Functional Morphology of the Hyolaryngeal Complex of the Harbor Porpoise (Phocoena phocoena): Implications for its Role in Sound Production and Respiration

Authors

  • Stefan Huggenberger,

    Corresponding author
    1. Department of Anatomy III (Dr. Senckenbergische Anatomie), Johann Wolfgang Goethe-University, Theodor-Stern-Kai 7, 60590 Frankfurt am Main, Germany
    • Zoological Institute II, University of Cologne, Weyertal 119, 50931 Köln, Germany
    Search for more papers by this author
    • Fax: +49-221-470-4889.

  • Michael A. Rauschmann,

    1. Orthopaedic University Department Friedrichsheim gGmbH, Johann Wolfgang Goethe-University, Marienburgstr. 2, 60528 Frankfurt am Main, Germany
    Search for more papers by this author
  • Helmut H.A. Oelschläger

    1. Department of Anatomy III (Dr. Senckenbergische Anatomie), Johann Wolfgang Goethe-University, Theodor-Stern-Kai 7, 60590 Frankfurt am Main, Germany
    Search for more papers by this author

Abstract

In several publications, it was shown that echolocation sound generation in the nasal (epicranial) complex of toothed whales (Odontoceti) is pneumatically driven. Modern hypotheses consider the larynx and its surrounding musculature to produce the initial air pressure: (a) contraction of the strong pipelike palatopharyngeal sphincter muscle complex, which connects the choanae with the epiglottic spout of the larynx, should provide much of the power for this process and (b) muscles suspending the larynx/hyoid complex from the skull base and the mandibles may support these pistonlike laryngeal movements. Here, we describe the morphology and topography of the larynx, the hyoid apparatus, and the relevant musculature in the harbor porpoise (Phocoena phocoena) with respect to odontocete vocalization and respiration. We demonstrate that the hyoid apparatus, reminiscent of a “swinging cage,” may not only be a stable framework in which the larynx can move but should support laryngeal actions by its own movements. Rostrocaudal relocations of the hyoid apparatus may thus support pistonlike actions of the larynx creating air flow into the nasal complex for sound production. The lift of the hyoid apparatus with the thick larynx in the direction of the skull base may squeeze the pharynx in the region of the piriform recesses and thus help to secure the (waterproof) tracheochoanal connection during respiration when the palatopharyngeal sphincter cannot be contracted maximally, because the air passage must remain open at the epiglottic spout. Anat Rec, 291:1262–1270, 2008. © 2008 Wiley-Liss, Inc.

Toothed whales (Odontoceti) use an active sonar system for orientation and hunting. Unique among the Mammalia, echolocation sounds are generated by structures in the nose of these animals (reviewed in Cranford, 2000; Cranford and Amundin, 2003). In different studies, it was shown that sound generation in the odontocete nasal (syn. epicranial) complex is pneumatically driven (Ridgway et al., 1980; Amundin and Andersen, 1983; Ridgway and Carder, 1988; Cranford, 2000; Cranford et al., 2001). Modern hypotheses consider the larynx and its surrounding musculature, that is, the palatopharyngeal sphincter and the powerful gular musculature, to produce the air pressure needed for nasal ultrasound and sound generation (Cranford et al., 1996, 2001; Houser et al., 2004). However, most of the fundamental investigations on odontocete laryngeal anatomy, however, were carried out long time ago (Tyson, 1680; Rawitz, 1900; Boenninghaus, 1903; Danois, 1910; Gallardo, 1913; Hain, 1914; Kernan and Schulte, 1918; Schulte and Smith, 1918) when the principles of echolocation and nasal sound generation were still unknown.

