Evolutionary Applications. 2017;1–20.  | 1wileyonlinelibrary.com/journal/eva

Received:22October2015  |  Accepted:29March2017DOI: 10.1111/eva.12484

O R I G I N A L A R T I C L E

Severe consequences of habitat fragmentation on genetic diversity of an endangered Australian freshwater fish: A call for assisted gene flow

Alexandra Pavlova1  | Luciano B. Beheregaray2 | Rhys Coleman3 | Dean Gilligan4 |  Katherine A. Harrisson1,5,6 | Brett A. Ingram7 | Joanne Kearns5 | Annika M. Lamb1 |  Mark Lintermans8 | Jarod Lyon5 | Thuy T. T. Nguyen9 | Minami Sasaki2 | Zeb Tonkin5 |  Jian D. L. Yen10 | Paul Sunnucks1

1SchoolofBiologicalSciences,ClaytonCampus,MonashUniversity,Clayton,VIC,Australia2SchoolofBiologicalSciences,FlindersUniversity,Adelaide,SA,Australia3AppliedResearch,MelbourneWater,Docklands,VIC,Australia4FreshwaterEcosystemsResearch,NSWDepartmentofPrimaryIndustries–Fisheries,BatemansBay,NSW,Australia5DepartmentofEnvironment,LandWaterandPlanning,ArthurRylahInstitute,Land,FireandEnvironment,Heidelberg,VIC,Australia6DepartmentofEcologyEnvironmentandEvolution,SchoolofLifeSciences,LaTrobeUniversity,Bundoora,Victoria,3083,Australia7DepartmentofEconomicDevelopment,Jobs,TransportandResources,FisheriesVictoria,Alexandra,VIC,Australia8InstituteforAppliedEcology,UniversityofCanberra,Canberra,ACT,Australia9AgricultureVictoria,AgriBio,CentreforAgriBioscience,Bundoora,VIC,Australia10SchoolofPhysicsandAstronomy,ClaytonCampus,MonashUniversity,Clayton,VIC,Australia

ThisisanopenaccessarticleunderthetermsoftheCreativeCommonsAttributionLicense,whichpermitsuse,distributionandreproductioninanymedium,providedtheoriginalworkisproperlycited.©2017TheAuthors.Evolutionary ApplicationspublishedbyJohnWiley&SonsLtd

CorrespondenceAlexandraPavlova,SchoolofBiologicalSciences,ClaytonCampus,MonashUniversity,Clayton,VIC,Australia.Email:alexandra.pavlova@monash.edu

Funding informationAustralianResearchCouncil,Grant/AwardNumber:LP110200017;VICDepartmentofSustainabilityandEnvironment(nowDELWP-DepartmentofEnvironment,Land,Water&Planning);IconWater(formerlyACTEWCorporation);MelbourneWater;FisheriesVictoria(nowwithinDEDJTR-DepartmentofEconomicDevelopment,Jobs,TransportandResources);Grant/AwardNumber:RecreationalFishingGrantsprogram;NSWDPI-Fisheries;GoulburnBrokenCatchmentManagementAuthority

AbstractGeneticdiversityunderpinstheabilityofpopulationstopersistandadapttoenviron-mentalchanges.Substantialempiricaldatashowthatgeneticdiversityrapidlydeterio-rates in small and isolatedpopulationsdue togeneticdrift, leading to reduction inadaptivepotentialandfitnessandincreaseininbreeding.Assistedgeneflow(e.g.viatranslocations)canreversethesetrends,butlackofdataonfitnesslossandfearofimpairingpopulation“uniqueness”oftenpreventsmanagersfromacting.Here,weusepopulation genetic and riverscape genetic analyses and simulations to explore theconsequencesofextensivehabitatlossandfragmentationonpopulationgeneticdi-versityandfuturepopulationtrajectoriesofanendangeredAustralianfreshwaterfish,MacquarieperchMacquaria australasica.Usingguidelines to assess the riskofout-breedingdepressionunderadmixture,wedeveloprecommendationsforpopulationmanagement,identifypopulationsrequiringgeneticrescueand/orgeneticrestorationandpotentialdonorsources.WefoundthatmostremainingpopulationsofMacquarieperchhavelowgeneticdiversity,andeffectivepopulationsizesbelowthethresholdrequiredtoretainadaptivepotential.Oursimulationsshowedthatundermanagementinaction, smaller populations of Macquarie perch will face inbreeding depression

2  |     PAVLAP et PVl.

1  | INTRODUCTION

Aprimarygoalofconservationmanagementistoimprovetheadaptivepotentialofpopulations,thatis,theabilityofpopulationstoadaptandpersistinthefaceofenvironmentalchanges(Sgrò,Lowe,&Hoffmann,2011). Genetic diversity underpins adaptive potential, but is lost insmallpopulationsthroughgeneticdrift,whichcan leadto lossoffit-ness,accumulationofgeneticloadandultimatelypopulationextinction(Harrisson,Pavlova,Telonis-Scott,&Sunnucks,2014;Reed&Frankham,2003;Willi,Van Buskirk, & Hoffmann, 2006). Through reduction orcessationofgeneflow,habitatfragmentationreducesspeciesrangestosmallpopulationsathighextinctionrisk,contributedtobyenviron-mental,demographicandgenetic factors (e.g. inbreedingdepression,genetic loadand inability torapidlyadapt toenvironmentalchanges)andtheirinteractions(Bensonetal.,2016;Fountain,Nieminen,Sirén,Wong,&Hanski,2016;Frankham,2005;Lopez,Rousset,Shaw,Shaw,&Ronce,2009).Giventhattaxaofconservationconcernhavetypicallyundergone habitat fragmentation and/or strong population contrac-tions,geneticissuesshouldbeaddressedinrecoveryormanagementplans,althoughthisisoftennotdone(Piersonetal.,2015,2016).

Effectiveandefficientallocationofconservationresourcesisbestassisted by information about the distribution of genetic variationacrosslandscapes(Hoffmannetal.,2015).Spatialpatternsofgeneticdiversity and genetic differentiation reflect stochastic and environ-mentaleffectsonkeydemographicandevolutionaryprocesses (e.g.populationsize,geneflow,adaptivepotential)linkedtoviability.Thus,geneticdatacaninformabouttheconservationstatusofpopulations,improve predictions about population responses to environmentalchangeandmanagementinterventions,andsupportthedevelopmentofeffectiveconservationstrategiesforenhancingpopulationviability(Willi&Hoffmann,2009).

Foragivenpopulationorspecies,adirectdemonstrationofgeneticproblems,suchaslossoffitnessduetodeclineingeneticdiversityorinbreeding,presentsastrongcaseforgeneticmanagement.Incontrast,

unavailabilityofevidenceforgeneticproblemsinaspecificcaseoftenpromptsanargumentagainstgeneticinterventionuntilsuchevidenceiscollected.Thisposition ignoresevidenceacquiredfromnumerouswildandcaptivesystemsthatsmallandisolatedpopulationswillmostlikelybe facinggeneticproblemswithnegativedemographicconse-quences (Keller &Waller, 2002; Saccheri etal., 1998;Woodworth,Montgomery, Briscoe, & Frankham, 2002), reflecting cultural ratherthanevidence-baseddecisionsconcerninggeneticmanagement(LoveStowell,Pinzone,&Martin,2017).Obtainingdataonpopulationfit-ness is challenging and expensive and may take years. Postponingdecision-makinguntilsuchdataareavailable,ordisregardinggeneticproblemscompletely,islikelytoresultinmanagingsmallpopulationsto extinction (Frankham etal., in press; Love Stowell etal., 2017;Weeks, Stoklosa,&Hoffmann, 2016). Instead, the augmentation ofgene flow isapowerfulconservationmanagementoption forcoun-teractinglossofgeneticdiversityresultingfromdriftinsmallpopula-tions,withthepotentialtopromotepositivedemographicoutcomes(Frankham,2015;Hufbaueretal.,2015;Whiteley,Fitzpatrick,Funk,&Tallmon,2015).Geneticaugmentationtoalleviatedetrimentaleffectsof inbreeding and/or genetic load in small, isolatedpopulations (i.e.geneticrescue)and/orincreaselevelsofgeneticdiversityandadaptivepotential(i.e.geneticrestoration)canbeachievedthroughreconnec-tionofhabitatorhuman-assistedtranslocations (Attardetal.,2016;Weeksetal.,2011;Whiteleyetal.,2015).Strongprecedentssupport-ing the overwhelmingly positive and long-lasting effects of geneticrescueonfitnessexistacrossabroadrangeoftaxa(Frankham,2015,2016).Bybringing innewgeneticdiversity,geneticaugmentation isexpectedtoimproveadaptivepotentialandprobabilityofpersistenceofsmallandisolatedpopulations,eveniftheydonotyetsufferfrominbreedingdepressionandgeneticload(Willietal.,2006).

Anothercommonargumentagainstgeneticaugmentationisthatgene flowbetweendivergent populationsmay “swamp” local adap-tation and distinctiveness, or contribute to outbreeding depression(Frankham etal., 2011; Le Cam, Perrier, Besnard, Bernatchez, &

withinafewdecades,butregularsmall-scaletranslocationswillrapidlyrescuepopula-tionsfrominbreedingdepressionandincreaseadaptivepotentialthroughgeneticres-toration. Despite the lack of data on fitness loss, based on our genetic data forMacquarieperchpopulations,simulationsandempiricalresultsfromothersystems,werecommend regular and frequent translocations among remnant populationswithincatchments.Thesetranslocationswillemulatetheeffectofhistoricalgeneflowandimprove population persistence through decrease in demographic and genetic sto-chasticity.Increasingpopulationgeneticconnectivitywithineachcatchmentwillhelptomaintainlargeeffectivepopulationsizesandmaximizespeciesadaptivepotential.Theapproachproposedherecouldbe readilyapplicable togeneticmanagementofotherthreatenedspeciestoimprovetheiradaptivepotential.

K E Y W O R D S

adaptivepotential,effectivepopulationsize,geneticrescue,geneticrestoration,inbreedingdepression,MacquarieperchMacquaria australasica,management,populationpersistence

     |  3PAVLAP et PVl.

Evanno, 2015; Love Stowell etal., 2017). Such concerns frequentlyleadtorecommendationstomaintainisolationofdifferentiatedpop-ulations,butwithoutappropriateassessmentoftherisksandbenefitsofgeneflow(Farrington,Lintermans,&Ebner,2014;Frankhametal.,2011;Roberts,Baker,&Perrin,2011).Geneticdifferentiationperseisnotaclearindicatorofriskthatpopulationsaredifferentlyadaptedand gene flowmight be harmful: some differentiation is inevitableunderrestricteddispersal,particularlyamongsmallpopulations,andisoftencausedbyrecenthumanimpacts(Coleetal.,2016;Coleman,Weeks, & Hoffmann, 2013; Faulks, Gilligan, & Beheregaray, 2011).Evenwheregeneticdifferentiationisassociatedwithadaptivediver-gence, locally adapted traits can still bemaintained in a populationin thepresenceof gene flow fromadifferently adaptedpopulation(Fitzpatrick,Gerberich,Kronenberger,Angeloni,&Funk,2015),unlessthelevelofgeneflowissolargethatitoverwhelmsnaturalselection(Lowe&Allendorf,2010).Thepotentialbenefitsofgeneflowarelarge,and any risks of outbreeding depression must be weighed againstthe risk of extinction due to inaction, for example from inbreedingdepression,geneticloadorinabilitytoadapttoanovelselectivepres-sure (Becker,Tweedie,Gilligan,Asmus,&Whittington,2013;Fisher,Garner,&Walker,2009;Frankhametal.,inpress;Harrisson,Pavlova,etal.,2016;Lintermans,2013).

Indevelopingplansforgeneticrescueand/orrestoration,fourkeyquestionsmustbeconsidered:

1. What level of genetic variation should trigger genetic rescue/restoration?

2. Whichpopulationscanbeusedassourcesfortranslocations?3. Whatistheriskofoutbreedingdepression?and4. Howmuchgeneflowisrequiredtorescuethepopulationsatriskandrestoretheiradaptivepotential?

Recentreviewsprovideguidelines:

1. Effective population size (Ne) ≥100 is required to prevent in-breeding depression in the short-term and limit loss of fitnessto ≤10% over 5 generations. A much larger global effectivepopulation size, Ne≥1,000, is required to retain adaptive po-tential, bymaintaining an equilibrium between the accumulationof genetic variation for a selectively neutral trait by mutation,and the loss variation by random genetic drift (Frankham,Bradshaw, & Brook, 2014). Based on a combination of theoryand empirical evidence, the numbers from previously proposed50/500 rule (Franklin, 1980) were found to be insufficient toprevent inbreeding depression/avoid loss of genetic variationfor fitness (Frankham etal., 2014).

2. Anypopulationonlyrecentlydivergedfromageneticallydepletedonecanbeusedasasource,butthemagnitudeofgeneticrescueisgreater if the sourcepopulation is outbred and/or differentiated(Frankham,2015).

3. Theriskofoutbreedingdepressionwillbelowifpopulationshavethesamekaryotype,wereisolatedfor<500yearsandareadaptedtosimilarenvironments(Frankhametal.,2011).

4. Iftheriskofoutbreedingdepressionislow,thenupto20%geneflowfromthesourcepopulationwillimprovepopulationadaptivepotential; ongoing monitoring within an adaptive managementframework(i.e.iterativeprocessofdecision-makingbasedonmon-itoringofoutcomes) is recommendedtoavoidgeneticswampingand outbreeding depression (Hedrick, 1995;Weeks etal., 2011;Whiteleyetal.,2015).

TheMacquarieperchMacquaria australasicaCuvier1830isanen-dangeredAustralianendemicfreshwaterfishspeciesthathasundergonesubstantialpopulationdeclineandrangecontractionandfragmentation.PriortoEuropeancolonizationinthelate18thcentury,theMacquarieperchwaswidespread, occurring in the inlandMurray–Darling Basin(MDB),andacrossthreecoastaldrainagebasins,Hawkesbury–Nepean(HNB),ShoalhavenandGeorges(Figure1).AlthoughdistincttaxawithinMacquaria australasica are not formally recognized, morphological, al-lozymeandDNAdataall indicatethattheinlandandcoastalformsofMacquarie perch diverged in the Pleistocene and probably representdifferentspecies (Dufty,1986;Faulks,Gilligan,&Beheregaray,2010a,2015;Faulksetal.,2011;Knight&Bruce,2010;Lintermans&Ebner,2010; Pavlova etal., 2017). Over the last two centuries, Macquarieperchsufferedstrongdeclines,particularlyinthelowlandreachesofriv-ers,andasaresultthespeciesisnowrestrictedtoisolatedlocationsintheheadwatersofLachlan,MurrumbidgeeandMurraytributariesintheMDB,HawkesburyandNepeanriversintheHNB,andGeorgesRiverintheGeorgesBasin(Figure1)(Gilligan,McGarry,&Carter,2010;Knight&Bruce,2010;Lintermans,2007).Ahighlydifferentiated form in theShoalhavenBasinisthoughttohavebecomeextinctby1998,whichun-derscorestheurgentneedtounderstandandmanageremaininggenomicdiversitywithinthespecies (Faulksetal.,2010a).Manyanthropogenicfactorshavebeenimplicatedinthespecies’declineacrossitsrange,in-cludinglossof95%ofhabitat,fragmentationbybarrierstomovement,increasedriversedimentation,alienfishspecies,overexploitationandal-teredflowregimesreducingpreviouslyconnectedhabitatstosmallfrag-ments(Cadwallader,1978;Ingram,Douglas,&Lintermans,2000;Ingrametal.,1990;Pollard,Ingram,Harris,&Reynolds,1990).Lossofphysicalandgeneticconnectivityacrossthespecies’rangeisexpectedtohavecontributedtoitsongoingdecline,althoughestablishingthatdirectlyischallenginggiventhelackofhistoricalgeneticsamples.

