factors affecting cashew processing by wild bearded

RESEARCH ARTICLE

Factors Affecting Cashew Processing by Wild Bearded Capuchin Monkeys(Sapajus libidinosus, Kerr 1792)

ELISABETTA VISALBERGHI1*, ALESSANDRO ALBANI1,2, MARIALBA VENTRICELLI1, PATRICIA IZAR3,GABRIELE SCHINO1, AND DOROTHY FRAGAZSY4

1Istituto di Scienze e Tecnologie della Cognizione, Consiglio Nazionale delle Ricerche, Rome, Italy2Dipartimento di Scienze, Universit�a degli Studi Roma Tre, Rome, Italy3Department of Experimental Psychology, University of S~ao Paolo, S~ao Paolo, Brazil4Department of Psychology, University of Georgia, Athens, Georgia

Cashew nuts are very nutritious but so well defended by caustic chemicals that very few species eatthem.We investigated howwild bearded capuchinmonkeys (Sapajus libidinosus) living at FazendaBoaVista (FBV; Piau�ı, Brazil) process cashew nuts (Anacardium spp.) to avoid the caustic chemicalscontained in the seedmesocarp.We recorded the behavior of 23 individuals toward fresh (N¼ 1282) anddry (N¼ 477) cashew nuts. Adult capuchins used different sets of behaviors to process nuts: rubbing forfresh nuts and tool use for dry nuts. Moreover, adults succeed to open dry nuts both by using teeth andtools. Age and body mass significantly affected success. Signs of discomfort (e.g., chemical burns,drooling) were rare. Young capuchins do not frequently closely observe adults processing cashew nuts,nor eat bits of nut processed by others. Thus, observing the behavior of skillful group members does notseem important for learning how to process cashew nuts, although being together with group memberseating cashews is likely to facilitate interest toward nuts and their inclusion into the diet. Thesefindingsdiffer from those obtained when capuchins crack palm nuts, where observations of others cracking nutsand encounters with the artifacts of cracking produced by others are common and support youngindividuals’ persistent practice at cracking. Cashew nut processing by capuchins in FBV appears todiffer from that observed in a conspecific population living 320km apart, where capuchins use tools toopen both fresh and dry nuts. Moreover, in the latter population, chemical burns due to cashew causticcompounds appear to be common. The sources of these differences across populations deserveinvestigation, especially given that social influences on youngmonkeys learning to open cashew nuts atFBV seem to be nonspecific. Am. J. Primatol. 78:799–815, 2016. © 2016 Wiley Periodicals, Inc.

Key words: Anacardium spp.; caustic compounds; age-related ability; social facilitation of eating;tool use

INTRODUCTIONA food item is worth eating when there is a

positive balance between the benefits of eating andthe costs of obtaining it. The benefits can beexpressed in terms of nutritional value of the food,whereas the costs involve the effort to find the foodand process it so as to obtain its edible part. Primateshave anatomical, physiological, and behavioraladaptations to overcome the mechanical and chemi-cal challenges of foods [Lambert, 2007].

Capuchin monkeys (genera Cebus and Sapajus)are omnivorous Neotropical primates that often feedon high quality foods that need to be manipulatedand extracted, such as nuts (even if protected by hardshells), invertebrates (including those hidden insidewoody substrates), and vertebrates [for a review seeFragaszy et al., 2004a]. The actions of capuchins’feeding repertoire are based on dexterity andphysical strength [Fragaszy et al., 2004a]. Both

genera are skilled extractive foragers, able to accessembedded foods and to overcome plant and animalresistance [Fragaszy et al., 2004a] and extensivelaboratory research has been carried out on their griptypes, grasping abilities and developmental changesin manipulation [Fragaszy and Adams-Curtis, 1997;

Contract grant sponsor: Leakey Foundation

�Correspondence to: Elisabetta Visalberghi, Istituto di Scienzee Tecnologie della Cognizione, Consiglio Nazionale delleRicerche, Via Aldrovandi, 16B,00197 Roma (Italy).E-mail: [email protected]

Received 12 October 2015; revised 2 March 2016; revisionaccepted 3 March 2016

DOI: 10.1002/ajp.22545Published online 23 March 2016 in Wiley Online Library(wileyonlinelibrary.com).

American Journal of Primatology 78:799–815 (2016)

© 2016 Wiley Periodicals, Inc.

Spinozzi et al., 2004; Truppa et al., 2016]. Moreover,the masticatory morphology of tufted capuchins(genus Sapajus), more robust than in the gracilecapuchins (genus Cebus), allows the former genus tobroaden its dietary niche to include mechanicallyprotected fruits, bromeliads, and other vegetationwhen fleshy fruit is relatively scarce [Wright, 2005;Wright et al., 2009]. Finally, a few species within thegenusSapajus are also habitual tool users in thewildand feed on food itemswhich they could not otherwiseobtain [S. libidinosus, Canale et al., 2009; DeMoraeset al., 2014; Ferreira et al., 2010; Fragaszy et al.,2004b; Mannu and Ottoni, 2009; Mendes et al., 2015;Moura and Lee, 2004; Spagnoletti et al., 2011; Wagaet al., 2006; S. xanthosternos, Canale et al., 2009;S. flavius, Ferreira et al., 2010; Souto et al., 2011].

Complex procedures to process food requirepractice to develop adult skills that eventuallycontribute to fitness [Janson and van Schaik,1993]. White-faced capuchin monkeys (Cebuscapucinus) become proficient atmulti-step extractiveforaging techniques following years of practice [Perryand Jim�enez, 2012]. White-faced capuchins openSloanea terniflora (a fruit thickly covered withurticating hairs that can be extremely irritating tothe eyes and skin) and Luehea candida (woodydehiscent pods 5–8 cm long, containing small wingedseeds) by performing rubbing, pounding, and fulcrumuse (i.e., applying force on an object working againsta substrate used as fulcrum) [O’Malley and Fedigan,2005a; Panger et al., 2002]. These relatively complexbehaviors are acquired slowly and age-related differ-ences in feeding competence are evident [O’Malleyand Fedigan, 2005a] as well as differences acrossfield sites and individuals [O’Malley and Fedigan,2005b; Panger et al., 2002]. Juvenile Sapajus apellaslowly learn to locate and extract invertebrate larvaeembedded in bamboo stalks [Gunst et al., 2008,2010a,b]. Finally, tufted capuchins become able touse stone hammers and stone (or wooden) anvils toopen palm nuts after a lot of practice when about4 years of age or older [Fragaszy et al., 2013; Resendeet al., 2008; Spagnoletti et al., 2011].

The cashew tree (Anacardium spp.), native to theNortheast of Brazil, produces “double fruits” consist-ing of an accessory hypocarp (also called pseudo-fruitor apple; hereafter, apple) and a hard drupe [Mitchelland Mori, 1987]. The drupe (hereafter, nut) containsthe cashew seed (hereafter, kernel) which is pro-tected by a mesocarp with a honeycomb structurecontaining the cashew nut shell liquid (hereafter,CNSL), strongly irritant to skin and mucosae[Mitchell and Mori, 1987; Rosen and Fordice,1994]. The mesocarp contains caustic chemicalsboth when nuts are immature, soft, and green(hereafter, fresh nuts) and when mature, dry, tough,and gray/brown (hereafter, dry nuts) (Fig. 1). In thefresh nut, the CNSL is more liquid, whereas in thedry nut, it is more viscous. Fresh cashew nuts appear

to present no more mechanical demands than otherfruits that capuchins bite open, while dry nuts areharder, though adult capuchins of both sexes can stillcrack them with their teeth [Chalk et al., 2016; JChalk. Unpublished data; E Visalberghi. Personalobservation].

Apparently, humans and bearded capuchins arethe only primate species that process cashew nuts toeat the kernels. Traditionally, humans roast or boilthe dry nuts to decarboxylate anacardic acid, themain component of CNSL [Gallina Toschi et al.,1993; Mohod et al., 2010]. In modern factories, nutsare cooked and protective clothing, such as glovesand masks, are worn by workers who handle them.So far, reports on S. libidinosus accessing the seed ofcashew nuts concern two populations living 320kmapart in the State of Piau�ı (Brazil): namely, thepopulation of Serra da Capivara National Park(hereafter SCNP) for which cashew nuts are themost common encased food exploited with tools[Fal�otico, 2011] and the population of Fazenda BoaVista (hereafter, FBV; [Sirianni and Visalberghi,2013]) for which the most common encased foodexploited with tools are several species of palm nuts[Spagnoletti et al., 2011].

Wild bearded capuchin monkeys of FBV exploitcashew trees by eating both the apple and the kernelcontained inside the nut. In a pilot study, Sirianni andVisalberghi [2013] reported that the monkeys at FBVuse different techniques to access the kernel of freshand dry nuts. When the nut is fresh, typicallycapuchins repeatedly rub it on rough surfaces untilthe shell is partially open and then they extractthe kernel with their index finger. In contrast,when the nut is dry, capuchins typically use a tool tocrack the shell and expose the kernel. Possibly, thesetwo techniques accommodate the different mechanicaland chemical challenges associated with nuts ofdiffering maturational stages. When the nut is fresh,it may be better to perforate the mesocarp by rubbingthe nut on a tree limb than cracking itwith a tool, sincethe latter behavior is likely to contaminate the kernelwith the caustic liquid. This contamination is indeedrisky: for the S. libidinosus living in the SCNP thatstrike fresh cashew nuts with stones, rather than rubthem, is not uncommon to have blisters on their lips[Fal�otico, 2011; Luncz et al., 2015]. So far, there are noreports concerning other populations of S. libidinosus,other species of Sapajus and Cebus, or other non-human primate species consuming the kernels ofcashew nuts, although reports of other species ofprimates eating the fleshy cashew apples are common(e.g.,Alouatta palliata [Glander, 1982];Ateles geoffroyi[Carpenter, 1935];Papio sp.,Colobus guereza [Lee andPriston, 2005]; Cebus olivaceus [John Robinson.Personal communication]; C. capucinus, [Perry andManson, 2008]). Also chimpanzees raiding cashewplantations eat only the cashewapple anddrop the nuton the ground [Hockings and Sousa, 2012].

