Capture model
The overall capture probability of birds that detected the decoy predator was 0.21 (SE = 0.022, n = 340 trials). The interaction between supplementary feeding and age of nestlings strongly affected the time-to-capture, our measure of mobbing intensity (Table 2). Food supplemented parents showed considerably shorter time-to-capture with old than young nestlings. This relationship disappeared in un-supplemented parents which showed constant time-to-capture throughout the nestling period. This resulted in increased time-to-capture in fed parents with young nestlings (Fig. 2a), and reduced time-to-capture in fed parents toward the end of the nestling period (Fig. 2b). Time-to-capture was shorter in parents with large than small broods (Table 2, Fig. 3). Control variables also affected the time-to-capture. Disturbance by bycatch, repeated capturing attempts at the same brood, and a decoy predator positioned close to the forest edge increased the time-to-capture (Table 2). Also, the weather variables wind and high temperatures resulted in increased time-to-capture.