In situ examination of boldnessshyness traits in the tropical poeciliid Brachyraphis episcopi - PDF document

Insituexaminationofboldness…shynesstraitsinthetropicalBrachyraphisepiscopiCULUMBROWN,FELICITYJONES&VICTORIABRAITHWAITEInstituteofEvolutionaryBiology,SchoolofBiologicalSciences,UniversityofEdinburgh(Received26October2004;initialacceptance6December2004;“nalacceptance21January2005;publishedonline2September2005;MS.number:8318) differencesalongtheshyness…boldnesscontinuum(etal.2004).Ageorsizeinparticularislikelytohaveasubstantialimpactonpersonalitytraitsinanimalssincemanyaspectsofhowananimalinteractswithitsenvi-ronmentchangeasitgrows(e.g.microhabitatchoice,diet,predationpressure).Furthermore,asanimalsagetheyhaveincreasingexperiencewiththeenvironmentandhaveeitheralteredtheirbehaviourtosuit(behaviouralplasticity)orhavebeeneliminatedfromthepopulation(selectivemortality).Laboratorystudieshaverevealedastrongrelationbetweenboldnessandbodysize(&Braithwaite2004)butthishasneverbeenveri“edin“eldstudies.Differencesbetweenthesexeswillalsohavesubstantialimpactsonpersonality,notleastbecausemalesandfemalesoftenhavedifferentlifehistoryprioritiesJennions&Telford2002Whilecontrolledlaboratorystudiescanprovideusefulinsights,theyareoftenhighlycontrivedandassumethatthebehaviouroftheanimalincaptivitywillbesimilartothatinitsnaturalenvironment.Owingtotheconstraintsofworkinginthelaboratory,experimentsinvestigatingboldnesshavebeensomewhatarti“cialandfewhaveattemptedtoinvestigatepersonalitytraitsinwildpopula-tions(seeWilsonetal.1993foranexception).Habitatvariablesknowntoaffecttheperceptionofrisk,suchasfoodavailability,predationpressureanddistancetocoverLima&Dill1990),aregenerallycontrolledbecauserecreatingrealisticenvironmentalconditionsinthelabo-ratoryisproblematic.Howthendoesananimalapplyexperiencesfromitsprevioushabitattothecontrivedhabitatinwhichitistested?Furthermore,becauseboldness…shynesstraitsarerelatively”exible(Sihetal.),movinganimalsfromthewildandsettlingthemintolaboratoryconditionsmaycausebehaviouralconver-gence(Wilsonetal.1993,1994).Howdowerelatetheresultsoflaboratorystudiestowhatisgoingonundernaturalconditions?Ourpreviousattempttostudyper-sonalitytraitsbybringingtheanimalsintothelaboratoryyieldedvariableresults,becausetheremighthavebeendifferentialresponsestohandlingstressandlifeincaptivity(Brown&Braithwaite2004).Thus,ifwearetounderstandfullytheroleoftheenvironmentinshapinganimalpersonalitiesitiscriticalthatwecarryoutsuchexperimentsundernaturalconditions.Wemeasuredtheboldnessof“shinsituinfourriverseachcontainingahigh-andlow-predationsite.Boldnesswasdeterminedbythetime“shtooktoemergefromablack,plasticchamberintotheirhomepools.Thisbehaviourisknowntobehighlycorrelatedwithothertestsofboldnessincludingthetendencytoleaveshoalmatesandapproachnovelobjects(C.Brown,unpublisheddata)andiseasytocarryoutunder“eldconditions.Weaddressedthreespeci“cquestionswiththisexper-iment.