Oxytocin reactivity during intergroup conflict in wild chimpanzees


Чтобы посмотреть этот PDF файл с форматированием и разметкой, скачайте его и откройте на своем компьютере.
Oxytocinreactivityduringintergroupconflictin
wildchimpanzees
LiranSamuni
a,b,1
,AnnaPreis
a,b
,RogerMundry
a
,TobiasDeschner
a
,CatherineCrockford
a,b,2
,andRomanM.Wittig
a,b,1,2
a
DepartmentofPrimatology,MaxPlanckInstituteforEvolutionaryAnthropology,04103Leipzig,Germany;and
b
TaChimpanzeeProject,CentreSuissede
RecherchesScientifiques,Abidjan01,IvoryCoast
EditedbyDorothyL.Cheney,UniversityofPennsylvania,Philadelphia,PA,andapprovedDecember1,2016(receivedforreviewOctober12,2016)
Intergroupconflictisevidentthroughoutthehistoryofourspecies,
evolutionofhumans

large-scalecooperativenature.Likehumans,
narysupportduringviolentintergroupconflicts.Inbothspecies,
cooperationamonggroupmembersisessentialforindividualsto
gainaccesstobenefitsfromengaginginintergroupconflict.Stud-
iessuggestthatacontributivemechanismregulatingin-groupco-
operationduringintergroupconflictsinhumansinvolvesthe
neuropeptidehormoneoxytocin,knowntoinfluencetrust,coor-
dination,andsocialcognition,althoughevidencefromnatural
vestigateoxytocinergicsysteminvolvementduringnaturalinter-
groupconflictsinwildchimpanzees.Wefoundthatchimpanzees
ofbothsexeshadsignificantlyhigherurinaryoxytocinlevelsim-
mediatelybeforeandduringintergroupconflictcomparedwith
controls.Also,elevatedhormonelevelswerelinkedwithgreater
cohesionduringintergroupconflicts,ratherthanwiththelevelof
potentialthreatposedbyrivalgroups,intragroupaffiliativesocial
interactions,orcoordinatedbehavioralone.Thus,theoxytociner-
gicsystem,potentiallyengenderingcohesionandcooperation
whenfacinganout-groupthreat,maynotbeuniquelyhuman
butratheramechanismwithevolutionaryrootssharedbyour
lastcommonancestorwithchimpanzees,likelyexpeditingfitness
gainsduringintergroupconflict.
Pantroglodytes
|
cooperation
|
groupcohesion
|
neuropeptide
|
parochialaltruism
R
ecentevolutionarymodelssuggestthatparochialaltruism,
othersatacosttooneself,iskeytounderstandingtheevolution
ofhumans

cooperativetraitsandpropensityforintergroupvi-
olence(1,2).Intergroupconflictisubiquitousacrosshuman
dice,war,andgenocide(2,3).Individualscontributetothese
patternsbothbysupportingin-groupmembersandactingwith
hostilitytowardtheout-group.Whensuchacombinationcon-
tributestosuccessinintergroupconflicts,parochialaltruismcould
haveevolved(1),andbiologicalmechanismsthatsustainand
promoteitarelikelyadaptive(4).
Onesuchproposedbiologicalmechanisminvolvestheneu-
ropeptidehormoneoxytocin,previouslylinkedwithvarious
aspectsofhumansociality,particularlythedevelopmentofmother

offspringbonds,butalsotolerance,coordination,andcooperation

7).Owingtoitsanxiolyticandprosocial
effects,oxytocinisproposedtofacilitatecooperationduringrisk,a
mechanismpotentiallyco-opted
frommaternaldefensecircuitry
(4).Intranasaladministrationof
oxytocinenhancesin-groupco-
operationandtrust(8,9)andout-groupdefensive,butnotoffen-
sive,competitioninmen(8).Thissuggeststhatoxytocintriggersa

tendanddefend

formofparochialaltruism,accentuatingco-
operativebehaviortowardthein-groupaswellasdefensivebe-
haviortowardout-groups(4).However,theseresultswereobtained
focusingonhumanmaleparticipants.Few,ifany,studieshave
involvedintergroupcontextsandoxytocinincaptiveorwildnon-
humananimals.Therefore,addition
alevidenceisessentialforcor-
roboratingoxytocinergicsysteminvolvementinanecologically
Wildchimpanzeesinalmostalllong-termfieldsitesengagein
MovieS1
),intergroupencountersand
borderpatrols.Intergroupencounters(directout-groupcontact)
arecharacterizedbycoordinatedattacks,withsynchronousvo-
calizationsandchargestowardandcombatagainstchimpanzees
fromrivalgroups(12,13).Inborderpatrols(nodirectout-group
contact),chimpanzees

typicalforagingandtravelingmovements
scouttheperipheralareasoftheirterritory,oftencontinuingfor
mal.Travelisslow,ofteninsinglefile,interruptedbyfrequent
pausesinwhichchimpanzeesmaysniffforestitems,suchasout-
groupchimpanzeefecesorfoodremnants,andareunusually
alerttosoundsbeyondtheimmediategroup(12,13).Individuals
appeartosearchforsignsindicativeofrecentrival-groupchim-
panzeepresence,suchasvocalpresence,orrecentphysicalpres-
ence,potentiallyassessingthestrengthoftheiropponents.
reducedforagingandincreasedtraveling(14),andrisky,asit
mayleadtoinjuryordeath(10).Successfulattacksonrivalgroups,
however,potentiallyincreasetheterritorysizeofthein-groupand
thereproductiveoutputofitsmembe
rs(14,15),therebyincreasing
Significance
Warfareisoneofthemostpervasiveproblemsamonghuman
cooperationandfavoritismisofparamountimportance.Wild
chimpanzeessharekeyfeaturesofhumans

