American Journal of Primatology 71:687–695 (2009)

RESEARCH ARTICLE

Effects of Positive Reinforcement Training Techniques on the PsychologicalWelfare of Zoo-Housed Chimpanzees (Pan troglodytes)

ORI POMERANTZ� AND JOSEPH TERKELDepartment of Zoology, The George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel-Aviv, Israel

Captive environments encompass various factors that can elevate stress levels and jeopardize thewellbeing of the captive animals. The use of positive reinforcement training (PRT) techniques enablesresearchers and caretakers to reduce tension directly associated with potentially stressful proceduresand states. The current study tested the general effect of PRT on the wellbeing of zoo-housedchimpanzees (Pan troglodytes) by measuring behaviors that reflect poor and good welfare and that werenot directly connected to the specific aim of the training session. The behavior of a group of twelvechimpanzees was measured throughout the day from the exhibition yard, at baseline (12 weeks) andduring the PRT period (10 weeks). The results show a significant decrease in abnormal and stress-related behaviors and a significant rise in prosocial affiliative behaviors following implementation of thetraining program. The training was shown to have a greater positive effect on low-ranking individualscompared with high-ranking ones. This research shows for the first time that PRT offers an enrichmenteffect whose general influence lasts throughout the day, irrespective of any direct link to a specifictrained behavior. Consequently, it can be claimed that PRT presents an effective enrichment tool thatcan be implemented with captive animals. Because of the above-noted differential effect between high-and low-ranking chimpanzees, however, this should be taken into consideration when combining PRTwith the non-human primates’ daily routine. Am. J. Primatol. 71:687–695, 2009. r 2009 Wiley-Liss, Inc.

Key words: chimpanzees; PRT; captivity; psychological welfare

INTRODUCTION

The psychological welfare of captive animals canbe compromised because of both biotic and abioticfactors in their surroundings [Morgan & Tromborg,2007]. Chimpanzees (P. troglodytes) held in zoos andresearch facilities, for instance, lack the ability tochoose what and when to eat, with whom they preferto share their immediate environment, or to freelyselect their mating partners. This limited abilityto choose, along with the lack of control overenvironmental factors and the restriction of naturalbehavioral requirements, impairs the animals’ psy-chological welfare [Dawkins, 1990]. Keeping chim-panzees under high density conditions, for example,was shown to result in high levels of stress that wereexpressed in elevated rates of yawning and scratching[Aureli & De Waal, 1997]; while attempts to introducenew individuals into an existing group of chimpanzeesoften resulted in injuries [Alford et al., 1995].

In addition, chimpanzees in captivity are highlyexposed to anthropogenic influences that can alsojeopardize their wellbeing and elevate their chronicstress levels. At the Los Angeles Zoo chimpanzeesexhibited fewer natural behaviors such as foragingand prosocial behaviors when a large crowd waspresent compared with when a small crowd was

present [Wood, 1998]. Similarly, Lambeth et al.[1997] reported that levels of aggression amongsocially housed chimpanzees increased on weekdays,when routine personnel activity in proximity to thechimpanzees was higher compared with such activ-ities on weekends.

Certain veterinarian procedures can also raisethe anxiety levels of chimpanzees. Anesthesia, forinstance, which is required in many procedures,results in the production of high levels of the ‘‘stresshormone’’—cortisol [Whitten et al., 1998].

Reducing the negative effects of captivitybecomes possible through the provisioning of variousstimuli according to the animals’ physiological andbehavioral needs [Bloomsmith & Else, 2005]. Fromthe behavioral aspect, the goal of environmental

Published online 11 May 2009 in Wiley InterScience (www.interscience.wiley.com).

DOI 10.1002/ajp.20703

Received 1 January 2009; revised 13 April 2009; revisionaccepted 14 April 2009

Contract grant sponsor: United Federation of Animal Welfare(UFAW SAWI).

�Correspondence to: Ori Pomerantz, Department of Zoology,The George S. Wise Faculty of Life Sciences, Tel Aviv University,Ramat Aviv 69978, Israel. E-mail: oripomer@post.tau.ac.il

rr 2009 Wiley-Liss, Inc.

enrichment is to reduce or eliminate abnormal andstress-related behaviors, as well as to increase thepotency of the natural behavioral repertoire, includ-ing prosocial behaviors. For example, as a result ofthe addition of straw and forage material to anindoor enclosure of socially housed chimpanzees,abnormal behaviors decreased and prosocial beha-viors, such as play, increased [Baker, 1997]. Inanother study, similarity between the behavior ofcaptive and wild chimpanzees increased as a result ofplacing uprooted trees in the outdoor part of thechimpanzees’ enclosure. This allowed the captivesubjects to express more species-typical behaviors,and thereby contributed to their welfare [Maki &Bloomsmith, 1989].

