Sex-specific syndromes and behavioral sexual dimorphism

1

Sex-specificbehavioralsyndromesallowtheindependentevolutionofbehavioraldimorphism

RaphaelRoyauté1,2,4AnnHedrick3

NedA.Dochtermann2

1MovementEcologyGroup,SenckenbergBiodiversityandClimateResearchCentre(SBiK-F),Frankfurt,Germany(currentaddress)2DepartmentofBiologicalSciences;NorthDakotaStateUniversity,Fargo,USA3DepartmentofNeurobiology,Physiology,andBehavior;UniversityofCalifornia,Davis,USA4correspondingauthor:[email protected]:Sex-specificsyndromesandbehavioralsexualdimorphismKeywords:Behavioralsyndromes,evolutionaryconstraints,Gmatrix,personality,behavioralecology,sexualdimorphism,intralocussexualconflictsAuthorContributions:Allauthorsconceivedtheprojectandsupervisedthegatheringofdata.R.Rconductedbehavioraltrialsandanalyzedthedata.R.R.wrotethefirstdraftofthemanuscriptandallauthorscontributedtosubsequentversions.Conflictsofinterests:TheauthorsdeclarenoconflictofinterestsDataAccessibilityStatement:Alldata,codeandanalysesareavailableat:https://osf.io/pnug5/Acknowledgements:WethankMonicaBerdal,KatelynCannon,JeremyDalos,SarahFelde,1

BradyKlock,IshanJoshi,HannahLambert,JennaLaCoursiereandAlondraNeunsingerfor2

assistanceinconductingbehavioraltrialsandinrearingandcareofthecricketsand3

MartoriFarms,DavidLightfoot,ScottBundy,NicoFranz,SangmiLee,CameronJones,4

KennyChapin,TiEriksson,MerandaFeagins,CharlotteMulloney,MelodyMartinez,Allyson5

Richins,MaurielRodriguez,HelenVesselsandDavidWikmanforassistanceincollecting6

thecrickets.WethankforJ.SztepanaczandA.J.Wilsonfortheiradviceonstatistical7

analysesandK.Laskowskiforcommentsonanearlydraftofthismanuscript.Thiswork8

wassupportedbyUSNSFIOSgrants1557951and1558069toN.A.D.andA.H.respectively. 9

2

ABSTRACT10

Whenselectiondiffersbysex,thecapacityforsexestoreachoptimalphenotypescanbe11

constrainedbythesharedgenomeofmalesandfemales.Becausephenotypictraitsare12

oftencorrelated,thisdifferenceextendsacrossmultipletraitsandunderlyinggenetic13

correlationscanfurtherconstrainevolutionaryresponses.Behaviorsarefrequently14

correlatedasbehavioralsyndromes,andthesecorrelationsoftenhaveageneticbasis.15

However,whethercross-sexandacrossbehaviorcorrelationsleadtoconstrainedevolution16

remainsunknown.Here,weshowthataboldness-activitysyndromeisstronglysex-17

specificatthegeneticlevelinthewesternfieldcricket(Gryllusinteger)andthatemergence18

fromashelterisgeneticallyindependentbetweenmalesandfemales.However,male19

activityisstronglyrelatedtofemaleshelteremergence,creatingthepotentialforbiased20

responsestoselection.Ourresultsshowthatthesex-specificgeneticarchitectureof21

behavioralsyndromescanshapetheevolutionofbehavioralphenotypes.22

23

3

INTRODUCTION24

Malesandfemalessharethesamegenomebutoftenrelyondifferentstrategiestoincrease25

fitness.Thiscanleadtointralocussexualconflictswhereselectionactingonasharedtrait26

displacesonesexfromitsoptimum(HedrickandTemeles1989;Bondurianskyand27

Chenoweth2009).Thissexualconflictisfurthercomplicatedbytheconstrainingpotential28

ofgeneticcorrelations(Lande1980;WalshandBlows2009).Forexample,behaviorsare29

oftencorrelatedascomponentsofabehavioralsyndrome(Sihetal.2004)andthese30

behavioralsyndromeshavethepotentialtoalterevolutionaryoutcomes(Dochtermann31

andDingemanse2013).32

Ingeneral,anysustainedselectionfavoringdifferentoptimabetweensexes,i.e.33

sexuallydiscordantselection,willeventuallyresolvethesexualconflictandattenuate34

geneticconstraintsovertime,thusallowingsexestoevolveindependently(Delphetal.35

2011).Thisconstraintcanbeformallymeasuredasthecross-sexcorrelationcoefficientrmf,36

withvalues<1increasingtherapidityatwhichsexualdimorphismcanevolve.Cross-sex37

correlationsaregenerallylargebuttendtodecreaseinspecieswithstrongersexual38

dimorphism(Poissantetal.2010).Additionalconstraintscanemergewhengeneralizingto39

multiplephenotypesexpressedinmalesandfemalesandtheirinteractions(Lande1980).40

Bydecomposingtheadditivegeneticcovariancematrixintoitssex-specific(Gm,Gf)and41

cross-sexsub-matrices(B),onecanestimateifgeneticcorrelationsacrosssexesandtraits42

