Sex Differences in Tool Use Acquisition in Bonobos (Pan Paniscus)

American Journal of Primatology 75:917–926 (2013)

RESEARCH ARTICLE Sex Differences in Tool Use Acquisition in Bonobos (Pan paniscus)

1 1 2 KLAREE J. BOOSE *, FRANCES J. WHITE , AND AUDRA MEINELT 1Department of Anthropology, University of Oregon, Eugene, Oregon 2Columbus Zoo and Aquarium, Powell, Ohio All the great ape species are known tool users in both the wild and captivity, although there is great variation in ability and behavioral repertoire. Differences in tool use acquisition between chimpanzees and gorillas have been attributed to differing levels of social tolerance as a result of differences in social structure. Chimpanzees also show sex differences in acquisition and both chimpanzees and bonobos demonstrate a female bias in tool use behaviors. Studies of acquisition are limited in the wild and between species comparisons are complicated in captivity by contexts that often do not reflect natural conditions. Here we investigated tool use acquisition in a captive group of naïve bonobos by simulating naturalistic conditions. We constructed an artificial termite mound fashioned after those that occur in the wild and tested individuals within a social group context. We found sex differences in latencies to attempt and to succeed where females attempted to fish, were successful more quickly, and fished more frequently than males. We compared our results to those reported for chimpanzees and gorillas. Males across all three species did not differ in latency to attempt or to succeed. In contrast, bonobo and chimpanzee females succeeded more quickly than did female gorillas. Female bonobos and female chimpanzees did not differ in either latency to attempt or to succeed. We tested the social tolerance hypothesis by investigating the relationship between tool behaviors and number of neighbors present. We also compared these results to those reported for chimpanzees and gorillas and found that bonobos had the fewest numbers of neighbors present. The results of this study do not support the association between number of neighbors and tool behavior reported for chimpanzees. However, bonobos demonstrated a similar sex difference in tool use acquisition, supporting the hypothesis of a female bias in tool use in Pan. Am. J. Primatol. 75:917–926, 2013. © 2013 Wiley Periodicals, Inc.

Key words: great apes; termite ﬁshing; social tolerance

INTRODUCTION ape species [Boesch & Boesch‐Achermann, 2000; Many species of non‐human primates use tools in Galdikas, 1982; Gruber et al., 2010; Goodall, 1986; both captivity [chimpanzee: Kohler, 1927; bonobo: Kano, 1992; McGrew, 1992; Nishida, 1990; Reynolds, Jordan, 1982; orangutan: Galdikas, 1982; sooty 2005; Whiten et al., 2001]. Observations such as these mangabey: Kyes, 1988; capuchin: Visalberghi, 1990; have resulted in the description of and hypotheses gorilla: Fontaine et al., 1995; golden lion tamarin: related to material culture in great apes [Hohmann & Stoinski & Beck, 2001; ring‐tailed lemur and brown Fruth, 2003; McGrew, 1992; van Schaik et al., 2003; lemur: Santos et al., 2005; gibbon: Cunningham Whiten et al., 2001]. et al., 2006; vervet: Santos et al., 2006; mandrill: Addressing these differences, van Schaik et al. Pansini & de Ruiter, 2011] and the wild [chimpanzee: [1999] proposed four conditions that favor the van Lawick‐Goodall, 1970; red colobus: Struhsaker, evolution of material culture: (1) ecological opportu- 1975; bonobo: Kano, 1982; orangutan: van Schaik nity, (2) motor dexterity (3), the cognitive ability to et al., 1996; capuchin: Phillips, 1998, Boinski et al., 2000; gorilla: Breuer et al., 2005; spider monkey: Contract grant sponsor: Office of the Vice President of Research Rodrigues & Lindshield, 2007; long‐tailed macaque: at the University of Oregon. Gumert et al., 2011]. Tool use by great apes is of Correspondence to: Klaree Boose, Department of Anthropology, 1218 University of Oregon, Eugene, OR 97403. particular interest for modeling and identifying E‐mail: [email protected] potential conditions that contributed to the evolution Received 20 November 2012; revised 2 February 2013; revision of hominin tool use [Byrne, 2004; McGrew, 1992; van accepted 1 March 2013 Schaik et al., 1999]. Research on this topic has DOI: 10.1002/ajp.22155 revealed a great deal of behavioral variation in tool Published online 19 April 2013 in Wiley Online Library use behaviors both within and between the different (wileyonlinelibrary.com).

