger habitats, the second the smaller ones. The P values of U-test are given in the parentheses

674 Conserv Genet (2007) 8:671–681

123

significant for T. lineola (Table 4). The FIS values

were relatively high for all species (T. acteon 15.5%,

20 individuals per population: 15.9%; T. lineola 4.7%,

20 individuals per population: 4.4%; T. sylvestris 5.0%,

20 individuals per population: 7.5%). The two out

groups showed a strong differentiation from the Central

European samples (T. sylvestris FCT 7.7% and T. lineola

FCT 14.0%).

We calculated Nei’s (1978) genetic distance

(means ± SD: T. acteon: 0.0402 ± 0.0096, T. sylvestris:

0.0176 ± 0.0023 and T. lineola: 0.0173 ± 0.0037). For

better comparison between species, we calculated

these distances on the basis of the first 20 individuals of

all populations, too (means ± SD: T. acteon: 0.0412 ±

0.009, T. sylvestris: 0.0324 ± 0.0026 and T. lineola:

0.0293 ± 0.0014). These means all differed significantly

from each other (U-test: all P < 0.0001).

We calculated phenograms based on Nei’s (1978)

genetic distances for the three species (not shown). The

tree topologies for T. acteon and T. lineola did not

coincide with the geographical distributions of the

sample sites and no isolation by distance exists (both

Mantel tests: P > 0.25). In contrast, geographical and

genetic data sets were correlated for T. sylvestris, and

20.3% of the genetic differentiation was explained by

isolation by distance (Mantel test: r = 0.452, P = 0.001,

Fig. 2).

Discussion

Genetic diversity and differentiation

The population genetic diversity of the three analysed

Thymelicus species was in the range of values typically

observed for butterflies and moths (cf. Graur 1985;

Packer et al. 1998; Vandewoestijne et al. 1999; Schmitt

et al. 2002). None of the parameters analysed reached

the average diversities found in the highly diverse

butterfly family of the Blues (Lycaenidae) (e.g.,

Lelievre 1992; Peterson 1995; Marchi et al. 1996;

Brookes et al. 1997; Schmitt and Seitz 2001; Schmitt

et al. 2003), but the observed values were considerably

higher than in genetically impoverished taxa as species

of the genus Yponomenta (Menken 1987) or Aglaope

infausta (Schmitt and Seitz 2004). In comparison with

Nymphalids, the studied Thymelicus species showed

population genetic diversities that were little less than

in common and widespread species of this family (e.g.,

Porter et al. 1995; Vandewoestijne et al. 1999; Schmitt

et al. 2005), but considerably higher than in rare and

very localised taxa (e.g., Debinski 1994; Britten et al.

1994, 1995; Pelz 1995). The observed genetic diversity

is in accordance with the commonness and broad dis-

tribution of T. sylvestris and T. lineola. However, it is

surprising that the localised T. acteon has similar or

sometimes even higher genetic diversities of its analy-

sed populations than the other skipper species (see

below).

The genetic differentiations among populations of

the three species analysed vary considerably. Thus, the

highest FST value and the highest mean genetic dis-

tance was found for T. acteon, which has a rather

scattered island-like distribution in the study region.

Comparable values were found in other regional

studies for species with similarly scattered distribution

patterns (Debinski 1994; Britten et al. 1994; Johanne-

sen et al. 1996; Gadeberg and Boomsma 1997; Schmitt

and Seitz 2004). Most probably, the gene flow among

the studied samples of T. acteon is very limited. On the

Table 4 Detailed varianceanalyses for the three skipperspecies for all individualsexcluding the out group (all)and for the first 20 individualsof each population excludingthe out group (20)

T. acteon T. sylvestris T. lineola

Variance among populations (all) 0.0718 0.0179 0.0081FST 5.1% 1.6% 0.9%P <0.0001 <0.0001 0.21Variance among populations (20) 0.0755 0.0256 0.0077FST 5.3% 2.3% 0.8%P <0.0001 0.011 0.86Variance among individuals

within populations (all)0.20926 0.05410 0.04134

FIS 15.5% 5.0% 4.7%P <0.0001 0.003 0.0068Variance among individuals

within populations (20)0.21342 0.08004 0.0386

FIS 16.0% 7.5% 4.4%P <0.0001 <0.0001 0.034Variance within individuals (all) 1.1386 1.0250 0.8433Variance within individuals (20) 1.1250 0.9868 0.8313Total variance (all) 1.4197 1.0970 0.8927Total variance (20) 1.4139 1.0924 0.8775

Conserv Genet (2007) 8:671–681 675

123

contrary, the values for T. sylvestris and especially

T. lineola were low or even non-significant as typically

found for other common and widespread species

(Eanes and Koehn 1978; Hughes and Zalucki 1984;

Daly and Gregg 1985; Zalucki et al. 1987; Goulson

1993; Korman et al. 1993; Bossart and Scriber 1995;

Porter and Geiger 1995; Nibouche et al. 1998).

