Populationfragmentationleadstoisolationby

distancebutnotgeneticimpoverishmentinthe philopatricLesserKestrel:acomparisonwith

thewidespreadandsympatricEurasianKestrel

MAlcaide1,DSerrano1,JJNegro1,JLTella1,TLaaksonen2,CMu¨ller3,AGal4 and EKorpima¨ki2

1DepartamentosdeEcolog´ıaEvolutivayBiolog´ıadelaConservacio´n,Estacio´n Biolo´gicadeDon˜ana(CSIC),Sevilla,Spain;2Department

ofBiology,SectionofEcology,UniversityofTurku,Turku,Finland;3SwissOrnithologicalInstitute, Sempach,Switzerlandand

4DepartmentofEvolution,SystematicsandEcology,TheHebrewUniversityofJerusalem,Jerusalem, Israel

Population fragmentationisawidespreadphenomenon usuallyassociatedwith humanactivity.Asaresult ofhabitat transformation,thephilopatric andsteppe-specialistLesser KestrelFalconaumanni underwentaseverepopulation decline during thelastcenturythatincreasedpopulation fragmentationthroughoutitsbreedingrange.In contrast,the ubiquitousEurasianKestrelFalcotinnunculusdidnotsuffer suchadverseeffects,itsbreedingrangestillremainingrather continuous.Usingmicrosatellites,wetestedtheeffectsof populationfragmentationonlarge-scalespatialpatternsof geneticdifferentiation anddiversitybycomparing thesetwo sympatricandphylogeneticallyrelatedspecies.Ourresults suggestthathabitatfragmentation hasincreasedgenetic differentiationbetweenLesserKestrelpopulations,following

anisolation-by-distance pattern,whilethepopulation of EurasianKestrelsispanmictic.Contrarytoexpectations, wedidnotdetectsignificantevidence ofreducedgenetic variationorincreasedinbreeding inLesserKestrels. Althoughthisstudyreportsgeneticdifferentiation ina speciesthathaspotentialforlong-distance dispersalbut philopatry-limited geneflow,largeenougheffectivepopula- tionsizesandmigrationmayhavebeensufficientto mitigate geneticdepauperation. Aseriousreductionofgenetic diversityinLesserKestrelswould,therefore, onlybe expectedafterseverepopulation bottlenecksfollowing extremegeographicisolation.

Keywords:geneticdiversity;habitatfragmentation;geneticstructure;steppelandbirds;dispersal;geneflow

Introduction

Humanactivities transformthenaturalhabitatsofmany species. Populationfragmentationoften leads tooverall reductionsinpopulation sizesand diminishesconnec- tivity amonghabitatpatches. Althoughpopulation fragmentationincreases extinctionrisksbecause of deterministic and stochasticeffectsondemographic parameters,restrictedgene flow may jeopardizelong- term persistence ofpopulations due toinbreeding depression and loss of genetic diversity. Both demo- graphicand geneticimpactsofpopulationfragmentation are believedtodependonthe number,size and spatial distributionofpopulations aswellasontimesince fragmentation.In this regard,dispersaland associated gene flowappearsoneofthemost critical factors influencing thegenetic structure and demographyof fragmentedpopulations(forexample,YoungandClarke,

2000; Frankham et al., 2002). However, restricted gene flow and the subsequent emergence of genetic

Correspondence:Dr M Alcaide,Estacio´nBiolo´gicadeDon˜ana(CSIC), Avda.MaLuisas/n,41013Sevilla,Spain.

E-mail:

Received 23 April 2008;revised2 September2008;accepted12

September2008;publishedonline 15October 2008

structuringisnot only the result ofphysicalbarriersor anatomicalimpedimentstolong-distance movement. Natal and breedingphilopatry(that is,the tendencyof individuals tobreed closetotheir birthplaceortheir previous breeding territory) is expected to enhance theeffectsofhabitatfragmentation (forexample, Greenwood,1980). Geneticdifferentiationamongfrag- ments ishence expectedtobeinverselycorrelatedwith thedispersalability ofthespecies.

