Chimpanzees overcome the tragedy of the commons with dominance

Abstract

Competition over common-pool resources (CPR) is a ubiquitous challenge for social animals. Many species face similar dilemmas, yet our understanding of the evolutionary trajectory of CPR social strategies remains unexplored. Here, we provide a first look at the social strategies of our closest living relatives, chimpanzees (Pan troglodytes), in two novel resource dilemma experiments. Dyads of chimpanzees were presented with renewable resource systems, collapsible at a quantity-dependent threshold. Dyads had to continuously resist overconsumption to maximize collective gains. In study 1, dyads of chimpanzees sustained a renewing juice source. Inequality of juice acquisition between partners predicted sustaining success, indicating that one individual dominated the task while the partner inhibited. Dyads in study 2 fed together on accumulating carrot pieces but could end the accumulation any time by grabbing an immediate selfish source of carrots. Dyads with low tolerance were more successful at collectively sustaining the resource than highly tolerant dyads. Further, the dominant individual was more likely to cause collapse in dyads with low tolerance than dyads with high tolerance. These results indicate that chimpanzees use a dominance-based monopolisation strategy moderated by social tolerance to overcome the tragedy of the commons.

Figure 1

Study 1 apparatus mechanisms portrayed in detail.

Figure 2

Study 1 setup in (a) the collective condition with one shared juice system, and (b) the parallel condition with two independent juice systems. The subject on the right has just collapsed her juice system.

Figure 3

Collapse latencies (x-axis) of all sessions, plotted in chronological order of sessions received by dyad from bottom to top on the y-axis. Vertical dashed lines represent collapse latency predictions of the model for each condition. The collective condition was more difficult than the parallel condition, yet three of these collective sessions clearly exceeded predicted latencies for the CPR condition.

Figure 4

Per-session proportions of self-distraction (pink diamonds) and drinking synchronicity (red dots) as a function of collapse latency on the x-axis in (a) the parallel condition and, (b) the collective condition.

Figure 5

The individual drinking rates of each dyad’s most successful collective condition session (also last session of this condition for all dyads). The dashed line indicates the average drinking rate of pre-test subjects. The yellow bar indicates the range of the estimated juice drip rate from the source into the cylinder where it became available for drinking.

Figure 6

Study 2 setup in (a) the collective condition (E1 present but not pictured) and (b) the parallel condition in which the subject on the left has just collapsed her carrot system.

Figure 7

Effect of trial number and condition on collapse latency in minutes (displayed on a log-scale). Dyads improved collapse latency with experience in the parallel condition (blue) and shortened collapse latency with experience in the collective condition (red).

Figure 8

Effect of dyadic co-feeding tolerance and condition on collapse latency in minutes (displayed on log-scale). More tolerant dyads performed better in the parallel condition and worse than low tolerance dyads in the collective condition.

Figure 9

Effect of co-feeding tolerance and condition on the proportion of subordinate-caused collapses. The interaction between tolerance and condition showed a non-significant trend: far fewer subordinate individuals in low tolerance dyads caused collapse in the collective condition than in the parallel condition.