Am. J. Primatol.

800 / Visalberghi et al.

Cashew nuts are high in energy. One hundredgrams of cashew kernels collected in FBV correspondto 580Kcal distributed in a balanced manner acrosscarbohydrates, fat, and protein (value calculatedfrom Peternelli [2015]). Moreover, the cashew kernelcontains all the essential amino acids in the rightproportions, and the United States Department ofAgriculture considers cashews a food of choice forhumans [USDA, 2015]. Therefore, we assume thatcashews are a food of choice for other omnivorousprimates, including capuchin monkeys.

Thus, cashews nuts are afirst-rate food protectedby a powerful chemical deterrent that can be avoidedby using particular processing techniques the acqui-sition and expression of which may vary in accordwith age, body mass, and social influences. The firstgoal of this study was to corroborate Sirianni andVisalberghi [2013] preliminary findings on the typesof processing behaviors used by capuchin monkeysfor cashew apples and nuts using targeted samplingof a larger number of individuals, including youngones. The second goal was to determine whether ageand body mass influenced success and time spentfeeding on cashew apples and extracting the kernels.The third goal was to find out whether socialinfluences were likely to support the acquisition ofnut processing skills. For this purpose, we assessedwhether the attention paid by youngsters wasselectively directed to proficient group members.We also monitored whether the monkeys collectedleftovers previously owned by group members andwhether they showed evidence of having contactedthe caustic CNSL.

Age and ExperienceCapuchin monkeys’ diet includes many foods

that require special handling and extraction techni-ques [e.g., Agostini and Visalberghi, 2005; Fragaszy,1986; Gunst et al., 2008, 2010a; Melin et al., 2014;O’Malley and Fedigan, 2005a,b; Panger et al., 2002;Perry and Jim�enez, 2012; Wright et al., 2009]. Sincecashewnuts require extractive processing aswell, weexpected the age of an individual (that reflects her/his experience in processing food), will correlate withefficiency, although it is likely that efficiency reaches

an asymptote at a certain age, as it does for otherfoods [Resende et al., 2014]. Many food processingbehaviors emerge early in life but need extendedpractice to be effective or properly executed. Alreadyin the first year of life, capuchins perform rubbingand pounding behaviors, though ineffectively, tomodify food items [Fragaszy et al., 2004a]. Althoughpounding objects on hard surfaces appears already inthe first year of life, the effective use of hammer andanvil tools to crack palm nuts does not occurroutinely until capuchins are about 4 years of ageor older [Eshchar et al., submitted a; Resende et al.,2008]. Finally, the effective behavior used by white-faced capuchins to obtain the seeds of Lueheacandida are performed more often and are betterorganized in adults than in youngsters [O’Malley andFedigan, 2005a].

Therefore, in the present study, we expected thatan individual’s probability of success at obtaining thekernel increases with age and time spent processingan individual cashew nut decreases with age.Moreover, since cracking a fresh nut with a stonecontaminates the kernel with caustic chemicalspresent in the liquid of the mesocarp [Sirianni andVisalberghi, 2013], we expected proficient capuchinsto use tools to exploit dry nuts, but not to exploit freshnuts. Finally, since rubbing breaches the mesocarpwhen the nut is fresh but not when it is dry, weexpected proficient capuchins to rub only fresh nuts.

Body MassTo breach the mesocarp and access the kernel of

fresh nuts, forceful rubbing is necessary. In contrast,to crack the mesocarp of dry nuts with a tool (andaccess its kernel) the strikes should not be forcefulsince the shell is not very hard. Studies carried out ontool use to crack palmnuts in FBVdemonstrated thatbody mass (which correlates positively with themaximum force of an individual’s strike) affectssuccess at cracking high resistance nuts but not atcracking low resistance nuts and that, for both typesof palm nuts, sex per se does not affect success[Fragaszy et al., 2010; Spagnoletti et al., 2011].Therefore, we expected body mass to positivelycorrelate with probability of success at opening fresh

Fig. 1. Sagittal section of a fresh cashew nut (left); note the white kernel and in the close-up the spongymesocarp whose tubules containthe CNSL. Cracked dry cashew nuts with the tough mesocarp containing caustic resin (right).

Am. J. Primatol.

Cashew Processing in Wild Capuchins / 801

cashew nuts that need to be rubbed but not tocorrelate with probability of success at opening drycashew nuts that do not require forceful strikes.

Attention to OthersIn many species, extractive foraging actions and

manipulative activities are socially influenced, andsocial influences appear to contribute to the acquisi-tion of skills in these activities [for a review seeRapaport and Brown, 2008]. The processing techni-ques used to exploit cashews in FBV are excellentcandidates for social influences. Although datacollection in the present study was not specificallydesigned to investigate this topic, as has been donefor stone tool use in bearded capuchin monkeys[Eshchar et al., in press; Fragaszy et al., 2013], ourmethod allows us to assess whether youngsters paymore attention to more proficient individuals than toless proficient ones.

Avoidance and Acceptance of Risky FoodAversive experiences with a novel food induce

food aversion [Garcia et al., 1955; for a discussion onthe phenomenon in primates, see Visalberghi, 1994].Therefore, individuals unable to process nuts with-out contacting the CNSL should be less interested ingetting cashew nuts than those who have acquiredthe proper processing technique. However, as indi-viduals become more proficient, they should over-come their reluctance and exploit this high energyfood. Information on wild primates feeding on itemscontaining caustic substances is absent. However,O’Malley and Fedigan [2005a] suggest that the rareconsumption of Sloanea by juvenile C. capucinusmay also reflect a form of risk-aversion since thehairs defending these fruits produce irritation.Therefore, we expected infants and young juveniles(i) to contact the “risky” nuts less than adults and (ii)to contact the “non-risky” apples and eat them to thesame extent as adults.

METHODSStudy Site, Groups, and Data Collection

Thedatawere collected inFBV(9°390 S,45°250W),in the State of Piau�ı, Brazil [for a detailed description,see Visalberghi et al., 2007 and Spagnoletti etal., 2012]. Cashew trees are common in the transitionarea between Cerrado (wooded savannah) andCaatinga (thorny bush) [Oliveira and Marquis,2002] where FBV is situated.

Systematic data were collected in 2012 and 2013betweenmid-August and lateOctober, that is, duringthe cashew season. Subjects belonged to the Chicaogroup, a fully habituated group of Sapajus libidino-sus (previously Cebus libidinosus, [see Lynch Alfaroet al., 2012a,b]). This group consisted of 19

individuals in 2012 and 23 individuals in 2013.Table I reports detailed information about the age ofeach subject when data were collected. Individualsbelow 18 months of age (which is the averageweaning age at FBV; EthoCebus. Unpublisheddata) were considered infants, individuals olderthan 5 years, or about this age in 2012 (as Doreewho was already mother [Fragaszy et al., 2016] andPati who had the same age of Doree) were consideredadults, finally individuals in-between were consid-ered juveniles. Note that hereafter when referring toboth infants and juveniles, we will use the termyoungsters.

The group was followed 6 days a week, fromdawn to dusk. Focal animal follows were conductedusing continuous recording of behaviors [Martin andBateson, 1993] on a single subject. Observationsstarted when the focal subject was in a cashew tree(or within 5m from it andmoved into the cashew treewithin 3min). The observations lasted 10�3min.Whenever the focal subject was out of sight for morethan 3min, the observation ended and the data wereexcluded. Whenever the subject was engaged incashewprocessingwithin 10min after the start of theobservation, the observation was continued until thefood item was exploited; focal sessions spanning lessthan 7min were discarded from the analysis. Foreach subject, we collected 15 observations in 2012and an average of 19.9 observations in 2013 (range3–27; all subjects with 18 or more sessions and Patiand Cangaceiro, two peripheral males, with threesessions). We sought to collect a focal observationfrom all group members before beginning a furtherobservation with any subject. When choosing thesubject of the next observation, we actively searchedfor the least observed individuals, but if the latterwere absent or away from a cashew tree, weproceeded to observe other individuals. Withineach round of observations, the order in whichindividuals were sampled was random.

In 2012, data were collected by A.A. and a fieldassistant (Marino Junior Fonseca de Oliveira). In2013, data were collected by M.V. and a fieldassistant (Marcio Fonseca de Oliveira). In 2013,when nuts were dry a second field assistant (RonePires de Oliveira) weighed the hammers and identi-fied their material (see below). Observational datawere collected using personal digital assistants(Hewlett Packard iPAQ Pocket) equipped with TheObserver XT 10 software package (Noldus Informa-tion Technology) after a training period duringwhichinter-observer reliability assessed by kappa coeffi-cient for behavioral patterns reached þ0.89 betweenA.A. andM.J.F.O.,þ0.87 betweenA.A. andM.V., andþ0.84 between M.V. and M.F.O.

The behaviors recorded with The Observer aredefined in Table II (see the videos provided asSupplemental Material S1–S7). We recorded thefeeding behavior performed by the focal subject and

Am. J. Primatol.


the food target (apple, fresh nut, or dry nut). Acashew nut is fresh when it has an immature greenexocarp and a spongy mesocarp. A cashew nut is drywhen it has a gray/brown exocarp and a toughmesocarp (see Fig. 1). Success was scored when theapple was eaten and the leftover discarded, or whenthe seed contained in the nut was eaten and theexocarp discarded. Processing time was the sum ofthe duration of all the behaviors directed to the targetfood from when it was collected until the food wasconsumed or discarded.