Comparing T. sylvestris and T. lineola, the genetic

differentiation among populations measured by FST

and genetic distances of the latter is less than in

T. sylvestris. This cannot be a function of habitat

availability because (i) both species thrive on a wide

variety of different grasslands (Ebert and Rennwald

1991; Asher et al. 2001) and (ii) co-occur in most of

their habitats. However, the dispersal ability (Bink

1992) differs between the two species with T. sylvestris

dispersing worse than T. lineola. Most probably,

T. lineola has a mostly panmictic population structure

all over our study region, whereas the relatively weak

but significant differentiation among the T. sylvestris

populations suggest a metapopulation system with

moderate to high gene flow rates between neighbour-

ing populations.

Thymelicus and isolation by distance

In our study, we found a moderate genetic differenti-

ation and isolation by distance explaining about 20%

of the observed genetic differentiation for T. sylvestris.

However, no correlation between geographic and ge-

netic distance was detected for the other species, nei-

ther for the highly isolated T. acteon with high genetic

differentiation among populations nor for the pan-

mictic T. lineola.

The high availability of suitable habitat patches

combined with the relatively low dispersal ability of

T. sylvestris cause that most probably only neighbour-

ing habitat patches are located within the normal dis-

persal distance of the butterflies (Fig. 3b). Therefore,

the genetic influence of one patch on another one de-

creases with distance. Over the generations, all patches

are interconnected by gene flow with the geographical

distances being at least partly responsible for the

amount of genetic exchange, hereby leading to an

isolation by distance system.

The situation of T. acteon differs from T. sylvestris

insofar as the former is a habitat specialist restricted to

dry and hot grasslands in Central Europe (Ebert and

Rennwald 1991; Settele et al. 1999). The habitat

availability for the Lulworth Skipper is therefore so

Fig. 2 Correlation between the geographical distances and therespective genetic distances (Nei 1978) of the populations of (a)Thymelicus lineola (r2 = 0.022; Mantel test: P = 0.282), (b) T.sylvestris (r2 = 0.204; Mantel test: P = 0.001) and (c) T. acteon(r2 = 0.013; Mantel test: P = 0.560). For T. sylvestris, theregression with 95% confidence intervals is given

676 Conserv Genet (2007) 8:671–681

123

restricted that the butterfly’s dispersal ability is not

sufficient to establish gene flow over a network of

habitats involving the whole study region; rather, even

sub-structuring of the analysed samples is probable,

maybe explaining the relatively high FIS value of 15%.

Therefore, isolation by distance might be possible in

well connected habitat networks on a local scale, but

will break down on a regional level if suitable habitats

are too distant from one another (Fig. 3c). As conse-

quence, isolated populations develop independently

resulting in relatively high rates of differentiation

among samples without mirroring the geographical

structures among populations (cf. Neve et al. 2000;

Jaggi et al. 2000; Schmitt and Seitz 2002). This is

underlined by the fact that small populations harbour

less genetic diversity in our studied Lulworth Skipper

populations, but not in the two other species.

As for T. acteon, no geographic structure was

detected for T. lineola, but due to completely different

reasons. First, the habitat availability of T. lineola is

much higher than for T. acteon, and second, the dis-

persal ability (Bink 1992) is higher than in the two

other skipper species so that dispersing Essex Skippers

might not only reach neighbouring patches but also

more distant ones with the result of massive gene flow

on a regional level (Fig. 3a).

Conservation implications

Two of the three analysed skipper species (T. sylvestris

and T. lineola) are widespread and rather common all

over Europe and are not of conservation concern (van

Swaay and Warren 1999). The situation of T. acteon is

completely different. This species is strongly declining

in most countries of Europe and is listed in the Euro-

pean Red Data Book as vulnerable (van Swaay and

Warren 1999), making T. acteon a species of major

conservation concern.

