Data from: The risk perception and response of Azure-winged magpies: On the aspect of feeding behaviour and alarm calls
Data files
Apr 02, 2025 version files 37.80 KB
-
raw_data.xlsx
31.45 KB
-
README.md
6.35 KB
Abstract
Prey are expected to effectively perceive predation cues, recognise predators, and adopt appropriate anti-predator strategies to enhance their chances of survival. Species with high cognitive abilities tend to be better at these processes, while empirical research is still lacking. The role of cognition in avoiding predation requires further investigation, especially in species with complex social structures and communication systems. Azure-winged magpies (Cyanopica cyanus) have demonstrated great talents in cognitive tasks. We conducted model presentation and playback experiments to test their predator detection and information transmission in the wild. We found that magpies exhibited distinct responses according to the model type (pigeon, falco,n, and cat) and eye condition of the models (covered or uncovered). Individuals postponed the visit to the feeder and took less food in response to predator and eye-uncovered models. The cat model was perceived as a higher risk, and magpies would emit alarm calls with a wider bandwidth, a higher frequency of 5%, and a higher frequency of 95%. We also found that the playback of alarm calls could induce different anti-predator behaviours from conspecifics. Our study examined how Azure-winged magpies perceive information and make decisions to avoid predators. It indicates that the alarm calls of Azure-winged magpies function in encoding and sharing information, providing an in-depth understanding of complex vocal communication and risk cognition in birds.
Dataset DOI: 10.5061/dryad.f1vhhmh73
Description of the data and file structure
This dataset tests the hypothesis that azure-winged magpies discriminate between predator types and eye gaze, using alarm calls with distinct acoustic structures to convey predator-specific information.
During the non-breeding season, five free-ranging magpie flocks were subjected to predator presentation and playback experiments. The predator presentation experiment followed a 3 × 2 factorial design (“model type” × “eye condition”), where model types included a domestic pigeon, a free-ranging cat, and a peregrine falcon. Each model was presented with eyes either covered (placed over the eyes) or uncovered (placed below the eyes). Mobbing calls elicited by the eye-uncovered cat and falcon models were later used in playback experiments.
Behavioral responses—latency, number of visits, and pieces of food taken—were recorded from video footage. Vocalizations were analyzed using Raven Pro 1.6, with spectrograms generated using a Hann window (512-sample DFT, 50% overlap, 0.0058 s time resolution, and 86.10 Hz frequency resolution). Five acoustic parameters were measured: Bandwidth 90%, Frequency 5%, Frequency 95%, peak frequency, and call duration.
Generalized linear mixed models (GLMMs) were used to analyze feeding behavior. Latency was modeled with a Gamma distribution, while the number of visits and pieces of food taken followed a Poisson distribution. Replicate and flock identity were included as random effects to control for habituation, with baseline measures treated as covariates. For predator presentations, fixed factors included model type, eye condition and their interaction. The mobbing call target (cat or falcon) was the fixed factor for playbacks . A posthoc pairwise comparison test with Bonferroni correction was conducted in each model to assess differences between all levels of the fixed factors and the interactions. Acoustic parameters (log-transformed) were analyzed using Gaussian GLMMs, with model type, eye condition, and their interaction as fixed effects. Results demonstrated that azure-winged magpies effectively perceive risk and the playback of alarm calls could induce different anti-predator behaviours from conspecifics.
Files and variables
File: raw_data.xlsx
Description:
This is a XLSX file containing raw data of article “The risk perception and response of Azure-winged magpies: on the aspect of feeding behaviour and alarm calls”, including three tables. The Table 1 is the raw data of model presentation experiments. The Table 2 is the raw data of acoustic parameters of mobbing calls. The Table 3 is the raw data of playback experiments.
Variables in sheet 1: Model presentation experiments
Number of variables: 10
Variable list:
-
Flock identity: five distinct flocks with a minimum distance of 500 m from the others ensuring that vocal signals from one flock could not reach others.
-
Replicate: each trial was replicated three times at every site.
-
Model type: “1”- the Pigeon model, “2”- the Falcon model, “3”- the Cat model.
-
Eye condition: “1”- the uncovered eyes, “2”- the covered eyes.
-
Latency in stimulus: seconds it takes magpies to visit the feeder since the start of stimulus period.
-
Number of visits in stimulus: the total number of visits of the focal flock within the 10-minute stimulus trail.
-
Pieces of food taken in stimulus: quantifying the amount of food taken by the birds within the 10-minute stimulus trial.
-
Latency in baseline: seconds it takes magpies to visit the feeder since the start of the baseline period.
-
Number of visits in baseline: the total number of visits of the focal flock within the 10-minute baseline trail.
-
Pieces of food taken in baseline: quantifying the amount of food taken by the birds within the 10-minute baseline trail.
Variables in sheet 2: Acoustic parameters
Number of variables: 8
Variable list:
-
Flock identity: five distinct flocks with a minimum distance of 500 m from the others ensuring that vocal signals from one flock could not reach others.
-
Model type: “1”- the Pigeon model, “2”- the Falcon model, “3”- the Cat model.
-
Eye condition: “1”- the uncovered eyes, “2”- the covered eyes.
-
BW 90% (Hz): Bandwidth 90%, the difference between the 5% and 95% frequencies.
-
Freq 5% (Hz): Frequency 5%, the frequency that divides the spectrum into two intervals which contain 5% and 95% of the energy.
-
Freq 95% (Hz): Frequency 95%, the frequency that divides the spectrum into two intervals which contain 95% and 5% of the energy.
-
Peak Freq (Hz): Peak frequency, the frequency at which peak power occurs; (5) Duration, the duration of the vocalisation.
-
Duration (s): the duration of the vocalisation.
Variables in sheet 3: Playback experiments
Number of variables: 9
Variable list:
-
Flock identity: five distinct flocks with a minimum distance of 500 m from the others ensuring that vocal signals from one flock could not reach others.
-
Replicate: each trial was replicated three times at every site.
-
Model type: “2”- the Falcon model, “3”- the Cat model.
-
Latency in stimulus: seconds it takes magpies to visit the feeder since the start of stimulus period.
-
Number of visits in stimulus: the total number of visits of the focal flock within the 10-minute stimulus trail.
-
Pieces of food taken in stimulus: quantifying the amount of food taken by the birds within the 10-minute stimulus trial.
-
Latency in baseline: seconds it takes magpies to visit the feeder since the start of the baseline period.
-
Number of visits in baseline: the total number of visits of the focal flock within the 10-minute baseline trail.
-
Pieces of food taken in baseline: quantifying the amount of food taken by the birds within the 10-minute baseline trail.
Code/software
The vocalisations of Azure-winged magpies were coded from the spectrograms of audio files using Raven Pro 1.6 (Cornell Laboratory of Ornithology, USA).
All the analyses were performed using SPSS 26.0 (IBM Corp., Armonk, NY, U.S.A.) and Origin 9.0 (OriginLab Corp., Northampton, MA, U.S.A.) was used for plotting.