Callapplenano

firstpublishedonline16January2012,doi:10.
1098/rsbl.
2011.
120082012Biol.
Lett.
BenO.
BrilotandMelissaBatesonWaterbathingaltersthreatperceptioninstarlingsReferenceshttp://rsbl.
royalsocietypublishing.
org/content/8/3/379.
full.
html#ref-list-1Thisarticlecites13articles,1ofwhichcanbeaccessedfreeSubjectcollections(696articles)behaviourArticlesonsimilartopicscanbefoundinthefollowingcollectionsEmailalertingservicehereright-handcornerofthearticleorclickReceivefreeemailalertswhennewarticlescitethisarticle-signupintheboxatthetophttp://rsbl.
royalsocietypublishing.
org/subscriptionsgoto:Biol.
Lett.
TosubscribetoonMarch12,2014rsbl.
royalsocietypublishing.
orgDownloadedfromonMarch12,2014rsbl.
royalsocietypublishing.
orgDownloadedfromAnimalbehaviourWaterbathingaltersthreatperceptioninstarlingsBenO.
Brilot*andMelissaBatesonCentreforBehaviourandEvolution,InstituteofNeuroscience,NewcastleUniversity,HenryWellcomeBuilding,FramlingtonPlace,NewcastleuponTyneNE24HH,UK*Authorforcorrespondence(ben.
brilot@ncl.
ac.
uk).
Themajorityofbirdtaxaperformwaterbathing,butlittleisknownabouttheadaptivevalueofthisbehaviour.
Ifbathingisimportantforfeathermain-tenancethenbirdsthathavenotbathedshouldhavepoorerfeathercondition,compromisedescapeabilityandthereforeincreasedresponsivenesstocuesofpredation.
Weconductedtwoexperimentsexaminingthebehaviourofcaptivestarlingsrespondingtoconspecicalarmcalls.
Birdsthathadnoaccesstobathingwatershowedadecreasedwillingnesstofeedandincreasedtheirvigilancebehaviourfollowinganalarmcall.
Wearguethatbirdsdeniedaccesstobathingwaterinterpretedanambiguouscueofthreatasrequiringmorecautionthanbirdsthathadaccess,consistentwithhigherlevelsofanxiety.
Ourresultssupporttheprovisionofbathingwaterforcaptivebirdsasanimportantwelfaremeasure.
Keywords:bathing;Europeanstarling;Sturnusvulgaris;threatperception;animalwelfare1.
INTRODUCTIONBathinginwaterisatraitcommontothemajorityofbirdtaxa[1,2],butlittleresearchhasbeenconductedintoitsadaptivevalue[1–6].
Ifbathingisessentialforthemaintenanceofplumagecondition,thenwecanderivesomepredictions.
Birdsthathavenotbathedshouldhaveimpairedightperformance,theirescapeabilityshouldbecompromisedandconsequently,theyshouldbemoreresponsivetosignalsofpredationthreat.
CaptiveEuropeanstarlings(Sturnusvulgaris)deniedaccesstobathingwatercollidewithmoreobjectsbutymorequicklyduringescapeights[7].
Separateexperimentshaveshownthatstarlingshousedincageswithoutenvironmentalenrichments(includingbathingwater)aremorelikelytointerpretambiguousstimuliasindicatinganegativefutureout-come[8,9].
Thesendingssuggestthatlackofaccesstobathingwatermayalterthreatperceptioninstar-lings.
Totestthishypothesismoredirectly,weexaminedthebehaviourofcagedstarlingsrespondingtoaconspecicalarmcall[10].
Thiscallsignalsthatapredatormaybepresentbutitisambiguousastothepredator'slocationoridentity.
Wepredictedthatstarlingspreviouslydeniedaccesstobathingwatershouldtakelongertobeginfeedingandhaveelevatedvigilancelevelsonhearingaconspecicalarmcall.
2.
MATERIALANDMETHODSWeused20starlingsforexperiment1and24forexperiment2.
Inbothexperiments,replicatesoffourbirdswerehousedindividuallyinvisuallyisolatedcages.
Bark-coveredkittenfoodanddrinkingwaterfromwall-mounteddrinkerswereprovidedadlibitum[10].
Allbirdsweregivenalargeplastictray;forhalfofthemthiswaslleddailywithcleanwater.