In odontocetes, the larynx is located in the medioventral vault below the neurocranium. It is characterized by a considerable elongation of the epiglottic and corniculate (cuneiform or “Wrisberg”) cartilages, thus forming a goose beaklike spout at its rostral end (Schneider, 1964; Blevins and Parkins, 1973; Green et al., 1980; Reidenberg and Laitman, 1987). The tip of the epiglottis forms a neck and protruding lip, which are surrounded and held in position by the prominent palatopharyngeal sphincter muscle, suspending the epiglottic spout from the bony nares (choanae) and the skull base, respectively (Negus, 1958; Moris, 1969; Purves and Pilleri, 1983; Reidenberg and Laitman, 1987). This sphincter muscle as well as the powerful gular musculature probably generates much of the pneumatic pressure necessary for sound production by forcing the closed larynx as a piston in the direction of the choanae (Ridgway et al., 1980; Amundin and Andersen, 1983; Cranford et al., 1996). An additional function of the palatopharyngeal sphincter is to prevent water in the oral cavity from entering the respiratory tract during feeding (Reidenberg and Laitman, 1987).

In this work, we outline a possible mechanism of air pressurization in the nasal cavity by synchronized actions of the larynx and hyoid musculature based on morphological studies. In addition, we present a hypothesis on how movements of the larynx supported by the hyoid apparatus may prevent water from penetrating into the airway during respiration.

MATERIAL AND METHODS

Our examination of the odontocete larynx and hyoid apparatus is based mainly on magnetic resonance imaging (MRI), cryosections, and macroscopical dissections of seven harbor porpoise heads (Phocoena phocoena; cf. Table 1). For comparison, heads of one bottlenose dolphin (Tursiops truncatus), one perinatal pantropical spotted dolphin (Stenella attenuata), and three La Plata dolphins (Pontoporia blainvillei) were studied with computer-assisted tomography (CT) and MRI. An additional head of the La Plata dolphin as well as the head of a pygmy sperm whale (Kogia breviceps) was dissected macroscopically (Table 1). The MRI scans were usually performed in all three planes, and the CT scans in the transverse plane. The horizontal plane was established in the skull base in parallel to the straight line between the two periotic bones (or between the eyes when the ear bones had been removed) and to the straight line from the tip of the rostrum to the ventralmost border of the foramen magnum. The midsagittal plane was determined along the line between the tip of the rostrum and the midpoint of the foramen magnum and perpendicular to the horizontal plane; the transverse plane standing at right angles to both of them. The lines and planes were set using prescans before the final scans were conducted. Finally, midsagittal histological sections through the complete larynx of a finless porpoise (Neophocaena phocaenoides; azocarmine and aniline blue stain; donated by Giorgio Pilleri, Courgeveaux, Switzerland; not included in Table 1) were used in the analysis, because its anatomy was found to be very similar to the larynx of the harbor porpoise.

Table 1. List of the material examined
SpeciesInstitutionaID numberAgeBody lengthSexFixationMethod applied
  • Methods applied: CS, (macroscopy) cryosectioning; CT, computer assisted tomography; DS, macroscopical dissection; MRI, magnetic resonance imaging.

  • a

    DMM, Deutsches Meeresmuseum, Stralsund, Germany; FTZ, Forschungs- und Technologiezentrum Westküste, Büsum, Germany; GP, collection of Prof. Dr. G. Pilleri, Courgeveaux, Switzerland (located in SMF); NMFS, National Marine Fishery Service, San Diego, CA, USA; NMNH, National Museum of Natural History, Washington, DC, USA; SAI; Dr. Senckenbergische Anatomie, Johann Wolfgang Goethe-University, Frankfurt a.M., Germany; SMF, Research Institute and Natural History Museum Senckenberg, Frankfurt a.M., Germany; ZMUC, Zoological Museum of the University of Copenhagen, Denmark.

