Landscape changes decrease genetic diversity in the Pallas’ long-tongued bat

(1)

SupportedbyBotic´arioGroupFoundationforNatureProtection

www.perspectecolconserv.com

Research Letters

Landscape changes decrease genetic diversity in the Pallas’

long-tongued bat

Rosane G. Collevatti

^a,∗

, Luciana C. Vitorino

^b

, Thiago B. Vieira

^c

, Monik Oprea

^a

, Mariana P.C. Telles

^a,d

aLaboratóriodeGenética&Biodiversidade,InstitutodeCiênciasBiológicas,UniversidadeFederaldeGoiás,CEP74001-970,Goiânia,Goiás,Brazil

bDepartamentodeBiodiversidadeeConservac¸ão,InstitutoFederalGoianocampusRioVerde,Cx.P.66,75901-970,RioVerde,Goiás,Brazil

cUniversidadeFederaldoPará,CampusAltamira,Altamira,Pará,Brazil

dEscoladeCiênciasAgráriaseBiológicas,PontifíciaUniversidadeCatólicadeGoiás,CEP74605-010,Goiânia,Goiás,Brazil

h i g h l i g h t s

•Thereplacementofnaturalvegeta- tiontoagriculture,pasturesorurban matricesdecreasedgeneticdiversity inthePalla’slong-tonguedbat.

•Pallas’slong-tonguedbathadaquasi- stable range distribution since the lastGlacialMaximum.

•TheQuaternaryclimatechangeshad nosigniﬁcanteffectsonthegenetic diversityofthePalla’slong-tongued

•bat.Populations of the Palla’s long-

tongued bat was differentiated by isolation-by-distance, but not by environment.

•Our ﬁndings point to the main- tenance of large areas of natural vegetationtoconservegeneticdiver- sityinthePalla’slong-tonguedbat.

g r a p h i c a l a b s t r a c t

a r t i c l e i n f o

Articlehistory:

Received5March2020 Accepted26June2020 Availableonline7August2020

Keywords:

Cerrado Connectivity

Ecologicalnichemodeling Habitatfragmentation Quaternaryclimatechanges Resistancesurface

a b s t r a c t

Specieswithhighmobilitymayhavelowgeneticdifferentiationamongpopulationsduetohistorical long-distancedispersal,butrecentstudiesshowthatbatdispersalmaybeaffectedbyhabitatlossand fragmentation.Here,weanalyzetheeffectsoflandscapecontemporarychangesanddynamicsinpale- odistributionduringtheQuaternaryongeneticdiversityanddifferentiationinthePallas’long-tongued bat,Glossophagasoricina.Wesampled18populationsinlandscapeswithdifferentlandcover,andused ninemicrosatellitelociandthesequenceofafragmentofthemitochondrialgenecytochromeb(CYB)to obtaingeneticdata.Weperformedecologicalnichemodellingandusedgenerallinearmixedmodelsand optimizationofmultipleresistancesurfacestoanalyzehowlandscapestructureandclimaticsuitability affectgeneticdiversityanddifferentiation.Ourresultsshowthattheconversionofnaturalvegetation, suchasforestsandsavannas,toagriculture,pasturesorurbanmatrices(unsuitablehabitats)decreases geneticdiversityandincreasesinbreeding,buthasnoeffectongeneticdifferentiationamongpopula- tions,thatwaslikelyaffectedbyspatialdistance.Ourﬁndingspointtotheimportanceofmaintenanceof largeareasofnaturalvegetationremnantstoconservegeneticdiversityofG.soricina,animportantbat pollinatorintheNeotropics.

).

∗ Correspondingauthor.

E-mailaddress:[email protected](R.G.Collevatti).

https://doi.org/10.1016/j.pecon.2020.06.006

creativecommons.org/licenses/by-nc-nd/4.0/).

(2)

170 R.G.Collevattietal./PerspectivesinEcologyandConservation18(2020)169–177 Introduction

Animalmobilityanddispersalmaybeaffectedbyhabitatloss andfragmentation(e.g.Umetsuetal.,2008;Puttkeretal.,2011), andthere isevidence ofdeleterious effectsof habitatfragmen- tationongeneticvariabilityanddifferentiationoncommonand widespreadvertebratespecieswithhighmobility(Johanssonetal., 2007;Delaneyetal.,2010;Levyetal.,2019).Landscapestructure, suchasmatrixqualityanddistanceamonghabitats,mayisolate populations,limitingtheirdispersal(Umetsuetal.,2008;Boscolo andMetzger,2011)andthus,populationconnectivityandgenetic diversity(e.g.Dixoetal.,2009;Balkenholetal.,2013;Jacksonand Fahrig,2016).

Moreover,climatechangesduringtheQuaternarymightalso haveaffectedspeciesgeneticdiversityduetothecyclesofpopu- lationexpansionandcontraction(Hewitt,2000).Populationrange contractionmaydecreaseeffectivepopulationsize,leadingtoloss ofgeneticdiversityduetogeneticdrift.Shiftsinspeciesgeographi- caldistributionsduetotheQuaternaryclimatechangeshaveplayed animportantroleinshapingthegeneticdiversityofanimalspecies intheCerradobiome(e.g.Mirandaetal.,2019).Thus,disentan- glingcontemporaryandhistoricalprocessesleadingtopopulation effectivesizereductionandlossofgeneticdiversityisimportant topredictspeciesresponsetochangesinlandscapeintheAnthro- pocene.

Batspeciesaredifferentiallyaffectedbyhabitatfragmentation (e.g. Gorresen and Willig, 2004; Meyer et al., 2009; Centeno- Cuadros et al., 2017). Some bats avoid ﬂying over open areas (Henryetal.,2007),watercourses(Albrechtetal.,2007),androads (Zurcheretal.,2010).Thus,fragmentationhasapotentialtoiso- latebatpopulationsandreducegeneticconnectivity(Rossiteretal., 2000)andcolonizationofnewareas(KerthandPetit,2005;Russell etal.,2005).Inadditiontothebarriersimposedbythelandscape, reproductivebehaviorandhistoricalpatternsofcolonizationmay alsoberesponsibleforrestrictedgeneﬂow(Wilmeretal.,1999).

Althoughhighlymobile,batsmayshowdifferentdispersalpatterns andbreedingsystems,includingvariationinroostingecology,and socialstructures(McCrackenandWilkinson,2000),thatmayaffect geneticstructure(e.g.Flandersetal.,2016).Indeed,severalstud- ieshaveshownsigniﬁcantpopulationsubstructureinbatspecies atdifferentspatialscales(e.g.Biollazetal.,2010;Rippergeretal., 2013;Flandersetal.,2016).