When monkeys cracked dry nuts with a tool, wescoredwhether the anvil they usedwas on the ground(e.g., a stone, a root) or above the ground (e.g., a treebranch). Moreover, in 92 episodes of tool use, werecorded the tool material (sandstone, siltstone, orpalm nut shell). In 80 episodes (out of the 92), we alsoweighed the tool.

For the purpose of evaluating possible socialinfluences, we recorded all the occurrences in whichthe focal subject (i) took a food item partiallyconsumed by another individual; (ii) looked (for anyduration) at a groupmember fromadistance of 1morless while the other processed a nut or an apple (theidentity of the groupmemberwas also noted). For thepurpose of evaluating possible responses to CNSL,we recorded if the focal subject scratched its hands or

face, and if she/he drooled. Finally, at the end of eachobservation, we noted whether the subject’s face orhands exhibited evidence of chemical burn, such asblisters.

We adhered to the American Society of Prima-tologists/Association for the Study of Animal Behav-iour guidelines for the treatment of animals inbehavioral research and teaching and to the Ameri-can Society of Primatologists’ principles for theethical treatment of primates. Permits to EthoCebusresearch were given by IBAMASISBIO: 28689-3 andCNPq: 002547/2011-2.

Data AnalysisMost analyses were based on generalized linear

mixed models. Depending on the nature of thedependent variable, we used logistic regressions (forbinary variables), negative binomial regressions (forcount variables), and linear regressions (for continu-ous variables). A first set of analyses compared theprobability of success and the time taken to open thefood item (apples, fresh nuts, and dry nuts). Otheranalyseswere runseparately for the twokindsofnuts.

A second set of analyses focused on the individualand behavioral characteristics that affected successin opening the nuts (binary dependent variable) and

TABLE I. Subjects. Name, Sex (F, Female;M,Male), BodyMass, Age (Years,Months, andDays), whenObservationsStarted on August 15th 2012 and 2013

Body mass (kg) Age (years; months; days)

Name Sex 2012 2013 2012 2013

Mansinho M 3.52 3.41 15; 2; 15a 16; 2; 15a

PiaScava (Alpha) F 1.86 1.86 13; 7; 15a 14; 7; 15a

Teninha F 2.02 2.10 12; 7; 15a 13; 7; 15a

Jatob�a (Alpha) M 4.12 4.16 12; 2; 15a 13; 2; 15a

Teimoso M 3.48 3.51 12; 2; 15a 13; 2; 15a

Chuchu F 1.98 2.04 10; 2; 15a 11; 2; 15a

Dita F 2.02 2.12 9; 2; 15a 10; 2; 15a

Tomate M 1.99 2.34 5; 7; 16 6; 7; 16Catu M 2.07 2.47 5; 5; 10 6; 5; 10Cangaceiro M 2.08 2.37 4; 10; 26 5; 10; 26Pati M 2.08 2.50 4; 8; 13 5; 8; 13Doree F 1.56 1.78 4; 8; 6 5; 8; 6PaScoca F 1.44 1.66 3; 7; 13 4; 7; 13Pamonha F 1.32 1.57 3; 7; 13 4; 7; 13Coco M 1.37 1.65 3; 1; 1 4; 1; 1Thais F 1.13 1.34 1; 6; 14 2; 6; 14Chani F 0.99 1.17 1; 6; 14 2; 6; 14Presente M 1.09 1.45 1; 5; 0 2; 5; 0CachaSca M 0.41 1.10 0; 5; 0 1; 5; 0Divina F – 0.97b – 0; 9; 8Titia F – 0.66 – 0; 7; 12Patr�ıcia F – 0.57 – 0; 7; 4Donzela F – 0.40 – 0; 6; 30

For the individuals for which birth dates are unknown, age was estimated. The alpha male and female are indicated.aThe age of these subjects was estimated on the basis of several parameters (EthoCebus, Unpublished data).bThe weight of this subject was estimated on the basis of infants’ growth curve in Fragaszy et al. (in press).

Am. J. Primatol.


the time taken to process and eat them (continuousdependent variable). Independent variables enteredinto analyses were either individual characteristics(sex, age, and body mass) or the presence/absence ofthe two most important behavior patterns used toopen the cashew nuts, rubbing and tool use. Thesubject identity was always entered as a randomeffect to avoid pseudo-replication.

For a third set of analyses, we calculatedindividual hourly rates of contact with the nuts(calculated separately for the 2 years of observationas a measure of the interest in the different kinds ofnuts) and entered them as the dependent variable ina linear regression that included sex and age as theindependent variables and the subject identity as arandom effect.

We also calculated individual counts of episodesof paying attention to others processing nuts andentered them in a negative binomial regression thatincluded sex and age as the independent variables,observation time as the exposure variable, and thesubject identity as a random effect. Finally, we usedwithin-subject conditional negative binomial regres-sions to evaluate if youngsters paidmore attention toother group members processing nuts depending onthe latters’ sex and age (youngster or adult; a proxyfor proficiency), and to evaluate if scratchingoccurred more often during the manipulation ofnuts than of apples. For the former analysis, weestimated the time each individual was available as apotential target of attention on the basis of the

proportion of observation time the individual spentprocessing cashewnuts, and entered this value as theexposure variable.

In the analysis concerning the effects of hammertype on the probability of success, a categoricalvariable with more than two levels (hammer type:siltstone, sandstone, palm nut shell) was includedamong the independent variables. We used a Waldtest to assess its overall significance.

RESULTSSuccess, Behaviors, and Processing Time

For the three types of food items and for eachsubject, Table III reports the number of food itemsprocessed, the percentage of successful episodes, theduration of processing time per item processed, andthe percentages of processing time spent performingeach behavior.

Success was achieved in 99% of the appleepisodes (Table IIIa). Individuals below 1 year ofage accounted for most of the failures; however, eventhem were successful in more than 90% of theirepisodes (Fig. 5a). The behaviors directed toward theapple consisted mainly in eating and sucking (Fig. 2and Supplemental Material S1). Other behaviors(pounding, rolling, rubbing, and tapping/shaking)were only occasionally performed, mostly by young-sters, and many were never observed (extractingwith fingers, processing with hands, processing withteeth, and tool using).

TABLE II. List and Description of the Behaviors Scored

Behavior Description

Foraging To search for food while being in a cashew tree, or on the ground below itBehaviors directed toward the food item

Eating To chew the food target and ingest itLicking/sucking To contact with the tongue or extract liquid by compressing the lipsRubbing To move the food target back and forth (or in one direction only) on a surfaceProcessing with teeth To take an encased item with teeth (usually canines) out of its case (without evident contact

between lips and food target)Processing with hands To exert force on the food target with the hand(s) in order to tear it apartExtracting with fingers To take an encased item with one finger/nail (usually index finger) out of its casePounding To bang the food target. The food is held with one or both hands and pounded one or several

times in succession against a hard surfaceTapping/shaking To strike gently with a light blow or blows. Either tap object with fingers or tap object on a

surface. Or to move to and fro in short, irregular movementsRolling To rotate the food target held between hands by moving them in opposite directions, or to roll the

food target over a surfaceTool using To use a percussor, usually a stone, to access the encased nut

Other behaviorsScratch To scratch hand(s) or to scratch faceAttention To attend to another individual who is feeding on cashew nut or apple. The focal subject looks at

the other from a distance of 1m or lessDrooling Abundant salivation due to food ingestionOther Includes behaviors not listed above (playing, grooming, resting, mating . . .)

Am. J. Primatol.


TABLE

III.

Apples(a),Fre

shNuts

(b),an

dDry

Nuts

(c)

Percentage

ofpr

ocessingtimesp

entpe

rformingea

chbe

hav

ior

Subject

#Item

spr

ocessed

%of

success

Processing

time/item

(sec)

Eating

Licking/su

cking

Rubb

ing

Teeth

processing

Han

dpr

ocess.

Finge

rex

tract.

Pou

nding

Tap

ping/sh

aking

Rolling

Tool

using

(a)App

les

Man

sinho

2210

036

10.7

89.3

00

00

00

00

PiaSca

va32

100

678.5

91.5

00

00

00

00

Jatob� a

4310

039

21.5

78.3

00

00

00

0.3

0Teimoso

2510

041

32.2

67.8

00

00

00

00

Ten

inha

3610

050

8.0

92.0

00

00

00

00

Chuch

u38

100

4620

.579

.50

00

00

00

0Dita

1810

039

14.8

85.2

00

00

00

00

Tom

ate

4410

047

13.8

85.6

00

00

00.6

00

Catu

3410

038

16.5

75.0

00

00

08.4

00

Can

gaceiro

1610

047

14.6

81.1

0.4

00

02.3

01.6

0Pati

1510

069

22.2

77.8

00

00

00

00

Doree

2195

4322

.277

.00

00

00

0.8

00

PaSc

oca

3710

039

14.5

85.5

00

00

00

00

Pam

onha

4710

056

21.0

78.2

0.3

00

00

0.5

00

Coco

6310

037

23.6

76.4

00

00

00

00

Ta� ıs

7099

3911

.386

.10

00

02.3

0.3

00

Chan

i57

100

6313

.986

.10

00

00

00

0Presente

6710

046

14.2

85.0

00

00

00.8

00

CachaSc

a58

9843

12.7

83.2

0.4

00

03.4

0.2

00

Divina

1894

6313

.484

.22.4

00

00

00

0Titia

1493

736.5

80.2

8.9

00

04.5

00

0Patr� ıcia

1292

117

16.6

65.0

17.1

00

00

01.2

0Don

zela

2295

9210

.772

.82.1

00

013

.30

0.9

0

(b)Fresh

nuts

Man

sinho

100

9775

5.3

036

.97.6

050

.10

0.1

00

PiaSca

va66

8884

3.3

038

.89.1

0.2

48.5

00.2

00

Jatob� a

134

9666

7.9

038

.117

.40

36.5

00

00

Teimoso

9297

687.6

030

.124

.80

37.4

00.1

00

Ten

inha

7895

896.5

033

.910

.10.6

48.2

00.7

00

Chuch

u11

691

634.9

035

.423

.70

35.8

00

0.1

0Dita

124

9891

4.8

028

.211

.60.5

54.9

00

00

Tom

ate

8291

835.8

040

.411

.70

41.1

00.7

0.2

0Catu

9594

924.7

033

.322

.70

38.9

00.3

00

Can

gaceiro

4510

095

4.6

032

.521

.10

41.8

00

00

Pati

3586

796.9

0.3

35.9

16.1

040

.60

00.1

0Doree

5686

127

4.8

029

.318

.30

47.5

0.1

00

0PaSc

oca

6289

133

2.7

034

.814

.90.1

47.4

0.1

00

0

Am. J. Primatol.


