The paradox of cooperative breeding ­whereby individuals assist others instead of reproducing independently­ is generally explained through ecological constraints, but experimental evidence is scant. Here we performed the crucial test of the role of habitat saturation through experimental creation of vacancies and find that despite abundant presence of potential mates, subordinates are reluctant to disperse into suitable vacant habitat where conspecifics are absent. We argue that sudden disappearance of multiple group members might indicate a heightened risk of predation. Thereby the results of this study are consistent with the ‘perceptual trap’ hypothesis: the avoidance of habitats because cues do not accurately reflect their quality. Interestingly, this hypothesis can also explain previous findings, which were widely interpreted as evidence for ecological constraints as a driver of cooperative breeding. Our results can have considerable implications for conservation as it means that opportunities for colonization might go unexploited.

We experimentally created vacancies using permanent translocations (>10km away) with two different treatments. First, to test whether the limited availability of suitable habitat itself can promote delayed dispersal, seven pair vacancies were created by removing all group members from their territory before the onset the breeding season. In our second treatment, single sex vacancies (N=4 male, N=3 female vacancies) were created by relocating either males or females from a territory, to confirm that dispersal can be experimentally initiated.