Songbirds tune their vocal tract to the fundamental frequency of their song
Abstract
In human speech, the sound generated by the larynx is modified by articulatory movements of the upper vocal tract, which acts as a variable resonant filter concentrating energy near particular frequencies, or formants, essential in speech recognition. Despite its potential importance in vocal communication, little is known about the presence of tunable vocal tract filters in other vertebrates. The tonal quality of much birdsong, in which upper harmonics have relatively little energy, depends on filtering of the vocal source, but the nature of this filter is controversial. Current hypotheses treat the songbird vocal tract as a rigid tube with a resonance that is modulated by the end-correction of a variable beak opening. Through x-ray cinematography of singing birds, we show that birdsong is accompanied by cyclical movements of the hyoid skeleton and changes in the diameter of the cranial end of the esophagus that maintain an inverse relationship between the volume of the oropharyngeal cavity and esophagus and the song’s fundamental frequency. A computational acoustic model indicates that this song-related motor pattern tunes the major resonance of the oropharyngeal–esophageal cavity to actively track the song’s fundamental frequency.
Sign up for PNAS alerts.
Get alerts for new articles, or get an alert when an article is cited.
Although the pulsating throat of a singing bird is apparent to even the casual observer, its importance in vocalization is unknown. Sound is generated deep in the thorax where the avian vocal organ, the syrinx, is located at the junction between the bronchi and trachea. This syringeal sound is modified by the filter properties of the suprasyringeal vocal tract before it radiates from the beak into the environment as song or call notes, but the anatomical basis of these vocal tract filters and their acoustic properties are not well understood.
The modulation of birdsong by vocal tract resonance was first demonstrated by Nowicki (1), who recorded nine species of songbirds singing in a light gas atmosphere (He–O2). Because the speed of sound in this gas mixture is nearly twice that in air, resonance frequencies of the vocal tract are shifted upward almost an octave, increasing the level of the second harmonic relative to that of the fundamental frequency (f0). In addition to providing evidence that the source-filter theory of human speech (2) also applies to birdsong, Nowicki (1) observed that second harmonics, which were apparent in light gas, were suppressed in air regardless of their absolute frequency. This finding indicated that the tuning of the resonance filter is not fixed, but can be adjusted to selectively support different fundamental frequencies.
The search for a tracking resonance filter in the songbird’s vocal tract has focused on song-related beak movements. Various investigators (e.g., refs. 3–6) have shown that birdsong is often accompanied by frequency-dependent beak movements in which high fundamental frequencies are associated with a wide beak opening, or gape, compared to that at low fundamentals. The fact that experimentally fixing the gape of a singing bird at a large value increases the relative level of the second harmonic when f0 is low (refs. 4 and 7, and R.A.S., F. Goller, R. Bermejo, M. Wild, and P. Zeigler, unpublished data) also implicates the beak as part of the vocal tract filter. It is hypothesized that by varying beak gape a bird alters the end-correction that must be applied to its vocal tract, considered as a rigid stopped tube (the trachea) that is flared at its open end by the beak. A wide beak gape should reduce the effective length of the vocal tract and raise its resonance frequency.
However, some data are inconsistent with the beak gape tracking filter hypothesis. The vocal tract filter in young songbirds learning to sing seems to develop before the appearance of frequency-related beak movements. When juvenile song sparrows (Melospiza melodia) begin to sing, their songs initially contain prominent higher harmonics. These harmonics become less prominent and song acquires its adult-like tonality between the mid and late phases of plastic song when the maximum beak gape is still less than ≈1 mm and is not coordinated with f0. Beak movements do not become coordinated with f0 until song becomes fixed in its adult form when the song sparrows are ≈400 days old, well after tonality appears (8). During late plastic phase of song development, these sparrows must be using something other than beak gape to suppress higher harmonics. In addition, theoretical models (9) suggest the acoustic effect of the maxilla and mandible per se is relatively small and limited to high fundamental frequencies.