In most primate species sociality embodies animportant part of life and, therefore, keeping nonhu-man primates (NHP) in social groups rather than inisolation is largely accepted as the most effective typeof enrichment for these species [Lutz & Novak, 2005].Sociality can also be expressed in the interactionsbetween NHP such as chimpanzees and their care-takers. For example, caretakers can play an impor-tant role in redirecting aggression by captivechimpanzees toward the caretakers themselves ratherthan toward other chimpanzees during the sensitiveshort time period following an intra-group aggression[Malone et al., 2000]. Positive, unstructured socialinteractions (e.g. play and grooming) between care-takers and NHP led to more prosocial behaviors andfewer abnormal behaviors within a group of chim-panzees [Baker, 2004], and a similar effect was foundin rhesus monkeys [Bayne et al., 1993].

It has been suggested that positive interactionsbetween caretakers and NHP are also likely to occurduring positive reinforcement training (PRT) sincethe training sessions require the caretakers to spendmore time with the animals, build more trustbetween the two, and provide rewards to the trainedsubjects [Buchanan-Smith, 2003]. It is thereforeassumed that training NHP promotes a closerrelationship, which benefits both the humans andthe NHP [Prescott & Buchanan-Smith, 2003].Trainers can identify specific situations that areassociated with poor psychological wellbeing, orstates that pose a behavioral management difficulty,and aim the training directly at redressing thosesituations. In one such case researchers identifiedthe traditional procedure of injection of anesthetic toNHP as highly stressful, and therefore trained thechimpanzees to voluntarily present a leg for injec-tion. The trained subjects had lower levels of bloodcomponents indicative of acute stress than chimpan-zees that were involuntarily anesthetized [Lambethet al., 2006]. In another example, moving chimpan-zees between different sections of an enclosure ispotentially time consuming, and training them tocooperate during this procedure may facilitate thework of the caretakers [Bloomsmith et al., 1998].

The aggression of an adult male toward its groupmembers during feeding time was another issue thatwas addressed by training. In this case Bloomsmithet al. [1994] reported that training this male to sitwhile others received food resulted in lower levels ofagonism during the feeding. Other aspects of socialbehavior have been the target of various trainingprograms: Desmond & Laule [1994] reported anincrease in reproductive behavior in drill baboonsthat received reinforcement to touch one another,and in another study female rhesus macaques weretrained to alter their levels of affiliative interactions,which resulted in an effect beyond that of thetraining sessions on those individuals that had beendescribed as low affiliators at baseline [Schapiroet al., 2001].

According to Baer et al. [1968] it is possible toconsider a behavioral change as general if it issustainable over time, if it is present in variousenvironments, or if it extends to various relatedbehaviors. Unlike previous studies that examinedthe direct influence of PRT, the aim of the currentstudy was to test its general effect by measuringvarious behaviors that reflect poor and good welfareand that were not directly connected to the specificaim of the training session. Measuring these beha-viors throughout the day, at baseline and during thetraining period, was intended to reveal whether PRTmay offer a broader significance than that shown todate, and thus allow us to assess the overall effect ofthe training on the chimpanzees.

METHODS

Subjects and Housing

Twelve adult and subadult chimpanzees(P. troglodytes), housed at the Tel Aviv-Ramat GanZoological Center in Israel, were evaluated in thisresearch between January and November 2007 (seeTable I). These NHP (ten females and two males)spend the first part of the day (08:00–17:00) in anoutdoor exhibition yard (1,100 square meters), and

TABLE I. Sex and age of study subjects

Individual Sex Age

Lili F 47Shoshi F 42Rachel F 37Luna F 21Liki F 13Livnat F 10Augusta F 36Avishag F 21Aviv M 14Avigail F 12Ella F 9Avshalom M 8

Am. J. Primatol.

