Limited social learning of a novel technical problem by ...animalcognitionlab.org/sites/default/files/limited social learning 2014... · dominance hierarchy that structures spotted

Page 1

L

SKa

b

c

d

a

ARRAA

KSINSS

1

aiVfia(1ioa&ci

W

h0

Behavioural Processes 109 (2014) 111–120

Contents lists available at ScienceDirect

Behavioural Processes

jo ur nal homep ag e: www.elsev ier .com/ locate /behavproc

imited social learning of a novel technical problem by spotted hyenas

arah Benson-Amrama,∗, Virginia K. Heinenb,c, Amelia Gessnerb, Mary L. Weldeled,ay E. Holekampb

Department of Zoology and Physiology, University of Wyoming, Laramie, Wyoming, 82071 USADepartment of Zoology, Michigan State University, East Lansing, Michigan 48824 USADepartment of Ecology, Evolution and Behavior, University of Minnesota, Saint Paul, Minnesota 55108 USADepartment of Psychology, University of California, Berkeley, California 94720 USA

r t i c l e i n f o

rticle history:eceived 25 May 2014eceived in revised form 8 September 2014ccepted 12 September 2014vailable online 20 September 2014

eywords:ocial learningnnovationeophobiatimulus enhancement

a b s t r a c t

Social learning can have profound evolutionary consequences because it drives the diffusion of novelbehaviours among individuals and promotes the maintenance of traditions within populations. Weinquired whether spotted hyenas (Crocuta crocuta), generalist carnivores living in complex, primate-like societies, acquire information from conspecifics about a novel problem-solving task. Previously, wepresented wild hyenas with a food-access puzzle and found that social learning opportunities did notaffect problem-solving success among observers, but did reduce observers’ neophobia. However, we hadlittle control over which individuals observed conspecifics solve the problem, and few wild hyenas weresuccessful. Therefore, we conducted an experiment in captivity where we controlled observer access totwo demonstration styles. Again, social learning opportunities did not affect problem-solving success,but tended to reduce neophobia among captive observers. Social learning opportunities also influenced

ocial facilitation problem-solving style. Captive hyenas showed limited evidence for directed social learning; low-rankingindividuals paid closer attention to demonstrators than high-ranking individuals, although this greaterattention did not result in greater success. We conclude that wild and captive hyenas exploit social learn-ing opportunities similarly, and that the limited social learning shown by hyenas on this task is likelybased on localized stimulus enhancement.

© 2014 Elsevier B.V. All rights reserved.

. Introduction

Advantages of group living include exchange of informationmong group members and opportunities to learn adaptive behav-ors from conspecifics (Lee, 1994; Giraldeau, 1997; Addessi &isalberghi, 2001). Social learning enables individuals to bene-t from the expertise and knowledge of other group members,nd represents the basis for formation of traditions and cultureHumphrey, 1976; Jolly, 1988; Russon, 1997; Whiten & Byrne,997; Day et al., 2003; Whiten & Van Schaik, 2007). Social learning

s hypothesized to be particularly adaptive for animals that foragepportunistically, use challenging techniques for food searchingnd handling, and are highly gregarious (Klopfer, 1959; Caldwell

Whiten, 2002). In this study, we examine the extent to whichaptive spotted hyenas use available social information when solv-ng a novel food-access problem, and then compare the abilities

∗ Corresponding author at: Department of Zoology and Physiology University ofyoming Laramie, WY 82071 USA, Tel.: +1 307 766 4207.

E-mail address: [email protected] (S. Benson-Amram).

ttp://dx.doi.org/10.1016/j.beproc.2014.09.019376-6357/© 2014 Elsevier B.V. All rights reserved.

of the captive hyenas to those previously found in wild hyenasconfronting the same task. The results presented here are particu-larly valuable because they provide a direct comparison betweenthe social learning abilities of wild and captive members of a singlespecies.

Spotted hyenas offer a good model system for investigating therole of social learning in the acquisition of innovative behaviorsfor several reasons. First, spotted hyenas are generalist carnivoresthat hunt and scavenge a diverse array of prey (Cooper et al.,1999). Social learning may be particularly adaptive when theseanimals explore novel food sources (Huber et al., 2001; Moscovice& Snowdon, 2006), and naïve hyenas would likely benefit greatlyfrom observing the foraging choices made by knowledgeable con-specifics. Second, spotted hyenas live in large, complex primate-likesocieties and exhibit patterns of competition and cooperation thatare remarkably similar to those seen in cercopithicine primates(Frank, 1986; Holekamp, 1999; Holekamp et al., 2007; Holekamp,

2007). Although social learning has received a great deal of atten-tion in studies of primate cognition (Call & Tomasello, 1995;Bugnyar & Huber, 1997; Whiten, 1998; Custance et al., 1999;Custance et al., 2001; Day et al., 2003; Caldwell & Whiten, 2004),

Page 2

1 ioural

a22bettawnlKledHuo(tamftmammfsl

oivBis2tlil(B

nw2iaBdoeo

ceseotwtv

12 S. Benson-Amram et al. / Behav

nd in studies outside of primates, such as fish (Brown & Laland,003), birds (Midford et al., 2000; Aplin et al., 2013; Boogert et al.,008), rats (Galef, 2009), and bats (Wright et al., 2011), it has rarelyeen investigated in carnivores (Nel, 1999; but see Mersmannt al., 2011). However, given their primate-like social complexity,heoretically social learning should be no less adaptive for spot-ed hyenas than it is for many primates. Third, spotted hyenasre especially interesting with regard to directed social learning,hich occurs when factors such as age, social rank, sex, related-ess, or patterns of association influence the likelihood of social

earning, and the likelihood of attending to a demonstrator (Coussi-orbel & Fragaszy, 1995). This is because spotted hyena societies,

ike those of cercopithicine primates, are structured by strict lin-ar dominance hierarchies, in which an individual’s social ranketermines its priority of access to resources (Frank, 1986; East &ofer, 2001). Most interestingly though, spotted hyenas are highlynusual among mammals in that adult females and their dependentffspring are socially dominant to adult males in the social groupHolekamp & Smale, 1991). This unique characteristic allows uso test whether spotted hyenas are biased towards learning fromdult females when compared to mammals living in more typicalale-dominated societies. Additionally, there may also be a sex dif-

erence among observers, where female hyenas pay more attentiono female demonstrators, but males do not. For example, long-tailed

acaques live in social groups with a dominant male, and therere some indications that male captive long-tailed macaques paidore attention to, and learned more from watching, the dominantale in the group demonstrate novel tool-using behavior than did

emales (Zuberbuhler et al., 1996). There is also evidence of directedocial learning within social groups of female-dominated ringtailemurs, Lemur catta (Kendal et al., 2010).

Whiten and Mesoudi (2008) noted that the validity of the fieldf social learning is limited by the overwhelming focus on studiesn captivity. Here, we compared social learning abilities of indi-iduals from both wild and captive populations of spotted hyenas.y testing individuals in both populations using the same exper-

mental apparatus, we were able to confirm our observations ofocial learning in a natural setting (Benson-Amram and Holekamp,012) with results from a controlled captive study. Our study showshat wild and captive hyenas acquire information through socialearning in the same way, and joins only a handful of other stud-es in using the same experimental apparatus to investigate socialearning in both wild and captive populations of a single speciesSeferta et al., 2001; Webster & Lefebvre, 2001; Gajdon et al., 2004;ouchard et al., 2007).