Inattemptstoreverseitsdecline,theMacquarieperchhasbeensubjecttovariousmanagementactionsincludinghabitatrestoration,threatamelioration,wild-to-wildtranslocationsandcaptivebreedingwithsubsequentstockingofcaptive-bredfishtore-establishextinctpopulations, augment threatened ones and support recreationalfisheries (Ho& Ingram, 2012; Lintermans, 2012; Lintermans, Lyon,Hammer,Ellis,&Ebner,2015).Despite substantial financial invest-ment and effort, these management actions have so far failed toreversethedeclineofthespecies.Indeed,waterscarcityunderrecentdroughtshasmadeseveraloftheremainingpopulationsexceedinglysmallandvulnerabletoenvironmental,demographicandgeneticsto-chasticity.Historically,vulnerableorextinctpopulationswouldhavebeenrecolonizedfromneighbouringpopulationswhenflowreturned,butthisisnolongerpossibleduetoinsufficientphysicalconnectivity

4  |     PAVLAP et PVl.

     |  5PAVLAP et PVl.

between remnant habitat patches. As others have stated (LoveStowell etal., 2017;Whiteleyetal., 2015),weargue that althoughthemainreasonsforpopulationdecline(e.g.habitatavailabilityanddegradation) should be addressed by conservation management,boostinggeneticdiversitythroughassistedgeneflowmightbenec-essary to improvepopulation fitnessand/oradaptivepotential anddecrease extinction probability of small populations. Managementdecisionsneedtobemadenowbasedonthebestavailableinforma-tion.Whendirectinformationongeneticproblemsinaspecificcaseisunavailable,drawingonaccumulatedevidencefromempirical,the-oreticalandsimulationstudiesofothersystemsisamoreevidence-basedapproachthanmakingtheunvalidatedassumptionthattherearenogeneticproblems(Frankhametal.,inpress;LoveStowelletal.,2017;Weeksetal.,2016).

Here,weexplorepatternsofgeneticdiversityandgeneticdiffer-entiationacrosstheentirerangeoftheMacquarieperchandexpandonearliergeneticworkthatrevealeddeepdivisionswithinthenamedspecies,hybridizationamonginlandandcoastalforms,strongpopula-tionsubstructureandenvironmentalfactorsaffectingthedistributionofpopulationgeneticdiversity (details inAppendixS1;Faulksetal.,2010a,2011;Pavlovaetal.,2017).Forthefirsttime,weestimateeffec-tivepopulationsizesacrossthespeciesrange,evaluatethemagainstestablished thresholds (Frankham etal., 2014) and demonstrate bysimulationsthatwithinafewdecades,smallerMacquarieperchpop-ulationsarelikelytosuffererosionofgeneticdiversityandinbreedingunlessgeneticdiversityisrestoredandmaintainedbytranslocations.Inaddition,weidentifynovelenvironmentalvariablesassociatedwithindividual-basedheterozygosityandimproveearlierestimatesofpop-ulation history, key divergence times, population genetic diversityanddifferentiation,andpopulationstructureatbasin-andcatchmentscales by analysing additional samples and populations, additionalmicrosatellitelociandlongermitochondrialsequences.Ourresultswillcontributetodevelopingseparatemanagementplansforappropriateevolutionaryunits,theselectionofbroodstocksourcepopulationsforthecurrenthatcherybreedingprogrammesandthe identificationofcriticallocationsforstockingandtranslocations.Theprocessesofhab-itatloss,degradationandfragmentationthathaveaffectedtheprevi-ouslycommonandwidespreadMacquariepercharesimilartothoseaffectingmanyspeciesincluding,butnotlimitedto,otherthreatenedMDB species that have experienced recent population extinctions,forexample troutcodMaccullochella macquariensis, southernpygmyperch Nannoperca australis and southern-purple spotted gudgeonMogurnda adspersa(Hammeretal.,2015;Lintermans,2007;Trueman,

2011);hence,theworkflowherecouldbereadilyapplicabletogeneticmanagement of other threatened species to improve their adaptivepotential.

2  | MATERIALS AND METHODS

2.1 | Population sampling, tissue sampling and DNA extraction

Macquarie perch samples were collected predominantly between2002and2014from20populations(Figure1,AppendixS2),includ-ing tenMDB, sevenHNB (note that the Cataract River populationconsistsofendemicHNBaswellasintroducedMDBfish),twocoastalpopulationsthathaveoriginatedfromtranslocationofMDBindividu-als(CataractDaminHNBandtheYarraRiverintheYarraRiverBasin)and a ShoalhavenBasin population (a single known genetic samplefromKangarooRiver). Individualswerecapturedusingvariouselec-trofishing, netting or angling techniques.Most fishweremeasuredforlengthandweightandimmediatelyreleasedatthesiteofcaptureaftercollectionoffinclips.Fintissuewaspreservedin100%ethanolandstoredat−20°C.DNAwasextractedusingaQiPg nDNeasytis-sueextractionkit,unlessavailablefromotherprojects (Faulksetal.,2010a,2011;Nguyen,Ingram,Lyon,Guthridge,&Kearns,2012).

2.2 | Mitochondrial control region sequencing

Sequencesofan844-bpmitochondrialDNAfragment includingthecomplete control regionwereanalysed for339 individuals from17populations (Appendix S2 Table S2A). For five samples, sequenceswere extracted from complete mitogenomes (GenBank accessionsKR152235, KR152240, KR152241, KR152248 and KR152253)(Pavlovaetal.,2017).Fortheremainingsamples,mtDNAfragmentswereamplifiedandsequencedusingprimersA(Lee,Conroy,Howell,& Kocher, 1995) and Dmod (GCCCATCTTAACATCTTCAGTG)(AppendixS3).Chromatogramswereeditedandaligned ing n iLus Pro6.1.3(Drummond,Ashton,etal.,2012).SequenceswereuploadedtoGenBank(accessionKT626048–KT626381).

2.3 | Microsatellite genotyping

Individualsweregenotypedfor19microsatellite loci (AppendixS3).Twentyindividualswithscoresmissingformorethanthreelociwereexcluded, as well as a single Dartmouth individual that returned a

F IGURE  1 GeographicdistributionofMacquarieperchsamplesanalysedforthisstudy(top),mitochondrialcontrolregionhaplotypenetworkshowingdistributionofhaplotypesacrosspopulations(middle;Shoalhavenhaplotypenotincluded)andindividualmembershipsintwoor12geneticclustersinferredfrommicrosatellitedata(majorstructureinferredbyK = 2 and K = 12 setrucetur analysisofallHNBandMDBsamples;seeAppendixS9;Shoalhavenindividualnotincluded).Twocolourstripesabovesetrucetur plotsarecolourcodesforbasins(top)andpopulations(bottom),asonmap.YarraandCataractDampopulationsaretranslocatedfromtheMDB;CataractRivercompriseshybridsbetweenendemicHNBlineageandfishdispersedfromCataractDam(detailsinAppendixS9).Onthehaplotypenetwork,populationsarecolouredasonthemap(populationlabelsarealsogivenforeachcolour);eachcircleisauniquehaplotype(numberwithin-haplotypeID,smallblackcircle–missinghaplotype);thesizeofeachcircleisproportionaltothehaplotypefrequency;connectionsbetweencirclesaresinglesubstitutionsunlessmarkedwithanumber(ofsubstitutions).Infivelocations,asinglehaplotypeisfixed(1inAbercrombie,10inCotter,41inMurrumbidgee,32inGlenbrook,42inWheeny)

6  |     PAVLAP et PVl.

triploidgenotype(confirmedbyrepeatedgenotypingofre-extractedDNA).Finalanalyses includedgenotypingscoresfor871individualsfrom20populations(AppendixS2;SupportingInformation);only0.3%ofscoresweremissing.TestsfordeviationfromHardy–Weinbergandlinkageequilibriawereperformedusingg n L4.2(Rousset,2008)(AppendixS3).

2.4 | Estimating population history and key divergence times

To evaluate the level of mitochondrial haplotype-sharing acrosspopulations, indicativeofpast gene flow, amedian-joiningnetwork(Bandelt,Forster,&Roehl,1999)wasconstructedinn etwLrk v4.6.1.0 (www.fluxus-engineering.com). Times of divergence among majormitochondrial lineages were previously estimated from completemitogenomes (Pavlova etal., 2017; summarized in Appendix S1);however, some lineageswerenot represented.Usingmitochondrialhaplotypes,weestimatedthetimesoflineagedivergencewithB Pset 1.8.2 (Drummond, Suchard, Xie, & Rambaut, 2012), assuming theHKY+G+Isubstitutionmodel,coalescenttreepriorandBayesiansky-line(with10groupsandpiecewise-constantskylinemodel)aspopu-lationsizeprior.Astrictmolecularclockwasused,basedonresultsofcompletemitogenomeanalysis (Pavlovaetal.,2017); thecontrolregionclockrate,estimatedbasedonawideclockratepriorinthatanalysis,wasusedhereasaprior(i.e.lognormalclockratewithmean(inrealspace)=0.05andSD=0.5;thiscorrespondsto95%ofratesbeing sampled between 0.017 and 0.118 substitutions per site permillionyears).Threereplicateswererunfor100millionstepssamplingevery100,000; runswere checked for convergence, combined andanalysedinetrPc r1.6.0(Rambaut&Drummond,2007)afterdiscard-ingfirst10%ofsamplesfromeachrunasburn-in.

2.5 | Genetic differentiation and tests for evolutionary timescale of divergence

PrV Quin3.5 (Excoffier&Lischer,2010)wasused tocalculatepair-wisepopulationFST,RST(microsatellites)andΦST(mtDNAsequences)between catchments. SPAGeDi ver 1.5 (Hardy&Vekemans, 2002)with 10,000permutations ofmicrosatellite allele sizeswas used totestifpopulationshavebeenevolvingindependentlylongenoughtoevolvenewmicrosatelliteallelesthroughmutation;locusAB009wasexcluded,becauseitwasmonomorphicinallbutonepopulation.

2.6 | Assessing population structure at basin- and catchment scales

setrucetur (Pritchard,Stephens,&Donnelly,2000)wasusedtoidentifygeneticclustersfrommicrosatellitegenotypes(dataonpopulationoforiginwerenotapplied).Wefirsttestedforgeneralpopulationstruc-tureacrossthespeciesrangeusingall870MDBandHNBindividu-als.Twenty replicatesof106 burn-in iterations followedby3×106 iterationswere run foreachK from1 to16 (themaximumwassetaccordingtopreliminaryrunsshowingnomeaningfulstructureabove

K=12)usingtheadmixturemodelwithcorrelatedallelefrequencies.Runs were summarized using the web server cVumPk (Kopelman,Mayzel,Jakobsson,Rosenberg,&Mayrose,2015).Thelargestnum-berofgeneticclusters resulting incoherentgeographicgroupswasconsideredtobestrepresentpopulationstructure.

Threesubsetsofthedatawerefurtheranalysedusingsetrucetur : (i)adatasetof671individualsofMDBorigin(10MDBpopulations,CataractDamandYarra)wasrunforKmax=16todetectmajorpopula-tionsubdivisionswithinMDB;(ii)adatasetof375individualsfromtheMurrayandYarracatchments(Dartmouth,Hollands,Sevens,Hughes,KingParrotandYarrapopulations)wasrunforKmax=7toexploretheoriginoftwopreviouslydetectedgeneticclusters inDartmouthandYarra(Nguyenetal.,2012);and(iii)adatasetof134individualsfromsix HNB populations (all except the introduced Cataract River andCataractDamsamples)was run forKmax=7 to test forpresenceofstructurewithintheHNB.

2.7 | Analyses of genetic diversity, proxy for historical effective population sizes

FormtDNAsequencingdata,PrV Quinwasusedtocalculatehaplotypediversity(Hd),nucleotidediversity(π),thenumberofsegregatingsites(S)aswellasTajima’sD (Tajima,1996)andFu’sFs (Fu,1997) testsforselectiveneutralityorpopulationsizechanges.Formicrosatellitedata,PrV Quin was used to calculate expected heterozygosity (He);g nPV x (Peakall and Smouse 2006), the number of private alleles(alleles restricted to a single location); and FsetPet (Goudet, 2001),allelic richness (AR)standardizedto theminimumsamplesizeof14individuals.He andAR reflect coalescent effectivepopulation sizes(Ne)undermutation–driftequilibriumandmaynot reliablyestimatecontemporaryNe associatedwith inbreeding and the probability ofpopulationpersistence(Hareetal.,2011).

