In size-structured groups, conflict over rank, resources or access to breeding opportunities is expected to be greatest among individuals that are similar in size. We tested this general prediction using the cooperatively breeding African cichlid, Neolamprologus pulcher. We predicted that, when size differences between group members were small, we would observe some or all of: increased aggression, increased submissive behaviour, increased help by subordinates or avoidance of dominants by subordinates. We created standardised groups each with a breeder male and female and a large and small helper (both males). The size of all group members was kept constant, with the exception of the breeder males, which were either only slightly larger than the largest helper or much larger. This created either large or small size differences between breeder males and the large helper (the 2nd ranked male in the group). We found that large helpers showed more submissive behaviours, reduced affiliative behaviour and kept further from breeding sites when male breeders were small. We did not find a consistent influence of breeder size on aggression. Together, these results support the prediction that conflict between breeder and helper is increased when breeders are small, but that this conflict is expressed through changes in submissive and affiliative behaviours and in space use rather than aggression. In contrast to our predictions, large helpers increased helping (territorial defence) when the male breeder was large; the reasons for this are unclear.
Group size has been shown to positively influence survival of group members in many cooperatively breeding vertebrates, including the Lake Tanganyika cichlid Neolamprologus pulcher, suggesting Allee effects. However, long-term data are scarce to test how these survival differences translate into changes in group extinction risk, group size and composition. We show in a field study of 117 groups from six different colonies (three from two populations each), that group size critically influences these parameters between years. Within one year, 34% of the groups went extinct. Group size correlated positively between years and large groups did not go extinct. The latter were more likely to contain small helpers the subsequent year, which is a cumulative measure of the previous months' reproductive success. Finally, there was a tendency that large groups were more likely to contain a breeding male and female still a year after the first check. The breeder male size, breeder female size, and largest helper size did not influence these parameters, and also did not correlate with the sizes of these categories of fish after one year. This suggests that group size, and not the body size or fighting ability of group members, was the critical variable determining the success of groups. In total, seven groups had fused with other groups between years. To our knowledge, this is the first study showing long-term benefits of large group size in a cooperatively breeding fish. We discuss the importance of differential survival and dispersal of group members for the demonstrated group size effects.
In cooperative breeders, between-group dispersal of helpers is expected to occur if it increases their fitness. Genetic data suggest that helpers in the cooperatively breeding Lake Tanganyika cichlid Neolamprologus pulcher occasionally migrate into nearby groups where they again become helpers. We studied in the field how and why helpers migrate between groups by recording their ranging and social behaviours. We found that helpers spent 5.3% of their time visiting other groups, where they received similar low levels of aggression as within their home group. Large helpers visited other groups more often than small helpers and helpers visited other groups more frequently when the queue in their home group was large, suggesting that helpers with low chances to inherit the territory search for alternatives. Our data show that helpers may use other groups' territories as a refuge, as helpers actively sought shelter within territories of neighbouring groups when we experimentally increased the perceived risk of staying in their home territory. We observed two attempted and one successful case of 'voluntary' (i.e., strategic) between-group dispersal, and experimentally induced three helpers to disperse into other groups. By regular visits, helpers appear to establish familiarity and social relationships with nearby groups, which serve as 'extended safe havens' to hide from predators. In the long run frequent visiting behaviour may facilitate between group dispersal.