solve problems, and (4) a high degree of social uals are routinely separated and reunited and must tolerance as a means to facilitate the social learning be highly socially tolerant to actively reform social of such complex behaviors. In the captive setting, bonds during fusion events. numerous tool use experiments and anecdotal In addition to the observed differences in tool evidence have demonstrated that all four species of repertoires between field sites, chimpanzees display great apes posses the motor dexterity and cognitive significant sex differences in complex foraging behav- ability to manufacture and use tools, even though ior. Females use tools to fish for termites more there is great variation in ability and behavioral frequently and for longer time periods than do males repertoires (gorillas: artificial fishing [Boysen [McGrew, 1979]. Females are also more efficient at et al., 1999]; orangutans: digging sticks and rain using stone tools to crack open nuts, whereas males covers [Galdikas, 1982]; chimpanzees and bonobos: hunt for prey and consume meat more frequently artificial fishing, aimed throwing, play, etc. [Gruber [Boesch & Boesch, 1984], although female chimpan- et al., 2010]). Even though chimpanzees and bonobos zees at Fongoli hunt with tools more frequently than differ in their abilities to solve physical and social males [Pruetz & Bertolani, 2007]. A study of tool use cognitive problems [Herrmann et al., 2010], recent acquisition in wild chimpanzees at Gombe found that studies also indicate that all of the great apes have a young females spent more time watching their similar understanding of the functional properties of mothers fish for termites, began fishing at an earlier tools [Herrmann et al., 2008]. age, and used tools more proficiently than did young Although ongoing research is continuing to males [Lonsdorf, 2005]. Although evidence of tool use reveal details and characteristics of tool use in all from bonobo field sites is sparse [Hohmann & of the great ape species, chimpanzees remain the Fruth, 2003; Ingmanson, 1996; Kano, 1982], a recent most prolific and diverse non‐human tool users (e.g.: survey of several captive groups found significant sex Bentley‐Condit & Smith [2010]), showing the highest differences in tool use behaviors where females used degree of behavioral variation and complexity tools more often and showed greater diversity in types [Boesch & Boesch‐Achermann, 2000; Goodall, 1986; of tools used [Gruber et al., 2010]. This raises the McGrew, 1992; Nishida, 1990; Reynolds, 2005; question of whether there are sex differences in tool Whiten et al., 2001]. This behavioral variation in use acquisition in bonobos and whether sex differ- tool use across chimpanzee study sites is thought to ences in tool behavior are similar or different across be the result of distinct cultural differences between the ape species. groups [McGrew, 1992], suggesting that social Differences in experimental methodology can learning, facilitated by having a high degree of social make direct comparisons among studies and species tolerance, plays a critical role in acquiring the tool use difficult. This study, therefore, examines the acquisi- behaviors present in a group [van Schaik et al., 1999]. tion of tool use behavior in bonobos under a Recent research focusing on differences in tool use naturalistic setting using a protocol modeled on the behavior between chimpanzees and gorillas supports experiment comparing tool use acquisition in chim- the hypothesis that social tolerance may play an panzees and gorillas at LPZ [Lonsdorf et al., 2009]. important role during acquisition [Lonsdorf et al., We presented a captive group of bonobos at the 2009]. A comparative study of tool use acquisition in Columbus Zoo and Aquarium (CZA) with an artificial these two species at the Lincoln Park Zoo (LPZ) termite mound and recorded latency to attempt and measured successful retrieval of bait from an artifi- to successfully use tools to extract bait. We first cial termite mound together with the number of predict that, similar to other groups of captive neighbors present during tool use, as a measure of bonobos [Gruber et al., 2010], this group of bonobos social tolerance [Lonsdorf et al., 2009]. They found will readily use tools. Chimpanzees that fish for that chimpanzees used tools more quickly and had a termites are selective in the types of tool material greater percentage of possible neighbors present used as well as the manner in which the raw material than did the gorillas. Later research on chimpanzee is modified [McGrew et al., 1979]. Chimpanzees often tool use acquisition demonstrated that the transmis- choose woody items in the form of tree and bush sion of novel tool use behaviors (the retrieval of bait branches that are modified through the employment from an apparatus called the “Panpipe” at the Yerkes of several behaviors specific to tool‐making, such as Primate Center) was facilitated by the high social detachment of raw material, side‐branch removal, tolerance exhibited by the two model individuals, leave stripping, and bark peeling [McGrew et al., each of whom were trained to use a different retrieval 1979]. Lonsdorf et al. [2009] describe a process of technique [Horner, 2010]. Higher social tolerance in allowing the chimpanzees and gorillas to select tool the chimpanzees at LPZ, therefore, may have material from the naturally occurring vegetation facilitated a faster transmission of tool use behaviors within the enclosure at LPZ. The bonobos at CZA than in the gorillas [Lonsdorf et al., 2009]. Lonsdorf were also allowed to select tool material that grew et al. [2009] hypothesized that chimpanzees are within their enclosure. The task of selecting tool especially primed for social learning because of the material from natural vegetation, versus being nature of their fission–fusion ranging, where individ- provisioned with tools by the keepers, provided a