2.8 | Analyses of contemporary effective population sizes

We estimatedNe using two single-samplemethods, both ofwhichassume that population size is stable over a few generations. First,we used the approximate Bayesian computationmethod based oneightsummarystatistics(includinglinkagedisequilibrium,LD)imple-mented in Ln sPm (Tallmon, Koyuk, Luikart, & Beaumont, 2008);50,000populationsweresimulatedusingapriorNerangeof4–1,000individuals.Ln sPm was also runwith a priorNe range of 4–500,to confirmestimate consistency. Second, anLD-basedmethodwasused(LDNe;Waples&Do,2008),implementedinn setimPetLrV2.0(Doetal.,2014);randommatingmodelandPCrit=0.02wereapplied.LDNeestimateswereinterpretedonlyforsamplesof≥50individuals,assmallersamplesarenotinformativewhentrueNe>100(Tallmonetal.,2010).Whenestimatedfromasinglecohortofaniteroparousspecies, anLD-basedestimateofNe reflects theharmonicmeanoftheeffectivenumberofbreedersinonereproductivecycle(Nb)andeffective sizeper generation (Ne), butwhenestimated frommixed-age sample, it approximates effective size per generation, albeit

     |  7PAVLAP et PVl.

consistently downwardly biased (50–90% of true Ne, least biasedwhen the number of cohorts in a sample is close to the length ofthegeneration time) (Waples,Antao,&Luikart, 2014).WhereasNb reflectsshort-termeffectivepopulationsizerelevanttoinbreeding,Ne isresponsibleforshapinglong-termevolutionaryprocesses(Waplesetal.,2014); thus,bothwouldbe required forevaluating twopartsof the100/1,000Ne threshold (Frankhametal.,2014).Usingsimu-lations,Waples, Luikart, Faulkner, and Tallmon (2013) showed thatthe Nb/Ne ratio can be approximated using two life-history traits.According to their formulae, for Macquarie perch Nb/Ne = 1.156 (assumingageatmaturityof3yearsandadult lifespanof23years(Appleford,Anderson,&Gooley,1998;Lintermans&Ebner,2010));thus, an LD-based estimate of Ne should reasonably approximatebothNb and Ne, provided that other biases can be accounted for.Macquarieperchsamplesincluding≥7cohortswouldbeleastbiased,assuming a generation time of 7years (calculated inALret x (Lacy,Miller,&Traylor-Holzer,2015),below).Accordingtotherangeofindi-vidualfishlengthsineachpopulation(SupportingInformation)andthemodelled relationshipbetween lengthandage forMacquarieperch(Figure3ofTodd&Lintermans,2015),ourMacquarieperchpopula-tionsamplesincludedfrom~3to>7cohorts,andweexpecttheleastdownwardbiasinLD-basedestimatesofNeforthesevenpopulationswitharangeofsampledlengths>200mm(Abercrombie,Adjungbilly,Dartmouth,Hughes,KingParrot,Lachlan,Sevens,Yarra).ForspecieswithNb/Ne~1,LD-basedestimatesofNb(andNe)areexpectedtobeunaffectedbytwo-lociWahlundeffectarisingwhenparentsfromdif-ferentcohortsarecombinedinasinglesample(Waplesetal.,2014).

2.9 | Identifying environmental variables associated with individual- based genetic diversity

Weusedanindividual-basedBayesianmodellingapproachtoidentifyenvironmental variables that were correlated with spatial distribu-tionofgeneticvariation(Harrisson,Yen,etal.,2016).Themodelwasbasedon the16populations from theMDBand theHNB (translo-catedYarraandCataractDamandadmixedCataractRiver sampleswereexcluded).

Theresponsevariablewashomozygosity-by-locus(HL),calculatedforeach individualusing theRhhpackage (Alho,Valimaki,&Merilä,2010) in R 3.2.0 (R Development Core Team 2014). HL measurestheproportionofhomozygouslociperindividual,attributinggreaterweight to homozygosity at more variable loci (Aparicio, Ortego, &Cordero, 2006). LowHL of an individual indicates low genetic sim-ilarity of its parents, which is expected in populations with highgeneticdiversity.Usinganindividual-basedmeasureofgeneticdiver-sity,ratherthanapopulation-basedone(suchasAR),enabledustoconsider environmental variation within populations (Appendix S4)ratherthanpopulationaverages,whichshouldresolveenvironmental-geneticassociationsatfinerspatialscales.

Eleven environmental variables reflecting flow regime, con-nectivity, habitat and climate, previously shown to be indicators ofMacquarie perch presence and population health (Appendix S4),wereoriginallyconsideredforinclusioninthemodelaspredictorsof

individualgeneticdiversity (Table1).Allbutonewere sourced fromtheNationalEnvironmentalStreamAttributesDatabasev1.1.5(http://www.ga.gov.au/metadata-gateway/metadata/record/gcat_75066)(GeoscienceAustralia2012; Stein, 2006) linkedvia stream segmentnumbertotheGeofabricstreamnetworklayer(BureauofMeterology2012); thedataonmeanNovember temperatureweredownloadedfrom http://www.bom.gov.au. Environmental data were extractedusingArcGIS10.2 (EnvironmentalSystemsResearch Institute1999-2014) foreach sampling site (1–8sitesperpopulation for16MDBandHNBpopulations;SupportingInformation).Environmentalpredic-torvariableswerestandardizedtozeromeanandunitvariance,andPearson’scorrelationsamongvariableswerecalculated(usingasinglerecordper samplingsite).Only sevenvariables thatwerenothighlycorrelated(Pearson’s|r|<0.7)wereincludedinthefinalmodel(Table1;AppendicesS4andS5).

We used a hierarchical Bayesian regression model to estimatetherelationshipsbetweenHLandenvironmentalvariables(AppendixS5).Weuseda reversible-jumpMarkovchainMonteCarlo (MCMC)algorithm to performmodel selection. Basin and sitewere includedasclusteringvariables (givenexchangeablepriors,equivalent to ran-domeffectsinastandardmixedmodel)toaccountforspatialcluster-ingofthesamples.ThemodelwasfittedusingWinBUGS1.4 (Lunn,Best,&Whittaker,2009;Lunn,Thomas,Best,&Spiegelhalter,2000;Lunn,Whittaker,&Best,2006).OutputsweremanagedinR3.1.2(RDevelopmentCoreTeam2014).Allparameterswereassignedvagueprior distributions. The primarymodel outputs are estimates of theprobabilityofinclusionforeachenvironmentalvariableandthefittedassociationsbetweenHLandeachenvironmentalvariable.Thepriorprobability ofvariable inclusionwas0.5, so that posterior probabili-tiesofinclusion>0.5provideevidenceinfavourofvariableinclusion.Posteriorprobabilitiesofinclusion>0.75providestrongevidenceforvariable inclusion (oddsratio>3).FullmodeldetailsandmodelcodeareinAppendixS5.

Fivefold cross-validation was used to estimate the predictivecapacityofthefittedmodelandtotestwhetherthemodelwaslikelytobeidentifyingtruerelationships.ToaccountforpossiblecorrelationofHLamongindividualsfromthesamegeneticclustersduetosharedhistory,cross-validationtestdatasetscomprisedninepopulationclus-ters,withapproximately120individualsineachtestdataset(1–5pop-ulationclusters;AppendixS5).Weusedmicrosatelliteclusters,ratherthanphylogenetic groups, because recent history (e.g. drift in smallpopulations)couldhaveastrongeffectonHL,whichcandifferacrossdistinctgeneticclusters.Modelpredictionsforcross-validationwerebasedonenvironmentalvariablesonly;clusteringvariables(basinandsite)wereusedinmodel-fittingbutnottomakepredictions.

2.10 | Simulating genetic rescue/genetic restoration of isolated Macquarie perch populations

Todemonstrategeneticrescueand/orgeneticrestorationeffectsoftranslocationstosmallpopulationsofMacquarieperch,wesimulatedtheoutcomesoftwomanagementscenarios(do-nothingvs.50yearsoftranslocation)fortwopairsofpopulationsusinganage-structured

8  |     PAVLAP et PVl.

populationmodelinALret x10.1.6(Lacy&Pollak,2014).ALret xsimu-lateseffectsofdeterministicforces,andthoseofdemographic,envi-ronmentalandgeneticstochasticityonpopulationparameters,suchas probability of survival, time to extinction, population size, allelicdiversity, heterozygosity and inbreeding (approximated as homozy-gosity atmodelled loci).Age-specificmortalitywasmodelledbasedonToddandLintermans (2015) (details inAppendixS6). Inbreedingdepressionwasmodelled by assuming that each founder individualhasuniquerecessivealleles(6.29lethalequivalentsperindividual,theempiricallyderivedALret xdefaultvalue).Fiftypercentofinbreedingdepressionwasassumedtobeduetorecessivelethalalleles:inbredoffspring with two copies of the same recessive alleles die beforereproduction.Thetranslocationscenarioassumedthatsixindividuals(3ofeachsex)weremovedfromlargertosmallerpopulationseachyear for50years to approximate the impactsof repeated,modest-sizedtranslocations;designingthemostcost-effectivetranslocationprogrammeisbeyondthescopeofthispaper.Alltranslocatedadultswere assumed to survive translocation. Genetic results were sum-marizedfor1mitochondrialand19microsatellitemarkerssimulatedbased on observed allele frequencies. The first pair of populationshad initial population sizesof3,000and500 (modelledusing allelefrequenciesofDartmouthandKingParrot,respectively),andthesec-ondpairhadinitialpopulationsizesof300and100(modelledusingCataractDamandMurrumbidgee).Fivehundredforward-in-timesim-ulationswere run for100years foreachscenariousingparameters

outlinedinAppendixS6.TocompensatefortheinabilityofALret x to modellargefemalefecundities,calculationswererestrictedtoadults(Lacyetal.,2015).

3  | RESULTS

3.1 | Population history, divergence time estimates and genetic differentiation

Themitochondrialhaplotypenetworkwasconsistentwithhistoricaliso-lationbetweentheMDBandtheHNB,historicalisolationoftheLachlancatchmentfromtherestoftheMDB,isolationofAdjungbilly—theonlyknownremnantMacquarieperchpopulationinthelowerMurrumbidgeecatchment—fromotherpopulations,andisolationamongHNBpopula-tions(Figure1).Lackofhaplotype-sharingamonganyHNBpopulations(AppendixS7)suggeststhatstronggeneticdriftoperatesinthisbasin,consistentwithearlierfindingsfromamitogenomestudy(Pavlovaetal.,2017). In contrast, haplotype-sharing acrossMurray River tributaries(Dartmouth,Hollands,Sevens,Hughes,KingParrot) indicatedhistori-calconnectivitywithinthesouthernMDB,consistentwithacontinuousreconstructed historical distribution in the southernMDB (Trueman,2011).Similarly,haplotype-sharingbetweenLachlanandAbercrombie(northernMDB)supportedpreviously inferredcontemporaryconnec-tivitybetweenthesetwopopulations(Faulksetal.,2011).Thetranslo-catedfishfromtheMDBpopulationintoCataractDamapparentlybore

TABLE  1 Detailsoftheenvironmentalmodelpredictinghomozygosity-by-locus(HL,aninverseofgeneticdiversity):finalsetofsevenvariables(firstcolumn)andvariableshighlycorrelatedtothem(secondcolumn),predictedrelationshipswithHL,probabilityofinclusioninthemodel(value>0.7indicateastrongassociationwithHL,showninbold)anddirectionoftheeffect(positivemeansthatavariableincreasesHL;negativemeansthatavariabledecreasesHL).EnvironmentalvariablesweresourcedfromNationalEnvironmentalStreamAttributesDatabase.Environsarevalleybottomsassociatedwiththestream

Environmental predictors Correlated variables Predicted relationships with HLProbability of inclusion

Direction of effect

Flowregimedisturbanceindexcalculatedforperiod1970–2000

Annualmeanaccumulatedsoilwatersurplus,Streamandenvironsaveragehottestmonthmaximumtemperature,Coefficientofvariationofmonthlytotalsofaccumulatedsoilwatersurplus

Highergeneticdiversity(lowerHL)withinlargerandmorepermanentstreamsand/ormorevarianceingeneticdiversityinsmallrivers

0.28 Neutral

Barrier-freeflow-pathlength Highergeneticdiversity(lowerHL)inlargerriverfragments

1 Positive

Maximumbarrier-freeflow-pathlengthupstream

Annualmeanaccumulatedsoilwatersurplus

Increaseingeneticdiversity(decreaseinHL)withdistancetoupstreamdam

0.89 Negative

Streamsegmentslope Highergeneticdiversity(lowerHL)instreamswithhigherslope

0.49 Neutral

Streamandvalleypercentageextantwoodlandandforestcover

Highergeneticdiversity(lowerHL)instreamswithmorevegetatedbanks

0.19 Positive

Streamandenvironsaveragecoldestmonthminimumtemperature

Meansegmentelevation Lowergeneticdiversity(higherHL)inwarmerstreams

0.73 Positive

MeanNovembertemperature Highergeneticdiversity(lowerHL)instreamswithcoolertempera-turesduringstartofthebreedingseason

0.47 Positive

     |  9PAVLAP et PVl.

twohaplotypes(thatwerethemostcommoninsouthernMDB),bothofwhichappeartohaveescapedintotheCataractRiver,wherethreeendemicHNBhaplotypesalsooccur.ThetranslocatedYarrapopulationsharedhaplotypeswithmostpopulationsfromthesouthernMDBandalsohadseveralhaplotypesnotsampledelsewhere.

Phylogenetic analysis inB Pset (Appendix S8) showed estimatesof lineage divergence times consistent with those from completemitogenomes (Pavlova etal., 2017): time to themost recent com-monancestor(TMRCA)ofShoalhavenandHNB+MDBlineageswasestimated at 1.061 (95% HPD 0.204–2.840) million years ago, ofMDBandHNB lineages (which formedcladeswithposteriorprob-abilities(PP)of0.85and1,respectively)at264(59–674)KY(thou-sandyears)ago,ofallHNBhaplotypesat90(16–240)KYandofallMDBhaplotypes at177 (45–464)KY.TheestimatesofTMRCAofLachlan+Abercrombieclade (PP=1)at27 (2–88)KYbring forwardthepreviousestimateofisolationoftheLachlancatchmentfromtherestoftheMDBbasedonpartialcontrolregionsequence(310KY;Faulksetal.,2010a).HistoricalisolationofthelowerMurrumbidgeecatchment(Adjungbilly)fromtheupperMurrumbidgeefor36(8–100)KY(TMRCAofAdjungbillyclade;PP=0.83)isanovelresultforapre-viouslyunstudiedpopulation.

StrongpopulationdifferentiationamongmostHNBpopulations,allthreeMDBcatchmentsandthreepopulationsoftheMurrumbidgeecatchment(Adjungbilly,CotterandMurrumbidgee)wasevidentfromgenerallyhighpairwisepopulationΦST(mtDNA)andFST(microsatel-lites)values(detailsinAppendicesS9andS10),consistentwithFaulks,Gilligan,&Beheregaray(2010b),Faulksetal.,(2010a,2011).Anevolu-tionarytimescaleofisolationwithintheHNBwassupportedbystrongmitochondrialdivergenceamongHNBpopulationsandbySPAGeDianalysis, which showed that evolution of microsatellite allele sizeshas contributed to divergence among northern HNB (Wollemi andWheeny)versussouthernHNBpopulations(rest)andwithinnorthernHNB(aswellasbetweentheHNBandMDB),butnotwithintheMDB(AppendixS10).Large,significant(p < 0.001)microsatelliteFSTvaluesacrossalmostallpairwisecomparisonssupportedcontemporaryiso-lationofpopulationsandstrongdrifteffects,whichisanovelresult.