In other experiments, in which the syrinx of eastern towhees (Pipilo erythrophthalmus) was replaced by a calibrated sound source, Nelson et al. (10) found that changes in beak gape had no significant effect on the vocal tract resonance for frequencies below ≈4 or 5 kHz, but frequencies between ≈4 and 7.5 kHz were attenuated by a small gape. These data suggest that beak movements control the attenuation, but not the tuning, of a vocal tract filter that suppresses higher harmonics when gape is small. The data do not support the hypothesis that beak movements cause the vocal tract resonance to track the f0. The concept of beak gape as controlling either a band-pass or a low-pass filter that attenuates high frequencies is consistent with Larsen and Dabelsteen’s (11) data from the European blackbird (Turdus merula) and with data from cardinals indicating that a large gape increases the level of higher harmonics for fundamental frequencies below ≈3 or 4 kHz (7).
In the following experiments, we use cineradiography to analyze changes in the configuration of the upper vocal tract between the glottis and the beak in spontaneously singing adult male northern cardinals (Cardinalis cardinalis) and a computational model (12) to assess the acoustic significance of these changes. We describe a resonance filter in the avian vocal tract, involving cyclical song-related changes in the volume of the oropharynx and cranial esophagus that track the f0 and can significantly influence the acoustic properties of vocalizations.
Results
Cardinal song is characterized by frequency modulated (FM) sweeps in which the f0 may be varied by as much as 2 or 3 octaves between ≈l and 9 kHz. Representative data for three of these syllable types, including one upward sweeping syllable (syllable type 1) and two downward sweeping syllables (types 3 and 4), are shown in Fig. 1.
Fig. 1.
Song is coordinated with cyclical movements of the hyoid skeleton (Fig. 1 C and D) that produce an air-filled supralaryngeal cavity, the volume of which is inversely related to f0 (see Movie 1, which is published as supporting information on the PNAS web site). The larynx moves with the hyoid skeleton to which it is attached by extrinsic muscles and ligaments (13–15). The avian hyoid apparatus is not attached to the skull, as it is in mammals, allowing greater freedom of movement and a much larger opening into the oropharyngeal cavity (16). The oropharyngeal cavity is enlarged during low fundamental frequencies by a dorsoventral (Fig. 1 C and D, LV; Syll 1, 7.2 ± 1.3 mm n ] 27; Syll 3, 6.8 ± 1.9 mm n ] 6; Syll 4, 6.1 ± 1.9 mm n ] 117; mean maximum movement ± SD; ≈5% of this movement is due to a backward movement of the vertebral column) and craniocaudal (Fig. 1 C and D, LH; Syll 1, 3.5 ± 1.1 mm n ] 23; Syll 3, 3.7 ± 1.6 mm n ] 6; Syll 4, 4.1 ± 2.0 mm n ] 113; mean maximum movement ± SD) movement of the basihyoid with a simultaneous lateral movement of the cornua (Fig. 1C Cornua; Syll 1, 6.4 ± 2.1 mm n ] 20; Syll 3, 6.8 ± 1.8 mm n ] 2; Syll 4, 8.3 ± 2.6 mm n ] 23; mean maximum distance between cornuae, each cornua moves laterally half of this distance). When f0 rises, these movements are reversed, reducing the volume of the oropharyngeal cavity (Fig. 1C). The magnitude of hyoid displacement is inversely correlated with changes in f0, regardless of the direction of the FM sweep (Fig. 2). Movements of the beak also tend to covary with f0, but are less consistent than those of the hyoid and cornua and seem to be acoustically important over only a limited range of frequencies (9, 10).
Fig. 2.
The volume of the supralaryngeal cavity is further increased during a low f0 by an expansion, visible in the x-ray images, of the cranial end of the esophagus, which merges with the oropharyngeal cavity to become a single chamber that is part of the supralaryngeal vocal tract (see Movie 1). The length of the resulting oropharyngeal–esophageal cavity (OEC) can be up to twice that of the oropharyngeal cavity alone. This combined laryngeal–hyoidal and esophageal motor pattern can vary the cavity volume several fold during the course of a syllable.
These cyclical, frequency-dependent changes in the dimensions of the OEC suggest that it has an important acoustic role in song, perhaps acting as a variable filter that tunes the vocal tract resonance to the f0 of the syringeal sound source. To gain data for a test of this hypothesis, we constructed a three-dimensional model of the OEC by tracing its outline in successive video frames (Fig. 3A and B; see Movie 1). These data were used to calculate the volume and geometry of the cavity and beak. The maximum volumes computed from this model agree with those measured from oropharyngeal–esophageal casts that were obtained from four dead birds. Both methods gave a maximum volume of ≈2.0 ml during the low frequencies (≈2 kHz) at the beginning of syllable 1. The minimum volume near the end of the high frequency upsweep (≈5 kHz) that terminates this syllable was ≈0.6 ml according to the three-dimensional model.