We previously reported that, among both wild and captive hye-as presented with a novel foraging task, those individuals thatere less neophobic (i.e. less adverse to novel objects (Greenberg,

003) and exhibited a greater diversity of initial exploratory behav-ors were more successful problem solvers than more neophobicnd less exploratory hyenas (Benson-Amram & Holekamp, 2012;enson-Amram et al., 2013). Although social learning opportunitiesid not result in better problem-solving success in those studies,ur findings led us to inquire whether hyenas acquired other ben-ficial information about the novel task by observing the behaviorf conspecifics.

Social learning is mediated by a variety of different psy-hological mechanisms including social facilitation, stimulusnhancement, emulation, and imitation (Byrne, 1995). Studies ofocial learning in captive capuchin monkeys, Cebus apella, providevidence for social facilitation (Visalberghi & Addessi, 2000), whichccurs when individuals are more likely to perform a behavior in

he presence of a conspecific performing the same behavior thanhen they are alone (Shettleworth, 2009). In contrast, domes-

ic dogs, Canis familiaris, acquire socially transmitted informationia stimulus enhancement (Mersmann et al., 2011), defined as

Processes 109 (2014) 111–120

an increased likelihood of an observer contacting or interactingwith an object as a result of observing another individual interactwith that object (Heyes, 1994; Shettleworth, 2009). More specifi-cally, localized stimulus enhancement occurs when an observer isattracted to the specific part of an object seen being manipulatedby a demonstrator (Huber et al., 2001). Unlike social facilitation,stimulus enhancement does not require the demonstrator to bepresent when the observer interacts with the object (Shettleworth,2009). Captive keas, Nestor notabilis, emulate the actions of demon-strators when interacting with a novel food-access puzzle (Huberet al., 2001). Emulation occurs when an individual copies ele-ments of a complex action, but does not fully imitate, or performthe same actions as, a demonstrator (Shettleworth, 2009). Finally,it appears that marmosets, Callithrix jacchus, (Voelkl and Huber,2000), chimpanzees, Pan troglodytes, (Whiten et al., 1996; Whiten,1998), gorillas, Gorilla gorilla, (Stoinksi et al., 2001), and bandedmongoose, Mungos mungo, (Müller and Cant, 2010) may imitate ademonstrator’s actions. Imitation, defined as “the copying of a novelor otherwise improbable act or utterance” (Thorpe, 1963), occurswhen an observer becomes more likely than otherwise to exhibitthe same novel action, action sequence, or combination of actionsthat it saw performed by a demonstrator (Hoppitt and Laland,2008). Imitation is thought to be more cognitively demanding thanother forms of social learning because it may require representa-tions of another individual’s intentions and perspective (Heyes andGalef Jr, 1996).

Here, we adopted an experimental approach to inquire which ofthese four social learning mechanisms are utilized by spotted hye-nas. We predicted that, if hyenas learn via social facilitation, thenindividuals should show reduced neophobia toward, and increasedinterest in, the puzzle, but only when in the presence of a demon-strator. Social facilitation should not lead to a decrease in neophobiawhen hyenas view a demonstrator interact with the puzzle, butare then tested alone. In contrast, if hyenas learn via localizedstimulus enhancement, then individuals observing a conspecificsolve a novel food-access puzzle should spend more time work-ing on relevant aspects of the puzzle than control individuals whohad no access to a demonstrator. If hyenas learn socially throughemulation, then observer hyenas should be more efficient problem-solvers and learn the solution faster than control hyenas. We alsospecifically tested whether observer hyenas showed evidence ofimitation by using two different demonstrator individuals, each ofwhich used a different behavioral strategy to solve the puzzle. Ifhyenas learn through imitation, then we expected that individualswould show behavior patterns similar to those of their demon-strator, and show patterns different from those of hyenas thatobserved a different demonstrator. Finally, we inquired whetherhyenas show directed social learning. If so, given the matriarchaldominance hierarchy that structures spotted hyena society, wepredicted that observer hyenas should attend more closely to ahigh-ranking female than to a low-ranking male demonstrator.

2. Methods

2.1. Subjects and study site

Experiments were conducted on members of a captive breed-ing colony of spotted hyenas maintained at the Field Station forBehavioral Research at the University of California, Berkeley. Datawere collected from June to August 2008 when the colony housed26 captive-born hyenas: 11 adult females, 11 adult males, and 4

juveniles (3 females and 1 male). Captive hyenas were consid-ered adults once they reached two years of age. The hyenas werehoused in outdoor or semi-outdoor enclosures, in groups of two orthree individuals. Social rank was known within dyads and triads

Page 3

S. Benson-Amram et al. / Behavioural

Fig. 1. Images of the puzzle boxes used in the experiments involving (a) wild and(b) captive subject populations. In (a) two wild hyenas are investigating the puzzlebop

omSdsdDaws

2

wbt

ox while in (b) a captive hyena is flipping a puzzle box. In (c) the image is a close-upf the latch bolt that hyenas needed to move laterally to access the meat inside theuzzle box.

f individuals housed together, but social ranks were not deter-ined for each individual relative to all other hyenas in the colony.

ocial ranks within dyads and triads were assessed indepen-ently by caretakers and were determined through observations ofubmissive and aggressive behavior and through observations ofisplacement in competitive feeding situations (Frank et al., 1989;rea et al., 2002b; Mathevon et al., 2010). Hyenas were categorizeds being either dominant or subordinate to the other hyenas withhich they were housed. Full details regarding the captive study

ubjects are given in Benson-Amram et al. (2013).

.2. Experimental apparatus

We presented captive hyenas with a novel food-access puzzle,hich was a metal puzzle box baited with raw meat (Fig. 1). The

ox had a single door on one long side with a simple bolt latch thathe hyenas needed to slide laterally for the door to swing open,

Processes 109 (2014) 111–120 113

and two rebar handles, one located centrally on each short side.The puzzle box was very similar to a novel food access puzzle wepreviously presented to wild hyenas from a study population in theMasai Mara National Reserve in Kenya (Fig. 1a; Benson-Amram &Holekamp, 2012) and had the following dimensions: L: 63.5 cm × H:33 cm × W: 33 cm with a 39 cm long door.

2.3. Experimental procedure

Eleven captive hyenas were ‘observers’ and had the opportunityto watch one conspecific demonstrator open the puzzle box imme-diately before they were given access to the puzzle box themselves.Six hyenas served as ‘controls’ and did not have any opportunity toobserve conspecifics interact with the puzzle box before their owntrials. Two hyenas, one subordinate adult male and one dominantadult female, served as demonstrators. The demonstrators were notspecifically trained to open the puzzle box, but were neverthelessconsistently successful in all of their trials. The experimental pro-tocol for the initial trial for each demonstrator hyena was identicalto the protocol for the control hyenas. For this reason, we includedthe two demonstrator hyenas as members of the control group inanalyses of behavior during the initial trial with the puzzle box, sosample size for the control group in these analyses was eight hye-nas. Five observer hyenas watched the adult male and six observerswatched the adult female. Hyenas were assigned to observer andcontrol groups before the experiments began in such a way as tobalance sex, age, and rank (Table 1).