TABLE

III.

Con

tinued

Percentage

ofpr

ocessingtimesp

entpe

rformingea

chbe

hav

ior

Subject

#Item

spr

ocessed

%of

success

Processing

time/item

(sec)

Eating

Licking/su

cking

Rubb

ing

Teeth

processing

Han

dpr

ocess.

Finge

rex

tract.

Pou

nding

Tap

ping/sh

aking

Rolling

Tool

using

Pam

onha

3985

131

2.0

034

.620

.70.1

40.7

1.9

00

0Coco

3275

111

4.6

045

.612

.10.7

34.8

1.0

01.0

0.2

Ta� ıs

2361

188

1.9

037

.122

.30

36.2

2.5

00

0Chan

i32

4714

81.8

049

.56.0

0.7

37.8

3.5

0.5

00.2

Presente

2665

125

2.6

035

.518

.00.1

42.7

0.4

0.6

00

CachaSc

a17

2410

50.8

2.8

44.3

23.7

016

.28.5

1.0

0.7

2.0

Divina

50

150

090

.60

00

9.4

00

0Titia

120

400

068

.66.3

00

21.9

03.2

0Patr� ıcia

50

220

012

.347

.40

16.0

24.3

00

0Don

zela

60

300

044

.03.0

00

46.8

4.7

1.5

0

(c)Dry

nuts

Man

sinho

3776

7420

01.1

20.8

4.5

32.4

1.3

211

.66.3

PiaSca

va42

7970

7.2

00

7.6

4.3

28.1

01.1

21.9

29.8

Jatob� a

3987

5114

.10

027

.42.8

36.4

0.2

1.3

13.3

4.5

Teimoso

1493

9514

.90

015

.93.5

52.9

02.9

6.3

3.5

Ten

inha

2882

9410

.80

013

.33.2

45.1

2.9

1.7

7.8

15.1

Chuch

u27

9392

9.7

00

23.4

0.1

39.1

1.7

2.0

10.2

13.7

Dita

3291

968.9

00

6.2

1.8

43.0

01.2

22.6

16.3

Tom

ate

1464

100

21.0

03.6

12.9

7.0

28.0

2.1

7.3

10.0

8.1

Catu

1872

125

21.1

00

10.9

6.0

39.8

1.2

2.1

6.2

12.6

Can

gaceiro

3291

8318

.50

07.0

2.5

35.0

4.1

0.9

11.2

20.8

Pati

3583

8110

.00

0.2

15.0

3.2

41.9

2.8

4.6

5.8

16.4

Doree

4869

126

7.0

00.1

8.8

3.1

37.4

4.8

1.8

9.8

27.2

PaSc

oca

3868

148

2.0

00

12.5

2.8

25.4

4.3

1.1

5.1

46.7

Pam

onha

2759

121

5.4

00.6

23.4

2.7

46.2

5.5

6.5

4.9

4.9

Coco

4370

146

6.5

00

8.3

4.2

22.7

6.9

1.1

13.8

36.4

Ta� ıs

1638

655.3

01.3

32.0

4.7

15.9

13.1

8.4

4.3

15Chan

i9

1110

50.2

00

18.4

08.7

20.9

3.7

1.8

46.3

Presente

1533

6811

.20

026

.60

24.1

12.3

17.5

8.3

0CachaSc

a12

3312

72.3

04.9

42.2

13.8

7.9

0.8

0.5

63.5

Divina

10

60

010

00

00

00

00

Titia

0–

–0

00

00

00

00

0Patr� ıcia

20

400

092

.67.4

00

00

00

Don

zela

30

370

05.1

28.5

00

66.5

00

0

For

each

subject,werepo

rtthenumbe

rof

food

item

spr

ocessed,

thepe

rcen

tage

ofsu

ccessfule

pisode

s,thedu

ration

ofpr

ocessingtimepe

ritem

processed(bas

edon

thetimeelap

seduntilsuccesswas

achieve

d,or

untiltheitem

was

discarde

d),a

ndthepe

rcen

tage

sof

processingtimesp

entpe

rformingea

chbe

hav

ior.

Am. J. Primatol.


Success with fresh nuts was achieved in 87% ofthe episodes (Table IIIb) and until 1 year of ageindividuals never succeeded (Fig. 5b). All behaviorswere performed: some very often (e.g., rubbing,processing with teeth, and extracting with fingers;[Fig. 3]; see Supplemental Material S2, S3, S5), andsome only by a few individuals (licking and to a lesserextent rolling), or only by younger individuals(pounding and tool using).

Success with dry nuts was achieved in 70% ofepisodes (Table IIIc) and until 1 year of ageindividuals never succeeded (Fig. 5c). All behaviors,except sucking/licking, were performed. Adultstended to use tools to crack the nut (Fig. 4); seeSupplemental Material S7) and they did not performrubbing. In contrast, until 1 year of age individualsdid not perform tool use, rolling, or tapping.

Overall, capuchins were clearly more successfulwith apples than with fresh or dry nuts (z¼9.54,P< 0.001 and z¼ 11.56, P<0.001, respectively).They were also more successful in opening freshnuts than dry nuts (z¼ 7.76, P<0.001). In successfulepisodes, the total processing time (time needed to

open and consume the nuts) was also longer for freshand dry nuts compared to apples (z¼10.59,P<0.001and z¼ 8.80, P<0.001, respectively), while timeneeded to open and consume fresh and dry nuts didnot differ (z¼ 0.82, P¼ 0.425).

Factors Affecting the Opening of Fresh NutsAs shown in Figure 5b, probability of success

gradually increased from 0 at 1 year of age to 0.85 at5 years, slightly increasing later in life. Among theindividual variables examined, only bodymass was asignificant predictor of probability of success, whilesex and age of the subject were unrelated toprobability of opening a fresh nut (Table IV). Giventhe strong correlation between age and body mass(year 2012: r¼ 0.803, N¼ 19, P<0.001; year 2013:r¼0.810, N¼ 23, P< 0.001), we repeated the analy-sis excluding the latter. Age became a significantpredictor of success (z¼5.56, P<0.001; Fig. 5b).Note that here, and below, we report both analyses inorder to make our results directly comparable withstudies that may have only one of the two variables.

We examined how the use of two behaviorscritical for opening nuts (rubbing and tool use)affected the probability of success. When dealingwith a fresh nut, success was more likely ifthe monkey rubbed it (z¼ 7.12, P< 0.001) than ifthe monkey did not; but success was as likely if themonkey used tools than if the monkey did not(z¼ 0.75, P¼ 0.456) (Fig. 6a).

In successful episodes, the total processing timewas unrelated to body mass and sex, butwas negatively related to age. Older subjects tookless time to open a fresh nut (Table V). When theanalysis was repeated excluding age, the effect ofbody mass was significant (z¼�5.16, P<0.001).

Factors Affecting the Opening of Dry NutsAs shown in Figure 5c, probability of success

gradually increased from 0.1 at 2 years of age to 0.69at 5 years of age; later in life, most of the valuesremained above this level. The notable exception wasthe very low probability of success (¼0.33) at 12 yearsof age; this value is based on only three episodesperformed by Chuchu, in which she opened one nutwith tools, and failed once using her teeth and onceusing a tool. Among the individual variables exam-ined, none significantly predicted the probability ofopening a dry nut (Table VI). When the analysis wasrepeated excluding body mass, age was a significantpredictor of success (z¼4.41, P<0.001; Fig. 5c).

When dealing with a dry nut, success was morelikely if the monkey used tools (z¼2.88, P¼ 0.004)than if the monkey did not; but success was as likelyif the monkey rubbed the nut than if the monkey didnot (z¼�0.89, P¼ 0.140) (Fig. 6b). When capuchinsused a tool, they succeeded in 79% of the episodes

Fig. 2. A young capuchin (S. libidinosus) is sucking the cashewapple with his head backward.

Fig. 3. An adult female is rubbing a fresh cashewnut on the barkof a cashew branch.

Am. J. Primatol.


(187 episodes out of 236); when capuchins did not usea tool, they succeeded in 61% of the episodes (148episodes of the 241).

Table VII reports the number of successful andunsuccessful episodes in which youngsters (of bothsexes), adult females and adult males used tools ornot, to access the kernel of dry nuts. Youngsters usedtools to process dry cashew nuts in 37% of theepisodes and were successful in 72% of them. Adultsused a tool in 54% of the dry nut episodes and weresuccessful in 81% of them. In particular, adultfemales used tools to process dry cashew nuts in78% of the episodes and were successful in 80% ofthem and adult males used tools to process drycashew nuts in 32% of the episodes and weresuccessful in 82% of them. When using techniquesother than tool use, youngsters, adult females, andadult males were successful in 28%, 75%, and 81% ofthe episodes, respectively.