Kogia brevicepsNMNHSubadultMFrozenDS
Phocoena phocoenaDMMAdultFFrozenCT, MRI, CS
Phocoena phocoenaFTZ1367Adult102 cmMFormalinDS
Phocoena phocoenaFTZ1369Adult116.5 cmFFormalinCT, MRI, DS
Phocoena phocoenaSAIAdultMFrozenCS (sagittal)
Phocoena phocoenaSAIAdultMFrozenCS (transverse)
Phocoena phocoenaZMUCCN 138Fetal42 cmEthanolCT, MRI
Phocoena phocoenaSMFNeonate72 cmEthanolCT, MRI
Pontoporia blainvilleiGPAdult(Skull length 44 cm)FormalinDS
Pontoporia blainvilleiGPAdultFormalinMRI
Pontoporia blainvilleiGPAdultFormalinCT, MRI
Pontoporia blainvilleiGPAdultFormalinCT, MRI
Stenella attenuataNMFSRBR 018Perinatal94.5 cmFormalinCT, MRI
Tursiops truncatusSAISAI 7928SubadultFormalinCT, MRI

RESULTS

In general, the larynx of the harbor porpoise (as well as that of the finless porpoise; Fig. 1c) shows the typical configuration and shape of components found in other toothed whales examined in this study (Table 1) and described in the literature (Rawitz, 1900; Boenninghaus, 1903; Hain, 1914; Hosokawa, 1950; Schneider, 1964; Lawrence and Schevill, 1965; Blevins and Parkins, 1973; Green et al., 1980; Reidenberg and Laitman, 1987, 1988, 1994; Figs. 1, 2). The same is true for the structural and topographical relations within the nasopharyngeal (choanae, esophagus, hyoid apparatus, and their musculature) and gular regions (Rawitz, 1900; Boenninghaus, 1903; Lawrence and Schevill, 1965; Reidenberg and Laitman, 1987, 1994; Figs. 1–3). Accordingly, we focus on the aspects relevant for a general understanding of these regions and for our functional interpretations. If not cited otherwise, the nomenclature for these hyolaryngeal structures follows Lawrence and Schevill (1965) and Reidenberg and Laitman (1987, 1988, 1994), respectively.

Figure 1.

(a) and (b) Schematic lateral view of the hyoid apparatus and the larynx as well as their associated musculature in the harbor porpoise (Phocoena phocoena). Note that the hyoid region is artificially expanded dorsoventrally to show the whole laryngeal region. The muscles are represented by strings, because in their full size, they would cover each other in lateral view (cf. Fig. 3 and text). The free end of the stylopharyngeus muscle (sp) inserts onto the pharynx, and the hyoid muscle is not shown. The blue arrow marks the dorsal entrance into the bony nasal passage (broken lines). (c) Midsagittal microscopic section of the larynx of the finless porpoise (Neophocaena phocaenoides) stained with AZAN. 1st, first tracheal cartilage; A, anterior; AC, arytenoid cartilage; AP, air passage; BB, basihyal bone; CB, ceratohyal bone; ch, ceratohyoid muscle; CO, corniculate cartilage; CR, cricoid cartilage; cr, cricothyroid muscle; cy, cricoarytenoid muscle; D, dorsal; dm, digastric muscle; EC, epiglottic cartilage; gh, geniohyoid muscle; he, hyoepiglottic muscle; hg, hyoglossus muscle; ia, interarytenoid muscle; LF, laryngeal furrows; MA, mandible; mh, mylohyoid muscle; OC, occipital condyle; ot, occipitothyroid muscle; P, posterior; PH, pterygoid hamulus; pp, palatopharyngeus muscle; PX, pharynx; RO, rostrum; SB, stylohyal bone; sd, sternothyroid muscle; sg, styloglossus muscle; sh, stylohyoid muscle; sm, sternohyoid muscle; ST, sternum; ta, thyroarytenoid muscle; TB, thyrohyal bone; TC, thyroid cartilage; th, thyrohyoid muscle; tl, thyropalatine muscle; TO, tongue; tp, thyropharyngeus muscle; TR, trachea; TY, tympanohyal cartilage; ZA, zygomatic arch; V, ventral.

Figure 2.