It is widely accepted that a species-specific level of

genetic diversity is necessary for the viability of its pop-

ulations (Frankham et al. 2002; Hansson and Westerberg

2002; Reed and Frankham 2003; Schmitt and Hewitt

2004) with many examples proffered (e.g., Saccheri et al.

1998; Westermeier et al. 1998; Bryant et al. 1999;

Madsen et al. 1999; Meagher 1999; Rowe et al. 1999;

Buza et al. 2000; Luijten et al. 2000; Ujvari et al. 2002).

Therefore, the results of T. acteon seem to carry an

optimistic message: The genetic diversity of the analysed

population is as high or even higher than for the two

common skipper species. As even more southern sam-

ples of T. sylvestris and T. lineola did not have higher

genetic diversity as the German T. acteon samples, major

bottlenecks are rather unlikely for our studied T. acteon

populations. Therefore, complete extinction due to

population genetic reasons is unlikely.

However, such conclusions should be taken with

caution. The high FST value indicates that genetic ex-

change between habitats is rare or missing for T. acteon

and strong decline of butterfly species was observed in

the study region over the last three decades due to

environmental changes (Wenzel et al. 2006). Further-

more, we detected genetic impoverishment of the

smaller populations, as frequently observed for other

organisms (e.g., Billington 1991; Buza et al. 2000;

Hudson et al. 2000; Jaggi et al. 2000; Madsen et al.

2000).

If one of the T. acteon populations becomes extinct

(e.g., by management mistakes, stochastic events like

parasitism, weather etc.), recolonisation from one of

the other patches is unlikely due to the high degree of

habitat fragmentation. Therefore, the conservation

status and demographic threats of T. acteon in our

study region might be worse than its genetic diversities

imply. The survival of each single population might

completely dependent on the habitat management, and

management mistakes might decrease biodiversity

without realistic perspectives of rapid recovery by re-

colonisation. A strict conservation of all remaining

calcareous grassland is therefore necessary for the

species’ survival, and the development of new or the

restoration of former calcareous grassland would fur-

ther safeguard its populations and the ones of other

rare and endangered animal and plant species.

Acknowledgements This work was supported by the GermanScience Foundation (Deutsche Forschungsgemeinschaft, Grant

suitable habitat

dispersal capacity of an individual

a c

patch of origin of an individual

b

suitable habitat

dispersal capacity of an individual

patch of origin of an individual

Fig. 3 Dispersal ability and habitat availability determine thegenetic structure of species. (a) High dispersal capacity and highhabitat availability result in intensive gene flow and a panmicticpopulation structure (T. lineola Type). (b) Lower dispersalcapacity in a landscape with high habitat availability reducesgene flow so that an isolation by distance system will establish(T. sylvestris Type). (c) Limited habitat availability in combina-tion with low dispersal capacity results in complete isolation withgenetic drift acting independently in each single population(T. acteon Type)

Conserv Genet (2007) 8:671–681 677

123

No. SCHM 1659/3-1 and 3-2). We also acknowledge the schol-arship for D. Louy of the Ministry for the environment and forestsof the Rhineland-Palatinate and local authorities in Saarbrucken,

Koblenz, Luxembourg and Metz for the permits to collect but-terflies and to work in several protected areas. We thank D. Kime(La Fontaine) for the correction of our English.

Table 6 Five parameters of genetic diversity for all populations analysed of Thymelicus lineola. The data for the out group Vallee desGlaciers (F) is given separately at the bottom of the table. For further information see Appendix Table 5

Sampling site A He (%) Ho (%) Ptot (%) P95 (%) N Sizecategory

Dates ofcapture

DRP Weinsheimbei Prum

2.22 (2.00) 11.3 (11.9) 11.7 (12.2) 72.2 (66.7) 50.0 (55.6) 39 1 10-VII-2003

DRP SchoneckerSchweiz

1.89 (1.61) 9.3 (9.0) 8.6 (8.4) 55.6 (44.4) 33.3 (38.9) 39 2 10-VII-2003

DRP Romerskopfchen 1.89 (1.50) 7.3 (6.2) 6.4 (5.8) 55.6 (38.9) 22.2 (27.8) 40 1 06-VII-2004DRP Scharren bei

Bettingen1.94 (1.72) 8.0 (7.5) 7.9 (7.0) 55.6 (44.4) 33.3 (27.8) 39 1 06-VII-2004