Total processing time of dry nuts was unrelatedto body mass and sex, but was negatively related toage (Table VIII). When the analysis was repeatedexcluding age, the effect of bodymass was significant(z¼�4.37, P< 0.001).

Most tool use episodes occurred when themonkeywas on the ground (68%). However, in 32% of theepisodes, the monkey was on a tree and used a treebranch as an anvil; the branch could belong to thecashew tree from which the nut was taken or toanother nearby tree. In 92 episodes performed by 11individuals, we were able to record detailed informa-tion about the hammers material. The monkeys usedsandstone (N¼31), siltstone (N¼16), and palm nutshell (N¼ 45) and reached success in 87%, 94%, and80% of the episodes, respectively. In 80 episodes, themassof the stonewasmeasured: onaveragesandstonehammersweighed131g (min–max: 37–316g,N¼ 29),siltstone hammers 496g (min–max: 170–877g,N¼ 16), and palm nut shell hammers 30g(min–max: 14–72g, N¼35). Neither the weight northe type of the hammer affected the probability ofopening the nut (hammer weight: z¼0.44, N¼80,P¼0.659; hammer type: x2¼ 0.00, df¼2, P¼0.999).

Chemical Burn, Drooling, and ScratchingDuring the study, we did not observe any blister

on the face or hands of the monkeys, although threeevents of drooling occurred. In 2012, Catu drooledwhile processing a fresh nut and in 2013 he drooled(and scratched) while rubbing a fresh nut; in 2013,Pamonha drooled while processing a dry nut (seeSupplemental Material S8). Scratching was infre-quent: out of the 26 episodes of scratching performedby 11 individuals, 18were directed to the hands and 8to the face. Scratching (directed to a hand or the face)occurred in 6 out of 809 apples episodes, 12 out of1,282 fresh nuts episodes, and 3 out of 477 dry nutsepisodes. Monkeys did not scratch more often whileprocessing a nut (either fresh or dry) than an apple(z¼0.59, N¼22, P¼ 0.555).

Interest Toward the Different FoodContact with a food item is a measure of interest

independently of success. Contact varied in relationto the type of food items. Younger monkeys contactedapples more often than older ones (z¼�2.19, N¼42,P¼ 0.028) regardless of sex (z¼0.23, N¼42,P¼ 0.821). In contrast, older subjects contactedboth fresh nuts and dry nuts more often than

Fig. 5. Probability of success (Mean�SEM) in relation to age for(a) cashew apples, (b) fresh nuts, and (c) dry nuts.

Fig. 4. An adult male uses a stone to crack a dry cashew nut.

Am. J. Primatol.


younger ones (fresh nuts: z¼4.98, N¼42, P<0.001;dry nuts: z¼2.55; N¼42, P¼ 0.011), again regard-less of sex (fresh nuts: z¼1.79, N¼ 42, P¼0.073; drynuts: z¼ 0.65; N¼42, P¼ 0.514).

Attending to Others Processing CashewsCashew trees produce many fruits at the same

time allowing several capuchins to feed concur-rently and (possibly) to attend to what others areprocessing. Notwithstanding this, only three adultsand ten youngsters attended to others processingfresh or dry cashews; this behavior was observed inonly 4.3% of the episodes (76 times out of 1,759 nutepisodes). In a first analysis, we considered all themonkeys as potential targets of attention and werelated the rate of paying attention to othersmanipulating nuts to the subjects’ age and sex.

Youngsters paid more attention to others manipu-lating nuts than did adults (z¼�5.10, N¼42,P< 0.001), while no sex differences emerged(z¼0.43, N¼ 42, P¼ 0.666). In a second analysis,we focused on the target of youngsters’ attention.Youngsters did not direct their attention preferen-tially to adults (z¼ 1.51, N¼346, P¼0.132) andpaid more attention to the behavior of females thanto the behavior of males (z¼1.51, N¼346,P¼ 0.033). We never observed monkeys payingattention to others manipulating apples.

We observed 66 events inwhich subjects collecteda food itempreviously ownedanddiscardedbyagroupmember. In 68% of them, the leftover was an apple, in1% a fresh nut, and in 30% a dry nut. Youngstersaccounted for 91% of the instances of collection andthey collected apples twice as often as nuts.

DISCUSSIONIn FBV, capuchins obtain the nutritious cashew

nuts, that are defended by caustic chemicals, byusing processing techniques that maximize successandminimize the risk of injury. Interestingly, and asexpected, when dealing with a fresh nut, success wasmore likely if the monkey rubbed the nut than if themonkey did not; in contrast, when dealing with a drynut, success was more likely if the monkey used toolsthan if the monkey did not. The use of these differentprocessing techniques also reduces the risk of injury(see below). Thus, we confirm Sirianni andVisalberghi [2013] initial description on how FBVadult capuchins process cashew nuts with a finermethodology.

In addition, we highlight the variation existingacross individuals in processing techniques andsuccess. Although capuchins of all ages are nearlyalways successful in eating the apple, only at 5 yearsof age, they reach a probability of success higher than0.85 and 0.68 of obtaining the kernel of fresh nutsand dry nuts, respectively. While adult individualsuse different techniques for the two food items,juveniles are less discriminating in their actionstoward them. Until 1 year of age, individuals do notperform the same actions as adults do, and theynever succeed in obtaining either the fresh or drynuts.

TABLE IV. Effect of Individual Variables on Proba-bility of Opening a Fresh Cashew Nut

Independent variable Coefficient z-value P-value

Sex �0.11 �0.17 0.862Body mass 1.42 2.46 0.014Age 0.11 1.16 0.246Intercept �1.75 2.10 0.036

Results of a random-effects logistic regression based on 1,282 fresh nuts.

TABLE V. Effect of Individual Variables on the TimeTaken to Open and Consume a Fresh Cashew Nut


Sex �0.17 �1.76 0.078Body mass �0.04 �0.37 0.713Age �0.07 �3.80 <0.001Intercept 5.40 36.58 <0.001

Results of a random-effects linear regression based on 1,121 successfullyopened fresh nuts.

Fig. 6. (a) Fresh nuts. Percent of success when rubbing or tooluse behaviors are performed. (b) Dry nuts. Percent of successwhen rubbing or tool use behaviors are performed.

Am. J. Primatol.


The Acquisition of Proficiency in ProcessingCashew Nuts

Cashew nuts are a challenging food for beardedcapuchin monkeys. Thus, foraging on cashew nutsfollows the ontogenetic pattern described for otherfoods requiring extraction (palm nuts in beardedcapuchins [Eshchar et al., submitted a]; Sloaneaseeds by white-faced capuchins [O’Malley andFedigan, 2005a]), with monkeys progressively mas-tering the skill needed to succeed routinely.

For fresh nuts, the probability of successincreases regularly with age, while for dry nuts,the increase is somehow irregular. Moreover, adultsopen fresh nuts more reliably and with greater inter-individual consistency than dry nuts. This couldreflect (i) our relatively small sample size of dry nuts;(ii) specific characteristics of dry nuts (e.g., resis-tance, dryness, CNSL viscosity) allowing differentstrategies to reach success; (iii) different availabilityof tools near the cashew trees, affecting probability ofsuccess when using tools; and (iv) difference in bitingforce and canine size between adult males andfemales, affecting the success while using teeth[Plavcan and Kay, 1988]. Further studies shouldexamine the role of these factors in a larger dataset.

Proficiency at processing cashew nuts increasesgradually until early adulthood (see Fig. 5b and c).The timeline for dry cashew nuts is similar to thatreported for cracking open resistant palm nuts usingstone hammers. In FBV striking a palm nut with astone, an obligatory action to crack a palm nut, firstappears in the second year or even later, and youngmonkeys do not regularly succeed at cracking nutsuntil their fourth year or later [Eshchar et al.,submitted a]. In cashew processing, ineffectiveactions gradually disappear from youngsters’ reper-toire as effective actions come to predominate.Similarly, youngsters attempting to crack palmnuts perform actions in a less discriminatingmannerthan do adults. In this respect, it is interesting to notethe similar pattern of acquisition despite the factthat cashews are available only for a few weeks,whereas palm nuts are available all year around[Spagnoletti et al., 2012]. Thus, young capuchinsgradually learn to tune a variety of species-typicalforaging actions (such as pounding and extraction toaccess embedded food) and use them to obtain

different food items that require similar processingtechniques. In thisway, learning to crack open a hardnut to overcome the resistance of its shell allows theacquisition of skills (such as good aim, regulation ofthe force, etc.) that are also necessary to crack acashew nut to avoid its caustic chemical.

Factors Affecting Success andProcessing Time

Age and body mass were highly correlated in oursample. However, these factors do not align perfectlybecause, for example, adult females that weigh onaverage 2.1 kg have more experience than juvenilemales thatweighmore than 2.1 kg fromabout 4 yearsof age [Fragaszy et al., 2016]. As expected, bodymasspredicted success at opening fresh nuts. Age pre-dicted duration of processing fresh nuts: olderindividuals processed fresh nuts more quickly. Inaddition, rubbing significantly contributed to successat opening fresh nuts. Also as expected, body massdid not predict success in opening dry nuts, whileduration of processing dry nuts was predicted by age.

Since adult monkeys opened dry nuts in roughlyequal percentages by processing techniques with orwithout tools, tool use appears unnecessary to accessthe kernel of dry cashew nuts. Moreover, successwith tools was reached also with light sandstones,(weighing on average 143 g), or a palm nut shells(weighing on average 32 g) and the average mass ofthese tools is 15–30 times less than that of the stonesused to crack a palm nut [Visalberghi et al., 2009];thus mass and material of the tool do not play acrucial role to crack open cashew nuts, although theydo so to crack open palm nuts at FBV [Visalberghiet al., 2009].

