Drivers of Dyadic Cofeeding Tolerance in Pan: A Composite Measure Approach

Abstract

This study aimed to construct a composite model of Dyadic Cofeeding Tolerance (DCT) in zoo-housed bonobos and chimpanzees using a validated experimental cofeeding paradigm and to investigate whether components resulting from this model differ between the two species or vary with factors such as sex, age, kinship and social bond strength. Using dimension reduction analysis on five behavioral variables from the experimental paradigm (proximity, aggression, food transfers, negative food behavior, participation), we found a two-factor model: "Tolerant Cofeeding" and "Agonistic Cofeeding". To investigate the role of social bond quality on DCT components alongside species effects, we constructed and validated a novel relationship quality model for bonobos and chimpanzees combined, resulting in two factors: Relationship Value and Incompatibility. Interestingly, bonobos and chimpanzees did not differ in DCT scores, and sex and kinship effects were identical in both species but biased by avoidance of the resource zone by male-male dyads in bonobos. Social bonds impacted DCT similarly in both species, as dyads with high Relationship Value showed more Tolerant Cofeeding, while dyads with higher Relationship Incompatibility showed more Agonistic Cofeeding. We showed that composite DCT models can be constructed that take into account both negative and positive cofeeding behavior. The resulting DCT scores were predicted by sex, kinship and social bonds in a similar fashion in both Pan species, likely reflecting their adaptability to changing socio-ecological environments. This novel operational measure to quantify cofeeding tolerance can now be applied to a wider range of species in captivity and the wild to see how variation in local socio-ecological circumstances influences fitness interdependence and cofeeding tolerance at the dyadic and group levels. This can ultimately lead to a better understanding of how local environments have shaped the evolution of tolerance in humans and other species.

Keywords: Pan paniscus; Pan troglodytes; principal component analysis; relationship quality; tolerance experiment.

Figure 1

The pasta plot paradigm. (a) Still image from video showing the pasta distributed over the resource zone in the apes’ outdoor enclosure. (b) Still image from video shortly after the apes were released into their outdoor enclosure. These images show the chimpanzees of group 1 at Beekse Bergen Safaripark.

Figure 2

Factors affecting composite measures of Pan Relationship Quality. Dyads of maternal kin scored higher on (a) Relationship Value and lower on (b) Relationship Incompatibility compared to unrelated dyads. Female–female (FF) dyads scored (c) higher on Relationship Value than female–male (FM) dyads and (d) lower than FM and male–male (MM) dyads on Relationship Incompatibility. Boxplot figure with lower and upper box boundaries at 25th and 75th percentiles, respectively. Line inside box shows median, black dots show raw data for bonobos while black triangles indicate raw chimpanzee data points. * indicates significant at p < 0.05 level.

Figure 3

The likelihood that a dyad is observed in the resource zone is dependent of its (a) Relationship Value and (b) Relationship Incompatibility. “Yes” indicates that the dyad was observed at least once in the resource zone, whereas “no” indicates it was never seen. Boxplot figure shows lower and upper box boundaries at 25th and 75th percentiles, respectively. Line inside box shows median, black dots show data points falling outside 10th and 90th percentiles. * indicates significant at p < 0.05 level.

Figure 4

Factors affecting composite measures of Pan Dyadic Cofeeding Tolerance. Scores on Tolerant Cofeeding were higher in dyads with (a) higher Relationship Value and in (b) male–male (MM) dyads versus all other sex combinations (FF = female–female; FM = female–male). Scores on Agonistic Cofeeding were higher in dyads with (c) higher Relationship Incompatibility and (d) dyads with maternal kin compared to unrelated dyads. Boxplot figures (b,d) show lower and upper box boundaries at 25th and 75th percentiles, respectively. Line inside box shows median, black dots show all data points for bonobos while black triangles indicate all chimpanzee data points. * indicates significant at p < 0.05 level.