(1)Withrespecttotherelationbetweenboldnessandstandardlength,wepredictedtwopossiblescenar-ios.Thepredationpressurehypothesisposesthatbe-causejuvenilesaremorepronetopredationthanadults(reviewedinSogard1997)theyoughttobemorecautiouswhenemergingfromshelter.Hence,wepre-dictedapositiverelationbetweenstandardlengthandboldness.Sucharelationoughttovarybetweensitesofhigh-andlow-predationpressure(i.e.thereshouldbeaninteractionbetweenpredationpressureandsize).Analternativehypothesis,themetabolichypothesis,posesthatgiventhatsmaller“shhaverelativelyhighermeta-bolicrates,fewerenergyreservesandhigherdragcoef-“cients(Wootton1994;Krauseetal.1998;Skalski&Gilliam2002)theyshouldbecompelledtoemergefromcoversoonerthanlarger“shinordertobeginforaging.Thishypothesispredictstheexactoppositerelationbetweenstandardlengthandboldnessandthereoughttobenodifferenceintherelationin“shfromhigh-andlow-predationsites.Variationinboldnessinre-lationtostandardlengthisinterestinginitsownright,butitalsoprovidesapotentialconfoundif“shfromeachsitedifferinstandardlength.Wethereforecon-trolledforstandardlength.(2)Thesecondpredictionrelatedtothehypothesisthatpredator-sympatricandpredator-allopatricpopulationsmayvaryintheirbold-nessscores.Speci“cally,predator-sympatricpopulationswouldbemorecautiousandemergefromshelterlaterthan“shfrompredator-allopatricpopulations.(3)Final-ly,wepredictedthatmalesoughttobebolderthanfemalessinceinmanyspeciesoffreshwater“shfemalestendtobemorerisksensitivethanmales(guppies,PoeciliareticulataGrif“ths&Magurran1998;minnows,PhoxinusphoxinusMagurran1986;rainbow“sh,Brown2000Wequanti“edtheboldnessofeightpopulationsoftheBrachyraphisepiscopi(alsoreferredtoasepiscopifromfourindependentriversalongthePanamaCanalinthe“eldduringthedryseasonof2003.Withineachoftheriversweselectedtwosamplesites,onewithhighandonewithlowpredatorabundance(seeTable1forGPSreferences).Thelow-predationsiteswerealllocatedabovesubstantialwaterfallsthatlimitupstreammovementofmost“shspecies.Thehigh-predationsitesbelowthefallsareknowntocontainmanypotentialpredatoryspecies(table1inBrown&Braithwaite2004).Allthe“shinagivenpoolwerecapturedwithsmalldipnetsandstoredindark20-litrebuckets.The“shinthesestreamsareveryin-quisitiveandvirtuallyswimintothenetwhenitisplacedinapool.Thenettingprocedure,therefore,isnotparticu-larlystressful.AllB.episcopiover1.5cmweresexedandtheirstandardlengthrecordedbeforebeingindividuallyplacedinasinglecontainerfor15minimmediatelypriortotesting.Totestforboldness,wegentlypouredeach“shfromtheholdingcontainerintoasmalldarkplasticbox10cmand19cmhigh)withanopaquelidplacedontopwheretheyremainedfor2min.Onlyone“shwastestedatatime.Theboxwaspositionedinapproximately5cmofwaterontheedgeofthepoolfromwhichtheindividualswerecollected.UnderneaththeboxawhiteplasticsheetintheformofaDprovidedastrongbackgroundcontrasttothecolourofthe“shandthesubstrate.TheDservedbothtoaidourabilitytotrackthe“shsmovementandtorepresentapotentiallydangerousbackgroundforthe“shtocross.WeusedapermanentpenANIMALBEHAVIOUR, tomarkablackarconthewhiteDinaradiusofapproximately5cmfromthetrapdoor(Fig.1).After2minsettlingtime,thetrapdoor(5cmwideand10cmhigh)wasopenedatthefrontoftheboxandthe“shwerefreetoemerge,travelacrosstheDandintothepool.Werecordedboththetimetakentoemergefromtheboxafterweopenedthetrapdoorandthetimetocrosstheblackarc.Ifthe“shhadnotemergedfromtheboxafter6minweremovedthelid,thusreducingthevalueoftherefugeandencouragingthe“shtoemergefromthebox.Ifthe“shstillhadnotemergedafter8minweterminatedthetrialandallocatedthe“shaceilingvalueof480s.Theexactnumberof“shtestedateachsitevarieddependingontheabundanceofavailabletestsubjects,althoughwetestedatleast20femalesfromeverysite.Intotalwetested290“sh.Bothmalesandfemalesweretested;however,maleswerefairlyuncommonbecauseofafemale-biasedsexratioatallsites(Table1Wede“nedtheboldnessscoreasthetimetakenforthe“shssnouttoemergefromthebox.Thistestdiffersfundamentallyfromtheopen“eldarenasthathavebeenusedtoexamineboldnessinvariousspecies(Waren&Callaghan1976;Fujitaetal.1994)inthatthe“shinthisexperimentwerereleasedintoanenvironmentwithwhichtheyweretotallyfamiliar,andtheywerethereforeawareoftherisksinvolvedinemergingfromthechamber.Thetimetoemergefromthechamberishighlycorrelatedwithothertestsofboldnessincludingthetendencytoleaveshoalmatesandapproachnovelobjects(C.Brown,unpublisheddata).Hesitancywasde“nedasthetimethe“shtooktocrosstheblackarcminusthetimeittookforthe“shtoemergefromthebox.Inthefewinstanceswhere“shhadnotemergedbytheendof8minnohesitancyscorewascalculated.Toinvestigatetheaffectofage/sizeonthetendencytoemergefromthebox,weconductedregressionanalysesonbothhesitancyandboldnessscoresagainststandardlength.WeanalysedstandardlengthbyANOVAtode-terminewhetherthereweredifferencesbetweensitesandsexes.Sinceastrongrelationbetweenboldnessscoresandstandardlengthwasapparentandthemeansizeof“shdifferedbetweencollectionsites,wethenanalysedtheboldnessdatausingANCOVAwithstandardlengthasacovariate,andriver,predatorregime(highorlow)andsexas“xedeffects.ThehesitancydatashowednorelationwithsizeandwasanalysedwithanANOVA.Boththehesitancydataandtheboldnessdatawerelogtransformedpriortoanalysis.Bothmalesandfemalesfromupstreamsitestendedtobelargerthanthosefromdownstreamsites(ANOVA Table1.Thecoordinatesateachofthelocations,percentageofB.episcopiandpotentialpredatorsandsexratiosSiteCoordinates%B.episcopi%PredatorsSexratio(M:F)No.offemales(totalQuebradaJuanGradeHighpredation79W,9N27.866.71:4.81830(40)Lowpredation79W,9N87.93.81:9.11129(38)AquaSaludHighpredation79W,9N70.429.61:6.92331(44)Lowpredation79W,9N97.70.01:30.50030(33)RioLimboHighpredation79W,9N7.579.11:5.63625(35)Lowpredation79W,9N85.60.01:14.50022(26)RiverMachoHighpredation79W,9N24.072.71:4.66724(36)Lowpredation79W,9N38.10.01:2.18830(38) Figure1.Diagramofthestartboxwiththelidremoved.Fishwereplacedintheboxwithalidontop.After2minthetrapdoorshownatthefrontoftheboxwasraisedandtheshwerefreetoemerge.Thehesitancyline(seetext)isshownasaringaroundtheopeningofthebox.BROWNETAL.:ENVIRONMENT,BOLDNESSANDPOECILIIDS 0.026).Malesweresmallerthanfemales(1,2740.001),andtherewasnosigni“cantdifferenceinthelengthof“shbetweenrivers3,2740.403).Therewerenosigni“cantinteractionsbetweenanyofthethreefactors.Femalesover40mmandmalesover30mmwereneverfoundindownstreamlocations.Thedataalsosuggestthatup-streammalesmaturelaterthandownstreammalessincenoupstreammalessmallerthan20mmwerecollected.Ouranalysisrevealedastrongcorrelationbetweenstandardlengthandboldness(1,28835.25,0.001)withsmall“shemergingfromtheboxsoonerthanlarger“sh(Fig.2).Atestofparallelismshowedthatthisrelationdidnotvarybetweenregionsofhigh-andlow-predationpressure(1,2720.675),betweenrivers3,2700.138)orbetweensexes(0.198).FurtheranalysisoftheboldnessdatausingANCOVAshowedlargedifferencesbetweenthefourriversinthemeantimetoemerge(3,2730.001),aswellasdifferencesbetweenthehigh-andlow-predationsites(1,2730.001).Posthocanalysis(Fisherspartialleastsquaresdifference,PLSD)revealedthat“shfromAquaSaludemergedmorequicklythan“shfromtheotherthreerivers.Contrarytoourexpectations,“shfromhigh-predationsitesweresigni“-cantlybolderthanlow-predation“sh(Fig.3).Therewasnosigni“cantinteractionbetweenriverandpredatorregime(1,2730.149),norwasthereasignif-icantdifferencebetweenthesexes(1,2730.478).Posthocanalysis(FishersPLSD)revealedsigni“cantdifferencesbetweenhigh-andlow-predationareasinallrivers(0.003).ThehesitancyscoreshowednorelationwithsizeandanANOVAcon“rmeddifferencesbetweenrivers(3,2510.001)andbetweensites(1,2510.008)similartothoseintheboldnessdata,aswellasrevealingdifferencesbetweenthesexes.Malesshowedlesshesitationwhencrossingthewhite,plasticsemicirclethanfemales(1,2510.019).Therewerenosigni“cantinteractionsbetweenanyofthevariables.DISCUSSIONAnalysisoftheboldnessscoreshighlightstheroleofexposuretodifferentenvironmentsinshapingpopulationdifferencesintemperamenttraits.First,therewasastrongnegativerelationbetweenboldnessandstandardlengthaspredictedbythemetabolichypothesisand,importantly,thisrelationdidnotvarybetweenupstreamanddown-streamlocations(Fig.2).Thissuggeststhatmetabolicrequirementsofsmall“shcompelthemtoemergefromsheltersoonerthanlarger“shandbeginforagingregard-lessofthelevelofpredationpressure.Weobtainedsimilarresultsinthelaboratory(Brown&Braithwaite2004suggestingthattheresulthaslittletodowiththeenvironmentintowhichthe“shwerereleased.Theexistenceofsimilarrelationbetweenbodysizeandtimetoemergeinbothupstreamanddownstream“shinthisstudydismissesthealternativeexplanationforheavyselectionpressureagainstboldindividuals,sincethiswouldpresumablyoccuronlyunderhigh-predationre-gimes.Thus,thedifferenceinthepropensitytotakerisksandemergefromcoverearlyseemstodiminishasthe“shage.Thisresultindicatesthatvariationinmetabolismmaybeoneofthephysiologicalmechanismsshapingperson-alitytraits.Fishfromhigh-predationareaswereconsistentlybolderthanthosefromlow-predationsites.Itiscommonlythoughtthatsincepredator-sympatricpopulationsaremorelikelytorespondtopredatorsthanpredator-allopat-ricpopulations,theyshouldbemorecautiousinthefaceofpredationhazards(Seghers1974;Pitcher&Parrish).Ourresultscon”ictwiththistheoryanddemon-stratethatinfacttheoppositeistrue.Fishfromhigh-predationareaswerefarbolderthanthosefromlow-predationareas.Predator-sympatric“shstillneedtoforageandreproducejustastheirupstreamcounterpartsdo.Tocarryoutthesebehavioursintheshadowofconstantpredationthreattheymustbehaverelatively 0.5 1 1.5 2 2.53Log (time to emerge) 10 15 20 25 30 35 40 45 5055 g th (mm)Low predation High predation Figure2.Therelationbetweenstandardlengthandthetimetoemergefromthestartboxforshfromhigh-andlow-predationregions.Regressionlinesareshownforshfromlow-predation(topline)andhigh-predation(bottomline)areas. Aqua saludLimboQuebrada Figure3.ThemeanSEtimetoemergefromthestartboxforshfromhigh-andlow-predationsitesfromeachofthefourrivers.ANIMALBEHAVIOUR, boldly.Undersuchcircumstancesany“shthatremainsinhidingforextendedperiodsislikelytoshowreduced“tnessowingtolostforagingormatingopportunitiesLima&Bednekoff1999).Thus,inhigh-predationareasthereisprobablystrongselectionfavouringboldindivid-uals,whereasthisselectionforceislackinginlow-pre-dationareas.Wilsonetal.