T1-weighted MRI scans of harbor porpoise (Phocoena phocoena) head showing the topographical relations of the larynx and its surrounding structures. (a) Midsagittal scan. Dotted line marks the approximate course of the alimentary tract, which runs bilaterally around the laryngeal spout (corniculate and epiglottic cartilages (CO, EC)); black arrow indicates the slit-like entrance into the laryngeal spout. White arrow indicates the level of the coronal section in (b) which is slightly oblique with respect to the transverse plane as defined in Materials and Methods. Scale applies to (a) and (b). BB, basihyal bone; BC, brain cavity; BH, blowhole; BL, blubber; ch, ceratohyoid muscle; CV, cervical vertebrae; D, dorsal; ES, esophagus; gg, genioglossus muscle; gh, geniohyoid muscle; he, hyoepiglottic muscle; hg, hyoglossus muscle; MA, mandible; mh, mylohyoid muscle; ME, melon; NA, nasal passage; P, posterior; PE, tympanoperiotic complex; pp, palatopharyngeus muscle; PR, piriform recess; RO, rostrum; sm, sternohyoid muscle; SB, stylohyal bone; TB, thyrohyal bone; TC, thyroid cartilage; TO, tongue; TR, trachea.

Figure 3.

Ventral semischematic drawing of the gular musculature in the harbor porpoise (Phocoena phocoena). The sphincter colli, digastric, hyoglossus, and the sternohyoid muscles as well as the basihyal and thyrohyal bones are removed. ch, ceratohyoid muscle; cr, cricothyroid muscle; gg, genioglossus muscle; gh, geniohyoid muscle; he, hyoepiglottic muscle; hm, hyoid muscle; L, left; LC, lacrimal; MA, mandible; mh, mylohyoid muscle; mo, mastohumeralis muscle; OC, occipital condyle; oo, orbicularis oris muscle; ot, occipitothyroid muscle; P, posterior; pg, palatoglossus muscle; PT, dorsal periosteum of thyrohyal bone; PX, pharynx; SB, stylohyal bone; sc, scalenus muscle; sd, sternothyroid muscle; sg, styloglossus muscle; sh, stylohyoid muscle; sp, stylopharyngeus muscle; sr, sternomastoid muscle; TC, thyroid cartilage; th, thyrohyoid muscle; tl, thyropalatine muscle; tp, thyropharyngeus muscle; TR, trachea; ZP, zygomatic process.

The hyoid apparatus of the harbor porpoise is located ventral to the skull base and C-shaped in lateral aspect. It consists of three paired and one unpaired bony elements. Its dorsalmost bone, the paired stylohyal (Fig. 1), is connected to the paroccipital process of the skull (Reidenberg and Laitman, 1994) by means of the tympanohyal cartilage (Fig. 1). The ventralmost elements of the hyoid are the unpaired basihyal bone rostrally (Figs. 1, 2) and the paired thyrohyal bone caudally (Fig. 1). This configuration is reminiscent of a whale fluke in dorsal or ventral view. On both sides, the anterior horn of the basihyal and the anterior tip of the stylohyal are linked by the short ceratohyal bone, which stands dorsoventrally (Fig. 1). All these hyoid bones are interconnected by small cartilaginous interfaces. One pair of these interfaces is the epihyals located rostrally of the stylohyals. Repeated careful movements of the hyoid apparatus by hand during the dissection of fresh specimens could relocate this complex a few centimeters ventrally and dorsally as well as rostrally and caudally. These movements were possible by the temporary deformation of the cartilaginous interfaces including the tympanohyal cartilage (synchondroses).

According to their topography and orientation, the components of the hyoid musculature may be divided into at least three different parts: a rostral, caudal, and a dorsoventral group of muscles with different potential functional implications.

The rostral group (hyoglossus, digastric, mylohyoid, geniohyoid muscles; Figs. 1, 2) pulls the hyoid apparatus in a rostral direction on both sides, so that it rotates around a transverse axis through its attachments at the skull base via the tympanohyal cartilages. We did not consider the styloglossus muscle in this group, because it has no bony insertion but participates in the movement of the tongue (Figs. 1a, 3).

The caudal group, consisting of the two sternohyoid muscles, may pull the ventral part of the hyoid apparatus caudally (Figs. 1, 2). This movement is antagonistic to the actions of the rostral group, and the resulting rostrocaudal movements of the hyoid apparatus are reminiscent of a swing (Fig. 1).