3.2 | Assessing population genetic structure at basin- and catchment scales

For all 19MDB andHNB populations (N = 870),setrucetur analysisassumingK=2 (Figure1)wasconsistentwiththeHNB-MDBdiver-genceandhybridizationoftranslocatedMDBfishwithendemicHNBfish in Cataract River (Faulks etal., 2011). Our analysis of a largeCataractRiversampleshowedthatindividualmemberships(Q)intheendemicHNBclusterrangedfrom0.68to0.001andwerenotassoci-atedwithmtDNA lineage, suggesting thatgene flow fromCataractDammightbeongoingandthathybridizationinCataractRiverisnotlimitedtothefirstgenerationandisnotsex-biased(AppendixS11).All other setrucetur analyses supported predominantly geographicstructureandatleastpartialcontemporarygeneticisolationofpopu-lations within basins. Analysis of all individuals (N = 870) assumingK=12yieldedgeographicallymeaningfulgeneticclusters (Figure1),

supported by separate analyses of 12 populations of MDB ori-gin (N = 671) and 6 populations from the southernMDB (N = 375).AnalysisofsixHNBpopulations(allexceptCataractRiverandCataractDam; N = 134)assignedallHNBpopulationstodifferentgeneticclus-ters,exceptasingleindividualintheLittleRiverhadQ>90%intheKowmungcluster,suggestingararelong-distancedispersaleventorundocumented translocation (detailsofallanalysesare inAppendixS11).setrucetur analysesdidnotsupportthepreviouslyreportedtwoclusterswithinDartmouthandYarra(Nguyenetal.,2012).

3.3 | Analyses of genetic diversity and effective population sizes

Mitochondrial (Hdandπ) andnuclear (ARandHe) geneticdiversityrangedwidelyacrosspopulations inbothbasins (Figure2;AppendixS2). Generally, populations with low nuclear genetic diversity(Wollemi,Whenny,GlenbrookintheHNBandAdjungbilly,CotterandMurrumbidgee intheMDB)alsohad lowmtDNAdiversity (Wollemiwasnotsequencedhere,buthadlowHdandπinthestudyofFaulksetal.,2010a),butmtDNAdiversitywasalsolowforsomeotherpopu-lations (Abercrombie, Lachlan). Tajima’sDwas significantly negativeforGlenbrookandAdjungbilly,suggestingpurifyingselectionand/orpastpopulationgrowth.Incomparison,bothtranslocatedpopulations(CataractDamandYarra)hadrelativelyhighlevelsofgeneticdiversity(Figure2).TheShoalhavenindividualwashomozygousat17of19loci.

Linkagedisequilibrium-basedestimatesofNe(LDNe)forfivepop-ulationswithN > 50weresmallestinCotter(meanNe=63,confidenceinterval35–134),slightlyhigherinCataractRiver(164,103–349)andCataractDam(128,68–501)andmoderateinDartmouth(307,201–593)andYarra (344,228–642;AppendixS2).Ln sPm Neestimatesforthesamefivepopulationswere~2.8×(range1.5–4.3)lowerthanLDNeestimates.Ln sPm Neestimateswerelow(meanNe<100)forallpopulationsexceptYarra(meanNe=132;Figure2),consistentlysoforbothsetsofpriors(AppendixS2).

3.4 | Estimated relationships between individual genetic diversity and environmental variables

Approximately66%ofthevariationinHLcouldbeexplainedbyenvi-ronmental and clustering variables (model r2=0.66). Environmentalvariables could predict only c. 17% of the variation in HL (cross-validated r2=0.17),suggestingthatstochasticprocesses(e.g.geneticdriftinsmallpopulations)haveastrongeffectongeneticdiversityofMacquarieperchand/orthatoneormorekeyvariableimpactingHLwasnot included in themodel (e.g.EHNvirus;Beckeretal.,2013).Threeofthesevenenvironmentalvariableshadprobabilitiesofinclu-sioninthefittedmodel>0.7,indicatingastrongassociationwithHL(oddsratios>2.3,Table1,AppendixS5).Maximumbarrier-freeflow-pathlengthupstreamhadanegativeassociationwithHL(i.e.distancefromtheupstreamdamwasassociatedwithhigherindividualgeneticdiversity),whereasbarrier-freeflow-pathlength,andminimumtem-perature of the coldest month, had positive associations with HL(Table1,Figure3).

10  |     PAVLAP et PVl.

3.5 | Simulating genetic rescue/genetic restoration of small isolated Macquarie perch populations

Fordo-nothingscenariosrunover100years,modelsforallfourpop-ulationsshowedoveralldecrease insurvivalprobability,decrease ingeneticdiversity and increase in inbreeding and inbreedingdepres-sionovertime(AppendixS6).Withdecreasinginitialpopulationsizes(3,000,500,300and100forDartmouth,KingParrot,CataractDamand Murrumbidgee, respectively) extinction probability increased,timetothefirstextinctiondecreased,inbreedingdepressionincreased(i.e.thenumberoflethalallelesperindividualatyear100decreased)and proportion of retained heterozygosity for extant populationsdecreased(Heyear100/Heyear0:0.98,0.89,0.85and0.79;Table2).FinalNe–valuesestimatedbasedonallelefrequencychangesresultedinamean Ne/Nratioof0.17(range0.1–0.26;Table2).

Resultsof50-years-of-translocationscenariosshowedthataddi-tionof individuals intoKingParrotandMurrumbidgee,comparedtodo-nothingscenarios,decreasedpopulationprobabilityofextinction35- and26-fold, respectively, increased time to first extinction anddecreasedinbreedingdepression1.1-and1.5-fold,andforpopulationsextantatyear100,increasedpopulationsizes(2-and3.8-fold),het-erozygosity(1.2-and2.2-fold)andnumberofalleles(2.4-and2.5-fold;Table2,Figure4).Thegeneticrestorationeffectwasrapid:inthefirst

10yearsoftranslocations,thenumberofallelesinKingParrotreached90%,andinMurrumbidgee>100%ofthoseofthesourcepopulations(AppendixS6).Geneticrescuefrominbreedingdepression(i.e.largernumber of lethal alleles per individual under 50-years-of-transloca-tion compared to the do-nothing scenario; Figure2)was apparentforthesmallMurrumbidgeepopulationatyear10andforthelargerKingParrotatyear40.Geneticrescue/restorationeffectswerelong-lasting:geneticdiversityremainedhigherandinbreedinglowerthanindo-nothingscenarioeven50yearsaftercessationof translocations.Effectofharvest(fortranslocation)onpopulationsize,heterozygosity,numberofallelesorinbreedingofbothsourcepopulationswasmini-mal,butharvestincreasedtheprobabilityofextinctionanddecreasedmeantimetoextinctionforthesmallerCataractDam(andtoalesserextent,thelargerpopulationatDartmouth;Table2,AppendixS6).

4  | DISCUSSION

4.1 | The need for specieswide genetic management to prevent further loss of genetic diversity and boost adaptive potential

Using novel genetic data, we investigated whether geneticaugmentation is likely to be beneficial and necessary for long-term

F IGURE  2 Estimatesofmitochondrialandnucleargeneticdiversityandeffectivepopulationsizes(Ne,Ln sPm:lower–upper95%confidencelimits,barsshowmeanNe)for19populations(colouredasonFigure1).CrossesindicatepopulationsforwhichmtDNAwasnotsequencedhere(estimatesfrompartialcontrolregion(Faulksetal.,2010a)werethefollowing:WollemiHd=0.111,π=0.0006;ErskineHd=0,π=0;KowmungHd=0.708,π=0.0076).InbreedingcoefficientFeforeachpopulationiscalculatedasFe=1−He/HeDartmouth;adjustedNeiscalculatedas4.5×Ln sPm Ne(seeSection4;lowerandupperboundsarenotadjusted)

5

7

9

11

13

15

17

19

Num

ber o

f pol

ymor

phic

loci

Polymorphic loci

00.10.20.30.40.50.60.70.80.9

1H

dHaplotype diversity, mtDNA

X X X

00.0010.0020.0030.0040.0050.0060.0070.0080.009

π

Nucleotide diversity, mtDNA

X X X0

200

400

600

800

1000

0

50

100

150

200

250

Ne

estim

ate

Ne, OneSamp

00.5

11.5

22.5

33.5

44.5

AR

Allelic richness, microsatellites-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.00

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Inbr

eedi

ng C

oeffi

cien

t, F

He

Average gene diversity, microsatellites

Adj

uste

d N

e

     |  11PAVLAP et PVl.

survivalofanendangeredfreshwaterfish,theMacquarieperch.Thespecieshasexperienceddramaticdeclines in rangeandnumbersofmostpopulations,intensehabitatfragmentation,levelsofinbreedingatwhichinbreedingdepressionisoftenobservedinwildpopulationsand continuing loss of populations despite intensive management.Assuminglevelsofinbreedingdepressionsimilartothosereportedforotherwildpopulations,oursimulationssupportadeclineinpopulationviabilityunlessthere isactivegeneticmanagement.Webelievethatthesecombinedlinesofevidenceprovidesufficientsupportforgeneticaugmentationtobeimplementednow,despitealackofdataonfitnessdecline,whichmaytaketoolongtocollect inordertobeusefulforpreventingextinctionsofmanypopulations(Weeksetal.,2016).

SignificantinvestmentinhabitatrestorationandprotectionhasfailedtoleadtodetectablerecoveryofsomeMacquarieperchpop-ulations (MarkTurner,GBCMApers. comm.).Althoughnodataon

fitnesslossexist,itisreasonabletoassumethatinbreedingdepres-sionmighthavecontributedtothisfailure.Oursimulationssuggestthat inbreeding depressionmay contribute to population declinesandimpedepopulationrecoveryinthenearfuture.OurestimatesofNesuggestthatmostpopulationswillhavelimitedcapacitytoadapttorapidenvironmentalchanges.Thus,geneticproblemsshouldbeaddressedinconjunctionwithotherkeythreatstopopulationper-sistence.Restoringthehistorical,within-basingeneticconnectivityofMacquarieperchviaassistedgeneflowwillimprovegeneticdiver-sityandadaptivepotentialacrossthespecies’range.Oursimulationsindicatethatsmall-scaletranslocations (6adults/year)willresult inrapidand long-lasting increase ingeneticdiversity (limitedonlybygenetic diversity and divergence of the source(s)) and decrease ininbreedingandprobabilityofpopulationextinction.Althoughlargesourcepopulationsshouldnotbeimpactedbyappropriatelylimitedharvest (e.g. for translocationorcaptivebreeding),smallersourcesmaybecomemorevulnerabletoextinction,lossofgeneticdiversityandinbreeding.Therefore,wherepossible,managementshouldcon-siderreciprocaltranslocationsandmultiplesources.

Our ALret xmodel has a numberof limitations: (i) offspring thatdonotsurvive to reproductionwerenotmodelled (due tosoftwarelimitation),but,givenextremefecundity(upto110,000eggs/female;Cadwallader & Rogan, 1977), strong purifying selection acting onjuveniles could purge deleterious mutations, reducing inbreedingdepression;(ii)individualsatthestartofsimulationswereassumedtobeunrelated,which, if incorrect,artificially increases the timewheninbreeding depression starts to impactviability; (iii)mutationswerenot modelled, and thus, novel genetic diversitywas not accountedfor;(iv)adaptivecapacitywasnotaddressed;(v)therearelargeuncer-tainties around environmental effects and population parameters,including an assumption that all individuals survive translocation.Nevertheless,oursimulationsareinagreementwithprevioussimula-tionstudies(Allendorf&Ryman,2002;Rieman&Allendorf,2001)thatundermanagementinaction,populationsof100–500adultswilllosegenetic diversity over time (>10%of alleles and>10%of heterozy-gosityin100years),andthelosswillbefasterinsmallerpopulations.InferenceoflowgeneticdiversityinMacquarieperch(meanobservedheterozygosityacrosspopulationsHo=0.371,estimatedfrom19lociwith13.5alleles/locusonaverage)wasconsistentwithsimilarinfer-encesforotherthreatenedAustralianfreshwaterfishes,Yarrapygmyperch Nannoperca obscura (Ho=0.318; 14 loci, 11.9 alleles/locus;Brauer, Unmack, Hammer, Adams, & Beheregaray, 2013), southernpygmyperchN. australis (Ho=0.47;12 loci,10.3alleles/locus;Coleetal.,2016)anddwarfgalaxiasGalaxiella pusilla(Ho=0.395;11loci,16.3 alleles/locus; Coleman etal., 2010), although estimates of Hothemselves might not be strictly comparablewhen estimated frommarker sets of different variability (Hedrick, 1999). Nevertheless,higher diversity reported for the golden perchMacquaria ambigua (Ho=0.52;8 loci,15.3alleles/locus;Faulksetal.,2010b), themorecommonandabundantMacquarieperchcongener,isconsistentwithanearlierfindingofameandecreaseof35%inheterozygosityin170threatenedtaxacomparedtotheirnonthreatenedrelatives(Spielman,Brook,&Frankham,2004).

F IGURE  3 FittedrelationshipsbetweenHLandenvironmentalvariablesforvariableswithprobabilityofinclusiongreaterthan0.7.Thex-axisshowsstandardizedvaluesforeachvariable(e.g.avalueof1means1standarddeviationabovethemeanvalueforthatvariable),andthey-axisshowsdeviationfromthemeanHLvalue(inunitsofHL)foragivenstandardizedvalueofthepredictorvariable.Greyshadingisonestandarddeviationofthefittedeffect.Modelsincludedbasinandsiteasclusteringvariables,sothesefittedeffectsaccountfordifferencesamongsitesorbetweenbasins

−1 0 2

−0.25

−0.20

−0.15

−0.10

−0.05

0.00

1

∆HL

(a)

−1.0 −0.5 0.0 1.0 1.5 2.0

0.0

0.1

0.2

0.3

0.5

∆HL

(b)

−2 −1 0 1

−0.05

0.00

0.05

0.10

0.15

0.20

Minimum temperature of the coldest month

∆HL

(c)

Maximum barrier-free flow-path length upstream

Barrier-free flow-path length

12  |     PAVLAP et PVl.

TABLE 2 SummaryofVortexsimulationsoftwomanagementscenarios.In50-years-of-translocationscenarios,KingParrotandMurrumbidgeepopulationsaresupplementedbyindividuals

fromDartmouthandCataractDam,respectively

Scen

ario

Do-

noth

ing

50-y

ears

-of-

tran

sloca

tion

Do-

noth

ing

50-y

ears

-of-

tran

sloca

tion

Popu

latio

nsD

artm

outh

King

Par

rot

Dar

tmou

thKi

ng P

arro

tCa

tara

ct D

amM

urru

mbi

dgee

Cata

ract

Dam

Mur

rum

bidg

ee

Initialpopulationsize,N

3,000

500

3,000

500

300

100

300

100

Probabilityofextinction

0.00

20.

210.

022

0.00

60.

360.

828

0.77

60.

