Journal of Human Evolution 56 (2009) 560–569
Contents lists available at ScienceDirect
Journal of Human Evolution journal homepage: www.elsevier.com/locate/jhevol
Complex tool sets for honey extraction among chimpanzees in Loango National Park, Gabon Christophe Boesch*, Josephine Head, Martha M. Robbins Max Planck Institute for Evolutionary Anthropology, Deutscher Platz 6, 04 103 Leipzig, Germany
a r t i c l e i n f o
a b s t r a c t
Article history: Received 20 May 2008 Accepted 5 April 2009
Homo faber was once proposed as a label for humans speciﬁcally to highlight their unique propensity for tool use. However, new observations on complex tool use by the chimpanzees of Loango National Park, Gabon, expand our knowledge about tool-using abilities in Pan troglodytes. Chimpanzees in Loango, when using tools to extract honey from three types of bee nests, were observed to regularly use three- to ﬁve-element tool sets. In other words, different types of tools were used sequentially to access a single food source. Such tool sets included multi-function tools that present typical wear for two distinct uses. In addition, chimpanzees exploited underground bee nests and used ground-perforating tools to locate nest chambers that were not visible from the ground surface. These new observations concur with others from Central African chimpanzees to highlight the importance of honey extraction in arguments favoring the emergence of complex tool use in hominoids, including different tool types, expanded tool sets, multifunction tools, and the exploitation of underground resources. This last technique requires sophisticated cognitive abilities concerning unseen objects. A sequential analysis reveals a higher level of complexity in honey extraction than previously proposed for nut cracking or hunting tools, and compares with some technologies attributed to early hominins from the Early and Middle Stone Age. A better understanding of similarities in human and chimpanzee tool use will allow for a greater understanding of tool-using skills that are uniquely human. Ó 2009 Elsevier Ltd. All rights reserved.
Keywords: Chimpanzee Tool use Complexity Cognition Evolution
Introduction Homo faber was once proposed as a label for humans in order to capture their unique aptitude for tool use, which placed them apart from all other animals (Oakley, 1956). Since then, researchers have imposed chimpanzees (Pan troglodytes) into this special category, as their tool use abilities have been shown to be universal and very ﬂexible (e.g., Goodall, 1970; Boesch and Boesch, 1990; McGrew, 1992; Whiten et al., 1999; Sanz and Morgan, 2007). The ubiquity of chimpanzee tool use has convinced most anthropologists that chimpanzees are part of this tool user category (Ambrose, 2001). Nevertheless, new criteria have frequently been proposed to qualify the differences between the two species. The classic criterion is that only humans modify natural objects to fashion tools (Leakey, 1961): a claim that was contradicted in the early 1960’s from observations of spontaneous use of sticks by the Gombe chimpanzees when ﬁshing for termites (Goodall, 1964). Next, it was proposed that only humans possess tool kits comprised of different tool types, of which
* Corresponding author. E-mail address: [email protected]
(C. Boesch). 0047-2484/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jhevol.2009.04.001
each fulﬁlls a different function (Oakley, 1956), however, the growing list of tool types used by the chimpanzee populations in Gombe, Mahale, Taı¨, Bossou, and Goualougo, all of whom were observed using between 15 to 25 different types of tools, have modiﬁed this claim (Sugiyama and Koman, 1979; Boesch and Boesch, 1990; Sanz et al., 2004). Another major claim was that only humans are dependent on tools for their survival (Trinkaus, 1992), but a more complete analysis has shown that, in some chimpanzee populations, the quantity of food that was acquired with tools is very important during some periods of the year (Boesch, 1996; Yamakoshi, 1998). Another proposition was that only humans fashion tools following arbitrary cultural rules rather than in a purely adaptive way (Klein, 2000). However, population differences in tool use, such as for ant dipping, were also shown to be cultural rather than adaptive (Boesch, 1996; Moebius et al., 2008). It has been proposed that only humans use tools to access underground food (Hatley and Kappelman, 1980; Wrangham et al., 1999; Laden and Wrangham, 2005), which is thought to be indicative of some higher cognitive abilities. However, new observations have documented underground food extraction with tools in different chimpanzee populations (Lanjouw, 2002; Sanz et al., 2004; Hernandez-Aguilar et al., 2007). Additionally, it was proposed that only
C. Boesch et al. / Journal of Human Evolution 56 (2009) 560–569
humans use tools to hunt (Leakey, 1961), but recent observations from Fongoli chimpanzees complemented those from Gombe and Taı¨ and help weaken this claim (Plooij, 1978; Boesch and BoeschAchermann, 2000; Pruetz and Bertolani, 2007). Finally, it has been suggested that humans are the only species to combine many tools to attain one goal, as well as use one tool for more than one purpose (Oakley, 1956). Only one type of tool is used in most cases of chimpanzee tool use, and tools are not often used for multiple purposes, but recently some evidence suggests that tool sets (i.e., different tool types that need to be used one after the other in order to reach one goal) might be regularly used by chimpanzees of Central Africa (Suzuki et al., 1995; Bermejo and Illero, 1999; Sanz et al., 2004). Combined, these studies suggest that much of what has been proposed as qualitative differences between these two species might instead be quantitative differences. Because tool use is now recognized in chimpanzees, adopting an integrated framework to describe the details and technological complexities observed in chimpanzees might help us in understanding the similarities and differences between chimpanzee and human tool-using skills. In addition, the identiﬁcation of qualitative differences in tool-using abilities may enable us to detect which cognitive and technical skills were likely to have been important in human evolution (McGrew, 1992; Byrne, 2004). Oswalt (1976) was one of the ﬁrst to propose a single framework to describe the complexity of material culture in different human civilizations, adopting a hierarchical and dichotomous taxonomy that allowed quantitative comparisons between forms and cultures. Semenov (1964) pioneered a detailed analysis of macro- and micro-traces on stones and bones to understand their functions and the way of life for the people that employed those tools. At the same time, the ‘‘chaıˆne ope´ratoire’’ framework was applied in archaeology to detail the dynamic interactions between