All hyenas were food deprived for 24 h prior to testing to bringall individuals to the same level of feeding motivation. All captivetrials were conducted between 1100 and 1630 h. Captive hyenassometimes have neophobic responses to novel situations (Drea &Carter, 2009), so we gave individuals a ten min habituation periodin the 39 m2 test enclosure just prior to their first trial of the day.This allowed hyenas to investigate the test enclosure in the absenceof the puzzle box, and minimized the amount of time hyenas spentinvestigating the enclosure during the experimental trial. After thehabituation period, hyenas were moved back into the holding penfor a few minutes while we set-up the puzzle box. All hyenas atthe field station were trained in infancy to move readily from oneenclosure to another (Drea & Carter, 2009).

In social learning trials, an observer hyena was moved into anenclosure adjacent to the test enclosure. These two enclosures wereseparated by a chain-link fence through which the hyenas could see,hear, and smell the adjacent enclosure. The puzzle box was set-upin the test enclosure such that the box was approximately 2 m fromthe chain-link fence separating the test and adjacent enclosure.Additionally, the door and latch of the puzzle box faced the adjacentenclosure. The puzzle box was baited with raw meat (approxi-mately 1 kg of beef ribs) and the latch handle was left protrudingat 90◦ from the box.

Demonstrator trials began when the demonstrator hyena leftthe holding pen and entered the test enclosure containing thepuzzle box. Demonstrators were allowed out of the holding penwhen the observer hyena was positioned within 2 m of the fence,or approximately 4 m from the puzzle box. Demonstrator trialsended when the demonstrator opened the box and removed themeat, at which point demonstrators were moved back into theholding pen. Following each demonstrator trial we immediatelyre-baited the puzzle box and set it back in the starting position. Theobserver hyena was then allowed into the test enclosure to beginits trial. Observer trials ended when the subject opened the box andremoved the meat, or when 30 min had passed, whichever came

first. We then repeated this sequence such that every observer trialwas immediately preceded by a demonstrator trial. We attemptedto conduct at least three trials per day on two consecutive days,such that each observer hyena had 6 opportunities to watch a

Page 4

114 S. Benson-Amram et al. / Behavioural Processes 109 (2014) 111–120

Table 1Details of the captive hyenas used in the experiments.

ID Treatment groupa Sex Age classb Social rankc Social groupd Prior exp. Success

Scooter (female) Demonstrator 1 F A Dominant 1 No YCass Observer-1 F A Dominant 2 No YKombo Observer-1 F A Dominant 5 No NNakuru Observer-1 F A Subordinate 3 Yes YRobie Observer-1 M A Subordinate 7 No YRocco Observer-1 M A Subordinate 10 No YZawadi Observer-1 M A Dominant 7 No Y

Bramble (male) Demonstrator 2 M A Subordinate 8 No YGremlin Observer-2 M A Subordinate 12 Yes YHaji Observer-2 F J Subordinate 4 No NHarley Observer-2 F J Dominant 9 No NUrsa Observer-2 F A Subordinate 2 No NWinnie Observer-2 M A Subordinate 5 Yes Y

BJ Control F A Dominant 11 No YDenali Control M A Subordinate 6 No YDusty Control M A Dominant 6 No YGulliver Control M A Subordinate 1 Yes YJambo Control F J Subordinate 4 No NNairobi Control F A Dominant 3 Yes Y

Each captive subject is listed along with its treatment group, sex, age, social rank, social group, whether or not it had previously participated in a study of cooperative problemsolving conducted in 1995, and whether it was ever successful in opening the puzzle box.

a Treatment group refers to whether the subject was a demonstrator, an observer of the female demonstrator, Scooter (Observer-1), an observer of the male demonstrator,Bramble (Observer-2), or in the control group that did not have access to a demonstrator.

iles (J)nas w8–12 w

dt

tsuwnt

oofiihsn(

2

eJiaJiar

zohee

te

b Hyenas were considered adults (A) when they were over 24-months and juvenc Social rank refers to the rank of the subject relative to the one or two other hyed Social group refers to the hyenas that were housed together. Hyenas in groups

emonstrator interact with the puzzle and 6 opportunities to solvehe problem itself.

Control animals also received a habituation period and werehen moved back into the holding area while the puzzle box waset-up. Like the observer hyenas, they participated in three trials ofp to 30 min each per day and were moved into the holding areahile the puzzle box was re-baited between trials. Control hyenasever observed a demonstrator interact with the puzzle box beforeheir trials.

In total, we conducted 170 trials on 19 captive hyenas. Allbserver and control hyenas had six trials each with the exceptionf four individuals: two hyenas observing the adult male only hadve trials, one control hyena had four trials, and one hyena observ-

ng the adult female had only one trial. One demonstrator hyenaad 38 total trials and the other had 39 total trials; both demon-trators opened the box in every trial. We conducted trials withine adult males, seven adult females, and three juvenile femalesTable 1).

.4. Data extraction from videotaped trials

We videotaped all puzzle box trials in their entirety, andxtracted behavioral data from these videos using the programWatcher (Blumstein et al., 2006). We extracted detailed behav-oral data from the initial trial of each observer and control hyena,nd from all demonstrator trials. SBA, VH, and AG extracted theWatcher data; inter-observer reliability was high across all behav-oral measures used (mean R = 0.98, range = 0.90–0.99). JWatcherutomatically recorded both the duration and number of occur-ences of each behavior observed in a trial.

A ‘successful’ individual was defined as one that opened the puz-le box in at least one trial. An ‘unsuccessful’ individual failed topen the puzzle box. The puzzle box was a novel stimulus for theyenas, so we measured ‘neophobia’ by examining the latency ofach focal hyena to contact the puzzle box in its initial trial once it

ntered a 5-m radius around the box.

We recorded the “work time” for each hyena as the length ofime it spent with its head down working on the puzzle box duringach trial. Along with total work time, we recorded both the amount

when they were under 24-months of age.ith whom they were housed.

ere housed with an individual that was not included in this study.

of work time a hyena spent on the side of the puzzle box with thelatch and on the sides of the box with the handles. We also extractedthe percent of the demonstrator’s work time during which eachobserver hyena was oriented toward the puzzle box, in order toexamine attentiveness of observer hyenas to the actions of theirdemonstrators.

We calculated both the number and frequency of exploratorybehaviors each focal hyena exhibited when interacting with thepuzzle box. We included ten different exploratory behaviors inthis measure (push, pull, dig, foot contact, mouth contact on latch,mouth contact on handles, lift, flip, rub, and lower). We defined‘digging’, ‘flipping’, and ‘pushing or pulling’ in the same manneras in the wild hyenas (Benson-Amram and Holekamp, 2012). ‘Footcontact’ occurred when a foot touched any part of the puzzle box,whereas ‘mouth contact on latch’ and ‘mouth contact on handles’included licking, biting, and otherwise making oral contact witheither the latch side or the two handle sides of the puzzle box.‘Lifting’ occurred when the hyena used its mouth to move the boxupward into the air. ‘Rubbing’ occurred when the hyena contactedthe box with its neck or torso and moved its body along the lengthof the box. We scored a behavior as ‘lowering’ when the hyenaput its face and/or body low to the ground on one or both forelegswithout fully lying down. We limited our analysis to these tenbehaviors because these were the behaviors that had the highestinter-observer reliability scores (all greater than 90%). We used thenumber of different exploratory behaviors hyenas exhibited wheninteracting with the puzzle box as a measure of ‘exploration diver-sity’. If a hyena demonstrated all 10 of these behaviors at least onceduring a trial it received the maximum exploration diversity scoreof 10. If a hyena demonstrated none of these behaviors, it receivedan exploration diversity score of 0.