intergroupconflict,
intermsofin-groupcoordination,coalitionarysupport,and
out-grouphostility.Thehormoneoxytocinmayregulatehu-
mans

lacking.Wefoundstrongevidencethat,likeinhumans,oxy-
tocinisinvolvedinchimpanzeeintergroupconflict.Bothin-
tergroupconflictanticipationandparticipationinvolvedhigh
urinaryoxytocinlevels,irrespectiveofintragroupaffiliationsor
potentialthreatbyrivals.Theseresultsareindicativeofsimilar
physiologicalprocessesinvolvedinintergroupviolenceand
intragroupsupportinbothspecies,likelysupportingbehavior
thatisadaptiveduringintergroupconflicts.
Authorcontributions:L.S.,T.D.,C.C.,andR.M.W.designedresearch;L.S.andA.P.per-
formedresearch;L.S.,A.P.,R.M.,C.C.,andR.M.W.analyzeddata;andL.S.,T.D.,C.C.,and
R.M.W.wrotethepaper.
Theauthorsdeclarenoconflictofinterest.
ThisarticleisaPNASDirectSubmission.
1
Towhomcorrespondencemaybeaddressed.Email:[email protected]@
eva.mpg.de.
2
C.C.andR.M.W.contributedequallytothiswork.
Thisarticlecontainssupportinginformationonlineat
www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1616812114/-/DCSupplemental
.
www.pnas.org/cgi/doi/10.1073/pnas.1616812114
PNASEarlyEdition
|
1of6
ANTHROPOLOGY
Thisstudyusedawithin-subjectsdesign,requiringrepeated
samplingofthesamewildchimpanzeeswhenengagedindiffer-
entnaturalevents.Weachievedthisbyusingperipheraloxytocin
measuresthatcouldbenoninvasivelycollected.Thebiological
validityofperipheraloxytocinmeasurementswithrespecttocentral
oxytocinpatternsnonet
helessisdebated(33

35).However,anin-
creasingbodyofevidenceshowsthatoxytocinpathwayscaninvolve
coordinatedcentralandperiphera
loxytocinrelease,indicatingthat
highperipheralmeasuresreflectthereleaseofcentraloxytocin.
Furthermore,anincreasingnumberofstudiesdemonstratethe
sameeffectsofbehavior,socialco
ntext,andsocialrelationshipson
bothcentralandperipher
aloxytocinmeasures(35

37).
Usinganecologicallyrelevantparadigm,wefoundthatthe
oxytocinergicsystem,ahighlycons
ervedphysiologicalmechanism
(5),isinvolvedincohesionduringintergroupconflictinchim-
panzees,assuggestedforhumans(8).Giventhatin-groupfavor-
itisminhumansandchimpanzeesmaybeunderpinnedbythesame
physiologicalmechanism,themos
tparsimoniousexplanationfor
suchsimilaritiesisthatthismechanismwaspresentinourcommon
ancestor,regulatingin-groupbias
.Accordingly,itmaybethatsome
aspectsthoughttoplayaroleinhumanparochialaltruismreston
moreancientevolutionaryoriginsthanhasbeenpresumed.
Thefundamentalneedforwithin-groupsupportintimesof
entlyalsopresentinoneofourclosestlivingrelatives,thechim-
groupdivision,categorization,andattributionclearlyisacurrent
andpressingtopic(3,38).Understandingtheevolutionarymech-
anismsunderlyingin-groupout-groupinteractions,thepressures
thatswitchintergroupcollaborat
iontoconflictandviceversa,and
mayeventuallyassistthebuildingofcooperationratherthande-
FieldworkwasconductedwiththeTaChimpanzeeProjectlocatedattheTa
NationalPark,IvoryCoast(552

N,720

habituatedEastandSouthneighboringchimpanzee(
P.t.verus
)groups.We
conductedall-dayfocalanimalsampling(23)on20individuals(5malesand
5parousfemalesineachgroup)foratotalof2,278observationhoursinEast
groupand2,164inSouthgroup,alongwithnoninvasivelycollectedurine
samples(analysisincluded
n
=
482samples,23.4

14.55samplesperindi-
vidual).Duringfocalfollows,wedocumentedchangesinthebehavior,social
interactions,andvocalizationsemittedbyanddirectedtowardthefocal
individual,usingCyberTrackersoftware(version3.389;
www.cybertracker.
org/
).Wecontinuouslyupdatedthesubgroupcompositionandsize.Every
occurrenceofaborderpatroloranintergroupencounterwasrecordedad
libitum.Duringthestudyperiod,67instancesofintergroupconflictwere
observedinEastgroup,arateof1every5d,outofwhich28involvedbotha
borderpatrol(withnodirectout-groupcontact)andanintergroupen-
counter(withdirectout-groupcontact;42%;1every11d);25instancesof
intergroupconflictwereobservedinSouthgroup,arateof1every12d,out
ofwhich17involvedbothaborderpatrolandanintergroupencounterand
asingleinstanceinvolvingonlyanintergroupencounter(72%;1every16d).
withineachgroup,weusedtheElo-rating(39),basedonunidirectional
submissivepantgruntvocalizations(40).Wecontinuouslyrecordedthelo-
cationofthefocalsubjectusingaGarminRino610globalpositioningsystem
forchangesinthechimpanzee

sendocrinologicalresponseinrelationtoproxi-
mitytoperipheralterritorialareasand,thus,thepotentialtoencounterrival
chimpanzeegroups.Wethenassessedtheproximityofthefocalindividual
totheborderareasoftheterritory.Weusedakerneldensityestimate(41)in
R(version3.2.3)(42)toconstructpolygonsrepresentingthepercentageof
home-rangeusekernelsrangingfrom5to99,with5representingthevery
coreofthehomerangeand99beingtheborderarea(
Fig.S4
).
Asanestimateofdefection,wemeasuredforeachadultindividualthe
numberoftimesitleftthesubgroupofthefocalindividualduringinstances
ofprolongedintergroupconflicts(Eastgroup
n
=
21;Southgroup
n
=
2;
duration,mean

