Individual contributions to territory defence in a cooperative breeder: weighing up the benefits and costs

Abstract

While investment in territory defence is expected to be influenced by its benefits, the additional role that costs may play is rarely considered. Here, we quantify both benefits and costs of repelling prospecting males in cooperative meerkats, and demonstrate that both are required to explain the substantial variation in individual contributions to the defence observed. Males benefit more from repelling prospectors than females, as males may lose dominance and be expelled during intrusions. Accordingly, males invest the most in repelling prospectors. We also show that males experience an associated cost in the form of reduced weight gain and, as such, heavier males contribute more to chasing prospectors. Finally, we show evidence of a cost not restricted to individuals engaged in chasing: both males and females reduce their contributions to feeding dependent pups when prospectors are present, resulting in a reduction in pup weight gain in this context. Males appear to adjust their contributions to chasing in light of this cost, chasing at lower rates when their group contains dependent young. Our findings support the view that investment in cooperative behaviours can be attributed to benefits and costs, and highlight the additional importance of considering trade-offs in investment between cooperative behaviours.

Figure 1.

Differences (a) between resident females and males in the proportion of chases of prospectors led (see electronic supplementary material, table S6; GLMM estimate, sex = 2.62, 95% CI 1.71–3.52); and (b) between resident males of different age category in their rates of chase leads (electronic supplementary material, table S7; GLMM estimate, subadult = −5.43, 95% CI −5.78 to −5.08, yearling = −4.61, 95% CI −4.98 to −4.24, adult = −4.62, 95% CI −5.03 to −4.21). Bars present predicted means ± s.e. from the GLMMs, estimated using means of predictor variables not graphed. Proportions in (a) do not sum to one because leaders were not identified in all of the chases recorded.

Figure 2.

Proportion of males in each age and social status (S, subordinate; D, dominant) category that permanently left their group (dark grey) or stayed but lost dominance (light grey), within the first week after a prospecting male immigrated into their group (n = 27 immigration events).

Figure 3.

Weight gain rates (g/h) of resident males across different age and social status categories when prospectors were absent (light grey) and present (dark grey) at their group (see electronic supplementary material, table S8a). LMM estimates with 95% CI of interaction with the presence of a prospector: juvenile = −0.14 (−1.35 to 1.06), subadult = −0.44 (−1.39 to 0.51), yearling = −0.38 (−1.32 to 0.56), adult subordinate = −1.55 (−2.60 to −0.49), adult dominant = −4.03 (−5.36 to −2.71). Bars present predicted means ± s.e. from the LMM, estimated using means of predictor variables not graphed.

Figure 4.

Pup feed rates (feeds h⁻¹) of (a) males and (b) females across different age and social status categories when prospectors were absent (light grey) and present (dark grey) at their group. See electronic supplementary material, table S9a for males; GLMM estimate with 95% CI: prospector present = −0.29 (−0.52 to −0.06). See electronic supplementary material, table S9b for females; GLMM estimates with 95% CI of interaction with the presence of a prospector: juvenile = 0.57 (0.07–1.07), subadult = −0.19 (−0.47 to 0.09), yearling = −0.22 (−0.49 to 0.04), adult subordinate = −0.52 (−1.02 to −0.02), adult dominant = −1.86 (−3.05 to −0.66). Bars present predicted means ± s.e. from the GLMMs, estimated using means of predictor variables not graphed.

Figure 5.

Weight gain rates (g h⁻¹) of pups when prospectors were absent and present at their group (see electronic supplementary material, table S10; LMM estimate, prospector present = −0.96, 95% CI −1.50 to −0.43). Bars present predicted means ± s.e. from the LMM, estimated using means of predictor variables not graphed.