Fig. 3.
A computational model based on the dimensions of the cardinal vocal tract (see Supporting Text and Fig. 5, which are published as supporting information on the PNAS web site) was then used to predict the vocal tract resonances (12). Computation shows that the frequency of the oropharyngeal–esophageal resonance can be varied by the bird from below that of the first tracheal resonance (≈1.6 kHz) to above that of the third tracheal resonance (≈8 kHz).
The predicted resonances for syllable type 1 are summarized by the solid curves in Fig. 3C. At the beginning of syllable 1, when the oropharyngeal–esophageal volume is 2.0 ml (Fig. 3A) and beak gape is small, the vocal tract has a major resonance peak between ≈1.5 and 2.0 kHz (solid purple curve) that coincides with the first tracheal resonance and matches the frequency of f0. As f0 rises, the oropharyngeal–esophageal volume decreases and the major resonance frequency increases, passing through 2.7 kHz (solid dark blue curve), for example, at a volume of ≈1.2 ml with a partially open beak. The oropharyngeal volume at the end of syllable 1 is 0.6 ml (Fig. 3B) and beak gape increases to ≈6 mm. These changes shift the resonance peak upward, tuning f0 to ≈5 kHz (Fig. 3C, solid light blue curve), which is a little above the second tracheal resonance and in close agreement with the terminal f0.
Other syllables are accompanied by similar oropharyngeal and esophageal motor patterns. In some of these f0 extends as high as 8 or 9 kHz. The bird could tune its vocal tract resonance to these frequencies by reducing the volume of the supralaryngeal vocal tract closer to its collapsed state (Fig. 3C, dashed resonance curves), while maintaining a relatively large beak gape. At high frequencies the cranial end of the esophagus collapses and the volume of the oropharyngeal cavity is reduced as the larynx returns toward its resting position near the base of the beak. An oropharyngeal volume of 0.2 ml, which is still more than twice the volume of the trachea, together with a large beak opening, for example, should shift the major resonance peak to ≈9 kHz (Fig. 3C, red curve).
The relationship between the linear dimensions of the oropharynx and f0 (Fig. 2) and the agreement between f0 and the predicted resonance of the upper vocal tract at the beginning and end of a FM syllable (Fig. 3) indicate that the first resonance of the oropharyngeal–esophageal filter tracks f0, at least during relatively long syllables with gradual FM. A higher sample rate is needed to determine the accuracy of tracking during the rapidly modulated portion of syllables 1 and 4.
Discussion
The experiments we report here provide several findings that contribute to a better understanding of how songbirds use their vocal tract to modulate their vocal signals. (i) We show that song is accompanied by previously undescribed frequency-dependent changes in the configuration of the upper vocal tract, involving movement of the hyoid skeleton, together with the larynx to which it is attached, and expansion of the cranial end of the esophagus. The larynx of the domestic cock descends during crowing, but its relationship to sound frequency is unknown (17). (ii) This song-related motor pattern maintains an inverse relationship between the volume of the OEC and the song’s f0. (iii) A computational acoustic model (12) of the upper vocal tract indicates that this motor pattern varies the major resonance of the OEC to track the f0, increasing its level relative to that of higher harmonics.
The vocal tract filter of songbirds is fundamentally different from that present in doves, which also expand their upper vocal tract during vocalization. Unlike songbirds, doves vocalize with their mouth and nares closed. They also use expiratory pressure to passively inflate their esophagus, rather than active muscular tension (18). Furthermore, the resonance of the dove’s filter is not critically dependent on the volume of the vocal tract, because there is a compensating effect on the physical properties of the esophageal wall (19, 20). However, the most crucial difference is that although sound radiation from songbirds is through the partly open beak, sound radiation in doves occurs by mechanical vibration of the thin walls of the esophagus.
Combining our findings with those of previous investigators, we conclude that the songbird’s vocal tract contains at least three different filter components, the trachea, the OEC, and the beak, that although they interact with each other to some degree, differ in the extent to which they can be tuned and the frequencies over which they are most effective.