createconstrainedevolutionaryoutcomes(Fig.1).43

Thisdecompositionapproachhasbeenusedformorphologicaltraits,revealingthat44

cross-sexcovariances(i.e.theBmatrix,(Lande1980))canprofoundlyaltertheevolution45

ofsexualdimorphism(Gosdenetal.2012;Bergeretal.2014;GosdenandChenoweth46

2014),especiallywhenselectionfavorsdifferentaveragetraitvaluesforeachsex(Long47

andRice2007;Bergeretal.2014).However,theGm,Gf,andBhaverarelybeenestimated48

forbehaviors.Howmuchthegeneticarchitectureofbehavioralsyndromescanaffect49

evolutionaryoutcomesunderintralocussexualconflictremainsunclear.Thisisan50

importantomissionbecausebehaviorfrequentlyimpactsbothsurvivalandfitness(Moiron51

etal.2020)andsexualdimorphismiscommonforbehaviors(Blanckenhorn2005;Aragón52

4

2011;Mainwaringetal.2011;Kokrasetal.2012).Despitemanyconceptualargumentsfor53

whybehavioralsyndromesshoulddifferbysexes(Schuettetal.2010;Hämäläinenetal.54

2018;Immonenetal.2018),sex-differencesintheexpressionofbehavioralcorrelations55

arefrequentlyignoredinpractice.Asaresult,thedegreetowhichcross-sexgenetic56

correlationsmightbeshapedbysexualconflictforbehavioralphenotypesremains57

unknownandtheimportanceofthesecorrelationsfortheevolutionofbehaviorsis58

similarlyunknown. 59

5

60

6

Figure1.Thegeneticarchitectureofcross-sexcorrelationscanhavenon-intuitive61 consequencesonevolutionaryresponsesandsexualdimorphism.A)Highvaluesofcross-62 sexcorrelationsforthesametraits(here,bodysize)indicatethatmaleandfemaleaverages63 arepositivelycorrelatedamongfamilies.B)Thecross-sexcovariancematrix(Gmf)allows64 tocomparepatternsofgeneticcovarianceswithinsexes(herebetweenbodymassand65 bodysize)aswellasunderstandthestrengthofcross-traitcorrelationswithinandamong66 traits.C)Cross-sexcross-traitcorrelation(rmf)canproducenon-intuitiveresponsesto67 selectivepressures.Hereweshow3scenariosindicativeofthevarietyofresponsesto68 selection(Δz)thatcanoccurdependingonthedirectionofselection(β)andthemagnitude69 andorientationofrmf.Theellipsesrepresentthebivariate(co)variationinfemurlength70 andmass,withthepopulationaverageatthevertexofthe(thinsolid)linesindicatingthe71 directionsinbivariatespacewiththemostvariation.Dashedlinesrepresentthedirection72 selection(β)ispushingapopulation’saverageandthesolidthickarrowsshowthe73 directionandmagnitude(arrowtip)ofselectionresponses.InbothscenariosC.1andC.2,74 selectionisconcordant—i.e.actingthesame—betweensexesandmalesandfemaleshave75 thesamefitnessoptimum.InscenarioC.3,selectionissexuallydiscordantandfavors76 increasedsexualdimorphism.ScenarioC.1showsthatstrongcross-sexcrosstrait77 correlationscanbiassexesawayfromtheiroptimum.InscenarioC.2,nullrmfresultsin78 independenttrajectoriesbysexes.InscenarioC.3,theeffectoffathersontheirdaughters’79 phenotypeisoppositeofthatofmother’sontheirsons’,resultinginasymmetricrmfand80 biasedresponsestoselectioninmales. 81

7

Intralocussexualconflictshouldbeparticularlystronginthefieldcrickets82

(Gryllidae).Inmanyfieldcrickets,malesandfemalesdifferintheirbehaviorsand83

reproductivestrategies(HedrickandKortet2012).Forexample,femalesexploretheir84

environmentstosamplemateswhilemalesremainatburrowsfromwhichtheysignal85

(FrenchandCade1987).Further,aboldness-activitysyndromehasaconservedgenetic86

basisinatleastonespecies,theWesternstutter-trillingfieldcricket(Gryllusinteger),87

whichseemstohaveconstrainedbehavioraldivergence(Royautéetal.2020).Whetherthis88

syndromeissex-specificiscurrentlyunknownbutthecombinationofconserved89

syndromestogetherwithalackofsex-specificitywouldfurtherconstraintheabilityof90

populationstorespondtolocalselectivepressures.91

Hereweusedbehavioralmeasurementsofover960Gryllusintegerindividualsto92

estimatetheinfluenceofcross-sexgeneticcorrelationsonevolutionaryresponses.We93

predictedthatcricketswouldexhibitdimorphisminaverageactivitygiventhemate-94

samplingbehavioroffemalesandthatgeneticvarianceforthistraitwouldbelowerfor95

femalecricketsduetoselectionfavoringincreasedtraveling.Wealsopredictedsex-96

specificityinthebehavioralsyndrome,withastrongerpositivegeneticcorrelation97

betweenantipredatorresponseandactivityinfemales.Sinceexitfromrefugiaandactivity98

shouldhaveastrongerinfluenceonfemalefitness,wealsoexpectedastrongercorrelation99

betweenshelteremergencewithactivityandantipredatorresponseinfemales.Although100

therearefewestimatesforcross-sexcorrelationsofbehaviors,thoseavailablesuggest101

behaviorsareundersimilarconstraintsasotherphenotypes(Poissantetal.2010).We102

thereforeexpectedthatcross-sexgeneticcorrelations(rmf)wouldnotdepartsignificantly103

from1.Wetestedthesepredictionsbyestimatingthecross-sexcovariancematrix,i.e.Gmf,104

decomposedintoitssex-specificandcross-sexsub-matrices(Gm,GfandBsubmatrices,105

Figure1).Wecomparedthestrengthofcovariancesamongsexesandtheeffectsofthese106

covariancesonresponsestoselectionusingrandomskeweranalysis.107

108

8

METHODS109

Wecollectedadultfemalecricketsfromfourpopulationsthroughoutthesouthwesternand110

westernUS:Socorro,NM;LasCruces,NM;Aguila,AZ;andDunnigan,CA(Figure2)during111

thesummerof2017andhousedthemandinourlaboratoryfacilitiesatNorthDakotaState112

University.Femaleswerehousedindividuallyin0.71Lcontainersandprovidedwithad113

libitumfood(PurinaChickStarter)andwater(waterwasprovidedinglassvialscapped114

withcotton).Eachcricketwasalsoprovidedwithasmallpieceofcardboardeggcartonfor115

shelter.Thecrickethousingroomwasmaintained~27Cona12:12dark:lightcycle116

reversedsuchthattheroomwasdarkduringdaytimehours.Werunoffspringofthis117

parentalgenerationthroughmultiplebehavioraltrialsbeforematingindividualsatrandom118

withineachpopulation.Werepeatedthisprocessfortwoadditionalgenerations.119

Behavioraltesting120

Latencytoemergefromshelter121

Gryllidcricketsusesmallburrowsandnaturalcracksasrefugestowhichtheyretreatwhen122

underthreat.Thetimetakentoemergefromashelterafterdisturbancecanthereforebe123

consideredaproxyforrisk-takingbehavioror“boldness”(KortetandHedrick2007).We124

transferredindividualsfromtheirhomecontainersintosmallartificialburrows(40cm3)125

placedwithina34.6×21cmarena.Weleftthecricketstorestfortwominutesafterwhich126

weremovedthecapfromtheburrowandletindividualsemerge.Wethenrecordedhow127

longittookforanindividualtoemerge(inseconds)foruptosixminutesandthirty128

seconds.Individualsthatdidnotemergeweregivenamaximumlatencyof390seconds.129

Openfieldexploratorybehavior130

Weusedopenfieldteststomeasureactivityandexploratorypropensityina30×30cm131

plexiglassarena.Thesetestsareclassicbehavioralassayacrosstaxa(WalshandCummins132

1976)andcanrevealstrongamong-individualdifferencesinexplorationpatterns,133

includingincrickets(Royautéetal.2015,2019;RoyautéandDochtermann2017).134

Individualsthatmovethroughmoreofthearenaareconsideredmorethoroughexplorers135

(Réaleetal.2007).Weintroducedindividualsintothearenaandleftthemtorestundera136

9

smallcontainerfor30seconds.Attheendofthis30seconds,weremovedthecontainer137

andthecricketwasallowedtoexplorethearenafor3minutesand40seconds.Thearena138

wascleanedwithisopropylalcoholbetweentrialstoremoveanychemosensorycuesfrom139

thearena.WeusedEthovisionXTtorecordthetotaldistancetheindividualmovedduring140

thetrial(cm).141

Responsetocuesofpredatorpresence142

Wemeasuredtheresponsetocuesofpredatorpresence,abehavioralassaycommonly143

usedinGryllidspeciestodetermineantipredatorresponse(RoyautéandDochtermann144

2017;Royautéetal.2019).Specifically,individualswereintroducedintoa15cmdiameter145

