The Ecology of Avian Acoustical Signals
87
Locatability is enhanced if signals are broad-band, frequency modulated, or
abrupt in onsets and interruptions (Marler 1955, 1956; Brown 1982; Klump and
Shalter 1984; Redondo and Arias de Reyna 1988). Conversely, pure tones (they
also should be high-pitched because receivers may detect such signals more easily
than the predator, Klump et a!. 1986) with gradual onset camouflage the sender's
position. Another aspect of locatability is the problem to direct the signal to a
reduced set of receivers over long distance. Directionality is mainly a signaler's
problem and cannot be discussed here for lack of room (see Witkin 1977; Larsen
and Dabelsteen 1990 for references).
The receiver wants to maximize signal detection while reducing the likelihood
of possibly costly false alarms (Richards 1981 aM Wiley and Richards 1982). This
depends very much on the uncertainty about the arrival of the signal, especially
when it is sent rarely. Preambles that are followed by the signal proper at a
predictable interval can solve this problem. Applied to bird songs, the introductory
notes should be easily detected and consist of pure tones, lack species-specific
characteristics, and they would be expected to precede signals rich in complex
notes with high repetition rates that are very species-specific (Richards 1981 a).
Since motivation affects discrimination as well (Johnsrude et a!. 1994), the
preamble's response eliciting power should be invariant across motivational
contexts.
The relative position of signaler and receiver with respect to each other and with
respect to features in the environment that are significant for sound propagation
has received some attention in the literature (Hunter 1980; Lemon et a!. 1981;
Wiley and Richards 1982; Wilczynski et al. 1989), but almost always treated from
the signaler's perspective. Others (Dabelsteen et al. 1993; Mathevon et al. 1996),
however, discuss the possibility that it might be more important for the signaler to
optimize his receiving rather than his broadcasting range. It is most important to
get high above the ground. Figures 1 and 2 present theoretical calculations based
on the models and algorithms described in Chessell ( 1977), Price et a!. ( 1988),
Pirinchieva (1991 ), Wempen ( 1991 ), and Sutherland and Daigle (1997). Figure I
might be relevant for antpittas (Sect. 3.2) that sing and receive on the ground and
thrushes that forage on the ground (receivers) and sing in middle and upper tiers of
the forest.
These calculations did not include the attenuation of higher frequencies through
vegetation (Price et a!. 1988). Moreover, with increasing signaler height the
portion of the sound rays that travel through the understory is reduced. Figure 2
represents the situation for a skylark for which the receiver also is assumed to be
at ground level. Here, I also included the effects of turbulence, following
Chessell's (1977) approach. The most significant result of these model calculations
is that singing high up during turbulent conditions, as would often occur in
grasslands (but not in forests), may be advantageous.
The above design principles are subject to constraints by the other demands on
the signals (Lambrechts 1996). The number of different classes of individuals
(sex, age, kin, individual, species) which should be identified by the receiver, the
87
Locatability is enhanced if signals are broad-band, frequency modulated, or
abrupt in onsets and interruptions (Marler 1955, 1956; Brown 1982; Klump and
Shalter 1984; Redondo and Arias de Reyna 1988). Conversely, pure tones (they
also should be high-pitched because receivers may detect such signals more easily
than the predator, Klump et a!. 1986) with gradual onset camouflage the sender's
position. Another aspect of locatability is the problem to direct the signal to a
reduced set of receivers over long distance. Directionality is mainly a signaler's
problem and cannot be discussed here for lack of room (see Witkin 1977; Larsen
and Dabelsteen 1990 for references).
The receiver wants to maximize signal detection while reducing the likelihood
of possibly costly false alarms (Richards 1981 aM Wiley and Richards 1982). This
depends very much on the uncertainty about the arrival of the signal, especially
when it is sent rarely. Preambles that are followed by the signal proper at a
predictable interval can solve this problem. Applied to bird songs, the introductory
notes should be easily detected and consist of pure tones, lack species-specific
characteristics, and they would be expected to precede signals rich in complex
notes with high repetition rates that are very species-specific (Richards 1981 a).
Since motivation affects discrimination as well (Johnsrude et a!. 1994), the
preamble's response eliciting power should be invariant across motivational
contexts.
The relative position of signaler and receiver with respect to each other and with
respect to features in the environment that are significant for sound propagation
has received some attention in the literature (Hunter 1980; Lemon et a!. 1981;
Wiley and Richards 1982; Wilczynski et al. 1989), but almost always treated from
the signaler's perspective. Others (Dabelsteen et al. 1993; Mathevon et al. 1996),
however, discuss the possibility that it might be more important for the signaler to
optimize his receiving rather than his broadcasting range. It is most important to
get high above the ground. Figures 1 and 2 present theoretical calculations based
on the models and algorithms described in Chessell ( 1977), Price et a!. ( 1988),
Pirinchieva (1991 ), Wempen ( 1991 ), and Sutherland and Daigle (1997). Figure I
might be relevant for antpittas (Sect. 3.2) that sing and receive on the ground and
thrushes that forage on the ground (receivers) and sing in middle and upper tiers of
the forest.
These calculations did not include the attenuation of higher frequencies through
vegetation (Price et a!. 1988). Moreover, with increasing signaler height the
portion of the sound rays that travel through the understory is reduced. Figure 2
represents the situation for a skylark for which the receiver also is assumed to be
at ground level. Here, I also included the effects of turbulence, following
Chessell's (1977) approach. The most significant result of these model calculations
is that singing high up during turbulent conditions, as would often occur in
grasslands (but not in forests), may be advantageous.
The above design principles are subject to constraints by the other demands on
the signals (Lambrechts 1996). The number of different classes of individuals
(sex, age, kin, individual, species) which should be identified by the receiver, the