DRP Our-Tal beiWallendorf

2.22 (1.83) 9.5 (8.6) 9.8 (9.2) 72.2 (50.0) 33.3 (38.9) 44 1 06/21/28-VII-2004

DRP Echternacherbruck 2.39 (1.94) 10.5 (9.9) 9.4 (8.9) 77.8 (61.1) 33.3 (27.8) 40 2 06-VII-2004L Aarnescht bei

Niederanven1.72 (1.72) 9.8 (9.8) 9.5 (9.5) 38.9 (38.9) 33.3 (33.3) 20 2 20-VII-2004,

05-VIII-2004DRP Perfeist bei

Wasserliesch2.00 (1.83) 11.8 (12.2) 10.0 (10.1) 66.7 (61.1) 33.3 (38.9) 24 2 28-VI-2004,

4/5/18-VII-2004

DRP Eiderberg beiFreudenburg

2.39 (2.06) 12.9 (13.3) 12.2 (12.5) 72.2 (61.1) 38.9 (50.0) 41 2 17/23-VII-2004

DSL Badstube beiMimbach

1.83 (1.61) 8.5 (8.8) 8.6 (10.3) 55.6 (50.0) 27.8 (33.3) 40 1 19/21-VII-2004

DSL Himsklammbei Niedergailbach

2.06 (1.72) 10.0 (8.8) 9.3 (7.3) 61.1 (55.6) 27.8 (27.8) 34 2 19/21-VII-2004

Average ± SD 2.05 ± 0.23 9.90 ± 1.67 9.40 ± 1.62 62.1 ± 11.3 33.3 ± 7.0 36.4 ± 7.5Average 20 ± SD 1.78 ± 0.17 9.64 ± 2.11 9.20 ± 2.07 52.0 ± 9.7 36.4 ± 9.4F Vallee des Glaciers 2.06 (1.78) 14.5 (13.3) 12.5 (10.9) 72.2 (50.0) 38.9 (38.9) 40 03-VIII-2003

Appendix

Table 5 Five parameters of genetic diversity for all populationsanalysed of Thymelicus acteon: Mean number of alleles per locus(A), percentage of expected heterozygosity (He), percentage ofobserved heterozygosity (Ho), total percentage of polymorphicloci (Ptot), percentage of polymorphic loci with the most

commonest allele not exceeding 95% (P95), the number ofindividuals analysed (N) and the dates of capture. Values inparenthesis refer to the first 20 individuals of the respectivesample. The ‘‘average 20’’ gives the average of the first 20individuals of each sample

Sampling site A He (%) Ho (%) Ptot (%) P95 (%) N Sizecategory

Dates ofcapture

DRP Romerskopfchen 1.72 (1.72) 10.9 (10.9) 10.0 (10.0) 61.1 (61.1) 27.8 (27.8) 20 1 7/8/12-VIII-2004DRP Scharren bei

Bettingen1.67 (1.67) 11.3 (11.3) 9.3 (9.3) 55.6 (55.6) 38.9 (38.9) 20 1 7/8/12-VIII-2004

DRP Our-Tal beiWallendorf

1.72 (1.72) 16.5 (16.5) 12.6 (12.6) 55.6 (55.6) 44.4 (44.4) 17 1 21/28-VII-2004,11-VIII-2004

DRPEchternacherbruck

1.78 (1.78) 12.5 (12.5) 10.5 (10.5) 72.2 (72.2) 50.0 (50.0) 19 2 21/28-VII-2004,11-VIII-2004

L Aarnescht beiNiederanven

2.06 (2.00) 17.2 (16.7) 14.5 (13.5) 77.8 (77.8) 50.0 (55.6) 29 2 05-VIII-2004

DRP Perfeist beiWasserliesch

2.00 (2.00) 16.2 (16.6) 12.5 (12.8) 61.1 (61.1) 44.4 (44.4) 22 2 24-VIII-2004,2/4-VIII-2004

F Montenach 2.17 (2.17) 18.9 (18.9) 17.5 (17.5) 77.8 (77.8) 66.7 (66.7) 19 2 17/23-VII-2004

DSL Himsklammbei Niedergailbach

1.94 (1.94) 15.5 (15.5) 13.6 (13.6) 66.7 (66.7) 66.7 (66.7) 20 2 17-VII-2004,03-VIII-2004

Average ± SD 1.88 ± 0.19 14.88 ± 2.94 12.56 ± 2.69 66.0 ± 9.1 48.6 ± 13.2 20.8 ± 3.6Average 20 ± SD 1.88 ± 0.18 14.86 ± 2.92 12.48 ± 2.61 66.0 ± 9.1 49.3 ± 13.4

