cartilages respond to increased circulating testosterone with a profound
growth spurt (Tuohimaa et al. 1981; Beckford et al. 1985). The result is a
typical f 0 for adult males that is about half that of adult females, despite an
average difference in body weights of only 20% (Hollien 1960). As mentioned in Section 2.2.2, hypertrophy of the male larynx, out of all proportion to body size, is carried to an absurd extreme in animals such as the
howler monkey (Allouatta seniculus, Schön 1971) and the hammerhead bat
(Hypsignathus monstrosus), which clearly illustrate that larynx size, within
broad limits, is unconstrained by body size. Although much less is known
about the relationship between body size and syrinx size in birds, it seems
likely that similar considerations apply. The syrinx, like the larynx, is free
from any skeletal constraints on its size and would be expected to respond
freely to selection for low voices. For example, both cranes and currassows
are groups with unusually large syringes (Delacour and Amadon 1973;
Johnsgard 1983; Fitch 1999), and both groups are typified by low-pitched,
loud voices. In contrast, other groups, such as Falconiformes (e.g., hawks,
eagles), have unusually high-pitched voices for their size. These observations suggest that the syrinx is not under any strong size constraints and can
respond to selection by either increasing or decreasing size.
When such developmental flexibility is present, there is clearly no a priori
reason to expect vocal fold size (and thus f 0 ) to be well-correlated with body
size (Fitch 1994; Fitch and Hauser 1995). Of course, between disparate
enough taxa some degree of correlation is inevitable simply due to the very
large differences in overall avian body sizes; the syrinx of an ostrich or emu
could contain the entire body of a hummingbird. Thus, various researchers
have found correlations between body size and some measure of vocal
frequency across different avian or mammalian taxa (birds: Ryan and
Brenowitz 1985; mammals: August and Anderson 1987; Hauser 1993).
Similarly, in species with large size differences between infants and adults,
we may expect some differences in pitch between young and old animals,
as indeed appears to be the case in humans, where the f 0 of infant cries averages around 500 Hz and adult speech between 100 and 200 Hz (Titze 1994).
However, the relevant information for many species in many communicative situations is not the size of young or of members of other species but
of conspecific adults. In this domain, and despite the common claim that
voice pitch provides an accurate cue to body size (e.g., Morton 1977), the
data reviewed above suggest that the voice source (larynx or syrinx) is
ill-suited to provide dependable cues to body size in adult terrestrial
vertebrates.
2.3.6. Vocal Tract Length and Acoustic Cues to Body Size
A different potential acoustic cue to body size comes from vocal tract length
and formant frequencies. If the cross-sectional area function of the vocal
tract is constant, the primary determinant of formant frequencies is the
102
W.T. Fitch and M.D. Hauser
growth spurt (Tuohimaa et al. 1981; Beckford et al. 1985). The result is a
typical f 0 for adult males that is about half that of adult females, despite an
average difference in body weights of only 20% (Hollien 1960). As mentioned in Section 2.2.2, hypertrophy of the male larynx, out of all proportion to body size, is carried to an absurd extreme in animals such as the
howler monkey (Allouatta seniculus, Schön 1971) and the hammerhead bat
(Hypsignathus monstrosus), which clearly illustrate that larynx size, within
broad limits, is unconstrained by body size. Although much less is known
about the relationship between body size and syrinx size in birds, it seems
likely that similar considerations apply. The syrinx, like the larynx, is free
from any skeletal constraints on its size and would be expected to respond
freely to selection for low voices. For example, both cranes and currassows
are groups with unusually large syringes (Delacour and Amadon 1973;
Johnsgard 1983; Fitch 1999), and both groups are typified by low-pitched,
loud voices. In contrast, other groups, such as Falconiformes (e.g., hawks,
eagles), have unusually high-pitched voices for their size. These observations suggest that the syrinx is not under any strong size constraints and can
respond to selection by either increasing or decreasing size.
When such developmental flexibility is present, there is clearly no a priori
reason to expect vocal fold size (and thus f 0 ) to be well-correlated with body
size (Fitch 1994; Fitch and Hauser 1995). Of course, between disparate
enough taxa some degree of correlation is inevitable simply due to the very
large differences in overall avian body sizes; the syrinx of an ostrich or emu
could contain the entire body of a hummingbird. Thus, various researchers
have found correlations between body size and some measure of vocal
frequency across different avian or mammalian taxa (birds: Ryan and
Brenowitz 1985; mammals: August and Anderson 1987; Hauser 1993).
Similarly, in species with large size differences between infants and adults,
we may expect some differences in pitch between young and old animals,
as indeed appears to be the case in humans, where the f 0 of infant cries averages around 500 Hz and adult speech between 100 and 200 Hz (Titze 1994).
However, the relevant information for many species in many communicative situations is not the size of young or of members of other species but
of conspecific adults. In this domain, and despite the common claim that
voice pitch provides an accurate cue to body size (e.g., Morton 1977), the
data reviewed above suggest that the voice source (larynx or syrinx) is
ill-suited to provide dependable cues to body size in adult terrestrial
vertebrates.
2.3.6. Vocal Tract Length and Acoustic Cues to Body Size
A different potential acoustic cue to body size comes from vocal tract length
and formant frequencies. If the cross-sectional area function of the vocal
tract is constant, the primary determinant of formant frequencies is the
102
W.T. Fitch and M.D. Hauser