Am. J. Primatol. Bonobo Sex Differences in Tool Use / 919

novel task for all of the bonobos at CZA. We, proved by the University of Oregon Institutional therefore, predict that the bonobos at CZA will select Animal Care and Use Committee (IACUC). The tool material and modify that material in a manner authors have no conflict of interest to declare. similar to what has been described for chimpanzees in both the wild [McGrew et al., 1979] and captive settings [Lonsdorf et al., 2009]. Following Gruber Subjects and Housing et al. [2010], we predict the bonobos will show similar All data were collected on the captive group of sex differences in tool use behaviors with females bonobos housed at CZA. Following Kano’s [1992] age learning to use tools more quickly and using tools class definitions (infant: 0–1; juvenile: 2–6; adoles- more frequently than males. Differences in social cent: 7–14; adult: 15), the group was composed of tolerance between species have been suggested to four adult males, four adult females, two adolescent explain why a fission–fusion species like chimpan- females, two adolescent males, one juvenile male, one zees use tools more readily than gorillas under juvenile female, one infant male, and one infant identical conditions [Lonsdorf et al., 2009]. Since female (Table I). The bonobos were housed in a bonobos are also a fission–fusion species [Kano, 1992; complex of areas consisting of two large indoor public White, 1996] and are, therefore, presumably more viewing exhibits (54.8 m2 each) with multiple climb- socially tolerant than gorillas, we predict that ing structures, two off‐exhibit indoor enclosures bonobos will be more similar to what has been (22.6 m2 each), two off‐exhibit outdoor enclosures reported for chimpanzees than for gorillas [Lonsdorf (18.5 m2 each) and a large naturalistic outdoor public et al., 2009] in both number of neighbors present and viewing exhibit (57.9 m 45.7 m, 2,647.7 m2) with tool use behaviors. grass, mature trees, and an artificial stream and waterfall. The keepers at CZA managed the bonobos fi – METHODS to simulate the species typical ssion fusion process of variable party composition. In the morning, Ethical Note bonobos were allowed access to each other and This study was conducted with approval at CZA, parties were set based mostly on individual bonobo an Association of Zoos and Aquariums (AZA) accred- association preferences rather than parties that were ited institution in Columbus, Ohio, USA. CZA predetermined by the keepers. Most individuals had adheres to the welfare and husbandry standards equal access to each other, but two of the adult males outlined by the AZA. All data were collected using (Jimmy and Donnie) were never allowed to be observations of spontaneous behavior. No animals together because of previous conflicts that had were separated from the group and no social groups resulted in serious injury. This management process were manipulated for the purpose of this study. All usually resulted in three parties that lasted for data collection methods adhered to the American 2–3 days, and rarely changed on a daily basis or Society of Primatologists Principles for the Ethical exceeded 4 days. Each sub‐group occupied different Treatment of Nonhuman Primates and were ap- public display areas.

TABLE I. Subjects (CZA), Number of Trials, and Latencies to Investigate, Attempt, and Succeed

Latencies

Subject Sex Age (years) Offspring No. of trials Investigate Attemptb Succeedb

Lola F 7 19 1 1 1 Gilda F 5 20 3 1 1 Unga F 18 Gander, Jerry 12 4 1 1 Ana Neema F 19 Bila Isia, Gilda, Wilbur 19 2 2 2 Donnie M 18 Jerry 14 1 3 3 JoT F 9 17 1 4 4 Maiko M 27 23 — 99 Gander M 8 15 1 10 10 Jerry M 3 12 1 12 12 Bila Isia M 10 20 2 11 20 Toby M 32a Lola 22 16 1 — Mary‐Rose F 1.5 16 2 14 — Susie F 29a Donnie, Lola, Mary‐Rose 16 1 —— Lady F 29a JoT 16 1 —— Wilbur M 0.5 19 5 —— Jimmy M 32a Donnie 21 ———

Note: aEstimated age for wild‐caught individuals; blatency to attempt and to succeed after ﬁrst investigation; italicized individuals are under 5 years old and were excluded from data analyses.

Am. J. Primatol. 920 / Boose et al.

During the study period, the bonobos were released into the outdoor exhibit each morning at approximately 0730 hr and were brought back into the indoor enclosures at approximately 1900 hr. The bonobos were rarely brought inside between these hours, except for extenuating circum- stances (i.e. dangerous inclement weather or an emergency). Each night the party that occupied the outdoor exhibit was brought into one of the indoor enclosures and the outdoor exhibit remained empty until the following morning. The bonobos were fed each morning and evening at approximately 0730 and 1900 hr. Additional supplemental enrichment feedings were sometimes given throughout the day. An artificial termite mound was installed in the outdoor enclosure near Viewing Area 1 (Fig. 1) for the purpose of this study and provided a novel structure and tool related task for this group at CZA. From the known histories of the study group (Table I), all but two individuals (Ana‐Neema and Unga) were Fig. 2. Juvenile female bonobo using a tool to fish bait from an fi artificial termite mound at the Columbus Zoo (Photo: Drew naïve to an arti cial extractive foraging structure. Enigk). Ana‐Neema had previously resided at the Milwaukee Zoo in Milwaukee, WI where she was exposed to an artificial structure and had learned to successfully Testing Apparatus and Bait fish for bait using tools provided by the staff as enrichment. Finding a tool from the naturalistic An artificial termite mound was constructed from habitat and modifying it for use on this new device concrete, rebar, wire mesh, and paint (Fig. 2). The provided this subject with a novel task for the purpose mound was constructed by the CZA staff and was of this study. Unga previously resided at the fashioned to resemble those naturally occurring in Planckendael Zoo in Belgium where she was pre- the wild. Eight bait holes were bored into the mound sented with several tool use opportunities, including and fitted with removable polyvinyl chloride (PVC) those involving extractive foraging using devices, but tubes that measured 15.0 cm in length and 4.5 cm in not an artificial termite mound. diameter. The mound had a small locked access door

Fig. 1. Composite maps of the a) bonobo enclosures at CZA and b) gorilla and chimpanzee enclosures at LPZ. Maps not to scale. (Original ﬁgures courtesy of the Columbus Zoo and Aquarium and the Lincoln Park Zoo.)