In spite of all these considerations, there is not necessarily a direct association between the spatial

distributionofpopulationsand the spatial distribution

ofgenetic diversity(forexample,Dannewitzetal.,2005; Jones et al., 2007; Koopman et al., 2007). Combined

demographic and genetic investigations are therefore being encouraged to rigorously evaluate the conse-

quencesofpopulationfragmentation(for example, Koenig and Dickinson,2004).Inthisrespect, elucidating

the demographicand ecological factors that determine thedistributionofgeneticvariationatdifferentscaleshas

become fundamentalto research in conservationand evolutionary biology. Polymorphicmolecularmarkers

and powerfulstatistical methodshave allowedthe investigation of the spatial distribution of genetic

variation in fragmented populations and provided a

measureof population connectivity.Such approaches,

combinedwith life-historyand demographic informa-

tion, have consistentlyprovidedrelevantdata tounder-

pin conservation and management initiatives aimed at preserving the genetic diversity of endangered

species (for example, Caizergues et al., 2003; Mart´ınez-Cruz et al., 2004; Hansson and Richardson,

2005;Koopmanetal.,2007).

Studies ofgenetic structureand diversityin birds of

preyareaccumulatingdue toanemergingconcern about the threatsderivedfrom populationfragmentationand

habitatalteration inthischarismaticavian group(for example,Godoy etal.,2004;Mart´ınez-Cruzetal.,2004; Helbig etal.,2005;Brownetal.,2007;Cadah´ıaetal.,2007; Hailer etal.,2007; Nittingeretal.,2007).Birdsofprey typicallyhavesmallpopulations,extendeddistributional rangesand they usuallyhave long-distance dispersal capabilities. Althoughraptorpopulationstend to be poorly structured (seereferencesabove),habitatfrag- mentationcould potentiallyincrease genetic divergence amongpopulationsand reducedpopulationsizewould initiate a loss of genetic variation. In this study, we

employedpolymorphicmicrosatellites toassess the influence ofpopulationfragmentationongenetic diver-

sity and large-scale (continental) spatial patterns of

genetic differentiation in two phylogenetically related and sympatricbirds of prey, the Lesser Kestrel Falco

naumanni and the Eurasian Kestrel Falcotinnunculus. Bothspecies breed inEurasia, acontinentalmass with a

broadtradition ofhuman-inducedlandscape transfor- mations, which have generated serious threats for

the conservationof many species (Goriupand Batten,

1990;McNeely, 1994).Althoughthe Lesser Kestrel isa

specialistfalconinhabiting steppeand pseudosteppe ecosystems(Crampand Simmons,1980),the Eurasian

Kestrelisconsideredatruly cosmopolitanfalconthatcan

liveinmost open-countryenvironments(Village, 1990). Open habitatsinEuropehave increaseddue toagricul-

ture and clear-cuttingofforests, afactthat may explain why the breedingrange ofthe EurasianKestrel has not

beendecisivelyaffected byhumanactivities. Incontrast, Lesser Kestrels have experienced a well-documented

populationdecline duringthe twentiethcenturythat is mostly explainedby humanperturbations,such as the

substitutionoftraditionalagriculturalpractices by intensive agricultureand irrigatedcrops that reduce foraginghabitats(Tella etal.,1998;Ursu´a etal.,2005). Such adramaticpopulationregressionled to the extirpationordisappearanceofthe Lesser Kestrel from several Europeancountries(Biber,1990).Itconsequently

hasapatchierdistributionalbreedingrange ascompared with its generalistcounterpart (Figure 1).In addition, long-term and extensive ringing studies of Lesser Kestrels inSpain have documentedhigh natal and breeding philopatryaswellasanegativeassociation betweeneffective dispersaland geographical distance (Negro etal.,1997; Serrano etal.,2001,2003,2008). Conversely,EurasianKestrels have shownalow philo- patryand frequenteffective long-distancedispersalsin populationsfrom Northernand WesternEurope(Korpi- ma¨ki,1988;Village, 1990;Korpima¨kietal.,2006;Vasko,

2007),althoughpreliminarydata fromaSpanishpopula-

tion suggesthigherphilopatryrates inSouthernEurope

(JAFargallo, personalcommunication).