Glossophagasoricina(Pallas1766)(Glossophaginae),thePallas’

long-tonguedbat,isoneofthemostcommonandwidespreadnec- tarivorousbatsintheNeotropics(Reisetal.,2005),andalsofeeds oninsectsandfruits(Willigetal.,1993),andusesalargevariety ofroosts(caves,trees,human-madestructures).Adultmalesmay defendharemsandjuvenilesandbothmalesandfemalesdisperse fromtheirnatalcolony(Pink,1996).Becauseofitsgeographicaldis- tributionandnaturalhistory,aloweffectofhabitatfragmentation onG.soricinageneticdiversitymaybeexpectedorevennoeffectat all.However,morethanhalfoftheCerradobiomehasbeenrapidly transformedintopasturesandcrops(Silvaetal.,2006),affecting speciesgeneticdiversityandinbreeding(e.g.Arrudaetal.,2017) jeopardizingspecieslong-termconservation.

Here,weaddresstheeffectsofcontemporarychangesinland- scapestructureand paleodistributionduringtheQuaternaryon geneticdiversityandpopulationdifferentiationinG.soricinainthe Cerradobiome.Weusedamultiscalespatialapproachtospecif- icallytestwhetherhabitatamount(forestsandsavanna),matrix (unsuitableland covers)typeand past climatechanges explain thecurrentgeneticdiversityand differentiationamongpopula- tions.Thegeneticdatawereobtainedfromnuclearbiparentally inheritedmicrosatelliteloci and thematernally inheritedmito- chondrialcytochromeBgene(CYB).Thus,usingdatafrommarkers withdifferentinheritance and evolutionaryrates ourdatamay

trackdifferencesindispersalbetweenmaleandfemalebatsand alsodifferencesintimeofresponsetoQuaternaryclimatechanges andcurrentlandscapechanges.Weexpecttoﬁndsigniﬁcantrela- tionshipsbetweenlandscapestructureandgeneticdiversityand differentiationformicrosatelliteloci,butnotforCYB.Conversely, weexpectthatCYBgeneticparameterswillbebetterexplained byclimaticchangesandhistoricaleffectivepopulationsizedueto lowermutationrate.

Materialsandmethods

Populationsandtissuessampling

We sampled 18 populations (229 individuals) of G. soricina throughout theBrazilianCerrado biome(Fig.1, TableS1). Pop- ulations were selected in landscapes with different land use composition, suchas habitat (savanna and forest)amount and matrix(unsuitablelandcover)type.Batswerecapturedwithmist netsinstalledatgroundlevelandwesampledapproximately3mm ofthewingmembrane(seedetailsinSupplementaryFileS1).

Geneticdata

Forgeneticdata,DNAwasextractedandindividualsweregeno- typedusingnine microsatelliteloci previouslydeveloped forG.

soricina(Opreaetal.,2012).Forsequencing,weusedtheprimers describedbyIrwinetal.(1991)toamplifya1140bpfragmentof thecytochrome-bmitochondrialgene(CYB).Detailsoftheproto- colsforgenotypingandsequencingaredescribedinSupplementary FileS1.

To obtain genetic parameters for microsatellite loci we estimated genetic diversity (heterozygosity expected under Hardy–Weinbergequilibrium, He),inbreedingcoefﬁcient(f)and allelicrichnessforeachpopulation.Forpopulationgeneticdiffer- entiationweestimatedWright’sF_ST(Wright,1951),G’_ST (Hedrick, 2005)andJost’sD(Jost,2008)usingPopGenReportpackage(Gruber andAdamack,2019)implementedinR3.6.1.(RCoreteam,2019).

WeusedaBayesianclusteringmethodimplementedinthesoft- wareStructure2.3.4(Pritchardetal.,2000),toidentifygenetically homogeneousgroupswithinoursamples.We estimateda pos- terioriprobabilityofK(numberofgroups)rangingfrom1to18 (correspondingtoeachsampledlocality).

ForCYB,weestimatednucleotide(␲)andhaplotype(h)diversity foreachpopulation,andpopulationdifferentiation(F_ST)usingthe softwareArlequin3.11(Excofﬁeretal.,2005).Populationstructure wasalsoassessedusingBayesianclusteringimplementedinthe softwareBAPSv6.0(Coranderetal.,2008)withanupperlimitofK

=18.

Tounderstandthegeographicaldistributionandphylogenetic relationshipsamongmtDNAhaplotypes,weinferredahaplotype networkusingthesoftwareNETWORK4.6.1.6(Forsteretal.,2004).

DetailsofgeneticanalysesareprovidedinSupplementaryFileS1.

Effectivepopulationsizeandconnectivity

Becausegeneticdiversitymaybeaneffectofchangesineffective populationsizeweestimatedhistoricalandcontemporaryeffec- tivepopulationsizes(Ne).Contemporary Newascalculated for microsatellitelociforeachpopulationwithNeEstimatorv.2.01 (Doetal.,2014)usingthemolecularcoancestrymethod(Nomura, 2008)toverifytheeffectofpopulationsizeongeneticparameters.

WeestimatedhistoricalNeusingcoalescentanalysisforboth microsatellitelociandCYB.Weestimatedthemutationparame- tertheta()andthenumberofimmigrantspergenerationamong

(3)

Fig. 1.Geographical distribution of the 18 samples of Glossophagasoricina in the Brazilian Cerrado and sampling design for landscapeanalysis. Two circular buffersdelimited eachpopulation.Theinner bufferhas2kmradius(red line)andtheouter(blackline)has5kmradius.Landusecategoriesare inlegends.

Other refers to mixing categories not detectableat the map scale. Landuse map was obtained from the MMA (Brazilian Ministry ofEnvironment) database http://mapas.mma.gov.br/mapas/aplic/probio/datadownload.htm.

allpopulationpairsusingLamarc2.1.10software(Kuhner,2006).

DetailsaredescribedinSupplementaryFileS1.

Ecologicalnichemodeling

Weobtained13,069occurrencerecordsofG.soricinaacrossthe Neotropics(Fig.S1).Afterdataﬁltering,theremaining2353occur- rencerecords(TableS2)weremappedinagridofcellsof2.5×2.5 (longitude×latitude),encompassingtheNeotropics,togenerate thematrixofpresenceusedtocalibratetheENMs.