the object and the technological activities (Leroi-Gourhan, 1964; Lemonnier, 1983; BarYosef et al., 1992; Roche et al., 1999). Such approaches have been applied to tool use in non-human primates with the aim of making comparisons between species (McGrew, 1987; Westergaard, 1994; Pruetz and Bertolani, 2007; Carvalho et al., 2008). Tool complexity has been described in terms of technounits (Oswalt, 1976), sequences or order (Wynn and McGrew, 1998; Carvalho et al., 2008), or hierarchical organization (Byrne, 2004). To better understand the similarities and differences in tool use behavior between humans and chimpanzees, we present evidence of tool use in a newly studied population of chimpanzees in Loango National Park, Gabon. Within this population, we have documented extensive use of tool sets of up to ﬁve different tools to extract honey from hives of different species of bees, including some with underground nests. By analyzing the complexity of this extractive technology using a chaıˆne ope´ratoire or operational sequence approach, which lists all possible sequential steps required from the selection of the raw material to the fulﬁllment of the goal, will allow us to make comparisons with observations of tool use in other chimpanzees and with tool use in Early and Middle Stone Age hominins. Methods Study site and habituation Loango National Park is located on the coast of Gabon, between the border with Congo and the capital city of Libreville, and was ofﬁcially declared a national park in 2002. Our study site is in the southwest section of the park between the Atlantic Ocean and a large lagoon. The site consists of several vegetation types including coastal forest, swamps, dry forest, and mangroves. The project started in February 2005 with the aim of habituating both chimpanzees and gorillas, with two teams of 2 to 3 observers
typically patrolling an area of about 80 km2 to search for traces of the apes and to try to contact them. The habituation of the apes in this area is still in its infancy; consequently, most direct observations are opportunistic and of short duration, depending upon the situation in which the animals are contacted (e.g., for the two ﬁrst years of study, the average contact duration with chimpanzees was 8.7 minutes, N ¼ 245 in year 1 and N ¼ 218 in year 2). In addition, the animals are often frightened by, or are interested in, our presence, and therefore unaltered observations of their behaviour are not yet possible. Tool collection and analysis During patrol, teams (which included the authors, research assistants, and local ﬁeld assistants) would collect all encountered tools that were believed to have been used by chimpanzees. The vast majority of tools were associated with honey extraction. We have not yet been able to conﬁrm the use of tools for termite eating by the Loango chimpanzees, as shown in Goualogo chimpanzees (Sanz and Morgan, 2007), but we could conﬁrm that nut cracking with hammers is absent, despite the abundant presence of both Coula edulis and Panda oleosa nuts (Boesch, pers. obs.). Most of the tools were found in association with obvious signs of chimpanzee activity next to or under a beehive, and covered by honey if fresh (or smelling strongly of honey). Only tools associated with such signs were collected. The teams revisited sites where tools were found by large hives, as chimpanzees would come back to a beehive as many as 5 times to extract more honey. Between February 2005 and September 2007, we collected 614 potential tools from 45 sites and stored them at our camp for later analysis. In the majority of the instances involving tree nests, chimpanzees would extract honey from nests located at least 20 m above the ground, and the tools they used were collected after they had already departed. Therefore, it was impossible to identify with conﬁdence the individual tools used by chimpanzees for different purposes from the many different tools found on the ground. Hence, we used an archaeological approach, whereby we used a clear morphological deﬁnition of tools (Table 1a and b) allowing for the classiﬁcation of the collected tools into different types, and avoiding problems of interobserver reliability. In doing so, we assumed that our morphological types corresponded to different functions, but we could conﬁrm this only in the few cases where direct observations were possible (see below). We set clear required criteria for an object to be classiﬁed under one of the six tool types. Those criteria include only objective physical properties that can be determined from impact signs on the sticks collected. From personal experience, we noticed that a stick broken forcefully and quickly from its substrate will break neatly, whereas a stick broken more progressively will exhibit ﬁbers that separate as the stick is progressively bent, producing something like a frayed end. To fulﬁll our criteria, a blunt end results from the pounding or hitting against a hard surface, bending the wood ﬁbers strongly at the end of the stick (Fig. 1a). Similarly, we considered an end to be frayed when, despite use, the ﬁbers remained extended and open (which would not happen if it was pounded). Goualougo chimpanzees have been described to intentionally produce a frayed end with their teeth (Sanz and Morgan, 2007). However, since we did not have direct observations to conﬁrm this, we did not imply such an intentional modiﬁcation. Furthermore, to prevent biased or subjectivity in tool classiﬁcations, we applied a ‘‘triple blind procedure.’’ Following the procedure used by Mercader et al. (2007), we implemented this procedure to independently assess if the tools we collected were, in fact, tools and showed signs that could be classiﬁed as tools with conﬁdence: we required a tool to show signs of modiﬁcation and wear. Three observers (CB, JH, and MR) independently determined
C. Boesch et al. / Journal of Human Evolution 56 (2009) 560–569
Table 1a Deﬁnition of tools and tool functions used by chimpanzees for honey extraction in Loango National Park. Name
Stick found under a beehive presenting at least one modiﬁcation and clear signs of wear from being used. Thick stick used to break open the protection of the beehive entrance so as to permit access. After use, the tool possesses at least one distinct blunt end. Stick used to perforate and enlarge the different compartments within the hive. After use, the tool presents at least one distinct blunt end and traces of levering sometimes on the side. Stick used to dip or scoop the honey out of the beehive. After use, the tool presents distinct signs of wear at one frayed end. Stick used to perforate through the ground to locate the honey chamber and to dig into the soil. After use, the tool presents at least one distinct blunt end with soil on it. Strips of bark used to dip and ‘‘spoon’’ the honey out of the opened beehive.