SD102.4

41.48min;Table1).Allintergroupconflict
periodsusedinthismodelincludedborderpatrolbehavior.Whereassome
periodsincludeddirectcontactwithrivalgroupsandwerelabeledasin-
tergroupencounters(
n
=
11),othersdidnotincludeout-groupcontactand
werelabeledasborderpatrols(
n
=
12).Wedefinedseparatefissionevents
asanyadultwholeftthesubgrouppermin,suchthatleavesthatoc-
curred

1minapartwerecountedasseparatefissionevents(Table1).Ac-
cordingly,
n
=
25fissioneventsoccurredduringintergroupconflictsatarateof
1every96min(fissionsperperiod,mean

SD1.089

1.276,with3.64

2.36
individualsleavingduringeachfissionevent).Wecomparedthiswithmatched
controlperiodsofthesamedurationandwithin1to3dbeforeorafterthe
intergroupconflict(Eastgroup
n
=
21;Southgroup
n
=
2),ondaysthatdidnot
includeintergroupconflictsorhuntingbehavior,andwithsimilarsubgroup
sizeandcomposition.Atotalof
n
=
65fissioneventsoccurredduringmatched
controlperiodsatarateof1every37min(fissionsperperiod,mean

SD
2.826

2.145,with4.8

3.23individualsleavingduringeachfissionevent).
UrineSampleCollectionandAnalysis.
Wetooktheclearanceofoxytocininto
ahumanclearancestudy(43).Urinaryoxytocinmeasuresshowbio-
behaviorallyrelevantlevelsfollowingtargetbehaviorsorsocialinteractions
thatoccurwithinthistimewindow(19,20).Samplecollection,extraction,
(19),incorporatingminorchanges(
).Analysiswasdoneusinga
commerciallyavailableenzymeimmunoassaykit(AssayDesigns;901-153A-
0001;
).Wemeasuredcreatininelevelsinallurinesamplesand
expressedurinaryoxytocinvaluesaspg/mgcreatinine,tocontrolforvaria-
tioninurinevolumeandconcentration(44).Becauseverylowcreatinine
valuesmayleadtooverestimationofurinaryoxytocinlevels,weexcludedall
urinesampleswithcreatininelevels

0.04mg/mL(
n
=
5,

1.2%ofthe
samplesincluded).
PlanckInstituteforEvolutionaryAnthropologyprimatologydepartment
).
StatisticalAnalysis.
Weconductedaseriesoflinearmixedmodels(24)with
Gaussianerrorstructureandidentitylinkfunction,andaPoissongeneralized
linearmixedmodel(24)withloglinkfunctioninR[version3.3.0(42)],using
thefunctionslmerandglmeroftheRpackagelme4(45).Ineachmodel,we
includedfactorsthatmightinfluencehormonelevels(asdescribedabove;
DatasetsS1

S3
).Furthermore,tokeeptypeIerrorrateatthenominal5%,we
includedrandomslopes(46,47)(
SIMethods
).Wecomparedthefitofthefull
modelswiththoseofarespectivenullmodellackingonlythetestpredictorsof
eventorperiodtypebutotherwiseidenticaltotherespectivefullmodelinall
otherterms(48),usingalikelihoodratiotest(
SIMethods
).
ACKNOWLEDGMENTS.
WethanktheCentreSuissedeRecherchesScientifi-
quesandstaffmembersoftheTaChimpanzeeProjectforsupport,Christophe
Boesch,LindaVigilant,andKevinLangergraberaswellasAlexanderMielke,
MichalSamuni-Blank,andtwoanonymousreviewersfortheirhelpfulcom-
ments.ThisstudywasfundedbytheMinervaFoundation(L.S.),Leakey
1.ChoiJ-K,BowlesS(2007)Thecoevolutionofparochialaltruismandwar.
Science
318(5850):636

640.
2.BowlesS(2009)Didwarfareamongancestralhunter-gatherersaffecttheevolutionof
humansocialbehaviors?
Science
324(5932):1293

1298.
3.CohenTR,InskoCA(2008)Warandpeace:Possibleapproachestoreducingin-
tergroupconflict.
PerspectPsycholSci
3(2):87

93.
groups:Anintegrativereviewandresearchagenda.
HormBehav
61(3):419

428.
ciality.
Science
322(5903):900

904.
6.InselTR(2010)Thechallengeoftranslationinsocialneuroscience:Areviewofoxy-
tocin,vasopressin,andaffiliativebehavior.
Neuron
65(6):768

779.
coordinationandcooperation.
JCognNeurosci
25(9):1418

1427.
8.DeDreuCKW,etal.(2010)Theneuropeptideoxytocinregulatesparochialaltruismin
intergroupconflictamonghumans.
Science
328(5984):1408

