Experimental evidence that feral cats cause local exti - PDF document

IntroductionAlmostaquarteroftheworldÕsterrestrialmammalspeciesarethreatenedwithextinction(Hoffmannetal.2011;Dietal.2012),mainlyduetohabitatloss,habitatdeg-radationandoverhunting(Schipperetal.2008).Small-andmedium-sizedinvasivepredators,notablyferalcatsFelissylvestriscatusandblackratsRattusrattus,havecauseddeclinesinsmall-mammalpopulations,butsofarthishasmostlybeendocumentedonsmallislands(Nogalesetal.2004;Bonnaudetal.2011).Thepowerfuleffectofpredatorsonislandfaunasmaybeexplainedbyseveralfac-tors:(1)small-to-medium-sizedintroducedpredatorsoftenoccurathighdensitiesonislands;(2)nativeanimalsonislandsmaybenaõvetopredators,and(3)nativespeciesmayhavelowtotalpopulationsizes;and(4)largermam-malianpredators,whichcanreducetheimpactofsmall-tomedium-sizedpredators(Ritchie&Johnson2009),areabsentfrommanyislands.Additionally,thedateofarrivalofnon-nativepredatorsisoftenbetterdocumentedonislands,allowingformoreaccuratedocumentationofcausesandresponsesthanonthemainland.Catshavebeenspreadworld-widebypeople(Denny&Dickman2010),andwhileitiscleartheyhavecausedextinctionofmanyvertebratespeciesonislands(Dickman1992;Bonnaudetal.2011;Medinaetal.2011;Loss,Will&Marra2013),thereismuchlessevidenceforimpactoncontinentalfaunas(Woods,McDonald&Harris2003;Loss,Will&Marra2013).ArecentstudybyLoss,Will&Marra(2013)intheUnitedStatesshowedthattheimpactoffree-rangingcatsislikelytobemuchhigherthanpreviouslythought,withmorethan20billionmam-malskilledbycatseachyearintheUnitedStatesalone.TranslatedtoAustralia,thisimpliesapotentiallysigniÞ-cantimpactofcatsoncurrentlywidespreadsmall-mammalspecies.Sincethe15thcentury,Australiahaslostmoremam-malspecies(atleast23)thananyothercontinent,accountingforapproximatelyone-thirdofallrecentglo-balextinctionsofmammals(Johnson2006).Manyoftheseextinctionshavebeenattributedtoinvasivepreda-tors,especiallytheredfoxVulpesvulpes(Johnson2006).AlthoughAustraliaÕsCommonwealthThreatAbatementPlanforpredationbyferalcats(2008)lists36mammalsasthreatenedbycats,evidencelinkingferalcatstopastmammalianextinctionsisequivocal(Denny&Dickman2010),exceptinafewinstancesonAustralianislands(Dickman1992).Untilrecently,thetropicalsavannasofnorthernAustraliaseemedtobeunaffectedbythedeclinesandextinctionsthatoccurredamongmammalsofthecentralaridzoneandsouthernAustralia(Woinarskietal.Thiswasattributedprimarilytotheabsencethereoffoxes.However,smallmammalsarecurrentlydecliningacrossmuchofnorthernAustralia(Woinarskietal.2011a;Fisheretal.2013).Sixspeciesareatimminentriskofextinction,andmorethan20species(of40speciesinthesmall-to-mediumbodysizerange)havesufferedseverereductionsingeographicrange(Fitzsimonsetal.Thereareindicationsthatferalcatsmaybeatleastpartlyresponsibleforthesedeclines(Fisheretal.2013).Forexample,thespeciesdecliningmostseverelyareinthepreferredprey-sizerangeofferalcats(1500g)andoccupyrelativelyopenhabitatswhereexposuretopreda-tionislikelytobehigh(Fisheretal.2013).SpeciesthataredecliningacrossthemainlandofnorthernAustraliamostlyhavestablepopulationsoncat-freeislands,buthavedeclinedonislandswherecatshaverecentlybeenintroduced(Dickman1992;Woinarskietal.2011b).ThisevidenceoftheeffectofcatpredationiscounteredbythefactthatcatshavebeenpresentinnorthernAustraliaforatleast100years,andaviewthattheabundanceofferalcatsinAustraliaÕstropicalsavannasisgenerallytoolowtoaccountformajordeclinesaffectingmanyspeciessimultaneouslyoverlargeareas(Bradshawetal.Here,weprovidetheÞrstexperimentaltestoftheeffectofcatpredationondemographyofsmallmammalsinnorthernAustralia.Weusedareplicatedexperimenttodirectlymeasuretheimpactofferalcatsonthesurvivalandpopulationgrowthratesofasmall-mammalspecies,whileaccountingforthepotentialeffectsoftheothermammalianpredatorinthearea,thedingoCanisdingousingcamerastomonitorpredatoractivity.Theexperi-mentalenclosureswerelocatedintropicalsavannatypicalofenvironmentsthathaverecentlyexperiencedsmall-mammaldeclines(Woinarskietal.2011a),andinwhichsmall-mammalabundanceisnowverylow(SarahLegge,unpublisheddata).Previousmonitoringintheareasug-gestedthatcatdensitywastypicalofthelow-densitypop-ulationsthatoccurelsewhereinnorthernAustralia03km,Ridpath1990;KatherineTuft&SarahLegge,unpublisheddata).Wemonitoredthefateofsmall-mam-malpopulationsthathadbeenintroducedintotwopaired5ha)enclosuresinnaturallandscapesthatwereeitherexposedtoorprotectedfromcats,testingthepredictionthatpopulationgrowthratesandsurvivalwouldbesignif-icantlylowerinthepredator-accessiblethanthepredator-inaccessibletreatments.MaterialsandmethodsSTUDYSITEThisstudywasconductedatWongalaraWildlifeSanctuaryS,134E),a192200hareserveintheNorthernTerri-tory(NT),Australia.WongalarawasapastoralpropertygrazedatlowintensitybycattleBostaurusxindicus,buffalo,horsesEquuscaballusanddonkeysEquusasinus,untilitwasacquiredbytheAustralianWildlifeConservancy(AWC)in2006.Managementhassincefocussedonremovalofgrazingungulates,andÞremanagementdesignedtoreducethesize,intensityandfrequencyofÞres.Theonlymammalianpredatorapartfromtheferalcatcurrentlypresentisthedingo;dingoesweresuppressedbypoisonbaitingbeforethepropertywasacquiredbytheAWC,butarenowuncontrolled.2014TheAuthors.JournalofAppliedEcology2014BritishEcologicalSociety,JournalofAppliedEcology,1486Ð1493Feralcatsextirpatesmallmammals I,trappingcommenced2monthsafterreleaseofrats.Seventrap-pingsessionsofbetweentwoandfourconsecutivetrapnightswerecarriedout(24trapnightsintotal).InEnclosureII,trap-pingcommenced6monthsafterthereleaseofratsbecauseoflimitedaccessduringthewetseason.Inthisenclosure,ninetrap-pingsessionsofbetweentwoandfourconsecutivenightsoccurred(29nightsintotal;TableS1,Supportinginformation).Ateachcapture,individualswereidentiÞed,weighedandsexed.BodyconditionwasscoredusingthemethodofUllmann-e&Foltz(1999),fromcategory1emaciatedtoobese.Recruits(i.e.individualstrappedintheenclosuresthatwerenotpartoftheinitialrelease)werePIT-taggedatÞrstcap-ture.Anyuncollaredratslargerthan100gwereradiocollared,resultingin49rats(EnclosureIlow:7,high:13;EnclosureIIlow:13,high:16)beingradiocollaredandtrackedoverthecourseofthestudy.DETECTIONOFPREDATORSWeusedheat-in-motioncameras(ReconyxPC800HyperÞre,Hol-men