circulararena(7.5cmheight),thefloorofwhichwascoveredwithdryfilterpaperthathad146

beensoakedwithdilutedexcretafromleopardgeckos(Eublepharismacularius).Crickets147

respondtoexposuretoleopardgeckocuesbyincreasingactivityandindividualswith148

higherdistancemovedareconsideredmoreresponsivetothecue(Royautéand149

Dochtermann2017;Royautéetal.2019).Weintroducedcricketstoaportionofthearena150

withoutpredatorcueandleftthemtorestunderasmallshelterfor30seconds.Wethen151

removedtheshelterandallowedtheindividualallowedtofreelymovethroughoutthe152

arenafor3minutesand40seconds.WethenusedEthovisionXTtorecordthetotal153

distanceanindividualmovedduringthetrial(cm).154

155

Statisticalanalyses156

AllanalyseswereperformedusingRversion4.0.3(Rcitation)usingtheMCMCglmm157

package (Hadfield2010).158

Estimationofcross-sexgeneticcovariances(Gmf)159

Weusedamulti-responsemixedeffectanimalmodels(Kruuk2004;Wilsonetal.2010)to160

estimategeneticvariancesandcovariances(i.e.theGmfmatrix).Weincludedtheeffectsof161

temperature,dayandtimeoftestinginthebehavioralarenaroomalongwithsex,life-stage162

andmassoftheindividualasfixedeffects.Weusedtheindividualrelatednessmatrix163

10

(basedontheknownpedigree)asarandomeffectandthefollowingbehavioraltraitswere164

includedasresponsevariables:(i)thelatencythatanindividualemergedfromtheshelter165

duringthetrial(modeledascensoredGaussian),(ii)thedistancemovedduringtheopen166

fieldtrial(Gaussian),(iii)thedistanceanindividualmovedduringthepredatorcue167

responsetrial(Gaussian).Toestimatebothsex-specificandcross-sexcovariances,we168

treatedthebehaviorofeachsexasaseparatetrait-resultingintheestimationofa6×6169

covariancematrix.Weranthemulti-responsemodelwithanMCMCchainof4.8×106170

iterations,withan800,000burn-inperiodandathinningintervalof4,000andweuseda171

parameterexpandedpriorthatwasminimallyinformativeforbothvariancesand172

covariances.Allvariancesandcovarianceswereestimatedattheadditivegeneticleveland173

onthelatentscale.174

175

Estimationofbehavioraldimorphism176

Wetestedfortheexistenceofsexualdimorphisminbehavioralexpressionbycomparing177

linearcoefficientforthesexfixedeffectincludedinourmultivariateanimalmodelandbase178

ourstatisticalinferenceontheirBayesianprobability(Pmcmc).Thismetricvariesbetween179

0.5and1andindicatestheprobabilityofasignificantdifferencebasedonthenumberof180

posteriorestimatesoverlappingwith0.Pmcmcvalues>0.95werejudgedassignificant.181

182

Comparisonofsex-specificcovariances(GmandGfmatrices)183

Weusedatwo-stepapproachtocomparetheintensityofthedifferenceingenetic184

covariancesamongsexes.First,wecalculatedthedifferenceincovariancebetweenmales185

andfemales(ΔCOVA)foreachpairofbehaviorsandtheirassociatedBayesianprobabilities186

(Pmcmc).Next,wetestedwhethersex-specificcovariancesweresimilarlyoriented.Todo187

so,wecalculatedthevectorcorrelation(r°)betweenaxescontainingthehighestamountof188

geneticvariationusingeigenvaluedecomposition.Wethenestimatedwhetherthe189

resultingvectorcorrelationsamongeigenvectorsofGdifferedsubstantiallyfrom190

expectationsof0(noalignmentofgeneticvariationamongsexes)and1(perfect191

11

alignment).Becausevectorcorrelationsareboundedby0and1,weestimatedtheRegion192

ofPosteriorEquivalence(ROPE)whichwedefinedastheintervals[0.0;0.1]and[0.9;1.0].193

EstimatesfallingwithintheseROPEregionsarejudged“practicallyequivalent”tovector194

correlationsof0and1respectively.Weconvertedtheproportionofestimatesfalling195

outsidetheseROPEregionsintoPmcmcvaluestoinfersignificance,withPmcmc>0.95196

indicatingsignificantdeparturefromthesenullhypotheses.197

198

Estimatingthegeneticconstraintimposedbythecross-sexcovariancematrix(B)199

Wefirsttestedwhethercross-sexcovarianceswithintraitsdifferedsignificantlyfrom0and200

1.Cross-sexcovariancesarerepresentedonthediagonalelementsoftheBmatrix.A201

covarianceof0indicatecompletegeneticindependenceamongsexes,whichwe202

determinedusingBayesianprobabilities(Pmcmc).Totestforadeparturefromcomplete203

geneticcouplingamongsexes,weconvertedthecovariancestocorrelationcoefficients204

(rmf)andusedtheROPEtestdescribedabove.Wetheninvestigatedwhethercross-sex205

cross-traitscovariances–representedontheoff-diagonalelementsofB–differed206

significantlyfromoneanotherbycalculatingtheposteriordifferenceincovarianceas207

described(ΔCOVA).208

Next,wecomparedtheevolutionarytrajectoriesofsexeswithascenarioweresexes209

evolvedindependentlywithoneweresexeswerefullyconstrained(Coxetal.2017).Todo210

so,wesimulated500selectiongradientsbasedonamultivariatenormaldistributionthat211

wescaledtounitlength.Wegeneratedconcordantselectiongradientsbyassigningthe212

samevaluestomaleandfemalegradients(βf=βm)anddiscordantselectionbysettingβf=-213

βm.Weappliedeachselectiongradienttoall1,000posteriorcovariancematricesofGmf214

estimatedfromourmultivariateanimalmodel,thusensuringwetooktheuncertaintyin215

estimatesforward.Wethencalculatedtheresultingresponsetoselection(𝛥𝑧̅)byapplying216