The larynx is suspended rostrally from the framework of the hyoid apparatus (drawn ventrocaudally in Fig. 1a to show hidden structures) by the short hyoepiglottic and thyrohyoid muscles. If the hyoid is pulled by the rostral muscle group (Fig. 1a), the larynx should be forced dorsorostrally in the direction of the choanae due to the rotation of the hyoid apparatus around the transverse axis through the attachment sites of tympanohyals at the skull base. The hyoepiglottic and thyrohyoid muscles may support the movement of the larynx rostrally (Figs. 1, 2). In contrast, when the hyoid apparatus is pulled caudally by the sternohyoid muscle, the larynx would be forced passively in the same direction. In addition, this caudal shift of the larynx may be supported by the sternothyroid muscle (Fig. 1a), which, therefore, can be included in the caudal muscle group mentioned earlier.

Actions of the dorsoventral group (stylohyoid, ceratohyoid, thyropalatine muscles; Figs. 1a, 3), should result in a bending of the C-shaped hyoid apparatus dorsally (rotating in the basihyal-ceratohyal and stylohyal-ceratohyal joints), so that the caudal end of the thyrohyal bone and the rostral end of the hyoid apparatus approach the skull base (Figs. 1a, 3). Moreover, according to its suspension within the hyoid apparatus, the larynx should also be forced in a dorsal direction during the bending of the hyoid through the action of laryngeal muscles (thyropharyngeus, occipitothyroid). The unfolding of the hyoid apparatus and the relocation of the larynx ventrally is facilitated by the simultaneous action of the rostral muscle group and the caudal group (sternohyoid and sternothyroid muscles; Figs. 1, 3).

The last muscle connecting the larynx with the skull base is a powerful choanal sphincter, the palatopharyngeus muscle (Figs. 1, 2), which arises within the bony nares and imposes itself on the epiglottic spout. This muscle, reminiscent of a strong hose and representing the “soft palate,” houses the distal part of the larynx (Fig. 1). In its dorsal part below the bony naris, the palatopharyngeus muscle largely displaces the lumen of the respective bony nasal passage where it originates. The pterygoid hamulus as a posterior elongation of the bony nasal passage may serve as an additional area of origin for this muscle.

DISCUSSION

The overall similarities of the larynx, hyoid apparatus, and the accessory musculature throughout the species investigated here and in the literature (Rawitz, 1900; Boenninghaus, 1903; Hein, 1914; Hosokawa, 1950; Schneider, 1964; Lawrence and Schevill, 1965; Green et al., 1980; Reidenberg and Laitman, 1987, 1988, 1994) imply that our functional interpretations may hold for the suborder Odontoceti, in general. Because the shape of the odontocete larynx is unique among the Mammalia, various hypotheses were put forward regarding its function. The major task of the larynx is probably to keep the respiratory and digestive tracts separate from one another during swallowing (Reidenberg and Laitman, 1987). Moreover, during the 1980s, the idea that the larynx of toothed whales could be the source of the pulsed sounds used in echolocation was established. This “larynx hypothesis” implied a similar mechanism of sound production as in other mammals (Blevins and Parkins, 1973; Purves and Pilleri, 1983; Reidenberg and Laitman, 1988, 1994). According to the modern sound generation hypotheses, however, the larynx and its associated musculature build up the pressure needed for air flow in the nasal complex used in the production of whistles and echolocation sounds in toothed whales (Norris, 1969; Norris and Harvey, 1972, 1974; Ridgway et al., 1980; Amundin and Andersen, 1983; Marten et al., 1988; Cranford et al., 1996, 2001; Cranford, 2000; Cranford and Amundin, 2003). The mechanism of click sound generation in the nose, as described by Cranford et al. (1996), does not require a large volume of air but high pressure. From the morphological point of view, the larynx and the palatopharyngeal sphincter muscles are very good candidates for the generation of the pressure needed for nasal vocalization. Here, the larynx seems to act as a piston, which is pulled forward and upward by the strong sphincter muscle in the direction of the choanae (Cranford et al., 1996; Houser et al., 2004, Huggenberger, 2004) and back by antagonistic muscles (see below). From these and other anatomical studies (Negus, 1958; Moris, 1969; Purves and Pilleri, 1983; Lawrence and Schevill, 1965; Reidenberg and Laitman, 1987), it is plausible that the action of the palatopharyngeal sphincter muscle should seal the larynx water tight at the tip of the epiglottic spout when pulling the latter in the direction of the choanae. In addition, the intranarial course of the palatopharyngeus muscle could be essential for diving maneuvers when shrinking air volumes may limit the sound generation process. Here, the shortening and thickening of the contracting muscle should replace most of the intranarial volume and thus transport air in the direction of the nasal complex.