,WI,USA)andweeklyvisualinspectionsfortracksandscatsinthesoftsoilaroundtheinsideandoutsideoftheperimeterfencestodetectpredators.Atleastthreecameras(butupto13andalwaysthesamenumberofcamerasatatime)wereplacedinsideeachhalfoftheenclosure,andonecamerawasinstalledateachsideontheoutsideofthefencesofeachenclosure(Fig.1).Cameraswereun-baited,toavoidattractingpredators.ANALYSISToestimatepopulationsizeandindividualsurvivalofrats,cap-recapturedatawereanalysedusingPollockÕsrobustsurveydesign(Pollocketal.1990).Trappingdatawereusedtocreateindividualencounterhistoriesforeachratindicatingcapture(1)ornon-capture(0)duringeachsamplingoccasion.DatawereanalysedusingaHugginsmodelwitharobustdesigninprogramv7.1(White&Burnham1999)andplottedusing(http://cran.r-project.org).Therearemultipleprimarysamplingperiods(i.e.repeatednightsofsamplingduringatrappingepisode)inPollockÕsrobustsurveydesign,whichareseparatedbylongerperiodsoftimeduringwhichthepopulationisopentogains(birthsandimmigration)andlosses(deathsandemigra-tion).Dataineachprimaryperiodarepooledtoestimatesur-vivalprobability,betweenprimarysamplingperiods.Withineachprimarysamplingperiod,thereareseveralsecondarysam-plingperiodsseparatedbyashortperiodduringwhichthepopu-lationisassumedclosed.DatafromeachsecondarysamplingperiodwereusedtoestimatetheprobabilityofÞrstcaptureandrecapture.Althoughrobustdesignmodelscanbeusedtoestimateprobabilitiesofmigrationintoandoutofthestudyarea,hereweÞxedthemigrationparameterstozero.Thiswasduetothepresenceofafencethatprecludedanimalmovement(i.e.probabilityofemigration,,andprobabilityofremainingawayfromthestudyareaiftheindividualhadpreviouslyemigrated,).ToensureparameteridentiÞcation,weassumedtheprobabil-ityofÞrstcaptureandprobabilityofrecapturewereconstant(.)withineachtrappingsession.Forwardmodelselectionwasused,wherebythesimplestnullmodelwasÞtted(i.e.(.)),representingconstantsur-(.),constantprobabilityofcapture(.)andconstantprobabilityofrecapture(.),aswellasnobehaviouraleffectonthecaptureprocess,thatis,chanceofÞrstcaptureisthesameaschanceofrecapture().ModelcomplexitywasintroducedbyÞrstinvestigatingwhethertherewasabehaviouraleffectinthecaptureprocess(i.e.)oraseasonaleffect(i.e.(session).(session)model),andthenwhethercovariatesinßuencedsurvival(i.e.thegrouptowhichtheindividualbelonged,eitherreleasedorrecruitedonsite,asessioneffect,sex,andwhetheritwasinthelow-orhigh-fencetreatmentoftheenclosure).Modelselec-tionusedAkaikeÕsinformationcriterionadjustedforsmallsam-plesize(AICc).ModelswereassumedequivalentiftheAICcvaluedifferedbylessthan2fromthemodelwiththelowestAICc(Burnham&Anderson2002).Todescribefactorsaffectingsurvival,15candidatemodelswerechosenaprioriandÞttedsep-aratelyforEnclosureIandEnclosureII(TableS2,Supportinginformation).Survivalofradiocollaredratswasanalysedasmeannumberofmonthstrackedwhileknowntobealive(indicatedbylastloca-tionchange),andthiswascomparedbetweencompartmentsforeachenclosureusingtwo-tailed-testsinMicrosoftExcel2010.Homogeneityofvariancewastestedusingan-test.Bodycondi-tionandreproductivestatuswerecomparedbetweentreatmentsbycomparingweightchangesbetweenthereleaseandÞrsttrap-pingsession,alsousingtwo-tailedUsingcameratrapdata,wecalculatedanactivityindexinv3.0.3formammalianpredatorsbydividingthenumberofencountersbythenumberofcamerastoaccountformalfunction-ingcameras.Thisindexwascalculatedforeachnight,thenaver-agedmonthlyandreportedwith95%binomialconÞdenceintervals.Wealsorecordedallinstancesofpredatoroccurrencesinsidetheenclosuresandaveragedhowregularlytheseoccurred.PREDATORACTIVITYANDINCURSIONSCatsorsignsofcatswereneverdetectedinsidethehigh-fencetreatmentofeitherenclosure.However,cattracksweredetectedinsidethelow-fencetreatmentofEnclosureIlessthan1monthafterratswerereleased(Fig.2).Atleasttwoindividualcatsweredetectedoncameraatthatsite4weekslater(Fig.2).Thereafter,catdetectionsrecurred,onaverage,everyweek(JuneOctober2012:9dayintervals;October2012April2013:60dayinter-vals)inthelow-fencetreatmentofEnclosureIuntiltheendofthestudy.Therewasalongerdelayofnearly1yearbeforecatswereÞrstdetectedinsidethelow-fencetreatmentofEnclosureII(Fig.3);duringtheentiresurveyperiod,therewereonlytwocat-detectioneventsoftwodifferentindividualsinsidethistreatment.DingoesorsignsofdingoeswereneverrecordedinsideeitherEnclo-sureIorIIineithertreatment.Cameratrapsontheoutsideoftheenclosuresindicatedthatingeneral,dingoactivitywassimilartocatactivityoutsideEnclosureI(Fig.S2a,Supportinginformation),whileoutsideEnclosureII,dingoactivitywasvariableandgreaterthancatactivityforthedurationofthestudy(Fig.S2b,Supportinginformation).CatactivitywassimilararoundEnclosureI(Fig.S2a,Supportinginforma-tion)andEnclosureII(Fig.S2b,Supportinginformation).2014TheAuthors.JournalofAppliedEcology2014BritishEcologicalSociety,JournalofAppliedEcology,1486Ð1493Feralcatsextirpatesmallmammals recordedbytheendofthestudyperiod,whilelittlerecruitmentwasrecordedinthelow-fencetreatment(Fig.2).Similarly,capturesofrecruitedratsatEnclosureIIwerealwaysmuchhigherinthehigh-fencecomparedtothelow-fencetreatment,withatotalof39comparedtosixwild-bornindividuals,respectively(Fig.3).Weestimatedtheabundanceofeachgroupofrats(i.e.releasedorrecruited)foreachtreatmentfromeachenclo-sureusingthemodelwiththelowestAIC.ForallratsatEnclosureI(Fig.2),abundanceinthelow-fencetreat-mentwasestimatedaszerofromtheÞrstsamplingperiod(2monthsafterrelease)untilourstudyceased.Theabun-danceofallreleasedratsinthehigh-fencetreatmentofEnclosureIdeclinedslowlyoverthe12monthsfollowingtheirrelease(Fig.2).TheabundanceofratsatEnclosureIIdeclinedmorerapidlyinthelow-fencethaninthehigh-fencetreatment(Fig.3).Threeofthefourradio-trackedfemalesinthelow-fencetreatmentweredeadwithin2monthsoftheirrelease,andneitheroftheuncollaredratswaseverrecaptured.AftertheÞrstsightingofcatsinthelow-fencetreatment,about1yearaftertheratswerereleased,thelastreleasedfemaleratwasmissingandthepopulationabundancedeclinedby75%.Afterthesecondcatdetection,noratswererecordedinthelow-fencetreat-ment(Fig.3).Radiocollaredratswereneverrecordedmovingoutsidetheenclosures.Whileweoccasionallyfoundrattracksalongtheperipheryoftheenclosures,werecordednoevi-denceofratstryingtoexittheenclosure.InEnclosureI,therewasasigniÞcantdifference(14,d.f.001)intheknownminimumsurvivaltimeofradio-collaredadultratsinthelow-fencetreatment(126months)comparedtothehigh-fencetreatmentSE034months).ForEnclosureII,therewasnodetecteddifferenceintheknownminimumsurvivaltime25,d.f.23)betweenthelow-fenceSE12months)andhigh-fence(69months)treatment(Fig.S4,Supportinginformation).Oftheradiocollaredratsassumedorknowntobedeadattheendofthestudy,thecorpsesofonly50%(threeoutofsix)wererecoveredfromthehigh-fencetreatmentand43%(threeoutofseven)fromthelow-fencetreatmentofEnclosureI,and55%(Þveoutofnine)inthehigh-fencetreatmentand25%(threeoutof12)inthelow-fencetreatmentofEnclosureII.Onaverage,thebodyconditionofratsintheenclosureswasÔwellconditionedÕ(categoryBC3;Ullmann-CullerFoltz1999).WeightsofreleasedrecapturedratsinEnclo-sureIincreased26%duringthe2monthsbeforetheÞrsttrappingsessioninthehigh-fencetreatment(Fig.S4,Supportinginformation);weightchangesinthelow-fencetreatmentofEnclosureIcouldnotbeassessedasratswerenotrecapturedafterrelease.Overthe6monthsbetweenthereleaseandtheÞrsttrappingsessioninEnclosureII,themeanweightsofreleasedrecapturedratsincreasedby71%inthelow-fencetreatmentand40%inthehigh-fencetreatment.ThemeanweightsofratsinthetwohalvesoftheenclosuredidnotdiffersigniÞcantly42,d.f.18).SurvivorshipoftheÞvewild-bornratswassimilartotheÞrst-generationanimalsbredincaptivity(threeoftheÞvewildratsdiedwithinamonthofrelease,onediedwithinsix,andthelastwildratdiedwithin8monthsofrelease).DiscussionWefoundamarkeddifferenceinthesurvivalandpopula-tiongrowthrateofratsbetweenthelow-andhigh-fencetreatments.Thiswasevidentlyduetopredationbycats.Populationdeclineswerecloselyrelatedtoincursionsbycats.Therewasnoevidencethatdingoeseverenteredanyenclosure,andavianandreptilianpredators(snakesandgoannas)wouldhavebeenequallylikelytoenterallcom-partments.Ratsinourexperimentalenclosuresmain-tainedgoodbodyconditionthroughoutthestudyperiod.Therefore,differencesinsurvivalwerenotrelatedtofoodlimitationorstressfromtheindirecteffectsofpredation.Inbothreplicates,ratswereextirpatedfromlow-fence(hereaftercalledÔcat-accessibleÕ)butnothigh-fence(here-aftercalledÔcat-inaccessibleÕ)treatments.However,thedemographicmechanismsleadingtoextirpationdifferedinourtworeplicates.InEnclosureI,reducedsurvivalinthecat-accessibletreatmentcausedrapidextirpationwithin2monthsofreleasingtherats;inEnclosureII,lowerlevelsofcatpredationandalowrecruitmentratebytheratsinthecat-accessibletreatmentcausedamoregradualdeclineintheratpopulationover15years.Althoughwewereunabletodeterminecauseofdeathofindividuals,itwashighlylikelythatcatswerethecauseoftheirultimateextirpation.AlthoughtheratpopulationswereconÞnedtofencedareas,ourenclosureswerelargerelativetothehomerangesizeofindividualrats(asindicatedfromourradiotrackingresults)andofferedsuf-Þcientspaceforratstoshelterindensevegetation,logsandhollowtrees.Thus,eventhoughcatsoccuratlowpopulationdensitiesintropicalsavannasofnorthern