Interestingly, about one-third of the tool episodesoccurred when the monkey was on a tree and she/heused a tree branch as an anvil to crack the dry nut.

TABLE VI. Effect of Individual Variables onProbability of Opening a Dry Cashew Nut


Sex �0.04 �0.08 0.936Body mass 0.967 1.82 0.069Age 0.08 0.96 0.337Intercept �1.67 �2.13 0.033

Results of a random-effects logistic regression based on 477 dry nuts.

TABLE VII. Dry Nuts. Number of Successful andUnsuccessful Episodes in Which Youngsters (of BothSex), Adult Females, and Adult Males Used Tools, orDid Not Use Tools

YoungstersAdultfemales

Adultmales

# of successful episodeswith tool use

37 103 47

# of unsuccessful episodeswith tool use

14 25 10

# of successful episodeswithout tool use

24 27 97

# of unsuccessful episodeswithout tool use

62 9 22

Total # of episodes 137 164 176

When not using tools, capuchins processed the dry nut only with teethand/or hands.

Am. J. Primatol.


This is the first report of habitual percussive tool useon tree branch anvils in the genus Sapajus. In orderto crack hard palm nuts, bearded capuchins in FBVmust use heavy stones that are very costly to useand especially to transport [Massaro et al., 2012],and stable wooden or stone anvils located on theground. The present data demonstrate that whenthe tool is light, capuchins readily transport and useit in a tree, as the Ta€ı chimpanzees do when crackingthe relatively soft coula nuts [Boesch and Boesch-Achermann, 2000].

Males and females processed cashew nutsequally often, and for fresh and dry nuts their rateof success and duration of processing did not differ.However, females used tools to process dry nutsmorethan twice as often as males, despite the sexes beingequally successful both while using tools and whileusing other processing techniques. The findings thatfemales process dry cashew nuts more often thanmales and that females use tools to a greater extentthan males is in sharp contrast with the patternreported for the same population using tools to crackpalm nuts. In fact, Spagnoletti et al. [2011] foundthat females performed fewer episodes of tool use tocrack palm nuts than males and were less successfulthan males at cracking high resistance palm nuts.Nut resistance explains these different patterns:while palm nuts require heavy tools and strength[Fragaszy et al., 2010; Spagnoletti et al., 2011],cashew nuts can be opened by both sexes withrelatively light tools as well as without them.

Therefore, since tool use is not mandatory toaccess the kernel of dry cashew nuts, we argue thatFBV capuchins use tools (i) to avoid contacting thecaustic resin with the lips more than to overcome theresistance of the shell and that (ii) adultmales do it toa lesser extent than females because their jaws andcanines are bigger and allow them to hold and crackthe nuts with their teeth without contacting theresin. These hypotheses could be tested by analyzingvideos of males and females orally processing drycashew nuts.

Social Influences on the Acquisition ofProcessing Skills

Proximity to and interest toward skilled individ-uals by others during foraging are common in many

species of birds and mammals [e.g., meerkats,Hoppitt et al., 2012] and much effort has beendevoted to evaluating the contribution of observingproficient others foraging on the development offoraging skills by the youngsters [Hoppitt andLaland, 2008]. Young capuchin monkeys are usuallyvery interested in adults’ processing activitieswith palm nuts [Coelho et al., 2015; Eshchar et al.,in press; Ottoni et al., 2005] and other difficultfoods, such as Sloanea fruits [O’Malley and Fedigan,2005a,b] and beetle larvae in bamboo canes [Gunstet al., 2008]. Finally, young-white faced capuchins(C. capucinus) show interest toward group membersforaging on food items that are rare, difficult toprocess, or large [Perry and Jim�enez, 2012].

However, despite requiring multi-step process-ing techniques, cashewnuts elicit little interest fromless proficient individuals. Young capuchins seldomattend closely to group members processing cashewnuts and very rarely collect discarded parts of thecashew for further inspection or ingestion. Althoughyoungsters pay more attention to others, particu-larly to females, than do adults, youngsters do notpreferentially observe more proficient individuals,that is, those from which something could be learnt.On the whole, this pattern of results indicates thatyoung capuchins do not pay preferential attention tothose performing behaviors that they themselvescan not yet perform. Finally, it is worthwhilementioning that we never observed adults intervenein any way with youngsters that were handlingcashew nuts.

Youngsters’ little attention to others processingcashews contrastswithwhat they dowhen adults usestone tools to crack palm nuts. Eshchar et al. [inpress] carried out a longitudinal study on stone tooluse acquisition in FBV. The study focused on theforaging behavior of youngsters (individuals up to6 years of age) in relation to others’ foragingbehavior. The data were collected while the groupmoved in its habitat including the “outdoor labora-tory,” an open area where capuchins receive palmnuts for experimental purposes [Visalberghi andFragaszy, 2013]. In Eshchar et al. [in press] study,youngsterswere three timesmore likely to be near ananvil when others were cracking palm nuts than atother times, and spent 3/4 of their time within 5m ofone or more monkeys when nut-cracking was takingplace. Scrounging often occurred: juveniles up to3 years of age frequently obtained bits of nuts or otherdesirable leftovers produced by nut-crackers [see alsoFragaszy et al., 1997]. Similarly, in a semi-free livinggroup of Sapajus spp. that use tools to crack Syagrusromazoffiana palm nuts, juveniles commonly ob-served others cracking nuts and scrounged bits ofnuts from others [Coelho et al., 2015].

Several factors can account for the strikingdifference in the attention devoted to dry cashewnut processing and to palm nut cracking events.

TABLE VIII. Effect of Individual Variables on theTime Taken to Open and Consume a Dry Cashew Nut


Sex �0.03 �0.20 0.840Body mass �0.15 �1.36 0.172Age �0.04 �2.93 0.003Intercept 5.24 33.94 <0.001

Results of a random-effects linear regression based on 335 dry nuts.

Am. J. Primatol.


First, tool use for cracking dry cashew nuts occursboth on trees and on the ground and only rarely is thesame anvil surface used repeatedly; in contrast, tooluse for cracking palm nut occurs over and over againon the same anvils leaving visually salient leftovers[Visalberghi et al., 2013]. Second, one or a few strikesare sufficient to crack the cashew nuts, while manystrikes and on average 2–3min are necessary forpalm nuts [Spagnoletti et al., 2011]. Moreover, palmnuts often shatter into multiple pieces with thebroken shells still containing small bits of kernel andthese leftovers prompt scrounging [Ramos da Silva,2008]. In contrast, cashew nuts are fully consumed.Cashew nuts can be cracked from a seated posturewhereas cracking palm nuts typically involves anerect bipedal posture. Cracking cashew nuts produ-ces less noise than cracking palm nuts, because thestrikes are less forceful. All these factors contributeto making cracking cashews with a tool less visuallysalient than cracking palm nuts with a tool. Finally,retrieving abandoned cracked cashew shells is notprofitable since they still contains the caustic resinbut little to no edible leftovers.

Nevertheless, social influences may still affectthe interest of youngsters toward cashew nuts eventhough they do not closely watch others processingthese nuts. Adults’ feeding on nuts may promptyoungsters to stay close by, and possibly facilitatestheir consumption of this food itemand/or helps themovercome their initial disinterest toward nuts (orpossible aversion, see Section below).

Risk Aversion and Acceptance of Risky FoodDuring our data collection, we rarely observed

signs of discomfort (drooling and scratching) and wenever observed individuals with blisters; however,in 2011, we observed blisters apparently due toCNSL on the lips of a young capuchin in FBV (seeSupplemental Material S9). Therefore, it is possiblethat capuchins that early in life had experiencedthe CNSL, later avoid this food item until theyacquire some of the skills or strength necessary toprocess it. We assumed that contact is an index ofinterest and that its absence indicates disinterest orcaution. As expected, we found that youngsterscontacted the “risky/difficult to process” cashewnuts significantly less than adults do, but bothadults and youngsters show equivalent interest inthe apples. Young capuchins usually mouth or bitefood items that they cannot yet consume, incontrast, the young monkeys in our study did noteven try to process the cashew nuts, or did so to avery limited extent. Thus, we propose that theyavoided them because of previous noxious experi-ence. In other words, young individuals might havebeen risk averse, that is, unwilling to consume afood type that had been associated with negativeconsequences.

This patternmatches that seen in juvenilewhite-faced capuchins, that rarely consume Sloanea seeds,often discard Sloanea fruits before opening themand, when they do it, they frequently rub their facesand sneeze [O’Malley and Fedigan, 2005a]. As doesCNSL, the urticating hairs present in Sloanea fruitsproduce long-lasting irritation to the skin; moreover,CNSL produces drooling. Perhaps, the initial reluc-tance of young capuchin monkeys to process Sloaneafruits and cashew nuts is due to a few events inwhicha young individual experiences the aversive effect ofthe hairs of Sloanea or CNSL of cashews. Thisphenomenon bears a resemblance to the foodaversion documented by Garcia et al. [1955] [Garciaand Koelling, 1966] in rats that associated gastroin-testinal illness with the novel food consumed earlier.

How juveniles overcome their initial reluctanceto handling and eating cashew nuts is not yetunderstood. We suggest that youngsters’ consump-tion of cashew nuts is socially facilitated by adultseating food in the vicinity. In fact, several studiesdemonstrated that eating is socially facilitated byseeing/hearing others eating food [Ferrari et al.,2005] and that social facilitation of eating allowstufted capuchin monkeys (Sapajus spp.) [Addessiet al., 2007; Visalberghi andAddessi, 2000, 2001] andhuman children [Galloway et al., 2005] to overcomecaution toward novel food.

