Although stable hierarchies have been documented in a wide variety of species, virtually nothing is known about what occurs when a hierarchy is disrupted, and the group is subsequently re-formed. How likely is it that the same hierarchy will re-form? This attribute, which we quantify using a measure we call replicability, may provide insight into the processes governing hierarchy formation and maintenance. We examined replicability in the cockroach Nauphoeta cinerea. A total of 11 groups, with four males in each group, were subjected to this formation and re-formation process. Aggressive interactions, leading to an eventual hierarchy formation, were observed for 3 d. Groups were then disbanded and individuals housed in isolation for 2 d. These same individuals were then reintroduced into the same 11 groups, and observed for another 3 d. Although 10 of the 11 groups formed stable hierarchies the first time they were put together, only six of the hierarchies were replicable.
While it is difficult to pinpoint why some hierarchies were replicable and others were not, one trend emerged that might help explain the difference. Alpha individuals in replicable hierarchies participated in a lower proportion of interactions the first time groups were together than alpha individuals in non-replicable hierarchies. Alpha individuals in non-replicable hierarchies thus had to participate in more interactions to maintain their alpha status. The combination of such exertion, group-dispansion, and re-establishment of position may preclude regaining alpha status for such individuals.