3.2.1. Community Studies
One of the first and most often cited tests of the acoustic adaptation hypothesis was performed by Eugene Morton, who compared the songs of birds
that reside in open and forested habitats in Panama (Morton 1975). Morton
first determined the amount of excess attenuation experienced by tones of
varying frequency transmitted at different heights within different environments. In all locations, higher frequencies generally experienced greater
attenuation than lower frequencies. However, near the ground in forest
there was a “sound window” of low excess attenuation between 1,500 and
2,500 Hz. This sound window did not exist at higher heights within forest or
in edge and grassland environments. Morton then compared the results of
these transmission studies with data on the song frequencies of different
species of birds living in these habitats. He found that birds that call in low
forest, but not birds that call in grassland or above the ground in forest, have
mean call frequencies in the range of the sound window. Morton concluded
that the songs of these forest species have evolved in response to selection
for decreased attenuation and thus increased transmission distance.
Since Morton’s original study, a number of similar comparative studies
have been conducted using the calls of birds and anurans. These studies
evaluated a number of different aspects of temporal and spectral fidelity in
addition to attenuation and frequency-dependent attenuation. Although
there is some disparity in the details of the analyses, community-level
studies in temperate and tropical birds generally support Morton’s findings,
at least for some signal characteristics (e.g., Richards and Wiley 1980;
Sorjonen 1986), but community-level studies in frogs do not (Zimmerman
1983; Penna and Solis 1996; Kime et al. 2000).
The interpretation of multispecies comparisons such as Morton’s can,
however, be easily confounded by other determinants of call structure such
as phylogenetic relationship or body size. As Ryan and Brenowitz (1985)
pointed out, the frequency differences that Morton found in his original
study could also be explained by differences in the body size of birds living
in the different locations or by differences in the background-noise composition of the different environments. Correcting for body size in temperate birds, Wiley (1991) failed to find differences in dominant frequency
among the songs of temperate-zone oscine birds in open and forested environments. For frogs, Zimmerman (1983) showed that the phylogenetic relationships among species in a community of tropical frogs were a better
predictor of signal structure than was habitat acoustics.
3.2.2. Studies of Single or Closely Related Species
Some of the more convincing tests of the acoustic adaptation hypothesis
compare signal structure among populations of a single species or among
closely related species with habitat shifts. At this level, it is often easier
to control for the effects of body size and other morphological or
5. Selection on Signals
235
One of the first and most often cited tests of the acoustic adaptation hypothesis was performed by Eugene Morton, who compared the songs of birds
that reside in open and forested habitats in Panama (Morton 1975). Morton
first determined the amount of excess attenuation experienced by tones of
varying frequency transmitted at different heights within different environments. In all locations, higher frequencies generally experienced greater
attenuation than lower frequencies. However, near the ground in forest
there was a “sound window” of low excess attenuation between 1,500 and
2,500 Hz. This sound window did not exist at higher heights within forest or
in edge and grassland environments. Morton then compared the results of
these transmission studies with data on the song frequencies of different
species of birds living in these habitats. He found that birds that call in low
forest, but not birds that call in grassland or above the ground in forest, have
mean call frequencies in the range of the sound window. Morton concluded
that the songs of these forest species have evolved in response to selection
for decreased attenuation and thus increased transmission distance.
Since Morton’s original study, a number of similar comparative studies
have been conducted using the calls of birds and anurans. These studies
evaluated a number of different aspects of temporal and spectral fidelity in
addition to attenuation and frequency-dependent attenuation. Although
there is some disparity in the details of the analyses, community-level
studies in temperate and tropical birds generally support Morton’s findings,
at least for some signal characteristics (e.g., Richards and Wiley 1980;
Sorjonen 1986), but community-level studies in frogs do not (Zimmerman
1983; Penna and Solis 1996; Kime et al. 2000).
The interpretation of multispecies comparisons such as Morton’s can,
however, be easily confounded by other determinants of call structure such
as phylogenetic relationship or body size. As Ryan and Brenowitz (1985)
pointed out, the frequency differences that Morton found in his original
study could also be explained by differences in the body size of birds living
in the different locations or by differences in the background-noise composition of the different environments. Correcting for body size in temperate birds, Wiley (1991) failed to find differences in dominant frequency
among the songs of temperate-zone oscine birds in open and forested environments. For frogs, Zimmerman (1983) showed that the phylogenetic relationships among species in a community of tropical frogs were a better
predictor of signal structure than was habitat acoustics.
3.2.2. Studies of Single or Closely Related Species
Some of the more convincing tests of the acoustic adaptation hypothesis
compare signal structure among populations of a single species or among
closely related species with habitat shifts. At this level, it is often easier
to control for the effects of body size and other morphological or
5. Selection on Signals
235