The climatic layers were represented by ﬁve bioclimatic variables:annualmeantemperature,meandiurnalrange,isother- mality,precipitationofwettestquarter,andprecipitationofdriest quarter,selectedbyfactorialanalysiswithVarimaxrotationfrom 19 bioclimatic variables obtained in the ecoClimate database (Lima-Ribeiro et al., 2015).Bioclimatic variables were obtained forpre-industrial(representingcurrentclimateconditions),mid- Holocene(6ka)andLastGlacialMaximum(LGM;21ka)fromﬁve Atmosphere-OceanGeneralCirculation Models(AOGCM) (Table S3).

Estimatesofthecurrentand pastpotentialdistributionofG.

soricinawereobtainedbymodelingitsecologicalnicheusingﬁve algorithms(TableS4).TheENMswerethenbuiltonthecurrentcli- maticscenarioandprojectedontotheclimaticconditionsduring boththemid-Holocene(6ka)andLGM(21ka).AllENMprocedures wererun in theintegrated computationalplatformBIOENSEM- BLES,whichprovidespredictionsbasedontheensembleapproach (Diniz-Filhoetal.,2009).

ForeachalgorithmandAOGCM,modelswerebuiltusing75%of occurrencesandtestedagainsttheremainingrecords(25%).Models withpoorperformancewereeliminated(TSS<0.7).Thefrequency ofpredictedpresencewasusedasameasureofconsensussuitabil- ityfrom50initialmodels.

Finally,thefullensemblewasobtainedforeach timeperiod usingthepredictiveperformances(TSS)tocomputea weighted meanofsuitabilities,fromwhichhistoricalrefugiaweremapped (i.e.allgridcellswithsuitabilityvalues≥0.5duringthethreetime periods).DetailsareprovidedinSupplementaryFileS1.

Landscapeandclimaticeffectsongeneticdiversity

Toanalyzetheeffectoflandscapestructureongeneticdiver- sitywequantiﬁedlandscapevariablesatnodelevel(Wagnerand Fortin,2013).Wedrewlandscape-buffersof2.0and5.0kmradius aroundeachsampledlocation(Fig.1)usinghigh-resolutionimages availableatthemapdatabaseoftheGeographicInformationSys- teminArcGisv.9.3(ESRI^®a).Wechosetheselandscape-buffersizes becauseofthehomerangeofG.soricina(e.g.Aguiaretal.,2014).

Ineachlandscape-bufferweclassiﬁedthelandcoverandcalcu- latedtheareaandthepercentageofeachlandcovertypeinside each buffer (Fig. 1, Table S5) using ArcGis v.9.3.We examined thecorrelation among land cover amounts and excludedthose withPearson’scorrelation≥

0.5

,toavoidcollinearity.Wekept percentageofnaturalvegetationremnants(forestsandsavanna), representinghabitatamount,andformatrixweusedmatrixtype

(4)

172 R.G.Collevattietal./PerspectivesinEcologyandConservation18(2020)169–177

Fig.2. BayesianclusteringformicrosatellitelociandCYBof18GlossophagasoricinapopulationsinBrazilianCerrado.Theblackfulllinesindicatepossiblegeneticdiscontinuity betweenlocalitiesbasedonDelaunay’striangulationformicrosatellitelociandgreenforCYB.Differentcolorswereassignedforeachclusteraccordingtotheﬁgurelegend.

Thecirclesectionsrepresenttheclusterfrequencyineachsampledpopulation.SeeTableS1forpopulationcodeandlocality.LandusemapwasobtainedfromtheMMA (BrazilianMinistryofEnvironment)databasehttp://mapas.mma.gov.br/mapas/aplic/probio/datadownload.htm.

(pasture,agriculture,urbanarea,waterbody,miningandoutcrops andmixeduse).

WeusedGeneralizedLinearMixedModel(GLMM)toanalyze theeffectsoflandscapestructureandpaleodistributionongenetic parameters.Weusedasexplanatoryvariablesmatrixtypeandper- centageofnaturalvegetationremnants,suitabilityatpresentday derivedfromENMandclimaticstability.Wealsousedhistorical (basedoncoalescence)andcontemporary(basedoncoancestry) Neineachlocalityasexplanatoryvariables.

We built several models with combinations of explanatory variablesandanull model.Analyseswerecarriedoutusingthe MCMCglmmpackage(Haldﬁeld,2019)implementedinR3.6.1.(R Coreteam,2019).Toselectwhichmodelbestexplainstheobserved variationingeneticparametersamonglandscapeswecalculated AICcorrectedfor small samples(AICc)and AICci, i.e.the dif- ferencebetweeneach model(i)andthebestmodels(Burnham andAnderson,2002).DetailsareprovidedinSupplementaryFile S1.

Landscapeandclimaticeffectsongeneticdifferentiation

Toperformtheanalysisatthelinklevel(WagnerandFortin, 2013), weused ResistanceGApackage(Peterman,2018), imple- mented in R 3.6.1. to optimize multiple resistance surfaces simultaneously.Forgeneticdistancematrixweusedthegenetic differentiation among all pairs of populations (Tables S6–S9):

F_ST (forboth microsatellites and CYB), and G_ST’ and Jost’ D for microsatellites.Forresistancematrixweextractedthelandcover between all pairs of populations and converted land use in resistancesurface.We alsousedtheconsensusmap of ENMat present-dayfor resistancedue toclimate, and thegeographical distancematrixtoaccountforisolation-by-distance.Optimization wasperformedwiththefunctiongdistance/commuteDistance.To selectthemostlikelymodel,weusedAICci.

We also analyzed genetic discontinuity among populations usingtheDelaunaynetwork(LegendreandLegendre,1998)imple-

mented inthe softwareSAM v.4(Rangelet al.,2010), for both microsatellitesandCYB.DetailsareprovidedinSupplementaryFile S1.

Results

Geneticdiversityandpopulationdifferentiation

Populationshad similarlevelsof geneticdiversityand poly- morphism(TableS10)despitethedifferencesinsamplesize.The inbreedingcoefficientsweresignificantlydifferentfromzerofor mostpopulations(TableS10).Mostpairsofpopulationshadlow valuesofpairwiseF_ST(TableS6),pairwiseJost’sDandpairwiseG’_ST (TablesS7andS9);theonlyexceptionwaspopulation18,which showedhighandsignificantvaluesforallcomparisons.Detailsof resultsareprovidedinSupplementaryFileS1.

Bayesianclusteringbasedonmicrosatelliteloci showedthat threeclusters(K=3,Fig.S2,TableS11)weremostlikely.Clus- ter1(green,Fig.2)comprisedmostpopulations.Cluster2(yellow, Fig.2)comprisedmainlypopulation18.PopulationsfromSouth- eastBrazil(14and15,Fig.2)wereequallyassignedtoclusters1 and3.