(1993)foundsubstantialdifferencesinthebehaviourofboldandshysun“sh,Lepomisgibbosus.Bold“shadjustedtolifeincaptivity5daysearlierthanshysun“sh.Inaddition,bold“shatemorecopepods,haddifferentparasitefaunaandtendedtoforagefurtherfromshoalmatesthanshy“sh.ItisclearfromWilsonetal.sstudyandourownthatvariationinpersonalitytraitswillhavesubstantialin”uencesonindividual“tnessinwildpopulations.Itmaybearguedthatbecausethe“shfromallpopulationswerereleasedintodifferentenvironments(intohigh-andlow-predationareas)theboldnessresultsmaybein”uencedinsomemanner.However,“shreleasedintolow-predationregionsoughttoemergefromcoversoonerthanthosereleasedintohigh-predationareas,whichistheoppositepatterntotheresultsreportedhere.Thedifferencesbetweenhigh-andlow-predation“shmayalsobetheresultofphylogeny(Johnson2001);however,itishighlyunlikelythatall“shfromhigh-predationareasaremorecloselyrelatedtooneanotherthantheyaretotheircounterpartsjustabovethewaterfallswithinthesameriver.Ourdataarefurthersupportedbylaboratorystudyinwhichguppiesandkilli“sh,Rivulushartiihigh-predationregionsweremoretenacious(de“nedasthefeedingrateinthepresenceofapredatordividedbythefeedingrateintheabsenceofapredator)than“shfromlow-predationregions(Fraser&Gilliam1987Similarly,Iberianrocklizards,Lacertamonticola,exposedtosimulatedpredationpressurealsodeveloppersonalitydifferencesthataremanifestedindifferentialpropensitytoemergefromcover(Lopezetal.2005).Weconclude,therefore,thatourresultsshowthatanimalstempera-mentsbecomeprimedforoptimalresponseswithinthecontextoftheprevailingecologicalconditionsandani-malsappeartobecomeincreasinglyshyastheygrowwithintheseconstraints.Whetherdifferentiationbetweenupstreamanddownstreamsitesoccursoveranevolution-aryorontogenetictimeframeremainstobeseenandrepresentsanareaforfutureresearch.Onealternativeexplanationforthedifferencesinthetimetoemergefromhidingisthatmotivationdiffersbetweenthetwopopulations.Itcouldbethat“shfromhigh-predationareasaresimplymorehungrybecausetheyhavelimitedaccesstopreyitems,owingtothepresenceofpredatorsandheterospeci“ccompetitors.Futureinvestigationscouldexaminethecorrelationbe-tweenbodyconditionandboldnessinthe“eldor,alternatively,onecouldprovidesupplementaryfoodbeforetestingthe“sh.Furtherresearchcouldalsobedirectedatidentifyingtheunderlyingphysiologicalorneuroendocrinologicalmechanismsthataredrivingthedivergenceinpersonalitytraits.Forexample,individualsmaydifferinhowtheyrespondtomildstressors,whichmayinturnaffecttheirpropensitytoexplorenovelenvironments(Brown&Braithwaite2005Whiletherewerenosigni“cantdifferencesbetweenthesexesintheboldnessscore,malesstilltendedtobefastertoemergethanfemales,aswepredicted.However,thereweredifferencesbetweenthesexesinthedegreeofhesitationasthe“shemergedfromtheshelterandenteredthepool.FemaleshesitatedforlongerbeforecrossingthewhiteDthanmales.Itislikelythatassoonasthemalesrealizedtheycouldgetbackintothestreamandchasefemalestheydidso,despitethedangerofcrossingthehigh-contrastbackground.Therewasnointeractionwithpredationregime,suggestingthatthesingle-mindednessofthemalesbehaviourdidnotchangeinthepresenceorabsenceofpredators.Similarobservationshavebeenmadeinjuvenilebrowntrout,Salmotrutta,wheremaleswerelesslikelytorespondtorepeatedpredatorattackthanfemalesandweremorethantwiceaslikelytoinstigateagonisticinteractions(etal.2001).Thesedifferencesinboldnessbetweenmalesandfemalesarelikelytohaveahormonalbasis.Muchofthemalemortalityintheseriversprobablyresultsfromfemalesdefendingthemselvesfromharassment,ratherthandirectlyfrompredators.Despitethedangersinvolvedwithcourtingfemales(theaveragefemaleweighs2.5timestheaveragemale),boldmalesarerewardedwithhigherratesofinseminationwithincreasingmatingattempts(Evansetal.2003).ThisnotionthatcourtshipisdangerousformalesalsotiesinwithourobservationofdecreasingnumbersofmalesastherelativedensitiesofB.episcopiincreasedandmayalsoexplainwhymalesmaturelaterandatlargersizesinlow-predationlocations.ItisnowwelldocumentedthatdifferencesinlifehistoryprioritiesresultindifferencesinthebehaviourofthesexesPoeciliareticulataReznicketal.2001Brachyraphisrhab-Reznicketal.1993;Johnson2001B.episcopiJennions&Telford2002).Malesmaximizetheir“tnessbyinseminatingasmanyfemalesaspossible,arepreoccupiedwithchasingfemalesandappeartoliveshort,dangerouslives.Females,ontheotherhand,maximize“tnessbyincreasinglongevityandspendmostoftheirtimeforag-ing.Thereisnobetterwaytoelucidatethisdifferenceinlifestylethanbyexaminingthebehaviourofpoeciliidsthatareconfrontedwithapredator.Femalesstopforaging,formshoalsand“xateonthepredator,whereasmalesmakethemostofthedistractionbyincreasingattemptsatsneakymating(Reznick&Endler1981;Evansetal.2003Finally,wefoundsigni“cantdifferencesinthebehav-iourofthe“shbetweenrivers.AquaSalud“shwerebolderthan“shfromtheotherthreerivers(Fig.3).Thisriveristhemostgeographicallyisolatedofthefourriverstestedand,ironically,hastheleastdifferenceinpredationpressurebetweenitstwocollectingsites(Table1).Al-thoughweattemptedtochooseriversandsiteswithinriversthatwereassimilaraspossibleintermsofthephysicalandsocialenvironments,clearlynoriverisidenticaltoanother.ThisenvironmentalvariationfromcatchmenttocatchmentevidentlyhasastrongimpactonthedevelopmentofboldnesstraitsthatisofasimilarmagnitudetothatexplainedbyvariationinpredationInconclusion,ourresultssuggestthatthetempera-mentsof“shvaryconsiderablywithage,sexandthelevelBROWNETAL.:ENVIRONMENT,BOLDNESSANDPOECILIIDS ofpredationpressureintheenvironmentindicatingthattemperamentsmaybemorelabilethanpreviouslysus-pected.Onemajordif“cultyremains,andthatisdistin-guishingbetweenshort-termbehaviouralplasticityandlong-term,stablepsychologicalstatesandbehaviouralsyndromes(Sihetal.2004).Thisisparticularlythecaseinshort-livedanimalsandmaybethesinglegreatestobstaclewhenattemptingtomarrypsychologywiththebehaviouralecologyliterature.Psychologistsmaypointoutthatboldness…shynessandotherpersonalitytraitsarede“nedaslong-term,stablepsychologicalstates.Never-theless,thepropensitytotakerisksandotherbehaviouraltraitsareknowntobeheavilyin”uencedbyinternalstates,suchashunger,aswellasdemographicvariablesincludingageandsex,allofwhichmaybein”uencedbybodysize(Wilsonetal.1994;Krauseetal.1998).Thisisalsothecasewithhumanstudieswheresex,ethnicity,ageandhealthstatusareallimportantvariableswhenex-plainingpersonalitydifferences(vanGestel&vanBroeck-hoven2003).Thetwonotionsmaynotnecessarilybeentirelyincompatible