The variability of song variability in zebra finch (Taeniopygia guttata) populations

Abstract

Birdsong is a classic example of a learned social behaviour. Song behaviour is also influenced by genetic factors, and understanding the relative contributions of genetic and environmental influences remains a major goal. In this study, we take advantage of captive zebra finch populations to examine variation in a population-level song trait: song variability. Song variability is of particular interest in the context of individual recognition and in terms of the neuro-developmental mechanisms that generate song novelty. We find that the Australian zebra finch Taeniopygia guttata castanotis (TGC) maintains higher song diversity than the Timor zebra finch T. g. guttata (TGG) even after experimentally controlling for early life song exposure, suggesting a genetic basis to this trait. Although wild-derived TGC were intermediate in song variability between domesticated TGC populations and TGG, the difference between domesticated and wild TGC was not statistically significant. The observed variation in song behaviour among zebra finch populations represents a largely untapped opportunity for exploring the mechanisms of social behaviour.

Keywords: song behaviour; song variability; zebra finch.

Figure 1.

(a) Fitted means and confidence intervals for population estimates of K-L distance for principal component 1 for unmanipulated T. guttata castanotis (TGC) and T. guttata guttata (TGG) sampled at the University of Illinois (UIUC) and East Carolina University (ECU) (χ² = 11.1, d.f. = 3, p = 0.01). (b) Fitted means and confidence intervals for population estimates of K-L distance for principal component 1 within populations of TGG and TGC after experimental tutoring by Bengalese finches (χ² = 14.1, d.f. = 1, p = 0.0001). (c) Scatterplots of song repertoires of cross-fostered individuals and tutors. Distinct clusters reflect unique song syllables. Each of the five scatter plots represents one Bengalese tutor (black points) and its respective TGC (shades of blue) and TGG tutees (shades of orange).

Figure 2.

(a) Scatterplots of song repertoires of four individuals per population (each colour represents a unique individual). Distinct clusters reflect unique song syllables. Three TGC populations are depicted: two domesticated TGC populations (one from Chicago and one from ECU), and one wild population of TGC from Macquarie. One TGG population is depicted from ECU (labelled ‘Timor’ on graph). (b) Fitted means and confidence intervals for population estimates of K-L distance for principal component 1 for each of the four populations described above reveals population differences in song variability (χ² = 11.2, d.f. = 3, p = 0.01).