Wefound171differentCYBhaplotypesforthe229individuals ofG.soricina(TableS12).Allpopulationsshowedhighhaplotype andnucleotidediversitiesandmosthaplotypeswereexclusiveto onepopulation(TablesS101,S12,Fig.S3).Haplotypesfrompop- ulation18(H158,Fig.S3),1(H01)and14(H117)werethemost differentiated and wereexclusive tothesepopulations (Fig. S3, TableS12).

PopulationsshowedhighpairwiseF_ST(TableS9).Bayesianclus- teringshowedanoptimalpartitionofK=3clusters(Fig.2,Table S13),butwithnocleargeographicalpattern.

Contemporaryand historical effective populationsizeswere lowforallpopulations(TableS10).Overall,thenumberofmigrants pergenerationwashigh(Nem>1.0)betweenallpairsofpopula- tions(TableS14)forbothmicrosatellitesandCYB.

(5)

Fig.3. Mapsofconsensusofthe75modelsofthepotentialdistributionofGlossophagasoricina,basedonecologicalnichemodellingduringtheLGM(21ka),mid-Holocene (6ka),present-dayandthehistoricalclimaticrefugiumthroughtime(fromtheLGMtopresent-day).Reddotsarethe18populationssampled.

Ecologicalnichemodeling

Glossophagasoricinashowedaclearpreferenceforhotclimate acrosstheNeotropics(Fig.3)withhighfrequencyofoccurrencein savannas,seasonallydryandrainforestsoftheNeotropics.Such climatespacewasequallyavailableattheLGM,mid-Holoceneand presentday(Fig.3,Figs.S4,S5,TableS15),andthusENMspre- dictedonlyaslightdifferenceinG.soricinapotentialdistribution throughtime,withaslightlylargerrangesizeattheLGM(Fig.3, Fig.S5)comparedtoboththemid-Holoceneandpresent-day.In addition,awideregionacrosseastern,northeasternBrazil,towards central-westAmazoniaandnorthernSouthAmericaandwestern MesoAmericaprobablyactedashistoricalrefugiamaintainingpop- ulationsofG.soricinaduringtheclimatechangesthroughoutthe lastglaciationcycle (Fig.3).ENMsalsoshoweda rangeshiftin geographicaldistributiontowardsthesoutheastBrazil,especially fromtheLGMtothemid-Holocene(Fig.3,Fig.S5).Fromthemid-

Holocenetopresent-day,ENMshowedquasi-stabilityinG.soricina geographicalrange(Fig.3,Fig.S5).SeedetailsinSupplementary FileS1.

Landscapeandclimaticeffectsongeneticdiversity

Overall,modelsincludingmatrixtypeexplainedbetterthevari- ationingeneticparametersformicrosatelliteloci(Figs.4A,S6A, TableS16)andCYB(Fig.4B,Fig.S6B,TableS16)atbothspatialscales.

Althoughmodelswitheffectivepopulationsize(Ne)andmatrix typehadsigniﬁcantcoefﬁcientsforkandh(TableS16),theywere lesslikelythanmodelscomprisingonlymatrixtype(Figs.4and S6).Modelsincludingclimaticsuitabilityorpercentageofnatural vegetationremnantswerelesslikelythanmodelswithmatrixtype (TableS16).

(6)

174 R.G.Collevattietal./PerspectivesinEcologyandConservation18(2020)169–177

Fig.4.ModelaveragingresultsforeachgeneticparameterbasedonthegenotypingofmicrosatellitelociandsequencingofCYBfor18GlossophagasoricinainBrazilian Cerradofor2kmspatialscale.PanelsshowtherelativeimportanceofeachvariablebasedonwAICc(Akaike’sweightofevidence)andthemodelcoefﬁcients(␤)for(A) microsatelliteloci,(B)CYB.Errorbarrepresentsstandarderror(seeTableS16forpvalues).SeeFig.S6for5kmspatialscale.He,geneticdiversity;f,inbreedingcoefﬁcient;

AR,allelicrichness;k,numberofhaplotypes;h,haplotypediversity;␲,nucleotidediversity.

Landscapeandclimaticeffectsongeneticdifferentiation

Geneticdifferentiationamongpopulationswasnotexplainedby landscapeorclimaticresistancetoG.soricinadispersal(TableS17, Figs.S7–S10).ThenullmodelwasthebestmodelforbothCYBand microsatellites,howevergeographicdistancecouldnotberejected asafactorexplainingthevariationingeneticdifferentiationamong populationsformicrosatellites(TableS17).

We found discontinuity between pairs of populations, i.e.

higherF_STthanexpectedbythespatialdistancebetweenpairsof populations.Formicrosatellitelocipopulation18wasthemostdif- ferentiated.Wefoundevidencefordiscontinuitybetweenﬁvepairs ofpopulations:4–7,2–18,3–18,7–18,and9–18(Fig.2,TableS17).

ForCYB,population17wasthemostdivergent.Wefounddiscon- tinuitybetween4–6,5–17,6–17,7–17,12–17and16–17(Fig.2, TableS18).

Discussion

Contemporarychangesinlandscapeandgeneticdiversity

Our ﬁndings show that contemporary changes in landscape affectsgeneticdiversityinpopulationsofGlossophagasoricinafrom theCerrradobiome.Thereplacementofnaturalvegetation,suchas forestsandsavannas,toagriculture,pasturesorurbanareasmay decreasegeneticdiversityandincreaseinbreeding.Theinﬂuence occursatbothspatialscalesanalyzedhere(2and5kmbuffers).

The effects of scale on therelationships betweenlandscape structureandbatoccurrence,abundanceandpopulationgenetic structurehavebeenreportedpreviously(e.g.Gorresenetal.,2005;

Razgouretal.,2014).Thelackofscaleeffectinourstudymaybedue toG.soricina’sdispersalabilityanditsforagingandmatingbehav- ior.Thehomerangeofthespeciesislargerthan2km,andthe maximumﬂightdistancetrackedsofarwas3.8km(Aguiaretal., 2014).However,G.soricinamakescoloniesandharemsdefended byadultmales,whichmayconstraintdispersalandcausethehigh inbreedingobserved.In addition,thedispersalconstraindueto matingbehaviormayalsoisolatecoloniesofG.soricina,explain- ingtheeffectofmatrixtypeoninbreeding.Smallerbats,suchas G.soricina,maybemoreaffectedbylandscapechangesindepen- dentofthegeographicaldistribution.Forinstance,geneticdiversity andgeneﬂowinCarolliacastanea,asmallwidelydistributedbat species,arenegativelyaffectedbylossofhabitat(forests)whereas thelargerArtibeusjamaicencisisnotaffectedbypercentageofforest andmatrix(Clearyetal.,2017).Likewise,Carolliaperspicillatashow lossofgeneticdiversityduetohabitatfragmentation,contrasting toUrodermabilobatum,alargermorevagilespecies(Meyeretal., 2009).

