Figure legends
Fig. 1 Probability of the decoy predator being detected in relation to time after the start of the predator exposure trial (in minutes) for fed (grey lines) and unfed (black lines) red kite breeding pairs (a) in situations of low rodent activity (10 % quantile), and (b) in situations of high rodent activity (90 % quantile). Solid lines represent means and dotted lines 95 % confidence intervals of model predictions. All other variables of the cox model were set to their mean values. As we did not find differences between years, the figure is based on a model without year effect. All numerical variables were set to their mean values and wind (categorical variable) was set to zero
Fig. 2 Capture probability in relation to the time since detection of the decoy predator for fed (grey lines) and unfed (black lines) red kite breeding pairs with (a) young nestlings (10 % quantile), and (b) old nestlings (90 % quantile). Solid lines represent means and dotted lines 95 % confidence intervals of model predictions. As we did not find differences between years, the figure is based on a model without year effect. The results are shown for first trials (no repetition) without disturbance, no wind, and no proximity to trees. All numerical variables were set to their mean values
Fig. 3 Capture probability in relation to the time since detection of the decoy predator for red kite breeding pairs with different brood sizes. Solid lines represent means and dotted lines 95 % confidence intervals of model predictions. As we did not find differences between years, the figure is based on a model without year effect. The results are shown for first trials (no repetition) without disturbance, no wind, and no proximity to trees. All numerical variables were set to their mean values