Experimental evidence that intruder and group member attributes affect outgroup defence and associated within-group interactions in a social fish

Abstract

In many social species, individuals communally defend resources from conspecific outsiders. Participation in defence and in associated within-group interactions, both during and after contests with outgroup rivals, is expected to vary between group members because the threat presented by different outsiders is not the same to each individual. However, experimental tests examining both the contributions to, and the consequences of, outgroup conflict for all group members are lacking. Using groups of the cichlid Neolamprologus pulcher, we simulated territorial intrusions by different-sized female rivals and altered the potential contribution of subordinate females to defence. Dominant females and subordinate females defended significantly more against size- and rank-matched intruders, while males displayed lower and less variable levels of defence. Large and small, but not intermediate-sized, intruders induced increased levels of within-group aggression during intrusions, which was targeted at the subordinate females. Preventing subordinate females from helping in territorial defence led to significant decreases in post-contest within-group and female-specific submissive and affiliative displays. Together, these results show that the defensive contributions of group members vary greatly depending both on their own traits and on intruder identity, and this variation has significant consequences for within-group social dynamics both during and in the aftermath of outgroup contests.

Keywords: Neolamprologus pulcher; defensive contributions; outgroup conflict; social interactions; territorial intrusions.

Figure 1.

Defence behaviour of resident group members towards different-sized female intruders (Experiment I). Fitted values (mean ± 95% confidence intervals) and partial residuals (black dots) from LMMs in electronic supplementary material, table S1 are shown. Statistics run on square-root transformed data. Significant treatment differences within individual categories highlighted. *p < 0.05, **p < 0.01. DF, dominant female; DM, dominant male; SF, subordinate female; SM, subordinate male. (Online version in colour.)

Figure 2.

Within-group acts of aggression (a) displayed and (b) received during intrusions by different-sized females (Experiment I). Fitted values (mean ± 95% confidence intervals) and partial residuals (black dots) from LMMs in electronic supplementary material, table S3 are shown. Significant treatment differences highlighted. *p < 0.05, **p < 0.01. DF, dominant female; DM, dominant male; SF, subordinate female; SM, subordinate male. (Online version in colour.)

Figure 3.

Post-intrusion change in within-group frequency of (a) submission displayed, (b) submission received, (c) affiliation displayed and (d) affiliation received depending on variation in subordinate female presence or absence and contribution to territorial defence (Experiment II). Fitted values (mean ± 95% confidence intervals) and partial residuals (black dots) from LMMs in electronic supplementary material, tables S12 and S13 are shown. Significant treatment differences highlighted. *p < 0.05, **p < 0.01. DF, dominant female; DM, dominant male; SF, subordinate female; SM, subordinate male. (Online version in colour.)