The Tracheal Filter.
The trachea behaves acoustically like a stopped tube with multiple resonances spaced at intervals across the range of the f0 (e.g., ref. 21). Tracheal resonances at ≈2, 5, 8, and 12 kHz, which are close to the frequencies predicted by the computational model, are detectable in FM syllables as a transient increase or decrease in sound level of the odd or even numbered harmonics, respectively, when they cross the predicted tracheal resonances (Fig. 4). The center frequency of the first few of these resonances changes very little during song; this is consistent with measurements indicating that tracheal length changes very little during zebra finch song (22) and with the finding that changes in the beak gape of the eastern towhee have minimal effect on tracheal resonance (10). In theory, the resonances of the songbird tracheal filter could be lowered without changing tracheal length, by constricting the glottis (19). We are unaware of evidence that songbirds normally constrict their glottis during song, but the contribution of the glottis to the vocal tract filter remains to be explored. Tracheal pressure in singing canaries does not increase more than ≈1 or 2 cm H2O above ambient atmospheric pressure (R.A.S., unpublished data), suggesting that the glottis is not severely constricted, although small changes in glottal opening might produce acoustic effects before changes in pressure or airflow become apparent.
Fig. 4.
The Oropharyngeal–Esophageal Filter.
The oropharyngeal–esophageal filter is the only one of these three filter components that tracks the fundamental frequency f0 of the song, perhaps through a reverse mode-locking effect on the syringeal vibrations, although its accuracy in tracking short steeply FM notes remains to be determined. Its tuning to low frequencies, where it may be particularly important, is made possible by lateral movements of the hyoid cornua, which expand the cranial end of the esophagus, greatly increasing vocal tract volume and allowing the bird to tune its major resonance to frequencies below 2 kHz. At higher f0, above ≈4 kHz, the esophagus is collapsed, the larynx returns toward its resting position, and beak gape becomes important. It remains to be determined which of ≈11 hyoidal and laryngeal muscles are primarily responsible for these movements.
These new song-related peri-oral motor patterns that control the shape of the OEC must be coordinated, by pathways yet to be identified, with the activity of the syringeal muscles that gate phonation and control the f0 (23). The computational model we used to calculate the curves in Fig. 3C predicts resonances with quite high Q values (ratio of resonance frequency to bandwidth at −3 dB, ≈30) and it could be thought that tuning of the song fundamental to an accurate match might present an insurmountable problem to the bird. There are actually two possible solutions. First, the numerical values used in the model assume ideally smooth rigid walls, which is clearly an oversimplification. Relaxation of this assumption would add to overall system losses and thus lower the Q value. Second, and perhaps more significantly, the model assumes a simple source–filter system in which the operation of the syringeal valve is not influenced by resonances of the upper vocal tract. An interaction of this sort has not been demonstrated in songbirds, but if present, it could automatically couple the syringeal vibration frequency to the OEC resonance to provide optimal tuning (24).
A direct measure of the contribution of the OEC filter by determining the effect of disabling it is difficult because the muscles involved are also involved in eating and swallowing. An estimate of its importance can be obtained, however, from the song of intact birds. The mean difference between 2f0 and f0 in the radiated spectrum of six cardinal syllables, similar to that in Fig. 4B, was −23 ± 7 dB (M ± SD). The mean difference between 3f0 and f0 in these same syllables was −34 ± 8 dB. Part of this difference is presumably due to a higher level of f0, compared 2f0 or 3f0, in the signal at the syringeal source. The cardinal’s source spectrum is not known, but source spectra of human modal speech decays according to (frequency)−2, which equals a relative attenuation of −12 and −19 dB for 2f0 and 3f0, respectively (25). Estimates of the syringeal source spectrum for the coos of ring doves are in close agreement with those reported for human speech (26). Assuming the source spectrum of a cardinal is subject to a similar decay, the boost in sound level due to the oropharyngeal–esophageal primary resonance may be in the range of ≈11–15 dB. This is less than the ≈20- to 30-dB difference, predicted by the model, between peak sound levels of the OEC primary resonances and those of tracheal resonances located some distance from that of the OEC (Fig. 3C).
The Effect of Beak Gape.