032

Timetofirstextinction,

years

8582

7692

7863

4290

NeatYear100

340

6125

7−221

4431

42−18

N(extant)atYear100

3,269

337

3,743

680

337

118

478

450

Ne/

N0.

104

0.18

1Notestimated

Notestimated

0.13

10.

26Notestimated

Notestimated

Heatyear0

0.491

0.42

60.491

0.42

60.

472

0.19

0.47

20.19

Heatyear100

0.48

10.379

0.47

80.

442

0.40

10.

151

0.398

0.33

1

Numberofallelesatyear0

5.68

3.52

5.68

3.52

3.36

1.8

3.36

1.8

Numberofallelesatyear

100

5.03

2.92

4.95

4.26

2.6

1.5

2.6

2.86

Numberofhaplotypesat

year

06

46

42

12

1

Numberofhaplotypesat

year

100

5.53

2.53

5.36

6.11

1.73

11.

752.

51

Numberoflethalalleles/

indi

vidu

al a

t yea

r 03.

153.

153.

153.

153.

153.

153.

153.

15

Numberoflethalalleles/

indi

vidu

al a

t yea

r 100

32.

532.94

2.8

2.28

1.83

2.44

2.72

     |  13PAVLAP et PVl.

Although insimulations inbreedingdepressiondidnotappeartoimpactthelargestpopulation(initialN = 3,000,ALret x Ne=257)foratleast50years(~7Macquarieperchgenerations),itstartedtoimpactthreesmallerpopulations(initialN ≤ 500,ALret x Ne<100)withinthefirst10years(i.e.<2generations),supportingtheshort-terminbreed-ing thresholdNe of 100 (Frankham etal., 2014). Inbreeding can beapproximated for a small population by scaling its genetic diversityby the diversity of a known outbred population using the effectiveinbreedingcoefficientFe: Fe=1−Heinbred/Heoutbred,whereHeinbred isheterozygosity(forneutralvariation)ofapopulationinquestionandHeoutbredisheterozygosityofanoutbredpopulation(Frankham,1998).If Dartmouth, one of the largest natural populations of Macquarieperch in theMDB, is used as anoutbred reference, thenFe coeffi-cientsforWollemi,Wheeny,Glenbrook,Adjungbilly,CotterandupperMurrumbidgee are>0.45, and for Little is 0.25 (i.e. is as high asFe for the offspring of full siblings; Figure2). At Fe of 0.2, inbreeding

depressionistypicallyobservedforpopulationsofnaturallyoutcross-ingspecies(Frankham,1995;Szulkin&Sheldon,2007;Wallingetal.,2011;Woodworth etal., 2002). Consistently, our simulations showthatwithinadecade,inbreedingdepressioncanbeexpectedtoactinMacquarieperchpopulationsof≤500adults,unlesstranslocationsareperformed.Althoughtheoreticallyoneeffectivemigrantpergenera-tionmaybesufficienttoprovideinbreedingconnectivity(i.e.signifi-cantlyreduceharmfuleffectsofinbreeding;Mills&Allendorf,1996),aswellasadaptiveconnectivity(i.e.potentialofhighlyadvantageousallelestospreadinarecipientpopulation;Rieseberg&Burke,2001),itisnotexpectedtodriveFebelow0.2(Lowe&Allendorf,2010).Inanycase,wellbeforeFeclimbsto0.2,manypopulationswillalreadybesuf-feringinbreedingdepression.Along-termstudyofinbreedingdepres-sioninthewildusinggenomicestimatesofinbreedingdetectedmajornegative impacts on lifetime reproductive success at F even below 0.1(Huisman,Kruuk,Ellis,Clutton-Brock,&Pemberton,2016).Thus,

F IGURE  4 ResultsofVortexsimulationsforKingParrot(initialpopulationsizeN = 500;blackline)andMurrumbidgee(N = 100;greyline)underdo-nothing(solidline)or50-years-of-translocation(dashedline)scenarios

1.11.62.12.63.13.64.14.65.15.66.1

0 10 20 30 40 50 60 70 80 90 100

Num

ber o

f alle

les

Year

1.7

1.9

2.1

2.3

2.5

2.7

2.9

3.1

3.3

0 10 20 30 40 50 60 70 80 90 100

Num

ber o

f let

hals

Year

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0 10 20 30 40 50 60 70 80 90 100

Het

eroz

ygos

ity

0.5

0.55

0.6

0.65

0.7

0.75

0.8

0.85

0 10 20 30 40 50 60 70 80 90 100

Inbr

eedi

ng c

oeffi

cien

t

00.10.20.30.40.50.60.70.80.9

1

0 10 20 30 40 50 60 70 80 90 100

Pro

babi

lity

of s

urvi

val

0

200

400

600

800

1000

1200

0 10 20 30 40 50 60 70 80 90 100

Abu

ndan

ce

Murrumbidgee do-nothingMurrumbidgee transloca�onsKing Parrot do-nothingKing Parrot transloca�ons

14  |     PAVLAP et PVl.

more thanoneeffectivemigrantpergenerationmaybe required todropFevaluessufficiently(e.g.below0.1)torescuegeneticallydepau-perate populations from strong inbreeding depression (Frankhametal.,inpress;Weeksetal.,2011).

ContemporaryNe related to inbreeding is best estimated usingtwoormoretemporalgeneticsamples(inbreedingNe;Luikart,Ryman,Tallmon,Schwartz,&Allendorf,2010),orusingasamplefromasingle-cohortandadjustingestimatesusinglife-historytraits(Waplesetal.,2014).Nerelevanttoadaptivepotentialisbestestimatedfromtwoormoregeneticsamplestakenseveralgenerationsapartandmeasuringthechangeinallelefrequenciesovertime(varianceNe;Luikartetal.,2010). Our datawere not suitable for temporal analyses: althoughsamplingwasspreadoveryears,temporalsamplesspanninggenera-tionsatindividualsiteswerelacking,butsamplingwasrandom,com-prisingmultiplegenerations,andoverlapped15-yeardrought,duringwhich many populations ceased regular breeding. Because Ne and populationsizearekeyparametersforevaluatingexistingthresholdsand predicting future population trajectories, here we used single-sampleNeestimatescorrectedforbiases,alongwithNe/Nratiosfromsimulations based on the observed allele frequencies, to estimateboundsonpopulationsizesofMacquarieperch.AlthoughwebelievethatourestimatesofNe and Nshouldbesufficientlyaccurateforthepurposesstatedabove,theyareassociatedwithanumberofuntestedorviolatedassumptions(seebelow)andhenceshouldbeinterpretedwithcautionandre-examinedwhengenomicdatabecomeavailable(Hollenbeck,Portnoy,&Gold, 2016). Ifweassume that (i)Ln sPm Neestimateissystematicallydownwardlybiasedandis~2.8×(range1.5–4.3)lowerthanLDNeNe(basedonfivesamplesof>50individu-alswithbothestimates;AppendixS2),(ii)LDNeNeestimateis~1.6×(1.1–2)lowerthanthetrueNe(Waplesetal.,2014)and(iii)Neis~5.9×(3.8–9.6)lowerthanthenumberofadultMacquarieperchN(Table2),then(i)thetrueNeforeachpopulationcouldbe~4.5×(1.7–8.6)higherthanLn sPm Neestimate(adjustedNe;Figure2),and(ii)adultpop-ulationsizeN couldbe~44.4× (6.27–82.6)higher thanLn sPm Ne estimate(AppendixS2).

Given that inbreedingNe is similar to varianceNe for this spe-cies (seeSection2), comparisonsof the aboveapproximationswith100/1,000minimumNethresholdsofFrankhametal.(2014)suggestthat(i)themajorityofpopulationshavemeanNe<600(lowerboundofNe<100, upper boundofNe<1,000), implying that genetic res-torationisrequiredtoboostpopulationadaptivepotentialandavoidinbreeding depression within a few decades; (ii) Wollemi, Wheenyand Erskine have mean Ne<100, implying they might experienceinbreedingdepressionwithina fewgenerationseven if theydonotalready;(iii)themajorityofpopulationshavemeansizes<3,000(lowerbound<500, upperbound<10,000),whichour simulations suggestmayleadtogeneticproblemsinthenextdecade;(iv)Little,CataractDam,DartmouthandYarrahavethelargestNeandmeansizes>3,000andthusarethebestsourcesfortranslocations.Wenotethatforfourpopulationswithsamplesizesof30individuals(UpperMurrumbidgee,Hollands,SevensandKingParrot)LDNeestimatesofNehadmeansandranges<100(e.g.outsidetherangeofTallmonetal.(2010)’ssim-ulations) andwere smaller thanLn sPm estimates, suggesting that

forsmallpopulationstheseLn sPm-basedapproximationscouldbestrongly upwardly biased (Appendix S2). For example, the numberofadultsmonitoredattheUpperMurrumbidgeesiteisestimatedat~20–50(Lintermans,unpublisheddata),whereasourLn sPm-basedapproximationprovidesamuchmoreoptimisticvalueof113–3,500(theLDNe-basedapproximationiscloser,13–307;AppendixS2).

For an idealWright–Fisher population of a stable size, differentmethodsofestimatingNe(e.g.fromLD,lossofgeneticdiversityovergeneration,driftamongreplicatesovergenerationsand/orinbreedingcoefficients)shouldyieldsimilarestimates.However,whenconditionsoftheWright–Fisherpopulationmodelareviolated(e.g.bytheeffectsof unequal sex-ratio, variance in family sizes and/or fluctuations inpopulation sizeovergenerations),differentmethodsyieldestimatesrelated to different time frames and spatial scales (Crow&Kimura,1970;Wang,2005).ThecalculationsabovedonotcorrectforbiasesrelatedtoviolationsoftheWright–Fishermodel.Indecliningpopula-tions,estimatesofNeareexpectedtobeupwardlybiased,reflectingapreviouslylargeNeonetotwogenerationsearlier(forLDNe)orseveralgenerationsinthecaseofLn sPm,anddeclinesareobservedinsomeMacquarie perch populations (e.g. Yarra; Tonkin, unpublished data).Despitemuchuncertainty aroundour approximationsof populationsizes,manyofourresultspointtotheneedtorestoregeneticdiversitybytranslocationsinmostMacquarieperchpopulations.Reconnectionof severalpopulationsbygene flowmightbe required to achieveaglobalNe>1,000necessary tomaintainpopulationadaptivepoten-tialandevolvabilityinthefaceofenvironmentalchange(Weeksetal.,2011).MeasuringNedynamicsusingmultipletemporalgeneticsam-pleswillallowmanagerstomonitortheeffectsofassistedgeneflowon Ne. When genomic data including those on physical linkage orgenomicpositionofgeneticmarkersbecomeavailableforMacquarieperch, contemporaryNe and its recent change could be estimatedfromcontemporarysamples(Hollenbecketal.,2016).

4.2 | Designing translocation strategies while considering risks of outbreeding depression and loss of local adaptation

Selecting appropriate sources of admixture requires simultaneousconsiderationof the timescaleofpopulationdifferentiationand therisks of inbreeding versus outbreeding depression and loss of localadaptation(Frankhametal.,2011;LoveStowelletal.,2017;Weeksetal., 2011, 2016). Long-termdivergence implies that reproductiveisolation may have evolved through adaptive differentiation and/or drift; hence, admixture between regions showing histories oflong-termdivergenceand/or localadaptationshouldbeundertakenwith caution (Frankham etal., 2011). Long-term divergence of theMDB and HNB Macquarie perch lineages (119–385 KY; Pavlovaetal., 2017) in different environments (e.g. Appendix S4) mighthave resulted in adaptive differences between basins. However,analysesofcompletemitogenomessuggestthatgeneticdriftwastheprominentforcedrivingmitochondrialdivergence;driftwasstrongerintheHNBeitherduetosmallerhistoricaleffectivepopulationsizes(approximated by allelic richness or haplotype diversity) or weaker

     |  15PAVLAP et PVl.

environmental pressures (Pavlova etal., 2017). In accordance withmitochondrialdifferentiation,themicrosatellitegenotypeofthesingleindividual from the Shoalhaven Basin population differedmarkedlyfromthoseof theHNBandMDB lineages (AppendicesS8andS9).ThefactthattheShoalhaven individualwashomozygousat90%oflocisuggeststhat lossofgeneticvariationand individual inbreedingaccompanied extinction of the endemic Shoalhaven lineage. Thewidespreadandnotsex-limitedhybridizationbeyondfirstgenerationof HNB and MDB lineages in lower Cataract River provides noevidence for outbreeding depression, but evidence for absence ofgenomicincompatibilitiesandunimpededfitnessofhybridsislacking(Verhoeven, Macel, Wolfe, & Biere, 2011). Separate managementofMacquarieperch lineages fromdifferentbasins,which is currentpractice,iswarranteduntilaninformedrisk–benefitanalysissuggestsotherwise (Frankhametal.,2011).Likewise, long-term (58–191KY;Pavlovaetal.,2017)divergencebetweenthenorthern(here:Wollemi,Wheeny) and southernHNBpopulations (here: the remainingHNBpopulations) suggests that genetic rescue/restorationbeconductedindependentlywithintheseregions,unlessan informedrisk–benefitanalysisindicatesotherwise.

Insomecases,nearby(presumablysimilarlyadapted)sourcepop-ulationsdonotexist,sotheremaybenooptionbuttosourceindivid-ualsfortranslocationfromamorestronglydivergentpopulation(LoveStowelletal.,2017;Weeksetal.,2011),assuggestedforthreatenedsouthern pygmyperch from theMDBbasedon genotype–environ-mentassociationanalyses(Brauer,Hammer,&Beheregaray,2016).IntheLachlancatchment(MDB),twosampledMacquarieperchpopula-tionsarerecentderivativesfromthesamesource(andareconnectedbyunidirectionalgeneflowfromtheAbercrombietoLachlan;Faulksetal., 2011); thus, little genetic rescue/restoration effectwould beexpectedfromtheiradmixture(Frankham,2015).However,theesti-mate of divergence of the Lachlan catchment from the otherMDBcatchments(2–88KYago)isnotrecentenoughtodismissthepossi-bilityofoutbreedingdepressionunderadmixturewithsouthernMDBpopulationsunderrecentdivergencecriteria(Frankhametal.,2011).TherangeoflikelyadjustedNevalues(41–360)andapproximatepop-ulationsizes(150–3,500)forthetwoLachlancatchmentpopulationssuggeststhatunderado-nothingscenariothesepopulationswillfaceinbreeding and loss of genetic diversity within a few generations.Althoughknowingwithcertaintythefitnessconsequencesofcross-catchmentadmixtureswouldbedesirable,theneedformanagementinterventiontopreventlossofgeneticdiversityandinbreedinginthenearfuturemaynecessitategeneflowfromlargerpopulationsoffishofMDBgeneticorigins(e.g.DartmouthorYarra)beforetheriskofout-breedingdepressioncanbeassessed throughbreedingexperiments(Frankham,2015;Frankhametal.,inpress).Thissuggestionisinlinewithcallsforfocussingonpreservationofgeneticdiversityatthespe-cies level, ratherthanpopulation,subspeciesorevolutionarysignifi-cantunit(LoveStowelletal.,2017;Weeksetal.,2016).