Table1.ModelselectionresultsforthetopthreeÞttedmodels(accordingtoAIC)toEnclosureIandEnclosureIIdata(fullresultsareavailableinTableS2,SupportingInformation).Fourcovariateswereinvestigated:thetreatment(lowf=low-fence),asessioneffect(,i.e.timebetweenprimaryoccasions),whethertheanimalsbelongedtothecaptive-bredorrecruitedgroup(andsexoftherat().Chanceofcapturewasassumedequaltothechanceofrecaptureandbothwereassumedtobeconstant(.).Migration()wasassumedtobezeroModelAICcdAICcNo.Par.DevianceEnclosureI(.)5147503508(sess)523298539504(.)523648893517EnclosureII(.)73293010712(.)7331502211710(.)7425896537362014TheAuthors.JournalofAppliedEcology2014BritishEcologicalSociety,JournalofAppliedEcology,1486Ð1493Feralcatsextirpatesmallmammals speciesinnorthernAustralia:aspotlightonWorldHeritageAreaKak-aduNationalPark.WildlifeResearch,419Breed,B.&Ford,F.(2007)NativeMiceandRats.CSIROPublishing,Collingwood,Victoria,Australia.Brook,L.A.,Johnson,C.N.&Ritchie,E.G.(2012)Effectsofpredatorcontrolonbehaviourofanapexpredatorandindirectconsequencesformesopredatorsuppression.JournalofAppliedEcology,1278Burnham,K.P.&Anderson,D.R.(2002)ModelSelectionandMultimodelInference:APracticalInformation-TheoreticApproach,2ndedn.Springer-Verlag,NewYork,USA.Childs,J.E.(1986)Size-dependentpredationonrats(Rattusnorvegicus)byhousecats(Feliscatus)inanurbansetting.JournalofMammalogy,196Cushing,B.S.(1985)Estrousmiceandvulnerabilitytoweaselpredation.,1976Denny,E.A.&Dickman,C.(2010)ReviewofCatEcologyandManage-mentStrategiesinAustralia.InvasiveAnimalsCooperativeResearchCentre,Canberra,ACT,Australia.DiMarco,M.,Cardillo,M.,Possingham,H.P.,Wilson,K.A,Blomberg,S.P.,Boitani,L.&Rondinini,C.(2012)Anovelapproachforglobalmammalextinctionriskreduction.ConservationLetters,134Dickman,C.R.(1992)ConservationofmammalsintheAustralasianregion:theimportanceofislands.AustraliaandtheGlobalEnvironmen-talCrisis(edsJ.N.Coles&J.M.Drew),pp.175214.AcademyPress,Canberra,ACT,Australia.Fisher,D.O.etal.(2013)ThecurrentdeclineoftropicalmarsupialsinAustralia:ishistoryrepeating?GlobalEcologyandBiogeographyFitzsimons,J.,Legge,S.,Traill,B.&Woinarski,J.(2010)IntoOblivion?TheDisappearingNativeMammalsofNorthernAustralia.NatureConservancy,Melbourne,VIC.Greenville,A.C.,Wardle,G.M.&Dickman,C.R.(2013)ExtremerainfalleventspredictirruptionsofratplaguesincentralAustralia.,754Hoffmann,M.,Belant,J.L.,Chanson,J.S.,Cox,N.A.,Lamoreux,J.,Rodrigues,A.S.L.,Schipper,J.&Stuart,S.N.(2011)ThechangingfatesoftheworldÕsmammals..PhilosophicalTransactionsoftheRoyalSocietyofLondon.SeriesB,BiologicalSciences,2598IUCN2013.IUCNRedListofthreatenedspecies.Version2013.2.Avail-ableat:www.iucnredlist.org.Downloadedon07April2014.Johnson,C.(2006)Australia’sMammalExtinctions:A50,000YearHis-.CambridgeUniversityPress,PortMelbourne,Victoria,Australia.Kennedy,M.,Phillips,B.L.,Legge,S.,Murphy,S.A.&Faulkner,R.A.(2012)DodingoessuppresstheactivityofferalcatsinnorthernAustra-AustralEcology,134Loss,S.R.,Will,T.&Marra,P.P.(2013)Theimpactoffree-rangingdomesticcatsonwildlifeoftheUnitedStates.NatureCommunicationsMedina,F.M.,Bonnaud,E.,Vidal,E.,Tershy,B.R.,Zavaleta,E.S.,Donlan,C.J.etal.