Am. J. Primatol. Bonobo Sex Differences in Tool Use / 921

to allow the staff to remove and replace the bait tubes. enclosures. We recorded latency to investigate the The unbaited mound was installed in the outdoor mound as the number of baited trails until first habitat in front of a large public display window investigation. Latency to attempt and latency to during November of 2010. The subjects were free to succeed were then recorded as the number of trials explore the unbaited mound both before and after after this initial investigation before they attempted June 29, 2011, the first day of baited trials. All the task or successfully fished. subjects were considered habituated (i.e. did not A tool use bout was defined as the time during display avoidance behavior and readily approached which an individual continuously fished or attempted the apparatus) to the mound before trials began. to extract bait. When an individual left 1 m proximity Beginning June 29, 2011, the mound was baited each of the mound, or stopped tool use behaviors for more morning by CZA staff during their regular morning than 30 sec but remained within 1 m proximity of check of the enclosure. The baited tubes remained in the mound, the bout was scored as finished. To the mound throughout the day until the following compare the time spent fishing between individuals, morning when they were replaced by new tubes with we first summed the duration of each successful fresh bait. Bait included honey and/or pureed blue- fishing bout for each individual and then divided this berries, raspberries, bananas, and cooked sweet total time by the number of trials for that individual potatoes and carrots. All subjects readily consumed to control for variation in number of trials (Table I). these food items as part of their regular diet. No tools To compare rate of fishing between individuals, we or raw material were explicitly provided to the first calculated the total number of bouts and then animals, instead, individuals were allowed to con- divided this number by the total number of trials for struct tools out of the naturally occurring vegetation that individual to again control for variation in growing in the enclosure (i.e. tree branches, grass, number of trials. bushes, etc.). Each group member within 1 m of the mound was scored as a “neighbor” [Lonsdorf et al., 2009] for each focal individual on each day during each bout. The fi Data Collection and De nitions number of neighbors was recorded for each individual Data were collected most days between June 29, that was within 1 m of the mound, regardless of 2011 and August 29, 2011. Data collection began whether or not that individual, or their “neighbors,” when the first individual was released into the were engaged in mound related behaviors. Group outdoor enclosure in the morning and ended when composition in the outdoor enclosure at CZA was the last individual was shut out of the enclosure for variable (see “Subjects and Housing Section”) and, the night. Each day that a subject had exposure to the following Lonsdorf et al. [2009], we controlled for baited mound was scored as one trial for that subject. differences in group sizes by calculating the number of Individuals within 1 m of the mound were scored as neighbors as a percentage based on presence data being within proximity of the mound. All presence (number of neighbors/number of group members in the data and behaviors within 1 m of the mound were enclosure 1). The percentage of neighbors was videotaped and coded by K.J.B. Continuous data calculated for all individuals for all approaches to were collected from the videotapes and included within 1 m proximity of the mound including fishing frequencies and durations of all predetermined behavior, attempts to fish, and non‐fishing approaches. mound‐related behaviors (i.e. investigate, poke, tool behavior, and fish as defined by Lonsdorf et al. [2009]) for each subject. Attempting the task was defined as Data Analyses inserting a tool into the bait tubes without accessing Frequency data were compared using G tests of the bait and successful fishing was defined as Goodness of Fit with Williams correction applied inserting a tool into the bait tubes and retrieving [Sokal & Rohlf, 2012] and ANOVAs were used to test bait [Lonsdorf et al., 2009]. Lonsdorf et al. [2009] for significant differences in means [Sokal & defined latency as the number of trials with the Rohlf, 2012]. All data were tested for fittothe baited mound experienced by each individual before assumptions of ANOVAs, including normality, and they attempted the task or successfully fished. no transformations were required. Data on individual The bonobos at CZA had access to the unbaited chimpanzee and gorilla latencies to attempt and to mound for approximately 6 months before the first succeed published in Table I of Lonsdorf et al. [2009] baiting and were housed outside of the visual and were used for comparison with the bonobo data auditory range of the mound during baiting. Because collected on mean latencies to attempt and to succeed the mound was baited out of sight of the bonobos, no for the purpose of this study. In addition, we obtained subjects had visual information that the mound was N’s, means, and standard errors for percentage of baited until they approached and investigated the possible neighbors for chimpanzees and gorillas structure. The mound was not baited unless observ- during baited trials using Figure 3 from Lonsdorf ers were present and the baited mound was observed et al. [2009]. We made the assumption that, although for the entire time that bonobos were in the the enclosures at CZA and LPZ differed in size, each

Am. J. Primatol. 922 / Boose et al.

was 5.67 SE 2.121 trials (range ¼ 1–20, N ¼ 9, individuals <5 years excluded: N ¼ 1) trials. After 60 days, 62.5% (N ¼ 10, including one individual <5 years) of the subjects had successfully performed fishing behavior (Table I). Of these subjects, 50.0% (N ¼ 5) successfully modified raw materials into tools and 50.0% (N ¼ 5) only used material that had been modified by another individual or material that did not require modification to be successfully used as a tool (i.e. straight piece of woody material with no side branches or leaves, etc.). A total of 901 tool use bouts were recorded, of which 99.7% (N ¼ 898) included tools made from woody material (tree and bush branches). Other tools used included grass, leaves, and wood–wool nesting material.

