Abstract
Recent studies have demonstrated that experimental increases in perceived predation risk can substantially impair breeding behavior and reduce reproductive success. Perceived predation risk may also occur in the context of sexual signaling, with potential consequences for sexual selection. In songbirds, singing at dawn is an important sexual signal but may also attract predators. Here, we report on 2 experiments designed to test whether perceived predation risk affects the occurrence and timing of dawn singing in a songbird community. In a pilot experiment, we broadcast predator playbacks intermittently across half a forest plot and nonpredator playbacks across the other half throughout early spring. In the second experiment, we repeated the treatments in 16 independent but smaller plots (8 with predator calls and 8 with nonpredator calls). In the predator treatment, most species were less likely to sing at dawn (small, nonsignificant effects) and started later if they did sing (significant for 2 species). Meta-analyses combining the data from both experiments showed an overall significant effect of the treatment on both the likelihood and timing of singing. Species that were less likely to sing also sang later if they did sing, corroborating that an increase in perceived predation risk was the common cause of the effects on both measures.