Pounding stick (1, 2) Hammer club (3)
Prying stick (2, 4) Dip stick (2, 5)
Bee probe (6), Fluid dip (7) Fishing probes (8, 4)
Punctuating stick (1, 5, 8) Digging stick (2) Perforating stick (4)
1 ¼ Sanz et al. (2004), 2 ¼ Fay and Carroll (1994), 3 ¼ Hicks et al. (2005), 4 ¼ Bermejo and Illero (1999), 5 ¼ Sanz and Morgan (2007), 6 ¼ Fowler and Sommer (2007), 7 ¼ Sanz and Morgan (2009), 8 ¼ Deblauwe et al. (2006).
the type of modiﬁcations (stripped of leaves, stripped of bark, end cut), as well as the type of wear (blunt or frayed end) for each of the potential tools. Furthermore, we tried to determine the possible function of the tool (Table 1a and b, Fig. 1a–c). We required unanimity between the three observers for a potential tool to be considered a tool, or in order to determine its function. This procedure is conservative, because items thought to be tools when collected, but did not show any signs of wear or modiﬁcation, were discarded from further analysis and their function was assigned only if all three observers agreed. In this double selection process, 178 of the 614 (29%) items collected in the forest were discarded during the triple blind test procedure. Of those considered as tools (N ¼ 436), we did not reach unanimity about the function for only 55 tools (12%) and these were therefore considered to have an ‘‘unknown’’ function. Bees and beehives in Loango National Park The chimpanzees in Loango were found to use tools mainly for extracting honey from hives of different species of bees. The most Table 1b Operational criteria for inferring function of tools used by chimpanzees for honey extraction in Loango National Park. Name
Pounder Enlarger Collector Enlarger/ Collector Swabber Perforator
Modiﬁcation Length Reduction
>3.1 cm <3 cm <3 cm <3 cm
f þ þ þ
f f f f
common hives were those of honeybees (Apis mellifera), large sweat bees or Melipone bees (Meliponula bocandei and Meliplebeia nebulata) located in trees, and small sweat bees (Meliplebeia lendliana) found in ground nests. It was difﬁcult to characterize beehives in trees, as they could be located very high off the ground. Honeybees generally make their nests in large tree trunks with a rather small opening, and many bees constantly guard the nest entrance. African honeybees can be extremely aggressive and would attack any aggressor in a swarm. We were therefore unable to conﬁrm if chimpanzees prefer to attack nests with a larger entrance and/or with less aggressive defenders. We found three intact tree nests of large sweat bees in fallen trees and dissected them (Fig. 2a and d). The entrance of the nest was either a small hole in the branch or tree trunk, or a larger hole that was partially closed by the bees with hard dried wax through which only a small entrance was seen and used by the bees. The arboreal nests had an internal diameter of about 30 cm and an internal length of as much as 1 m. Such nests are normally found in large branches of tall trees. Many bee nests are situated in branches with a very small entrance, or just a tiny fold of the bark, and cannot be accessed by the chimpanzees, while those attacked by the chimpanzees possess larger entrances or are in large branches that break off. The internal structure was difﬁcult to judge, but from the three we opened, we observed that wax chambers structured the interior of the nests so that honey was not directly accessible even after breaking open the main entrance (Fig. 2b and c). The entrance to ground nests is typically a very small and fragile tube made of wax and resin (Fig. 2e). These can descend as far as 100 cm deep underground, where small Meliplebeia bees construct one chamber that is rarely located directly vertical beneath the tube entrance (Fig. 2f). As these tubes follow an irregular and circumvented route to the underground chambers, we believe that chimpanzees would typically need to perforate the ground with sticks to ﬁnd out where those chambers are, and we regularly found many tools left behind that were still inserted in the ground. The sticks selected to perforate the ground are very straight (Fig. 1c). When perforated, the thin wax layer surrounding the chamber breaks into an open space where the honeycombs are located. This action alerts the bees, and their buzzing can be heard from the surface. Personal tests conﬁrmed that perforating the ground with a stick enables one to locate underground chambers. Statistical analysis The morphological characteristics, length, diameter, and number of modiﬁcations of the ﬁve types of tools for the three different bee nests were compared using the Kruskal-Wallis test. Due to repeated testing, we required a p < 0.01 to be considered as signiﬁcant. Results Direct observations of tool use by chimpanzees
f f f f
þ þ þ
Length reduction is observed when, after breaking off a stick, the other end is cut to reduce its length. þ ¼ required to be present; ¼ required to be absent; f ¼ facultative (as it can be seen, but is not used as a criterion in the classiﬁcation of tools).
Habituation of chimpanzees takes ﬁve years on average, so that it is not surprising that our direct observations of tool use are limited (Goodall, 1986; Boesch and Boesch-Achermann, 2000). However, we did make some clear observations of tool use in chimpanzees for the purpose of honey extraction. On July 6th, 2005, a group of 9 chimpanzees was seen gathered around a large sweat bee nest located approximately 40 m high in a tree. J.H. and an assistant heard a heavy pounding sound and saw a large pounder fall to the ground as they arrived, most likely after the nest opening was enlarged. The chimpanzees stayed for 79 minutes before becoming aware of the presence of the observers,
C. Boesch et al. / Journal of Human Evolution 56 (2009) 560–569
Figure 1. Illustrations of some of the different tool types used by chimpanzees in Loango National Park, Gabon, to extract honey from 3 different types of bee nests: a) Group of tools including pounders (right pounder diameter ¼ 4.8 cm) and enlargers; b) Closer view of three collectors (left) and two enlargers (right) (left enlarger diameter ¼ 2.0 cm); c) Ground nest tools showing two collectors (right collector diameter ¼ 1.9 cm) and some perforators with their special straight line shape; d) A female chimpanzee using an enlarger in an Apis tree nest and, e) then taking a piece of comb with her hand.
after which they quickly departed. Four individuals were observed making new tools by breaking branches from a tree, modifying them, and then inserting them into the tree hole to extract honey. Once inserted, the chimpanzees were seen to rapidly and forcefully rotate them inside the hole, suggesting that they were ﬁrst breaking up the internal chambers (with the enlarger) before ‘‘dipping’’ for the honey. One adult male was observed to ﬁrst remove the bark of a Garcinia sp. branch with its teeth and then put one end of the stick in its mouth and chew on it, as if to fray it before using it to extract honey. A short while later, an adult female with two fabricated sticks approached the nest entrance and let her juvenile offspring lick the end of the ﬁrst stick, while she used the second one to extract more honey. In another instance, on the 23rd
of July 2005, C.B. and an assistant were attracted to a site by a loud pounding sound, at which two adult males were seen to use tools to extract honey for 13 minutes from a big fallen branch. The ﬁrst chimpanzee used a large broken shrub with leaves and branches still intact, and tried to push it into the hole. He then broke the shrub, rejected the end with the leaves, and shoved the newly modiﬁed tool into the hole and performed sweeping semi-circular motions before removing it and putting it into his mouth. He then pulled the tool through his mouth to remove the honey, and reinserted it into the hole. The second male used a ready-made tool lying on the ground. In eight other instances, we saw twelve chimpanzees using tools to extract honey from both honeybee and sweat bee nests in trees or in fallen branches.