Computational models indicate that the beak has most effect on the filter when it is nearly closed (9). Assessing the specific contribution of beak gape to the vocal tract filter is difficult in living birds because neither the power spectrum of the sound generated by the syrinx nor the extent to which it is modulated by concurrent changes in the configuration of other parts of the suprasyringeal vocal tract are known. Nelson et al. (10) attempted to minimize these potential sources of error by measuring the effect of beak gape on a known signal from a miniature speaker inserted into the base of the trachea at the position of the syrinx in dead eastern towhees. In this experiment, opening or closing the beak amplified or attenuated, respectively, frequencies between ≈4 and 7.5 kHz by up to 20 or 30 dB, but had little effect on frequencies below ≈4 kHz. Thus, the beak seems to exert its major acoustic effect by attenuating frequencies above ≈4 kHz, when gape is small. The towhee experiment has the limitation that it eliminates possible feedback or coupling between the oscillating syringeal source and the upper vocal tract. However, it remains to be determined whether source-filter interactions are acoustically important. It is reassuring that data from live cardinals support the conclusion that birds use beak gape to suppress high frequencies. When a singing cardinal is prevented from reducing his beak gape, the attenuation of the second and third harmonics relative to the fundamental is reduced ≈20 dB for f0 below ≈3 kHz (R.A.S., F. Goller, R. Bermejo, M. Wild, and P. Zeigler, unpublished data; see ref. 7).
Relationship to Communication.
Tonal sounds, i.e., those with relatively little energy in overtones, are common in the songs of many species (27, 28), and there is evidence that birds attend to the tonality of their vocalizations (29–31). The reason for the predominance of tonal sounds in avian communication is not clear. Nelson et al. (10) suggest that it may be advantageous to suppress harmonics of fundamental frequencies below 4 kHz that might otherwise degrade the temporal pattern of song elements with an f0 between 4 and 8 kHz by adding acoustic “clutter” to this high-frequency portion of the bird’s hearing range.
A vocal tract filter that adjusts its major resonance to track f0 while suppressing higher harmonics may also facilitate acoustic communication by increasing the loudness of the fundamental and uniformity of tone in birdsong. Songbirds thus employ an acoustic strategy similar to that used by human sopranos who adjust the resonance of their vocal tract to match high values of f0 to be heard over the orchestra (32, 33).
Methods
Cineradiography.
X-ray imaging was performed with an x-ray unit having a mobile C-arm (Cardiac Digital Mobile Imaging System; Series 9800). This system provides pulsed digital cine (up to 150 mA) at 30 pulses per second with a 10-ms pulse width and true 1,000 × 1,000 imaging resolution, allowing digital recording at 30 frames per s. The digital signal from the fluoroscope was recorded on a S-VHS (NTSC standard) video recorder (JVC Super VHS ET, model HR-S2901U) together with the sound recorded by a directional microphone (Sennheiser ME88) aimed at the bird from a distance of 1 m, as described elsewhere (19). All experiments were in compliance with the guidelines of the National Institutes of Health and were reviewed and approved by the Institutional Animal Care and Use Committee of Indiana University.
Eight different syllable types were recorded from three male northern cardinals as they sang while sitting in the x-ray beam of the C-arm. Data were obtained from all three birds for syllables 1 and 4, and from two birds in the case of syllables 2 and 3. These syllables were selected because they were sung by at least two birds, covered a relatively large frequency range, including both upward and downward sweeps, and provided a reasonable sample size. The data presented here are representative of all birds that sang the syllable. Four syllables were excluded because they had a small sample size from only one bird and/or had a duration <60 ms and so were represented by only one video frame. Excluded syllables also had a limited bandwidth (<3 kHz). Based on the available data, hyoid movements in the excluded syllables are consistent with those of the syllables reported here.
Two distances in the lateral (LV and LH) and one in the ventro–dorsal view (Cornua) were measured. LV is the distance between the larynx and the mid-point of the second vertebra. LH is the distance between the larynx and the dorsal edge of the beak-skull transition. Cornua refers to the distance between the most ventral point of the cornua of the hyoid apparatus. Distances were computed (tracker, http://sunflower.bio.indiana.edu/∼emartins/tracker.html) from the coordinates of two points selected manually in each frame. Only cineradiographs with an essentially perfect lateral or ventrodorsal view were used. A reference of known length positioned at the mid-sagittal level of the bird allowed accurate calibation of distance measurements. Ten repeated measures of each of the same three distances in a single frame had a standard deviation of 0.5 mm.