Matingbehaviorandphilopatrymayaffectpopulationstructure inbatsandincreaserelatednesswithinpopulationsandinbreeding (Biollazetal.,2010;Rippergeretal.,2013;Flandersetal.,2016).

Our ﬁndings show higher geneticdifferentiation for mitochon- drialCYBthanfornuclearmicrosatellites,whichmaybeevidence of female philopatry, i.e., malesdisperse more frequently than

(7)

females.However,itisimportanttonotethatthemitochondrial genome(mtDNA)ismaternallyinheritedwithoutrecombination andtheeffectivepopulationsizeofmtDNAissmallerthanthatof thenucleargenome,therebyleadingtoahigherrateofalleleﬁxa- tionbygeneticdriftformtDNAthanfornuclearDNA(Birkyetal., 1983).Effectivepopulationsizemaybeoneofthemostimpor- tantfactorsinﬂuencingratesofmolecularevolution(Lanfearetal., 2014).Thus,differencesbetweenmicrosatellitelociandCYBmay beduetodifferencesindispersalbehaviorbutalsoinevolution- aryrates.Moreover,geneticdifferentiationatmicrosatelliteloci andCYBshoweddifferentresponsestogeographicaldistance.CYB showednoeffectofspatialdistancelikelybecausemalesdisperse morefrequentlythanfemales.

Herewefoundnoeffectoflandcoveringeneticdifferentiation atboth CYBand microsatelliteloci. However,wefounddiscon- tinuityin severalpairsofpopulations,meaningthatsomepairs ofpopulationshavehigherF_ST thanexpectedbytheirgeograph- icaldistance.Thegeneticdiscontinuitymaybeduetotherecent fragmentationandanthropogenicdisturbancebecausepopulations showinggeneticdiscontinuityformicrosatellitelociareimbedded inurbanareas,agricultureorpasturethatmayaffectthedispersal ofG.soricinaamonglocalities,restrictinggeneﬂowandincreasing inbreeding.Forinstance,population18iswithinaconservation unit,whichhaspristinevegetationandawidevariation infood resources,surroundedbyextensiveagricultureareas(mainlysoy- beanandcorn).Theregionpresentshighnumberofwell-preserved caves,whichareusedasroostsbythespecies.Allthesecharacteris- ticsmayfavorbatsstayingwithintheconservationunit,decreasing theconnectivityofthispopulation(18)evenwithpopulationsat closedistance,suchaspopulations2,3and9(seeFig.2),leadingto ageneticboundary.Infact,largerandmorestablecavesandsur- roundedbynaturalvegetationcanhavemorebatspeciesandlarger populations(e.g.Barrosetal.,2020).Population9isalsowithina conservationunit,however,withextensiveareasofsavannaand nocaves,whichmayfavorG.soricinadispersal,decreasingdiscon- tinuity.Population17issurroundedbycropsandminingindustry, whichmayisolatetheG.soricinapopulation.

Inaddition,CYBshoweddifferentpatternsofdiscontinuityand Bayesianclusteringcomparedtomicrosatelliteloci.Forinstance, population18showedhighconnectivitywithotherpopulations andnogeneticdiscontinuityforCYB.Ontheotherhand,popula- tion17hadhighdifferentiationanddiscontinuityforCYB,butnot formicrosatellites.Thisresultmaybeduetodifferentdispersal patternbetweenmalesandfemalesorduetodifferencesinevolu- tionaryrates.Microsatellitesmayshowsignalsofcurrentlandscape changeswhileCYBmayshowtheeffectsofhistoricalconnectivity andpaleodistribution(seebelow).Indeed,batsmayshowmale- biaseddispersalthatcanbehighlyaffectedbylandscapechanges andhabitatloss,leadingtolossofgeneticdiversity(Halczoketal., 2018).

Despite the high level of fragmentation and anthropogenic disturbances in the Brazilian Cerrado, the studied populations ofG.soricinapresented relativelyhighgeneticdiversity atboth microsatellitelociandCYB,similartothelevelsofotherPhyllosto- midaebatssuchasLophostomasilviculum(He=0.490;Dechmann etal.,2002),Carolliabrevicauda(Herangingfrom0.430to0.940;

Bardelebenetal.,2007)andDermanurawatsoni(hrangingfrom 0.902to0.991,andnumberofhaplotypesrangingfrom11to21;

Rippergeretal.,2013).Ourresultsalsoshowlowbutsigniﬁcantdif- ferentiationformicrosatelliteloci(F_ST =0.056,P<0.0001)among populationsofG.soricinaintheBrazilianCerrado,similartoother bat species,such asPlecotus auritus(F_ST ranging from0.011to 0.026;BurlandandWilmer,2001)andNyctalusazoreum(F_STrang- ingfrom0.009to0.036;Salgueiroetal.,2008;).Thelowgenetic differentiationamongpopulationsmaybetheoutcomeofhistorical long-distancedispersal.Infact,Bayesianclusteringshowedthree

geneticgroupsindicatingweakgeneticstructureamongsiteswith highlevelsofgeneﬂow.Theonlyexceptionwaspopulation18, whichshowedhighandsigniﬁcantvaluesofpairwiseF_ST forall comparisons.

Paleodistributiondynamicsaffectedgeneticboundariesbetween populations

Wefoundnoevidencethatpaleodistributionchangesduring theQuaternarysignificantlyaffectedthegeneticparametersana- lyzed.Thelackofsignificanteffectmaybeduetothequasistability ofG.soricina’sgeographicalrangesincetheLGM.Infact,ourfind- ingsshowaslightlylargergeographicalrangeduringtheLGMthan atthemid-Holoceneorpresent-day,showingthatG.soricinamay haveexpanded itsdistributionduringtheglacialperiodsof the Quaternary,whenseasonally dryforestsandopenedvegetation expandedtheirdistributions(Penningtonetal.,2000).