Hence, themain questionthat thisstudywilladdress iswhetherhabitatalterationhas resultedinpopulation

Figure1Breedingdistributional ranges(grey areas) ofLesser (a) and Eurasian(b)Kestrels across theWesternPaleartic. Populations analysedinthis studyareindicatedbyblack dots. Lesser Kestrels were sampled from southwestern Spain (SWS), central-western Spain (CWS),northeasternSpain (NES),France (FRA),Italy (ITA), Israel (ISR)and Kazakhstan(KAZ).The continentalsubspeciesof the EurasianKestrel was sampledfrom SWS,CWS,NES,Switzer- land (SWI),Finland(FIN)and ISR.Inaddition,two subspeciesof theEurasian Kestrel inhabitingtheCanaryIslands(indicatedby asterisks)weresampled(FVforFalcotinnunculusdacotiaeandTFfor Falcotinnunculuscanariensis).

differentiationand lossofgenetic diversityinthehighly philopatric LesserKestrelcomparedwith thewidely distributedand highly dispersiveEurasianKestrel. The suitability of the genetic methodswe used here was tested bymeansofadditionalanalyses oftwo insular subspeciesoftheEurasianKestrel inhabitingtheCanary Islands.Weexpectedthepopulationsofthesesubspecies to hold comparably lower levels of genetic variation because ofthe well-documentedeffects ofinsularityon demographyand genetic diversity (forexample,Bollmer etal.,2005).

Materialsandmethods

Studyspeciesandpopulations

The Lesser Kestrel isa small trans-Saharianmigratory falcon whose breedingrange covers mid-latitudeand

low elevationsofEurasia (Crampand Simmons,1980). This colonial falcon originally occupied small cliffs

surroundedby naturalsteppes(Tella etal.,2004),but most pairs breed nowadaysin humanstructuressur-

roundedby traditionalagriculturalland. The Eurasian

Kestrel is a sedentaryor partiallymigratoryfalcon of

slightly larger size that is widespread in Eurasia, normally showing a territorial breeding behaviour

(Crampand Simmons,1980).In Europe,the estimated

populationsizeofLesser Kestrels isabout 25000–42000 breedingpairs,whereasthatofEurasianKestrelsisabout

300000–500000 breeding pairs. We analysedbreeding

populationsoftheLesserKestrel insouthwesternSpain,

central-westernSpain, northeasternSpain, France, Italy,

Greece and Israel (seeFigure 1a).Thecontinental subspecies of the Eurasian Kestrel (Falco tinnunculus

tinnunculus) wassampledinsouthwesternSpain,central- westernSpain, northeasternSpain, Switzerland, Finland

and Israel (seeFigure 1b).Twoinsularsubspeciesofthe

Eurasian Kestrel inhabiting the Canary Islands, Falco

tinnunculus canariensisand Falcotinnunculus dacotiae(see

Figure 1b),were also investigatedtoprovidecompara-

tive data. Estimatedpopulationsizes are about 400–500 breeding pairs for F. t. dacotiae and less than 4000

breedingpairs forF.t.canariensis(Madron˜oetal.,2004). The majority of sampled individuals (490%) were

nestlings, and we only analysed one individualper brood to minimize problems associated with close

relatedness.Extra-pairpaternityinLesser and Eurasian

Kestrels has showntoberare (below 7.5%ofnestlings,

see Korpima¨kietal.(1996)and Alcaide etal.(2005)for details), and thus, theprobabilityforadultmales toraise theirownoffspringishigh.Estimatedpopulationsizesof thegeographicallydistinctpopulationsofLesserKestrels investigatedin this studyare shownin Table 1. The numberofLesserand EurasianKestrels sampledateach location isshownin3and 4,respectively.

