Predator exposure trials
Predator exposure trials were carried out from early May to mid-July, during the rearing time of the nestlings. To trigger mobbing reactions in the breeding pairs, a human-habituated eagle owl was used as decoy predator. The eagle owl was set on a perch at forest edges in the vicinity of the focal red kite nests. To capture mobbing red kites, a Dho-gaza net (height: 4 m; length: 6 m; 60 mm mesh) was set up next to the eagle owl, just out of its reach to avoid entanglement (Bloom et al. 2007). The Dho-gaza net was mounted perpendicular to the treeline and to the wind direction in order to guide the course of the swooping. We selected breeding pairs with nestlings older than 6 days to avoid disturbance during the very early nestling period. Exposure trials were carried out between 6 am and 9 pm. The trial generally ended 90 minutes after the placement of the decoy predator. We managed to conduct a maximum of five trials per day at different nests. Other bird species also reacted to the eagle owl and had to be removed from the net. Because this could have negative effects on the mobbing behaviour of red kites, predator exposure trials were categorized into trials disturbed by bycatch and undisturbed trials (resulting in a binary variable denoted as “disturbance”). Moreover, mobbing behaviour might differ with vicinity to the forest edge due to limited manoeuvrability. This was recorded in a binary variable denoted as “proximity to trees” (decoy predator closer than 5 meters to the forest edge).
In a subset of trials, we recorded the number of swoops and calling intensity (binary: few callings = 0; frequent calling = 1). Capture success was highly associated with the number of swoops (Estimate: 0.32; 95 % CI: 0.19 – 0.44), and with calling intensity (Estimate: 1.75; 95 % CI: 1.14 - 2.37). As variation in mobbing intensity is expected to result in different time-to-capture, we recorded the time-to-capture if capture occurred. Therefore, we used capture success as a measure of mobbing intensity. To account for differences between pairs in the time-to-detection of the decoy predator, we split the trials into two periods: detection time (the time from predator placement until the first reaction of one of the members of the breeding pair) and mobbing time (the time from detection until capture). It was not possible to record data blindly because our study involved focus animals in the field and included capturing of focus individuals. No individual identification of the parents was possible; thus, mobbing behaviour reflects the behaviour of breeding pairs.