Abbreviations: Sampling site: D, Germany; L, Luxembourg; F, France; RP, Rhineland-Palatinate; SL, Saarland; BG, Bulgaria

678 Conserv Genet (2007) 8:671–681

123

References

Asher J, Warren M, Fox R, Harding P, Jeffcoate G, Jeffcoate S(2001) The millennium atlas of butterflies in Britain andIreland. Oxford University Press, Oxford

Billington HL (1991) Effects of population size on geneticvariation in a dioecious conifer Halocarpus bidwillii. Con-serv Biol 5:115–119

Bink FA (1992) Ecologische Atlas van de Dagvlinders vanNoordwest-Europa. Schuyt & Co. Uitgevers en Importeurs,Haarlem

Bossart JL, Scriber JM (1995) Maintenance of ecologicallysignificant genetic variation in the tiger swallowtail butterflythrough differential selection and gene flow. Evolution49:1163–1171

Britten HB, Brussard PF, Murphy DD, Austin GT (1994)Colony isolation and isozyme variability of the western seepfritillary, Speyeria nokomis apacheana (Nymphalidae), inthe western Great Basin. Great Basin Nat 54:97–105

Britten HB, Brussard PF, Murphy DD, Ehrlich PR (1995) A testfor isolation-by-distance in Central Rocky Mountain andGreat Basin populations of Edith’s Checkerspot Butterfly(Euphydryas editha). J Heredity 86:204–210

Brookes MI, Graneau YA, King P, Rose OC, Thomas CD,Mallet JLB (1997) Genetic analysis of founder bottlenecksin the rare british butterfly Plebejus argus. Conserv Biol11:648–661

Buza L, Young A, Thrall P (2000) Genetic erosion, inbreedingand reduced fitness in fragmented populations of theendangered tetraploid pea Swainsona recta. Biol Conserv93:177–186

Bryant EH, Backus VL, Clark ME, Reed DH (1999) Experi-mental tests of captive breeding for endangered species.Conserv Biol 13:1487–1496

Daly JC, Gregg P (1985) Genetic variation in Heliothis inAustralia: species identification and gene flow in the twopest species H. armigera (Hubner) and H. punctigeraWallengren (Lepidoptera: Noctuidae). Bull Entomol Res75:169–184

Debinski DM (1994) Genetic diversity assessment in a meta-population of the butterfly Euphydryas gillettii. Heredity 70:25–30

Eanes WF, Koehn RK (1978) An analysis of genetic structure inthe monarch butterfly, Danaus plexippus L. Evolution32:784–797

Ebert G, Rennwald E (Hrsg.) (1991) Die Schmetterlinge Baden-Wurttembergs, Band 2. Verlag Eugen Ulmer, Stuttgart

Felsenstein J (1993) PHYLIP (Phylogeny inference package)Ver. 3.5.c. Department of Genetics, University of Washing-ton, Seattle, Washington

Frankham R, Ballou JD, Briscoe DA (2002) Introduction toconservation genetics. Cambridge University Press, Cam-bridge

Gadeberg RME, Boomsma JJ (1997) Genetic population struc-ture of the large blue butterfly Maculinea alcon in Denmark.J Insect Conserv 1:99–111

Goulson D (1993) Allozyme variation in the butterfly Maniolajurtina (Lepidoptera: Satyrinae) (L.): evidence for selection.Heredity 71:386–393

Graur D (1985) Gene diversity in Hymenoptera. Evolution39:190–199

Hanski I (1999) Metapopulation ecology. Oxford UniversityPress, Oxford

Hansson B, Westerberg L (2002) On the correlation betweenheterozygosity and fitness in natural populations. Mol Ecol11:2467–2474

Harris H, Hopkinson DA (1978) Handbook of enzyme electro-phoresis in human genetics. North-Holland, Amsterdam

Table 7 Five parameters of genetic diversity for all populations analysed of Thymelicus sylvestris. The data for the out group KresnaGorge (BG) is given separately at the bottom of the table. For further information see Appendix Table 5

Sampling site A He (%) Ho (%) Ptot (%) P95 (%) N Sizecategory

Dates of capture

DRP Weinsheimbei Prum

1.94 (1.72) 12.7 (14.4) 11.1 (12.6) 38.9 (38.9) 27.8 (38.9) 40 1 10-VII-2003

DRP SchoneckerSchweiz

1.83 (1.67) 12.0 (11.5) 12.4 (11.3) 50.0 (44.4) 27.8 (33.3) 40 2 10-VII-2003

DRP Romerskopfchen 2.11 (1.83) 12.8 (13.5) 11.0 (10.8) 55.6 (50.0) 27.8 (38.9) 38 1 06-VII-2004DRP Scharren

bei Bettingen2.00 (1.72) 12.3 (12.1) 10.8 (9.3) 38.9 (33.3) 27.8 (27.8) 40 1 01/06-VII-2004