All data are corrected for a delay between the recorded audio and video signals due to the processing time of the x-ray image. The remaining margin of error between the alignment of the vocalization and the x-ray image is estimated not to exceed one-half frame (±17 ms). Segments of song recorded on the S-VHS video tape were digitized at 30 frames per s (video) and 44.1 kHz (audio) sampling rate (Vegas Video, Sonic Foundry, Madison, WI) and displayed on a computer monitor as individual video frames. Data points were selected with an on-screen cursor. Acoustic measurements were performed by using sound analysis software (praat, version 4.1; www.praat.org).
Computing the Volume of the Supralaryngeal Vocal Tract.
The volume of the OEC was estimated by two independent methods. Four male cardinals were killed with an overdose of isoflurane and a cast of the cavity at its maximum volume was obtained by injecting it with impression medium. In two of these birds, threads were attached to the larynx and hyoidal cornua, which were pulled caudo-ventrally or caudo-laterally, respectively, to their positions of maximum displacement during song, before injection of the impression medium. The volume of the cast, or parts of it, was determined by measuring the volume of water it displaced.
The volume of the OEC as it changes during song was also computed from x-ray images of its lateral and ventrodorsal cross-sectional area in each of 10 successive images at 33-ms intervals during the course of a type 1 syllable. Air provides negative contrast in radiological images, making it possible to clearly delineate the borders of the pharynx and the inflated esophagus (see Movie 1). The extents of the oropharyngeal cavity were outlined on both the lateral and ventrodorsal x-ray images using a two-dimensional painting package (photoshop, Adobe Systems, San Jose). The outlined lateral and ventrodorsal views were imported into a three-dimensional modeling package (ac3d, version 4.08; www.ac3d.org), mapped to perpendicular planes, and registered with each other. A three-dimensional polygonal model was then constructed to closely approximate the outlines in both views while also honoring the anatomical constraints of the throat and cavity. Finally, the polygonal model was exported to a three-dimensional geometric analysis tool (rapidform, 2004 version; www.rapidform.com) to compute the volume of the enclosed space. Maximum OEC volumes computed in this way agreed closely with those obtained from casts. These volumetric data were used in a computational model (12) to predict the resonances of the cardinal’s vocal system (see Supporting Text and Fig. 5).
Abbreviations:
- FM
- frequency modulation
- OEC
- oropharyngeal–esophageal cavity
- f0
- fundamental frequency.
Acknowledgments
We thank Drs. H. Herzel, M. J. Owren, and B. Nelson for their comments on a draft of this article and D. Homberger for discussions on hyoid anatomy. Three-dimensional modeling, volume measurements, and animation were provided by Eric A. Wernert (Advanced Visualization Lab, Indiana University, Bloomington). We are indebted to S. Ronan for technical assistance and to S. A. Zollinger for artwork in Fig. 1. This work was supported by a National Institutes of Health/National Institute of Neurological Disorders and Stroke grant (to R.A.S.) and a fellowship from the Postdoctoral Program of the German Academic Exchange Service (DAAD) (to T.R.).