(2011)Aglobalreviewoftheimpactsofinvasivecatsonislandendangeredvertebrates.GlobalChangeBiology,3503Nogales,M.,Martõn,A.,Tershy,B.R.,Donlan,C.J.,Veitch,D.,Puerta,N.,Wood,B.&Alonso,J.(2004)AreviewofferalcateradicationonConservationBiology,310Pollock,K.H.,Nichols,J.D.,Brownie,C.&Hines,J.E.(1990)Statisticalinferenceforcapture-recaptureexperiments.WildlifeMonographsPuan,C.L.,Goldizen,A.W.,Zakaria,M.&Baxter,G.S.(2011)Under-standingofrelationshipsbetweengroundcoverandratabundances:anintegrativeapproachformanagementoftheoilpalmagroecosystem.CropProtection,1263Read,J.&Eldridge,S.(2010)Anoptimisedrapiddetectiontechniqueforsimultaneouslymonitoringactivityofrabbits,cats,foxesanddingoesintherangelands.RangelandJournal,389Ridpath,M.G.(1990)Feralmammalsandtheirenvironment.Australia:Landscape,EcologyandManintheNorthernLowlandsC.D.Haynes,M.G.Ridpath&M.A.J.Williams),pp.169191.A.A.Balkema,Rotterdem.Ritchie,E.G.&Johnson,C.N.(2009)Predatorinteractions,mesopredatorreleaseandbiodiversityconservation.EcologyLetters,982Robley,A.,Purdey,D.,Johnston,M.,Lindeman,M.,Busana,F.&Long,K.(2007)Experimentaltrialstodetermineeffectivefencedesignsforferalcatandfoxexclusion.EcologicalManagement&RestorationSchipper,J.etal.(2008)ThestatusoftheworldÕslandandmarinemam-mals:diversity,threat,andknowledge.,225Stokes,V.L.,Pech,R.P.,Banks,P.B.&Arthur,A.D.(2004)ForagingbehaviourandhabitatusebyAntechinusavipesSminthopsismuri-(Marsupialia:Dasyuridae)inresponsetopredationriskineucalyptBiologicalConservation,331Sutherland,D.R.&Predavec,M.(1999)TheeffectsofmoonlightonmicrohabitatusebyAntechinusagilis(Marsupialia:Dasyuridae).Austra-lianJournalofZoologye,M.H.&Foltz,J.(1999)Bodyconditionscoring:arapidandaccuratemethodforassessinghealthstatusinmice.LaboratoryAnimalScience,319White,G.C.&Burnham,K.P.(1999)ProgramMARK:survivalestima-tionfrompopulationsofmarkedanimals.BirdStudy,S120Woinarski,J.C.Z.,Mackey,B.,Nix,H.&Trail,B.(2007)ThenatureofNorthernAustralia:naturalvalues,ecologicalprocessesandfutureper-.ANUe-press,Canberra,ACT,Australia.Woinarski,J.C.Z.,Legge,S.,Fitzsimons,J.A.,Traill,B.j.&Burbidge,etal.(2011a)ThedisappearingmammalfaunaofnorthernAustralia:context,cause,andresponse.ConservationLetters,192Woinarski,J.C.Z.,Ward,S.,Mahney,T.,Bradley,J.,Brennan,K.,Ziembicki,M.&Fisher,A.(2011b)ThemammalfaunaoftheSirEdwardPellewislandgroup,NorthernTerritory,Australia:refugeandWildlifeResearch,307Woods,M.,McDonald,R.A.&Harris,S.(2003)PredationofwildlifebydomesticcatsFeliscatusinGreatBritain.MammalReviewReceived16January2014;accepted31July2014HandlingEditor:ChrisFridSupportingInformationAdditionalSupportingInformationmaybefoundintheonlineversionofthisarticle.Fig.S1.ThefencedesigninEnclosureIandEnclosureII.Fig.S2.DingoandcatactivitysurroundingEnclosureIandEnclosureII.Fig.S3.Meanvegetationcoverinpercentat10cmaboveFig.S4.Survivalofradio-collaredreleasedrats.Fig.S5.WeightchangesofallratsbetweentheirreleaseandtheÞrsttrappingsession.TableS1.Summaryofreleaseinformationandtrappingoccasions.TableS2.Resultsofallselectedmodels.2014TheAuthors.JournalofAppliedEcology2014BritishEcologicalSociety,JournalofAppliedEcology,1486Ð1493Feralcatsextirpatesmallmammals