,however,aslongasthedifferencesbetweenindividualsaremaintainedinavarietyofcon-texts,whilestillallowingforadegreeofshort-termbehaviouralplasticityinresponsetointernalorexternalenvironmental”uctuations(Sihetal.2004).Thestudyofanimalpersonalitiesandtheirpotentialevolutionaryconsequencesisstillinitsearlystages.Westillknowlittleabouttheproximateandultimatecausesofinterindivid-ualbehaviouralvarianceinpopulationsfromanecolog-icalandevolutionaryperspective.Disentanglingtherelativerolesofheritability,ecologicalandsocialforcesthatcreateandmaintainpersonalitytraits,aswellasidentifyingtheunderlyingphysiologicalandneuroendo-crinologicalmechanismsresponsibleforthesetraits,remainsamajorchallenge.AcknowledgmentsOurworkwasfundedbyNERCresearchgrant#(NER/A/S/01/00608)andwasconductedunderANAMpermit#SEX/A-88-03.WethanktheSmithsonianTropicalResearchInstitute,inparticularBiffBirmingham,fortheircontinu-ingcooperationandsupport.Wealsothanktheanony-mousrefereesfortheirhelpfulcomments.ReferencesBrown,C.2000.Thebehaviouralecologyofpredatoravoidanceinrainbowsh(Melanotaeniaspp.).Ph.D.thesis,UniversityofQueensland.http://geocities.com/culumbrown/thesisBrown,C.&Braithwaite,V.A.2004.Sizematters:atestofboldnessineightpopulationsofbishop,BrachyraphisepiscopiAnimalBehaviour,1325–1329.doi:10.1016/j.anbehav.2004.04.004.Brown,C.&Braithwaite,V.A.2005.EffectsofpredationpressureonthecognitiveabilityofthepoeciliidBrachyraphisepiscopiBehavioralEcology,482–497.doi:10.1093/beheco/ari016.Budaev,S.V.1997.‘Personality’intheguppy(Poeciliareticulataacorrelationalstudyofexploratorybehaviorandsocialtendency.JournalofComparativePsychology,399–411.Endler,J.A.1995.Multipletraitcoevolutionandenvironmentalgradientsinguppies.TrendsinEcologyandEvolution,22–29.Evans,J.P.,Pilastro,A.&Ramnarine,I.W.2003.Spermtransferthroughforcedmatingsanditsevolutionaryimplicationsinnaturalguppy(Poeciliareticulata)populations.BiologicalJournaloftheLinnaeanSociety,605–612.doi:10.1046/j.0024-4066.2002.00193.x.Fraser,D.F.&Gilliam,J.F.1987.Feedingunderpredationhazard,responseoftheguppyandHart’srivulusfromsiteswithcontrastingpredationhazard.BehavioralEcologyandSociobiology,203–209.Fraser,D.F.&Huntingford,F.A.1986.Feedingandavoidingpredationhazard:thebehaviouralresponseofprey.Fujita,O.,Annen,Y.&Kitaoka,A.1994.Tsukubahigh-emotionalandlow-emotionalstrainsofrats(Rattusnorvegicus):anoverview.BehaviorGenetics,389–415.vanGestel,S.&vanBroeckhoven,C.2003.Geneticsofpersonality:arewemakingprogress?MolecularPsychiatry840–852.doi:10.1038/sj.mp.4001367.Goddard,M.E.&Bilharz,R.G.1985.Amultivariateanalysisofthegeneticsoffearfulnessinpotentialguidedogs.BehaviorGenetics,69–89.Godin,J.-G.J.&Smith,S.A.1988.Atnesscostofforaginginthe,69–71.Gosling,S.2001.Frommicetomen:whatwecanlearnaboutpersonalityfromanimalresearch.PsychologicalBulletinGrifÞths,S.W.&Magurran,A.E.1998.SexandschoolingbehaviourintheTrinidadianguppy.AnimalBehaviour,689–Jennions,M.D.&Telford,S.R.2002.Life-historyphenotypesinpopulationsofBrachyrhaphisepiscopi(Poeciliidae)withdifferentpredatorcommunities.,44–50.doi:10.1007/s00442-002-0942-4.Johnson,J.B.2001.Adaptivelife-historyevolutioninthelivebearingBrachyrhaphisrhabdophora:geneticbasisforparalleldi-vergenceinageandsizeatmaturityandatestofpredator-inducedplasticity.,1486–1491.Johnsson,J.I.,Sernland,E.