Fig. 3. Mean number of trials in latency to attempt and to succeed for male and female bonobos. Significant differences Sex Differences in Bonobos (P < 0.05). Attempt means for five males (N ¼ 34 total trials) and five females (N ¼ 9 total trials). Succeed means for four males There was a significant sex difference in mean (N ¼ 42 total trials) and five females (N ¼ 9 total trials). latency to attempt to use tools to fish in this bonobo group. Females attempted the task after fewer trials of the groups had equal access to their respective (mean ¼ 1.80 SE 0.583 trials, N ¼ 5, individuals artificial termite mounds. Where there were signifi- <5 years excluded: N ¼ 1) than did the males cant differences among the three species, we used a (mean ¼ 6.80 SE 2.010 trials, N ¼ 5, individuals priori orthogonal multiple comparisons [Sokal & <5 years excluded: N ¼ 1, ANOVA: F ¼ 5.71, df Rohlf, 2012] to first compare gorillas to the Pan genus ¼ 1,8, P < 0.05, Fig. 3). There was also a significant (both chimpanzees and bonobos) as a single unit and sex difference in mean latency to succeed in this then to compare bonobos to chimpanzees. Following bonobo group, where females succeeded in fishing the data analysis methodology in Lonsdorf et al. after significantly fewer trials (mean ¼ 1.80 SE [2009], all individuals from all three species that were 0.583 trials, N ¼ 5, individuals <5 years excluded: under 5 years at the time of data collection were not N ¼ 1) than did the males (mean ¼ 10.50 SE included in the analysis and results. Analyses were 3.524 trials, N ¼ 4, ANOVA: F ¼ 7.56, df ¼ 1,7, run using SAS© and BIOMstat (version 3.3). P < 0.05, individuals <5 years excluded: N ¼ 1). There was no significant difference in the number trials experienced by males and females (F ¼ 0.66, ¼ ¼ RESULTS df 1,14, P 0.4301, Table I). Females engaged in significantly more fishing bouts than did the males Group Results (488 female bouts, 74 male bouts, Goodness of Fit: The first individual to both attempt and success- G ¼ 325.737, P < 0.001). There was no significant fully fish was a naïve juvenile female, Lola, who difference in the mean duration of bouts per trial investigated the mound, manufactured a tool, between females (mean ¼ 408.6 SE 159.7 sec/tri- attempted, and successfully fished on the first baited al) and males (mean ¼ 223.2 SE 194.3 sec/trial, trial presented to the bonobos. Group mean latency to ANOVA: F ¼ 0.55, df ¼ 1,11, P ¼ 0.472). investigate the mound was 3.00 SE 1.335 trials. Seven individuals (four females and three males) investigated the mound on their first trial, and seven Sex Differences Across Species individuals (four females and three males) investi- The comparison of results for males and females gated the mound after their first trial (range ¼ 2–16 among the three ape species (excluding individuals trials, mean ¼ 4.86 trials, Table I). Two adult males, under 5 years old) demonstrated that there were no Jimmy and Maiko, never investigated the mound significant differences among the species in mean during the study period and although Maiko success- latency to attempt or to succeed in males (attempt— fully learned to fish for bait, he was never observed to ANOVA: F ¼ 0.64, df ¼ 2,5, P ¼ 0.5636; succeed— investigate the mound. Group mean latency to ANOVA: F ¼ 0.73, df ¼ 2,4, P ¼ 0.5364) or in latency attempt after investigation was 4.300 SE 1.291 to attempt in females (ANOVA: F ¼ 3.61, df ¼ 2,12, trials (range ¼ 1–11, N ¼ 10, individuals <5 years P ¼ 0.0593, Table II). There were significant differ- excluded: N ¼ 2), although some individuals never ences among the species in latency to succeed in attempted the task (N ¼ 3, one male and two females (ANOVA: F ¼ 12.34, df ¼ 2,12, P < 0.005). females, individuals <5 years excluded: N ¼ 1). The multiple comparisons within this statistically Group mean latency to succeed after investigation significant ANOVA show that there was no