2.5. Statistical analyses

To examine the effects of social learning and minimize the

effects of individual trial-and-error learning we limited our anal-yses with the captive hyenas to their initial trials with the puzzlebox. The only exception was the analysis of learning across all trialsfor all subjects (Fig. 3). Additionally, previous analyses showed that

Page 5

S. Benson-Amram et al. / Behavioural Processes 109 (2014) 111–120 115

Table 2Results of generalized linear models examining the effect of Demonstrator ID and treatment group on box-oriented behavior of control and observer hyenas.

Behavior Demo 1 (Female) vs. Demo 2 (Male) a Observer 1 vs. observer 2 b Observer 1 vs. control b Observer 2 vs. control b

Latch Duration c P = 0.073 P = 0.016 P = 0.036 NSLatch Number d NS P = 0.066 P = 0.071 NSHandle Duration c NS NS NS NSHandle Number d P = 0.0026 NS NS NSDig Duration e NS NS P = 0.080 NSDig Number f NS NS P = 0.089 NSExploration Diversity g P = 0.0040 NS NS NSPull Duration e P = 0.053 NS NS NSPull Number f NS NS NS NSLift Number f P = 0.040 NS NS NSFlip Numberf P = 0.054 NS NS NS

All response variables were corrected for work time in the statistical model. Significant results and non-significant trends (p ≤ 0.10) are reported in detail. Other non-significantresults are denoted by the symbol NS.

a Demonstrator data were taken from the first trial each demonstrator had with each observer. Demonstrator 1 is Scooter, the female demonstrator, and had n = 6 firsttrials in front of adult observers. Demonstrator 2 is Bramble, the male demonstrator, and had n = 3 first trials in front of adult observers.

b Observer and control data were taken from each focal hyena’s initial trial with the puzzle box. Observer 1 refers to hyenas that watched Demonstrator 1 open the puzzlebox (n = 6 adults). Observer 2 refers to hyenas that watched Demonstrator 2 open the puzzle box (n = 3 adults). The sample size for control hyenas was 5 adults.

c Latch and handle duration are measures that combine the time spent biting and licking the sides of the puzzle box with either the latch or handles, respectively.d Latch and handle number denote all instances when a focal hyena initiated oral contact with either the latch or handle sides of the puzzle box.

e puz initiat a foc

cc2ntit

tcpOchrra

hctwww

ofiaedwwpwoctbovftr

e Dig and pull duration refer to the total time spent digging around the base of thf Dig, pull, lift, and flip number refer to the total number of times the focal hyenag Exploration diversity is the total number of different exploratory behaviors tha

aptive juveniles were significantly more neophobic and less suc-essful on this task than were captive adults (Benson-Amram et al.,013). In fact, all captive juveniles were unsuccessful in solving theovel food-access puzzle (Benson-Amram et al., 2013). Therefore,o get the clearest picture of social learning in this species we onlyncluded trials on adults in our analyses, unless we were specificallyesting the effect of age.

We used generalized linear models (GLM; R 2.13.0) to evaluatehe effect of social learning opportunities on problem-solving suc-ess, exploration diversity, and neophobia. Latency to approach theuzzle box was log-transformed to achieve a normal distribution.ne outlier was excluded from our analyses on neophobia in theaptive hyenas because this individual did not receive a 10-minabituation period prior to the start of its first trial. However, wean all tests on neophobia with and without this outlier, and theelative significance and direction of the results were the same inll cases.

We used generalized linear mixed models (GLMM) to examineow work time changed over successive trials for observer versusontrol hyenas. We included whether the subject was in the con-rol or observer treatment group as a fixed covariate to determinehether hyenas in these two treatments differed in the rate athich they learned to solve the puzzle box task. Focal hyena IDas included as a random effect.

To examine whether observer hyenas showed similar box-riented behavior to that of the demonstrator they observed, werst ensured that the two demonstrator hyenas differed from onenother in their box-oriented behavior. To do this, we examinedach response variable using generalized linear models withemonstrator ID as the predictor variable. In these analyses, weere interested in determining whether an observer’s behavioras influenced by the demonstration they actually witnessedrior to the observer’s first interaction with the puzzle box, soe only used data from the first demonstration trial for each

bserver hyena. For the response variables that differed signifi-antly between the two demonstrators, we then tested whetherhe observer and control hyenas differed in their exploratoryehavior and whether the observer hyenas matched the behaviorf the particular demonstrator each watched. All of the response

ariables used in these analyses are listed in Table 2. To controlor work time, we used residuals from a generalized linear modelhat included work time as the sole predictor variable. Theseesiduals were then used as our response variable when examining

zle box or pulling the box.ted these behaviors during a trial.al hyena exhibited during a puzzle-box trial.

differences between demonstrators and treatment groups. Weused a quasi-poisson distribution for analyses of count data.

To examine the factors affecting the amount of attentionobservers paid to their demonstrators, assessed as the percentof time an observer hyena spent oriented toward its demonstra-tor while the demonstrator was working on the puzzle box, weused generalized linear models with residuals corrected for thedemonstrator’s work time as the response variable, and with thefollowing predictor variables: the sex, age, and social rank of theobserver hyena and the ID of the demonstrator. Alternative modelswere compared using Akaike’s Information Criterion (AIC) values. Asmaller AIC value indicates a better-fitting model (Crawley, 2007),and the results from the model with the lowest AIC value arereported here.

Mean values are given ± standard error. Differences betweengroups were considered significant when P ≤ 0.05.

3. Results

3.1. Effect of social learning opportunity on problem-solvingsuccess

The likelihood of success in opening the puzzle box did notdiffer between observer and control hyenas (�2

1 = 2.52, P = 0.11).In fact, there was a trend for controls to be more successful thanobservers. Only two captive adults failed to solve the problem, andboth of these individuals were observers. We also found no differ-ence in exploration diversity or work time between observers andcontrols (exploration diversity: F1,16 = 0.060, P = 0.81; work time:F1,17 = 2.82, P = 0.11). Likewise, the percent of time observers spentoriented toward their demonstrator had no effect on problem-solving success (F1,7 = 3.034, P = 0.13). In fact, unsuccessful hyenasactually tended to spend a greater percent of their demonstrator’swork time oriented toward their demonstrators than did successfulhyenas. However, we did find that observers showed a trend towardbeing less neophobic than controls (F1,17 = 4.0, P < 0.062; Fig. 2).Specifically, hyenas that observed a conspecific open the puzzlebox approached the box more quickly in their initial trial than didcontrol hyenas without access to a demonstrator. Despite reducing

neophobia, opportunities for social learning did not improve thespeed at which observers learned the problem-solving task. Indi-viduals in both control and observer groups became significantlyfaster at opening the puzzle box over successive trials (F1,56 = 17.52,

Page 6

116 S. Benson-Amram et al. / Behavioural Processes 109 (2014) 111–120

Fig. 2. Mean ± SE latency to approach the puzzle box during the initial trial in a) captive adults (N = 15) and b) wild individuals (N = 49), divided by social learning context.One captive hyena was excluded from this analysis because this individual did not receiveWild hyenas were tested in Kenya, as described in Benson-Amram & Holekamp (2012) a* = P < 0.05.