C. Boesch et al. / Journal of Human Evolution 56 (2009) 560–569
Figure 2. Illustrations of the three types of bee nests exploited by the chimpanzees in Loango National Park, Gabon: a) A large sweat bee (Meliponula sp.) tree nest in a fallen branch with the pile of tools used by chimpanzees found on the ground; b) Close-up of the intact nest entrance of a Melipone tree nest showing the cross-shaped wax entrance; c) Close-up of the entrance once the hard wax is removed, showing the softer wax layer containing the honey; d) Tree nest of honeybees (Apis mellifera) with a pounder still inserted in the nest entrance; e) Entrance to a ground nest of small sweat bees (Meliplebeia lendliana) visible only thanks to a small yellow wax tunnel and, f) Close up of one chamber containing the honey of an underground sweat bee nest.
Loı¨c Mackaga made a detailed observation of chimpanzee tool use at an arboreal honeybee nest in October 2008. An adult male chimpanzee was observed to ﬁrst break the entrance of the nest with a large pounder and then remove honeycombs with his hand. Two adult females, up to 5 m away from the hive, made tools and used them to enlarge the nest and remove pieces of honeycomb from its wall. Once successful, they let the tool fall down and removed the comb with their hands (Fig. 1d and e). One of the females made a second tool to allow her to break off more of the comb. Two juveniles and another adult female then came to remove honey and pieces of comb without using tools.
Number and type of tools used for extracting honey More tools were used on average at the tree nests of Melipone bees than at both Apis bee nests or Melipone ground nests (Table 2). We revisited some of the large Melipone nests up to 4 times and each time we found fresh tools under the nest. One of the large nests had 125 potential tools lying on the ground under the tree following 4 visits by the chimpanzees. Apis nests were also found in trees, seemed to possess smaller entrances, and fewer tools were used to extract their honey. Melipone ground nests were much more difﬁcult to detect, and we had the impression that sometimes
C. Boesch et al. / Journal of Human Evolution 56 (2009) 560–569
Table 2 Measurements, with standard deviations in brackets, for the different tool types used by the Loango chimpanzees to extract honey from the three different nest types. Nest species
Number of modiﬁcations
Pounder Collector Enlarger Pounder Collector Enlarger Collector/Enlarger Swabbera Perforator Collector
6 6 5 32 178 75 22 55 10 26
4.2 cm (2.20) 1.1 cm (0.34) 1.0 cm (0.16) 3.7 cm (1.23) 1.9 cm (6.59) 1.2 cm (0.41) 1.2 cm (0.27)
88.8 cm (20.6) 95.7 cm (43.0) 64.9 cm (16.3) 77.4 cm (32.7) 67.7 cm (31.6) 77.5 cm (24.8) 74.1 cm (18.7)
0.66 (1.15) 1.66 (1.96) 4.20 (1.30) 1.92 (1.26) 3.44 (1.55) 4.34 (1.05) 5.09 (0.81)
1.1 cm (0.24) 1.1 cm (0.31)
69.3 cm (24.0) 48.1 cm (23.1)
3.37 (0.61) 2.95 (0.85)
Melipone tree nest
Melipone ground nest a
Swabbers are bark strips that are dried and have rapidly lost their original shape making reliable measurements of length and diameter impossible.
tools were used to search for the underground chambers, but that this quest was not always successful, which could explain the lower number of tools used. The difference in the number of tools resulted not only from the fact that they used more tools for Melipone tree nests but also that they used more types of tools (Fig. 3). Tool sets (i.e., different types of tools used chronologically to access a single food source) seemed to be used for all three types of beehives (Fig. 3). In addition, tool sets were larger for Melipone tree nests, including as many as 5 types of tools, than for the other two types of nests, with three types of tools used at Apis nests and 2 types at Melipone ground nests (Table 2). Chimpanzees were seemingly using 5 types of tools in a sequential order to access the honey within Melipone tree nests: ﬁrst, a pounder was used to break open the nest entrance, then enlargers created an opening to access the different chambers within the nest, and ﬁnally, collectors were used to remove the honey. Additionally, at two nests, swabbers (elongated strips of bark) were also used as collectors. Finally, some tools presented obvious wear at both ends, with one end of the tool being blunt and the other having been frayed, indicating that they had been used for two functions, such as an enlarger as well as a collector. For Apis nests, only three different types of tools were found in the tool set: pounders, enlargers, and collectors. It is possible that painful bee stings limit the amount of time chimpanzees can stay at nests, which would in turn limit the number of tools they use. Two types of tools were found at Melipone ground nests, including one that was unique to this type of nest:
Percentage of tools
perforators, which were most likely used to vertically penetrate the ground around the nest entrance and locate the underground honey resources. Tests conducted at nest sites, and the inspection of nests previously exploited by chimpanzees, indicated that the perforators must be inserted into the ground to a depth of 20–90 cm to locate the precise position of the single underground chamber. No signs of the chambers are visible on the ground level (Fig. 1d). Furthermore, we never found a pounder at these ground nests, since dried wax was not blocking the entrance of the nest and no pounding was required to access them. Tool size and functions Different tool types within each tool set were characterized by different lengths and diameters, as measured with a ruler at the wear end for diameter (Table 2, Fig. 4a and b). Generally, tools used for the same function were rather homogenous in size, as the length and diameter did not differ between nest types (comparisons of each tool type for each of the three nest types using KruskalWallis tests found no signiﬁcant difference in diameter, length, and modiﬁcation number). The only exception was that the collectors were longer for Melipone tree nests than for Apis tree nests, and the shortest for Melipone ground nests (p < 0.001). However, tools used for different functions also looked different. Pounders used to break open nests are thicker (comparisons of pounders to other tool types across the three nest types using Kruskal-Wallis tests: p < 0.001 for diameter). We deﬁned pounders as being thicker, and this result is only a conﬁrmation of the initial impression that led us to deﬁne pounders in the ﬁrst place. In contrast, tool length is similar for each tool type regardless of nest type (p > 0.05), as they all are used for the function of accessing the honey and, therefore, the length seems more determined by the depth of the nest in which the honey is found, rather than by the function of the tool. Tool making and number of modiﬁcations
Nest type Unknown Swabber Enlarger/Collector Perforator
Collector Enlarger Pounder
Figure 3. Tool sets in chimpanzees of the Loango National Park, Gabon, used to extract honey from three different types of nests. Tool sets included 3 to 5 different tool types. Grey-marked tool types were found only for one speciﬁc type of nest.