The Different Approach to Cashews Nuts inTwo Populations of Bearded Capuchins

Understanding the benefits and costs of feedingon particular embedded foods requires attention tomany facets of the foraging process, including thestrength and dexterity required to breach the plant’sdefenses and the possible costs of doing so inef-fectively. Cashew nut processing has an obviousadaptive value, but the nutritious kernel requiresprocessing techniques that minimize the risk ofinjury. To achieve this goal human beings use fire[Mohod et al., 2010] or, more rarely, tools [ElisabettaVisalberghi, Personal observation]. The ability ofFBV capuchins to avoid the caustic liquid/resinpresent in the cashew mesocarp, which effectivelydeters most other seed predators, is remarkable.

This study demonstrates that capuchin monkeysat FBV process cashew nuts using different sets ofbehaviors according to the ripeness of the fruit andstate (liquid/resin) of the CNSL present in itsmesocarp. Namely, they access the kernel of freshnuts by rubbing the nut on substrates anduse tools ortheir teeth to open dry nuts. In contrast, thepopulation of bearded capuchins living in SCNP,320km from FBV, uses tools to smash/crack bothfresh and dry nuts [Fal�otico, 2011]. Since blisters onthe lips are apparently common in SCNP [Fal�otico,2011] and rare in FBV (see above), it can be arguedthat the use of different techniques to access fresh

Am. J. Primatol.


and dry nuts reduces the risk of chemical burns forFBV capuchins. Thus, cashew nut processing inbearded capuchin monkeys may join the list ofbehaviors related to feeding and to social interactions[e.g., Panger et al., 2002; Perry, 2011] that aredistinctively different across groups and that developover a period of years in supportive social contexts.Future studies should investigate how differencesacross populations come about, particularly as closeobservation does not seem a likely contributor tolearning how to process cashew nuts for the monkeysat FBV.

ACKNOWLEDGMENTSWe thank Maria ConceiSc~ao Fonseca de Oliveira

and the whole Family M for logistic support andpermission to work at FBV. We are grateful toMarino Junior Fonseca de Oliveira, Marcio Fonsecade Oliveira, and Rone Pires de Oliveira, our excellentfield assistants, and to Valentina Truppa for usefulsuggestions. We are grateful to the two anonymousreferees whose advices were very helpful.

REFERENCESAddessi E, Chiarotti F, Visalberghi E, Anzenberger G. 2007.

Response to novel food and the role of social influences incommon marmosets (Callithrix jacchus) and Goeldi’smonkeys (Callimico goeldii). American Journal of Primatol-ogy 69:1–13.

Agostini I, Visalberghi E. 2005. Social influences on theacquisition of sex-typical foraging patterns by juveniles in agroup of wild tufted capuchin monkeys (Cebus nigritus).American Journal of Primatology 65:335–351.

Boesch CH, Boesch-Achermann H. 2000. The chimpanzees ofthe Ta€ı forest. Oxford: Oxford University Press. p 316.

Canale GR, Guidorizzi CE, Kierulff MCM, Gatto C. 2009. Firstrecord of tool use by wild populations of the yellow-breastedcapuchinmonkey (Cebus xanthosternos) and new records forthe bearded capuchin (Cebus libidinosus). American Jour-nal of Primatology 71:1–7.

Carpenter CR. 1935. Behavior of red spider monkeys inPanama. Journal of Mammalogy 16:171–180.

Chalk J, Wright B, Lucas P, et al. 2016. Age-related variationin the mechanical properties of foods processed by Sapajuslibidinosus. American Journal of Physical Anthropology159:199–209.

Coelho CG, Fal�otico T, Izar P, et al. 2015. Social learningstrategies for nut-cracking by tufted capuchin monkeys(Sapajus spp.). Animal Cognition 18:911–919.

DeMoraesBLC,DaSilva Souto A, SchielN. 2014. Adaptabilityin stone tool use by wild capuchin monkeys (Sapajuslibidinosus). American Journal of Primatology 76:967–977.

Eshchar Y, Izar P, Visalberghi E, Resende B, Fragaszy D.Submitted a. Ontogeny of nut-cracking behavior in wildbearded capuchins (Sapajus libidinosus): Links to growth,weaning and foraging. Animal Behaviour.

Eshchar Y, Izar P, Visalberghi E, Resende B, Fragaszy D. Inpress. When and where to practice: social influence on thedevelopment of nut cracking in bearded capuchins (Sapajuslibidinosus). Animal Cognition.

Fal�oticoT. 2011.Usode ferramentaspormacacos-prego (Sapajuslibidinosus) do Parque Nacional Serra da Capivara � PI.(Doctoral dissertation). Retrieved from http://www.teses.usp.br/teses/disponiveis/47/47132/tde-04112011-171428/

Ferrari PF, Maiolini C, Addessi E, Fogassi L, Visalberghi E.2005. The observation and hearing of eating actionsactivates motor programs related to eating in macaquemonkeys. Behavioural Brain Research 161:95–101.

Ferreira RG, Emidio RA, Jerusalinsky L. 2010. Three stonesfor three seeds: natural occurrence of selective tool use byCapuchinmonkeys (Cebus libidinosus) based on an analysisof theweight of stones found at nut-cracking sites. AmericanJournal of Primatology 72:270–275.

Fragaszy DM. 1986. Time budgets and foraging behavior inwedge-capped capuchins (Cebus olivaceous): age and sexdifferences. In: King F, Taub D, editors. Current perspec-tives in primate social dynamics. New York: Van Nostrand.p 159–174.

Fragaszy D, Izar P, Liu Q, et al. 2016. Insights into life historyof bearded capuchins (Sapajus libidinosus). Longitudinalrecords of bodymass and reproduction. American Journal ofPrimatology. DOI: 10.1002/ajp.22509

FragaszyD, Izar P, Visalberghi E, Ottoni E, OliveiraM. 2004b.Wild capuchin monkeys (Cebus libidinousus) use anvils andstone pounding tools. American Journal of Primatology64:359–366.

FragaszyDM,Adams-Curtis LE. 1997. Developmental changesinmanipulation in tufted capuchins (Cebusapella) frombirththroughtwoyears,andtheirrelationtoforagingandweaning.Journal of Comparative Psychology 111:201–211.

Fragaszy DM, Biro D, Eshchar Y, et al. 2013. The fourthdimension of tool use: temporally enduring artefacts aidprimates learning to use tools. Philosophical Transactions ofthe Royal Society B: Biological Sciences 368:20120410.

Fragaszy DM, Feuerstein J, Mitra D. 1997. Transfers of foodfrom adults to infants in tufted capuchins (Cebus apella).Journal of Comparative Psychology 111:194–200.

Fragaszy DM, Pickering T, Liu Q, et al. 2010. Beardedcapuchin monkeys’ and a human’s efficiency at crackingpalm nuts with stone tools: field experiments. AnimalBehaviour 79:321–332.

Fragaszy DM, Visalberghi E, Fedigan LM. 2004a. Thecomplete capuchin: the biology of the genus Cebus. Cam-bridge: Cambridge University Press. p 339.

Gallina Toschi T, CaboniMF, Penazzi G, Lercker G, Capella P.1993. A study on cashew nut oil composition. Journal of theAmerican Oil Chemists’ Society 70:1017–1020.

Galloway AT, Addessi E, Fragaszy D, Visalberghi E. 2005.Social facilitation of eating familiar food in tufted capuchinmonkeys (Cebus apella): does it involve behavioral coordi-nation? International Journal of Primatology 26:181–189.

Garcia J, Kimeldorf DJ, Koelling RA. 1955. Conditionedaversion to saccharin resulting from exposure to gammaradiation. Science 122:157–158.

Garcia J, Koelling RA. 1966. Relation of cue to consequence inavoidance learning. Psychonom Science 4:123–124.

Glander KE. 1982. The impact of plant secondary compoundson primate feeding behavior. American Journal of PhysicalAnthropology 25:1–18.

GunstN, Boinski S, FragaszyDM. 2008. Acquisition of foragingcompetence in wild brown capuchins (Cebus apella), withspecial reference to conspecifics’ foraging artefacts as anindirect social influence. Behaviour 145:195–229.

Gunst N, Boinski S, Fragaszy DM. 2010a. Development ofskilled detection and extraction of embedded prey by wildbrown capuchin monkeys (Cebus apella apella). Journal ofComparative Psychology 124:194–204.

Gunst N, Leca JB, Boinski S, Fragaszy D. 2010b. The ontogenyof handling hard-to-process food in wild brown capuchins(Cebus apella apella): evidence from foraging on the fruit ofMaximiliana maripa. American Journal of Primatology72:960–973.

Hockings KJ, Sousa C. 2012. Differential utilization ofcashew—a low-conflict crop—by sympatric humans andchimpanzees. Oryx 46:375–381.

Am. J. Primatol.


http://www.teses.usp.br/teses/disponiveis/47/47132/tde-04112011-171428/


Hoppitt W, Laland KN. 2008. Social processes influencinglearning in animals: a review of the evidence. Advances inthe Study of Behavior 38:105–165.

Hoppitt W, Samson J, Laland KN, Thornton A. 2012.Identification of learning mechanisms in a wild meerkatpopulation. PLoS ONE 7:e42044.

Janson CH, van Schaik CP. 1993. Ecological risk aversion injuvenile primates: slow and steady wins the race. In:Pereira ME, Fairbanks LA, editors. Juvenile primates: lifehistory, development and behavior. Chicago: University ofChicago Press. p 57–74.

Lambert JE. 2007. Primate nutritional ecology. In:Campbell CJ, Fuentes A, MacKinnon KC, Panger M,Bearder SK, editors. Primates in perspective. New York:Oxford University Press. p 482–495.