DNAisolationandmicrosatellitegenotyping

About 100ml of blood preserved in 96% ethanol or growingfeathersthat were pulledfromthebirds’dorsal plumagewere digestedbyincubationwith proteinaseK

for at least 3h. DNA purificationwas carriedout by

using 5M LiCl organic extractionmethodwith chloro- form/isoamylicalcohol (24:1)and furtherDNA precipi-

tation using absolute ethanol. Pellets obtained were driedand washedtwice with 70% ethanol,and later

storedat —201C in 0.1ml of TE buffer. We amplified seven microsatellitesthat were isolatedoriginallyinthe

peregrinefalcon Falcoperegrinus by Nesje etal.(2000) (Fp5, Fp13, Fp31, Fp46-1, Fp79-4, Fp89 and Fp107). In

addition,wedesignedtwo setsofprimersflanking two microsatellitesequences also isolatedin the peregrine

falcon that were availablein GenBank (AF448412and

AF448411, respectively). Locus Cl347 was amplified

using primers Cl347Fw: tgtgtgtgtaaggttgccaaaand

Cl347Rv:cgttctcaacatgccagttt.Locus Cl58was amplified

using primers Cl58Fw: tgtgtctcagtggggaaaaaand

Cl58Rv: tgctttggtgctgaagaaac.For each locus, the PCR

was carriedout in a PTC-100 Programmable Thermal

Controller(MJResearch Inc.,Waltham,MA,USA)using

thefollowingPCRprofile:35cyclesof40sat941C,40sat

Table1EstimatedpopulationsizesofLesser Kestrels sampledfor thisstudy

551C,40sat721Cand finally, 4min at721C.Each11ml

reaction contained 0.2U of Taq polymerase (Bioline,

London,UK), 1xPCR manufacturer-supplied buffer,

1.5mM MgCl2, 0.02%gelatine(AmershamLifeSciences, Buckinghamshire,UK),0.12mM ofeachdNTP, 5pmol of eachprimerand, approximately,10ngofgenomicDNA. Forwardprimers were50-endlabelledwithHEX,NEDor

6-FAM.Amplified fragmentswere resolvedonanABI Prism 3100 Genetic Analyser (Applied Biosystems, Foster City,CA,USA).

Geneticanalyses

Polymorphism statisticsat each microsatellitemarker

(that is, the numberof alleles and range size of the

amplified fragments)were calculatedusing thepro- grammeGenetix 4.04(Belkhir et al., 1996–2004).Con-

formity toHardy–Weinberg equilibriumwasanalysed throughGENEPOP (Raymondand Rousset, 1995),using

a single locus and a global multilocustest for hetero- zygositydeficit orexcess bythe MarkovChain Method

(Raymondand Rousset, 1995).

WeemployedthesoftwareSTRUCTURE2.2(Pritchard

etal.,2000)totestforthepresenceofgeneticallydistinct clusterswithinour studysystem. We did not use any

prior informationabout thegeographic origin ofthe individuals,and we assumedcorrelatedallele frequen-

ciesand theadmixturemodel. Tensimulations were performedforeach ofthe Kvalues rangingfrom 1to6

(that is, the number of putatively different genetic

clusters), and probability values of the data, that is ln Pr(X/K), were plotted. Values of K¼1 indicate a

geneticallyuniform population, while values of K¼2 and soonindicatethe existence ofgeneticallydifferent

arrays of individuals. Analyseswere carriedout with

100000iterations,followingaburn-inperiodof10000

iterations. Nonetheless, testing fordifferencesinallele frequenciesbetweengeographicallydistinctpopulations

may bemore useful than clusteringanalysesperformed in STRUCTURE when genetic differentiation is weak

(forexample,Latch etal.,2006)oraffected byisolation by distance (see software documentation in

pritch.bsd.uchicago.edu/software/structure22/readme. pdf). Thus, we employed the programme GENETIX

4.04to calculate FST values betweengroupsof indivi- duals sampled from different locations of the Lesser Kestrel breedingdistribution.Althoughthe distribution range oftheEurasianKestrelisrelativelycontinuous,we alsocalculatedFST values between distantsampled locations tocontrastFST pair-wisevalues with STRUC- TUREresults.ThesignificanceofFST pair-wisecompar- isons was given by a P-value calculated using 10000 randompermutation teststhatwasfurtheradjusted accordingtosequentialBonferronicorrectionsformulti- ple tests (Rice,1989).Isolation bydistancewas investi-

gated through Mantel tests based on the traditional

Location Code Populationsize(breedingpairs)

Spanishcorearea SWSand CWS 12000–19000

EbroValley NES 1000

France FRA o100

Italy ITA 3640–3840

Greece GRE 2000–3480

Israel ISR o1000

Data were taken from BirdLifeInternational(2007),Prugnolleetal. (2003)and Liven-Schulmanetal.(2004).SeeFigure 1forgeographic locations.

