Cofeeding tolerance in chimpanzees depends on group composition: A longitudinal study across four communities

Abstract

Summary Social tolerance is generally treated as a stable, species-specific characteristic. Recent research, however, has questioned this position and emphasized the importance of intraspecific variation. We investigate the temporal stability of social tolerance in four groups of sanctuary-housed chimpanzees over eight years using a commonly employed measure: experimental cofeeding tolerance. We then draw on longitudinal data on the demographic composition of each group to identify the factors associated with cofeeding tolerance. We find appreciable levels of variation in cofeeding tolerance across both groups and years that correspond closely to changes in group-level demographic composition. For example, cofeeding tolerance is lower when there are many females with young infants. These results suggest that social tolerance may be a “responding trait” of chimpanzee sociality, reflecting individual-level behavioral responses to social changes. Additional, experimental research is needed to better model the causal drivers of social tolerance within and among species.

Group‐level variation in co‐feeding tolerance between two sanctuary‐housed communities of chimpanzees (Pan troglodytes)

“…More specifically, by measuring co‐feeding tolerance over time it becomes possible to see whether demographic changes may influence co‐feeding tolerance or whether there might be additional group‐level processes at play that induce certain levels of social tolerance contrary or beyond expectations based on demographics factors alone. In fact, a contemporary study investigating longitudinal expressions of co‐feeding tolerance in chimpanzees with one of the applied measures – the peanut swing – provisionally reports that whereas a large part of the variation can be explained by demographic variables like the number of mothers with dependent offspring, a significant part of group‐level stability remains unexplained (DeTroy et al., 2021). Whether this unexplained portion might be best accounted for by unexplored social factors (like the state of female receptivity or the influence of high‐status or dominant individuals) or possibly by reaction norms that are shaped by within‐group social learning processes (e.g.,Boesch et al., 2020; Cantor & Whitehead, 2015; van Leeuwen et al., 2018; Sapolsky, 2006)) is an exciting avenue for future research.…”

“…In contrast to the peanut assays in which resources deplete over time (and thus the relevance of being in the resource zone decreases), the juice pipe provides a continued incentive for all group members to be in the resource zone – an incentive that requires tolerance to be satisfied. Hence, the juice pipe assay necessitates a continuous level of social tolerance throughout the session in order to obtain access to the resource (also see (DeTroy et al., 2021)). As such, the juice pipe assay revealed group differences in the sampled groups whilst the established assays (the peanut swing and peanut plot (Cronin De Groot & Stevens, 2015; Cronin, van Leeuwen, et al., 2014)) did not, or did so less delineated (different slopes in the peanut swing, see previous paragraph).…”

DieTryin: An R package for data collection, automated data entry, and post-processing of network-structured economic games, social networks, and other roster-based dyadic data

“…The second widely-used approach is a roster-based design, whereby the researcher generates a list of all members of a population and then asks each participant to report whether they have a specific kind of relationship with each-and-every individual on the roster (Marsden, 2005). In exceptional cases, social networks can also be created from long-term ethnography or observation (see Ready et al, 2020a;DeTroy et al 2021), focal or scan sampling (see Altmann, 1974;Amato et al 2013), and direct GPS tracking or proximity detection (see Davis et al 2018;Wood et al, 2021). Each of these methods carries their own costs and benefits.…”

Reliable network inference from unreliable data: A tutorial on latent network modeling using STRAND

“…Social network analysis allows researchers to parse the potential mechanisms that influence social relationships at different levels-i.e., the individual level (e.g., popularity), dyadic level (e.g., attribute-similarity), or higherorder levels (e.g., the formation of transitive groups where 'friends of friends become friends'). Given the promise of such an approach, network-based frameworks are now widely used for assessing the formation and maintenance of friendships (Ball and Newman 2013;Krackhardt and Kilduff 1999;Selfhout et al 2010), examining theories of social support/cooperation (Koster and Leckie 2014;Lakey and Cohen 2000;Nolin 2012;Power 2017;von Rueden et al 2019;Zhu et al 2013) and/or animus (Gervais 2017;Pisor and Ross 2021), describing the dynamics of drug and alcohol use (Knecht et al 2011;Ready et al 2020), and measuring interaction rates, spatial proximity, and tolerance in human and nonhuman animals (Crofoot et al 2011;Farine et al 2016;Eagle et al 2009;DeTroy et al 2021).…”