Tools with different functions were also modiﬁed in different ways (Table 2, Fig. 5). Generally, enlargers were modiﬁed in more ways than the other tool types (Kruskal-Wallis tests: p < 0.001, for the number of modiﬁcations). Pounders showed the least signs of modiﬁcations (Kruskal-Wallis test: p < 0.001, for number of modiﬁcations). In addition, the type of bee nest also inﬂuenced the number of modiﬁcations made to a tool: collectors for Melipone tree nests had more signs of modiﬁcation than collectors used for the other two nest types (p < 0.001). Discussion Tool sets, which have been proposed to be uniquely important in human tool use, have rarely been observed in wild chimpanzee
C. Boesch et al. / Journal of Human Evolution 56 (2009) 560–569
Melipone tree Melipone ground
Tool length (cm)
80 70 60 50 40 30 20 Pounder
Tool Type 6
Melipone tree Melipone ground
Tool diameter (cm)
Apis tree 4
Tool Type Figure 4. Tool length (a) and diameter (b) of 5 different tool types used for the three bee nests by chimpanzees in Loango National Park.
populations until recently. One notable exception involved Taı¨ chimpanzees, who were observed using hammers to break the nuts of Panda oleosa, Parinari excelsa, or Detarium senegalense, and were often seen inserting a stick in order to extract pieces of kernel from inside the shells (Boesch and Boesch, 1990). Furthermore, a reintroduced group of chimpanzees in Liberia was also observed using similar sets of sticks to occasionally break open beehives (Hannah 6 Melipone tree Melipone ground Apis tree
Number of Modifications
Figure 5. Number of modiﬁcations done to 5 different tool types used for three bee nest species by chimpanzees of the Loango National Park.
and McGrew, 1987). In recent years, data from central African chimpanzees have shown the prevalence of tool sets for honey extraction and termite ﬁshing (Bai Hokou: Fay and Carroll, 1994; Lossi: Bermejo and Illero, 1999; Goualougo: Sanz et al., 2004; Sanz and Morgan, 2007; Ngotto: Hicks et al., 2005; Dja Biosphere Reserve: Deblauwe et al., 2006). The tool sets found in Loango conﬁrm that Central African chimpanzees use tool sets more systematically than chimpanzees from other African regions. More tools have been found under bee nests in Loango than in Goualougo, but this difference could be explained by possible differences in nest structure: one of the two Melipone bee species raided in Loango is not the same as the two exploited by the chimpanzees in Goualougo (Sanz and Morgan, 2009). Speciﬁcally, Loango chimpanzees did not conduct raids on mason bees that construct their nests on tree trunks, which most likely require fewer tools than when raiding lodger bees that use deep tree cavities. Chimpanzees demonstrate an elaborate ﬂexibility in their tool use and can readily use complex tool sets whenever necessary. Such tool sets are used to extract honey from both tree hives and underground hives. Gathering honey from underground hives, similar to underground termite ﬁshing in Goualougo, is special in the sense that chimpanzees cannot see where the resource is hidden and use the ﬁrst tool, the perforator, as an exploratory tool to ‘‘feel’’ where the resource is located underground. In both cases, external indirect signs of food sources are visible (e.g., large termite mounds or small fragile Melipone-made tubes), but the nest itself is not visible and its exact location cannot be inferred. Therefore, chimpanzees have to investigate the soil in order to locate food that can be, in the case of Melipone underground nests, as much as 1 m deep and 70 cm lateral to the visible tube. Locating the underground chamber can take a human between 20 to 40 minutes (Boesch, pers. obs.). The successful locating of honey is apparent from honey sticking to the ends of perforators. To extract honey, a tunnel needs to be dug sideways so as to reach the underground chamber and prevent soil from getting mixed with the honey once the membrane of the chamber is broken (in general, the intact upper membrane of the chamber in the emptied hole can be felt). We think that such tunnels are dug with the help of perforators to loosen the soil. These tunnels are sometimes barely large enough to let a human arm through, and therefore indicate that chimpanzees know exactly where they are aiming. This cannot be done by simply following the bee tube, as it is much too fragile to resist the toolassisted digging process. Thus, an elaborate understanding of unseen nest structure, combined with a clear appreciation that tools permit the location of unseen resources, and a precise threedimensional sense of geometry for reaching the honey chamber from the correct angle, is demonstrated by the chimpanzees when extracting underground honey. It has been proposed that an elaborate understanding of causal relationships between external objects is required for ﬂexible tool use to evolve (Boesch and Boesch-Achermann, 2000), and the fact that such exploratory tools are only seen in chimpanzees and humans supports this proposition. Multiple function tools, in which both ends have been used for two different functions, have previously been proposed to be a uniquely human invention. Their use has now been conﬁrmed in Loango chimpanzees for the purpose of gathering sweat bee honey. Obvious wear signs are still recognizable on both ends of 10% of the collected tools, even months after their use, indicating different functions for each end (see Sanz and Morgan  for a similar case in Goualougo chimpanzees). In the absence of direct observations, it is too early to determine how chimpanzees used these tools speciﬁcally, but it is important to note that this type of tool use is not known from other chimpanzee populations in East and West Africa.