1411.
9.IsraelS,WeiselO,EbsteinRP,BornsteinG(2012)Oxytocin,butnotvasopressin,increases
bothparochialanduniversalaltruism.
Psychoneuroendocrinology
37(8):1341

1344.
Pan
strategiesthanhumanimpacts.
Nature
513(7518):414

417.
PNASEarlyEdition
|
5of6
ANTHROPOLOGY
contactwithrivalgroupsoccurredwithinthetimewindowfor
encounters,andthosethatdidnotincludeanycontactasborder
patrols.Bothtypesofintergroupconflicthadasignificantposi-
tiveeffectonlog-transformedurinaryoxytocinlevels(pg/mg
creatinine),incontrasttocontrolsituationswithnopositive
socialinteractions[

type

model;linearmixedmodel(LMM),
likelihoodratiotest:

2
=
44.600,df
=
2,
P

0.001;
Fig.S1
and
TableS2
].However,log-transformedurinaryoxytocinlevelsdid
TableS3
;
see
fortestsseparatingthesetwocontexts).Thus,
althoughborderpatrolsrepresentonlyacovertout-groupthreat,
theyaresubjecttosimilarurinaryoxytocinandcohesionlevelsas
timeperiodsincludingintergroupencounters.Hence,wecom-
binethemasonecontextinthefollowingeventmodel(see
SI
,
Fig.S2
,and
TableS4
fortestsonlyincludingintergroup
encounters).
Herewetestedtheimpactofdifferenteventsandconfounding
factorsonurinaryoxytocinlevels(eventmodel).Becauseaffili-
ationfrequentlyoccursduringin-groupout-groupcontextsand
mightimpacturinaryoxytocinlevels(19),wedividedintergroup
conflictsamplesintotwocategories:groupmembersparticipat-
inginintergroupconflict(
i
)withoutin-groupaffiliation(11
subjects,
n
=
103samplesof37events),or(
ii
)within-group
affiliation(groomingorplaywithmultiplepartners;15subjects,
n
=
64samplesof24events).Wecontrastedintergroupconflict
withthreecontroleventsexcludingintergroupconflictorfood
sharing,asthelattershowsassociationwithhighoxytocinlevels
inchimpanzees(20):(
i
)90-mintimeperiodsinwhichnopositive
socialinteractions,exceptvocalizations,occurred(

controlwith-
outaffiliation

;20subjects,
n
=
178samplesof150events);(
ii
)
multipartnergroomingofatleast10-minduration(

controlwith
affiliation

;19subjects,
n
=
100samplesof87events);and(
iii
)
participationingrouphuntingofmonkeys,acoordinatedbehavior
(22)thatdoesnotinvolvein-groupout-groupcontexts(

control
withcoordination

;9subjects,
n
=
23samplesof17events).The
twolattercontextsallowedustocontrolforthein-groupaffiliative
andcoordinatedbehavioroftenobservedduringintergroupcon-
atleast15minafterthestartandupto60minaftertheendof
interactions(19,20).WefittedanLMM(eventmodel)(24)totest
fortheinfluenceofintergroupconflict,coordination,andaffilia-
tiononlog-transformedurinaryoxytocinlevels(pg/mgcreatinine).
Tocontrolforotherfactorsthatmightinfluencehormonelevels,
weincludedindividuals

sexandrank,subgroupsize,groupiden-
tity,andproximitytoborderareastoevaluatepotentialrisk.Our
ferentindividualsfrom296differentevents.
Overall,thefull-nullmodelcomparisonwassignificant(LMM,
likelihoodratiotest:

2
=
51.253,df
=
4,
P

0.001;Fig.1
A
and
Table2).Morespecifically,intergroupconflictswithandwithout
affiliationwereassociatedwithhigherurinaryoxytocinlevels
thanthethreecontrols(Table2and
TablesS5
,
S6
,and
S7
).We
alsofoundapositiveeffectofgroupcoordinatedhuntingbe-
havioronurinaryoxytocinlevelscomparedwithboththecontrol
withandwithoutaffiliation,althoughlesspronouncedthanthe
effectofintergroupconflict(
TableS7
).Theseeffectswerenei-
therdrivenbyindividualrankorsex,subgroupsize,norprox-
imitytoborderareas.However,wefoundagroupeffect,with
EastgrouphavinghigherurinaryoxytocinlevelsthanSouth
group(Table2),despitehavingsimilargroupsizesandlittle
posthocanalysescomparingintergroupconflictwithandwithout
affiliation,wefoundnosignificanteffectofaffiliationonurinary
oxytocinlevelswithinthiscontext(LMM:

2
=
0.136,df
=
1,
P
=
0.334;
TableS5
).
Whenfacingrivalgroups,chimpanzeein-groupbehaviorwas
positivelylinkedwithurinaryoxytocinlevels.Thiswastrueeven
whenaccountingforaffiliativeinteractionsandpotentialthreat
fromrivalgroups,suggestingthat,similartohumans,theoxy-
tocinergicsystemisaninfluentialmechanisminvolvedinchim-
panzeein-groupout-groupcontexts.Thestimulusthattriggers
oxytocinreleaseinintergroupcontexts,however,remainsun-
knownforeitherhumansorchimpanzees.Affiliativecontacthas
beenproposedasanoxytocintrigger(26),butourresultsconcur
withotherstudies,ofbothhumansandchimpanzees,suggesting
thatphysicalcontactisnotnecessarilyrequired(20,27)norsuf-
affiliationduringintergroupconflictnormultipartneraffiliation
withoutintergroupconflict(Table2)ledtourinaryoxytocinlevels
thatdifferedfromnonaffiliativeintergroupconflictorcontrol
samples,respectively.Thisisinagreementwithrecentevidence
thatthemereactofgroomingisnotlinkedwithanoxytocin
Urinary oxytocin pg/mg creatinine
10
100
1000
n = 178n = 100n = 23n = 103n = 64
ControlControlControlIntergroup ConflictIntergroup Conflict
with Affiliation
without Affiliation
with Affiliation
without Affiliationwith Coordination
***
*
**
*
10
100
1000
n = 38n = 14
Control
with Affiliation
Preborder Patrol
with Affiliation
***
AB
Fig.1.
(
A
)Effectsofintergroupconflictwithandwithoutin-groupaffiliationonurinaryoxytocinlevelsinwildchimpanzeesinEastandSouthgroups(
n
=
468samples,20subjects,296events).(
B
)EffectsofimminentintergroupconflictinEastgroupchimpanzeesonurinaryoxytocinlevels(
n
=
52samples,
9subjects,43events).Shownaremedians(thinhorizontallines),quartiles(boxes),percentiles(2.5and97.5%;verticallines),minimumandmaxi
mum(laying
crosses),aswellasthefittedmodelandits95%confidenceintervals(thicklines).***
P

0.001,**
P

0.01,*
P

0.05.
PNASEarlyEdition
|
3of6
ANTHROPOLOGY
AmJPrimatol
70(6):519

532.
12.BoeschC,Boesch-AchermannH(2000)
TheChimpanzeesoftheTaForest:
BehaviouralEcologyandEvolution
(OxfordUnivPress,NewYork).
13.WattsDP,MitaniJC(2001)Boundarypatrolsandintergroupencountersinwild
chimpanzees.
Behaviour
138(3):299

327.
14.WilsonML(2013)Chimpanzees,warfare,andtheinventionofpeace.
War,Peace,and
HumanNature:TheConvergenceofEvolutionaryandCulturalViews
,edFryDP
(OxfordUnivPress,NewYork),pp361

388.
expansioninwildchimpanzees.
CurrBiol
20(12):R507

R508.
16.AxelrodR,HamiltonWD(1981)Theevolutionofcooperation.
Science
211(4489):
1390

1396.
ingandfuturechallenges.
EvolAnthropolIssuesNewsRev
18(5):215

227.
18.MawsonAR(2005)Understandingmasspanicandothercollectiveresponsestothreat
anddisaster.
Psychiatry
68(2):95

113.
relatedwildchimpanzees.
ProcBiolSci
280(1755):20122765.
20.WittigRM,etal.(2014)Foodsharingislinkedtourinaryoxytocinlevelsandbonding
inrelatedandunrelatedwildchimpanzees.
ProcBiolSci
281(1778):20133096.

fusiondynamics:Newresearchframeworks.
Curr
Anthropol
49(4):627

654.
22.BoeschC(1994)Cooperativehuntinginwildchimpanzees.
AnimBehav
48(3):
653

667.
Behaviour
49(3):227

267.
24.BaayenRH(2008)
AnalyzingLinguisticData:APracticalIntroductiontoStatistics
UsingR
(CambridgeUnivPress,Cambridge,UK),1stEd.
neighboringchimpanzeecommunities.
CurrBiol
22(10):922

926.
26.LimMM,YoungLJ(2006)Neuropeptidergicregulationofaffiliativebehaviorand
socialbondinginanimals.
HormBehav
50(4):506

517.
27.SeltzerLJ,ZieglerTE,PollakSD(2010)Socialvocalizationscanreleaseoxytocinin
humans.
ProcBiolSci
277(1694):2661

2666.
28.SobolewskiME,BrownJL,MitaniJC(2012)Territoriality,toleranceandtestosterone
inwildchimpanzees.
AnimBehav
84(6):1469

1474.
inbonobosandchimpanzees.
ProcNatlAcadSciUSA
107(28):12457

12462.
30.DurrantR(2011)Collectiveviolence:Anevolutionaryperspective.
AggressViolent
Behav
16(5):428

436.
31.RuschH(2014)Theevolutionaryinterplayofintergroupconflictandaltruisminhu-
mans:Areviewofparochialaltruismtheoryandprospectsforitsextension.
ProcBiol
Sci
281(1794):20141539.
32.TenVeldenFS,DaughtersK,DeDreuCKW(June8,2016)Oxytocinpromotesintuitive
ratherthandeliberatedcoop
erationwiththein-group.
HormBehav
,10.1016/
j.yhbeh.2016.06.005.
33.LengG,LudwigM(2016)Intranasaloxytocin:Mythsanddelusions.
BiolPsychiatry
79(3):243

250.
34.MacdonaldK,FeifelD(2013)Helpingoxytocindeliver:Considerationsinthedevel-
opmentofoxytocin-basedtherapeuticsforbraindisorders.
FrontNeurosci
7:35.
35.McCulloughME,ChurchlandPS,MendezAJ(2013)Problemswithmeasuringpe-
ripheraloxytocin:Canthedataonoxytocinandhumanbehaviorbetrusted?
Neurosci
BiobehavRev
37(8):1485

1492.
36.CrockfordC,DeschnerT,ZieglerTE,WittigRM(2014)Endogenousperipheraloxy-
tocinmeasurescangiveinsightintothedynamicsofsocialrelationships:Areview.
FrontBehavNeurosci
8:68.
37.RossHE,YoungLJ(2009)Oxytocinandtheneuralmechanismsregulatingsocial
cognitionandaffiliativebehavior.
FrontNeuroendocrinol
30(4):534