Lee PC, Priston NEC. 2005. Human attitudes to primates:perceptions of pests, conflict and consequences for primateconservation. In: Paterson JD, Wallis J, editors. Commen-salism and conflict: the human-primate interface. Madison:American Society of Primatologists. p 1–23.

Luncz LV, Tiago Fal�otico, Eduardo Ottoni, Michael Haslam.2015. Seasonal variation of food properties influences toolselection in wild capuchin monkeys (Sapajus libidinosus) atSerra da Capivara National Park, Brazil. Folia Primatolog-ica 86:312.

Lynch Alfaro JW, De Sousa e Silva J, Rylands AB. 2012. Howdifferent are robust and gracile capuchin monkeys? Anargument for the use of Sapajus and Cebus. AmericanJournal of Primatology 74:273–286.

Lynch Alfaro JW, De Sousa e Silva J, Rylands AB. 2012. Howdifferent are robust and gracile capuchin monkeys? Anargument for the use of Sapajus and Cebus. AmericanJournal of Primatology 74:273–286.

Mannu M, Ottoni EB. 2009. The enhanced tool-kit of twogroups of wild bearded capuchin monkeys in the Caatinga:tool making, associative use, and secondary tools. AmericanJournal of Primatology 71:242–251.

Martin P, Bateson P. 1993. Measuring behaviour—an intro-ductory guide. 2nd edition. Cambridge: Cambridge Univer-sity Press. p 187.

Massaro L, Liu Q, Visalberghi E, Fragaszy D. 2012. Wildbearded capuchin select hammer tools on the basis of bothstone mass and distance from the anvil. Animal Cognition15:1065–1074.

Melin AD, Young HC, Mosdossy KN, Fedigan LM. 2014.Seasonality, extractive foraging and the evolution ofprimate sensorimotor intelligence. Journal of HumanEvolution 71:77–86.

Mendes FDC, Cardoso RM, Ottoni EB, Izar P, Villar DNA,Marquezan RF. 2015. Diversity of nutcracking tool sitesused bySapajus libidinosus in Brazilian Cerrado. AmericanJournal of Primatology 77:535–546.

Mitchell JD, Mori SA. 1987. The cashew and its relatives(Anacardium: Anacardiaceae). Memoirs of the New YorkBotanical Garden 42:1–76.

Mohod A, Jain S, Powar AG. 2010. Pollution sources andstandards of cashew nut processing. American Journal ofEnvironmental Sciences 6:324–328.

Moura ADA, Lee PC. 2004. Capuchin stone tool use inCaatinga dry forest. Science 306:1909–1909.

O’Malley RC, Fedigan LM. 2005. Variability in food-processingbehavior amongwhite-faced capuchins (Cebus capucinus) inSanta Rosa National Park, Costa Rica. American Journal ofPhysical Anthropology 128:63–73.

O’Malley RC, Fedigan LM. 2005. Evaluating social influenceson food processing behavior in white-faced capuchins (Cebuscapucinus). American Journal of Physical Anthropology127:481–491.

Oliveira PS, Marquis RJ. 2002. The cerrados of Brazil. Ecologyand natural history of a neotropical savanna. New York:Columbia University Press. p 400.

Ottoni EB, Resende BD, Izar P. 2005. Watching the bestnutcrackers: what capuchin monkeys (Cebus apella) knowabout others’ tool-using skills. Animal Cognition24:215–219.

Panger MA, Perry S, Rose L, et al. 2002. Cross-site differencesin foraging behavior of white-faced capuchins (Cebuscapucinus). American Journal of Physical Anthropology119:52–66.

Perry S, Jim�enez JCO. 2012. The effects of food size, rarity, andprocessing complexity on white-faced capuchins’ visualattention to foraging conspecifics. In: Hohmann G,Robbins M, Boesch C, editors. Feeding ecology in apes andother primates. Cambridge: Cambridge University Press. p203–234.

Perry S, Manson JH. 2008. Manipulative monkeys: Thecapuchins of Lomas Barbudal. Cambridge: Harvard Uni-versity Press. p 368.

Perry S. 2011. Social traditions and social learning in capuchinmonkeys (Cebus). Philosophical Transactions of the RoyalSociety B: Biological Sciences 366:988–996.

Peternelli Dos Santos L. 2015. Parametros nutricionais dadieta de duas populaSc~oes de macacos-prego: Sapajuslibidinosus no ec�otono Cerrado/Caatinga e Sapajus nigritusna Mata Atlantica. (Doctoral dissertation). Retrieved fromhttp://www.teses.usp.br/teses/disponiveis/47/47132/tde-10082015-110633/

Plavcan JM, Kay RF. 1988. Sexual dimorphism and dentalvariability in platyrrhine primates. International Journal ofPrimatology 9:169–178.

Ramos da Silva ED. 2008. Escolha de alvos coespec�ıficos naobservaSc~ao do uso de ferramentas pormacacos-prego (Cebuslibidinosus) selvagens. (Master’s dissertation). Retrievedfrom http://www.teses.usp.br/teses/disponiveis/47/47132/tde-12022009-152828/

Rapaport LG, Brown GR. 2008. Social influences on foragingbehavior in young nonhuman primates: learning what,where, and how to eat. Evolutionary Anthropology: Issues,News, and Reviews 17:189–201.

Resende BD, Ottoni EB, Fragaszy DM. 2008. Ontogeny ofmanipulative behavior and nut-cracking in young tuftedcapuchin monkeys (Cebus apella): A perception-actionperspective. Developmental Science 11:828–840.

Resende BD, Nagy-Reis MB, Lacerda FN, Pagnotta M,Savalli C. 2014. Tufted capuchin monkeys (Sapajus sp)learning how to crack nuts: does variability declinethroughout development? Behavioural Processes 109:89–94.

Rosen T, Fordice DB. 1994. Cashew nut dermatitis. SouthernMedical Journal 87:543–546.

Sirianni G, Visalberghi E. 2013. Wild bearded capuchinsprocess cashew nuts without contacting caustic compounds.American Journal of Primatology 75:387–393.

SoutoA, BioneCB,BastosM, et al. 2011. Critically endangeredblonde capuchins fish for termites and use new techniques toaccomplish the task. Biology Letters 7:532–535.

Spagnoletti N, Visalberghi E, Ottoni E, Izar P, Fragaszy D.2011. Stone tool use by adult wild bearded capuchinmonkeys (Cebus libidinosus). Frequency, efficiency, andtool selectivity. Journal of Human Evolution 61:97–107.

Spagnoletti N, Visalberghi E, VerderaneMP, et al. 2012. Stonetool use in wild bearded capuchin monkeys, Cebus libid-inosus. Is it a strategy to overcome food scarcity? AnimalBehaviour 83:1285–1294.

Spinozzi G, Truppa V, Lagan�a T. 2004. Grasping behavior intufted capuchin monkeys (Cebus apella): Grip types andmanual laterality for picking up small food item. AmericanJournal of Physical Anthropology 125:30–41.

Truppa V, Spinozzi G, Lagan T, et al. 2016. Versatile graspingability in power grip actions by tufted capuchin monkeys(Sapajus spp.). American Journal of Physical Anthropology159:63–72.

Am. J. Primatol.






USDA (United States Department of Agriculture). 2015.Composition of foods raw, processed, prepared. USDANational nutrient database for standard reference, Release27. Documentation and user guide. Available from: http://www.usda.gov/wps/portal/us

Visalberghi E. 1994. Learning processes and food preferencesin monkeys. In: Galef BG, Mainardi M, Valsecchi P, editors.Behavioral aspects of feeding. Basic and applied research onmammals. Chur, Switzerland: Harwood Academic. p257–270.

Visalberghi E, Addessi E, Truppa V, et al. 2009. Selection ofeffective stone tools by bearded capuchin monkeys. CurrentBiology 19:213–217.

Visalberghi E, Addessi E. 2000. Seeing group members eatinga familiar food enhances the acceptance of novel foods incapuchin monkeys. Animal Behaviour 60:69–76.

Visalberghi E, Addessi E. 2001. Acceptance of novel foods inCebus apella: do specific social facilitation and visualstimulus enhancement play a role? Animal Behaviour62:567–576.

Visalberghi E, Fragaszy D. 2013. The EthoCebus Project.Stone tool use by wild capuchin monkeys. In: Sanz C, Call J,Boesch C, editors. Tool use in animals: cognition andecology. NewYork: Cambridge University Press. p 203–222.

Visalberghi E, Haslam M, Spagnoletti N, Fragaszy D. 2013.Use of stone hammer tools and anvils by bearded capuchin

monkeys over time and space: dynamic construction of anarcheological record of tool use. Journal of ArcheologicalScience 40:3222–3232.

Visalberghi E, Fragaszy D, Ottoni E, et al. 2007. Character-istics of hammer stones and anvils used by wild beardedcapuchin monkeys (Cebus libidinosus) to crack open palmnuts. American Journal of Physical Anthropology 132:426–444.

Waga IC, Dacier AK, Pinha PS, Tavares MCH. 2006.Spontaneous tool use by wild capuchin monkeys (Cebuslibidinosus) in the cerrado. Folia Primatologica 77:337–344.

Wright BW,Wright KA, Chalk J, et al. 2009. Fallback foragingas a way of life: using dietary toughness to compare thefallback signal among capuchins and implications forinterpreting morphological variation. American Journal ofPhysical Anthropology 140:687–699.

Wright BW. 2005. Craniodental biomechanics and dietarytoughness in the genus Cebus. Journal of Human Evolution48:473–492.

SUPPORTING INFORMATIONAdditional supporting information may be found inthe online version of this article at the publisher’sweb-site.

Am. J. Primatol.


http://www.usda.gov/wps/portal/us

http://www.usda.gov/wps/portal/us

factors affecting cashew processing by wild bearded

Documents