Chimpanzee females queue but males compete for social status.

Steffen Foerster,Carson M Murray,Anne E Pusey,Joseph T Feldblum,Mathias Franz,Ian C Gilby,Kara K Walker

doi:10.1038/srep35404

Abstract

Dominance hierarchies are widespread in animal social groups and often have measureable effects on individual health and reproductive success. Dominance ranks are not static individual attributes, however, but instead are influenced by two independent processes: 1) changes in hierarchy membership and 2) successful challenges of higher-ranking individuals. Understanding which of these processes dominates the dynamics of rank trajectories can provide insights into fitness benefits of within-sex competition. This question has yet to be examined systematically in a wide range of taxa due to the scarcity of long-term data and a lack of appropriate methodologies for distinguishing between alternative causes of rank changes over time. Here, we expand on recent work and develop a new likelihood-based Elo rating method that facilitates the systematic assessment of rank dynamics in animal social groups, even when interaction data are sparse. We apply this method to characterize long-term rank trajectories in wild eastern chimpanzees (Pan troglodytes schweinfurthii) and find remarkable sex differences in rank dynamics, indicating that females queue for social status while males actively challenge each other to rise in rank. Further, our results suggest that natal females obtain a head start in the rank queue if they avoid dispersal, with potential fitness benefits.

Highlights

Methods for calculating dominance rank[20,21] often require aggregating data over arbitrary time periods, which can make it difficult to keep track of changes in group composition across matrices and distinguish between the influence of demographic change and successful rank challenges on rank changes over time
Outcomes of observed agonistic interactions are not used to further update Elo scores, and individual Elo scores remain constant at their initial, estimated Elo scores. This finding indicates that the frequent rank changes observed in Model 1 (Fig. 1d) and Model 2 (Fig. 1e) are largely an artifact of assigning individuals inaccurate initial Elo scores that need to be corrected based on subsequently observed wins and losses
Our findings indicate that there were no successful rank challenges and that relative rank relationships among females did not change over their lifetimes

Summary

Introduction

Methods for calculating dominance rank[20,21] often require aggregating data over arbitrary time periods, which can make it difficult to keep track of changes in group composition across matrices and distinguish between the influence of demographic change and successful rank challenges on rank changes over time. Elo rating[22] can overcome several shortcomings of matrix-based approaches to quantifying rank positions[23,24] This method calculates and continuously updates so called Elo scores at each observed interaction, avoiding the need for selecting arbitrary time windows. We extend a recently developed modification of the Elo rating method that uses maximum likelihood fitting of Elo model parameters[25] We apply this new method to characterize long-term rank trajectories in wild eastern chimpanzees (P. troglodytes schweinfurthii), with emphasis on sex differences in the dynamics of rank relationships. We test the related predictions that (1) females tend to enter the hierarchy at low dominance ranks, (2) successful rank challenges among females are rare and/or temporary, and (3) females mainly rise in rank when senior high-ranking females die

Methods

Results

Conclusion