themultivariatebreeder’sequationforcross-sexcovariance(Lande1980):217

$𝛥𝑧!̅𝛥𝑧"̅

% = #$$𝐆𝐟 𝐁′𝐁 𝐆𝐦

% $𝛽!𝛽"% (equation1)218

12

Wecontrastedtheseresponsestoselectiontocaseswherewesetallcross-sexcovariances219

0andbyfullyconstrainingcross-sexcorrelationsto1.Toachievethislaststep,we220

convertedGmftoacorrelationmatrixandreplacedBelementsby1.Wethenback-221

convertedthismodifiedmatrixintocovariancesbyreplacingthediagonalelementsofBby222

thegeometricmeanofmaleandfemalegeneticvariances(i.e.thediagonalelementsofGf223

andGmrespectively).Wethenestimatedthevectorcorrelationbetweentheresponse224

calculatedfromourestimatedmatrixandthoseestimatedwithunconstrained(rB=0)and225

constrainedmatrices(rB=1).Next,wecomparedhowconsistentmaleandfemaleresponse226

waswhenevolvingtowardthesameoptimum(concordantselection)andwhenselection227

wasdiscordant.Thiswasachievedbycalculatingthevectorcorrelationbetweenthe228

directionofselectionforeachsexandthecorrespondingresponse(rβ×Δz).229

RESULTS230

Malesandfemalesshowedlittleevidenceforsexualdimorphisminaveragebehavior(all231

Pmcmc<0.57,TableS1,S2).WedidfindevidenceofaG×Sexinteraction,andfemaleshad232

lowerheritabilityandevolvabilitycomparedtomales(meanfemaleh2=0.10;meanmale233

h2=0.47;meanfemaleI=3.04%;meanmaleI=15.73%).Thisisconfirmedbythefact234

thattheadditivegeneticvariancewaslowerinfemalescomparedtomalesforallthree235

behaviors(posteriormedian[89%CI];shelteremergence:ΔVA(females–males)=-50.12[-236

110.18;8.18],Pmcmc=0.92;activity:ΔVA=-43.38[-89.01;-9.19],Pmcmc=0.99;237

antipredatorresponse:ΔVA=-17.63[-44.47;5.45],Pmcmc=0.92)(Figure2,3;TableS3).238

Wealsofoundstrongevidenceforthesex-specificexpressionofbehavioral239

syndromes,withweakergeneticcorrelationsbetweenbehaviorsinfemales.Malesand240

femalesdifferedprimarilyinhowshelteremergencerelatedtoopen-fieldactivityand241

antipredatorresponse(Figure2),whiletherelationshipbetweenopen-fieldactivityand242

antipredatorresponsewasstablebetweensexes(females:rmf=0.65[0.20;0.98],Pmcmc=243

0.94,Pmcmc=0.91;males:rmf=0.64[0.27;0.92],Pmcmc=0.98).Femalesthatstayed244

longerintheshelterhadlowerantipredatorresponseandtravelledfurtherintheopen-245

field(shelteremergence×open-field:rmf=-0.45[-0.92;0.11],Pmcmc=0.86;shelter246

emergence×antipred.:rmf=-0.48[-0.90;-0.01],Pmcmc=0.91).Incontrast,maleswith247

13

slowshelteremergencehadhigheractivityandantipredatorresponse(shelteremergence248

×open-field:rmf=0.47[0.06;0.87],Pmcmc=0.93;shelteremergence×antipred.:rmf=249

0.50[-0.01;0.84],Pmcmc=0.92).Inaddition,thecorrelationbetweenmajoraxesof250

geneticvariation(gmax)wassignificantly<1,providinganotherlineofevidenceforthe251

presenceofasex-specificsyndrome(vectorcorrelationr°=0.36[0.00;0.76],Pmcmc≠1=252

0.97)(TableS2,S3).253

254

14

255

Figure2.Geneticcorrelationmatrix(Gmf)indicatingsex-specificandcross-sexgenetic256 correlations.Heritabilities(h2)areindicatedonthemaindiagonalandgeneticcorrelations257 (r)ontheoff-diagonalelements.Off-diagonalelementsrepresenteitherthesex-specific258 geneticcorrelations(rfandrminlightgrey)orthecross-sexgeneticcorrelations(rmfin259 darkgrey).BoldindicatesignificantcorrelationsbasedonPmcmc>0.95.Correlationswith260 Pmcmc>0.90areindicatedinitalics.Cross-sexcorrelationssignificantlydifferentfrom1261 areindicatedwithanasterisk.262

263

15

Allbehaviorsshowedweaktomoderatecross-sexcorrelationsandwere264

significantlybelow1(shelteremergence:rmf=-0.09[-0.59;0.39],Pmcmc≠1=1.00;activity:265

rmf=0.65[0.22,0.97],Pmcmc≠1=0.94;antipredatorresponse:rmf=0.64[0.26;0.94],266

Pmcmc≠1=0.96)(TableS3).Inaddition,sexeshadbiasedexpressionofdifferenttrait267

combinations,asindicatedbythehighproportionofasymmetryinB—thecross-sex268

covariancematrix(proportionofskew-symmetry=0.25[0.04,0.42]).Thisdegreeof269

asymmetrywasmostpronouncedforthecross-sexcorrelationsbetweenshelter270

emergenceandactivity(ΔCOVA=-21.39[-43.72;-0.56],Pmcmc=0.97)andbetween271

shelteremergenceandantipredatorresponse(ΔCOVA=-14.8[-32.33;1.46],Pmcmc=272

0.95).Incontrast,thecorrelationbetweenactivityandantipredatorresponsedidnotdiffer273

amongsexes(ΔCOVA=-3.02[-11.73;6.73],Pmcmc=0.71)(Figure3).274

Thismeansthathighlyactivefathersproduceddaughterswithfastershelter275

emergence(rmf=-0.50[-0.85;-0.06],Pmcmc=0.94)andhigherantipredatorresponse(rmf276

=0.57[0.12;0.95],Pmcmc=0.93).Incontrast,activemothersonlyweaklycontributedto277

theirsons’shelteremergence(rmf=0.39[-0.14;0.85],Pmcmc=0.85)andantipredator278

response(rmf=0.42[-0.01;0.85],Pmcmc=0.90).279

280

16

281

Figure3.Thegeneticstructureoftheboldness-activitysyndromedifferedbothintermsof282 itssex-specificgeneticvariances(A)andcorrelations(B)aswellasitscross-sex283 correlationswithintraits(C)andcross-sexcross-traitscorrelations(D).284

285

17

Finally,bysimulatingresponsestoselection,wefoundthatevolutionarytrajectories286

weremoreconsistentwithbehaviorsbeingsexuallyindependentthanwithaconstrained287

expressionofbehaviors.Thiswasthecaseregardlessofwhetherselectionfavoredsex-288

specificoptima(i.e.discordantselection:rB=0= 0.95 [0.73;1.00];rB=1=0.28[0.00;0.56];Δr289