FST/1—FST approach.Weintroducedinthe programme GENETIXamatrix containing values ofgenetic differ- entiationbetweeneachpairofsampledpopulations(that is, FST/1—FST values representedin the yaxis) plus a matrix containing the geographical distance in kilo- metres betweeneach pair ofsampledlocations (repre- sentedinthexaxis).Geographical distanceswere calculatedaccordingto a straightline connectingthe geometricalcentre ofeach pair ofsampledpopulations. Calculationswere accomplishedbyusing ascaled map

and aruler. Thesignificanceofthe correlationbetween

genetic differentiation and geographicaldistancewas

tested in GENETIX 4.04througha P-valuecalculated using 10000permutations.

Allelic richness, average observed heterozygosities and the inbreeding coefficient FIS among groups of samplesencompassingindividuals from differentspe- ciesorsubspecieswere comparedusing thepermutation test (N¼10000)implementedinFSTAT(Goudet,2001). Theallelicrichnessestimate,which iscalculated from randompermutationsofaminimumsharednumberof individualsbetweengroups,isespeciallyuseful inthis

study ashighly polymorphiclocisuch asFp79-4may decisivelybias estimatesofgenetic diversityinrelation to samplesize. The non-parametricWilcoxon test was also employedtodetect significantdifferencesbetween sampledlocations inpolymorphismstatisticsobtainedat eachlocus(that is,allelicrichnessand averageobserved heterozygosities).Finally, microsatellitediversityateach pairoflocations,measuredasthemeannumberofalleles per individual,was comparedusing Student’st-tests.

Results

Hardy–Weinbergequilibriumandgeneticdiversity Overall,103alleleswere found in320LesserKestrels, 75 alleles in128mainlandEurasianKestrels and 46alleles in28island EurasianKestrels (seeTable2).LocusFp107 departed significantlyfrom Hardy–Weinberg expecta- tions, showingheterozygositydeficits in most popula- tionsthat areprobablyexplainedbythepresenceofnull alleles (seealsoAlcaide etal.,2005).Asnull alleles may violate several assumptionsofthe genetic methodswe intendedtoapply,locusFp107wasremovedfromfurther analysis.MainlandpopulationsfrombothKestrelspecies fitted to Hardy–Weinberg expectationsafter excluding this locus. Wefound,incontrast,statisticallysignificant heterozygositydeficits, evenafterBonferronicorrections

for multiple tests, in the smallestinsular population

correspondingtoF.t.dacotiae.

Populationdifferentiation

InLesserKestrels, theBayesian analysis ofpopulation structureexcludingany apriori informationabout the

origin ofindividualsindicatedpanmixia(that is,K¼1,

seeFigure2) asthemostlikelyscenario. Nevertheless, traditionalestimatesofpopulationdifferentiationrelying

ondifferencesinallele frequencies revealedweak (FSTo0.055) but significantpatternsofgenetic differen- tiation, even after Bonferroni corrections for multiple tests, when wecomparedgeographically distinct populations(Table 3).Infact,genetic divergenceacross thestudyareaconformedsignificantlytoanisolation-by- distancepattern(Figure 3).

Figure 2Bayesian clustering analysis of 320 Lesser Kestrels sampledin different regions of the WesternPaleartic. For each value of K (that is, the number of putatively different genetic clusterstested), 10simulationswere carried outtoobtain the probabilityofthedata (yaxis).