Nevertheless, thelarge areasof high suitabilitythrough the Quaternarymayhave beenhistoricalclimaticrefugia,maintain- ing populationsthroughtime. The historical refugiaprovided a sourceofmigrantstocolonizenewsuitableareaswhilethespecies wastrackingclimaticniche.ENMshowedthatseveralpopulations ofG.soricina inCentral Brazilwereoutside theclimaticrefugia (seeFig.3)fromtheLGMtopresent-day,explainingthegenetic discontinuityat CYB betweensome populations,suchas popu- lation17.Ontheotherhand,somepopulationswereconnected inthepast,suchaspopulation18,andcontemporarylandscape changes mayhave isolated populations leadingto low connectivity and genetic discontinuityat microsatelliteloci. Thus, the geneticboundariesamongpopulationsmaybedue toboth, the paleodistributiondynamicsintheQuaternary,andthecontempo- rarylandscapechanges.

Concludingremarks

Our ﬁndingsshowlossofgeneticdiversityinpopulationsof G.soricinaintheBrazilianCerradoduetocontemporarychanges in landscape,despitethehighdispersal ability.Our resultsalso showisolationofpopulationssurroundedbypasture,agriculture orurbanarea, increasinginbreeding.Lessdistanceamongnatu- ralvegetationremnantsmayalsoincreasepopulationconnectivity, decreasinggeneticdifferentiation.Althoughpaleodistributiondid notexplainvariationingeneticdiversityanddifferentiationamong populations,historicalrefugiamayexplainthepatternsofgenetic boundariesbetweenpairsofpopulations.

Ourﬁndingshighlighttheeffectsofhabitatlossandspatialdis- tanceamongremnantsingeneticdiversityanddifferentiationof highlymobileandwidelydistributedspecies,suchasthePallas’

long-tonguedbat,highlightingtheharmfuleffectsofhabitatlossfor batspeciesandgeneticdiversity(e.g.Meyeretal.,2009;Muylaert etal.,2016;Clearyetal.,2017)Ecologicalinteractionsmayalsobe affectedinthewayfragmentationaffectsG.soricinasinceecolog- icalprocesses,suchaspollinationandseeddispersalmayalsobe affected,modifyingecosystemdynamicsandregeneration.

Moreover,ourresultssuggestthat conservationplanningfor G.soricinamaybebasedonthelocalpatternoflandscapestruc- ture,protectingnaturalvegetationremnantstoincreasepopulation geneticdiversityandconnectivity.

Authorcontributions

RGCandMOconceivedthework.RGCandMPCTfunded the work.MOobtainedthedata.RGC,LCVandTBVcarriedoutanaly- ses.RGCwrotetheoriginaldraftandallauthorsapprovedtheﬁnal version.

(8)

176 R.G.Collevattietal./PerspectivesinEcologyandConservation18(2020)169–177 DNAsequences

CYBsequencedataareavailableatGenBank(https://www.ncbi.

nlm.nih.gov/),accessnumberfromMN719187toMN719369.

Conﬂictofintereststatement

Theauthorsdeclare thatthe researchwasconductedin the absenceofanycommercialorﬁnancialrelationshipsthatcouldbe construedasapotentialconﬂictofinterest.

Acknowledgments

ThisworkwassupportedbytheprojectCNPq/FAPEG/AUXPESQ CH007/2009andCAPES/PROCAD(projectno.88881.068425/2014- 01).M.Oprea receivedaCAPESfellowship.WethankACPavan, LSalles,LGeise,LAguiar,MCNascimento,KCFaria,WUieda,A Pedrozo,TBVieiraandPMendesfortissuesamples,andMatheus SLima-Ribeiroforhelpingwithecologicalnichemodelling.RGC andMPCThavebeencontinuouslysupportedbyCNPqandCAPES grants and fellowships,which we gratefully acknowledge. The authorizationfor sampling was granted by the Brazilian Insti- tute for Environment (ICMBio/MMA number 20982-1). Current researchisfundedbytheNationalInstituteforScienceandTech- nology(INCT)inEcology,EvolutionandBiodiversityConservation (MCTIC/CNPq/FAPEGprojectno.465610/2014-5).

AppendixA. Supplementarydata

Supplementarymaterialrelated tothis article canbefound, in the online version, at doi:https://doi.org/10.1016/j.pecon.

2020.06.006.

References

Aguiar,L.M.S.,Bernard,E.,Machado,R.B.,2014.Habitatuseandmovementsof GlossophagasoricinaandLonchophylladekeyseri(Chiroptera:Phyllostomidae) inaNeotropicalsavannah.Zoologia31,223–229.

Albrecht,L.,Meyer,C.F.J.,Kalko,E.K.V.,2007.Differentialmobilityintwosmall phyllostomidbats,ArtibeuswatsoniandMicronycterismicrotis,inafragmented neotropicallandscape.ActaTheriol.52,141–149.

Arruda,M.P.,Costa,W.P.,Recco-Pimentel,S.M.,2017.GeneticdiversityofMorato’s diggertoad,Proceratophrysmoratoi:spatialstructure,geneﬂow,effectivesize andtheneedfordifferentialmanagementstrategiesofpopulations.Genet.

Mol.Biol.40,502–514.

Balkenhol,N.,Pardini,R.,Cornelius,C.,Fernandes,F.,Sommer,S.,2013.

Landscape-levelcomparisonofgeneticdiversityanddifferentiationinasmall mammalinhabitingdifferentfragmentedlandscapesoftheBrazilianAtlantic Forest.Conserv.Genet.14,355–367.

Bardeleben,C.,Campbell,P.,Lara,M.,Moore,R.L.,2007.Isolationofpolymorphic tetranucleotidemicrosatellitemarkersforthesilkyshort-tailedbatCarollia brevicauda.Mol.Ecol.Resour.7,63–65.

Barros,J.S.,Bernard,E.,Ferreira,R.L.,2020.EcologicalpreferencesofNeotropical cavebatsinroostsiteselectionandtheirimplicationsforconservation.Basic Appl.Ecol.45,31–41.

Biollaz,F.,Bruyndonckx,N.,Beuneux,G.,Mucedda,M.,Goudet,J.,Christe,P.,2010.

GeneticisolationofinsularpopulationsoftheMaghrebianbat,Myotispunicus, intheMediterraneanbasin.J.Biogeogr.37,1557–1569.

Birky,C.W.,Maruyama,T.,Fuerst,P.,1983.Anapproachtopopulationand evolutionarygenetictheoryforgenesinmitochondriaandchloroplasts,and someresults.Genetics103,513–527.

Boscolo,D.,Metzger,J.P.,2011.Isolationdeterminespatternsofspeciespresence inhighlyfragmentedlandscapes.Ecography34,1–12.