688 / Pomerantz and Terkel

the second part (17:00–08:00) in their night cages(3.1� 3.5� 3 m3), with 2–4 chimpanzees, usually ofthe same social composition, in each cage. The cagesfeature solid cement floors, sidewalls, and cage frontswith 3 cm spaces between the bars, as well as restingboards, and (in some of them) hammocks made fromfire hoses. After the chimpanzees exit their cages tothe yard, the doors are closed and they cannot returnto them. All chimpanzees were fed five meals a day,consisting mainly of fruits and vegetables. Water wasavailable ad lib both in the night cages and in theexhibition yard.

Procedures

The protocol of the study was approved by theTel-Aviv University Committee for Animal Care andUse, and the research complied with the protocolaccording to Israeli law. The research was dividedinto three stages. During Stage 1, the baseline stage,(12 weeks), no interruption was made to thechimpanzees’ daily routine. The NHP woke up atabout 07:00 and by 08:00 were usually in theexhibition yard after receiving breakfast in theircages. This was the only interaction between thecaretakers and the chimpanzees and it lasted a fewseconds with each individual. The experimentalphase of the study was divided into stages 2 and 3.During Stage 2 (10 weeks), six of the chimpanzeesunderwent a program, training them to presentvarious areas of their body for inspection. Trainingconsisted of brief morning sessions (5–10 min witheach individual), 5 days per week. Training wasconducted according to the principles of operantconditioning [Laule et al., 2003]. The first sessionswere aimed at training the subjects to hold twoobjects (targets) in their hands. Once this behaviorhad been shaped, the chimpanzees were furthertrained to present their feet, abdomen, back, hands,arms, shoulders, mouth, head, and ears for inspec-tion. Training sessions were carried out in thepresence of all the inhabitants of the cage. Twotrainers worked simultaneously, first with the twohigher-ranking chimpanzees in the cage, and thenwith the low-ranking ones, to prevent aggression. Inthe early stages of the experimental phase thoseindividuals that had finished their training sessionneeded to be further occupied (usually by providingfood and reinforcing them for not interfering) inorder to allow the rest of the chimpanzees to betrained without interference. This ensured that allsubjects received an equal amount of attention andtraining time. During Stage 3 (10 weeks), the trainingprogram continued for the first six individuals, whilesix more chimpanzees began their training sessionsin exactly the same manner. The first six chimpan-zees were trained for a total of 20 weeks, while theother six received training for 10 weeks.

In addition, the caretaker who cotrained thechimpanzees along with the first author, and hasbeen working with them for over 20 years, was askedto fill in a questionnaire at the end of Stages 1 and 3,to evaluate the levels of aggression of the chimpan-zees toward her, before and after the trainingprogram. The answers were given on a scale of1–10, where 1 represents the lowest level of aggres-sion toward the caretaker and 10 the most aggressivelevel. The caretaker’s evaluation was based onattempts at grabbing her, feces throwing, threats,and displays.

Data Collection

During the baseline phase (Stage 1) 89 hr of datawere collected, which represented approximately7.5 hr per subject (an average of 88 five-minute-longobservations per chimpanzee). In the second part ofthe experimental phase (Stage 3), 75 hr of data werecollected. This represented roughly 6 hr of data persubject (an average of 75 observations per chimpan-zee). During both Stages 1 and 3 abnormal andstress-related behaviors indicative of poor psycholo-gical welfare as well as prosocial affiliative behaviorsindicative of good welfare were assessed (seeTable II). The chimpanzees were randomly video-taped from the roof of the night cages while theywere in the exhibition yard, which enabled coveringmost of the yard area. When the focal subject left theframe, the videotaping was ceased and the datacollector repositioned to other viewing points aroundthe exhibition yard. Data collection was done onSunday, Tuesday, and Saturday, three times a day:morning (08:00–10:00), noon (12:00–14:00), andafternoon (15:00–17:00), using continuous focalanimal sampling [Martin & Bateson, 1993]. Videoswere analyzed using the Observer XT software andstatistical analyses were done using the SPSSversion 14 software. Data collected during Stage 1were compared with those collected during Stage 3.

Statistical Analysis

Data for each chimpanzee were pooled acrossfocal observations and analysis was performed usingpercentage of presented behavior out of the totalobservation time for each data-collection period (i.e.if a subject exhibited 500 sec of behavior X out of atotal pool of 21,000 sec from all observations, thenthe behavior occurred for 2.4% of the time).Mann–Whitney tests revealed that the results forthe six chimpanzees that were trained for 20 weeksdid not differ from those that were trained for 10weeks in all three behavioral categories. Accordingly,the subsequent analyses were performed for all 12chimpanzees. One-tailed Wilcoxon matched-pairsigned rank tests were performed to evaluate thesubjects for significant differences before and duringthe training program. a was set at 0.05.

Am. J. Primatol.

Effect of PRT on Chimpanzee Welfare / 689

RESULTS

Abnormal Behaviors

Abnormal behaviors were significantly reducedduring implementation of the training program(n 5 12; Z 5�2.4; Po0.01) (Fig. 1). The chimpanzeesexhibited negative behaviors for (6.471.1)% of thetime during baseline, and for (3.870.6)% of the timeduring Stage 3.

Examining the change in these behaviors inthe individual chimpanzees reveals differencesin the magnitude of their responses. Figure 2 showsthe level of behaviors in this category for eachchimpanzee.

During the baseline period abnormal behaviorsranged from 1.3 to 13.7% of the observed time forthe different chimpanzees and in the experimentalphase for 1.5–9.1% of the time. Some individualsexhibited a stronger reduction in these behaviors(e.g. Augusta, Ella, and Avishag), while othersshowed a more moderate response (e.g. Aviv, Shoshi,and Lili).

Stress-Related Behaviors

Similarly to the previous behavioral category,the stress-related behaviors were also significantlyreduced in the experimental period (n 5 12;Z 5�2.6; Po0.01) (Fig. 3). The chimpanzees had

TABLE II. Definitions of the behaviors evaluated in the research [based on: Ross and Lukas, 2001; Baker, 1997;Baker, 2004; Baker and Aureli, 1997]

Abnormal behaviors

Coprophagy Ingestion of fecesRegurgitation and reingestion Deliberate regurgitation accomplished by various methods including lowering head

to the ground, bobbing head, or more subtle techniques.The vomitus was eitherretained within the mouth or expelled into hand or substrate before beingreingested.

Hair plucking Pulling out own or another animal’s hair; may be ingested.Idiosyncratic movement or posture Sustained movement of body, such as rocking or head-bobbing, with a definitive

repetitive pattern.Eye poking Poking finger(s) into the eyeEar poking Poking finger(s) into one or both earsUrine drinking Ingesting urineThumb sucking Common usageStress-related behaviorsScratching Raking of fingernails over skin; either smaller movements of hand, or larger

sweeping scratching involving arm movement.Yawning Involuntary wide opening of the mouth accompanied by deep inhalation.Self-grooming Picking through and/or slow brushing aside of animals’ own hair or skin with the

mouth and/or one or both handsProsocial affiliative behaviorsSocial grooming Picking through hair or at skin of another individual and removing debris with hands

and/or mouth.Does not include pulling hair.Received grooming Receiving social groomingSocial playing Non-aggressive interactions involving two or more animals. Never accompanied by

pilo-erection or antagonism; may be accompanied by play-face and/or laughing.Includes rough-and-tumble play (fast-paced, vigorous locomotion, wrestling,hitting, pulling, chasing, biting, etc.), quiet play (slower-paced, gentle-tickling,finger and toe manipulation, etc.) and also includes social play initiation.

Tandem walking (Side by side) Walking while putting one or both arms around the body of the partnerfrom the side so that bodily contact is made. (Dorsal-ventral) walking immediatelybehind partner while simultaneously resting arms on the partner’s back orclutching arms around the partner’s thighs. Not necessary for both animals tohave arms resting on one another.

0

1

2

3

4

5

6

7

8

PRTBaseline

Mea

n %

tim

e (+

SE

)

Fig. 1. Levels of abnormal behaviors before and during imple-mentation of the training program (�Po0.01).

Am. J. Primatol.

690 / Pomerantz and Terkel

shown these behaviors for 15.0% of the time beforethe training program was implemented, and thisdropped to 10.2% of the time in Stage 3.

As before, individual differences are prominentboth in the time spent in these behaviors, and thelevel of change. While low-ranking chimpanzees suchas Liki, Avigail, and Livnat exhibited high levels ofstress-related behaviors during the baseline and amajor decrease during the experimental phase, thea-male, Aviv, showed an increase of 55.2% in thesebehaviors (Fig. 4).

