Seasonality in song behaviour revisited: Seasonal and annual variants and invariants in the song of the domesticated canary (Serinus canaria)

Abstract

The song of the domesticated canary (Serinus canaria) is one of the most widely used models to study the neural correlates of behavioural plasticity and the mechanisms of female mate choice. However, only few studies have described the song behaviour in detail and monitored their changes throughout the year, and these data are restricted to the “Waterslager” strain. Here, we studied the song characteristics of the male common domesticated canary at different times of the year, the spring breeding and autumnal non-breeding season, and monitored the birds' songs up to the following breeding season. During breeding, males have increased plasma levels of testosterone, and songs are on average longer and consist of fewer non-repeated syllable types compared to the non-breeding season. When subsequent seasons are compared, song duration and the proportion of non-repeated syllable types change seasonally but not across years. Repertoire size remains constant throughout seasons although syllable types are exchanged. Syllable carry-over is significantly higher from one breeding season to the next than between the breeding and non-breeding season. Further, the repertoire of the breeding season contains more potentially sexually attractive syllable types than that of the non-breeding season. These data show that overall song structure is retained throughout the year while seasonality occurs in the temporal pattern and in repertoire composition.

Introduction

Males of many temperate zone species exhibit seasonality in their song behavior, which is correlated with seasonal reproduction (Catchpole and Slater, 2008). Therefore, singing activity is usually higher during the breeding season associated with increased levels of circulating gonadal steroid hormones (Wingfield and Farner, 1993, Ball et al., 2003). Despite song activity other aspects of vocal performance such as song duration and song stereotypy can differ seasonally but their extent is subject to variation between species (Nottebohm et al., 1986, Rost, 1990, Smith et al., 1997, Leitner et al., 2001a, Smulders et al., 2006). For example, European Starlings (Sturnus vulgaris) sing throughout the year, but songs are shorter in autumn than in spring (Riters et al., 2000). Song in male Nuttall's white-crowned sparrows (Zonotrichia leucophrys nuttalli) ceases completely during the non-breeding period while male song sparrows (Melospiza melodia) continue singing in autumn but with greater variability compared to spring (Baker et al., 1984, Smith et al., 1997).

The song behaviour and the neural song control system of songbirds has become an important model to study adult brain plasticity (Tramontin and Brenowitz, 2000, Ball et al., 2003). The initial studies on seasonal changes in song production were performed in the domesticated canary (Serinus canaria). It has been described that the “stable” breeding season song changes into an “unstable” autumnal non-breeding song. This song instability is thought to reflect a phase during which new syllable types are added to the song repertoire while others are lost (Nottebohm et al., 1986, Nottebohm et al., 1987). Furthermore, annual comparisons revealed that repertoire composition changes between years and that its size increases with age (Nottebohm and Nottebohm, 1978, Nottebohm et al., 1986, Nottebohm et al., 1987). These data, however, were collected from an inbred strain of Belgian Waterslager canaries and therefore cannot be generalized for other canary strains. Waterslager canaries, which were especially bred for song, differ in song production and perception from outbred canary strains. They possess a hereditary hearing loss, anatomically related to inner ear abnormalities, at frequencies above 2000 Hz and poor frequency discrimination ability in that region (Okanoya and Dooling, 1985, Gleich et al., 1994, Lauer et al., 2007). Furthermore, Nottebohm and Nottebohm (1976) reported a strong left-side dominance of the syrinx during song production, whereas in common domesticated canaries both sides of the syrinx make important contributions to the song repertoire (Suthers et al., 2004). Although numerous studies concerning the mechanisms of song control have been conducted using canary strains different from the original one (Brenowitz and Arnold, 1992, Hidalgo et al., 1995, Fusani et al., 2000, Appeltants et al., 2001, Singh et al., 2003), the early results on seasonal changes in song have not been confirmed in other strains. A recent study performed on an outbred strain of 1 and 2-year old male common domesticated canaries revealed contradictory results concerning age-related changes of song behavior and neuroanatomy (Leitner and Catchpole, 2004). Furthermore, free-living island canaries (S. canaria), the ancestors of the domesticated canaries, continue to sing “stable” songs, i.e. highly stereotyped syllable types, with constant repertoire size throughout the seasons, and only exchange some syllables in autumn, which are reintegrated in the repertoire of the next breeding season (Leitner et al., 2001a). These subtle changes are not associated with changes in song system anatomy (Leitner et al., 2001b).

Altogether, these findings led us to reconsider the subject of seasonal and annual changes in the song of the domesticated canary. Therefore, the aim of the present study is to revisit the early results of Nottebohm and Nottebohm (1978) and Nottebohm et al., 1986, Nottebohm et al., 1987, obtained from an inbred strain of Waterslager canaries and to test their validity for an outbred strain of common domesticated canaries.