Fig. 3. Average learning curves for successful observer (N = 7) and control (N = 5)hyenas in captivity when they interacted with the puzzle box. The learning curvesran

Potliof

3

aaFsptt(

epresent the mean ± SE latency to solve for all individuals who were successful in given trial. Sample sizes are indicated above each data point and varied becauseot all hyenas were successful in every trial.

= 0.0001; Fig. 3). The learning curves were nearly identical, andbservers and controls did not differ in their work time across allrials with the puzzle box (F1,12 = 2.26, P = 0.16; Fig. 3). When weimited our analysis to the initial trial, we similarly saw no signif-cant difference in work time to open the puzzle box between thebserver and control hyenas, although controls tended to succeedaster (F1,14 = 2.80, P = 0.12; Fig. 3).

.2. Predictors of attention paid to demonstrators

Both subordinate hyenas and juveniles paid significantly morettention to their demonstrator than did dominant hyenas ordults (Rank: F1,8 = 6.55, P = 0.038; Fig. 4a; Age: F1,9 = 6.76, P = 0.035;ig. 4b). However, our sample size for the age comparison is verymall, as we tested only three juveniles. Male and female observers

aid equal attention to the demonstrators (F1,7 = 2.09, P = 0.20). Fur-hermore, we found no evidence that observers paid more attentiono a dominant female than to a subordinate male demonstratorF1,6 = 0.078, P = 0.79).

a 10-min habituation period in the test enclosure prior to the start of its first trial.nd Benson-Amram et al. (2013). Asterisks represent significant differences where

3.3. Effect of demonstrators on exploratory behavior

The two demonstrators showed significant differences in manyaspects of their box-oriented behavior (Table 2). In general, thefemale demonstrator spent more time focused on the latch (LatchDuration: t9 = 2.028, P = 0.073; Fig. 5a) and generally did not movethe box, whereas the male demonstrator frequently used thehandles (Handle Number: t9 = −4.13, P = 0.0026; Table 2) to pull(Pull duration: t9 = −2.23, P = 0.053), lift (Lift number: t9 = −2.40,P = 0.040), and flip (Flip Number: t9 = −2.22, P = 0.054) the puz-zle box. Overall, the male demonstrator exhibited substantiallymore exploratory behaviors when interacting with the puzzle boxthan did the female demonstrator (Exploration diversity: t9 = −3.84,P = 0.0040; Table 2). The demonstrators did not differ in theirlatency to contact either the latch (F2,13 = 0.15, P = 0.86) or the han-dles (F2,13 = 1.33, P = 0.30) of the puzzle box.

Despite the consistently different behavioral strategiesemployed by the demonstrators, observers only differed sig-nificantly from controls with respect to the amount of time theyspent working on the latch side of the puzzle box (Fig. 5a; Table 2).Observers of the female demonstrator spent significantly moretime working on the latch than did either observers of the maledemonstrator (t13 = 2.76, P = 0.016) or individuals in the controlgroup (t13 = 2.34, P = 0.036). Observers of the female demonstratoralso showed a trend toward contacting the latch more frequentlythan did hyenas in other treatment groups (vs. observers of themale demonstrator: t13 = 2.012, P = 0.066; vs. Control: t13 = 1.96,P = 0.071; Fig. 5b; Table 2). Surprisingly, there was no correlationbetween the proportion of time an observer spent watching thedemonstrator and the similarity of the observer’s actions to thoseof the demonstrator for any behavior that we examined (WorkTime: r = −0.095, N = 11, P = 0.78; Latch duration: r = 0.19, N = 11,P = 0.58; Latch Number: r = 0.35, N = 11, P = 0.28; Flip Number:r = 0.44, N = 11, P = 0.18; Push number: r = 0.033, N = 11, P = 0.93;Lift number: r = 0.42, N = 11, P = 0.20; Pull number: r = 0.30, N = 11,P = 0.38; Handle Duration: r = 0.14, N = 11, P = 0.69; Handle number:r = 0.36, N = 11, P = 0.28).

4. Discussion

Our results provide evidence for limited social learning of anovel technical problem by spotted hyenas. Specifically, observ-ing a conspecific open the puzzle box through manipulating the

Page 7

S. Benson-Amram et al. / Behavioural Processes 109 (2014) 111–120 117

Fig. 4. Mean ± SE percent of demonstrator work time observers spent oriented toward demonstrators during the first demonstration depending on a) social rank (eitherdominant or subordinate) and b) age class of the observer. All captive individuals includedsignificant differences where * = p < 0.05.

Fig. 5. Mean ± SE (a) percent of work time spent at the latch, and (b) number oftimes per min the focal hyena contacted the latch, among both demonstrators andobservers. Bar color represents social treatment group. The black and white bars onthe left side of the figure represent the male and female demonstrator respectively.The black and white bars on the right side of each figure represent observers ofeither the male (Bramble) or the female (Scooter) demonstrator and the grey barrepresents the control group without access to a demonstrator. All captive adultstested in the experiment are included here (N = 16). Asterisks represent significantdifferences where * = P < 0.05.

in the observer treatment group are represented here (N = 11). Asterisks represent

latch led observers to focus more attention on the latch side ofthe box, which suggests that hyenas are learning to focus on thefunctionally relevant aspect of the problem via localized stimulusenhancement (Huber et al., 2001; Caldwell & Whiten, 2004). AsCaldwell and Whiten (2004) point out, localized stimulus enhance-ment can serve as a powerful force in social learning; by focusingthe attention of observers on the functionally important part of theproblem, individuals can theoretically learn about the problem andfind the solution via trial-and-error learning faster than they wouldwithout this focus.

Observing a conspecific open the puzzle box also tended todecrease neophobia. In both wild and captive hyenas, naïve individ-uals approached the puzzle box more quickly when they observedanother hyena present at the box than when they were alone(Benson-Amram and Holekamp, 2012). Neophobia can be a stronginhibitor of problem-solving success and behavioral flexibilitybecause individuals that fail to engage with a novel problem or foodresource cannot possibly succeed in solving the problem (Sefertaet al., 2001; Webster & Lefebvre, 2001; Greenberg, 2003; Bouchardet al., 2007; Boogert et al., 2008; Cole et al., 2011). Social facilitationmay also have played a role in reducing neophobia of wild hye-nas (Benson-Amram and Holekamp, 2012). Previous experimentalwork has shown that observer hyenas are more likely to eat, drink,scent mark, greet conspecifics, investigate olfactory stimuli, andplay in the presence of a conspecific demonstrating the same behav-ior than hyenas that have not just been exposed to a conspecificexhibiting these behaviors (Glickman et al., 1997). Social facilitationmay also play a role in the acquisition of cooperative problem-solving behavior in this species (Drea & Carter, 2009). However, ourresults from the captive hyenas cannot be attributed to social facil-itation, which requires the presence of the demonstrator. Captivecontrol and observer hyenas were tested individually, such that noother hyena was present in either the test enclosure or an adjacentenclosure during the focal hyena’s trial. It would be interesting toinvestigate further whether social facilitation improves the abili-ties of spotted hyenas to solve novel food-access puzzles by alsotesting captive hyenas in groups.