Prosocial Affiliative Behaviors

Behaviors in this category increased followingthe initiation of PRT (n 5 12; Z 5�2.5; Po0.01).The chimpanzees had exhibited prosocial affiliativebehaviors for (14.972.4)% of the time during

baseline, and increased the time spent in engagingin those behaviors to (24.273.4)% during the secondpart of the experimental phase (Fig. 5).

As in the previous category, although a signifi-cant overall change (in this case an increase) in thesepositive behaviors was detected, some chimpanzeesrevealed a much higher increase in prosocial affilia-tive behaviors, while others exhibited a reduction inthese behaviors during the experimental phase.Figure 6 shows the scores for prosocial affiliativebehaviors for each individual. While some chimpan-zees showed large increases in these behaviors (Liki,Avishag, and Avigail), the a-male (Aviv), as well asthe oldest individual in the group (Lili), showed adecrease in this category.

Aggression Toward the Caretaker

The level of aggression of the chimpanzeestoward their caretaker/trainer was evaluated usinga questionnaire on a scale of 1–10 during the baselineand at the end of the experimental phase (where 1represents the lowest level of aggression and 10 thehighest).

The mean evaluated level of aggression was5.870.8 before any intervention was applied, andsignificantly decreased to 370.3 after the applicationof PRT (n 5 12; Z 5�2.83; Po0.01). A reduced levelof aggression was scored for all the subjects duringthe experimental phase except for two females(Livnat and Luna), which had been scored very lowat baseline and remained so throughout the experi-mental phase.

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

Augusta

Aviv

Avigail

Avshalom

Avishag El

laLiki

Luna Lil

i

Rachel

Livnat

Shoshi

% o

f ti

me

BaslinePRT

Fig. 2. Levels of abnormal behaviors before and during implementation of the training program for each chimpanzee.

0

2

4

6

8

10

12

14

16

18

PRTBaseline

Mea

n %

tim

e (+

SE

)

Fig. 3. Levels of stress-related behaviors before and duringimplementation of the training program (�Po0.01).

Am. J. Primatol.

Effect of PRT on Chimpanzee Welfare / 691

DISCUSSION

In their review, Lutz & Novak [2005] presenttwo hypotheses for promoting species-typical beha-viors: the first refers to the direct impact of the usageof enrichment objects; while the second has a widerand more generalized intent, as it refers to the abilityof enrichment to increase species-typical behaviorsother than those resulting from direct use of theenrichment object. For example, the addition of aforaging device is assumed to increase foragingbehavior in the animals, but it is also predicted toinfluence their daily activities and time budget.

In an analogy to the above, studies that haveexamined the impact of PRT on animal welfare todate have referred to its ability to decrease specificunwanted behaviors as a direct result of the ‘‘usage’’of the trained behavior [e.g. Bloomsmith et al., 1994;Reinhardt, 2003], whereas the current research

shows for the first time that, in addition, PRT alsoinfluences certain behaviors that are not directlylinked to such ‘‘usage,’’ thus indicating a generalityof the behavioral change [Baer et al., 1968]. In thisstudy the researchers did not attempt to identifyspecific points that are associated with poor welfareand direct the training at them, but, rather,examined the general enrichment effect of thetraining and its ability to improve the chimpanzees’welfare, which is chronically compromised by manydifferent factors in their captive surroundings.

Examining the behavior of the chimpanzees inthree distinct categories (abnormal behaviors, stress-related behaviors, and prosocial affiliative behaviors)revealed a major improvement in their wellbeingfollowing implementation of the PRT. This improve-ment was expressed in a significant reduction inabnormal behaviors and stress-related behaviors anda significant increase in prosocial affiliative beha-viors. Although the training sessions that wereaimed at presenting areas of the subjects’ body forinspection took place in the mornings, when thechimpanzees were still in their cages, the positiveeffect was seen throughout the day, in the exhibitionyard, during routine situations that are not relatedto any veterinarian procedure. This suggests thatPRT has a broader significance on the chimpanzees’psychological welfare than that considered until now,and that its effect lasts throughout the day.

As a result of the training, more positive socialinteractions occurred among group members. Simi-lar to the effect of unstructured positive interactionsbetween NHP and their caretakers [e.g. Bloomsmithet al., 1999], the more structured interaction that isestablished during PRT also reduces stress andabnormalities among the chimpanzees and enables

0.0

5.0

10.0

15.0

20.0

25.0

% o

f ti

me

Basline

PRT

Augusta

Aviv

Avigail

Avshalom

Avishag El

laLiki

Luna Lil

i

Rachel

Livnat

Shoshi

Fig. 4. Levels of stress-related behaviors before and during implementation of the training program for each chimpanzee.

