Proc.Nati.Acad.Sci.USAVol.89,pp.12137-12139,December1992PsychologyCostsofdeception:Cheatersarepunishedinrhesusmonkeys(Macacamulatta)MARCD.HAUSER*AnimalCommunicationLaboratory,DepartmentofZoology,UniversityofCalifornia,Davis,CA95616-8761CommunicatedbyPeterMarler,September14,1992ABSTRACTFromafunctionalperspective,deceptioncanevolveinanimalpopulationsbutshouldbeconstrainedbythecostsassociatedwithdetection.Itthenfollowsthatwithholdinginformationshouldbemoreprevalentasaformofdeceptionthanactivefalsificationofinformationbecauseoftherelativedifficultiesassociatedwithdetectingcheaters.Empiricalstud-iesofdeceptionhavefocusedonthebenefitsofcheatingbuthaveprovidednodataonthecostsasatedwithbeingdetectedasacheater.Ipresentresultsfromfieldexperimentsonrhesusmonkeys(Macacamulatta)whichshowthatindivid-ualsdiscoveringfoodannouncetheirdiscoveriesbycallingon45%ofalltrials.Discovererswhofailedtocall,butweredetectedwithfoodbyothergroupmembers,receivedsignifi-cantlymoreaggressionthanvocaldiscoverers.Moreover,silentfemalediscoverersatesignificantlylessfoodthanvocalfemales.Thisdemonstratesthattherearesignitcoststowithholdinginformation.Suchcostsmayconstrainthefre-quencywithwhichdeceptionoccursinthisandotherpopula-tions.Theoreticaltreatmentsofanimalcommunication(e.g.,refs.1-4)suggestthatwithholdinginformationfromothergroupmemberscanleadtosubstantialfitnessgains.Thisformofdeception,itisclaimed,shouldoccurmorefrequentlythanactivefalsificationbecauseitismoredifficulttodetectcheaters.However,studiesofdeceptioninnonhumanani-mals(5-10)havenotprovidedevidenceofthecostsassoci-atedwithcheating.Consequently,wehavelittleunderstand-ingofthefactorsthatconstrainthefrequencyofdeception.Thispaperprovidesexperimentaldataonthecostsofde-ceptioninafree-rangingmonkey.Undernonexperimentalconditions,rhesusmonkeys(Macacamulatta)onCayoSantiago,PuertoRico,produceacomplexoffiveacousticallydistinctcallsupondiscoveringfood(Fig.1).Callproduction,however,isnotanecessaryconsequenceofencounteringfood.Rather,productionorsuppressionoffood-associatedcallsisaffectedbygenderandthenumberofcloselyrelatedkin:femalescallmore(mean=0.10callperhr)thanmales(mean=0.03callperhr;t=3.78,df=20,P<0.001),andfemaleswithlargematrilinescallmorethanfemaleswithsmallmatrilines(p=0.81,n=6,P<0.05).Experimentswereconductedonagroup(L)offree-rangingrhesusmacaquesonCayoSantiago.Over50%oofthedietcomesfromprovisioning(i.e.,chow).Theremainingportionofthedietismadeupofnaturalfoliage,fruits(e.g.,coconut),insects,andsoil(12).Withintheconfinesoftheisland(15ha),movementisunrestrictedand,consequently,malestransferbetweengroups.Therearenopredatorsontheisland(12).Experimentswereconductedon28adultmalesand21adultfemales,includinghigh-,middle-,andlow-rankingindividuals;onlyonetrialwasconductedperindividual.Threeobserverswereinvolvedineachexperiment.Oneobserver,positioned10mawayfromthediscoverer,wasresponsibleforvideotapingthetrial.Asecondobserverpositionedherselfata900anglebetweenthediscovererandthevideocameraandwasresponsiblefordroppingoneofthefoodorcontrolstimuliinfrontofthediscovererandthenmoving10maway.Boththesecondandthirdobserversprovideddataoninteractionsoccurringoutsideofthecam-era'srange.Thetrialstartedwitha30-secbaselineperiodandcontinuedfor20minthereafter.Thepresentationofstimulialternatedbetweentrials.Individualsweretestedwhentheywere>10mfromothergroupmembersand,fromourperspective,hiddenfromview;onCayoSantiago,individu-alsoftenforageatadistanceof10mfromothergroupmembersandthisisparticularlythecasepriortothedispen-sationofchowandinthelateafternoon.Alltrialsweretranscribedtoquantifythebehavioralresponse.Threevariableswereexplicitlyexaminedfortheireffectsoncallproduction:timeofday,foodquality,andgender.Trialswereconductedeitherbetween0600and0700orbetween1300and1500.Thefirsttimeblockprecedesthestrongestpeakinfeedingactivityandthedeliveryofchowintothedispensers.Consequently,individualstestedduringthisperiodwereunlikelytohaveeatenduringtheprevious10-12hr.Thesecondtimeblockoccursafter=80%oftheday'sfeedinghasbeencompletedand,ingeneral,amajorityofthechowhasbeendepletedfromthedispensers.Thetwofooditemswerechowandcoconut.Coconutisahighlypreferredfoodandofsufficientrarityontheislandthatcompetitionforaccesstoevensmallpiecesisintense,oftenescalatingtoaggressivechasesandsubsequentinjury.Inagiventrial,15piecesofchoworcoconutweredroppedfromanopaquebag,=5mfromthetargeteddiscoverer;eachpieceoffoodwasapproximatelythesamesizeandstandardizedtoanaveragepieceofchow(i.e.,3x6x2cm).Asacontrol,15sticksofwoodwerepresented,eachpieceapproximatingthedimensionsofapieceofchow.Uponseeingthesticksdrop,discoverers(n=9)nevercallednordidtheyapproachthedroparea.Whendiscoverersnoticedeitherchow(n=20)orcoconut(n=20),theytypicallyscannedtheareaaroundthemfor-2.0sec(range,0.0-5.5),apparentlytoassessthesocialcompositionofnearbyindividuals,andthenmovedtowardthefood.Onaverage,discoverersate4.9piecesoffood(SD=2.3;range,0.25-9.0).Timeofday,foodtype,gender,anddominancerankdidnothaveastatisticallysignificanteffectontheamountoffoodconsumed.On18ofthe40foodtrials(45%),thediscovererproducedatleastonefood-associatedcall.Femalescalled(n=15trials)significantlymoreoftenthanmales(n=3trials;x2=14.34,P