Habitual well-digging in a rainforest-living group of East African chimpanzees, which may have been imported into the community’s behavioural repertoire by an immigrant female

Well-digging in a community of forest-living wild East African chimpanzees (Pan troglodytes schweinfurthii). Hella Péter, Klaus Zuberbühler & Catherine Hobaiter. Primates, Jun 6 2022. https://rd.springer.com/article/10.1007/s10329-022-00992-4

Abstract: Access to resources shapes species’ physiology and behaviour. Water is not typically considered a limiting resource for rainforest-living chimpanzees; however, several savannah and savannah-woodland communities show behavioural adaptations to limited water. Here, we provide a first report of habitual well-digging in a rainforest-living group of East African chimpanzees (Pan troglodytes schweinfurthii) and suggest that it may have been imported into the community’s behavioural repertoire by an immigrant female. We describe the presence and frequency of well-digging and related behaviour, and suggest that its subsequent spread in the group may have involved some degree of social learning. We highlight that subsurface water is a concealed resource, and that the limited spread of well-digging in the group may highlight the cognitive, rather than physical, challenges it presents in a rainforest environment.

Discussion

Chimpanzees living in water-restricted areas are able to exploit subsurface water by digging wells to access it (Nishida et al. 1999; McGrew et al. 2003; Hunt and McGrew 2002). Waibira chimpanzees were observed digging wells by hand next to a pool with stagnant surface water, their main water source during the dry season. We did not observe any use of tools for well-digging. Over a period of 377 days across seven annual dry seasons we documented habitual well-digging in four female chimpanzees. Importantly, this happened while stagnant surface water was available at the same time, suggesting they preferred the well water. Other individuals subsequently exploited the pre-dug wells for their own water access, either directly or by using sponges, again while (stagnant) surface water was available nearby.

Although digging behaviour (e.g. play digging) had been observed in the Waibira community prior to ONY’s immigration in 2015, we recorded no observations of well-digging in the community prior to 2015, despite camera trap video recording at the site (ongoing since January 2013) and direct observations during focal follows (ongoing since 2012). We also recorded no indirect evidence of well-digging at the water hole prior to 2015. While it is impossible to rule out that we missed this behaviour or failed to recognise the indirect traces of it, we consider it likely that ONY introduced the behaviour into the Waibira group. First, her competence and frequency of well-digging were remarkable from the beginning, suggesting that she knew the behaviour prior to immigration. Second, and equally remarkable, were the behavioural responses of other adult individuals who closely observed her well-digging behaviour and then exploited her wells over several years, suggesting that the behaviour was previously unknown to other Waibira adults.

A similar pattern was recorded in adult chimpanzees in Bossou, who closely observed (‘peering’) previously unknown nut-cracking behaviour in a field experiment (Biro et al. 2003). It has been argued that peering is a good indicator of ongoing social learning in apes (Schuppli et al. 2016), which is in line with our observations. Since its introduction in 2015, well-digging has now been observed repeatedly and in multiple individuals, suggesting it has spread–potentially by social learning–within the Waibira community. No similar behaviour has ever been observed in the well-studied neighbouring Sonso community (Reynolds 2005), despite three decades of careful observations and the fact that both groups occupy the same continuous forest habitat. Sonso chimpanzees, however, benefit from year-round access to a small river (the Sonso river), which flows across core areas of their territory and provides continuous access to fresh water.

Over the seven-season study period, we observed eight individuals to dig wells, but all four of the habitual well-diggers were females (three adults and one juvenile). While some younger males were observed to dig a well on at least one occasion, no adult male has so far been seen to dig one (although they were observed exploiting wells dug by others, suggesting a preference for the well water over the stagnant water that was still freely available). This female-biased pattern of spread is similar to that observed in Japanese macaques (Macaca fuscata) for potato washing (Nakamichi et al. 1998). Our camera trap coverage of the water hole is incomplete and it was not possible to observe all drinking events or social behaviours. Thus, it is likely that some well-digging and other digging-related behaviour were not captured in our video dataset. However, there is characteristic physical evidence of well-digging that is relatively easy to identify, such as the hole having clearly defined sides; the presence of separate marks made by fingers at the hole’s lip (where the fingers are initially dug into the soil) or in the area where the tailings remain; the presence of a small pile of substrate where the direction of digging with the fingers is consistent; in addition, the likelihood that well-digging has occurred is further supported by the presence of drinking tools in or around the hole (Video S4) (McGrew et al. 2007, 2013). During the same period of observation, the water hole and the surrounding area was surveyed regularly (at least once a week) for another study, and no indirect physical evidence of well-digging was observed prior to 2015. The failure to acquire this–very easily performed–behaviour by more individuals is puzzling, particularly since they have been observed to exploit the wells dug by other individuals. One possible explanation is that, while the physical act of well-digging for subsurface water is easy, the cognitive puzzle presented by its status as a concealed resource is more challenging, particularly given the (short) delay between the action of digging and the appearance of clean water, and the likely absence of clear cues to its presence. For example, in a muddy rainforest water hole olfactory cues to subsurface water presence are likely obscured by those from stagnant surface water. Delayed rewards or trace conditioning is shown to negatively impact the speed of learning compared to direct stimulus association (Kamin 1961; Beylin et al. 2001), and due to its significant cognitive demands, has been proposed as a possible test for animal consciousness (Shea and Heyes 2010). Additional reinforcing factors, such as observing a conspecific well-digging, may facilitate recognition of the connection between the action of manual digging and acquiring clean subsurface water.

Given wider trends in the spread and maintenance of group-specific behaviour in chimpanzees, we predict that in the future we will see (1) further spread of well-digging between adult females and immature individuals of both sexes; (2) matrilineal spread, as three of the habitual well-diggers were mature females, two of them with offspring; and (3) the possible spread to adult males with the maturation and rise in rank of immature male well-diggers.

In conclusion, we describe a new case of well-digging in chimpanzees–the first described for a rainforest-living group. We describe the apparent spread of this behaviour, which was potentially introduced by an immigrant female. The repeated innovation of well-digging across four communities of chimpanzees (McGrew et al. 2003, 2007; Hunt and McGrew 2002) could be explained by individual learning in response to a strong ecological necessity; however, the apparent absence of this behaviour in Waibira prior to ONY’s immigration, followed by its subsequent rapid acquisition by a few–but not many–group members, suggests that there was a socially mediated component to its spread in Waibira. Our observations support previous evidence suggesting that social transmission typically occurs to other younger or low or similarly ranked individuals (Horner et al. 2010). The now habitual use of a technique previously associated with communities that live in savannah or savannah-woodland highlights the importance of seasonal variation as well as broad ecological variation in resource availability for forest-dwelling chimpanzees (Wessling et al. 2018). Irrespective of the typical availability of water as a resource, the limited presence of water during at least some periods of the year appears sufficient to shape chimpanzee behaviour in the Waibira community. Taken together, these observations highlight both the striking variation and flexibility of chimpanzee behavioural repertoires.