DRP Our-Tal beiWallendorf

2.06 (2.00) 12.1 (11.4) 11.6 (10.8) 55.6 (55.6) 33.3 (33.3) 32 1 06/21/28-VII-2004

DRP Echternacherbruck 2.39 (1.83) 12.8 (12.8) 10.7 (10.3) 77.8 (50.0) 33.3 (33.3) 40 2 6/28-VII-2004L Aarnescht bei

Niederanven2.33 (1.89) 11.8 (10.5) 11.7 (10.7) 61.1 (50.0) 33.3 (27.8) 39 2 20-VII-2004,

05-VIII-2004DRP Perfeist bei

Wasserliesch2.06 (1.83) 10.9 (11.3) 10.8 (11.7) 44.4 (33.3) 27.8 (27.8) 40 2 28-VI-2004,

5/18-VII-2004DRP Eiderberg bei

Freudenburg2.11 (1.89) 11.6 (12.4) 11.2 (11.4) 50.0 (38.9) 27.8 (33.3) 39 2 17/23-VII-2004

DSL Badstube beiMimbach

1.83 (1.67) 9.7 (8.9) 9.6 (8.2) 50.0 (44.4) 27.8 (33.3) 35 1 19/21-VII-2004

DSL Himsklamm beiNiedergailbach

2.22 (1.78) 13.3 (11.8) 14.4 (13.3) 61.1 (33.3) 27.8 (27.8) 35 2 19/21-VII-2004

Average ± SD 2.08 ± 0.18 12.00 ± 1.01 11.39 ± 1.22 53.0 ± 11.2 29.3 ± 2.6 38.0 ± 2.8Average 20 ± SD 1.80 ± 0.10 11.87 ± 1.47 10.95 ± 1.41 42.9 ± 7.9 32.3 ± 4.2BG Kresna Gorge 2.33 (1.78) 12.7 (11.8) 13.9 (12.8) 72.2 (55.6) 38.9 (33.3) 39 16-VI-2004

Conserv Genet (2007) 8:671–681 679

123

Hebert PDN, Beaton MJ (1993) Methodologies for allozymeanalysis using cellulose acetat electrophoresis. Helena Lab-oratories, Beaumont, TX

Hudson QJ, Wilkins RJ, Waas JR, Hogg ID (2000) Low geneticvariability in small populations of New Zealand kokakoCallaeas cinerea wilsoni. Biol Conserv 96:105–112

Hughes JM, Zalucki MP (1984) Genetic variation in a contin-uously breeding population of Danaus plexippus L. (Lepi-doptera: Nymphalidae). Heredity 52:1–7

Jaggi C, Wirth T, Baur B (2000) Genetic variability in subpop-ulations of the asp viper (Vipera aspis) in the Swiss Juramountains: implications for a conservation strategy. BiolConserv 94:69–77

Johannesen J, Veith M, Seitz A (1996) Population geneticstructure of the butterfly Melitaea didyma (Nymphalidae)along a northern distribution range border. Mol Ecol 5:259–267

Johannesen J, Samietz J, Wallaschek M, Seitz A, Veith M (1999)Patch connectivity and genetic variation in two congenericgrasshopper species with different habitat preferences. JInsect Conserv 3:201–209

Keller I, Largiader CR (2003) Recent habitat fragmentationcaused by major roads leads to reduction of gene flow andloss of genetic variability in ground beetles. Proc Roy SocLondon B, Biol Sci 270:417–423

Korman AK, Mallet J, Goodenough JL, Graves JB, Hayes JL,Hendricks DE, Luttrell R, Pair SD, Wall M (1993) Popu-lation structure in Heliothis virescens (Lepidoptera: Noctui-dae): an estimation of gene flow. Ann Entomol Soc Am86:182–188

Kraus W (1993) Verzeichnis der Großschmetterlinge (Insecta:Lepidoptera) der Pfalz. Pollichia Buch 27 Selbstverlag derPollichia, Bad Durkheim