On the other hand, the larynx is framed by the hyoid apparatus and suspended from it by thyrohyoid membranes and ligaments as well as by a set of muscles (Fig. 1a; Lawrence and Schevill, 1965; Reidenberg and Laitman, 1994). Thus, the strong gular musculature connecting the hyoid apparatus with the mandibles and the skull base (Fig. 1a) may likewise force the larynx in the direction of the choanae (Cranford et al., 1996). Figure 1a demonstrates schematically the potential direction of preferential muscle traction on the hyoid apparatus and the larynx (in Fig. 1, the laryngeal apparatus is displaced ventrally for didactical reasons). Interestingly, a set of very strong muscles of the rostral group (eight muscles) forces both the hyoid and larynx dorsorostrad (protraction), but only four muscles (sternohyoid and sternothyroid) seem to act as retractors. Therefore, it is likely that the presumed laryngeal piston mechanism for sound generation is driven by the strong palatopharyngeal sphincter muscle (Cranford et al., 1996) and, in addition, by the protractors of the hyoid and the larynx.

In earlier works (Norris et al., 1971 [cited after Norris, 1980]), it was suggested that after each sound generation cycle, air in the nasal system can be recycled. In this step, the sternohyoid and sternothyroid muscles can retract the larynx from its intranarial position caudally (Fig. 1) and thus help to bring back air from the nasal air sacs into the nasal passage below the nasal plugs (not shown), together with the blowhole musculature (Huggenberger, 2004). This retraction is all the more plausible, because, on both sides, the hyoid apparatus is “articulated” with the skull via the stylohyal and tympanohyal elements. Thus, the movements of the larynx/hyoid complex resemble that of a swing oscillating between extreme dorsorostral and ventrocaudal positions. These functional implications support the idea that, within the framework of the hyoid complex, the odontocete larynx may “swing” from its infratemporal to a choanal (intranarial) position and, as a piston, shifts quanta of air beyond the choanae into the nasal passages. However, it is unclear whether this characteristic hyoid/larynx configuration can be added to the complex of sonar-related skull features as, for example, the so-called facial depression, which harbors the nasal complex, because hyoid bones are rarely seen in the fossil record (Fordyce and de Muizon, 2001; Fordyce, 2002), so that it is difficult to estimate at which geological era the typical size and shape of the hyoid “swing” occurred.

Next to sonar clicks, so-called bangs or jaw pops are other types of loud impulse sound frequently recorded in dolphins (but not in porpoises as yet). These low-frequency sounds (peak frequency around 1–8 kHz), generally lasting for several milliseconds, are thought to represent (apart from whistles, etc.) signals in social interactions (Connor and Smolker, 1996) and perhaps also serve in the debilitation of prey animals (Norris and Møhl, 1983; Marten et al., 2001). Although the bangs are likely to be generated in the nasal complex, the underlying mechanism seems to be correlated with a rapid closure of the jaws (Marten et al., 1988). Cranford et al. (1993) suggested a generation mechanism similar to that used for echolocation pulses, except that the (nasal) tissue region involved should be larger and the air pressure considerably higher. Our inspection of the gular musculature in porpoises and its overall similarity with that of dolphins supports this hypothesis, because via these muscles, the mandible is massively connected to the hyoid apparatus and the larynx, respectively. Opening of the jaws, therefore, would include the retraction of the larynx far caudally by the sternohyoid and sternothyroid muscles. This movement, in turn, should enlarge the respiratory air space between the nasal plugs (bony nares) dorsally and the tip of the epiglottic spout ventrally. The following simultaneous contraction of (1) the hoselike palatopharyngeal sphincter complex, (2) the muscles connecting the larynx/hyoid with the mandible (Fig. 1), and (3) the musculature for jaw closure (particularly, the temporal and pterygoid muscles) should violently press this enlarged air quantum into the nasal complex to produce the bang.