=0.64[0.24;0.98],Pmcmc=0.96)orwhenbothsexeshadthesameoptimum(i.e.290

concordantselection:rB=0=0.95[0.74;1.00],Δr=0.23[-0.04;0.66],Pmcmc=0.95)(Figure291

4).Wenextcomparedtheagreementbetweensimulatedselectiongradientsandpredicted292

responsetoselection(rβ×Δz)betweensexes.Malesandfemaleshadequallyconsistent293

responsestoconcordantselection(femalerβ×Δz=0.72[0.29;1.00];malerβ×Δz=0.75[0.47;294

1.00];Δrβ×Δz=-0.19[-0.67;0.07],Pmcmc=0.88).However,femaleresponsetendedtobe295

biasedawayfromselectioncomparedtomaleswhenselectionwasdiscordant(femalerβ×Δz296

=0.50[0.06;0.95];malerβ×Δz=0.77[0.46;1.00],Δrβ×Δz=-0.27[-0.75;0.11],Pmcmc=0.86).297

Thiscouldmeanthatthemagnitudeofthegeneticconstraintisexpressedmorestronglyin298

males,whichisconsistentwiththeasymmetricnatureofthecross-sexcovariancematrix.299

However,thislastresultshouldbeinterpretedwithcautiongiventhewidecredible300

intervalsaroundtheseestimates. 301

18

302

Figure4.Theresponsetoselectionismoreconsistentwithmodelsweresexesevolve303

independently(B=0)thanmodelswheresexesarefullyconstrained(B=1).Thistrendis304

mostpronouncedwhenselectionfavoursoppositeoptimaamongsexes(discordant305

selection).Positivevaluesindicatethattheobservedresponsetoselection(Δz)ismore306

stronglycorrelatedwithresponsestoselectionwhereelementsofthecross-sexcovariance307

matrixaresetto0(B=0)comparedtoresponsestoselectionwherethecross-sex308

covarianceisfullyconstrained(B=1),basedon1,000randomselectiongradientsapplied309

toeach1,000posteriorcovariancematrices. 310

19

DISCUSSION311

Ourresultsshowthatmalesandfemalesdiffersubstantiallyintheirbehavioralsyndromes312

atthegeneticlevel.Shelteremergencewasgeneticallyindependentbetweenmalesand313

females,whereasthegeneticconstraintforactivityandantipredatorresponsewas314

strongerbutstilldepartedfrom1.Theabsenceofastronggeneticconstraintlinkingthe315

sexesissurprisinggiventhatthebehaviorswemeasuredarenottypicallyconsidered316

distinctsex-specifictraits.Inthecaseofshelteremergence,across-sexcorrelation317

approachingzeroimpliesthatthesamebehaviorisunderpinnedbycompletely318

independentsetsofgenesinmalesandfemales.Suchgeneticuncouplingmeansthatthis319

behaviorcanfulfilldifferentfunctionsineachsexandcanevolveindependently.Our320

resultssuggestthatsexualconflictmayhavebeenresolvedinthisspecieseveninabsence321

ofobservablebehavioraldimorphism.322

Byapplyingaquantitativegeneticapproach,wewereabletouncovermultipleways323

inwhichsexesdifferedintheirbehaviors.Whilemalesandfemalesdidnotdifferinmean324

behaviors,weuncoveredasignatureofbehavioraldimorphismintheamountofgenetic325

variationexpressedbyeachsex.ThisG×Sexeffectwascharacterizedbyfemalebehavior326

beinglessheritablethanobservedformales.Inourcase,traitswithlowerheritabilities327

alsohadlowerevolvabilities,indicatingthatfemalesarelessresponsivetoselective328

pressuresthanmalesinthisspecies.Thisstrongdifferenceingeneticvariancebetween329

sexescouldresultfromstrongerstabilizingselectioninfemaleserodinggeneticvariation330

intraitsrelatedtoexplorationandrisktakingwhilemaintainingasimilaroptimumineach331

sex.Multivariatestabilizingselectionhasindeedbeenshowntobeanimportantdriverof332

traitevolution,forexampleingenitalmorphologyinarthropods(Arnqvist1997;Houseet333

al.2020).Anotherpossibilityisthatrearingindividualsinlaboratoryconditionswithad-lib334

accesstofoodandsheltercouldalleviatetheresourceacquisitionandallocationtrade-offs335

thatmaybenegotiateddifferentlybyeachsex.Thisexplanationisespeciallyplausibleif336

malesandfemalesmodifytheirphenotypestodifferentdegreeinresponsetocaptivity.337

Notetoothattheenvironmentalcontributiontofemales’behavioralphenotypewasmuch338

greaterthanthatofmales,leavingspacefordifferentialadjustmentofbehaviorbysexesto339

occur.340

20

Malesandfemalesdifferedsubstantiallyinhowtheyexpressedaboldness-activity341

syndrome.Contrarytoourpredictions,inbothsexesactivegenotypeswerealsomore342

sensitivetopredatorycues,withnochangetothemagnitudeofthegeneticcorrelation.The343

primarydifferencewasthereforeinhowshelteremergencerelatedtoactivityand344

antipredatorresponse.“Bold”females–i.e.femaleswithfastemergencefromtheshelter–345

hadhigheractivityandantipredatorresponse.Incontrast,activemalesweremore346

sensitivetopredatorcuesandbehavedcautiouslywhenemergingfromtheshelter.This347

resultprovidesimportantinsightintohowmaleandfemalecricketshandlerisky348

situations.Femalemusttravelthroughriskyenvironmentsinordertolocatemates.349

Therefore,boldandactivegenotypesmayneedtocompensatefortheseriskybehaviorsby350

beingmorereactivetothepresenceofpredatorcues.Males,incontrast,stayclosetotheir351

shelterbutproducecourtshipsignalsthatmakethemthetargetofpredatorsandparasites.352

