1. Buoyancy control
The gas bladder of actinopterygian fishes is probably homologous to the paired lungs of tetrapods despite its dorsal connection with the gut (the lungs develop from a ventral diverticulum, or budding). See Kardong pp. 399-401 for an outline of how the physoclistous gas bladder changes its volume (which, in turn, regulates buoyancy). The figure below shows the evolution of the lung from a respiratory to a buoyancy control organ.
2. Vocalization
(be sure to visit the bird and frog call links at the bottom of the page)
Sound is a traveling pressure wave that is propogated away from the source like the concentric ring of ripples propogating away from a stone dropped into water. At some distance from the source of the sound, the pressure will alternatingly rise and drop as the air compresses and expands (rarefaction). This is why objects vibrate when a sound is loud enough.
The basic idea of vocalization, then, is to generate a pressure wave. In general, the larynx is the source of the the pressure wave in tetrapods, but the nasal cavity, the lips, and tongue can also create sound. Toothed whales (Ondontoceti), for example, vocalize with structures in the nasal cavity.
There are at least three ways to do generate sound in the larynx, but all have to do with air moving past the vocal structures (vocal cords in frogs and humans, the syrinx in birds). The part of the vocal system inferior or posterior to the vocal structure is called infralaryngeal (or sublaryngeal or subglottal). The part superior or anterior is supralaryngeal (or supraglottal). The space that is constricted is the glottis.
A human larynx
|
 |
Below: action of the human vocal cord muscles as illustrated by the dean of American medial illustrators, Frank Netter |
 |
The three ways to generate sound are:
1) Vibration hypothesis. Vibrating structures alternatingly compress and expand an adjacent volume of fluid (air or water) and, therefore, produce sound waves. This is how a guitar string or a tuning fork works. The frequency of the vibration determines the pitch of the sound, and can be controlled by the geometry and tension of the vibrating structure. By constricting the air passage, the air passing a vocal structure speeds up and the pressure drops (Bernouilli effect). This drop in pressure causes the vocal structure to bow into the air passage. The elasticity of the vocal structure pulls the structure back (unbows it) and what results is a vibration of the structure resulting from the competing forces positioning the structure.
2) Pulse hypothesis. Closing the air passage during exhalation of the lung increases the infralaryngeal pressure as the thorax contracts due to elastic rebound or active muscle contraction. When the pressure rises, the force on the vocal structure increases to a point that will open the structure and allow air to pass. The sudden increase in flow past the vocal structure decreases the pressure and the vocal structure will bow in and restrict the space. Again, the sublaryngeal pressure rises enough to open the glottis and allow air to pass. This cycle continues. The pulses of high and low pressure create the sound wave and vibrate the vocal structure (this differs somewhat from the vibrating vocal structure creating the pressure wave as in #1).
3) The aerodynamic hypothesis. This hypothesis suggests that air passing the vocal structure separates from the surface of the structure, forming a von Karmen vortex street in its wake (which would be supralaryngeal). The wake represents a series of high and low pressure areas. The growth and shedding of vortices on the vocal structures may cause the structure to vibrate. Importantly, the sound is produced by the geometry of the vortices in the wake and not by a vibrating vocal structure. If vibration occurs, it would simply be the result of the vortex shedding (attached vortices detaching and moving downstream in the wake; the presence and absence of attached vortices on the downstream size cause pressure variation on the vocal structure which results in vibration). Were the vocal structures extremely stiff, they may not vibrate at all but the sound would still be produced.
Summary. The vibration and pulse hypothesis both require a vibration of the vocal structure but in the vibration hypothesis the vibrating structure is the source of the sound, by compression and rarefaction of the adjacent air, while in the pulse hypothesis, the pulses of flow are the source of the sound. These two hypotheses are really at the extremes of a continum depending on the magnitude of the constricted airsace (the pulse hypothesis would be a completely restricted space followed by a small opening, while the vibration hypothesis could work with only a slightly constriced airspace. The aerodynamic hypothesis is like the pulse hypothesis, in that a vibration is not the cause of the sound. The difference is that vibration is necessary in the pulse hypothes (because this vibration is both a cause and a consequence of the pulses of high and low pressure) but is not necessary to generate the sound in the aerodynamic hypothesis and in fact, may actually interfere with it..