Beyond its high significance in sound production, the larynx/hyoid complex probably takes part in suction feeding behavior (Reidenberg and Laitman, 1994; Heyning and Mead, 1996). Here, the gular region and the tongue represent a “hydraulic piston,” which is retracted caudally by the laryngohyoid musculature (Bloodworth and Marshall, 2005, 2007). During such negative pressure events in the oral cavity and pharynx, the full contraction of the palatopharyngeal sphincter muscle around the laryngeal spout shuts down the choanolaryngeal connection as mentioned by Reidenberg and Laitman (1987, 1994). Moreover, because rapid respiration is typical for surfacing animals (Kooyman, 1973 ; Pabst et al., 1999), a mechanism is needed that prevents the animals from choking even during rapid inspiration. Strong negative pressure in the thorax and lungs, as the prerequisite for rapid inspiration, could draw water from the alimentary tract (pharynx) into the respiratory tract. However, at this time, the sphincter muscle cannot be contracted maximally around the laryngeal spout, because the air passage must remain open during inspiration. Such partial contraction of the palatopharyngeus muscle around the laryngeal spout may also be important for expiration when there is positive pressure in the respiratory tract, so that air cannot escape into the pharynx. According to our morphological examinations, it seems likely that, during respiration, the larynx stays in place in relation to the choanae (Fig. 4). However, to guarantee the watertight sealing of the airway during strong negative (or positive) pressure in the respiratory tract, the pharynx should be squeezed against the skull base and collapse due to a dorsal shift of the hyoid apparatus by its dorsoventral muscle group (flexing of the hyoid apparatus dorsally; Figs. 1a, 2), so that the alimentary tract is closed at the level of the piriform recesses (Fig. 4). This may be an additional mechanism for the waterproof separation of the digestive tract and airway during respiration (Fig. 4).

Figure 4.

Schematic transverse section (cf. Fig. 2b) of a harbor porpoise head at the laryngeal spout [corniculate and epiglottic cartilages (CO, EC)] demonstrating the closure of the pharynx at the piriform recesses (PR) due to the lift of the hyoid apparatus (right, gray arrows) by the dorsoventral muscle group. This collapse of the recesses is a potential mechanism to separate the pharynx and the esophagus, respectively, from the airway during respiration. BC, brain cavity; BL, blubber; SB, stylohyal bone; SK, skull base; TB, thyrohyal bone.

Acknowledgements

The authors thank Dr. Harald Benke (Stralsund, Germany), Dr. James G. Mead (Washington, DC, USA), Prof. Dr. Giorgio Pilleri (Courgeveaux, Switzerland), Dr. William F. Perrin (San Diego, CA, USA), Dr. Charles Potter (Washington, DC, USA), Dr. Ursula Siebert (Büsum, Germany), and Dr. Gerhard Storch (Frankfurt a.M., Germany) for the donation of toothed whale material under their care. They also thank Dr. Jan Schmitt, Prof. Dr. Thomas J. Vogl, and the staff of the Institute for Diagnostic and Interventional Radiology (Frankfurt a.M., Germany) for the help in the capture of the MRI and CT scans. The authors also thank two anonymous referees for their valuable comments on this work. Prof. Dr. Jürgen Winckler (Frankfurt a.M., Germany), who passed away in May 2004, is gratefully remembered for his generous and continous support for this project.

Ancillary