Am. J. Primatol. Bonobo Sex Differences in Tool Use / 923

TABLE II. Latencies to Attempt and to Succeed for manufacture and use appropriate tools. Although Males and Females in Chimpanzeesa, Gorillasa, and recent studies also indicate that all of the great apes Bonobos have a similar understanding of the functional properties of tools [Herrmann et al., 2008; but see Males Females Herrmann et al., 2010], other studies have suggested Latency to Mean SE N Mean SE N that species as well as sexes may show differing levels of propensity or aptitude in constructing and using Attempt tools [Gruber et al., 2010; Lonsdorf et al., 2004; Chimpanzee 5.00 4.00 2 1.00 0.00 4 Lonsdorf et al., 2009; McGrew, 1992; Reynolds, 2005; Gorilla 1.00 0.00 1 12.33 4.85 6 Whiten et al., 2001]. We found that, comparable to the Bonobo 6.80 2.01 5 1.80 0.58 5 chimpanzees at LPZ, several individuals in this Succeed group of bonobos rapidly became proficient in Chimpanzee 8.00 5.00 2 1.00 0.00 4 manufacturing tools and extracting bait from the Gorilla 1.00 0.00 1 22.83 5.15 6 artificial termite mound. Tools were mostly con- Bonobo 10.50 3.52 4 1.80 0.58 5 structed out of woody material harvested primarily aNote: Data for chimpanzees and gorillas from Lonsdorf et al. [2009]. from the nearby trees and bushes within the naturalistic outdoor enclosure. Tool construction was remarkably similar to what has been described significant difference within Pan, between female for termite fishing in wild chimpanzees [McGrew bonobos and female chimpanzees, in latency to et al., 1979], including detachment of raw material, succeed (F ¼ 0.02, df ¼ 1, P ¼ 0.8864), but there side branch removal, leaf stripping, and bark peeling. was a significant difference between female gorillas Both the individual that constructed the tool as well and female Pan in latency to succeed (F ¼ 24.63, as other members of the group often reused tools that df ¼ 1, P < 0.001). were constructed out of woody branch‐type material. On several occasions, individuals were observed carrying tools away from the mound and returning Numbers of Neighbors some time later with the same tool to begin another We recorded the number of neighbors for 1,192 fishing bout. observations of 15 focal animals engaged in all Following Lonsdorf et al. [2009], “investigation” mound‐related behaviors. We tested for differences of the mound was defined as using visual and in the number of neighbors during mound related olfactory senses to examine the contents of the bait behaviors by focal bonobos in this group using a holes in order to identify when subjects were first Nested ANOVA. There were significant differences in clearly aware that the termite mound contained bait. number of neighbors between individuals (F ¼ 10.61, This definition appeared appropriate for most indi- df ¼ 13,1177, P < 0.0001) but there was no signifi- viduals, except for one. Maiko, a low‐ranking male, cant sex difference in number of neighbors (F ¼ 0.21, did not investigate the mound under this definition, df ¼ 1,13, P ¼ 0.6521). There was, however, a but was in the vicinity of fishing individuals on significant difference in mean number of neighbors several occasions prior to his first attempt. In this present between bonobos (CZA), chimpanzees (LPZ), single case, it appears that Maiko used information and gorillas (LPZ) (ANOVA: F ¼ 25.793, df ¼ 2,26, from observing the behaviors of other group members P < 0.001, Table III). Bonobos had the fewest mean to identify that the mound contained bait. Also percentage of possible neighbors present at the following Lonsdorf et al. [2009], “attempt” was mound. defined as non‐successful use of tools, however, all of the bonobos except Maiko frequently poked at the holes in the termite mound with their fingers before DISCUSSION their first attempt to use tools to extract bait. It is The results presented in this study confirm possible, therefore, that Maiko learned to successful- previous studies and our prediction that bonobos ly fish by observing other group members from a posses the motor dexterity and cognitive ability to distance rather than through a standard method of trial‐and‐error that involved very close proximity to fi TABLE III. Mean Percentage of Number of Neighbors group members that were shing and frequently for Chimpanzeesa, Gorillasa, and Bonobos poking. It is not surprising that, because the bonobos were tested within social groups, a low‐ranking male Mean (%) SE N would not approach the mound while multiple individuals were engaged in mound related behaviors Chimpanzees 75.81 5.58 6 surrounding a high‐value food source. Maiko’s Gorillas 36.28 6.05 10 behavior highlights the importance of conducting Bonobos 21.71 3.44 13 research within the social context in order to observe aNote: Data for chimpanzees and gorillas from Lonsdorf et al. [2009]. potential variation in behavioral strategies.

Am. J. Primatol. 924 / Boose et al.

As has been observed in chimpanzees more frequently than did the gorillas. In addition, the [Lonsdorf, 2005; McGrew, 1979], there were similar chimpanzees were faster to use tools to successfully significant sex differences in tool use behaviors in this retrieve bait and had a greater percentage of possible group of bonobos. We found that the female bonobos neighbors present than did the gorillas. Lonsdorf were much quicker to attempt to use tools and et al. [2009] suggested that social tolerance, as successfully fished more quickly than did the males. measured by number of neighbors present, might This result is similar to the description of the impact the social learning necessary for successful acquisition of termite fishing in chimpanzees tool use acquisition. The authors concluded that [Lonsdorf, 2005]. Although Lonsdorf [2005] observed differences in tool use behavior among species may, learning in young chimpanzees, whereas we mea- therefore, be related to differences in social structure sured tool acquisition in bonobos ranging in age from as party or group sizes, stability, and social dynamics 6 months to 32 years, the result that mean time to can differ significantly both within and between acquisition in females was significantly shorter than species. We tested this hypothesis with the inclusion in males is similar to what we observed in the of number of neighbors data from the CZA bonobos bonobos. In addition, once successful, the female during their first 60 days of exposure to the baited bonobos returned to the artificial termite mound and termite mound. Bonobos also exhibit a fission‐fusion fished significantly more frequently than did the social system and previous studies have demonstrat- male bonobos. Together our results support previous ed that they have a high degree of social tolerance in research demonstrating a female bias in tool use in the form of lower aggression, higher cohesion, and Pan [Gruber et al., 2010]. greater cofeeding during food sharing opportunities The observed sex differences in this group of compared to chimpanzees [Hare et al. 2007; Kano, bonobos highlight important variations in tool use 1992; White, 1996]. Because we found that bonobo aptitude and propensity across the ape species and females, like chimpanzee females, were faster than between the sexes. When comparing our results with gorilla females to succeed in this tool use experiment, those published by LPZ, we found that there was no we expected that the mean percentage of number of significant difference among the males of the three neighbors in bonobos to be similar to the number species in latency to attempt or to succeed. Although reported in chimpanzees. Instead, we found that the caution must be applied when considering small bonobos had the lowest mean percentage of number of sample sizes, the males in these groups of chimpan- neighbors than either chimpanzees or gorillas after zees, bonobos, and gorillas demonstrated similar controlling for differences in group and party sizes. aptitudes in tool use acquisition. Further research is Our data, therefore, do not support the hypothesis needed to determine if these results accurately reflect that this measure of social tolerance is a facilitator of aptitude among the males and the degree to which the transmission of tool use behaviors in bonobos. propensity to use tools affects successful acquisition However, it is possible that differences in the social in males. There were also important similarities and management of the groups, at CZA fission–fusion is some differences among the females of the three simulated whereas at LPZ all the chimpanzee and species. Females from all three species had similar gorilla groups are stable, may have impacted the mean latencies to attempt to use tools to extract bait results. suggesting that female chimpanzees, bonobos, and It is interesting that, given the sex differences in gorillas share a similar propensity to use tools. the bonobos in latency to attempt and to succeed, we However, while chimpanzee and bonobo females found no significant difference between the sexes in also had similar latencies to successfully use tools, social tolerance at the mound as measured by the the two Pan species were significantly quicker to number of neighbors present. There were, however, succeed than were the gorilla females. Our results significant individual differences indicating that demonstrate similarity between chimpanzees and some males and females were more social than bonobos in these attributes of tool use acquisition, others. For example, the adult male that fished the and verify a key difference in tool use aptitude most frequently and for the longest duration per bout between the Pan genus and gorillas. (Maiko) rarely fished at the mound when any other Differences in ecological constraints have been individuals were present, exhibiting a preference to suggested to explain why chimpanzees and gorillas fish alone. Maiko also seemingly learned to fish by differ in their tool using abilities [Breuer et al., 2005]. observing others from a distance. One female (Unga), Chimpanzees use tools to extract resources that are in contrast, rarely fished alone, and when approach- otherwise difficult to attain, whereas gorillas are ing the mound, would often loud call to out‐of‐sight more able to use physical strength to acquire hard to members of her party who would then join her at the obtain food items [Breuer et al., 2005]. Lonsdorf et al. mound a short time after. This further suggests that [2009] reported significant differences in propensity the social tolerance hypothesis does not universally to investigate the artificial termite mound at LPZ. function at the individual level in this bonobo group. Prior to baiting the structure, the chimpanzees Observations of this group, in contrast, suggest to us engaged in mound‐related behaviors significantly that there may be inter‐individual differences in