Table2NumberofallelesacrossninemicrosatellitemarkersintheLesserKestrel(Falconaumanni),theEuropeansubspecies oftheEurasian Kestrel (Falcotinnunculus tinnunculus) and the two subspeciesofthe EurasianKestrel inhabitingthe CanaryIslands(Falcotinnunculus canariensisand Falcotinnunculusdacotiae)

Locusrangesize(bp) / Falconaumanni
(n¼320) / Falcot.tinnunculus
(n¼128) / Falcot.canariensis
(n¼12) / Falcot.dacotiae
(n¼16)
Fp5 / 7 / 8 / 7 / 7
99–111 / 101–115 / 101–113 / 101–113
Fp13 / 5 / 4 / 2 / 4
86–106 / 92–98 / 92–94 / 92–98
Fp31 / 8 / 7 / 3 / 2
124–142 / 128–142 / 134–138 / 134–138
Fp46-1 / 10 / 6 / 4 / 6
115–139 / 117–127 / 119–125 / 115–125
Fp79-4 / 35 / 19 / 6 / 8
125–192 / 129–154 / 137–149 / 137–152
Fp89 / 4 / 5 / 2 / 4
116–122 / 116–124 / 118–120 / 116–122
Fp107 / 17 / 17 / 5 / 5
185–231 / 195–233 / 193–221 / 193–221
Cl347 / 11 / 9 / 5 / 5
96–116 / 100–116 / 100–112 / 100–112
Fp5 / 6 / NA / NA / NA
118–123 / NA / NA / NA

Abbreviation:NA,notapplicable.

Thenumberofindividualsanalysedforeachspecies orsubspeciesisshowninparentheses.

Onthe other hand,theclusteringanalysisimplemen-

ted in STRUCTURE detected only two genetically

distinctclusterswithinEurasianKestrels (that is,K¼2)

Table3Pair-wiseFST values (above diagonal)and corresponding

P-values(belowdiagonal)betweenLesserKestrelpopulationsfrom

t heWesternPaleartic

NESCWSSWSFRAITAGREISR

NES(68) / 0.008 / 0.008 / 0.014 / 0 / 0.009 0.035
CWS(76) o0.001 / 0.001 / 0.019 / 0.016 / 0.014 0.041
SWS(69)0.0012 / 0.19 / 0.023 / 0.013 / 0.013 0.038
FRA(26) / 0.0021 o0.001 / o0.001 / 0.009 / 0.041 0.034
ITA(26) / 0.56o0.001 / 0.0048 0.0664 / 0.017 0.021

GRE(21) 0.002 0.0026 0.002 0.001 0.005 0.054

ISR(34) o0.001 o0.001 o0.001 0.001 0.006 o0.001

Sample sizes at each location are indicatedin parentheses. Significant values after Bonferroni corrections for multiple tests are outlinedinbold. Non-Bonferroni-correctedP-valuesare given below thediagonal.SeeFigure 1forgeographicallocations.

that distinguishedthe mainlandsubspeciesagainstthe

two insular subspecies.This finding agrees with the

comparablyhigh and statisticallysignificantpair-wise FST values reported betweenEurasia and the Canary Islands (FST40.075,allBonferroni-correctedP-values o0.05;Table4). Conversely,therewasnoevidencefor genetic subdivision within Eurasia, as none of the pair-wise FST values were significantlydifferentfrom zero (FSTo0.015, all non-Bonferroni-corrected P-values

40.05), or within the Canarian Archipelago (FST¼

—0.018, P¼0.87) (see Table 5). Contrary to Lesser

Kestrels, our set of genetic markers did not reveal

significant evidenceofisolationbydistanceinthe mainlandsubspeciesofthe EurasianKestrel (Figure 3).

Tocompare data frombothspecies, weperformeda generalized linear model with FST as the response variableand species identityand Euclidean distance betweenthepopulationsasindependentvariables.After conservatively adjustingthedenominatordegreesof freedom to compensate for the non-independence

FST

1-FST

0.06

0.05

0.04

0.03

0.02

0.01

0.00

-0.01

5.65.86.06.26.46.66.87.07.27.47.67.88.08.28.48.6

In geographical distance (km)

Figure3Relationshipsbetweenthe extent ofgenetic differentiationand geographicaldistanceinthe Lesser Kestrel (open dots, r¼0.50,

P¼0.04)and EurasianKestrel (blackdots, r¼—0.44,P¼0.84)populationssampledacross theWesternPaleartic.

