Predator type shapes call-type composition and post-threat vocal dynamics in Japanese quail (Coturnix japonica): a pilot case-series

Abstract

Japanese quail (Coturnix japonica) rely on an extensive vocal repertoire to coordinate social behaviour and respond to predation risk, yet how individual call types are modulated by predator identity remains poorly characterized. In this pilot case-series, two sexually mature Japanese quail (each contributing more than 2,000 classified vocalizations across repeated predator-exposure trials) were exposed to simulated aerial (hawk) and terrestrial (cat) predator stimuli across four social contexts, before, during, and after stimulus presentation. Vocalizations were classified, via blind dual-observer review, into 18 spectrographically distinct call types (an exploratory repertoire estimate, not independently statistically validated), of which 10 were acoustically characterized. Chi-square analyses of the full call-type composition revealed highly significant differences between hawk and cat exposure in both individuals (both p < 0.0001); call types 5 and 10 were significantly cat-biased and call type 2 was significantly hawk-biased in both individuals. An exploratory correlation suggested that call types with a higher maximum frequency tended to be relatively hawk-biased (Spearman rho = -0.67, p = 0.033, uncorrected for multiple comparisons). Independently, call-type composition shifted significantly between the pre- and post-exposure phases in all four individual-by-predator combinations (all p < 0.0001), driven substantially by a consistent post-exposure increase in call type 9 across both individuals and both predator classes. Additional exploratory analyses of social-context (mirror and concealed-conspecific) effects on call-type composition are presented in the supporting information. Given the small sample size (n = 2), these findings are presented as case-based, hypothesis-generating observations intended to justify and inform a fully powered follow-up study, rather than as population-level conclusions.