Lelievre T (1992) Phylogenie des Polyommatinae et structuregenetique de six especes du genre Lysandra, HEMMING(Lepidopteres Lycaenidae). Doktorarbeit (Universite deProvence), unpublished data, 221 S

Louis EJ, Dempster ER (1987) An exact test for Hardy–Weinberg and multiple alleles. Biometrics 43:805–811

Luijten SH, Dierick A, Gerard J, Oostermeijer B, RaijmannLEL, Den Nijs HCM (2000) Population size, geneticvariation, and reproductive success in a rapidly declining,self-compatible perennial (Arnica montana) in The Neth-erlands. Conserv Biol 14:1776–1787

Madsen T, Olsson M, Wittzell H, Stille B, Gullberg A, Shine R,Andersson S, Tegelstrom H (2000) Population size andgenetic diversity in sand lizards (Lacerta agilis) and adders(Vipera berus). Biol Conserv 94:257–262

Madsen T, Shine R,Olsson M, Wittzell H (1999) Restoration ofan inbred adder population. Nature 402:34–35

Marchi A, Addis G, Hermosa VE, Crnjar R (1996) Geneticdivergence and evolution of Polyommatus coridon gennar-genti (Lepidoptera, Lycaenidae) in Sardinia. Heredity77:16–22

Meagher S (1999) Genetic diversity and Capillaria hepatica(Nematoda) prevalence in Michigan deer mouse popula-tions. Evolution 53:1318–1324

Menken SBJ (1987) Is the extremely low heterozygosity level inYponomeuta rorellus caused by bottlenecks? Evolution41:630–637

Nei M (1978) Estimation of average heterozygosity and geneticdistance from a small number of individuals. Genetics89:583–590

Negedank J (1974) Trier und Umgebung. In: Gwinner M. (Hrsg.)Sammlung geologischer Fuhrer 60. Gebruder Borntraeger,Berlin

Neve G, Barascud B, Descimon H, Baguette M (2000) Geneticstructure of Proclossiana eunomia populations at theregional scale (Lepidoptera, Nymphalidae). Heredity84:657–666

Nibouche S, Bues R, Toubon JF, Poitout S (1998) Allozymepolymorphism in the cotton bollworm Helicoverpa armigera(Lepidoptera: Noctuidae): comparision of African andEuropean populations. Heredity 80:438–445

Packer L, Taylor JS, Savignano DA, Bleser CA, Lane CP,Sommers LA (1998) Population biology of an endangeredbutterfly, Lycaeides melissa samuelis (Lepidoptera; Lycae-nidae): genetic variation, gene flow, and taxonomic status.Can J Zool 76:320–329

Pedersen AA, Loeschcke V (2001) Conservation genetics ofperipheral populations of the mygalomorph spider Atypusaffinis (Atypidae) in northern Europe. Mol Ecol 10:1133–1142

Pelz V (1995) Biosystematik der europaischen Arten des TribusMelitaeini Newman, 1870. Oedippus 11:1–62

Peterson MA (1995) Phenological isolation, gene flow anddevelopmental differences among low- and high-elevationpopulations of Euphilotes enoptes (Lepidoptera: Lycaeni-dae). Evolution 49:446–455

Porter AH, Geiger H (1995) Limitations to the inference of geneflow at regional geographioc scales – an example from thePieris napi group (Lepidoptera: Pieridae) in Europe. Biol JLinnean Soc 54:329–348

Porter AH, Schneider RW, Price BA (1995) Wing pattern andallozyme relationship in the Coenonympha arcania group,emphasising the C. gardetta-darwiniana contact area atBellwald, Switzerland (Lepidoptera, Satyridae). NotaLepidopterologica 17:155-174

Reed DH, Frankham R (2003) Correlation between fitness andgenetic diversity. Conserv Biol 17:230–237

Richardson B, Baverstock PR, Adams M (1986) Allozymeelectrophoresis. A handbook for animal systematics andpopulation studies. Academic Press, San Diego

Rosenzweig ML (1995) Species diversity in space and time.Cambridge University Press, Cambridge

Rowe G, Beebee TJC, Burke T (1999) Microsatellite heterozy-gosity, fitness, and demography in natterjack toads Bufocalamita. Animal Conserv 2:85–92

Saccheri I, Kuussaari M, Kankare M, Vikman P, Fortelius W,Hanski I (1998) Inbreeding and extinction in a butterflymetapopulation. Nature 392:491–494

Saitou N, Nei M (1987) The neighbor-joining method: a newmethod for reconstructing phylogenetic trees. Mol Biol Evol4:406–425