&Blixt,M.2001.Sex-specicaggressionandantipredatorbehaviourinyoungbrowntrout.,587–599.Kamil,A.C.&Balda,R.P.1990.Spatialmemoryinseed-cachingPsychologyofLearningandMotivation,1–25.Krause,J.,Loader,S.P.,McDermott,J.&Ruxton,G.D.Refugeusebyshasafunctionofbodylength-relatedmetabolicexpenditureandpredationrisks.ProceedingsoftheRoyalSocietyofLondon,SeriesB,2373–2379.doi:10.1098/rspb.1998.0586.Lima,S.L.1998.Stressanddecisionmakingundertheriskofpredation:recentdevelopmentsfrombehavioural,reproductiveandecologicalperspectives.AdvancesintheStudyofAnimal,215–290.Lima,S.L.&Bednekoff,P.A.1999.Temporalvariationindangerdrivesantipredatorbehavior:thepredationriskallocationhypoth-AmericanNaturalist,649–659.Lima,S.L.&Dill,L.M.1990.Behavioraldecisionsmadeundertheriskofpredation,areviewandprospectus.CanadianJournalof,619–640.Lopez,P.,Hawlena,D.,Polo,V.,Amo,L.&Martin,J.Sourcesofindividualshy–boldvariationsinantipredatorbehaviourofmaleIberianrocklizards.AnimalBehaviour,1–9.doi:10.1016/j.anbehav.2004.05.010.Magurran,A.E.1986.Predatorinspectionbehaviourinminnowshoals:differencesbetweenpopulationsandindividuals.ioralEcologyandSociobiology,267–273.ANIMALBEHAVIOUR, Pitcher,T.J.&Parrish,J.1993.Thefunctionofshoalingbehaviour.In:TheBehaviourofTeleostFishes.2ndedn.(Ed.byT.J.Pitcher),pp.363–439.London:Chapman&Hall.Pitcher,T.J.,Lang,S.H.&Turner,J.A.1988.Ariskbalancingtradeoffbetweenforagingrewardsandpredationhazardinashoalingsh.BehavioralEcologyandSociobiology,225–Reznick,D.&Endler,J.A.1981.TheimpactofpredationonlifehistoryevolutioninTrinidadianguppies(Poeciliareticulata,160–177.Reznick,D.,Meyer,A.&Frear,D.1993.Life-historyofBrachyraphisrhabdophora(Pisces:Poeciliidae).,103–111.Reznick,D.,Butler,M.J.&Rodd,H.2001.Life-historyevolutioninguppies.VII.Thecomparativeecologyofhigh-andlow-predationenvironments.AmericanNaturalist,126–140.Seghers,B.1974.Schoolingbehaviourintheguppy(),anevolutionaryresponsetopredation.486–489.Sih,A.1980.Optimalbehaviour:canforagersbalancetwoconictingdemands?,1041–1043.Sih,A.,Bell,A.&Johnson,J.C.2004.Behavioralsyndromes:anecologicalandevolutionaryoverview.TrendsinEcologyand,372–378.doi:10.1016/j.tree.2004.04.009.Skalski,G.T.&Gilliam,J.F.2002.Feedingunderpredationhazard:testingmodelsofadaptivebehaviorwithstreamsh.,158–172.doi:10.1086/341012.Sogard,S.M.1997.Sizeselectivemortalityinthejuvenilestageofteleostshes:areview.BulletinofMarineScience,1129–1157.Waren,E.W.&Callaghan,S.1976.Individualdifferencesinresponsetoanopeneldtestbytheguppy,PoeciliareticulataJournalofFishBiology,499–513.Werner,E.E.&Anholt,B.R.1993.Ecologicaltrade-offbetweengrowthandmortalityratesmediatedbyforagingactivity.AmericanNaturalist,242–272.Wilson,D.S.,Coleman,K.,Clark,A.B.&Biederman,L.Shy–boldcontinuuminpumpkinseedsunsh(Lepomisgibbosusanecologicalstudyofapsychologicaltrait.JournalofComparative,250–260.Wilson,D.S.,Clark,A.B.,Coleman,K.&Dearstyne,T.Shynessandboldnessinhumansandotheranimals.TrendsinEcologyandEvolution,442–446.doi:10.1016/0169-5347(94)90134-1.Wootton,R.J.1994.Energyallocationinthethreespinestickleback.TheEvolutionaryBiologyoftheThree-spinedStickleback(Ed.byA.Belt&S.A.Foster),pp.250–260.Oxford:OxfordUniversityBROWNETAL.:ENVIRONMENT,BOLDNESSANDPOECILIIDS