The results from our study indicate that the social learning abil-ities of spotted hyenas may be relatively limited, at least whenlearning the solution to a novel foraging task. Our results imply

that hyenas are not gaining an understanding of the problem, orits solution, through social learning. As we found earlier with wildhyenas (Benson-Amram and Holekamp, 2012), observing a conspe-cific open the puzzle box did not lead to enhanced problem-solving

Page 8

1 ioural

sfbtAsmtiMfClc

mhpTspfialoce

awawsFwh

togsciamfwdoahcdf

4

Kpiethnn

18 S. Benson-Amram et al. / Behav

uccess among observers in the captive population. Opportunitiesor social learning also did not improve the rate at which hyenasecame proficient at opening the puzzle box, nor did they decreasehe time it took observers to solve the problem in their first trial.dditionally, the attentiveness of the observer during the demon-trator’s trial did not affect the extent to which an observer’s actionsatched those of the demonstrator. These results are similar to

hose obtained in keas (Huber et al., 2001), vervet monkeys, Cercop-thecus aethiops (van de Waal & Bshary, 2011), Japanese macaques,

acaca fuscata (Leca et al., 2010), long-tailed macaques, Macacaascicularis (Zuberbuhler et al., 1996), and common marmosets,allithrix jacchus (Caldwell & Whiten, 2004), where subjects also

earned via localized stimulus enhancement without enhanced suc-ess.

In contrast to previous findings in keas (Gajdon et al., 2004) andarmosets (Halsey et al., 2006), we saw no evidence that captive

yenas have more advanced cognitive abilities, or use more com-lex mechanisms of social learning, than their wild counterparts.his result is particularly interesting because captive hyenas areignificantly more successful than wild hyenas at solving this novelroblem (Benson-Amram et al., 2013). The results of this study con-rm that the greater success rate in the captive hyenas cannot bettributed to more advanced social learning abilities and, instead, isikely due to factors such as reduced neophobia to man-made metalbjects and higher-quality work time due to fewer distractions andompeting interests in the captive environment (Benson-Amramt al., 2013).

Individuals should be strategic about when to rely on sociallycquired knowledge. If the problem at hand can easily be solvedith trial-and-error learning, then there is little need to observe the

ctions of others (Kendal et al., 2009). Thus, it is possible that weould observe a greater effect of social learning on problem-solving

uccess if we tested captive hyenas on a more difficult problem.urther work exposing hyenas to a diverse set of novel problemsould help clarify the circumstances under which hyenas rely moreeavily on socially acquired than individually learned information.

Additionally, it is possible that the identity of the demonstra-ors had a dampening effect on the levels of social learning that webserved. None of the observers were housed in the same socialroup as their demonstrator. Although all hyenas are housed in theame facility, and can communicate vocally, they are not all physi-ally housed together in the same enclosure. Hyenas were housedn groups of 2 or 3 animals, thus the majority of their social inter-ctions are with one or two other individuals. We might have seenore social learning had we paired observers with demonstrators

rom their social group. It is also possible that, because observersere not paired with demonstrators from their social group, theemonstration phase was actually even somewhat stressful for thebservers. If this was the case, then the demonstration might actu-lly have had a negative effect on observers compared to controlyenas. If this were the case, however, we might have expectedontrol hyenas to perform better than observers, but found no evi-ence to support this possibilit Nevertheless, it is worth consideringor future work.

.1. Directed social learning

We found limited evidence for directed social learning (Coussi-orbel & Fragaszy, 1995) in captive spotted hyenas. This isarticularly interesting because vervet monkeys have a very sim-

lar social system to spotted hyenas, and vervets were found toxhibit directed social learning by paying more attention to female

han male demonstrators (van de Waal et al., 2010). However,ere, male and female observers did not differ in their attentive-ess to the demonstrator and the identity of the demonstrator didot affect attentiveness in observers, although these results might

Processes 109 (2014) 111–120

change with larger sample sizes or with use of a demonstrator fromthe observer’s social group. We did find that captive juvenile andsubordinate hyenas spent a significantly greater proportion of thedemonstrator’s trial oriented toward the demonstrator than didadults and dominant hyenas, although this increased attention didnot lead to enhanced problem-solving success. Unfortunately, oursample size of juveniles was small. Thus, additional studies with alarger sample size of juveniles need to be conducted to determinewhether this finding is robust.

Increased social learning and attentiveness amongst subordi-nate individuals has also been shown in tits (Aplin et al., 2013),guppies (Laland and Reader, 1999), and deer mice (Kavaliers et al.,2005). If social learning enables individuals to benefit from theexpertise of more experienced group members (Russon, 1997;Nicol & Pope, 1994), then juveniles would likely benefit more fromattending to conspecifics than would adults. Additionally, lower-ranking individuals likely need to monitor the actions of dominantgroup members in order to avoid aggressive interactions and tocapitalize on opportunities to scrounge resources. Dominant indi-viduals can usually control access to resources and thus may notneed to be as attentive to the behavior of subordinates. Of course,dominance in relation to a single cage-mate, as it was measuredin this study of captive hyenas, is quite different from dominancerelationships in a complex social hierarchy, as would be the casein a wild population. Thus, further investigations in wild popula-tions are needed to fully understand the effects of dominance onattentiveness.

5. Conclusion

Living in a complex society requires that individuals recognize,interact with, and monitor the states of other group members. Ifintelligence evolved in response to these selection pressures, then itis probable that natural selection also favored the ability to observeand learn from the actions of conspecifics (Bugnyar & Huber, 1997).Our results potentially challenge this view by demonstrating thatthe social learning abilities of spotted hyenas are rather limited,at least when dealing with a novel technical challenge. Indeed, itappears that hyenas fail to acquire nearly as much informationfrom social learning opportunities as many other animals withsimilarly complex social systems (Whiten et al., 1996; Bugnyar& Huber, 1997; Huber et al., 2001; Stoinksi et al., 2001; Kendalet al., 2010; Reader & Biro, 2010). However, further tests of thesocial learning abilities of hyenas in different contexts are needed todetermine whether the social learning capabilities of spotted hye-nas are indeed inferior to those documented in these other species.In particular, hyenas need to be tested on more difficult problems,where there should be a greater incentive to rely on socially ratherthan individually acquired information.