In cases when it is unclear whether mitochondrial diver-gence confers selective advantage on local mitogenomes (such asAdjungbilly, inferredtohavedivergedfromtheotherpopulationsoftheMurrumbidgeecatchment~8–100KYago),geneticrescuecould

involvewithin-catchment translocation of males. Assuming that onthistimescalemitochondrialdivergencedidnotdrivemito-nuclearco-evolution(Wolff,Ladoukakis,Enriquez,&Dowling,2014),thistrans-locationapproachwouldpreserveanypotential selectiveadvantageof localmtDNAwhileprovidingnucleargene flow for increasingNe (Weeks&Corrigan,2011).

Incontrasttocasesoflong-termdivergence,admixtureofpopula-tionsthathavedivergedonlyrecently(includingduetorecentanthro-pogenic barriers tomovement) should be associatedwithvery lowriskofoutbreedingdepressionorlossoflocaladaptation(Frankham,2015),whereasmanagingthesepopulationsseparatelywilldecreasetheiradaptivepotentialandincreasespeciesextinctionrisk(Colemanetal., 2013;Weeks etal., 2016). In the HNB, extreme divergencebetween the two northern Macquarie perch populations (WollemiandWheeny)islikelyduetorecentisolationbyanthropogenicbarri-ers(Faulksetal.,2011)andstrongeffectsofgeneticdriftinverysmallpopulations(meanadjustedNe<100).IntheMDB,historicalrecordsindicate that prior to arrival of Europeans in Australia, MacquarieperchwerecommonalongthelengthoftheMurrumbidgeeRiverandinmanytributariesoftheMurrayRiver(includinglowlandzoneoftheCentralMurrayRiver),suggestingthattheMurrumbidgeeandMurraycatchmentswerepreviouslyconnectedbygeneflow(Trueman,2011).Historical cross-catchment connectivity is supported by the mostcommon MDB haplotype (10 on Figure1) being shared betweenMurrumbidgee and Murray catchments. Current differentiationamongpopulationswithintheMurrumbidgeeandMurraycatchments(likely resulting fromveryrecent isolatingeffectsof largedamsandimpoundments)shouldbereversedbywithin-catchmentadmixtures.Occasionalcross-catchmenttranslocationsshouldalsobeconsideredtoimprovepopulationadaptivepotential.Designingcross-catchmenttranslocations (e.g. from the Murray catchment to the Lachlan orMurrumbidgeecatchments)wouldbenefitfromfurthermodellingandcaptivebreedingexperimentsmeasuringrelativefitnessofintercatch-ment crosses. Similar recommendations were suggested for otherthreatenedAustralianfisheswithanthropogenicallyimpededgeneticconnectivity(Braueretal.,2013,2016;Coleetal.,2016).

Historicallytranslocatedpopulations,thathavebeenisolatedforsufficientgenerationstoaccumulateallelicdifferentiation, representviablesourcesfor translocationsback intotheirpopulationoforiginiftheycontaindiversitylostfromtherecipientpopulation(Frankham,2015). Two translocated populations of Macquarie perch, CataractDam,established fromtheMurrumbidgeeRiver in1916 (LegislativeAssemblyofNewSouthWales1916)andtheYarra,sourcedbetween1907 and1943 frommultiple southernMDBpopulations includingKingParrotCreek,BrokenRiverand theGoulburnRiver (where thespecies is now locally extinct; Cadwallader, 1981; Trueman, 2011;Ho&Ingram,2012),couldbeusedassourcepopulationsforlow-risktranslocationstoMurrumbidgeeandMurraycatchments,respectively.Giventheirhighgeneticdiversity,CataractDamandYarrapopulationsmayconstituteparticularlyviabletranslocationoptions,providedtheirNe and Ncansupportit.

Finally, our simulations suggest that in cases when conditionsofaremnantpopulationareunlikelyto improveenoughtofacilitate

16  |     PAVLAP et PVl.

long-termsustainability (as increeks lackingconnectivitywithmorestableriversystems,andwithinsufficientwateravailabilityundercli-matechangescenarios),translocatingallremainingindividualsintoamore sustainable systemmight bemore beneficial than attemptingtomaintaintheoriginalpopulations,fortheoverallchanceofspeciessurvival and adaptive potential. Successful examples of fish rescueduring extremedrought and bushfire runoff already exist (includingMacquarieperchrescue;Kearns,2009;Hammeretal.,2013,2015).ManagementplansincorporatingsuchactionscouldhavepreventedhighmortalityandlossofgeneticdiversityoftheBrokenRiverpop-ulationofMacquarieperchduringtheMillenniumDroughtof1995–2009; only three individuals have beendetected in the reach since2011,whereasprior to thedrought this populationwas consideredstable(Kearns,unpublisheddata).

4.3 | Environmental correlates of genetic diversity

EnvironmentalmodellingsuggestedthatMacquarieperchweremoregeneticallydiverse (i.e.had lowerHL) if theywere (i) furtherdown-streamfromabarrier(damwall,spillwayorlargedam)orhadlargerflow-path lengthsupstream, (ii) in a short-to-medium,butnot long,stretchofabarrier-free fragmentof thestream,and (iii) in streamsthatarecolderinwinter.Lowergeneticdiversityinsitesimmediatelydownstreamfromdamscouldindicatethatcold-waterpollutionand/orotheraspectsofwatersupplyreleases(e.g.changesinhabitatqual-ity/availability)havealong-lastingeffectonMacquarieperchpopula-tions,aswaspreviouslyinferredforMurraycod(Todd,Ryan,Nicol,&Bearlin,2005). Indeed,theMacquarieperchpopulationintheMittaMitta River downstream of Dartmouth Dam in Victoria was extir-pated following the dam’s construction, with cold-water pollutionthemost likely cause (Koehn,Doeg,Harrington,&Milledge,1995).Alternatively, the correlation betweenbarrier-free flow-path lengthupstreamandmeanannualflow(r=0.78,AppendixS5)suggeststhatlarger streams (with highermean annual flow)might be harbouringlargerpopulations,forexamplebyprovidingmorehabitatand/orbybeingmoreresilienttodroughtsorotherdisturbances.Unfortunately,few reaches with these attributes remain, with the overwhelmingmajorityofmid-anduplandreachesoflargeriverswherethespecieswereonceabundant (Trueman,2011)nowcontaining, or regulatedheavilyby,majorweirsanddams.High levelsof correlationamongenvironmental variables make it challenging to identify underlyingcausaldrivers.

Our second result, that individuals are less diverse in very longconnectedstreamfragments(>1.5standarddeviationsfromthemeanlength; Figure3), is counterintuitive, as one would expect longerbarrier-free paths to providemore habitat.The longest barrier-freeflow-pathlengthsarerestrictedtothreeMDBpopulations:Adjungbilly,upperMurrumbidgeeandSevens (AppendixS4),andare largest (>2standarddeviations)inAdjungbillyandupperMurrumbidgee,thetwopopulations with low genetic diversity. Thus, a significant relation-ship between barrier-free flow-path length with HL could be spu-rious anddrivenbypopulation history or other correlatedvariables(e.g. limitedaccesstosuitablespawningsitesorpresenceofnatural

barriersthatsubdividepopulation–thatis,waterfalls,notconsideredfor thismetric).The final finding that individuals are less diverse instreamsthatarewarmerinwinterraisesthepossibilitythatsomelifestagesofMacquarieperchcouldbesensitivetowarmerwintersandthatclimatewarmingmayfurtherexacerbatelossofgeneticdiversity.Althoughwe foundsomepotentiallymeaningfulenvironmentalcor-relatesof individual genetic diversity, thedatawere limited in theirability to distinguish whether environmental factors or populationhistory(geneticdrift/geneflow/mutation)arethemaindeterminantsofgeneticdiversity,mainlybecausethespeciesislimitedtoveryfewsmall and isolated upstream reaches, without much environmentalvariationacrosssitesineach(AppendixS4).

4.4 | Summary of management recommendations

WefoundthatmostMacquarieperchpopulationsareassociatedwithlow levels of genetic diversity and effective population size. GiventhatMacquarieperchcontinuetoexperienceongoingdeclinesinspiteofhabitatrestorationandthreatameliorationefforts, thesepopula-tionsareatriskofextinction,atleastinpartduetonegativegeneticeffects(e.g.inbreedingdepression,lowadaptivepotential).Toelevategeneticdiversity,Neandpopulationsizes,werecommendreconnect-ingpopulationswithinthefollowingregionswithassistedgeneflow:(i)northernHNB;(ii)southernHNB;(iii)theMurraycatchment,whereYarraisrecommendedasanadditionalsourceoftranslocations;and(iv) the Murrumbidgee catchment, where Cataract Dam is recom-mendedasanadditionalsource.Severallinesofevidencesuggestthatthemixingrecommendedaboveposesanegligibleriskofoutbreedingdepressionandswampingoflocaladaptation:(i)manypopulationsarederivedfromalargerancestralpopulationwithgeneticdiversitypri-marilylostthroughdriftinpopulationsthataresmallthroughhumanactions;(ii)thesepopulationshaveonlyrecentlybeenisolated;thus,genomicincompatibilitiesareunlikely;(iii)theadmixedYarrapopula-tionappearstobehealthyandself-sustainable.Managingpopulationswithintheseregionsasametapopulationislikelytobebeneficialforlong-termpopulationviability.

Genetic rescue/restoration is a powerful management tool fora numberof reasons: (i) it canhave a large and long-lasting impacton genetic diversity and adaptive potential for relatively little addi-tionalworkandexpense(Frankham,2015),(ii)itdoesnotnecessarilyrequire ongoing intervention, although if populations remain small,assisted gene flowwill need to be performed periodically (Hedrick,Peterson,Vucetich,Adams,&Vucetich,2014), (iii) itcouldworkatarange of scales from single species rescue (Hedrick & Fredrickson,2010) to landscape reconnections (Smith, van der Ree, & Rosell,2015). Population viability analysis can assist in developing objec-tiveandmeasurablecriteriaofmetapopulationconnectivity (Carroll,Fredrickson,&Lacy,2014;Lacy,2000),providedsufficientknowledgeofspeciesbiologyexists.Empiricalresultsfromothersystemscouldbeusedtohelpprovidepriorsforunknownparameters(e.g.inbreed-ingdepression)inpopulationviabilitymodelsusedtopredictpopula-tionpersistenceundermanagementinactionandvariousscenariosofgeneticintervention.Translocationsofjuveniles,inadditiontoadults,

     |  17PAVLAP et PVl.

overseveralconsecutiveyearshavebeenadvocatedwhentherearelimitedadultfishavailableinasourcepopulations(Todd&Lintermans,2015).Forthesecases,wearguefortranslocatinggeneticallydiversesetsofindividuals,achievedbycollectingasmallnumberofjuvenilesfromeachofmanysamplingsites.Numbersoftranslocatedjuvenileswouldneedtoaccountforhighjuvenilemortality(e.g.forMacquarieperch~561-year-oldjuvenileswouldberequiredtoproduce6adults;AppendixS6;ToddandLintermans(2015)).

Even ifdirectevidenceofgeneticproblems (fitness loss) in spe-cificcases is regardedasdesirable, itshouldnotberequiredto jus-tifygeneticinterventionforsmallpopulations.Geneticaugmentationaloneoftenmaynotbesufficienttoreversepopulationdeclines,butitmustaccompanyotherthreatmanagementactionsforsmallpopu-lations(Frankhametal.,inpress;LoveStowelletal.,2017).Performedandmonitoredcarefullywithinarisk-assessmentframework(e.g.afterconsidering the risk of outbreeding depression), in most instancesgeneticaugmentationwillbebeneficial,andnotharmful,forlong-termpopulationviability.

ACKNOWLEDGEMENTS

Thisworkwas supported by ARCGrant LP110200017 toMonashUniversity,FlindersUniversityofSouthAustraliaandtheUniversityof Canberra, with Partner Organization University of Montana.Funding and other support were contributed by industry PartnerOrganizationsIconWater(formerlyACTEWCorporation),MelbourneWater,DepartmentofSustainabilityandEnvironment (Victoria)andFisheries Victoria (nowwithin DEDJTR - Department of EconomicDevelopment, Jobs, Transport and Resources). Goulburn BrokenCatchmentManagementAuthority (GBCMA)provided the funds toundertakefishsurveysandcollectgeneticmaterial.Additionalfund-ingwasprovidedtoAPandPSbytheFisheriesVictoriaRecreationalFishing Grants programme (DEDJTR) and to DG by NSW DPI—Fisheries. Computationally intensive analyseswere performed on aMonash computer cluster.We thankM.Asmus, R. Ayres, P. Boyd,B. Broadhurst, A. Bruce, T. Cable, R. Clear, J. Doyle, B. Ebner, L.Farrington, L. Faulks, D. Hartwell, P. Heath, J. Knight, J.Mahoney,A.McDonald,T.McGarry,J.O’Mahony,A.Pickworth,K.Pitman,S.Raymond,M.Rodgers,J.Stanger,D.Stoessel,B.Talbot,M.Timminsand P. Unmack for sample collection,M. Roitman, A. Sunnucks, R.Alburyforlaboratoryassistance,D.TallmonforprovisionofthelocallyinstallableversionofLn sPm, L. Faulks for curatorial assistanceofearliersamplesandJ.Thomsonforstatisticaladvice.WearegratefultotheChiefEditorProf.LouisBernatchez,AssociateEditorProf.CraigPrimmerandthreeanonymousreviewers,whosecommentshelpedustoimprovethemanuscript.

DATA ARCHIVING STATEMENT

Data for this study are available as Appendices and SupportingInformation (collection details, lengths, microsatellite genotypes,mitochondrial control region sequences and HL values for 872Macquarie perch samples from 20 populations and 11 raw and

standardizedenvironmentalvariables for16populations).SequencedataarealsodepositedtoGenBank(accessionKT626048–KT626381).

REFERENCES

Alho,J.S.,Valimaki,K.,&Merilä,J. (2010).Rhh:AnRextensionforesti-mating multilocus heterozygosity and heterozygosity-heterozygositycorrelation. Molecular Ecology Resources,10,720–722.

Allendorf,F.W.,&Ryman,N.(2002).The role of genetics in population viabil-ity analysis. Population viability analysis(pp.50–85).Chicago:UniversityofChicagoPress.

Aparicio,J.M.,Ortego,J.,&Cordero,P.J.(2006).Whatshouldweweighto estimate heterozygosity, alleles or loci? Molecular Ecology, 15,4659–4665.

