Neighbour-stranger discrimination in an African wood dove inhabiting equatorial rainforest

Abstract

We investigated within- and between-individual song variation and song-based neighbour-stranger discrimination in a non-learning bird species, the blue-headed wood-dove (Turtur brehmeri), which inhabits lowland rainforests of West and Central Africa. We found that songs of this species are individually specific and have a high potential for use in individual recognition based on the time-frequency pattern of note distribution within song phrases. To test whether these differences affect behaviour, we conducted playback experiments with 19 territorial males. Each male was tested twice, once with the songs of a familiar neighbour and once with the songs of an unfamiliar stranger. We observed that males responded more aggressively to playback of a stranger's songs: they quickly approached close to the speaker and spent more time near it. However, no significant differences between treatments were observed in the vocal responses. In addition, we explored whether responses differed based on the song frequency of the focal male and/or that of the simulated intruder (i.e., playback), as this song parameter is inversely related to body size and could potentially affect males' decisions to respond to other birds. Song frequency parameters (of either the focal male or the simulated intruder) had no effect on the approaching response during playback. However, we found that the pattern of response after playback was significantly affected by the song frequency of the focal male: males with lower-frequency songs stayed closer to the simulated intruder for a longer period of time without singing, while males with higher-frequency songs returned more quickly to their initial song posts and resumed singing. Together, these results depict a consistently strong response to strangers during and after playback that is dependent on a male's self-assessment rather than assessment of a rival's strength based on his song frequency. This work provides the first experimental evidence that doves (Columbidae) can use songs for neighbour-stranger discrimination and respond according to a "dear enemy" scheme that keeps the cost of territory defence at a reasonable level.

Keywords: Broadcast vocalisation; Columbidae vocalisation; Dear enemy phenomenon; Neighbour–stranger discrimination; Playback experiment; Song.

Figure 1

Spectrogram illustrating the song of the blue-headed wood-dove. Letters indicate: N₁, N₂—durations of following notes; P₁, P₂—pauses between following notes, based on manually selected beginning and end of each note; PF₃, PF₄—peak frequency of notes measured with the One-dimensional transformation function of Avisoft SASLab Pro; PTP_3–4—pauses between syllables measured as time duration between peak frequency of following notes. The figure shows example measurements that were taken for all notes and pauses within each song.

Figure 2

Percentage of correct classification of the blue-headed wood-dove song based on discriminant function analysis (DFA) with note peak frequency (PF) and pulse-to-pulse duration (PTP). Numbers on X-axis indicate sequences analysed with DFAs: 1—PF₁ + PTP_1–2, 2—PF₁ + PF₂ + PTP_1–2 + PTP_2–3 and so on. H_s—Beecher's statistics for the three chosen sequences of PF + PTP. The red dotted lines indicate the data points chosen to showcase the values.

Figure 3

The blue-headed wood-dove male responses to playback of neighbour and stranger songs measured with PC1-approaching and PC2-after playback response compound response measurements (N = 19). The dotted lines represent the same tested male responding to neighbour and stranger songs.

Figure 4

Relationships (linear regression models) between response to playbacks (measured with PC1-approaching and PC2-after playback response) and peak frequencies (PF) of focal male and playback songs. Different colours indicate neighbour and stranger treatments (N = 19 males tested twice).