Acknowledgements

The research presented here was described in Animal ResearchApplication No. 07/08-099-00, approved most recently on 4 June,2010 by the All University Committee on Animal Use and Careat Michigan State University. The experimental procedures wereapproved by the Institutional Animal Care and Use Committee ofthe University of California, Berkeley. This research was supportedby National Science Foundation grants IOS0819437, IOS1121474,and DEB1353110 to KEH and IOS0920793 to Stephen E. Glickmanfor support of the captive colony. Michigan State University, the

Animal Behaviour Society and Sigma Xi provided funding to SBA.We thank the staff at the FSBR for their generous access to the cap-tive facility and for their help in running the captive trials. KatherineHelms and Matt Maksimoski made valuable contributions to the

Page 9

ioural

Jebmf

R

A

A

B

B

B

B

B

B

B

B

C

C

C

C

C

C

CC

C

D

D

D

E

F

F

G

G

GG

G

H

H

S. Benson-Amram et al. / Behav

Watcher analyses. We thank Roger Carr for his instruction, gen-rous sharing of equipment, and help in constructing the puzzleox that was used in the captive study. We also thank Steve Glick-an, Alan Bond, Neeltje Boogert, and Jeff Clune for their valuable

eedback.

eferences

ddessi, E., Visalberghi, E., 2001. Social facilitation of eating novel food in tuftedcapuchin monkeys (Cebus apella): input provided by group members andresponses affected in the observer. Anim. Cogn. 4, 297–303.

plin, Lucy, M., Ben, C., Sheldon, Morand-Ferron, J., 2013. Milk bottles revisited:social learning and individual variation in the blue tit Cyanistes caeruleus. Anim.Behav. 85.6, 1225–1232.

enson-Amram, S., Holekamp, K.E., 2012. Innovative problem solving by wild spot-ted hyenas. Proc. Biol. Sci. 279, 4087–4095.

enson-Amram, S., Weldele, M.L., Holekamp, K.E., 2013. A comparison of problem-solving abilities between wild and captive spotted hyenas (Crocuta crocuta).Anim. Behav. 85, 349–356.

lumstein, D.T., Evans, C.S. & Daniel, J.C. 2006. JWatcher 1.0. See http://www.jwatcher.ucla.edu

oogert, N.J., Reader, S.M., Hoppitt, W., Laland, K.N., 2008. The origin and spread ofinnovations in starlings. Anim. Behav., 75.

ouchard, J., Goodyer, W., Lefebvre, L., 2007. Social learning and innovation arepositively correlated in pigeons (Columba livia). Anim. Cogn. 10, 259–266.

rown, C., Laland, K.N., 2003. Social learning in fishes: a review. Fish Fisheries 4 (3),280–288.

ugnyar, T., Huber, L., 1997. Push or pull: an experimental study on imitation inmarmosets. Anim. Behav. 54, 817–831.

yrne, R.W., 1995. The Thinking Ape: Evolutionary Origins of Intelligence. OxfordUniversity Press, Oxford.

aldwell, C.A., Whiten, A., 2002. Evolutionary perspectives on imitation: is a com-parative psychology of social learning possible? Anim. Cogn. 5, 193–208.

aldwell, C.A., Whiten, A., 2004. Testing for social learning and imitation in com-mon marmosets, Callithrix jacchus, using an artificial fruit. Anim. Cogn. 7,77–85.

all, J., Tomasello, M., 1995. Use of social information in the problem solving oforangutans (Pongo pygmaeus) and human children (Homo sapiens). J. Comp.Psychol. 109, 308–320.

ole, E.F., Cram, D.L., Quinn, J.L., 2011. Individual variation in spontaneous problem-solving performance among wild great tits. Anim. Behav. 81, 491–498.

ooper, S.M., Holekamp, K.E., Smale, L., 1999. A seasonal feast: long-term analysis offeeding behavior in the spotted hyaena, Crocuta crocuta (Erxleben). Afr. J. Ecol.37, 149–160.

oussi-Korbel, S., Fragaszy, D.M., 1995. On the relation between social dynamics andsocial learning. Anim. Behav. 50, 1441–1453.

rawley, M.J., 2007. The R book. John Wiley & Sons Inc.ustance, D., Whiten, A., Fredman, T., 1999. Social learning of an artificial fruit task

in capuchin monkeys (Cebus apella). J. Comp. Psychol. 113, 13–23.ustance, D., Whiten, A., Sambrook, T., Galdikas, B., 2001. Testing for social learn-

ing in the “artificial fruit” processing of wildborn orangutans (Pongo pygmaeus),Tanjung Puting, Indonesia. Anim. Cogn. 4, 305–313.

ay, R.L., Coe, R.L., Kendal, J.R., Laland, K.N., 2003. Neophilia, innovation and sociallearning: a study of intergeneric differences in callitrichid monkeys. Anim.Behav. 65, 559–571.

rea, C.M., Carter, A.N., 2009. Cooperative problem solving in a social carnivore.Anim. Behav. 78, 967–977.

rea, C.M., Vignieri, S.N., Cunningham, S.B., Glickman, S.E., 2002. Responses toolfactory stimuli in spotted hyenas (Crocuta crocuta): I. Investigation of envi-ronmental odors and the function of rolling. Journal of Comparative Psychology116, 331e341.

ast, M.L., Hofer, H., 2001. Male spotted hyenas (Crocuta crocuta) queue for status insocial groups dominated by females. Behav. Ecol. 12, 558–568.

rank, L.G., 1986. Social organization of the spotted hyaena Crocuta crocuta. II. Dom-inance and reproduction. Anim. Behav. 34, 1510–1527.

rank, L.G., Glickman, S.E., Zabel, C.J., 1989. Ontogeny of female dominance in thespotted hyaena: perspectives from nature and captivity. Symposia of ZoologicalSociety of London 61, 127e146.

ajdon, G.K., Fijn, N., Huber, L., 2004. Testing social learning in a wild mountainparrot, the kea (Nestor notabilis). Learn. Behav. 32, 62–71.

alef, B.G., 2009. Strategies for social learning: testing predictions from formal the-ory. Adv. Study Behav. 39, 117–151.

iraldeau, L.A., 1997. The ecology of information use. Behav. Ecol. 4, 42–68.lickman, S.E., Zabel, C.J., Yoerg, S.I., Weldele, M.L., Drea, C.M., Frank, L.G., 1997.

Social facilitation, affiliation, and dominance in the social life of spotted hyenas.Ann. NY Acad. Sci. 807, 175–184.

reenberg, R., 2003. The role of neophobia and neophilia in the development of inno-vative behaviour of birds. In: Reader, S.M., Laland, K.N. (Eds.), Animal innovation.

Oxford University Press, New York, pp. 175–196.

alsey, L.G., Bezerra, B.M., Souto, A.S., 2006. Can wild common marmosets (Callithrixjacchus) solve the parallel strings task? Anim. Cogn. 9, 229–233.

eyes, C.M., 1994. Social learning in animals: categories and mechanisms. Biol. Rev.69, 207–231.

Processes 109 (2014) 111–120 119

Heyes, C.M., Galef Jr., B.G., 1996. Social Learning in Animals: The Roots of Culture.Academic Press, Inc, San Diego.

Holekamp, K., Sakai, S., Lundrigan, B., 2007. The spotted hyena (Crocuta crocuta) as amodel system for study of the evolution of intelligence. J. Mammal. 88, 545–554.

Holekamp, K.E., 1999. Vocal recognition in the spotted hyaena and its possible impli-cations regarding the evolution of intelligence. Anim. Behav. 58, 383–395.

Holekamp, K.E., 2007. Questioning the social intelligence hypothesis. Trends Cogn.Sci. 11, 65–69.