547.
38.TroppLR,ed(2012)
TheOxfordHandbookofIntergroupConflict
(OxfordUnivPress,
Oxford),1stEd.
ofprogressiveevaluationwithElo-rating.
AnimBehav
82(4):911

921.
femalechimpanzeesoftheTaNationalPark.
IntJPrimatol
24(4):847

867.
rangestudies.
Ecology
70(1):164

168.
42.RCoreTeam(2016)R:ALanguageandEnvironmentforStatisticalComputing
(RFoundStatComput,Vienna).https://www.R-project.org/.
43.AmicoJA,UlbrechtJS,RobinsonAG(1987)Clearancestudiesofoxytocininhumans
usingradioimmunoassaymeasurementsofthehormoneinplasmaandurine.
JClin
64(2):340

345.
44.BahrNI,PalmeR,MhleU,HodgesJK,HeistermannM(2000)Comparativeaspectsof
Gen
CompEndocrinol
117(3):427

438.
45.BatesB,MchlerM,BolkerBM,WalkerSC(2015)Fittinglinearmixed-effectsmodels
usinglme4.
JournalofStatisticalSoftware
67(1):1

48.
46.BarrDJ,LevyR,ScheepersC,TilyHJ(2013)Randomeffectsstructureforconfirmatory
hypothesistesting:Keepitmaximal.
JMemLang
68(3):255

278.
matesinmixedmodels.
BehavEcol
20(2):416

420.
Overestimatedeffectsizesandthewinner

scurse.
BehavEcolSociobiol
65(1):47

55.
efficients.
1(2):103

113.
50.FoxJ,WeisbergS(2010)
AnRCompaniontoAppliedRegression
(SAGE,Thousand
Oaks,CA).
51.QuinnGP,KeoughMJ(2002)
ExperimentalDesignandDataAnalysisforBiologists
(CambridgeUnivPress,Cambridge,UK).
6of6
|
www.pnas.org/cgi/doi/10.1073/pnas.1616812114
fitness.Accesstobenefitsofintergroupconflictcanbemain-
tainedthroughcooperation,providingbenefitstobothactor
andrecipientregardlessofpotentialshort-termcoststothe
actor(16).Chimpanzeein-groupbehaviorduringintergroup
conflictisconsideredtobecooperative(11,17),asitencompasses
prolongedcoordinationandcohesion,aswellascoalitionary
support,whichcaninvolveindividualssafefromtheout-group
runningforwardtodefendanin-groupmemberunderattackfrom
theout-group(11).Inhumans,cohesionandsupportalsoin-
tensifyunderlife-threateningsituations,suchaswar(18).These
parallelshighlightkeyfeaturessharedbychimpanzeeandhuman
groupdefenseandintergroupconflict(14).Alsosimilartohu-
mans,theoxytocinergicsysteminchimpanzeesisinvolvedin
groupmembers,suchasgrooming(19)andfoodsharing(20),
suggestingparallelsinoxytocinergicsysteminvolvementbetween
humansandchimpanzees.

in-groupbehavior
duringborderpatrolsandintergroupencounters(intergroup
conflict)enhancesgroupcohesionandinvolvestheoxytocinergic
zeeintergroupinteractionsandinaccordancewiththehuman
literature,weusetheterminte
rgroupconflict(2)todescribe
borderpatrolsandintergroupencounters.Weassessed(
i
)

groupcohesionisgreaterduringin-
tergroupconflictsthancontrolperiods(

defection

model).
Chimpanzeesliveinafission

fusionsocialsysteminwhichin-
dividualsfromthesamegroupsplitintosmallanddynamic
subgroupsofvaryingsize,composition,andduration(12).Ac-
cordingly,wemeasuredfissionsofadultindividualsasanesti-
mateofdefection.Fissionmaybeaffectedbyecological,social,
quirecohesion,weexpectedfewerfissionsduringintergroup
conflictsthanduringcontrolperiods,regardlessofsubgroupsize
andcomposition,whichmayaffectfissionpatterns,orproximity
toterritoryborderareas,whereintergroupencountersoften
occur.Wealsoinvestigatedwh
borderpatrolsandintergroupencountersengagedtheoxy-
tocinergicsystem.Theoxytocinergicsysteminfluencesattributes
likelytoassistcooperationandhencesuccessfulintergroup
conflict,suchasin-grouptrustandcoordination(5

7).Accord-
ingly,wehypothesizedthathighoxytocinlevelsimmediately
beforeandduringintergroupconflictswouldbeadaptivewhen
influencinggroupcooperation.Weexpectedboth(
ii
)high
oxytocinlevelsduringintergroupconflicts(