Burland,T.M.,Wilmer,J.W.,2001.Seeinginthedark:molecularapproachesto studyofbatpopulations.BiologicalReview76,389–409.

Burnham,K.P.,Anderson,D.R.,2002.ModelSelectionandMultimodelInference:A PracticalInformation-theoreticApproach,2ndedn.Springer,NewYork.

Centeno-Cuadros,A.,Hulva,P.,Romportl,D.,Santoro,S.,Stríbná,T.,Shohami,D., Evin,A.,Tsoar,A.,Benda,P.,Horácek,I.,Nathan,R.,2017.Habitatuse,butnot geneﬂow,isinﬂuencedbyhumanactivitiesintwoecotypesofEgyptianfruit bat(Rousettusaegyptiacus).Mol.Ecol.26,6224–6237.

Cleary,K.,Waits,L.P.,Finegan,B.,2017.Comparativelandscapegeneticsoftwo frugivorousbatsinabiologicalcorridorundergoingagricultural

intensiﬁcation.Mol.Ecol.26,4603–4617.

Corander,J.,Marttinen,P.,Sirén,J.,Tang,J.,2008.EnhancedBayesianmodelingin BAPSsoftwareforlearninggeneticstructuresofpopulations.BMC Bioinformatics9,1–14.

Dechmann,D.K.N.,Garbely,E.,Kerth,G.,Garner,T.W.J.,2002.Highlypolymorphic microsatellitesforthestudyoftheround-earedbat,Tonatiasilvicola (d’Orbigny).Conserv.Genet.3,455–458.

Delaney,K.S.,Riley,S.P.D.,Fisher,R.N.,2010.Arapid,strong,andconvergent geneticresponsetourbanhabitatfragmentationinfourdivergentand widespreadvertebrates.PLoSOne5,e12767.

Diniz-Filho,J.A.F.,Bini,L.M.,Rangel,T.F.,Loyola,R.D.,Hof,C.,Nogues-Bravo,D., Araujo,M.B.,2009.Partitioningandmappinguncertaintiesinensemblesof forecastsofspeciesturnoverunderclimatechange.Ecography32,897–906.

Dixo,M.,Metzger,J.P.,Morgante,J.S.,Zamudio,K.R.,2009.Habitatfragmentation reducesgeneticdiversityandconnectivityamongtoadpopulationsinthe BrazilianAtlanticCoastalForest.Biolog.Conserv.142,1560–1569.

Do,C.,Waples,R.S.,Peel,D.,Macbeth,G.M.,Tillett,B.J.,Ovenden,J.R.,2014.

NeEstimatorV2:re-implementationofsoftwarefortheestimationof contemporaryeffectivepopulationsize(Ne)fromgeneticdata.Mol.Ecol.

Resour.14,209–214.

Excofﬁer,L.,Laval,G.,Schneider,S.,2005.Arlequinver.3.0:anintegratedsoftware packageforpopulationgeneticsdataanalysis.Evol.Bioinform.Online1,47–50.

Flanders,J.,Inoue-Murayama,M.,Rossiter,S.J.,Hill,D.A.,2016.Femalephilopatry andlimitedmale-biaseddispersalintheUssuritube-nosedbat,Murina ussuriensis.J.Mammal.97,545–553.

Forster,P.,Bandelt,H.J.,Rohl,A.,etal.,Softwareavailablefreeat2004.Network 4.2.0.1.FluxusTechnologyLtdawww.ﬂuxus-engineering.com.

Gorresen,P.M.,Willig,M.R.,2004.Landscaperesponsesofbatstohabitat fragmentationinAtlanticForestofParaguay.J.Mammal.85,688–697.

Gorresen,P.M.,Willig,M.R.,Strauss,R.E.,2005.Multivariateanalysisof scale-dependentassociationsbetweenbatsandlandscapestructure.Ecol.

Appl.15,2126–2136.

Gruber,B.,Adamack,A.,Freeavailableat2019.Package‘PopGenReport’.

http://www.popgenreport.org/.

Halczok,T.K.,Brändel,S.D.,Flores,V.,Puechmaille,S.J.,Tschanpka,M.,Page,R.A., Kerth,G.,2018.Male-biaseddispersalandthepotentialimpactof human-inducedhabitatmodiﬁcationsontheNeotropicalbatTrachops cirrhosus.Ecol.Evol.8,6065–6080.

Haldﬁeld,J.,Freeavailableat2019.PackageMSMCglmm:MCMCGeneralised LinearMixedModels.

https://cran.rproject.org/web/packages/MCMCglmm/index.html.

Hedrick,P.W.,2005.Astandardizedgeneticdifferentiationmeasure.Evolution59, 1633–1638.

Henry,M.,Pons,J.M.,Cosson,J.F.,2007.Foragingbehaviourofafrugivorousbat helpsbridgelandscapeconnectivityandecologicalprocessesinafragmented rainforest.J.Anim.Ecol.3,801–813.

Hewitt,G.,2000.ThegeneticlegacyoftheQuaternaryiceages.Nature405, 907–913.

Irwin,D.M.,Kocher,T.D.,Wilson,A.C.,1991.Evolutionofthecytochromebgeneof mammals.J.Mol.Evol.32,128–144.

Jackson,N.D.,Fahrig,L.,2016.Habitatamount,nothabitatconﬁguration,best predictspopulationgeneticstructureinfragmentedlandscapes.Landsc.Ecol.

31,951–968.

Johansson,M.,Primmer,C.R.,Merilä,J.,2007.Doeshabitatfragmentationreduce ﬁtnessandadaptability?Acasestudyofthecommonfrog(Ranatemporaria).

Mol.Ecol.16,2693–2700.

Jost,L.,2008.G_STanditsrelativesdonotmeasuredifferentiation.Mol.Ecol.17, 4015–4026.

Kerth,G.,Petit,E.,2005.Colonizationanddispersalinasocialspecies,the Bechstein’sbat(Myotisbechsteinii).Mol.Ecol.14,3943–3950.

Kuhner,M.K.,2006.Lamarc2.0:maximumlikelihoodandBayesianestimationof populationparameters.Bioinformatics22,768–770.

Lanfear,R.,Kokko,H.,Eyre-Walker,A.,2014.Populationsizeandtherateof evolution.TrendsEcol.Evol.(Amst.)29,33–41.

Legendre,P.,Legendre,L.,1998.NumericalEcology.Elsevier,Oxford.