Appleford,P.,Anderson,T.,&Gooley,G. (1998).Reproductivecycleandgonadal development of Macquarie perch, Macquaria australasica Cuvier (Percichthyidae), in Lake Dartmouth and tributaries of theMurray–DarlingBasin,Victoria,Australia.Marine & Freshwater Research,49,163–169.

Attard, C., Möller, L., Sasaki, M., Hammer, M., Bice, C., Brauer, C., …Beheregaray,L. (2016).Anovelholistic frameworkforgenetic-basedcaptive-breeding and reintroduction programs. Conservation Biology,30,1060–1069.

Bandelt,H.-J.,Forster,P.,&Roehl,A.(1999).Median-joiningnetworksforinferring intraspecific phylogenies.Molecular Biology & Evolution, 16,37–48.

Becker, J. A., Tweedie, A., Gilligan, D., Asmus, M., &Whittington, R. J.(2013).Experimental infectionofAustralianfreshwaterfishwithepi-zootichaematopoieticnecrosisvirus(EHNV).Journal of Aquatic Animal Health,25,66–76.

Benson,J.F.,Mahoney,P.J.,Sikich,J.A.,Serieys,L.E.,Pollinger,J.P.,Ernest,H.B.,&Riley,S.P.(2016).Interactionsbetweendemography,genetics,andlandscapeconnectivityincreaseextinctionprobabilityforasmallpopulationoflargecarnivoresinamajormetropolitanarea. Proceedings of the Royal Society B, 283,20160957.

Brauer, C. J., Hammer, M. P., & Beheregaray, L. B. (2016). Riverscapegenomics of a threatened fish across a hydroclimatically heteroge-neousriverbasin.Molecular Ecology,25,5093–5113.

Brauer,C.J.,Unmack,P.J.,Hammer,M.P.,Adams,M.,&Beheregaray,L.B.(2013).Catchment-scaleconservationunits identified for the threat-enedYarraPygmyPerch(Nannoperca obscura)inhighlymodifiedriversystems.PLoS One,8,e82953.

BureauofMeterology. (2012).Australian hydrological geospatial fabric ver-sion 2, Bureau of Meterology.Retrievedfromhttp://www.bom.gov.au/water/geofabric

Cadwallader,P.L.(1978).Somecausesofthedeclineinrangeandabun-danceofnativefishintheMurray–Darlingriversystem.Proceedings of the Royal Society of Victoria,90,211–224.

Cadwallader,P. (1981).Pastandpresentdistributionsand translocationsof Macquarie perch Macquaria australasica (Pisces: Percichthyidae),withparticularreferencetoVictoria.Proceedings of the Royal Society of Victoria,93,23–30.

Cadwallader,P.,&Rogan,P.(1977).TheMacquarieperch,Macquria austral-asica(Pisces:Percichthyidae),ofLakeEildon,Victoria.Australian Journal of Ecology,2,409–418.

Carroll,C.,Fredrickson,R.J.,&Lacy,R.C.(2014).Developingmetapopula-tionconnectivitycriteriafromgeneticandhabitatdatatorecovertheendangeredMexicanwolf.Conservation Biology,28,76–86.

Cole,T.L.,Hammer,M.P.,Unmack,P.J.,Teske,P.R.,Brauer,C.J.,Adams,M.,&Beheregaray,L.B.(2016).Range-widefragmentationinathreat-ened fish associatedwith post-European settlementmodification intheMurray–DarlingBasin,Australia.Conservation Genetics,17,1377.

Coleman,R.,Pettigrove,V.,Raadik,T.,Hoffmann,A.,Miller,A.,&Carew,M.(2010).MicrosatellitemarkersandmtDNAdata indicatetwodistinct

18  |     PAVLAP et PVl.

groupsindwarfgalaxias,Galaxiella pusilla(Mack)(Pisces:Galaxiidae),athreatenedfreshwaterfishfromsouth-easternAustralia.Conservation Genetics,11,1911–1928.

Coleman,R.,Weeks,A.,&Hoffmann,A.(2013).Balancinggeneticunique-nessandgeneticvariationindeterminingconservationandtransloca-tionstrategies:Acomprehensivecasestudyofthreateneddwarfgalax-ias,Galaxiella pusilla(Mack)(Pisces:Galaxiidae).Molecular Ecology,22,1820–1835.

Crow,J.F.,&Kimura,M.(1970).An introduction to population genetic theory. NewYork,NY:Harper&Row.

Do,C.,Waples,R.S.,Peel,D.,Macbeth,G.,Tillett,B.J.,&Ovenden,J.R.(2014).NeEstimatorv2:Re-implementationofsoftwarefortheestima-tionofcontemporaryeffectivepopulationsize(Ne)fromgeneticdata.Molecular Ecology Resources,14,209–214.

Drummond,A.,Ashton,B.,Buxton,S.,Cheung,M.,Cooper,A.,Duran,C.,…Wilson,A.(2012).Geneious v5.6.Retrievedfromhttp://www.geneious.com

Drummond,A.J.,Suchard,M.A.,Xie,D.,&Rambaut,A.(2012).BayesianphylogeneticswithBEAUtiandtheBEAST1.7.Molecular Biology and Evolution,29,1969–1973.

Dufty, S. (1986). Genetic andmorphological divergence between popu-lationsofMacquarieperch (Macquaria australasica) east andwestoftheGreatDividingRange.UnpublishedBSc.(Hons)thesis,UniversityofNewSouthWales,Kensington.

Environmental Systems Research Institute. (1999-2014). ArcGIS version 10.2inR.California,ed.RedlandsCalifornia.

Excoffier,L.,&Lischer,H.E.L.(2010).Arlequinsuitever3.5:AnewseriesofprogramstoperformpopulationgeneticsanalysesunderLinuxandWindows.Molecular Ecology Resources,10,564–567.

Farrington, L.W., Lintermans,M., & Ebner, B. C. (2014). CharacterisinggeneticdiversityandeffectivepopulationsizeinonereservoirandtworiverinepopulationsofthethreatenedMacquarieperch.Conservation Genetics,15,707–716.

Faulks,L.K.,Gilligan,D.M.,&Beheregaray,L.B. (2010a).Evolutionandmaintenanceofdivergent lineages inanendangered freshwater fish,Macquaria australasica. Conservation Genetics,11,921–934.

Faulks,L.K.,Gilligan,D.M.,&Beheregaray,L.B.(2010b).Islandsofwaterinaseaofdryland:Hydrologicalregimepredictsgeneticdiversityanddispersal in awidespread fish fromAustralia’s arid zone, the goldenperch(Macquaria ambigua).Molecular Ecology,19,4723–4737.

Faulks,L.K.,Gilligan,D.M.,&Beheregaray,L.B.(2011).Theroleofanthro-pogenic vs. natural in-stream structures in determining connectivityand genetic diversity in an endangered freshwater fish, Macquarieperch(Macquaria australasica).Evolutionary Applications,4,589–601.

Faulks, L., Gilligan, D., & Beheregaray, L. B. (2015). ‘Ragged mountainranges,droughtsandfloodingrains’:Theevolutionaryhistoryandcon-servationofAustralianfreshwaterfishes.InA.Stow,N.Maclean,&G.I.Holwell (Eds.),Austral ark: The state of wildlife in Australia and New Zealand.Cambridge:CambridgeUniversityPress.

Fisher,M.C.,Garner,T.W.,&Walker,S.F. (2009).GlobalemergenceofBatrachochytrium dendrobatidis and amphibian chytridiomycosis inspace,time,andhost.Annual Review of Microbiology,63,291–310.

Fitzpatrick, S., Gerberich, J., Kronenberger, J., Angeloni, L., & Funk,W.(2015).Locallyadaptedtraitsmaintainedinthefaceofhighgeneflow.Ecology Letters,18,37–47.

Fountain, T., Nieminen, M., Sirén, J., Wong, S. C., & Hanski, I. (2016).Predictableallele frequencychangesduetohabitat fragmentation intheGlanvillefritillarybutterfly.Proceedings of the National Academy of Sciences,113,2678–2683.

Frankham, R. (1995). Inbreeding and extinction: A threshold effect.Conservation Biology,9,792–799.

Frankham, R. (1998). Inbreeding and extinction: Island populations.Conservation Biology,12,665–675.

Frankham,R.(2005).Geneticsandextinction.Biological Conservation,126,131–140.

Frankham, R. (2015).Genetic rescue of small inbred populations:Meta-analysisreveals largeandconsistentbenefitsofgeneflow.Molecular Ecology,24,2610–2618.

Frankham,R. (2016).Geneticrescuebenefitspersist toat theF3gener-ation,basedonameta-analysis.Biological Conservation,195,33–36.

Frankham,R.,Ballou,J.D.,Eldridge,M.D.B.,Lacy,R.C.,Ralls,K.,Dudash,M.R.,&Fenster,C.B.(2011).Predictingtheprobabilityofoutbreedingdepression.Conservation Biology,25,465–475.

Frankham, R., Ballou, J. D., Ralls, K., Eldridge,M. D. B., Dudash,M. R.,Fenster,C.B.,…Sunnucks,P. (inpress).Genetic management of frag-mented animal and plant populations.Oxford:OxfordUniversityPress.

Frankham,R.,Bradshaw,C.J.A.,&Brook,B.W.(2014).Geneticsincon-servation management: Revised recommendations for the 50/500rules, Red List criteria and population viability analyses. Biological Conservation,170,56–63.

Franklin, I. R. (1980). Evolutionary change in small populations. InM. E.Soule,&B.A.Wilcox(Eds.),Conservation biology: An evolutionary-eco-logical perspective(pp.135–148).Sunderland,MA:Sinauer.

Fu,Y.-X.(1997).Statisticaltestsofneutralityofmutationsagainstpopulationgrowth,hitchhikingandbackgroundselection.Genetics,147,915–925.

GeoscienceAustralia.(2012).National environmental stream attributes data-base v1.1.5.GeoscienceAustralia.Retrievedfromhttp://www.ga.gov.au/metadata-gateway/metadata/record/gcat_75066

Gilligan, D., McGarry, T., & Carter, S. (2010). A scientific approach todeveloping habitat rehabilitation strategies in aquatic environments:AcasestudyontheendangeredMacquarieperch(Macquaria austral-asica) in the Lachlan catchment.A report to the LachlanCatchmentManagement Authority. Department of Industry and Investment(Industry&InvestmentNSW).

Goudet,J.(2001).FSTAT, a program to estimate and test gene diversities and fixation indices (version 2.9.3).

Hammer,M.P.,Bice,C.M.,Hall,A.,Frears,A.,Watt,A.,Whiterod,N.S.,…Zampatti,B.P.(2013).Freshwaterfishconservationinthefaceofcrit-icalwatershortagesinthesouthernMurray–DarlingBasin,Australia.Marine & Freshwater Research,64,807–821.

Hammer,M. P., Goodman,T. S.,Adams,M., Faulks, L. F., Unmack, P. J.,Whiterod,N.S.,&Walker,K.F.(2015).Regionalextinction,rediscoveryandrescueofa freshwater fish fromahighlymodifiedenvironment:Theneedforrapidresponse.Biological Conservation,192,91–100.

Hardy,O.J.,&Vekemans,X. (2002).SPAGeDi:Aversatilecomputerpro-gramtoanalysespatialgeneticstructureattheindividualorpopulationlevels.Molecular Ecology Notes,2,618–620.

Hare, M. P., Nunney, L., Schwartz, M. K., Ruzzante, D. E., Burford, M.,Waples,R.S.,…Palstra,F.(2011).Understandingandestimatingeffec-tivepopulationsizeforpracticalapplicationinmarinespeciesmanage-ment. Conservation Biology,25,438–449.

Harrisson, K.A., Pavlova,A., Gonçalves da Silva,A., Rose, R., Bull, J. J.,Lancaster,M.,…Sunnucks,P. (2016).Scopeforgeneticrescueofanendangeredsubspeciesthoughre-establishingnaturalgeneflowwithanothersubspecies.Molecular Ecology,25,1242–1258.

Harrisson,K.A.,Pavlova,A.,Telonis-Scott,M.,&Sunnucks,P.(2014).Usinggenomics to characterize evolutionary potential for conservation ofwildpopulations.Evolutionary Applications,7,1008–1025.

Harrisson,K.A.,Yen, J.D. L., Pavlova,A., Rourke,M. L.,Gilligan,D.M.,Ingram,B.,…Sunnucks,P.(2016).Identifyingenvironmentalcorrelatesofintra-specificgeneticvariation.Heredity,117,155–164.

Hedrick,P.W.(1995).Geneflowandgeneticrestoration:TheFloridapan-therasacasestudy.Conservation Biology,9,996–1007.

Hedrick,P.W.(1999).Perspective:Highlyvariablelociandtheirinterpreta-tioninevolutionandconservation.Evolution,53,313–318.

Hedrick, P.W.,& Fredrickson, R. (2010).Genetic rescue guidelineswithexamples from Mexican wolves and Florida panthers. Conservation Genetics,11,615–626.

Hedrick, P. W., Peterson, R. O., Vucetich, L. M., Adams, J. R., &Vucetich,J.A. (2014).Genetic rescue in IsleRoyalewolves:Genetic

     |  19PAVLAP et PVl.

analysisandthecollapseofthepopulation.Conservation Genetics,15, 1111–1121.

Ho, H. K., & Ingram, B. A. (2012). Genetic risk assessment for stocking Macquarie perch into Victorian waterways 2011. Fisheries VictoriaInternal Report No. 45 (p. 24). Alexandra: Department of Primaryindustries,FisheriesVictoria.

Hoffmann,A.,Griffin,P.,Dillon,S.,Catullo,R.,Rane,R.,Byrne,M.,…Sgrò,C. (2015).Aframeworkfor incorporatingevolutionarygenomics intobiodiversityconservationandmanagement.Climate Change Responses,2,1–24.

Hollenbeck, C., Portnoy, D., & Gold, J. (2016). A method for detect-ing recent changes in contemporary effective population size fromlinkage disequilibrium at linked and unlinked loci. Heredity, 117, 207–216.

Hufbauer,R.A.,Szucs,M.,Kasyon,E.,Youngberg,C.,Koontz,M.J.,Richards,C.,…Melbourne,B.A.(2015).Threetypesofrescuecanavertextinc-tioninachangingenvironment.Proceedings of the National Academy of Sciences,112,10557–10562.

Huisman, J., Kruuk, L. E., Ellis, P.A., Clutton-Brock, T., & Pemberton, J.M. (2016). Inbreedingdepressionacross the lifespan inawildmam-malpopulation.Proceedings of the National Academy of Sciences,113,3585–3590.