Bathingwasnotdirectlyobserved,butwasevincedbywetcagepapersandreducedwaterlevels.
Birdsweregiven3daystosettleandthen,onatestday,deprivedoffoodfor2h.
Thelaboratorylightsweresubsequentlyswitchedoffandabark-lledfoodbowlcontaining10mealworms(apreferredfood)wasplacedineachcage.
Thebarkincreasedthedifcultyoftheforagingtasktoinduceaforaging-vigilancetrade-off.
Theexper-imenterlefttheroomandafter5minanacousticstimuluswasplayed;onitscompletion,thelightswereswitchedonandthebirds'behaviourrecordedondigitalvideo.
Forbothexperiments,theacousticstimuluscomprisedastarlingalarmcall[10]playedusinganAppleNanoipod(frequencyresponse:20Hzto20kHz+3dB)andYamahaYST-M20DSPactivespeakers(frequencyresponse:70Hzto20kHz+3dB).
Thesoundpressurelevelwasstandardizedtoapeakamplitudeof75dB,measuredattheperchineachcagethatwasfurthestfromthespeakers(allwereequidistantfromthespeakers).
Birdsinexper-iment2werealsosubjectedtoacontrolstarling'threat'call,asignalgiveninmildagonisticconspecicencounters.
Calltypeswerepre-sentedindividuallyonconsecutivedaysinacounterbalancedrepeated-measuresdesign.
Additionally,inexperiment2,allbathswereremovedpriortothelightsbeingswitchedofftoensurethattherewasnomotivationforbirdstomoveinordertobathe.
WescoredthefollowingbehavioursusingTheObserver(XTv.
8.
0,Noldus):latencytomove;latencytobeginfeeding;durationoftherstfeedingbout;durationofeachperiodspentwiththebillcontinuouslybelowhorizontalduringthisbout(head-downboutduration);thedurationofeachperiodspentwiththebillcontinu-ouslyabovehorizontalduringtherstfeedingbout(head-upboutduration);thefrequencyoftransitionofthebillfrombelowtoabovehorizontalduringtherstfeedingbout(head-uprate).
Unfortunately,thebirdscouldnotbeacousticallyisolatedandaudi-torydisturbancesoccurredbothoutsideandwithinthelaboratory(e.
g.
somebirdsemittedalarmcallsinresponsetotheexperimenter).
Anybirdsthatexperiencedsuchdisturbancebeforetrialsorduringthetrialswereexcluded.
Therecordingsfortwobirdsforoneofthecall-typesinexperiment2allowedlatenciestobescored,butthevideoqualitywasnotsatisfactoryforscoringvigilance.
Hence,thenalsamplesizeswere14birdsforexperiment1and10forexperiment2.
Toreduceourdependentvariables,wedroppedlatencytomovesincethiswashighlycorrelatedwithlatencytofeed(experiment1:r0.
530,p0.
051(thestrengthofthiscorrelationwasgreatlyreducedbythedatafromonesubject);experiment2,alarmcall:r0.
999,p,0.
001;experiment2,'threat'call:r0.
978,p,0.
001).
Weenteredtheremainingmeasures(transformedtoensurenormality)intoaprincipalcomponentanalysis(PCA,usingPASWStatisticsforMacv.
18.
0.
3,SPSSInc.
)assumingnorotation(theresultsalsoheldunderanassumptionoforthogonal/obliquerelationshipbetweenfactors).
3.
RESULTSThePCAyieldedtwofactorsforbothexperiments1and2(table1).
Forexperiment1,weemployedamulti-variateanalysisofvariancewiththetwofactorsasthedependentvariables.
Bathinghadasignicanteffectonthesubjects'behaviour(F2,115.
503,p0.
022;gure1).
Thiswaslimitedtotherstfactorwherebathinghadalargeeffectasjudgedbytheeffectsizeestimate,Hedges'unbiasedestimatord[11](factor1:nobathgroup,X0.
68+0.
76;bathgroup,X0.
68+0.
72;F1,1211.
565,p0.
005,d1.
702;factor2:nobathgroup,X0.
09+1.
09,bathgroup,X0.
09+0.
98;F1,120.
113,p0.
742,d0.
168).
Becauseofthemixeddesigninexper-iment2,weconductedseparatelinearmixedmodelanalysesusinganunstructuredcovariancematrixforeach(log-transformed)PCAfactor(gure2).