Supporting Information
01262Movie1.mov
- Download
- 9.65 MB
01262Fig5.jpg
- Download
- 31.82 KB
References
1
S. Nowicki Nature 325, 53–55 (1987).
2
G. Fant Acoustic Theory of Speech Production (Mouton, The Hague, 1970).
3
M. W. Westneat, J. Long, H. John, W. Hoese, S. Nowicki J. Exp. Biol 182, 147–171 (1993).
4
W. J. Hoese, J. Podos, N. C. Boetticher, S. Nowicki J. Exp. Biol 203, 1845–1855 (2000).
5
F. Goller, M. J. Mallinckrodt, S. D. Torti J. Neurobiol 59, 289–303 (2004).
6
J. Podos, J. A. Southall, M. R. Rossi-Santos J. Exp. Biol 207, 607–619 (2004).
7
R. A. Suthers, F. Goller Curr. Ornithol 14, 235–288 (1997).
8
J. Podos, J. K. Sherer, S. Peters, S. Nowicki Anim. Behav 50, 1287–1296 (1995).
9
N. H. Fletcher, A. Tarnopolsky J. Acoust. Soc. Am 105, 35–49 (1999).
10
B. S. Nelson, G. J. L. Beckers, R. A. Suthers J. Exp. Biol 208, 297–308 (2005).
11
O. L. Larsen, T. Dabelsteen Ornis Scand 21, 37–45 (1990).
12
N. H. Fletcher, T. Riede, R. A. Suthers J. Acoust. Soc. Am 119, 1005–1011 (2006).
13
J. C. George, A. J. Berger Avian Myology (Academic, New York, 1966).
14
D. G. Homberger Am. Zool 19, 988 (1979).
15
D. G. Homberger The Lingual Apparatus of the African Grey Parrot, Psittacus erithacus Linne (Aves: Psittacidae): Description and Theoretical Mechanical Analysis (American Ornithologist’s Union, Washington, DC, 1986).
16
D. G. Homberger, eds N. J. Adams, R. H. Slotow (BirdLife South Africa, Durban), pp. 94–113 (1999).
17
S. S. White J. Anat 103, 390–392 (1968).
18
A. S. Gaunt, S. L. L. Gaunt, R. M. Casey Auk 99, 474–494 (1982).
19
T. Riede, G. J. L. Beckers, W. Blevins, R. A. Suthers J. Exp. Biol 207, 4025–4036 (2004).
20
N. H. Fletcher, T. Riede, G. J. L. Beckers, R. A. Suthers J. Acoust. Soc. Am 116, 3750–3756 (2004).
21
W. T. Fitch J. Zool 248, 31–48 (1999).
22
M. Daley, F. Goller J. Neurobiol 59, 319–330 (2004).
23
F. Goller, R. A. Suthers J. Neurophysiol 76, 287–300 (1996).
24
N. H. Fletcher J. Theor. Biol 135, 455–481 (1988).
25
J. R. Flanagan Speech Analysis Synthesis and Perception (Springer, New York, 1972).
26
G. J. L. Beckers, R. A. Suthers, C. ten Cate Proc. Natl. Acad. Sci. USA 100, 7372–7376 (2003).
27
C. H. Greenewalt Bird Song: Acoustics and Physiology (Smithsonian Inst. Press, Washington, DC, 1968).
28
P. Marler Bird Vocalizations, ed R. A. Hinde (Cambridge Univ. Press, Cambridge, U.K.), pp. 5–27 (1969).
29
J.-C. Bremond Behaviour 58, 99–116 (1976).
30
J. B. Falls, ed C. G. Sibley (American Ornithologist’s Union, Ithaca, NY), pp. 259–271 (1963).
31
J. Strote, S. Nowicki Behaviour 133, 161–172 (1996).
32
J. Sundberg Acustica 32, 89–96 (1975).
33
E. Joliveau, J. Smith, J. Wolfe Nature 427, 116 (2004).
Information & Authors
Information
Published in
Classifications
Copyright
© 2006 by The National Academy of Sciences of the USA.
Submission history
Received: July 28, 2005
Published online: April 4, 2006
Published in issue: April 4, 2006
Keywords
Acknowledgments
We thank Drs. H. Herzel, M. J. Owren, and B. Nelson for their comments on a draft of this article and D. Homberger for discussions on hyoid anatomy. Three-dimensional modeling, volume measurements, and animation were provided by Eric A. Wernert (Advanced Visualization Lab, Indiana University, Bloomington). We are indebted to S. Ronan for technical assistance and to S. A. Zollinger for artwork in Fig. 1. This work was supported by a National Institutes of Health/National Institute of Neurological Disorders and Stroke grant (to R.A.S.) and a fellowship from the Postdoctoral Program of the German Academic Exchange Service (DAAD) (to T.R.).
Authors
Competing Interests
Conflict of interest statement: No conflicts declared.
Metrics & Citations
Metrics
Citation statements
Altmetrics
Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
Cited by
Loading...
View Options
View options
PDF format
Download this article as a PDF file
DOWNLOAD PDFGet Access
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginRecommend to a librarian
Recommend PNAS to a LibrarianPurchase options
Purchase this article to get full access to it.