The vocal tract
The sound wave generated by the vocal structure actually has multiple frequencies: a fundamental frequency and a series of frequencies that are integer multiples of the fundamental. For example, 400 Hz, 800 Hz, 1200 Hz, 1600 Hz, 2000 Hz, etc, where 400 Hz is the fundamental frequency and the other frequencies are 2X, 3X, 4X, and 5X this fundamental. These higher frequencies are the harmonics. The actual sound that leaves the mouth is determined by the relative amplitude of each of the harmonics. The vocal tract, the part of the respiratory system between the glottis and the mouth/nasal cavity, acts as a bandpass filter, a device that allows certain frequencies to pass and dampens (lowers the amplitude) other frequencies. The vocal tract, then, is a resonance chamber and its shape determines the perceived pitch (not the real pitch) of the sound. For a great page explaining this, including "helium voice" and some sound files of normal and "helium voice" speech, visit the physics in speech page. Understanding "helium voice" is more than just fun, its a tool to test models of the vocal system and a mixture of helium and oxygen called heliox has been used to test models of frog and bird calls.
Frog calls
Frogs have vocal cords in the larynx similar to that of mammals. Additionally, male frogs have a vocal sac that balloons out during vocalization. The figure below shows the openings (apertures) to the vocal sacs in the floor of the mouth of a frog.
The figure below shows the respiratory cycle in a frog. Panel 2C and 3 is the vocalization loop. An older idea on the function of the vocal sacs was that it acts as a resonance chamber, like the vocal tract of a human. A study with heliox (a mixture of helium and oxygen) showed that the vocal sac is not a resonance chamber. I don't have that paper but I'm guessing the idea is that the helium mixture should change the resonance properties of the vocal sac (shift the filter to high frequencies - see the "helium voice" page). The tissue of the vocal sac is full of elastic fibers so its possible the vocal sac acts as a low energy method for repeated vocalization (see Jaramillow et al. 1997). That is, the frog uses energy to inflate the lungs, which being elastic, will rebound to their resting size and expel the air through the larynx. By keeping the mouth closed and opening the vocal sac apertures, the air will rush into the vocal sac, which will expand. The sacs will contract due to the elastic properties of the tissue and push the air out back through the larynx into the lungs and reinflating these. This cycle can continue with very little enerby input because both structures, the sacs and the lungs are highly elastic and store and release the original energy used to prime the system.
bird calls
Birds do not have vocal folds but instead have a structure in the trachea above the branching of the primary bronchi (non-songbirds) or at the top of each primary bronchus (songbirds) called a syrinx.
Classical model of avian phonation |
|
| In this figure of a syrinx, the internal (or medial) tympaniform (not typaniform) membrane is labelled. The "classical" model of bird phonation argued that the source of bird vocalization resulted from the vibration of this membrane. Later, the "pulse-tone hypothesis" suggested that the Medial tympaniform membrane's constriction against the lateral walls was the source. Finally, its been proposed in the "whistle hypothesis" that the constriction of the bronchi or trachea by this membrane creates a vortex flow that produces a whistle sound. The commonality of all these models is that they involve the medial (internal on the figure) tympaniform membrane. |  |
Revised model of avian phonation |
|
| In a series of papers, Goller and Larsen (1997a, 1997b, 1999) have shown that it is in fact other vocal structures that are the source of bird phonations in both songbirds (oscines) and non-songbirds. When the medium tympaniform membrane is experimentally kept from vibrating, the bird still vocalizes as it would with an intact, vibratable medium tympaniform membrane. | |
|
In the pigeon (a non-songbird), there is a lateral (or external) tympaniform membrane that spans two of the tracheal rings (see figure above) superior to the medial tympaniform membranes. It is these membranes that constrict the trachea, vibrate, and produce the sounds (probably by the pulse method since they completely constrict the tracheal space). From Goller and Larsen, (1997b). |
 |
| In songbirds, it is the lateral labium (LL), that bows out toward the medial labium (ML) and constricts the passage. Again, the large constriction is support for the pulse hypothesis of sound generation. From Goller and Larsen (1997a). |
|