C. Boesch et al. / Journal of Human Evolution 56 (2009) 560–569
The sequential organization of the techniques requires many steps to be performed to have the appropriate tools ready in the correct order to extract the honey (Fig. 6a,b). Not only are different actions and modiﬁcations necessary to produce tools or results, but those tools are then combined in a precise order to achieve the goal. Different sequences can be produced, for example, using only two tools and a hand for extracting honey from tree nests (Fig. 6a), or using three or four tools. In other words, different sequential routes can be taken, either with two tools or with three or four tools, to extract honey from tree nests. In addition, parts of sequences can be repeated, e.g., collectors, or enlargers and collectors, can be refashioned after having been used previously for honey eating. All of this is done high up in the trees, often while suspending in acrobatic positions, because nest entrances are frequently oriented downwards or located away from convenient suspensory branches (Fig. 1d, see also Sanz et al.  for pictures), tools are often discarded once honey can be reached, and, if needed again, tools have to be remade. This explains why we found an average of 18 tools under such tree nests. A similar sequential analysis for ground bee nests revealed that the operational sequence is less complex, as a maximum of two different types of tools was used but only one was actually mandatory (Fig. 6b). Thus, we observed Loango chimpanzees following a hierarchical sequence of steps embedded within a sequential organization of tool use, when extracting honey from beehives. A similar operational sequence approach revealed lower levels of complexity in chimpanzees, both for nut cracking (Boesch and Boesch-Achermann, 2000; Carvalho et al., 2008), and for hunting tools (Pruetz and Bertolani, 2007), as these are basically linear sequences. The same will probably be true if applied to termite ﬁshing or dipping, as seen in Gombe chimpanzee (McGrew, 1987). However, an operational sequence analysis of thistle-eating techniques in mountain gorillas revealed a non-linear sequence, and was therefore proposed to be of greater complexity than chimpanzee tool use (Byrne, 2004). However, since the non-linearity results from alternative routes used to achieve the same goal, and since only one of them will be used at a time, this sequence is less complex than the case of mandatory tool sets where different parallel sequences need to be performed to reach a goal, as is the case for honey extraction. Central African forests are especially rich in many different species of bees, and this may have lead to the use of these specialized tool sets by Central African chimpanzees for honey extraction (see also Sanz and Morgan, 2009). The forests of West Africa have many bees as well, but Taı¨ chimpanzees were seen to concentrate mainly on honey from Apis tree nests, often using only their hands; tools were rarely used, perhaps because entrances in trees are often large enough to allow direct access (Boesch and Boesch, 1990). This conﬁrms that tools are routinely used to access underground resources by some chimpanzee populations (see also Sanz et al., 2004; Hernandez-Aguilar et al., 2007). Therefore, as in the use of hammers to crack nuts, chimpanzees are signiﬁcantly increasing the food sources they can exploit within a given habitat. Humans have been proposed to be unique in shaping the resources they can extract from their environment through technology (Wrangham et al., 1999; Ambrose, 2001; Laland et al., 2007). Chimpanzees clearly show a similar propensity with these unique tool use techniques. In addition, chimpanzees can also routinely use the same tools for multiple functions, which illustrates another facet of their ﬂexibility in complex tool use. Some of the sequential actions we observed in honey gathering by chimpanzees are reminiscent of what has been proposed for early hominin tool use during the Early and Middle Stone Age (Wynn and McGrew, 1998; Roche et al., 1999; Henshilwood et al., 2001; Goren-Inbar et al., 2002; Wynn, 2002). This includes: an
Find tree beehive Selection of branches/saplings Break off branch Trim leaves/side branches Transport Cut to length Strip bark Trim one/both ends Pounder
Goal reached Mandatory action
Facultative action Hierarchical sequence Repeated sequence
Find ground beehive
Selection of branches/saplings
Break off branch
Trim leaves/side branches
Cut to length
Strip bark Perforator Trim one/both ends
Figure 6. Schematic diagram of the operational sequence structure representing the possibilities of the sequential use of tools for honey extraction in Loango chimpanzees, a) for Melipone bee tree nests, and b) for Melipone bee ground nests. Thick arrows present the obligatorily sequential order between tools to obtain honey from an intact beehive. Normal arrows present additional tools, actions, or the reuse of tools that can occur when extracting honey from a beehive. Thin arrows present actions required (solid arrows) or facultative ones (dashed arrows), on the raw material in order for it to become a tool. Tools are in a dashed light gray box, actions on raw material are in a thin dashed dark gray box, and actions are in a box. The tool type ‘‘enlarger/collector’’ is not represented in Fig. 6a, as they simply are sticks for which each end represents two tools.
C. Boesch et al. / Journal of Human Evolution 56 (2009) 560–569
appreciation of the quality of the raw material, sometimes before even being at the food source; material selectivity; transport of raw material and tools; reduction and shaping of raw material before use (reduction in length, removal of lateral branches and leaves, and intentional shaping of brush in some cases); retouching during usage; a notion of order when using sequential tools; a notion of geometry; uniformity of tool forms; and an important cultural component in tool use (e.g., Loango chimpanzees live in a forest full of Coula edulis and Panda oleaso nuts, but do not crack them open with tools as Taı¨ chimpanzees do in Coˆte d’Ivoire). However, the repeated reduction (up to 30) of one stone core in the Middle Stone Age indicates an elaborate de´bitage scheme that has not been found in chimpanzees. Another difference is that the use of tool sets as seen as in chimpanzees, has now to been found in that period of human evolution, while the use of tools to make other tools (e.g., stone knapping by Early and Middle Stone Age hominins), has not yet been observed in chimpanzees. There is no direct evidence of tool manufacture and tool use before 2.5 Ma (Ambrose, 2001). Does that mean our ancestors at that time period did not use tools? If we assume that brain size is related to tool use, then we should expect to ﬁnd plenty of evidence for tool use in Australopithecus. However, direct evidence of tool use is restricted to the occurrence of bone tools or inferred indirectly from the presence of isotopes showing evidence of protein consumption (d’Errico et al., 2001). If, however, we