Forfactor1,theminimalmodelincludedsignicanteffectsofbathingtreatment,acousticstimulustypeandacousticstimuluspresentationorder(table2).
Forfactor2,therewasasignicantinteractioneffectforacousticstimulustypeacousticstimuluspresentationorder(table2).
Biol.
Lett.
(2012)8,379–381doi:10.
1098/rsbl.
2011.
1200Publishedonline16January2012Received8December2011Accepted13December2011379Thisjournalisq2012TheRoyalSocietyonMarch12,2014rsbl.
royalsocietypublishing.
orgDownloadedfromBathingappearedtoincreasefactor1scoresinexper-iment1anddecreasetheminexperiment2.
However,factor1weightingsforbothexperimentswerequalitat-ivelyequivalent(durationofrstfeedingboutaside):latencytofeedvariedpositivelywiththeaveragedurationofeachhead-upboutbutvariedinverselywiththehead-uprateperminuteandtheaveragedurationofeachhead-downbout(table1).
Hence,bathinghadqualitativelythesameeffectinbothexperiments.
4.
DISCUSSIONAccesstobathingwaterhadalargeandsignicanteffectonabehaviouralfactorthatcapturessensitivitytothreatincaptivestarlings.
Bathingcausedbirdstodecreasetheirlatencytofeed,decreasetheaveragedurationofeachhead-upscanningbout,increasetheaveragedur-ationofeachhead-downfeedingboutandincreasetheirhead-uprate.
Thus,birdsgivenaccesstobathingwaterweremorewillingtofeedinthefaceofanambiguousthreatperformingshorter,albeitmorefrequent,vigilancebouts.
Thisindicatestwopossibilities:eitherbirdsthatbathedinterpretedtheambiguousthreatsignalledbytheacousticstimuliasbeinglessdangerous,ortheyweremoremotivatedtomove/feed.
Thelatterisunlikelyasallbirdswerefedadlibitumuntilthedayoftesting.
Takingawaywaterbathsduringtestinginexperiment2alsoremovedanypotentialconfoundofmotivationtobatheinthegroupgivenaccesstobathingwater(thoughnosubjectsdidsoduringtestinginexperiment1).
Wearguethatbirdsdeniedaccesstobathingwaterinterpretedanambiguouscueofthreatasrequiringmorecautionthanbirdsthathadaccessbecausetheirabilitytocopewiththreatswasimpaired.
Thisisconsist-entwithighttrials[7],whichsuggestedthatbirdswithnoaccesstobathingwaterconsideredescapingfrompotentialthreattobemoreimportantthanavoidingphysicalharmfromcollisions.
Wetentativelyproposethatthendingsfrombothstudiesmaybeowingtodifferencesinfeatherconditioncausedbyacombinationofbathingandpreening.
Inanycase,theeffectofbathingmustbeshort-termasbathingwaterwasonlyremovedfor3daysandhadpreviouslybeenprovidedadlibitum.
Thealarmcallelicitedagreaterdefensiveresponsethanthe'threat'call,butthebathingmanipulationhadasignicanteffectontheresponsetoboth.
Aprioriwepredictedthatthe'threat'callwouldnotbeperceivedasasignofimminentphysicaldangersothebathingmanipulationshouldhavehadnoeffect.
Therearetwopossibilities:eitherthe'threat'callcontainssomeconno-tationofpotentialharm;orthebathingmanipulationmoregenerallychangedtheperceptionoftheexperimen-talcontext.
Previousexperimentsshowedthatstarlingsalsorespondmoredefensivelytowhitenoisethantothesame'threat'call[10].
Thus,itmaybetheexper-imentalcontextthatthebirdsperceiveasambiguouslythreatening,ratherthanthe'threat'callperse.
Further3.
02.
01.
0factor1scores0–1.
0–2.
0threatcallalarmcallFigure2.
Birdswithaccesstobathingwaterhadsignican-tlylowerfactor1scoresindicatingreducedvigilanceinexperiment2.
Notethattheseareuntransformedscores(log-transformedscoreswereemployedfortheanalysis).
Errorbarsrepresent95%CI.
Darkgreyboxrepresentsnoaccesstobathingwater;lightgreyrepresentsaccesstobathingwater.
1.
51.
00.
5factor1scores0–0.
5–1.
0–1.