0

5

10

15

20

25

30

PRTBaseline

Mea

n %

tim

e(+S

E)

Fig. 5. Levels of prosocial affiliative behaviors before and duringimplementation of the training program (�Po0.01).

Am. J. Primatol.

692 / Pomerantz and Terkel

more positive expressions of social behaviors withinthe group.

Previously it was found that positive interactionsbetween NHP and their caretakers promote goodrelationships between them [Waitt et al., 2002] andalter their relations from predominantly aggressive tofriendly [Baker, 2004]. In this study the caretaker’sevaluation of aggression toward her revealed similareffects. Increasing the time spent with each individualfrom a few seconds in the baseline period to 5–10 minper day during the test phase, and filling it withpositive, rewarding interactions, led to fewer attemptsat grabbing, scratching, and feces throwing. It shouldbe noted that these results must be viewed withcaution since they are based on the evaluation of onlyone caretaker who was not blinded to the study’spurpose. Nevertheless, this particular caretaker hasmany years experience of caring for this group ofchimpanzees and spends significantly more timeobserving and interacting with them compared withthe other caretakers. A further assessment based on alarger sample size of evaluators is however required toestablish a more firm conclusion.

In the current study, however, even though theoverall psychological welfare of the chimpanzeegroup was demonstrated to improve to a large extent,it is not surprising that profound individual differ-ences were found among the different members.

In high-ranking individuals (Aviv, Shoshi, andLili) it was shown either to improve only mildly or, insome cases, to worsen, while in lower-rankingindividuals (i.e. Avigail, Liki, and Ella), it improvedsubstantially.

Stress-related behaviors have been shown topositively correlate with increased intra-group

tension, and with the risk of aggression [Baker &Aureli, 1997], while prolonged exposure to stressorsand the lack of appropriate sensory input have beenshown to promote the appearance of abnormalbehaviors [Hediger, 1964; Moodie & Chamove,1990]. Dominant individuals have a naturally higherlevel of good welfare than that of low-rankingchimpanzees, since they are less prone to sufferfrom lack of stimulation [Sapolsky, 2005]. Thus,their need for further stimulation in the form of arewarding interaction with the caretaker (such asthose provided in PRT sessions) is possibly weakerthan that of low-ranking individuals. Therefore, itcan be suggested that the impact of such interactionon the welfare of the high-ranking individuals wasmilder than that observed among the low-rankingchimpanzees. Abnormal behaviors may increasebecause of stressful procedures (e.g. veterinarian),isolation [Baker, 1996], and insufficient sensorystimulation [Warniment & Brent, 1997]. Providingthe chimpanzees with mental and social stimulationthrough the use of PRT, and enabling the low-ranking ones the same amount of stimulation as thehigh-ranking individuals, resulted in a higher impacton the former than on the latter, which usually,because of their lower social status, would havedenied this [Celli et al., 2003].

Moreover, the training sessions were held whileall the inhabitants of the cage were present, andindividuals that were not being trained werereinforced not to interfere with the training of theircage mates. It is reasonable to assume that trainingthe chimpanzees together, while concomitantlyreinforcing those in the cage that were not beingtrained, not to be aggressive, contributed to the

0.0

10.0

20.0

30.0

40.0

50.0

60.0

Augusta Av

iv

Avigail

Avshalom

Avishag El

la Liki

Luna Lil

i

Rachel

Livnat

Shoshi

% o

f ti

me Basline

PRT

Fig. 6. Levels of prosocial affiliative behaviors before and during implementation of the training program for each chimpanzee.

Am. J. Primatol.

Effect of PRT on Chimpanzee Welfare / 693

significant increase in affiliative behaviors [as wasshown by Schapiro et al., 2001], as well as to thereduction in stress. The stress reduction wasparticularly prominent among the low-rankingchimpanzees, which are usually the target of aggres-sion rather than the high-ranking ones.

It was previously reported that long-term ex-posure to stressors often results in lower levels ofpro-social behaviors [Moodie and Chamove, 1990],and presumably the greater stress-reduction amongthe low-ranking chimpanzees promoted the greaterincrease in pro-social behaviors that was observed inthese low-ranking subjects.