consider our evidence from chimpanzees, we need to take into account the environmental conditions experienced by our ancestors. Nuts, termites, and beehives need to be available in order for individuals to use tools to feed on them. However, to exploit those resources, an understanding of external object causality is required, as both edible parts, the nut and the honey, are not always directly visible. Boesch and Boesch-Achermann (2000) proposed that it is through hunting that this cognitive ability (the basis for ﬂexible tool use) is acquired. While tool use per se has been observed in many birds and mammal species, ﬂexible and complex tool use distinguishes humans and chimpanzees from all other animal species (Boesch and Boesch-Achermann, 2000). The fact that our early ancestors lived in more forested regions than previously thought (Rayner et al., 1993) suggests that they might have hunted on small mammals, like chimpanzees, and this practice may have opened the way to ﬂexible and complex tool use. The divide originally proposed between human and chimpanzee tool use continues to narrow as we discover new ways that tools are used in chimpanzee populations. Similarities in tool use between these two species lie in the fact that tools are used to reach some speciﬁc resource. It seems to be the presence of resources, combined with an understanding of how to reach them, that determines how tools are used. It is interesting that, for both humans and chimpanzees, honey, which is a rich and abundant food source in the tropical forest, seems to be an important resource in eliciting complex tool use (Bahuchet, 1985; Bailey, 1991; Hill and Hurtado, 1996). Important differences still remain, as tools used by humans have a much larger scope of function, which makes humans so dependent upon tools. But the fact remains that the dependence on tools in some human populations is much less important than in others (McGrew, 1987), suggesting that the key differences between both species in this domain might be more one of quantity rather than quality. The Savanna Model proposed that leaving the forest stimulated the acquisition of human-like behavior in our ancestors, however, tool use by chimpanzees is more frequent and diverse in forest dwelling populations rather than in the savanna (see also Boesch and Boesch, 1990; Sanz and Morgan, 2007), and thus speciﬁc ecological challenges seem more powerful in explaining the presence of different tool sets in chimpanzees. We anticipate that more tool use and tool use abilities will
be discovered in chimpanzees, as more populations living in different habitats will be studied. Hence, we need to learn more about chimpanzee tool use to resolve questions of what is speciﬁc about human tool use. Acknowledgements We thank the Centre National des Parcs Nationaux (CNPN) and the Centre National de la Recherche Scientiﬁque et Technique (CENAREST) of Gabon for permission to conduct our research in Loango National Park. This project is ﬁnancially supported by the Socie´te´ pour la Conservation et le De´veloppement (SCD) and the Max Planck Society, and logistically supported by SCD and Wildlife Conservation Society (WCS). We are very grateful to Rombout Swanborn, Tomo Nishihara, and Edward Truter for their constant logistical support. We are indebted to Loı¨c Mackaga, Luisa Rabanal, Nikki Tagg, Beke Graw, Emilie Fairet, Erick Reteno Guizard, Kharl Remanda, Aime´ Relonga, and Pierre Boukoussou for collecting tools in the forest, for their invaluable tracking assistance in the forest and contributing to the long term data collection. Special thanks to David Roubik and Crickette Sanz for the identiﬁcation of the bee species. References Ambrose, S., 2001. Paleolithic technology and human evolution. Science 291, 1748– 1753. Bahuchet, S., 1985. Les Pygme´es Aka et la Foreˆt Centrafricaine. Selaf, CNRS, Paris. Bailey, R., 1991. The Behavioral Ecology of Efe Pygmy Men in the Ituri Forest, Zaire. Anthropological Papers, Museum of Anthropology, Ann Arbor, Michigan. Bar-Yosef, O., Vandermeersch, B., Arensburg, B., Belfer-Cohen, A., Goldberg, P., Laville, H., Meignen, L., Rak, Y., Speth, J., Tchernov, E., Tillier, A.-M., Weiner, S., 1992. The excavations in Kebara Cave, Mt. Carmel. Curr. Anthropol. 33, 497–534. Bermejo, M., Illero, G., 1999. Tool-set for termite-ﬁshing and honey extraction by wild chimpanzees in the Lossi Forest, Congo. Primates 40, 619–627. Boesch, C., 1996. Three approaches for assessing chimpanzee culture. In: Russon, A., Bard, K., Parker, S. (Eds.), Reaching into Thought. Cambridge University Press, Cambridge, pp. 404–429. Boesch, C., Boesch, H., 1990. Tool use and tool making in wild chimpanzees. Folia. Primatol. 54, 86–99. Boesch, C., Boesch-Achermann, H., 2000. The Chimpanzees of the Taı¨ Forest: Behavioural Ecology and Evolution. Oxford University Press, Oxford. Byrne, R., 2004. The manual skills and cognition that lie behind hominid tool use. In: Russon, A., Begun, D. (Eds.), The Evolution of Thought: Evolutionary Origin of Great Ape Intelligence. Cambridge University Press, Cambridge, pp. 31–44. Carvalho, S., Cunha, E., Sousa, C., Matsuzawa, T., 2008. Chaıˆnes ope´ratoires and resource-exploitation strategies in chimpanzee (Pan troglodytes) nut cracking. J. Hum. Evol. 55, 148–163. Deblauwe, I., Guislan, P., Dupain, J., van Elsacker, L., 2006. Use of a tool-set by Pan troglodytes troglodytes to obtain termites (Macrotermes) in the periphery of the Dja Biosphere Reserve, Southeast Cameroon. Am. J. Primatol. 68, 1191–1196. d’Errico, F., Backwell, L., Berger, L., 2001. Bone tool use in termite foraging by early hominids and its impact on our understanding of early hominid behaviour. S. Afr. J. Sci. 97, 71–75. Fay, M., Carroll, R., 1994. Chimpanzee tool use for honey and termite extraction in Central Africa. Am. J. Primatol. 34, 309–317. Fowler, A., Sommer, V., 2007. Subsistence technology of Nigerian chimpanzees. Int. J. Primatol. 28, 997–1023. Goodall, J., 1964. Tool-using and aimed throwing in a community of free-living chimpanzees. Nature 201, 1264–1266. Goodall, J., 1970. Tool-using in primates and other vertebrates. In: Lehrmann, D., Hinde, R., Shaw, E. (Eds.), Advances in the Study of Behavior, vol. 3. Academic Press, New York, pp. 195–249. Goodall, J., 1986. The Chimpanzees of Gombe: Patterns of Behavior. Belknap Press, Cambridge. Goren-Inbar, N., Sharon, G., Melamed, Y., Kislev, M., 2002. Nuts, nut cracking, and pitted stones at Gesher Benot Ya’aqov, Israel. Proc. Natl. Acad. Sci. U S A 99, 2455–2460. Hannah, A., McGrew, W., 1987. Chimpanzees using stones to crack open oil palm nuts in Liberia. Primates 28 (1), 31–46. Hatley, T., Kappelman, J., 1980. Bears, pigs, and Plio-Pleistocene hominids: A case for the exploitation of belowground food resources. Hum. Ecol. 8, 371–387. Henshilwood, C., d’Errico, F., Marean, C., Milo, R., Yates, R., 2001. An early bone tool industry from the middle stone age at Blombos Cave, South Africa: implications for the origins of modern human behaviour, symbolism and language. J. Hum. Evol. 41, 631–678.