Asaresult,boldermalesmayignorethepresenceofpredatorycuesifthisstrategyyields353

higherfrequencyofencounterswithfemales.Thistypeofriskcompensationstrategyhas354

alreadybeenshowninpreviousstudieswheremaleswithmoreattractivesongsstay355

longerinshelters(Hedrick2000).356

Ourfindingsthatmalesandfemalesdifferedinthegeneticexpressionbehavioral357

syndromesandthatcross-sexcorrelationsareuncoupledsuggestthatrisk-takingis358

regulatedbydifferentphysiologicalpathwaysineachsex.Incrickets,severalmonoamine359

neurohormonesareinvolvedintheregulationofmultiplebehaviorstiedtoa“fightor360

flight”response,includingaggression,courtship,dispersalandresponsetosimulated361

predationexposure(Adamoetal.1995,2013;StevensonandRillich2016;Adamo2017;362

Lundgrenetal.2021).However,experimentstendtoeitherfocusonmalesexclusivelyor363

arenotdesignedtoaddresssex-differencesinphysiologicalmechanisms.However,sex-364

differencesinasimilaraxisofvariation,theproactive/reactiveaxis,havealsobeen365

describedinseveralvertebratetaxa(Kokrasetal.2012;Immonenetal.2018).366

Thegeneticstructureofbehavioralsyndromesishighlyconservedinthisspecies,367

evenacrossisolatedpopulations(Royautéetal.2020).Thisindicatesthatbehavioral368

syndromesmaynotvarymuchwhenexposedtodifferentselectivepressures.Instead,369

geneticconstraintsresultingfrompleiotropywasthemorelikelyexplanationsforthe370

21

presenceofbehavioralsyndromeinthisspecies.Here,weshowthatbehavioralsyndromes371

differedmorestronglybetweensexesthanamongpopulations.Alikelyexplanationisthat372

malesandfemalesexpresssex-specificbehaviorsregardlessofthepopulationoforigin.As373

aresult,selectionmaybemorelikelytodifferbetweensexesthanbetweenpopulations.374

Whilecomparisonsofdifferencesinselectionamongsexesandpopulationsarescarce,375

previousstudiessuggestthatsexuallydiscordantselectioniscommoninthewildand376

tendstobestrongerintraitsinwhichsexualdimorphismispronounced(Coxetal.2009).377

Whilerarelyinvestigated,sexspecificdifferencesinbehavioralsyndromeshave378

beendemonstratedinsomespecies.However,mostofthesestudieshavefocusedon379

comparisonsofbehavioralrepeatabilitiesbetweenmalesandfemalesinsingletraits380

(Jenkins2011;Debeffeetal.2015)ratherprovidingabroaderexplorationofmultivariate381

patternsofcross-sexcorrelationinbehavioraltraits,aswedidhere.Interestingly,(Hedrick382

andKortet2012)previouslyidentifiedsexdifferencesintherepeatabilityofshelter383

emergenceinaseparatepopulationofG.integer.Contrarytowhatwedetected,maleshad384

lowerrepeatabilitycomparedtofemales.Notethatrepeatabilityandheritabilityarenot385

directlycomparablebecauserepeatabilityalsoincludessourcesofvariationduetothe386

“permanentenvironment”whileheritabilityonlyincludesadditivegeneticvariation.More387

recentlyseveralstudieshavealsocomparedsyndromestructureamongsexeswithmixed388

results.Somestudiesreportedlargedifferencesinbehavioralsyndromesbetweenmales389

andfemales(Fresneauetal.2014;Hanetal.2015;Royauté2015;Wayetal.2015)while390

otherssupportaconservedsyndromestructurebetweensexes(Michelangelietal.2016;391

Gouletetal.2021).Ourresultssupplyanadditionallineofevidenceinfavorofsex-specific392

syndromes.393

Thereisonlylimitedexplorationofmultivariatepatternsofcross-sexcorrelationin394

behavioraltraits.Moststudiesreportedfrom(Poissantetal.2010)concernsingletraits395

andpointtostrongcross-sexcorrelationswithrelativelyweakdimorphismforbehavior.396

Themajorityoftheseestimateswere,however,fromstudiesinterestedinsexualselection,397

with<4%ofestimatescomingfromexplicitlybehavioralstudies.Becausewealso398

estimatedthecross-sexcorrelationattheadditivegeneticlevel,wewereabletodetermine399

thatbehavioraltraitsmaybelessconstrainedbysexthanpreviouslythought.Indeed,our400

22

meanestimateforrmfwasmuchweakerthanobservedinpreviousstudiesevaluatingintra-401

locussexualconflictsinbehavior(meanrmf=0.46vs.0.77,Poissantetal.2010).Our402

resultssuggestthatbehavioraltraitsmaycommonlyfulfilldifferentfunctionsbetween403

sexesandshouldthereforebeanalyzedasseparatetraitsbetweenthesexesinbehavioral404

syndromeandanimalpersonalitystudies.405

Veryfewstudieshavebeenabletoestimatethegeneticcontributionofsexesto406

behavioralsyndromesaswehavedonehere.Intheorb-weavingspiderNuctenea407

umbratical,cross-sexcorrelationsforactivityandaggressionwereinthesamerangeasour408

estimatesforactivityandantipredatorresponse(rmf=0.50)(Kralj-Fišeretal.2019).In409

Gryllusbimaculatus,whichiscloselyrelatedtoG.integer,explorationandaggressionare410

morestronglycorrelatedinfemalesandthesetraitshaveasymmetriccontributionsacross411

sexes(Hanetal.2019).However,cross-sexcorrelationswithintraitswerequalitatively412

strongerthanobservedinourstudy(meanrmf=0.70vsmeanrmf=0.46),pointingtoa413

strongergeneticconstraintbetweensexesthanweobservedinGryllusinteger.Finally,414

Whiteetal.(2019)showeddifferencesbetweensex-specificcovariancesinguppiesto415

similarG.integer.However,theyreportedamuchweakersignalofasymmetryinB.The416

factthatthesefewstudiesdifferwidelyinconclusionsshowsthatthereisnoconsensusyet417

onhowbehavioralsyndromesdifferbetweensexesandevencloselyrelatedspeciesmay418

showstrongdifferencesintheexpressionofbehavioraldimorphism.419

Otherstudies,conductedprimarilyonmorphologicaltraits,pointtostrong420

constraintsimposedbygeneticcross-sexcovariances(Gosdenetal.2012;Gosdenand421

Chenoweth2014;SztepanaczandHoule2019).Inourcase,theconstrainingeffectofcross-422

sexcovarianceswasweakatbestandresponsestoselectionweremoreconsistentwith423

independentevolutionarytrajectoriesforeachsex.Thisissimilartoapreviousstudyon424

sexuallydimorphicornamentsinbrownanoleswhichshowedthatthecross-sex425

covarianceamongornamentsdidnotsubstantiallyconstraintheevolutionofdimorphism426

(Coxetal.2017).427

Wefoundahighdegreeofasymmetryinthecross-sexcross-traitscomponentsof428

thegeneticcovariancematrix.Inparticular,maleactivityhadastrongercontributionto429

