3.3.1. Field Studies
The function of male song in maintaining and establishing songbird territories is well-known. For example, removing a male songbird’s ability to
sing has dramatic effects on his success at holding a territory (Peek 1972,
Smith and Reid 1979; McDonald 1989), and simply broadcasting a conspecific song from an unoccupied territory leads to significantly lower rates of
settlement in that territory than in others’ territories where control sounds
or no sounds are played (Krebs et al. 1978; Yasukawa 1981; Falls 1987).
Moreover, territory residents often respond weakly, or not at all, to a neighbor singing from a familiar location but more strongly to a stranger singing
from that same location. Using these facts along with a variety of clever
song-playback techniques in the field, a very large number of studies have
demonstrated that males in many (at least 23) songbird species are capable
of discriminating among neighbors and strangers on the basis of song alone.
Furthermore, for several (at least 8) species, listeners are capable of recognizing individual singers on the basis of their songs (see Stoddard 1996).
3.3.2. Signal Variation
There are several ways that singer identity could be represented in the
acoustic variation of male song. In the simplest case, individual males might
sing a unique song or sets of songs (i.e., repertoires), and recognition would
follow by the association of specific songs with specific singers. This strategy appears to be used by song sparrows, Melospiza melodia (Beecher
et al. 1994), and European starlings (Gentner et al. 2000). One obvious
feature of this strategy, potentially worth exploring, is that it may be heavily
constrained by the memory capacity of the recognition system. Although
this question deserves further attention, initial results suggest that the
capacity of these systems is, in fact, quite high (see Gentner et al. 2000) and,
at least for song sparrows, exceeds the number of exemplars that an individual is likely to face at a single time in the wild (Stoddard et al. 1992).
A second recognition strategy employed relies on morphological differences in the acoustics of shared song types. In both white-throated sparrows and field sparrows, the songs of neighboring territorial males share
several acoustic features but vary slightly in frequency. Neighbors rely on
these subtle frequency differences to recognize one another (Brooks and
Falls 1975; Nelson 1989). Related to this is a third possible strategy for
recognition. If the morphology of the vocal-production apparatus varies
slightly between individuals, then this variation might impart unique spectral features, or so-called “voice characteristics,” to all of an individual’s
vocalizations. The use of voice characteristics has been suggested for great
tits, Parus major (Weary and Krebs 1992), but does not appear to be a
relevant cue for either song sparrows (Beecher et al. 1994) or starlings
(Gentner et al. 2000). Finally, vocal recognition might also rely on the
7. Neuroethology of Vocal Communication
357
The function of male song in maintaining and establishing songbird territories is well-known. For example, removing a male songbird’s ability to
sing has dramatic effects on his success at holding a territory (Peek 1972,
Smith and Reid 1979; McDonald 1989), and simply broadcasting a conspecific song from an unoccupied territory leads to significantly lower rates of
settlement in that territory than in others’ territories where control sounds
or no sounds are played (Krebs et al. 1978; Yasukawa 1981; Falls 1987).
Moreover, territory residents often respond weakly, or not at all, to a neighbor singing from a familiar location but more strongly to a stranger singing
from that same location. Using these facts along with a variety of clever
song-playback techniques in the field, a very large number of studies have
demonstrated that males in many (at least 23) songbird species are capable
of discriminating among neighbors and strangers on the basis of song alone.
Furthermore, for several (at least 8) species, listeners are capable of recognizing individual singers on the basis of their songs (see Stoddard 1996).
3.3.2. Signal Variation
There are several ways that singer identity could be represented in the
acoustic variation of male song. In the simplest case, individual males might
sing a unique song or sets of songs (i.e., repertoires), and recognition would
follow by the association of specific songs with specific singers. This strategy appears to be used by song sparrows, Melospiza melodia (Beecher
et al. 1994), and European starlings (Gentner et al. 2000). One obvious
feature of this strategy, potentially worth exploring, is that it may be heavily
constrained by the memory capacity of the recognition system. Although
this question deserves further attention, initial results suggest that the
capacity of these systems is, in fact, quite high (see Gentner et al. 2000) and,
at least for song sparrows, exceeds the number of exemplars that an individual is likely to face at a single time in the wild (Stoddard et al. 1992).
A second recognition strategy employed relies on morphological differences in the acoustics of shared song types. In both white-throated sparrows and field sparrows, the songs of neighboring territorial males share
several acoustic features but vary slightly in frequency. Neighbors rely on
these subtle frequency differences to recognize one another (Brooks and
Falls 1975; Nelson 1989). Related to this is a third possible strategy for
recognition. If the morphology of the vocal-production apparatus varies
slightly between individuals, then this variation might impart unique spectral features, or so-called “voice characteristics,” to all of an individual’s
vocalizations. The use of voice characteristics has been suggested for great
tits, Parus major (Weary and Krebs 1992), but does not appear to be a
relevant cue for either song sparrows (Beecher et al. 1994) or starlings
(Gentner et al. 2000). Finally, vocal recognition might also rely on the
7. Neuroethology of Vocal Communication
357