Since social stimulus is considered to be a highlyimportant part in the lives of many primate species[e.g. Dettmer & Fragaszy, 2000], increased positivesocial interactions within the group are believed toconduce to a greater reduction in abnormal beha-viors among low-ranking chimpanzees.

To conclude: The rise in prosocial affiliativebehaviors and the decrease in abnormal and stress-related behaviors found in this study are in accor-dance with the findings of previous studies, whichshowed that positive interactions with their care-takers contributed to the welfare of NHP. Waitt et al.[2002], for instance, reported that a friendly relation-ship between caretakers and stump-tailed macaques(Macuca arctoides) led to reduced tension in themonkeys, and as a result to an increase in positivesocial interactions.

It should be noted that in the current researchthe chimpanzees were held in the indoor area forabout 15 hr a day. These night cages are fairly smalland barren, and are not as complex and stimulatingas the settings in other zoos may be. Thus, it ispossible that under better social and physicalconditions the effect of PRT on the psychologicalwelfare of the chimpanzees may be weaker. Thiswould undoubtedly be interesting to study in futureresearch.

In the present study, rather than examining theendpoint of the training (i.e. the specific shapedbehavior), the entire process of the training, whichincluded the strengthening of the caretaker–chim-panzee relationship, mental stimulation, and arewarding atmosphere, was examined. The findingsshow for the first time that PRT offers an enrich-ment effect whose general influence lasts throughoutthe day, irrespective of any direct link to the specifictrained behaviors. Consequently, it can be claimedthat PRT presents an effective enrichment tool thatcan be implemented with captive animals.

Because of the demonstrated differential effectbetween high- and low-ranking chimpanzees, how-ever, this should be taken into consideration whencombining PRT with the NHP daily routine.

Finally, in addition to reducing stress levels andaggression in specific situations, PRT has also beenshown here to possess qualities that engender a more

general effect, leading to an overall improvement inthe psychological welfare of zoo-housed chimpanzees.

ACKNOWLEDGMENTS

We express our gratitude to Mr. Yehuda Bar,Dr. Amelia Terkel, Mr. Ofer Matalon and Mrs. NatiKovatz-Tam of the Tel Aviv-Ramat Gan ZoologicalCenter for their willing cooperation during theresearch. We also thank Dr. Virginia Landau andMs. Erica Metelovski at Chimpanzoo for their much-appreciated assistance. We gratefully acknowledgeMs. Naomi Paz for her help in preparing and editingthis manuscript. This research complied with thelaws and regulations applicable to animal care inIsrael.

REFERENCES

Alford PL, Bloomsmith MA, Keeling ME, Beck TF. 1995.Wounding aggression during the formation and mainte-nance of captive multimale chimpanzee groups. Zoo Biol14:347–359.

Aureli F, De Waal FBM. 1997. Inhibition of social behavior inchimpanzees under high density conditions. Am J Primatol41:213–228.

Baer DM, Wolf MM, Risley TR. 1968. Some current dimen-sions of applied behavioral analysis. J Appl Behav Anal1:91–97.

Baker KC. 1996. Chimpanzees in single cages and small socialgroups: effect of housing on behavior. Contemp Top LabAnim Sci 35:71–74.

Baker KC. 1997. Straw and forage material ameliorateabnormal behaviors in adult chimpanzees. Zoo Biol16:225–236.

Baker KC. 2004. Benefits of positive human interaction forsocially housed chimpanzees. Anim Welf 13:239–245.

Baker KC, Aureli F. 1997. Behavioral indicators of anxiety: anempirical test in chimpanzees. Behaviour 134:1031–1050.

Bayne K, Dexter SL, Strange GM. 1993. The effects of foodtreat provisioning and human interaction on the behavioralwell-being of rhesus monkeys (Macaque mulatta). ContempTop Lab Anim Sci 32:6–9.

Bloomsmith MA, Laule GE, Alford PL, Thurston RH. 1994.Using training to moderate chimpanzee aggression duringfeeding. Zoo Biol 13:557–566.

Bloomsmith MA, Stone AM, Laule GE. 1998. Positivereinforcement training to enhance the voluntary movementof group-housed chimpanzees within their enclosures. ZooBiol 17:333–341.