Holekamp, K.E., Smale, L., 1991. Dominance acquisition during mammalian socialdevelopment: the “Inheritance” of maternal rank. Am. Zool. 31, 306–317.

Hoppitt, W., Laland, K.N., 2008. Social processes influencing learning in animals: areview of the evidence. Adv. Study Behav. 38, 105–165.

Huber, L., Rechberger, S., Taborsky, M., 2001. Social learning affects objectexploration and manipulation in keas Nestor notabilis. Anim. Behav. 62,945–954.

Humphrey, N.K., 1976. The social function of intellect. In: Bateson, P.G., Hinde, R.A.(Eds.), Growing Points in Ethology. Cambridge University Press, Cambridge, pp.303–317.

Jolly, A., 1988. The evolution of purpose. In: Byrne, R.W., Whiten, A. (Eds.), Machi-avellian intelligence: social expertise and the evolution of intellect in monkeys,apes, and humans. Oxford University Press, New York, pp. 363–378.

Kavaliers, M., Colwell, D.D., Choleris, E., 2005. Kinship, familiarity and social sta-tus modulate social learning about “micropredators”(biting flies) in deer mice.Behav. Ecol. Sociobiol. 58 (1), 60–71.

Kendal, R.L., Coolen, I., Laland, K.N., 2009. Adaptive Trade-offs in the use of Socialand Personal Information. In: Dukas, R., Ratcliffe, J. (Eds.), Cognitive Ecology:The evolutionary ecology of learning, memory and information use. Universityof Chicago Press.

Kendal, R.L., Custance, D., Kendal, J.R., Vale, G., Stoinksi, T.S., Rakotomalala, N.L.,Rasamimanana, H., 2010. Evidence for social learning in wild lemurs (Lemurcatta). Learn. Behav. 38, 220–234.

Klopfer, P.H., 1959. Social interactions in discrimination learning with special refer-ence to feeding behaviour in birds. Behaviour 14, 282–299.

Laland, K.N., Reader, S.M., 1999. Foraging innovation in the guppy. Anim. Behav. 57(2), 331–340.

Leca, J.B., Gunst, N., Huffman, M.A., 2010. Indirect social influence in the maintenanceof the stone-handling tradition in Japanese macaques Macaca fuscata. Anim.Behav. 79, 117–126.

Lee, P., 1994. Social Structure and Evolution. In: Slater, P.J.B., Halliday, T.R. (Eds.),Behaviour and Evolution. Cambridge University Press, pp. 266–303.

Mathevon, N., Koralek, A., Weldele, M., Glickman, S.E., Theunissen, F.E., 2010. Whatthe hyena’s laugh tells: sex, age, dominance and individual signature in thegiggling call of Crocuta crocuta. BMC Ecology 10, 9e24.

Mersmann, D., Tomasello, M., Call, J., Kaminski, J., Taborsky, M., 2011. Simple mech-anisms can explain social learning in domestic dogs (Canis familiaris). Ethology117, 1–16.

Midford, P.E., Hailman, J.P., Woolfenden, G.E., 2000. Social learning of a novel for-aging patch in families of free-living Florida scrub-jays. Anim. Behav. 59 (6),1199–1207.

Moscovice, L.R., Snowdon, C.T., 2006. The role of social context and individual expe-rience in novel task acquisition in cottontop tamarins Saguinus oedipus. Anim.Behav. 71, 933–943.

Müller, C.A., Cant, M.A., 2010. Imitation and traditions in wild banded mongooses.Curr. Biol. 20 (13), 1171–1175.

Nel, J.A.J., 1999. Social learning in canids: an ecological perspective. MammalianSocial Learning, 259–277.

Nicol, C.J., Pope, S.J., 1994. Social learning in small flocks of laying hens. Anim. Behav.47 (6), 1289–1296.

Reader, S.M., Biro, D., 2010. Experimental indentification of social learning in wildanimals. Learn. behav. 38, 265–283.

Russon, A.E., 1997. Exploiting the expertise of others. In: Whiten, A., Byrne, R.W.(Eds.), Machiavellian intelligence II: Extensions and evaluations. Cambridge Uni-versity Press, Cambridge, United Kingdom, pp. 174–206.

Seferta, A., Guay, P.J., Marzinotto, E., Lefebvre, L., 2001. Learning differences betweenferal pigeons and zenaida doves: the role of neophobia and human proximity.Ethology 107, 281–293.

Shettleworth, S.J., 2009. Cognition, Evolution, and Behavior. Oxford University Press,USA.

Stoinksi, T.S., Wrate, J.L., Ure, N., Whiten, A., 2001. Imitative learning by captivewestern lowland gorillas (Gorilla gorilla gorilla) in a simulated food-processingtask. J. Comp. Psychol. 115, 272–281.

Thorpe, W.H., 1963. Learning and Instinct in Animals, Second ed. Methuen, London.van de Waal, E., Bshary, R., 2011. Social-learning abilities of wild vervet monkeys in

a two-step task artificial fruit experiment. Anim. Behav. 81, 433–438.van de Waal, E., Renevey, N., Favre, C.M., Bshary, R., 2010. Selective attention to

philopatric models causes directed social learning in wild vervet monkeys. Proc.Biol. Sci. 277, 2105–2111.

Visalberghi, E., Addessi, E., 2000. Seeing group members eating a familiar foodenhances the acceptance of novel foods in capuchin monkeys. Anim. Behav. 60,69–76.

Voelkl, B., Huber, L., 2000. True imitation in marmosets. Anim. Behav. 60 (2),

195–202.

Webster, S.J., Lefebvre, L., 2001. Problem solving and neophobia in acolumbiform–passeriform assemblage in Barbados. Anim. Behav. 62, 23–32.

Whiten, A., 1998. Imitation of the sequential structure of actions by chimpanzees(Pan troglodytes). J. Comp. Psychol. 112, 270–281.

Page 10

1 ioural

W

W

W

20 S. Benson-Amram et al. / Behav

hiten, A., Byrne, R.W., 1997. Machiavellian intelligence II: Extensions and evalua-tions. Cambridge University Press.

hiten, A., Mesoudi.F A., 2008. Establishing an experimental science of culture: ani-mal social diffusion experiments. Phil. Trans. R. Soc. B 363 (no. 1509), 3477–3488.

hiten, A., Custance, D.M., Gomez, J.C., Teixidor, P., Bard, K.A., 1996. Imitative learn-ing of artificial fruit processing in children (Homo sapiens) and chimpanzees (Pantroglodytes). J. Comp. Psychol. 110, 3–14.

Processes 109 (2014) 111–120

Whiten, A., Van Schaik, C.P., 2007. The evolution of animal ‘cultures’ and socialintelligence. Philos. Trans. R. Soc. Lond. B Biol. Sci. 362, 603–620.

Wright, G.S., Wilkinson, G.S., Moss, C.F., 2011. Social learning of a novel foraging taskby big brown bats Eptesicus fuscus. Anim. Behav. 82 (5), 1075–1083.

Zuberbuhler, K., Gygax, L., Harley, N., Kummer, H., 1996. Stimulus enhancement andspread of a spontaneous tool use in a colony of long-tailed macaques. Primates37, 1–12.