Schmidt-Koehl W (1977) Die Groß-Schmetterlinge des Saarlan-des (Insecta, Lepidoptera): Tagfalter, Spinner und Schwar-mer. Abhandlungen der Arbeitsgemeinschaft fur tier- undpflanzengeografische Heimatforschung im Saarland 7:1–231

Schmitt T, Hewitt G (2004) The genetic pattern of populationthreat and loss: a case study of butterflies. Mol Ecol 13:21–31

Schmitt T, Seitz A (2001) Allozyme variation in Polyommatuscoridon (Lepidoptera: Lycaenidae): identification of ice-agerefugia and reconstruction of post-glacial expansion. JBiogeogr 28:1129–1136

Schmitt T, Seitz A (2002) Influence of habitat fragmentation onthe genetic structure of Polyommatus coridon (Lepidoptera:Lycaenidae): implications for conservation. Biol Conserv107:291–297

Schmitt T, Seitz A (2004) Low diversity but high differentiation:the population genetics of Aglaope infausta (Zygaenidae:Lepidoptera). J Biogeogr 31:137–144

Schmitt T, Gießl A, Seitz A (2002) Postglacial colonisation ofwestern Central Europe by Polyommatus coridon (Poda

680 Conserv Genet (2007) 8:671–681

123

1761) (Lepidoptera: Lycaenidae): evidence from populationgenetics. Heredity 88:26–34

Schmitt T, Gießl A, Seitz A (2003) Did Polyommatus icarus(Lepidoptera: Lycaenidae) have distinct glacial refugia insouthern Europe? – evidence from population genetics. BiolJ Linnean Soc 80:529–538

Schmitt T, Rober S, Seitz A (2005) Is the last glaciation the onlyrelevant event for the present genetic population structureof the Meadow Brown butterfly Maniola jurtina (Lepidop-tera: Nymphalidae)? Biol J Linnean Soc 85:419–431

Schneider S, Roessli D, Excoffier L (2000) Arlequin ver. 2.000 –a software for population genetics data analysis. Anthro-pology, University of Geneve

Settele J, Feldmann R, Reinhardt R (1999) Die TagfalterDeutschlands – Ein Handbuch fur Freilandokologen, Um-weltplaner und Naturschutzer. Ulmer, Stuttgart

Siegismund HR (1993) G-Stat, ver. 3, Genetical statisticalprograms for the analysis of population data. The Arbore-tum, Royal Veterinary and Agricultural University, Hors-holm, Denmark

Tolman T, Lewington R (1998) Die Tagfalter Europas undNordwestafrikas. Franckh-Kosmos Verlag, Stuttgart

Ujvari B, Madsen T, Kotenko T, Olsson M, Shine R, Wittzell H(2002) Low genetic diversity threatens imminent extinctionfor the Hungarian meadow viper (Vipera ursinii rakosien-sis). Biol Conserv 105:127–130

Vandewoestijne S, Neve G, Baguette M (1999) Spatial andtemporal population genetic structure of the butterfly Aglaisurticae L. (Lepidoptera, Nymphalidae). Mol Ecol 8:1539–1543

van Swaay CAM, Warren M (1999) Red data book of Europeanbutterflies (Rhopalocera). Nature and environment 99,Council of Europe Publishing, Strasbourg

Weir BS (1991) Genetic data analysis. Sinauer, Sunderland, MAWenzel M, Schmitt T, Weitzel M, Seitz A (2006) The severe

decline of butterflies on western German calcareous grass-lands during the last 30 years: a conservation problem. BiolConserv 128:542–552

Westemeier RL, Brawn JD, Simpson SA, Esker TL, Jansen RW,Walk JW, Eric L, Kershner EL, Bouzat JL, Paige KN (1998)Tracking the long-term decline and recovery of an isolatedpopulation. Science 282:1695–1698

Williams BL, Brawn JD, Paige KN (2003) Landscape scalegenetic effects of habitat fragmentation on a high gene flowspecies: Speyeria idalia (Nymphalidae). Mol Ecol 12:11–20

Young A, Boyle T, Brown T (1996) The population geneticconsequences of habitat fragmentation for plants. TrendsEcol Evol 11:413–418

Zalucki MP, Hughes JM, Cater PA (1987) Genetic variation inDanaus plexippus L.: habitat selection or differences inactivity time? Heredity 59:213–221

Conserv Genet (2007) 8:671–681 681

123