Am. J. Primatol. Bonobo Sex Differences in Tool Use / 925

propensity that did not appear to be strictly sex‐ Gruber T, Clay Z, Zuberbühler K. 2010. A comparison of bonobo based. Although females in this group fished signifi- and chimpanzee tool use: evidence for a female bias in the – cantly more frequently than did the males, some Pan lineage. Anim Behav 80:1023 1033. Gumert MD, Hoong LK, Malaivijitnond S. 2011. Sex differ- females showed little interest in the mound (e.g. ences in the stone tool‐use behavior of a wild population of Lady, Susie, and JoT), whereas some males fished burmese long‐tailed macaques (Macaca fascicularis aurea). regularly (e.g. Maiko, Gander, and Jerry). Further Am J Primatol 73:1239–1249. studies are needed to test whether it may be Hare B, Melis A, Woods V, Hastings S, Wrangham R. 2007. Tolerance allows bonobos to outperform chimpanzees on a differences in propensity and aptitude, rather than cooperative task. Curr Biol 17:619–623. social tolerance, which are driving sex differences, Herrmann E, Wobber V, Call J. 2008. Great apes’ (Pan and possibly species differences, in tool use troglodytes, Pan paniscus, Gorilla gorilla, Pongo pygmaeus) behaviors. understanding of tool functional properties after limited experience. J Comp Psychol 122:220–230. Herrmann E, Hare B, Call J, Tomasello M. 2010. Differences in ACKNOWLEDGMENTS the cognitive skills of bonobos and chimpanzees. PLoS ONE 5(8):e12438. We gratefully acknowledge Karen Huebel, Kelly Hohmann G, Fruth B. 2003. Culture in bonobos? Between‐ Vineyard, Shannon Morarity, Shelly Roach, and the species and within‐species variation in behavior. Curr Anthropol 44:563–571. entire African Forest staff at the Columbus Zoo for Horner V. 2010. The cultural mind of chimpanzees: how social their institutional support, construction of the tolerance can shape the transmission of culture. In: Lonsdorf termite mound, and for allowing us to collect EV, Ross SR, Matsuzawa T, Goodall J, editors. The mind of behavioral observations. We thank Stephany Harris, the chimpanzee: ecological and experimental perspectives. – Drew Enigk, Mary Fisher, Ashley Thoerner, Nicole Chicago: University of Chicago Press. p 101 115. Ingmanson EJ. 1996. Tool‐using behavior in wild Pan paniscus Thoerner, and Patrick Hoehn for their assistance in social and ecological considerations. In: Russon AE, Bard video recording, and Thomas Dick for providing KA, Parker ST, editors. Reaching into thought: the minds of graphic design work. We also wish to thank two the great apes. New York: Cambridge University Press. p anonymous reviewers for their helpful comments and 190–210. Jordan C. 1982. Object manipulation and tool‐use in captive insightful suggestions. pygmy chimpanzees (Pan paniscus). J Hum Evol 11:35–39. Kano T. 1982. The social group of pygmy chimpanzees (Pan REFERENCES paniscus) of Wamba. Primates 23:171–188. Kano T. 1992. The last ape: pygmy chimpanzee behavior and Bentley‐Condit VK, Smith E. 2010. Animal tool use: current ecology. Stanford: Stanford University Press. p 248. definitions and an updated comprehensive catalog. Behav- Kohler W. 1927. The mentality of apes. London: Routledge & iour 147:185–221. Kegan Paul. p 286. Boesch C, Boesch H. 1984. Possible causes of sex differences in Kyes RC. 1988. Grooming with a stone in sooty mangabeys the use of natural hammers by wild chimpanzees. J Hum (Cercocebus atys). Am J Primatol 16:171–175. Evol 13:415–440. Lonsdorf E, Eberly L, Pusey A. 2004. Sex differences in Boesch C, Boesch‐Achermann H. 2000. The chimpanzees of the learning in chimpanzees. Nature 428:715–716. Tai forest: behavioural ecology and evolution. New York: Lonsdorf E, Ross S, Linick S, Milstein M, Melber T. 2009. An Oxford University Press. p 316. experimental, comparative investigation of tool use in Boinski S, Quatrone RP, Swartz H. 2000. Substrate and tool chimpanzees and gorillas. Anim Behav 77:1119–1126. use by brown capuchins in Suriname: ecological contexts and Lonsdorf EV. 2005. Sex differences in the development of cognitive bases. Am Anthropol 102:741–761. termite‐fishing skills in the wild chimpanzees, Pan troglo- Boysen ST, Kuhlmeier VA, Halliday P, Halliday YM. 1999. Tool dytes schweinfurthii, of Gombe National Park, Tanzania. use in captive gorillas. In: Parker ST, Mitchell RW, Miles HL, Anim Behav 70:673–683. editors. The mentalities of gorillas and orangutans: compar- McGrew WC. 1979. Evolutionary implications of sex differ- ative perspectives. Cambridge: Cambridge University Press. ences in chimpanzee predations and tool use. In: Hamburg p 179–187. DA, McCown ER, editors. The great apes. Menlo Park: Breuer T, Ndoundou‐Hockemba M, Fishlock V. 2005. First Benjamin Cummings. p 441–463. observation of tool use in wild gorillas. PLoS Biol 3:2041– McGrew WC. 1992. Chimpanzee material culture. Cambridge: 2043. Cambridge University Press. p 424. Byrne RW. 2004. The manual skills and cognition that lie McGrew WC, Tutin CEG, Baldwin PJ. 1979. Chimpanzees, behind hominid tool use. In: Russon AE, Begun DR, editors. tools, and termites: cross‐cultural comparisons of Senegal, The evolution of thought: evolutionary origins of great ape Tanzania, and Rio Muni. Man 14:185–214. intelligence. Cambridge: Cambridge Univ Press. p 31–44. Nishida T. 1990. The chimpanzees of the Mahale Mountains: Cunningham CL, Anderson JR, Mootnick AR. 2006. Object sexual and life history strategies. Tokyo: University of Tokyo manipulation to obtain a food reward in hoolock gibbons Press. p 328. (Bunopithecus hoolock). Anim Behav 71:621–629. Pansini R, de Ruiter J. 2011. Observation of tool use and Fontaine B, Moisson PY, Wickings EJ. 1995. Observations of modification for apparent hygiene purposes in a mandrill. spontaneous tool making and tool use in a captive group of Behav Process 88:53–55. western lowland gorillas (Gorilla gorilla gorilla). Folia Phillips KA. 1998. Tool use in wild capuchin monkeys (Cebus Primatol 65:219–223. albifrons trinitatis). Am J Primatol 46:259–261. Galdikas BMF. 1982. Orang‐utan tool‐use at Tanjung Puting Pruetz JD, Bertolani P. 2007. Savanna chimpanzees, Pan Reserve, central Indonesian Borneo (Kalimantan Tengah). J troglodytes verus, hunt with tools. Curr Biol 17:412–417. Hum Evol 11:19–33. Reynolds V. 2005. The chimpanzees of the Budongo Forest: Goodall J. 1986. The chimpanzees of Gombe: patterns of ecology, behaviour, and conservation. Oxdord: Oxford behavior. Cambridge: Harvard University Press. p 673. University Press. p 297.