5noaccesstobathingwateraccesstobathingwaterFigure1.
Birdswithaccesstobathingwaterhadsignicantlyhigherfactor1scoresindicatingreducedvigilanceinexper-iment1.
Errorbarsrepresent95%CI.
Table1.
Principalcomponentanalysisresultsforbothexperiments.
behaviouralmeasureexperiment1experiment2effectofbathingonfactor1(experiment1/experiment2)factor1loadingfactor2loadingfactor1loadingfactor2loadinglatencytofeed20.
64220.
2050.
77220.
155/head-upboutduration20.
6880.
6960.
91720.
077/head-uprate0.
83620.
07020.
3670.
903/head-downboutduration0.
8230.
47720.
39220.
453/durationofrstfeedingvisit0.
0140.
9180.
7450.
462/380B.
O.
Brilot&M.
BatesonBathingandthreatperceptionBiol.
Lett.
(2012)onMarch12,2014rsbl.
royalsocietypublishing.
orgDownloadedfromexperimentsarerequiredusingnoacousticstimulustoaddressthispossibility.
EuropeanUnionlegislationregardinglaboratorybirdsadvisesthatbathingwatershouldbeavailableeithercon-stantlyoronaregularbasis,dependingonthespeciesconcerned(revisedAppendixAofETS123,CouncilofEuropeConvention).
Morespecicguidelinesexistsuggestingtheconstantprovisionofshallowwaterbathsforstarlings[12].
Thisrecommendationisbasedonanotionthatbathingisimportantforfeathermain-tenanceandontheanecdotalobservationthatstarlingsareenthusiasticbathers.
However,of106researcharticlesfeaturingcaptivestarlings,only15reportedpro-visionofwaterforbathing[13].
Ourndingssuggestthatwhenbathsarenotprovided,starlingsmayhaveacontin-ualbiasintheirperceptionofambiguouslythreateningsituations(e.
g.
ambientnoises)arisingfromaperceivedincreaseintheirvulnerabilitytopredation.
Wethereforehypothesizethatlong-termlackofaccesstobathingwatermaybeacauseofchronicstressand/oranxiety-likesymptomsincaptivestarlings[14].
However,furtherexperimentsarerequiredinordertodemonstrateanypotentiallong-termimpact(e.
g.
permanentchangesinwillingnesstoalarmcall;changesinbaselineandstress-inducedcorticosteronelevels).
Whateverthelong-termconsequences,intheshort-termatleast,theprovisionofbathingwaterforstarlings(andarguably,otherwater-bathingbirdspecies)isclearlyofwelfareimportancegiventhelargeeffectofbathingwateravail-abilityonthreatperceptionthatwehavedemonstrated.
WeadheredtotheAssociationfortheStudyofAnimalBehaviour's"GuidelinesfortheUseofAnimalsinResearch".
TheexperimentalprotocolwasalsosubjecttointernalethicalreviewbytheNamedAnimalCareWelfareOfcer.
Thebirdswereinspected,fed,wateredandcleanedoutonadailybasisbytheexperimenter.
Nobirdsshowedlong-termalterationsintheirbehaviouraftertheexperimentalmanipulationsandonreleasebackintofree-ightaviariesshowednoothersignsofadverseeffects.
ThesubjectswereoriginallycapturedfromthewildunderlicencefromNaturalEngland.
Atthecloseofallexperimentsthebirdsreceivedahealthinspectionbyaqualiedvetandwerethenreleasedbacktothewildatthesiteoftheirinitialcapture.
WethankMichelleWaddle,MarkWhittingham,DanBlumsteinandananonymousreviewer.
ThisworkwassupportedbyagrantawardedtoM.
B.
bytheBiotechnologyandBiologicalSciencesResearchCouncil(BB/E012000/1).
1Simmons,K.
E.
L.
1964Feathermaintenance.
InAnewdictionaryofbirds(ed.
A.
L.
Thomson),pp.
278–286.
NewYork,NY:McGraw-Hill.
2Slessers,M.
1970Bathingbehavioroflandbirds.
Auk87,91–99.
3VanRhijn,J.
G.
1977Processesinfeatherscausedbybathinginwater.
Ardea65,126–147.
4Elowson,A.
M.
1984Spread-wingposturesandthewaterrepellencyoffeathers:atestofRijke'shypothesis.
Auk101,371–383.