event

model)
and(
iii
)highanticipatoryoxytocinlevelsbeforeborderpatrol
initiation(

anticipation

model).Weexpectedhighoxytocin
levelstopersist,evenwhencontrollingfortheoccurrenceof
in-groupaffiliativebehavior,proximitytoborderareas,orbe-
haviorinvolvingin-groupcoordinationintheabsenceofout-
groupthreat,specificallyhuntingeventswherechimpanzees
coordinatetocapturemonkeys(22).
Weinvestigatedourhypothesesusingawithin-subjectsdesign,
samplingnaturallyoccurringeventsduringintra-andintergroup
interactionsinwildchimpanzees(
Pantroglodytesverus
)inthe
TaNationalPark,IvoryCoast.Weconductedfocalanimal
sampling(23)for20adultmaleandfemalechimpanzeesoftwo
neighboringgroups,andmeasuredtheoxytocinconcentrationof
samplespecificevents(19,20).
ResultsandDiscussion
Aswithhumans,chimpanzeegroupdefenserequirescoordina-
tionandcoalitionarysupporttobeeffective(11,17),andwhen
suchcooperativebehaviorismaintained,accesstobenefitsis
olenceoccurspredominantlyattimesofpowerimbalancein
favorofattackers(10),andthusgroupcohesionmayreduce
thelikelihoodofsufferingcosts.Toinvestigatetheinfluenceof
intergroupconflictongroupcohesion(defectionmodel),we
countedfissionsperindividualasanestimateofdefection(Table
conflict(
n
=
23),orduringamatchedcontrolperiod(
n
=
23).
Controlsweredefinedasperiodswithsimilarduration,subgroup
size,andcompositionondaysthatdidnotincludeintergroup
conflictsorhuntingbehavior.Wefoundsimilarnumbersof
leavesperindividualduringintergroupconflictperiodsthatin-
cludeddirectinteractionswithrivalgroups(intergroupencoun-
ters
n
=
11)andthosethatdidnot(borderpatrols
n
=
12)(Table
1).Accordingly,wefittedaPoissongeneralizedlinearmixed
bothtypesofintergroupconflictandcontrols.Wecontrolledfor
groupidentity,proximitytoborderareaswhereencounterswith
rivalsaremorelikely,andsubgroupduration.
Wefoundthatduringintergroupconflicts,individualswere
significantlylesslikelytoleavethesubgroupthanduringcontrol
periods(GLMM,likelihoodratiotest:

2
=
13.484,df
=
1,
P

0.001;estimate

SE

1.732

0.446;
TableS1
).Thiseffectwas
notdrivenbyproximitytoborderareas,whereencounterswith
rivalsaremorelikely.Wealsofoundaneffectofgroupidentity
(
TableS1
),withSouthgroupindividualsbeingsignificantlyless
likelytoleavethesubgroupthanEastgroupindividuals;however,
thisparameterwasassociatedwithsomeinstability.Contextual
variationindefectionnumbersshowedthat,incomparisonwith
controlcontexts,intergroupconf
lictpromotedgroupcohesion.
Similarly,humansoldiersincreasetheircohesionandaffiliation
whengoingintobattle(18).Althoughcohesionisfundamentalin
tergrouphostilitieshavecontrib
utedtotheproliferationofchim-
panzeecooperativecapacities,asistheorizedforhumans.
panzeein-groupactivityduringborderpatrolsandintergroup
encountersengagedtheoxytocinergicsystem.Incaseadirect
Fissionevents*Individualfissions

Subgroupsize

Proportionof
adultleaves

Events
n
MeanSD
n
MeanSDMeanSDMeanSD
Control,
n
=
23652.8262.1453121.0790.97410.0433.2680.6970.329
Intergroupconflict,
n
=
23251.0891.276910.3650.69410.3043.0960.2570.279
Borderpatrol,
n
=
12161.4540.934440.3820.5399.3332.2690.3410.234
Intergroupencounter,
n
=
1190.7501.484470.3500.80611.3633.6130.1730.298
*Adultindividualsleavingthesubgroupwithinatmost1minwerecountedashavingleftduringthesamefissionevent.

Numberoftimeseachadultindividualleftthesubgroup.

Subgroupsizeatthestartofeachevent.

Theproportionofadultindividualsthatleftthesubgroupatleastoncetothetotaladultindividualspresentinthesubgroup.
2of6
|
www.pnas.org/cgi/doi/10.1073/pnas.1616812114
responsebutratherthesocialcontextinwhichthegroomingoc-
curs(19).Also,proximitytoborderareas,whereencounterswith
rivalgroupsaremorelikely,didnotaffecturinaryoxytocinlevels.
Thelattersuggeststhatapotentialthreatbyout-groupsaloneis
bothcoordinatedactivities(i.e.,huntingandintergroupconflict)
showedhigherurinaryoxytocinleve
lsthancontrolwithandwithout
affiliationsuggeststhatcoordina
tedbehaviorislinkedtotheoxy-
tocinergicsystem.Giventhatintergroupconflictshadsignificantly
higheroxytocinlevelsthanhuntingindicatesthatthiseffectislikely
reinforcedinthecontextofout-grou
pthreat.Itisthereforepossible
thatin-groupcoordinatedactivityandtheperceptionofanin-group
out-groupcontextactinsynergyduringintergroupconflicts.How-
ever,becausewearelackingabeh
avioralmeasureofthedegreeof
coordination,wecannotruleoutthattheoxytocinergicsystemre-
activityobservedinintergroupconflictsisamerefunctionofa
greaterlevelofcoordinationwhenfacinghostilerivalgroups.
coordination,oxytocinsecretion,orboth,ourresultssuggestthat
chimpanzeein-groupcohesivebehaviorinthefaceofout-group
threatislikelysupportedbythes
amephysiologicalmechanism
suggestedforhumanparochialaltruism,theoxytocinergicsystem.
precipitatesparticipationinintergroupconflicts.Owingtothe
beneficialvalueofcooperativegroupactionduringintergroup
conflict,weexpectthatananticipatoryoxytocinincreasewould
beadaptivewheninfluencinggroupcohesion.Becauseinthe
TaForest,chimpanzeeborderpatrolsareoftenprecededby
groomingwithmultiplegroupmembers,themajorityofpreborder
patrolsamplescollectedinvolvedgroominginteractions.Accord-
ingly,toinvestigateanticipatoryoxytocinincrease(anticipation
model),wecontrastedurinesamplescollectedaftermultipartner
groomingsessions:(
i
)shortlybeforetheinitiationofborderpa-
trols(