23

femaleshelteremergencethanfemaleactivityonmaleshelteremergence.Thistypeof430

asymmetriccontributiontothephenotypeisexpectedtobiasresponsesawayfrom431

selection.InDrosophilaforexample,wingshapeevolutionisprimarilyshapedbycross-sex432

covariancesforthesametrait(i.e.thediagonalelementsofB)ratherthancross-sexcross-433

traitcovariances(SztepanaczandHoule2019).Thiswasthecaseeventhoughupperand434

lowerelementsoftheBmatrixfrequentlyhadoppositesignsanddespiteageneral,though435

weaker,signalofasymmetrysuchaswereporthere(5%vs.25%).Despitestrong436

asymmetryinourestimateofcross-sexcovariances,wefoundonlyweakevidencefor437

biasedresponsestoselection.Instead,oursimulationsoftheresponsestoselections438

showedthattheevolutionofbehavioraldimorphismswasmoreconsistentwithscenarios439

inwhichsexesaregeneticallyuncoupledandallowedtoevolveindependentlycomparedto440

afullyconstrainedcross-sexgeneticarchitecture.Thismeansthattheindependent441

evolutionarytrajectoriesbysexeswereportarelikelyaresultoftheweakcross-sex442

correlationswithintraitswedetected(meanrmf<0.50).443

Weshowthattraitsinvolvedinstudiesofanimalpersonalityhaveasex-specific444

geneticarchitecture.Thesebehaviorsaregeneticallyuncoupledbetweensexesandallow445

sexestofollowindependentevolutionarytrajectories.Ourresultsareintriguingbecause446

theorypredictsthatcross-sexcovarianceswillactasconstraintsandlimittheabilitiesof447

sexestoreachtheiroptimum.Thetypeofgeneticindependencewefoundsuggestsapast448

historyofdiscordantselectionthathasstronglyshapedbehavioraldimorphismandhas449

resolvedsexualconflictinthisspecies.450

451

24

LITTERATURECITED452

Adamo,S.A.2017.Thestressresponseandimmunesystemshare,borrow,andreconfigure453

theirphysiologicalnetworkelements:Evidencefromtheinsects.Horm.Behav.454

88:25–30.455

Adamo,S.A.,I.Kovalko,andB.Mosher.2013.Thebehaviouraleffectsofpredator-induced456

stressresponsesinthecricket(Gryllustexensis):theupsideofthestressresponse.457

J.Exp.Biol.216:4608–4614.TheCompanyofBiologistsLtd.458

Adamo,S.A.,C.E.Linn,andR.R.Hoy.1995.Theroleofneurohormonaloctopamineduring459

“fightorflight”behaviourinthefieldcricketGryllusbimaculatus.J.Exp.Biol.460

198:1691–1700.TheCompanyofBiologistsLtd.461

Aragón,P.2011.Theresponsetothesocialenvironmentrevealssex-dependent462

behaviouralsyndromesintheBosca’snewt(Lissotritonboscai).J.Ethol.29:79–83.463

Arnqvist,G.1997.Theevolutionofanimalgenitalia:distinguishingbetweenhypothesesby464

singlespeciesstudies.Biol.J.Linn.Soc.60:365–379.465

Berger,D.,E.C.Berg,W.Widegren,G.Arnqvist,andA.A.Maklakov.2014.Multivariate466

intralocussexualconflictinseedbeetles.Evolution68:3457–3469.467

Blanckenhorn,W.U.2005.BehavioralCausesandConsequencesofSexualSize468

Dimorphism.Ethology111:977–1016.469

Bonduriansky,R.,andS.F.Chenoweth.2009.Intralocussexualconflict.TrendsEcol.Evol.470

24:280–288.471

Cox,R.M.,R.A.Costello,B.E.Camber,andJ.W.McGlothlin.2017.Multivariategenetic472

architectureoftheAnolisdewlaprevealsbothsharedandsex-specificfeaturesofa473

sexuallydimorphicornament.J.Evol.Biol.30:1262–1275.474

25

Debeffe,L.,J.F.Lemaître,U.A.Bergvall,A.J.M.Hewison,J.M.Gaillard,N.Morellet,M.475

Goulard,C.Monestier,M.David,H.Verheyden-Tixier,L.Jäderberg,C.Vanpé,andP.476

Kjellander.2015.Short-andlong-termrepeatabilityofdocilityintheroedeer:sex477

andagematter.Anim.Behav.109:53–63.478

Delph,L.F.,J.C.Steven,I.A.Anderson,C.R.Herlihy,andE.D.B.Iii.2011.Eliminationofa479

GeneticCorrelationBetweentheSexesViaArtificialCorrelationalSelection.480

Evolution65:2872–2880.481

Dochtermann,N.A.,andN.J.Dingemanse.2013.Behavioralsyndromesasevolutionary482

constraints.Behav.Ecol.24:806–811.OxfordAcademic.483

French,B.W.,andW.H.Cade.1987.Thetimingofcalling,movement,andmatinginthe484

fieldcricketsGryllusveletis,G.pennsylvanicus,andG.integer.Behav.Ecol.485

Sociobiol.21:157–162.486

Fresneau,N.,E.Kluen,andJ.E.Brommer.2014.Asex-specificbehavioralsyndromeina487

wildpasserine.Behav.Ecol.25:359–367.488

Gosden,T.P.,andS.F.Chenoweth.2014.TheEvolutionaryStabilityofCross-Sex,Cross-489

TraitGeneticCovariances.Evolution68:1687–1697.490

Gosden,T.P.,K.-L.Shastri,P.Innocenti,andS.F.Chenoweth.2012.TheB-MatrixHarbors491

SignificantandSex-SpecificConstraintsontheEvolutionofMulticharacterSexual492

Dimorphism.Evolution66:2106–2116.493

Goulet,C.T.,W.Hart,B.L.Phillips,J.Llewelyn,B.B.M.Wong,andD.G.Chapple.2021.No494

behavioralsyndromesorsex-specificpersonalitydifferencesinthesouthern495

rainforestsunskink(Lampropholissimilis).Ethology127:102–108.496

26

Hadfield,J.D.2010.MCMCmethodsformulti-responsegeneralizedlinearmixedmodels:497

theMCMCglmmRpackage.J.Stat.Softw.33:1–22.498

Hämäläinen,A.,E.Immonen,M.Tarka,andW.Schuett.2018.Evolutionofsex-specificpace-499

of-lifesyndromes:causesandconsequences.Behav.Ecol.Sociobiol.72:50.500

Han,C.S.,T.P.Gosden,andN.J.Dingemanse.2019.Proteindeprivationfacilitatesthe501

independentevolutionofbehaviorandmorphology.Evolution73:1809–1820.502

Han,C.S.,P.G.Jablonski,andR.C.Brooks.2015.Intimidatingcourtshipandsexdifferences503