Levy,E.,Byrne,M.,Huey,J.A.,Firman,R.C.,Ottewell,K.M.,2019.Limitedinﬂuence oflandscapeonthegeneticstructureofthreesmallmammalsina

heterogeneousaridenvironment.J.Biogeogr.46,539–551.

Lima-Ribeiro,M.S.,Varela,S.,González-Hernández,J.,Oliveira,G.,Diniz-Filho, J.A.F.,Terribile,L.C.,2015.EcoClimate:adatabaseofclimatedatafrommultiple modelsforpast,present,andfutureformacroecologistsandbiogeographers.

Biodivers.Inform.10,1–21.

McCracken,G.F.,Wilkinson,G.S.,2000.Batmatingsystems.In:Crichton,E.G., Krutzsch,P.H.(Eds.),ReproductiveBiologyofBats.AcademicPress,NewYork, pp.321–362.

Meyer,C.F.J.,Kalko,E.K.V.,Kerth,G.,2009.Small-scalefragmentationeffectson localgeneticdiversityintwophyllostomidbatswithdifferentdispersal abilitiesinPanama.Biotropica41,95–102.

Miranda,N.E.O.,Maciel,N.M.,Lima-Ribeiro,M.S.,Colli,G.R.,Haddad,C.F.B., Collevatti,R.G.,2019.DiversiﬁcationofthewidespreadNeotropicalfrog PhysalaemuscuvieriinresponsetoNeogene-Quaternarygeologicaleventsand climatedynamics.Mol.Phylogenet.Evol.132,67–80.

Muylaert,R.L.,Stevens,R.D.,Ribeiro,M.C.,2016.Thresholdeffectofhabitatlosson batrichnessincerrado-forestlandscapes.Ecol.Appl.26,1854–1867.

Nomura,T.,2008.Estimationofeffectivenumberofbreedersfrommolecular coancestryofsinglecohortsample.Evol.Appl.1,462–474.

(9)

Oprea,M.,Peixoto,F.P.,Resende,L.V.,Collevatti,R.G.,Telles,M.P.C.,2012.Isolation andcharacterizationof10microsatellitelociforPallas’long-tonguedbat Glossophagasoricina(Phyllostomidae).Genet.Mol.Resour.11,3518–3521.

Pennington,R.T.,Prado,D.E.,Pendry,C.A.,2000.Neotropicalseasonallydryforests andQuaternaryvegetationchanges.J.Biogeogr.27,261–273.

Peterman,W.E.,2018.ResistanceGA:anRpackagefortheoptimizationof resistancesurfacesusinggeneticalgorithms.MethodsEcol.Evol.9,1638–1647.

Pink,B.,1996.ReproductiveandSocialBehavioroftheBatGlossophagasoricina (Phyllostomidae;Glossophaginae).UnpublishedPhil.Thesis.Universityof Erlangen-Nuremberg.

Pritchard,J.K.,Stephens,M.,Donnelly,P.,2000.Inferenceofpopulationstructure usingmultilocusgenotypedata.Genetics155,945–959.

Puttker,T.,Bueno,A.A.,Barros,C.S.,Sommer,S.,Pardini,R.,2011.Immigration ratesinfragmentedlandscapes—empiricalevidencefortheimportanceof habitatamountforspeciespersistence.PLoSOne6,e27963.

RCoreTeam,Availableat2019.R:ALanguageandEnvironmentforStatistical Computing.RFoundationforStatisticalComputing,Vienna,Austria http://www.R-project.org.

Rangel,T.F.,Diniz-Filho,J.A.F.,Bini,L.M.,2010.SAM:acomprehensiveapplication forspatialanalysisinmacroecology.Ecography33,46–50.

Razgour,O.,Rebelo,H.,Puechmaille,S.J.,Juste,J.,Ibá ˜nez,C.,Kiefer,A.,Burke,T., Dawson,D.A.,Jones,G.,2014.Scale-dependenteffectsoflandscapevariables ongeneﬂowandpopulationstructureinbats.Divers.Distrib.20,1173–1185.

Reis,N.R.,Peracchi,A.L.,Pedro,W.A.,Lima,I.P.,2005.MorcegosdoBrasil.

UniversidadeEstadualdeLondrina,Londrina.

Ripperger,S.P.,Tschapka,M.,Kalko,E.K.V.,Rodriguez-Herrera,B.,Mayer,F.,2013.

Lifeinamosaiclandscape:anthropogenichabitatfragmentationaffectsgenetic populationstructureinafrugivorousbatspecies.Conserv.Genet.14,925–934.

Rossiter,S.J.,Jones,G.,Ransome,R.D.,Barratt,E.M.,2000.Parentage,reproductive successandbreedingbehaviourinthegreaterhorseshoebat(Rhinolophus ferrumequinum).Proc.R.Soc.B267,545–551.

Russell,A.L.,Medellín,R.A.,McCracken,G.F.,2005.Geneticvariationandmigration intheMexicanfree-tailedbat(Tadaridabrasiliensismexicana).Mol.Ecol.14, 2207–2222.

Salgueiro,P.,Palmeirim,J.M.,Ruedi,M.,Coelho,M.M.,2008.Geneﬂowand populationstructureoftheendemicAzoreanbat(Nyctalusazoreum)basedon microsatellites:implicationsforconservation.Conserv.Genet.9,1163–1171.

Silva,J.F.,Farinas,M.R.,Felﬁli,J.M.,Klink,C.A.,2006.Spatialheterogeneity,landuse andconservationinthecerradoregionofBrazil.J.Biogeogr.33,536–548.

Umetsu,F.,Metzger,J.P.,Pardini,R.,2008.Importanceofestimatingmatrixquality formodelingspeciesdistributionincomplextropicallandscapes:atestwith Atlanticforestsmallmammals.Ecography31,359–370.

Wagner,H.,Fortin,M.J.,2013.Aconceptualframeworkforthespatialanalysisof landscapegeneticdata.Conserv.Genet.14,253–261.

Willig,M.R.,Camilo,G.R.,Nobile,S.J.,1993.Dietaryoverlapinfrugivorousand insectivorousbatsfromedaphiccerradohabitatsofBrazil.J.Mammal.74, 117–128.

Wilmer,J.W.,Hall,L.,Barratt,E.,Moritz,C.,1999.Geneticstructureand male-mediatedgeneﬂowintheghostbat(Macrodermagigas).Evolution53, 1582–1591.

Wright,S.,1951.Thegeneticstructureofpopulations.Ann.Eugen.15,323–354.

Zurcher,A.A.,Sparks,D.W.,Bennett,V.J.,2010.Whythebatdidnotcrosstheroad?

ActaChiroptera12,337–340.