preborderpatrolwithaffiliation

;6subjects,
n
=
14sam-
plesof10intergroupconflictevents)or,asacontrol,(
ii
)ondays
withoutintergroupconflict(controlwithaffiliation;9individuals,
n
=
38samplesof34events).Thismodelincludedsamplesfroma
singlechimpanzeegroup,Eastgroup,becausenopreborderpatrol
withaffiliationsampleswereattainedforSouthgroup.Wefitted
anLMMcontrollingfordurationofgrooming,subgroupsize,
individuals

sexandrank,andproximitytoborderareas.Wefound
asignificantpositiveimpactofimminentborderpatrolsonurinary
oxytocinlevels(LMM,
likelihoodratiotest:

2
=
11.132,df
=
1,
P

0.001;Fig.1
B
and
TableS8
),aneffectthatwasnotdrivenbythe
durationofmultipartnergroomin
gorothercontrolpredictors.
Ourresultsdemonstrateananticipatoryincreaseinurinaryoxy-
tocininasimilarmannertotheanti
cipatorytestos
teroneincrease
inchimpanzees.Theobservedhighurinaryoxytocinlevelsbefore
borderpatrolinitiationsuggestthatindividualsmayanticipate
imminentintergroupconflict.Moreover,whencomparingpreborder
patrolwithaffiliationwithintergroupconflictwithaffiliation,we
foundnosignificanteffectonurinaryoxytocinlevels(LMM:

2
=
0.181,df
=
1,
P
=
0.669;
Fig.S3
and
TableS9
).Thissuggests
thattheobservedanticipatoryincreaseremainshighthroughout
theintergroupconflict.Theanxiolyticeffectofoxytocinispro-
posedtofacilitatesocialcohesionduringhighlyriskysituations
thatmightotherwiseprecipitatedefectionawayfromthreat(4).
Accordingly,whengroupdefenseprovidesindividualswithfitness
advantages,mechanismsinvolvinganticipatoryhighoxytocinpo-
tentiallymaintaincooperationandsafeguardagainstdefection.
Possiblyresultingfromamalepropensitytoparticipateinin-
tergrouphostility(30),experimentalapproachesinvestigatingco-
operationduringintergroupconflicthavebeenmainlylimitedto
malebehavior(31).However,inarecentstudy,intranasalad-
ministrationofoxytocinenhancedin-groupcooperationduringan
suggestsoxytocinasaninfluentialphysiologicalmechanisminboth
sexesinin-groupout-groupcontexts.Accordingly,weinvestigated
sexdifferencesduringnaturalintergroupconflictsinchimpanzees.
IntheTaForest,bothsexesparticipateinintergroupconflict(12),
andfemalesparticipatedin91%ofintergroupconflictsinthis
study(proportionofalladultfemalesandparousfemales0.44

0.21and0.35

andsexwasnotsignificant,showingthatfemaleandmalechim-
panzeeshavesimilaroxytocinreactivityacrossevents.Male
andfemalehostagetakingbyrivalgroupsaresubstantialthreats
forfemales(10,11)andmayincreasethelikelihoodoffemale
avoidanceofintergroupconflicts(13).However,intheTaForest,
otherchimpanzeefieldsites(10).Whetherthisisacauseoran
effectofthelikelihoodofbothsexestocooperateinthreatening
thatselectivepressuresmayhaveledtosimilaroxytocinergic
systeminvolvementinintergroupconflictinbothsexes.Future
investigationsofthehormonalmechanismsinvolvingin-groupout-
ofthereportedsexdifferences(30)inhumanintergroupviolence.
Table2.Eventmodel:Effectofintergroupconflictandin-groupaffiliationonurinaryoxytocinlevels,log-transformed
TermEstimateSECI
lower
CI
upper

2
*
P
Intercept4.3530.2123.9434.783

Testpredictorlevels
Controlwithaffiliation

0.1220.160

0.1990.4110.5490.459
Controlwithcoordination

0.6980.2220.2541.1376.5560.010
Intergroupconflictwithoutaffiliation

1.1970.1710.8401.57636.044

0.001
Intergroupconflictwithaffiliation

1.3330.1930.9281.71827.298

0.001
Controlpredictors
Group


0.3190.127

0.579

0.0485.9160.015
Sex


0.3070.200

0.7170.0542.2920.130
Proximity
{

0.0350.057

0.1530.0700.3780.538
Subgroupsize
#
0.0820.049

0.0100.1862.7140.099
Rank
jj
0.0770.100

0.1050.2690.5740.449
Statisticallysignificantresults(
P

0.05)appearinbold.CI,confidenceinterval.
*Degreesoffreedomare1.

,

,

2
and
P
valuesrefertocomparisonwiththereferencecategories:

controlwithoutaffiliation,

Eastgroup,and

female.
{
,#,
jj
z
-transformed,mean

SDoftheoriginalvariables:
{
61.59

29.21(range5to99),
#
11.77

5.87,and
jj
0.62

0.24(range0to1,with1beingthehighest
socialrank).
4of6
|
www.pnas.org/cgi/doi/10.1073/pnas.1616812114

Приложенные файлы

  • pdf 14794828
    Размер файла: 585 kB Загрузок: 0

Добавить комментарий