Table4Pair-wiseFST values (abovediagonal) and correspondingP-values(belowdiagonal) betweenEurasianKestrelpopulationsfromthe

WesternPalearticand theCanaryIslands

FV

0.083

0.121

0.107

0.099

0.105

0.105

—0.018

Sample sizesateach location areindicatedinparentheses.Significant values after Bonferronicorrectionsformultipletests areoutlinedin bold. Non-Bonferroni-correctedP-valuesaregiven below thediagonal.SeeFigure 1forgeographicallocations.

Table5Comparisonofaveragegenetic estimatesamonggroupsof Kestrel populationsthat wasperformedusing thepermutationtest (N¼10000)implementedintheprogrammeFSTAT

Discussion

We studied the genetic implications of habitat frag- mentation by comparing the generalist, continuously

Allelic

richness

Observed

heterozygosity

Inbreeding

coefficient(FIS)

distributedmainlandsubspeciesoftheEurasianKestrel

and the steppe-specialist, patchily distributed Lesser

Lesser Kestrel5.820.660.024

Kestrel. Our findingsindicatesimilar levels of genetic

EurasianKestrel

(Mainland) EurasianKestrel (CanaryIslands)

5.280.660.084

4.240.460.265

variationinboth thespecies, but lower levelsofgenetic diversityintwo island subspeciesofEurasianKestrels.

With respect topopulationdifferentiation,the Bayesian

clusteringmethodseparatedthe mainlandpopulation

Allelic richness was calculatedover a minimumnumber of 12 individuals.

Table6Genetic diversityacross eight microsatellitemarkersinsix geographicallydistinctpopulationsofLesser Kestrels

ofEurasianKestrels from their island counterparts. Likewise, FST analysesshowedsignificant genetic differentiationbetween,but not within,these clusters. InLesserKestrels, STRUCTUREassignedallindividuals to a single putative population.Nonetheless,the esti- mates ofpopulationdifferentiationthat made useofthe additionalinformationonthegeographicdistributionof allele frequenciesrevealedlow but significantlevels of

Allelic

richness

Averageobserved

heterozygosity

Inbreedingcoefficient

(FIS)

genetic differentiationfollowinganisolation-by-distance model.

NES6.60.630.07

CWS+SWS7.060.650.05

FRA6.020.600.04

ITA6.890.67—0.06

GRE6.880.640.01

ISR7.420.660.03

Itiscurrentlyassumedthat species thrivingwithina range of environmental conditionsare more sensitive

to habitat transformations, their distributional ranges

becomingpatchierand the risk forgenetic drift within fragments increasing(for example,Ferrer and Negro,

2004).Our empiricalapproach exemplifiesa situation

Allelicrichnessestimateswere adjustedtoaminimumsamplesize

of21individuals.SeeFigure 1forgeographicallocations.

between sampling locations (see Bailey et al., 2007), the interaction term remained significant (F1,9¼9.11, P¼0.015).

Geneticdiversity

ThepermutationtestperformedinFSTATdid notreveal statistically significant differences in genetic diversity

(allelic richnessand averageobservedheterozygosity) or inbreeding (FIS) between the Lesser Kestrel and the mainland subspecies of the Eurasian Kestrel (all two-sidedP-values40.05, Table 5).Incontrast,average observed heterozygosity was significantly lower in island than inthecontinentalsubspeciesoftheEurasian Kestrel (0.46vs0.66,two-sidedP-value¼0.009;Table5), and thedifferenceinallelicrichness wasmarginally significant (4.24 vs 5.28, two-sided P-value¼0.08;

Table 5).Furthermore,we foundstatisticallysignificant evidencesofincreased inbreeding(FIS)intheKestrel genotypes from the Canary Islands (0.265 vs 0.084, two-sidedP-value¼0.02;Table5).