Am. J. Primatol. 926 / Boose et al.

Rodrigues MR, Lindshield SL. 2007. Scratching the surface: van Lawick‐Goodall J. 1970. Tool‐using in primates and other observations of tool use in wild spider monkeys. Am J Phys vertebrates. Adv Study Behav 3:195–249. Anthropol 44:201–202. van Schaik CP, Fox E, Sitompul A. 1996. Manufacture and use Santos LR, Mahajan N, Barnes JL. 2005. How prosimian of tools in wild Sumatran orangutans. Naturwissenschaften primates represent tools: experiments with two lemur 83:186–188. species (Eulemur fulvus and Lemur catta). J Comp Psychol van Schaik CP, Deaner RO, Merrill MY. 1999. The conditions 119:394–403. for tool use in primates: implications for the evolution of Santos LR, Pearson HM, Spaepen GM, Tsao F, Hauser M. 2006. material culture. J Hum Evol 36:719–741. Probing the limits of tool competence: experiments with two van Schaik CP, Ancrenaz M, Borgen G, et al. 2003. Orangutan non‐tool‐using species (Cercopithecus aethiops and Saguinus cultures and the evolution of material culture. Science oedipus). Anim Cognit 9:94–109. 299:102–105. Stoinski TS, Beck BB. 2001. Spontaneous tool use in captive, Visalberghi E. 1990. Tool use in Cebus. Folia Primatol 54:146– free‐ranging golden lion tamarins (Leontopithecus rosalia 154. rosalia). Primates 42:319–326. White FJ. 1996. Comparative socioecology of Pan paniscus. Sokal RR, Rohlf FJ. 2012. Biometry: the principles and practice In: McGrew WC, Nishida T, Marchant L, editors. Great of statistics in biological research. New York: W. H. Freeman ape societies. Cambridge: Cambridge University Press. and Co. p 937. p29–41. Struhsaker TT. 1975. The red colobus monkey. Chicago: Whiten A, Goodall J, McGrew W, et al. 2001. Charting cultural University of Chicago Press. p 326. variation in chimpanzees. Behaviour 138:1481–1516.

Am. J. Primatol.