C. Boesch et al. / Journal of Human Evolution 56 (2009) 560–569 Hernandez-Aguilar, A., Moore, J., Pickering, T., 2007. Savanna chimpanzees use tools to harvest the underground storage organs of plants. Proc. Natl. Acad. Sci. U S A 104, 19210–19213. Hicks, T., Fouts, R., Fouts, D., 2005. Chimpanzee (Pan troglodytes troglodytes) tool use in the Ngotto Forest, Central African Republic. Am. J. Primatol. 65, 221–237. Hill, K., Hurtado, M., 1996. Ache Life History: the Ecology and Demography of a Foraging People. Walter de Gruyter, New York. Klein, R., 2000. Archeology and the evolution of human behaviour. Evol. Anthropol. 9, 17–36. Laden, G., Wrangham, R., 2005. The rise of the hominids as an adaptive shift in fallback foods: plant underground storage organs (USOs) and australopith origins. J. Hum. Evol. 49, 482–498. Laland, K., Kendal, J., Brown, G., 2007. The niche construction perspective: implications for evolution and human behaviour. J. Evol. Psychol. 5, 51–66. Lanjouw, A., 2002. Behavioural adaptations to water scarcity in Tongo chimpanzees. In: Boesch, C., Hohmann, G., Marchant, L. (Eds.), Behavioural Diversity in Chimpanzees and Bonobos. Cambridge University Press, Cambridge, pp. 52–60. Leakey, L.S.B., 1961. The Progress and Evolution of Man in Africa. Oxford University Press, London. Lemonnier, P., 1983. L’e´tude des syste`mes techniques: une urgence en technologie culturelle. Technique et Culture 1, 11–34. Leroi-Gourhan, A., 1964. Le Geste et la Parole. I: Technique et Language. Albin Michel, Paris. McGrew, W., 1987. Tools to get food: the subsistants of Tasmanian aborigines and Tanzanian chimpanzees compared. J. Anthropol. Res. 43, 247–258. McGrew, W., 1992. Chimpanzee Material Culture: Implications for Human Evolution. Cambridge University Press, Cambridge. Mercader, J., Barton, H., Gillespie, J., Harris, J., Kuhn, S., Tyler, R., Boesch, C., 2007. 4,300-year-old chimpanzee sites and the origins of percussive stone technology. Proc. Natl. Acad. Sci. U S A 104, 3043–3048. Moebius, Y., Boesch, C., Koops, K., Matsuzawa, T., Humle, T., 2008. Cultural differences in army ant predation by West African chimpanzees? A comparative study of microecological variables. Anim. Behav. 76, 37–45. Oakley, K., 1956. Man the Tool-Maker. British Museum, London. Oswalt, W., 1976. An Anthropological Analysis of Food-Getting Technology. John Wiley and Sons, New York. Plooij, F.X., 1978. Tool use during chimpanzees’ bushpig hunt. Carnivore 1, 103–106.
Pruetz, J., Bertolani, P., 2007. Savanna chimpanzees, Pan troglodytes verus, hunt with tools. Curr. Biol. 17, 412–417. Rayner, R.J., Moon, B.P., Masters, J.C., 1993. The Makapansgat australopithecine environment. J. Hum. Evol. 24, 219–231. Roche, H., Delagnes, A., Brugal, J.-P., Feibel, C., Kibunjias, M., Mourre, V., Texier, P.-J., 1999. Early hominid stone tool production and technical skill 2.34 Myr ago in West Turkana, Kenya. Nature 399, 57–60. Sanz, C., Morgan, D., 2007. Chimpanzee tool technology in the Goualougo Triangle, Republic of Congo. J. Hum. Evol. 52, 420–433. Sanz, C., Morgan, D., 2009. Flexible and persistent tool-using strategies in honeygatherring by wild chimpanzees. Int. J. Primatol.. doi:10.1007/s10764-0099350-5. Sanz, C., Morgan, D., Gulick, S., 2004. New insights into chimpanzees, tools and termites from the Congo Basin. Am. Nat. 164, 567–581. Semenov, S., 1964. Prehistoric Technology: an Experimental Study of the Oldest Tools and Artefacts from Traces of Manufacture and Wear. Cory, Adams and Mackay, London. Sugiyama, Y., Koman, J., 1979. Tool-using and -making behavior in wild chimpanzees at Bossou, Guinea. Primates 20, 513–524. Suzuki, S., Kuroda, S., Nishihara, T., 1995. Tool-set for termite-ﬁshing by chimpanzees in the Ndoki Forest, Congo. Behaviour 132, 219–234. Trinkaus, E., 1992. Evolution of human manipulation. In: Jones, S., Martin, R., Pilbeam, D. (Eds.), The Cambridge Encyclopedia of Human Evolution. Cambridge University Press, Cambridge, pp. 346–349. Westergaard, G., 1994. The subsistence technology of capuchins. Int. J. Primatol. 15, 899–906. Whiten, A., Goodall, J., McGrew, W., Nishida, T., Reynolds, V., Sugiyama, Y., Tutin, C., Wrangham, R., Boesch, C., 1999. Cultures in chimpanzees. Nature 399, 682– 685. Wrangham, R., Jones, J., Laden, G., Pilbeam, D., Conklin-Brittain, N., 1999. The raw and the stolen: cooking and the ecology of human origins. Curr. Anthropol. 40, 567–594. Wynn, T., 2002. Archaeology and cognitive evolution. Behav. Brain. Sci. 25, 426–432. Wynn, T., McGrew, W., 1998. An ape’s view of the Oldowan. Man 24, 383–398. Yamakoshi, G., 1998. Dietary responses to fruit scarcity of wild chimpanzees at Bossou, Guinea: possible implications for ecological importance of tool use. Am. J. Phys. Anthropol. 106, 283–295.