Ingram,B.,Barlow,C.,Burchmore,J.,Gooley,G.,Rowland, S.,&Sanger,A.(1990).ThreatenednativefreshwaterfishesinAustralia—Somecasehistories.Journal of Fish Biology,37,175–182.

Ingram,B.A.,Douglas,J.W.,&Lintermans,M.(2000).Threatenedfishesof the world: Macquaria australasica Cuvier, 1830 (Percichthyidae).Environmental Biology of Fishes,59,68–68.

Kearns,J.(2009).Translocation of Macquarie Perch, Macquaria australasica, from King Parrot Creek and Hughes Creek, Victoria 2009. Confidential client report prepared for the Murray–Darling Basin Authority (MDBA). Heidelberg,VIC:DepartmentofSustainabilityandEnvironment.

Keller,L.F.,&Waller,D.M.(2002).Inbreedingeffectsinwildpopulations.Trends in Ecology & Evolution,17,230–241.

Knight,J.,&Bruce,A.(2010).Threatenedfishprofile:‘Eastern’MacquarieperchMacquaria australasica Cuvier 1830.Australian Society for Fish Biology Newsletter,40(2),76–78.

Koehn,J.D.,Doeg,T.J.,Harrington,D.J.,&Milledge,G.A. (1995).The effects of Dartmouth Dam on the aquatic fauna of the Mitta Mitta River.ReporttoMurray–DarlingBasinCommission (p.151).Victoria:DepartmentofConservationandNaturalResources.

Kopelman,N.M.,Mayzel,J.,Jakobsson,M.,Rosenberg,N.A.,&Mayrose,I. (2015). Clumpak: A program for identifying clustering modes andpackagingpopulationstructureinferencesacrossK.Molecular Ecology Resources,15,1179–1191.

Lacy,R.C.(2000).StructureoftheVORTEXsimulationmodelforpopula-tionviabilityanalysis.Ecological Bulletins,48,191–203.

Lacy,R.C.,Miller,P.S.,&Traylor-Holzer,K.(2015).Vortex 10 user’s man-ual. 15 April 2015 update.IUCNSSCConservationBreedingSpecialistGroup, and Chicago Zoological Society Apple Valley, Minnesota, USA.

Lacy,R.,&Pollak,J.(2014).VORTEX: A stochastic simulation of the extinction process, version 10.0.Brookfield,IL:ChicagoZoologicalSociety.

LeCam,S.,Perrier,C.,Besnard,A.-L.,Bernatchez,L.,&Evanno,G.(2015).Genetic and phenotypic changes in an Atlantic salmon populationsupplemented with non-local individuals: A longitudinal study over21 years. Proceedings of the Royal Society B: Biological Sciences, 282,20142765.

Lee,W.-J., Conroy, J.,Howell,W.H.,&Kocher,T. (1995). Structure andevolutionofteleostmitochondrialcontrolregions.Journal of Molecular Evolution,41,54–66.

LegislativeAssemblyofNewSouthWales(1916).Report on the fisheries of New South Wales for the year 1915.

Lintermans,M.(2007).Fishes of the Murray–Darling Basin: An introductory guide.Canberra,ACT:Murray–DarlingBasinCommission.

Lintermans,M. (2012).Managingpotential impacts of reservoir enlarge-mentonthreatenedMacquaria australasica and Gadopsis bispinosus in southeasternAustralia.Endangered Species Research,16,1–16.

Lintermans,M.(2013).TheriseandfallofatranslocatedpopulationoftheendangeredMacquarieperch,Macquaria australasica,insouth-easternAustralia.Marine & Freshwater Research,64,838–850.

Lintermans, M., & Ebner, B. (2010). Threatened fish profile: ‘Western’MacquarieperchMacquaria australasicaCuvier1830.Australian Society for Fish Biology Newsletter,40(2),76–78.

Lintermans,M.,Lyon,J.P.,Hammer,M.P.,Ellis,I.,&Ebner,B.C.(2015).Underwater, out of sight: Lessons from threatened freshwater fishtranslocations in Australia. In D. P. Armstrong, M. W. Hayward,D.Moro, & P. J. Seddon (Eds.),Advances in reintroduction biology of Australian and New Zealand Fauna(pp.237–253).Collingwood:CSIROPublishing.

Lopez,S.,Rousset,F.,Shaw,F.H.,Shaw,R.G.,&Ronce,O. (2009).Jointeffectsofinbreedingandlocaladaptationontheevolutionofgeneticloadafterfragmentation.Conservation Biology,23,1618–1627.

Love Stowell, S.M., Pinzone, C.A., &Martin,A. P. (2017). Overcomingbarriers to active interventions for genetic diversity.Biodiversity and Conservation.https://doi.org/10.1007/s10531-017-1330-z.

Lowe,W.H.,&Allendorf,F.W. (2010).What cangenetics tell us aboutpopulationconnectivity?Molecular Ecology,19,3038–3051.

Luikart, G., Ryman, N., Tallmon, D. A., Schwartz, M. K., & Allendorf, F.W. (2010). Estimation of census and effective population sizes:The increasing usefulness of DNA-based approaches. Conservation Genetics,11,355–373.

Lunn, D. J., Best, N., &Whittaker, J. C. (2009). Generic reversible jumpMCMCusinggraphicalmodels.Statistics and Computing,19,395–408.

Lunn,D.J.,Thomas,A.,Best,N.,&Spiegelhalter,D.(2000).WinBUGS—ABayesianmodellingframework:Concepts,structure,andextensibility.Statistics and Computing,10,325–337.

Lunn,D.J.,Whittaker,J.C.,&Best,N.(2006).ABayesiantoolkitforgeneticassociationstudies.Genetic Epidemiology,30,231–247.

Mills,S.L.,&Allendorf,F.W.(1996).Theone-migrant-per-generationruleinconservationandmanagement.Conservation Biology,10,1509–1518.

Nguyen,T.T.T.,Ingram,B.A.,Lyon,J.,Guthridge,K.,&Kearns,J.(2012).Genetic diversity of populations of Macquarie perch, Macquaria aus-tralasica, in Victoria.FisheriesVictoria InternalReportNo.40 (p.14).Queenscliff:FisheriesVictoria.

Pavlova,A.,Gan,H.M.,Lee,Y.P.,Austin,C.M.,Gilligan,D.,Lintermans,M.,&Sunnucks,P.(2017).PurifyingselectionandgeneticdriftshapedPleistocene evolution of the mitochondrial genome in an endan-geredAustralianfreshwaterfish.Heredity,118,466–476.https://doi.org/10.1038/hdy.2016.120

Peakall,R.,&Smouse,P.E. (2006).GenAlEx6:Geneticanalysis inExcel.Population genetic software for teaching and research. Molecular Ecology Notes 6,288–295.

Pierson,J.C.,Beissinger,S.R.,Bragg,J.G.,Coates,D.J.,Oostermeijer,J.G.B.,Sunnucks,P.,…Young,A.G.(2015).Incorporatingevolutionaryprocesses into population viability models. Conservation Biology, 29,755–764.

Pierson,J.C.,Coates,D.,Oostermeijer,J.G.B.,Beissinger,S.R.,Bragg,J.G.,Sunnucks,P.,…Young,A.G.(2016).Geneticfactorsinthreatenedspeciesrecoveryplansonthreecontinents.Frontiers in Ecology and the Environment,14,433–440.

Pollard,D.,Ingram,B.,Harris,J.,&Reynolds,L.(1990).ThreatenedfishesinAustralia—Anoverview.Journal of Fish Biology,37,67–78.

Pritchard,J.K.,Stephens,M.,&Donnelly,P.(2000).Inferenceofpopulationstructureusingmultilocusgenotypedata.Genetics,155,945–959.

RDevelopmentCoreTeam(2014).R: A language and environment for statisti-cal computing.Vienna,Austria:RFoundationforStatisticalComputing.Retrievedfromhttp://www.R-project.org/ISBN3-900051-07-0.

Rambaut,A.,&Drummond,A.J.(2007).Tracer v1.4.Retrievedfromhttp://beast.bio.ed.ac.uk/Tracer

20  |     PAVLAP et PVl.

Reed, D. H., & Frankham, R. (2003). Correlation between fitness andgeneticdiversity.Conservation Biology,17,230–237.

Rieman,B.E.,&Allendorf,F.(2001).Effectivepopulationsizeandgeneticconservationcriteriaforbulltrout.North American Journal of Fisheries Management,21,756–764.

Rieseberg, L.H.,&Burke,J. (2001).Agenicviewof species integration.Journal of Evolutionary Biology,14,883–886.

Roberts,D.G.,Baker,J.,&Perrin,C.(2011).PopulationgeneticstructureoftheendangeredEasternBristlebird,Dasyornis brachypterus;implica-tionsforconservation.Conservation Genetics,12,1075–1085.

Rousset, F. (2008). genepop’007: A complete re-implementation of thegenepopsoftwareforWindowsandLinux.Molecular Ecology Resources,8,103–106.

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

Sgrò,C.M.,Lowe,A.J.,&Hoffmann,A.A. (2011).Buildingevolutionaryresilienceforconservingbiodiversityunderclimatechange.Evolutionary Applications,4,326–337.

Smith,D.J.,vanderRee,R.,&Rosell,C.(2015).Wildlife crossing structures. Handbook of road ecology (pp. 172–183). Chichester: JohnWiley &Sons,Ltd.

Spielman,D.,Brook,B.W.,&Frankham,R. (2004).Mostspeciesarenotdriven toextinctionbeforegenetic factors impact them.Proceedings of the National Academy of Sciences of the United States of America,101,15261–15264.

Stein,J.L.(2006).A continental landscape framework for systematic conser-vation planning for Australian rivers and streams. Canberra:AustralianNationalUniversity.

Szulkin,M., & Sheldon, B. C. (2007).The environmental dependence ofinbreedingdepressioninawildbirdpopulation.PLoS One,2,e1027.

Tajima, F. (1996). The amount of DNA polymorphism maintained in afinitepopulationwhen theneutralmutation ratevariesamongsites.Genetics,143,1457–1465.

Tallmon,D.A., Gregovich,D.,Waples, R. S., Scott Baker, C., Jackson, J.,Taylor,B.L.,…Schwartz,M.K.(2010).Whenaregeneticmethodsuse-fulforestimatingcontemporaryabundanceanddetectingpopulationtrends?Molecular Ecology Resources,10,684–692.

Tallmon,D.A.,Koyuk,A.,Luikart,G.,&Beaumont,M.A.(2008).ONeSAMP:A program to estimate effective population size using approximateBayesiancomputation.Molecular Ecology Resources,8,299–301.

Todd,C.R.,&Lintermans,M.(2015).Whodoyoumove?Astochasticpopu-lationmodeltoguidetranslocationstrategiesforanendangeredfresh-waterfishinsouth-easternAustralia.Ecological Modelling,311,63–72.

Todd,C.R.,Ryan,T.,Nicol,S.J.,&Bearlin,A.R.(2005).Theimpactofcoldwaterreleasesonthecriticalperiodofpost-spawningsurvivalanditsimplicationsforMurraycod (Maccullochella peelii peelii):Acasestudyof the Mitta Mitta River, southeastern Australia. River Research and Applications,21,1035–1052.

Trueman, W. T. (2011). True tales of the trout cod: River histories of the Murray–Darling Basin,MDBAPublicationNo.215/11.

Verhoeven,K.J.,Macel,M.,Wolfe, L.M.,&Biere,A. (2011).Populationadmixture, biological invasions and thebalancebetween local adap-tation and inbreeding depression. Proceedings of the Royal Society of London B: Biological Sciences,278,2–8.

Walling,C.A.,Nussey,D.H.,Morris,A.,Clutton-Brock,T.H.,Kruuk,L.E.,&Pemberton,J.M.(2011).Inbreedingdepressioninreddeercalves.BMC Evolutionary Biology,11,318.

Wang, J. (2005). Estimation of effective population sizes from data ongenetic markers. Philosophical Transactions of the Royal Society B: Biological Sciences,360,1395–1409.

Waples,R.S.,Antao,T.,&Luikart,G.(2014).Effectsofoverlappinggener-ationsonlinkagedisequilibriumestimatesofeffectivepopulationsize.Genetics,197,769–780.

Waples,R.S.,&Do,C.(2008).LDNE:Aprogramforestimatingeffectivepopulationsizefromdataonlinkagedisequilibrium.Molecular Ecology Resources,8,753–756.

Waples,R.S.,Luikart,G.,Faulkner,J.R.,&Tallmon,D.A.(2013).Simplelife-historytraitsexplainkeyeffectivepopulationsizeratiosacrossdiversetaxa.Proceedings of the Royal Society of London B: Biological Sciences,280,20131339.

Weeks,A. R., & Corrigan, T. (2011). Translocation strategy for the Mount Buller population of the mountain pygmy possum.Melbourne:VictorianGovernmentDepartmentofSustainabilityandEnvironment(DSE).

Weeks,A.R.,Sgro,C.M.,Young,A.G.,Frankham,R.,Mitchell,N.J.,Miller,K.A.,…Hoffmann,A.A. (2011).Assessing thebenefits and risksoftranslocations in changing environments: A genetic perspective.Evolutionary Applications,4,709–725.

Weeks, A. R., Stoklosa, J., & Hoffmann, A. A. (2016). Conservation ofgeneticuniquenessofpopulationsmay increaseextinction likelihoodof endangered species:The case ofAustralianmammals.Frontiers in Zoology,13,31.

Whiteley,A. R., Fitzpatrick, S.W., Funk,W.C.,&Tallmon,D.A. (2015).Genetic rescue to the rescue. Trends in Ecology & Evolution, 30, 42–49.

Willi,Y.,&Hoffmann,A.A.(2009).Demographicfactorsandgeneticvari-ation influence population persistence under environmental change.Journal of Evolutionary Biology,22,124–133.

Willi,Y.,VanBuskirk,J.,&Hoffmann,A.A.(2006).Limitstotheadaptivepotentialofsmallpopulations.Annual Review of Ecology, Evolution, and Systematics,37,433–458.

Wolff, J. N., Ladoukakis, E. D., Enriquez, J.A., &Dowling, D. K. (2014).Mitonuclear interactions: Evolutionary consequences over multi-ple biological scales.Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences,369,20130443.

Woodworth, L. M., Montgomery, M. E., Briscoe, D. A., & Frankham, R.(2002).Rapidgeneticdeteriorationincaptivepopulations:Causesandconservationimplications.Conservation Genetics,3,277–288.

SUPPORTING INFORMATION

Additional Supporting Informationmay be found online in the sup-portinginformationtabforthisarticle.

How to cite this article:PavlovaA,BeheregarayLB,ColemanR,etal.SevereconsequencesofhabitatfragmentationongeneticdiversityofanendangeredAustralianfreshwaterfish:Acallforassistedgeneflow.Evol Appl. 2017;00:1–20. https://doi.org/10.1111/eva.12484