5Oswald,S.
A.
,Bearhop,S.
,Furness,R.
W.
,Huntley,B.
&Hamer,K.
C.
2008Heatstressinahigh-latitudeseabird:effectsoftemperatureandfoodsupplyonbathingandnestattendanceofgreatskuasCatharactaskua.
J.
AvianBiol.
39,163–169.
(doi:10.
1111/j.
2008.
0908-8857.
04187.
x)6Murphy,S.
M.
,Braun,J.
V.
&Millama,J.
R.
2011Bathingbehaviorofcaptiveorange-wingedAmazonpar-rots(Amazonaamazonica).
Appl.
Anim.
Behav.
Sci.
132,200–210.
(doi:10.
1016/j.
applanim.
2011.
04.
010)7Brilot,B.
O.
,Asher,L.
&Bateson,M.
2009Waterbath-ingaltersthespeed-accuracytrade-offofescapeightsinEuropeanstarlings.
Anim.
Behav.
78,801–807.
(doi:10.
1016/j.
anbehav.
2009.
07.
022)8Bateson,M.
&Matheson,S.
M.
2007Removalofenvironmentalenrichmentinduces'pessimism'incaptiveEuropeanstarlings(Sturnusvulgaris).
Anim.
Welfare16,33–36.
9Matheson,S.
M.
,Asher,L.
&Bateson,M.
2008Larger,enrichedcagesareassociatedwith'optimistic'responsebiasesincaptiveEuropeanstarlings(Sturnusvulgaris).
Appl.
Anim.
Behav.
Sci.
109,374–383.
(doi:10.
1016/j.
applanim.
2007.
03.
007)10Brilot,B.
O.
,Normandale,C.
L.
,Parkin,A.
&Bateson,M.
2009Canweusestarlings'aversiontoeyespotsasthebasisforanovel'cognitivebias'taskAppl.
Anim.
Behav.
Sci.
118,182–190.
(doi:10.
1016/j.
applanim.
2009.
02.
015)11Nakagawa,S.
&Cuthill,I.
C.
2007Effectsize,con-denceintervalandstatisticalsignicance:apracticalguideforbiologists.
Biol.
Rev.
82,591–605.
(doi:10.
1111/j.
1469-185X.
2007.
00027.
x)12Hawkins,P.
etal.
2001Laboratorybirds:renementsinhusbandryandprocedures.
Lab.
Anim.
35,S1–S163.
13Asher,L.
&Bateson,M.
2008UseandhusbandryofcaptiveEuropeanstarlings(Sturnusvulgaris)inscienticresearch:areviewofcurrentpractice.
Lab.
Anim.
42,111–126.
(doi:10.
1258/la.
2007.
007006)14Bateson,M.
,Brilot,B.
O.
&Nettle,D.
2011Anxiety:anevolutionaryapproach.
Can.
J.
Psychiat.
56(12),707–715.
Table2.
Linearmixedmodelanalysisresultsforexperiment2.
signicantmodelterms(minimalmodel)afactor1factor2F-ratio(d.
f.
)coefcientestimatebp-valueF-ratio(d.
f.
)coefcientestimatebp-valuebathing17.
0621,7.
420.
2210.
004c0.
0041,70.
0020.
953acousticstimulustype24.
2961,80.
1930.
001c5.
0891,720.
1750.
059bathingacousticstimulustype0.
7771,80.
0840.
4040.
0011,70.
0010.
978acousticstimuluspresentationorder14.
3211,720.
1470.
007c4.
4551,720.
2340.
073acousticstimulustypeacousticstimuluspresentationordern.
s.
andexcludedfrommodel10.
2461,720.
2050.
015caFullmodelincluded:acousticstimulustype,bathingtreatmentandacousticstimulipresentationorder,andalltwo-wayinteractions.
Termswereremovedsequentiallybyhighestp-value,buttheexperimentalfactorsandtheirinteractionwereretained.
bCoefcientcomparisonsformaineffectsaregivenas:nobathingwaterversusbathingwater;threatcallversusalarmcall;alarmcallheardsecondversusalarmcallheardrst.
cIndicatessignicanceatthea0.
05level.
BathingandthreatperceptionB.
O.
Brilot&M.
Bateson381Biol.
Lett.
(2012)onMarch12,2014rsbl.
royalsocietypublishing.
orgDownloadedfrom