Bloomsmith MA, Baker KC, Ross SK, Lambeth SP. 1999.Comparing animal training to non-training human interac-tions as environmental enrichment for chimpanzees. Am JPrimatol 49:35–36 (Abstract).

Bloomsmith MA, Else JG. 2005. Behavioral management ofchimpanzees in biomedical facilities: the state of the science.ILAR J 46:192–201.

Buchanan-Smith HM. 2003. The benefits of positive reinforce-ment training and its effects on human non-human animalinteractions. Proceedings of the Fifth Annual Symposiumon Zoo Research. p 21–26.

Celli ML, Tomonage M, Udono T, Teramoto M, Nagano K.2003. Tool use task as environmental enrichment for captivechimpanzees. Appl Anim Behav Sci 81:171–182.

Dawkins MS. 1990. From an animal’s point of view: motiva-tion, fitness, and animal welfare. Behav Brain Sci 13:1–61.

Am. J. Primatol.

694 / Pomerantz and Terkel

Desmond T, Laule G. 1994. Use of positive reinforcementtraining in the management of species for reproduction. ZooBiol 13:471–477.

Dettmer E, Fragaszy D. 2000. Determining the value of socialcompanionship to captive tufted capuchin monkeys (Cebusapella). J Appl Anim Wel Sci 3:293–304.

Hediger H. 1964. Wild animals in captivity. New York: DoverPress.

Lambeth SP, Bloomsmith MA, Alford PL. 1997. Effects of humanactivity on chimpanzee wounding. Zoo Biol 16:327–333.

Lambeth SP, Hau J, Perlamn JE, Martino M, Schapiro SJ.2006. Positive reinforcement training affects hematologicand serum chemistry values in captive chimpanzees (Pantroglodytes). Am J Primatol 68:245–256.

Laule GE, Bloomsmith MA, Schapiro SJ. 2003. The use ofpositive reinforcement training techniques to enhance thecare, management, and welfare of primates in the labora-tory. J Appl Anim Welf Sci 6:163–173.

Lutz CK, Novak MA. 2005. Environmental enrichment fornonhuman primates: theory and application. ILAR J46:178–191.

Maki S, Bloomsmith MA. 1989. Uprooted trees facilitate thepsychological well-being of captive chimpanzees. Zoo Biol8:79–87.

Malone N, Vaughan L, Fuentes A. 2000. The role of humancaregivers in the post-conflict interactions of captivechimpanzees (Pan troglodytes). Lab Anim Newsl 39:1–3.

Martin P, Bateson P. 1993. Measuring behaviour, 2nd ed.Cambridge: Cambridge University Press. 222p.

Moodie EM, Chamove AS. 1990. Brief threatening eventsbeneficial for captive tamarines? Zoo Biol 9:275–286.

Morgan K, Tromborg C. 2007. Sources of stress in captivity.Appl Anim Behav Sci 102:262–302.

Prescott MJ, Buchanan-Smith HM. 2003. Training nonhumanprimates using positive reinforcement techniques. J ApplAnim Welf Sci 6:157–161.

Reinhardt V. 2003. Working with rather than againstmacaques during blood collection. J Appl Anim Welf Sci6:189–197.

Ross SK, Lukas KE. 2001. Exhibit use by great apes as part ofpost-occupancy evaluation. Am J Primatol 54:31.

Sapolsky RM. 2005. The influence of social hierarchy onprimate health. Science 308:648–652.

Schapiro SJ, Perlman JE, Boudreau BA. 2001. Manipulatingthe affiliative interactions of group-housed rhesus macaquesusing positive reinforcement training techniques. Am JPrimatol 55:137–149.

Waitt C, Buchanan-Smith HM, Morris K. 2002. The effects ofcaretaker–primate relationship on primates in the labora-tory. J Appl Anim Welf Sci 5:309–319.

Warniment A, Brent L. 1997. Abnormal behavior in a captivechimpanzee colony. Newsletter 8:1–3.

Whitten PL, Stavisky R, Aureli F, Russel E. 1998. Responseof fecal cortisol to stress in captive chimpanzees(Pan troglodytes). Am J Primatol 44:57–69.

Wood W. 1998. Interactions among environmental enrich-ment, viewing crowds, and zoo chimpanzees (Pantroglodytes). Zoo Biol 17:211–230.

Am. J. Primatol.

Effect of PRT on Chimpanzee Welfare / 695