inpredationriskleadtosex-specificbehaviouralsyndromes.Anim.Behav.109:177–504

185.505

Hedrick,A.V.,andR.Kortet.2012.Sexdifferencesintherepeatabilityofboldnessover506

metamorphosis.Behav.Ecol.Sociobiol.66:407–412.507

Hedrick,A.V.,andE.J.Temeles.1989.Theevolutionofsexualdimorphisminanimals:508

Hypothesesandtests.TrendsEcol.Evol.4:136–138.509

House,C.,P.Tunstall,J.Rapkin,M.J.Bale,M.Gage,E.delCastillo,andJ.Hunt.2020.510

Multivariatestabilizingsexualselectionandtheevolutionofmaleandfemalegenital511

morphologyintheredflourbeetle*.Evolution74:883–896.512

Immonen,E.,A.Hämäläinen,W.Schuett,andM.Tarka.2018.Evolutionofsex-specificpace-513

of-lifesyndromes:geneticarchitectureandphysiologicalmechanisms.Behav.Ecol.514

Sociobiol.72:60.515

Jenkins,S.H.2011.Sexdifferencesinrepeatabilityoffood-hoardingbehaviourofkangaroo516

rats.Anim.Behav.81:1155–1162.517

27

Kokras,N.,C.Dalla,A.C.Sideris,A.Dendi,H.G.Mikail,K.Antoniou,andZ.Papadopoulou-518

Daifoti.2012.Behavioralsexualdimorphisminmodelsofanxietyanddepression519

duetochangesinHPAaxisactivity.Neuropharmacology62:436–445.520

Kortet,R.,andA.Hedrick.2007.AbehaviouralsyndromeinthefieldcricketGryllusinteger:521

intrasexualaggressioniscorrelatedwithactivityinanovelenvironment.Biol.J.522

Linn.Soc.91:475–482.523

Kralj-Fišer,S.,K.L.Laskowski,andF.Garcia-Gonzalez.2019.Sexdifferencesinthegenetic524

architectureofaggressivenessinasexuallydimorphicspider.Ecol.Evol.9:10758–525

10766.526

Kruuk,L.E.B.2004.Estimatinggeneticparametersinnaturalpopulationsusingthe‘animal527

model.’Philos.Trans.R.Soc.Lond.B.Biol.Sci.359:873–890.528

Lande,R.1980.SexualDimorphism,SexualSelection,andAdaptationinPolygenic529

Characters.Evolution34:292–305.[SocietyfortheStudyofEvolution,Wiley].530

Long,T.A.F.,andW.R.Rice.2007.Adultlocomotoryactivitymediatesintralocussexual531

conflictinalaboratory-adaptedpopulationofDrosophilamelanogaster.Proc.R.Soc.532

BBiol.Sci.274:3105–3112.RoyalSociety.533

Lundgren,K.A.,R.N.Abbey-Lee,L.C.Garnham,A.Kreshchenko,S.Ryding,andH.Løvlie.534

2021.Manipulatingmonoaminesreducesexplorationandboldnessof535

Mediterraneanfieldcrickets.Behav.Processes183:104298.536

Mainwaring,M.C.,J.L.Beal,andI.R.Hartley.2011.Zebrafinchesarebolderinanasocial,537

ratherthansocial,context.Behav.Processes87:171–175.538

28

Michelangeli,M.,D.G.Chapple,andB.B.M.Wong.2016.Arebehaviouralsyndromessex539

specific?Personalityinawidespreadlizardspecies.Behav.Ecol.Sociobiol.70:1911–540

1919.541

Moiron,M.,K.L.Laskowski,andP.T.Niemelä.2020.Individualdifferencesinbehaviour542

explainvariationinsurvival:ameta-analysis.Ecol.Lett.23:399–408.543

Poissant,J.,A.J.Wilson,andD.W.Coltman.2010.Sex-SpecificGeneticVarianceandthe544

EvolutionofSexualDimorphism:ASystematicReviewofCross-SexGenetic545

Correlations.Evolution64:97–107.546

Réale,D.,S.M.Reader,D.Sol,P.T.McDougall,andN.J.Dingemanse.2007.Integrating547

animaltemperamentwithinecologyandevolution.Biol.Rev.82:291–318.548

Royauté,C.,R.,.Buddle,C.M.,.Vincent.2015.Undertheinfluence:Sublethalexposuretoan549

insecticideaffectspersonalityexpressioninajumpingspider.29.550

Royauté,R.,andN.A.Dochtermann.2017.Whenthemeannolongermatters:551

Developmentaldietaffectsbehavioralvariationbutnotpopulationaveragesinthe552

housecricket(Achetadomesticus).Behav.Ecol.28:337–345.553

Royauté,R.,C.Garrison,J.Dalos,M.A.Berdal,andN.A.Dochtermann.2019.Currentenergy554

stateinteractswiththedevelopmentalenvironmenttoinfluencebehavioural555

plasticity.Anim.Behav.148:39–51.556

Royauté,R.,K.Greenlee,M.Baldwin,andN.A.Dochtermann.2015.Behaviour,metabolism557

andsize:phenotypicmodularityorintegrationinAchetadomesticus?Anim.Behav.558

110:163–169.559

29

Royauté,R.,A.Hedrick,andN.A.Dochtermann.2020.Behaviouralsyndromesshape560

evolutionarytrajectoriesviaconservedgeneticarchitecture.Proc.R.Soc.BBiol.Sci.561

287:20200183.RoyalSociety.562

Schuett,W.,T.Tregenza,andS.R.X.Dall.2010.Sexualselectionandanimalpersonality.563

Biol.Rev.85:217–246.564

Sih,A.,A.Bell,andJ.C.Johnson.2004.Behavioralsyndromes:anecologicaland565

evolutionaryoverview.TrendsEcol.Evol.19:372–378.566

Stevenson,P.A.,andJ.Rillich.2016.Controllingthedecisiontofightorflee:therolesof567

biogenicaminesandnitricoxideinthecricket†.Curr.Zool.62:265–275.568

Sztepanacz,J.L.,andD.Houle.2019.Cross-sexgeneticcovarianceslimittheevolvabilityof569

wing-shapewithinandamongspeciesofDrosophila.Evolution73:1617–1633.John570

Wiley&Sons,Ltd.571

Walsh,B.,andM.W.Blows.2009.AbundantGeneticVariation+StrongSelection=572

MultivariateGeneticConstraints:AGeometricViewofAdaptation.Annu.Rev.Ecol.573

Evol.Syst.40:41–59.574

Way,G.P.,A.L.Kiesel,N.Ruhl,J.L.Snekser,andS.P.McRobert.2015.Sexdifferencesina575

shoaling-boldnessbehavioralsyndrome,butnolinkwithaggression.Behav.576

Processes113:7–12.577

Wilson,A.J.,D.Réale,M.N.Clements,M.M.Morrissey,E.Postma,C.A.Walling,L.E.B.578

Kruuk,andD.H.Nussey.2010.Anecologist’sguidetotheanimalmodel.J.Anim.579

Ecol.79:13–26.580

581