Finally,pair-wiseanalyses comparinglocusbylocus failed todetect statistically significantdifferencesin genetic diversity between any of the geographically distinctpopulationsofLesserKestrels investigated(non- parametricWilcoxontest,allP-values40.05;seeTable6). Averagemicrosatellitediversityper individualwas not statisticallydifferentamongpopulations either (t-tests, allP-values40.05), exceptforacouple ofcomparisons involvingthesmallestand themost geographically isolatedpopulationfromSouthernFrance.Suchcompar- isons involvedthe less geneticallydiversepopulation (France) and two ofthe most geneticallydiverse(Italy and Israel, seeTable6)populations.

whereby genetic differentiation reflects the spatial

distributionofpopulations,which, inturn, isdelimited by habitat requirements. Thus, genetic differentiation

between Lesser Kestrel populations increases with geographicaldistance(see also Alcaide etal.,2008for

data on MHC genes). Even thoughLesser Kestrel isa long-distancemigratoryspecies, gene flow isrestricted

over short distances due to high natal and breeding philopatry(Negro etal.,1997;Serrano etal.,2001;Serrano

and Tella,2003).Elsewhere,wefound,however, alackof fine-scale patterns ofgenetic differentiationinaspatially

structured population of Lesser Kestrels located in northeasternSpain (Alcaide etal.,inpress). Thisfinding

was attributedtothefactthat populationsubdivisionat the geographicalscale studied(about 10000km2) may

not have been sufficient, given the long-distancedis- persal capabilities displayed by the species; conse-

quently,gene flow had homogenizedallele frequencies. Nonetheless,effective long-distancedispersalbyLesser

Kestrels (4100km) has rarely been documented by direct observations(Prugnolleetal.,2003;Serrano etal.,

2003;PPilard and FMart´ın, personalcommunication;D Serrano et al., unpublisheddata; M Alberdi,personal

communication),afactthat wouldbeinagreementwith theemergenceofgenetic structuringatlarge geographi-

cal scales. In contrast, it has been shown in several

Europeanpopulations of Eurasian Kestrels that natal

dispersal regularly occurs over large distances (for

example,Snow, 1968;seealsoKorpima¨ki,1988;Village,

1990; Korpima¨ki et al., 2006; Vasko, 2007). This high amplitudeofdispersal,combinedwiththelowincidence

ofhabitatfragmentationinthe EurasianKestrel, would

thereforeexplain itsgenetic uniformity.

Populationgenetics theory predictsthat reductionsin

population size and limited migration decrease local genetic variation,triggeringnegativegenetic processes

such as inbreeding depression and loss of adaptive potential (Frankham et al., 2002). Following these

predictions, recent studiesin the Lesser Kestrel have

repeatedly found weak positive correlationsbetween

fitness componenttraits and individualgenetic diversity at11polymorphicmicrosatellitemarkers(Ortego etal.,

2007b,c).However,our genetic analyses,relying on at least sixmicrosatellitespreviouslyamplifiedbyOrtego

and co-workers(Fp5, Fp13, Fp31, Fp46-1, Fp79-4 and

Fp89), have not revealed comparably low levels of

microsatellitediversityorincreasedinbreedinginLesser

Kestrels inrelationtotheputativelyoutbredsubspecies

oftheEurasianKestrel. Genetic variationatfunctionally and evolutionaryrelevantMHC loci have also shown

extraordinarylevelsofpolymorphism(4100alleles ata single locus) and heterozygositiesabove 95%in Lesser

Kestrels (Alcaide etal.,2008).

We believe that additional analyses of the pre-

bottlenecked population are needed to evaluate the degreeof genetic depauperationin the Lesser Kestrel.

Inanycase,itappearsincautious toassumethat the populationdeclineexperiencedbythisspeciesislikelyto

have translated into reduced levels of contemporary genetic variationand increasedinbreeding.Forinstance,

Brown et al. (2007) have recently failed to detect signaturesof a genetic bottleneckin peregrine falcons

after adevastatingdecline inthemid-twentiethcentury due to organochlorine contaminants. Similarly, some

Lesser Kestrel populationshave been knowntoexperi- encedemographicgrowth,either naturally(forexample,

Tellaetal.,1998; Ortego etal.,2007a)orbymeansof reintroduction or supplementation programmes (for

example,Pomarol,1993).Yeteven in the bottlenecked and geographicallyisolatedpopulation from Southern

France, from where wereportthelowest levelsof microsatellite polymorphism (Table 6), there is no