The study was conducted in the Zuojia Nature Reserve (44°01′–45°00′N, 126°00′–126°08′E) of Jilin Province, Northeast China. Both sparrowhawks and owls are distributed in this region, although they exhibit significant differences in their activities and dietary preferences. The sparrowhawk is primarily diurnal, whereas the owl hunts during twilight and night. The oriental turtle dove Streptopelia orientalis was also part of this study and served as a control species (Yu et al., 2017).
Prior to the breeding season, artificial nest boxes were placed in the Zuojia Nature Reserve to attract Japanese tits for nesting. Starting in mid-April, nest boxes within the study plots were monitored weekly. Upon detecting signs of nest occupation, the observational frequency was increased to every 2–3 days. Breeding parameters were recorded, including the date of the first-egg laying, their number, clutch size, as well as incubation and hatching times.
Experimental trials with dummy sparrowhawks, owls and doves were performed during nesting periods from May 18 to June 18, 2022, and when nestlings were 5 to 10 days old. Experiments were performed on clear, windless days between 8:00 am and 5:00 pm. The presentation of these dummies was randomized with at least 1-hour intervals between trials. Upon confirming the absence of adult Japanese tits near the nest, dummies were promptly positioned on the nest box lids. The human observer was hiding about 10 to 15 m from the nest and recorded responsive behavior for 5 minutes following the appearance of birds within 10 m of the nest. All trials were recorded using a video-recorder (SONY TD20E, Sony Corporation, Tokyo, Japan) for subsequent behavioral analysis. Dummy birds were maintained in natural standing postures, with two dummies per species randomly selected for trials to minimize posture-related experimental bias. Trials were aborted if adult birds failed to return within 30 minutes post-placement, with subsequent attempts at ~1-hour intervals. All dummy trials were completed in a single day in order to control for meteorological variations.
The researcher assessed the birds' responsive behavior using a 5-point scale with the following scoring criteria: 1 point, the tit ignored the dummy and continued feeding nestlings; 2 points, the tit remained stationary while monitoring the dummy; 3 points, the tit exhibited vigilance behavior, such as hopping or wing-flashing; 4 points, the tit approached the dummy, spread its wings and tail, displayed body swaying (accompanied by hissing calls), and/or made aggressive postures without physical encounter (e.g., hovering), and 5 points, the tit directly engaged using physical aggression toward the dummy. If multiple behaviors were observed during a given trial, the highest recorded score was recorded (Møller et al., 2015; Yu et al., 2017)).
Statistical analysis
The Cumulative Linear Mixed Model (CLMM, clmm in R package ordinal) was used to analyze significant differences in the response behavior scores. The 2-tailed likelihood ratio test was used to obtain P-values. In this model, the response score was set as the dependent variable, whereas the dummy type, brood size, and nestling age were included as fixed independent terms. Nest box number and presentation order were treated as random independent terms. When significant differences were detected in multiple comparisons, post-hoc pairwise comparisons were conducted using the same method, with P-values adjusted by the false discovery rate correction.